comparative safety profiles of intravitreal bevacizumab, ranibizumab and pegaptanib: the analysis of...

PHARMACOEPIDEMIOLOGYAND PRESCRIPTION

Comparative safety profiles of intravitreal bevacizumab,ranibizumab and pegaptanib: the analysis of the WHO databaseof adverse drug reactions

C. Biagi & V. Conti & N. Montanaro & M. Melis & E. Buccellato &

M. Donati & A. Covezzoli & R. Amato & L. Pazzi & M. Venegoni &A. Vaccheri & D. Motola

Received: 22 July 2014 /Accepted: 11 September 2014 /Published online: 20 September 2014# Springer-Verlag Berlin Heidelberg 2014

AbstractPurpose The purpose of this study is to conduct a compara-tive analysis of the suspected adverse drug reactions (ADRs)associated with intravitreal bevacizumab, ranibizumab andpegaptanib in the WHO database in order to have a real-lifeinformation on these drugs, which now is only based on datacoming from clinical trials.Methods ADR reports for intravitreal use of bevacizumab,ranibizumab and pegaptanib from January 2002 toDecember 2012 were selected from the WHO-VigiBase.Reporting odds ratio (ROR) with confidence interval of95 % and p value was calculated. The analysis was performedfor drug-reaction pairs. The Medical Dictionary forRegulatory Activities (MedDRA) terminology for ADRswas used.Results The analysis was performed on 3180 reports corre-sponding to 7753 drug-reaction pairs. Significant RORs forendophthalmitis and uveitis (1.90, 95 % confidence interval(CI) 1.48–2.43, and 10.62, 6.62–17.05, respectively) wereretrieved for bevacizumab, and cerebrovascular accident andmyocardial infarction produced significant ROR (1.54, 1.14–2.10 and 1.73, 1.18–2.53, respectively) for ranibizumab.Pegaptanib was significantly associated with visual

impairment (1.98, 1.12–3.5, p=0.02), nausea (3.29, 1.57–6.86, p<0.001), vomiting (2.91, 1.2–7.07, p=0.01) and drughypersensitivity (8.75, 3.1–24.66, p<0.001).Conclusions Our data showed an elevated disproportionalityfor cardiovascular ADRs in patients treated with ranibizumaband for infective ocular reactions in those treated withbevacizumab. No relevant safety issues were identified forpegaptanib. These findings suggest bevacizumab as a suitablechoice for AMD therapy due to its effectiveness similar to thatof ranibizumab, its favourable safety profile and for its lowercost.

Keywords Bevacizumab . Ranibizumab . Pegaptanib .

Intravitreal use . Safety profile . Spontaneous reportingdatabase

Introduction

Age-related macular degeneration (AMD) is the major causeof loss of central vision and visual impairment among indi-viduals older than 50 years in industrialized countries [1].Neovascular AMD is the most aggressive form, and its prev-alence increases with age. Until 2005, treatments of AMDwere laser photocoagulation and photodynamic therapy. In thelast decade, significant advances in the treatment ofneovascular AMD and related diseases have been achievedthrough the marketing of agents against the vascular endothe-lial growth factor (VEGF) such as pegaptanib, ranibizumaband bevacizumab. Ranibizumab received marketing approvalfor AMD from the Food and Drug Administration (FDA) in2006 and from the European Medicines Agency (EMA) in2007, whereas bevacizumab was approved in the USA in2005 and in Europe in 2006 as a systemic anticancer therapy.However, before ranibizumab was licensed, the literature

C. Biagi :N. Montanaro :M. Melis : E. Buccellato :M. Donati :A. Vaccheri :D. Motola (*)Unit of Pharmacology, Department of Medical and SurgicalSciences, University of Bologna, via Irnerio 48, 40126Bologna, Italye-mail: [email protected]

V. Conti :M. VenegoniPharmacovigilance Regional Centre of Lombardy,Via Rosellini 17, 20124 Milan, Italy

A. Covezzoli :R. Amato : L. PazziHealth Care Systems Department, CINECA - InteruniversityConsortium, via Magnanelli 6/3, 40033 Casalecchio di Reno, Italy

Eur J Clin Pharmacol (2014) 70:1505–1512DOI 10.1007/s00228-014-1755-1

offered widespread evidence for the use of off-label intravit-real bevacizumab [2-6]. So notwithstanding ranibizumab reg-ulatory approval, off-label intravitreal bevacizumab in theUSA accounted for 58 % of all AMD injections in 2008, alsoin consideration of its lower cost [7].

Until a few years ago, no head-to-head studies evaluatingthe relative efficacy of the two major anti-VEGF drugs inneovascular AMDwere available. This situation changed withthe results of two largemulticentre clinical trials CATT [8] andIVAN [9] and with theMANTA trial [10] that confirmed thosefindings.

The relative ocular and systemic safety of bevacizumaband ranibizumab are still debated. The CATT trial showedcomparable safety profiles at 2 years. Although the rate ofhospitalizations was higher in the bevacizumab group, thatstudy was not powered to assess the comparative safety of thetwo drugs and probably the mentioned finding was not drug-related but due to patients’ comorbidities. Indeed, an analysisof the baseline patient characteristics showed that almost asmany as twice of patients on bevacizumab than onranibizumab were older, less healthy (e.g. had had a previousTIA) and a less favourable socioeconomic status. On thecontrary, 1-year data from the IVAN study [11] suggest anincreased risk of stroke with ranibizumab, whereas no differ-ences were found in the 2-year IVAN report, as well as inGEFAL and MANTA efficacy trials. An analysis of pooleddata from both CATT and IVAN studies confirmed no signif-icant differences between the two drugs in terms of mortality,heart attacks or strokes but a higher rate of other seriousadverse drug reactions (ADRs) in patients receivingbevacizumab. The authors themselves interpreted these resultsas unrelated to bevacizumab since events were more commonin patients treated less frequently, and they arose mainly fromhospitalizations for a wide variety of conditions not previouslyassociated with VEGF inhibition. Although none of theabovementioned studies was powered to detect differencesin harm between the treatments, their results have been usedto suggest doubts about the safety of bevacizumab.

More recently a systematic review compared the safety ofbevacizumab and ranibizumab in AMD, but the validity of itsresults was undermined by several biases in data collectionand in statistical analysis [12].

Differences in the pharmacokinetic properties have beenproposed to justify the supposed distinct systemic safety pro-file of the two drugs [13, 14]: the longer intravitreal (9.8 vs.7.1 days) [15] and systemic half-life (20 days vs. 2–6 hfollowing intravenous injections) of bevacizumab [16, 17]could result in a different systemic exposure and thus in agreater risk of ADR [18]. The clinical relevance of thesefindings is still a question.

Recently, the Italian Competition Authority fined Novartisand Roche a total of € 180 million for striking an alliance toprevent the use of bevacizumab (Roche) in favour of the more

expensive ranibizumab (Novartis) through a complex collu-sive media strategy. According to Italy’s antitrust, “The eco-nomic rationale of firms’ conduct stems from the relationshipbetween the Roche and Novartis groups: while Roche collectssignificant royalties from the sales of Lucentis, which hasbeen developed by its subsidiary Genentech, Novartis benefitsdirectly from Lucentis’ sales and holds a >30 % share inRoche. Genentech was not considered liable for the infringe-ment” [19].

In order to contribute with real-life data to the debate on thesafety profiles of these drugs, that now is only based oninformation from clinical trials, we aimed to conduct a com-parative analysis of the reports of suspected ADRs associatedwith the intravitreal injections of the three anti-VEGF com-pounds in the database of the World Health Organization(WHO).

Evenmore recently, the Italian National Health Council hasstated the therapeutic equivalence between bevacizumab andranibizumab in the AMD therapy [20].

Methods

Data were obtained from the global individual case reportsdatabase “VigiBase” of the WHO. VigiBase is maintained bythe Uppsala Monitoring Centre (UMC) [21] and containsnational data of ADRs reporting from 107 countries (May2012). Starting from 1968, the centre receives summary clin-ical reports about individual suspected ADRs from the nation-al centres of the countries participating in the collaborativeprogramme. National centres (in Italy, the Italian MedicinesAgency—AIFA) are WHO-approved pharmacovigilance cen-tres in countries participating in the WHO Programme forInternational Drug Monitoring; they collect spontaneousADRs that occurred in their country and submit them to theUMC on a regular basis. The participating countries mayaccess and analyze the data in order to investigate potentialADR signals and statistics. For our purpose, we selected allreports in which bevacizumab, ranibizumab or pegaptanib werelisted as suspected drugs. In order to detect and exclude asmany duplicates as possible in the database, we performed ananalysis using a record-linkage strategy by grouping the over-lapping records in eight key fields: country-text, gender, age-reaction, re-outcome, preferred-base-name, reported-term, on-set-date, and start-date. The records having seven out of eightoverlapping information and a single missing data in the rele-vant key fields were considered as duplicates. ADRs due to theoncological use of bevacizumab use were also excluded.

We performed the analysis in terms of drug-reaction pairs,comparing each of the three drugs versus the other two.

Reporting odds ratios (ROR), with 95 % confidence inter-vals (CIs) and p value was calculated as a measure ofdisproportionality [22].

1506 Eur J Clin Pharmacol (2014) 70:1505–1512

Data management and statistical analysis were carried outusing SAS statistical package (SAS Institute, Cary, NC),version 9.2: the results were considered significant at p<0.05.

Results

ADR reports for bevacizumab, ranibizumab and pegaptanibfrom January 2002 to December 2012 were selected from theWHO VigiBase. We retrieved 20,965 ADR reports (16,820related to bevacizumab, 3881 to ranibizumab and 264 topegaptanib), corresponding to 69,864 drug-reaction pairs.

Following removal of duplicates and of reports concerningthe oncological use of bevacizumab, we obtained 56,617pairs. The final analysis was therefore performed on 3180reports, corresponding to 7753 drug-reaction pairs, 56.8 %(1806 reports) involved females, 36.1% (1147)males while in7.1 % (227) the information on gender was missing. Twothousand sixty nine drug-reaction pairs (27 %) concernedbevacizumab, 5130 (66 %) ranibizumab and 554 (7 %)pegaptanib. The ADR was classified as serious (death,hospitalization, life-threatening, permanent disability) in85.2 % of the cases (734 out of 861 reports) for bevacizumabincluding 41 fatal cases (5.6 % of the serious cases), in 81.9 %of the cases (1756 out of 2145) for ranibizumab including 266

deaths (15.1 % of the serious cases) and in 89.1 % (155 out of174 reports) for pegaptanib, including eight fatal (5.2 %)cases.

In more than half of the reports (57 %), drugs were used forthe treatment of AMD; indication was missing in 14 % of thereports.

Table 1 lists the first 20 ADRs, and the correspondingpercentages, most frequently reported in patients with AMDand treated with one of the three anti-VEGF drugs. Amongthese, we found ocular ADRs, including reduced visual acuity,endophthalmitis, eye pain, vision impairment, uveitis andretinal haemorrhage, and systemic ADRs, such as myocardialinfarction, cerebrovascular accident and death. Reducedvisual acuity was theADRwith the largest number of reportsboth for bevacizumab and ranibizumab, with 121 reports(5.85 %) and 322 (6.28 %), respectively, whereas it was thesecond ADR for pegaptanib (19, 3.43 %). Endophthalmitisranked second for bevacizumab, with 111 reports (5.36 %),whereas the ADR was third for ranibizumab (145, 2.83 %)and first for pegaptanib (20, 3.61 %). Cerebrovascular acci-dent ranked second for ranibizumab (167, 3.26 %) whereaswas tenth for bevacizumab and fifth for pegaptanib, 43(2.08%) and 13 (2.35 %), respectively. Reports of blindnesswere42 (2.03%) forbevacizumab,whichputs suchADRstillamong the top 20, 49 for ranibizumab (0.96 %) and 5 for

Table 1 Number and percentages of the first 20 ADRs most reported for bevacizumab, ranibizumab and pegaptanib

ADRs

Bevacizumab N (%) Ranibizumab N (%) Pegaptanib N (%)

1 Visual acuity reduceda 121 (5.85) Visual acuity reduceda 322 (6.28) Endophthalmitisa 20 (3.61)

2 Endophthalmitisa 111 (5.36) Cerebrovascular accidenta 167 (3.26) Visual acuity reduceda 19 (3.43)

3 Uveitisa 82 (3.96) Endophthalmitisa 145 (2.83) Eye paina 14 (2.53)

4 Detachment of retinal pigment epa 78 (3.77) Myocardial infarctiona 117 (2.28) Visual impairmenta 14 (2.53)

5 Eye inflammationa 62 (3.00) Eye paina 112 (2.18) Cerebrovascular accident 13 (2.35)

6 Intraocular pressure increaseda 52 (2.51) Death 93 (1.81) Detachent of retinal pigment epa 11 (1.99)

7 Eye paina 49 (2.37) Drug ineffectivea 73 (1.42) Myocardial infarction 9 (1.62)

8 Retinal pigment epithelial teara 47 (2.27) Transient ischaemic attacka 72 (1.40) Nauseaa 9 (1.62)

9 Vitritisa 45 (2.17) Visual impairmenta 71 (1.38) Intraocular pressure increaseda 8 (1.44)

10 Cerebrovascular accidenta 43 (2.08) Retinal haemorrhagea 68 (1.33) Eye haemorrhagea 8 (1.44)

11 Blindnessa 42 (2.03) Blood pressure increaseda 58 (1.13) Ocular hyperaemiaa 8 (1.44)

12 Product quality issue 41 (1.98) Vision blurreda 55 (1.07) Blood pressure increaseda 7 (1.26)

13 Drug ineffectivea 31 (1.50) Retinal pigment epithelial teara 54 (1.05) Retinal detachmenta 7 (1.26)

14 Death 30 (1.45) Vitreous haemorrhagea 52 (1.01) Retinal haemorrhagea 6 (1.08)

15 Vision blurreda 30 (1.45) Eye haemorrhagea 50 (0.97) Vision blurreda 6 (1.08)

16 Retinal haemorrhagea 29 (1.40) Blindnessa 49 (0.96) Vomitinga 6 (1.08)

17 Toxic anterior segment syndrome 29 (1.40) Headachea 46 (0.90) Drug hypersensitivitya 6 (1.08)

18 Myocardial infarctiona 26 (1.26) Vitreous floatersa 43 (0.84) Pneumonia 5 (0.90)

19 Transient ischaemic attacka 23 (1.11) Pneumonia 43 (0.84) Product quality issue 5 (0.90)

20 Vitreous haemorrhagea 23 (1.11) Malaise 43 (0.84) Blindness unilateral 5 (0.90)

a Statistically significant

Eur J Clin Pharmacol (2014) 70:1505–1512 1507

pegaptanib (0.90 %). Other widely reported ADRs forbevacizumabwereuveitis (82,3.96%),detachmentof retinalpigment (78, 3.77%) and eye inflammation (62, 3.00%). Forranibizumab, the most frequent reports were myocardial in-farction (117, 2.28%), eye pain (112, 2.18%) and death (93,1.81 %) and for pegaptanib eye pain (14, 2.53 %), visualimpairment (14, 2.53) and detachment of retinal pigment(11, 1.99 %). Overall, 75 % of bevacizumab ADRs involvedthe eye and 25%were systemic ADRs, whereas this propor-tion was 60 vs 40 % for ranibizumab and 65 vs 35 % forpegaptanib.

Disproportionality analysis

Tables 2, 3 and 4 show the reporting odds ratios, 95 %confidence intervals and p values of the adverse reactionsrelated to the intravitreal use of bevacizumab, ranibizumaband pegaptanib. For bevacizumab, a signif icantdisproportionality was observed for endophthalmitis (ROR1.90, 95 % CI 1.48–2.43, p<0.001), uveitis (10.62, 6.62–17.05, p<0.001), eye inflammation (7.95, 4.88–12.97,p<0.001), blindness (2.24, 1.49–3.38, p<0.001) and others.No significant disproportionality was detected for visualacuity reduced (0.97, 0.79–1.21, p=0.80), death (0.87,0.57–1.31, p=0.49) and transient ischaemic attack (0.86,0.54–1.38, p=0.54), whereas an inverse significantdisproportionality was observed for other systemic ADRs,such as cerebrovascular accident (0.65, 0.46–0.91, p=0.01)and myocardial infarction (0.56, 0.37–0.86, p=0.007): thelatter findings do not indicate a protective effect ofbevacizumab but a significantly lower reporting of thoseADRs than expected. Most ocular ADRs for ranibizumabdid not show a significant disproportionality, including visualacuity reduced (1.19, 0.97–1.46, p=0.10), visual impairment(1.01, 0.67–1.51, p=0.97), and others. On the contrary, asignificant disproportionality was found for serious systemicADR associated with ranibizumab: cerebrovascular accident(1.54, 1.14–2.1, p=0.005), myocardial infarction (1.73, 1.18–2.53, p=0.004), cardiac failure congestive (3.17, 1.34–7.52,p=0.006), pneumonia (2.21, 1.11–4.4, p=0.02) and others.Pegaptanib showed disproportionality only for visual impair-ment (1.98, 1.12–3.5, p=0.02), nausea (3.29, 1.57–6.86,

Table 2 Main ADRs for bevacizumab with reporting odds ratio (95 %confidence intervals) and p value against ranibizumab and pegaptanib

ADR ROR (95 % CI) p value

Visual acuity reduced 0.97 (0.79–1.21) 0.80

Endophthalmitis 1.90 (1.48–2.43) <0.001

Uveitis 10.62 (6.62–17.05) <0.001

Detachment of retinal pigment ep 4.60 (3.2–6.61) <0.001

Eye inflammation 7.95 (4.88–12.97) <0.001

Intraocular pressure increased 3.03 (2.04–4.5) <0.001

Eye pain 1.07 (0.77–1.49) 0.69

Retinal pigment epithelial tear 2.29 (1.55–3.39) <0.001

Vitritis 5.47 (3.3–9.07) <0.001

Cerebrovascular accident 0.65 (0.46–0.91) 0.01

Blindness 2.24 (1.49–3.38) <0.001

Product quality issue 11.47 (5.74–22.94) <0.001

Drug ineffective 1.12 (0.74–1.71) 0.59

Death 0.87 (0.57–1.31) 0.49

Vision blurred 1.36 (0.87–2.11) 0.17

Retinal haemorrhage 1.08 (0.7–1.66) 0.73

Toxic anterior segment syndrome 80.79 (11–593.44) <0.001

Myocardial infarction 0.56 (0.37–0.86) 0.007

Transient ischaemic attack 0.86 (0.54–1.38) 0.54

Vitreous haemorrhage 1.19 (0.73–1.95) 0.48

Vitreous floaters 1.35 (0.82–2.23) 0.24

Visual impairment 0.71 (0.44–1.13) 0.15

Retinal detachment 1.2 (0.71–2.03) 0.50

Iritis 5.26 (2.44–11.33) <0.001

Blindness unilateral 2.07 (1.12–3.82) 0.02

Transmission of an infectious ag. 49.88 (6.65–373.84) <0.001

Iridocyclitis 8.85 (3.24–24.19) <0.001

Ocular hyperaemia 0.84 (0.47–1.5) 0.55

Blood pressure increased 0.59 (0.33–1.05) 0.07

Pain 1.75 (0.9–3.43) 0.10

Dizziness 0.79 (0.43–1.47) 0.46

Photophobia 3.59 (1.57–8.19) <0.001

Eye haemorrhage 0.57 (0.3–1.06) 0.07

Macular hole 4.14 (1.69–10.14) <0.001

Hypertension 1.08 (0.54–2.17) 0.83

Confusional state 1.84 (0.82–4.09) 0.13

Corneal oedema 3.45 (1.36–8.74) 0.006

Eye infection 1.46 (0.65–3.27) 0.36

Eye swelling 1.46 (0.65–3.27) 0.36

Vitreous opacities 6.2 (1.91–20.17) <0.001

Acute myocardial infarction 0.78 (0.36–1.72) 0.54

Thrombosis 1.38 (0.59–3.22) 0.46

Inflammation 1.83 (0.75–4.49) 0.18

Hypopyon 2 (0.8–4.98) 0.13

Retinal artery occlusion 2.45 (0.94–6.35) 0.06

Macular oedema 1.75 (0.68–4.52) 0.24

Anterior chamber inflammation 1.93 (0.73–5.07) 0.18

Pseudoendophthalmitis 2.41 (0.87–6.65) 0.08

Table 2 (continued)


Haemorrhagic stroke 2.75 (0.96–7.86) 0.05

Cardiac failure congestive 0.44 (0.19–1.05) 0.06

Macular degeneration 0.53 (0.22–1.27) 0.15

Cerebral infarction 0.59 (0.24–1.42) 0.23

Chest pain 0.69 (0.28–1.68) 0.41

Retinal vascular disorder 5.51 (1.38–22.04) 0.007


Table 3 Main ADRs for ranibizumab with ROR (95 % confidentialintervals) and p value against ranibizumab and pegaptanib




Endophthalmitis 0.55 (0.43–0.7) <0.001


Eye pain 0.91 (0.66–1.24) 0.54

Death 1.49 (1–2.24) 0.05

Drug ineffective 1.1 (0.73–1.66) 0.65

Transient ischaemic attack 1.54 (0.97–2.45) 0.07




Vision blurred 0.78 (0.51–1.19) 0.24

Retinal pigment epithelial tear 0.55 (0.37–0.81) 0.002

Vitreous haemorrhage 1.11 (0.68–1.8) 0.68

Eye haemorrhage 1.28 (0.76–2.16) 0.35

Blindness 0.55 (0.37–0.83) 0.004

Vitreous floaters 0.81 (0.5–1.32) 0.40

Pneumonia 2.21 (1.11–4.4) 0.02

Malaise 4.43 (1.75–11.19) <0.001

Dizziness 1.35 (0.75–2.4) 0.31

Fall 10.82 (2.62–44.72) <0.001


Intraocular pressure increased 0.34 (0.22–0.5) <0.001


Cataract 5.02 (1.79–14.05) <0.001

Detachment of retinal pigment ep 0.21 (0.14–0.3) <0.001

Cardiac failure congestive 3.17 (1.34–7.52) 0.006

Atrial fibrillation 19.05 (2.61–138.91) <0.001

Cerebral infarction 2.39 (0.99–5.79) 0.05

Macular degeneration 1.38 (0.67–2.86) 0.38

Acute myocardial infarction 1.54 (0.72–3.27) 0.26

Hypertension 1.02 (0.52–1.99) 0.95

Vitritis 0.24 (0.14–0.4) <0.001

Retinal oedema 2.26 (0.85–5.96) 0.09

Uveitis 0.13 (0.08–0.2) <0.001

Eye inflammation 0.16 (0.09–0.26) <0.001

Chest pain 1.02 (0.48–2.19) 0.95

Blindness unilateral 0.42 (0.23–0.77) 0.004

Choroidal neovascularisation 1.22 (0.53–2.78) 0.64

Eye irritation 1.15 (0.5–2.65) 0.74

Subretinal fibrosis 2.9 (0.85–9.92) 0.07

Diabetes mellitus 8.72 (1.16–65.54) 0.01

Eye infection 0.82 (0.37–1.8) 0.62

Thrombosis 0.85 (0.37–1.95) 0.70

Retinal disorder 1.92 (0.64–5.79) 0.24

Vitreous detachment 2.56 (0.74–8.85) 0.12

Staphylococcal infection 3.84 (0.88–16.82) 0.05

Cardiac disorder 7.69 (1.02–58.24) 0.02

Renal failure 7.69 (1.02–58.24) 0.02

Table 3 (continued)


Eye disorder 0.55 (0.26–1.17) 0.11

Aphasia 1.79 (0.59–5.45) 0.30

Cardiac failure 3.59 (0.81–15.79) 0.07

Bronchitis 7.18 (0.94–54.59) 0.03

Urinary tract infection 7.18 (0.94–54.59) 0.03

Eye swelling 0.51 (0.24–1.1) 0.08

Deep vein thrombosis 1.33 (0.47–3.74) 0.59

Oedema peripheral 1.33 (0.47–3.74) 0.59

Pulmonary embolism 1.33 (0.47–3.74) 0.59

Cerebral haemorrhage 2.22 (0.63–7.79) 0.20

Ischaemic stroke 3.33 (0.75–14.76) 0.09

Scotoma 2.05 (0.58–7.26) 0.26

Age-related macular degeneration 6.15 (0.8–47.31) 0.05

Cataract operation 6.15 (0.8–47.31) 0.05

Inflammation 0.62 (0.26–1.51) 0.29

Hypopyon 0.7 (0.28–1.75) 0.44

Myocardial ischaemia 5.63 (0.73–43.66) 0.06

Iritis 0.27 (0.12–0.58) <0.001

Photophobia 0.39 (0.17–0.9) 0.02

Anterior chamber inflammation 0.73 (0.28–1.92) 0.52

Vitreous disorder 1.71 (0.47–6.2) 0.41

Angina pectoris 5.12 (0.66–40.03) 0.08

Macular oedema 0.51 (0.2–1.29) 0.15

Retinal tear 0.92 (0.31–2.75) 0.88

Hypoaesthesia 1.15 (0.35–3.74) 0.81

Sudden death 1.15 (0.35–3.74) 0.81

Anaemia 1.53 (0.42–5.67) 0.52

Metamorphopsia 2.3 (0.5–10.67) 0.27

Respiratory failure 2.3 (0.5–10.67) 0.27

Corneal oedema 0.41 (0.16–1.04) 0.05

Retinal artery occlusion 0.45 (0.17–1.18) 0.10

Pseudoendophthalmitis 0.58 (0.21–1.61) 0.29

Ocular hypertension 1.02 (0.31–3.4) 0.97

Choroidal haemorrhage 2.05 (0.43–9.65) 0.35

Conjunctival haemorrhage 2.05 (0.43–9.65) 0.35

Macular hole 0.27 (0.11–0.69) 0.03

Haemorrhagic stroke 0.51 (0.18–1.46) 0.20

Glaucoma 0.89 (0.26–3.06) 0.86

Macular scar 1.19 (0.31–4.62) 0.80

Gastrointestinal haemorrhage 1.79 (0.37–8.63) 0.46

Visual field defect 1.79 (0.37–8.63) 0.46

Eye infection bacterial 0.61 (0.19–2.01) 0.41

Renal failure acute 0.77 (0.22–2.72) 0.68

Anterior chamber cell 3.07 (0.37–25.52) 0.27

Arrhythmia 3.07 (0.37–25.52) 0.27

Arteriosclerosis coronary artery 3.07 (0.37–25.52) 0.27

Cardiac arrest 3.07 (0.37–25.52) 0.27

Cardiovascular disorder 3.07 (0.37–25.52) 0.27


p<0.001), vomiting (2.91, 1.2–7.07, p=0.01) and drug hyper-sensitivity (8.75, 3.1–24.66, p<0.001). As far as “death” isconcerned, it was statistically significant only for ranibizumab(1.49, 1–2.24, p=0.05), and bevacizumab did not yield asignificant ROR (0.87, 0.57–1.31, p=0.49) whereas no caseswere retrieved for pegaptanib.

Discussion

The present study provides important information about therelative safety of AMD intravitreal treatments with biologicaldrugs and is the first analysis comparing bevacizumab,ranibizumab and pegaptanib by means of data obtained froma worldwide spontaneous reporting database. The observedevent rates highlight disproportionality for cardiovascularADRs in patients treated with ranibizumab and for ocularreactions in those treated with bevacizumab. No relevantsafety issues were identified for pegaptanib.

The largest disproportionality observed for ocular infectionADRs reported for bevacizumab, in particular endophthalmitis,uveitis, eye inflammation and blindness, could be attributableto the compounding procedures used to prepare multiple syrin-ges starting from a single vial containing the drug, which arenot always performed under standard aseptic techniques, so thatcontamination can occur during preparation or administration.

Concerning the higher disproportionality detected for car-diovascular ADRs from ranibizumab, we cannot rule out thathigher-risk patients were more likely to receive this druginstead of bevacizumab owing to the regulatory warnings

about the increased risk of thromboembolic events associatedwith intravenous bevacizumab.

Besides clinical trials, a number of observational studies aswell as systematic reviews and meta-analyses have been pub-lished on this topic in the literature.

A retrospective cohort study published on 2010 and con-ducted on 146,942 Medicare beneficiaries did not show anydifference in the risk ofmortality, heart attack, stroke or bleedingbetween patients with AMD treated with bevacizumab orranibizumab and those who received photodynamic therapy orpegaptanib [23].

Later, a systematic review of 278 articles by Van Der Reiset al. [24] concluded that the rates of serious ocular and sys-temic ADRs after anti-VEGF injections used for any ophthal-mic indication were low, without differences in inci-dences between the assessed drugs. As a matter of fact,systemic adverse events had a lower incidence rates for intra-vitreal bevacizumab in comparison with ranibizumab andpegaptanib.

In contrast, the results of a systematic review [25] investi-gating the relative safety of bevacizumab and ranibizumabshowed a significantly higher rate of ocular and systemicadverse effects with bevacizumab compared to ranibizumab.It should be noted, however, that most of patients came fromthe CATT trial, which was not statistically powerful enough toidentify safety differences between the two compounds.

In their nested case-control study [26], Campbell and col-leagues showed that intravitreal injections of VEGF inhibitorswere not associated with significant risks of ischaemic stroke,acute myocardial infarction, congestive heart failure or venousthromboembolism.

Pharmacovigilance studies based on spontaneous reportinghave several limitations and our findings should be interpretedin this context. Under-reporting and selective reporting areconsidered the main limitations of a spontaneous reporting-based pharmacovigilance system [27]. Moreover, estimationof the risk of a given ADR with a drug requires adequatedenominator information on drug utilization: this is common-ly obtained from sales data, which may not accurately reflectusage and prescribing levels. Another limit is the multiplecomparison issue (i.e. the propensity to generate false positiveresults as the number of comparison increases), consideredendemic in the analysis of spontaneous report databases [28].Also, the voluntary nature of the cases creates inherent limi-tations in data interpretation, and population differencesacross the participating countries have not been taken intoaccount in the present study. In addition, clinical informationcontained in a spontaneous reporting database such as WHOVigiBase is necessarily limited and, on the other hand, casescan be redundantly reported, although our method allowed usto eliminate as much as possible potential duplicates. Finally,causality assessment is frequently not reported in the retrievedrecords.

Table 4 ADRs for pegaptanib with reporting odds ratio (95 % confi-dence intervals) and p value against ranibizumab and pegaptanib


Endophthalmitis 1.02 (0.64–1.61) 0.95


Eye pain 1.13 (0.65–1.97) 0.66



Detachment of retinal pigment ep 1.25 (0.67–2.33) 0.49


Nausea 3.29 (1.57–6.86) <0.001

Intraocular pressure increased 1.13 (0.55–2.34) 0.74

Eye haemorrhage 1.69 (0.8–3.54) 0.16





Vision blurred 0.92 (0.4–2.11) 0.84

Vomiting 2.91 (1.2–7.07) 0.01

Drug hypersensitivity 8.75 (3.1–24.66) <0.001


Advantages are that VigiBase covers patients from mostcountries worldwide and the reports contained therein reflectboth real-life events and real-life prescribing, therefore includ-ing drug use patterns that cannot be studied in clinical trials forethical reasons, such as excessive dosage or inappropriate co-medication. Most importantly, spontaneous reports can beuseful in determining adverse drug reactions when medicinesare used off-label, as in our research.

In addition to the intravitreal anti-VEGF agents examinedin the present analysis, new products have entered the market,such as aflibercept, and further anti-VEGF drug developmentis ongoing. In the future, it will be interesting to compare all ofthese drugs currently available to get a more accurate pictureof their relative safety.

Conclusions

In contrast to the findings of previous studies, our data highlightan increase in disproportionality for cardiovascular ADRs inpatients treated with ranibizumab and for ocular infective reac-tions in those treated with bevacizumab (the latter probably dueto the repackaging of bevacizumab without proper aseptictechnique) and support the use of bevacizumab in the treatmentof AMD. These findings add to the already available evidenceand suggest bevacizumab as a suitable choice for AMD therapydue to its effectiveness similar to that of ranibizumab, itsfavourable safety profile and, last but not the least, for itsextraordinarily lower cost. Further head-to-head observationalstudies are warranted to verify this issue.

Acknowledgments The information does not represent the opinion ofthe WHO.We are very grateful to Mrs Anita Grabham for providing datafrom the Uppsala Monitoring Centre—WHO Collaborating Centre forInternational Drug Monitoring, Uppsala (Sweden).

Conflict of interest Chiara Biagi, Valentino Conti, Nicola Montanaro,Mauro Melis, Monia Donati, Elena Buccellato, Domenico Motola,Mauro Venegoni and Alberto Vaccheri declare no conflict of interest forthe submitted work. Roberta Amato, Linda Pazzi and Anna Covezzolireport grants from Novartis, grants from Roche and grants from Pfizer,outside the submitted work.

Funding source None.

Authors’ contribution Chiara Biagi, Domenico Motola, AlbertoVaccheri, Nicola Montanaro, Mauro Melis and Mauro Venegoni partici-pated in the study conception and planning. Chiara Biagi, DomenicoMotola, Valentino Conti, Monia Donati, Elena Buccellato, Linda Pazzi,Roberta Amato, Anna Covezzoli and Mauro Venegoni took part in theacquisition, analysis or interpretation of data. Chiara Biagi, DomenicoMotola, Nicola Montanaro, MauroMelis, Mauro Venegoni and ValentinoConti drafted the manuscript. Monia Donati, Elena Buccellato, LindaPazzi, Roberta Amato, Anna Covezzoli and Alberto Vaccheri criticallyrevised the manuscript for important intellectual content. All the authorsapproved the final version of the manuscript. The manuscript does notcontain clinical studies or patient data.

References

1. Kodjikian L, Souied EH, Mimoun G, Mauget-Faysse M, Behar-Cohen F, Decullier E et al (2013) Ranibizumab versusBevacizumab for Neovascular Age-related Macular Degeneration:results From the GEFAL Noninferiority Randomized Trial.Ophthalmology 120(11):2300–2309

2. Rosenfeld PJ, Moshfeghi AA, Puliafito CA (2004) Optical coherencetomography findings after an intravitreal injection of bevacizumab(Avastin) for neovascular age-related macular degeneration.Ophthalmic Surg, Lasers Imaging: Off J Int Soc Imaging Eye36(4):331–335

3. Bashshur ZF, Schakal A, Hamam RN, El Haibi CP, Jaafar RF,Baha’N N (2007) Intravitreal bevacizumab vs verteporfin photody-namic therapy for neovascular age-related macular degeneration.Arch Ophthalmol 125(10):1357–1361

4. Chen CY, Wong TY, Heriot WJ (2007) Intravitreal bevacizumab(Avastin) for neovascular age-related macular degeneration: a short-term study. Am J Ophthalmol 143(3):510–512

5. Lazic R, Gabric N (2007) Verteporfin therapy and intravitrealbevacizumab combined and alone in choroidal neovascularizationdue to age-related macular degeneration. Ophthalmology 114(6):1179–1185

6. Wu L, Mart_nez-Castellanos MA, Quiroz-Mercado H, Arevalo JF,Berrocal MH, Farah ME et al (2008) Twelve-month safety of intra-vitreal injections of bevacizumab (Avastin): results of the Pan-American Collaborative Retina Study Group (PACORES). GraefesArch Clin Exp Ophthalmol 246(1):81–87

7. Brechner RJ, Rosenfeld PJ, Babish JD, Caplan S (2011)Pharmacotherapy for neovascular age-related macular degeneration:an analysis of the 100 2008 medicare fee-for-service part B claimsfile. Am J Ophthalmol 151(5):887–895

8. Martin DF, Maguire MG, Ying GS, Grunwald JE, Fine SL, Jaffe GJ(2011) Ranibizumab and bevacizumab for neovascular age-relatedmacular degeneration. N Engl J Med 364(20):1897–1908

9. Chakravarthy U, Harding SP, Rogers CA, Downes SM, Lotery AJ,Culliford LA et al (2013) Alternative treatments to inhibit VEGF inage-related choroidal neovascularisation: 2-year findings of theIVAN randomised controlled trial. Lancet 382(9900):1258–1267

10. Ehlers JP (2013) The MANTA 1-year results: the anti-VEGF debatecontinues. Br J Ophthalmol 97(3):248–250

11. Chakravarthy U, Harding SP, Rogers CA, Downes SM, LoteryAJ, Wordsworth S et al (2012) Ranibizumab versusbevacizumab to treat neovascular age-related macular degeneration:one-year findings from the IVAN randomized trial. Ophthalmology119(7):1399–1411

12. Zhang XY, Guo XF, Zhang SD, He JN, Sun CY, Zou Y et al (2014)Comparison of bevacizumab and ranibizumab in age-related maculardegeneration: a systematic review and meta-analysis. Int JOphthalmol 7(2):355–364

13. Ghazi NG, Kirk TQ, Knape RM, Tiedeman JS, Conway BP (2010) Ismonthly retreatment with intravitreal bevacizumab (Avastin) neces-sary in neovascular age-related macular degeneration? ClinOphthalmol 4:307–314

14. Ventrice P, Leporini C, Aloe JF, Greco E, Leuzzi G, Marrazzo G et al(2013) Anti-vascular endothelial growth factor drugs safety andefficacy in ophthalmic diseases. J Pharmacol Pharmacother 4(1):S38–S42

15. Krohne TU, Eter N, Holz FG, Meyer CH (2008) Intraocular phar-macokinetics of bevacizumab after a single intravitreal injection inhumans. Am J Ophthalmol 146(4):508–512

16. Lucentis (2007) EPAR - Scientific Discussion, (14/03/2007).Available on: http://www.ema.europa.eu/ema/index.jsp?curl=pages/medicines/human/medicines/000715/human_med_000890.jsp&mid=WC0b01ac058001d124. Accessed 10 September 2014


http://www.ema.europa.eu/ema/index.jsp?curl=pages/medicines/human/medicines/000715/human_med_000890.jsp&mid=WC0b01ac058001d124



17. Semeraro F, Morescalchi F, Duse S, Parmeggiani F, Gambicorti E,Costagliola C (2013) Aflibercept in wet AMD: specific role andoptimal use. Drug design, dev and ther 7:711–722

18. Avery RL, Alessandro AC, Nathan CS, Dilsher SD, Pieramici DJ,See R, Couvillion S et al (2014) Systemic pharmacokinetics follow-ing intravitreal injections of ranibizumab, bevacizumab or afliberceptin patients with neovascular AMD. Br J Ophthalmol. doi:10.1136/bjophthalmol-2014-305252

19. Authority IC (2014) Pharmaceuticals and Antitrust: the ItalianCompetition Authority fines Roche and Novartis over eur 180 mil-lion for cartelizing the sales of two major ophthalmic drugs, Avastinand Lucentis

20. Health IM (2014) Ministry of Heatlh - Consiglio Superiore di SanitàSessione XLIX

21. WHO (2012) WHO collaborating centre for international drug mon-itoring. Uppsala Monitoring Centre, Uppsala, Sweden UMC

22. Van Puijenbroek EP, Diemont WL, van Grootheest K (2003)Application of quantitative signal detection in the Dutch spontaneousreporting system for adverse drug reactions. Drug Saf 26(5):293–301

23. Curtis LH,Hammill BG, SchulmanKA, Cousins SW (2010) Risks ofmortality, myocardial infarction, bleeding, and stroke associated with

therapies for age-related macular degeneration. Arch Ophthalmol128(10):1273–1279

24. der Reis V, Margriet I, La Heij EC, De Jong-Hesse Y, Ringens PJ,Hendrikse F, Schouten JS (2011) A systematic review of the adverseevents of intravitreal anti-vascular endothelial growth factor injec-tions. Retina 31(8):1449–1469

25. Schmucker C, Ehlken C, Agostini HT, Antes G, Ruecker G,Lelgemann M et al (2012) A safety review and meta-analyses ofbevacizumab and ranibizumab: off-label versus goldstandard. PLoSONE 7(8):e42701

26. Campbell RJ, Gill SS, Bronskill SE, Paterson JM,WhiteheadM, BellCM (2012) Adverse events with intravitreal injection of vascularendothelial growth factor inhibitors: nested case-control study. BMJ345:e4203

27. Biagi C, Montanaro N, Buccellato E, Roberto G, Vaccheri A, MotolaD (2013) Underreporting in pharmacovigilance: an intervention forItalian GPs (Emilia–Romagna region). Eur J Clin Pharmacol 69(2):237–244

28. Huang L, Guo T, Zalkikar JN, Tiwari RC (2014) A Review ofStatistical Methods for Safety Surveillance. Ther Innov Regul Sci48(1):98–108


http://dx.doi.org/10.1136/bjophthalmol-2014-305252

http://dx.doi.org/10.1136/bjophthalmol-2014-305252

comparative safety profiles of intravitreal bevacizumab, ranibizumab and pegaptanib: the analysis of...

Documents