Eyüpcan ŞENSOY, Mehmet ÇITIRIK, Burhan BAŞKAN, Hasan ÖNCÜL, Abdullah BEYOĞLU
Turkish Journal of Medical Sciences - 2026;56(5):1373-1383
Background/aim: Bevacizumab biosimilars are increasingly used as alternatives to reference bevacizumab in ophthalmic practice, but comparative real-world evidence in treatment-naïve neovascular age-related macular degeneration (nAMD) remains limited. In this study, the short-term anatomical, functional, and documented safety outcomes associated with bevacizumab, bevacizumab-bvzr (Zirabev), and bevacizumab-awwb (Mvasi) were evaluated in treatment-naïve patients with nAMD in a multicentre real-world setting. Materials and methods: This multicentre retrospective cohort included 250 treatment-naïve eyes with nAMD from 16 centres in Türkiye that completed a 3-month loading regimen with the same bevacizumab agent. Central macular thickness (CMT) was the prespecified primary longitudinal outcome; best-corrected visual acuity (BCVA) was secondary, while intraretinal fluid (IRF) and subretinal fluid (SRF) were repeated binary outcomes. CMT and BCVA were analysed using longitudinal linear mixed-effects models, whereas IRF and SRF were analysed using binomial generalised linear mixed-effects models with a logit link. Documented adverse events were assessed descriptively. Results: In the primary longitudinal CMT analysis, time (p < 0.001) and the agent-by-time interaction (p = 0.016) were significant, whereas the overall agent effect was not (p = 0.932). At the third follow-up, adjusted mean CMT was 295.1 (95% CI, 276.5-313.6), 302.4 (277.1-327.8), and 306.2 (281.7-330.7) µm for bevacizumab, bevacizumab-bvzr, and bevacizumab-awwb, respectively. For BCVA, time (p < 0.001), agent (p = 0.042), and the agent-by-time interaction (p = 0.031) were significant; corresponding adjusted mean BCVA values were 0.721 (95% CI, 0.632-0.810), 0.605 (0.483-0.726), and 0.897 (0.779-1.016) logMAR. IRF decreased over time (p = 0.001), without significant agent or interaction effects. Time-averaged adjusted SRF probability was lower with bevacizumab-bvzr than with bevacizumab and bevacizumab-awwb. No systemic adverse events were documented; ocular adverse-event observations were infrequent. Conclusion: Short-term anatomical and visual outcomes changed during the three-injection loading phase, with adjusted longitudinal trajectories differing for CMT and BCVA. These findings represent exploratory observational associations and do not demonstrate superiority or equivalence among the agents.
