PRIMA Trial: Percutaneous PFO Closure vs Medical Therapy in Migraine With Aura (Mattle et al., 2016)
In patients with migraine with aura, does percutaneous PFO closure with the Amplatzer occluder reduce monthly migraine days compared with medical therapy alone? The trial was unblinded, with no sham arm.
Mattle HP et al. (European Heart Journal 2016;37(26):2029-2036) · doi:10.1093/eurheartj/ehw027 · 107 patients
Population
Included
- Migraine with aura
- Patent foramen ovale
Verifier-confirmed inclusion
- Migraine with aura
- Patent foramen ovale
Primary Outcome
Reduction in monthly migraine days, months 9 to 12 versus a 3-month baseline
PRIMARY ENDPOINT NOT MET (P=0.17). This is a continuous outcome measured in days per month, not a responder rate and not a percentage. The figures below are days.
Amplatzer PFO closure
2.9
fewer migraine days per month, vs its own 3-month baseline
Medical therapy
1.7
fewer migraine days per month, vs its own 3-month baseline
Between-arm comparison
P = 0.17
Not significant. Do not subtract the two figures above: each is a reduction measured against that arm's own baseline, and the trial's between-arm comparison was model-based, so the difference of the two is not the trial's estimate. The publication reports no confidence interval we could verify, and no NNT is computed. The comparison the trial actually made is the P value.
Read both figures against the design. PRIMA was unblinded against a self-reported diary endpoint, was stopped early for slow enrolment, and the device was implanted in 41 of the 53 patients assigned to closure (77%), with closure adjudicated at 6 months in 35 (66%).
Study Arms
- Agent
- Amplatzer PFO occluder
- Route
- Percutaneous transcatheter device implantation
- Frequency
- Single procedure
Delivered incompletely. Of the 53 patients randomized to this arm, the device was implanted in 41 (77%) and successful closure was adjudicated at 6 months in 35 (66%).
No sham procedure. The trial was unblinded, so patients in both arms knew their assignment while recording the headache diary that generated the primary endpoint.
Trial Design
Type
- Multicenter randomized controlled trial, UNBLINDED, with no sham arm
- Percutaneous PFO closure with the Amplatzer device vs medical therapy
- Primary endpoint: reduction in monthly migraine days over months 9 to 12 against a 3-month baseline
- Primary endpoint NOT met (reduction of 2.9 days vs 1.7 days, P=0.17)
- Terminated early for slow enrolment
- Incomplete delivery: device implanted in 41 of 53 assigned to closure (77%), successful closure adjudicated at 6 months in 35 (66%)
Timeline
Primary comparison at months 9 to 12 against a 3-month baseline; closure adjudicated at 6 months; published European Heart Journal 2016;37(26):2029-2036
N
107
Enrollment
107 patients randomized, 53 assigned to closure (NCT00505570). Amplatzer PFO Occluder. Open-label with no sham procedure. Stopped early for slow enrolment. Published European Heart Journal 2016.
ClinicalTrials.gov
NCT00505570Bedside Pearl
PRIMA (Eur Heart J 2016) MISSED its primary endpoint: monthly migraine days fell 2.9 with Amplatzer closure vs 1.7 with medical therapy over months 9 to 12, P=0.17. Overall attacks were not significantly reduced either (P=0.09). The aura-specific endpoints and the 38% vs 15% responder rate are secondary. The trial was unblinded with no sham, stopped early for slow enrolment, and of 53 randomized to closure only 41 (77%) got a device and 35 (66%) had closure adjudicated at 6 months. No NNT.