Decoding the long-term safety of anti-CGRP (receptor) mAbs. A meta-analysis and systematic review

Objective To evaluate the long-term safety (≥12 months) of monoclonal antibodies (mAbs) targeting the calcitonin gene-related peptide (CGRP) or its receptor in migraine prevention by synthesising evidence from clinical trials and real-world studies. We focus on drug discontinuation due to adverse events and the type and frequency of adverse events. This is the first review to analyse the effects of the long-term use of all anti-CGRP (receptor) mAbs, aiming to provide novel insights for clinical practice and future treatment strategies. Methods We systematically searched PubMed, Cochrane Library, and ClinicalTrials.gov for studies with ≥12 months of anti-CGRP (receptor) mAb use between January 2013 and April 2025. A random-effects meta-analysis of proportions (logit transformation, inverse variance weighting, restricted maximum likelihood) was performed to estimate pooled discontinuation and adverse event rates. Risk of bias was assessed using the ROBINS-I score. Results From a total of 1,499 records, 14 met the inclusion criteria and were eligible for data analysis. These 14 records corresponded to 11 individual studies with observational durations all exceeding 12 months. Seven studies investigated erenumab, two eptinezumab, and one each fremanezumab and galcanezumab. All studies were judged to have a severe risk of bias due to their underlying design. The overall pooled proportion of treatment discontinuation for any reason among patients receiving anti-CGRP (receptor) mAbs was 23%, whereas the pooled proportion of discontinuation specifically due to adverse events was substantially lower at 3%. Time-trend analysis showed that adverse event–related discontinuation remained low (<5%) beyond the first year, while overall adverse event incidence was high at baseline (>70%) but did not further increase with prolonged follow-up. Conclusion Evidence on long-term use of anti-CGRP (receptor) mAbs over 12 months remains limited, but our analysis indicates good tolerability with consistently low adverse event-related discontinuation, no emergent safety signals, and largely non-serious, stable adverse event profiles. However, heterogeneity and study-level bias warrant cautious interpretation, highlighting the need for long-term clinical studies and continued real-world surveillance (e.g. registries).