PDF: MOTS
(2022)Objective
To determine whether migraine preventive therapy without switching or limiting the overused medication was noninferior to migraine preventive therapy with switching from the overused medication to an alternative medication used ≤2 days per week for treating chronic migraine with medication overuse
Study Summary
• Both strategies produced clinically meaningful reductions in headache frequency (baseline 22.5 → week 12: 15.4 vs 16.5 days)
• Switching resulted in lower prevalence of medication overuse at weeks 9-12 (53% vs 73%, p<0.001)
Intervention
Two treatment strategies: (1) Switching = migraine preventive medication plus switching from the overused symptomatic medication to an alternative used ≤2 days/week; (2) No switching = migraine preventive medication plus continuation of the overused medication with no maximum limit
Inclusion Criteria
Adults ≥21 years meeting ICHD-3 beta criteria for chronic migraine and medication overuse; no change in migraine preventive therapy or dose within 4 weeks before randomization
Study Design
Arms: Two arms: (1) Switching — preventive medication plus switching to alternative symptomatic medication used ≤2 days/week; (2) No switching — preventive medication plus continuation of overused medication without a maximum limit
Patients per Arm: 720 randomized (361 switching, 359 no switching); primary analysis (multiple imputation) included all 720; available-case analysis at weeks 9-12 included 294 switching and 291 no switching
Outcome
• Secondary: Lower MO prevalence with switching (53% vs 73%, p<0.001); fewer symptomatic medication days with switching (10.2 vs 14.7, p<0.001)
• Quality of life, disability, anxiety, and depression improved similarly in both groups
Bottom Line
For patients with chronic migraine with medication overuse, migraine preventive medication without switching or limiting the overused symptomatic medication is noninferior to migraine preventive medication with switching to a different symptomatic medication limited to ≤2 days per week when the outcome of interest is reduction in moderate to severe headache days over 12 weeks. Both strategies resulted in clinically meaningful improvements. Switching strategy resulted in lower medication overuse prevalence (53% vs 73%) but similar overall outcomes. The optimal strategy may differ based on individual patient characteristics such as baseline medication use frequency and treatment goals.
Major Points
- First large pragmatic randomized trial directly comparing two commonly used treatment strategies for chronic migraine with medication overuse
- 720 participants randomized from 34 US clinics (6 primary care, 14 general neurology, 14 headache specialty) between February 2017 and December 2020
- Open-label, patient-centered design with patient partners involved in all aspects of trial design and conduct
- Population had severe disease burden: mean 22.5 headache days per 28 days, 12.8 moderate-severe headache days, 21.4 days of symptomatic medication use at baseline
- Primary outcome: moderate to severe headache day frequency during weeks 9-12, with non-inferiority testing (margin 1.5 days per 28 days); primary analysis used multiple imputation and included all 720 participants
- Non-inferiority met at weeks 9-12 (multiple imputation, primary): switching 9.3 vs no switching 9.1 days per 28 d; difference 0.2, 95% CI -0.9 to 1.3, p=0.71 (available case: 9.3 [SD 7.2, n=294] vs 9.1 [SD 6.8, n=291]; difference 0.2, 95% CI -1.0 to 1.3, p=0.75)
- Superiority not demonstrated during weeks 1-2 (per 14 d): switching 6.6 vs no switching 6.4; multiple imputation difference 0.2, 95% CI -0.4 to 0.8, p=0.55 (available case: 6.6 [SD 3.7, n=313] vs 6.4 [SD 3.6, n=308], p=0.57)
- Both strategies resulted in clinically meaningful reductions: 3.5-5.3 days reduction in switching arm, 3.6-4.9 days reduction in no switching arm
- Switching strategy achieved lower prevalence of medication overuse at weeks 9-12: 53% vs 73%, difference -19%, 95% CI -27% to -11%, p<0.001
- Switching strategy had fewer symptomatic medication days at week 12: 10.2 vs 14.7 days, difference -4.5, 95% CI -5.9 to -3.1, p<0.001
- No difference in total headache days (not just moderate-severe) at week 12: 15.4 vs 16.5 days, p=0.14
- Quality of life measures improved similarly in both groups: HIT-6, PROMIS Pain Interference, GAD-7, PHQ-9, EQ-5D-5L all showed no significant differences
- Subgroup analysis suggested patients with higher baseline symptomatic medication use (>23 days per 28 days) might benefit more from switching (interaction p=0.048, unadjusted)
- Patients with higher baseline anxiety (GAD-7 >7) showed a trend toward better outcomes without switching (interaction p=0.16, hypothesis-generating only)
- No difference in outcomes based on class of overused medication (opioid/barbiturate vs others)
- Outcomes similar across clinic types: headache specialty, general neurology, and primary care
- Adverse events uncommon and similar between groups (~17% in each group; 63/361 vs 62/359 per eTable), with 2% serious AEs per group
- Only 44% of this severely affected population was using preventive medication at enrollment, highlighting undertreatment
- Complete-case data available for 81% of participants in each group (294/361 switching, 291/359 no switching); withdrawal similar between groups (51 vs 49)
- Results consistent with prior studies (COMOESTAS, Carlsen et al.) showing efficacy of preventive medication with or without withdrawal
Study Design
- Study Type
- Prospective, multicenter, open-label, parallel-group, randomized, pragmatic clinical trial
- Randomization
- Yes
- Blinding
- Open-label, no blinding. Participants, site personnel, and investigators were aware of treatment assignments. However, outcome assessment was based on prospective electronic diary completed by participants
- Sample Size
- 720
- Follow-up
- 12 weeks for the primary and secondary outcome time points (weeks 9-12 and weeks 1-2). AEs were collected through 48 weeks
- Centers
- 34
- Countries
- United States
Primary Outcome
Definition: Frequency of days with moderate to severe headache during weeks 9-12 after randomization (primary) and during weeks 1-2 after randomization (secondary primary endpoint tested only if non-inferiority was demonstrated for weeks 9-12). A moderate-severe headache day was defined as a day during which a headache lasted ≥2 hours and at any time peaked with at least moderate intensity. Data collected via daily electronic headache diaries. Non-inferiority margin: 1.5 days per 28 days. Primary analysis used multiple imputation and included all 720 randomized participants at all time points; available-case analysis also reported
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| Not applicable — two active treatment strategies compared | Multiple imputation (primary, n=361 switching vs 359 no switching): Weeks 9-12 (per 28 d) 9.3 vs 9.1; Weeks 1-2 (per 14 d) 6.6 vs 6.4. Available case: Weeks 9-12 (per 28 d) 9.3 (SD 7.2, n=294) vs 9.1 (SD 6.8, n=291); Weeks 1-2 (per 14 d) 6.6 (SD 3.7, n=313) vs 6.4 (SD 3.6, n=308) | - (Multiple imputation — Weeks 9-12: difference 0.2, 95% CI -0.9 to 1.3; Weeks 1-2: difference 0.2, 95% CI -0.4 to 0.8. Available case — Weeks 9-12: difference 0.2 (SD 7.0), 95% CI -1.0 to 1.3; Weeks 1-2: difference 0.2 (SD 3.6), 95% CI -0.4 to 0.7) | Multiple imputation — Weeks 9-12 p=0.71 (non-inferiority met); Weeks 1-2 p=0.55. Available case — Weeks 9-12 p=0.75; Weeks 1-2 p=0.57 |
Limitations & Criticisms
- Open-label design without blinding, though outcome assessment was objective via prospective electronic diary
- No baseline run-in diary phase; baseline headache frequency determined by participant recall which may be subject to recall bias
- Did not include treatment arm with discontinuation of overused medication without switching to alternative symptomatic medication and without preventive medication (though patient stakeholders strongly recommended against this)
- Primary outcome assessed at 12 weeks; longer-term follow-up needed to assess durability of effects and recidivism rates
- High screen failure rate and selective enrollment: most common reason for non-randomization was unwillingness to be randomized to switching strategy (n=77), limiting generalizability
- Population predominantly female (88%) and White (81%; White non-Hispanic 75%) and from the United States; generalizability to other populations uncertain
- Only 44% were using preventive medication at baseline despite severe disease, highlighting selection of particularly treatment-resistant or undertreated population
- Did not track adherence to preventive medications; cannot determine if adherence differed between groups or contributed to outcomes
- Cannot determine whether continuing overused medication with reduced frequency to ≤2 days/week (rather than switching to different class) would provide similar outcomes
- Cannot determine whether switching to medication in same class with frequency limits would provide similar outcomes
- No comparison to complete withdrawal without any alternative symptomatic medication available
- Pragmatic design allowed flexibility in preventive medication selection; cannot determine comparative effectiveness of specific preventive medications
- Multiple imputation used for missing data (~19% missing primary outcome data per group); assumes missing at random
- Noninferiority margin of 1.5 days per 28 days chosen based on clinical consensus and recent trials; some might consider this too liberal
- Subgroup analyses were exploratory with limited power; findings regarding high baseline medication use and anxiety need confirmation
- No assessment of patient satisfaction with treatment strategy or preference for future treatment
- Greater awareness about medication overuse from trial participation likely contributed to improvements in both groups; magnitude of this effect unknown
- Did not assess specific withdrawal symptoms during early weeks
- Population had very severe baseline disease (22.5 headache days per month); results may not generalize to less severe chronic migraine with medication overuse
- Did not assess cognitive or other potential benefits of reduced medication exposure beyond headache frequency
- Trial conducted 2017-2020; newer preventive options (CGRP monoclonal antibodies, gepants) were not widely used
- No assessment of cost-effectiveness of the two strategies
- Follow-up limited to 12 weeks for primary analysis; optimal duration of preventive treatment and timing of symptomatic medication liberalization unknown
Citation
Neurology 2022;98:e1409-e1421