Simultaneous Comparisons of 25 Acute Migraine Medications: A Big Data Analysis of 10 Million Patient Self-Reported Treatment Records From A Migraine Smartphone Application
Chia-Chun Chiang1, Xuemin Fang2, Zsolt Horvath3, Francois Cadiou3, Alexandre Urani3, Weijie Poh3, Hiroto Narimatsu2, Yu Cheng4, David Dodick5
1Neurology, Mayo Clinic, 2Graduate School of Health Innovation, Kanagawa University of Human Services, 3Healint Pte. Ltd, 4Statistics, University of Pittsburgh, 5Neurology, Mayo Clinic Arizona
Objective:

To simultaneously compare patient-reported treatment effectiveness of 25 acute migraine medications using a big-data approach.

Background:

Many acute treatment options exist for migraine. However, large-scale, head-to-head comparisons of treatment effectiveness from real-world patient experience reports are lacking.

Design/Methods:

This is a cross-sectional analysis of 10,842,795 migraine attack records extracted from an e-diary smartphone application between June 30, 2014, and July 2, 2020. We analyzed 25 acute medications among seven classes- acetaminophen, NSAIDs, triptans, combination analgesics, ergots, anti-emetics, and opioids. Gepants and ditan were not included due to relatively lower numbers of usage when data was extracted. We employed a two-level nested logistic regression model to analyze the odds ratio(OR) of treatment effectiveness of each medication by adjusting concurrent medications and the covariance within the same user. Subgroup analyses were conducted for users in the United States(US), the United Kingdom(UK), and Canada(CA).

Results:

Our final analysis included 4,777,524 medication-outcome pairs from 3,119,517 migraine attacks among 278,006 users. Triptans(mean OR 4.8), ergots(mean OR 3.02), and anti-emetics(mean OR 2.67) were the top three classes of medications with the highest effectiveness, followed by opioids(mean OR 2.49), NSAIDs(mean OR 1.94), acetaminophen/acetylsalicylic acid/caffeine(OR 1.69), others(OR 1.49), and acetaminophen(OR 0.83), using ibuprofen as the reference. Individual medications with the highest ORs were eletriptan(OR 6.1), zolmitriptan(OR 5.7), and sumatriptan(OR 5.2). The ORs for 24 medications, except that of acetylsalicylic acid, achieved statistical significance, and our nested logistic regression model achieved an AUC of 0.849. Country-specific subgroup analyses revealed similar ORs of each medication and AUCs(US 0.849, UK 0.864, and CA 0.842).

Conclusions:

We analyzed patient-generated real-time records of 10 million migraine attacks and conducted simultaneous head-to-head comparisons of 25 acute migraine medications. Our findings that triptans, ergots and anti-emetics are the most effective classes of medications align with the guideline recommendations and offer generalizable insights to complement clinical practice.

10.1212/WNL.0000000000202426