Reports

This report is based on medical evidence presented at sanctioned medical congress, from peer reviewed literature or opinion provided by a qualified healthcare practitioner. The consumption of the information contained within this report is intended for qualified Canadian healthcare practitioners only.

MEDICAL FRONTIERS - Association for Research in Vision and Ophthalmology (ARVO) 2012 Annual Meeting

Fort Lauderdale, Florida / May 6-10, 2012

Fort Lauderdale - The controversy regarding the relative efficacy and safety of vascular endothelial growth factor inhibitors in the treatment of age-related macular degeneration was addressed but not resolved by 2 large, independent studies that compared ranibizumab, which is licensed for this indication, to bevacizumab, which is being used off-label. On the basis of 2 years of follow-up in the CATT trial and one year of follow-up in the IVAN trial, ranibizumab and bevacizumab were characterized by the trial authors as providing similar efficacy. However, the greater number of serious adverse events in the bevacizumab arm in both studies complicated interpretation of the results. Moreover, several surrogate markers of drug activity favoured ranibizumab, including protection against lesion growth. The debate, perhaps fuelled rather than modified by the studies, encompasses a range of issues, including those regarding cost as a driver of off-label medication use and the safeguards of drug licensing.

Chief Medical Editor: Dr. Léna Coïc, Montréal, Quebec

Inhibition of the vascular endothelial growth factor (VEGF) protein slows and in some cases halts the vision loss that defines the clinical expression of wet age-related macular degeneration (AMD), which is largely driven by the growth of new blood vessels. Ranibizumab has demonstrated efficacy in the inhibition of AMD angiogenesis in phase III studies and is approved for this application. Recently, physicians in many countries, particularly the US, have been employing bevacizumab, another VEGF inhibitor currently indicated in the treatment of colon cancer, in off-label AMD treatment. The impetus is a lower acquisition cost. The CATT (Comparison of Age-related Macular Degeneration Treatments Trial) and IVAN (Inhibition of VEGF in Age-related choroidal Neovascularisation) studies were launched in the US and UK, respectively, to evaluate whether bevacizumab is a reasonable substitution.

Findings from CATT

Presenting new results from CATT at the 2012 ARVO meeting, Dr. Daniel F. Martin, Chair, Cole Eye Institute, Cleveland Clinic, Ohio, concluded that the non-inferiority of bevacizumab for visual acuity observed at the end of one year persisted through a second year. However, due to differences in safety and signals of disease progression, 2-year data produced many of the same dilemmas as the one-year data, according to Dr. Martin.

“I have to say that it has been pretty entertaining because, in a sentence, [ranibizumab] users see differences while [bevacizumab] users see similarities,” Dr. Martin remarked. He was referring to the one-year CATT but he may have been referring just as easily to the most recent findings. In essence, the visual acuity results, although generally providing ranibizumab with a non-significant numeric advantage, were similar, but many other important end points, particularly safety, also favoured ranibizumab. For a chronic therapy with the potential for differences to enlarge over time, there is room for concern.

In CATT, the 1185 participating patients were initially randomized into 4 groups: ranibizumab monthly, bevacizumab monthly, ranibizumab PRN and bevacizumab PRN. This permitted comparison of both the drugs and common dosing strategies. At one year, non-inferiority was defined as an average difference in visual acuity of less than 5 letters. Interestingly, there was a notable difference when comparing PRN vs. monthly dosing for each agent. Ranibizumab administered PRN was non-inferior to the same agent given monthly. In contrast, PRN bevacizumab failed to meet non-inferiority vs. the monthly regimen, the results being deemed “non-conclusive.” Overall, the authors concluded that bevacizumab was non-inferior to ranibizumab for visual acuity when administered according to the same schedule, even though the rate of serious adverse events (SAEs) on ranibizumab was significantly lower (P=0.04) and some signs of drug activity, particularly the absence of fluid on optical coherence tomography (P<0.001), were significantly greater.

The goals of the comparison at 2 years in the CATT trial were altered slightly, but the results were similar. Most important of the differences in the 2- vs. the one-year analysis, 549 patients who had been receiving monthly injections during the first year were randomized to remain on monthly injections or switch to PRN in the second. As a result, the 2-year study allowed investigators not only to continue comparing the relative efficacy and safety of ranibizumab and bevacizumab but also to gather new information on the relative importance of the dosing strategy. Safety and efficacy data from the 1107 evaluable patients were presented in respect to drug assignment and regimen assignment.

Figure 1. 

Again, the average differences in visual acuity were modest (Figure 1), both in regard to the agent and the regimen. Although ranibizumab produced consistently better visual acuity than bevacizumab (whether delivered monthly or PRN), and monthly regimens did better than PRN regimens (whether patients received bevacizumab or ranibizumab), there was debate about the clinical value of these differences. It is notable, for example, that while visual acuity at 104 weeks was similar to that observed at 52 weeks in 3 of the arms, there was a modest but steady decline in average visual acuity starting at about week 64 in the PRN bevacizumab group. One of the problems in showing a relative advantage for ranibizumab was the end-of-first-year re-randomization of monthly patients to continue receiving monthly or PRN therapy. This subverted the effort to compare monthly and PRN therapy because of loss of statistical power.

However, the rate of SAEs had widened at the end of 2 years when compared to one year. When those on either regimen of ranibizumab were compared to those on either regimen of bevacizumab, the adjusted risk ratio was 30% greater (RR 1.3; 95% CI, 1.07-1.57; P=0.009) for bevacizumab. When rendered graphically, the relative increase in SAEs over the course of the 2-year study was steady and progressive.

Figure 2.

In addition, several measures suggested that ranibizumab may exert greater control over the underlying pathology. As in the first year, the mean change in lesion area was significantly smaller (P=0.006) among those taking ranibizumab when compared to bevacizumab, whether on a monthly (-0.4 vs. +1.6 mm²) or PRN (+1.9 vs. +3.0 mm²) schedule (Figure 2). Similarly, the proportion of patients with no fluid at the end of 2 years was significantly greater (P=0.0003) on ranibizumab whether on a monthly (46% vs. 30%) or a PRN (22% vs. 14%) regimen (Figure 3). Geographic atrophy, although greater on monthly (26% vs. 18%) or PRN (15% vs. 13%) ranibizumab, was not significantly different (P=0.13). The difference in mean change in total retinal thickness, although smaller on ranibizumab than bevacizumab, did not reach statistical significance by drug assignment; the greater protection offered by monthly over PRN injections did approach statistical significance (P=0.08).

The average number of injections on PRN therapy was only one fewer on ranibizumab than bevacizumab (13 vs. 14), but this difference did reach statistical significance (P=0.01). As PRN injections were driven by persistent disease, this difference supports greater activity from ranibizumab. Although comparisons in CATT were limited to 2 years, it is relevant to consider the effect of even modest differences in a lifelong condition.

Figure 3. 

IVAN Results and Other Discussions

The one-year results of IVAN, which had a similar 4-group randomization, were highly consistent with those of CATT. Presented by Prof. Usha Chakravarthy, Queen’s University, Belfast, UK, the 1.99-letter advantage of ranibizumab over bevacizumab at one year approached statistical significance (P=0.056) but did not exceed the predefined margin of non-inferiority. The odds ratio for any systemic adverse event was increased by 35% in the bevacizumab group (OR 1.35; 95% CI, 0.8-2.27; P=0.25), even though the difference did not reach statistical significance. However, it is notable that the VEGF serum concentrations were significantly lower on bevacizumab than on ranibizumab in the IVAN study, suggesting that the former agent was exerting a systemic effect. The migration of bevacizumab out of the eye is one potential explanation for the greater rate of adverse events on this agent. All of the morphological changes, including lesion area and dye leakage, favoured ranibizumab but these were not significant. Due to the lack of significance, the authors concluded that bevacizumab was non-inferior and that PRN treatment was comparable to monthly therapy.

Several objections to the claim of non-inferiority were raised during the discussion period after the presentation as well as in other forums over the course of the ARVO meeting. One objection was whether dose equivalents were compared. The authors of CATT and IVAN responded that no effort was made to define dose equivalence primarily because there are no prior dose-finding or efficacy studies with bevacizumab in AMD. Rather, the decision was made to employ the dose of bevacizumab most commonly used by US physicians for AMD. The potential challenges of reconstitution of bevacizumab, which is not marketed in ophthalmic doses, concern the relative safety of these agents. In IVAN, a rigorous protocol of reconstitution by trained staff using glass vials contrasts with the less rigorous practices in day-to-day care in which sterilized glass vials may not be employed with the same degree of diligence. Indeed, a recent cluster of cases of infectious endophthalmitis in patients treated with bevacizumab that was attributed to reconstitution (Gonzalez et al. Am J Ophthalmol 2012;153:196-203) indicates that the relative risks of these agents may differ in a real-world setting.

The consistent but generally non-significant advantage of monthly anti-VEGF therapy over PRN management, which reduced the number of injections over 2 years by almost 50%, led authors of both CATT and IVAN to note that a PRN schedule appears to be preferable for most AMD patients. In addition to a lower cost and reduced risk of injection-related complications, they speculated that most patients would prefer to forego monthly injections if visual acuity were not significantly affected. The second year of the CATT study, in which patients previously on monthly injections were randomized to remain on monthly injections or switch to PRN, demonstrated that there was no disadvantage from starting on a monthly regimen and then switching to PRN at a later time. In CATT, those who remained on monthly injection for the full 2 years did have the best average visual acuity; but the advantage over PRN when started initially or started after one year, which provided similar results at 2 years, was not significant.

Summary

Within the predefined criteria and time limits of the trials, both sets of investigators declared bevacizumab to be non-inferior to ranibizumab, but vigorous discussion after the presentation of the data suggested that these conclusions were not uniformly embraced. The differences in safety and the potential for the advantages associated with ranibizumab at one year to progressively increase over time led many in the audience to conclude that the 2 trials support the conclusion that the agents are not interchangeable.