We thank all the patients and staff from all the centres participating in the REALITY trial. Joint Clinical Research Centre, Kampala, Uganda: P Mugyenyi, C Kityo, V Musiime, P Wavamunno, E Nambi, P Ocitti, M Ndigendawani; JCRC, Fort Portal, Uganda: S Kabahenda, M Kemigisa, J Acen, D Olebo, G Mpamize, A Amone, D Okweny, A Mbonye, F Nambaziira, A Rweyora, M Kangah, V Kabaswahili; JCRC. Gulu, Uganda: J Abach, G Abongomera, J Omongin, I Aciro, A Philliam, B Arach, E

Ocung, G Amone, P Miles, C Adong, C Tumsuiime, P Kidega, B Otto, F Apio; JCRC. Mbale, Uganda: K Baleeta, A Mukuye, M Abwola, F Ssennono, D Baliruno, S Tuhirwe, R Namisi, F Kigongo, D Kikyonkyo, F Mushahara, D Okweny, J Tusiime, A Musiime, A Nankya, D Atwongyeire, S Sirikye, S Mula, N Noowe; JCRC. Mbarara, Uganda: A Lugemwa, M Kasozi, S Mwebe, L Atwine, T Senkindu, T Natuhurira, C Katemba, E Ninsiima, M Acaku, J Kyomuhangi, R Ankunda, D Tukwasibwe, L Ayesiga; University of

Zimbabwe, Harare, Zimbabwe: J Hakim, K Nathoo, M Bwakura-Dangarembizi, A Reid, E Chidziva, T Mhute, GC Tinago, J Bhiri, S Mudzingwa, M Phiri, J Steamer, R Nhema, C Warambwa, G Musoro, S Mutsai, B Nemasango, C Moyo, S Chitongo, K Rashirai, S Vhembo, B Mlambo, S Nkomani, B Ndemera, M Willard, C Berejena, Y Musodza, P Matiza, B Mudenge, V Guti; KEMRI Wellcome Trust Research Programme, Kilifi, Kenya: A Etyang, C Agutu, J Berkley, K Maitland, P Njuguna, S Mwaringa, T Etyang, K

Awuondo, S Wale, J Shangala, J Kithunga, S Mwarumba, S Said Maitha, R Mutai, M Lozi Lewa, G Mwambingu, A Mwanzu, C Kalama, H Latham, J Shikuku, A Fondo, A Njogu, C Khadenge, B Mwakisha; Moi University Clinical Research Centre, Eldoret, Kenya: A Siika, K Wools-Kaloustian, W Nyandiko, P Cheruiyot, A Sudoi, S Wachira, B Meli, M Karoney, A Nzioka, M Tanui, M Mokaya, W Ekiru, C Mboya, D Mwimali, C Mengich, J Choge, W Injera, K Njenga, S Cherutich, M Anyango Orido, G Omondi Lwande, P

Rutto, A Mudogo, I Kutto, A Shali, L Jaika, H Jerotich, M Pierre; Department of Medicine and MLW Clinical Research Programme, College of Medicine, Blantyre, Malawi: J Mallewa, S Kaunda, J Van Oosterhout, B O'Hare, R Heydermann, C Gonzalez, N Dzabala, C Kelly, B Denis, G Selemani, L Nyondo Mipando, E Chirwa, P Banda, L Mvula, H Msuku, M Ziwoya, Y Manda, S Nicholas, C Masesa, T Mwalukomo, L Makhaza, I Sheha, J Bwanali, M Limbuni; MRC Clinical Trials Unit at UCL, London, UK:D Gibb,

M Thomason, AS Walker, S Pett, A Szubert, A Griffiths, H Wilkes, C Rajapakse, M Spyer, A Prendergast, N Klein. Independent REALITY Trial Monitors: F Kyomuhendo, S Nakalanzi, J Peshu, S Ndaa, J Chabuka, N Mkandawire, L Matandika, C Kapuya. Social Science Group: F Cowan, J Seeley, S Bernays, R Kawuma, Z Mupambireyi. Trial Steering Committee: I Weller (Chair), E Malianga, C Mwansambo, F Miiro, P Elyanu, E Bukusi, E Katabira, O Mugurungi, D Gibb, J Hakim, A Etyang, P Mugyenyi, J Mallewa.

Data Monitoring Committee: T Peto (Chair), P Musoke, J Matenga, S Phiri. Endpoint Review Committee: H Lyall (Co-Chair), V Johnston (Co-Chair), F Fitzgerald, F Post, F Ssali, A Prendergast, A Turkova, A Bamford, A Arenas-Pinto.

Funding: REALITY was funded by the UK Department for International Development (DFID), the Wellcome Trust and the Medical Research Council (MRC). Additional funding support was provided by the PENTA foundation. Merck Sharp & Dohme, Gilead Sciences, Cipla Ltd, ViiV Healthcare/GlaxoSmithKline donated drugs for REALITY and Valid International supplied Ready-to-Use-Supplementary-Food (RUSF).

Abraham Siika1, Leanne McCabe2, Mutsa Bwakura-Dangarembizi3, Cissy Kityo4, Jane Mallewa5, Kath Maitland6, Anna Griffiths2, Keith Baleeta7, Shepherd Mudzingwa3, James Abach8, Kusum Nathoo3, Andrew J. Prendergast9, A Sarah Walker2, Diana M Gibb2 and the REALITY Trial Team 1Moi University School of Medicine, Eldoret, Kenya; 2MRC Clinical Trials Unit at UCL, London, UK; 3University of Zimbabwe Clinical Research Centre, Harare, Zimbabwe; 4Joint Clinical Research Centre, Kampala, Uganda;

5Department/College of Medicine and Malawi-Liverpool-Wellcome Trust Clinical Research Programme, Blantyre, Malawi; 6KEMRI Wellcome Trust Research Programme, Kilifi, Kenya; 7Joint Clinical Research Centre, Mbale, Uganda; 8Joint Clinical Research Centre, Gulu, Uganda; 9Queen Mary University of London, London, UK

Contact details

Professor Diana Gibb,

MRC Clinical Trials Unit at UCL

90 High Holborn, Second floor

London WC1V 6LJ

email: diana.gibb@ucl.ac.uk

Poster 896

Background: In sub-Saharan Africa, severely immunocompromized individuals are at high risk

of mortality during the first few months after starting ART. We aimed to determine predictors of

this early mortality and whether such “late presenters” could be grouped into phenotypes with

different mortality risks.

Methods: ART-naïve adults/children ≥5y with CD4<100 cells/ul initiating ART in Uganda,

Zimbabwe, Malawi and Kenya were included in the REALITY trial (ISRCTN43622374).

Baseline predictors of mortality through 48 weeks on ART were identified using Cox

regression with backwards elimination (exit p>0.1). Late presenter phenotypes were identified

using hierarchical clustering.

Results: Final multivariable models included 1711 participants (26<13y) of whom 203(12%)

died. Mortality was independently higher in those who were older (p=0.002), with lower CD4s

(p<0.001), lower albumin (p=0.001), lower haemoglobin (p=0.01) and weaker grip strength

(p=0.03); those in whom physicians reported WHO stage 3/4 weight loss (p=0.04); and in

those patients reporting fever (p=0.001), vomiting (p=0.02), some problems with mobility

(p=0.005) and inability to wash or dress themselves (p=0.003) at baseline. Five “late

presenter” groups were identified (figure), with mortality ranging from 4-25%. Group-1 had the

highest mortality (25%) and median CD4 28 cells/ul; they had a high burden of

symptoms/signs other than rash, weight loss, and problems with mobility and self-care; they

also had lower albumin and haemoglobin. Group-2 (11% mortality; median CD4 43 cells/ul)

had higher white blood cells, platelets and neutrophils, despite only 31% reporting infections

at baseline. Group-3 (10% mortality) were mainly younger women; they had similarly low

CD4s (median 27 cells/ul), haemoglobin and BMI to Group-1, but low symptom burden and

maintained fat mass. The remaining two groups had lower (4-6%) mortality, higher CD4

(median 42 and 48 cells/ul) and had predominantly maintained their weight within normal

ranges. Of note, the effect of the randomized enhanced prophylaxis bundle on mortality was

similar across all groups (interaction p=0.32).

Conclusions: Clinical and laboratory characteristics identified groups at highest risk of

mortality following ART initiation. A screening tool appropriate to the level of facility

could therefore help identify which patients with low CD4 counts should be prioritized,

e.g. for same-day ART initiation, more intensive follow-up, and enhanced prophylaxis.

ABSTRACT (updated) RESULTS

BACKGROUND

~20-25% of people starting ART in Africa have CD4 <100 cells/µL

- ~10% die within three months of starting treatment

Studies in “late presenters” are important to identify their key

characteristics, particularly if baseline CD4 count testing is not always

available, in order to identify them earlier at healthcare facilities and

prioritize them for additional interventions

In adults and older children enrolled in the REALITY trial, we

investigate in detail:

– Predictors of mortality in the first 48 weeks on ART

- Different patterns of “late presentation”

METHODS

REALITY (ISRCTN43622374) recruited 1805 HIV-

infected ART-naïve adults and children ≥5 years (98%

aged ≥13 years) from Zimbabwe, Uganda, Malawi and

Kenya with CD4<100 cells/µL

Patients initiated standard WHO-recommended

2NRTI+NNRTI with cotrimoxazole prophylaxis, and were

randomized using a factorial design to three different

additional 12 week interventions that might reduce early

mortality (enhanced antimicrobial prophylaxis, raltegravir,

ready-to-use supplementary food)

Baseline demographics, laboratory results (including VL

assayed retrospectively on stored samples), physical

LATE PRESENTATION WITH HIV IN AFRICA: PHENOTYPES, RISK AND RISK STRATIFICATION

http://www.ctu.mrc.ac.uk/our_research/research_areas/hiv/studies/reality/

ISRCTN43622374

measurements (including height, weight, MUAC,

body composition, blood pressure), physician

reported diagnoses (WHO stage 3/4 events),

participant-reported symptoms, EQ-5D scores and

randomized groups were all considered as

predictors of mortality during the first 48 weeks on

ART (using Cox models), after which participants

exited the trial.

– Factors were selected using backwards elimination (exit p>0.1

for an explanatory model). Clinical centre (reflecting

management and access to diagnostic facilities) was included

in all models which were restricted to complete cases.

Different patterns of “late presenters” were

identified using hierarchical cluster analysis.

Factor (effect in Cox models) Median (IQR) or

n (%)

Univariable effects of

factors selected for

multivariable model

Final multivariable

model in REALITY trial

participants (n=1711)

Multivariable model from

Bisson et al fitted to

REALITY trial participants

HR 95% CI P HR (95% CI P HR (95% CI P

Age (per 5 years older) 36 (29,42) 1.09 (1.03, 1.16) 0.005 1.12 (1.04, 1.20) 0.002 1.13 (1.06, 1.21) <0.001

CD4 (per 10 cells/mm3 higher) 37 (16,63) 0.89 (0.85, 0.94) <0.001 0.90 (0.85, 0.95) <0.001 0.87 (0.82, 0.92) <0.001

Haemoglobin (per g/dl higher) 11.2 (9.6,12.7) 0.77 (0.72, 0.82) <0.001 0.90 (0.83, 0.99) 0.01 0.87 (0.80, 0.94) <0.001

Albumin (per g/l higher) 35 (30,40) 0.92 (0.90, 0.93) <0.001 0.96 (0.94, 0.98) 0.001 0.94 (0.92, 0.97) <0.001

White blood cells (per 109/L higher) 3.5 (2.7,4.7) 1.17 (1.08, 1.26) <0.001 1.08 (1.00, 1.17) 0.054 1.16 (1.06, 1.25) 0.001

Grip strength (per kg higher) 24.5 (19.3,31.0) 0.93 (0.91, 0.95) <0.001 0.98 (0.96,1.00) 0.03 -

Previous healthcare contact † 163 (9%) 1.16 (0.75, 1.79) 0.50 0.63 (0.39, 1.03) 0.067 -

WHO 3/4 weight loss 327 (18%) 1.90 (1.42, 2.53) <0.001 1.43 (1.01, 2.03) 0.04 -

Patient-reported fever 240 (13%) 3.37 (2.54, 4.47) <0.001 1.67 (1.19,2.35) 0.003 -

Patient-reported vomiting 114 (6%) 3.72 (2.64, 5.26) <0.001 1.55 (1.00, 2.39) 0.048 -

EQ5D Mobility:

Some problems (vs none)

364 (20%)

4.42

(3.36, 5.81)

<0.001

1.88

(1.21, 2.93)

0.005

-

Confined to bed (vs none) 28 (2%) 11.87 (7.00, 20.1) <0.001 2.47 (1.00, 6.08) 0.050 -

EQ5D Self-care:

Some problems (vs none)

286 (16%)

3.73

(2.78, 5.02)

<0.001

1.32

(0.82, 2.13)

0.26

-

Unable to wash/dress (vs

none)

77 (4%)

8.59

(5.91, 12.5)

<0.001

2.78

(1.42, 5.43)

0.003

-

Enhanced prophylaxis 906 (50%) 0.75 (0.57, 0.97) 0.03 0.72 (0.54, 0.96) 0.02 -

Sex (female vs male) 961 (53%) 1.23 (0.95, 1.60) 0.12 - 1.19 (0.89, 1.60) 0.23

BMI: <18.5 (vs >25) 731 (41%) 2.41 (1.30, 4.48) 0.005 - 0.90 (0.52, 1.55) 0.70

18.5-25 (vs >25) 909 (51%) 1.38 (0.74, 2.58) 0.32 - 0.70 (0.41, 1.19) 0.19

HIV VL (per log10 higher) 5.4 (5.0,5.8) 1.52 (1.22, 1.89) <0.001 - 1.37 (1.09, 1.74) 0.008

Neutrophil % (per 5% higher) 50% (40,62) 1.07 (1.02, 1.12) 0.003 - 0.99 (0.95, 1.04) 0.64

87 45 25 12 15

268

349 217

206 383

0

50

100

150

200

250

300

350

400

450

Group-1 Group-2 Group-3 Group-4 Group-5

Nu

mb

er

of

pa

rtic

ipa

nts

Died AliveMortality

25% 11% 10% 6% 4%

Table 1: Independent predictors of mortality in the first 48 weeks on ART

initiated with CD4 <100 cells/mm3

† chronic health conditions or prescribed medications more than 14 days prior to screening visit. Note: final multivariable model in REALITY participants also adjusted for centre.

Figure 2: Characterisation of late presenter

“phenotypes” by baseline factors

Figure 1: Mortality by late presenter “phenotype”

CONCLUSIONS

Clinical and laboratory characteristics

identified groups of patients (all with

CD4<100 cells/mm3) at highest risk of

mortality following ART initiation

A screening tool appropriate to the level

of facility could therefore help identify

which patients with low CD4 counts

should be prioritized, e.g. for same-day

ART initiation, more intensive follow-up,

and enhanced prophylaxis

• 225 (12%) of 1805 participants starting ART with

CD4<100 cells/mm3 died before 48w

– median 8 weeks (IQR 3-18) on ART

• Mortality was independently higher in those who

were older, with lower CD4s, lower albumin,

lower haemoglobin and weaker grip strength;

those in whom physicians reported WHO stage 3/4

weight loss; and in those patients reporting fever,

vomiting, some problems with mobility and

inability to wash or dress themselves at

baseline (Table 1)

– Receiving enhanced-prophylaxis independently reduced

mortality (p=0.02) as in the trial overall (Hakim et al NEJM

2017 PMID 28723333), but receiving raltegravir (p=0.60)

and receiving ready-to-use-supplementary food (p=0.37)

did not

– The area under the receiver operating curve was 80%

(95% CI 76-83%) supporting good model discrimination

• VL was not an independent predictor once

problems with mobility were included as a

prognostic factor, suggesting mobility may be a

broad physical marker for the detrimental impact of

ongoing viral replication

• There was a marginal association between WBC

and mortality in our final model, supporting

infection and/or inflammation playing a role in early

increased mortality risk

• We were able to identify a more detailed and

slightly different set of clinical predictors to a recent

analysis of a smaller trial with lower mortality (7%)

(Bisson et al, AIDS 2017, PMID 28742529)

– High HIV VL and WBC, but not neutrophil percent, were

independent predictors of mortality in REALITY participants

when fitting the prognostic model from this previous study

(Table 1)

• However, other clinical factors were more

predictive than these laboratory parameters in

REALITY participants

• We then considered whether risk factors clustered

in specific groups of patients, that is, whether late

presenters could be phenotyped

• We identified 5 groups of patients (all with CD4<100

cells/mm3) whose mortality ranged from 25%

(Group-1) to 4% (Group-5) (Figure 1)

• Group-1 (25% mortality; median CD4 28 cells/mm3) had high

burden of problems reported on the EQ-5D, particularly for mobility,

self-care and usual activities (Figure 2). Excluding rash, they also

had the highest burden of symptoms and illness, especially weight

loss, difficulty walking and poor appetite

• Group-2 (11% mortality; median CD4 43 cells/mm3) had high levels

of WBC, platelets and neutrophils, despite lower levels of reported

fever (12%) than Group-1 (28%).

– Only 122 (31%) had an infection reported at enrolment,

suggesting they may have had underlying infections/inflammation

– This group also had substantial weight loss and low fat mass,

and low BMI, albumin and haemoglobin, similar to Group-1.

However, median CD4 was higher in Group-2 vs Group-1 as was

grip strength (median 24.5 vs 20.0kg)

• Group-3 (10% mortality; median CD4 27 cells/mm3) were mainly

(non-pregnant non-breastfeeding) females (76%) and younger

(median 30 vs 36 years overall). They had low CD4 counts, WBC,

neutrophils, haemoglobin, BMI, MUAC and grip strength, but

despite this had low burden of symptoms/ illnesses and maintained

a reasonable fat mass and higher albumin.

• Group-4 (6% mortality; median CD4 48 cells/mm3) was also mostly

female (97%), but older than Group-3 and much higher fat mass

BMI and MUAC

• Group-5 (4% mortality; median CD4 42 cells/mm3) was

predominately male (86%) and older (median 39 years) with higher

fat mass, fat free mass and grip strength