Predicting Inadequate Response to Acute Migraine Medication: Results From the American Migraine Prevalence and Prevention (AMPP) Study

Background: Pain freedom at 2 hours and sustained pain response at 24 hours are important outcomes of acute migraine therapy. Some studies have examined rates and predictors of successful treatment outcomes for single attacks in clinical trials. However, little is known about predictors of typical response to acute treatment over multiple attacks in the population. Objectives: To identify sociodemographic features, headache characteristics, comorbidities and treatment-related factors that predict acute treatment success or failure at 2 hours and 24 hours post dose in a US population sample of persons with episodic migraine. Methods: Eligible respondents completed the 2006 American Migraine Prevalence and Prevention Study survey, met criteria for episodic migraine, reported the use of acute treatment for migraine and answered questions on acute treatment outcomes from the Migraine Treatment Optimization Questionnaire (mTOQ). One question focused on 2 hour pain free (2hPF) response and the other focused on pain relief at 2 hours and 24 hours (24hPR). For each question, responses were considered adequate if they were achieved “half the time or more” and inadequate if they were achieved “never,” “rarely,” or “less than half the time.” Models were run to identify predictors of outcomes in relation to usual acute treatment: (1) Inadequate 2hPF response; (2) Inadequate 24hPR response; and (3) Inadequate 24 hour Sustained Pain Freedom (24hSPF), which was a conditional analysis of pain freedom at 24 hours among those who initially had an adequate pain freedom response at 2 hours. Binary logistic regression models were used to separately predict each of the 3 outcomes adjusting for covariates. Potential predictor variables were analyzed independently and variables that did not contribute significantly to outcome prediction were trimmed. The retained variables were entered into a final multivariable binary logistic regression that included age, sex, and the covariates that survived the trimming process. Odds ratio (OR) and 95% confidence interval (CI) statistics were generated for each variable. A value of P ≤.05 was used to identify statistically significant variables. Results: Among 8233 eligible respondents with episodic migraine, 56.0% of respondents reported Inadequate 2hPF response to usual acute treatment and 53.7% reported Inadequate 24hPR. Among the 44.0% of individuals who achieved Adequate 2hPF (N = 3621), 25.7% reported Inadequate 24hSPR or recurrence. Significant predictors of Inadequate 2hPF response included demographic variables (male sex and higher body mass index [BMI]), headache features (higher headache pain intensity, cutaneous allodynia, more headache days per month), comorbidity (depression), and migraine pharmacologic treatment factors (not using preventive migraine medications). Significant predictors of Inadequate 24hPR included headache features (more headache days per month, higher migraine symptom severity composite score, cutaneous allodynia), comorbidity (depression), and migraine pharmacologic treatment factors (medication overuse). Significant predictors of Inadequate 24hSPR were greater monthly headache day frequency, cutaneous allodynia, depression, medication overuse, and higher migraine symptom severity composite score. Conclusions: These results provide a snapshot of the effectiveness of usual acute treatment as well as predictors of inadequate acute treatment response in a large population sample of people with episodic migraine. These results highlight the high rates of unmet treatment needs in people with migraine. Similarities and differences in predictors vary with outcome for reasons that are discussed.