Response to “Beyond headache days: Refining the framework for understanding CGRP monoclonal antibody discontinuation”

We thank the authors for their thoughtful and constructive comments on our work. We greatly appreciate their careful and critical appraisal of our study and are grateful for the opportunity to address the issues raised and to clarify several aspects of our analysis.

Tavasoli et al. ¹ correctly pointed out that our early outcome analyses were limited to the first calcitonin gene-related peptide (CGRP) monoclonal antibody (mAb) and did not capture outcomes across the entire treatment pathway. They highlighted that among patients who discontinued due to ineffectiveness, 55.9% had received more than one CGRP mAb, compared with 35.4% among those continuing therapy. However, this observation has the inherent limitation that it does not provide information on treatment response to the second or third CGRP mAb. Real-world evidence suggests that switching between CGRP mAb may still be effective in patients who did not respond to a previous CGRP-targeting mAb.^2,3 We acknowledge that the absence of post-switch outcome data limits agent-specific and class-level conclusions. At the same time, although standardized outcome analyses were limited to the first three months following initiation of the first CGRP mAb, persistence and treatment continuation were assessed across the entire CGRP-targeted treatment pathway, including treatment switches and at least one year of follow-up. This broader perspective provides insight into longer-term treatment decisions, while early real-world patterns of effectiveness may offer an initial indication of the subsequent treatment trajectory in these patients. Future longitudinal studies are needed to determine whether early response patterns can reliably predict treatment outcomes across subsequent CGRP-targeted therapies.

The authors also highlight the lack of validated patient-reported outcome measures (PROMs), such as MIDAS, HIT-6 or PGIC, which we acknowledged as a limitation in the original manuscript and which were unfortunately not systematically available in our dataset. Although these measures were not available, our study provides a complementary real-world perspective by demonstrating that qualitative patient-reported improvements, particularly reductions in attack intensity and duration, appear to influence treatment continuation decisions in routine clinical practice. Importantly, the studies cited by the authors^4,5 support our central conclusion that frequency-based response thresholds alone may underestimate clinically meaningful treatment benefits. Nevertheless, we agree that the absence of validated PROMs limits standardization, reproducibility, and direct comparison with other cohorts. Future studies should therefore systematically incorporate and report validated PROMs alongside conventional effectiveness measures to capture the full spectrum of patient-relevant treatment benefits.

The authors further noted the relatively small number of patients experiencing secondary loss of effectiveness (SLE) in our cohort. As acknowledged in the manuscript, the number of SLE cases was limited, and our observations should therefore be interpreted with caution. In our cohort, 76.7% of patients who discontinued treatment due to ineffectiveness did so within the first 24 weeks of therapy. We therefore believe that the low number of SLE cases should primarily be understood in the context of the high rate of early treatment discontinuation due to insufficient effectiveness. Because treatment response was assessed after three months, delayed treatment responses beyond 24 weeks, as reported by Barbanti et al., ⁶ and later occurrences of SLE may have remained undetected.

We also believe that the comparison with the study by Raffaelli et al. ⁷ should be interpreted with caution. That study primarily investigated the clinical course following treatment discontinuation and focused on the recurrence or worsening of migraine after a previously successful therapeutic response. In contrast, our analysis focused on reasons for discontinuation and early response patterns during treatment. These represent distinct clinical scenarios.

We acknowledge that nonparticipation among patients with unclear treatment status may have introduced selection bias, as their actual treatment trajectories and reasons for possible discontinuation remain unknown. This may have led to under- or overestimation of discontinuation rates and reasons, including SLE. At the same time, this challenge highlights the inherent difficulties of retrospectively assessing long-term treatment trajectories and treatment outcomes in routine clinical practice. Comprehensive evaluation of treatment persistence, delayed responses, and SLE will likely require prospective longitudinal data collection.

Finally, we agree that pediatric populations require dedicated studies. As CGRP mAb is currently not approved for pediatric migraine in Germany, our study was intentionally restricted to adults with drug-resistant migraine, and our conclusions should not be generalized to children and adolescents. Nevertheless, the age-related differences in treatment response observed for oral preventive therapies, as highlighted by the authors, represent an important research question that should be specifically addressed once sufficient pediatric data become available.

Taken together, the points raised by the authors underscore the need for prospective, multidimensional studies incorporating validated PROMs, structured assessment of discontinuation reasons, post-switch outcomes, and sensitivity analyses for missing follow-up. At the same time, they support the central implication of our study: in routine clinical care, the perceived effectiveness of CGRP mAb therapy extends beyond headache-day response thresholds and is shaped by qualitative patient-reported benefits that influence treatment continuation and discontinuation decisions.