Treat and extend versus pro re nata regimens of ranibizumab and aflibercept in neovascular age-related macular degeneration: a comparative study

Age-related macular degeneration (AMD) is a common, chronic retinal disease that may cause dramatic impairment of visual acuity (VA). Neovascular AMD (nvAMD), a subtype, may be associated with the development of defunct macular blood vessels. NvAMD is characterized by leakage of fluid or blood into the macula, which severely affects central VA and may lead to legal blindness. Recent years have seen dramatic improvements in the therapeutic options of nvAMD, as anti-vascular endothelial growth factors (anti-VEGF), applied by intravitreal injection, inhibit development of these novel, aberrant blood vessels and the ensuing macular fluid leakage successfully.

Anti-VEGF treatment options have upended treatment of several retinal diseases, yet most notably of nvAMD, where treatment has to be applied often indefinitely and in short time intervals. Thus, disease management, albeit highly effective, constitutes a high burden for patients and healthcare providers alike. Moreover, every single injection bears a small yet significant risk of endophthalmitis, a feared complication threatening eyesight, making any attempt of reducing injection frequency a crucial, patient-relevant objective. [1]

Consequently, there is still some uncertainty on the optimal treatment regimen that manages to balance all pros and cons optimally. Importantly, results of the pivotal phase III trials on the use of ranibizumab and aflibercept were based on regular, uninterrupted dosing protocols. [2‐5] As a first attempt to alleviate high treatment burden, the pro re nata (PRN) regimen was introduced, a treatment protocol where follow-up intervals remain fixed, while decisions to carry out an injection are based on the anatomic findings at each respective visit. [6‐8] Notably, by following such a PRN treatment protocol, injections are based on prior pathological OCT findings and, therefore, fall short of preventing the damaging yo-yo effect of retinal fluid accumulation. In consequence, functional PRN results fared worse as compared with the pivotal phase III trials; in particular, initial VA gains could not be maintained as well during a longer follow-up time. [7‐9] In order to develop a protocol with a reduced treatment burden, while at the same time ensuring timely treatment before any macular fluid accumulation occurs, another treatment regimen, the treat and extend (T&E) protocol, was developed. In T&E regimens, the time interval of follow-up visits will be adjusted based on the patient’s clinical course—i.e. if a patient shows no sign of an active disease (e.g. the macula remains dry, without any leakage), intervals will be extended; if there is fluid accumulation, the next interval will be shortened. Importantly, at each patient’s visit following T&E, an injection will be performed; the current clinical status only has an impact on the duration of the next injection interval. Therefore, T&E may also be called “pro-active” in contrast to the more “reactive” PRN protocol. Prior studies have shown efficacy of T&E, but are mostly still limited to smaller studies. [10‐15] As pointed out by the American Academy of Ophthalmology, larger studies on the efficacy of T&E regimens are clearly warranted. [16]

The primary aim of this study was therefore to compare efficacy of PRN vs. T&E treatment protocols using ranibizumab, as well as comparing the effect of ranibizumab and aflibercept in clinical practice using the T&E protocol, both with regard to visual acuity (evaluated using logMAR i.e. the logarithm of the minimum angle of resolution) and with regard to the injection frequency.

Despite ongoing efforts to determine optimal treatment regimens of anti-VEGF treatment, a gold standard has not been determined yet, and the trade-off between treatment efficacy versus patient burden and injection complications remains controversial.

The results of this study largely confirmed available data of the published literature [10‐15, 17]. Notably, all three treatment regimens yielded favourable logMAR stabilization over a substantial observational phase of more than 40 months, and the initial gains of logMAR were for the most part maintained. The T&E treatment protocol proved to be overall more favourable in terms of speed and strength of logMAR improvement. As expected, annual numbers of examinations were higher in the PRN group than in the two T&E groups. However, yearly numbers of injections were lowest in the PRN group. This may be partly due to the more rigid and standardized injection scheme in the T&E regimen. Another reason may be better patient compliance in light of the reduced patient burden in the T&E groups. According to the authors’ experience, patients treated according to T&E appeared to be less anxious during the visit, knowing in advance that they would definitely receive an injection. Patients also expressed their appreciation for the prolonged intervals between consultations. As no ancillary data were collected on the patients’ subjective experience, no statistical evaluation could be performed to further support this observation.

Interestingly, ranibizumab and aflibercept T&E injection interval duration, while initially of similar length, started to diverge 2 years after the loading phase. We observed longer injection intervals on average in patients treated with ranibizumab than with aflibercept 3 years after the loading phase. These results are surprising, as aflibercept can be dosed less frequently than ranibizumab (every 8 weeks after loading phase) according to the Summary of Product Characteristics (SmPC). In fact, less frequent treatment was seen as a key advantage of aflibercept in comparison with other drugs. Further studies on the possible advantage of ranibizumab over aflibercept with respect to frequency of treatment will be therefore coveted, as any reduction in injections is highly relevant for patients, healthcare providers and national healthcare systems alike.

A strength of this study is certainly its high numbers of examined eyes and injections, along with its comparative results not only on different treatment protocols (T&E vs. PRN), but also on different drugs. To exclude effects of prior treatment, only patients receiving initial anti-VEGF injections were included. Thus, all data after a treatment change were excluded. The observational, retrospective design of this study implicates, naturally, certain limitations of the data quality, and a lower evidence level than those from a clinical trial. Despite the consecutive patient recruitment, implicit selection bias cannot be ruled out entirely. Moreover, treating physicians displayed changing preferences for ranibizumab vs. aflibercept over time: initially, they tended to choose ranibizumab for intraretinal leakage i.e. for somewhat milder cases, and aflibercept for patients with more severe findings, including subretinal leakage and signs of subretinal fibrosis. In a second phase, physicians treated all patients with aflibercept. In the last, third phase, physicians attempted to choose between both substances randomly, albeit in a subjective way. Thus, a certain degree of selection bias must be taken into account. In summary, these changing preferences may be the cause for the differing size of the patient groups, and it suggests that the aflibercept patients tended to be the more severe cases. Notwithstanding, the initial VAs in the ranibizumab T&E and aflibercept T&E groups were comparable, which again may relativize that assumption. However, as other previous publications described a higher efficacy of aflibercept than ranibizumab, this potential weakness of this study’s design must be taken into consideration.

Overall, the differences in mean VA values between all treatment groups were relatively small. However, our results showed clear trends supporting the prevailing hypothesis of T&E being superior to PRN treatment. Notably, in the PRN group, there was a considerable rate of patients who experienced a worsening of logMAR (> 0.2, equivalent to a 2-line change on the Snellen chart); whereas in the T&E groups, logMAR worsened only in very few patients.

In summary, this large observational study confirms prior data on the efficacy of anti-VEGF treatment, provides additional data on the potential superiority of T&E over PRN protocols, and raises the hypothesis of ranibizumab warranting fewer injections than aflibercept on the long run. Further studies will be required to further elucidate these findings.