Adherence to intravitreal aflibercept in neovascular age-related macular degeneration: 24-month results from the ANDROMEDA study.

Purpose: The ANDROMEDA study was planned to assess adherence to intravitreal aflibercept 2mg (IVT-AFL) treatment over 24 months for neovascular age-related macular degeneration (nAMD) and to identify patient- and/or physician-related factors for non-adherence (NA) in clinical settings.

Design: Prospective, observational, non-controlled, multicenter cohort study.

Subjects: Treatment-naïve and previously treated adult patients with nAMD under IVT-AFL treatment from 41 centers in Germany.

Methods: Time to first occurrence of NA was analyzed descriptively using Kaplan-Meier methods followed by a Cox model to explore the potential impact of patient- and physician-related factors on NA. Participants reported reasons for NA in standardized telephone interviews.

Main outcome measures: Primary endpoints were time to first occurrence of and reasons for NA. Secondary endpoints included change in best-corrected visual acuity and central retinal thickness from baseline to months 4, 12, and 24.

Results: The median time to first NA was 180 days in the 509 study participants (mean age: 77.2 years; 57.2% female). Among them, 44.0% were treatment-naïve, 22.6% previously treated with IVT-AFL, and 33.4% with other anti-VEGF agents (VEGF, vascular endothelial growth factor). Adherence rates, particularly among treatment-naive participants, fell early and markedly: 69.6% at 4 months (95% CI: 62.9% - 75.3%), 53.0% at 12 months (95% CI: 46.0% - 59.5%), and 40.1% at 24 months (95% CI: 33.3% - 46.9%). Key risk factors for NA (hazard ratio (HR) < 1 indicating a higher risk of NA for the mentioned factor, HR > 1 indicating a lower risk for the mentioned factor) included presence of any 'other diseases' in the treatment-naïve cohort at baseline [HR 0.57, 95% CI: 0.37-0.88], and patient-reported 'lack of information for accompanying persons' [HR 1.36, 95% CI: 1.03-1.79], involvement of referrals / multiple providers [HR 0.76, 95% CI: 0.60-0.97], and 'bilateral AMD' [HR 0.73, 95% CI: 0.57-0.93] in the total cohort.

Conclusions: In the ANDROMEDA study key determinants of NA were comorbidities, bilateral disease, treatment by multiple providers and a lack of patient caregiver education. Thus, better adherence may be achieved through holistic patient management considering additional disease parameters, single-center treatment, and improved (caregiver) education.