Long-term atogepant for treatment-resistant migraine in real-world clinical practice: 12-month results (gema project)

Abstract Background Short- and mid-term real-world data support the effectiveness of atogepant for migraine prevention, but evidence beyond 6 months remains scarce. This study evaluated the 12-month effectiveness, safety, persistence, and predictors of response in a large treatment-resistant migraine cohort. Methods This prospective multicentre study enrolled patients from 15 tertiary Headache Units in Spain within the GEMA Project. Clinical and demographic characteristics, previous and concomitant preventive treatments, monthly headache days (MHD), monthly migraine days (MMD), and adverse events were systematically collected at baseline and after 3, 6, and 12 months. In a subset of patients, Headache Impact Test (HIT-6), Hospital Anxiety and Depression Scale (HADS), and Insomnia Severity Index (ISI) were also assessed. Results A total of 513 patients were included. The cohort showed marked treatment resistance, with a median of 6 previous preventive failures; 74% ( n = 380) had chronic migraine, 84% ( n = 423) had previously failed onabotulinumtoxinA, and more than half had prior exposure to anti-CGRP monoclonal antibodies ( n = 270). Treatment persistence was approximately 80% at 6 months ( n = 409) and 60% at 12 months ( n = 304), with lack of effectiveness as the main reason for discontinuation. Among patients remaining on treatment with complete data at all follow-up visits, median MHD decreased from 20 days at baseline to 10 days at 3 months and 9 days at 6 months, remaining stable at 9 days at 12 months. Median MMD decreased from 13 days to 6 days at 3 months, 5 days at 6 months, and remained at 5 days at 12 months ( p < 0.0001 versus baseline for all time points). At 12 months, ≥ 50% response rates were 52.3% for MHD and 57.8% for MMD in the observed-case analysis, and approximately 31% and 33%, respectively, using non-responder imputation in the full baseline cohort. Improvements in disability and medication overuse were observed, while exploratory analyses in smaller subsets suggested improvements in anxiety, depression, and insomnia. A greater burden of previous preventive treatment failures independently predicted lower odds of achieving a ≥ 50% response at 12 months. The frequency of most reported adverse events was lower at later follow-up visits, and no new safety signals emerged. Conclusions Atogepant was associated with maintained clinical benefit over 12 months among patients who remained on treatment, with a favourable long-term safety profile in a highly treatment-resistant migraine population. Greater previous preventive failure burden predicted lower response rates, reinforcing the importance of earlier preventive optimisation. Trial registration ClinicalTrials.gov, NCT07699549, retrospectively registered on 8th of July 2026.