Medical comorbidities and patient demographics might be considered when selecting amongst migraine preventive medications.
This is a real-world electronic health record (EHR) study using two comprehensive and longitudinally established EHR databases, one including patients with migraine seen by any provider at Mayo Clinic and another with those seen by headache specialists at Mayo Clinic with documented treatment outcomes. We analyzed demographics, comorbidities, number of preventive medication trials, and responses to beta-blockers, topiramate, tricyclic antidepressants (TCA), onabotulinumtoxinA (BoNT/A), and CGRP monoclonal antibodies (mAbs).
We included 71,863 patients (female, 58,953 [82%]; chronic migraine, 21,640 [30%]) with migraine for prescription analysis and 4620 for response analysis (female, 3356 [73%]; chronic migraine, 3217 [75%]). Male patients were more likely to receive beta-blockers (odds ratio [95% confidence interval]: 1.24 [1.17–1.30]) and TCA (1.29 [1.22–1.36]), but less likely to be prescribed topiramate (0.66 [0.62–0.70]) and BoNT/A (0.73 [0.67–0.80]); they were less likely to respond to TCA (0.68 [0.50–0.92]) and BoNT/A (0.54 [0.39–0.74]). Patients with chronic migraine were less likely to be prescribed beta-blockers (0.50 [0.47–0.53]), topiramate (0.66 [0.62–0.69]), and TCA (0.47 [0.44–0.49]), but more likely to be prescribed BoNT/A (5.71 [5.27–6.18]) and CGRP mAbs (2.25 [2.09–2.41]). The positive predictors of response included thyroid disorders (1.32 [1.01–1.74]) for BoNT/A and migraine with aura for CGRP mAbs (1.60 [1.07–2.40]). Patients with hypertension were less likely to receive (0.60 [0.57–0.64]), but more likely to respond to topiramate (1.41 [1.03–1.93]). Patients with depression/anxiety were more likely to receive (1.28 [1.23–1.34]) but less likely to respond to TCA (0.68 [0.53–0.88]).
Using our comprehensive EHR databases, we found that the presence of certain demographics and medical comorbidities influence prescription patterns and response to migraine preventives. The mechanisms by which comorbidities affect treatment response should be further explored.