Functional versus functional and anatomical criteria-guided ranibizumab treatment in patients with neovascular age-related macular degeneration – results from the randomized, phase IIIb OCTAVE study

OCTAVE was an open-label, phase IIIb, randomized, double-masked (with regard to the retreatment criteria) study that was originally planned to have a 24-month duration (Fig. 1), but was terminated earlier than planned (October 2014). The decision to discontinue the study was made by the sponsor after OCT-guided monitoring of disease activity has been included in the posology for ranibizumab and due to the updated label approved for disease activity (including VA and anatomical parameters). The OCTAVE study was initiated in June 2013 and completed in July 2015. The study was conducted at 92 sites in 24 countries.

OCTAVE was conducted according to the principles of the Declaration of Helsinki and the study protocol was reviewed by the Independent Ethics Committee or Institutional Review Board for each study center (see Additional file 1: Table S1, which lists all the IECs/IRBs). Patients provided written informed consent. The manuscript reporting adheres to the CONSORT guidelines for the reporting of randomized trials.

Eligible patients were randomized (1:1) via Interactive Response Technology to either of the two treatment groups, using a validated system that automated the random assignment of patient numbers to randomization numbers: Group I (VA-guided): three initial monthly ranibizumab 0.5 mg injections with retreatment thereafter at the investigator’s discretion based on best-corrected visual acuity (BCVA) loss due to nAMD; Group II (VA and/or OCT-guided): three initial monthly ranibizumab 0.5 mg injections with retreatment thereafter at the investigator’s discretion based on BCVA loss due to nAMD and/or signs of nAMD disease activity on OCT. Both groups had the same monthly assessments, including OCT. The VA assessor was masked to the retreatment strategy assignment and performed only the BCVA assessment. The evaluating and/or treating investigator (who could be the same person) was not masked to the retreatment strategy and performed all other study related activities. The decision to retreat was solely at the investigator’s discretion as per the retreatment criteria. During the study, no rescue medication was permitted for the treatment of nAMD.

Patients included were ≥ 50 years of age with visual impairment due to nAMD; with active, newly diagnosed, angiographically documented choroidal neovascularization (CNV) secondary to AMD in a previously untreated eye; and with CNV or its sequelae (i.e., pigment epithelium detachment, subretinal or subretinal pigment epithelial hemorrhage, blocked fluorescence, macular edema, or subretinal, subretinal pigment epithelial or intraretinal fluid [IRF]) involving the center of the fovea, and a BCVA score at both screening and baseline between 78 and 23 letters inclusive (approximate Snellen equivalent of 20/32 and 20/320). If both eyes were eligible for the study, a single study eye was selected at screening by the investigator based on clinical judgement and the decision was confirmed at the baseline visit.

Key exclusion criteria applied at screening or baseline were active periocular or ocular infection or inflammation; uncontrolled glaucoma (intraocular pressure [IOP] ≥ 30 mmHg on medication or according to the investigator’s judgment); neovascularization of the iris or neovascular glaucoma; cataract (if causing significant visual impairment), aphakia, vitreous hemorrhage, rhegmatogenous retinal detachment, proliferative retinopathy or CNV of any other cause than nAMD; irreversible structural damage within 0.5 disc diameter of the center of the macula (e.g., vitreomacular traction, epiretinal membrane, scar, laser burn, macular hole) that in the investigator’s opinion could substantially impact visual function improvement with treatment; atrophy or fibrosis involving the center of the fovea; and total area of fibrosis comprising more than 50% of the lesion area (see Additional file 2 for detailed exclusion criteria).

The following efficacy variables were assessed: mean change from baseline over 12 months in BCVA and central subfield thickness (CSFT); mean change from baseline at Month 12 in BCVA and CSFT categorized by presence at baseline of extrafoveal CNV lesions (defined by the Vienna Reading Center [VRC] as lesions where the CNV border was more than 200 μm away from the foveal centerpoint), subretinal fluid (SRF) or cysts, and by thickness of cysts in the central millimeter subfield; proportion of patients with a BCVA improvement of ≥5, ≥10, and ≥ 15 letters from baseline at Month 12; and proportion of patients with BCVA ≥73 letters (20/40 Snellen equivalent) at Month 12.

BCVA was assessed in the study eye at all visits in a sitting position using Early Treatment Diabetic Retinopathy Study (ETDRS)-like VA testing charts at an initial testing distance of 4 m or at 1 m (if it was not possible to assess at 4 m). OCT images were taken using different spectral-domain OCT devices certified for multicenter study use by the VRC. Each patient was assessed using the same machine throughout the study. The investigator used the information collected to assess the status of disease activity. The images taken monthly from screening to end-of-study visit were sent to the VRC for further evaluation. The captured images contained quantitative parameters (CSFT and the presence or absence of qualitative parameters (intraretinal cysts (further referred to as cysts; defined as round, minimally reflective spaces within the neurosensory retina [15]) and SRF (identified as a non-reflective space between the posterior boundary of the neurosensory retina and the retinal pigment epithelium/choriocapillaris reflection [15]). The CSFT represented the average retinal thickness between the Bruch’s membrane and the inner limiting membrane of the circular area within 1 mm diameter around the foveal center. Fluorescein angiography was performed in conjunction with color fundus photography. The investigator could use these images at his/her discretion to confirm their decision to retreat. The investigator also sent these images to the VRC. Efficacy data are summarized descriptively, except for the explorative BCVA analysis investigating the impact of the presence of baseline cysts. A linear regression model for change in BCVA from baseline to Month 12 was fitted, including treatment, presence of baseline cysts as factors, baseline BCVA as covariate and presence of baseline cysts/treatment as interaction term.

Treatment exposure was evaluated in both groups over the entire study period. Safety assessments included the type, frequency, relationship, and severity of ocular and non-ocular adverse events (AEs). Safety assessments were performed in the safety set which consisted of all patients who received at least one application of study treatment and had at least one post-baseline safety assessment. Safety data for all safety-analyzable patients are summarized descriptively.

Mean (SD) BCVA increased from baseline to Month 12 (Fig. 2). At Month 12, mean (SD) gain was numerically lower in Group I compared with Group II (6.7 [13.48] vs 8.3 [13.53] letters; Fig. 2). BCVA gain of ≥10 or ≥ 15 letters at Month 12 was observed in both groups, but there was a trend towards more patients in Group II versus Group I achieving these gains (Fig. 3). Similarly, a numerically higher proportion of patients in Group II achieved BCVA ≥73 letters at Month 12 (54% vs 49%) than Group I. The mean (SD) CSFT decreased from baseline over time (Fig. 4) and at Month 12 was − 161.3 (163.48) μm in Group I and − 175.3 (170.45) μm in Group II.

In both groups, there were differences in baseline BCVA and CSFT between patients with extrafoveal CNV lesions versus those with foveal (centerpoint is occupied by CNV) and juxtafoveal (border of CNV is 1–200 μm away from centerpoint) CNV lesions. There was a trend towards higher mean (SD) BCVA gains in patients with foveal and juxtafoveal lesions than those with extrafoveal CNV lesions in both groups; the gains were numerically higher in Group II than in Group I (see Additional file 3: Figure S1,). Mean (SD) decrease in CSFT from baseline at Month 12 was observed irrespective of the presence or absence of extrafoveal CNV lesions at baseline; the reduction was numerically higher in Group II in those with extrafoveal CNV lesions (see Additional file 3: Figure S1). In both groups, the mean change in BCVA or CSFT from baseline at Month 12 did not differ based on the presence or absence of SRF at baseline; the change was numerically higher in Group II irrespective of SRF (see Additional file 4: Figure S2).

Patients with cysts had a lower BCVA at baseline that stayed low over time compared with BCVA in patients without cysts at baseline, exploratory p < 0.05 for the difference in BCVA gain between presence/absence of cysts in both groups (linear regression, see, Additional file 5: Table S2). The presence of cysts at baseline was associated with slightly lower BCVA gain from baseline at Month 12 in Group I versus Group II, whereas in those without baseline cysts, the BCVA gains from baseline at Month 12 were similar between both groups (see, Additional file 4: Figure S2). The difference in mean BCVA change from baseline to Month 12 by presence or absence of cysts at baseline was higher in Group I compared with Group II (7.64 vs 4.80 letters), as estimated by linear regression. The mean CSFT decrease from baseline to Month 12 was higher in both groups in patients with cysts at baseline than those without. The absolute reduction in CSFT was comparable between groups in those with baseline cysts and was numerically higher in Group II in those without baseline cysts (see, Additional file 4: Figure S2). The BCVA gain from baseline at Month 12 and reduction in mean CSFT from baseline at Month 12 were higher in both groups in those with cysts larger than 400 μm at baseline (see, Additional file 4: Figure S2).

The mean (SD) number of injections was 8.2 (2.46) and 8.4 (2.87) in Groups I and II, respectively. The frequencies of ranibizumab injections administered over 12 months were generally comparable in both treatment groups (Fig. 5).

Ocular serious AEs (SAEs) were reported in a small but comparable proportion of patients in both groups (Table 2). Non-ocular SAEs were reported in 8.1% of patients in Group I and in 12.8% of patients in Group II (Table 2). Seven patients died during the study; one patient in Group I (due to cardiopulmonary failure) and six patients in Group II (due to myocardial infarction, cardiovascular insufficiency, pneumonia, squamous cell carcinoma, congestive cardiac failure, and unknown causes [n = 1 each]). One death reported in Group II due to unknown causes was suspected to have a causal relationship with the study drug.

Overall, ocular and non-ocular AEs were reported in a comparable proportion of patients in both groups (Table 3). The ocular AE with the highest incidence in Group I was conjunctival hemorrhage (5.7%) and IOP increase in Group II (6.3%, Table 3). There were no cases of endophthalmitis.

Ocular and non-ocular AEs suspected to be related to the study drug and/or ocular injection were also reported in a comparable proportion of patients in both groups. Conjunctival hemorrhage in Group I and IOP in Group II were the most common ocular AEs (see, Additional file 6: Table S3). Ocular AEs leading to study drug discontinuation were reported by 2 patients, one in each group (Group 1: retinal detachment and VA reduced; Group II: vitreous hemorrhage). Nine patients discontinued from study drug due to non-ocular AEs, one from Group I (asthma) and eight from Group II (cardiac failure, presyncope, migraine, cholestasis, chronic myeloid leukemia, ischemic stroke, neutropenic sepsis, squamous cell carcinoma, and death [n = 1 each]).