who treatment guidelines for drug- resistant … mdr tb...otsuka for the clinical trial of...
TRANSCRIPT
ISBN: 978-92-4-154963-9
2016 update
WHO treatment guidelines for drug-resistant tuberculosis
OCTOBER 2016 REVISION
These guidelines were developed in compliance with the process for evidence gathering, assessment and formulation of recommendations, as outlined in the WHO Handbook for Guideline Development (version March 2014; available at http://www.who.int/kms/handbook_2nd_ed.pdf ).
WHO Library Cataloguing-in-Publication Data
WHO treatment guidelines for drug-resistant tuberculosis, 2016 update. October 2016 revision.
Contents: Web Annex 4: GRADE tables – Web Annex 5: Evidence to decision tables – Web Annex 6: Summaries of unpublished data used for the recommendations
1.Tuberculosis, Multidrug-Resistant – drug therapy. 2.Antitubercular Agents. 3.Guideline. I.World Health Organization.
ISBN 978 92 4 154963 9 (NLM classification: WF 310)
© World Health Organization 2016
All rights reserved. Publications of the World Health Organization are available on the WHO website (http://www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; email: [email protected]).
Requests for permission to reproduce or translate WHO publications –whether for sale or for non-commercial distribution– should be addressed to WHO Press through the WHO website (http://www.who.int/about/licensing/copyright_form/index.html).
The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use.
Printed by the WHO Document Production Services, Geneva, Switzerland
Design by Inis Communication – www.iniscommunication.com
WHO/HTM/TB/2016.04
Contents
Abbreviations & acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Declarations of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Executive summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
WHO policy recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Annex 1. Agenda for the Guideline Development Group meeting on the WHO treatment guidelines for drug-resistant TB, 2016 update, 9–11 November 2015 . . . . . . . 46
Annex 2. Experts involved in the development of the WHO treatment guidelines for drug-resistant TB, 2016 update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Annex 3. PICO questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Annex 4. GRADE tables
Annex 5. Evidence to decision tables
Annex 6. Summaries of unpublished data used for the recommendations
Annexes 4, 5 and 6 are only available online at: www.who.int/entity/tb/areas-of-work/drug-resistant-tb/treatment/resources/en/
iv
Abbreviations & acronyms1
aDSM active TB drug safety monitoring and managementaOR adjusted odds ratioAIDS acquired immunodeficiency syndromeaIPD adult individual patient dataCDC United States Centers for Disease Control and PreventionCL confidence limitsCNS central nervous systemCPTR Critical Path to TB Drug RegimensDSMB Data and Safety Monitoring BoardDST drug susceptibility testingEBA early bactericidal activityERG External Review GroupGDF Global Drug FacilityGDG Guideline Development GroupGRADE Grading of Recommendations Assessment, Development and EvaluationGRC WHO Guideline Review CommitteeGTB WHO Global TB ProgrammeHIV human immunodeficiency virusIPD individual patient dataKNCV KNCV Tuberculosis FoundationLSHTM London School of Hygiene and Tropical MedicineLTBI latent TB infectionMDR-TB multidrug-resistant tuberculosisMIC minimum inhibitory concentrationMSF Médecins sans FrontièresMTBDRsl GenoType Mycobacterium tuberculosis drug-resistant second-line assayNTM non-tuberculous mycobacteriaOR odds ratioPICO Patients, Intervention, Comparator and OutcomespIPD paediatric individual patient dataRCT randomized controlled trialRR-TB rifampicin-resistant TBSAE serious adverse eventSIAPS Systems for Improved Access to Pharmaceuticals and ServicesTAG Treatment Action GroupTB tuberculosisTB-PRACTECAL Pragmatic Clinical Trial for a More Effective Concise and Less Toxic MDR-TB Treatment
Regimen(s)UNION International Union Against Tuberculosis and Lung DiseaseUSAID United States Agency for International DevelopmentWHO World Health OrganizationXDR-TB extensively drug-resistant tuberculosis
1 See also page 23 for the abbreviations of the names of TB medicines.
1
Acknowledgements
The recommendations and remarks contained in this document resulted from the discussions of an ad hoc Guideline Development Group (GDG) convened by the Global TB Programme (GTB) of the World Health Organization (WHO) in Geneva, Switzerland from 9 to 11 November 2015 (Annex 1). WHO gratefully acknowledges the contributions made by this Group ahead of, during and after this meeting by its members, namely Holger Schünemann (Chair and GRADE methodologist) and Charles L Daley (Co-Chair) and other experts: Farhana Amanullah, José A Caminero, Tsira Chakhaia, Daniela Cirillo, Kelly Dooley, Luis Gustavo do Valle Bastos, Michel Gasana, Agnes Gebhard, Armen Hayrapetyan, Antonia Kwiecien, Sundari Mase, Lindsay McKenna, Nguyen Viet Nhung, Maria Rodriguez, James Seddon, Tom Shinnick, Alena Skrahina, and Carlos Torres-Duque (Annex 2). The experts on the External Review Group (ERG) who provided comments in preparation of the meeting and on the draft document before its finalization – Chen-Yuan Chiang, Dalene von Delft, Vaira Leimane, Guy Marks, Norbert Ndjeka, Lee Reichman and Rohit Sarin – are also acknowledged.
The writing of these guidelines was coordinated by Dennis Falzon and Elizabeth Harausz (Consultant), under the guidance and supervision of Ernesto Jaramillo and Karin Weyer, and the overall direction of Mario Raviglione, Director of the GTB. The authors acknowledge the contribution of other WHO staff in the production of these guidelines making up the WHO Guideline Steering Committee, namely Nathan Ford, Giuliano Gargioni, Haileyesus Getahun, Malgorzata Grzemska, Avinash Kanchar, Soleil Labelle, Christian Lienhardt, Knut Lönnroth, Fuad Mirzayev, Linh Nhat Nguyen, Marco Antonio Vitoria, Fraser Wares, Diana Weil and Matteo Zignol. The following WHO staff from the regional offices received a final draft of the guideline document for review: Masoud Dara (Europe), Mirtha del Granado (Americas), Daniel Kibuga (Africa), Hyder Khursid (South-East Asia), Mohammed Abdel Aziz (Eastern Mediterranean) and Nobuyuki Nishikiori (Western Pacific). The document was finalized following an iteration of comments in early 2016 from members of the GDG, the ERG, and the the WHO Guideline Steering Committee, ahead of submission to the WHO Guideline Review Committee (GRC) in March 2016 following the WHO internal clearance process.
The funding for the update of the guidelines was made available by the United States Agency for International Development (USAID), through the USAID–WHO Consolidated Grant No. GHA-G-00–09–00003/US-2014–735.
2
Declarations of interest
Guideline Development Group (GDG)
The scope of the guidelines update, and the composition of the GDG, including their biographies, were made public for comment ahead of the meeting in line with WHO’s conflict of interest policy. All GDG members completed the WHO Declaration of Interest forms. The WHO Guideline Steering Committee in preparation for the update of the guidelines and the GDG meeting reviewed the completed forms.
The following GDG members declared no interests: Luis Gustavo do Valle Bastos, José A Caminero, Tsira Chakhaia, Michel Gasana, Armen Hayrapetyan, Antonia Kwiecien, Sundari Mase, Nguyen Viet Nhung, Maria Rodriguez, Holger Schünemann, James Seddon and Alena Skrahina.
The following GDG members declared interests that were judged not to be in conflict with the objectives of the meeting:
Farhana Amanullah declared having received funding for consultancies (US$500/day) and travel from WHO; and grants from the Global Fund and TB-REACH to cover her salary (10% full time equivalent).
Daniela Cirillo declared having received funding from FIND to conduct evaluation of drug susceptibility testing (DST) for new drugs (US$16 000), and from Otsuka to evaluate DST for delamanid (US$25 000). She also declared being the head of a supranational TB reference laboratory in Italy involved in country capacity building in DST technologies for second-line drugs and new diagnostics for drug-resistant TB; and being a member of the Italian national committee for the use of bedaquiline.
Charles L Daley declared having received funding from Otsuka to serve as chair of the data monitoring committee for trials of delamanid (US$47 000 over 7 years–current).
Kelly Dooley declared having received funding to provide expert advice on a trial design for TB/HIV (US$2000/year paid to the university/employer); she also declared the following activities and roles: co-chair AIDS Clinical Trials Group (ACTG) study assessing bedaquiline and delamanid; principal investigator for adjuvant paclitaxel and trastuzumab (APT) trial assessing pretomanid (PA-824); and investigator in trials assessing high-dose isoniazid for MDR-TB, rifapentine for pregnant women and children with latent TB infection (LTBI), high-dose rifampicin and levofloxacin for paediatric TB meningitis, as well as bedaquiline and delamanid for children with MDR-TB and HIV infection.
3
Declarations of interest
Agnes Gebhard declared that she works for the KNCV TB Foundation, which has two projects funded by the Eli Lilly and Company Foundation: (i) engaging the private sector in diagnosis and treatment of TB and MDR-TB with quality assured second-line TB drugs, and (ii) the roll-out of QuanTB (a drug forecasting tool) in countries not supported by the Systems for Improved Access to Pharmaceuticals and Services (SIAPS) implemented by Management Sciences for Health. In addition, she declared that the KNCV TB Foundation has a collaborative project with Cepheid in two countries (Nigeria, Viet Nam) with KCNV providing services for the installation and initial training on the use of GeneXpert machines.
Carlos Torres-Duque declared having received honoraria from Janssen Pharmaceuticals for presentations on TB prevention and WHO policy on bedaquiline at a Latin American Meeting on MDR-TB held in 2014 (US$2000).
Tom Shinnick declared being an employee of the United States Centers for Disease Control and Prevention (CDC). CDC supports his travel and research related to his work on laboratory services needed for TB control. He declared having often represented CDC’s position on laboratory services needed for TB diagnosis, treatment and control. As part of his official duties for CDC, he served on the Data and Safety Monitoring Board (DSMB) organized by Otsuka for the clinical trial of delamanid. He did not receive any remuneration for serving on the DSMB nor for travel expenses (CDC paid for all travel expenses related to serving on the DSMB). The DSMB has completed its work for the trial.
The following GDG members declared interests that were judged to be in conflict with some of the objectives of the meeting and were thus recused from some of the discussions:
Lindsay McKenna declared non-commercial support to Treatment Action Group (TAG), her employer, from Stop TB Partnership; Bill & Melinda Gates Foundation; the US Department of Veteran Affairs (on behalf of CDC); Janssen Therapeutics for Hepatitis C and HIV projects and the Global Alliance for TB Drug Development (a public-private entity developing new drugs and regimens for TB treatment). She was thus recused from participating in the 9 November 2015 meeting session on Patients, Intervention, Comparator and Outcomes (PICO) question 1 on MDR-TB regimen composition for adults and children.
José A Caminero stated in his biosketch that he is a staff consultant of the International Union Against Tuberculosis and Lung Disease (UNION), an agency directly involved in the implementation and evaluation of programmes using shorter MDR-TB regimens. He was therefore recused from the 10 November 2015 meeting session on PICO question 3 on shorter regimens for MDR-TB.
4
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
External Review Group (ERG)
The following ERG members declared no interest related to the objects of this meeting: Chen-Yuan Chiang, Celine Garfin, Michael Kimerling, Vaira Leimane, Gao Mengqiu, Norbert Ndjeka, Ejaz Qadeer, Lee Reichman, Rohit Sarin and Irina Vasilyeva.
The following two ERG members declared interests which were judged not to be in conflict with the objects of the guidelines development.
Guy Marks declared research support from AERAS (US$450 000) related to the evaluation of latent TB infection and the rate of recurrence of TB after initial treatment in Viet Nam. He also declared being the Vice-President (and a board member) of the UNION and Editor-in-Chief of the International Journal of Tuberculosis and Lung Disease (for which he receives an honorarium).
Dalene von Delft declared having received support from TAG, USAID, UNITAID, Janssen Pharmaceuticals, Critical Path to TB Drug Regimens (CPTR) and AERAS to cover travel costs and accommodation to give presentations/speeches on drug-resistant TB. She declared that in 2011 she received bedaquiline as part of her MDR-TB treatment through a compassionate use access programme.
Evidence reviewers
The researchers who undertook the systematic reviews of evidence for this revision were the following (Annex 2):
McGill University, Montréal, Canada [Mayara Bastos, Gregory J Fox, Faiz Ahmad Khan, Richard (Dick) Menzies]
London School of Hygiene and Tropical Medicine (LSHTM), London, UK [Katherine Fielding, Rebecca Harris, Mishal Khan, David Moore]
Stellenbosch University, Cape Town, South Africa [Anneke Hesseling]
The evidence reviewers did not participate in the formulation of the policy recommendations.
The following reviewers declared no interest related to the objectives of this meeting: Mayara Bastos, Faiz Ahmad Khan, Richard (Dick) Menzies and Mishal Khan.
The following reviewers declared interests that were judged not to be in conflict with the objectives of the meeting:
Gregory J Fox declared having received research and non-monetary support from the UNION (sponsored by Otsuka) valued at about US$5000 to attend the 2015 International Union Conference and to receive the Young Innovator Award (he declares no work for Otsuka nor any relationship of this award with any commercial or research activities with Otsuka).
5
Declarations of interest
Katherine Fielding declared that her employer (LSHTM) was a recipient of an award from Médecins Sans Frontières (MSF) (GB£26 890) for the period February–December 2015 to provide statistical support for the TB-PRACTECAL study on which she is a co-investigator. The study is a Phase II-III randomized controlled trial (RCT) to evaluate the efficacy and safety of shorter MDR-TB regimens for adults.
Rebecca Harris declared she is consulting for a clinical research organization (Cromsource) working for GlaxoSmithKline (GSK) vaccines (~GB£90 000 in 2013); and on GSK vaccines not related to TB (~GB£10 000 since 2013) for Manpower Solutions.
David Moore declared receiving research support from the Wellcome Trust Research Training Fellowship Programme to supervise a PhD student to study MDR-TB in Peru (GB£ 207 056 in 2014).
Anneke Hesseling declared that her employer (Stellenbosch University) is a recipient of an award from Otsuka Pharmaceutical (~US$70 000 to date) for her work on the Phase III delamanid clinical trials in children.
6
Executive summary
In November 2015, the World Health Organization (WHO) convened a meeting of a Guideline Development Group (GDG) for the update of policy recommendations on the treatment of drug-resistant TB. The GDG was composed of a multidisciplinary group of tuberculosis (TB) and drug-resistant TB experts external to WHO. Before the meeting, the members of the GDG and the WHO Guideline Steering Committee had decided upon the priority questions in the treatment and care of patients with drug-resistant TB to be considered for the update of the guidelines. The scope of the 2016 update comprised the following:
i. The optimal combination of medicines and approach towards regimen design for TB patients (both adults and children) with isoniazid-resistant, rifampicin-resistant (RR-TB), multidrug-resistant (MDR-TB), and extensively drug-resistant (XDR-TB) forms of TB, as well as for patients with M. bovis disease.
ii. The effectiveness and safety of standardized regimens lasting up to 12 months for the treatment of patients with MDR-TB (“shorter regimens”) when compared with longer treatment.
iii. The effect of delay in starting treatment on treatment outcomes for patients with drug-resistant TB.
iv. The effect of surgical interventions on treatment outcomes for patients with drug-resistant TB.
The scope of the WHO treatment guidelines for drug-resistant tuberculosis, 2016 update thus differed from the one that guided the previous update of the WHO policy recommendations on the programmatic management of drug-resistant TB in 2011 (1). It did not cover aspects of policy guidance on the programmatic management of drug-resistant TB that were of lesser priority or for which no new evidence has emerged since the 2011 revision. These included questions relating to the use of rapid diagnostics for RR-TB, the monitoring of response to treatment, the duration of longer (“conventional”) MDR-TB regimens, the delay in starting antiretroviral therapy in MDR-TB patients with human immunodeficiency virus (HIV) and models of care. The GDG considered that the 2011 recommendations relating to these areas would continue to apply until future evidence reviews show a need for revision of current WHO policy.
In preparation for the GDG meeting, systematic reviews were conducted to answer questions formulated in PICO format (Patients, Intervention, Comparator and Outcomes) that addressed all domains of the guidelines scope. The treatment regimen recommendations for adults in the 2016 update were based in part on individual patient data meta-analysis (of 9153 patients who were mostly adults) that informed the 2011 guidance, supplemented with additional evidence published until August 2015, which was summarized in a study-level meta-analysis. Treatment regimen recommendations for children were based on a paediatric individual patient data
7
executive summary
(pIPD) meta-analysis, which included data on 974 children in cohorts and studies published until September 2014. The data for shorter MDR-TB treatment regimens (up to 12 months) were from an analysis of individual patient data and aggregated data from observational studies conducted in Asia and Africa. Surgical recommendations for MDR-TB patients were based on individual patient data analysis and a study-level meta-analysis.
The evidence available on the treatment of isoniazid-resistant TB and on the delay in starting MDR-TB treatment could not address the respective PICO questions. There were very few published studies on the treatment of M. bovis and the regimens differed too much, precluding any attempt at formulating recommendations of clinical use.
The recommendations that address the other PICO questions are summarized below.
1. Shorter MDR-TB regimen for adults and childrenIn patients with RR-TB or MDR-TB who were not previously treated with second-line drugs and in whom resistance to fluoroquinolones and second-line injectable agents was excluded or is considered highly unlikely, a shorter MDR-TB regimen of 9–12 months may be used instead of the longer regimens (conditional recommendation, very low certainty in the evidence).
2. Longer MDR-TB regimens for adults and children2a) In patients with RR-TB or MDR-TB, a regimen with at least five effective TB medicines during the intensive phase is recommended, including pyrazinamide and four core second-line TB medicines – one chosen from Group A, one from Group B, and at least two from Group C2 (conditional recommendation, very low certainty in the evidence). If the minimum number of effective TB medicines cannot be composed as given above, an agent from Group D2 and other agents from Group D3 may be added to bring the total to five.3
2b) In patients with RR-TB or MDR-TB, it is recommended that the regimen be further strengthened with high-dose isoniazid and/or ethambutol (conditional recommendation, very low certainty in the evidence).
It is recommended that any patient – child or adult – with RR-TB in whom isoniazid resistance is absent or unknown be treated with a recommended MDR-TB regimen. It could either be a shorter MDR-TB regimen, or a longer MDR-TB regimen to which isoniazid is added.
As a result of the update, the grouping of medicines used in the treatment of MDR/RR-TB has been revised from the one used in the last guidance to reflect the updated evidence on the efficacy and safety of the different agents. This reclassification of medicines has a bearing on the choice of medicines when users design longer, individualized regimens for patients with
2 Group A=levofloxacin, moxifloxacin, gatifloxacin; Group B=amikacin, capreomycin, kanamycin, (streptomycin); Group C= ethionamide (or prothionamide), cycloserine (or terizidone), linezolid, clofazimine; in children with non-severe disease Group B medicines may be excluded (see guidelines text for how disease severity was assessed).
3 Group D2=bedaquiline, delamanid; Group D3=p-aminosalicylic acid, imipenem–cilastatin, meropenem, amoxicillin clavulanate, (thioacetazone). Bedaquiline is only recommended for adults; delamanid may also be used in patients aged 6–17 years.
8
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
drug-resistant TB. There is no change in the recommended use of bedaquiline and delamanid from those defined by the WHO interim guidance. These two new medicines now occupy a unique subgroup within the add-on agents used to treat MDR/RR-TB.
3. Surgical interventions in patients with MDR-TBIn patients with RR-TB or MDR-TB patients, elective partial lung resection (lobectomy or wedge resection) may be used alongside a recommended MDR-TB regimen (conditional recommendation, very low certainty in the evidence).
MAIN CHANGES TO THE WHO POLICY RECOMMENDATIONS FOR THE TREATMENT OF DRUG-RESISTANT TB
These guidelines update the previous evidence-informed recommendations on the treatment of drug-resistant TB issued by the World Health Organization (WHO) in 2011. The current priorities in the management of drug-resistant TB have been reflected in the scope of the current guidance. For the 2016 update, the Guideline Development Group convened to update the guidelines proposed priority questions focused on the composition of treatment regimens for rifampicin-resistant (RR-TB) and multidrug-resistant TB (MDR-TB), the effectiveness and safety of shorter MDR-TB regimens, the treatment of isoniazid-resistant and M. bovis TB, the role of surgery, and the impact of delays in starting treatment for RR-TB. In contrast to the 2011 recommendations the current guidance did not update the policy on the use of rapid diagnostics for RR-TB, the monitoring of response to treatment, the duration of longer MDR-TB regimens, the delay in starting antiretroviral therapy in MDR-TB patients with HIV infection and models of care. For these aspects of the programmatic management of drug-resistant TB, the 2011 recommendations continue to apply until future evidence reviews conducted for the purpose of updating WHO policy show a need for revision.
The main changes in the 2016 recommendations are as follows:
� A shorter MDR-TB treatment regimen is recommended under specific conditions.
� Medicines used in the design of longer MDR-TB treatment regimens are now regrouped differently based upon current evidence on their effectiveness and safety. Clofazimine and linezolid are now recommended as core second-line medicines in the MDR-TB regimen while p-aminosalicylic acid is an add-on agent.
� MDR-TB treatment is recommended for all patients with RR-TB, regardless of confirmation of isoniazid resistance.
� Specific recommendations are made on the treatment of children with RR-TB or MDR-TB.
� Clarithromycin and other macrolides are no longer included among the medicines to be used for the treatment of MDR/RR-TB.
� Evidence-informed recommendations on the role of surgery are now included.
Bedaquiline and delamanid have now been assigned to a specific subgroup of add-on agents. In October 2016, WHO published its new policy on delamanid. Delamanid may now also be used alongside longer MDR-TB regimens in patients aged 6–17 years. Bedaquiline is still only recommended for adults.
9
Introduction
The WHO treatment guidelines for drug-resistant tuberculosis, 2016 update aims to support health professionals worldwide to respond to the continued challenge posed by multidrug-resistant and extensively drug-resistant tuberculosis in the post-2015 period of the End TB Strategy (2). It includes important policy changes made following a review of the latest available evidence on the medical and surgical treatment of both adults and children. This revision updates several of the evidence-informed recommendations released by WHO in 2011 (1). Until such time as future evidence reviews conducted for the purpose of updating WHO policy guidance show a need for further revision, the previous recommendations which were not revised in the 2016 update continue to apply (see also Table 1).
Methods
Preparation for revisionThe WHO Guideline Steering Committee met regularly from November 2014 through November 2015 to draft the scope and the corresponding PICO (Patients, Intervention, Comparator and Outcomes) questions, and to follow up the development of the guidelines. An application for the revision of the guidelines was submitted to the WHO Guideline Review Committee (GRC) in August 2015 that received final approval in September 2015.
Seven webinars (using WebEx) were held between May and November 2015 (on May 20, July 17, August 7, August 28, September 16, October 6, and November 5) to discuss with the GDG members the scoping, the PICO questions, the scoring of the outcomes, and progress with the evidence reviews ahead of the meeting. For certain sessions, the groups assessing the evidence were invited to these discussions in their capacity as resource persons. In between the webinars, discussions were continued via email. Two WebEx discussions were also held in 2015 with the External Review Group (ERG) members (on 7 September and 29 October), during which they were briefed about their roles and expectations as peer-reviewers.
10
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Tabl
e 1. Sum
mar
y of
cha
nges
in t
he e
vide
nce
base
d re
com
men
dati
ons
betw
een
the
20
11
(1)
and
20
16
gui
delin
esa
2011 G
UID
ELIN
ES (
1)
2016 G
UID
ELIN
ES
Use
of
rap
id d
iagn
osti
cs f
or r
ifa
mp
icin
re
sist
an
ce
Rap
id d
rug
susc
eptib
ility
tes
ting
(DS
T) o
f is
onia
zid
and
rifam
pici
n or
of rif
ampi
cin
alon
e is
rec
omm
ende
d ov
er c
onve
ntio
nal t
estin
g or
no
test
ing
at t
he t
ime
of T
B
diag
nosi
s, s
ubje
ct t
o av
aila
ble
reso
urce
s (c
ondi
tiona
l rec
omm
enda
tion,
ver
y lo
w
qual
ity e
vide
nce)
.
UPD
ATED
[fr
om (
3)]
Rap
id D
ST
of a
t le
ast
rifam
pici
n is
rec
omm
ende
d in
adu
lts a
nd
child
ren
over
con
vent
iona
l tes
ting
or n
o te
stin
g at
the
tim
e of
TB
dia
gnos
is (st
rong
rec
omm
enda
tion,
hig
h ce
rtai
nty
in t
he
evid
ence
).
Use
of
sput
um-s
mea
r m
icro
scop
y a
nd c
ultu
re t
o m
onit
or r
espo
nse
to t
rea
tmen
t
The
use
of s
putu
m-s
mea
r m
icro
scop
y an
d cu
lture
rat
her
than
spu
tum
-sm
ear
mic
rosc
opy
alon
e is
rec
omm
ende
d fo
r th
e m
onito
ring
of p
atie
nts
with
MD
R-T
B
durin
g tr
eatm
ent
(con
ditio
nal r
ecom
men
datio
n, v
ery
low
qua
lity
evid
ence
).
REM
AIN
S V
ALID
b
Use
of
a s
hor
ter
MD
R-T
B t
rea
tme
nt
regi
me
n
NO
SPE
CIF
IC R
ECO
MM
END
ATIO
N
In p
atie
nts
with
RR
-TB
or
MD
R-T
B w
ho w
ere
not
prev
ious
ly
trea
ted
with
sec
ond-
line
drug
s an
d in
who
m r
esis
tanc
e to
flu
oroq
uino
lone
s an
d se
cond
-line
inje
ctab
le a
gent
s w
as e
xclu
ded
or is
con
side
red
high
ly u
nlik
ely, a
sho
rter
MD
R-T
B r
egim
en o
f 9–1
2
mon
ths
may
be
used
inst
ead
of t
he lo
nger
reg
imen
s (c
ondi
tiona
l re
com
men
datio
n, v
ery
low
cer
tain
ty in
the
evi
denc
e).c
Com
pos
itio
n o
f lo
nge
r M
DR
-TB
tre
atm
en
t re
gim
en
s
In t
he t
reat
men
t of
pat
ient
s w
ith M
DR
-TB
, a fl
uoro
quin
olon
e sh
ould
be
used
(st
rong
re
com
men
datio
n, v
ery
low
qua
lity
evid
ence
).
In t
he t
reat
men
t of
pat
ient
s w
ith M
DR
-TB
, a la
ter-g
ener
atio
n flu
oroq
uino
lone
ra
ther
tha
n an
ear
lier-g
ener
atio
n flu
oroq
uino
lone
sho
uld
be u
sed
(con
ditio
nal
reco
mm
enda
tion,
ver
y lo
w q
ualit
y ev
iden
ce).
In t
he t
reat
men
t of
pat
ient
s w
ith M
DR
-TB
, eth
iona
mid
e (o
r pr
othi
onam
ide)
sho
uld
be u
sed
(str
ong
reco
mm
enda
tion,
ver
y lo
w q
ualit
y ev
iden
ce).
In t
he t
reat
men
t of
pat
ient
s w
ith M
DR
-TB
, reg
imen
s sh
ould
incl
ude
at
leas
t py
razi
nam
ide,
a fl
uoro
quin
olon
e, a
par
ente
ral a
gent
, eth
iona
mid
e (o
r pr
othi
onam
ide)
, and
eith
er c
yclo
serin
e or
p-a
min
osal
icyl
ic a
cid
(PAS
) if
cycl
oser
ine
cann
ot b
e us
ed (co
nditi
onal
rec
omm
enda
tion,
ver
y lo
w q
ualit
y ev
iden
ce).
In t
he t
reat
men
t of
pat
ient
s w
ith M
DR
-TB
, fou
r se
cond
-line
ant
i-tub
ercu
losi
s dr
ugs
likel
y to
be
effe
ctiv
e (in
clud
ing
a pa
rent
eral
age
nt),
as w
ell a
s py
razi
nam
ide,
sho
uld
be in
clud
ed in
the
inte
nsiv
e ph
ase
(con
ditio
nal r
ecom
men
datio
n, v
ery
low
qua
lity
evid
ence
).
UPD
ATED
d
In p
atie
nts
with
RR
-TB
or
MD
R-T
B, a
reg
imen
with
at
leas
t fiv
e ef
fect
ive
TB m
edic
ines
dur
ing
the
inte
nsiv
e ph
ase
is
reco
mm
ende
d, in
clud
ing
pyra
zina
mid
e an
d fo
ur c
ore
seco
nd-li
ne
TB m
edic
ines
– o
ne c
hose
n fr
om G
roup
A, o
ne fro
m G
roup
B, a
nd
at le
ast
two
from
Gro
up C
e (c
ondi
tiona
l rec
omm
enda
tion,
ver
y lo
w
cert
aint
y in
the
evi
denc
e). If
the
min
imum
num
ber
of e
ffec
tive
TB
med
icin
es c
anno
t be
com
pose
d as
giv
en a
bove
, an
agen
t fr
om
Gro
up D
2 a
nd o
ther
age
nts
from
Gro
up D
3 m
ay b
e ad
ded
to b
ring
the
tota
l to
five.
f
In p
atie
nts
with
RR
-TB
or
MD
R-T
B, i
t is
rec
omm
ende
d th
at t
he
regi
men
be
furt
her
stre
ngth
ened
with
hig
h-do
se is
onia
zid
and/
or
etha
mbu
tol (
cond
ition
al r
ecom
men
datio
n, v
ery
low
cer
tain
ty in
the
ev
iden
ce).
11
methoDs
2011 G
UID
ELIN
ES (
1)
2016 G
UID
ELIN
ES
Tre
atm
en
t of
pa
tie
nts
wit
h R
R-T
B
NO
SPE
CIF
IC R
ECO
MM
END
ATIO
N
It is
rec
omm
ende
d th
at a
ny p
atie
nt –
chi
ld o
r ad
ult
– w
ith R
R-T
B in
w
hom
ison
iazi
d re
sist
ance
is a
bsen
t or
unk
now
n be
tre
ated
with
a
reco
mm
ende
d M
DR
-TB
reg
imen
, eith
er a
sho
rter
MD
R-T
B r
egim
en,
or if
thi
s ca
nnot
be
used
, a lo
nger
MD
R-T
B r
egim
en t
o w
hich
is
onia
zid
is a
dded
.
Du
rati
on o
f lo
nge
r M
DR
-TB
tre
atm
en
t re
gim
en
s
In t
he t
reat
men
t of
pat
ient
s w
ith M
DR
-TB
, an
inte
nsiv
e ph
ase
of e
ight
mon
ths
is
sugg
este
d fo
r m
ost
patie
nts;
the
dur
atio
n m
ay b
e m
odifi
ed a
ccor
ding
to
the
patie
nt’s
re
spon
se t
o th
erap
y (c
ondi
tiona
l rec
omm
enda
tion,
ver
y lo
w q
ualit
y ev
iden
ce).
In t
he t
reat
men
t of
pat
ient
s ne
wly
dia
gnos
ed w
ith M
DR
-TB
(i.e
. no
t pr
evio
usly
tr
eate
d fo
r M
DR
-TB
), a
tota
l tre
atm
ent
dura
tion
of 2
0 m
onth
s is
sug
gest
ed for
m
ost;
the
dur
atio
n m
ay b
e m
odifi
ed a
ccor
ding
to
the
patie
nt’s
res
pons
e to
the
rapy
(c
ondi
tiona
l rec
omm
enda
tion,
ver
y lo
w q
ualit
y ev
iden
ce).
REM
AIN
S V
ALID
b
Sta
rt o
f a
nti
retr
ovir
al t
he
rapy
wit
h M
DR
-TB
tre
atm
en
t
Antir
etro
vira
l the
rapy
is r
ecom
men
ded
for
all p
atie
nts
with
HIV
and
dru
g-re
sist
ant
TB
requ
iring
sec
ond-
line
anti-
tube
rcul
osis
dru
gs, i
rres
pect
ive
of C
D4 c
ell-c
ount
, as
early
as
pos
sibl
e (w
ithin
the
firs
t ei
ght
wee
ks) fo
llow
ing
initi
atio
n of
ant
i-tub
ercu
losi
s tr
eatm
ent
(str
ong
reco
mm
enda
tion,
ver
y lo
w q
ualit
y ev
iden
ce).
REM
AIN
S V
ALID
b
Use
of
surg
ery
as
pa
rt o
f M
DR
-TB
tre
atm
en
t
NO
SPE
CIF
IC R
ECO
MM
END
ATIO
N
In p
atie
nts
with
RR
-TB
or
MD
R-T
B, e
lect
ive
part
ial l
ung
rese
ctio
n (lo
bect
omy
or w
edge
res
ectio
n) m
ay b
e us
ed a
long
side
a
reco
mm
ende
d M
DR
-TB
reg
imen
(co
nditi
onal
rec
omm
enda
tion,
ver
y lo
w c
erta
inty
in t
he e
vide
nce)
.
Mod
els
of
MD
R-T
B c
are
(a
mb
ula
tory
/hos
pit
ali
zati
on)
Patie
nts
with
MD
R-T
B s
houl
d be
tre
ated
usi
ng m
ainl
y am
bula
tory
car
e ra
ther
tha
n m
odel
s of
car
e ba
sed
prin
cipa
lly o
n ho
spita
lizat
ion
(con
ditio
nal r
ecom
men
datio
n,
very
low
qua
lity
evid
ence
).
REM
AIN
S V
ALID
b
a Th
ese
reco
mm
enda
tions
nee
d to
be
read
alo
ng w
ith the
acc
ompa
nyin
g re
mar
ks in
the
rel
evan
t se
ctio
ns o
f thi
s do
cum
ent,
whi
ch a
re c
ritic
al to
thei
r pr
oper
impl
emen
tatio
n.b
This
rec
omm
enda
tion
cont
inue
s to
app
ly. It
will
be
revi
sed
if a
futu
re e
vide
nce
revi
ew c
ondu
cted
for
the
pur
pose
of
upda
ting
WH
O p
olic
y gu
idan
ce s
how
s su
ch a
nee
d.c
See
tex
t fo
r th
e de
finiti
on o
f th
e sh
orte
r M
DR
-TB
tre
atm
ent
regi
men
, and
for
oth
er c
ondi
tions
tha
t ap
ply
whe
n im
plem
entin
g th
is r
ecom
men
datio
n.d
No
chan
ges
to t
he W
HO
inte
rim p
olic
ies
on t
he u
se o
f be
daqu
iline
and
del
aman
id in
the
May
2016 u
pdat
e (4
,5).
In O
ctob
er 2
01
6, W
HO
pub
lishe
d its
rev
ised
pol
icy
on
dela
man
id. D
elam
anid
may
now
als
o be
use
d al
ongs
ide
long
er M
DR
-TB
reg
imen
s in
pat
ient
s ag
ed 6
–17 y
ears
(6
). B
edaq
uilin
e is
onl
y re
com
men
ded
for
adul
ts.
e G
roup
A=
levo
floxa
cin,
mox
iflox
acin
, gat
iflox
acin
; G
roup
B=am
ikac
in, c
apre
omyc
in, k
anam
ycin
, (st
rept
omyc
in);
Gro
up C
=et
hion
amid
e (o
r pr
othi
onam
ide)
, cyc
lose
rine
(or
teriz
idon
e), l
inez
olid
, clo
fazi
min
e; in
chi
ldre
n w
ith n
on-s
ever
e di
seas
e G
roup
B m
edic
ines
may
be
excl
uded
(se
e gu
idel
ines
tex
t fo
r ho
w d
isea
se s
ever
ity w
as a
sses
sed)
.f
Gro
up D
2=
beda
quili
ne, d
elam
anid
; G
roup
D3=
p-am
inos
alic
ylic
aci
d, im
ipen
em–c
ilast
atin
, mer
open
em, a
mox
icill
in–c
lavu
lana
te, (
thio
acet
azon
e).
12
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
ScopeThe 2016 update of the WHO treatment guidelines for drug-resistant tuberculosis, 2016 update aimed to revise the previous evidence-informed policy recommendations from 2011 (1). The scope of the current guidelines differed from that of the 2011 guidance in a number of ways. In 2011, the scope of the guidelines was broader and included programmatic aspects, such as rapid diagnostics for RR-TB, patient monitoring with culture and sputum microscopy during treatment, length of the intensive phase and total duration of treatment in longer (“conventional”) regimens, use of antiretroviral therapy and ambulatory/inpatient models of care. In deciding the scope of the 2016 update, the GDG and the WHO Guideline Steering Committee considered priority questions at the time of the update (2014–2015). The scope did not cover other aspects of policy guidance on the programmatic management of drug-resistant TB for which no new evidence has been published since the 2011 revision.
The GDG agreed to limit the scope of these guidelines to the following priority areas within the current debates on the treatment and care of patients with drug-resistant TB:
i. The optimal combination of medicines and approach towards regimen design for TB patients with isoniazid-resistant, rifampicin-resistant (RR-TB), multidrug-resistant (MDR-TB), and extensively drug-resistant (XDR-TB) forms of TB as well as for patients with M. bovis disease.
ii. The effectiveness and safety of standardized regimens lasting up to 12 months for the treatment of patients with MDR-TB (“shorter regimens”) when compared with longer treatment.
iii. The effect of delay in starting treatment on treatment outcomes for patients with drug-resistant TB.
iv. The effect of surgical interventions on treatment outcomes for patients with drug-resistant TB.
As far as possible and where evidence exists, the guidelines also aimed to formulate recommendations which would be relevant to patients of all ages as well as individuals with key comorbidities (e.g. HIV, diabetes).
The target audience of the guidelines includes staff and medical practitioners working in prevention and care of TB, managers implementing the programmatic management of drug-resistant TB within their centres and national programmes, and organizations providing technical and financial support for drug-resistant TB. Although primarily intended for use in resource-limited countries, the recommendations are also applicable in other settings.
Key questionsThe PICO questions were grouped into four sets (see full versions in Annex 3). PICO questions 1 and 2 were devoted to the first area of the guidelines scope (see i above). PICO question 3 was devoted entirely to the second area (see ii above) and PICO question 4 covered both the third and fourth areas (see iii and iv above).
The outcomes were defined and scored by the GDG (Table 2). The mean scores for the nine responses received were all in the “Critical” range (7–9 points).
13
methoDs
Table 2. Scoring of outcomes considered relevant by the GDG for evidence reviews related to the WHO treatment guidelines for drug-resistant TB 2016 updatea
OUTCOMES MEAN SCORE
Adherence to TB treatment (treatment interruption due to non-adherence) 6.8
Avoiding adverse reactions from TB medicines 7.0
Avoiding the acquisition or amplification of drug resistance 7.9
Cure or successful completion by the end of treatment 9.0
Culture conversion by month six 7.4
Death (survival) by the end of projected treatment 8.1
Treatment failure 8.7
Relapse 7.7
a Relative importance was rated on an incremental scale:1–3 points: Not important for making recommendations on the treatment of drug-resistant TB.4–6 points: Important but not critical for making recommendations on the treatment of drug-resistant TB.7–9 points: Critical for making recommendations on the treatment of drug-resistant TB.
Certainty of evidence and strength of recommendationsThe recommendations in these guidelines qualify their strength as well as the certainty of evidence on which they are based. The text of the recommendation itself should be read along with the accompanying remarks that summarize the evidence upon which the recommendation was made, the anticipated desirable and undesirable effects of the interventions to assess the balance of expected benefits to risks, and other considerations which are important for the implementation of the policy. The GDG also made a statement about research priorities within the different dimensions covered by each of the PICO questions (see Section E below).
The certainty of evidence is categorized into four levels (Table 3). The criteria used by the evidence reviewers to qualify the quality of available evidence are summarized in the Grading of Recommendations Assessment, Development and Evaluation (GRADE) tables annexed to these guidelines (online Annex 4). A number of factors may increase or decrease the certainty of evidence (see Figure 9.1 of (7)). The highest rating is usually assigned to data from randomized controlled trials (RCT) while evidence from observational studies is usually assigned a low or very low quality value at the start.
14
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Table 3. Certainty of evidence and definitions (8)
CERTAINTY OF EVIDENCE DEFINITION
High ( ) Further research is very unlikely to change our confidence in the estimate of effect.
Moderate ( ) Further research is likely to have an important impact on our confidence in the effect and may change the estimate.
Low ( ) Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low ( ) Any estimate of effect is very uncertain.
A recommendation may be strong or conditional. Apart from the quality of evidence, the strength of a recommendation is determined by the balance between desirable and undesirable effects, values and preferences, and costs or resource allocation (online Annex 5; (8)). For strong recommendations, the GDG is confident that the desirable effects of adherence to the recommendation outweigh the undesirable effects. For conditional recommendations, the GDG considers that desirable effects probably outweigh the undesirable effects. The strength of a recommendation has different implications for the individuals affected by these guidelines (Table 4).
Table 4. Implications of the strength of a recommendation for different users (adapted from (8))
PERSPECTIVE STRONG RECOMMENDATION CONDITIONAL RECOMMENDATION
For patients Most individuals in this situation would want the recommended course of action and only a small proportion would not. Formal decision aids are not likely to be needed to help individuals make decisions consistent with their values and preferences.
The majority of individuals in this situation would want the suggested course of action, but many would not.
For clinicians Most individuals should receive the intervention. Adherence to this recommendation according to the guidelines could be used as a quality criterion or performance indicator.
Recognize that different choices will be appropriate for individual patients, and that patients must be helped to arrive at a management decision consistent with their values and preferences. Decision aids may be useful in helping individuals to make decisions consistent with their values and preferences.
For policy-makers The recommendation can be adopted as policy in most situations.
Policy-making will require substantial debate and involvement of various stakeholders.
15
methoDs
DefinitionsRifampicin-resistant TB (RR-TB) refers to TB strains that are considered eligible for treatment with MDR-TB regimens (9). Rifampicin-resistant TB strains may be susceptible to isoniazid, or resistant to isoniazid (i.e. MDR-TB), or resistant to other medicines from the first-line group (poly-resistant) or from the second-line medicine group (e.g. XDR-TB) (10).
Drug-susceptibility testing (DST) refers to in vitro testing using either phenotypic methods to determine susceptibility or molecular techniques to detect resistance-conferring mutations to a particular medicine. New policy guidance on the use of line probe assay for the detection of resistance to second-line anti-TB drugs are now available (11).
A second-line TB medicine (drug or agent) is used to treat drug-resistant TB (see also Section B under WHO policy recommendations in these guidelines). For the treatment of RR-TB and MDR-TB, streptomycin is included as a substitute for second-line injectable agents when aminoglycosides or capreomycin cannot be used and susceptibility is highly likely. The core second-line TB medicines (or agents) refer to those in Groups A, B or C.
A shorter MDR-TB regimen refers to a course of treatment for RR-TB or MDR-TB lasting 9–12 months, which is largely standardized, and whose composition and duration follows closely the one for which there is documented evidence from different settings (12–14). The features and indications of this regimen are further elaborated in Section A under WHO policy recommendations in these guidelines.
Longer MDR-TB regimens are treatments for RR-TB or MDR-TB which last 18 months or more and which may be standardized or individualized. These regimens are usually designed to include a minimum number of second-line TB medicines considered to be effective based on patient history or drug-resistance patterns (1,9). These regimens were previously qualified as “conventional”, having been the mainstay of MDR-TB treatment before the 2016 update. The features and indications of longer regimens are further elaborated in Section B of the current document.
The treatment outcome categories used in these guidelines and the term relapse were applied according to the definitions agreed for use by TB programmes, unless otherwise specified (10,15).
For the purposes of the reviews conducted for these guidelines, a serious adverse event (SAE) is defined as one which was classified as Grade 3 (severe) or Grade 4 (life-threatening or disabling) (16), or which led to the medicine being stopped permanently.
Assessment of evidence and its gradingTeams of experts were commissioned to assess the evidence for the PICO questions and their outcomes through systematic literature reviews following a standard methodology (17). Evidence reviewers are listed in Annex 2; more details on the methods used in unpublished studies are presented in Annex 6 (online) and in published studies referenced under the respective sections. Titles, abstracts and full text of potentially relevant literature were screened using key subject words and text words. Authors in the field and members of the GDG were
16
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
contacted to identify missing studies or studies in progress. Individual patient-level data were used to address PICO 1 (adults (18) and children; see also Section B), for PICO 3 (shorter MDR-TB regimens; see Section A) and PICO 4 (use of surgery (19); see Section D).
Relative effects (relative risks or odds ratios of an event) were calculated from pooled data in individual or aggregated formats from the included studies. Absolute effects and risk differences were used to express the magnitude of an effect or difference between the intervention and comparator groups. Where possible, adjustments were made to reduce risk of bias and confounding. More details are provided in the notes on the GRADE evidence profiles that were used to summarize the results of systematic reviews done for each question (online Annex 4). The evidence profiles were prepared using GRADEPro software – an online tool to create guideline materials (see http://gdt.guidelinedevelopment.org). The certainty of the evidence was assessed using the following criteria: study design, limitations in the studies (risk of bias), imprecision, inconsistency, indirectness, publication bias, magnitude of effect, dose–effect relations and residual confounding (7).
The GDG membership represented a broad cross-section of future users of the guidelines as well as affected persons (including the patient). Ahead of the GDG meeting held at the WHO headquarters in Geneva, Switzerland, between 9 and 11 November 2015, one or more discussants were identified from among the GDG members to assess the evidence for each of the PICO questions and to present his or her perspective on the implications of the findings during the meeting. Drafts of the review reports were shared with the GDG members ahead of the meeting (Annexes 4 and 6). During the days of the meeting and in the following weeks additional analyses were shared with the group upon their demand. The GRADE evidence profiles were discussed by the GDG ahead of formulating the recommendations. The group used the “Evidence to Decision” tables via the GRADEPro interface to capture the content of the discussions, make judgements, annotate the different considerations, develop the wording and strength of the recommendations, and add the remarks that accompany each recommendation (online Annex 5).
Apart from the quality of evidence, the strength of a recommendation was determined by assessing the balance between desirable and undesirable effects, values and preferences, considerations on equity, resource use and feasibility. In the preparation of PICO questions and outcomes, and in the discussions of the evidence before, during and after the meeting, the GDG members paid particular attention to the spectrum of values and preferences attached to the recommendations by the different users. One important factor that lowered the strength of all recommendations made in these guidelines was the variability in values and preferences of those affected by these policies as perceived by the GDG members. Resource use was not assessed by means of formal cost-effectiveness studies, and the GDG assessed it from the perspective of the patient and the health services, in terms of feasibility and opportunity cost. Decisions on the certainty of evidence and on the wording of a recommendation and of its strength were largely made through moderated discussion. Any disagreements were resolved by a group decision on an acceptable position. For the recommendation on surgery (part of PICO 4), the final wording was agreed through voting. None of the recommendations for these guidelines were strong and all the certainty in the evidence was rated as very low.
17
methoDs
External reviewThe ERG commented on the questions during their formulation (in mid-2015) and on a draft text of the guidelines, including the recommendations, following comments from the GDG (in February 2016). Six reviewers provided substantive comments on the draft of the guidelines.
Publication, implementation, evaluation and expiryThese guidelines were published on the World Health Organization Global TB Programme (WHO/GTB) website (http://www.who.int/tb/areas-of-work/drug-resistant-tb/treatment/resources/en/) as freely downloadable pdf files from 13 May 2016. The main text of the guidelines (without Annexes 4, 5 and 6) will also be made available in print version in late 2016. The evidence reviews as well as the recommendations are also being published separately in peer-reviewed journals to improve the dissemination of the main messages. The changes to the policy guidance will also be reflected in a forthcoming revision of the WHO implementation handbook for programmatic management of drug-resistant TB planned later in 2016 (9).
WHO will work closely with its regional and country offices, as well as technical and funding agencies and partners, to ensure wide communication of the updated guidance in technical meetings and training activities. WHO/GTB will review and update these guidelines within four to five years after their publication, or earlier if new evidence becomes available (e.g. on bedaquiline and delamanid use). These changes will also be reflected in a forthcoming revision of the implementation handbook (9).
18
WHO policy recommendations
A. The effectiveness and safety of standardized regimens lasting up to 12 months for the treatment of patients with MDR-TB when compared with longer treatment
Recommendation
In patients with RR-TB or MDR-TB who were not previously treated with second-line drugs and in whom resistance to fluoroquinolones and second-line injectable agents was excluded or is considered highly unlikely, a shorter MDR-TB regimen of 9–12 months may be used instead of the longer regimens (conditional recommendation, very low certainty in the evidence).
Justification
The interest in reducing the duration of treatment for MDR-TB has motivated a number of initiatives to treat patients with shorter regimens under programmatic as well as trial conditions. Experience and data on the effect of shorter MDR-TB regimens was limited until recently and before the 2016 update of the guidelines WHO advised that shorter regimens were to be used only under operational research conditions and with close monitoring for effectiveness and safety during and after the end of treatment. In the past few years, results from three studies of MDR-TB patients on shorter regimens have been published and other observational studies as well as a randomized controlled trial in different settings have begun (12–14, 20). Early results from observational studies in Bangladesh, Cameroon and Niger using regimens lasting 12 months or less have shown much higher likelihood of treatment success compared with longer regimens when treating patients with specific inclusion criteria (such as lack of previous exposure to second-line anti-TB medications). Given the published data and potential impact of shorter regimens on treatment cost and affordability, WHO proceeded with the evidence assessment. A PICO question was developed to assess the effectiveness of the shorter MDR-TB treatment regimens lasting up to 12 months and to inform a possible policy change with respect to their use and application (Annex 3; Question 3).
The evidence reviewed for this question compared the treatment outcomes for confirmed RR-TB or MDR-TB patients treated with these regimens with those of patients treated with longer regimens (online Annex 4; Section 1). The shorter MDR-TB treatment regimens were standardized in content and duration and split into two distinct parts. The first was an intensive phase of four months (extended up to a maximum of six months in case of lack of sputum smear conversion) and included the following drugs: gatifloxacin (or moxifloxacin), kanamycin, prothionamide, clofazimine, high-dose isoniazid, pyrazinamide and ethambutol. This was followed by a continuation phase of five months with the following medicines:
19
Who policy recommenDations
gatifloxacin (or moxifloxacin), clofazimine, pyrazinamide and ethambutol (prothionamide was kept in the continuation phase in earlier studies). In the studies, patients were placed on these regimens based on a set of criteria, and individuals who had prior exposure to second-line TB drugs were excluded from the analysis. No modifications were made to the shorter MDR-TB regimen if previously unknown drug resistance was detected after start of treatment. The recommendation made on the shorter MDR-TB regimen applies only to a regimen profile with similar characteristics of the ones studied. This is because the substitution or exclusion of one or more of the medicines of this regimen may affect its overall performance which is not possible to predict given the lack of evidence of the impact of such modifications (see “Implementation considerations” below).
All data used to assess the shorter MDR-TB treatment regimens were derived from observational studies (see online Annex 6 for background, methods and summary of findings). Individual patient data from Bangladesh (n=493; supported by the Damien Foundation), Uzbekistan (n=65; supported by Médecins sans Frontières (MSF)) and Swaziland (n=24; MSF) as well as aggregated data from Cameroon (n=150) (13), Niger (n=65) (14) and seven sub-Saharan African countries (n=408; supported by the International Union Against Tuberculosis and Lung Disease (UNION)) were included in the analysis (total number of observations=1205, of whom 89 cases were lost to follow-up and were therefore excluded in certain analyses). These were compared with the outcomes of patients without previous exposure to second-line TB drugs who were included in the adult individual patient data (aIPD) analysis (n=7665) (18) (see also Section B below for more details on the aIPD). The standard outcomes used in the intervention and comparator arms largely complied with the standardized outcomes used by TB programmes (10,15,21).
The analyses performed for the evidence assessment showed that patients who met specific inclusion criteria for receiving the shorter MDR-TB treatment regimens had a statistically significant higher likelihood of treatment success than those who received longer regimens – 90% versus 78% when success was compared with treatment failure/relapse/death (Table 5) and 84% versus 62% when compared with treatment failure/relapse/death/loss to follow-up (see also online Annex 4). The number of relapses was very low, although this may be due to the relatively small number of patients followed up. As expected, treatment success was lower among patients with additional resistance to pyrazinamide and/or fluoroquinolones on shorter MDR-TB regimens, even if in general it remained high and exceeded that in the patients on longer regimens (although the differences were not statistically significant).
20
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Table 5. Treatment success in patients treated with a shorter MDR-TB regimen vs longer MDR-TB regimensa
RESISTANCE PATTERN SHORTER MDR-TB REGIMEN LONGER MDR-TB REGIMEN
N % (95% CL) N % (95% CL)
All cases regardless of pyrazinamide and fluoroquinolone susceptibility
1008/1116 90.3% (87.8%– 92.4%)
4033/5850 78.3% (71.2%–84%)
Pyrazinamide resistant; fluoroquinolone resistant
19/28 67.9% (47.6%–84.1%)
81/137 59.1% (50.6%–67.1%)
Pyrazinamide resistant; fluoroquinolone susceptible
90/100 88.8% (47.3%–98.6%)
840/1075 81.4% (71.6%–88.4%)
Pyrazinamide susceptible; fluoroquinolone resistant
12/15 80.0% (50.0%–94.1%)
72/120 64.4% (49.6%–76.9%)
Pyrazinamide susceptible; fluoroquinolone susceptible
121/125 96.8% (77.3%–99.6%)
890/1119 83.5% (75.7%–89.2%)
a Treatment success (cured or treatment completed (10,15)) versus treatment failure/relapse/death in patients not previously treated with second-line TB medications; percentages shown have been adjusted where possible (see also online Annex 4; Section 1 for more details).
Until more evidence is available, WHO recommends that the shorter MDR-TB regimen not be used in patients who have been previously treated with second-line drugs for more than one month or who have documented or are likely to have strains resistant to medicines in the regimen. Preferably, resistance to at least fluoroquinolones and the injectable agent used in the regimen is excluded before starting treatment by in vitro testing. In the absence of such testing, patients who are highly unlikely to be infected with resistant strains based on history of exposure, use of second-line medicines at country level or recent representative surveillance data may also be eligible for the shorter MDR-TB regimen (see “Implementation considerations” below).
Subgroup considerations
Rifampicin-resistant TB (RR-TB) without MDR-TB. All patients – children or adult – with RR-TB in whom isoniazid resistance is not confirmed may be treated with the shorter MDR-TB treatment regimen.
Resistance additional to MDR-TB. For patients infected with strains known or strongly suspected of being resistant to one or more drugs in the shorter MDR-TB treatment regimen (e.g. pyrazinamide) it is recommended not to use the shorter regimen until more evidence becomes available about its performance in such a situation.
21
Who policy recommenDations
People living with HIV need to be given the same consideration for treatment with the shorter MDR-TB treatment regimen as people who are HIV seronegative.
Children were generally excluded from studies of shorter MDR-TB treatment regimens. However, given that the same medicines have been in use in paediatric MDR-TB regimens for many years, there is no plausible biological reason to believe that these regimens are less effective or safe in children than in adults. As a result, it is recommended that children with confirmed RR-TB or MDR-TB be given the same consideration for treatment with a shorter MDR-TB treatment regimen as adults.
Pregnancy was an exclusion criterion for the shorter MDR-TB treatment regimen studies. Two of the core components of the shorter MDR-TB regimens – the injectable agent and ethionamide (or prothionamide) – are usually contraindicated in pregnancy (9). Withholding these medicines from the shorter MDR-TB treatment regimen could seriously compromise its effectiveness. Thus for pregnant women, it is recommended that a longer individualized regimen be used which can allow the inclusion of four or more effective second-line TB medicines with no known teratogenic properties (see Section B below).
Extrapulmonary disease. The findings from studies of shorter MDR-TB regimen were limited to patients with pulmonary disease, and they cannot be extrapolated directly to extrapulmonary TB cases. No recommendation is thus possible at this stage to use the shorter regimen in patients with extrapulmonary MDR-TB.
Implementation considerations
In order to reproduce the high cure rates achieved by the studies included in the reviews for this guidance, all efforts need to be made to avoid the acquisition of additional resistance, by ensuring careful selection of patients to be enrolled, and effective patient support to enable full adherence to treatment. It is recommended that patients be tested for susceptibility or resistance to fluoroquinolones and to second-line injectable agents used in the regimen before being started on a shorter MDR-TB regimen. Patients with strains resistant to any of the two groups of medicines are to be transferred to a longer MDR-TB regimen (see Section B below).
The availability of reliable and rapid tests would be valuable in deciding (within a few days) which patients would be eligible for the shorter MDR-TB regimen, and what modifications to longer, individualized MDR-TB regimens are necessary based on the resistance detected. In patients with confirmed RR-TB or MDR-TB, WHO now recommends that the GenoType M. tuberculosis drug-resistant second-line assay (MTBDRsl) be used as an initial direct test, instead of phenotypic culture-based DST, to detect resistance to fluoroquinolones and second-line injectable drugs (conditional recommendation; certainty of evidence low to moderate (11),(22)). This applies to testing in both children and adults. While resistance-conferring mutations to fluoroquinolones detected by the MTBDRsl assay are highly correlated with phenotypic resistance to ofloxacin and levofloxacin, the correlation with moxifloxacin and gatifloxacin is less clear and the inclusion of moxifloxacin or gatifloxacin in a MDR-TB regimen is best guided by phenotypic DST results.
22
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
In settings in which laboratory capacity for DST to fluoroquinolones and injectable agents is not yet available, treatment decisions would need to be guided by the likelihood of resistance to these medicines, informed by the patient’s clinical history and recent representative surveillance data.
The evidence for the effectiveness and safety of the shorter MDR-TB regimen derives from studies where treatment was administered under fairly standardized conditions with relatively little variation in the content and duration, and with close monitoring. Thus, the recommendation for the shorter MDR-TB regimen is premised on the use of a regimen similar in composition and duration to those used in observational studies. Any replacement of medicines or any changes to the duration are only to be considered within the parameters applied in these studies (e.g. gatifloxacin replaced by moxifloxacin; prothionamide replaced by ethionamide; intensive phase is prolonged up to six months in case of no sputum conversion).
Two staples of the regimen, clofazimine and high-dose isoniazid may be difficult to procure in some countries. Moreover, there are no good paediatric formulations of clofazimine and dividing the capsule into smaller doses is almost impossible, making dosing in children uncertain. Given the global shortage in the supply of quality-assured gatifloxacin in recent years, the sites where observational studies have been conducted have had to substitute this agent with moxifloxacin. This led to an increase in the overall price of the regimen, although the costs for quality-assured moxifloxacin have since declined. The implementation of these guidelines at the national level needs to ensure that sufficient quantities of these medicines are available to meet the demand and that no stock-outs occur.
Monitoring and evaluation
Patients who receive a shorter MDR-TB treatment regimen need to be monitored during treatment and after completion of treatment using schedules of relevant clinical and laboratory testing which have been successfully applied in the studies under field conditions. The WHO framework for active TB drug-safety monitoring and management (aDSM) needs to be applied to ensure appropriate action to monitor and respond promptly to adverse events (23,24). This could be conducted alongside the routine programme monitoring for patient response and for treatment outcomes that has been conducted worldwide for many years (10,25).
Continued efforts to reduce MDR-TB treatment duration, both under observational and trial conditions, is ongoing and is expected to increase the knowledge base for the effectiveness/efficacy and safety of the regimens under different field conditions, patient subgroups and composition – including new medicines.
B. The optimal combination of medicines and approach towards regimen design for TB patients with RR-TB and MDR-TB
As part of the GDG discussion on the design of MDR-TB regimens for adults and children, a regrouping of TB medicines from that being formerly used is proposed (1,9). These include medicines used in first-line TB treatment that may also have a role in strengthening MDR-TB
23
Who policy recommenDations
regimens (Table 6). When reclassifying these medicines, the GDG assessed the available evidence and the associated level of certainty, as well as other considerations relating to the balance between anticipated desirable and undesirable effects, and feasibility of implementation. WHO considers that currently only the medicines shown in Table 6 have a role in MDR-TB treatment under programmatic conditions.4
Table 6. Medicines recommended for the treatment of RR-TB and MDR-TBa
Group A. Fluoroquinolonesb Levofloxacin
Moxifloxacin
Gatifloxacin
Lfx
Mfx
Gfx
Group B. Second-line injectable agents Amikacin
Capreomycin
Kanamycin
(Streptomycin)c
Am
Cm
Km
(S)
Group C. Other core second-line agentsb Ethionamide / prothionamide
Cycloserine / terizidone
Linezolid
Clofazimine
Eto / Pto
Cs / Trd
Lzd
Cfz
Group D. Add-on agents (not part of the core MDR-TB regimen)
D1 Pyrazinamide
Ethambutol
High-dose isoniazid
Z
E
Hh
D2 Bedaquiline
Delamanid
Bdq
Dlm
D3 p-aminosalicylic acid
Imipenem–cilastatind
Meropenemd
Amoxicillin-clavulanated
(Thioacetazone)e
PAS
Ipm
Mpm
Amx-Clv
(T)
a This regrouping is intended to guide the design of longer regimens; the composition of the recommended shorter MDR-TB regimen is standardized (see Section A).
b Medicines in Groups A and C are shown by decreasing order of usual preference for use (subject to other considerations; see text).
c Refer to the text for the conditions under which streptomycin may substitute other injectable agents. Resistance to streptomycin alone does not qualify for the definition of XDR-TB (26).
d Carbapenems and clavulanate are meant to be used together; clavulanate is only available in formulations combined with amoxicillin.
e HIV-status must be confirmed to be negative before thioacetazone is started.
4 Other medicines than those in Table 6 are currently being investigated for use in TB (see Figure 8.3 of reference (25)).
24
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
B1. Longer treatment regimens for RR-TBThe recommendations in this section cover all forms of RR-TB, including also patients with strains susceptible to isoniazid, or with additional resistance to isoniazid (i.e. MDR-TB), or resistant to other medicines from the first-line group (poly-resistant) or from the second-line group (e.g. XDR-TB) (online Annex 4; Section 2).
Recommendations5
• In patients with RR-TB or MDR-TB, a regimen with at least five effective TB medicines during the intensive phase is recommended, including pyrazinamide and four core second-line TB medicines – one chosen from Group A, one from Group B, and at least two from Group C6 (conditional recommendation, very low certainty in the evidence). If the minimum number of effective TB medicines cannot be composed as given above, an agent from Group D2 and other agents from Group D3 may be added to bring the total to five.7
• In patients with RR-TB or MDR-TB, it is recommended that the regimen be further strengthened with high-dose isoniazid and/or ethambutol (conditional recommendation, very low certainty in the evidence).
Justification
Treatment of MDR-TB in adults and children with longer second-line regimens has been known to increase the likelihood of cure and lower the risk of chronicity and death (18,27). This section refers to MDR-TB treatment regimens that are of longer duration than the shorter MDR-TB regimen described in Section A. The composition and duration of longer regimens are based on a number of factors, including the combination of sufficient agents considered to be effective, the balance of expected benefits to harms, and the response or reactions to treatment in the individual patient. Recommendations for the design of these regimens have been issued for a number of years and have been implemented in many countries worldwide.
The evidence base for the effectiveness of many of the medicines used in MDR-TB regimens relies heavily on observational studies with only a few having been studied under randomized controlled conditions. As a result, the overall quality of the evidence is graded as low or very low.
Adults. The evidence that informed the adult treatment recommendations is based on two main sources (see GRADE tables in online Annex 4; Section 2): (i) an IPD meta-analysis including data on 9153 mostly adult patients (only 76 were under <15 years) from studies that incorporated three systematic reviews of MDR-TB treatment outcomes published until 2010 (18); and (ii) additional evidence published until August 2015 that summarized a study-level meta-analysis conducted expressly for the revision of the current guidelines (see online Annex 6 for background, methods and summary of findings). All studies included had to
5 See (6) for the new recommendation on the use of delamanid in patients aged 6-17 years.6 Group A=levofloxacin, moxifloxacin, gatifloxacin; Group B=amikacin, capreomycin, kanamycin, (streptomycin); Group
C=ethionamide (or prothionamide), cycloserine (or terizidone), linezolid, clofazimine; in children with non-severe disease Group B medicines may be excluded.
7 Group D2=bedaquiline, delamanid; Group D3=p-aminosalicylic acid, imipenem-cilastatin, meropenem, amoxicillin-clavulanate, (thioacetazone). Bedaquiline is only recommended for adults. Following the revised WHO policy published in October 2016, delamanid may now also be used in patients aged 6–17 years.
25
Who policy recommenDations
report treatment outcomes, have less than 10% extrapulmonary cases (unless pulmonary and extrapulmonary cases were reported separately), and include at least 25 adult patients with bacteriologically confirmed MDR-TB.
The best available evidence has been used to construct recommendations for a regimen that has high relapse-free cure rates, reduced likelihood of death and low emergence of additional resistance while minimizing SAEs. In the case of high-dose isoniazid, the results from a separate, paediatric individual patient data (pIPD) meta-analysis were extrapolated to adults.
Children. These treatment regimen recommendations are based on the pIPD meta-analysis that included both published and unpublished data on 974 children up until September 30, 2014 (see GRADE tables in online Annex 4 Section 3; and online Annex 6, Section 3 for background, methods and summary of findings). Datasets were eligible if they included a minimum of three children (aged <15 years) within a defined treatment cohort who were treated for clinically diagnosed or bacteriologically confirmed pulmonary or extrapulmonary MDR-TB, and for whom treatment outcomes were reported, using standard WHO TB case definitions. Eligible study designs included controlled and uncontrolled retrospective and prospective studies and case series. No randomized control trials were included (or known to exist) and as a result the overall certainty of the evidence is very low.
Children with XDR-TB were excluded from the analysis (n=36) as their treatment regimens were not considered to be comparable with those of other MDR-TB patients and their numbers were too low to analyse independently. For analysis, children were split in two different cohorts: (i) those who were bacteriologically confirmed as having MDR-TB, and (ii) those who were clinically diagnosed with MDR-TB. When making treatment recommendations, preference was given to the results in the bacteriologically confirmed cohort, as this group had a higher certainty of diagnosis. The children with bacteriologically confirmed MDR-TB were more likely to have severe disease; they had statistically significant higher levels of malnutrition, severe disease on chest radiography, severe extrapulmonary disease and were more likely to be HIV positive. Children without these features were considered to have milder forms of disease.
Where data on children were unavailable, evidence from adults was extrapolated to children. The best available evidence was used to construct recommendations for a regimen that has high relapse-free cure rates, reduces the likelihood of death and of the emergence of additional resistance while minimizing SAEs.
Remarks
Based on the evidence reviews, it is recommended that the MDR-TB regimen be composed of at least five drugs that are likely to be effective, i.e. four core second-line drugs plus pyrazinamide. If a minimum of four core second-line TB medicines cannot be reached by using agents from Groups A to C alone, drugs from Group D2 (in adults; delamanid may also be used in patients aged 6–17 years) or, if not possible, from Group D3 are added. Pyrazinamide is added routinely unless there is confirmed resistance from reliable DST, or well-founded reasons to believe that the strain is resistant, or there is risk of significant toxicity. If pyrazinamide is compromised or cannot be used, the regimen may also be strengthened with an additional agent from Groups
26
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
C or D (preferably D2, or if not possible, from D3). Other agents from Group D1 are included if they are considered to add benefit (e.g. high-dose isoniazid in patients without high-level isoniazid resistance). The total number of TB medicines to include in the regimen needs to balance expected benefit with risk of harms and non-adherence when the pill-burden is high.
The recommendations for children are mostly identical to those of adults. However, in children with mild forms of disease, the harms associated with Group B medications (second-line injectable agents) outweigh potential benefits and therefore Group B medications may be excluded in this group of children. The GDG based this decision on the observation that treatment success in children with clinically diagnosed disease (which was associated with less severe clinical or radiological manifestations) was high and not significantly different in patients treated with and without a Group B medication (93.5% versus 98.1%; n=219; see online Annex 4). In October 2016, WHO published its revised policy on the use of delamanid in patients aged 6–17 years following the advice of a separate Guideline Development Group convened expressly to review evidence for its effectiveness and safety in children and adolescents (6). Delamanid is now conditionally recommended for use in these age-groups as an add-on agent to a longer MDR-TB regimen. Lack of data at this stage on the use of bedaquiline in children precludes a recommendation in patients aged under 18 years.
WHO recommends that all TB patients – children or adult – diagnosed with strains shown to be resistant to rifampicin be placed on a MDR-TB treatment regimen. In such cases, isoniazid is added alongside the rest of the MDR-TB regimen until susceptibility results are confirmed. If isoniazid susceptibility cannot be tested, isoniazid may still be added to the regimen unless there are well-founded grounds to consider the drug ineffective.
Desirable and undesirable effects
Group A. FluoroquinolonesBased on the evidence reviews, the GDG concluded that treatment with later-generation fluoroquinolones (defined for these guidelines as high-dose levofloxacin,8 moxifloxacin, and gatifloxacin) significantly improves treatment outcomes in adults with RR-TB and MDR-TB. This group of medicines is considered to be the most important component of the core MDR-TB regimen and the benefits from their use outweigh potential risks. They should therefore always be included unless there is evidence for absolute contraindication for their use. The order of preference for the inclusion of later-generation fluoroquinolones in longer MDR-TB regimens is: high-dose levofloxacin, moxifloxacin and gatifloxacin. It is recommended that ofloxacin be phased out from MDR-TB regimens and ciprofloxacin never used due to the limited evidence of their effectiveness. Although the pIPD had high levels of confounding and insufficient numbers to discern the treatment effect of high-dose levofloxacin, moxifloxacin and gatifloxacin, data from adults with MDR-TB show a treatment benefit. Therefore these recommendations have been extrapolated to children.
8 For levofloxacin, high-dose is usually defined as 750 mg/day or more. The definition of high-dose will be the subject of discussion of another WHO consultation planned in early 2017.
27
Who policy recommenDations
Fluoroquinolones in general have a good safety profile and considering the seriousness of RR-TB and MDR-TB, the potential for drug-related harms is offset by the benefits from their use. Although adverse events were poorly recorded, in the study-level meta-analysis the frequency of SAEs attributed to fluoroquinolones was low (1.2%–2.8%; Table 7). Moxifloxacin and other fluoroquinolones carry a risk of QT prolongation, which is a cause for concern when used in combination with medications that have a similar effect, such as bedaquiline, delamanid and clofazimine.
Table 7. Serious adverse events (SAEs) in patients on MDR-TB treatment regimens
MEDICINE COHORTS USING THE DRUG AND
REPORTING SAEs (N)
PATIENTS RECEIVING MEDICINE
(N)
SAEs ATTRIBUTED TO INDIVIDUAL MEDICINE
N PATIENTS % (95%CL)a
Pyrazinamide 19 2023 56 2.8% (2.1%–3.7%)
Ethambutol 16 1325 6 0.5% (0.2%–1.1%)
Second-line injectable agent
19 2538 184 7.3% (6.2%–8.4%)
Ofloxacin or ciprofloxacin
9 1408 40 2.8% (1.9%–4.1%)
Other fluoroquinolones 13 827 10 1.2% (0.6%–2.4%)
Ethionamide/prothionamide
17 2106 173 8.2% (7.0%–9.6%)
Cycloserine 16 2140 96 4.5% (3.6%–5.5%)
p-aminosalicylic acid 16 1706 208 12.2% (10.6%–13.9%)
Linezolid 8 190 28 14.7% (10.0%–20.6%)
a values from fixed effects meta-analysis.
Source: study-level meta-analysis (Bastos M, Lan Z, Menzies R. An updated systematic review and meta-analysis for treatment of multidrug-resistant tuberculosis, 2016. [under review, 25 July 2016]); 43/73 studies reported adverse events, but only 5/43 studies reported Grade 3–4 adverse events, and 28/43 studies reported TB drugs being stopped due to adverse events; for linezolid estimate is based on an aggregated analysis of eight observational studies (28–35) (see also online Annex 4; Section 2 for the respective GRADE tables).
Concerns about dysglycaemia reported in 2006 in patients treated with gatifloxacin for conditions other than TB led the parent company to stop manufacturing the medicine (36), and a global shortage in quality-assured formulations of this drug ensued. A trial of a four-month standardized regimen for drug-susceptible TB which included gatifloxacin (400 mg once daily) published in 2014 reported no significant risk of hyperglycaemia associated with exposure to gatifloxacin (37). Although in general adverse events were poorly recorded in the studies assessed for this review, the data showed that there was a lower risk of SAEs in patients taking gatifloxacin than in those who did not, including those receiving no fluoroquinolones (3.6% vs 8%, not statistically significant; see online Annex 4; Section 2). The frequency of SAEs
28
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
associated with gatifloxacin was thus comparable to the one associated with fluoroquinolones in the study-level meta-analysis (Table 7).
Group B. Second-line injectable agentsBased on the available evidence, second-line injectable agents were associated with an increased likelihood of treatment success when included in a longer MDR-TB treatment regimen (the small size of the population not receiving an injectable agent in the aIPD limited the power to quantify the impact of this class of agents). It is therefore recommended that adults with RR-TB or MDR-TB always receive a second-line injectable agent as part of their regimen unless there is an important contraindication. In children with mild forms of disease, however, the harms associated with this group of medications may outweigh potential benefits and therefore injectable agents may be excluded for children. The GDG based this decision upon the observation that in children with clinically-diagnosed disease – which was associated with less severe clinical manifestations – treatment success was in general high and did not differ significantly between patients who received Group B medication and those who did not (see above and online Annex 4; Section 3). For children with additional resistance to fluoroquinolones, Group B medications are best retained.
The choice from among amikacin, capreomycin or kanamycin would be determined by the likelihood of effectiveness and implementation considerations. While streptomycin is not usually included with the second-line drugs it can be used as the injectable agent of the core MDR-TB regimen if none of the other three agents can be used and if the strain is unlikely to be resistant to it. Streptomycin resistance by itself does not qualify to define XDR-TB (26) and DST methods for it are not considered accurate or reproducible (38).
Adverse events need to be carefully monitored while using second-line injectable agents. Hearing loss and nephrotoxicity are the most frequent and serious adverse reactions. However, skin rash, hypersensitivity and peripheral nephropathy may also occur. The risk of adverse reactions increases with the total cumulative dose of second-line injectable agents, so particular caution should be given to people who have previously received these medications, including streptomycin as part of a regimen for drug-susceptible TB. In children especially, hearing loss can have a profound impact on their quality of life, affecting acquisition of language and the ability to learn at school.
Although adverse events are poorly reported, the data for this review found that 7.3% of adult patients (10.1% in children) had SAEs attributed to second-line injectable agents (Table 7). In a study focused on hearing loss in children with TB (30% of the children were HIV-infected), 24% of children treated for MDR-TB with an injectable agent had hearing loss and 64% of children had progression of hearing loss after completing the treatment (39).
Group C. Other core second-line agentsWhen designing the core MDR-TB treatment regimen, two or more of the following four medicines are to be included: ethionamide (or prothionamide), cycloserine (or terizidone), linezolid and clofazimine, usually in this order of preference, unless the balance of benefits-to-harms for the individual patient demands otherwise. Group C agents are included to bring the total number of effective second-line TB medicines in the core regimen to at least four during
29
Who policy recommenDations
the intensive phase. If pyrazinamide cannot be included or counted upon, another agent is added. Ethionamide can be used interchangeably with prothionamide, and terizidone can be used instead of cycloserine.
Given the lack of reliable DST for drugs in Group C, the choice of which ones to include is determined by the balance of desirable to undesirable effects and implementation considerations. The aIPD and pIPD meta-analyses showed an increase in the likelihood of treatment success when MDR-TB treatment regimens included cycloserine (marginally statistically significant) and ethionamide/prothionamide (statistically significant only in adults). In the pIPD, the vast majority of children received ethionamide or prothionamide and significance testing was therefore not always possible for want of sufficient number of controls. In contrast to cycloserine/terizidone and ethionamide/prothionamide, RCT data from a few recent studies are now available for clofazimine and linezolid (32,40,40). Linezolid has shown a statistically significant treatment benefit in both RCT and in cohort studies in adult patients, with this benefit being most pronounced in patients with additional resistance to fluoroquinolones and with XDR-TB (41). Both the aIPD and pIPD showed no significant increase in treatment success associated with the use of clofazimine, while linezolid was used too sparingly in the cohorts included to allow a conclusive analysis (18).
Ethionamide and prothionamide cause gastrointestinal disturbance, in particular vomiting, which can limit tolerability. Hypothyroidism may occur, especially in combination with PAS, but is reversible upon cessation of drugs. This review found that 8.2% of patients had SAEs due to ethionamide or prothionamide, although adverse events were poorly reported across the individual studies (Table 7).
Cycloserine has a well-established association with neuropsychiatric adverse reactions. However, the aIPD meta-analysis in adults revealed low levels of SAEs (4.5% in the study-level meta-analysis conducted for this update). A meta-analysis published in 2013 comparing the adverse effects of cycloserine with terizidone found that terizidone had no to little benefit over cycloserine with regard to adverse reactions (42).
Adverse reactions of linezolid include lactic acidosis, thrombocytopenia and anaemia. These can be severe and life threatening, although they are reversible with cessation of the drug or on some occasions by lowering its dose (usually from 600 mg daily to 300 mg daily) (9). Haematologic toxicities are less common with current strategies of once-daily dosing. Peripheral neuropathy may or may not improve with cessation of the drug. Optic neuropathy should be treated as a medical emergency. Given the potential seriousness of the adverse reactions associated with linezolid the decision to use it must balance its risks and benefits, and the availability of other TB medicines. Its use needs to be accompanied by close monitoring for adverse events. Where this is not possible, linezolid would best be reserved for MDR-TB patients who have additional drug resistance, or XDR-TB patients, or those who are intolerant to other components of the core regimen.
Clofazimine probably contributes to the sterilizing function of MDR-TB regimens where pyrazinamide is not effective. Although the single published for clofazimine use in MDR-TB had serious methodological concerns, it showed a statistically significant treatment
30
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
benefit associated with clofazimine use (40). However, much of the evidence for the effect of clofazimine in MDR-TB is based upon observational studies, which showed conflicting or inconclusive findings (43). One of the main adverse effects of clofazimine is skin discoloration/darkening, which may be distressing to patients. In the RCT, the adverse events reported were mostly limited to skin conditions and discoloration, and did not lead to discontinuation in the use of the drug. Overall, small rates of adverse events were noted in the observational studies and SAEs appear to be relatively uncommon. Clofazimine may prolong the QT interval, so caution is advised when using this medication in combination with other drugs also known to have the same effect.
Group D. Add-on agentsThis group of medicines includes drugs that do not form part of the core second-line agents. It is split into three subgroups:
Group D1 consists of pyrazinamide, ethambutol and high-dose isoniazid. These agents are usually added to core second-line medications, unless confirmed resistance, pill burden, intolerance or drug–drug interaction outweigh their potential benefits.
The aIPD showed improved likelihood of success (versus treatment failure, relapse or death combined) in patients who had pyrazinamide included in their regimens. This effect was significant both statistically and in absolute terms. The pIPD did not show a significant treatment effect with use of pyrazinamide. In many settings, RR-TB strains frequently have additional resistance to pyrazinamide. While it would be desirable to avoid giving pyrazinamide to patients whose strains are resistant to the drug, it is acknowledged that reliable DST for pyrazinamide is very often unavailable in resource-constrained settings. Although adverse events were poorly reported, data from the study-level meta-analysis showed that 2.8% of patients who received pyrazinamide had SAEs (Table 7). The balance of desirable to undesirable effects favours the addition of pyrazinamide to the core second-line MDR-TB regimen by default, unless there is confirmed resistance from reliable DST, or well-founded reasons to believe that the strain is resistant, or there are other contraindications for its use, particularly risk of significant toxicity. As for the drugs from the core regimen, if pyrazinamide is compromised or cannot be used, more agents from Group C and subsequently Group D are added until at least five effective medicines are available in the intensive phase.
The recommendation for the inclusion of isoniazid9 in adult MDR-TB regimens is largely based on evidence from the analysis of pIPD. This analysis showed a statistically significant increased likelihood of treatment success (versus treatment failure, relapse or death combined) in children with bacteriologically confirmed MDR-TB, even after adjustment for age, HIV status, sex, TB disease severity and treatment centre (treatment with high-dose isoniazid was almost exclusively done in South African sites). An RCT of high-dose isoniazid therapy for MDR-TB in adults found no increased risk of hepatotoxicity (44). Additionally, high-dose isoniazid was very well tolerated in children with drug susceptible tuberculous meningitis in a large cohort study from the Western Cape (45).
9 For isoniazid, the definition of high-dose will be the subject of discussion of another WHO consultation planned in early 2017.
31
Who policy recommenDations
Isoniazid is recommended alongside a full MDR-TB regimen in patients with rifampicin-resistant strains confirmed or suspected to be susceptible to isoniazid. High-dose isoniazid is one of the core components of the shorter MDR-TB treatment regimen (see Section A above). Strains bearing mutations in the promoter region of the inhA gene may have a minimum inhibitory concentration (MIC) to isoniazid that is low enough to be overcome by high-dose isoniazid. In such settings the drug may still add benefit (46). However, this mutation has been associated with high-level ethionamide resistance (47) and therefore, if present, ethionamide (or prothionamide) may have to be replaced in the regimen. In settings with elevated prevalence of high-level isoniazid resistance associated with katG mutations, high-dose isoniazid may be less effective and therefore its routine use may not be warranted. In such a situation the susceptibility to ethionamide (or prothionamide) is not affected and it can be used in combination with high-dose isoniazid.
The aIPD did not show any statistically significant association between the use of ethambutol and likelihood of treatment success. Ethambutol may cause ocular toxicity, which can be difficult to diagnose in young children, although this risk is reduced if the dose does not exceed recommended limits. SAEs were reportedly associated in 0.5% of cases in the meta-analysis conducted for this review although the reporting of adverse events data was incomplete (Table 7). Special care is needed when renal function is compromised. RR-TB and MDR-TB strains may also be resistant to ethambutol, particularly in those patients who have been treated with this drug previously. However, DST for this drug is not considered reliable and reproducible (38). The potential benefit that ethambutol may add to a core MDR-TB regimen needs to be balanced carefully with the inconvenience of adding another medicine to the regimen and the risks of associated harms.
Group D2 is made up of two new drugs released in recent years – bedaquiline and delamanid. WHO issued an interim policy on the use of these medicines in 2013 and 2014 (4,5). In October 2016, WHO published its revised policy on delamanid following the advice of a separate GDG which reviewed its use in children and adolescents (6) (see also above). At this point, bedaquiline remains only recommended for use in adults. When the results from ongoing studies and the Phase III trials become available the evidence for the effectiveness of these two new drugs will be re-evaluated with respect to the other medicines making up the MDR-TB regimen.
Group D3 consists of p-aminosalicylic acid (PAS), imipenem–cilastatin, meropenem, clavulanate and thioacetazone. These drugs are only to be used when a MDR-TB regimen with at least five effective drugs in the intensive phase (i.e. four core second-line medicines plus pyrazinamide) cannot be otherwise composed.
The aIPD (18), as well as the study-level meta-analysis conducted for the current guidelines revision, found no significant effect of PAS on treatment success. PAS use is associated with a high frequency of adverse reactions (12.2% SAEs in the meta-analysis undertaken for this study) and is thus reserved for situations when there is no option to use other drugs.
Carbapenems (imipenem–cilastin or meropenem) appear to be hydrolyzed more slowly by M. tuberculosis when combined with clavulanic acid (48,49). Amoxicillin-clavulanate has
32
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
shown poor results in in vitro studies and in early bactericidal activity (EBA) studies (50–52). The aIPD showed that patients treated with amoxicillin- clavulanate were more likely to have poor treatment outcomes, although this may be due to confounding by the higher likelihood that patients receiving this drug tended to have more severe disease (not all confounding could be adjusted for in the analysis). WHO recommends that whenever amoxicillin-clavulanate and carbapenems are included in regimens they are always to be used together. Clavulanate is only available as a combination preparation containing amoxicillin. The spectrum of adverse reactions associated with amoxicillin-clavulanate and carbapenems is to a large extent identical to that associated with penicillins (53).
Thioacetazone has been used extensively in the past as part of first-line combination therapy for TB, based on RCT evidence of effectiveness (54). Use of the drug in TB treatment has however been restricted since the early 1990s due to the severe skin reactions it causes (including Stevens–Johnson syndrome and toxic epidermal necrolysis that can lead to death, especially in people living with HIV (55)), as well as the widespread availability of safer, affordable alternatives for combination TB regimens. If thioacetazone is being considered as part of a MDR-TB treatment regimen, close monitoring for severe skin reactions is required and it is imperative that the patient be tested for HIV, and the drug not used if the patient is HIV seropositive.
M. tuberculosis is intrinsically resistant to the macrolide class of antibiotics (56). The evidence reviews for the current guidelines did not show any effectiveness of drugs of this class (clarithromycin, azithromycin) (57), which have at times been included in MDR-TB regimens in both adults and children. In addition, the aIPD showed an increased risk – although not statistically significant – for poor outcomes in patients receiving macrolides although macrolides appeared to be safe during prolonged use. Macrolides are associated with QT prolongation (58), which would be of particular concern if patients are receiving other TB drugs with a similar risk, such as moxifloxacin, clofazimine, bedaquiline or delamanid. WHO therefore no longer recommends the use of clarithromycin or azithromycin as part of regimens for the treatment of MDR/RR-TB.
Adverse reactions linked to PAS include gastrointestinal disturbance and hypothyroidism (in particular when given in combination with ethionamide/prothionamide). Hypothyroidism is reversible upon cessation of drugs. Although adverse events were poorly reported in studies assessed, the data for this review found that 12.2% of patients had SAEs attributed to PAS (Table 7). The pIPD showed the possibility of harm associated with the use of PAS (not statistically significant). However, PAS is frequently given to children when few other treatment options remain, and therefore this effect may be due to confounding by indication (sites that had poorer outcomes with PAS also had significantly higher rates of children who were HIV seropositive and malnourished, as well as with severe pulmonary disease, and additional resistance to fluoroquinolones and second-line injectable medicines).
Subgroup considerations
Rifampicin-resistant TB/MDR-TB with additional resistance to fluoroquinolones, second-line injectable agents and XDR-TB. In patients with RR-TB and MDR-TB, if there is confirmed or well-founded belief of resistance to medications from Group A (fluoroquinolones) or Group
33
Who policy recommenDations
B (second-line injectable agents), the medicines in the regimen that belong to these classes are substituted as detailed in the beginning of section B1. If any of the components of the regimen – the four core second-line medicines and pyrazinamide – is not considered to be effective, additional agents from Groups D2 or D3 are added. This is almost always necessary when resistance to both Groups A and B drugs (i.e. XDR-TB) is present. Analysis of additional individual patient data collected for the update of the WHO drug-resistant TB treatment guidelines of 2011 concluded that regimens containing more drugs were associated with the highest odds of success for MDR-TB patients who had additional resistance to fluoroquinolones and/or second-line injectable agents (59). The current WHO advice continues to apply when designing regimens for patients with resistance to fluoroquinolones and second-line injectable medications, as well as those with XDR-TB (9).
Access to rapid diagnostic testing, which could reliably identify resistance to Group A or Group B agents, would help clinicians decide on how to modify longer MDR-TB regimens. The Genotype MTBDRsl line probe assay (22) may now be used as an initial test, over phenotypic culture-based DST, to detect resistance to fluoroquinolones and to the second-line injectable drugs among patients with MDR/RR-TB (conditional recommendation; certainty of evidence low to moderate for direct testing (12)). Genotype MTBDRsl can be used in both children and adults and as a direct and indirect test (it could thus be used on extrapulmonary samples). While resistance-conferring mutations to fluoroquinolones detected by the MTBDRsl assay are highly correlated with phenotypic resistance to ofloxacin and levofloxacin, the correlation with moxifloxacin and gatifloxacin is less clear and the inclusion of moxifloxacin or gatifloxacin in a MDR-TB regimen is best guided by phenotypic DST results.
TB of the central nervous system. The treatment of tuberculous meningitis related to rifampicin-resistant or MDR strains is best guided by drug susceptibility results and the known properties of TB drugs to penetrate the central nervous system (CNS) (9). In patients with MDR/RR-TB meningitis, it is recommended that medications selected for the regimen have good CNS penetration properties.
The fluoroquinolones recommended by these guidelines have good CNS penetration (60), as do ethionamide (or prothionamide), cycloserine (or terizidone) and linezolid (61,62). Pyrazinamide has good CNS penetration, although caution should be exercised, as a large percentage of MDR-TB strains may be resistant. Isoniazid penetrates the CNS very well, with higher doses reaching adequate MICs in the cerebrospinal fluid. Due to its good CNS penetration, high-dose isoniazid is recommended as part of the treatment regimen unless high-level resistance is known to exist.
PAS and ethambutol do not penetrate the CNS well and should not be counted on as effective drugs to treat MDR-TB meningitis. Kanamycin, amikacin and streptomycin only penetrate the cerebrospinal fluid in the presence of meningeal inflammation. There are little data on the CNS penetration of capreomycin, clofazimine, bedaquiline or delamanid.
People living with HIV. The composition of the treatment regimen for MDR-TB does not differ for people living with HIV. However, thioacetazone should not be given to patients who
34
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
are HIV positive. If it is being considered as part of a treatment regimen, then HIV infection needs to be reliably excluded in the patient.
Implementation considerations
The implementation of MDR-TB chemotherapy is feasible under programmatic conditions, as has been amply shown by the global expansion in the use of MDR-TB regimens worldwide, particularly in the past decade (25,63). Changes made by the current revision to the grouping of medicines and composition of longer MDR-TB regimens are not expected to have a major impact on their continued use. Most of the fluoroquinolones and injectable agents are readily available as are the majority of the Group C and Group D agents. The latest WHO Model Lists of Essential Medicines (August 2015) include most of the agents in Table 6 except for gatifloxacin and thioacetazone (64,65). However, clofazimine, meropenem, imipenem–cilastatin and amoxicillin-clavulanate are listed for indications other than TB, while bedaquiline and delamanid are only included in the adult list. Other specific factors important for implementation are discussed in the respective sections below.
Where possible a patient’s rifampicin-resistant or MDR-TB strain needs to be tested for susceptibility to medicines planned for inclusion in the regimen. The availability of reliable, rapid tests for susceptibility to fluoroquinolones and second-line injectable drugs which would give results within a few days is valuable to ensure that longer MDR-TB regimens are strengthened as necessary (11,38). Where reliable DST is not an option, proof of the effectiveness of a medicine needs to be based on careful clinical history of the patient’s previous exposure to the medicine, of significant contact with another MDR/RR-TB patient whose antibiogramme is documented, and knowledge of the prevalent resistance patterns based on representative drug-resistance surveillance. Both the DST and the individual clinical history should be considered when constructing a treatment regimen. The only reliable laboratory tests for TB drug susceptibility (or resistance) which are widely used today are those for isoniazid, rifampicin, fluoroquinolones and second-line injectable agents.
The recommendations made by the current guidelines envisage a more widespread application of the shorter MDR-TB regimen among MDR/RR-TB patients. This implies that a larger proportion of the patients to whom longer MDR-TB regimens will be given would have additional resistance to core second-line medications than is the case today. For this reason additional care will need to be taken to ensure that regimens are adequately strengthened to ensure the best possible outcomes for these patients.
The current revision of the guidelines did not re-analyse the optimal duration of treatment (intensive and continuation phases). Thus the recommendations from the 2011 guidelines which were based on the aIPD meta-analysis continue to apply (1,18). The 2011 guidelines conditionally recommended an intensive phase of eight months for most MDR-TB patients and total treatment duration of 20 months in patients who had not been previously treated. The duration may need to be modified according to the patient’s response to therapy (9). The association between treatment success and the total length of treatment was less clear in patients who had been previously treated compared with those who had not, although the likelihood of treatment success appeared to peak between 27.6 and 30.5 months. The number
35
Who policy recommenDations
of observations was also far fewer than for those who had no previous MDR-TB treatment. As a result no recommendation on total duration was made in the 2011 revision for previously-treated patients. Many of the MDR/RR-TB patients who will be ineligible for the shorter MDR-TB regimen and referred for treatment with longer regimens would have been treated with second-line medication in the past. In these patients, uncertainties will remain on the optimal duration of treatment and therefore the duration of therapy would need to be guided primarily by their response to therapy.
Group A. Fluoroquinolones. Both levofloxacin and moxifloxacin are commonly used to treat MDR-TB. Levofloxacin is more widely available than moxifloxacin, which is more expensive although a reduction in its price is expected in the coming years.
Gatifloxacin is an affordable drug that was commonly used by TB treatment programmes until the concerns about its dysglycaemic effects led to a global shortage in its supplies. If manufacture of quality-assured formulations of the drug restarts, it could provide more options for regimen design and could lower the costs of regimens by substituting more expensive fluoroquinolones.
Moxifloxacin is relatively easy to administer to older children. However, the tablet must be split to accommodate dosing in younger children and it is highly unpalatable once split or crushed. Levofloxacin is available as a suspension.
Group B. Second-line injectable agents. These agents present problems to administer intramuscularly or intravenously on a daily basis for several months, often necessitating hospitalization. Giving injections to children and underweight adults is particularly unpleasant and unwelcome.
Group C. Other core second-line agents. Ethionamide and prothionamide are inexpensive, readily available worldwide and easily administered.
Cycloserine has been one of the standard inexpensive drugs for the treatment of MDR-TB for several years and therefore experience in its use is widespread. Terizidone is less widely used but is available on the Global Drug Facility (GDF) Products List.
Clofazimine is relatively inexpensive but it can be difficult to procure. The implementation of the recommendation on the shorter MDR-TB regimen, of which this medicine is an irreplaceable core component, needs to ensure that sufficient quantities of this medicine are available to meet the demand and that no stock-outs occur.
When linezolid is used, there needs to be close monitoring for adverse effects, particularly anaemia, thrombocytopenia, lactic acidosis, peripheral neuropathy and optic neuropathy, as these can be severe and life threatening. Historically linezolid has been very expensive, however, it has recently come off patent and the availability of generic products has hugely reduced its market price and it may become even more affordable in future.
Group D. Add-on agents. Pyrazinamide is inexpensive, readily available and easy to administer. Isoniazid is inexpensive. It is important to consider the epidemiology of high-level versus low-level isoniazid resistance in a population before standard treatment regimens including
36
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
high-dose isoniazid are recommended. Ethambutol is inexpensive and readily available. All of these three medicines are core components of first-line regimens for drug-susceptible TB.
PAS may be difficult to obtain although it is available through the GDF. Otherwise it is relatively inexpensive and easy to administer.
Amoxicillin-clavulanate is inexpensive and easily obtainable. However, the carbapenems are expensive and are difficult to administer as they must be given two or three times per day via an intravenous line.
Thioacetazone is inexpensive but it has limited availability and is not currently available through the GDF.
Monitoring and evaluation
Patients on longer MDR-TB treatment regimens need to be monitored for response to treatment and for safety using reasonable schedules of relevant clinical and laboratory testing (9,24). Frameworks for the surveillance of bacteriological status, drug-resistance and outcomes have been fairly standardized over the past decade. The systematic monitoring of adverse events during and after the end of treatment is a recent introduction in TB programmes and experience in their implementation is still developing in many countries. Its rationale is largely defined by frequent use of new and re-purposed medications in MDR-TB treatment regimens in the world, at times in combinations for which there has been very limited experience of use.
B2. Treatment regimens for isoniazid-resistant TB and M. bovisIn the review for isoniazid-resistant TB, no cohorts or RCTs were found which included fluoroquinolones as part of standardized combination TB regimens intended primarily for isoniazid-resistant TB. Fluoroquinolones, when used, were individualized and introduced at varying points in a patient’s regimen. These studies thus did not allow meaningful pooling. In three recent RCTs that investigated the potential for fluoroquinolones to shorten first-line TB regimens (37,66,67) over 240 patients with non-MDR, isoniazid-resistant strains were placed on fluoroquinolone-containing regimens. Data for 66 of these patients enrolled in one of these RCTs showed similar levels of unfavourable outcome (treatment failure/relapse/death/loss to follow-up) in patients on fluoroquinolone-containing four-month regimens (20.7%) compared with the standard 2HRZE/4HR10 regimen (21.6%) (37) (personal communication, Merle C). In a second trial, success rates in patients treated with four-month fluoroquinolone-containing regimens were similar in subgroups with isoniazid-resistant strains and those with fully susceptible strains (66) (personal communication, Gillespie SH). In conclusion, the evidence reviews of published studies on isoniazid-resistant TB could not address the PICO question.
Only eight studies identified by the literature search provided information on treatment and treatment outcomes of patients with confirmed M. bovis disease. Of these only three studies
10 2HRZE/4HR = two months of isoniazid, rifampicin, pyrazinamide and ethambutol followed with four months of rifampicin and isoniazid.
37
Who policy recommenDations
included 20 or more subjects – a minimum criterion for the review. In the three case series retained, treatment regimens were very different and tended to be individualized. It was thus impossible to group the different case series for pooled analysis.
Owing to the lack of data to address the questions directly, no clinically useful recommendations could be made for these two forms of TB.
C. The effect of delay in starting treatment on treatment outcomes for patients with drug-resistant TB
Global monitoring of the response to MDR/RR-TB shows that several countries have successfully expanded diagnostic services for RR-TB without matching it with complementary capacity to enrol patients on adequate treatment (63,69). This has led to patients with confirmed drug-resistant TB waiting for months or even years to initiate treatment. It is widely held, based largely upon findings from TB patients without drug-resistant disease, that prolonging the time to initiate treatment in TB patients is undesirable and predisposes to unfavourable clinical and public health consequences, such as increased disease progression with higher bacillary load in sputum, more lung damage and continued transmission. A PICO question was thus developed to inform any policy recommendation to be made in support of earlier start of treatment (see Annex 3; PICO 4). Evidence was reviewed to assess whether starting an adequate treatment regimen within four weeks of diagnosis, or a strong presumption of MDR/RR-TB, was associated with positive outcome, and to quantify any such effect.
An initial search of the literature yielded 1978 references of which 64 underwent full text review (70). None of these articles fulfilled the inclusion criteria. A supplementary full text review of the 64 references was undertaken with the explicit aim of determining whether any articles described treatment outcomes in MDR-TB patients stratified by delay to initiation of treatment. The original parameters were subsequently broadened from those in the PICO question to allow for the use of other time delay categorizations and to look for other relevant outcomes such as culture conversion. Sixteen articles were identified from which scant data could be abstracted. None of these articles addressed the independent effect of delay in start of treatment upon treatment outcomes with a meaningful comparator group.
A major obstacle to finding published evidence to support the assumption that shorter delays lead to better outcomes is the lack of studies reporting outcome in which treatment delay could be analysed as a dependent variable in groups which were otherwise comparable or in which other covariates could be adjusted for.
Differences in time to treatment initiation rarely occur in isolation. Programmatic changes related to delivery of care and modifications in drug regimen are common in the literature reviewed. Attribution of variations in delay to treatment outcomes is thus a significant challenge. Even if such data were available, an additional constraint is that the interval from RR-TB or MDR-TB diagnosis to start of treatment does not account for any delay in diagnosis, the magnitude of which may dominate overall delay and overshadow any benefits that could accrue from reducing the time to start treatment once the disease is diagnosed.
38
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Despite the absence of a discrete evidence base, it is reasonable to advise national programmes to adhere to the general standard of TB care which promotes an early start of appropriate therapy when MDR/RR-TB are diagnosed or strongly suspected (71). Studies to address this question are not a priority and intentionally withholding or delaying treatment presents ethical concerns. Nonetheless, this should not preclude from attempts to quantify the effect of delay using data from studies – observational or otherwise – mounted to answer other questions.
D. The effect of surgical interventions on treatment outcomes for patients with drug-resistant TB
Recommendation
In patients with RR-TB or MDR-TB, elective partial lung resection (lobectomy or wedge resection) may be used alongside a recommended MDR-TB regimen (conditional recommendation, very low certainty in the evidence).
Justification
Surgery has been employed in treating TB patients since before the advent of chemotherapy. In many countries it remains one of the treatment options for TB. With the challenging prospect in many settings of inadequate regimens to treat MDR-/XDR-TB, and the risk for serious sequelae, the role of pulmonary surgery is being re-evaluated as a means to reduce the amount of lung tissue with intractable pathology, to reduce bacterial load and thus improve prognosis.
The review for this question was based on both an individual patient-level meta-analysis to evaluate the effectiveness of different forms of elective surgery as an adjunct to combination medical therapy for MDR-TB (19), as well as a systematic review and study-level meta-analysis (72) (online Annex 4; Section 4). Demographic, clinical, bacteriological, surgical and outcome data of MDR-TB patients on treatment were obtained from the authors of 26 cohort studies participating in the aIPD (18). The analyses summarized in the GRADE tables consist of three strata comparing treatment success (cure and completion) with different combinations of treatment failure, relapse, death and loss to follow-up. Two sets of such tables were prepared for (i) partial pulmonary resection, and (ii) pneumonectomy.
In the study-level meta-analysis that examined all forms of surgery together, there was a statistically significant improvement in cure and successful treatment outcomes among patients who received surgery. However, when the aIPD meta-analysis examined patients who underwent partial lung resection and those who had a more radical pneumonectomy, versus patients who did not undergo surgery, those who underwent partial lung resection had statistically significantly higher rates of treatment success. Those patients who underwent pneumonectomy did not have better outcomes than those who did not undergo surgery. Prognosis appeared to be better when partial lung resection was performed after culture conversion. This effect was not observed in patients who underwent pneumonectomy.
There are several important caveats to these data. Substantial bias is likely to be present given that only patients judged to be fit for surgery would have been operated upon. No patient
39
Who policy recommenDations
with HIV co-infection in the aIPD underwent lung resection surgery. Therefore the effects of surgery among HIV-infected patients with MDR-TB could not be evaluated.
Rates of death did not differ significantly between those who underwent surgery versus those who received medical treatment only. However, the outcomes could be biased because the risk of death could have been much higher among patients in whom surgery was prescribed had they not been operated upon.
Subgroup consideration
The relative benefits of surgery are expected to depend substantially on the population subgroups that are targeted. The analysis could not provide a refined differentiation of the type of patient who would be best suited to benefit from the intervention or the type of intervention that would bear most benefit. The effect is expected to be moderate in the average patient considered appropriate for surgery.
The odds of success for patients with XDR-TB were statistically significantly lower when they underwent surgery compared with other patients (adjusted OR 0.4, 0.2–0.9). This effect is likely to be biased given that patients who underwent surgery would have had other factors predisposing to poor outcomes, which could not be adjusted for.
Implementation considerations
Partial lung resection for patients with MDR-TB is only to be considered under conditions of good surgical facilities, trained and experienced surgeons and with careful selection of candidates.
Monitoring and evaluation
The rates of death in the IPD for surgical outcomes did not differ significantly between patients who underwent surgery and those who received medical treatment only.
There were not enough data on adverse events, surgical complications or long term sequelae – some of which may be fatal – to allow a meaningful analysis.
Despite the unknown magnitude of perioperative complications the GDG assumed that overall there is a net benefit from surgery.
E. Research priorities
In addition to summarizing the available evidence, the reviews undertaken for this update revealed a number of gaps in current knowledge about critical areas for the treatment for MDR/RR-TB. Where evidence was available it was usually assigned a very low quality rating. This was one of the main reasons why all the recommendations made in this guidelines revision are conditional.
The GDG discussed research priorities and highlighted a number of them. They identified some problem areas which had already been singled out by earlier efforts to define research priorities for MDR-TB treatment, such as preventive therapy for MDR-TB and improving evidence on reduction of regimen duration (1,73,74).
40
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
The optimal combination of medicines and approach towards regimen-design for patients (both adults and children) with isoniazid-resistant TB, RR-TB, MDR-TB and XDR-TB, as well as for patients with M. bovis disease.
• More randomized controlled trials, especially involving the new drugs and regimens, but also for patients with isoniazid-resistant forms of TB who are placed on fluoroquinolone-containing regimens.
• Inclusion and separate reporting of outcomes for key subgroups, especially children and HIV-positive individuals on treatment, in randomized controlled studies.
• Complete recording of adverse events and standardized data recording on organ class, seriousness, severity, and certainty of association, to allow reliable comparison of the association between adverse events and exposure to different medicines.
• Identification of factors that determine the optimal duration of treatment (e.g. previous treatment history, baseline resistance patterns, site of disease, child/adult).
• Determination of the minimum number of drugs and treatment duration (especially in patients previously treated for MDR-TB).
• Determination of conditions under which injectable-sparing regimens can be used in both children and adults (e.g. surrogates for severity / extent of disease, alternative medication).
• Pharmacokinetic studies to determine optimal drug dosing and safety (especially in pregnancy).• Improved diagnostics and drug-susceptibility testing methods (e.g. which test for
pyrazinamide).• Randomized controlled trials to define the benefits and harms of chemoprophylaxis for
child and adult contacts of MDR/RR-TB (with and without additional resistance patterns) (9,75). The composition, dosages and duration of the latent TB infection (LTBI) regimen for MDR-TB needs to be optimized and the potential role of newer drugs with good sterilization properties investigated. Studies are needed to examine the adverse reactions of the long-term use of fluoroquinolones in preventive treatment.
• Palliative and end-of-life care in patients with very advanced resistance patterns.
The effectiveness and safety of standardized regimens lasting up to 12 months for the treatment of patients with MDR-TB (“shorter regimens”) when compared with longer treatment
• Future research needs to include the effectiveness/safety of the shorter MDR-TB treatment regimen in subgroups which have been systematically excluded from study protocols (e.g. children, patients with different forms of extrapulmonary TB) and in settings where background resistance to drugs other than fluoroquinolones and second-line injectable agents is high (e.g. pyrazinamide or high-level isoniazid resistance).
• Implementation research on the introduction of the shorter MDR-TB regimen.• More studies on cost effectiveness and health-related quality of life.
The effect of surgical interventions on treatment outcomes for patients with drug-resistant TB
• Better definition of the role of surgery (i.e. decisions about when to operate and the type of surgical intervention, drug-resistance patterns), needs to be better examined.
• Improved collection, reporting, standardization of data on surgery including long-term survival post surgery.
41
Who policy recommenDations
References1. Guidelines for the programmatic management of drug-resistant tuberculosis, 2011 Update. (WHO/
HTM/TB/2011.6) Geneva: World Health Organization; 2011 (http://whqlibdoc.who.int/publications/2011/9789241501583_eng.pdf, accessed 13 May 2016).
2. Resolution WHA67.1. Global strategy and targets for tuberculosis prevention, care and control after 2015. Geneva: World Health Organization. In 2014. Available from: http://apps.who.int/gb/ebwha/pdf_files/WHA67/A67_R1-en.pdf, accessed 19 July 2016).
3. Automated real-time nucleic acid amplification technology for rapid and simultaneous detection of tuberculosis and rifampicin resistance: Xpert MTB/RIF assay for the diagnosis of pulmonary and extrapulmonary TB in adults and children. Policy Update (WHO/HTM/TB/2013.16). Geneva: World Health Organization; 2013 (http://apps.who.int/iris/bitstream/10665/112472/1/9789241506335_eng.pdf, accessed 13 May 2016).
4. The use of bedaquiline in the treatment of multidrug-resistant tuberculosis. Interim policy guidance. (WHO/HTM/TB/2013.6). Geneva: World Health Organization; 2013. (http://apps.who.int/iris/bitstream/10665/84879/1/9789241505482_eng.pdf, accessed 13 May 2016).
5. The use of delamanid in the treatment of multidrug-resistant tuberculosis. Interim policy guidance (WHO/HTM/TB/2014.23). Geneva: World Health Organization; 2014 (http://apps.who.int/iris/bitstream/10665/137334/1/WHO_HTM_TB_2014.23_eng.pdf, accessed 13 May 2016).
6. The use of delamanid in the treatment of multidrug-resistant tuberculosis in children and adolescents: interim policy guidance (WHO/HTM/TB/2016.14). Geneva: World Health Organization; 2016 (http://apps.who.int/iris/bitstream/10665/250614/1/9789241549899-eng.pdf, accessed 30 October 2016).
7. WHO handbook for guideline development. 2nd ed. Geneva: World Health Organization; 2014 (http://www.who.int/kms/handbook_2nd_ed.pdf, accessed 13 May 2016).
8. Guyatt GH, Oxman AD, Kunz R, Falck-Ytter Y, Vist GE, Liberati A, et al. Rating quality of evidence and strength of recommendations: GRADE: Going from evidence to recommendations. BMJ 2008;336(7652):1049–51.
9. Companion handbook to the WHO guidelines for the programmatic management of drug-resistant tuberculosis. (WHO/HTM/TB/2014.11). Geneva: World Health Organization; 2014 (http://apps.who.int/iris/bitstream/10665/130918/1/9789241548809_eng.pdf, accessed 13 May 2016).
10. Definitions and reporting framework for tuberculosis – 2013 revision (WHO/HTM/TB/2013.2). Geneva: World Health Organization; 2013 (http://www.who.int/iris/bitstream/10665/79199/1/9789241505345_eng.pdf, accessed 13 May 2016).
11. The use of molecular line probe assays for the detection of resistance to second-line anti-tuberculosis drugs. Policy guidance. Geneva: World Health Organization; 2016 (http://www.who.int/tb/publications/lpa-mdr-diagnostics/en/, accessed 13 May 2016).
12. van Deun A, Maug AKJ, Salim MAH, Das PK, Sarker MR, Daru P, et al. Short, highly effective, and inexpensive standardized treatment of multidrug-resistant tuberculosis. Am J Respir Crit Care Med 2010;182(5):684–92.
13. Kuaban C, Noeske J, Rieder HL, Aït-Khaled N, Abena Foe JL, Trébucq A. High effectiveness of a 12-month regimen for MDR-TB patients in Cameroon. Int J Tuberc Lung Dis. 2015;19(5):517–24.
14. Piubello A, Harouna SH, Souleymane MB, Boukary I, Morou S, Daouda M, et al. High cure rate with standardised short-course multidrug-resistant tuberculosis treatment in Niger: no relapses. Int J Tuberc Lung Dis. 2014;18(10):1188–94.
15. Laserson KF, Thorpe LE, Leimane V, Weyer K, Mitnick CD, Riekstina V, et al. Speaking the same language: treatment outcome definitions for multidrug-resistant tuberculosis. Int J Tuberc Lung Dis. 2005;9(6):640–5.
42
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
16. National Institutes of Health (National Cancer Institute). Common Terminology Criteria for Adverse Events (CTCAE). Version 4.0 [website]. U.S. Department of Health and Human Services; 2009 (http://eversus.nci.nih.gov/ftp1/CTCAE/CTCAE_4.03_2010-06-14_QuickReference_5x7.pdf, accessed 13 May 2016).
17. Cochrane handbook for systematic reviews of interventions. 2011. (Version 5.1.0) [website] (http://handbook.cochrane.org/, accessed 13 May 2016).
18. Ahuja SD, Ashkin D, Avendano M, Banerjee R, Bauer M, Bayona JN, et al. Multidrug resistant pulmonary tuberculosis treatment regimens and patient outcomes: an individual patient data meta-analysis of 9,153 patients. PLoS Med. 2012;9(8):e1001300.
19. Fox GJ, Mitnick CD, Benedetti A, Chan ED, Becerra M, Chiang C-Y, et al. Surgery as an adjunctive treatment for multidrug-resistant tuberculosis: an individual patient data metaanalysis. Clin Infect Dis 2016;62(7):887–95.
20. Nunn AJ, Rusen ID, Van Deun A, Torrea G, Phillips PPJ, Chiang C-Y, et al. Evaluation of a standardised treatment regimen of anti-tuberculosis drugs for patients with multi-drug-resistant tuberculosis (STREAM). Trials. 2014; 15: 353. Published online 2014 September 9. doi: 10.1186/1745–6215–15–353 (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4164715/, accessed 13 May 2016).
21. Medecins Sans Frontieres – Field Research. Research Protocol – Effectiveness of a simplified short regimen for Multidrug Resistant Tuberculosis treatment in Karakalpakstan, Uzbekistan – MSF Field Research [website] (http://fieldresearch.msf.org/msf/handle/10144/322296, accessed 13 May 2016).
22. Rapid diagnostics of tuberculosis and its resistances. Nehren: Germany: Hain Lifescience GmbH; 2015 (http://www.hain-lifescience.de/uploadfiles/file/produkte/mikrobiologie/mykobakterien/tb_eng.pdf, accessed 13 May 2016).
23. World Health Organization. Active TB drug-safety monitoring and management (aDSM) [website] (http://www.who.int/tb/areas-of-work/drug-resistant-tb/treatment/pharmacovigilance/en/, accessed 13 May 2016).
24. Active tuberculosis drug-safety monitoring and management (aDSM). Framework for implementation. (WHO/HTM/TB/2015.28). Geneva: World Health Organization; 2015 (http://apps.who.int/iris/bitstream/10665/204465/1/WHO_HTM_TB_2015.28_eng.pdf, accessed 13 May 2016).
25. Global tuberculosis report 2015. (WHO/HTM/TB/2015.22). Geneva: World Health Organization; 2015 (http://apps.who.int/iris/bitstream/10665/191102/1/9789241565059_eng.pdf, accessed 13 May 2016).
26. Case definition for extensively drug-resistant tuberculosis. Wkly Epidemiol Rec. 2006;81(42):408.
27. Seddon JA, Hesseling AC, Godfrey-Faussett P, Schaaf HS. High treatment success in children treated for multidrug-resistant tuberculosis: an observational cohort study. Thorax. 2014;69(5):458–64.
28. Koh W-J, Kwon OJ, Gwak H, Chung JW, Cho S-N, Kim WS, et al. Daily 300 mg dose of linezolid for the treatment of intractable multidrug-resistant and extensively drug-resistant tuberculosis. J Antimicrob Chemother. 2009;64(2):388–91.
29. Schecter GF, Scott C, True L, Raftery A, Flood J, Mase S. Linezolid in the treatment of multidrug-resistant tuberculosis. Clin Infect Dis. 2010;50(1):49–55.
30. Udwadia ZF, Sen T, Moharil G. Assessment of linezolid efficacy and safety in MDR- and XDR-TB: an Indian perspective. Eur Respir J. 2010;35(4):936–938–940.
31. Singla R, Caminero JA, Jaiswal A, Singla N, Gupta S, Bali RK, et al. Linezolid: an effective, safe and cheap drug for patients failing multidrug-resistant tuberculosis treatment in India. Eur Respir J. 2012;39(4):956–62.
32. Lee M, Lee J, Carroll MW, Choi H, Min S, Song T, et al. Linezolid for Treatment of Chronic Extensively Drug-Resistant Tuberculosis. N Engl J Med. 2012;367(16):1508–18.
43
Who policy recommenDations
33. De Lorenzo S, Alffenaar JW, Sotgiu G, Centis R, D’Ambrosio L, Tiberi S, et al. Efficacy and safety of meropenem-clavulanate added to linezolid-containing regimens in the treatment of MDR-/XDR-TB. Eur Respir J. 2013;41(6):1386–92.
34. Jiang H, Tang R-N, Wang J. Linezolid versus vancomycin or teicoplanin for nosocomial pneumonia: meta-analysis of randomised controlled trials. Eur J Clin Microbiol Infect Dis. 2013;32(9):1121–8.
35. Padayatchi N, Mac Kenzie WR, Hirsch-Moverman Y, Feng P-J, Villarino E, Saukkonen J, et al. Lessons from a randomised clinical trial for multidrug-resistant tuberculosis. Int J Tuberc Lung Dis. 2012;16(12):1582–7.
36. Park-Wyllie LY, Juurlink DN, Kopp A, Shah BR, Stukel TA, Stumpo C, et al. Outpatient gatifloxacin therapy and dysglycemia in older adults. N Engl J Med. 2006;354(13):1352–61.
37. Merle CS, Fielding K, Sow OB, Gninafon M, Lo MB, Mthiyane T, et al. A four-month gatifloxacin-containing regimen for treating tuberculosis. N Engl J Med. 2014;371(17):1588–98.
38. Implementing tuberculosis diagnostics: A policy framework. (WHO/HTM/TB/2015.11). Geneva: World Health Organization. 2015 (http://apps.who.int/iris/bitstream/10665/162712/1/9789241508612_eng.pdf, accessed 13 May 2016).
39. Seddon JA, Thee S, Jacobs K, Ebrahim A, Hesseling AC, Schaaf HS. Hearing loss in children treated for multidrug-resistant tuberculosis. J Infect. 2013;66(4):320–9.
40. Tang S, Yao L, Hao X, Liu Y, Zeng L, Liu G, et al. Clofazimine for the treatment of multidrug-resistant tuberculosis: prospective, multicenter, randomized controlled study in China. Clin Infect Dis. 2015;60(9):1361–7.
41. Tang S, Yao L, Hao X, Zhang X, Liu G, Liu X, et al. Efficacy, safety and tolerability of linezolid for the treatment of XDR-TB: a study in China. Eur Respir J. 2015;45(1):161–70.
42. Hwang TJ, Wares DF, Jafarov A, Jakubowiak W, Nunn P, Keshavjee S. Safety of cycloserine and terizidone for the treatment of drug-resistant tuberculosis: a meta-analysis. Int J Tuberc Lung Dis. 2013;17(10):1257–66.
43. Hwang TJ, Dotsenko S, Jafarov A, Weyer K, Falzon D, Lunte K, et al. Safety and availability of clofazimine in the treatment of multidrug and extensively drug-resistant tuberculosis: analysis of published guidance and meta-analysis of cohort studies. BMJ Open. 2014;4(1):e004143.
44. Katiyar SK, Bihari S, Prakash S, Mamtani M, Kulkarni H. A randomised controlled trial of high-dose isoniazid adjuvant therapy for multidrug-resistant tuberculosis. Int J Tuberc Lung Dis. 2008;12(2):139–45.
45. van Toorn R, Schaaf HS, Laubscher JA, van Elsland SL, Donald PR, Schoeman JF. Short intensified treatment in children with drug-susceptible tuberculous meningitis. Pediatr Infect Dis J. 2014;33(3):248–52.
46. Niehaus AJ, Mlisana K, Gandhi NR, Mathema B, Brust JCM. High prevalence of inhA promoter mutations among patients with drug-resistant tuberculosis in KwaZulu-Natal, South Africa. PLoS One. 2015;10(9):e0135003.
47. Banerjee A, Dubnau E, Quemard A, Balasubramanian V, Um KS, Wilson T, et al. inhA, a gene encoding a target for isoniazid and ethionamide in Mycobacterium tuberculosis. Science. 1994;263(5144):227–30.
48. Hugonnet J-E, Blanchard JS. Irreversible inhibition of the Mycobacterium tuberculosis beta-lactamase by clavulanate. Biochemistry (Mosc). 2007;46(43):11998–2004.
49. Solapure S, Dinesh N, Shandil R, Ramachandran V, Sharma S, Bhattacharjee D, et al. In vitro and in vivo efficacy of β-lactams against replicating and slowly growing/nonreplicating Mycobacterium tuberculosis. Antimicrob Agents Chemother. 2013;57(6):2506–10.
50. Chambers HF, Kocagöz T, Sipit T, Turner J, Hopewell PC. Activity of amoxicillin/clavulanate in patients with tuberculosis. Clin Infect Dis. 1998;26(4):874–7.
51. Donald PR, Sirgel FA, Venter A, Parkin DP, Van de Wal BW, Barendse A, et al. Early bactericidal activity of amoxicillin in combination with clavulanic acid in patients with sputum smear-positive pulmonary tuberculosis. Scand J Infect Dis. 2001;33(6):466–9.
44
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
52. Gonzalo X, Drobniewski F. Is there a place for β-lactams in the treatment of multidrug-resistant/extensively drug-resistant tuberculosis? Synergy between meropenem and amoxicillin/clavulanate. J Antimicrob Chemother. 2013;68(2):366–9.
53. Prescott, Jr. WA, DePestel DD, Ellis JJ, Regal RE. Incidence of carbapenem-associated allergic-type reactions among patients with versus patients without a reported penicillin allergy. Clin Infect Dis. 2004;38(8):1102–7.
54. Isoniazid with thiacetazone (thioacetazone) in the treatment of pulmonary tuberculosis in East Africa–fifth investigation. A co-operative study in East African hospitals, clinics and laboratories with the collaboration of the East African and British Medical Research Councils. Tubercle. 1970;51(2):123–51.
55. Nunn P, Kibuga D, Gathua S, Brindle R, Imalingat A, Wasunna K, et al. Cutaneous hypersensitivity reactions due to thiacetazone in HIV-1 seropositive patients treated for tuberculosis. Lancet. 1991;337(8742):627–30.
56. Andini N, Nash KA. Intrinsic macrolide resistance of the Mycobacterium tuberculosis complex is inducible. Antimicrob Agents Chemother. 2006;50(7):2560–2.
57. Winters N, Butler-Laporte G, Menzies D. Efficacy and safety of World Health Organization group 5 drugs for multidrug-resistant tuberculosis treatment. Eur Respir J. 2015;46(5):1461–70.
58. Iannini PB. Cardiotoxicity of macrolides, ketolides and fluoroquinolones that prolong the QTc interval. Expert Opin Drug Saf. 2002;1(2):121–8.
59. Falzon D, Gandhi N, Migliori GB, Sotgiu G, Cox H, Holtz TH, et al. Resistance to fluoroquinolones and second-line injectable drugs: impact on MDR-TB outcomes. Eur Respir J. 2013;42(1):156–68.
60. Thwaites GE, Bhavnani SM, Chau TTH, Hammel JP, Torok ME, Van Wart SA, et al. Randomized pharmacokinetic and pharmacodynamic comparison of fluoroquinolones for tuberculous meningitis. Antimicrob Agents Chemother. 2011;55(7):3244–53.
61. Donald PR. The chemotherapy of tuberculous meningitis in children and adults. Tuberculosis. 2010;90(6):375–92.
62. Sun F, Ruan Q, Wang J, Chen S, Jin J, Shao L, et al. Linezolid manifests a rapid and dramatic therapeutic effect for patients with life-threatening tuberculous meningitis. Antimicrob Agents Chemother. 2014;58(10):6297–301.
63. Falzon D, Mirzayev F, Wares F, Baena IG, Zignol M, Linh N, et al. Multidrug-resistant tuberculosis around the world: what progress has been made? Eur Respir J. 2015;45(1):150–60.
64. WHO Model List of Essential Medicines. 19th list. Geneva: World Health Organization; 2015 (http://www.who.int/selection_medicines/committees/expert/20/EML_2015_FINAL_amended_AUG2015.pdf, accessed 13 May 2016).
65. WHO Model List of Essential Medicines for Children. 5th list. Geneva: World Health Organization; 2015 (http://www.who.int/medicines/publications/essentialmedicines/EMLc_2015_FINAL_amended_AUG2015.pdf, accessed 13 May 2016).
66. Gillespie SH, Crook AM, McHugh TD, Mendel CM, Meredith SK, Murray SR, et al. Four-month moxifloxacin-based regimens for drug-sensitive tuberculosis. N Engl J Med. 2014;371(17):1577–87.
67. Jawahar MS, Banurekha VV, Paramasivan CN, Rahman F, Ramachandran R, Venkatesan P, et al. Randomized clinical trial of thrice-weekly 4-month moxifloxacin or gatifloxacin containing regimens in the treatment of new sputum positive pulmonary tuberculosis patients. PloS One. 2013;8(7):e67030.
68. Global tuberculosis report 2014. (WHO/HTM/TB/2014.08). Geneva: World Health Organization. 2014. (http://www.who.int/tb/publications/global_report/gtbr14_main_text.pdf, accessed 13 May 2016).
69. Virenfeldt J, Rudolf F, Camara C, Furtado A, Gomes V, Aaby P, et al. Treatment delay affects clinical severity of tuberculosis: a longitudinal cohort study. BMJ Open. 2014 Jun 1;4(6):e004818.
45
Who policy recommenDations
70. Harris RC, Grandjean L, Martin LJ, Miller AJP, Nkang J-EN, Allen V, et al. The effect of early versus late treatment initiation after diagnosis on the outcomes of patients treated for multidrug-resistant tuberculosis: a systematic review. BMC Infectious Diseases 2016 Dec;16(1). (http://bmcinfectdis.biomedcentral.com/articles/10.1186/s12879-016-1524-0, accessed 19 July 2016).
71. TB CARE 1. International standards for tuberculosis care. 3rd ed. The Hague: The Netherlands; 2014 (http://www.who.int/tb/publications/ISTC_3rdEd.pdf, accessed 13 May 2016).
72. Harris RC, Khan MS, Martin LJ, Allen V, Moore DAJ, Fielding K, et al. The effect of surgery on the outcome of treatment for multidrug-resistant tuberculosis: a systematic review and meta-analysis. BMC Infectious Diseases. 2016 Dec;16(1). (http://bmcinfectdis.biomedcentral.com/articles/10.1186/s12879-016-1585-0, accessed 19 July 2016).
73. Priorities for tuberculosis research. A report of the Disease reference group on TB, leprosy and Buruli ulcer. Geneva: Tropical Disease Research; 2013 (http://apps.who.int/iris/bitstream/10665/ 85888/1/9789241505970_eng.pdf, accessed 13 May 2016).
74. Mitnick CD, Rodriguez CA, Hatton ML, Brigden G, Cobelens F, Grobusch MP, et al. Programmatic Management of Drug-Resistant Tuberculosis: An Updated Research Agenda. PLOS ONE. 2016 May 25;11(5):e0155968. (http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0155968, accessed 26 July 2016).
75. Guidelines on the management of latent tuberculosis infection. (WHO/HTM/TB/2015.01). Geneva: World Health Organization; 2014 (http://apps.who.int/iris/bitstream/10665/136471/1/9789241548908_eng.pdf, accessed 13 May 2016).
46
Annex 1. Agenda for the Guideline Development Group meeting on the WHO treatment guidelines for drug-resistant TB, 2016
update, 9–11 November 2015
Chair: Holger Schünemann | Co-chair: Charles L Daley
Day 18:30–9:00 Registration
9:00–9:15 Welcome and introductions Karin Weyer
9:15–9:30 Meeting objectives and expected outcomes, agenda and working methods
Declarations of interest
Ernesto Jaramillo
Dennis Falzon
9:30–10:00 WHO requirements for evidence-based guidelines, GRADE methodology
Holger Schünemann
10:00–10:45 Plenary – Presentation of draft GRADE tables
PICO 1: MDR-TB REGIMEN COMPOSITION – SYSTEMATIC REVIEWS OF INDIVIDUAL DRUGS
Dick Menzies, Mayara Bastos
10:45–11:00 Coffee break
11:00–11:30 Plenary – Presentation of draft GRADE tables
PICO 1: MDR-TB REGIMEN COMPOSITION – PAEDIATRIC IPD
Anneke Hesseling
11:30–11:40 Plenary – Discussants present their perspectives on the implications of the findings for the approach to the composition and duration of MDR-TB regimens in adults and children.
Discussants: Charles L Daley (adults), Farhana Amanullah (children)
11:40–13:00 Plenary – Development of decision tables to formulate draft recommendations based on quality of the evidence, and other considerations (balance between desirable and undesir-able effects, resources, feasibility, values and preferences).
Facilitated discussion
13:00–14:00 Lunch break
14:00–15:30 Continued – Development of decision tables to formulate draft recommendations based on quality of the evidence, and other considerations (balance between desirable and undesirable effects, resources, feasibility, values and preferences).
Facilitated discussion
15:30–15:45 Coffee break
15:45–17:45 Continued – Finalization of draft recommendations Facilitated discussion
17:45–18:00 Summary of the day Chair
47
annex 1
Day 28:30–9:15 Plenary – Presentation of draft GRADE tables
PICO 2: REGIMENS FOR ISONIAZID RESISTANCE and M. bovis
Dick Menzies, Mayara Bastos
9:15–9:30 Plenary – Discussants present their perspectives on the implications of the findings for the approach to the composition and duration of regimens in adults and children.
Discussants:
Daniela Cirillo; Carlos Torres (isoniazid-resistant); Jose Caminero; Agnes Gebhard (M.bovis)
9:30–10:45 Plenary – Development of decision tables to formulate draft recommendations based on quality of the evidence, and other considerations (balance between desirable and undesirable effects, resources, feasibility, values and preferences).
Facilitated discussion
10:45–11:00 Coffee break
11:00–13:00 Continued – Finalization of draft recommendations Facilitated discussion
13:00–14:00 Lunch break
14:00–14:45 Plenary – Presentation of GRADE tables
PICO 3: SHORTER REGIMENS FOR MDR-TB
Dick Menzies,
Faiz A Khan
14:45–15:00 Plenary – Discussants present their perspectives on the implications of the findings for the treatment of MDR-TB using shorter regimens.
Discussants:
Sundari Mase, Tsira Chakhaia, Michel Gasana
15:00–16:00 Plenary – Development of decision tables to formulate draft recommendations based on quality of the evidence, and other considerations (balance between desirable and undesirable effects, resources, feasibility, values and preferences).
Facilitated discussion
16:00–16:15 Coffee break
16:15–17:00 Continued – Finalization of draft recommendations Facilitated discussion
17:00–17:45 Implications of the findings from reviews of PICO 1 and PICO 3 for the approach to the composition and duration of MDR-TB regimens.
Facilitated discussion
17:45–18:00 Wrap-up and summary of the day Chair
48
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Day 38:30–9:30 Plenary – Presentation of draft GRADE tables
PICO 4: DELAYS IN STARTING MDR TREATMENT, THE ROLE OF SURGERY
Mishal Khan,
Rebecca Harris,
Greg Fox
9:30–9:40 Plenary – Discussant presents perspectives on the implications of the findings for the approach to the management of MDR-TB.
Discussant :
Armen Hayrapetyan (role of surgery)
9:40–10:45 Plenary – Development of decision tables to formulate draft recommendations based on quality of the evidence, and other considerations (balance between desirable and undesirable effects, resources, feasibility, values and preferences).
Facilitated discussion
10:45–11:00 Coffee break
11.00–11.30 Levels of resistance to pyrazinamide and fluoroquinolones Matteo Zignol
11:30–13:00 Review of the recommendations for the four PICOs combined (continued)
Facilitated discussion
13:00–14:00 Lunch break
14:00–15:00 Research priorities on treatment of drug-resistant TB Dick Menzies
Christian Lienhardt
15:00–15:30 Next steps and closure Chair &
Karin Weyer
49
Annex 2. Experts involved in the development of the WHO treatment guidelines for drug-
resistant TB, 2016 update
GUIDELINE DEVELOPMENT GROUP MEMBERS
Dr Farhana AmanullahAssociate DirectorPediatric TB ProgramIndus HospitalKarachiPakistan
Dr Tsira ChakhaiaACSM Advisor, Civil Society University Research Co., LLCTbilisiGeorgia
Dr Daniela Maria CirilloHeadEmerging Bacterial Pathogens UnitSan Raffaele del Monte Tabor Foundation (hSR)San Raffaele Scientific InstituteMilanoItaly
Dr Charles L Daley (Co-Chair)ChiefDivision of Mycobacterial and Respiratory InfectionsNational Jewish HealthDenver, COUSA
Dr Luis Gustavo do Valle BastosSenior Technical AdvisorCenter for Pharmaceutical ManagementManagement Sciences for HealthArlington, VAUSA
Dr Kelly DooleyAssociate Professor of MedicinePharmacology & Molecular Science Divisions of Clinical Pharmacology & Infectious DiseasesJohns Hopkins University School of MedicineCenter for Tuberculosis ResearchBaltimore, MDUSA
Dr Carlos A Torres DuqueDirectorTuberculosis DepartmentLatin American Thoracic SocietyBogotáColombia
Dr Michel GasanaManagerNational TB & Other Respiratory Communicable Diseases DivisionMinistry of HealthKigaliRwanda
Dr Agnes GebhardSenior ConsultantTeam Leader of ACCESS TeamTechnical DivisionKNCV Tuberculosis FoundationThe HagueNetherlands
Dr Armen HayrapetyanDirectorNational TB Control CentreMinistry of HealthAbovyan cityArmenia
50
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Dr Antonia KwiecienSenior Technical AdvisorSystems for Improved Access toPharmaceuticals and Services Program (SIAPS)Management Sciences for Health (MSH)Arlington, VAUSA
Dr José A Caminero LunaCoordinatorMDR-TB Unit, International Union Against Tuberculosis & Lung Disease (UNION)General Hospital of Gran Canaria “Dr. Negrin”Las Palmas de Gran CanariaSpain
Dr Sundari MaseMedical Team LeadField Services and Evaluation Branch Division of Tuberculosis Elimination National Center for HIV, Hepatitis, STD and TB PreventionCenters for Disease Control and PreventionAtlanta, GAUSA
Dr Lindsay McKennaTB/HIV Project OfficerTreatment Action GroupNew York, NYUSA
Dr Nguyen Viet NhungDirectorNational Tuberculosis Control ProgrammeHanoiViet Nam
Dr Maria RodriguezCoordinatorMDR-TB National Technical UnitMinistry of HealthSanto DomingoDominican Republic
Dr Holger Schünemann (Chair)Chair and ProfessorDepartments of Clinical Epidemiology & Biostatistics and of MedicineMcMaster UniversityHamilton, ONCanada
Dr James SeddonClinical LecturerDepartment of PaediatricsImperial CollegeLondonUnited Kingdom
Dr Thomas ShinnickAssociate DirectorGlobal Laboratory Activities Mycobacteriology Laboratory Branch, Division of Tuberculosis Elimination Centers for Disease Control and PreventionAtlanta, GAUSA
Dr Alena SkrahinaScientific DirectorRepublican Scientific & Practical Centre for Pulmonology & TuberculosisBelarus Research Institute of Pulmonology and TuberculosisMinskBelarus
51
OBSERVERS (at the GDG Meeting in Geneva in November 2015)
Dr J Peter CegielskiTeam LeaderMDR TB International Programs and Research Branch, Division of Tuberculosis EliminationCenters for Disease Control & Prevention (CDC)Atlanta, GAUSA
Mrs Janet Kristen GinnardTechnical Officer Strategy & ResultsUNITAIDGenevaSwitzerland
Dr Giovanni Battista MiglioriDirectorWHO Collaborating Centre for Tuberculosis and Lung DiseasesFondazione Salvatore MaugeriTradate, VAItaly
Dr Payam NahidProfessor of MedicineUniversity of California, San FranciscoSan Francisco General Hospital Division of Pulmonary and Critical Care MedicineSan Francisco, CAUSA
Dr Nobuyuki NishikioriRegional Adviser, TBWHO Western Pacific Regional OfficeManilaThe Philippines
Dr Thomas W PiggottResident PhysicianMcMaster UniversityHamilton ONCanada
Dr Anna ScardigliDisease Advisor, TuberculosisTechnical Advice and PartnershipThe Global Fund to Fight AIDS, Tuberculosis and MalariaGenevaSwitzerland
Dr Barbara SeaworthProfessor of MedicineDirector Heartland National TB CenterUniversity of Texas Health Science Center, TylerSan Antonio, TXUSA
Dr Mohammed YassinTechnical Advisor, TuberculosisThe Global Fund to Fight AIDS, Tuberculosis and MalariaGenevaSwitzerland
Dr Ya Diul MukadiSenior TB Technical AdvisorInfectious Disease Division, Global Health BureauUS Agency for International Development (USAID)Washington, DCUSA
RESOURCE PERSONS (at the GDG Meeting in Geneva in November 2015)
Dr Philipp Du CrosTB AdvisorMédecins sans Frontières (MSF)LondonUnited Kingdom
Dr Michael L RichMedical OfficerPartners in HealthHarvard Medical SchoolBoston, MAUSA
annex 2
52
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Dr Abdul Hamid SalimAdvisor, NTP BangladeshTB Gate, Leprosy hospital Compound MohakhaliDhakaBangladesh
Dr Valérie SchwoebelMedical OfficerInternational Union Against Tuberculosis & Lung Disease (UNION)ParisFrance
Dr Francis VaraineInternational Medical CoordinatorMédecins sans Frontières (MSF)ParisFrance
Dr Askar B. YedilbayevMedical OfficerPartners in HealthBoston, MAUSA
EXTERNAL REVIEW GROUP
Dr Chen-Yuan ChiangInternational Union Against Tuberculosis and Lung Disease (UNION)ParisFrance
Dr Celine GarfinInfectious Diseases for Prevention and Control DivisionDisease Prevention and Control BureauDepartment of HealthManilaThe Philippines
Dr Michael KimerlingKNCV Tuberculosis FoundationThe HagueNetherlands
Dr Vaira LeimaneNational TB ProgrammeMinistry of HealthRigaLatvia
Dr Guy MarksInternational Union Against Tuberculosis and Lung Disease (UNION)ParisFrance
Dr Gao MengqiuBeijing Chest HospitalCapital Medical University, Beijing Tuberculosis and Thoracic Tumor Research InstituteBeijingChina
Dr Norbert NdjekaDrug Resistant TB, TB & HIVDepartment of HealthPretoriaSouth Africa
Dr Ejaz QadeerNational TB ProgrammeMinistry of HealthIslamabadPakistan
Dr Lee ReichmanRutgers UniversityNew Jersey, NJUSA
Dr Rohit SarinLRS Institute of TB and Respiratory DiseasesDelhiIndia
Dr Irina VasilyevaCentral Tuberculosis Research Institute (CTRI)MoscowRussian Federation
Dr Dalene von DelftTB ProofSouth Africa
53
EVIDENCE REVIEWERS (at the GDG Meeting in Geneva in November 2015)
Dr Mayara Lisboa Soares de BastosMcGill UniversityMontreal, QcCanada
Dr Gregory J Fox* University of SydneySpit Junction, NSWAustralia
Ms Rebecca HarrisLondon School of Hygiene and Tropical MedicineLondonUnited Kingdom
Dr Anneke HesselingPaediatric TB Research ProgrammeDesmond Tutu TB Centre Department of Paediatrics and Child HealthFaculty of Medicine and Health SciencesStellenbosch UniversityCape TownSouth Africa
Dr Faiz KhanFaculty of Medicine, McGill UniversityMontreal Chest InstituteMcGill University Health CentreMontreal, QcCanada
Dr Mishal KhanLondon School of Hygiene and Tropical MedicineLondonUnited Kingdom
Dr Dick MenziesMontreal Chest InstituteMcGill University Health CentreMontreal, QcCanada
WHO GUIDELINE STEERING COMMITTEE
Dr Dennis Falzon, LDR/GTBDr Nathan Ford, TAC/HIVDr Giuliano Gargioni, TSC/GTBDr Haileyesus Getahun, THC/GTBDr Malgorzata Grzemska, TSC/GTBDr Ernesto Jaramillo, LDR/GTBDr Avinash Kanchar, THC/GTBMs Soleil Labelle, TSC/GTBDr Christian Lienhardt, PSI/GTB
Dr Knut Lönnroth, PSI/GTBDr Alberto Matteelli, THC/GTBDr Fuad Mirzayev, LDR/GTBDr Linh Nhat Nguyen, LDR/GTBDr Marco Antonio Vitoria, TAC/HIVDr Fraser Wares, LDR/GTBMrs Diana Weil, PSI/GTBDr Karin Weyer, LDR/GTBDr Matteo Zignol, TME/GTB
WHO CONSULTANT
Dr Elizabeth Harausz
annex 2
* Affiliated with McGill University for the evidence reviews done for these guidelines
54
Ann
ex 3
. P
ICO
que
stio
nsa
Q1
. In
patie
nts
with
RR
-TB
or
MD
R-T
B, w
hat
indi
vidu
al d
rugs
in the
reg
imen
s ar
e lik
ely
to le
ad to
the
out
com
es
liste
d be
low
?
PO
PU
LATI
ON
INTE
RVEN
TIO
N
CO
MPA
RAT
OR
OU
TCO
MES
Rifa
mpi
cin-
resi
stan
t TB
onl
y
MD
R-T
B w
ithou
t re
sist
ance
or
seve
re
into
lera
nce
to t
he s
econ
d-lin
e dr
ugs
MD
R-T
B w
ith r
esis
tanc
e or
sev
ere
into
lera
nce
to a. F
luor
oqui
nolo
nes
b. S
econ
d-lin
e in
ject
able
dru
gs, b
oth
clas
ses
c. P
yraz
inam
ide
d. 2
or
3 G
roup
4 d
rugs
b
e. F
luor
oqui
nolo
nes
+ s
econ
d-lin
e in
ject
able
s (i.
e. X
DR
-TB
+/-
oth
er r
esis
tanc
e)
Chi
ldre
n (0
–4, 5
–14 y
ears
), pe
rson
s w
ith H
IV,
preg
nant
wom
en, p
erso
ns w
ith d
iabe
tes
A se
cond
-line
reg
imen
c whi
ch IN
CLU
DES
:• C
ured
/com
plet
ed b
y en
d of
tre
atm
ent
• C
ultu
re c
onve
rsio
n by
si
x m
onth
s
• Fa
ilure
• R
elap
se
• S
urvi
val (
or d
eath
)
• Ad
vers
e re
actio
ns
(sev
erity
, typ
e, o
rgan
cl
ass)
–pyr
azin
amid
e –n
o py
razi
nam
ide
–inj
ecta
ble
agen
ts (ka
nam
ycin
/am
ikac
in/c
apre
omyc
in)
–no
inje
ctab
le a
gent
s (k
anam
ycin
/am
ikac
in/c
apre
omyc
in)
–pro
thio
nam
ide/
ethi
onam
ide
–no
prot
hion
amid
e/et
hion
amid
e
–cyc
lose
rine
or t
eriz
idon
e –n
o cy
clos
erin
e or
ter
izid
one
–PAS
–no
PAS
–lat
er-g
ener
atio
n flu
oroq
uino
lone
d –n
o la
ter-g
ener
atio
n flu
oroq
uino
lone
d
–hig
h-do
se is
onia
zid
–no
high
-dos
e is
onia
zid
–clo
fazi
min
e –n
o cl
ofaz
imin
e
–lin
ezol
id –n
o lin
ezol
id
–bed
aqui
line
–no
beda
quili
ne
–del
aman
id –n
o de
lam
anid
–oth
er in
divi
dual
Gro
up 5
b dr
ugs
–no
othe
r in
divi
dual
Gro
up 5
b dr
ugs
a P
ICO
= P
opul
atio
n, In
terv
entio
n, C
ompa
rato
r, O
utco
mes
b G
roup
4 d
rugs
ref
er t
o: e
thio
nam
ide,
pro
thio
nam
ide,
cyc
lose
rine,
ter
izid
one,
PAS
. G
roup
5 d
rugs
ref
er t
o: a
mox
icill
in-c
lavu
lana
te, b
edaq
uilin
e, c
larit
hrom
ycin
, clo
fazi
min
e,
dela
man
id, h
igh-
dose
ison
iazi
d, im
ipen
em–c
ilast
atin
, lin
ezol
id, m
erop
enem
, thi
oace
tazo
ne.
For
beda
quili
ne a
nd d
elam
anid
, defi
nitiv
e re
com
men
datio
ns o
n th
eir
role
in
trea
tmen
t w
ill o
nly
be p
ossi
ble
once
the
res
ults
of
the
Phas
e III
tria
ls b
ecom
e av
aila
ble.
c D
ata
from
reg
imen
s la
stin
g up
to
12 m
onth
s w
ill n
ot b
e in
clud
ed in
thi
s qu
estio
n bu
t in
Que
stio
n 3.
d M
oxifl
oxac
in o
r ga
tiflox
acin
; hi
gh-d
ose
levo
floxa
cin
may
be
incl
uded
but
res
ults
to
be m
ade
avai
labl
e se
para
tely.
55
Ann
ex 3
. P
ICO
que
stio
nsa
Q1
. In
patie
nts
with
RR
-TB
or
MD
R-T
B, w
hat
indi
vidu
al d
rugs
in the
reg
imen
s ar
e lik
ely
to le
ad to
the
out
com
es
liste
d be
low
?
PO
PU
LATI
ON
INTE
RVEN
TIO
N
CO
MPA
RAT
OR
OU
TCO
MES
Rifa
mpi
cin-
resi
stan
t TB
onl
y
MD
R-T
B w
ithou
t re
sist
ance
or
seve
re
into
lera
nce
to t
he s
econ
d-lin
e dr
ugs
MD
R-T
B w
ith r
esis
tanc
e or
sev
ere
into
lera
nce
to a. F
luor
oqui
nolo
nes
b. S
econ
d-lin
e in
ject
able
dru
gs, b
oth
clas
ses
c. P
yraz
inam
ide
d. 2
or
3 G
roup
4 d
rugs
b
e. F
luor
oqui
nolo
nes
+ s
econ
d-lin
e in
ject
able
s (i.
e. X
DR
-TB
+/-
oth
er r
esis
tanc
e)
Chi
ldre
n (0
–4, 5
–14 y
ears
), pe
rson
s w
ith H
IV,
preg
nant
wom
en, p
erso
ns w
ith d
iabe
tes
A se
cond
-line
reg
imen
c whi
ch IN
CLU
DES
:• C
ured
/com
plet
ed b
y en
d of
tre
atm
ent
• C
ultu
re c
onve
rsio
n by
si
x m
onth
s
• Fa
ilure
• R
elap
se
• S
urvi
val (
or d
eath
)
• Ad
vers
e re
actio
ns
(sev
erity
, typ
e, o
rgan
cl
ass)
–pyr
azin
amid
e –n
o py
razi
nam
ide
–inj
ecta
ble
agen
ts (ka
nam
ycin
/am
ikac
in/c
apre
omyc
in)
–no
inje
ctab
le a
gent
s (k
anam
ycin
/am
ikac
in/c
apre
omyc
in)
–pro
thio
nam
ide/
ethi
onam
ide
–no
prot
hion
amid
e/et
hion
amid
e
–cyc
lose
rine
or t
eriz
idon
e –n
o cy
clos
erin
e or
ter
izid
one
–PAS
–no
PAS
–lat
er-g
ener
atio
n flu
oroq
uino
lone
d –n
o la
ter-g
ener
atio
n flu
oroq
uino
lone
d
–hig
h-do
se is
onia
zid
–no
high
-dos
e is
onia
zid
–clo
fazi
min
e –n
o cl
ofaz
imin
e
–lin
ezol
id –n
o lin
ezol
id
–bed
aqui
line
–no
beda
quili
ne
–del
aman
id –n
o de
lam
anid
–oth
er in
divi
dual
Gro
up 5
b dr
ugs
–no
othe
r in
divi
dual
Gro
up 5
b dr
ugs
a P
ICO
= P
opul
atio
n, In
terv
entio
n, C
ompa
rato
r, O
utco
mes
b G
roup
4 d
rugs
ref
er t
o: e
thio
nam
ide,
pro
thio
nam
ide,
cyc
lose
rine,
ter
izid
one,
PAS
. G
roup
5 d
rugs
ref
er t
o: a
mox
icill
in-c
lavu
lana
te, b
edaq
uilin
e, c
larit
hrom
ycin
, clo
fazi
min
e,
dela
man
id, h
igh-
dose
ison
iazi
d, im
ipen
em–c
ilast
atin
, lin
ezol
id, m
erop
enem
, thi
oace
tazo
ne.
For
beda
quili
ne a
nd d
elam
anid
, defi
nitiv
e re
com
men
datio
ns o
n th
eir
role
in
trea
tmen
t w
ill o
nly
be p
ossi
ble
once
the
res
ults
of
the
Phas
e III
tria
ls b
ecom
e av
aila
ble.
c D
ata
from
reg
imen
s la
stin
g up
to
12 m
onth
s w
ill n
ot b
e in
clud
ed in
thi
s qu
estio
n bu
t in
Que
stio
n 3.
d M
oxifl
oxac
in o
r ga
tiflox
acin
; hi
gh-d
ose
levo
floxa
cin
may
be
incl
uded
but
res
ults
to
be m
ade
avai
labl
e se
para
tely.
Q2
. In
TB p
atie
nts
with
dru
g-re
sist
ance
pat
tern
s ot
her
than
with
rifa
mpi
cin
resi
sita
nce
or m
ultid
rug
resi
stan
ce,
wha
t dr
ug r
egim
en c
ompo
sitio
n an
d du
ratio
n is
like
ly to
lead
to the
out
com
es li
sted
bel
ow?
PO
PU
LATI
ON
aIN
TER
VEN
TIO
Nb
CO
MPA
RAT
OR
OU
TCO
MES
Res
ista
nce
to
ison
iazi
dc
RZE
+ F
Qd fo
r 6–9
mon
ths
–HR
ZE for
6–9
mon
ths
–RZE
for
6–9
mon
ths
–8-m
onth
firs
t-lin
e dr
ug r
etre
atm
ent
regi
men
(“C
ateg
ory
2”)
–6-m
onth
firs
t-lin
e dr
ug t
reat
men
t re
gim
en
(“C
ateg
ory
1 o
r 3”)
–oth
er
• C
ured
/com
plet
ed b
y en
d of
tr
eatm
ent
• C
ultu
re c
onve
rsio
n by
6 m
onth
s
• Fa
ilure
• R
elap
se
• S
urvi
val (
or d
eath
)
• Ad
vers
e re
actio
ns (se
verit
y,
type
, org
an c
lass
)
• Ac
quis
ition
(am
plifi
catio
n) o
f ad
ditio
nal d
rug
resi
stan
ce
RZ
+ F
Qd fo
r 9–1
2 m
onth
s –8
-mon
th fi
rst-l
ine
drug
ret
reat
men
t re
gim
en (“C
ateg
ory
2”)
–6-m
onth
firs
t-lin
e dr
ug in
itial
tre
atm
ent
regi
men
(“
Cat
egor
y 1 o
r 3”)
–oth
er
Sec
ond-
line
inje
ctab
le/R
/Eto
/FQ
d fo
r 3 m
onth
s + R
/Eto
/FQ
d fo
r 15 m
onth
s
–8-m
onth
firs
t-lin
e dr
ug r
etre
atm
ent
regi
men
(“C
ateg
ory
2”)
–6-m
onth
firs
t-lin
e dr
ug in
itial
tre
atm
ent
regi
men
(“
Cat
egor
y 1 o
r 3”)
–oth
er
M. b
ovis
6- o
r 8-m
onth
firs
t-lin
e dr
ug t
reat
men
t re
gim
ens
(“C
ateg
ory
1, 2
or
3”)
–oth
er
–2H
RE/
7H
R
a If
dat
a ar
e av
aila
ble
the
effe
cts
will
be
stra
tified
by
key
subp
opul
atio
ns:
child
ren
(0–4
, 5–1
4 y
ears
), pe
rson
s w
ith H
IV, p
regn
ant
wom
en, a
nd p
eopl
e w
ith d
iabe
tes.
b T
he t
reat
men
t m
odal
ities
des
crib
ed in
the
MD
R-T
B t
reat
men
t ha
ndbo
ok (
8).
c In
clud
ing
case
s w
ith o
r w
ithou
t ad
ditio
nal r
esis
tanc
e to
eth
ambu
tol a
nd p
yraz
inam
ide.
The
ass
essm
ent
of e
vide
nce
and
reco
mm
enda
tions
on
thes
e re
sist
ance
pat
tern
s w
ill b
e co
nditi
oned
by
the
fact
tha
t D
ST
for
etha
mbu
tol a
nd p
yraz
inam
ide
is o
ften
unr
elia
ble
(alth
ough
rel
iabl
e D
ST
to p
yraz
inam
ide
may
be
avai
labl
e in
som
e se
ttin
gs),
but
whi
ch n
onet
hele
ss m
ay b
ring
abou
t th
e us
e of
fluo
roqu
inol
ones
, and
tha
t th
e ev
iden
ce m
ay b
e ba
sed
on e
xper
ienc
e fr
om p
atie
nts
who
may
hav
e be
en s
witc
hed
to
fluor
oqui
nolo
ne-c
onta
inin
g re
gim
ens
afte
r a
perio
d of
tim
e on
firs
t-lin
e re
gim
ens.
d If
dat
a ar
e av
aila
ble
the
effe
cts
will
be
stra
tified
by
the
type
of
fluor
oqui
nolo
ne (
early
ver
sus
late
r ge
nera
tion)
.
R =
rifa
mpi
cin
Z = p
yraz
inam
ide
E = e
tham
buto
l; FQ
= fl
uoro
quin
olon
e; E
to =
eth
iona
mid
e; H
= is
onia
zid.
annex 3
56
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATEQ
3. I
n M
DR
-TB
pat
ient
s, a
re tre
atm
ent
regi
men
s la
stin
g up
to 1
2 m
onth
s lik
ely
to le
ad to
the
out
com
es li
sted
be
low
whe
n co
mpa
red
with
tho
se r
ecom
men
ded
in the
WH
O g
uide
lines
of 201
1?
PO
PU
LATI
ON
INTE
RVEN
TIO
Na
CO
MPA
RAT
OR
aO
UTC
OM
ES
MD
R-T
B p
atie
nts
Prev
ious
ly t
reat
ed w
ith s
econ
d-lin
e dr
ugs
or n
ot
Sev
erity
of di
seas
e (m
ild/e
xten
sive
ra
diog
raph
ic le
sion
s)
Dru
g re
sist
ance
pat
tern
s (for
firs
t-lin
e dr
ugs
and
seco
nd-li
ne d
rugs
)
His
tory
of pa
tient
use
of et
ham
buto
l or
pyra
zina
mid
e
Chi
ldre
n (0
–14 y
ears
) vs
adu
lts
Pers
ons
with
HIV
, pre
gnan
t w
omen
, and
peo
ple
with
dia
bete
s
Dur
atio
n of
9–1
2 m
onth
s
Inje
ctab
le a
gent
for
4–6
m
onth
s
Com
bina
tion
of d
rugs
(u
sual
ly 7
in t
he in
tens
ive
phas
e an
d 4–5
in t
he
cont
inua
tion)
Use
of at
leas
t 4 e
ffec
tive
seco
nd-
line
drug
s pl
us p
yraz
inam
ide
Inje
ctab
le a
gent
giv
en for
abo
ut
8 m
onth
s, a
t le
ast
4 m
onth
s af
ter
cultu
re c
onve
rsio
n
Tota
l tre
atm
ent
for
at le
ast
18
mon
ths
past
the
dat
e of
cul
ture
co
nver
sion
to
nega
tive
Inje
ctab
le a
gent
giv
en u
ntil
smea
r co
nver
sion
and
tot
al t
reat
men
t fo
r at
leas
t 12 m
onth
s af
ter
smea
r co
nver
sion
• C
ured
/com
plet
ed b
y en
d of
tr
eatm
ent
• C
ultu
re c
onve
rsio
n by
6 m
onth
s
• Fa
ilure
• R
elap
se
• S
urvi
val (
or d
eath
)
• Ad
vers
e re
actio
ns (se
verit
y,
type
, org
an c
lass
)
• Ac
quis
ition
(am
plifi
catio
n) o
f ad
ditio
nal d
rug
resi
stan
ce
• Ad
here
nce
to t
reat
men
t (o
r tr
eatm
ent
inte
rrup
tion
due
to
non-
adhe
renc
e)
a R
egim
ens
last
ing
>12 a
nd <
18 m
onth
s w
ill n
ot b
e in
clud
ed in
the
inte
rven
tion
or c
ompa
rato
r.
Q4. A
mon
g pa
tient
s on
MD
R-T
B tre
atm
ent, a
re the
follo
win
g tw
o in
terv
entio
ns (d
elay
in s
tart
of tr
eatm
ent
and
elec
tive
surg
ery)
like
ly to
lead
to the
out
com
es li
sted
bel
ow?
PO
PU
LATI
ON
aIN
TER
VEN
TIO
N
CO
MPA
RAT
OR
O
UTC
OM
E
Patie
nts
on M
DR
-TB
tre
atm
ent
Patie
nts
on X
DR
-TB
tre
atm
ent
Chi
ldre
n (0
–14 y
ears
) vs
adu
lts
Pers
ons
with
HIV
(on
ant
iretr
ovira
ls)
Preg
nant
wom
en, a
nd p
eopl
e w
ith
diab
etes
Sta
rt o
f ad
equa
te t
reat
men
t w
ithin
fo
ur w
eeks
of di
agno
sis
(or
stro
ng
pres
umpt
ion)
Trea
tmen
t st
arte
d be
yond
fou
r w
eeks
of di
agno
sis
(or
stro
ng
pres
umpt
ion)
• C
ured
/com
plet
ed b
y en
d of
tre
atm
ent
• C
ultu
re c
onve
rsio
n by
6 m
onth
s
• Fa
ilure
• R
elap
se
• S
urvi
val (
or d
eath
)
• Ad
vers
e re
actio
ns (se
verit
y, t
ype,
or
gan
clas
s)
• Ad
here
nce
to t
reat
men
t (o
r tr
eatm
ent
inte
rrup
tion
due
to n
on-a
dher
ence
)
Elec
tive
surg
ery
(diff
eren
t ty
pes
/ st
ages
of di
seas
e)N
o el
ectiv
e su
rger
y
a T
he p
opul
atio
ns a
re e
xpec
ted
to d
iffer
for
the
tw
o su
bque
stio
ns;
patie
nts
who
are
sur
gica
lly o
pera
ted
are
mor
e lik
ely
to b
e XD
R-T
B a
nd p
erso
ns w
ith H
IV w
ho a
re h
avin
g an
tiret
rovi
rals
wou
ld b
e pa
rtic
ular
ly im
port
ant
for
the
first
sub
ques
tion.
ISBN: 978-92-4-154963-9
2016 update
WHO treatment guidelines for drug-resistant tuberculosis
OCTOBER 2016 REVISION
2016 update
WHO treatment guidelines for drug-resistant tuberculosis
ISBN: 978-92-4-154963-9
Annexes 4, 5 and 6
These guidelines were developed in compliance with the process for evidence gathering, assessment and formulation of recommendations, as outlined in the WHO Handbook for Guideline Development (version March 2014; available at http://www.who.int/kms/handbook_2nd_ed.pdf ).
WHO Library Cataloguing-in-Publication Data
WHO treatment guidelines for drug-resistant tuberculosis, 2016 update.
Contents: Web Annex 4: GRADE tables – Web Annex 5: Evidence to decision tables – Web Annex 6: Summaries of unpublished data used for the recommendations
1.Tuberculosis, Multidrug-Resistant – drug therapy. 2.Antitubercular Agents. 3.Guideline. I.World Health Organization.
ISBN 978 92 4 154963 9 (NLM classification: WF 310)
© World Health Organization 2016
All rights reserved. Publications of the World Health Organization are available on the WHO website (http://www.who.int) or can be purchased from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; email: [email protected]).
Requests for permission to reproduce or translate WHO publications –whether for sale or for non-commercial distribution– should be addressed to WHO Press through the WHO website (http://www.who.int/about/licensing/copyright_form/index.html).
The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted and dashed lines on maps represent approximate border lines for which there may not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.
All reasonable precautions have been taken by the World Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied. The responsibility for the interpretation and use of the material lies with the reader. In no event shall the World Health Organization be liable for damages arising from its use.
Printed by the WHO Document Production Services, Geneva, Switzerland
Design by Inis Communication – www.iniscommunication.com
WHO/HTM/TB/2016.04
Contents
Annex 4. GRADE tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1. Shorter regimens for MDR-TB (PICO 3) . . . . . . . . . . . . . . . . . . . . . . . . . . 22. MDR-TB regimen composition – systematic reviews of individual medicines
in adults (PICO 1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123. MDR-TB regimen composition – paediatric individual patient data
meta-analysis (PICO 1). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324. The role of surgery (PICO 4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Annex 5. Evidence to decision tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
1. Standardized shorter regimens versus longer regimens for the treatment of MDR-TB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2. Regimens with individualized composition and duration for adults and children with MDR-TB in whom a shorter MDR-TB regimen cannot be used . . . . . . . . . 55
3. Elective partial lung resection versus no surgery for patients on treatment for MDR-TB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Annex 6. Summaries of unpublished data used for the recommendations . . . . . . . . . . . 70
1. Short MDR-TB regimens: meta-analyses of data from published and unpublished studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
2. An updated systematic review and meta-analysis for treatment of multidrug-resistant tuberculosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3. A systematic review and individual patient data meta-analysis of treatment and outcomes among children with multidrug- resistant tuberculosis . . . . . . . . . . . . 77
Appendix 6A. Search strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Appendix 6B. Table of included studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
1
AN
NE
X 4
GR
AD
E ta
bles
The G
RA
DE
tabl
es ar
e ord
ered
as fo
llow
s:1
1.
Shor
ter r
egim
ens f
or M
DR
-TB
(PIC
O 3
)
2.
MD
R-T
B re
gim
en co
mpo
sitio
n –
syst
emat
ic re
view
s of i
ndiv
idua
l med
icin
es in
adul
ts (P
ICO
1)
3.
MD
R-T
B re
gim
en co
mpo
sitio
n –
paed
iatr
ic in
divi
dual
pat
ient
dat
a met
a-an
alys
is (P
ICO
1)
4.
The r
ole o
f sur
gery
(PIC
O 4
)
1 Ev
iden
ce fr
om st
udie
s ide
ntifi
ed d
urin
g th
e rev
iew
s per
form
ed to
answ
er P
ICO
que
stion
2 (t
reat
men
t of i
soni
azid
-resis
tant
TB
and
Myc
obac
teriu
m b
ovis)
as w
ell a
s par
t of P
ICO
que
stion
4 (d
elay
in
star
ting M
DR
-TB
trea
tmen
t) co
uld
not b
e sum
mar
ized
as G
RA
DE
tabl
es an
d th
us ar
e not
incl
uded
her
e. Se
e also
Ann
ex 6
for a
sum
mar
y of fi
ndin
gs fr
om st
udie
s tha
t wer
e not
pub
lishe
d at
the
time o
f the
rele
ase o
f the
se g
uide
lines
.
2
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
1. S
hort
er r
egim
ens
for
MD
R-T
B (
PIC
O 3
)
Aut
hor(
s): A
hmad
Kha
n F,
Ham
id S
alim
MA
, Sch
woe
bel V
, Tré
bucq
A, D
uCro
s P, C
asas
E, F
alzo
n D
, Men
zies
D (1
0 N
ovem
ber 2
015)
Que
stio
n: S
tand
ardi
zed
shor
ter
regi
men
s co
mpa
red
to lo
nger
reg
imen
s fo
r th
e tr
eatm
ent
of M
DR
-TB
(all
case
s; re
gard
less
of
pyra
zina
mid
e or
flu
oroq
uino
lone
susc
eptib
ility
)
Sett
ing:
Am
ong
patie
nts w
ho h
ad n
o hi
stor
y of
pre
viou
s tre
atm
ent w
ith se
cond
-line
dru
gs; s
hort
er re
gim
ens r
efer
to th
ose
lasti
ng u
p to
12
mon
ths;
long
er re
gim
ens l
ast 1
8 m
onth
s or m
ore.
Not
e tha
t the
“lon
ger r
egim
ens”
gro
up p
ools
data
from
stud
ies t
hat d
iffer
in th
e com
bina
tion
and
num
ber o
f dr
ugs,
in th
e dur
atio
n of
trea
tmen
t, an
d in
the u
se o
f a st
anda
rdiz
ed ve
rsus
an in
divi
dual
ized
appr
oach
. Hen
ce th
e poo
led
estim
ates
do
not n
eces
saril
y re
flect
the
outc
omes
ass
ocia
ted
with
the
regi
men
reco
mm
ende
d in
the
2011
WH
O G
uide
lines
for t
he p
rogr
amm
atic
man
agem
ent o
f dru
g-re
sista
nt
tube
rcul
osis.
Bib
liogr
aphy
: Res
ults
for s
hort
er re
gim
ens f
rom
agg
rega
te m
eta-
anal
ysis
com
bini
ng p
relim
inar
y da
ta fr
om th
ree
serie
s (1–
3), w
ith d
ata
from
thre
e pu
blish
ed st
udie
s (4–
6). R
esul
ts fo
r lon
ger r
egim
ens f
rom
aggr
egat
e met
a-an
alys
is us
ing d
ata f
rom
31
stud
ies o
f lon
ger M
DR
regi
men
s (7)
.
(1) M
édec
ins S
ans F
ront
ière
s Sw
azila
nd, p
relim
inar
y ou
tcom
es, u
npub
lishe
d da
ta. (
2) M
édec
ins S
ans F
ront
ière
s Uzb
eksit
an, p
relim
inar
y ou
tcom
es,
unpu
blish
ed d
ata.
(3) T
rébu
cq A
, Sch
woe
bel V
, Ghi
slain
Kou
ra K
, Rog
gi A
, Rie
der H
L. O
bser
vatio
nal s
tudy
on
the
eval
uatio
n of
the
toler
ance
and
eff
ectiv
enes
s of a
shor
t 9 m
onth
s tre
atm
ent f
or m
ultid
rug
resis
tant
tube
rcul
osis
patie
nts:
prel
imin
ary
repo
rt fo
r the
Wor
ld H
ealth
Org
aniz
atio
n. Th
e In
tern
atio
nal U
nion
Aga
inst
Tub
ercu
losis
and
Lung
Dise
ases
(UN
ION
). O
ctob
er 1
6 20
15. (
4) A
ung
KJ,
Van
Deu
n A
, Dec
lercq
E, S
arke
r MR
, Das
PK
, Hos
sain
MA
, et a
l. Suc
cessf
ul ‘9
-mon
th B
angl
ades
h re
gim
en’ f
or m
ultid
rug-
resis
tant
tube
rcul
osis
amon
g ove
r 500
cons
ecut
ive p
atie
nts.
Int J
Tub
erc
Lung
Dis.
201
4;18
(10)
:118
0–7.
(5) P
iube
llo A
, Har
ouna
SH
, Sou
leym
ane M
B, B
ouka
ry I,
Mor
ou S
, Dao
uda M
, et a
l. Hig
h cu
re ra
te w
ith st
anda
rdise
d sh
ort-c
ours
e mul
tidru
g-re
sista
nt tu
berc
ulos
is tr
eatm
ent i
n N
iger
: no
rela
pses
. Int
J Tu
berc
Lun
g Dis.
201
4;18
(10)
:118
8–94
. (6)
Kua
ban
C, N
oesk
e J,
Rie
der H
L, A
ït-K
hale
d N
, Abe
na F
oe JL
, Tré
bucq
A. H
igh
effec
tiven
ess o
f a 1
2-m
onth
regi
men
for M
DR
-TB
patie
nts i
n C
amer
oon.
Int J
Tub
erc L
ung
Dis.
201
5;19
(5):5
17–2
4. (7
) Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug
resis
tant
pul
mon
ary
tube
rcul
osis
trea
tmen
t reg
imen
s and
pat
ient
out
com
es: a
n in
divi
dual
pat
ient
dat
a met
a-an
alys
is of
9,1
53 p
atie
nts.
PLoS
Med
. 201
2;9(
8):1
212.
33
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
QU
ALI
TYIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se (a
sses
sed
with
: ind
irect
com
paris
on o
f two
agg
rega
te d
ata
met
a-an
alys
es (o
ne o
f sho
rter r
egim
ens
and
one
of lo
nger
regi
men
s)a
37b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
1008
/103
3 (9
7.6%
)c40
33/4
639
(86.
9%)d
not
estim
able
e
e
VERY
lOW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo a
ggre
gate
dat
a m
eta-
anal
yses
(one
of s
horte
r reg
imen
s an
d on
e of
long
er re
gim
ens)
a
37b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
1008
/111
6 (9
0.3%
)f40
33/5
850
(68.
9%)g
not
estim
able
e
e
VERY
lOW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
/los
s to
follo
w-up
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo p
oole
d in
divi
dual
pat
ient
met
a-an
alys
es)a
37b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
1008
/120
5 (8
3.7%
)h40
33/7
665
(52.
6%)i
not
estim
able
e
e
VERY
lOW
CRIt
ICAl
Cls:
con
fiden
ce li
mits
; RE:
rand
om e
ffect
sa I
n th
e sh
orte
r reg
imen
met
a-an
alys
is, d
ata
on re
laps
e we
re o
nly
avai
labl
e fro
m th
e pu
blis
hed
stud
ies
(refe
renc
es 4
–6);
in th
e lo
nger
regi
men
stu
dies
rela
pse
was
asce
rtain
ed in
14
coho
rts o
vera
ll (re
fere
nce
7).
b six
stud
ies
of s
horte
r reg
imen
s, 3
1 st
udie
s of
long
er re
gim
ens.
c Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 97.
6% (9
5% C
ls: 9
2.4%
–99.
2%).
d Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 91.
2% (9
5% C
ls: 8
6.1%
–94.
6%).
e Due
to m
etho
dolo
gica
l diff
eren
ces
in th
e st
udie
s th
e re
lativ
e an
d ab
solu
te ri
sks
are
not s
hown
. the
sho
rter M
DR-t
b re
gim
ens
data
set c
onsi
sts
of re
cent
ly c
ondu
cted
stu
dies
– s
ome
ongo
ing
– in
whi
ch p
atie
nts
were
ca
refu
lly s
elec
ted,
and
all
data
wer
e pr
ospe
ctive
ly c
olle
cted
as
part
of a
rese
arch
pro
toco
l. Pa
tient
s we
re u
nifo
rmly
trea
ted
with
a s
tand
ardi
zed
regi
men
. In
cont
rast
, stu
dies
with
long
er re
gim
ens
data
set w
ere
on
aver
age
olde
r, an
d m
any
were
retro
spec
tive
serie
s, a
nd m
any
used
dat
a co
llect
ed fo
r clin
ical
pur
pose
s. th
e la
rge
maj
ority
of p
atie
nts
in th
e co
nven
tiona
l reg
imen
s gr
oup
rece
ived
indi
vidu
alize
d th
erap
y, wi
th m
any
regi
men
s th
at d
iffer
ed fr
om o
ne a
noth
er in
num
ber a
nd ty
pe o
f dru
gs u
sed,
and
the
dura
tion
of tr
eatm
ent.
f Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 90.
3% (9
5% C
ls: 8
7.8%
–92.
4%).
g Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 78.
3% (9
5% C
ls: 7
1.2%
–84%
).h U
nwei
ghte
d pr
opor
tion;
wei
ghte
d pr
opor
tion
from
RE
met
a-an
alys
is: 8
3.7%
(95%
Cls
: 79.
2%–8
7.4%
).i U
nwei
ghte
d pr
opor
tion;
wei
ghte
d pr
opor
tion
from
RE
met
a-an
alys
is: 6
1.7%
(95%
Cls
: 53.
1%–6
9.6%
).
4
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): A
hmad
Kha
n F,
Ham
id S
alim
MA
, Sch
woe
bel V
, Tré
bucq
A, D
uCro
s P, C
asas
E, F
alzo
n D
, Men
zies
D (1
0 N
ovem
ber 2
015)
Que
stio
n: S
tand
ardi
zed
shor
ter r
egim
ens c
ompa
red
to lo
nger
regi
men
s for
the
trea
tmen
t of M
DR
-TB
(pyr
azin
amid
e su
scep
tible
; fluo
roqu
inol
one
susc
eptib
le)
Sett
ing:
Am
ong
patie
nts w
ho h
ad n
o hi
stor
y of
pre
viou
s tre
atm
ent w
ith se
cond
-line
dru
gs; s
hort
er re
gim
ens r
efer
to th
ose
lasti
ng u
p to
12
mon
ths;
long
er re
gim
ens l
ast 1
8 m
onth
s or m
ore.
Not
e tha
t the
“lon
ger r
egim
ens”
grou
p po
ols d
ata f
rom
stud
ies t
hat d
iffer
ed in
the c
ombi
natio
n an
d nu
mbe
r of
drug
s, in
the d
urat
ion
of tr
eatm
ent,
and
in th
e use
of a
stan
dard
ized
vers
us an
indi
vidu
aliz
ed ap
proa
ch. H
ence
the p
oole
d es
timat
es d
o no
t nec
essa
rily
refle
ct th
e ou
tcom
es a
ssoc
iate
d w
ith th
e re
gim
en re
com
men
ded
in th
e 20
11 W
HO
Gui
delin
es fo
r the
pro
gram
mat
ic m
anag
emen
t of d
rug-
resis
tant
tu
berc
ulos
is.
Bib
liogr
aphy
: Res
ults
for s
hort
er re
gim
ens f
rom
indi
vidu
al p
atie
nt d
ata m
eta-
anal
ysis
of u
npub
lishe
d (1
,2) a
nd p
ublis
hed
(3) d
ata.
Res
ults
for l
onge
r re
gim
ens f
rom
indi
vidu
al p
atie
nt d
ata m
eta-
anal
ysis
usin
g dat
a fro
m st
udy (
4).
(1) M
édec
ins S
ans F
ront
ière
s Sw
azila
nd, p
relim
inar
y ou
tcom
es, u
npub
lishe
d da
ta. (
2) M
édec
ins S
ans F
ront
ière
s Uzb
eksit
an, p
relim
inar
y ou
tcom
es,
unpu
blish
ed d
ata.
(3)
Aun
g K
J, Va
n D
eun
A, D
ecler
cq E
, Sar
ker
MR
, Das
PK
, Hos
sain
MA
, et a
l. Su
cces
sful ‘
9-m
onth
Ban
glad
esh
regi
men
’ for
m
ultid
rug-
resis
tant
tube
rcul
osis
amon
g ove
r 500
cons
ecut
ive p
atie
nts.
Int J
Tub
erc L
ung D
is. 2
014;
18(1
0):1
180–
7. (4
) Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug r
esist
ant p
ulm
onar
y tub
ercu
losis
trea
tmen
t reg
imen
s and
pat
ient
out
com
es: a
n in
divi
dual
pat
ient
da
ta m
eta-
anal
ysis
of 9
,153
pat
ient
s. PL
oS M
ed. 2
012;
9(8)
:121
2.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se (a
sses
sed
with
: ind
irect
com
paris
on o
f two
poo
led
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
es)a
26b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
dos
e re
spon
se g
radi
entc
121/
121
(100
.0%
)d89
0/97
9 (9
0.9%
)eno
t es
timab
lef
f
VERY
lOW
CRIt
ICAl
55
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo p
oole
d in
divi
dual
pat
ient
dat
a m
eta-
anal
yses
)a
26b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
dos
e re
spon
se g
radi
entc
121/
125
(96.
8%)g
890/
1119
(7
9.5%
)hno
t es
timab
lef
f
VERY
lOW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
/los
s to
follo
w-up
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo p
oole
d in
divi
dual
pat
ient
dat
a m
eta-
anal
yses
)a
26b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
dos
e re
spon
se g
radi
entc
121/
132
(91.
7%)i
890/
1666
(5
3.4%
)jno
t es
timab
lef
f
VERY
lOW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; R
E: ra
ndom
effe
cts
a In
the
shor
ter r
egim
en in
divi
dual
pat
ient
met
a-an
alys
is, d
ata
on re
laps
e we
re o
nly
avai
labl
e in
the
bang
lade
sh s
erie
s, in
whi
ch s
ix pa
tient
s ex
perie
nced
trea
tmen
t fai
lure
and
thre
e ot
hers
rela
psed
.b t
hree
stu
dies
of s
horte
r reg
imen
s; 2
3 st
udie
s of
long
er re
gim
ens.
c Dos
e-re
spon
se g
radi
ent r
efer
s to
the
inve
rse
rela
tions
hip
obse
rved
bet
ween
incr
easi
ng re
sist
ance
and
dec
reas
ing
effe
ctive
ness
of t
reat
men
t.d C
onfid
ence
lim
its c
ould
not
be
com
pute
d us
ing
met
a-an
alyt
ical
met
hods
. Exa
ct b
inom
ial 9
5%Cl
s: 9
7.0%
–100
%.
e Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 94.
5% (9
5% C
ls: 8
8.9%
–97.
4%).
f Due
to m
etho
dolo
gica
l diff
eren
ces
in th
e st
udie
s th
e re
lativ
e an
d ab
solu
te ri
sks
are
not s
hown
. the
sho
rter M
DR-t
b re
gim
ens
data
set c
onsi
sts
of re
cent
ly c
ondu
cted
stu
dies
– s
ome
ongo
ing
– in
whi
ch p
atie
nts
were
ca
refu
lly s
elec
ted,
and
all
data
wer
e pr
ospe
ctive
ly c
olle
cted
as
part
of a
rese
arch
pro
toco
l. Pa
tient
s we
re u
nifo
rmly
trea
ted
with
a s
tand
ardi
zed
regi
men
. In
cont
rast
, stu
dies
with
long
er re
gim
ens
data
set w
ere
on
aver
age
olde
r, an
d m
any
were
retro
spec
tive
serie
s, a
nd m
any
used
dat
a co
llect
ed fo
r clin
ical
pur
pose
s. th
e la
rge
maj
ority
of p
atie
nts
in th
e lo
nger
regi
men
s gr
oup
rece
ived
indi
vidu
alize
d th
erap
y, wi
th m
any
regi
men
s th
at d
iffer
ed fr
om o
ne a
noth
er in
num
ber a
nd ty
pe o
f dru
gs u
sed,
and
the
dura
tion
of tr
eatm
ent.
g Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 96.
8% (9
5% C
ls: 7
7.3%
–99.
6%).
h Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 83.
5% (9
5% C
ls: 7
5.7%
–89.
2%).
i Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 91.
7% (9
5% C
ls: 7
3.9%
–97.
7%).
j Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 68.
2% (9
5% C
ls: 5
6.2%
–78.
1%).
6
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): A
hmad
Kha
n F,
Ham
id S
alim
MA
, Sch
woe
bel V
, Tré
bucq
A, D
uCro
s P, C
asas
E, F
alzo
n D
, Men
zies
D (1
0 N
ovem
ber 2
015)
Que
stio
n: S
tand
ardi
zed
shor
ter r
egim
ens c
ompa
red
to lo
nger
regi
men
s for
the
trea
tmen
t of M
DR
-TB
(pyr
azin
amid
e su
scep
tible
; fluo
roqu
inol
one
resis
tant
)
Sett
ing:
Am
ong
patie
nts w
ho h
ad n
o hi
stor
y of
pre
viou
s tre
atm
ent w
ith se
cond
-line
dru
gs; s
hort
er re
gim
ens r
efer
to th
ose
lasti
ng u
p to
12
mon
ths;
long
er re
gim
ens l
ast 1
8 m
onth
s or m
ore.
Not
e tha
t the
“lon
ger r
egim
ens”
grou
p po
ols d
ata f
rom
stud
ies t
hat d
iffer
ed in
the c
ombi
natio
n an
d nu
mbe
r of
drug
s, in
the d
urat
ion
of tr
eatm
ent,
and
in th
e use
of a
stan
dard
ized
vers
us an
indi
vidu
aliz
ed ap
proa
ch. H
ence
the p
oole
d es
timat
es d
o no
t nec
essa
rily
refle
ct th
e ou
tcom
es a
ssoc
iate
d w
ith th
e re
gim
en re
com
men
ded
in th
e 20
11 W
HO
Gui
delin
es fo
r the
pro
gram
mat
ic m
anag
emen
t of d
rug-
resis
tant
tu
berc
ulos
is.
Bib
liogr
aphy
: Res
ults
for s
hort
er re
gim
ens f
rom
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
is of
unp
ublis
hed
(1) a
nd p
ublis
hed
(2) d
ata.
Res
ults
for l
onge
r re
gim
ens f
rom
indi
vidu
al p
atie
nt d
ata m
eta-
anal
ysis
usin
g dat
a fro
m st
udy (
3).
(1) M
édec
ins S
ans F
ront
ière
s Sw
azila
nd, p
relim
inar
y out
com
es, u
npub
lishe
d da
ta. (
2) A
ung K
J, Va
n D
eun
A, D
ecler
cq E
, Sar
ker M
R, D
as P
K, H
ossa
in
MA
, et a
l. Su
cces
sful ‘
9-m
onth
Ban
glad
esh
regi
men
’ for
mul
tidru
g-re
sista
nt tu
berc
ulos
is am
ong o
ver 5
00 co
nsec
utiv
e pat
ient
s. In
t J T
uber
c Lun
g Dis.
20
14;1
8(10
):118
0–7.
(3) A
huja
SD
, Ash
kin
D, A
vend
ano
M, B
aner
jee
R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug
resis
tant
pul
mon
ary
tube
rcul
osis
trea
tmen
t reg
imen
s and
pat
ient
out
com
es: a
n in
divi
dual
pat
ient
dat
a met
a-an
alys
is of
9,1
53 p
atie
nts.
PLoS
Med
. 201
2;9(
8):1
212.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se (a
sses
sed
with
: ind
irect
com
paris
on o
f two
poo
led
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
es)a
18b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
do
se re
spon
se
grad
ient
c
12/1
4 (8
5.7%
)d72
/95
(75.
8%)e
not
estim
able
f
f
VERY
lOW
CRIt
ICAl
77
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo p
oole
d in
divi
dual
pat
ient
dat
a m
eta-
anal
yses
)a
18b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
dos
e re
spon
se g
radi
entc
12/1
5 (8
0.0%
)g72
/120
(6
0.0%
)hno
t es
timab
lef
f
VERY
lOW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
/los
s to
follo
w-up
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo p
oole
d in
divi
dual
pat
ient
dat
a m
eta-
anal
yses
)a
18b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
dos
e re
spon
se g
radi
entc
12/1
8 (6
6.7%
)i72
/155
(4
6.5%
)jno
t es
timab
lef
f
VERY
lOW
CRIt
ICAl
Cls:
con
fiden
ce li
mits
; RE:
rand
om e
ffect
sa F
luor
oqui
nolo
ne re
sist
ance
was
an
excl
usio
n cr
iterio
n fo
r enr
olm
ent i
nto
MsF
’s Uz
beki
stan
sho
rter r
egim
en c
ohor
t. In
the
abov
e in
divi
dual
pat
ient
met
a-an
alys
es fo
r the
sho
rter r
egim
ens,
eac
h gr
oup
cons
ists
of 1
pa
tient
from
the
swaz
iland
coh
ort w
ith th
e re
mai
nder
con
sist
ing
of p
atie
nts
from
the
bang
lade
sh s
tudy
(13
for s
ucce
ss v
ersu
s fa
ilure
; 14
for s
ucce
ss v
ersu
s fa
ilure
or d
eath
; and
17
for s
ucce
ss v
ersu
s fa
ilure
, dea
th,
or lo
ss to
follo
w-up
). In
the
shor
ter r
egim
en in
divi
dual
pat
ient
met
a-an
alys
is, d
ata
on re
laps
e we
re o
nly
avai
labl
e in
the
bang
lade
sh s
erie
s.b t
wo s
tudi
es o
f sho
rter r
egim
ens;
16
stud
ies
of lo
nger
regi
men
s.c D
ose-
resp
onse
gra
dien
t ref
ers
to th
e in
vers
e re
latio
nshi
p ob
serv
ed b
etwe
en in
crea
sing
resi
stan
ce a
nd d
ecre
asin
g ef
fect
ivene
ss o
f tre
atm
ent.
d Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m F
E m
eta-
anal
ysis
: 85.
7% (9
5% C
ls: 5
3.5%
-96.
9%).
e Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 55.
7% (9
5% C
ls: 4
0.8%
-69.
8%).
f Due
to m
etho
dolo
gica
l diff
eren
ces
in th
e st
udie
s th
e re
lativ
e an
d ab
solu
te ri
sks
are
not s
hown
. the
sho
rter M
DR-t
b re
gim
ens
data
set c
onsi
sts
of re
cent
ly c
ondu
cted
stu
dies
– s
ome
ongo
ing
– in
whi
ch p
atie
nts
were
ca
refu
lly s
elec
ted,
and
all
data
wer
e pr
ospe
ctive
ly c
olle
cted
as
part
of a
rese
arch
pro
toco
l. Pa
tient
s we
re u
nifo
rmly
trea
ted
with
a s
tand
ardi
zed
regi
men
. In
cont
rast
, stu
dies
with
long
er re
gim
ens
data
set w
ere
on
aver
age
olde
r, an
d m
any
were
retro
spec
tive
serie
s, a
nd m
any
used
dat
a co
llect
ed fo
r clin
ical
pur
pose
s. th
e la
rge
maj
ority
of p
atie
nts
in th
e lo
nger
regi
men
s gr
oup
rece
ived
indi
vidu
alize
d th
erap
y, wi
th m
any
regi
men
s th
at d
iffer
ed fr
om o
ne a
noth
er in
num
ber a
nd ty
pe o
f dru
gs u
sed,
and
the
dura
tion
of tr
eatm
ent.
g Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m F
E m
eta-
anal
ysis
: 80.
0% (9
5% C
ls: 5
0.0%
–94.
1%).
h Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 64.
4% (9
5% C
ls: 4
9.6%
–76.
9%).
i Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m F
E m
eta-
anal
ysis
: 66.
7% (9
5% C
ls: 4
1.1%
–85.
2%).
j Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 56.
1% (9
5% C
ls: 4
0.7%
–70.
4%).
8
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): A
hmad
Kha
n F,
Ham
id S
alim
MA
, Sch
woe
bel V
, Tré
bucq
A, D
uCro
s P, C
asas
E, F
alzo
n D
, Men
zies
D (1
0 N
ovem
ber 2
015)
Que
stio
n: S
tand
ardi
zed
shor
ter
regi
men
s co
mpa
red
to lo
nger
reg
imen
s fo
r th
e tr
eatm
ent
of M
DR
-TB
(pyr
azin
amid
e re
sista
nt; fl
uoro
quin
olon
e su
scep
tible
)
Sett
ing:
Am
ong
patie
nts w
ho h
ad n
o hi
stor
y of
pre
viou
s tre
atm
ent w
ith se
cond
-line
dru
gs; s
hort
er re
gim
ens r
efer
to th
ose
lasti
ng u
p to
12
mon
ths;
long
er re
gim
ens l
ast 1
8 m
onth
s or m
ore.
Not
e tha
t the
“lon
ger r
egim
ens”
grou
p po
ols d
ata f
rom
stud
ies t
hat d
iffer
ed in
the c
ombi
natio
n an
d nu
mbe
r of
drug
s, in
the d
urat
ion
of tr
eatm
ent,
and
in th
e use
of a
stan
dard
ized
vers
us an
indi
vidu
aliz
ed ap
proa
ch. H
ence
the p
oole
d es
timat
es d
o no
t nec
essa
rily
refle
ct th
e ou
tcom
es a
ssoc
iate
d w
ith th
e re
gim
en re
com
men
ded
in th
e 20
11 W
HO
Gui
delin
es fo
r the
pro
gram
mat
ic m
anag
emen
t of d
rug-
resis
tant
tu
berc
ulos
is.
Bib
liogr
aphy
: Res
ults
for s
hort
er re
gim
ens f
rom
indi
vidu
al p
atie
nt d
ata m
eta-
anal
ysis
of u
npub
lishe
d (1
,2) a
nd p
ublis
hed
(3) d
ata.
Res
ults
for l
onge
r re
gim
ens f
rom
indi
vidu
al p
atie
nt d
ata m
eta-
anal
ysis
usin
g dat
a fro
m st
udy (
4).
(1) M
édec
ins S
ans F
ront
ière
s Sw
azila
nd, p
relim
inar
y ou
tcom
es, u
npub
lishe
d da
ta. (
2) M
édec
ins S
ans F
ront
ière
s Uzb
eksit
an, p
relim
inar
y ou
tcom
es,
unpu
blish
ed d
ata.
(3)
Aun
g K
J, Va
n D
eun
A, D
ecler
cq E
, Sar
ker
MR
, Das
PK
, Hos
sain
MA
, et a
l. Su
cces
sful ‘
9-m
onth
Ban
glad
esh
regi
men
’ for
m
ultid
rug-
resis
tant
tube
rcul
osis
amon
g ove
r 500
cons
ecut
ive p
atie
nts.
Int J
Tub
erc L
ung D
is. 2
014;
18(1
0):1
180–
7. (4
) Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug r
esist
ant p
ulm
onar
y tub
ercu
losis
trea
tmen
t reg
imen
s and
pat
ient
out
com
es: a
n in
divi
dual
pat
ient
da
ta m
eta-
anal
ysis
of 9
,153
pat
ient
s. PL
oS M
ed. 2
012;
9(8)
:121
2.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se (a
sses
sed
with
: ind
irect
com
paris
on o
f two
poo
led
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
es)a
26b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
dos
e re
spon
se g
radi
entc
90/9
6 (9
3.8%
)d84
0/96
2 (8
7.3%
)eno
t es
timab
lef
f
VERY
lOW
CRIt
ICAl
99
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
s (a
sses
sed
with
: ind
irect
com
paris
on o
f two
poo
led
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
es)a
26b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
dos
e re
spon
se g
radi
entc
90/1
00
(90.
0%)g
840/
1075
(7
8.1%
)hno
t es
timab
lef
f
VERY
lOW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
s/lo
ss to
follo
w-up
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo p
oole
d in
divi
dual
pat
ient
dat
a m
eta-
anal
yses
)a
26b
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
serio
us
stro
ng a
ssoc
iatio
n al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
th
e de
mon
-st
rate
d ef
fect
dos
e re
spon
se g
radi
ent3
90/1
07
(84.
1%)i
840/
1392
(6
0.3%
)jno
t es
timab
lef
f
VERY
lOW
CRIt
ICAl
Cls:
con
fiden
ce li
mits
; RE:
rand
om e
ffect
sa I
n th
e sh
orte
r reg
imen
indi
vidu
al p
atie
nt m
eta-
anal
ysis
, dat
a on
rela
pse
were
onl
y av
aila
ble
in th
e ba
ngla
desh
ser
ies.
b thr
ee s
tudi
es o
f sho
rter r
egim
ens;
23
stud
ies
of lo
nger
regi
men
s.c D
ose-
resp
onse
gra
dien
t ref
ers
to th
e in
vers
e re
latio
nshi
p ob
serv
ed b
etwe
en in
crea
sing
resi
stan
ce a
nd d
ecre
asin
g ef
fect
ivene
ss o
f tre
atm
ent.
d Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 93.
5% (9
5% C
ls: 4
0.4%
–99.
7%).
e Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 90.
1% (9
5% C
ls: 8
3.5%
–94.
2%).
f Due
to m
etho
dolo
gica
l diff
eren
ces
in th
e st
udie
s th
e re
lativ
e an
d ab
solu
te ri
sks
are
not s
hown
. the
sho
rter M
DR-t
b re
gim
ens
data
set c
onsi
sts
of re
cent
ly c
ondu
cted
stu
dies
– s
ome
ongo
ing
– in
whi
ch p
atie
nts
were
ca
refu
lly s
elec
ted,
and
all
data
wer
e pr
ospe
ctive
ly c
olle
cted
as
part
of a
rese
arch
pro
toco
l. Pa
tient
s we
re u
nifo
rmly
trea
ted
with
a s
tand
ardi
zed
regi
men
. In
cont
rast
, stu
dies
with
long
er re
gim
ens
data
set w
ere
on
aver
age
olde
r, an
d m
any
were
retro
spec
tive
serie
s, a
nd m
any
used
dat
a co
llect
ed fo
r clin
ical
pur
pose
s. th
e la
rge
maj
ority
of p
atie
nts
in th
e lo
nger
regi
men
s gr
oup
rece
ived
indi
vidu
alize
d th
erap
y, wi
th m
any
regi
men
s th
at d
iffer
ed fr
om o
ne a
noth
er in
num
ber a
nd ty
pe o
f dru
gs u
sed,
and
the
dura
tion
of tr
eatm
ent.
g Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 88.
8% (9
5% C
ls: 4
7.3%
–98.
6%).
h Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 81.
4% (9
5% C
ls: 7
1.6%
–88.
4%).
i Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 83.
3% (9
5% C
ls: 2
7.3%
–98.
5%).
j Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m R
E m
eta-
anal
ysis
: 64.
0% (9
5% C
ls: 5
3.0%
–73.
8%).
10
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): A
hmad
Kha
n F,
Ham
id S
alim
MA
, Sch
woe
bel V
, Tré
bucq
A, D
uCro
s P, C
asas
E, F
alzo
n D
, Men
zies
D (1
0 N
ovem
ber 2
015)
Que
stio
n: S
tand
ardi
zed
shor
ter
regi
men
s co
mpa
red
to lo
nger
reg
imen
s fo
r th
e tr
eatm
ent
of M
DR
-TB
(pyr
azin
amid
e re
sista
nt; fl
uoro
quin
olon
e re
sista
nt)
Sett
ing:
Am
ong
patie
nts w
ho h
ad n
o hi
stor
y of
pre
viou
s tre
atm
ent w
ith se
cond
-line
dru
gs; s
hort
er re
gim
ens r
efer
to th
ose
lasti
ng u
p to
12
mon
ths;
long
er re
gim
ens l
ast 1
8 m
onth
s or m
ore.
Not
e tha
t the
“lon
ger r
egim
ens”
grou
p po
ols d
ata f
rom
stud
ies t
hat d
iffer
ed in
the c
ombi
natio
n an
d nu
mbe
r of
drug
s, in
the d
urat
ion
of tr
eatm
ent,
and
in th
e use
of a
stan
dard
ized
vers
us an
indi
vidu
aliz
ed ap
proa
ch. H
ence
the p
oole
d es
timat
es d
o no
t nec
essa
rily
refle
ct th
e ou
tcom
es a
ssoc
iate
d w
ith th
e re
gim
en re
com
men
ded
in th
e 20
11 W
HO
Gui
delin
es fo
r the
pro
gram
mat
ic m
anag
emen
t of d
rug-
resis
tant
tu
berc
ulos
is.
Bib
liogr
aphy
: Res
ults
for s
hort
er re
gim
ens f
rom
one
pub
lishe
d st
udy (
1). R
esul
ts fo
r lon
ger r
egim
ens f
rom
indi
vidu
al p
atie
nt d
ata m
eta-
anal
ysis
usin
g da
ta fr
om st
udy (
2).
(1) A
ung K
J, Va
n D
eun
A, D
ecler
cq E
, Sar
ker M
R, D
as P
K, H
ossa
in M
A, R
iede
r HL.
Suc
cess
ful ‘9
-mon
th B
angl
ades
h re
gim
en’ f
or m
ultid
rug-
resis
tant
tu
berc
ulos
is am
ong
over
500
con
secu
tive
patie
nts.
Int J
Tub
erc
Lung
Dis.
201
4;18
(10)
:118
0–7.
(2) A
huja
SD
, Ash
kin
D, A
vend
ano
M, B
aner
jee
R,
Baue
r M, B
ayon
a JN
, et a
l. M
ultid
rug
resis
tant
pul
mon
ary
tube
rcul
osis
trea
tmen
t reg
imen
s and
pat
ient
out
com
es: a
n in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of 9
,153
pat
ient
s. PL
oS M
ed. 2
012;
9(8)
:121
2.a
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se (a
sses
sed
with
: ind
irect
com
paris
on o
f two
poo
led
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
es)b
19c
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
very
ser
ious
dst
rong
ass
ocia
tion
all p
laus
ible
resi
d-ua
l con
foun
ding
wo
uld
redu
ce
the
dem
on-
stra
ted
effe
ct d
ose
resp
onse
gra
dien
te
19/2
6 (7
3.1%
)f81
/112
(7
2.3%
)gno
t es
timab
leh
h
VERY
lOW
CRIt
ICAl
1111
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
STAN
DARD
IZED
SH
ORTE
R RE
GIM
ENS
LONG
ER
REGI
MEN
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo p
oole
d in
divi
dual
pat
ient
dat
a m
eta-
anal
yses
)b
19c
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
very
ser
ious
dst
rong
ass
ocia
tion
all p
laus
ible
resi
d-ua
l con
foun
ding
wo
uld
redu
ce
the
dem
on-
stra
ted
effe
ct d
ose
resp
onse
gra
dien
te
19/2
8 (6
7.9%
)i81
/137
(5
9.1%
)jno
t es
timab
leh
h
VERY
lOW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
/los
s to
follo
w-up
(ass
esse
d wi
th: i
ndire
ct c
ompa
rison
of t
wo p
oole
d in
divi
dual
pat
ient
dat
a m
eta-
anal
yses
)b
19c
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
very
ser
ious
dst
rong
ass
ocia
tion
all p
laus
ible
resi
d-ua
l con
foun
ding
wo
uld
redu
ce
the
dem
on-
stra
ted
effe
ct d
ose
resp
onse
gra
dien
te
19/3
2 (5
9.4%
)k81
/193
(4
2.0%
)lno
t es
timab
leh
h
VERY
lOW
CRIt
ICAl
Cls:
con
fiden
ce li
mits
; RE:
rand
om e
ffect
sa I
n th
e st
udy
by A
ung,
et a
l. (1
) rep
ortin
g re
sults
from
the
sam
e ba
ngla
desh
coh
ort,
high
-leve
l gat
iflox
acin
-resi
stan
ce (d
efine
d as
MIC
≥2m
g/m
l) w
as a
ssoc
iate
d wi
th u
nsuc
cess
ful t
reat
men
t, bu
t not
low-
leve
l ga
tiflox
acin
-resi
stan
ce. I
n th
e ab
ove
tabl
e, a
ll pe
rson
s in
the
shor
t reg
imen
gro
up h
ad o
floxa
cin-
resi
stan
t MDR
-tb,
and
am
ongs
t the
se, h
igh-
leve
l gat
iflox
acin
resi
stan
ce w
as d
ocum
ente
d in
15;
low-
leve
l gat
iflox
acin
-re
sist
ance
in 1
3; a
nd g
atifl
oxac
in M
IC w
as n
ot m
easu
red
in 4
.b I
n th
e sh
orte
r reg
imen
indi
vidu
al p
atie
nt m
eta-
anal
ysis
, all
data
are
from
ban
glad
esh
(i.e.
no
patie
nts
from
swa
zilan
d or
Uzb
ekis
tan)
.c O
ne s
tudy
of s
horte
r reg
imen
s; 1
8 st
udie
s of
long
er re
gim
ens.
d Con
fiden
ce li
mits
are
wid
e fo
r sho
rter r
egim
en; a
ll sh
orte
r reg
imen
resu
lts a
re fr
om o
ne s
tudy
onl
y (A
ung,
et a
l.), a
nd fe
w pa
tient
s in
volve
d.e D
ose-
resp
onse
gra
dien
t ref
ers
to th
e in
vers
e re
latio
nshi
p ob
serv
ed b
etwe
en in
crea
sing
resi
stan
ce a
nd d
ecre
asin
g ef
fect
ivene
ss o
f tre
atm
ent.
f Unw
eigh
ted
prop
ortio
n; e
xact
bin
omia
l 95%
Cls
: 52.
2%–8
7.1%
.g U
nwei
ghte
d pr
opor
tion;
wei
ghte
d pr
opor
tion
from
RE
met
a-an
alys
is: 5
9.4%
(95%
Cls
: 41.
2%–7
5.3%
).h D
ue to
met
hodo
logi
cal d
iffer
ence
s in
the
stud
ies
the
rela
tive
and
abso
lute
risk
s ar
e no
t sho
wn. t
he s
horte
r MDR
-tb
regi
men
s da
tase
t con
sist
s of
rece
ntly
con
duct
ed s
tudi
es –
som
e on
goin
g –
in w
hich
pat
ient
s we
re
care
fully
sel
ecte
d, a
nd a
ll da
ta w
ere
pros
pect
ively
col
lect
ed a
s pa
rt of
a re
sear
ch p
roto
col.
Patie
nts
were
uni
form
ly tr
eate
d wi
th a
sta
ndar
dize
d re
gim
en. I
n co
ntra
st, s
tudi
es w
ith lo
nger
regi
men
s da
tase
t wer
e on
av
erag
e ol
der,
and
man
y we
re re
trosp
ectiv
e se
ries,
and
man
y us
ed d
ata
colle
cted
for c
linic
al p
urpo
ses.
the
larg
e m
ajor
ity o
f pat
ient
s in
the
long
er re
gim
ens
grou
p re
ceive
d in
divi
dual
ized
ther
apy,
with
man
y re
gim
ens
that
diff
ered
from
one
ano
ther
in n
umbe
r and
type
of d
rugs
use
d, a
nd th
e du
ratio
n of
trea
tmen
t.i U
nwei
ghte
d pr
opor
tion;
exa
ct b
inom
ial 9
5% C
ls: 4
7.6%
–84.
1%.
j Unw
eigh
ted
prop
ortio
n; w
eigh
ted
prop
ortio
n fro
m F
E m
eta-
anal
ysis
: 59.
1% (9
5% C
ls: 5
0.6%
–67.
1%).
k Unw
eigh
ted
prop
ortio
n; e
xact
bin
omia
l 95%
Cls
: 40.
6%–7
6.3%
.l U
nwei
ghte
d pr
opor
tion;
wei
ghte
d pr
opor
tion
from
RE
met
a-an
alys
is: 4
9.9%
(95%
Cls
: 30.
6%–6
9.2%
).
12
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
2. M
DR
-TB
reg
imen
com
posi
tion
– s
yste
mat
ic r
evie
ws
of in
divi
dual
med
icin
es in
adu
lts
(PIC
O 1
)
Aut
hor(
s): B
asto
s M, L
an Z
, Men
zies
R (1
1 N
ovem
ber 2
015)
Que
stio
n: A
late
r gen
erat
ion
fluor
oqui
nolo
ne co
mpa
red
to n
o la
ter g
ener
atio
n flu
oroq
uino
lone
for a
dults
with
rifa
mpi
cin-
resis
tant
TB
or M
DR
-TBa
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing c
onve
ntio
nal r
egim
ens l
astin
g abo
ut 2
4 m
onth
s, in
low
and
high
re
sour
ce se
tting
s, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are
Bib
liogr
aphy
: Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R
, Bau
er M
, Bay
ona
JN, e
t al.
Mul
tidru
g re
sista
nt p
ulm
onar
y tu
berc
ulos
is tr
eatm
ent
regi
men
s and
pat
ient
out
com
es: a
n in
divi
dual
pat
ient
dat
a met
a-an
alys
is of
9,1
53 p
atie
nts.
PLoS
Med
. 201
2;9(
8):1
212.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
A LA
TER
GENE
RATI
ON
FLUO
ROQU
I-NO
LONE
NO L
ATER
GE
NERA
TION
FL
UORO
QUI-
NOLO
NERE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on la
ter g
ener
atio
n flu
oroq
uino
lone
ver
sus
no fl
uoro
quin
olon
e, a
s pa
rt of
a M
DR-t
b re
gim
en (a
sses
sed
with
: ind
ivid
ual p
atie
nt d
ata
met
a-an
alys
is (A
huja
sD,
et a
l. Pl
Os M
ed. 2
012)
32
obse
rvat
iona
l st
udie
s se
rious
bno
t ser
ious
no
t ser
ious
no
t ser
ious
al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
691/
833
(83.
0%)
301/
678
(44.
4%)
OR 2
.5
(1.0
to
5.9)
c
390
mor
e pe
r 1,0
00
(from
30
fewe
r to
640
mor
e)
lO
W
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on la
ter g
ener
atio
n flu
oroq
uino
lone
ver
sus
oflox
acin
, as
part
of a
MDR
-tb
regi
men
(ass
esse
d wi
th: i
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
(Ahu
ja
sD, e
t al.
PlOs
Med
. 201
2)
32
obse
rvat
iona
l st
udie
s se
rious
bno
t ser
ious
no
t ser
ious
no
t ser
ious
al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
691/
833
(83.
0%)
3386
/462
4 (7
3.2%
) OR
1.9
(1
.0 to
3.
6)c
100
mor
e pe
r 1,0
00
(from
15
fewe
r to
240
mor
e)
lO
W
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se in
pat
ient
s on
late
r gen
erat
ion
fluor
oqui
nolo
ne v
ersu
s no
fluo
roqu
inol
one
or c
ipro
floxa
cin
or o
floxa
cin,
as
part
of a
MDR
-tb
regi
men
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)d
48
obse
rvat
iona
l st
udie
s se
rious
eno
t ser
ious
no
t ser
ious
no
t ser
ious
no
ne
4270
/497
8 (8
5.8%
)f33
97/4
046
(84.
0%)g
10 fe
wer
pe
r 1,0
00
(from
78
fewe
r to
57 m
ore)
VE
RY l
OW
CRIt
ICAl
1313
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
A LA
TER
GENE
RATI
ON
FLUO
ROQU
I-NO
LONE
NO L
ATER
GE
NERA
TION
FL
UORO
QUI-
NOLO
NERE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on la
ter g
ener
atio
n flu
oroq
uino
lone
ver
sus
no fl
uoro
quin
olon
e or
cip
roflo
xaci
n or
oflo
xaci
n, a
s pa
rt of
a M
DR-t
b re
gim
en (a
sses
sed
with
: ag
greg
ated
dat
a m
eta-
anal
ysis
201
5)
47
obse
rvat
iona
l st
udie
s se
rious
eno
t ser
ious
no
t ser
ious
no
t ser
ious
no
ne
4270
/547
4 (7
8.0%
)h33
97/4
958
(68.
5%)i
23 m
ore
per 1
,000
(fr
om 6
0 fe
wer t
o 10
8 m
ore)
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s (G
rade
3 o
r 4, o
r dru
gs s
topp
ed d
ue to
adv
erse
eve
nts)
in p
atie
nts
on la
ter-g
ener
atio
n flu
oroq
uino
lone
13
obse
rvat
iona
l st
udie
s se
rious
no
t ser
ious
no
t ser
ious
no
t ser
ious
no
nej
10/8
27
(1.2
%)k
not
estim
able
j
VERY
lOW
CRIt
ICAl
serio
us a
dver
se e
vent
s (G
rade
3 o
r 4, o
r dru
gs s
topp
ed d
ue to
adv
erse
eve
nts)
in p
atie
nts
on o
floxa
cin
or c
ipro
floxa
cin
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)
9 ob
serv
atio
nal
stud
ies
serio
us
not s
erio
us
not s
erio
us
not s
erio
us
none
j40
1/14
08
(28.
5%)l
not
estim
able
j
VERY
lOW
CRIt
ICAl
Cls:
con
fiden
ce li
mits
; FE:
fixe
d ef
fect
s; O
R: o
dds
ratio
a Use
of l
ater
gen
erat
ion
fluor
oqui
nolo
nes
(mox
iflox
acin
, gat
iflox
acin
or l
evofl
oxac
in) i
s co
mpa
red
with
use
of o
floxa
cin
or n
o flu
oroq
uino
lone
alo
ngsi
de o
ther
dru
gs in
the
MDR
-tb
regi
men
; one
out
com
e re
late
d to
se
vere
adv
erse
eve
nts
of o
floxa
cin
also
incl
uded
in th
is ta
ble.
b In
the
indi
vidu
al p
atie
nt d
ata
anal
ysis
(Ahu
ja s
D, e
t al.)
, mos
t pat
ient
s re
ceive
d in
divi
dual
ized
treat
men
t, wi
th s
ubst
antia
l ris
k of
con
foun
ding
by
indi
catio
n (a
s we
ll as
sel
ectio
n bi
as).
c Odd
s ra
tio a
djus
ted
for a
ge, H
IV s
tatu
s, s
putu
m s
mea
r pos
itivi
ty, c
avita
tion
on c
hest
radi
ogra
ph, a
nd p
rior t
reat
men
t with
firs
t-lin
e an
d se
cond
-line
tb d
rugs
.d A
djus
tmen
t for
indi
vidu
al p
atie
nt c
hara
cter
istic
s no
t pos
sibl
e; th
e ad
just
ed v
alue
s of
the
pool
ed p
ropo
rtion
s (w
ith th
eir 9
5% C
l) s
hown
in fo
otno
tes
belo
w.e I
n 20
stu
dies
the
patie
nts
were
give
n st
anda
rdize
d re
gim
ens,
but
in th
e re
mai
ning
stu
dies
ther
apy
was
indi
vidu
alize
d, le
adin
g to
risk
of c
onfo
undi
ng b
y in
dica
tion.
f Adj
uste
d pr
opor
tion:
91%
(95%
Cl:
85%
–95%
).g A
djus
ted
prop
ortio
n: 9
2% (9
5% C
l: 8
7%–9
6%).
h Adj
uste
d pr
opor
tion:
80%
(95%
Cl:
74%
–85%
).i A
djus
ted
prop
ortio
n: 7
8% (9
5% C
l: 7
4%–8
5%).
j ser
ious
adv
erse
eve
nts
(sAE
s) re
porte
d in
pat
ient
s we
re a
ttrib
uted
to a
med
icin
e by
the
auth
ors
who
were
unb
linde
d an
d us
ed n
on-s
tand
ardi
zed
met
hods
to d
efine
, asc
erta
in a
nd re
port
sAEs
. No
valid
com
paris
ons
are
poss
ible
with
pat
ient
s no
t on
the
targ
et m
edic
ine,
bec
ause
sAE
s in
thes
e pa
tient
s co
uld
be d
ue to
oth
er d
rugs
rece
ived.
k Poo
led
prop
ortio
n: F
E 95
% C
l: 0
.6%
–2.4
%.
l Poo
led
prop
ortio
n: F
E 95
% C
l: 1
.9%
–4.1
%.
14
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): B
asto
s M, L
an Z
, Men
zies
R (1
1 N
ovem
ber 2
015)
Que
stio
n: G
atifl
oxac
in co
mpa
red
to n
o ga
tiflox
acin
for t
he tr
eatm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B or
MD
R-T
B
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith r
ifam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing
conv
entio
nal r
egim
ens l
astin
g ab
out 2
4 m
onth
s and
shor
ter
MD
R-T
B re
gim
ens,
in lo
w an
d hi
gh re
sour
ce se
tting
s, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are
Bib
liogr
aphy
: (1)
Van
Deu
n A
, Mau
g A
KJ,
Salim
MA
H, D
as P
K, S
arke
r MR
, Dar
u P,
et a
l. Sh
ort,
high
ly e
ffect
ive,
and
inex
pens
ive
stand
ardi
zed
trea
tmen
t of m
ultid
rug-
resis
tant
tube
rcul
osis.
Am
J R
espi
r Crit
Car
e Med
. 201
0;18
2(5)
:684
–92.
(2) B
utov
DA
, Efre
men
ko Y
V, P
rihod
a ND
, Yur
chen
ko
LI, S
okol
enko
NI,
Arja
nova
OV,
et al
. Adj
unct
imm
une t
hera
py o
f firs
t-dia
gnos
ed T
B, re
laps
ed T
B, tr
eatm
ent-f
aile
d T
B, m
ultid
rug-
resis
tant
TB
and
TB/
HIV
. Im
mun
othe
rapy
201
2;4(
7):6
87–6
95. (
3) X
u H
B, Ji
ang R
H, X
iao
HP.
Clo
fazi
min
e in
the t
reat
men
t of m
ultid
rug-
resis
tant
tube
rcul
osis.
Clin
M
icro
biol
Infe
ct. 2
012;
18(1
1):1
104–
1110
. (4)
Xu
HB,
Jian
g RH
, Li L
, Xia
o H
P. L
inez
olid
in th
e tre
atm
ent o
f MD
R-T
B: a
retr
ospe
ctiv
e clin
ical
stud
y. In
t J T
uber
c Lun
g Dis.
201
2;16
(3):3
58–3
63. (
5) C
arro
ll M
W, L
ee M
, Cai
Y, H
alla
han
CW
, Sha
w P
A, M
in JH
, et a
l. Fr
eque
ncy o
f adv
erse
reac
tions
to
first-
and
seco
nd-li
ne an
ti-tu
berc
ulos
is ch
emot
hera
py in
a K
orea
n co
hort
. Int
J Tu
berc
Lun
g Dis.
201
2;16
(7):9
61–9
66. (
6) Ja
wah
ar M
S, B
anur
ekha
VV,
Pa
ram
asiv
an C
N, R
ahm
an F
, Ram
acha
ndra
n R
, Ven
kate
san
P, e
t al.
Ran
dom
ized
clin
ical
tria
l of t
hric
e-w
eekl
y 4-
mon
th m
oxifl
oxac
in o
r gat
iflox
acin
co
ntai
ning
regi
men
s in
the t
reat
men
t of n
ew sp
utum
pos
itive
pul
mon
ary t
uber
culo
sis p
atie
nts.
PLoS
One
201
3;8(
7):e
6703
0. (7
) Jo
KW
, Lee
SD
, Kim
W
S, K
im D
S, S
him
TS.
Tre
atm
ent o
utco
mes
and
mox
iflox
acin
susc
eptib
ility
in o
floxa
cin-
resis
tant
mul
tidru
g-re
sista
nt tu
berc
ulos
is. In
t J T
uber
c Lun
g D
is. 2
014:
18(1
):39–
43. (
8) R
usto
mje
e R, L
ienh
ardt
C, K
anyo
k T,
Dav
ies G
R, L
evin
J, M
thiy
ane T
, et a
l. A
Pha
se II
stud
y of t
he st
erili
sing a
ctiv
ities
of
oflox
acin
, gat
iflox
acin
and
mox
iflox
acin
in p
ulm
onar
y tub
ercu
losis
. Int
J Tu
berc
Lun
g Dis.
200
8;12
(2):1
28–1
38.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
GATI
FLOX
ACIN
NO
GATI
FLOX
ACIN
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
(ass
esse
d wi
th: V
an D
eun
2010
; but
ov 2
011;
Xu
2012
a, 2
012b
)a
4 ob
serv
atio
nal
stud
ies
very
se
rious
bse
rious
no
t ser
ious
se
rious
st
rong
ass
ocia
tion
189/
225
(84.
0%)
174/
268
(64.
9%)
191
mor
e pe
r 1,0
00
(116
mor
e to
265
m
ore)
VE
RY l
OW
CRIt
ICAl
Deat
h ve
rsus
all
othe
r out
com
es (a
sses
sed
with
: Van
Deu
n 20
10, b
utov
201
1, X
u 20
12a,
201
2b)a
4 ob
serv
atio
nal
stud
ies
very
se
rious
bse
rious
no
t ser
ious
se
rious
no
ne
6/22
5 (2
.7%
) 23
/268
(8.6
%)
59 fe
wer
pe
r 1,0
00
(20
fewe
r to
99
fewe
r)
VE
RY l
OW
CRIt
ICAl
1515
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
GATI
FLOX
ACIN
NO
GATI
FLOX
ACIN
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
serio
us a
dver
se e
vent
s (G
rade
3 o
r 4, o
r dru
gs s
topp
ed d
ue to
adv
erse
eve
nts)
in p
atie
nts
on g
atifl
oxac
in v
ersu
s no
gat
iflox
acin
(ass
esse
d wi
th: c
ompa
rativ
e ob
serv
atio
nal s
tudi
es: C
arol
l 201
2; Ja
waha
r 20
13; J
o 20
14; R
usto
mje
e 20
08; V
an D
eun
2010
)a
5 ob
serv
atio
nal
stud
ies
very
se
rious
se
rious
no
t ser
ious
se
rious
no
nec
15/4
22
(3.6
%)d
137/
1711
(8
.0%
)eno
t es
timab
lec
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; F
E: fi
xed
effe
cts
a In
the
no g
atifl
oxac
in g
roup
the
othe
r fluo
roqu
inol
one
used
was
eith
er o
floxa
cin,
levo
floxa
cin
or m
oxifl
oxac
in.
b sm
all o
bser
vatio
nal s
tudi
es u
sing
indi
vidu
alize
d re
gim
ens
with
sub
stan
tial p
oten
tial f
or b
ias;
in th
e Va
n De
un, e
t al.
stud
y ga
tiflox
acin
was
use
d as
par
t of s
horte
r MDR
-tb
regi
men
s re
serv
ed fo
r pat
ient
s se
lect
ed
upon
spe
cific
crit
eria
.c s
erio
us a
dver
se e
vent
s (s
AEs)
repo
rted
in p
atie
nts
were
attr
ibut
ed to
a m
edic
ine
by th
e au
thor
s wh
o we
re u
nblin
ded
and
used
non
-sta
ndar
dize
d m
etho
ds to
defi
ne, a
scer
tain
and
repo
rt sA
Es. N
o va
lid c
ompa
rison
s ar
e po
ssib
le w
ith p
atie
nts
not o
n th
e ta
rget
med
icin
e, b
ecau
se s
AEs
in th
ese
patie
nts
coul
d be
due
to o
ther
dru
gs re
ceive
d.d P
oole
d pr
opor
tion:
FE
95%
Cl:
2.0
%–5
.8%
.e P
oole
d pr
opor
tion:
FE
95%
Cl:
6.8
%–9
.4%
.
16
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): B
asto
s M, L
an Z
, Men
zies
R (1
1 N
ovem
ber 2
015)
Que
stio
n: A
seco
nd-li
ne in
ject
able
com
pare
d to
no
seco
nd li
ne in
ject
able
for a
dults
with
rifa
mpi
cin-
resis
tant
TB
or M
DR
-TBa
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith r
ifam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing
conv
entio
nal r
egim
ens l
astin
g ab
out 2
4 m
onth
s and
shor
ter
MD
R-T
B re
gim
ens,
in lo
w an
d hi
gh re
sour
ce se
tting
s, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are
Bib
liogr
aphy
: (1)
Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug
resis
tant
pul
mon
ary
tube
rcul
osis
trea
tmen
t re
gim
ens a
nd p
atie
nt o
utco
mes
: an
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
is of
9,1
53 p
atie
nts.
PLoS
Med
. 201
2;9(
8):e
1001
300.
(2) B
asto
s M, L
an Z
, M
enzi
es R
. An
upda
ted
syst
emat
ic re
view
and
met
a-an
alys
is fo
r tre
atm
ent o
f mul
tidru
g-re
sista
nt tu
berc
ulos
is, 2
016
(und
er re
view
, 28
May
201
6).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
A SE
COND
-LIN
E IN
JECT
ABLE
NO S
ECON
D LI
NE
INJE
CTAB
LERE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on k
anam
ycin
or a
mik
acin
, as
part
of a
MDR
-tb
regi
men
(ass
esse
d wi
th: i
ndivi
dual
pat
ient
dat
a m
eta-
anal
ysis
(Ahu
ja s
D, e
t al.
PlOs
Med
. 201
2)
32
obse
rvat
iona
l st
udie
s se
rious
bno
t ser
ious
no
t ser
ious
no
t ser
ious
al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
2572
/346
7 (7
4.2%
) 55
7/98
1 (5
6.8%
) aO
R 1.
6 (1
.2 to
2.
0)c
170
mor
e pe
r 1,0
00
(from
55
mor
e to
28
0 m
ore)
lO
W
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on c
apre
omyc
in, a
s pa
rt of
a M
DR-t
b re
gim
en (a
sses
sed
with
: ind
ivid
ual p
atie
nt d
ata
met
a-an
alys
is (A
huja
sD,
et a
l. Pl
Os M
ed. 2
012)
32
obse
rvat
iona
l st
udie
s se
rious
bno
t ser
ious
no
t ser
ious
no
t ser
ious
al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
733/
1018
(7
2.0%
) 55
7/98
1 (5
6.8%
) aO
R 1.
3 (0
.5 to
3.
7)c
150
mor
e pe
r 1,0
00
(from
75
fewe
r to
310
mor
e)
lO
W
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se in
pat
ient
s on
kan
amyc
in o
r am
ikac
in, a
s pa
rt of
a M
DR-t
b re
gim
en (a
sses
sed
with
: agg
rega
ted
data
met
a-an
alys
is 2
015)
d
43
obse
rvat
iona
l st
udie
s se
rious
eno
t ser
ious
no
t ser
ious
no
t ser
ious
no
nef
3336
/393
5 (8
4.8%
)g,h
3378
/394
2 (8
5.7%
)g,i
not
estim
able
36 m
ore
per 1
,000
(fr
om 3
8 fe
wer t
o 11
0 m
ore)
VE
RY l
OW
CRIt
ICAl
1717
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
A SE
COND
-LIN
E IN
JECT
ABLE
NO S
ECON
D LI
NE
INJE
CTAB
LERE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on k
anam
ycin
or a
mik
acin
, as
part
of a
MDR
-tb
regi
men
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)
43
obse
rvat
iona
l st
udie
s se
rious
eno
t ser
ious
no
t ser
ious
no
t ser
ious
no
nef
3336
/474
1 (7
0.4%
)g,j
3378
/428
2 (7
8.9%
)g,k
not
estim
able
21 m
ore
per 1
,000
(fr
om 9
0 fe
wer t
o 13
1 m
ore)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se in
pat
ient
s on
cap
reom
ycin
ver
sus
no o
ther
sec
ond-
line
inje
ctab
le d
rug,
as
part
of a
MDR
-tb
regi
men
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)
43
obse
rvat
iona
l st
udie
s se
rious
eno
t ser
ious
no
t ser
ious
no
t ser
ious
no
nef
3960
/465
8 (8
5.0%
)l27
54/3
219
(85.
6%)m
not
estim
able
5 fe
wer
pe
r 1,0
00
(from
73
fewe
r to
62 m
ore)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on c
apre
omyc
in v
ersu
s no
oth
er s
econ
d-lin
e in
ject
able
dru
g, a
s pa
rt of
a M
DR-t
b re
gim
en (a
sses
sed
with
: agg
rega
ted
data
met
a-an
alys
is 2
015)
43
obse
rvat
iona
l st
udie
s se
rious
eno
t ser
ious
no
t ser
ious
no
t ser
ious
no
nef
3960
/514
1 (7
7.0%
)n27
54/3
882
(70.
9%)o
not
estim
able
69 m
ore
per 1
,000
(fr
om 3
1 fe
wer t
o 16
8 m
ore)
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s (G
rade
3 o
r 4, o
r dru
gs s
topp
ed d
ue to
adv
erse
eve
nts)
in p
atie
nts
on a
mik
acin
, cap
rem
ycin
or k
anam
ycin
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)
19
obse
rvat
iona
l st
udie
s se
rious
fno
t ser
ious
no
t ser
ious
no
t ser
ious
no
nep
184/
2538
(7
.2%
)q-
not
estim
able
p-
VE
RY l
OW
CRIt
ICAl
Cls:
con
fiden
ce li
mits
; FE:
fixe
d ef
fect
sa I
n th
is a
naly
sis,
the
use
of a
spe
cific
inje
ctab
le a
gent
(am
ikac
in, k
anam
ycin
or c
apre
omyc
in) i
s co
mpa
red
with
no
use
of th
at p
artic
ular
age
nt, a
lthou
gh a
noth
er s
econ
d-lin
e in
ject
able
age
nt m
ay h
ave
been
use
d as
pa
rt of
the
MDR
-tb
regi
men
.b I
ndiv
idua
l pat
ient
dat
a ta
ken
from
32
obse
rvat
iona
l stu
dies
in w
hich
mos
t pat
ient
s re
ceive
d in
divi
dual
ized
treat
men
t. Ri
sk o
f sel
ectio
n bi
as, a
nd c
onfo
undi
ng b
y in
dica
tion.
c aOR
: Odd
s ra
tio a
djus
ted
for a
ge, H
IV, p
ositi
vity
on
sput
um-s
mea
r mic
rosc
opy,
ches
t rad
iogr
aph
cavi
tatio
n, a
nd p
rior t
reat
men
t with
firs
t-lin
e an
d se
cond
-line
tb d
rugs
.d I
n th
e ag
greg
ated
dat
a m
eta-
anal
ysis
pat
ient
s wi
th X
DR-t
b we
re e
xclu
ded
wher
e po
ssib
le.
e In
tota
l, 61
coh
orts
pro
vide
d en
d-of
-trea
tmen
t out
com
e in
form
atio
n: in
23
coho
rts th
e pa
tient
s we
re g
iven
stan
dard
ized
regi
men
s an
d in
38
coho
rts th
erap
y wa
s in
divi
dual
ized,
lead
ing
to ri
sk o
f con
foun
ding
by
indi
catio
n. O
f the
61
coho
rts, 1
8 co
horts
did
not
spe
cify
whi
ch s
econ
d-lin
e in
ject
able
age
nt w
as u
sed,
and
ther
efor
e on
ly th
e re
mai
ning
43
coho
rts w
ere
reta
ined
for t
his
anal
ysis
.f P
oten
tial c
onfo
undi
ng fr
om p
refe
rent
ial i
nclu
sion
of c
apre
omyc
in in
the
indi
vidu
alize
d re
gim
ens
of p
atie
nts
with
mor
e ad
vanc
ed re
sist
ance
pat
tern
s or
dis
ease
.g G
iven
that
am
ikac
in o
r kan
amyc
in w
ere
used
in a
lmos
t all
stud
ies,
the
com
paris
on is
mad
e be
twee
n st
udie
s in
whi
ch 7
2% –1
00%
of p
atie
nts
rece
ived
the
inje
ctab
le a
gent
(int
erve
ntio
n gr
oup)
ver
sus
a co
mpa
rato
r gr
oup
of s
tudi
es in
whi
ch 0
%–7
1% o
f pat
ient
s re
ceive
d on
e of
thes
e ag
ents
.h A
djus
ted
prop
ortio
n: 9
4% (9
5% C
l: 9
0%–9
7%).
i Adj
uste
d pr
opor
tion:
89%
(95%
Cl:
83%
–96%
).
18
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
j Adj
uste
d pr
opor
tion:
82%
(95%
Cl:
75%
–88%
).k A
djus
ted
prop
ortio
n: 7
8% (9
5% C
l: 7
0%–8
6%).
l Adj
uste
d pr
opor
tion:
92%
(95%
Cl:
87%
–97%
).m A
djus
ted
prop
ortio
n: 9
3% (9
5% C
l: 8
6%–9
7%).
n Adj
uste
d pr
opor
tion:
77%
(95%
Cl:
69%
–84%
).o A
djus
ted
prop
ortio
n: 8
3% (9
5% C
l: 7
6%–8
9%).
p ser
ious
adv
erse
eve
nts
(sAE
s) re
porte
d in
pat
ient
s we
re a
ttrib
uted
to a
med
icin
e by
the
auth
ors
who
were
unb
linde
d an
d us
ed n
on-s
tand
ardi
zed
met
hods
to d
efine
, asc
erta
in a
nd re
port
sAEs
. No
valid
com
paris
ons
are
poss
ible
with
pat
ient
s no
t on
the
targ
et m
edic
ine,
bec
ause
sAE
s in
thes
e pa
tient
s co
uld
be d
ue to
oth
er d
rugs
rece
ived.
q Poo
led
prop
ortio
n: F
E 95
% C
l: 6
.2%
–8.4
%.
1919
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): M
enzi
es R
, Bas
tos M
, Lan
Z (1
1 N
ovem
ber 2
015)
Que
stio
n: E
thio
nam
ide/
prot
hion
amid
e com
pare
d to
no
ethi
onam
ide/
prot
hion
amid
e for
adul
ts w
ith ri
fam
pici
n-re
sista
nt T
B or
MD
R-T
B
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith r
ifam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing
conv
entio
nal r
egim
ens l
astin
g ab
out 2
4 m
onth
s and
shor
ter
MD
R-T
B re
gim
ens,
in lo
w an
d hi
gh re
sour
ce se
tting
s, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are
Bib
liogr
aphy
: (1)
Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug
resis
tant
pul
mon
ary
tube
rcul
osis
trea
tmen
t re
gim
ens a
nd p
atie
nt o
utco
mes
: an
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
is of
9,1
53 p
atie
nts.
PLoS
Med
. 201
2;9(
8):e
1001
300.
(2) B
asto
s M, L
an Z
, M
enzi
es R
. An
upda
ted
syst
emat
ic re
view
and
met
a-an
alys
is fo
r tre
atm
ent o
f mul
tidru
g-re
sista
nt tu
berc
ulos
is, 2
016
(und
er re
view
, 28
May
201
6).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
ETHI
ONAM
IDE/
PR
OTHI
ONA-
MID
E
NO
ETHI
ONAM
IDE/
PR
OTHI
ONA-
MID
E RE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on e
thio
nam
ide/
prot
hion
amid
e as
par
t of a
MDR
-tb
regi
men
(ass
esse
d wi
th: i
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
(Ahu
ja s
D, e
t al.
PlOs
M
ed. 2
012)
)a
32
obse
rvat
iona
l st
udie
s se
rious
bno
t ser
ious
no
t ser
ious
no
t ser
ious
al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
4101
/566
7 (7
2.4%
) 87
8/14
87
(59.
0%)
aOR
1.9
(1.5
to
2.3)
c
130
few
er
per 1
000
(from
65
mor
e to
18
5 m
ore)
lO
W
CRIt
ICAl
serio
us a
dver
se e
vent
s (G
rade
3 o
r 4, o
r dru
gs s
topp
ed d
ue to
adv
erse
eve
nts)
in p
atie
nts
on e
thio
nam
ide/
prot
hion
amid
e as
par
t of a
MDR
-tb
regi
men
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)
17
obse
rvat
iona
l st
udie
s se
rious
no
t ser
ious
no
t ser
ious
no
t ser
ious
no
ned
173/
2106
(8
.2%
)e-
not
estim
able
d
VERY
lOW
CRIt
ICAl
Cl: c
onfid
ence
lim
it; F
E: fi
xed
effe
cts
a In
this
ana
lysi
s, u
se o
f eth
iona
mid
e is
com
bine
d wi
th p
roth
iona
mid
e, a
nd c
ompa
red
to re
sults
in p
atie
nts
who
did
not g
et e
ither
of t
hese
dru
gs, b
ut re
ceive
d m
ultip
le o
ther
dru
gs.
b thi
s is
indi
vidu
al p
atie
nt d
ata
take
n fro
m 3
2 ob
serv
atio
nal s
tudi
es in
whi
ch m
ost p
atie
nts
rece
ived
indi
vidu
alize
d tre
atm
ent.
ther
e is
risk
of s
elec
tion
bias
and
con
foun
ding
by
indi
catio
n.c a
OR: O
dds
ratio
adj
uste
d fo
r age
, HIV
, aci
d fa
st b
acill
us s
mea
r, ch
est r
adio
grap
h ca
vita
tion,
and
prio
r tre
atm
ent w
ith fi
rst-l
ine,
and
sec
ond-
line
tb d
rugs
.d s
erio
us a
dver
se e
vent
s (s
AEs)
repo
rted
in p
atie
nts
were
attr
ibut
ed to
a m
edic
ine
by th
e au
thor
s wh
o we
re u
nblin
ded
and
used
non
-sta
ndar
dize
d m
etho
ds to
defi
ne, a
scer
tain
and
repo
rt sA
Es. N
o va
lid c
ompa
rison
s ar
e po
ssib
le w
ith p
atie
nts
not o
n th
e ta
rget
med
icin
e, b
ecau
se s
AEs
in th
ese
patie
nts
coul
d be
due
to o
ther
dru
gs re
ceive
d.e P
oole
d pr
opor
tion:
FE
95%
Cl:
7.0%
–9.6
%.
20
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): M
enzi
es R
, Bas
tos M
, Lan
Z (1
1 N
ovem
ber 2
015)
Que
stio
n: C
yclo
serin
e/te
rizid
one c
ompa
red
to n
o cy
clos
erin
e/te
rizid
one f
or ad
ults
with
rifa
mpi
cin-
resis
tant
TB
or M
DR
-TB
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith M
DR
-TB
regi
men
s, in
low
and
high
reso
urce
sett
ings
, with
in h
ospi
tal o
r am
bula
tory
mod
els o
f car
e
Bib
liogr
aphy
: (1)
Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug
resis
tant
pul
mon
ary
tube
rcul
osis
trea
tmen
t re
gim
ens a
nd p
atie
nt o
utco
mes
: an
indi
vidu
al p
atie
nt d
ata m
eta-
anal
ysis
of 9
,153
pat
ient
s. PL
oS M
ed. 2
012;
9(8)
:e10
0130
0. (2
) Hw
ang T
J, W
ares
DF,
Ja
faro
v A, J
akub
owia
k W
, Nun
n P,
Kes
havj
ee S
. Saf
ety o
f cyc
lose
rine a
nd te
rizid
one f
or th
e tre
atm
ent o
f dru
g-re
sista
nt tu
berc
ulos
is: a
met
a-an
alys
is. In
t J
Tube
rc L
ung D
is. 2
013;
17(1
0):1
257–
66. (
3) B
asto
s M, L
an Z
, Men
zies
R. A
n up
date
d sy
stem
atic
revi
ew an
d m
eta-
anal
ysis
for t
reat
men
t of m
ultid
rug-
resis
tant
tube
rcul
osis,
201
6 (u
nder
revi
ew, 2
8 M
ay 2
016)
.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NSCY
CLOS
ERIN
E/
TERI
ZIDO
NE
NO
CYCL
OSER
INE/
TE
RIZI
DONE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
succ
ess
vers
us fa
ilure
/rel
apse
/dea
th fo
r cyc
lose
rine
and
teriz
idon
e fro
m In
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
(Ahu
ja s
D, e
t al.
PlOs
Med
201
2)a
32
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
3115
/424
0 (7
3.5%
) 18
64/2
914
(64.
0%)
OR 1
.5
(0.9
to
2.2)
a
95 m
ore
per 1
,000
(fr
om 7
3 m
ore
to
117
mor
e)
VE
RY l
OW
CRIt
ICAl
succ
ess
vers
us fa
ilure
/rel
apse
for c
yclo
serin
e an
d te
rizid
one
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)
53ob
serv
atio
nal
stud
ies
serio
us
serio
us
not s
erio
us
serio
us
none
44
74/5
285
(84.
7%)2
1969
/247
9 (7
9.4%
)3no
t es
timab
le49
mor
e pe
r 1,0
00
(from
56
fewe
r to
155
mor
e)
VE
RY l
OW
CRIt
ICAl
succ
ess
vers
us fa
ilure
/rel
apse
/dea
th fo
r cyc
lose
rine
and
teriz
idon
e (a
sses
sed
with
: agg
rega
ted
data
met
a-an
alys
is 2
015)
53ob
serv
atio
nal
stud
ies
serio
us
serio
us
not s
erio
us
serio
us
none
44
74/5
916
(75.
6%)4
1969
/282
3 (6
9.7%
)5no
t es
timab
le5
few
er
per 1
,000
(fr
om 1
39
fewe
r to
129
mor
e)
VE
RY l
OW
CRIt
ICAl
2121
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NSCY
CLOS
ERIN
E/
TERI
ZIDO
NE
NO
CYCL
OSER
INE/
TE
RIZI
DONE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
Drug
dis
cont
inue
d du
e to
maj
or p
sych
iatri
c to
xicity
from
cyc
lose
rine
used
to tr
eat M
DR-t
b (a
sses
sed
with
: Hwa
ng, e
t al.
Int J
tube
rc l
ung
Dis.
201
2 (s
yste
mat
ic re
view
))b
26
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
se
rious
no
nec
144/
1923
(7
.5%
) -
not
estim
able
c
VERY
lOW
CRIt
ICAl
Drug
dis
cont
inue
d du
e to
toxic
ity (a
ll ty
pes)
from
cyc
lose
rine
used
to tr
eat M
DR-t
b (a
sses
sed
with
: Hwa
ng tJ
, et a
l. In
t J tu
berc
lun
g Di
s. 2
012
(sys
tem
atic
revi
ew))
b
27
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
se
rious
no
nec
201/
2164
(9
.3%
) -
not
estim
able
c
VERY
lOW
CRIt
ICAl
serio
us a
dver
se e
vent
s (G
rade
3 o
r 4, o
r dru
gs s
topp
ed d
ue to
adv
erse
eve
nts)
in p
atie
nts
on c
yclo
serin
e as
par
t of a
MDR
-tb
regi
men
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)
16
obse
rvat
iona
l st
udie
s se
rious
no
t ser
ious
no
t ser
ious
no
t ser
ious
no
nec
96/2
140
(4.5
%)d
-no
t es
timab
lec
VE
RY l
OW
CRIt
ICAl
Drug
dis
cont
inue
d du
e to
toxic
ity (a
ll ty
pes)
from
teriz
idon
e us
ed to
trea
t MDR
-tb
(ass
esse
d wi
th: H
wang
tJ, e
t al.
Int J
tube
rc l
ung
Dis.
201
2 (s
yste
mat
ic re
view
))b
10
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
se
rious
no
nec
111/
707
(15.
7%)
-no
t es
timab
lec,
e
VERY
lOW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; F
E: fi
xed
effe
cts;
OR:
odd
s ra
tioa A
djus
ted
for a
ge, e
xten
t of d
isea
se, H
IV, a
nd p
rior t
reat
men
t with
firs
t-lin
e or
sec
ond-
line
tb d
rugs
. Pat
ient
s on
cyc
lose
rine
and
teriz
idon
e we
re c
ombi
ned
toge
ther
for t
his
anal
ysis
.b N
o re
gion
al d
iffer
ence
s ob
serv
ed.
c ser
ious
adv
erse
eve
nts
(sAE
s) re
porte
d in
pat
ient
s we
re a
ttrib
uted
to a
med
icin
e by
the
auth
ors
who
were
unb
linde
d an
d us
ed n
on-s
tand
ardi
zed
met
hods
to d
efine
, asc
erta
in a
nd re
port
sAEs
. No
valid
com
paris
ons
are
poss
ible
with
pat
ient
s no
t on
the
targ
et m
edic
ine,
bec
ause
sAE
s in
thes
e pa
tient
s co
uld
be d
ue to
oth
er d
rugs
rece
ived.
d Poo
led
prop
ortio
n: F
E 95
% C
l: 3
.6%
–5.5
%.
e ter
izido
ne a
nd c
yclo
serin
e we
re c
ompa
red
in th
ree
of th
e st
udie
s. A
utho
rs re
porte
d no
diff
eren
ces
and
conc
lude
d th
at th
e ef
fect
of t
erizi
done
var
ied
from
not
bei
ng d
iffer
ent t
o be
ing
mod
erat
ely
bette
r tha
n cy
clos
erin
e.
22
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): M
enzi
es R
, Bas
tos M
, Lan
Z (1
1 N
ovem
ber 2
015)
Que
stio
n: L
inez
olid
com
pare
d to
no
linez
olid
for a
dult
patie
nts o
n tr
eatm
ent f
or M
DR
-TB/
XD
R-T
B
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B/M
DR
-/X
DR
-TB
usin
g co
nven
tiona
l reg
imen
s las
ting
abou
t 24
mon
ths,
in lo
w a
nd h
igh
reso
urce
setti
ngs,
with
in h
ospi
tal o
r am
bula
tory
mod
els o
f car
e
Bib
liogr
aphy
: (1)
Alte
t MN
, Vid
al R
, Milá
C, R
odrig
o T,
Cas
als M
, Mir
I, et
al. M
onito
ring c
hang
es in
anti-
tube
rcul
osis
trea
tmen
t: as
soci
ated
fact
ors
dete
rmin
ed at
the t
ime o
f dia
gnos
is. In
t J T
uber
c Lun
g Dis.
201
3;17
(11)
:143
5–41
. (2)
Car
roll
MW
, Lee
M, C
ai Y
, Hal
laha
n C
W, S
haw
PA
, Min
JH, e
t al
. Fre
quen
cy o
f adv
erse
reac
tions
to fi
rst-
and
seco
nd-li
ne an
ti-tu
berc
ulos
is ch
emot
hera
py in
a K
orea
n co
hort
. Int
J Tu
berc
Lun
g Dis.
201
2;16
(7):9
61–
966.
(3)
De
Lore
nzo
S, A
lffen
aar
JW, S
otgi
u G
, Cen
tis R
, D’A
mbr
osio
L, T
iber
i S, e
t al
. Effi
cacy
and
safe
ty o
f mer
open
em-c
lavul
anat
e ad
ded
to
linez
olid
-con
tain
ing
regi
men
s in
the
trea
tmen
t of M
DR
-/X
DR
-TB.
Eur
Res
pir J
. 201
3;41
(6):1
386–
92. (
4) Ji
ang
R-H
, Xu
H-B
, Li L
. Com
para
tive
role
s of m
oxifl
oxac
in an
d le
voflo
xaci
n in
the t
reat
men
t of p
ulm
onar
y mul
tidru
g-re
sista
nt tu
berc
ulos
is: a
retr
ospe
ctiv
e stu
dy. I
nt J
Ant
imic
rob
Age
nts.
2013
;42(
1):3
6–41
. (5)
Koh
W-J,
Kw
on O
J, G
wak
H, C
hung
JW
, Cho
S-N
, Kim
WS,
et
al. D
aily
300
mg
dose
of l
inez
olid
for
the
trea
tmen
t of
in
trac
tabl
e m
ultid
rug-
resis
tant
and
ext
ensiv
ely
drug
-resis
tant
tube
rcul
osis.
J A
ntim
icro
b C
hem
othe
r. 20
09;6
4(2)
:388
–91.
(6) L
ee M
, Lee
J, C
arro
ll M
W, C
hoi H
, Min
S, S
ong T
, et a
l. Li
nezo
lid fo
r Tre
atm
ent o
f Chr
onic
Ext
ensiv
ely D
rug-
Res
istan
t Tub
ercu
losis
. N E
ngl J
Med
. 201
2;36
7(16
):150
8–18
. (7)
Mig
none
F, C
odec
asa L
R, S
colfa
ro C
, Raff
aldi
I, L
ance
lla L
, Fer
rare
se M
, et a
l. Th
e spr
ead
of d
rug-
resis
tant
tube
rcul
osis
in ch
ildre
n: an
Ital
ian
case
serie
s. Ep
idem
iol I
nfec
t. 20
14;1
42(1
0):2
049–
56. (
8) P
aday
atch
i N, M
ac K
enzi
e WR
, Hirs
ch-M
over
man
Y, F
eng P
-J, V
illar
ino
E, S
aukk
onen
J, et
al.
Less
ons f
rom
a ra
ndom
ised
clin
ical
tria
l for
mul
tidru
g-re
sista
nt tu
berc
ulos
is. In
t J T
uber
c Lun
g Dis.
201
2;16
(12)
:158
2–7.
(9) S
ingl
a R, C
amin
ero
JA,
Jaisw
al A
, Sin
gla N
, Gup
ta S
, Bal
i RK
, et a
l. Li
nezo
lid: a
n eff
ectiv
e, sa
fe an
d ch
eap
drug
for p
atie
nts f
ailin
g mul
tidru
g-re
sista
nt tu
berc
ulos
is tr
eatm
ent
in In
dia.
Eur
Res
pir J
. 201
2;39
(4):9
56–9
62. (
10) S
chec
ter G
F, S
cott
C, T
rue L
, Raft
ery
A, F
lood
J, M
ase S
. Lin
ezol
id in
the t
reat
men
t of m
ultid
rug-
resis
tant
tube
rcul
osis.
Clin
Infe
ct D
is. 2
010;
50(1
):49–
55. (
11) T
ang
S, Y
ao L
, Hao
X, Z
hang
X, L
iu G
, Liu
X, e
t al.
Effica
cy, s
afet
y an
d to
lerab
ility
of
line
zolid
for t
he tr
eatm
ent o
f XD
R-T
B: a
stud
y in
Chi
na. E
ur R
espi
r J. 2
015;
45(1
):161
–70.
(12)
Udw
adia
ZF,
Sen
T, M
ohar
il G
. Ass
essm
ent o
f lin
ezol
id effi
cacy
and
safe
ty in
MD
R- a
nd X
DR
-TB
: an
Indi
an p
ersp
ectiv
e. Eu
r Res
pir J
. 201
0;35
(4):9
36–9
38–9
40.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
LINE
ZOLI
DNO
LIN
EZOL
IDRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in X
DR-t
b pa
tient
s gi
ven
linez
olid
(ass
esse
d wi
th: R
Ct in
Chi
na, 2
009–
2011
(tan
g, e
t al,
2015
))a
1 ra
ndom
ized
trial
s se
rious
no
t ser
ious
no
t ser
ious
se
rious
st
rong
ass
ocia
tion
23/2
9 (7
9.3%
)b11
/29
(37.
9%)c
not
estim
able
414
mor
e pe
r 1,0
00
(from
184
m
ore
to
644
mor
e)
M
ODER
AtE
CRIt
ICAl
2323
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
LINE
ZOLI
DNO
LIN
EZOL
IDRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus.
failu
re/r
elap
se/d
eath
/def
ault
in M
DR-t
b or
XDR
-tb
patie
nts
give
n lin
ezol
id (a
sses
sed
with
: 1RC
t +
6 ob
serv
atio
nal s
tudi
es c
ombi
ned)
7 ob
serv
atio
nal
stud
iesd
very
se
rious
se
rious
no
t ser
ious
se
rious
no
ne
153/
198
(77.
3%)e
387/
606
(63.
9%)f
not
estim
able
134
mor
e pe
r 1,0
00
(from
64
mor
e to
20
4 m
ore)
VE
RY l
OW
CRIt
ICAl
Deat
h (v
ersu
s al
l oth
er o
utco
mes
) in
MDR
-tb
and
XDR-
tb p
atie
nts
give
n lin
ezol
id (a
sses
sed
with
: 1RC
t +
6 ob
serv
atio
nal s
tudi
es c
ombi
ned)
7 ob
serv
atio
nal
stud
iesd
very
se
rious
se
rious
no
t ser
ious
se
rious
no
ne
21/2
12
(9.9
%)
65/4
68
(13.
9%)
not
estim
able
40 fe
wer
pe
r 1,0
00
(91
fewe
r to
11 m
ore)
VE
RY l
OW
CRIt
ICAl
Grad
e 3–
4 se
rious
adv
erse
eve
nts
and/
or d
rugs
sto
pped
due
to li
nezo
lid (a
sses
sed
with
: int
erna
l com
para
tor g
roup
s)g,
h
4 ob
serv
atio
nal
stud
iesh,
ive
ry
serio
us
serio
us
not s
erio
us
serio
us
none
11
/49
(22.
4%)
112/
1305
(8
.6%
) no
t es
timab
le13
9 m
ore
per 1
,000
(2
1 m
ore
to
257
mor
e)
VE
RY l
OW
CRIt
ICAl
Grad
e 3–
4 se
rious
adv
erse
eve
nts
and/
or d
rugs
sto
pped
due
to li
nezo
lid 6
00 m
g/da
y (a
sses
sed
with
: lar
gely
unc
ontro
lled
obse
rvat
iona
l stu
dies
)j
8 ob
serv
atio
nal
stud
iesi,j
very
se
rious
se
rious
no
t ser
ious
se
rious
no
ne28
/190
(1
4.7%
)kno
t es
timab
le
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
it; R
Ct: r
ando
mize
d co
ntro
lled
trial
a Met
hod
of ra
ndom
izatio
n no
t des
crib
ed, h
ence
risk
of a
lloca
tion
bias
unk
nown
. stu
dy w
as n
ot b
linde
d, h
ence
risk
of a
scer
tain
men
t bia
s, a
nd s
mal
l num
ber o
f sub
ject
s.b 9
5% C
l: 6
5%–9
4%.
c 95%
Cl:
20%
–56%
.d A
ll we
re s
mal
l stu
dies
. the
1 R
Ct w
as v
ery
smal
l and
unb
linde
d wi
th u
ncle
ar ra
ndom
izatio
n. th
e 6
obse
rvat
iona
l had
indi
vidu
alize
d re
gim
ens.
e 95%
Cl:
73%
–84%
.f 9
5% C
l: 4
6%–9
0%.
g Not
sho
wing
the
effe
cts
in tw
o st
udie
s fo
r pat
ient
s re
ceiv
ing
1200
mg
per d
ay (9
/51;
18%
).h A
ltet 2
013;
Car
roll
2012
; Mig
none
201
4; P
aday
atch
i 201
2 (o
nly
Pada
yatc
hi re
porte
d th
e do
se).
i the
inte
rven
tion
grou
p wa
s gi
ven
linez
olid
at a
sta
rt do
se o
f 120
0 m
g pe
r day
for 4
–6 w
eeks
and
follo
wed
by a
dos
e of
300
–600
mg
per d
ay.
j Koh
200
9; s
chec
ter 2
010;
Udw
adia
201
0; s
ingl
a 20
12; l
ee 2
012;
De
lore
nzo
2013
; Jia
ng 2
013;
Pad
ayat
chi 2
012
(onl
y Pa
daya
tchi
repo
rted
sAE
in g
roup
not
rece
ivin
g lin
ezol
id; s
ingl
a (6
00 m
g vs
120
0 m
g) a
nd
De l
oren
zo (6
00 m
g vs
>60
0 m
g) c
ompa
red
sAE
at d
iffer
ent d
oses
).k 9
5% C
l: 1
0%–2
1%.
24
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): R
onal
d L,
Cer
igo
H, F
ox G
, Men
zies
R (1
1 N
ovem
ber 2
015)
Que
stio
n: C
lofa
zim
ine c
ompa
red
to n
o cl
ofaz
imin
e for
the t
reat
men
t of a
dults
with
rifa
mpi
cin-
resis
tant
TB
or M
DR
-TB
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith r
ifam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing
conv
entio
nal r
egim
ens l
astin
g ab
out 2
4 m
onth
s and
shor
ter
MD
R-T
B re
gim
ens,
in lo
w an
d hi
gh re
sour
ce se
ttin
gs, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are (
as w
ell a
s non
-tube
rcul
ous m
ycob
acte
ria (N
TM
) in
som
e out
com
es fo
r SA
E)
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
CLOF
AZIM
INE
NO
CLOF
AZIM
INE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in M
DR-t
b pa
tient
s on
clo
fazim
ine
(ass
esse
d wi
th: i
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
(201
0))a
31
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
not s
erio
us
none
45
9/80
6 (5
6.9%
)b32
92/4
970
(66.
2%)c
adju
sted
OR
1.4
(0
.4 to
4.
0)
10 m
ore
per 1
,000
(fr
om 2
20
fewe
r to
340
mor
e)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/d
eath
in n
on-X
DR M
DR-t
b pa
tient
s wi
th c
lofa
zimin
e in
thei
r reg
imen
(ass
esse
d wi
th: 1
RCt
201
0–20
11 (t
ang
s, e
t al.
2015
))
1 ra
ndom
ized
trial
s se
rious
dno
t ser
ious
eno
t ser
ious
se
rious
est
rong
ass
ocia
tion
39/4
9 (7
9.6%
)f28
/47
(59.
6%)g
not
estim
able
200
mor
e pe
r 1,0
00
(from
60
fewe
r to
450
mor
em
M
ODER
AtE
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
(ass
esse
d wi
th: 1
RCt
+ 5
coh
orts
of M
DR/X
DR p
atie
nts)
h
6 ob
serv
atio
nal
stud
iesi
very
se
rious
se
rious
no
t ser
ious
se
rious
no
ne
75/1
02
(73.
5%)j
68/9
2 (7
3.9%
)kno
t es
timab
le10
few
er
per 1
,000
(fr
om 2
10
fewe
r to
170
mor
e)
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s re
sulti
ng in
dru
g di
scon
tinua
tion
in M
DR-/
XDR-
tb p
atie
nts
on c
lofa
zimin
e (a
sses
sed
with
: com
para
tive
stud
ies)
l
5 ob
serv
atio
nal
stud
ies
very
se
rious
se
rious
no
t ser
ious
se
rious
no
ne
2/81
(2.5
%)
281/
658
(42.
7%)
not
estim
able
VE
RY l
OW
CRIt
ICAl
2525
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
CLOF
AZIM
INE
NO
CLOF
AZIM
INE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
serio
us a
dver
se e
vent
s re
sulti
ng in
dru
g di
scon
tinua
tion
in N
tM p
atie
nts
on c
lofa
zimin
e (a
sses
sed
with
: unc
ontro
lled
stud
ies)
l
6 ob
serv
atio
nal
stud
ies
very
se
rious
se
rious
se
rious
se
rious
no
ne
25/1
95
(12.
8%)
not
estim
able
VERY
lOW
CR
ItIC
Al
serio
us a
dver
se e
vent
s re
sulti
ng in
dru
g di
scon
tinua
tion
in N
tM p
atie
nts
on c
lofa
zimin
e (a
sses
sed
with
: com
para
tive
stud
ies
only
)l
4 ob
serv
atio
nal
stud
ies
very
se
rious
se
rious
se
rious
se
rious
no
ne
6/18
1 (3
.3%
) 15
/167
(9.0
%)
not
estim
able
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; R
E: ra
ndom
effe
cts
a Out
com
es w
ere
com
pare
d in
per
sons
who
rece
ived
clof
azim
ine
vers
us th
ose
who
rece
ived
no G
roup
5 d
rugs
. Adj
uste
d es
timat
e fro
m p
rope
nsity
sco
re m
atch
ing
was
done
, pat
ient
s wi
th c
lofa
zimin
e m
atch
ed to
pa
tient
s fro
m c
entre
s wh
ere
clof
azim
ine
was
not u
sed.
b RE
valu
e on
poo
led
met
a-an
alys
is: 6
3% (9
5% C
l: 4
9%–7
8%).
c RE
valu
e on
poo
led
met
a-an
alys
is: 6
2% (9
5% C
l: 4
5%–7
9%).
d Met
hod
of ra
ndom
izatio
n no
t des
crib
ed, a
nd n
o bl
indi
ng, i
ncre
asin
g ris
k of
allo
catio
n bi
as a
nd a
scer
tain
men
t bia
s.e O
ne s
tudy
in fi
ve c
entre
s in
one
cou
ntry
(Chi
na) o
nly.
f 95%
Cl:
68%
–91%
.g 9
5% C
l: 4
6%–7
4%.
h ben
efit w
as s
een
in o
ne R
Ct, b
ut in
5 s
mal
l obs
erva
tiona
l stu
dies
pat
ient
s re
ceiv
ing
clof
azim
ine
had
wors
e ou
tcom
es. t
hese
regi
men
s we
re in
divi
dual
ized
so th
ere
is ri
sk o
f bia
s (c
onfo
undi
ng b
y in
dica
tion)
.i o
ne ra
ndom
ized
cont
rol t
rial +
5 c
ohor
ts.
j Adj
uste
d pr
opor
tion
73%
; 95%
Cl:
64%
–82%
.k A
djus
ted
prop
ortio
n 89
%; 9
5% C
l: 7
3%–1
00%
.l A
dver
se e
vent
s re
porte
d in
pat
ient
s ta
king
clo
fazim
ine
were
attr
ibut
ed to
the
drug
by
auth
ors
who
were
unb
linde
d an
d us
ed n
on-s
tand
ardi
zed
met
hods
to d
efine
, asc
erta
in a
nd re
port
adve
rse
even
ts. N
o va
lid
com
paris
ons
are
poss
ible
with
pat
ient
s no
t tak
ing
clof
azim
ine,
bec
ause
adv
erse
eve
nts
in p
atie
nts
not r
ecei
ving
clo
fazim
ine
coul
d be
due
to o
ther
dru
gs re
ceive
d co
ncom
itant
ly.m
P=0
.04;
trea
tmen
t fai
lure
als
o si
gnifi
cant
ly lo
wer t
han
in c
ontro
l (11
% v
ersu
s 29
%; P
=0.0
3).
26
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): W
inte
rs N
, But
ler-L
apor
te G
, Men
zies
D (1
1 N
ovem
ber 2
015)
Que
stio
n: M
acro
lides
(cla
rithr
omyc
in, a
zith
rom
ycin
) com
pare
d to
no
mac
rolid
es fo
r tre
atm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B or
MD
R-T
B.
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing c
onve
ntio
nal r
egim
ens l
astin
g abo
ut 2
4 m
onth
s, in
low
and
high
re
sour
ce se
tting
s, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are (
as w
ell a
s non
-tube
rcul
ous m
ycob
acte
ria (N
TM
) in
som
e out
com
es fo
r SA
E)
Bib
liogr
aphy
: Win
ters
N, B
utler
-Lap
orte
G, M
enzi
es D
. Effi
cacy
and
safe
ty o
f Wor
ld H
ealth
Org
aniz
atio
n gr
oup
5 dr
ugs f
or m
ultid
rug-
resis
tant
tu
berc
ulos
is tr
eatm
ent.
Eur R
espi
r J. 2
015;
46(5
):146
1–70
.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
MAC
ROLI
DES
(CLA
RITH
RO-
MYC
IN,
AZIT
HROM
YCIN
)NO
M
ACRO
LIDE
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
in M
DR-t
b pa
tient
s on
cla
rithr
omyc
in (H
IV u
ninf
ecte
d)
2 ob
serv
atio
nal
stud
iesa
serio
us
not s
erio
us
not s
erio
us
serio
us
none
20
/61
(32.
8%)
59/1
91
(30.
9%)
not
estim
able
19
mor
e pe
r 1,0
00
(from
10
fewe
r to
11 m
ore)
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s in
NtM
pat
ient
s on
cla
rithr
omyc
in (H
IV u
ninf
ecte
d) (a
sses
sed
with
: ran
dom
ized
cont
rolle
d tri
als)
3 ra
ndom
ized
trial
s no
t se
rious
se
rious
se
rious
bse
rious
no
nec
31/1
74
(17.
8%)
26/1
75
(14.
9%)
not
estim
able
10
mor
e pe
r 1,0
00
(from
60
fewe
r to
70 m
ore)
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s in
NtM
pat
ient
s on
cla
rithr
omyc
in (H
IV u
ninf
ecte
d) (a
sses
sed
with
: unc
ontro
lled
coho
rts)
15
obse
rvat
iona
l st
udie
sdse
rious
ese
rious
se
rious
bno
t ser
ious
no
ne
41/6
15
(6.7
%)
-no
t es
timab
le
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s in
NtM
pat
ient
s on
cla
rithr
omyc
in (H
IV in
fect
ed) (
asse
ssed
with
: ran
dom
ized
cont
rolle
d tri
als)
8 ra
ndom
ized
trial
s no
t se
rious
no
t ser
ious
se
rious
bse
rious
no
nec,
f10
8/10
88
(9.9
%)
118/
1111
(1
0.6%
) no
t es
timab
le7
few
er
per 1
,000
(fr
om 2
0 fe
wer t
o 20
mor
e)
lO
W
CRIt
ICAl
2727
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
MAC
ROLI
DES
(CLA
RITH
RO-
MYC
IN,
AZIT
HROM
YCIN
)NO
M
ACRO
LIDE
SRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
serio
us a
dver
se e
vent
s in
NtM
pat
ient
s on
cla
rithr
omyc
in (H
IV in
fect
ed) (
asse
ssed
with
: unc
ontro
lled
coho
rts)
6 ob
serv
atio
nal
stud
iesd
serio
use
not s
erio
us
serio
usb
not s
erio
us
none
12
2/58
4 (2
0.9%
)g-
not
estim
able
e
VERY
lOW
CR
ItIC
Al
serio
us a
dver
se e
vent
s in
NtM
pat
ient
s on
azit
hrom
ycin
(HIV
uni
nfec
ted)
(ass
esse
d wi
th: u
ncon
trolle
d co
horts
)
5 ob
serv
atio
nal
stud
iesd
serio
use
serio
us
serio
usb
not s
erio
us
none
7/
197
(3
.6%
)hno
t es
timab
lee
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s in
NtM
pat
ient
s on
azit
hrom
ycin
(HIV
infe
cted
) (as
sess
ed w
ith: r
ando
mize
d co
ntro
lled
trial
s)
7 ra
ndom
ized
trial
s no
t se
rious
se
rious
se
rious
bse
rious
no
nec,
f11
3/12
15
(9.3
%)
57/1
196
(4.8
%)
not
estim
able
40 m
ore
per 1
,000
(fr
om 3
0 fe
wer t
o 10
0 m
ore)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in M
DR-t
b pa
tient
s on
mac
rolid
es (a
sses
sed
with
: ind
ivid
ual p
atie
nt d
ata
met
a-an
alys
is (A
huja
sD,
et a
l. 20
12; F
ox G
, et a
l. 20
15))
31
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
us
not s
erio
us
not s
erio
us
none
i25
4/39
6 (6
4.1%
)j32
92/4
970
(66.
2%)k
adju
sted
OR
0.7
(0
.3 to
1.
9)l
20 m
ore
per 1
,000
(fr
om 1
20
fewe
r to
150
mor
e)
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a Con
trolle
d co
horts
.b b
ased
on
stud
ies
of p
atie
nts
on p
reve
ntive
or c
urat
ive tr
eatm
ent f
or n
on-tu
berc
ulou
s m
ycob
acte
rial d
isea
se.
c Pat
ient
s wi
th a
dvan
ced
HIV,
and
stu
dies
from
pre
-ant
iretro
vira
ls e
ra.
d Un-
cont
rolle
d co
horts
.e U
nblin
ded
stud
ies;
adv
erse
eve
nts
attri
bute
d to
stu
dy d
rugs
by
auth
ors
with
non
-sta
ndar
dize
d m
etho
ds.
f ser
ious
adv
erse
eve
nts
expe
cted
to b
e m
ore
frequ
ent i
n th
ese
patie
nts
(adv
ance
d HI
V di
seas
e an
d no
ant
iretro
vira
l tre
atm
ent).
g 95%
Cl:
12%
–27%
.h 9
5% C
l: 0
%–8
%.
i Adj
uste
d es
timat
es u
sing
pro
pens
ity s
core
mat
chin
g.j A
djus
ted
estim
ate:
75%
(95%
Cl:
69%
–81%
).k A
djus
ted
estim
ates
73%
(95%
Cl:
66%
–81%
).l A
djus
ted
odds
ratio
est
imat
ed u
sing
pro
pens
ity s
core
mat
chin
g. R
efer
ence
pop
ulat
ion
for t
his
estim
ate
is p
atie
nts
in c
entre
s wh
ere
this
dru
g wa
s no
t use
d at
all.
28
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): F
ox G
, Men
zies
R, e
t al.
(11
Nov
embe
r 201
5)
Que
stio
n: Th
ioac
etaz
one c
ompa
red
to n
o th
ioac
etaz
one f
or tr
eatm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B an
d M
DR
-TB.
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing c
onve
ntio
nal r
egim
ens l
astin
g abo
ut 2
4 m
onth
s, in
low
and
high
re
sour
ce se
tting
s, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are.
Bib
liogr
aphy
: (1)
Fox
G, e
t al.
Gro
up 5
dru
gs fo
r mul
tidru
g-re
sista
nt tu
berc
ulos
is: in
divi
dual
pat
ient
dat
a met
a-an
alys
is (u
nder
revi
ew).
(2) A
huja
SD
, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. Mul
tidru
g res
istan
t pul
mon
ary t
uber
culo
sis tr
eatm
ent r
egim
ens a
nd p
atie
nt o
utco
mes
: an
indi
vidu
al p
atie
nt d
ata m
eta-
anal
ysis
of 9
,153
pat
ient
s. PL
oS M
ed. 2
012;
9(8)
:e10
0130
0.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
THIO
ACET
AZON
ENO
TH
IOAC
ETAZ
ONE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on th
ioac
etaz
one
as p
art o
f MDR
-tb
treat
men
t (as
sess
ed w
ith: i
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
)
31a
obse
rvat
iona
l st
udie
s ve
ry
serio
us
serio
usb
not s
erio
us
not s
erio
us
all p
laus
ible
resi
d-ua
l con
foun
ding
wo
uld
redu
ce th
e de
mon
stra
ted
effe
ct
491/
612
(80.
2%)c
3670
/564
7 (6
5.0%
)dad
just
ed
OR 2
.1
(0.8
to
5.5)
e
22 m
ore
per 1
,000
(fr
om 3
1 le
ss to
74
mor
e)f
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; R
E: ra
ndom
effe
cts
a In
7 of
thes
e st
udie
s at
leas
t one
per
son
rece
ived
thio
acet
azon
e (ra
nge:
1–6
71 p
er s
tudy
).b I
- squa
red
= 0%
(95%
Cl:
0%
–71%
).c R
E ad
just
ed %
= 8
0% (9
5% C
l: 7
7%–8
3%).
d RE
adju
sted
% =
72%
(95%
Cl:
63%
–80%
), am
ong
cont
rols
who
did
not
rece
ive th
ioac
etaz
one
in s
tudi
es w
here
thio
acet
azon
e wa
s no
t give
ne A
djus
ted
usin
g RE
mul
tivar
iabl
e an
alys
is w
ith p
rope
nsity
sco
re m
atch
ing
to a
djus
t for
pot
entia
l con
foun
ding
bet
ween
pat
ient
s ta
king
thio
acet
azon
e an
d m
atch
ed c
ontro
ls in
stu
dies
whe
re th
ioac
etaz
one
was
not u
sed
f RE
anal
ysis
, onl
y in
clud
ing
7 st
udie
s wh
ere
thio
acet
azon
e wa
s us
ed.
2929
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): B
asto
s M, L
an Z
, Men
zies
R (1
1 N
ovem
ber 2
015)
Que
stio
n: p-
amin
osal
icyl
ic ac
id co
mpa
red
to n
o p-
amin
osal
icyl
ic ac
id fo
r tre
atm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B or
MD
R-T
B.
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith ri
fam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing c
onve
ntio
nal r
egim
ens l
astin
g abo
ut 2
4 m
onth
s, in
low
and
high
re
sour
ce se
tting
s, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are
Bib
liogr
aphy
: (1)
Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug
resis
tant
pul
mon
ary
tube
rcul
osis
trea
tmen
t re
gim
ens a
nd p
atie
nt o
utco
mes
: an
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
is of
9,1
53 p
atie
nts.
PLoS
Med
. 201
2;9(
8):e
1001
300.
(2) B
asto
s M, L
an Z
, M
enzi
es R
. An
upda
ted
syst
emat
ic re
view
and
met
a-an
alys
is fo
r tre
atm
ent o
f mul
tidru
g-re
sista
nt tu
berc
ulos
is, 2
016
(und
er re
view
, 28
May
201
6).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
P-AM
INOS
ALI-
CYLI
C AC
IDNO
P-A
MIN
O-SA
LICY
LIC
ACID
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on p
-am
inos
alic
ylic
aci
d (P
As),
as p
art o
f a M
DR-t
b re
gim
en (a
sses
sed
with
: ind
ivid
ual p
atie
nt d
ata
met
a-an
alys
is (2
012)
)
32
obse
rvat
iona
l st
udie
s se
rious
ano
t ser
ious
no
t ser
ious
no
t ser
ious
al
l pla
usib
le re
sid-
ual c
onfo
undi
ng
woul
d re
duce
the
dem
onst
rate
d ef
fect
2162
/287
1 (7
5.3%
) 28
17/4
283
(65.
8%)
aOR
1.0
(0.8
to
1.4)
b
105
mor
e pe
r 1,0
00
(from
110
fe
wer t
o 12
0 m
ore)
lO
W
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se in
pat
ient
s on
PAs
as
part
of a
MDR
-tb
regi
men
(ass
esse
d wi
th: a
ggre
gate
dat
a m
eta-
anal
ysis
(201
5)
55
obse
rvat
iona
l st
udie
s se
rious
cno
t ser
ious
no
t ser
ious
no
t ser
ious
no
ned
4981
/574
4 (8
6.7%
)e29
68/3
595
(82.
6%)f
49 m
ore
per 1
,000
(fr
om
7 fe
wer
to 1
07
mor
e)l l
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on P
As a
s pa
rt of
a M
DR-t
b re
gim
en (a
sses
sed
with
: agg
rega
te d
ata
met
a-an
alys
is (2
015)
g
55
obse
rvat
iona
l st
udie
s se
rious
cno
t ser
ious
no
t ser
ious
no
t ser
ious
no
ned
4981
/627
6 (7
9.4%
)h29
68/4
521
(65.
6%)i
54 m
ore
per 1
,000
(fr
om 3
4 fe
wer t
o 14
4 m
ore)
l
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s (G
rade
3 o
r 4, o
r dru
gs s
topp
ed d
ue to
adv
erse
eve
nts)
in p
atie
nts
on P
As, a
s pa
rt of
a M
DR-t
b re
gim
en (a
sses
sed
with
: agg
rega
ted
data
met
a-an
alys
is 2
015)
16
obse
rvat
iona
l st
udie
s se
rious
no
t ser
ious
no
t ser
ious
no
t ser
ious
no
nej
208/
1706
(1
2.2%
)kno
t es
timab
le
VE
RY l
OW
CRIt
ICAl
30
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Cl: c
onfid
ence
lim
its; F
E: fi
xed
effe
cts
a Ind
ivid
ual p
atie
nt d
ata
take
n fro
m 3
2 ob
serv
atio
nal s
tudi
es in
whi
ch m
ost p
atie
nts
rece
ived
indi
vidu
alize
d tre
atm
ent.
Risk
of s
elec
tion
bias
, and
con
foun
ding
by
indi
catio
n.b a
OR: O
dds
ratio
adj
uste
d fo
r age
, HIV
, aci
d-fa
st b
acill
us s
mea
r, ch
est r
adio
grap
h ca
vita
tion,
and
prio
r tre
atm
ent w
ith fi
rst l
ine,
and
sec
ond
line
tb d
rugs
.c V
ery
serio
us li
mita
tions
– a
ll st
udie
s we
re o
bser
vatio
nal –
lead
ing
to ri
sk o
f sel
ectio
n an
d in
form
atio
n bi
as. I
n 20
stu
dies
the
patie
nts
were
give
n st
anda
rdize
d re
gim
ens,
but
in th
e re
mai
ning
40
stud
ies
ther
apy
was
indi
vidu
alize
d, le
adin
g to
risk
of c
onfo
undi
ng b
y in
dica
tion.
d Una
djus
ted
anal
ysis
.e P
oole
d pr
opor
tion:
93%
(95%
Cl:
83%
–96%
).f P
oole
d pr
opor
tion:
90%
(95%
Cl:
85%
–95%
).g F
rom
agg
rega
te d
ata
met
a-an
alys
is: P
atie
nts
with
XDR
-tb
excl
uded
from
ana
lyse
s, w
here
pos
sibl
e.h P
oole
d pr
opor
tion:
81%
(95%
Cl:
75%
–87%
).i P
oole
d pr
opor
tion:
78%
(95%
Cl:
71%
–85%
).j s
erio
us a
dver
se e
vent
s (s
AEs)
repo
rted
in p
atie
nts
were
attr
ibut
ed to
a m
edic
ine
by th
e au
thor
s wh
o we
re u
nblin
ded
and
used
non
-sta
ndar
dize
d m
etho
ds to
defi
ne, a
scer
tain
and
repo
rt sA
Es. N
o va
lid c
ompa
rison
s ar
e po
ssib
le w
ith p
atie
nts
not o
n th
e ta
rget
med
icin
e, b
ecau
se s
AEs
in th
ese
patie
nts
coul
d be
due
to o
ther
dru
gs re
ceive
d.k P
oole
d pr
opor
tion:
FE
95%
Cl:
10.
6%–1
3.9%
.l R
isk
diffe
renc
e fro
m a
djus
ted
anal
ysis
.
3131
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): B
asto
s M, L
an Z
, Men
zies
R (1
1 N
ovem
ber 2
015)
Que
stio
n: P
yraz
inam
ide c
ompa
red
to n
o py
razi
nam
ide f
or ad
ults
with
rifa
mpi
cin-
resis
tant
TB
or M
DR
-TB.
Sett
ing:
Tre
atm
ent o
f adu
lts w
ith r
ifam
pici
n-re
sista
nt T
B/M
DR
-TB/
XD
R-T
B us
ing
conv
entio
nal r
egim
ens l
astin
g ab
out 2
4 m
onth
s and
shor
ter
MD
R-T
B re
gim
ens,
in lo
w an
d hi
gh re
sour
ce se
tting
s, w
ithin
hos
pita
l or a
mbu
lato
ry m
odel
s of c
are.
Bib
liogr
aphy
: (1)
Ahu
ja S
D, A
shki
n D
, Ave
ndan
o M
, Ban
erje
e R, B
auer
M, B
ayon
a JN
, et a
l. M
ultid
rug
resis
tant
pul
mon
ary
tube
rcul
osis
trea
tmen
t re
gim
ens a
nd p
atie
nt o
utco
mes
: an
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
is of
9,1
53 p
atie
nts.
PLoS
Med
. 201
2;9(
8):e
1001
300.
(2) B
asto
s M, L
an Z
, M
enzi
es R
. An
upda
ted
syst
emat
ic re
view
and
met
a-an
alys
is fo
r tre
atm
ent o
f mul
tidru
g-re
sista
nt tu
berc
ulos
is, 2
016
(und
er re
view
, 28
May
201
6).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
PYRA
ZINA
MID
ENO
PY
RAZI
NAM
IDE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se/d
eath
in p
atie
nts
on p
yraz
inam
ide
as p
art o
f a M
DR-t
b re
gim
en (a
sses
sed
with
: ind
ivid
ual p
atie
nt d
ata
met
a-an
alys
is (A
huja
sD,
et a
l. Pl
Os M
ed. 2
012)
20
obse
rvat
iona
l st
udie
s se
rious
no
t ser
ious
no
t ser
ious
no
t ser
ious
no
ne
2454
/377
5 (6
5.0%
) 55
/89
(61.
8%)
aOR
1.3
(1.1
to
1.6)
a
32 m
ore
per 1
000
(from
10
mor
e to
60
mor
e)
VE
RY l
OW
CRIt
ICAl
serio
us a
dver
se e
vent
s (G
rade
3–4
eve
nts,
or d
rugs
sto
pped
due
to a
dver
se e
vent
s) in
pat
ient
s on
pyr
azin
amid
e as
par
t of a
MDR
-tb
regi
men
(ass
esse
d wi
th: a
ggre
gate
d da
ta m
eta-
anal
ysis
201
5)
19
obse
rvat
iona
l st
udie
s se
rious
no
t ser
ious
no
t ser
ious
no
t ser
ious
no
neb
56/2
023
(2.8
%)c
not
estim
able
VERY
lOW
CR
ItIC
Al
Cl: c
onfid
ence
lim
its; F
E: fi
xed
effe
cts
a aOR
: odd
s ra
tio a
djus
ted
for a
ge, H
IV, a
cid-
fast
bac
illus
sm
ear,
ches
t rad
iogr
aph
cavi
tatio
n, a
nd p
rior t
reat
men
t with
firs
t-lin
e an
d se
cond
-line
tb d
rugs
.b s
erio
us a
dver
se e
vent
s (s
AEs)
repo
rted
in p
atie
nts
were
attr
ibut
ed to
a m
edic
ine
by th
e au
thor
s wh
o we
re u
nblin
ded
and
used
non
-sta
ndar
dize
d m
etho
ds to
defi
ne, a
scer
tain
and
repo
rt sA
Es. N
o va
lid c
ompa
rison
s ar
e po
ssib
le w
ith p
atie
nts
not o
n th
e ta
rget
med
icin
e, b
ecau
se s
AEs
in th
ese
patie
nts
coul
d be
due
to o
ther
dru
gs re
ceive
d.c P
oole
d pr
opor
tion:
FE
95%
Cl:
2.1
%–3
.7%
.
32
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
3. M
DR
-TB
reg
imen
com
posi
tion
– p
aedi
atric
indi
vidu
al p
atie
nt d
ata
met
a-an
alys
is (
PIC
O 1
)A
utho
r(s)
: Eliz
abet
h H
arau
sz, A
ntho
ny G
arci
a-Pr
ats,
Sim
on S
chaa
f, St
epha
nie
Law,
Dic
k M
enzi
es, J
enni
fer
Furin
, Tam
ara
Kre
do a
nd A
nnek
e C
. H
esse
ling o
n be
half
of th
e Pae
diat
ric M
DR
-TB
IPD
Gro
up (1
1 N
ovem
ber 2
015)
Que
stio
n: L
ater
-gen
erat
ion
fluor
oqui
nolo
nes c
ompa
red
to n
o la
ter-
gene
ratio
n flu
oroq
uino
lone
s for
chi
ldre
n w
ith M
DR
-TB
(exc
ludi
ng c
onfir
med
X
DR
-TB)
.
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
LATE
R-
GENE
RATI
ON
FLUO
ROQU
I-NO
LONE
S
NO L
ATER
- GE
NERA
TION
FL
UORO
QUI-
NOLO
NES
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– co
nfirm
ed c
ases
(IPD
ana
lysi
s): n
= 6
23
12
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
480/
551
(87.
1%)
36/4
5 (8
0.0%
) OR
0.7
10
(0.0
94 to
5.
370)
a
37 fe
wer
pe
r 100
0 (fr
om 1
80
fewe
r to
110
mor
e)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die/
lost
to fo
llow
up –
unc
onfir
med
cas
es (I
PD a
naly
sis)
: n =
219
b
3 ob
serv
atio
nal
stud
ies
serio
us
serio
us
not s
erio
us
not s
erio
us
none
19
/21
(90.
5%)
169/
184
(91.
8%)
OR 0
.667
(0
.064
to
6.96
6)a,
b
47 fe
wer
pe
r 100
0 (fr
om 1
3 fe
wer t
o 10
8 m
ore)
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a All
effe
ct e
stim
ates
sho
wn a
re a
djus
ted
for a
ge, H
IV s
tatu
s, g
ende
r, tb
dis
ease
sev
erity
and
site
(ran
dom
effe
cts
mod
el w
ith c
lust
erin
g by
site
).b U
ncon
firm
ed c
ases
incl
ude
lost
to fo
llow
up in
this
ana
lysi
s on
ly.
3333
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): E
lizab
eth
Har
ausz
, Ant
hony
Gar
cia-
Prat
s, Si
mon
Sch
aaf,
Step
hani
e La
w, D
ick
Men
zies
, Jen
nife
r Fu
rin, T
amar
a K
redo
and
Ann
eke
C.
Hes
selin
g on
beha
lf of
the P
aedi
atric
MD
R-T
B IP
D G
roup
(11
Nov
embe
r 201
5)
Que
stio
n: S
econ
d-lin
e inj
ecta
ble a
gent
com
pare
d to
no
seco
nd-li
ne in
ject
able
agen
t for
child
ren
with
MD
R-T
B (e
xclu
ding
confi
rmed
XD
R-T
B)
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
SECO
ND-L
INE
IN
JECT
ABLE
AG
ENT
NO S
ECON
D LI
NE
INJE
CTAB
LE
AGEN
TRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– co
nfirm
ed c
ases
(IPD
ana
lysi
s): n
= 6
23
25
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
493/
566
(87.
1%)
41/5
7 (7
1.9%
) OR
3.3
2 (1
.53
to
7.21
)a
43 m
ore
per 1
000
(from
107
fe
wer t
o 19
4 m
ore)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– un
confi
rmed
cas
es (I
PD a
naly
sis)
: n =
219
12
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
154/
157
(98.
1%)
58/6
2 (9
3.5%
) OR
1.3
8 (0
.14
to
13.5
0)a
11 m
ore
per 1
000
(from
108
fe
wer t
o 12
9 m
ore)
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
it; O
R: o
dds
ratio
a All
effe
ct e
stim
ates
sho
wn a
re a
djus
ted
for a
ge, H
IV s
tatu
s, g
ende
r, tb
dis
ease
sev
erity
and
site
(ran
dom
effe
cts
mod
el w
ith c
lust
erin
g by
site
).
34
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): E
lizab
eth
Har
ausz
, Ant
hony
Gar
cia-
Prat
s, Si
mon
Sch
aaf,
Step
hani
e La
w, D
ick
Men
zies
, Jen
nife
r Fu
rin, T
amar
a K
redo
and
Ann
eke
C.
Hes
selin
g on
beha
lf of
the P
aedi
atric
MD
R-T
B IP
D G
roup
(11
Nov
embe
r 201
5)
Que
stio
n: E
thio
nam
ide/
prot
hion
amid
e com
pare
d to
no
ethi
onam
ide/
prot
hion
amid
e for
child
ren
with
MD
R-T
B (e
xclu
ding
confi
rmed
XD
R-T
B)
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
ETHI
ONAM
IDE/
PROT
HION
A-M
IDE
NO
ETHI
ONAM
IDE/
PROT
HION
A-M
IDE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– co
nfirm
ed c
ases
(IPD
ana
lysi
s): n
= 6
23
24
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
493/
574
(85.
9%)
41/4
9 (8
3.7%
) OR
2.0
4 (0
.29
to
14.6
0)a
59 fe
wer
pe
r 100
0 (fr
om 1
80
fewe
r to
60 m
ore)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– un
confi
rmed
cas
es (I
PD a
naly
sis)
: n =
219
11
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
181/
187
(96.
8%)
31/3
2 (9
6.9%
) OR
1.0
8 (0
.05
to
21.9
0)a
19 fe
wer
pe
r 100
0 (fr
om 1
39
fewe
r to
102
mor
e)
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a All
effe
ct e
stim
ates
sho
wn a
re a
djus
ted
for a
ge, H
IV s
tatu
s, g
ende
r, tb
dis
ease
sev
erity
and
site
(ran
dom
effe
cts
mod
el w
ith c
lust
erin
g by
site
).
3535
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): E
lizab
eth
Har
ausz
, Ant
hony
Gar
cia-
Prat
s, Si
mon
Sch
aaf,
Step
hani
e La
w, D
ick
Men
zies
, Jen
nife
r Fu
rin, T
amar
a K
redo
and
Ann
eke
C.
Hes
selin
g on
beha
lf of
the P
aedi
atric
MD
R-T
B IP
D G
roup
(11
Nov
embe
r 201
5)
Que
stio
n: C
yclo
serin
e/te
rizid
one c
ompa
red
to n
o cy
clos
erin
e/te
rizid
one f
or in
child
ren
with
MD
R-T
B (e
xclu
ding
confi
rmed
XD
R-T
B)
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NSCY
CLOS
ERIN
E/
TERI
ZIDO
NE
NO
CYCL
OSER
INE/
TE
RIZI
DONE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die/
lost
– c
onfir
med
cas
es o
nly
(IPD
anal
ysis
): n
= 70
1
24
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
307/
339
(90.
6%)
227/
284
(79.
9%)
OR 1
.70
(0.9
1 to
3.
19)a
3 fe
wer
pe
r 100
0 (fr
om 9
0 fe
wer t
o 97
mor
e)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– un
confi
rmed
cas
es (I
PD a
naly
sis)
: n =
219
10
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
132/
134
(98.
5%)
80/8
5 (9
4.1%
) OR
0.3
8 (0
.01
to
28.9
0)a
13 fe
wer
pe
r 100
0 (fr
om 1
06
fewe
r to
81 m
ore)
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a All
effe
ct e
stim
ates
sho
wn a
re a
djus
ted
for a
ge, H
IV s
tatu
s, g
ende
r, tb
dis
ease
sev
erity
and
site
(ran
dom
effe
cts
mod
el w
ith c
lust
erin
g by
site
).
36
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): E
lizab
eth
Har
ausz
, Ant
hony
Gar
cia-
Prat
s, Si
mon
Sch
aaf,
Step
hani
e La
w, D
ick
Men
zies
, Jen
nife
r Fu
rin, T
amar
a K
redo
and
Ann
eke
C.
Hes
selin
g on
beha
lf of
the P
aedi
atric
MD
R-T
B IP
D G
roup
(11
Nov
embe
r 201
5)
Que
stio
n: C
lofa
zim
ine c
ompa
red
to n
o cl
ofaz
imin
e for
child
ren
with
MD
R tu
berc
ulos
is (e
xclu
ding
confi
rmed
XD
R-T
B)
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
CLOF
AZIM
INE
NO
CLOF
AZIM
INE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– co
nfirm
ed c
ases
(IPD
ana
lysi
s): n
= 6
23
9 ob
serv
atio
nal
stud
ies
serio
us
serio
us
not s
erio
us
serio
us
none
18
/23
(78.
3%)
516/
600
(86.
0%)
OR 0
.46
(0.0
2 to
10
.00)
a
47 m
ore
per 1
000
(from
81
fewe
r to
170
mor
e)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– un
confi
rmed
cas
es (I
PD a
naly
sis)
: n =
219
2 ob
serv
atio
nal
stud
ies
serio
us
serio
us
not s
erio
us
serio
us
none
4/
4 (1
00.0
%)
208/
215
(96.
7%)
OR 0
.25
(0.1
2 to
5.
30)b
47 m
ore
per 1
000
(from
14
fewe
r to
107
mor
e)
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a Effe
ct e
stim
ates
for t
he c
onfir
med
are
adj
uste
d fo
r age
, HIV
sta
tus,
gen
der,
tb d
isea
se s
ever
ity a
nd s
ite (r
ando
m e
ffect
s m
odel
with
clu
ster
ing
by s
ite).
b Effe
ct e
stim
ate
is n
ot a
djus
ted.
3737
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): E
lizab
eth
Har
ausz
, Ant
hony
Gar
cia-
Prat
s, Si
mon
Sch
aaf,
Step
hani
e La
w, D
ick
Men
zies
, Jen
nife
r Fu
rin, T
amar
a K
redo
and
Ann
eke
C.
Hes
selin
g on
beha
lf of
the P
aedi
atric
MD
R-T
B IP
D G
roup
(11
Nov
embe
r 201
5)
Que
stio
n: P
yraz
inam
ide c
ompa
red
to n
o py
razi
nam
ide f
or ch
ildre
n w
ith M
DR
tube
rcul
osis
(exc
ludi
ng co
nfirm
ed X
DR
-TB)
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
PYRA
ZINA
MID
ENO
PY
RAZI
NAM
IDE
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– co
nfirm
ed c
ases
(IPD
ana
lysi
s): n
= 6
23
26
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
499/
582
(85.
7%)
35/4
1 (8
5.4%
) OR
0.4
5 (0
.01
to
33.4
0)a
66 fe
wer
pe
r 100
0 (fr
om 1
60
fewe
r to
26 m
ore)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– un
confi
rmed
cas
es (I
PD a
naly
sis)
: n =
219
12
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
187/
194
(96.
4%)
25/2
5 (1
00.0
%)
OR 0
.490
(0
.027
to
8.84
0)b
50 fe
wer
pe
r 100
0 (fr
om 1
14
fewe
r to
14 m
ore)
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a Effe
ct e
stim
ates
for c
onfir
med
are
adj
uste
d fo
r age
, HIV
sta
tus,
gen
der,
tb d
isea
se s
ever
ity a
nd s
ite (r
ando
m e
ffect
s m
odel
with
clu
ster
ing
by s
ite)
b OR
for u
ncon
firm
ed c
ases
is n
ot a
djus
ted.
38
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): E
lizab
eth
Har
ausz
, Ant
hony
Gar
cia-
Prat
s, Si
mon
Sch
aaf,
Step
hani
e La
w, D
ick
Men
zies
, Jen
nife
r Fu
rin, T
amar
a K
redo
and
Ann
eke
C.
Hes
selin
g on
beha
lf of
the P
aedi
atric
MD
R-T
B IP
D G
roup
(11
Nov
embe
r 201
5)
Que
stio
n: H
igh
dose
ison
iazi
d co
mpa
red
to n
o hi
gh d
ose i
soni
azid
for c
hild
ren
with
MD
R-T
B (e
xclu
ding
confi
rmed
XD
R-T
B)a
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
HIGH
DOS
E
ISON
IAZI
DNO
HIG
H DO
SE
ISON
IAZI
DRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– co
nfirm
ed c
ases
(IPD
ana
lysi
s): n
= 6
23
6 ob
serv
atio
nal
stud
ies
serio
usa
serio
us
not s
erio
us
not s
erio
us
none
13
0/13
3 (9
7.7%
) 40
4/49
0 (8
2.4%
) OR
6.9
7 (2
.11
to
23.0
0)b
120
mor
e pe
r 100
0 (fr
om 5
9 m
ore
to
187
mor
e)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– un
confi
rmed
cas
es (I
PD a
naly
sis)
: n =
219
c
1 ob
serv
atio
nal
stud
ies
serio
usa
serio
us
not s
erio
us
not s
erio
us
none
85
/85
(100
.0%
) 12
7/13
4 (9
4.8%
) OR
10.
06
(0.5
6 to
17
8.40
)c
—
VERY
lOW
CR
ItIC
Al
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a Mos
t of t
he c
ases
rece
ivin
g hi
gh-d
ose
ison
iazid
wer
e fro
m c
ohor
ts in
sou
th A
frica
, so
desp
ite a
djus
ting
for s
tudy
site
, the
re m
ay s
till b
e so
me
resi
dual
con
foun
ding
.b E
ffect
est
imat
es s
hown
are
adj
uste
d fo
r age
, HIV
sta
tus,
gen
der,
tb d
isea
se s
ever
ity a
nd s
ite (r
ando
m e
ffect
s m
odel
with
clu
ster
ing
by s
ite).
c OR
for t
he u
ncon
firm
ed c
ases
is n
ot a
djus
ted.
3939
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): E
lizab
eth
Har
ausz
, Ant
hony
Gar
cia-
Prat
s, Si
mon
Sch
aaf,
Step
hani
e La
w, D
ick
Men
zies
, Jen
nife
r Fu
rin, T
amar
a K
redo
and
Ann
eke
C.
Hes
selin
g on
beha
lf of
the P
aedi
atric
MD
R-T
B IP
D G
roup
(11
Nov
embe
r 201
5)
Que
stio
n: p-
amin
osal
icyl
ic ac
id co
mpa
red
to n
o p-
amin
osal
icyl
ic ac
id fo
r chi
ldre
n w
ith M
DR
-TB
(exc
ludi
ng co
nfirm
ed X
DR
-TB)
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NSP-
AMIN
OSAL
I-CY
LIC
ACID
NO
P-AM
INOS
ALI-
CYLI
C AC
ID
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– co
nfirm
ed c
ases
(IPD
ana
lysi
s): n
= 6
23
20
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
no
t ser
ious
no
ne
115/
135
(85.
2%)
419/
488
(85.
9%)
OR 0
.52
(0.2
6 to
1.
07)a
5 fe
wer
pe
r 100
0 (fr
om 1
10
fewe
r to
95 m
ore)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die/
lost
to fo
llow
up –
unc
onfir
med
cas
es (I
PD a
naly
sis)
: n =
237
b
8 ob
serv
atio
nal
stud
ies
serio
us
serio
us
not s
erio
us
serio
us
none
69
/75
(92.
0%)
143/
162
(88.
3%)
OR 0
.18
(0.0
2 to
1.
76)a,
b
27 fe
wer
pe
r 100
0 (fr
om 6
0 fe
wer t
o 11
5 m
ore)
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a All
effe
ct e
stim
ates
for c
onfir
med
cas
es a
re a
djus
ted
for a
ge, H
IV s
tatu
s, g
ende
r, tb
dis
ease
sev
erity
and
site
(ran
dom
effe
cts
mod
el w
ith c
lust
erin
g by
site
).b O
R fo
r the
unc
onfir
med
cas
es in
clud
es lo
st to
follo
w up
in th
is c
alcu
latio
n on
ly.
40
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Aut
hor(
s): E
lizab
eth
Har
ausz
, Ant
hony
Gar
cia-
Prat
s, Si
mon
Sch
aaf,
Step
hani
e La
w, D
ick
Men
zies
, Jen
nife
r Fu
rin, T
amar
a K
redo
and
Ann
eke
C.
Hes
selin
g on
beha
lf of
the P
aedi
atric
MD
R-T
B IP
D G
roup
(11
Nov
embe
r 201
5)
Que
stio
n: C
larit
hrom
ycin
com
pare
d to
no
clar
ithro
myc
in fo
r chi
ldre
n w
ith M
DR
-TB
(exc
ludi
ng co
nfirm
ed X
DR
-TB)
Sett
ing:
Inte
rnat
iona
l
Bib
liogr
aphy
: Ref
er to
Ann
ex 6
, Sec
tion
3 fo
r a su
mm
ary o
f thi
s unp
ublis
hed
stud
y (H
arau
sz E
, Gar
cia-
Prat
s AJ,
Scha
af S
, Law
S, F
urin
J, K
redo
T, e
t al
., for
The C
olla
bora
tive G
roup
for M
eta-
Ana
lysis
of P
aedi
atric
Indi
vidu
al P
atie
nt D
ata i
n M
DR
-TB.
A sy
stem
atic
revi
ew an
d in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
of tr
eatm
ent a
nd o
utco
mes
amon
g chi
ldre
n w
ith m
ulti-
drug
resis
tant
tube
rcul
osis.
A p
relim
inar
y rep
ort f
or th
e Gui
delin
e Dev
elop
men
t G
roup
Mee
ting o
f the
Wor
ld H
ealth
Org
aniz
atio
n, N
ovem
ber 9
–11
2015
).
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
TC
ERTA
INTY
O
F EV
IDEN
CE
IMP
OR
TAN
CE
NO. O
F ST
UDIE
SST
UDY
DESI
GNRI
SK O
F BI
ASIN
CONS
ISTE
NCY
INDI
RECT
NESS
IMPR
ECIS
ION
OTHE
R CO
NSID
ERAT
IONS
CLAR
ITHR
OMYC
INNO
CL
ARIT
HROM
YCIN
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– co
nfirm
ed c
ases
(IPD
ana
lysi
s): n
= 6
23
11
obse
rvat
iona
l st
udie
s se
rious
se
rious
no
t ser
ious
se
rious
no
ne
22/3
2 (6
8.8%
) 51
2/59
1 (8
6.6%
) OR
0.2
4 (0
.04
to
1.51
)a
24 fe
wer
pe
r 100
0 (fr
om 2
20
fewe
r to
170
mor
e)
VE
RY l
OW
CRIt
ICAl
treat
men
t suc
cess
ver
sus
fail/
rela
pse/
die
– un
confi
rmed
cas
es (I
PD a
naly
sis)
: n =
219
2 ob
serv
atio
nal
stud
ies
serio
us
serio
us
not s
erio
us
serio
us
none
3/
3 (1
00.0
%)
209/
216
(96.
8%)
not
estim
able
—
VE
RY l
OW
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a All
effe
ct e
stim
ates
sho
wn a
re a
djus
ted
for a
ge, H
IV s
tatu
s, g
ende
r, tb
dis
ease
sev
erity
and
site
(ran
dom
effe
cts
mod
el w
ith c
lust
erin
g by
site
).
4141
ANNE
X 4:
GRA
DE tA
blEs
4.
The
role
of s
urge
ry (
PIC
O 4
)
Aut
hor(
s): H
arris
RC
, Kha
n M
S, A
llen
V, M
oore
DA
J, Fi
eldi
ng K
, Gra
ndje
an L
, and
the
LSH
TM
MD
R-T
B su
rger
y sy
stem
atic
revi
ew g
roup
(11
Nov
embe
r 201
5)
Que
stio
n: S
urge
ry co
mpa
red
to n
o su
rger
y for
trea
tmen
t of M
DR
or X
DR
TB
Sett
ing:
Geo
rgia
, Lat
via,
Rus
sia, S
outh
Afri
ca, S
outh
Kor
ea an
d Tu
rkey
Bib
liogr
aphy
: (1)
Har
ris R
C, K
han
MS,
Mar
tin L
J, A
llen
V, M
oore
DA
J, Fi
eldi
ng K
, et a
l. an
d th
e LS
HT
M M
DR
-TB
surg
ery
syst
emat
ic re
view
gr
oup.
The e
ffect
of s
urge
ry o
n th
e out
com
e of t
reat
men
t for
mul
tidru
g-re
sista
nt tu
berc
ulos
is: a
syst
emat
ic re
view
and
met
a-an
alys
is. B
MC
Infe
ct D
is.
2016
;16(
1). (
2) D
ravn
iece
G, C
ain
KP,
Hol
tz T
H, R
ieks
tina V
, Lei
man
e V, Z
ales
kis R
. Adj
unct
ive r
esec
tiona
l lun
g sur
gery
for e
xten
sivel
y dru
g-re
sista
nt
tube
rcul
osis.
Eur
Res
pir J
. 200
9;34
(1):1
80–1
83. (
3) G
egia
M, K
alan
dadz
e I,
Kem
pker
RR
, Mag
ee M
J, Bl
umbe
rg H
M. A
djun
ctiv
e su
rger
y im
prov
es
trea
tmen
t out
com
es am
ong p
atie
nts w
ith m
ultid
rug-
resis
tant
and
exte
nsiv
ely d
rug-
resis
tant
tube
rcul
osis.
Int J
Infe
ct D
is. 20
12;1
6:e3
91–3
96. (
4) K
arag
öz
T, Y
azic
ioğl
u M
oçin
O, P
azar
li P,
Sen
ol T
, Yet
iş D
uman
D, D
uman
G, e
t al.
The t
reat
men
t res
ults
of p
atie
nts w
ith m
ultid
rug r
esist
ant t
uber
culo
sis an
d fa
ctor
s affe
ctin
g tre
atm
ent o
utco
me.
Tube
rk T
orak
s. 20
09;5
7:38
3–39
2. (5
) Kes
havj
ee S
, Gel
man
ova I
Y, F
arm
er P
E, M
ishus
tin S
P, S
trel
is A
K, A
ndre
ev
YG, e
t al.
Trea
tmen
t of e
xten
sivel
y dr
ug-re
sista
nt tu
berc
ulos
is in
Tom
sk, R
ussia
: a re
tros
pect
ive c
ohor
t stu
dy. L
ance
t 200
8;37
2:14
03–1
409.
(6) K
im
H-R
, Hw
ang S
S, K
im H
J, Le
e SM
, Yoo
C-G
, Kim
YW
, et a
l. Im
pact
of e
xten
sive d
rug r
esist
ance
on
trea
tmen
t out
com
es in
non
-HIV
-infe
cted
pat
ient
s w
ith m
ultid
rug-
resis
tant
tube
rcul
osis.
Clin
Infe
ct D
is. 2
007;
45(1
0):1
290–
1295
. (7)
Kim
DH
, Kim
HJ,
Park
S-K
, Kon
g S-J,
Kim
YS,
Kim
T-H
, et a
l. Tr
eatm
ent o
utco
mes
and
long
-term
surv
ival
in p
atie
nts w
ith ex
tens
ivel
y dru
g-re
sista
nt tu
berc
ulos
is. A
m J
Res
pir C
rit C
are M
ed. 2
008;
178:
1075
–108
2.
(8) K
wak
N, K
im H
R, Y
oo C
G, K
im Y
W, H
an S
K, Y
im JJ
. Cha
nges
in tr
eatm
ent o
utco
mes
of m
ultid
rug-
resis
tant
tube
rcul
osis.
Int J
Tub
erc
Lung
D
is. 2
015;
19:5
25–5
30. (
9) K
won
YS,
Kim
YH
, Suh
GY,
Chu
ng M
P, K
im H
, Kw
on O
J, et
al.
Trea
tmen
t out
com
es fo
r HIV
-uni
nfec
ted
patie
nts w
ith
mul
tidru
g-re
sista
nt an
d ex
tens
ivel
y dru
g-re
sista
nt tu
berc
ulos
is. C
lin In
fect
Dis.
200
8;47
:496
–502
. (10
) Lei
man
e V, R
ieks
tina V
, Hol
tz T
H, Z
arov
ska
E, S
krip
cono
ka V
, Thor
pe L
E, et
al. C
linic
al o
utco
me o
f ind
ivid
ualis
ed tr
eatm
ent o
f mul
tidru
g-re
sista
nt tu
berc
ulos
is in
Lat
via:
a re
tros
pect
ive c
ohor
t st
udy.
Lanc
et 2
005;
365:
318–
326.
(11)
Mitn
ick
CD
1, S
hin
SS, S
eung
KJ,
Ric
h M
L, A
twoo
d SS
, Fur
in JJ
, et a
l. Com
preh
ensiv
e tre
atm
ent o
f ext
ensiv
ely
drug
-resis
tant
tube
rcul
osis.
New
Eng
l J M
ed 2
008;
359:
563–
574.
(12)
She
an K
P, W
illco
x PA
, Siw
endu
SN
, Las
erso
n K
F, G
ross
L, K
amm
erer
S, e
t al
. Tre
atm
ent o
utco
me
and
follo
w-u
p of
mul
tidru
g-re
sista
nt tu
berc
ulos
is pa
tient
s, W
est C
oast
/Win
elan
ds, S
outh
Afri
ca, 1
992–
2002
. Int
J Tu
berc
Lu
ng D
is. 2
008;
12(1
0):1
182–
1189
. (13
) Skl
yuev
S, L
evin
A, T
chei
mac
h E,
Kra
snov
D. P
C-6
58–0
2 C
ompl
ex tr
eatm
ent a
ppro
ach
for p
atie
nts w
ith
destr
uctiv
e pul
mon
ary
tube
rcul
osis
by ap
plic
atio
n of
endo
bron
chia
l val
ve. I
nt J
Tube
rc L
ung
Dis.
201
3;17
(12,
Sup
p.2)
:S32
9–33
0. (1
4) T
ahao
ğlu
K,
Tör
ün T
, Sev
im T
, Ata
ç G, K
ir A
, Kar
asul
u L,
et a
l. Th
e tre
atm
ent o
f mul
tidru
g-re
sista
nt tu
berc
ulos
is in
Tur
key.
N E
ngl J
Med
. 200
1;34
5:17
0–17
4.
(15)
Tör
ün T
, Tah
aoğl
u K
, Ozm
en I,
Sev
im T
, Ata
ç G, K
ir A
, et a
l. Th
e rol
e of s
urge
ry an
d flu
oroq
uino
lone
s in
the t
reat
men
t of m
ultid
rug-
resis
tant
tu
berc
ulos
is. In
t J T
uber
c Lun
g Dis.
200
7;11
(9):9
79–9
85.
42
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
QU
ALI
TYIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NSSU
RGER
YNO
SUR
GERY
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
Cure
d (fo
llow
up: r
ange
0.5
to 1
0 ye
ars;
ass
esse
d wi
th: W
HO d
efini
tion)
5 ob
serv
atio
nal
stud
ies
serio
usa–
gno
t ser
ious
hno
t ser
ious
ino
t ser
ious
no
nej
118/
157
(75.
2%)
308/
561
(54.
9%)
OR 3
.03
(1.5
9 to
5.
78)
238
mor
e pe
r 1,0
00
(from
110
m
ore
to
327
mor
e)
VE
RY l
OW
CRIt
ICAl
succ
essf
ul o
utco
me
(follo
w up
: ran
ge 0
.25
to 7
yea
rs; a
sses
sed
with
: cur
e or
trea
tmen
t suc
cess
, WHO
defi
nitio
n)
14
obse
rvat
iona
l st
udie
s se
rious
a–
g,k,
lno
t ser
ious
mno
t ser
ious
nno
t ser
ious
no
nej,o
371/
453
(81.
9%)p
1197
/200
6 (5
9.7%
) OR
2.6
2 (1
.94
to
3.54
)p
198
mor
e pe
r 1,0
00
(from
145
m
ore
to
243
mor
e)
VE
RY l
OW
CRIt
ICAl
Deat
h (fo
llow
up: r
ange
0.5
to 1
0 ye
ars;
ass
esse
d wi
th: a
ll-ca
use
mor
talit
y or
tb m
orta
lity)
5 ob
serv
atio
nal
stud
ies
serio
usa–
f,k,q
–sno
t ser
ious
mse
rious
tse
rious
sno
nej
11/1
91
(5.8
%)
52/7
20 (7
.2%
) OR
0.8
2 (0
.41
to
1.64
)
12 fe
wer
pe
r 1,0
00
(from
41
fewe
r to
41 m
ore)
VE
RY l
OW
CRIt
ICAl
loss
to fo
llow
up (p
revi
ousl
y de
faul
t) (fo
llow
up: r
ange
0.5
to 1
0 ye
ars;
ass
esse
d wi
th: W
HO d
efini
tion)
4 ob
serv
atio
nal
stud
ies
seri-
ousa–
f,uno
t ser
ious
mno
t ser
ious
vno
t ser
ious
no
nej,w
6/15
6
(3.8
%)
77/6
13
(12.
6%)
OR 0
.35
(0.1
5 to
0.
81)x
78 fe
wer
pe
r 1,0
00
(from
21
fewe
r to
105
fewe
r)
VE
RY l
OW
CRIt
ICAl
treat
men
t fai
lure
(fol
low
up: r
ange
0.5
to 1
0 ye
ars;
ass
esse
d wi
th: W
HO d
efini
tion
)
5 ob
serv
atio
nal
stud
ies
serio
us
a–g,
kno
t ser
ious
mno
t ser
ious
vno
t ser
ious
no
nej,w
8/19
1
(4.2
%)
82/7
20
(11.
4%)
OR 0
.38
(0.1
8 to
0.
81)
67 fe
wer
pe
r 1,0
00
(from
20
fewe
r to
91 fe
wer)
VE
RY l
OW
CRIt
ICAl
trans
fer o
ut (f
ollo
w up
: not
repo
rted)
2 ob
serv
atio
nal
stud
ies
serio
us
a–c,
f,y,z
not s
erio
usaa
not s
erio
us
not s
erio
usaa
none
z,bb
0/13
9
(0.0
%)
6/30
5
(2.0
%)
not
estim
able
VERY
lOW
CR
ItIC
Al
4343
ANNE
X 4:
GRA
DE tA
blEs
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
QU
ALI
TYIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NSSU
RGER
YNO
SUR
GERY
RELA
TIVE
(9
5% C
L)AB
SOLU
TE
(95%
CL)
Rela
pse
or re
laps
e/fa
ilure
– n
ot re
porte
d
- -
- -
- -
- -
- -
see
com
men
t -
Adve
rse
even
ts fr
om s
urge
ry (f
ollo
w up
: ran
ge 1
.5 to
10
year
s)
1 ob
serv
atio
nal
stud
ies
serio
usa,
b,f
not s
erio
uscc
not s
erio
us
not s
erio
uscc
publ
icat
ion
bias
stro
ngly
su
spec
teddd
2/66
(3%
) sur
gica
l pat
ient
s di
ed d
ue
to s
urgi
cal c
ompl
icat
ions
.
VERY
lOW
CR
ItIC
Al
Cl: C
onfid
ence
lim
its; O
R: O
dds
ratio
a Do
not a
ddre
ss o
r adj
ust f
or c
onfo
unde
rs a
nd s
ome
stud
ies
do n
ot fu
lly d
escr
ibe
the
popu
latio
n –
Drav
niec
e, e
t al.
2009
; Kar
agoz
, et a
l. 20
09; K
im e
t al.
2007
; Kwa
k et
al.
2015
; Kwo
n et
al.
2008
; Mitn
ick,
et a
l. 20
08; s
hean
, et a
l. 20
08; s
klyu
ev, e
t al.
2013
; tah
aogl
u, e
t al.
2001
; and
toru
n, e
t al.
2007
.b R
etro
spec
tive
obse
rvat
iona
l stu
dies
do
not h
ave
rand
omiza
tion
and
have
inhe
rent
bia
s in
who
is o
ffere
d su
rger
y –
Drav
niec
e, e
t al.
2009
; Kar
agoz
, et a
l. 20
09; K
esha
vjee
, et a
l. 20
08; K
im e
t al.
2007
; Kim
, et a
l. 20
08; K
wak,
et a
l. 20
15; K
won,
et a
l. 20
08; l
eim
ane,
et a
l. 20
05; M
itnic
k, e
t al.
2008
; she
an, e
t al.
2008
; tah
aogl
u, e
t al.
2001
; and
toru
n, e
t al.
2007
.c U
ncer
tain
ty in
repr
esen
tativ
enes
s of
stu
dy p
opul
atio
n –
Drav
niec
e, e
t al.
2009
; Kar
agoz
, et a
l. 20
09; K
im e
t al.
2007
; Kwa
k, e
t al.
2015
; Kwo
n et
al.
2008
; she
an, e
t al.
2008
; and
taha
oglu
, et a
l. 20
01d N
o es
timat
e of
var
iabi
lity
give
n –
Drav
niec
e, e
t al.
(200
9) a
nd ta
haog
lu, e
t al.
(200
1).
e Num
ber o
f “lo
st to
follo
w-up
” rep
orte
d, b
ut c
hara
cter
istic
s no
t des
crib
ed –
taha
oglu
, et a
l. (2
001)
.f l
engt
h of
follo
w up
not
des
crib
ed o
r adj
uste
d fo
r in
anal
ysis
– D
ravn
iece
, et a
l. 20
09; K
im, e
t al.
2007
; Kwa
k, e
t al.
2015
; Kwo
n, e
t al.
2008
; lei
man
e, e
t al.
2005
; Mitn
ick,
et a
l. 20
08; s
hean
, et a
l. 20
08; t
ahao
glu,
et
al.
2001
; and
toru
n, e
t al.
2007
.g I
n su
rgic
al s
tudi
es, i
t is
not p
ossi
ble
to b
lind
patie
nts
or th
e st
udy
team
. Out
com
e as
sess
ors
coul
d be
blin
ded,
and
is s
omew
hat i
mpo
rtant
for a
sses
sing
cur
e us
ing
smea
r as
an o
utco
me
indi
cato
r. Ho
weve
r, pe
rson
nel
othe
r tha
n th
e di
agno
sing
phy
sici
an, g
ener
ally
con
duct
labo
rato
ry a
sses
smen
t. Fo
r tre
atm
ent s
ucce
ss/f
ailu
re th
ere
is a
risk
of r
epor
ting
bias
due
to la
ck o
f blin
ding
whe
re d
ata
are
prog
ram
mat
ic, a
s th
ere
may
be
over
-repo
rting
due
to p
rogr
amm
atic
targ
ets
and
coul
d be
bia
sed
by k
nowl
edge
of s
urgi
cal s
tatu
s.h M
oder
ate
I-squ
ared
(54.
2%) a
nd o
verla
ppin
g Cl
s be
twee
n st
udie
s, a
nd a
re th
us n
ot d
owng
rade
d.i s
ome
varia
tion
in d
urat
ion
of fo
llow-
up in
out
com
e de
finiti
on, h
owev
er it
is n
ot d
owng
rade
d as
alo
ne it
is n
ot c
lass
ified
as
serio
us fo
r thi
s ou
tcom
e.j A
ll st
udie
s ar
e co
hort
base
d, a
nd th
eref
ore
ther
e m
ay b
e so
me
conf
ound
ing
due
to p
atie
nt a
lloca
tion
to s
urge
ry o
r no
surg
ery.
Patie
nts
who
are
mor
e un
well
may
be
mor
e lik
ely
to b
e re
com
men
ded
for s
urge
ry
(ther
efor
e ca
usin
g un
dere
stim
ate
of e
ffect
size
). Ho
weve
r, th
e m
ost s
ick
are
ofte
n no
t offe
red
surg
ery
as th
ey m
ay b
e to
o un
well
or th
e di
seas
e m
ay b
e to
o di
ssem
inat
ed to
allo
w su
rger
y (th
eref
ore
over
estim
atin
g ef
fect
si
ze).
In a
dditi
on, t
here
may
be
varia
tion
in th
e po
pula
tion
offe
red
surg
ery
by s
ettin
g or
sur
geon
. As
ther
e is
a s
peci
fic w
indo
w fo
r sur
gery
, the
se b
iase
s m
ay h
ave
an im
pact
on
estim
atio
n of
effe
ct s
ize, t
houg
h it
is
uncl
ear w
heth
er th
ey w
ould
bia
s th
e es
timat
ion
in a
par
ticul
ar d
irect
ion,
and
are
a re
flect
ion
of th
e re
ality
of t
he p
atie
nt g
roup
offe
red
surg
ery.
ther
efor
e, th
e re
view
ers
deci
ded
not t
o up
grad
e or
dow
ngra
de th
e ra
ting.
k Rep
orts
num
ber,
but n
ot s
umm
ary
stat
istic
s or
pre
cisi
on fo
r thi
s sp
ecifi
c ou
tcom
e –
leim
ane,
et a
l. (2
005)
and
Mitn
ick,
et a
l. (2
008)
l Abs
tract
onl
y, ou
tcom
e an
d pa
tient
cha
ract
eris
tics
not c
lear
ly d
escr
ibed
– D
ravn
iece
, et a
l. (2
009)
m l
ow I-
squa
red
and
over
lapp
ing
Cls
betw
een
stud
ies,
so
not d
owng
rade
d.n M
ost s
tudi
es fo
llowe
d W
HO o
utco
me
defin
ition
s. s
ome
varia
tion
in d
urat
ion
of fo
llow
up to
ass
ess
outc
ome
but n
ot d
owng
rade
d as
alo
ne is
not
cla
ssifi
ed a
s se
rious
issu
e fo
r thi
s ou
tcom
e.o E
mpt
y lo
wer r
ight
qua
dran
t of f
unne
l plo
t. Ho
weve
r, it
seem
s th
at s
mal
ler (
less
pre
cise
) stu
dies
are
repo
rting
lowe
r effe
ct e
stim
ate
so if
pub
licat
ion
bias
wer
e to
exis
t thi
s wo
uld
sugg
est t
he c
urre
nt e
stim
ate
effe
ct
mea
sure
is c
onse
rvat
ive. P
er p
roto
col,
stud
ies
with
<10
sur
gica
l par
ticip
ants
wer
e ex
clud
ed, t
here
fore
the
very
sm
alle
st o
f stu
dies
wer
e no
t inc
lude
d. P
lot i
s no
t suf
ficie
ntly
asy
mm
etric
al to
rais
e se
rious
con
cern
s, a
nd
any
bias
wou
ld a
ppea
r to
caus
e an
und
eres
timat
e of
effe
ct, t
here
fore
qua
lity
is n
ot d
owng
rade
d.
44
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
p n=1
3 fo
r OR
estim
ates
, but
n=1
1 fo
r num
bers
of p
atie
nts
sum
mar
ized
in th
e ta
ble,
as
only
two
stud
ies
repo
rt ef
fect
est
imat
e ra
ther
than
the
num
ber o
f pat
ient
s wi
th th
e ou
tcom
e an
d th
e de
nom
inat
or.
q In
surg
ical
stu
dies
, it i
s no
t pos
sibl
e to
blin
d pa
tient
s or
stu
dy te
am. O
utco
me
asse
ssor
s co
uld
be b
linde
d, b
ut u
nim
porta
nt in
mor
talit
y ou
tcom
e as
no
subj
ectiv
ity in
ass
essm
ent.
r tim
e pe
riod
of fo
llow
up v
ery
varia
ble,
and
for p
atie
nts
with
follo
w up
for <
2 ye
ars
the
follo
w up
per
iod
is p
oten
tially
insu
ffici
ent f
or m
orta
lity
outc
ome
– sh
ean,
et a
l. (2
008)
and
toru
n, e
t al.
(200
7).
s Po
oled
Cls
cro
ss th
e nu
ll. E
vent
rate
is lo
w an
d po
st h
oc o
ptim
al in
form
atio
n si
ze c
alcu
latio
n in
dica
ted
num
ber i
nclu
ded
in a
sses
smen
t of t
his
outc
ome
is to
o lo
w to
give
suf
ficie
nt p
ower
.t V
aria
tion
betw
een
stud
ies
in o
utco
me
defin
ition
use
d (a
ll-ca
use
vers
us tb
-onl
y). U
ncle
ar/v
aria
ble
perio
d ov
er w
hich
dea
th w
as a
sses
sed
(e.g
. die
d du
ring
treat
men
t, wi
thin
six
mon
ths
of c
ompl
etio
n, o
r afte
r two
ye
ars)
.u I
n su
rgic
al s
tudi
es, i
t is
not p
ossi
ble
to b
lind
patie
nts
or s
tudy
team
. Out
com
e as
sess
ors
coul
d be
blin
ded,
but
whe
re d
ata
are
prog
ram
mat
ic th
ey a
re u
nlik
ely
to b
e. th
is c
ould
intro
duce
und
eres
timat
e in
repo
rting
of
defa
ult,
but t
his
bias
is u
nlik
ely
to v
ary
betw
een
stud
y gr
oups
.v M
ostly
use
WHO
defi
nitio
n, m
inor
var
iatio
n in
defi
nitio
n in
som
e st
udie
s, b
ut s
uffic
ient
ly d
irect
not
to d
owng
rade
.w
OR (s
imila
r to
rela
tive
risk
give
n th
e in
frequ
ency
of t
he e
vent
) is
<0.5
and
the
uppe
r con
fiden
ce li
mit
woul
d st
ill p
rovi
de a
clin
ical
ly s
igni
fican
t ben
efit,
ther
efor
e th
is w
ould
be
cons
ider
ed a
larg
e ef
fect
size
. How
ever
, th
e qu
ality
is n
ot u
pgra
ded
as a
ccor
ding
to G
RADE
met
hodo
logy
this
sho
uld
not b
e do
ne if
the
risk
of b
ias
is s
erio
us.
x n=2
stu
dies
had
no
patie
nts
lost
to fo
llow-
up in
the
surg
ery
grou
p, s
o 0.
5 ha
s be
en a
dded
to a
ll ce
lls in
ord
er th
at a
Cl
can
be c
alcu
late
d. th
e su
mm
ary
OR re
stric
ted
to th
e 2
stud
ies
that
had
at l
east
one
pat
ient
lo
st to
follo
w-up
in e
ach
grou
p is
0.4
7 (9
5% C
l: 0
.18,
1.2
4).
y Alth
ough
repo
rted
sepa
rate
ly, u
nlik
ely
that
cle
ar d
iffer
entia
tion
has
been
mad
e be
twee
n “lo
ss to
follo
w-up
” and
“tra
nsfe
r out
”.z s
uspe
cted
und
erre
porti
ng o
f out
com
e, b
ut u
ncer
tain
as
to h
ow th
is w
ould
impa
ct th
e co
nclu
sion
s.aa
No
pool
ed e
stim
ate,
so
insu
ffici
ent e
vide
nce
to a
sses
s.bb
Onl
y tw
o pu
blic
atio
ns, s
o no
t pos
sibl
e to
ass
ess
publ
icat
ion
bias
, but
give
n ho
w fe
w re
port
this
out
com
e pu
blic
atio
n bi
as m
ay b
e pl
ausi
ble.
cc O
ne s
tudy
and
no
com
para
tor g
roup
so
not p
ossi
ble
to e
stim
ate.
dd l
ikel
y th
at c
ompl
icat
ions
occ
urre
d in
oth
er s
tudi
es, b
ut h
ave
eith
er n
ot b
een
repo
rted
or h
ave
been
incl
uded
in a
ll-ca
use
deat
hs.
4545
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): F
ox G
J, M
itnic
k C
D, B
ened
etti
A, C
han
ED, B
ecer
ra M
, Chi
ang
C-Y
, Kes
havj
ee S
, Koh
W-J,
Shi
raish
i Y, V
iiklep
p P,
Yim
J-J,
Pasv
ol G
, R
ober
t J, S
him
TS,
Shi
n SS
, Men
zies
R (1
1 N
ovem
ber 2
015)
Que
stio
n: E
lect
ive p
artia
l lun
g res
ectio
n co
mpa
red
to n
o su
rger
y for
pat
ient
s on
trea
tmen
t for
MD
R-T
B
Sett
ing:
Whi
ch t
ypes
of s
urge
ry e
ncom
pass
ed (
lobe
ctom
y, se
gmen
tect
omy,
wed
ge re
sect
ion)
? D
efini
tion
of n
on-re
spon
se a
nd a
dver
se o
utco
me
of
surg
ery;
defi
nitio
n of
ext
ensiv
e di
seas
e; h
ow s
peci
aliz
ed w
ere
the
cent
res/
prac
titio
ners
whi
ch p
rovi
ded
surg
ery
(ext
erna
l val
idity
)? U
nder
whi
ch
cond
ition
s to
indi
cate
rese
ctio
n su
rger
y and
whe
n to
cont
rain
dica
te; b
efor
e or a
fter c
ultu
re co
nver
sion.
Bib
liogr
aphy
: Fox
GJ,
Mitn
ick
CD
, Ben
edet
ti A
, Cha
n ED
, Bec
erra
M, C
hian
g C-Y
, et a
l. Su
rger
y as a
n ad
junc
tive t
reat
men
t for
mul
tidru
g-re
sista
nt
tube
rcul
osis:
an in
divi
dual
pat
ient
dat
a met
a-an
alys
is. C
lin In
fect
Dis.
201
6;62
(7):8
87–9
5.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
ELEC
TIVE
PA
RTIA
L
LUNG
RE
SECT
ION
NO S
URGE
RYRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
succ
ess
vers
us tr
eatm
ent f
ailu
re o
r rel
apse
(ass
esse
d wi
th: I
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
)
26a
obse
rvat
iona
l st
udie
sbno
t se
rious
cno
t ser
ious
dno
t ser
ious
eno
t ser
ious
fno
ne
185/
204
(90.
7%)g
1134
/139
8 (8
1.1%
)hOR
2.4
(0
.4 to
15
.6)i
100
mor
e pe
r 100
0 (fr
om 1
74
mor
e to
17
9 fe
wer)
lO
W
CRIt
ICAl
succ
ess
vers
us tr
eatm
ent f
ailu
re o
r rel
apse
or d
eath
(ass
esse
d wi
th: I
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
)
26a
obse
rvat
iona
l st
udie
sbno
t se
rious
cno
t ser
ious
dno
t ser
ious
eno
t ser
ious
fno
ne
185/
214
(86.
4%)j
1134
/170
2 (6
6.6%
)kOR
2.0
(0
.4 to
9.
5)i
133
mor
e pe
r 100
0 (fr
om 2
22
fewe
r to
284
mor
e)
lO
W
CRIt
ICAl
succ
ess
vers
us tr
eatm
ent f
ailu
re o
r rel
apse
or d
eath
or l
oss
to fo
llow-
up (a
sses
sed
with
: Ind
ivid
ual p
atie
nt d
ata
met
a-an
alys
is)
26a
obse
rvat
iona
l st
udie
sbno
t se
rious
cno
t ser
ious
dno
t ser
ious
eno
t ser
ious
fno
ne
185/
229
(80.
8%)l
1134
/219
3 (5
1.7%
)mOR
3.5
(1
.5 to
8.
1)i
272
mor
e pe
r 100
0 (fr
om 9
9 m
ore
to
380
mor
e)
lO
W
CRIt
ICAl
46
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
ELEC
TIVE
PA
RTIA
L
LUNG
RE
SECT
ION
NO S
URGE
RYRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
Deat
h ve
rsus
trea
tmen
t fai
lure
or r
elap
se o
r suc
cess
(ass
esse
d wi
th: I
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
)
26a
obse
rvat
iona
l st
udie
sbno
t se
rious
cno
t ser
ious
dno
t ser
ious
eno
t ser
ious
fno
ne
10/2
14
(4.7
%)
304/
1702
(1
7.9%
) OR
0.6
(0
.2 to
2.
2)i
63 fe
wer
pe
r 100
0 (fr
om 1
37
fewe
r to
145
mor
e)
lO
W
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a 26
stud
ies
incl
ude
18 s
tudi
es w
here
sur
gery
was
per
form
ed, a
nd e
ight
stu
dies
whe
re s
urge
ry w
as n
ot p
erfo
rmed
.b L
imita
tions
. All
data
are
from
obs
erva
tiona
l stu
dies
. the
bac
kgro
und
med
icat
ion
regi
men
and
the
qual
ity o
f sur
gery
and
oth
er c
are
are
expe
cted
to d
iffer
bet
ween
the
stud
ies.
bia
s ex
pect
ed b
ecau
se th
e de
cisi
on to
op
erat
e an
d th
e ty
pe o
f sur
gery
are
usu
ally
clo
sely
link
ed to
pro
gnos
tic fa
ctor
s su
ch a
s se
verit
y/se
rious
ness
of t
he c
ondi
tion,
the
exte
nt o
f res
ista
nce
patte
rn, e
ffect
ivene
ss o
f the
med
ical
opt
ions
ava
ilabl
e an
d th
e pa
tient
resp
onse
to tr
eatm
ent.
c Ris
k of
bia
s. A
ll in
clud
ed s
tudi
es a
re o
bser
vatio
nal,
and
sele
ctio
n bi
as is
a s
ubst
antia
l ris
k. P
atie
nt s
elec
tion
for s
urge
ry m
ay b
e bi
ased
towa
rds
patie
nts
with
mor
e fa
vour
able
pro
gnos
tic fa
ctor
s or
the
oppo
site
. le
ngth
of t
reat
men
t diff
ered
sub
stan
tially
bet
ween
sur
gica
l and
non
-sur
gica
l pat
ient
s, s
ugge
stin
g th
at d
iffer
ence
s in
the
back
grou
nd m
edic
al re
gim
ens
may
als
o af
fect
out
com
es; a
lthou
gh th
is a
nd o
ther
mea
sure
d po
tent
ial c
onfo
unde
rs w
ere
incl
uded
in th
e ad
just
ed a
naly
sis
of e
ffect
.d I
ncon
sist
ency
. bas
ed o
n es
timat
ed I
2 R .
Estim
ates
for t
he fi
rst t
wo o
utco
mes
(suc
cess
ver
sus
treat
men
t fai
lure
or r
elap
se +
/- d
eath
) wer
e ve
ry s
imila
r but
OR
for s
ucce
ss in
crea
ses
when
indi
vidu
als
who
were
lost
to
follo
w-up
wer
e in
clud
ed in
the
anal
ysis
.e I
ndire
ctne
ss. N
o in
dire
ctne
ss e
xpec
ted
give
n th
at a
ll pa
tient
s we
re o
n tre
atm
ent f
or M
DR-/
XDR-
tb. t
he o
utco
mes
(suc
cess
, tre
atm
ent f
ailu
re, r
elap
se a
nd d
eath
) wer
e am
ong
thos
e sc
ored
as
criti
cal b
y th
e Gu
idel
ine
Deve
lopm
ent G
roup
; los
s to
follo
w up
was
not
one
of t
he s
peci
fied
outc
omes
but
is re
leva
nt to
the
ques
tion.
f Im
prec
isio
n. 9
5% c
onfid
ence
lim
its fo
r effe
ct e
stim
ate
appl
ied
with
adj
ustm
ent.
g Poo
led
prop
ortio
n 93
% (8
9%–9
7%).
h Poo
led
prop
ortio
n 77
% (6
9%–8
5%).
i Adj
uste
d ef
fect
est
imat
es. t
he m
etho
d of
adj
ustm
ent w
as o
ne to
one
pro
pens
ity s
core
mat
chin
g be
twee
n su
rgic
al p
atie
nts
and
non-
surg
ical
pat
ient
s, fr
om n
on-s
urgi
cal s
tudi
es.
j Poo
led
prop
ortio
n 90
% (8
6%–9
4%).
k Poo
led
prop
ortio
n 64
% (5
4%–7
3%).
l Poo
led
prop
ortio
n 66
% (6
2%–7
0%).
m P
oole
d pr
opor
tion
51%
(40%
–62%
).
4747
ANNE
X 4:
GRA
DE tA
blEs
Aut
hor(
s): F
ox G
J, M
itnic
k C
D, B
ened
etti
A, C
han
ED, B
ecer
ra M
, Chi
ang
C-Y
., K
esha
vjee
S, K
oh W
-J, S
hira
ishi Y
, Viik
lepp
P, Y
im J-
J, Pa
svol
G,
Rob
ert J
, Shi
m T
S, S
hin
SS, M
enzi
es R
Que
stio
n: E
lect
ive p
neum
onec
tom
y com
pare
d to
no
surg
ery f
or p
atie
nts o
n tr
eatm
ent o
f MD
R-T
B.
Sett
ing:
Bef
ore o
r afte
r cul
ture
conv
ersio
n; w
hich
com
paris
on gr
oup
wou
ld h
ave (
a) fa
ilure
/ re
lapse
, (b)
failu
re /
relap
se /
deat
h, an
d (c
) fai
lure
/ re
lapse
/ d
eath
/ lo
ss to
follo
w u
p.
Bib
liogr
aphy
: Fox
GJ,
Mitn
ick
CD
, Ben
edet
ti A
, Cha
n ED
, Bec
erra
M, C
hian
g C-Y
, et a
l. Su
rger
y as a
n ad
junc
tive t
reat
men
t for
mul
tidru
g-re
sista
nt
tube
rcul
osis:
an in
divi
dual
pat
ient
dat
a met
aana
lysis
. Clin
Infe
ct D
is. 2
016;
62(7
):887
–95.
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
ELEC
TIVE
PN
EUM
ONEC
-TO
MY
NO S
URGE
RYRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
succ
ess
vers
us tr
eatm
ent f
ailu
re o
r rel
apse
(ass
esse
d wi
th: i
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
)
26a
obse
rvat
iona
l st
udie
s no
t se
rious
bno
t ser
ious
cno
t ser
ious
dno
t ser
ious
eno
ne
72/9
1 (7
9.1%
)f11
34/1
398
(81.
1%)g
OR 0
.8
(0.1
to
6.0)
h
4 fe
wer
pe
r 100
(fr
om 1
5 m
ore
to 5
1 fe
wer)
lO
W
CRIt
ICAl
succ
ess
vers
us tr
eatm
ent f
ailu
re o
r rel
apse
or d
eath
(ass
esse
d wi
th: i
ndiv
idua
l pat
ient
dat
a m
eta-
anal
ysis
)
26a
obse
rvat
iona
l st
udie
s no
t se
rious
bno
t ser
ious
cno
t ser
ious
dno
t ser
ious
eno
ne
72/1
05
(68.
6%)i
1134
/170
2 (6
6.6%
)jOR
0.7
(0
.1 to
3.
0)h
8 fe
wer
pe
r 100
(fr
om 1
9 m
ore
to 5
0 fe
wer)
lO
W
CRIt
ICAl
succ
ess
vers
us tr
eatm
ent f
ailu
re o
r rel
apse
or d
eath
or l
oss
to fo
llow-
up (a
sses
sed
with
: ind
ivid
ual p
atie
nt d
ata
met
a-an
alys
is)
26a
obse
rvat
iona
l st
udie
s no
t se
rious
cno
t ser
ious
cno
t ser
ious
dno
t ser
ious
eno
ne
72/1
17
(61.
5%)k
1134
/219
3 (5
1.7%
)lOR
1.4
(0
.7 to
3.
2)h
83 m
ore
per 1
000
(from
89
fewe
r to
257
mor
e)
lO
W
CRIt
ICAl
48
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
QU
ALI
TY A
SSES
SM
ENT
NO
. O
F PA
TIEN
TSEF
FEC
T
CER
TAIN
TY
OF
EVID
ENC
EIM
PO
RTA
NC
ENO
. OF
STUD
IES
STUD
Y DE
SIGN
RISK
OF
BIAS
INCO
NSIS
TENC
YIN
DIRE
CTNE
SSIM
PREC
ISIO
NOT
HER
CONS
IDER
ATIO
NS
ELEC
TIVE
PN
EUM
ONEC
-TO
MY
NO S
URGE
RYRE
LATI
VE
(95%
CL)
ABSO
LUTE
(9
5% C
L)
Deat
h ve
rsus
suc
cess
or t
reat
men
t fai
lure
or r
elap
se (a
sses
sed
with
: ind
ivid
ual p
atie
nt d
ata
met
a-an
alys
is)
26a
obse
rvat
iona
l st
udie
s no
t se
rious
bno
t ser
ious
cno
t ser
ious
dno
t ser
ious
eno
ne
14/1
05
(13.
3%)
304/
1702
(1
7.9%
) OR
1.8
(0
.6 to
5.
1)h
103
mor
e pe
r 100
0 (fr
om 6
3 fe
wer t
o 34
7 m
ore)
lO
W
CRIt
ICAl
Cl: c
onfid
ence
lim
its; O
R: o
dds
ratio
a 26
stud
ies
incl
ude
18 s
tudi
es w
here
sur
gery
was
per
form
ed, a
nd e
ight
stu
dies
whe
re s
urge
ry w
as n
ot p
erfo
rmed
.b R
isk
of b
ias.
All
incl
uded
stu
dies
are
obs
erva
tiona
l, an
d se
lect
ion
bias
is a
sub
stan
tial r
isk.
Pat
ient
sel
ectio
n fo
r sur
gery
may
be
bias
ed to
ward
s pa
tient
s wi
th m
ore
favo
urab
le p
rogn
ostic
fact
ors
or th
e op
posi
te.
leng
th o
f tre
atm
ent d
iffer
ed s
ubst
antia
lly b
etwe
en s
urgi
cal a
nd n
on-s
urgi
cal p
atie
nts,
sug
gest
ing
that
diff
eren
ces
in th
e ba
ckgr
ound
med
ical
regi
men
s m
ay a
lso
affe
ct o
utco
mes
; alth
ough
this
and
oth
er m
easu
red
pote
ntia
l con
foun
ders
wer
e in
clud
ed in
the
adju
sted
ana
lysi
s of
effe
ct.
c Inc
onsi
sten
cy. b
ased
on
estim
ated
I-sq
uare
d R.
Est
imat
es fo
r the
firs
t two
out
com
es (s
ucce
ss v
ersu
s tre
atm
ent f
ailu
re o
r rel
apse
+/-
dea
th) w
ere
very
sim
ilar b
ut O
R fo
r suc
cess
incr
ease
s wh
en in
divi
dual
s wh
o we
re
lost
to fo
llow-
up w
ere
incl
uded
in th
e an
alys
is.
d Ind
irect
ness
. No
indi
rect
ness
exp
ecte
d gi
ven
that
all
patie
nts
were
on
treat
men
t for
MDR
-/XD
R-tb
. the
out
com
es (s
ucce
ss, t
reat
men
t fai
lure
, rel
apse
and
dea
th) w
ere
amon
g th
ose
scor
ed a
s cr
itica
l by
the
Guid
elin
e De
velo
pmen
t Gro
up; l
oss
to fo
llow
up w
as n
ot o
ne o
f the
spe
cifie
d ou
tcom
es b
ut is
rele
vant
to th
e qu
estio
n.e I
mpr
ecis
ion.
95%
con
fiden
ce li
mits
for e
ffect
est
imat
e ap
plie
d wi
th a
djus
tmen
t.f P
oole
d pr
opor
tion
79%
(71%
–88%
).g P
oole
d pr
opor
tion
77%
(6%
–85%
).h E
ffect
est
imat
es. A
djus
ted
effe
ct e
stim
ates
app
lyin
g on
e to
one
pro
pens
ity s
core
mat
chin
g be
twee
n su
rgic
al p
atie
nts
and
non-
surg
ical
pat
ient
s fro
m n
on-s
urgi
cal s
tudi
es.
i Poo
led
prop
ortio
n 69
% (6
0%–7
8%).
j Poo
led
prop
ortio
n 64
% (5
4%–7
3%).
k Poo
led
prop
ortio
n 62
% (5
4%–7
1%).
l Poo
led
prop
ortio
n 51
% (4
0%–6
2%).
49
AN
NE
X 5
Evid
ence
to
deci
sion
tab
les
1. S
tand
ardi
zed
shor
ter r
egim
ens
vers
us lo
nger
regi
men
s fo
r the
tre
atm
ent
of M
DR
-TB
Popu
latio
n:Ad
ults
or c
hild
ren
with
mul
tidru
g or
rifa
mpi
cin-
resi
stan
t tb
(MDR
/RR-
tb)
Back
grou
nd:
the
inte
rest
in re
duci
ng th
e du
ratio
n of
trea
tmen
t for
MDR
-tb
has
mot
ivate
d a
num
ber o
f ini
tiativ
es to
trea
t pat
ient
s wi
th s
horte
r reg
imen
s un
der p
rogr
amm
atic
as
well
as tr
ial c
ondi
tions
. In
the
past
few
year
s, re
sults
from
thre
e st
udie
s of
pat
ient
s on
sh
orte
r reg
imen
s ha
ve b
een
repo
rted
and
othe
r stu
dies
hav
e be
gun,
incl
udin
g bo
th
obse
rvat
iona
l coh
orts
and
RCt
s in
diff
eren
t set
tings
. Ear
ly re
sults
from
obs
erva
tiona
l st
udie
s in
ban
glad
esh,
Cam
eroo
n an
d Ni
ger u
sing
regi
men
s la
stin
g 12
mon
ths
or
less
hav
e sh
own
muc
h hi
gher
trea
tmen
t suc
cess
com
pare
d wi
th lo
nger
con
vent
iona
l re
gim
ens
when
trea
ting
patie
nts
with
spe
cific
incl
usio
n cr
iteria
. Give
n th
e lim
ited
expe
rienc
e in
the
use
of th
ese
shor
ter M
DR-t
b re
gim
ens,
WHO
’s po
sitio
n ha
s un
til
now
reco
mm
ende
d su
ch re
gim
ens
to o
nly
be u
sed
with
in a
con
text
of o
pera
tiona
l re
sear
ch a
nd u
nder
clo
se m
onito
ring
for e
ffect
ivene
ss a
nd s
afet
y du
ring
and
afte
r th
e en
d of
trea
tmen
t. th
e fir
st fi
ndin
gs fr
om o
ngoi
ng R
Cts
eval
uatin
g th
is re
gim
en in
di
ffere
nt c
ount
ries
are
not e
xpec
ted
befo
re th
e en
d of
201
7.
Inte
rven
tion:
stan
dard
ized
shor
ter r
egim
ens
Com
paris
on:
long
er re
gim
ens
Mai
n ou
tcom
es:
treat
men
t suc
cess
ver
sus
failu
re/r
elap
se; t
reat
men
t suc
cess
ver
sus
failu
re/
rela
pse/
deat
h; tr
eatm
ent s
ucce
ss v
ersu
s fa
ilure
/rel
apse
/dea
th/l
oss
to
follo
w-up
Sett
ing:
Amon
g pa
tient
s wh
o ha
d no
his
tory
of p
revi
ous
treat
men
t with
sec
ond-
line
drug
s; s
horte
r reg
imen
s re
fer t
o th
ose
last
ing
up to
12
mon
ths;
long
er
regi
men
s la
st 1
8 m
onth
s or
mor
e. N
ote
that
the
“long
er re
gim
ens”
gro
up
pool
s da
ta fr
om s
tudi
es th
at d
iffer
ed in
the
com
bina
tion
and
num
ber o
f dr
ugs,
in th
e du
ratio
n of
trea
tmen
t, an
d in
the
use
of a
sta
ndar
dize
d ve
rsus
an
indi
vidu
alize
d ap
proa
ch. H
ence
the
pool
ed e
stim
ates
do
not n
eces
saril
y re
flect
the
outc
omes
ass
ocia
ted
with
the
regi
men
reco
mm
ende
d in
the
2011
W
HO d
rug-
resi
stan
t tb
guid
elin
es.
Pers
pect
ive:
Mor
e ex
tens
ive u
se o
f sho
rter M
DR-t
b re
gim
ens
for p
atie
nts
who
are
elig
ible
, wi
th c
onse
quen
t im
prov
ed p
atie
nt q
ualit
y of
life
thro
ugh
redu
ctio
n of
tre
atm
ent d
urat
ion,
bet
ter a
dher
ence
and
out
com
es, a
nd lo
wer r
esou
rce
use.
50
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Ass
essm
ent
JUD
GEM
ENT
RES
EAR
CH
EVID
ENC
EA
DD
ITIO
NA
L C
ON
SID
ERAT
ION
S
PROBLEMIs
the
prob
lem
a p
riorit
y?
No
P
roba
bly
no
P
roba
bly
yes
Ye
s
Var
ies
D
on’t
know
An e
stim
ated
hal
f a m
illio
n ne
w ca
ses
of
MDR
-tb
emer
ge e
ach
year
nec
essi
tatin
g tre
atm
ent.
Only
abo
ut o
ne fo
urth
of t
hese
we
re re
porte
d to
be
plac
ed o
n tre
atm
ent
in re
cent
yea
rs.
Outc
omes
of M
DR-t
b tre
atm
ent o
n a
glob
al le
vel a
re p
oor w
ith m
uch
loss
to
follo
w up
and
dea
th; o
nly
abou
t one
hal
f of
cas
es h
ave
a su
cces
sful
out
com
e at
th
e en
d of
trea
tmen
t.
MDR
-tb
is a
glo
bal c
halle
nge
and
acce
ss to
trea
tmen
t ofte
n pr
oble
mat
ic, w
ith re
gi-
men
s be
ing
typi
cally
of l
ong
dura
tion,
toxic
and
exp
ensi
ve.
ther
e is
cle
arly
an
inte
rest
in re
duci
ng th
e du
ratio
n of
trea
tmen
t, si
mpl
ifyin
g th
e ad
min
istra
tion
of th
e re
gim
en, a
nd p
rovi
ding
pat
ient
s wi
th a
saf
er c
ombi
natio
n of
m
edic
ines
that
can
cur
e th
em in
the
larg
e m
ajor
ity o
f ins
tanc
es.
DESIRABLE EFFECTS
How
sub
stan
tial a
re th
e de
sira
ble
antic
ipat
ed e
ffec
ts?
tr
ivia
l
sm
all
M
oder
ate
l
arge
Var
ies
D
on’t
know
In c
ontra
st to
long
er re
gim
ens,
sho
rter
MDR
-tb
regi
men
s ha
ve b
een
repo
rted
to
give
rela
pse-
free
cure
rate
s of
ove
r 85%
am
ong
sele
cted
pat
ient
s. th
e ev
iden
ce
sum
mar
ized
for t
he u
pdat
e of
thes
e gu
idel
ines
has
sho
wn s
ucce
ss ra
tios
to
be s
tatis
tical
ly s
igni
fican
tly h
ighe
r am
ong
patie
nts
treat
ed w
ith th
e sh
orte
r reg
imen
co
mpa
red
with
thos
e tre
ated
with
long
er
regi
men
s (e
ven
when
adj
uste
d fo
r cer
-ta
in fa
ctor
s).
Very
few
obse
rvat
ions
are
ava
ilabl
e up
to n
ow o
n th
e pe
rform
ance
of s
horte
r M
DR-t
b re
gim
ens
in th
e pr
esen
ce o
f add
ition
al re
sist
ance
.Ex
clus
ion
crite
rion:
Pre
viou
sly
treat
ed M
DR-t
b pa
tient
s wi
th s
econ
d-lin
e dr
ugs
(this
m
ay n
ot o
nly
be a
fact
or o
f dru
g re
sist
ance
but
bec
ause
thes
e pa
tient
s m
ay d
iffer
in
beha
viou
r, ad
here
nce)
. In
east
ern
Euro
pean
/cen
tral A
sian
set
tings
whe
re re
sist
ance
pa
ttern
s ar
e m
ore
wide
-rang
ing
and
wher
e Ds
t to
som
e of
the
drug
s is
cha
lleng
ing,
th
e re
gim
en m
ay b
e ex
pect
ed to
be
less
effe
ctive
. so
prev
ious
trea
tmen
t with
a re
g-im
en c
onta
inin
g se
cond
-line
dru
gs is
an
excl
usio
n cr
iterio
n (a
ccur
ate
info
rmat
ion
on p
revi
ous
drug
his
tory
may
be
diffi
cult
to g
et fr
om p
atie
nt o
r med
ical
file
s).
UNDESIRABLE EFFECTS
How
sub
stan
tial a
re th
e un
desi
rabl
e an
ticip
ated
ef
fect
s?
lar
ge
M
oder
ate
s
mal
l
triv
ial
V
arie
s
Don
’t kn
ow
Use
of th
e sh
orte
r reg
imen
s ha
s be
en a
ssoc
iate
d wi
th lo
wer l
evel
s of
adv
erse
ev
ents
, eve
n wh
en th
ese
were
col
lect
ed m
ore
syst
emat
ical
ly w
ithin
a fr
amew
ork
of
oper
atio
nal r
esea
rch.
Ga
tiflox
acin
, the
fluo
roqu
inol
one
of c
hoic
e fo
r the
sho
rter M
DR-t
b re
gim
en, u
ntil
rece
ntly
was
repo
rted
to b
e as
soci
ated
with
dys
glyc
aem
ia in
eld
erly
pat
ient
s tre
ated
for c
ondi
tions
oth
er th
an tb
. sin
ce th
en g
atifl
oxac
in w
as s
hown
to n
ot
incr
ease
dys
glyc
aem
ia w
hen
used
as
part
of fo
ur-m
onth
regi
men
s fo
r tb
treat
men
t. th
e be
nefit
s fo
r its
use
are
exp
ecte
d to
out
weig
h th
e ris
ks w
hen
the
drug
has
a
mai
nsta
y ro
le in
the
treat
men
t of a
con
ditio
n as
ser
ious
as
MDR
-tb.
the
shor
ter r
egim
ens
do n
ot in
clud
e a
num
ber o
f dru
gs th
at a
re m
ost o
ften
asso
ci-
ated
with
ser
ious
or d
istre
ssin
g ad
vers
e ev
ents
(suc
h as
cyc
lose
rine,
PAs
, lin
ezol
id).
thes
e ca
n th
us b
e re
serv
ed to
be
used
as
part
of a
sal
vage
regi
men
sho
uld
the
patie
nt n
ot re
spon
d to
a s
horte
r reg
imen
.
CERTAINTY OF EVIDENCE
Wha
t is
the
over
all c
erta
inty
of t
he e
vide
nce
of e
ffec
ts?
V
ery
low
l
ow
M
oder
ate
H
igh
N
o in
clud
ed s
tudi
es
All d
ata
anal
ysed
for t
his
upda
te w
ere
deriv
ed fr
om o
bser
vatio
nal s
tudi
es. t
he
resu
lts o
f ran
dom
ized
cont
rolle
d tri
al
data
are
not
exp
ecte
d be
fore
the
end
of
2017
.
5151
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
JUD
GEM
ENT
RES
EAR
CH
EVID
ENC
EA
DD
ITIO
NA
L C
ON
SID
ERAT
ION
S
VALUES
Is th
ere
impo
rtan
t unc
erta
inty
abo
ut o
r var
iabi
lity
in
how
muc
h pe
ople
val
ue th
e m
ain
outc
omes
?
Impo
rtant
unc
erta
inty
or v
aria
bilit
y
Pos
sibl
y im
porta
nt u
ncer
tain
ty o
r var
iabi
lity
P
roba
bly
no im
porta
nt u
ncer
tain
ty o
r var
iabi
lity
N
o im
porta
nt u
ncer
tain
ty o
r var
iabi
lity
N
o kn
own
unde
sira
ble
outc
omes
BALANCE OF EFFECTS
Does
the
bala
nce
betw
een
desi
rabl
e an
d un
desi
rabl
e ef
fect
s fa
vour
the
inte
rven
tion
or th
e co
mpa
rison
?
Fav
ours
the
com
paris
on
P
roba
bly
favo
urs
the
com
paris
on
D
oes
not f
avou
r eith
er th
e in
terv
entio
n or
the
c
ompa
rison
Pro
babl
y fa
vour
s th
e in
terv
entio
n
Fav
ours
the
inte
rven
tion
V
arie
s
Don
’t kn
ow
RESOURCES REQUIRED
How
larg
e ar
e th
e re
sour
ce re
quire
men
ts (
cost
s)?
l
arge
cos
ts
M
oder
ate
cost
s
Neg
ligib
le c
osts
and
sav
ings
Mod
erat
e sa
ving
s
lar
ge s
avin
gs
V
arie
s
Don
’t kn
ow
No re
sear
ch e
vide
nce
was
iden
tified
.
CERTAINTY OF EVIDENCE OF REQUIRED RESOURCES
Wha
t is
the
cert
aint
y of
the
evid
ence
of r
esou
rce
requ
irem
ents
(co
sts)
?
Ver
y lo
w
low
Mod
erat
e
Hig
h
No
incl
uded
stu
dies
No re
sear
ch e
vide
nce
was
iden
tified
.
52
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
JUD
GEM
ENT
RES
EAR
CH
EVID
ENC
EA
DD
ITIO
NA
L C
ON
SID
ERAT
ION
S
COST EFFECTIVENESS
Does
the
cost
-eff
ectiv
enes
s of
the
inte
rven
tion
favo
ur
the
inte
rven
tion
or th
e co
mpa
rison
?
Fav
ours
the
com
paris
on
P
roba
bly
favo
urs
the
com
paris
on
D
oes
not f
avou
r eith
er th
e in
terv
entio
n or
the
com
paris
on
P
roba
bly
favo
urs
the
inte
rven
tion
F
avou
rs th
e in
terv
entio
n
Var
ies
N
o in
clud
ed s
tudi
es
No re
sear
ch e
vide
nce
was
iden
tified
.
EQUITY
Wha
t wou
ld b
e th
e im
pact
on
heal
th e
quity
?
Red
uced
Pro
babl
y re
duce
d
Pro
babl
y no
impa
ct
P
roba
bly
incr
ease
d
Incr
ease
d
Var
ies
D
on’t
know
No re
sear
ch e
vide
nce
was
iden
tified
.Al
thou
gh n
o re
liabl
e da
ta a
re a
vaila
ble
on th
e co
sts
of s
horte
r tb
regi
men
s, it
is
expe
cted
that
bot
h dr
ug c
osts
and
pro
gram
me
cost
s wo
uld
not b
e hi
gher
than
lo
nger
regi
men
s. th
is w
ould
mea
n th
at m
ore
reso
urce
s wo
uld
be a
vaila
ble
for t
he
treat
men
t of m
ore
patie
nts.
ACCEPTABILITY
Is th
e in
terv
entio
n ac
cept
able
to k
ey s
take
hold
ers?
N
o
Pro
babl
y no
Pro
babl
y ye
s
Yes
V
arie
s
Don
’t kn
ow
the
shor
ter M
DR-t
b re
gim
ens
have
bee
n su
cces
sful
ly im
plem
ente
d in
a n
umbe
r of
setti
ngs
in A
frica
and
Asi
a in
rece
nt y
ears
thro
ugh
the
effo
rts o
f a n
umbe
r of t
echn
i-ca
l age
ncie
s an
d na
tiona
l pro
gram
mes
. the
inte
rven
tion
is a
ccep
tabl
e to
clin
icia
ns
and
patie
nts.
5353
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
JUD
GEM
ENT
RES
EAR
CH
EVID
ENC
EA
DD
ITIO
NA
L C
ON
SID
ERAT
ION
S
FEASIBILITY
Is th
e in
terv
entio
n fe
asib
le to
impl
emen
t?
No
P
roba
bly
no
P
roba
bly
yes
Ye
s
Var
ies
D
on’t
know
the
inte
rven
tion
has
been
suc
cess
fully
impl
emen
ted
unde
r a n
umbe
r of s
ettin
gs,
and
even
sup
porte
d by
maj
or d
onor
s su
ch a
s th
e Gl
obal
Fun
d to
Fig
ht A
IDs,
tb a
nd
Mal
aria
.su
pply
of c
lofa
zimin
e, w
hich
is in
dica
ted
as a
lepr
osy
drug
, is
a pr
oble
m in
lat
in
Amer
ica
and
else
wher
e.
No q
ualit
y-as
sure
d so
urce
of g
atifl
oxac
in –
a c
heap
fluo
roqu
inol
one
(whi
ch w
as th
e co
rner
ston
e of
the
shor
ter M
DR-t
b re
gim
en u
ntil
rela
tivel
y re
cent
ly),
is a
vaila
ble
toda
y. th
ere
has
been
a g
loba
l sho
rtage
in m
anuf
actu
ring
follo
wing
the
repo
rted
risk
of a
ssoc
iate
d dy
sgly
caem
ia. t
his
has
sinc
e be
en s
hown
to b
e m
uch
less
ser
ious
an
d th
e be
nefit
s wo
uld
likel
y ou
twei
gh ri
sks
when
the
drug
is u
sed
to tr
eat a
con
di-
tion
as s
erio
us a
s M
DR-t
b.Ho
weve
r, th
e W
HO re
com
men
datio
n fo
r the
use
of s
horte
r MDR
-tb
regi
men
and
an
upd
ate
of th
e W
HO M
odel
Ess
entia
l Med
icin
es l
ist (
whic
h as
yet
doe
s no
t fea
-tu
re c
lofa
zimin
e an
d ga
tiflox
acin
as
tb d
rugs
) will
be
expe
cted
to h
ave
a fa
vour
able
im
pact
on
drug
man
ufac
ture
rs a
nd fu
el th
eir i
nter
est t
o in
vest
in th
e pr
oduc
tion
of
thes
e tw
o dr
ugs.
Con
clus
ions
SH
OU
LD S
TAN
DA
RD
IZED
SH
OR
TER
REG
IMEN
S B
E U
SED
FO
R T
HE
TREA
TMEN
T O
F M
DR
-TB
IN
STE
AD
OF
LON
GER
REG
IMEN
S (
ALL
CA
SES
; R
EGA
RD
LES
S O
F P
YR
AZ
I-N
AM
IDE
OR
FLU
OR
OQ
UIN
OLO
NE
SU
SC
EPTI
BIL
ITY)?
Type
of re
com
men
dati
onst
rong
reco
mm
enda
tion
agai
nst t
he in
terv
entio
n
Cond
ition
al re
com
men
datio
n ag
ains
t the
inte
rven
tion
Cond
ition
al re
com
men
datio
n fo
r eith
er th
e in
terv
entio
n or
th
e co
mpa
rison
Cond
ition
al re
com
men
datio
n fo
r the
inte
rven
tion
stro
ng re
com
men
datio
n fo
r the
in
terv
entio
n
Rec
omm
enda
tion
In p
atie
nts
with
rifa
mpi
cin-
resi
stan
t tb
or M
DR-t
b wh
o ha
ve n
ot b
een
prev
ious
ly tr
eate
d wi
th s
econ
d-lin
e dr
ugs
and
in w
hom
resi
stan
ce to
fluo
roqu
inol
ones
and
se
cond
-line
inje
ctab
le a
gent
s ha
s be
en e
xclu
ded
or is
con
side
red
high
ly u
nlik
ely,
the
WHO
Gui
delin
e De
velo
pmen
t Gro
up re
com
men
ds th
at a
sho
rter M
DR-t
b re
gi-
men
may
be
used
inst
ead
of a
con
vent
iona
l reg
imen
(con
ditio
nal r
ecom
men
datio
n, v
ery
low
certa
inty
in th
e ev
iden
ce)
Just
ifica
tion
All d
ata
used
to a
sses
s th
e sh
orte
r MDR
-tb
treat
men
t reg
imen
s we
re d
erive
d fro
m o
bser
vatio
nal s
tudi
es. I
ndiv
idua
l pat
ient
dat
a fro
m b
angl
ades
h (s
uppo
rted
by
the
Dam
ien
Foun
datio
n), U
zbek
ista
n (s
uppo
rted
by M
édec
ins
sans
Fro
ntiè
res
(MsF
)) a
nd s
wazil
and
(MsF
) as
well
as a
ggre
gate
d da
ta fr
om s
ub-s
ahar
an A
frica
n co
untri
es (s
uppo
rted
by th
e UN
ION
and
Actio
n Da
mie
n; b
enin
, bur
kina
Fas
o, b
urun
di, C
amer
oon,
Cen
tral A
frica
n Re
publ
ic, D
R Co
ngo,
Nig
er) w
ere
incl
uded
in
the
anal
ysis
. the
se w
ere
com
pare
d wi
th th
e ou
tcom
es o
f pat
ient
s wi
thou
t pre
viou
s ex
posu
re to
sec
ond-
line
tb d
rugs
who
wer
e in
clud
ed in
the
adul
t ind
ivid
ual
patie
nt d
ata
(aIP
D) a
naly
sis.
the
stan
dard
out
com
es u
sed
in th
e in
terv
entio
n an
d co
mpa
rato
r arm
s la
rgel
y co
mpl
ied
with
the
stan
dard
ized
outc
omes
use
d by
tb
prog
ram
mes
.th
e an
alys
es p
erfo
rmed
for t
he u
pdat
e of
the
guid
elin
es s
howe
d th
at p
atie
nts
who
rece
ived
shor
ter M
DR-t
b tre
atm
ent r
egim
ens
had
a st
atis
tical
ly s
igni
fican
t hig
her
likel
ihoo
d of
trea
tmen
t suc
cess
than
thos
e wh
o re
ceive
d lo
nger
con
vent
iona
l reg
imen
s. th
e nu
mbe
r of r
elap
ses
was
very
low,
alth
ough
this
may
hav
e be
en th
e re
sult
of th
e re
lativ
ely
smal
l num
ber o
f pat
ient
s fo
llowe
d up
. As
expe
cted
, tre
atm
ent s
ucce
ss w
as lo
wer i
n pa
tient
s wi
th a
dditi
onal
resi
stan
ce to
pyr
azin
amid
e an
d/or
fluo
-ro
quin
olon
es, e
ven
if in
gen
eral
it re
mai
ned
high
and
exc
eede
d th
at in
the
patie
nts
on in
divi
dual
ized,
long
er re
gim
ens
(alth
ough
the
diffe
renc
es w
ere
not s
tatis
tical
ly
sign
ifica
nt).
54
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Sub
grou
p co
nsid
erat
ions
Until
mor
e ev
iden
ce is
ava
ilabl
e, W
HO re
com
men
ds th
at th
e sh
orte
r MDR
-tb
regi
men
not
be
used
in p
atie
nts
who
have
bee
n pr
evio
usly
trea
ted
with
sec
ond-
line
drug
s fo
r mor
e th
an o
ne m
onth
or w
ho h
ave
know
n re
sist
ance
to m
edic
ines
in th
e re
gim
en. t
his
reco
mm
enda
tion
is s
ubje
ct to
pat
ient
s ha
ving
bee
n te
sted
for i
n vi
tro re
sist
ance
to a
t lea
st fl
uoro
quin
olon
es a
nd th
e in
ject
able
age
nt u
sed
in th
e re
gim
en b
efor
e st
artin
g tre
atm
ent.
In th
e ab
senc
e of
relia
ble
test
ing,
pat
ient
s wh
o ar
e hi
ghly
unl
ikel
y to
be
infe
cted
with
resi
stan
t stra
ins
base
d on
clin
ical
or r
ecen
t rep
rese
ntat
ive s
urve
illan
ce d
ata
may
als
o be
elig
ible
for t
he s
horte
r MDR
-tb
regi
men
.Pe
ople
livin
g wi
th H
IV n
eed
to b
e gi
ven
the
sam
e co
nsid
erat
ion
for t
reat
men
t with
the
shor
ter M
DR-t
b tre
atm
ent r
egim
en a
s pe
ople
who
are
HIV
ser
oneg
ative
.Ch
ildre
n we
re g
ener
ally
exc
lude
d fro
m s
tudi
es o
f sho
rter M
DR-t
b tre
atm
ent r
egim
ens.
How
ever
, the
re is
no
plau
sibl
e bi
olog
ical
reas
on to
bel
ieve
that
thes
e re
gi-
men
s ar
e le
ss e
ffect
ive in
chi
ldre
n th
an in
adu
lts. A
s a
resu
lt, it
is re
com
men
ded
that
chi
ldre
n wi
th p
ulm
onar
y rif
ampi
cin-
resi
stan
t tb/
MDR
-tb
be g
iven
the
sam
e co
nsid
erat
ion
for t
reat
men
t with
a s
horte
r MDR
-tb
treat
men
t reg
imen
as
adul
ts.
Preg
nanc
y wa
s an
exc
lusi
on c
riter
ion
for s
horte
r MDR
-tb
treat
men
t reg
imen
stu
dies
. two
of t
he c
ore
com
pone
nts
of th
e sh
orte
r MDR
-tb
regi
men
s –
the
inje
ctab
le
agen
t and
eth
iona
mid
e (o
r pro
thio
nam
ide)
– a
re u
sual
ly c
ontra
indi
cate
d in
pre
gnan
cy. W
ithho
ldin
g th
ese
med
icin
es fr
om th
e sh
orte
r MDR
-tb
treat
men
t reg
imen
co
uld
howe
ver s
erio
usly
com
prom
ise
its e
ffect
ivene
ss. t
hus
for p
regn
ant w
omen
it is
reco
mm
ende
d th
at a
n in
divi
dual
ized,
long
er re
gim
en b
e us
ed w
hich
can
allo
w th
e in
clus
ion
of fo
ur o
r mor
e ef
fect
ive m
edic
ines
with
no
know
n te
rato
geni
c pr
oper
ties.
Extra
pulm
onar
y di
seas
e. t
he fi
ndin
gs fr
om s
tudi
es o
f sho
rter M
DR-t
b re
gim
en w
ere
limite
d to
pat
ient
s wi
th p
ulm
onar
y di
seas
e, a
nd th
ey c
anno
t be
extra
pola
ted
dire
ctly
to a
ll di
ffere
nt fo
rms
of e
xtra
pulm
onar
y tb
. No
reco
mm
enda
tion
is th
us p
ossi
ble
at th
is s
tage
to u
se th
e sh
orte
r reg
imen
in p
atie
nts
with
ext
rapu
lmon
ary
MDR
-tb.
Resi
stan
ce a
dditi
onal
to is
onia
zid
and
rifam
pici
n. In
pat
ient
s in
fect
ed w
ith s
train
s kn
own
or s
trong
ly s
uspe
cted
of b
eing
resi
stan
t to
one
or m
ore
drug
s in
the
shor
ter M
DR-t
b tre
atm
ent r
egim
en (e
.g. p
yraz
inam
ide)
, it i
s re
com
men
ded
that
the
shor
ter r
egim
en n
ot b
e us
ed u
ntil
mor
e ev
iden
ce b
ecom
es a
vaila
ble
abou
t its
pe
rform
ance
in s
uch
a si
tuat
ion.
Impl
emen
tati
on
cons
ider
atio
nsIn
ord
er to
repr
oduc
e th
e hi
gh c
ure
rate
s ac
hiev
ed b
y th
e st
udie
s in
clud
ed in
the
revi
ews
for t
his
guid
ance
, all
effo
rts n
eed
to b
e m
ade
to a
void
the
acqu
isiti
on o
f ad
ditio
nal r
esis
tanc
e, th
roug
h ca
refu
l sel
ectio
n of
pat
ient
s to
be
enro
lled,
and
effe
ctive
pat
ient
sup
port
to e
nabl
e fu
ll ad
here
nce
to tr
eatm
ent.
It is
reco
mm
ende
d th
at p
atie
nts
be te
sted
for s
usce
ptib
ility
or r
esis
tanc
e to
fluo
roqu
inol
ones
and
to th
e se
cond
-line
inje
ctab
le a
gent
use
d in
the
regi
men
bef
ore
bein
g st
arte
d on
a
shor
ter M
DR-t
b re
gim
en. P
atie
nts
with
stra
ins
resi
stan
t to
any
of th
e tw
o gr
oups
of m
edic
ines
sho
uld
be tr
ansf
erre
d to
trea
tmen
t with
a lo
nger
, ind
ivid
ualis
ed re
gi-
men
. the
ava
ilabi
lity
of re
liabl
e an
d ra
pid
test
s wo
uld
be v
alua
ble
to d
ecid
e (w
ithin
a fe
w da
ys) w
hich
pat
ient
s wo
uld
be e
ligib
le fo
r sho
rter M
DR-t
b re
gim
ens,
and
wh
at m
odifi
catio
ns to
long
er M
DR-t
b re
gim
ens
are
nece
ssar
y ba
sed
on th
e re
sist
ance
det
ecte
d. In
pat
ient
s wi
th c
onfir
med
rifa
mpi
cin-
resi
stan
t tb
or M
DR-t
b, th
e M
tbDR
sl a
ssay
may
be
used
as
the
initi
al te
st, o
ver c
ultu
re a
nd p
heno
typi
c Ds
t, to
det
ect r
esis
tanc
e to
fluo
roqu
inol
ones
and
to th
e se
cond
-line
inje
ctab
le d
rugs
(c
ondi
tiona
l rec
omm
enda
tions
; cer
tain
ty o
f evi
denc
e fo
r dire
ct te
stin
g of
spu
tum
from
low
to m
oder
ate)
. thi
s ap
plie
s to
test
ing
in b
oth
child
ren
and
adul
ts. I
ndire
ct
test
ing
may
incl
ude
biol
ogic
al s
ampl
es fr
om e
xtra
pulm
onar
y si
tes.
Whi
le re
sist
ance
-con
ferri
ng m
utat
ions
to fl
uoro
quin
olon
es d
etec
ted
by th
e M
tbDR
sl a
ssay
are
hi
ghly
cor
rela
ted
with
phe
noty
pic
resi
stan
ce to
oflo
xaci
n an
d le
voflo
xaci
n, th
e co
rrela
tion
with
mox
iflox
acin
and
gat
iflox
acin
is le
ss c
lear
and
the
incl
usio
n of
mox
i-flo
xaci
n or
gat
iflox
acin
in a
MDR
-tb
regi
men
is b
est g
uide
d by
phe
noty
pic
Dst
resu
lts.
In s
ettin
gs in
whi
ch la
bora
tory
cap
acity
for D
st to
fluo
roqu
inol
ones
and
inje
ctab
le a
gent
s is
not
yet
ava
ilabl
e, th
e cl
inic
ian
and
the
tb p
rogr
amm
e m
anag
er w
ould
ne
ed to
dec
ide
on th
e ba
sis
of th
e lik
elih
ood
of re
sist
ance
to th
ese
med
icin
es, i
nfor
med
by
the
patie
nt’s
clin
ical
his
tory
and
rece
nt re
pres
enta
tive
surv
eilla
nce
data
.th
e ev
iden
ce fo
r the
effe
ctive
ness
and
saf
ety
of th
e sh
orte
r MDR
-tb
regi
men
der
ives
from
stu
dies
whe
re th
is tr
eatm
ent w
as a
dmin
iste
red
unde
r fai
rly s
tand
ardi
zed
cond
ition
s wi
th re
lativ
ely
little
var
iatio
n in
the
cont
ent a
nd d
urat
ion.
thus
, the
reco
mm
enda
tion
on th
e us
e of
the
shor
ter M
DR-t
b re
gim
en is
mad
e un
der t
he p
rem
-is
e th
at it
is im
plem
ente
d as
per
the
com
posi
tion
and
dura
tion
used
in th
e ob
serv
atio
nal s
tudi
es. R
epla
cem
ent o
f med
icin
es a
nd p
rolo
ngat
ion/
shor
teni
ng o
f the
du
ratio
n wo
uld
only
be
perm
issi
ble
with
in th
e pa
ram
eter
s ap
plie
d in
thes
e st
udie
s (e
.g. g
atifl
oxac
in re
plac
ed b
y m
oxifl
oxac
in; p
roth
iona
mid
e re
plac
ed b
y et
hion
a-m
ide;
inte
nsive
pha
se p
rolo
nged
up
to s
ix m
onth
s in
cas
e of
no
sput
um c
onve
rsio
n).
two
stap
les
of th
e re
gim
en, c
lofa
zimin
e an
d hi
gh-d
ose
ison
iazid
, may
be
diffi
cult
to p
rocu
re in
som
e co
untri
es. M
oreo
ver,
ther
e ar
e no
goo
d pa
edia
tric
form
ula-
tions
of c
lofa
zimin
e an
d di
vidi
ng th
e ca
psul
e in
to s
mal
ler d
oses
is a
lmos
t im
poss
ible
, mak
ing
dosi
ng in
chi
ldre
n un
certa
in. G
iven
the
glob
al s
horta
ge in
the
supp
ly
of q
ualit
y-as
sure
d ga
tiflox
acin
in re
cent
yea
rs, t
he s
ites
wher
e ob
serv
atio
nal s
tudi
es h
ave
been
con
duct
ed h
ave
had
to s
ubst
itute
this
age
nt w
ith m
oxifl
oxac
in. t
his
has
led
to a
n im
porta
nt in
crea
se in
the
over
all p
rice
of th
e re
gim
en, w
ith m
oxifl
oxac
in ty
pica
lly a
ccou
ntin
g fo
r abo
ut o
ne h
alf o
f ove
rall
drug
cos
ts. t
he im
plem
enta
-tio
n of
thes
e gu
idel
ines
at t
he n
atio
nal l
evel
nee
ds to
ens
ure
that
suf
ficie
nt q
uant
ities
of t
hese
med
icin
es a
re a
vaila
ble
to m
eet t
he d
eman
d an
d th
at n
o st
ock-
outs
oc
cur.
5555
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
Mon
itor
ing
and
eval
uati
onPa
tient
s wh
o re
ceive
a s
horte
r MDR
-tb
treat
men
t reg
imen
nee
d to
be
mon
itore
d du
ring
treat
men
t and
afte
r com
plet
ion
usin
g sc
hedu
les
of re
leva
nt c
linic
al a
nd
labo
rato
ry te
stin
g wh
ich
have
bee
n su
cces
sful
ly a
pplie
d in
the
stud
ies
unde
r fiel
d co
nditi
ons.
the
WHO
fram
ewor
k fo
r act
ive tb
dru
g-sa
fety
mon
itorin
g an
d m
anag
e-m
ent (
aDsM
) nee
ds to
be
appl
ied
to e
nsur
e ap
prop
riate
act
ion
to re
spon
d pr
ompt
ly to
adv
erse
eve
nts
and
an a
ccep
tabl
e le
vel o
f mon
itorin
g fo
r the
m, a
long
side
th
e m
onito
ring
for t
reat
men
t out
com
es.
Res
earc
h pr
ioriti
es• s
tREA
M s
tudy
resu
lts w
ill b
e av
aila
ble
in a
few
year
s bu
t the
pan
el fe
lt co
mfo
rtabl
e to
mak
e th
e co
nditi
onal
reco
mm
enda
tions
.
• the
WHO
Gui
delin
e De
velo
pmen
t Gro
up d
iscu
ssed
the
rese
arch
prio
ritie
s fo
r red
ucin
g th
e du
ratio
n of
MDR
-tb
regi
men
s an
d hi
ghlig
hted
the
follo
wing
prio
ritie
s: –Fu
ture
rese
arch
nee
ds to
incl
ude
the
effe
ctive
ness
/saf
ety
of th
e sh
orte
r MDR
-tb
treat
men
t reg
imen
in s
ubgr
oups
whi
ch h
ave
been
sys
tem
atic
ally
exc
lude
d fro
m
stud
y pr
otoc
ols
(e.g
. chi
ldre
n, p
atie
nts
with
diff
eren
t for
ms
of e
xtra
pulm
onar
y di
seas
e) a
nd in
set
tings
whe
re b
ackg
roun
d re
sist
ance
to d
rugs
oth
er th
an fl
uoro
-qu
inol
ones
and
sec
ond-
line
inje
ctab
le a
gent
s is
hig
h (e
.g. p
yraz
inam
ide
or h
igh-
leve
l iso
niaz
id re
sist
ance
). –Im
plem
enta
tion
rese
arch
on
the
intro
duct
ion
of th
e sh
orte
r MDR
-tb
regi
men
. –st
udie
s on
cos
t effe
ctive
ness
.
2. R
egim
ens
wit
h in
divi
dual
ized
com
posi
tion
and
dur
atio
n fo
r ad
ults
and
chi
ldre
n w
ith
MD
R-T
B
in w
hom
a s
hort
er M
DR
-TB
reg
imen
can
not
be u
sed
Popu
latio
n:Ad
ults
and
chi
ldre
n wi
th ri
fam
pici
n-re
sist
ant t
b/M
DR-t
b in
wh
om a
sho
rter M
DR-t
b re
gim
en c
anno
t be
used
back
grou
nd:
A nu
mbe
r of r
ifam
pici
n-re
sist
ant t
b/M
DR-t
b pa
tient
s (c
hild
ren
and
adul
ts) a
re e
xpec
ted
not t
o be
elig
ible
for t
he s
horte
r MDR
-tb
regi
men
reco
mm
ende
d el
sewh
ere
in th
ese
guid
elin
es. t
hese
in
clud
e pa
tient
s wh
o we
re p
revi
ousl
y tre
ated
with
sec
ond-
line
tb m
edic
ines
for m
ore
than
one
m
onth
, ind
ivid
uals
infe
cted
with
stra
ins
resi
stan
t to
one
or m
ore
drug
s in
the
shor
ter M
DR-t
b re
gim
en, a
nd p
atie
nts
with
ext
rapu
lmon
ary
dise
ase.
In th
ese
patie
nts,
a lo
nger
regi
men
is u
su-
ally
indi
cate
d, w
ith a
com
posi
tion
and
dura
tion
indi
vidu
alize
d to
incr
ease
the
likel
ihoo
d of
the
regi
men
’s ef
fect
ivene
ss a
nd a
chie
ve a
goo
d ba
lanc
e of
exp
ecte
d be
nefit
s to
har
ms.
thes
e re
g-im
ens
have
bee
n in
use
for s
ever
al y
ears
in m
any
diffe
rent
geo
grap
hica
l set
tings
but
thei
r use
ha
s be
en li
mite
d to
pub
lishe
d ob
serv
atio
nal s
tudi
es o
f pat
ient
s fo
llowe
d up
und
er p
rogr
am-
mat
ic c
ondi
tions
, with
onl
y so
litar
y RC
ts d
esig
ned
and
cond
ucte
d to
ass
ess
the
bene
fit /
saf
ety
of th
e lo
nger
regi
men
s.Fo
r the
201
6 up
date
of t
hese
gui
delin
es, W
HO h
as u
sed
thre
e di
ffere
nt s
ourc
es o
f evi
denc
e to
su
mm
arize
the
effe
cts,
nam
ely:
i)
A sy
stem
atic
revi
ew a
nd s
tudy
-leve
l met
a-an
alys
is fo
r the
effe
ct o
f ind
ivid
ual s
econ
d-lin
e dr
ugs
in M
DR-t
b tre
atm
ent (
see
Anne
x 6
for s
umm
ary
of u
npub
lishe
d st
udy)
.ii)
An
indi
vidu
al-p
atie
nt d
ata
anal
ysis
for 9
153
MDR
-tb
patie
nts
near
ly a
ll of
who
m a
re a
dults
(u
p to
201
0) (A
huja
sD,
et a
l. Pl
os M
ed. 2
012;
9(8)
:e10
0130
0).
iii) A
n in
divi
dual
-pat
ient
dat
a an
alys
is fo
r 974
pae
diat
ric M
DR-t
b pa
tient
s (s
ee A
nnex
6 fo
r su
mm
ary
of u
npub
lishe
d st
udy)
.
Inte
rven
tion:
A lo
nger
regi
men
of i
ndiv
idua
lized
com
posi
tion
and
dura
tion
Com
paris
on:
Othe
r
Mai
n ou
tcom
es:
Cure
d/co
mpl
eted
by
the
end
of tr
eatm
ent;
cultu
re c
onve
rsio
n by
six
mon
ths;
trea
tmen
t fai
lure
; rel
apse
; sur
viva
l (or
dea
th);
adve
rse
reac
tions
(sev
erity
, typ
e, o
rgan
cla
ss)
setti
ng:
treat
men
t adm
inis
tere
d to
pat
ient
s in
bot
h ho
spita
l and
am
bula
tory
set
tings
; the
dis
tribu
tion
of th
e st
udie
s an
d da
ta
was
glob
al
Pers
pect
ive:
the
long
er re
gim
en is
rese
rved
for a
dult
and
paed
iatri
c pa
tient
s wh
o ar
e in
elig
ible
for t
he s
horte
r MDR
-tb
regi
men
du
e to
add
ition
al re
sist
ance
, ext
rapu
lmon
ary
dise
ase,
or o
ther
co
ntra
indi
catio
ns. t
he d
esig
n of
the
regi
men
’s co
mpo
sitio
n is
revi
sed
to o
ptim
ize th
e us
e of
ava
ilabl
e m
edic
ines
bas
ed
on a
vaila
ble
evid
ence
and
thus
it m
axim
izes
the
likel
ihoo
d of
pa
tient
s ha
ving
a s
ucce
ssfu
l out
com
e at
its
end.
56
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Ass
essm
ent
JUD
GEM
ENT
RES
EAR
CH
EVID
ENC
EA
DD
ITIO
NA
L C
ON
SID
ERAT
ION
SDESIRABLE EFFECTS
How
sub
stan
tial a
re th
e de
sira
ble
antic
ipat
ed e
ffec
ts?
tr
ivia
l
sm
all
M
oder
ate
l
arge
Var
ies
D
on’t
know
treat
men
t of M
DR-t
b in
adu
lts a
nd c
hild
ren
with
lo
nger
sec
ond-
line
regi
men
s is
kno
wn to
incr
ease
th
e lik
elih
ood
of c
ure
and
lowe
r the
risk
of c
hro-
nici
ty a
nd d
eath
(Ahu
ja s
D, e
t al.
Plos
Med
. 20
12;9
(8):e
1001
300;
sed
don
JA, e
t al.
thor
ax.
2014
;69(
5):4
58–6
4.).
Rece
nt re
view
s ha
ve s
hown
th
at s
ucce
ss ra
tios
aver
agin
g to
abo
ut 6
0% in
ad
ults
with
MDR
-tb
and
90%
in c
hild
ren
are
pos-
sibl
e am
ong
patie
nts
treat
ed u
nder
pro
gram
mat
ic
cond
ition
s.A
num
ber o
f the
sec
ond-
line
med
icat
ions
are
ass
oci-
ated
with
und
esira
ble
adve
rse
effe
cts
in b
oth
adul
ts
and
child
ren
with
MDR
-tb,
whi
ch a
t tim
es le
ad to
se
rious
out
com
es a
nd d
isco
ntin
uatio
n or
sub
stan
-tia
l cha
nge
in re
gim
ens
(blo
ss E
, et a
l. In
t J tu
berc
lu
ng D
is. 2
010;
14(3
):275
–81;
sed
don
JA, e
t al.
Jour
nal o
f Inf
ectio
n. 2
013;
66(4
):320
–9; s
ee a
lso
body
of g
uide
lines
(inc
ludi
ng ta
ble
7), G
RADE
tabl
es
in A
nnex
4 a
nd s
umm
arie
s of
unp
ublis
hed
stud
ies
in
Anne
x 6
for m
ore
deta
ils o
n ef
fect
ivene
ss a
nd s
afet
y of
long
er M
DR-t
b re
gim
ens
in a
dults
and
chi
ldre
n).
the
likel
ihoo
d of
suc
cess
is e
xpec
ted
to v
ary
depe
ndin
g on
a n
umbe
r of
patie
nt fa
ctor
s (s
ever
ity o
f dis
ease
, res
ista
nce
patte
rns)
and
hea
lth c
are
serv
ices
(acc
ess
to d
iffer
ent m
edic
atio
ns o
f goo
d-qu
ality
, pat
ient
mon
itor-
ing
and
supp
ort).
UNDESIRABLE EFFECTS
How
sub
stan
tial a
re th
e un
desi
rabl
e an
ticip
ated
eff
ects
?
lar
ge
M
oder
ate
s
mal
l
triv
ial
V
arie
s
Don
’t kn
ow
the
likel
ihoo
d of
har
ms
is e
xpec
ted
to v
ary
depe
ndin
g on
a n
umbe
r of
patie
nt fa
ctor
s (c
omor
bidi
ty, d
isea
se s
ever
ity) a
nd th
e he
alth
inte
rven
-tio
n (c
hoic
e of
dru
gs, p
ill-b
urde
n an
d dr
ug–d
rug
inte
ract
ions
, ade
quac
y of
sa
fety
mon
itorin
g an
d su
ppor
t, op
tions
to s
witc
h dr
ugs
in c
ase
of a
dver
se
reac
tions
). th
e fa
ct th
at lo
nger
regi
men
s ar
e co
mpo
sed
of a
t lea
st fi
ve
med
icat
ions
in th
e in
tens
ive p
hase
incr
ease
s th
e lik
elih
ood
of a
dditi
ve
adve
rse
effe
cts
and
inte
ract
ions
.th
e re
clas
sific
atio
n of
PAs
to G
roup
D3
impl
ies
that
this
med
icat
ion
that
is
ofte
n re
spon
sibl
e fo
r man
y un
desi
rabl
e ef
fect
s wo
uld
be u
sed
less
ofte
n.M
oreo
ver,
it is
exp
ecte
d th
at a
larg
er p
ropo
rtion
of p
atie
nts
will
be p
lace
d on
the
shor
ter M
DR-t
b re
gim
en la
stin
g 9–
12 m
onth
s, w
hich
con
tain
s le
ss
med
icat
ions
ass
ocia
ted
with
maj
or a
dver
se e
ffect
s (c
yclo
serin
e, li
nezo
lid,
PAs;
eth
iona
mid
e /
prot
hion
amid
e lim
ited
to th
e in
tens
ive p
hase
whi
ch is
sh
orte
r tha
n in
mos
t lon
ger r
egim
ens)
.
CERTAINTY OF EVIDENCE
Wha
t is
the
over
all c
erta
inty
of t
he e
vide
nce
of e
ffec
ts?
V
ery
low
l
ow
M
oder
ate
H
igh
N
o in
clud
ed s
tudi
es
Mos
t of t
he d
ata
revi
ewed
for b
oth
effe
ctive
ness
and
sa
fety
der
ive fr
om o
bser
vatio
nal s
tudi
es w
ith li
mite
d po
ssib
ilitie
s of
adj
ustm
ent.
the
qual
ity o
f the
evi
-de
nce
from
the
few
rand
omize
d tri
als
was
mar
ked
down
ward
s be
caus
e of
impr
ecis
ion
(sin
gle
stud
ies
with
rela
tivel
y sm
all n
umbe
rs o
f obs
erva
tions
).
Effo
rts w
ere
mad
e to
use
indi
vidu
al-le
vel p
atie
nt d
ata
wher
e po
ssib
le
in b
oth
adul
ts a
nd c
hild
ren
to a
djus
t for
cov
aria
tes
that
cou
ld in
fluen
ce
outc
omes
. How
ever
, res
idua
l con
foun
ding
is v
ery
likel
y to
hav
e be
en s
ub-
stan
tial i
n m
any
of th
e an
alys
es.
5757
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
JUD
GEM
ENT
RES
EAR
CH
EVID
ENC
EA
DD
ITIO
NA
L C
ON
SID
ERAT
ION
S
BALANCE OF EFFECTSDo
es th
e ba
lanc
e be
twee
n de
sira
ble
and
unde
sira
ble
effe
cts
favo
ur th
e in
terv
entio
n or
the
com
paris
on?
F
avou
rs th
e co
mpa
rison
Pro
babl
y fa
vour
s th
e co
mpa
rison
Doe
s no
t fav
our e
ither
the
inte
rven
tion
or th
e co
mpa
rison
Pro
babl
y fa
vour
s th
e in
terv
entio
n
Fav
ours
the
inte
rven
tion
V
arie
s
Don
’t kn
ow
No re
sear
ch e
vide
nce
was
iden
tified
.FEASIBILITY
Is th
e in
terv
entio
n fe
asib
le to
impl
emen
t?
No
P
roba
bly
no
P
roba
bly
yes
Ye
s
Var
ies
D
on’t
know
ther
e ha
s be
en a
ste
ady
incr
ease
in th
e nu
mbe
r of
rifam
pici
n-re
sist
ant t
b/M
DR-t
b pa
tient
s pl
aced
on
seco
nd-li
ne tr
eatm
ent g
loba
lly a
nd re
ports
to W
HO
show
that
in 2
014,
111
000
pat
ient
s we
re s
tarte
d on
trea
tmen
t (Gl
obal
tube
rcul
osis
repo
rt 20
15
(WHO
/HtM
/tb/
2015
.22)
; app
s.wh
o.in
t/iri
s/bi
tstr
eam
/106
65/1
9110
2/1/
9789
2415
6505
9_en
g.pd
f)
long
er, i
ndiv
idua
lized
regi
men
s ha
ve b
een
succ
essf
ully
use
d to
trea
t M
DR-t
b pa
tient
s fo
r the
pas
t few
dec
ades
, und
er a
var
iety
of s
ettin
gs in
bo
th h
igh-
and
low-
reso
urce
situ
atio
ns. E
fforts
in lo
w- a
nd m
iddl
e-in
com
e se
tting
s ha
ve b
een
supp
orte
d th
roug
h do
mes
tic fu
ndin
g an
d al
so e
xter
-na
lly b
y m
ajor
don
ors
like
UsAI
D an
d th
e Gl
obal
Fun
d to
Fig
ht A
IDs,
tb
and
Mal
aria
.th
e av
aila
bilit
y of
cor
e se
cond
-line
dru
gs re
quire
d to
com
pose
the
long
er
regi
men
s ha
s im
prov
ed in
rece
nt y
ears
. the
Glo
bal D
rug
Faci
lity
now
incl
udes
mos
t of t
he d
rugs
on
its c
atal
ogue
. the
pric
e of
the
drug
s ha
s al
so d
ecre
ased
ove
r tim
e, in
clud
ing
that
of l
inez
olid
and
mox
iflox
acin
, as
the
gene
ric m
anuf
actu
re o
f the
se a
gent
s ha
s in
crea
sed
(mos
t dru
gs in
Gr
oups
A to
D w
ith th
e ex
cept
ion
of th
e ne
w dr
ugs
beda
quili
ne a
nd d
ela-
man
id a
re n
ow o
ff pa
tent
).th
ere
are
still
cha
lleng
es in
the
proc
urem
ent o
f cer
tain
dru
gs. b
edaq
uilin
e an
d de
lam
anid
rem
ain
expe
nsive
alth
ough
initi
ative
s ha
ve s
ucce
ssfu
lly
dona
ted
beda
quili
ne in
the
past
few
year
s an
d m
ade
dela
man
id a
vaila
ble
at a
lowe
r cos
t to
low-
inco
me
coun
tries
. the
re is
no
qual
ity-a
ssur
ed s
ourc
e of
gat
iflox
acin
(a c
heap
late
r-gen
erat
ion
fluor
oqui
nolo
ne),
whic
h is
ava
ila-
ble
toda
y gi
ven
a gl
obal
sho
rtage
in m
anuf
actu
re fo
llowi
ng a
repo
rted
risk
of a
ssoc
iate
d dy
sgly
cem
ia (t
his
risk
has
sinc
e be
en s
hown
to b
e m
uch
lowe
r in
tb p
atie
nts
and
bene
fits
woul
d lik
ely
outw
eigh
risk
s wh
en th
e dr
ug
is u
sed
to tr
eat a
con
ditio
n as
ser
ious
as
MDR
-tb)
. Clo
fazim
ine
supp
lies
are
also
lim
ited
and
this
dru
g is
indi
cate
d pr
imar
ily fo
r lep
rosy
and
use
d “o
ff-la
bel”
for t
he tr
eatm
ent o
f MDR
-tb.
bot
h cl
ofaz
imin
e an
d ga
tiflox
acin
do
not
as
yet f
eatu
re o
n th
e W
HO M
odel
lis
ts o
f Ess
entia
l Med
icin
es a
s tb
dru
gs.
the
prog
ram
mat
ic m
anag
emen
t of d
rug-
resi
stan
t tb
has
beco
me
a m
ain-
stay
com
pone
nt o
f man
y na
tiona
l tb
prog
ram
mes
and
sev
eral
cou
ntrie
s ha
ve s
ucce
ssfu
lly in
trodu
ced
a so
und
mon
itorin
g fra
mew
ork
to fo
llow
up
patie
nts
for r
espo
nse
to tr
eatm
ent a
nd to
saf
egua
rd p
atie
nt s
afet
y.
58
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Con
clus
ions
SH
OU
LD A
REG
IMEN
OF
IND
IVID
UA
LIZ
ED C
OM
PO
SIT
ION
AN
D D
UR
ATIO
N B
E U
SED
IN
AD
ULT
S A
ND
CH
ILD
REN
WIT
H R
IFA
MP
ICIN
-RES
ISTA
NT
TB/
MD
R-T
B I
N W
HO
M
A S
HO
RTE
R M
DR
-TB
REG
IMEN
CA
NN
OT
BE
USED
?
Type
of
reco
mm
enda
tion
stro
ng re
com
men
datio
n ag
ains
t th
e in
terv
entio
n
Cond
ition
al re
com
men
datio
n ag
ains
t the
inte
rven
tion
Cond
ition
al re
com
men
datio
n fo
r eith
er th
e in
terv
entio
n or
the
com
paris
on
Cond
ition
al re
com
men
datio
n fo
r th
e in
terv
entio
n
stro
ng re
com
men
datio
n fo
r the
in
terv
entio
n
the
reco
mm
enda
tions
app
ly to
chi
ldre
n an
d ad
ults
with
rifa
mpi
cin-
resi
stan
t tb/
MDR
-tb
(see
als
o un
der s
ubgr
oup
cons
ider
atio
ns).
In c
hild
ren
with
out s
ever
e di
seas
e (s
ever
ity d
efine
d in
the
paed
iatri
c in
divi
dual
pat
ient
dat
a (p
IPD)
on
the
basi
s of
poo
r nut
ritio
nal s
tatu
s, e
xten
sive
dis
ease
on
ches
t rad
iogr
a-ph
y, pr
esen
ce o
f sev
ere
form
s of
ext
rapu
lmon
ary
dise
ase
and
HIV
sero
-pos
itivi
ty),
the
inje
ctab
le a
gent
s m
ay b
e ex
clud
ed fr
om th
e re
gim
en.
Rec
omm
enda
tion
a) In
pat
ient
s wi
th ri
fam
pici
n-re
sist
ant t
b or
MDR
-tb,
the
Guid
elin
e De
velo
pmen
t Gro
up re
com
men
ds a
regi
men
with
at l
east
five
effe
ctive
tb m
edic
ines
dur
ing
the
inte
nsive
ph
ase,
incl
udin
g py
razin
amid
e pl
us fo
ur c
ore
seco
nd-li
ne tb
med
icin
es, o
ne c
hose
n fro
m G
roup
A, o
ne fr
om G
roup
b, a
nd a
t lea
st tw
o fro
m G
roup
C [1
] (co
nditi
onal
reco
mm
en-
datio
n, v
ery
low
certa
inty
in th
e ev
iden
ce).
If th
e m
inim
um o
f effe
ctive
tb m
edic
ines
can
not b
e co
mpo
sed
as a
bove
, an
agen
t fro
m G
roup
D2
and
othe
r age
nts
from
Gro
up D
3 m
ay b
e ad
ded
to b
ring
the
tota
l to
five
[2].
b) In
pat
ient
s wi
th ri
fam
pici
n-re
sist
ant t
b or
MDR
-tb,
the
Guid
elin
e De
velo
pmen
t Gro
up re
com
men
ds th
at th
e re
gim
en b
e fu
rther
stre
ngth
ened
with
hig
h-do
se is
onia
zid a
nd/o
r et
ham
buto
l (co
nditi
onal
reco
mm
enda
tion,
ver
y lo
w ce
rtain
ty in
the
evid
ence
).[1
] Gro
up A
=lev
oflox
acin
, mox
iflox
aci,
gatifl
oxac
in; G
roup
b=a
mik
acin
, cap
reom
ycin
, kan
amyc
in, (
stre
ptom
ycin
); Gr
oup
C= e
thio
nam
ide
(or p
roth
iona
mid
e), c
yclo
serin
e (o
r ter
-izi
done
), lin
ezol
id, c
lofa
zimin
e; in
chi
ldre
n wi
th n
on-s
ever
e di
seas
e Gr
oup
b m
edic
ines
may
be
excl
uded
[2] G
roup
D2=
beda
quili
ne, d
elam
anid
; Gro
up D
3=p-
amin
osal
icyl
ic a
cid,
Imip
enem
-cila
stat
in, m
erop
enem
, am
oxic
illin
-cla
vula
nate
, (th
ioac
etaz
one)
5959
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
Just
ifica
tion
Desi
rabl
e an
d un
desi
rabl
e ef
fect
s
A. F
luor
oqui
nolo
nes
base
d on
the
evid
ence
revi
ews,
the
GDG
conc
lude
d th
at tr
eatm
ent w
ith la
ter-g
ener
atio
n flu
oroq
uino
lone
s (d
efine
d fo
r the
se g
uide
lines
as
high
-dos
e le
voflo
xaci
n, m
oxifl
oxac
in,
and
gatifl
oxac
in) s
igni
fican
tly im
prov
es tr
eatm
ent o
utco
mes
in a
dults
with
rifa
mpi
cin-
resi
stan
t tb
or M
DR-t
b. th
is g
roup
of d
rugs
is c
onsi
dere
d to
be
the
mos
t im
porta
nt c
ompo
-ne
nt o
f the
cor
e M
DR-t
b re
gim
en a
nd th
e be
nefit
s fro
m th
eir u
se o
utwe
ighs
pot
entia
l ris
ks. t
hey
shou
ld th
eref
ore
alwa
ys b
e in
clud
ed u
nles
s th
ere
is a
n ab
solu
te c
ontra
indi
catio
n fo
r the
ir us
e. th
e or
der o
f pre
fere
nce
for t
he in
clus
ion
of th
e la
ter-g
ener
atio
n flu
oroq
uino
lone
s in
MDR
-tb
regi
men
s is
as
follo
ws: h
igh-
dose
levo
floxa
cin,
mox
iflox
acin
and
gat
i-flo
xaci
n. It
is re
com
men
ded
that
oflo
xaci
n be
pha
sed
out f
rom
MDR
-tb
regi
men
s an
d th
at c
ipro
floxa
cin
is n
ever
use
d du
e to
the
limite
d ev
iden
ce fo
r the
ir ef
fect
ivene
ss. A
lthou
gh
the
pIPD
had
hig
h le
vels
of c
onfo
undi
ng a
nd in
suffi
cien
t num
bers
to a
dequ
atel
y an
alys
e th
e tre
atm
ent e
ffect
of h
igh-
dose
levo
floxa
cin,
mox
iflox
acin
and
gat
iflox
acin
, dat
a fro
m
adul
ts w
ith M
DR-t
b sh
ows
a tre
atm
ent b
enefi
t. th
eref
ore
thes
e re
com
men
datio
ns h
ave
been
ext
rapo
late
d to
chi
ldre
n.
Fluo
roqu
inol
ones
in g
ener
al h
ave
a go
od s
afet
y pr
ofile
and
con
side
ring
the
serio
usne
ss o
f rifa
mpi
cin-
resi
stan
t tb/
MDR
-tb,
the
pote
ntia
l for
dru
g-re
late
d ha
rms
is o
ffset
by
the
bene
fits
from
thei
r use
. Alth
ough
adv
erse
eve
nts
were
poo
rly re
cord
ed, i
n th
e st
udy-
leve
l met
a-an
alys
is, t
he fr
eque
ncy
of s
AEs
(defi
ned
as G
rade
3–4
adv
erse
eve
nts
or m
edi-
cine
s st
oppe
d pe
rman
ently
due
to a
dver
se e
vent
) attr
ibut
ed to
fluo
roqu
inol
ones
was
low
(1.2
%–2
.8%
). M
oxifl
oxac
in c
arrie
s a
risk
of Q
t pr
olon
gatio
n, a
cau
se fo
r con
cern
whe
n us
ed in
com
bina
tion
with
med
icat
ions
that
hav
e a
sim
ilar e
ffect
(inc
ludi
ng b
edaq
uilin
e an
d de
lam
anid
). th
ere
are
fewe
r con
cern
s ab
out t
he c
ardi
otox
icity
of l
evofl
oxac
in a
nd g
at-
iflox
acin
, an
impo
rtant
con
side
ratio
n gi
ven
that
sev
eral
oth
er s
econ
d-lin
e dr
ugs
have
Qt-
prol
ongi
ng p
oten
tial.
Conc
erns
abo
ut d
ysgl
ycae
mia
repo
rted
in 2
006
in p
atie
nts
treat
ed w
ith g
atifl
oxac
in fo
r con
ditio
ns o
ther
than
tb le
d th
e pa
rent
com
pany
to s
top
man
ufac
ture
of t
he d
rug,
and
a
glob
al s
horta
ge in
qua
lity-
assu
red
form
ulat
ions
of t
his
drug
ens
ued.
A tr
ial o
f a fo
ur-m
onth
sta
ndar
dize
d re
gim
en fo
r dru
g-su
scep
tible
tb th
at in
clud
ed g
atifl
oxac
in (4
00 m
g on
ce
daily
) pub
lishe
d in
201
4 re
porte
d no
sig
nific
ant r
isk
of h
yper
glyc
aem
ia a
ssoc
iate
d wi
th e
xpos
ure
to g
atifl
oxac
in. A
lthou
gh a
dver
se e
vent
s we
re p
oorly
reco
rded
the
data
for t
his
revi
ew s
howe
d th
at th
ere
was
a lo
wer r
isk
of s
erio
us a
dver
se e
vent
s (s
AEs;
defi
ned
as G
rade
3–4
adv
erse
eve
nts
or d
rugs
sto
pped
due
to a
dver
se e
vent
) in
patie
nts
taki
ng g
ati-
floxa
cin
(3.6
%) t
han
in th
ose
who
did
not,
incl
udin
g th
ose
rece
ivin
g no
fluo
roqu
inol
ones
(8%
; not
sta
tistic
ally
sig
nific
ant).
the
frequ
ency
of s
AEs
asso
ciat
ed w
ith g
atifl
oxac
in w
as
thus
com
para
ble
to th
e on
e as
soci
ated
with
fluo
roqu
inol
ones
in th
e st
udy-
leve
l met
a-an
alys
is.
B. S
econ
d-lin
e in
ject
able
age
nts
base
d on
the
avai
labl
e ev
iden
ce, s
econ
d-lin
e in
ject
able
age
nts
were
ass
ocia
ted
with
an
incr
ease
d lik
elih
ood
of tr
eatm
ent s
ucce
ss w
hen
incl
uded
in a
long
er M
DR-t
b tre
atm
ent
regi
men
(the
sm
all s
ize o
f the
pop
ulat
ion
not r
ecei
ving
an
inje
ctab
le a
gent
in th
e aI
PD li
mite
d th
e po
wer t
o de
tect
an
impa
ct o
f thi
s cl
ass
of a
gent
s). I
t is
ther
efor
e re
com
men
ded
that
adu
lts w
ith ri
fam
pici
n-re
sist
ant t
b or
MDR
-tb
alwa
ys re
ceive
a s
econ
d-lin
e in
ject
able
age
nt a
s pa
rt of
thei
r reg
imen
unl
ess
ther
e is
an
impo
rtant
con
train
dica
tion.
In c
hil-
dren
with
mild
form
s of
dis
ease
, how
ever
, the
har
ms
asso
ciat
ed w
ith th
is g
roup
of m
edic
atio
ns m
ay o
utwe
igh
pote
ntia
l ben
efits
and
ther
efor
e in
ject
able
age
nts
may
be
excl
uded
in
this
gro
up. t
he G
DG b
ased
this
dec
isio
n up
on th
e ob
serv
atio
n th
at tr
eatm
ent s
ucce
ss in
chi
ldre
n wi
th c
linic
ally
-dia
gnos
ed d
isea
se (w
hich
was
ass
ocia
ted
with
less
sev
ere
clin
ical
man
ifest
atio
ns) w
as in
gen
eral
hig
h an
d di
d no
t diff
er s
igni
fican
tly b
etwe
en p
atie
nts
who
rece
ived
a Gr
oup
b m
edic
atio
n (9
8.1%
) and
thos
e wh
o di
d no
t (93
.5%
). Fo
r ch
ildre
n wi
th a
dditi
onal
resi
stan
ce to
fluo
roqu
inol
ones
, Gro
up b
med
icat
ion
is b
est r
etai
ned.
the
choi
ce o
f whi
ch o
f the
thre
e st
anda
rd a
gent
s to
use
– a
mik
acin
, cap
reom
ycin
or k
anam
ycin
– w
ould
be
dete
rmin
ed b
y th
e lik
elih
ood
of e
ffect
ivene
ss a
nd im
plem
enta
tion
cons
ider
atio
ns. W
hile
stre
ptom
ycin
is n
ot u
sual
ly in
clud
ed w
ith th
e se
cond
-line
dru
gs it
can
be
used
as
the
inje
ctab
le a
gent
of t
he c
ore
MDR
-tb
regi
men
if n
one
of th
e th
ree
othe
r age
nts
can
be u
sed
and
if th
e st
rain
is u
nlik
ely
to b
e re
sist
ant t
o it.
60
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Adve
rse
effe
cts
need
to b
e ca
refu
lly m
onito
red
for w
hile
usi
ng s
econ
d-lin
e in
ject
able
age
nts.
Hea
ring
loss
and
nep
hrot
oxic
ity a
re a
mon
g th
e m
ost f
requ
ent a
nd m
ost s
ever
e si
de e
ffect
s. H
owev
er, s
kin
rash
, hyp
erse
nsiti
vity
and
per
iphe
ral n
ephr
opat
hy m
ay a
lso
occu
r. th
e ris
k of
adv
erse
effe
cts
incr
ease
s wi
th th
e to
tal c
umul
ative
dos
e of
sec
ond-
line
inje
ctab
le a
gent
s, s
o ca
utio
n ha
s to
be
exer
cise
d wh
en g
iven
to p
eopl
e wh
o ha
ve p
revi
ousl
y re
ceive
d th
ese
med
icat
ions
, inc
ludi
ng s
trept
omyc
in a
s pa
rt of
a re
gim
en fo
r dru
g-su
scep
tible
tb. I
n ch
ildre
n es
peci
ally,
hea
ring
loss
can
hav
e a
prof
ound
impa
ct o
n qu
ality
of l
ife, a
ffect
ing
acqu
isiti
on o
f lan
guag
e an
d th
e ab
ility
to le
arn
at s
choo
l.
Alth
ough
adv
erse
eve
nts
are
poor
ly re
porte
d, th
e da
ta fo
r thi
s re
view
foun
d th
at 7
.3%
of a
dult
patie
nts
(10.
1% in
chi
ldre
n) h
ad s
AEs
attri
bute
d to
sec
ond-
line
inje
ctab
le a
gent
s.
In a
stu
dy fo
cuse
d on
hea
ring
loss
in c
hild
ren
with
tb, 2
4% o
f chi
ldre
n tre
ated
for M
DR-t
b wi
th a
n in
ject
able
age
nt h
ad h
earin
g lo
ss a
nd 6
4% o
f chi
ldre
n ha
d pr
ogre
ssio
n of
he
arin
g lo
ss a
fter c
ompl
etin
g it
(in th
is s
tudy
, 30%
of t
he c
hild
ren
were
HIV
-infe
cted
).
C. O
ther
cor
e se
cond
-line
age
nts
Whe
n de
sign
ing
the
core
MDR
-tb
treat
men
t reg
imen
, two
or m
ore
of th
e fo
llowi
ng fo
ur m
edic
ines
are
to b
e in
clud
ed: e
thio
nam
ide
(or p
roth
iona
mid
e), c
yclo
serin
e (o
r ter
-izi
done
), lin
ezol
id a
nd c
lofa
zimin
e, u
sual
ly in
this
ord
er o
f pre
fere
nce,
unl
ess
the
bala
nce
of b
enefi
ts-to
-har
ms
for t
he in
divi
dual
pat
ient
dem
ands
oth
erwi
se. G
roup
C a
gent
s ar
e in
clud
ed to
brin
g th
e to
tal e
ffect
ive s
econ
d-lin
e tb
med
icin
es in
the
core
regi
men
to a
t lea
st fo
ur d
urin
g th
e in
tens
ive p
hase
of t
he re
gim
en. I
n ad
ditio
n, if
pyr
azin
amid
e ca
nnot
be
incl
uded
or c
ount
ed u
pon,
ano
ther
age
nt is
add
ed. E
thio
nam
ide
can
be u
sed
inte
rcha
ngea
bly
with
pro
thio
nam
ide,
and
teriz
idon
e in
stea
d of
cyc
lose
rine.
Give
n th
e la
ck o
f rel
iabl
e Ds
t fo
r dru
gs b
elon
ging
to G
roup
C, t
he c
hoic
e of
whi
ch o
nes
to in
clud
e is
det
erm
ined
by
the
bala
nce
of d
esira
ble
to u
ndes
irabl
e ef
fect
s an
d by
impl
e-m
enta
tion
cons
ider
atio
ns. t
he a
dult
and
paed
iatri
c IP
D m
eta-
anal
yses
sho
wed
an in
crea
se in
the
likel
ihoo
d of
trea
tmen
t suc
cess
whe
n M
DR-t
b tre
atm
ent r
egim
ens
incl
uded
cy
clos
erin
e (m
argi
nally
sta
tistic
ally
sig
nific
ant)
and
ethi
onam
ide/
prot
hion
amid
e (s
tatis
tical
ly s
igni
fican
t onl
y in
adu
lts; i
n th
e pI
PD th
e va
st m
ajor
ity o
f chi
ldre
n di
d no
t rec
eive
et
hion
amid
e or
pro
thio
nam
ide
and
sign
ifica
nce
test
ing
was
ther
efor
e no
t alw
ays
poss
ible
for w
ant o
f a s
uffic
ient
num
ber o
f con
trols
). In
con
trast
to c
yclo
serin
e/te
rizid
one
and
ethi
onam
ide/
prot
hion
amid
e, R
Ct d
ata
from
a fe
w re
cent
stu
dies
are
now
ava
ilabl
e fo
r clo
fazim
ine
and
linez
olid
. lin
ezol
id h
as s
hown
a s
tatis
tical
ly s
igni
fican
t tre
atm
ent b
ene-
fit in
bot
h RC
t an
d co
hort
stud
ies
in a
dult
patie
nts,
with
this
ben
efit b
eing
mos
t pro
noun
ced
in p
atie
nts
with
add
ition
al re
sist
ance
to fl
uoro
quin
olon
es a
nd w
ith X
DR-t
b. b
oth
the
adul
t and
pae
diat
ric IP
D sh
owed
no
sign
ifica
nt in
crea
se in
trea
tmen
t suc
cess
ass
ocia
ted
with
the
use
of c
lofa
zimin
e, w
hile
line
zolid
was
use
d to
o sp
arin
gly
in th
e co
horts
in
clud
ed to
allo
w a
conc
lusi
ve a
naly
sis.
Ethi
onam
ide
and
prot
hion
amid
e ca
use
gast
roin
test
inal
dis
turb
ance
, in
parti
cula
r vom
iting
, whi
ch c
an li
mit
tole
rabi
lity.
Hypo
thyr
oidi
sm m
ay o
ccur
, esp
ecia
lly in
com
bina
tion
with
PA
s. H
ypot
hyro
idis
m is
reve
rsib
le u
pon
cess
atio
n of
dru
gs. A
lthou
gh a
dver
se e
vent
s ar
e po
orly
repo
rted,
the
data
for t
his
revi
ew fo
und
that
8.2
% o
f pat
ient
s ha
d sA
Es d
ue to
et
hion
amid
e or
pro
thio
nam
ide.
Cycl
oser
ine
has
a we
ll-es
tabl
ishe
d as
soci
atio
n wi
th n
euro
psyc
hiat
ric a
dver
se e
ffect
s. H
owev
er, t
he a
IPD
met
a-an
alys
is re
veal
ed lo
w le
vels
of s
AEs,
alth
ough
dat
a on
adv
erse
ev
ents
wer
e po
orly
repo
rted
(4.5
% in
the
stud
y-le
vel m
eta-
anal
ysis
con
duct
ed fo
r thi
s up
date
). A
met
a-an
alys
is p
ublis
hed
in 2
013
com
parin
g th
e ad
vers
e ef
fect
s of
cyc
lose
rine
with
teriz
idon
e fo
und
that
teriz
idon
e ha
d lit
tle to
no
bene
fit o
ver c
yclo
serin
e wi
th re
gard
to a
dver
se e
ffect
s.
Adve
rse
effe
cts
of li
nezo
lid in
clud
e th
rom
bocy
tope
nia
and
anae
mia
. the
se c
an b
e se
vere
and
life
thre
aten
ing,
alth
ough
thes
e ad
vers
e ef
fect
s ar
e re
vers
ible
with
ces
satio
n of
dru
g or
on
som
e oc
casi
ons
with
lowe
ring
the
drug
dos
e (u
sual
ly fr
om 6
00 m
g da
ily to
300
mg
daily
). Ha
emat
olog
ic to
xiciti
es a
re le
ss c
omm
on w
ith c
urre
nt s
trate
gies
of o
nce-
daily
do
sing
. Per
iphe
ral n
euro
path
y m
ay o
r may
not
impr
ove
with
ces
satio
n of
dru
g. th
e ou
tcom
e of
opt
ic n
euro
path
y up
on c
essa
tion
of li
nezo
lid is
less
cle
ar, a
nd s
houl
d be
trea
ted
as a
med
ical
em
erge
ncy.
Give
n th
e po
tent
ially
ser
ious
adv
erse
effe
cts
of li
nezo
lid –
par
ticul
arly
ana
emia
, thr
ombo
cyto
peni
a, la
ctic
aci
dosi
s, p
erip
hera
l neu
ropa
thy
and
optic
neu
-ro
path
y –
the
deci
sion
to u
se li
nezo
lid m
ust b
alan
ce it
s ris
ks a
nd b
enefi
ts a
nd th
e av
aila
bilit
y of
oth
er tb
med
icin
es. D
ue to
the
pote
ntia
l for
sev
ere
adve
rse
even
ts, l
inez
olid
use
ne
eds
to b
e ac
com
pani
ed b
y cl
ose
mon
itorin
g fo
r adv
erse
eve
nts.
Whe
re th
is is
not
pos
sibl
e, li
nezo
lid w
ould
bes
t be
rese
rved
for M
DR-t
b pa
tient
s wh
o ha
ve a
dditi
onal
dru
g re
sist
ance
, or X
DR-t
b pa
tient
s, o
r for
thos
e wh
o ar
e in
tole
rant
to o
ther
com
pone
nts
of th
e co
re re
gim
en.
6161
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
Clof
azim
ine
prob
ably
con
tribu
tes
to th
e st
erili
zing
func
tion
of M
DR-t
b re
gim
ens
wher
e py
razin
amid
e is
not
effe
ctive
. the
sin
gle
rand
omize
d co
ntro
l tria
l, al
thou
gh it
had
ser
i-ou
s m
etho
dolo
gica
l con
cern
s, s
howe
d a
stat
istic
ally
sig
nific
ant t
reat
men
t ben
efit a
ssoc
iate
d wi
th th
e us
e of
clo
fazim
ine.
How
ever
, muc
h of
the
evid
ence
for i
ts e
ffect
in M
DR-t
b is
bas
ed o
n ob
serv
atio
nal s
tudi
es, w
hich
sho
wed
confl
ictin
g or
inco
nclu
sive
find
ings
. One
of t
he m
ain
adve
rse
effe
cts
of c
lofa
zimin
e is
ski
n di
scol
orat
ion/
dark
enin
g, w
hich
may
be
dis
tress
ing
to p
atie
nts.
In th
e RC
t, th
e ad
vers
e ev
ents
repo
rted
were
mos
tly li
mite
d to
ski
n co
nditi
ons
and
disc
olor
atio
n, a
nd d
id n
ot le
ad to
dis
cont
inua
tion
in th
e us
e of
the
drug
. Ove
rall,
sm
all r
ates
of a
dver
se e
vent
s we
re n
oted
in o
bser
vatio
nal s
tudi
es. s
AEs
appe
ar to
be
rela
tivel
y un
com
mon
. the
re h
as b
een
som
e ev
iden
ce th
at c
lofa
zimin
e m
ay
prol
ong
the
Qt in
terv
al, s
o ca
utio
n is
adv
ised
whe
n us
ing
this
med
icat
ion
in c
ombi
natio
n wi
th o
ther
dru
gs a
lso
know
n to
hav
e th
e sa
me
effe
ct.
D. A
dd-o
n ag
ents
this
gro
up o
f med
icin
es in
clud
es d
rugs
that
do
not f
orm
par
t of t
he c
ore
seco
nd-li
ne a
gent
s. It
is s
plit
into
thre
e su
bgro
ups:
Grou
p D1
con
sist
s of
pyr
azin
amid
e, e
tham
buto
l and
hig
h-do
se is
onia
zid. t
hese
age
nts
are
usua
lly a
dded
to c
ore
seco
nd-li
ne m
edic
atio
ns, u
nles
s th
e ris
ks fr
om c
onfir
med
resi
st-
ance
, pill
bur
den,
into
lera
nce
or d
rug–
drug
inte
ract
ions
out
weig
h po
tent
ial b
enefi
ts.
the
aIPD
sho
wed
impr
oved
like
lihoo
d of
suc
cess
(ver
sus
treat
men
t fai
lure
, rel
apse
or d
eath
com
bine
d) in
pat
ient
s wh
o ha
d py
razin
amid
e in
clud
ed in
thei
r reg
imen
s. th
is e
ffect
wa
s si
gnifi
cant
bot
h st
atis
tical
ly a
nd in
abs
olut
e te
rms.
the
pIPD
did
not
sho
w a
sign
ifica
nt tr
eatm
ent e
ffect
with
use
of p
yraz
inam
ide.
In m
any
setti
ngs,
rifa
mpi
cin-
resi
stan
t tb
stra
ins
frequ
ently
hav
e ad
ditio
nal r
esis
tanc
e to
pyr
azin
amid
e (in
the
orde
r of 5
0%–6
0%).
Whi
le it
wou
ld b
e de
sira
ble
to a
void
giv
ing
pyra
zinam
ide
to p
atie
nts
whos
e st
rain
s ar
e re
sist
ant t
o th
e dr
ug, i
t is
ackn
owle
dged
that
relia
ble
Dst
for p
yraz
inam
ide
is v
ery
ofte
n un
avai
labl
e in
reso
urce
-con
stra
ined
set
tings
. Alth
ough
adv
erse
eve
nts
are
poor
ly
repo
rted,
the
data
from
the
stud
y-le
vel m
eta-
anal
ysis
sho
wed
that
2.8
% o
f pat
ient
s wh
o re
ceive
d py
razin
amid
e ha
d sA
Es a
ttrib
uted
to it
. the
bal
ance
of d
esira
ble
to u
ndes
ira-
ble
effe
cts
favo
urs
the
addi
tion
of p
yraz
inam
ide
to th
e co
re s
econ
d-lin
e M
DR-t
b re
gim
en b
y de
faul
t, un
less
resi
stan
ce is
con
firm
ed fr
om re
liabl
e Ds
t, or
ther
e ar
e we
ll-fo
unde
d re
ason
s to
bel
ieve
that
the
stra
in is
resi
stan
t, or
ther
e ar
e ot
her c
ontra
-indi
catio
ns fo
r its
use
, par
ticul
arly
risk
of s
igni
fican
t tox
icity
. As
for t
he d
rugs
from
the
core
regi
men
, if
pyra
zinam
ide
is c
ompr
omis
ed o
r can
not b
e us
ed, m
ore
agen
ts fr
om G
roup
C a
nd s
ubse
quen
tly G
roup
D a
re a
dded
unt
il fiv
e ef
fect
ive d
rugs
are
pre
sent
in th
e in
tens
ive p
hase
of
the
regi
men
.
the
reco
mm
enda
tion
for t
he in
clus
ion
of h
igh-
dose
ison
iazid
in a
dult
MDR
-tb
regi
men
s is
larg
ely
base
d on
evi
denc
e fro
m th
e an
alys
is o
f pIP
D. th
is a
naly
sis
show
ed a
sta
tisti-
cally
sig
nific
ant i
ncre
ased
like
lihoo
d of
trea
tmen
t suc
cess
(ver
sus
treat
men
t fai
lure
, rel
apse
or d
eath
com
bine
d) in
chi
ldre
n wi
th b
acte
riolo
gica
lly c
onfir
med
MDR
-tb,
eve
n af
ter
adju
stm
ent f
or a
ge, H
IV s
tatu
s, s
ex, t
b di
seas
e se
verit
y an
d tre
atm
ent c
entre
(tre
atm
ent w
ith h
igh-
dose
ison
iazid
was
alm
ost e
xclu
sive
ly d
one
in s
outh
Afri
can
site
s). A
n RC
t of
hi
gh-d
ose
ison
iazid
ther
apy
for M
DR-t
b in
adu
lts fo
und
no in
crea
sed
risk
of h
epat
otox
icity
. Add
ition
ally,
hig
h-do
se is
onia
zid w
as v
ery
well
tole
rate
d in
chi
ldre
n wi
th d
rug
susc
epti-
ble
tube
rcul
ous
men
ingi
tis in
a la
rge
coho
rt st
udy
from
the
Wes
tern
Cap
e (v
an to
orn
R, e
t al.
Pedi
atr I
nfec
t Dis
J. 2
014;
33(3
):248
–52)
.
Ison
iazid
is re
com
men
ded
alon
gsid
e a
full
MDR
-tb
regi
men
in p
atie
nts
with
rifa
mpi
cin-
resi
stan
t tb
stra
ins
confi
rmed
or s
uspe
cted
to b
e su
scep
tible
to is
onia
zid. H
igh-
dose
is
onia
zid is
one
of t
he c
ore
com
pone
nts
of th
e sh
orte
r MDR
-tb
treat
men
t reg
imen
. stra
ins
bear
ing
mut
atio
ns in
the
prom
oter
regi
on o
f the
inhA
gen
e m
ay h
ave
a m
inim
um in
hib-
itory
con
cent
ratio
n (M
IC) t
o is
onia
zid, w
hich
is lo
w en
ough
to b
e ov
erco
me
by h
igh-
dose
ison
iazid
; and
in s
uch
setti
ngs
the
drug
may
stil
l add
ben
efit.
Howe
ver,
this
mut
atio
n ha
s be
en a
ssoc
iate
d wi
th h
igh-
leve
l eth
iona
mid
e re
sist
ance
and
ther
efor
e, if
pre
sent
, eth
iona
mid
e (o
r pro
thio
nam
ide)
may
hav
e to
be
repl
aced
in th
e re
gim
en. I
n se
tting
s wi
th e
le-
vate
d pr
eval
ence
of h
igh-
leve
l iso
niaz
id re
sist
ance
ass
ocia
ted
with
kat
G m
utat
ions
, hig
h-do
se is
onia
zid m
ay b
e le
ss e
ffect
ive a
nd th
eref
ore
its ro
utin
e us
e m
ay n
ot b
e wa
rrant
ed.
susc
eptib
ility
to e
thio
nam
ide
(or p
roth
iona
mid
e) is
not
affe
cted
by
thes
e m
utat
ions
and
can
be
used
in c
ombi
natio
n wi
th h
igh-
dose
ison
iazid
if th
e is
onia
zid re
sist
ance
mut
atio
n is
not
kno
wn.
62
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
the
aIPD
did
not
sho
w an
y st
atis
tical
ly s
igni
fican
t ass
ocia
tion
betw
een
use
of e
tham
buto
l and
like
lihoo
d of
suc
cess
. Eth
ambu
tol m
ay c
ause
ocu
lar t
oxic
ity, w
hich
can
be
diffi
cult
to d
iagn
ose
in y
oung
chi
ldre
n, a
lthou
gh th
is ri
sk is
redu
ced
if th
e do
se d
oes
not e
xcee
d re
com
men
ded
limits
(0.5
% o
f sAE
s re
porte
d as
soci
ated
with
the
met
a-an
alys
is c
on-
duct
ed fo
r thi
s re
view
alth
ough
the
repo
rting
of a
dver
se e
vent
s da
ta is
ofte
n in
com
plet
e). s
peci
al c
are
is n
eede
d wh
en re
nal f
unct
ion
is c
ompr
omis
ed. R
ifam
pici
n-re
sist
ant t
b/M
DR-t
b st
rain
s m
ay a
lso
be re
sist
ant t
o et
ham
buto
l, pa
rticu
larly
in th
ose
patie
nts
who
have
bee
n tre
ated
with
this
dru
g pr
evio
usly.
How
ever
Dst
for t
his
drug
is n
ot c
onsi
dere
d re
liabl
e an
d re
prod
ucib
le. t
he p
oten
tial b
enefi
t tha
t eth
ambu
tol m
ay a
dd to
a c
ore
MDR
-tb
regi
men
nee
ds to
be
bala
nced
car
eful
ly w
ith th
e in
conv
enie
nce
of a
ddin
g an
othe
r m
edic
ine
to th
e re
gim
en a
nd th
e ris
ks fo
r ass
ocia
ted
harm
s.
Grou
p D2
is m
ade
up o
f bed
aqui
line
and
dela
man
id, t
wo n
ew d
rugs
that
hav
e be
en re
leas
ed in
rece
nt y
ears
. WHO
has
issu
ed in
terim
pol
icy
on th
e us
e of
thes
e m
edic
ines
in
2013
and
201
4. th
e cu
rrent
gui
delin
es m
ake
no c
hang
e to
the
prev
ious
reco
mm
enda
tions
on
how
beda
quili
ne a
nd d
elam
anid
may
be
adde
d to
a c
ore
MDR
-tb
regi
men
in
adul
ts (n
o re
com
men
datio
n fo
r chi
ldre
n). t
he W
HO p
olic
y on
the
role
of D
2 ag
ents
, inc
ludi
ng th
eir p
oten
tial u
se in
chi
ldre
n, w
as u
nder
revi
ew a
t the
tim
e of
the
prod
uctio
n of
th
ese
guid
elin
es.
Grou
p D3
con
sist
s of
p-a
min
osal
icyl
ic a
cid
(PAs
), im
ipen
em–c
ilast
atin
, mer
open
em, c
lavu
lana
te a
nd th
ioac
etaz
one.
thes
e dr
ugs
are
only
to b
e us
ed w
hen
a M
DR-t
b re
gim
en
with
at l
east
five
effe
ctive
dru
gs (i
.e. p
rimar
ily fo
ur c
ore
seco
nd-li
ne m
edic
ines
plu
s py
razin
amid
e) c
anno
t be
othe
rwis
e co
mpo
sed.
the
aIPD
, as
well
as th
e st
udy-
leve
l met
a-an
alys
is c
ondu
cted
for t
he c
urre
nt g
uide
lines
revi
sion
, fou
nd n
o si
gnifi
cant
effe
ct o
f PAs
on
treat
men
t suc
cess
. In
addi
tion,
PAs
use
is
asso
ciat
ed w
ith a
hig
h fre
quen
cy o
f adv
erse
effe
cts
(12.
2% s
AEs
in th
e m
eta-
anal
ysis
und
erta
ken
for t
his
stud
y). P
As is
thus
rese
rved
for s
ituat
ions
whe
n th
ere
is n
o op
tion
to
use
othe
r dru
gs.
Carb
apen
ems
(imip
enem
–cila
stin
or m
erop
enem
) app
ear t
o be
hyd
roly
zed
mor
e sl
owly
by
M. t
uber
culo
sis
when
com
bine
d wi
th c
lavu
lani
c ac
id. C
lavu
lana
te h
as s
hown
poo
r re
sults
in in
vitr
o st
udie
s an
d in
ear
ly b
acte
ricid
al a
ctiv
ity (E
bA) s
tudi
es. t
he a
IPD
show
ed th
at p
atie
nts
treat
ed w
ith c
lavu
lana
te w
ere
mor
e lik
ely
to h
ave
poor
trea
tmen
t out
-co
mes
, alth
ough
this
may
be
due
to c
onfo
undi
ng b
y th
e hi
gher
like
lihoo
d th
at p
atie
nts
rece
ivin
g th
is d
rug
tend
ed to
hav
e m
ore
seve
re d
isea
se (n
ot a
ll co
nfou
ndin
g co
uld
be
adju
sted
for i
n th
e an
alys
is).
WHO
reco
mm
ends
that
whe
neve
r cla
vula
nate
and
car
bape
nem
s ar
e in
clud
ed in
regi
men
s th
ey a
re to
be
alwa
ys u
sed
toge
ther
. Cla
vula
nate
is o
nly
avai
labl
e as
com
bina
tion
prep
arat
ions
con
tain
ing
amox
icill
in. t
he s
pect
rum
of a
dver
se e
ffect
s as
soci
ated
with
am
oxac
illin
–cla
vula
nate
and
car
bape
nem
s is
to a
larg
e ex
tent
id
entic
al to
that
ass
ocia
ted
with
the
peni
cilli
ns.
thio
acet
azon
e ha
s be
en u
sed
exte
nsive
ly in
the
past
as
part
of fi
rst-l
ine
com
bina
tion
ther
apy
for t
b, b
ased
on
RCt
evid
ence
of e
ffect
ivene
ss. U
se o
f the
dru
g in
tb tr
eatm
ent h
as
howe
ver b
een
rest
ricte
d si
nce
the
early
199
0s d
ue to
the
seve
re s
kin
reac
tions
it c
ause
s, in
clud
ing
stev
ens-
John
son
synd
rom
e an
d to
xic e
pide
rmal
nec
roly
sis
(whi
ch c
an le
ad
to d
eath
, esp
ecia
lly in
peo
ple
livin
g wi
th H
IV),
and
the
wide
spre
ad a
vaila
bilit
y of
saf
er, a
fford
able
alte
rnat
ives
for t
he c
ombi
natio
n tb
regi
men
s. If
thio
acet
azon
e is
bei
ng c
onsi
d-er
ed a
s pa
rt of
a M
DR-t
b tre
atm
ent r
egim
en, c
lose
mon
itorin
g fo
r sev
ere
skin
reac
tions
is re
quire
d an
d it
is im
pera
tive
that
the
patie
nt b
e te
sted
for H
IV, a
nd th
at th
e dr
ug n
ot b
e us
ed if
the
patie
nt is
HIV
ser
opos
itive
.
M. t
uber
culo
sis
is in
trins
ical
ly re
sist
ant t
o th
e m
acro
lide
clas
s of
ant
ibio
tics.
the
evid
ence
revi
ews
for t
he c
urre
nt g
uide
lines
sho
wed
no in
dica
tion
of th
e ef
fect
ivene
ss o
f dru
gs
of th
is c
lass
(cla
rithr
omyc
in, a
zithr
omyc
in),
whic
h ha
ve a
t tim
es b
een
incl
uded
in M
DR-t
b re
gim
ens
in b
oth
adul
ts a
nd c
hild
ren.
In a
dditi
on, t
he a
IPD
show
ed a
n in
crea
sed
risk,
al
thou
gh n
ot s
tatis
tical
ly s
igni
fican
t, fo
r poo
r out
com
es in
pat
ient
s re
ceiv
ing
mac
rolid
es a
lthou
gh m
acro
lides
app
eare
d to
be
safe
in p
rolo
nged
use
. Mac
rolid
es a
re a
ssoc
iate
d wi
th Q
t pr
olon
gatio
n, w
hich
wou
ld b
e of
par
ticul
ar c
once
rn if
pat
ient
s ar
e re
ceiv
ing
othe
r tb
drug
s th
at m
ay h
ave
a si
mila
r ris
k, s
uch
as m
oxifl
oxac
in, c
lofa
zimin
e, b
edaq
uilin
e or
de
lam
anid
. WHO
ther
efor
e re
com
men
ds th
at c
larit
hrom
ycin
and
azit
hrom
ycin
not
be
incl
uded
in M
DR-t
b re
gim
ens.
Adve
rse
effe
cts
of P
As in
clud
e ga
stro
inte
stin
al d
istu
rban
ce a
nd h
ypot
hyro
idis
m (i
n pa
rticu
lar w
hen
give
n in
com
bina
tion
with
eth
iona
mid
e/pr
othi
onam
ide)
. Hyp
othy
roid
ism
is
reve
rsib
le u
pon
cess
atio
n of
the
drug
s. A
lthou
gh a
dver
se e
vent
s ar
e po
orly
repo
rted,
the
data
for t
his
revi
ew fo
und
that
12.
2% o
f pat
ient
s ha
d sA
Es (d
efine
d as
Gra
de 3
–4
adve
rse
even
ts o
r dru
gs s
topp
ed d
ue to
adv
erse
eve
nt) a
ttrib
uted
to P
As. t
he p
IPD
show
ed p
ossi
bilit
y of
trea
tmen
t har
m a
ssoc
iate
d wi
th th
e us
e of
PAs
(not
sta
tistic
ally
sig
nif-
ican
t). H
owev
er, P
As is
freq
uent
ly g
iven
to c
hild
ren
with
few
othe
r tre
atm
ent o
ptio
ns, a
nd th
eref
ore
this
effe
ct m
ay b
e du
e to
con
foun
ding
by
indi
catio
n (s
ites
that
had
poo
rer
outc
omes
with
PAs
als
o ha
d si
gnifi
cant
ly h
ighe
r rat
es o
f chi
ldre
n wh
o we
re H
IV s
erop
ositi
ve, m
alno
uris
hed,
had
sev
ere
pulm
onar
y di
seas
e an
d wh
o ha
d ad
ditio
nal r
esis
tanc
e to
flu
oroq
uino
lone
s an
d th
e se
cond
-line
inje
ctab
le m
edic
ines
).
6363
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
Sub
grou
p co
nsid
erat
ions
Rifa
mpi
cin-
resi
stan
t TB/
MDR
-TB
with
add
ition
al re
sist
ance
to fl
uoro
quin
olon
es, s
econ
d-lin
e in
ject
able
age
nts
and
XDR-
TBIn
rifa
mpi
cin-
resi
stan
t tb/
MDR
-tb
patie
nts
with
con
firm
ed o
r wel
l-fou
nded
bel
ief o
f res
ista
nce
to m
edic
atio
ns fr
om G
roup
A (fl
uoro
quin
olon
es) o
r Gro
up b
(sec
ond-
line
inje
cta-
ble)
, sub
stitu
tion
of d
rugs
from
thes
e cl
asse
s pr
ocee
ds a
s de
taile
d be
low.
If a
ny o
f the
com
pone
nts
of th
e re
gim
en –
the
four
cor
e se
cond
-line
med
icin
es a
nd p
yraz
inam
ide
– is
co
nsid
ered
not
to b
e ef
fect
ive, a
dditi
onal
age
nts
from
Gro
ups
D2 o
r D3
are
adde
d. th
is is
alm
ost a
lway
s ne
cess
ary
when
resi
stan
ce to
bot
h Gr
oups
A a
nd b
dru
gs (i
.e. X
DR-t
b)
is p
rese
nt. A
n an
alys
is o
f ind
ivid
ual d
ata
colle
cted
for t
he u
pdat
e of
the
WHO
dru
g-re
sist
ant t
b tre
atm
ent g
uide
lines
of 2
011
conc
lude
d th
at re
gim
ens
cont
aini
ng m
ore
drug
s we
re a
ssoc
iate
d wi
th th
e hi
ghes
t odd
s of
suc
cess
for M
DR-t
b pa
tient
s wh
o ha
d ad
ditio
nal r
esis
tanc
e to
fluo
roqu
inol
ones
and
/or s
econ
d-lin
e in
ject
able
age
nts.
the
curre
nt W
HO
advi
ce w
hen
desi
gnin
g re
gim
ens
for p
atie
nts
with
resi
stan
ce to
fluo
roqu
inol
ones
, sec
ond-
line
inje
ctab
le m
edic
atio
ns a
nd X
DR-t
b co
ntin
ues
to a
pply.
Acce
ss to
rapi
d di
agno
stic
test
ing
whic
h co
uld
relia
bly
iden
tify
resi
stan
ce to
fluo
roqu
inol
ones
or i
njec
tabl
e m
edic
atio
ns w
ould
hel
p cl
inic
ians
to d
ecid
e ho
w to
mod
ify lo
nger
M
DR-t
b re
gim
ens.
the
Geno
type
Mtb
DRsl
line
pro
be a
ssay
may
now
be
used
as
an in
itial
test
, ove
r phe
noty
pic
cultu
re-b
ased
Dst
, to
dete
ct re
sist
ance
to fl
uoro
quin
olon
es
and
seco
nd-li
ne in
ject
able
dru
gs (c
ondi
tiona
l rec
omm
enda
tion;
cer
tain
ty o
f evi
denc
e lo
w to
mod
erat
e fo
r dire
ct te
stin
g). G
enot
ype
Mtb
DRsl
can
be
used
in b
oth
child
ren
and
adul
ts a
nd a
s a
dire
ct a
nd in
dire
ct te
st (f
or e
xtra
pulm
onar
y sa
mpl
es).
Whi
le re
sist
ance
-con
ferri
ng m
utat
ions
to fl
uoro
quin
olon
es d
etec
ted
by th
e M
tbDR
sl a
ssay
are
hig
hly
corre
-la
ted
with
phe
noty
pic
resi
stan
ce to
oflo
xaci
n an
d le
voflo
xaci
n, th
e co
rrela
tion
with
mox
iflox
acin
and
gat
iflox
acin
is le
ss c
lear
and
the
incl
usio
n of
mox
iflox
acin
or g
atifl
oxac
in in
a
MDR
-tb
regi
men
is b
est g
uide
d by
phe
noty
pic
Dst
resu
lts.
TB o
f the
cen
tral
ner
vous
sys
tem
the
treat
men
t of t
uber
culo
us m
enin
gitis
rela
ted
to ri
fam
pici
n-re
sist
ant o
r MDR
stra
ins
is b
est g
uide
d by
dru
g su
scep
tibili
ty re
sults
and
the
know
n pr
oper
ties
of tb
dru
gs to
pen
e-tra
te th
e ce
ntra
l ner
vous
sys
tem
(CNs
). In
pat
ient
s wi
th ri
fam
pici
n-re
sist
ant t
b/M
DR-t
b m
enin
gitis
, it i
s re
com
men
ded
that
the
med
icat
ions
sel
ecte
d fo
r the
regi
men
hav
e go
od
CNs
pene
tratio
n pr
oper
ties.
the
fluor
oqui
nolo
nes
reco
mm
ende
d by
thes
e gu
idel
ines
hav
e go
od C
Ns p
enet
ratio
n, a
s do
eth
iona
mid
e (o
r pro
thio
nam
ide)
, cyc
lose
rine
(or t
erizi
done
) and
line
zolid
. Py
razin
amid
e ha
s go
od C
Ns p
enet
ratio
n, a
lthou
gh c
autio
n sh
ould
be
exer
cise
d, a
s a
larg
e pe
rcen
tage
of M
DR-t
b st
rain
s m
ay b
e re
sist
ant.
Ison
iazid
pen
etra
tes
the
CNs
very
we
ll, w
ith h
ighe
r dos
es re
achi
ng a
dequ
ate
MIC
s in
the
cere
bros
pina
l flui
d. D
ue to
its
good
CNs
pen
etra
tion,
hig
h-do
se is
onia
zid is
reco
mm
ende
d as
par
t of t
he tr
eatm
ent r
egi-
men
unl
ess
high
-leve
l res
ista
nce
is k
nown
to e
xist.
PAs
and
etha
mbu
tol d
o no
t pen
etra
te th
e CN
s we
ll an
d sh
ould
not
be
coun
ted
upon
am
ong
the
num
ber o
f effe
ctive
dru
gs to
trea
t MDR
-tb
men
ingi
tis. K
anam
ycin
, am
ikac
in
and
stre
ptom
ycin
onl
y pe
netra
te th
e ce
rebr
ospi
nal fl
uid
in th
e pr
esen
ce o
f men
inge
al in
flam
mat
ion.
ther
e ar
e lit
tle d
ata
on th
e CN
s pe
netra
tion
of c
apre
omyc
in, c
lofa
zimin
e,
beda
quili
ne o
r del
aman
id.
Peop
le li
ving
with
HIV
the
com
posi
tion
of th
e tre
atm
ent r
egim
en fo
r MDR
-tb
does
not
diff
er fo
r peo
ple
livin
g wi
th H
IV. H
owev
er, t
hioa
ceta
zone
sho
uld
not b
e gi
ven
to p
atie
nts
who
are
HIV
posi
tive.
If
thio
acet
azon
e is
bei
ng c
onsi
dere
d as
par
t of a
trea
tmen
t reg
imen
HIV
infe
ctio
n ne
eds
to b
e re
liabl
y ex
clud
ed in
the
patie
nt.
64
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
Impl
emen
tati
on
cons
ider
atio
nsth
e im
plem
enta
tion
of M
DR-t
b ch
emot
hera
py is
feas
ible
und
er p
rogr
amm
atic
con
ditio
ns, a
s ha
s be
en a
mpl
y sh
own
by th
e gl
obal
exp
ansi
on in
the
use
of lo
nger
MDR
-tb
regi
-m
ens
world
wide
, par
ticul
arly
in th
e pa
st d
ecad
e. C
hang
es m
ade
by th
e cu
rrent
revi
sion
to th
e gr
oupi
ng o
f the
med
icin
es a
nd to
the
com
posi
tion
of th
e lo
nger
regi
men
are
not
ex
pect
ed to
hav
e m
ajor
impa
ct o
n th
eir c
ontin
ued
use.
Mos
t of t
he fl
uoro
quin
olon
es a
nd th
e in
ject
able
age
nts
are
read
ily a
vaila
ble,
as
are
the
maj
ority
of t
he G
roup
C a
nd G
roup
D
agen
ts. t
he la
test
WHO
Mod
el l
ists
of E
ssen
tial M
edic
ines
(Aug
ust 2
015)
incl
udes
mos
t of t
he a
gent
s in
Gro
ups
A to
D e
xcep
t for
gat
iflox
acin
and
thio
acet
azon
e. H
owev
er, c
lo-
fazim
ine,
mer
open
em, i
mip
enem
–cila
stat
in a
nd a
mox
icill
in–c
lavu
lana
te a
re li
sted
for i
ndic
atio
ns o
ther
than
tb, w
hile
bed
aqui
line
and
dela
man
id a
re o
nly
incl
uded
in th
e ad
ult
list.
Othe
r spe
cific
fact
ors
impo
rtant
for i
mpl
emen
tatio
n ar
e di
scus
sed
in th
e re
spec
tive
sect
ions
bel
ow.
Whe
re p
ossi
ble
a pa
tient
with
rifa
mpi
cin-
resi
stan
t tb/
MDR
-tb
stra
in n
eeds
to b
e te
sted
for s
usce
ptib
ility
to m
edic
ines
pla
nned
for i
nclu
sion
in th
e re
gim
en. t
he a
vaila
bilit
y of
re
liabl
e te
sts
for s
usce
ptib
ility
to fl
uoro
quin
olon
es a
nd to
the
seco
nd-li
ne in
ject
able
dru
gs (w
hich
wou
ld g
ive re
sults
with
in a
few
days
) is
valu
able
to e
nsur
e th
at lo
nger
MDR
-tb
regi
men
s ar
e st
reng
then
ed a
s ne
cess
ary
(refe
renc
e is
mad
e to
the
reco
mm
enda
tions
on
the
use
of li
ne p
robe
ass
ay fo
r sec
ond-
line
drug
s –
the
Mtb
DRsl
ass
ay).
Whe
re re
liabl
e Ds
t is
not
an
optio
n, p
roof
of t
he e
ffect
ivene
ss o
f a m
edic
ine
need
s to
be
base
d on
a c
aref
ul c
linic
al h
isto
ry o
f the
pat
ient
’s pr
evio
us e
xpos
ure
to th
e m
edic
ine,
of
sign
ifica
nt c
onta
ct w
ith a
noth
er ri
fam
pici
n-re
sist
ant t
b/M
DR-t
b pa
tient
who
se a
ntib
iogr
amm
e is
doc
umen
ted,
and
from
kno
wled
ge o
f the
pre
vale
nt re
sist
ance
pat
tern
s ce
ntre
d on
repr
esen
tativ
e dr
ug-re
sist
ance
sur
veill
ance
. bot
h th
e Ds
t an
d th
e in
divi
dual
clin
ical
his
tory
sho
uld
be c
onsi
dere
d wh
en c
onst
ruct
ing
a tre
atm
ent r
egim
en. t
he o
nly
relia
ble
labo
rato
ry te
sts
for t
b dr
ug s
usce
ptib
ility
(or r
esis
tanc
e) th
at a
re w
idel
y us
ed to
day
are
thos
e fo
r iso
niaz
id, r
ifam
pici
n, fl
uoro
quin
olon
es a
nd s
econ
d-lin
e in
ject
able
age
nts.
A. F
luor
oqui
nolo
nes
both
levo
floxa
cin
and
mox
iflox
acin
are
com
mon
ly u
sed
to tr
eat M
DR-t
b. l
evofl
oxac
in is
mor
e wi
dely
ava
ilabl
e th
an m
oxifl
oxac
in, w
hich
is m
ore
expe
nsive
alth
ough
a re
duct
ion
in
its p
rice
is e
xpec
ted
in th
e co
min
g ye
ars.
Gatifl
oxac
in w
as a
n af
ford
able
dru
g an
d ha
d be
en c
omm
only
use
d by
tb tr
eatm
ent p
rogr
amm
es u
ntil
the
conc
erns
abo
ut it
s dy
sgly
caem
ic e
ffect
s le
d to
a g
loba
l sho
rtage
in
this
med
icin
e. If
man
ufac
ture
of q
ualit
y-as
sure
d fo
rmul
atio
ns o
f the
dru
g re
star
ts, i
t cou
ld s
ubst
antia
lly lo
wer t
he c
osts
of r
egim
ens
by s
ubst
itutin
g m
ore
expe
nsive
opt
ions
in
fluor
oqui
nolo
nes.
Mox
iflox
acin
is re
lativ
ely
easy
to a
dmin
iste
r to
olde
r chi
ldre
n. H
owev
er, t
he ta
blet
mus
t be
split
to a
ccom
mod
ate
dosi
ng in
you
nger
chi
ldre
n an
d it
is h
ighl
y un
pala
tabl
e on
ce s
plit
or c
rush
ed. l
evofl
oxac
in is
ava
ilabl
e as
a s
uspe
nsio
n.B.
Sec
ond-
line
inje
ctab
le a
gent
sth
ese
agen
ts p
rese
nt p
robl
ems
to a
dmin
iste
r par
ente
rally
on
a da
ily b
asis
for s
ever
al m
onth
s, o
ften
nece
ssita
ting
hosp
italiz
atio
n. G
ivin
g in
ject
ions
to c
hild
ren
and
unde
rwei
ght
adul
ts is
par
ticul
arly
pai
nful
and
unw
elco
me.
C. O
ther
age
nts
Ethi
onam
ide
and
prot
hion
amid
e ar
e in
expe
nsive
, rea
dily
ava
ilabl
e wo
rldwi
de a
nd e
asily
adm
inis
tere
d.Cy
clos
erin
e ha
s be
en o
ne o
f the
sta
ndar
d dr
ugs
for t
he tr
eatm
ent o
f MDR
-tb
for s
ever
al y
ears
and
ther
efor
e ex
perie
nce
in it
s us
e is
wid
espr
ead.
It is
inex
pens
ive.
teriz
idon
e is
less
wid
ely
used
but
is a
vaila
ble
on th
e GD
F Pr
oduc
ts l
ist.
Clof
azim
ine
is in
expe
nsive
but
it c
an b
e di
fficu
lt to
pro
cure
.th
e im
plem
enta
tion
of th
ese
guid
elin
es a
t nat
iona
l lev
el n
eeds
to e
nsur
e th
at s
uffic
ient
qua
ntiti
es o
f the
se m
edic
ines
are
ava
ilabl
e to
mee
t the
dem
and
and
that
no
stoc
k-ou
ts
occu
r. M
oreo
ver,
give
n th
at th
ere
are
no g
ood
paed
iatri
c fo
rmul
atio
ns th
e ca
psul
e co
nten
ts n
eed
to b
e ex
pres
sed
man
ually
and
div
ided
into
sm
alle
r dos
es, w
ith ri
sks
of in
corre
ct
dosi
ng in
chi
ldre
n.W
hen
linez
olid
is u
sed,
ther
e ne
eds
to b
e cl
ose
mon
itorin
g fo
r sid
e ef
fect
s, p
artic
ular
ly a
naem
ia, t
hrom
bocy
tope
nia,
lact
ic a
cido
sis,
per
iphe
ral n
euro
path
y an
d op
tic n
euro
path
y, as
thes
e ca
n be
sev
ere
and
life
thre
aten
ing.
His
toric
ally
line
zolid
has
bee
n ve
ry e
xpen
sive
, how
ever
, it h
as re
cent
ly c
ome
off p
aten
t and
the
avai
labi
lity
of g
ener
ic p
rodu
cts
has
redu
ced
its m
arke
t pric
e su
bsta
ntia
lly a
nd it
may
dec
reas
e ev
en fu
rther
.
6565
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
D. A
dd-o
n ag
ents
Pyra
zinam
ide
is in
expe
nsive
, rea
dily
ava
ilabl
e an
d ea
sy to
adm
inis
ter.
Ison
iazid
is in
expe
nsive
. It i
s im
porta
nt to
con
side
r the
epi
dem
iolo
gy o
f hig
h-le
vel v
ersu
s lo
w-le
vel i
soni
azid
mut
atio
ns in
a p
opul
atio
n be
fore
sta
ndar
d tre
atm
ent r
egim
ens
incl
udin
g hi
gh-d
ose
ison
iazid
are
reco
mm
ende
d.Et
ham
buto
l is
inex
pens
ive a
nd re
adily
ava
ilabl
e.PA
s m
ay b
e di
fficu
lt to
obt
ain
alth
ough
it is
ava
ilabl
e th
roug
h th
e GD
F. Ot
herw
ise
it is
rela
tivel
y in
expe
nsive
and
eas
y to
adm
inis
ter.
Amox
acill
in-c
lavu
lana
te is
inex
pens
ive a
nd e
asily
obt
aina
ble.
How
ever
, the
car
bape
nem
s ar
e ex
pens
ive a
nd a
re d
ifficu
lt to
adm
inis
ter a
s th
ey m
ust b
e gi
ven
two
or th
ree
times
pe
r day
via
an
intra
veno
us li
ne.
thio
acet
ezon
e is
inex
pens
ive b
ut it
has
lim
ited
avai
labi
lity
and
it is
not
cur
rent
ly a
vaila
ble
thro
ugh
the
GDF.
the
curre
nt re
visi
on o
f the
gui
delin
es d
id n
ot re
-ana
lyse
the
optim
al d
urat
ion
of tr
eatm
ent (
inte
nsive
and
con
tinua
tion
phas
es).
the
reco
mm
enda
tions
from
the
2011
gui
de-
lines
that
wer
e ba
sed
on th
e aI
PD m
eta-
anal
ysis
, thu
s co
ntin
ue to
app
ly. th
e 20
11 g
uide
lines
con
ditio
nally
reco
mm
ende
d an
inte
nsive
pha
se o
f eig
ht m
onth
s fo
r mos
t MDR
-tb
patie
nts
and
tota
l tre
atm
ent d
urat
ion
of 2
0 m
onth
s in
pat
ient
s wh
o ha
d no
t bee
n pr
evio
usly
trea
ted.
the
dura
tion
may
nee
d to
be
mod
ified
acc
ordi
ng to
the
patie
nt’s
resp
onse
to
ther
apy.
the
asso
ciat
ion
betw
een
treat
men
t suc
cess
and
the
tota
l len
gth
of tr
eatm
ent w
as le
ss c
lear
in p
atie
nts
who
had
been
pre
viou
sly-
treat
ed c
ompa
red
with
thos
e wh
o ha
d no
t, al
thou
gh th
e lik
elih
ood
of tr
eatm
ent s
ucce
ss a
ppea
red
to p
eak
betw
een
27.6
and
30.
5 m
onth
s. th
e nu
mbe
r of o
bser
vatio
ns w
as a
lso
far f
ewer
than
for t
hose
who
had
no
pre
viou
s M
DR-t
b tre
atm
ent.
As a
resu
lt no
reco
mm
enda
tion
on to
tal d
urat
ion
was
mad
e in
the
2011
revi
sion
for p
revi
ousl
y tre
ated
pat
ient
s. M
any
of th
e rif
ampi
cin-
resi
stan
t tb
/MDR
-tb
patie
nts
who
will
be in
elig
ible
for t
he s
horte
r MDR
-tb
regi
men
and
refe
rred
for t
reat
men
t with
long
er re
gim
ens
woul
d ha
ve b
een
treat
ed w
ith s
econ
d-lin
e m
edic
a-tio
n in
the
past
; in
thes
e pa
tient
s un
certa
intie
s wi
ll re
mai
n on
the
optim
al d
urat
ion
of tr
eatm
ent a
nd th
eref
ore
the
leng
th o
f the
rapy
wou
ld n
eed
to b
e gu
ided
prim
arily
by
the
resp
onse
to th
erap
y.
Mon
itor
ing
and
eval
uati
onPa
tient
s on
long
er M
DR-t
b tre
atm
ent r
egim
ens
need
to b
e m
onito
red
for r
espo
nse
to tr
eatm
ent a
nd fo
r saf
ety
usin
g re
ason
able
sch
edul
es o
f rel
evan
t clin
ical
and
labo
rato
ry
test
ing.
Fra
mew
orks
for t
he s
urve
illan
ce o
f bac
terio
logi
cal s
tatu
s, d
rug
resi
stan
ce a
nd o
utco
mes
hav
e be
en fa
irly
stan
dard
ized
over
the
past
dec
ade.
the
syst
emat
ic m
onito
ring
of a
dver
se e
vent
s du
ring
and
afte
r the
end
of t
reat
men
t is
a m
ore
rece
nt in
trodu
ctio
n in
tb p
rogr
amm
es a
nd e
xper
ienc
e in
thei
r im
plem
enta
tion
is s
till d
evel
opin
g in
man
y co
un-
tries
. Its
ratio
nale
is la
rgel
y de
fined
by
mor
e fre
quen
t use
of n
ew a
nd re
-pur
pose
d m
edic
atio
ns in
MDR
-tb
treat
men
t reg
imen
s in
the
world
, at t
imes
in c
ombi
natio
ns fo
r whi
ch
ther
e ha
s be
en v
ery
limite
d ex
perie
nce
of u
se.
Res
earc
h pr
ioriti
es• A
nee
d fo
r mor
e ra
ndom
ized
cont
rol s
tudi
es, e
spec
ially
invo
lvin
g th
e ne
w dr
ugs
and
regi
men
s.
• Incl
usio
n an
d se
para
te re
porti
ng o
f out
com
es fo
r key
sub
grou
ps in
suc
h st
udie
s, e
spec
ially
chi
ldre
n an
d HI
V-po
sitiv
e in
divi
dual
s on
trea
tmen
t.
• Mor
e co
mpl
ete
reco
rdin
g of
adv
erse
eve
nts
and
stan
dard
ized
data
reco
rdin
g on
org
an c
lass
, ser
ious
ness
, sev
erity
, and
cer
tain
ty o
f ass
ocia
tion,
to a
llow
relia
ble
com
paris
on o
f th
e as
soci
atio
n be
twee
n ad
vers
e ev
ents
and
exp
osur
e to
diff
eren
t med
icin
es.
• Iden
tifica
tion
of fa
ctor
s th
at d
eter
min
e th
e op
timal
dur
atio
n of
trea
tmen
t (e.
g. p
revi
ous
treat
men
t his
tory
, bas
elin
e re
sist
ance
pat
tern
s, s
ite o
f dis
ease
, chi
ld/a
dult)
.
• Det
erm
inat
ion
of th
e m
inim
um n
umbe
r of d
rugs
and
trea
tmen
t dur
atio
n (e
spec
ially
in p
atie
nts
prev
ious
ly tr
eate
d fo
r MDR
-tb)
.
• Con
ditio
ns u
nder
whi
ch in
ject
able
-spa
ring
regi
men
s ca
n be
use
d in
bot
h ch
ildre
n an
d ad
ults
(e.g
. sur
roga
tes
for s
ever
ity /
ext
ent o
f dis
ease
, alte
rnat
ive m
edic
atio
n).
• Pha
rmac
okin
etic
stu
dies
to d
eter
min
e op
timal
dru
g do
sing
and
saf
ety
(esp
ecia
lly in
pre
gnan
cy).
• Impr
oved
dia
gnos
tics
and
drug
-sus
cept
ibili
ty te
stin
g m
etho
ds (e
.g. w
hich
test
for p
yraz
inam
ide)
.
• Palli
ative
and
end
-of-l
ife c
are
in p
atie
nts
with
ver
y ad
vanc
ed re
sist
ance
pat
tern
s.
66
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
3. E
lect
ive
part
ial l
ung
rese
ctio
n ve
rsus
no
surg
ery
for
pati
ents
on
trea
tmen
t fo
r M
DR
-TB
Popu
latio
n:Pa
tient
s on
trea
tmen
t for
MDR
-tb
back
grou
nd:
surg
ery
has
been
use
d to
trea
t tb
patie
nts
sinc
e be
fore
the
adve
nt o
f che
mot
her-
apy.
With
the
chal
leng
ing
pros
pect
of i
nade
quat
e re
gim
ens
to tr
eat M
DR/X
DR-t
b an
d th
e ris
k fo
r ser
ious
seq
uela
e, th
e ro
le o
f pul
mon
ary
surg
ery
is b
eing
re-e
valu
ated
as
a m
eans
to “d
ebul
k” in
tract
able
pat
holo
gy in
the
lung
and
to re
duce
bac
teria
l loa
d an
d th
us im
prov
e pr
ogno
sis.
the
revi
ew fo
r thi
s qu
estio
n is
bas
ed u
pon
an in
divi
dual
, pat
ient
-leve
l met
a-an
aly-
sis
to e
valu
ate
the
effe
ctive
ness
of d
iffer
ent f
orm
s of
ele
ctive
sur
gery
as
an a
djun
ct
to c
ombi
natio
n m
edic
al th
erap
y fo
r MDR
-tb
(sup
plem
ente
d by
a li
tera
ture
revi
ew).
Dem
ogra
phic
, clin
ical
, bac
terio
logi
cal,
surg
ical
and
out
com
e da
ta o
n M
DR-t
b pa
tient
s on
trea
tmen
t wer
e ob
tain
ed fr
om th
e au
thor
s of
26
coho
rt st
udie
s, id
entifi
ed fr
om
thre
e sy
stem
atic
revi
ews
of M
DR-t
b tre
atm
ent.
the
anal
yses
sum
mar
ized
in th
e GR
ADE
tabl
es c
onsi
st o
f thr
ee s
trata
com
parin
g tre
at-
men
t suc
cess
(cur
e an
d co
mpl
etio
n) w
ith d
iffer
ent c
ombi
natio
ns o
f tre
atm
ent f
ailu
re,
rela
pse,
dea
th a
nd lo
ss to
follo
w-up
. two
set
s of
suc
h ta
bles
wer
e pr
epar
ed fo
r: (i)
pa
rtial
pul
mon
ary
rese
ctio
n an
d (ii
) pne
umon
ecto
my.
Parti
al p
ulm
onar
y re
sect
ion
was
sign
ifica
ntly
ass
ocia
ted
with
trea
tmen
t suc
cess
whe
n co
mpa
red
with
all
othe
r ou
tcom
es p
ut to
geth
er (t
reat
men
t fai
lure
or r
elap
se o
r dea
th o
r los
s to
follo
w-up
) (P
<0.0
5). P
rogn
osis
app
eare
d to
be
bette
r whe
n su
rger
y wa
s pe
rform
ed a
fter c
ultu
re
conv
ersi
on. N
o ef
fect
was
obs
erve
d in
pne
umon
ecto
my.
Desp
ite a
num
ber o
f pot
entia
l bia
ses
and
limita
tions
whi
ch c
ould
not
be
adju
sted
for,
parti
al lu
ng re
sect
ion
surg
ery
afte
r cul
ture
con
vers
ion
may
hel
p im
prov
e ou
tcom
es in
se
lect
ed p
atie
nts
who
do n
ot re
spon
d to
app
ropr
iate
med
icat
ion.
Refe
renc
e: F
ox G
J, M
itnic
k CD
, ben
edet
ti A,
Cha
n ED
, bec
erra
M, C
hian
g C-
Y, et
al.
surg
ery
as a
n ad
junc
tive
treat
men
t for
mul
tidru
g-re
sist
ant t
uber
culo
sis:
An
indi
vidu
al
patie
nt d
ata
met
aana
lysi
s. C
lin In
fect
Dis
. 201
6;62
(7):8
87–9
5.
Inte
rven
tion:
Elec
tive
parti
al lu
ng re
sect
ion
Com
paris
on:
No s
urge
ry
Mai
n ou
tcom
es:
succ
ess
vers
us tr
eatm
ent f
ailu
re o
r rel
apse
; suc
cess
ver
sus
treat
men
t fai
lure
or
rela
pse
or d
eath
; suc
cess
ver
sus
treat
men
t fai
lure
or r
elap
se o
r dea
th o
r lo
ss to
follo
w-up
; dea
th v
ersu
s tre
atm
ent f
ailu
re o
r rel
apse
or s
ucce
ss.
setti
ng:
Whi
ch ty
pes
of s
urge
ry e
ncom
pass
ed (l
obec
tom
y, se
gmen
tect
omy,
wedg
e re
sect
ion)
; defi
nitio
n of
non
-resp
onse
and
adv
erse
out
com
e of
sur
gery
; def
-in
ition
of e
xten
sive
dis
ease
; how
spe
cial
ized
were
the
cent
res/
prac
titio
ners
wh
ich
prov
ided
sur
gery
(ext
erna
l val
idity
); un
der w
hich
con
ditio
ns to
indi
-ca
te re
sect
ion
surg
ery
and
when
to c
ontra
indi
cate
; bef
ore
or a
fter c
ultu
re
conv
ersi
on.
Pers
pect
ive:
Defin
ing
bette
r the
role
of s
urge
ry; d
ecis
ion
when
to o
pera
te a
nd ty
pe o
f in
terv
entio
n; a
nd it
s im
pact
in p
atie
nts
on tr
eatm
ent f
or M
DR-t
b or
XDR
-tb.
6767
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
Ass
essm
ent
JUD
GEM
ENT
RES
EAR
CH
EVID
ENC
EA
DD
ITIO
NA
L C
ON
SID
ERAT
ION
SDESIRABLE EFFECTS
How
sub
stan
tial a
re th
e de
sira
ble
antic
ipat
ed e
ffec
ts?
tr
ivia
l
sm
all
M
oder
ate
l
arge
Var
ies
D
on’t
know
In th
e su
rgic
al m
eta-
anal
ysis
that
exa
min
ed a
ll fo
rms
of s
urge
ry to
geth
er,
ther
e wa
s a
stat
istic
ally
sig
nific
ant i
mpr
ovem
ent i
n cu
re a
nd s
ucce
ssfu
l tre
atm
ent o
utco
mes
am
ong
patie
nts
who
rece
ived
surg
ery.
Howe
ver,
when
th
e in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
exa
min
ed p
atie
nts
that
und
er-
went
par
tial l
ung
rese
ctio
n an
d th
ose
that
und
erwe
nt p
neum
onec
tom
y, ve
rsus
pat
ient
s th
at d
id n
ot u
nder
go s
urge
ry, t
hose
that
und
erwe
nt p
artia
l lu
ng re
sect
ion
had
stat
istic
ally
sig
nific
ant h
ighe
r rat
es o
f tre
atm
ent s
ucce
ss.
thos
e th
at u
nder
went
pne
umon
ecto
my
did
not h
ave
bette
r out
com
es th
an
thos
e wh
o di
d no
t und
ergo
sur
gery
.th
ere
are
seve
ral c
avea
ts to
this
dat
a. s
elec
tion
bias
may
be
an is
sue,
as
patie
nts
who
were
det
erm
ined
to b
e he
alth
y en
ough
to u
nder
go s
urge
ry w
ere
the
only
peo
ple
who
unde
rwen
t sur
gery
. Peo
ple
livin
g wi
th H
IV w
ere
excl
uded
fro
m th
e IP
D. A
dditi
onal
ly, a
lthou
gh th
ere
is li
kely
qui
te a
bit
of c
onfo
undi
ng,
patie
nts
with
XDR
-tb
were
foun
d to
hav
e si
gnifi
cant
ly w
orse
out
com
es w
hen
they
und
erwe
nt s
urge
ry.
Rate
s of
dea
th d
id n
ot d
iffer
sig
nific
antly
bet
ween
thos
e wh
o un
derw
ent s
ur-
gery
ver
sus
thos
e wh
o re
ceive
d m
edic
al tr
eatm
ent o
nly.
ther
e wa
s no
t eno
ugh
data
on
adve
rse
even
ts o
r sur
gica
l com
plic
atio
ns to
do
an
anal
ysis
.Re
fere
nce:
Fox
GJ,
Mitn
ick
CD, b
ened
etti
A, C
han
ED, b
ecer
ra M
, Chi
ang
C-Y,
et a
l. su
rger
y as
an
adju
nctiv
e tre
atm
ent f
or m
ultid
rug-
resi
stan
t tub
er-
culo
sis:
An
indi
vidu
al p
atie
nt d
ata
met
aana
lysi
s. C
lin In
fect
Dis
. 201
6;
62(7
):887
–95.
Effe
ct e
xpec
ted
to b
e m
oder
ate
in th
e av
erag
e pa
tient
con
side
red
appr
opria
te fo
r sur
gery
.
UNDESIRABLE EFFECTS
How
sub
stan
tial a
re th
e un
desi
rabl
e an
ticip
ated
eff
ects
?
lar
ge
M
oder
ate
s
mal
l
triv
ial
V
arie
s
Don
’t kn
ow
Unce
rtain
ty a
bout
per
iope
rativ
e or
pos
t-ope
rativ
e co
mpl
icat
ions
.
CERTAINTY OF EVIDENCE
Wha
t is
the
over
all c
erta
inty
of t
he e
vide
nce
of e
ffec
ts?
V
ery
low
l
ow
M
oder
ate
H
igh
N
o in
clud
ed s
tudi
es
No re
sear
ch e
vide
nce
was
iden
tified
.su
bsta
ntia
l het
erog
enei
ty e
xpec
ted
in a
num
-be
r of p
aram
eter
s in
clud
ing
the
crite
ria u
sed
to
sele
ct c
andi
date
s fo
r sur
gery
, the
type
/qua
lity
of in
terv
entio
n, th
e ef
fect
ivene
ss o
f con
com
itant
ch
emot
hera
py a
nd o
ther
sup
porti
ve m
easu
res,
and
th
e rig
our a
nd le
ngth
of t
ime
durin
g wh
ich
effe
cts
(ben
efici
al o
r adv
erse
) wer
e m
onito
red.
68
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
JUD
GEM
ENT
RES
EAR
CH
EVID
ENC
EA
DD
ITIO
NA
L C
ON
SID
ERAT
ION
S
BALANCE OF EFFECTSDo
es th
e ba
lanc
e be
twee
n de
sira
ble
and
unde
sira
ble
effe
cts
favo
ur th
e in
terv
entio
n or
the
com
paris
on?
F
avou
rs th
e co
mpa
rison
Pro
babl
y fa
vour
s th
e co
mpa
rison
Doe
s no
t fav
our e
ither
the
inte
rven
tion
or th
e co
mpa
rison
Pro
babl
y fa
vour
s th
e in
terv
entio
n
Fav
ours
the
inte
rven
tion
V
arie
s
Don
’t kn
ow
No re
sear
ch e
vide
nce
was
iden
tified
.lo
ng te
rm s
eque
lae,
som
e of
whi
ch m
ay b
e ul
ti-m
atel
y fa
tal,
may
be
unkn
own.
the
bene
fits
vary
dep
endi
ng o
n th
e po
pula
tion
sele
ctio
n; h
owev
er, o
vera
ll th
e ef
fect
s of
sur
gery
ap
pear
to b
e be
nefic
ial a
s lo
ng a
s pa
tient
s ar
e se
lect
ed c
aref
ully
for s
urge
ry.
Desp
ite th
e un
know
n m
agni
tude
of p
erio
pera
tive
com
plic
atio
ns th
e pa
nel a
ssum
ed th
at o
vera
ll th
ere
is a
net
ben
efit f
rom
sur
gery
.
FEASIBILITY
Is th
e in
terv
entio
n fe
asib
le to
impl
emen
t?
No
P
roba
bly
no
P
roba
bly
yes
Ye
s
Var
ies
D
on’t
know
No re
sear
ch e
vide
nce
was
iden
tified
.su
rger
y is
use
d ex
tens
ively
in m
any
coun
tries
but
th
e qu
ality
var
ies.
If pr
ogra
mm
es in
vest
in s
urge
ry in
pre
fere
nce
to
othe
r com
pone
nts
ther
e ar
e op
portu
nity
cos
ts.
Equi
ty is
sues
: the
re is
an
issu
e of
acc
ess
to h
igh
qual
ity s
urge
ry s
o eq
uity
is “p
roba
bly
redu
ced”
and
in
equi
ty is
a p
ossi
bilit
y.Ac
cept
abili
ty to
sta
keho
lder
s an
d pa
tient
s: “v
arie
s”
depe
ndin
g on
the
stak
ehol
der.
Con
clus
ions
ELEC
TIV
E PA
RTI
AL
LUN
G R
ESEC
TIO
N V
ERSU
S N
O S
UR
GER
Y F
OR
PAT
IEN
TS O
N T
REA
TMEN
T FO
R M
DR
-TB
Type
of
reco
mm
enda
tion
stro
ng re
com
men
datio
n ag
ains
t th
e in
terv
entio
n
Cond
ition
al re
com
men
datio
n ag
ains
t the
inte
rven
tion
Cond
ition
al re
com
men
datio
n fo
r eith
er th
e in
terv
entio
n or
the
com
paris
on
Cond
ition
al re
com
men
datio
n fo
r th
e in
terv
entio
n
stro
ng re
com
men
datio
n fo
r the
in
terv
entio
n
Votin
g re
sults
of t
he G
DG fo
r con
ditio
nal r
ecom
men
datio
n on
ele
ctive
par
tial s
urge
ry: 1
5 in
favo
ur; 1
aga
inst
; 1 a
bste
ntio
n; 2
no
reco
mm
enda
tion;
2 n
ot a
vaila
ble
Rec
omm
enda
tion
In p
atie
nts
with
rifa
mpi
cin-
resi
stan
t tb
or M
DR- t
b, th
e W
HO G
uide
line
Deve
lopm
ent G
roup
sug
gest
s th
at e
lect
ive p
artia
l lun
g re
sect
ion
(lobe
ctom
y or
wed
ge re
sect
ion)
may
be
used
alo
ngsi
de a
n ap
prov
ed M
DR-t
b re
gim
en (c
ondi
tiona
l rec
omm
enda
tion,
ver
y lo
w ce
rtain
ty in
the
evid
ence
).
6969
ANNE
X 5:
EVI
DENC
E tO
DEC
IsIO
N tA
blEs
Just
ifica
tion
In th
e su
rgic
al m
eta-
anal
ysis
that
exa
min
ed a
ll fo
rms
of s
urge
ry to
geth
er, t
here
was
a s
tatis
tical
ly s
igni
fican
t im
prov
emen
t in
cure
and
suc
cess
ful t
reat
men
t out
com
es a
mon
g pa
tient
s wh
o re
ceive
d su
rger
y. Ho
weve
r, wh
en th
e in
divi
dual
pat
ient
dat
a m
eta-
anal
ysis
exa
min
ed p
atie
nts
that
und
erwe
nt p
artia
l lun
g re
sect
ion
and
thos
e th
at u
nder
went
pne
u-m
onec
tom
y ve
rsus
pat
ient
s th
at d
id n
ot u
nder
go s
urge
ry, t
hose
that
und
erwe
nt p
artia
l lun
g re
sect
ion
had
stat
istic
ally
sig
nific
ant h
ighe
r rat
es o
f tre
atm
ent s
ucce
ss. t
hose
who
un
derw
ent p
neum
onec
tom
y di
d no
t hav
e be
tter o
utco
mes
than
thos
e wh
o di
d no
t und
ergo
sur
gery
.th
ere
are
seve
ral c
avea
ts to
thes
e da
ta. s
elec
tion
bias
may
be
an is
sue,
as
patie
nts
who
were
det
erm
ined
to b
e he
alth
y en
ough
to u
nder
go s
urge
ry w
ere
the
only
peo
ple
who
unde
rwen
t sur
gery
. Peo
ple
livin
g wi
th H
IV w
ere
excl
uded
from
the
IPD.
Add
ition
ally,
alth
ough
ther
e is
pos
sibl
y qu
ite a
bit
of c
onfo
undi
ng, p
atie
nts
with
XDR
-tb
were
foun
d to
hav
e si
gnifi
cant
ly w
orse
out
com
es w
hen
they
und
erwe
nt s
urge
ry.
Rate
s of
dea
th d
id n
ot d
iffer
sig
nific
antly
bet
ween
thos
e wh
o un
derw
ent s
urge
ry v
ersu
s th
ose
who
rece
ived
med
ical
trea
tmen
t onl
y.th
ere
was
not e
noug
h da
ta o
n ad
vers
e ev
ents
or s
urgi
cal c
ompl
icat
ions
to d
o an
ana
lysi
s.
Sub
grou
p co
nsid
erat
ions
the
data
sho
w th
at X
DR p
atie
nts
who
unde
rwen
t sur
gery
did
wor
se th
an o
ther
pat
ient
s (a
OR 0
.4, 0
.2–0
.9) a
nd th
eref
ore
the
reco
mm
enda
tion
does
not
app
ly to
XDR
-tb
patie
nts
Impl
emen
tati
on
cons
ider
atio
nsth
e re
com
men
datio
n is
lim
ited
to p
artia
l res
ectio
n, c
ondu
cted
as
an e
lect
ive in
terv
entio
n. M
inim
al s
urgi
cal i
nter
vent
ions
suc
h as
dra
inag
e of
abs
cess
es w
ere
not i
nclu
ded.
Mor
e ra
dica
l pne
umon
ecto
my
is n
ot in
clud
ed.
Parti
al lu
ng re
sect
ion
for p
atie
nts
with
MDR
-tb
is re
com
men
ded
only
und
er c
ondi
tions
of g
ood
surg
ical
faci
litie
s, tr
aine
d an
d ex
perie
nced
sur
geon
s an
d wi
th c
aref
ul s
elec
tion
of
appr
opria
te s
urgi
cal c
andi
date
s.
Mon
itor
ing
and
eval
uati
on• D
efini
tion
of e
xten
sive
dis
ease
• Defi
nitio
n of
the
para
met
ers
for m
onito
ring
resp
onse
, non
-resp
onse
, adv
erse
out
com
e an
d se
quel
ae o
f sur
gery
• Mon
itorin
g fo
r the
opt
imal
tim
ing
afte
r cul
ture
con
vers
ion
to u
nder
take
sur
gery
Res
earc
h pr
ioriti
es• W
hich
con
ditio
ns in
dica
te re
sect
ion
surg
ery
and
when
to c
ontra
indi
cate
(sel
ectio
n of
pat
ient
s an
d ty
pe o
f dis
ease
)
70
ANNEX 6
Summaries of unpublished data used for the recommendations
1. Short MDR-TB regimens: meta-analyses of data from published and unpublished studies
F Ahmad Khan1, E C Casas2, P DuCros3, MA Hamid Salim4, V Schwoebel5, A Trébucq5, Atadjan Khamraev6, Welile Sikhondze7, D Falzon8, D Menzies1
1 Respiratory Epidemiology and Clinical Research Unit & McGill International TB Centre, McGill University, Montreal, Canada; 2 Médecins Sans Frontières-Operational Center Amsterdam, The Netherlands & Médecins Sans Frontières-OCA Swaziland; 3 Manson Unit, Médecins Sans Frontières, London, UK; 4 National TB Programme of Bangladesh, USAID-MDR-TB Project, Dhaka, Bangladesh; 5 International Union Against Tuberculosis and Lung Disease, Paris, France; 6 Ministry of Health, Nukus, Karakalpakstan, Uzbekistan; 7 National TB Control Program, Ministry of Health, Swaziland Ministry of Health, Swaziland; 8 World Health Organization, Global TB Programme, Geneva, Switzerland
BackgroundFor the treatment of multidrug-resistant tuberculosis (MDR-TB), the World Health Organization (WHO) recommends an intensive phase of at eight months, and total treatment duration of at least 20 months. Shortening the treatment duration without compromising efficacy would substantially reduce the burden that prolonged therapy places on patients and programmes. To inform updated policy recommendations, we synthesized data from published and unpublished studies of MDR-TB patients treated with standardized regimens based on the 9-month “Bangladesh regimen” initially described by Van Deun and co-workers in 2010 (1).
MethodsAn expert committee identified published and ongoing studies of MDR-TB patients treated with standardized regimens of up to 12 months in duration (“short MDR-TB regimens”) that were based on the Bangladesh regimen. We sought to: (i) estimate the probability of treatment success (cure or treatment completion) versus an unfavourable outcome (failure/relapse, death, or default); (ii) identify baseline characteristics associated with these outcomes; and (iii) compare outcomes to those reported in patients treated with regimens of conventional duration (at least 18 months). Patients were included if they had MDR-TB confirmed by culture or molecular drug-susceptibility testing (DST). We also included patients with rifampicin-resistant TB in whom isoniazid DST had not been performed. Aggregate (study-level) meta-analyses were performed to estimate pooled proportions using data from all studies. To identify patient characteristics associated with outcomes, we conducted individual patient data meta-analyses stratified by characteristics of interest. Lastly, we compared outcomes with short MDR-TB regimens to those with longer duration regimens. To do so, we used data from MDR-TB patients treated with regimens of at least 18 months (“longer regimens”) taken from a previous individual patient meta-analysis. Hence, the comparison group included many different regimens, some of which were individualized, and not all
71
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
of which met existing WHO recommendations for MDR-TB treatment. Meta-analyses used random effects models with the exact binomial likelihood method.
ResultsSix studies were identified. Three are ongoing and shared interim data for this analysis. One published and two ongoing studies provided individual patient data. The description of the studies and patients is reported in Table A6.1.1.
Exclusion criteria
Five studies excluded patients that had previously been treated with second-line anti-TB medications. The following exclusion criteria were used in some of the studies: pregnancy, age <14 years, severe liver or renal co-morbidity, baseline XDR-TB, baseline resistance to moxifloxacin, baseline resistance to ofloxacin, resistance to at least two second-line injectables, severe clinical condition, baseline QT prolongation and extrapulmonary TB.
Regimens
In all six studies, the minimum duration of the intensive phase was four months. The intensive phase could be extended by two months – exceptionally up to four months in the Swaziland series – in the absence of conversion. The duration of the continuation phase was five months in four studies, and eight months in two studies. The intensive phase regimens typically consisted of kanamycin, moxifloxacin (usual dose) or gatifloxacin (high or usual dose), high-dose isoniazid, prothionamide, clofazimine, pyrazinamide and ethambutol. In all studies, the continuation phase regimen included the same fluoroquinolone (moxifloxacin or gatifloxacin), clofazimine, pyrazinamide and ethambutol; prothionamide was also continued in three studies. All treatment was under direct observation, and in most studies either some or all patients were hospitalized for a portion of the treatment.
Outcome definitions
Outcomes of cure, treatment completion, failure, death and default were reported in all studies. Relapse was defined as a positive culture, post treatment completion. Because this outcome was rare and only reported in the three published studies, the few relapse cases were counted as failures.
Aggregate meta-analyses
Rates of successful and unsuccessful treatment are reported in Table A6.1.2. The proportion of those successfully treated was higher with standardized short regimens. When death and “loss to follow-up” were included as unsuccessful outcomes along with failure/relapse, the percentage success was significantly higher in patients on the shorter regimens compared with those on the longer regimen (confidence limits not overlapping).
Individual patient data meta-analyses
When stratified by baseline susceptibility to fluoroquinolone and pyrazinamide, the proportion of patients successfully treated remained consistently greater with standardized short MDR-TB regimens; however, the confidence limits overlapped. There was a trend towards worsening treatment outcomes in
72
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
both the short regimens group, and the longer regimen group, in patients with fluoroquinolone- and/or pyrazinamide-resistant MDR-TB.
ConclusionShort MDR-TB regimens based on the “Bangladesh regimen” have shown promising results in patients that have never been treated with second-line drugs and with baseline susceptibility to fluoroquinolones and pyrazinamide. There is a paucity of data on relapse; however, the available evidence suggests relapse is rare.
Table A6.1.1. Description of studies and patients
BANGLADESH 2005–2011
NIGER 2008–2010
CAMEROON 2008–2011
UZBEKISTAN 2013–2015
MULTIPLE 2013–2015
SWAZILAND 2014–2015
Status Published Published Published Ongoing Ongoing Ongoing
Data available for meta-analysis
Individual patient Aggregate Aggregate Individual patient Aggregate Individual patient
Data available on relapse at 2 years post-end of treatment
Yes Yes No No No No
Patients eligible for initiation of MDR treatment
640 124 323 NR 1169 114
Patients with MDR-TB or rifampicin-resist-ant TB confirmed
527† 97† 237† 117* 1169** 76*
Excluded from analysis, n (%)
34 (6.4%) 32 (33.0%) 87 (36.7%) 52 (44.4%)Ω 761 (65.1%)*** 52 (68.4%) Ω
Included, n (%) 493 (93.5%) 65 (67.0%) 150 (63.3%) 65 (55.6%) 408 (34.9%) 24 (31.5%)
Age (± standard deviation, or IQR)
33.6 (±12.9) 31 (27–38) 35.1 34.1 (±14.3) 35.1 35.2 (±14.4)
Female, n (%) 150/493 (30.4%) 12/65 (18.5%) 73/150 (48.7%) 35/65 (53.8%) 152/408 (37.3%) 13/24 (54.2%)
Primary MDR, n (%) 4/493 (0.8%) 1/65 (1.5%) 1/150 (0.1%) 47/61 (77.0%) 59/407 (14.5%) 20/24 (83.3%)
HIV, n (%) 0 1/58 (1.7%) 30/150 (20%) 0/44 91/407 (22.4%) 16/24 (66.7%)
Smear positive, n (%) 475/493 (96.3%) 54/65 (83.1%) 150/150 (100%) 28/62 (45.2%) 354/406 (87.2%) 12/23 (52.2%)
Chest radiograph cavities, n (%)
99/493 (20.1%) 23/65 (35.4%) NR 26/61 (42.6%) NR NR
Pyrazinamide resist-ance, n (%)
99/240 (41.3%) NR NR 33/39 (84.6%) 80/150 (51.3%) 10/14 (71.4%)
Ethambutol resistance, n (%)
321/493 (65.1%) 45/65 (69.2%) NR 31/44 (70.5%) NR 12/17 (70.6%)
MDR: multidrug-resistant; tb: tuberculosis; NR: not reportedMultiple: benin, burkina Faso, burundi, Cameroon, Central Africa Republic, Democratic Republic of Congo, Niger.† Isoniazid and rifampicin resistance confirmed in all participants.* Includes participants with rifampicin-resistant tb in whom Dst to isoniazid was not performed (Uzbekistan, n=7; swaziland, n=6).** Includes participants with rifampicin-resistant tb in whom Dst to isoniazid was not performed (n=137) or with Dst-confirmed susceptibility to
isoniazid (n=22).*** 409/761 never initiated the short MDR-tb regimen: 65 with prior exposure to second-line drugs; 1 with XDR-tb; 112 lost prior to initiation; 34
died prior to initiation; 197 other (pregnancy, children, medical/social contra-indications, refusals, non-residents).Ω Majority of exclusions were accounted for by participants in whom short MDR-tb treatment was ongoing, or had ended recently: Uzbekistan,
39/52; swaziland, 47/52.
73
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
Table A6.1.2. Pooled treatment outcomes from aggregate data meta-analyses
OUTCOMESTANDARDIZED SHORT MDR-TB REGIMENS,
6 STUDIES LONGER REGIMENS, 31 STUDIES
SUCCESS/NWEIGHTED PROPORTION
(95% CL) SUCCESS/NWEIGHTED PROPORTION
(95% CL)
success versus failure or relapse
1008/1033 97.5% (92.4%–99.2%) 4033/4639 91.2% (86.1%–94.6%)
success versus failure, relapse or death
1008/1116 90.3% (87.8%–92.4%) 4033/5850 78.3% (71.2%–84.0%)
success versus failure, relapse, death, or loss to follow up
1008/1205 83.7% (79.2%–87.4%) 4033/7665 61.7% (53.1%–69.6%)
Cl: Confidence limitsMeta-analyses used random-effects models.In the shorter regimens, data on relapse were only available in the three published studies.Bold indicates that 95% Cl does not overlap with the longer regimens group.
Reference1. Van Deun A, Maug AKJ, Salim MAH, Das PK, Sarker MR, Daru P, et al. Short, highly effective, and inexpensive standardized treatment of multidrug-resistant tuberculosis. Am J Respir Crit Care Med. 2010;182(5):684–692.
74
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
2. An updated systematic review and meta-analysis for treatment of multidrug-resistant tuberculosis
Mayara Lisboa Bastos1, Zhiyi Lan2, Dick Menzies2
1 Internal Medicine Graduate Program, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil2 Respiratory Epidemiology and Clinical Research Unit, Montreal Chest Institute, McGill University, Montreal, Canada
IntroductionTreatment for MDR-TB or XDR-TB requires lengthy use of second-line TB drugs, although the regimens used vary widely. The WHO 2011 guidelines recommended that MDR-TB treatment include as a minimum pyrazinamide, one second-line injectable (kanamycin, amikacin or capreomycin), one later generation fluoroquinolone (levofloxacin or moxifloxacin), and at least two Group 4 drugs (ethionamide/prothionamide, cycloserine/terizidone or p-aminosalicylic acid). We performed a systematic review to update the evidence for MDR-TB treatment to inform the WHO Guideline Development Group.
Methods
Literature search and study selectionThe PICO (Patients, Intervention, Comparator and Outcomes) questions were developed by the WHO Guideline Development Group in 2014–2015 to assist evidence reviews to inform its 2016 update of the guidance on MDR-TB treatment. The main focus of this review was the efficacy and safety of available drugs for the treatment of MDR-TB patients. The following groups of drugs were analyzed: first-line drugs (pyrazinamide, ethambutol, and high dose isoniazid), injectable drugs (streptomycin, kanamycin, amikacin, and capreomycin), fluoroquinolones (ofloxacin, levofloxacin and moxifloxacin), drugs from Group 4 (ethionamide/prothionamide, cycloserine /terizidone and p-aminosalicylic acid (PAS)), and the new drug bedaquiline. Drugs from Group 5 were not included in our review since at least four independent systematic reviews were recently conducted for these drugs.
Three major databases were used for our search: MEDLINE (through OVID), EMBASE (through OVID) and The Cochrane Library. The search strategy used a combination of Medical Subject Heading (MeSH) terms and free-text words in titles, abstracts and key words. Terms related to MDR-TB and XDR-TB, drugs of interest and treatment outcomes were included. Since this is an update from previous reviews which included studies published up to December 2008 our search was limited to the period from January 2009 to August 2015. The detailed search strategy is available in the supplemental material.
Two independent reviewers screened titles, abstracts and full texts, with consensus in each stage. A third reviewer was consulted to resolve possible disagreements. We included studies published in English, French, Chinese, Portuguese and Spanish. All studies that met the following inclusion criteria were selected: (i) MDR-TB confirmed by phenotypic tests (GeneXpert® was not adequate unless confirmed); (ii) pulmonary TB (studies that had more than 10% extrapulmonary patients and did not report the outcomes separately were excluded); (iii) cohorts or RCTs with a minimum of 25 MDR-TB (or XDR-TB) patients treated; (iv) a clear regimen specifying the drugs received; and (v) at least reported one of the following outcomes: end-of-treatment outcomes, six-month culture conversion, adverse events due to MDR-TB treatment. Studies that evaluated short regimens (<18 months) were excluded.
75
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
Data abstractionData from eligible studies were abstracted using a standardized data abstraction form (see supplemental material). We recorded information of age, sex, HIV (and use of antiretroviral treatment), acid-fast bacillus smear results, chest radiograph cavitation, prior TB treatment (with first-line drugs or second-line drugs), drug susceptibility test results, number of patients that received each drug, duration of treatment, and whether the regimen was standardized or individualized. Outcomes abstracted included: end of treatment outcomes defined according to published criteria, six-month sputum culture conversion and serious adverse events (SAEs; defined as Grade 3–4 events, or defined operationally as drugs discontinued permanently). For SAEs, we recorded the study definition of severity and the drug responsible for the event, if identified.
Data synthesis and statistical analysisFor end of treatment outcomes, we compared success (defined as cured or treatment completed) to: (i) failure or relapse; or (ii) failure or relapse or death. We examined the relationship between end of treatment outcomes and six-month culture conversion; the number of patients receiving each specific drug, average number of drugs used, and duration of treatment; as well as the average value for each cohort of major clinical and demographic characteristics of the patients. If HIV information or age were missing, values were estimated using information from other studies in this review from the same country, and if no such study was available, from data published by the World Bank or WHO. Variables were categorized according to the distribution observed (i.e. in median, terciles or quartiles).
Occurrence of adverse events was pooled if the study identified the drug responsible for the event and if the event was classified as Grade 3 to 4 severity, or the drug of interest was permanently stopped.
All statistical analyses were performed using SAS (version 9.2 Institute, Cary, NC, USA). Linear mixed models were used to pool the proportion with events (NLIMIXED procedure in SAS). For pooling the proportions of adverse events, we used generalized linear mixed model (GLIMMIX procedure in SAS).
ResultsA total of 2336 titles were identified, and after eliminating duplicates and non-relevant publications based on review of titles and abstracts, 250 were selected for full text review, of which 74 met the review inclusion criteria. 19 studies reported adverse events that were classified as Grade 3 or 4, or required permanent discontinuation of the drug, and identified the drug responsible.
Pooled treatment success rate was 26% (CI 95%, 23%–30%) in XDR-TB patients, compared to 60% in all cohorts of MDR-TB patients (with or without additional second-line resistance). The occurrence of SAE ranged from 0.5% to 12.2% (Table A6.2.1). Less than 3% of patients receiving fluoroquinolones or pyrazinamide experienced an SAE, compared to more than 5% of patients receiving second-line injectables or a thiamide (ethionamide or prothionamide).
ConclusionThis review identified 74 studies, with 84 distinct cohorts, published since January 2009 that reported treatment regimens and outcomes in 17 494 MDR-TB and XDR-TB patients. These studies reported
76
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
adverse events, six-month culture conversion and end of treatment outcomes. Treatment outcomes were substantially worse in patients with XDR-TB, and somewhat worse in patients who received standardized regimens for MDR-TB. However, despite the large number of studies and patients, no other treatment parameter, including number or duration of drugs, and individual drugs were associated with improved six-month culture conversion, or end of treatment outcomes. This may reflect the limitations and difficulties of pooling this data rather than true lack of differences in efficacy of regimens or individual drugs. This review highlights the need for more standardized reporting as well as evidence from well-designed randomized trials, or from meta-analysis of pooled individual patient data set from multiple observational studies.
Table A6.2.1. Occurrence of serious adverse events (SAEs), attributed to specific drugs in treatment of MDR-TB or XDR-TB
(Results from 19 studies (20 cohorts) that reported Grade 3–4 adverse events, or drugs permanently stopped due to adverse events, and identified the drug responsible for the adverse events.)
DRUG
ARMS/COHORTS REPORTING SAE AND
USED THE DRUG
N PATIENTS RECEIVED THE
DRUG
SERIOUS ADVERSE EVENTS DUE TO DRUG
N PATIENTS WITH SAE RELATED TO
THE DRUG POOLED ESTIMATE1
(CI 95%)
Pyrazinamide 19 2023 56 2.8% (2.1%-3.7%)
Ethambutol 16 1325 6 0.5% (0.2%-1.1%)
Injectable 19 2538 184 7.3% (6.2%-8.4%)
later gen. FQN 13 827 10 1.2% (0.6%-2.4%)
Ofx/Cfx 9 1408 40 2.8% (1.9%-4.1%)
thiamide 17 2106 173 8.2% (7.0%-9.6%)
Cycloserine 16 2140 96 4.5% (3.6%-5.5%)
PAs 16 1706 208 12.2% (10.6%-13.9%)
later gen. FQN: later generation fluoroquinolone (includes gatifloxacin /levofloxacin /moxifloxacin), Ofx/Cfx: ofloxacin/ciprofloxacin, PAs: p-aminosalicylic acid1 Pooled using Proc Glimmix in sAs – fixed effects meta-analysis.
77
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
3. A systematic review and individual patient data meta-analysis of treatment and outcomes among children with multidrug- resistant tuberculosis
Elizabeth Harausz1, Anthony J. Garcia-Prats2, H. Simon Schaaf2, Stephanie Law3, Jennifer Furin4, Tamara Kredo5, Anna Turkova6, Dick Menzies3 and Anneke C. Hesseling2 for The Collaborative Group for Meta-Analysis of Paediatric Individual Patient Data in MDR-TB
1 Military HIV Research Program, Bethesda, USA2 Desmond Tutu TB Centre, Department of Paediatrics and Health Sciences, Faculty of Medicine and Health Sciences, Stellenbosch
University, Tygerberg, South Africa3 Montreal Chest Institute, McGill University, Montreal, Canada4 Department of Global Health and Social Medicine, Harvard Medical School, USA5 South African Medical Research Council, Cape Town, South Africa6 Imperial College Healthcare NHS Trust, Institute of Clinical Trials and Methodology, London, United Kingdom,
IntroductionMultidrug-resistant tuberculosis (MDR-TB) in children is under-recognized, under-diagnosed and under-reported. Despite approximately 32 000 children developing MDR-TB each year (1) and historical studies showing mortality rates from TB of 40%, 16% and 5% for infants, toddlers and young children, respectively (2), very little is known about optimal treatment for children with MDR-TB. Treatment of MDR-TB is difficult, requiring use of toxic medications for at least 18 months with formulations and regimens not suited to children. However, individual studies have reported successful treatment outcomes in at least 80% of children treated for MDR-TB (3). A more rigorous evidence base is needed to help inform the management of MDR-TB treatment in children. A systemic review in 2012 sought to better quantify treatment outcomes in children, however, many questions remain on how to optimize successful treatment outcomes and minimize adverse events (3).
In order to address key questions regarding the treatment of MDR-TB and to inform paediatric-specific guidelines, we undertook a systematic review and individual patient data meta-analysis (IPD) of children with MDR-TB. The objective was to provide information on the management of children with MDR-TB by analysing determinants of key treatment outcomes among children treated for MDR-TB, and addressing questions specifically relevant to the paediatric population with MDR-TB (Table A6.3.1).
Methods
Eligibility criteriaData sets were eligible if they included a minimum of three children (aged <15 years) within a defined treatment cohort who were treated for clinically diagnosed or bacteriologically confirmed pulmonary or extrapulmonary MDR-TB, and for whom treatment outcomes were reported, using standard World Health Organization (WHO) TB case definitions (4,5). Eligibility criteria were applied at the individual level, so that studies reporting on both adults and children could be considered eligible if they otherwise met the specified criteria. Both published and unpublished data were included, without date restriction. Eligible study designs included controlled and non-controlled retrospective and prospective studies and
78
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
case series. All cohorts containing children included in a previous systematic review and individual patient data meta-analysis of MDR-TB were considered eligible (6). Only reports written in Dutch, English, French, Russian and Spanish were included. We excluded studies that utilized only combinations of rifampicin, isoniazid (INH), pyrazinamide (PZA), ethambutol (EMB) or streptomycin to treat MDR-TB, as this is now considered inadequate therapy.
Identifying primary reportsTo identify eligible reports, including conference abstracts, we searched PubMed, LILACS, Embase, The Cochrane Library, PsychINFO, and BioMedCentral databases up to 30 September 2014, with a search strategy, using a combination of the search terms, viz. “tuberculosis”, “multidrug resistance”, “MDR-TB”, “multidrug-resistant”, and “children”, both as exploded MESH headings and free-text terms, and without language restriction. The specific search strategies for Pubmed and Embase are presented in Appendix 6A. We also reviewed conference abstracts from the annual meeting of the International Union Against Tuberculosis and Lung Diseases.
To identify additional published and unpublished data we contacted experts in the field of paediatric MDR-TB. We also requested additional data through multiple routes, such as at national and international conferences and training events, and through international and in-country organizations working in paediatric MDR-TB, including the Sentinel Project on Pediatric Drug-Resistant Tuberculosis, the WHO Childhood TB sub-Group, Médecins Sans Frontières (MSF), the United States and European CDC, International Union Against Tuberculosis and Lung Disease (UNION), National Institutes of Health (NIH) and others.
Report selection and reviewAll abstracts were screened by EH and a researcher with the South African Cochrane Centre to select full text reports to review. All full text reports were reviewed independently by two reviewers (EH, AGP, HSS, JF, ACH) to assess for eligibility, except reports in Dutch, French, Russian and Spanish, which were reviewed by a single reviewer (from among AT, EH, ACH and JF). A third reviewer resolved any disagreements about study selection. If report eligibility was unclear, two attempts were made to contact the authors of the primary report; and if we could not make contact after two unsuccessful attempts, these reports were excluded.
Individual patient data abstractionThe authors of all eligible studies were contacted to access individual patient data. Individual patient data were used following a written agreement with the study team by the original authors, which included confirmation of ethical approval according to local guidelines.
Data were collected on multiple factors which could influence treatment decision and outcome, including: demographic characteristics, nutritional status, HIV status and antiretroviral usage, adult MDR-TB source case information, culture confirmed versus clinical diagnosis, information on disease location (pulmonary or extrapulmonary) and severity (using a standard approach), drug susceptibility test results, the use of individual drugs, and the duration of drug use within the treatment regimen. Data were collected on acid-fast bacillus (smear) microscopy and culture conversion, adverse effects, as well as WHO-defined treatment
79
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
outcomes including cure, treatment completion, culture conversion by six months, treatment failure, relapse, loss to follow-up and mortality. Severity of disease on chest radiograph, based on a standardized disease severity classification developed for an international paediatric TB randomized control trial (Palmer M, personal communication), was graded independently by two reviewers (EH, ACH) as either severe or non-severe; disagreements were arbitrated by a third reviewer (HSS). The primary authors of all included reports were contacted as needed to resolve any queries.
In order to contextualize the clinical data, information was also requested from each primary author on site-level characteristics, including but not limited to methods for TB diagnosis, availability and type of drug-susceptibility testing performed, how treatment outcomes were defined, and how adverse effects were assessed.
A database was created, and primary data from each study were entered into the database.
AnalysisThe analysis was planned to address PICO question 1, as per Table A6.3.1. Primary analyses focused on success versus failure/relapse/death in children with confirmed MDR-TB only. There were no paediatric data available to address the section in PICO question 1 regarding rifampicin mono-resistant TB.
Table A6.3.1. WHO-defined PICO question 1, in HIV-infected and uninfected children aged 0–14 years with MDR-TB; and which individual drugs in the regimens are likely to lead to the outcomes listed below?
POPULATION INTERVENTION COMPARATOR OUTCOMES
MDR-tb without resist-ance to the second-line drugs
A second-line regimena which includes: • Cured/completed by end of treatment
• Failure
• Relapse
• survival (or death)
• Adverse reactions from tb drugs (severity, type, organ class)
– pyrazinamide – no pyrazinamide
– injectable agents (Km/Am/Cm)
– no injectable agents (Km/Am/Cm)
– prothionamide/ethionamide – no prothionamide/ethionamide
– cycloserine or terizidone – no cycloserine or terizidone
– PAs – no PAs
– later-generation fluoroquinolone2 – no later-generation fluoroquinoloneb
– high-dose isoniazid – no high-dose isoniazid
– clofazimine – no clofazimine
– linezolid – no linezolid
– other individual Group 5 drugs – no other individual Group 5 drugs
a Data from regimens lasting up to 12 months were not included in this question.b Moxifloxacin or gatifloxacin; any use of standard or high-dose levofloxacin was included as levofloxacin use.
For all analyses, treatment outcomes were dichotomized as either successful or unsuccessful. Successful outcome was defined as when cure was achieved or treatment was completed, and unsuccessful outcome was defined as failure, relapse or death. There were inadequate numbers of events to support analysis of
80
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
failure/relapse. The primary analyses estimated the odds of treatment success (versus fail/relapse/death) associated with the use of each drug among patients with bacteriologically confirmed MDR-TB but without confirmed XDR-TB. To assess the effect of fluoroquinolones, the use of any later generation fluoroquinolone was compared to a regimen excluding the use of a later generation fluoroquinolone.
All analyses were repeated on patients with clinically diagnosed MDR-TB (i.e. not bacteriologically confirmed MDR-TB), where the data supported analyses. Children with confirmed XDR-TB were excluded from the primary analysis.
For all adjusted analyses, we fitted random-effects logistic regressions (random intercept and random slopes, when possible, and only random intercept when not) by maximum likelihood with quadrature approximation, using PROC GLIMMIX in SAS software (version 9.4, SAS Institute, Cary, North Carolina). Patients were considered to be clustered within studies, and intercepts and slopes of the main exposure variables were allowed to vary across studies. This was to account for unmeasured differences between patient populations across studies, as well as site-specific differences in data ascertainment, measurement and other factors. Estimates were adjusted for four covariates: age (dichotomized as under five years old and 5–15 years old), sex, HIV infection and severe TB disease (defined as being underweight or malnourished, having oedema, having low weight for age, having severe extrapulmonary disease, or having severe disease on chest radiograph). In order to improve data modelling, given some missing data on HIV status, children from countries with very low HIV prevalence who did not have an HIV test done were assumed to be HIV negative, following consultation with the study investigators. For the main analyses, single imputation (as opposed to multiple imputation) was performed where missing values for the four covariates used in multivariable analyses were substituted with the mean value from the other participants of the same study to which the individual belonged. In sensitivity analyses, multiple imputation using chained equations was used for missing values. All statistical analyses were performed using SAS 9.4.
Data on adverse events were sparse, and therefore we chose to provide descriptive analysis only for key toxicities in studies consistently reporting adverse events; in particular the incidence of ototoxicity (descriptive analysis only) because it is a frequent and serious side effect of aminoglycosides, which are a cornerstone of treatment, and of particular interest to health care providers and patients with MDR-TB.
Assessment of overall quality of evidenceThe quality of studies was described using a modified Newcastle-Ottawa tool (Appendix 6B) adapted for use in paediatric MDR-TB. We assessed the quality of evidence across the studies with Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology (7) defining the quality of evidence for each outcome as “the extent to which one can be confident that an estimate of effect or association is close to the quantity of specific interest” (8). The quality rating across studies has four levels: high, moderate, low or very low. Randomized controlled trials are initially categorized as providing high quality evidence, but the quality can be downgraded. Similarly, other types of controlled trials and observational studies are categorized as providing low quality evidence but the quality can be upgraded if justified. Factors that decrease the quality of evidence include limitations in design, indirectness of evidence, unexplained heterogeneity or inconsistency of results, imprecision of results or high probability of publication bias. Factors that can increase the quality level of a body of evidence include studies with a large magnitude of effect, and studies in which all plausible confounding would lead to an underestimation of effect.
81
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
EthicsThe Health Research Ethics Committee of the Faculty of Medicine and Health Sciences and Stellenbosch University provided ethical approval for this study.
Results
Search results and report selectionFigure A6.3.1 presents a summary of the search results and report selection. Results from searching the database and other sources yielded 2771 search results, which were narrowed down to 242 results, after screening of abstracts (Figure A6.3.1). Of these 242 papers reviewed and 210 were excluded. Included in these excluded studies were 89 studies in which the authors were contacted when eligibility criteria were unclear; of those authors, 48 never replied, 18 confirmed that their study did not meet inclusion criteria and 1 study was rejected as it did not include primary data (a systematic review). It’s important to note that the vast majority of these queries were for studies that were primarily adult studies but may have possibly contained a small number of children, however the precise number was often not specified. It is therefore unlikely that a large number of children were missed from these excluded studies.
Twenty-seven studies (from 32 studies requested) provided individual patient data (9–26) that included data from 974 patients. Two authors declined to share data, two authors could not get Internal Review Board permission in time to share their data and one author no longer had access to the primary data (Figure A6.3.1).
82
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Figure A6.3.1. Flow diagram of study selection for systematic review and individual patient meta-analysis of children with multidrug-resistant tuberculosis
Report characteristicsGeographic distribution of the data received from sites from 18 countries is shown in Figure A6.3.2. Patients from six continents were included; the majority were from Africa. Four countries (India, Pakistan, Russia and South Africa) are among the 30 high-burden TB countries. Details of the included studies are presented in Appendix 6B.
No. of studies for which IPD was not provided
n=5n=2: authors declinedn=2: ethical approval not obtainedn=1: data no longer available
No. of studies identified through other sources
n=10
No. of studies identified through database searching
n=2,761
No. of studies screenedn=2,771
No. of studies excludedn=2,529
No. of studies excludedn=210
n=125: <3 children includedn=18: Duplicate study populationn=48: unclear; no reply from authorn=18: did not meet eligibilityn=1: not primary data
No. of studies for which IPD was requested
n=32
No. of studies for which IPD was provided, n=27
No. of participants for which IPD was provided, n=974
No. of full text studies screened for eligibility
n=242
83
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
Figure A6.3.2. Details of geographic locations of patients included in the individual patient data meta-analysis (the number indicates the number of participants included at each location)
Summary patient data and outcomesData from 974 children that were included in the analysis showed median age of 7.1 years and 44% males. The HIV status was known in 822 children, of whom 44% had HIV infection. Two-hundred thirty seven children had clinically diagnosed MDR-TB, and 737 children had bacteriologically confirmed MDR-TB. Of those with a confirmed diagnosis, 35 had MDR-TB with additional resistance to a fluoroquinolone, 28 had MDR-TB with additional second-line injectable resistance and 36 had XDR-TB (i.e. resistance to both a fluoroquinolone and a second-line injectable). Children with XDR-TB were not included in the primary analysis and data from children with additional resistance to a fluoroquinolone or a second-line injectable agent were combined with MDR-TB (see methods). Key clinical and demographic characteristics, stratified by clinically diagnosed versus bacteriologically confirmed MDR-TB are shown in Table A6.3.2. Treatment outcomes summarized for the entire data set are shown in Table A6.3.3. Some children were listed as “cured”, because they were bacteriologically diagnosed with TB disease, but their MDR-TB was not confirmed by drug-susceptibility testing.
84
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
Table A6.3.2. Key demographic and clinical characteristics among children with clinically diagnosed or bacteriologically confirmed multidrug-resistant tuberculosis
CLINICALLY DIAGNOSED MDR-TB
N = 237 (%)
BACTERIOLOGICALLY CONFIRMED MDR OR PRE-XDR
N = 701 (%)
Age
• Under 5 years
• 5 to 15 years156 (66)81 (34)
231 (33)470 (67)
Malnourished*
• Yes
• No
• Unknown
47 (19.8)67 (28.3)
123 (51.9)
274 (39.1)366 (52.2)
61 (8.7)
severe Disease on chest radiograph**
• Yes
• No
• Unknown
68 (28.7)126 (53.2)43 (18.1)
519 (74.0)163 (23.3)
19 (2.7)
severe Extrapulmonary Disease*** 24 (10.1) 103 (14.7)
HIV status
• HIV-infected
• HIV-uninfected
• HIV status unknown
36 (15.2)141 (59.5)60 (25.3)
318 (45.4)300 (42.8)83 (11.8)
A combined variable of “severe disease: included the presence of malnutrition* or severe disease on chest radiograph** or severe extrapulmonary disease***, given missing data on these three variables, individually.
Table A6.3.3. Summary of treatment outcomes for children with multidrug-resistant tuberculosis
CLINICALLY DIAGNOSED MDR-TB
N = 237
CONFIRMED MDR-TB WITH-OUT CONFIRMED XDR-TB
N = 701
CONFIRMED XDR-TB
N = 36
Cured 46 (19.3%) 327 (46.6%) 23 (64%)
Completed treatment 166 (69.7%) 209 (29.8%) 7 (19%)
Fail or relapse 0 14 (1.9%) 1 (3%)
Death 7 (2.9%) 73 (10.4%) 3 (8%)
lost to follow-up 18 (8%) 77 (11%) 2 (6%)
The results for primary outcome analysis of treatment benefit for individual drugs are presented in GRADE tables (see Annex 4). Note that although numbers are small, children who did not receive a second-line injectable, but who had clinically diagnosed/unconfirmed disease (and in general had less severe disease, Table A6.3.2) did well (93.5% successful outcomes, see GRADE table for second-line injectables). Children who did not receive a second-line injectable tended to have less severe TB disease overall, and have lower rates of malnutrition, severe disease (on chest radiograph) and severe extrapulmonary TB disease (see Table A6.3.4).
85
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
Adverse events were in general poorly reported. Only nine datasets consistently reported toxicities, and these data are presented in Table A6.3.5.
Table A6.3.4. Characteristics of children who received a second-line injectable (SLI) versus those who did not
TREATED WITH SLI NOT TREATED WITH SLI
N % N %
severe disease 296 91% 30 9%
Malnourished 239 93% 18 7%
severe pulmonary disease on chest radiograph 479 92% 40 8%
severe extrapulmonary disease 85 83% 18 17%
Table A6.3.5. Frequency of specific toxicities in studies that reported adverse events (N=306 subjects in 9 datasets where adverse events were consistently reported)
SIDE EFFECT NUMBER OF CHILDREN DENOMINATOR % EVENTS
Hearing loss 39 383 10.1%
Peripheral neuropathy 2 383 0.5%
Optic neuropathy 11 383 2.9%
thyroid dysfunction 50 383 13.1%
liver toxicity 16 383 4.2%
Arthropathy 20 383 5.2%
Nephrotoxicity 2 383 0.5%
Note: secondary analyses are planned in the future, including analyses of outcomes in children with shorter durations of injectable treatment, stratified by severe and non-severe tb disease, and by bacteriological status.
DiscussionThis first ever systematic review and IPD of paediatric MDR-TB, which included a large number of children, shows overall good treatment outcome with 76.4% of children with bacteriologically confirmed MDR-TB, and 89% of children with unconfirmed MDR-TB, having successful treatment outcomes. More than two-thirds of children had bacteriologically confirmed MDR-TB, which strengthens the quality of data from this review, given the fact that TB in children is typically paucibacillary in nature. The overall mortality was low (10.4% in children with confirmed disease, and 2.9% in the unconfirmed group). Children with confirmed XDR-TB although a small number (n=36) had favourable treatment outcomes in 83% of cases.
Most children (77%) received injectable drugs. There was a high prevalence of HIV infection, with 45% of children in the confirmed and 15% in the unconfirmed MDR-TB group, being HIV-infected; HIV testing was relatively complete. A high proportion of children had chest radiographic or features of
86
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
extrapulmonary disease compatible with severe TB, as well as malnutrition. The overall good treatment outcomes should therefore be seen in the context of HIV co-infection and more severe disease, indicating that good outcomes are achievable in children with MDR-TB.
There are a number of notable findings from the analysis of the impact of individual second-line drugs. The use of second-line injectable drugs significantly predicted the treatment success versus failure/relapse/death (OR: 3.32; 95% CI: 1.53–7.21) in children with confirmed MDR-TB. This has to be seen in the context of high toxicity with hearing loss reported in 10.1% of the studies that reported on safety outcomes. However, children with clinically (unconfirmed) diagnosed MD-TB (who tended to have less severe disease) had good outcomes when not treated with a second-line injectable. These data are supportive of the practice of using injectable sparing regimens in children with less severe disease in order to spare children from SAEs associated with second-line injectables, without an adverse impact on treatment outcomes.
High-dose isoniazid (used in approximately 25% of subjects, at a dose of 15–20 mg/kg), predicted treatment success (versus failure/relapse/death; OR: 6.97; 95% CI: 2.11–23.03), even after adjusting for site in the analysis (high-dose isoniazid was most frequently used in South African sites). Another consideration to this finding is that high-dose isoniazid is also typically used in combination with ethionamide/prothionamide.
Later-generation fluoroquinolones (primarily moxifloxacin; since there was virtually no reported use of gatifloxacin) did not appear to offer a treatment benefit. However, only 10% of cases in the dataset received later-generation fluoroquinolones, which may have masked any benefit associated with their use.
In general, these findings of individual drug effects should be taken in the context of the use of multiple drugs as part of MDR-TB treatment regimens.
LimitationsThis individual patient data (IPD) is limited by the overall low quality of studies; most studies included were retrospective or prospective observational cohort studies and there were no trials that could be included. The overall sample size was modest compared to adult IPD datasets, and estimates were frequently imprecise while some associations were not estimated due to limited data (e.g. relapse/failure). There were no data available yet on the novel drugs, delamanid and bedaquiline, for inclusion in this IPD. We also had a very small sample of children with confirmed XDR-TB. Toxicity was frequently poorly assessed while missing data regarding HIV status was handled by imputation. Missing data on individual variables of nutritional status and disease severity were handled using a composite disease severity variable.
In summary, this first paediatric specific IPD has provided data for guideline development of high clinical and programmatic relevance. Overall, the proportion of children with favourable treatment outcome, even with severe TB and with HIV co-infection, was good. Data regarding the use of novel drugs, bedaquiline and delamanid in children, are urgently needed. Future questions should focus on the use of shorter (<18 months) regimens and injectable-sparing regimens in children. Paediatric-specific treatment evidence is important to allow for inclusion of children in evidence-based MDR-TB treatment guidelines.
87
ANNEX 6: sUMMARIEs OF UNPUblIsHED DAtA UsED FOR tHE RECOMMENDAtIONs
References1. Jenkins HE, Tolman AW, Yuen CM, Parr JB, Keshavjee S, Pérez-Vélez CM, et al. Incidence of multidrug-resistant
tuberculosis disease in children: systematic review and global estimates. Lancet. 2014; 383(9928):1572–9.
2. Wallgren A. Primary tuberculous infections in young adult life and in childhood. Am J Dis Child. 1941; 61: 577–589
3. Ettehad D, Schaaf HS, Seddon JA, Cooke GS, Ford N. Treatment outcomes for children with multidrug resistant tuberculosis: a systemic review and meta-analysis. Lancet Infec Dis. 2012; 12: 449–56.
4. Guidance for national tuberculosis programmes on the management of tuberculosis in children: second edition. Geneva: World Health Organization; 2014 (http://www.who.int/tb/publications/childtb_guidelines/en/, accessed 14 May 2016).
5. Definitions and reporting framework for tuberculosis: 2013 revision (updated December 2014). Geneva: World Health Organization; 2013 (http://apps.who.int/iris/bitstream/10665/79199/1/9789241505345_eng.pdf, accessed 14 May 2016).
6. Ahuja SD, Ashkin D, Avendano M, Banerjee R, Bauer M, Bayona JN, et al; Collaborative Group for Meta-Analysis of Individual Patient Data in MDR-TB. Multidrug resistant pulmonary tuberculosis treatment regimens and patient outcomes: an individual patient data meta-analysis of 9,153 patients. PLoS Med. 2012;9(8):e1001300. Epub 2012 Aug 28. PMID: 22952439
7. Guyatt GH, Oxman AD, Schunemann HJ, Tugwell P, Knottnerus A. GRADE guidelines: a new series of articles in the Journal of Clinical Epidemiology. J Clin Epidemiol. 2011;64(4):380–2.
8. Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011 (http://www.cochrane-handbook.org, accessed 14 May 2016).
9. Chan ED, Laurel V, Strand MJ, Chan JF, Huynh ML, Goble M, et al. Treatment and outcome analysis of 205 patients with multidrug-resistant tuberculosis. Am J Respir Crit Care Med. 2004;169(10):1103–9.
10. Drobac PC, Mukherjee JS, Joseph JK, Mitnick C, Furin JJ, del Castillo H, et al. Community-based therapy for children with multidrug-resistant tuberculosis. Pediatrics. 2006;117(6):2022–9.
11. Fairlie L, Beylis NC, Reubenson G, Moore DP, Madhi SA. High prevalence of childhood multi-drug resistant tuberculosis in Johannesburg, South Africa: a cross sectional study. BMC Infect Dis. 2011;11:28.
12. Gegia M, Jenkins HE, Kalandadze I, Furin J. Outcomes of children treated for tuberculosis with second-line medications in Georgia, 2009–2011. Int J Tuberc Lung Dis. 2013;17(5):624–9.
13. Banerjee R, Allen J, Westenhouse J, Oh P, Elms W, Desmond E, et al. Extensively drug-resistant tuberculosis in California, 1993–2006. Clin Infect Dis. 2008;47(4):450–7.
14. Hicks RM, Padayatchi N, Shah NS, Wolf A, Werner L, Sunkari VB, et al. Malnutrition associated with unfavorable outcome and death among South African MDR-TB and HIV co-infected children. Int J Tuberc Lung Dis. 2014;18(9):1074–83.
15. Isaakidis P, Paryani R, Khan S, Mansoor H, Manglani M, Valiyakath A, et al. Poor outcomes in a cohort of HIV-infected adolescents undergoing treatment for multidrug-resistant tuberculosis in Mumbai, India. PLoS One. 2013; 8(7):e68869.
16. Datta BS, Hassan G, Kadri SM, Qureshi W, Kamili MA, Singh H, et al. Multidrug-resistant and extensively drug resistant tuberculosis in Kashmir, India. J Infect Dev Ctries. 2009;4(1):19–23.
17. Méndez Echevarría A, Baquero Artigao F, García Miguel MJ, Rojo Conejo P, Ballesteros Díez Y, Rubio Gribble B, et al. Multidrug-resistant tuberculosis in the pediatric age group. An Pediatr (Barc). 2007;67(3):206–11.
18. Moore BK, Anyalechi E, van der Walt M, Smith S, Erasmus L, Lancaster J, et al. Epidemiology of drug-resistant tuberculosis among children and adolescents in South Africa, 2005–2010. Int J Tuberc Lung Dis. 2015;19(6):663–9.
19. Kuksa L, Riekstina V, Leimane V, Ozere I, Skenders G, Van den Bergh R, et al. Multi- and extensively drug-resistant tuberculosis in Latvia: trends, characteristics and treatment outcomes. Public Health Action. 2014; 4(Suppl 2):S47–53.
20. Santiago B, Baquero-Artigao F, Mejías A, Blázquez D, Jiménez MS, Mellado-Peña MJ; EREMITA Study Group. Pediatric drug-resistant tuberculosis in Madrid: family matters. Pediatr Infect Dis J. 2014;33(4):345–50.
21. Seddon JA, Hesseling AC, Willemse M, Donald PR, Schaaf HS. Culture-confirmed multidrug-resistant tuberculosis in children: clinical features, treatment, and outcome. Clin Infect Dis. 2012;54(2):157–66.
88
WHO TREATMENT GUIDELINES FOR DRUG-RESISTANT TUBERCULOSIS, 2016 UPDATE
22. Seddon JA, Hesseling AC, Godfrey-Faussett P, Schaaf HS. High treatment success in children treated for multidrug-resistant tuberculosis: an observational cohort study. Thorax. 2014;69(5):458–64.
23. Kim DH, Kim HJ, Park SK, Kong SJ, Kim YS, Kim TH, et al. Treatment outcomes and long-term survival in patients with extensively drug-resistant tuberculosis. Am J Respir Crit Care Med. 2008;178(10):1075–82.
24. Geerligs WA, Van Altena R, De Lange WCM, Van Soolingen D, Van Der Werf TS. Multidrug-resistant tuberculosis: long-term treatment outcome in the Netherlands. Int J Tuberc Lung Dis. 2000;4(8):758–64.
25. Williams B, Ramroop S, Shah P, Anderson L, Das S, Riddell A, et al. Multidrug-resistant tuberculosis in UK children: presentation, management and outcome. Eur Respir J. 2013;41(6):1456–8.
26. Kim HR, Hwang SS, Kim HJ, Lee SM, Yoo CG, Kim YW, et al. Impact of extensive drug resistance on treatment outcomes in non-HIV-infected patients with multidrug-resistant tuberculosis. Clin Infect Dis. 2007;45(10):1290–5.
89
APPENDIX 6A
Search strategies
DATABASE: PUBMED; SEARCH DATE: 01 OCTOBER 2014
SEARCH QUERYITEMS FOUND
#25 search ((#22 AND #23) NOt (animals[mh] NOt humans[mh])) 1653
#24 search (#22 AND #23) 1653
#23 search (infant[mh] OR infant[tiab] OR infants[tiab] OR infancy[tiab] OR toddler*[tiab] OR preterm*[tiab] OR prematur*[tiab] OR postmatur*[tiab] OR baby[tiab] OR babies[tiab] OR neonat*[tiab] OR newborn[tiab] OR preschool*[tiab] OR pre-school*[tiab] OR child[mh] OR child*[tiab] OR kindergar*[tiab] OR pupil*[tiab] OR schoolchild*[tiab] OR teen*[tiab] OR youth[tiab] OR youths[tiab] OR youngster*[tiab] OR young person*[tiab] OR young people[tiab] OR minors[mh] OR minors[tiab] OR puberty[mh] OR puberty[tiab] OR pubescen*[tiab] OR prepubescen*[tiab] OR paediatric*[tiab] OR pediatric*[tiab] OR peadiatric*[tiab] OR schools[mh:noexp] OR school*[tiab] OR kid[tiab] OR kids[tiab] OR boy*[tiab] OR girl*[tiab] OR creche*[tiab] OR highschool*[tiab] OR juvenil*[tiab] OR adolescent[mh] OR adolescen*[tiab] OR under ag*[tiab] OR underag*[tiab])
3521601
#22 search (tuberculosis, multidrug-resistant[mh] OR multidrug resistant tuberculosis[tiab] OR drug resistant tuberculosis[tiab] OR multiple drug resistant tuberculosis[tiab] OR MDR tuberculosis[tiab] OR MDR tb[tiab] OR MDRtb[tiab] OR ((drug resistance[tiab] OR multidrug resistance[tiab] OR multiple drug resistance[tiab] OR multiresistant[tiab] OR multi resistant[tiab]) AND (tuberculosis[tiab] OR tb[tiab]))
8600
DATABASE: EMBASE; SEARCH DATE: 01 OCTOBER 2014
NO. QUERY RESULTS
#8 #3 NOt #7 1837
#7 #4 NOt #6 5002895
#6 #4 AND #5 1303481
#5 ‘human’/de OR ‘normal human’/de OR ‘human cell’/de 15207023
#4 ‘animal’/de OR ‘animal experiment’/de OR ‘invertebrate’/de OR ‘animal tissue’/de OR ‘animal cell’/de OR ‘nonhuman’/de
6306376
#3 #1 AND #2 1889
#2 ‘infant’/exp OR infant:ab,ti OR infants:ab,ti OR infancy:ab,ti OR toddler*:ab,ti OR preterm*:ab,ti OR prematur*:ab,ti OR postmatur*:ab,ti OR baby:ab,ti OR babies:ab,ti OR neonat*:ab,ti OR newborn:ab,ti OR preschool*:ab,ti OR pre+school*:ab,ti OR ‘child’/exp OR child*:ab,ti OR kindergar*:ab,ti OR pupil*:ab,ti OR schoolchild*:ab,ti OR teen*:ab,ti OR youth:ab,ti OR youths:ab,ti OR youngster*:ab,ti OR ‘young person’:ab,ti OR ‘young persons’:ab,ti OR ‘young people’:ab,ti OR ‘minors’/exp OR minors:ab,ti OR ‘puberty’/exp OR puberty:ab,ti OR pubescen*:ab,ti OR prepubescen*:ab,ti OR paediatric*:ab,ti OR pediatric*:ab,ti OR peadiatric*:ab,ti OR ‘schools’/exp OR school*:ab,ti OR kid:ab,ti OR kids:ab,ti OR boy*:ab,ti OR girl*:ab,ti OR creche*:ab,ti OR highschool*:ab,ti OR ‘juvenile’/exp OR juvenil*:ab,ti OR ‘adolescent’/exp OR adolescen*:ab,ti OR (under NEXt/1 ag*):ab,ti OR underag*:ab,ti
4649411
#1 ‘multidrug resistant tuberculosis’/exp OR ‘multidrug resistant tuberculosis’:ab,ti OR ‘drug resist-ant tuberculosis’:ab,ti OR ‘multiple drug resistant tuberculosis’:ab,ti OR ‘mdr tuberculosis’:ab,ti OR ‘mdr tb’:ab,ti OR (‘drug resistance’:ab,ti OR ‘multidrug resistance’:ab,ti OR ‘multiple drug resistance’:ab,ti OR multiresistant:ab,ti AND (tuberculosis:ab,ti OR tb:ab,ti))
9249
90
APP
END
IX 6
B
Tabl
e of
incl
uded
stu
dies
SITE
/AU
THO
RC
OU
NTR
YN
UM
BER
PR
EVIO
US I
PD
IN
CLU
SIO
NN
UM
BER
OF
SU
BJE
CTS
PU
BLI
CAT
ION
STA
TUS
TITL
E O
F PA
PER
STU
DY
DES
IGN
STU
DY
PO
PU
LATI
ON
Acha
rUz
beki
stan
1No
20Un
publ
ishe
dN/
ARe
trosp
ectiv
e co
hort
bact
erio
logi
cally
con
-fir
med
and
clin
ical
ly
diag
nose
d
Aman
ulla
hPa
kist
an2
No29
Unpu
blis
hed
N/A
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
bane
rjee
UsA
9M
enzie
s IP
D 20
123
Publ
ishe
dEx
tens
ively
dru
g-re
sist
ant t
uber
cu-
losi
s in
Cal
iforn
ia, 1
993–
2006
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
Chan
taiw
an4
No4
Unpu
blis
hed
N/A
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
Chan
UsA
5M
enzie
s IP
D 20
123
Publ
ishe
dtre
atm
ent a
nd o
utco
me
anal
ysis
of
205
patie
nts
with
mul
tidru
g-re
sist
-an
t tub
ercu
losi
s
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
Chio
tan
Rom
ania
16No
17Un
publ
ishe
dN/
ARe
trosp
ectiv
e co
hort
bact
erio
logi
cally
co
nfirm
ed
Datta
Indi
a12
No3
Publ
ishe
dM
ultid
rug-
resi
stan
t and
ext
en-
sive
ly d
rug
resi
stan
t tub
ercu
losi
s in
Ka
shm
ir, In
dia
Pros
pect
ive c
ohor
tba
cter
iolo
gica
lly
confi
rmed
Drob
acPe
ru6
No36
Publ
ishe
dCo
mm
unity
-bas
ed th
erap
y fo
r ch
ildre
n wi
th m
ultid
rug-
resi
stan
t tu
berc
ulos
is
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly c
on-
firm
ed a
nd c
linic
ally
di
agno
sed
Fairl
ieso
uth
Afric
a 7
No10
Publ
ishe
dHi
gh p
reva
lenc
e of
chi
ldho
od
mul
ti-dr
ug re
sist
ant t
uber
culo
sis
in Jo
hann
esbu
rg, s
outh
Afri
ca: a
cr
oss
sect
iona
l stu
dy
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
9191
APPE
NDIX
6b:
tAbl
E OF
INCl
UDED
stU
DIEs
SITE
/AU
THO
RC
OU
NTR
YN
UM
BER
PR
EVIO
US I
PD
IN
CLU
SIO
NN
UM
BER
OF
SU
BJE
CTS
PU
BLI
CAT
ION
STA
TUS
TITL
E O
F PA
PER
STU
DY
DES
IGN
STU
DY
PO
PU
LATI
ON
Gegi
aGe
orgi
a8
No55
Publ
ishe
dOu
tcom
es o
f chi
ldre
n tre
ated
for
tube
rcul
osis
with
sec
ond-
line
med
-ic
atio
ns in
Geo
rgia
, 200
9–20
11
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly c
on-
firm
ed a
nd c
linic
ally
di
agno
sed
Hick
sso
uth
Afric
a 22
No82
Publ
ishe
dM
alnu
tritio
n as
soci
ated
with
unf
a-vo
urab
le o
utco
mes
and
dea
th
amon
g so
uth
Afric
an M
DR-t
b an
d HI
V co
-infe
cted
chi
ldre
n
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
Isaa
kidi
sIn
dia
11No
8Pu
blis
hed
Poor
out
com
es in
a c
ohor
t of H
IV-
infe
cted
ado
lesc
ents
und
ergo
ing
treat
men
t for
mul
tidru
g-re
sist
ant
tube
rcul
osis
in M
umba
i, In
dia
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
Mar
yand
yshe
vRu
ssia
17No
38Un
publ
ishe
dN/
ARe
trosp
ectiv
e co
hort
bact
erio
logi
cally
con
-fir
med
and
clin
ical
ly
diag
nose
d
Men
dez-
Eche
varri
asp
ain
13No
8Pu
blis
hed
Mul
tidru
g-re
sist
ant t
uber
culo
sis
in
the
pedi
atric
age
gro
upRe
trosp
ectiv
e co
hort
bact
erio
logi
cally
con
-fir
med
and
clin
ical
ly
diag
nose
d
Moo
reso
uth
Afric
a 14
No33
9Pu
blis
hed
Epid
emio
logy
of d
rug-
resi
st-
ant t
uber
culo
sis
amon
g ch
ildre
n an
d ad
oles
cent
s in
sou
th A
frica
, 20
05–2
010
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly c
on-
firm
ed a
nd c
linic
ally
di
agno
sed
Ozer
e/ K
uksa
latv
ia15
No53
som
e pu
blis
hed
Mul
ti an
d ex
tens
ively
dru
g-re
sist
-an
t tub
ercu
losi
s in
lat
via:
tren
ds,
char
acte
ristic
s an
d tre
atm
ent
outc
omes
.
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly c
on-
firm
ed a
nd c
linic
ally
di
agno
sed
Ryba
kUk
rain
e18
No6
Unpu
blis
hed
N/A
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly c
on-
firm
ed a
nd c
linic
ally
di
agno
sed
sant
iago
spai
n19
No10
som
e pu
blis
hed
Pedi
atric
dru
g-re
sist
ant t
uber
culo
-si
s in
Mad
rid: f
amily
mat
ters
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly c
on-
firm
ed a
nd c
linic
ally
di
agno
sed
92
WHO
TREA
TMEN
T GU
IDEL
INES
FOR
DRU
G-RE
SIST
ANT T
UBER
CULO
SIS,
201
6 UP
DATE
SITE
/AU
THO
RC
OU
NTR
YN
UM
BER
PR
EVIO
US I
PD
IN
CLU
SIO
NN
UM
BER
OF
SU
BJE
CTS
PU
BLI
CAT
ION
STA
TUS
TITL
E O
F PA
PER
STU
DY
DES
IGN
STU
DY
PO
PU
LATI
ON
sedd
on (1
)so
uth
Afric
a 20
No88
Publ
ishe
dCu
lture
con
firm
ed m
ultid
rug
resi
st-
ant t
uber
culo
sis
in c
hild
ren:
cl
inic
al fe
atur
es, t
reat
men
t and
ou
tcom
e
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
sedd
on (2
)so
uth
Afric
a 21
No13
1Pu
blis
hed
High
trea
tmen
t suc
cess
in c
hil-
dren
trea
ted
for m
ultid
rug
resi
stan
t tu
berc
ulos
is: a
n ob
serv
atio
nal
coho
rt st
udy
Pros
pect
ive c
ohor
tba
cter
iolo
gica
lly c
on-
firm
ed a
nd c
linic
ally
di
agno
sed
shar
ma
Indi
a23
No8
Unpu
blis
hed
N/A
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
shim
/Kim
Kore
a24
Men
zies
IPD
2014
2Pu
blis
hed
treat
men
t out
com
es a
nd lo
ng-te
rm
surv
ival i
n pa
tient
s wi
th e
xten
sive
ly
drug
-resi
stan
t tub
ercu
losi
s
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
skra
hina
bela
rus
25No
5Un
publ
ishe
dN/
A
bact
erio
logi
cally
con
-fir
med
and
clin
ical
ly
diag
nose
d
Van
Der W
erf/
Ge
erlin
gsNe
ther
land
s26
Men
zies
IPD
2014
2Pu
blis
hed
Mul
tidru
g-re
sist
ant t
uber
culo
sis:
lo
ng-te
rm tr
eatm
ent o
utco
me
in
the
Neth
erla
nds
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
Will
iam
sUn
ited
King
dom
3No
Pu
blis
hed
Mul
tidru
g-re
sist
ant t
uber
culo
sis
in
UK c
hild
ren:
pre
sent
atio
n, m
an-
agem
ent a
nd o
utco
me
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed
Yim
/ Ki
mKo
rea
27M
enzie
s IP
D 20
141
Publ
ishe
dIm
pact
of e
xten
sive
dru
g re
sist
-an
ce o
n tre
atm
ent o
utco
mes
in
non-
HIV-
infe
cted
pat
ient
s wi
th
mul
tidru
g-re
sist
ant t
uber
culo
sis
Retro
spec
tive
coho
rtba
cter
iolo
gica
lly
confi
rmed