Beyond headache days: Refining the framework for understanding CGRP monoclonal antibody discontinuation

Dear Editor,

We read with great interest the article by Menzel et al. examining reasons for discontinuation of calcitonin gene-related peptide (CGRP) monoclonal antibody (mAb) therapy in a real-world cohort of drug-resistant migraine patients. ¹ The study addresses a clinically relevant topic, and the authors are to be commended for their effort to characterize discontinuation patterns beyond predefined response thresholds. The finding that subjective improvements in attack intensity and duration may drive treatment continuation among formal non-responders represent a valuable contribution to the field. We wish to raise several methodological considerations that may refine the interpretation of these findings and inform future research.

First, the handling mAbs switching introduces an analytical inconsistency that warrants further discussion. Outcome measures including monthly headache days (MHDs), monthly migraine days (MMDs), and acute medication days were analyzed exclusively for the first initiated mAb, whereas treatment duration and persistence were calculated across the entire treatment course with mAbs including sequential switches. Figure 3 in the study by Menzel et al. illustrates this discrepancy clearly: 55.9% of patients who discontinued treatment due to ineffectiveness had trialed two or more mAbs, compared to 35.4% of those continuing therapy. Consequently, discontinuation reasons reflect the cumulative experience across multiple agents, while efficacy outcome reflect only the initial treatment. Without separate analysis of post-switch outcomes, the study may overestimate the true rate of CGRP pathway non-responsiveness and obscure potentially modifiable factors in the discontinuation decision. We recommend that future studies should report switching outcomes independently to distinguish agent-specific from class-level failure.

Second, the definitions of “subjective benefit” and “treatment effectiveness” represent a central element of the study's conclusions, but they were not anchored to any validated patient-reported outcome measure (PROM). Although the absence of instruments such as the Migraine Disability Assessment Scale (MIDAS) and Headache Impact Test-6 (HIT-6) has been acknowledged as a study limitation, the implications for the study's interpretation merit further emphasis. Evidence suggest that combining validated outcome measures with migraine frequency substantially enhances detection of treatment response: Göbel et al. reported response rate of 84.8% in episodic and 77.7% in chronic migraine using MMDs plus MIDAS/HIT-6, compared with 50.3% and 58.2% using frequency alone. Similarly, Muñoz-Vendrell et al. found that 91.5% of apparent non-responders demonstrated improvement in at least one outcome domain, including MIDAS improvement in 60.2% and Patient Global Impression of Change (PGIC) ≥5 in 54.5%.^2,3 Incorporation of validated PROMs would allow to capture qualitative benefits in a standardized, reproducible way, to facilitate cross-study comparisons and strengthening the conclusion that treatment effectiveness extends beyond the reduction of headache frequency.

Third, the discontinuation framework limited to perceived ineffectiveness, adverse events, and subjective remission does not encompass several factors commonly encountered in real-world practice, including insurance or access barriers, treatment fatigue, patient preference, and comorbidity-driven discontinuation. These unmeasured factors may have contributed to some discontinuations classified under “perceived ineffectiveness,” potentially inflating this category's reported proportion of 67.2%. Additionally, the reported secondary loss of effectiveness (SLE) rate of 9.3% (4/43) among patients discontinuing for ineffectiveness contrasts notably with higher rates reported by Kaltseis et al. ⁴ Beyond acknowledged design differences, the small absolute numbers (n = 4) yield a 95% confidence interval of approximately 2.6–22.1%, reflecting substantial statistical imprecision. More importantly, because 76.7% of discontinuations for ineffectiveness occurred within 6 months, most patients were not observed long enough for SLE to manifest. This temporal constraint likely biased the SLE estimate downward. Supporting this interpretation, Raffaelli et al. prospectively demonstrated that migraine burden worsens progressively within 8–16 weeks following CGRP mAb discontinuation in patients who had initially responded, underscoring that treatment response dynamics in this drug class evolve over longer timeframes than those captured in many discontinuation cohorts. ⁵ Furthermore, the 31.9% loss to follow-up (60/188 who were uncontactable or declining to participation) introduces potential selection bias, as patients experiencing SLE may have been more likely to disengaged from care.

Finally, the recent FDA approval of fremanezumab for pediatric episodic migraine (ages 6–17) has expanded the use of CGRP monoclonal antibodies into younger populations. The SPACE trial demonstrated significant reductions in monthly migraine days versus placebo, while emerging real-world data showed 6-month treatment persistence of 88.2% for CGRP mAb versus only 30.6% for conventional oral preventives in adolescents. These findings highlight the increasingly important limitation of adult-only discontinuation studies.^6,7 Future research should extend discontinuation analyses to pediatric populations, where treatment expectations, decision-making dynamics, and reasons for cessation may differ substantially from adult cohorts.

In summary, Menzel et al. provide valuable real-world insights into CGRP mAb discontinuation in a challenging clinical population. Future studies would benefit from separating switching outcomes from first-agent responses, incorporate validated PROMs, and adopt comprehensive frameworks capturing both clinical and non-clinical reasons for discontinuation. Additionally, sensitivity analyses addressing loss to follow-up and inclusion of pediatric populations would further enhance the reproducibility and clinical applicability of findings in this evolving area of migraine therapeutics.