Antivascular endothelial growth factor (VEGF) agents have revolutionised our treatment paradigms and therapeutic approach to numerous retinal diseases. Intravitreal anti-VEGF therapies are now at the forefront of our armamentarium for the treatment of neovascular age-related macular degeneration (AMD). Three anti-VEGF drugs are now available for use: bevacizumab (Avastin), ranibizumab (Lucentis), aflibercept (Eylea). Two of these drugs are FDA-approved in the USA for the treatment of neovascular AMD: aflibercept and ranibizumab.1–3 On the other hand, bevacizumab, although approved for systemic treatment of various malignancies, is used off-label for ophthalmic applications.4

The comparative efficacy and safety of these drugs in numerous ophthalmic diseases, including neovascular AMD, continues to be actively debated as new data accumulates. The use of bevacizumab, in particular, remains among the most hotly contested topics. Questions about functional and anatomic efficacy, treatment regimens, systemic safety, and compounding safety fuel the energy within the debate.5–8 With those issues in mind, relative cost difference and cost burden maintains this discussion among clinicians, payers and governments. The cost difference in the USA is nearly 35–40-fold less for bevacizumab (∼US$50) compared with ranibizumab (∼US$2000), aflibercept (∼US$1800).9 ,10 For perspective, gold (as of now) is trading at ∼US$1800 an ounce which, as one recent article has mentioned, places the equivalent volume of liquid gold at approximately 5–6 times less than ranibizumab or aflibercept.11 What we want as a medical community is the best investment for our patients’ future vision and health; and to that end, our search continues for answers related to the relative efficacy, safety and value of these drugs so that we can help guide our patients towards the best therapeutic decision for them.

In this issue of British Journal of Ophthalmology, Krebs and colleagues add to the growing body of evidence of comparative medicine with the 1-year results of the Multicenter Anti-VEGF Trial in Austria (MANTA).12 This is the third major head-to-head randomised multicentre trial to release 1-year results for comparative effectiveness of bevacizumab and ranibizumab in neovascular AMD. The Comparison of AMD Treatment Trial (CATT) in the USA, and the Inhibit VEGF in Age-related choroidal Neovascularisation (IVAN) trial in the UK, have both provided 1-year results, and CATT has released 2-year results (table 1).9 ,12–14

MANTA is a prospective randomised double-blind multicentre comparative study examining the effectiveness of bevacizumab and ranibizumab when used in an as-needed (pro re nata (PRN)) optical coherence tomography guided regimen in 321 patients. All patients were treatment-naïve with subfoveal choroidal neovascularisation (CNV) secondary to AMD. The treatment protocol for MANTA included 3 consecutive monthly injections at study onset and PRN treatment following the loading phase. Retreatment was dictated based on vision loss of ≥5 letters with associated evidence of active CNV, new macular haemorrhage, new classic CNV, increase in central retinal thickness on OCT of ≥100 µm, or persistent fluid on OCT. The primary outcome in this study was the mean change in best-corrected visual acuity (BCVA) at 1 year. The study was powered to detect a difference of seven letters with 95% confidence, more letters than both CATT (five letters with 99.2% confidence) and IVAN (3.5 letters with 95% confidence).9 ,12–14

The results of MANTA support much of the previous published literature regarding comparative efficacy. Change in BCVA was equivalent between the two drugs (4.9 letters for bevacizumab and 4.1 letters for ranibizumab). There were no statistically significant differences in 1-line and 3-line gainers or losers. Interestingly, though not statistically significant, the number of patients with 3-line gainers was slightly less in the ranibizumab group. Additionally, the ranibizumab group had a slightly higher number of patients losing 3 lines (again not statistically significant).12 As noted in the 2-year data of CATT, there is a slight fall-off in visual acuity with PRN treatment over time. In CATT, this was significant compared with monthly treatment at 2 years.14 The visual acuity curve in MANTA appears to peak at 4 months (following initial monthly induction) at approximately a 6.5-letter gain for bevacizumab and a 6.0-letter gain for ranibizumab. Though not statistically significant, over the next 8 months, the mean change in visual acuity begins to trend down to 4.9 letters for bevacizumab, and 4.1 letters for ranibizumab supporting caution with a PRN regimen over time.12

In both IVAN and CATT, ranibizumab was superior to bevacizumab in reducing retinal thickness and removing fluid.9 ,13 ,14 In MANTA, the anatomic response was similar between the two drugs.12 The central retinal thickness decreased by 86.30 µm in the bevacizumab group, and 89.86 µm in the ranibizumab group (less than 4 µm different). Though not statistically significant, the absolute central retinal thickness in the ranibizumab group (275.14 µm) was slightly less than the bevacizumab group (288.29 µm). However, the baseline thickness of the bevacizumab group (374.59 µm) was also slightly thicker than the ranibizumab group (365.0 µm), again not statistically significant.12 The number of injections was similar in both groups. Both groups required approximately three additional injections during the year following the initial three injection-loading protocols, mean total of 6.1 bevacizumab injections and 5.8 ranibizumab injections.12 In CATT, the PRN group required a mean of 7.7 bevacizumab injections and 6.9 injections during the first year.9 In CATT, there was no loading regimen; and although it is difficult to compare, it is interesting that in MANTA, the number of injections needed was nearly 1–2 injections less (per year), even after mandating an initial three injections.

The difference in systemic adverse events between the two drugs has garnered significant attention. In MANTA, there was no significant difference between the two drugs in adverse events. Bevacizumab had 19 adverse events compared with 15 for ranibizumab (p=0.37).12 However, MANTA (321 participants) is significantly smaller than both IVAN (610 participants) and CATT (1208 participants).9 ,12 ,13 Even with the large size of CATT, it was not powered to find small differences in rare adverse events, so only limited conclusions can be drawn regarding the adverse events in MANTA.9 ,12–14 All three trials have found overall systemic adverse events to be higher in the bevacizumab group (although only CATT was statistically significant), but none of the studies have found bevacizumab to be a higher risk for arteriothrombotic events (eg, non-fatal myocardial infarction, stroke). In fact, in IVAN, ranibizumab had a higher rate of arteriothrombotic events or heart failure compared with bevacizumab (2.9% vs 0.7%, p=0.03).13

In examining the systemic effects of localised ocular VEGF suppression, IVAN was the first study to report systemic VEGF levels which were lower in monthly treatment versus discontinuous treatment, and higher in the ranibizumab group, suggesting increased systemic VEGF suppression with bevacizumab therapy and with monthly therapy.13 The implications of the changes in serum VEGF levels are still not known, however, it is clear that ocular anti-VEGF therapy impacts systemic VEGF levels. Regarding local adverse events, there were no cases of endophthalmitis in MANTA.12 At 1 year, CATT reported six cases of endophthalmitis (all in the monthly treatment groups).9 No cases of endophthalmitis were reported in IVAN.13

Krebs and colleagues add important data to the ever-growing body of literature on comparative data for anti-VEGF therapy in AMD with the 1-year MANTA results. In both CATT and IVAN, ranibizumab was slightly more favourable at nearly all endpoints for visual acuity (though in most cases not statistically significant).9 ,13 ,14 MANTA is the first study to report bevacizumab with a slightly more favourable trend (again not statistically significant).12 These variations may simply reflect noise within the trials. However, the consistent message among all the comparative trials is that we have two drugs that have impressive and very similar visual acuity results.

With the addition of aflibercept and the outstanding results in the VIEW trials, we have three excellent drugs in our armamentarium that have transformed the AMD landscape.3 Concerns certainly still persist regarding relative systemic safety, relative treatment regimen efficacy, development of geographic atrophy, the persistence of residual fluid, and a long-term solution to treatment burden for both physicians and patients.5 ,13 ,14 Recent outbreaks of endophthalmitis following intravitreal bevacizumab therapy also highlight the safety concerns regarding drug integrity (eg, counterfeit drug) and issues surrounding drug compounding.6–8

The ideal treatment regimen also remains unknown. All these trials, to date, have examined either the monthly regimen or a modified PRN approach. Yet, with the perceived improvement in treatment/visit burden, and the importance of maintaining a fluid-free macula, treat-and-extend has become one of the most common treatment protocols utilised for any of these drugs, though it has the least evidence of comparative efficacy.15–18

One final issue is cost. How does cost fit in with management decisions? If a system of outcomes-based reimbursement is put in place, will cost be considered? Will a value-based payer system be utilised where cost and outcomes are incorporated into physician reimbursement? And if so, how do we apply these trial results and others to provide an optimal approach that protects independent clinician decision making and maintains an individualised approach to patient care based on the best available clinical evidence? Future comparative trial findings, including long-term results, will help answer many of these questions. Ultimately, continued drug development and novel approaches for sustained therapeutic delivery to enhance long-term results while reducing treatment burdens for our patients are needed.