Participants were recruited from the community and outpatient clinics in New South Wales and Queensland, Australia. Participants were eligible if they were 18 years or older; had 2–14 days of migraine per month over the past 3 months and had an office systolic BP (SBP) of ≥120 mm Hg and a diastolic BP (DBP) of ≥75 mm Hg. These thresholds were selected to minimize the risk of symptomatic hypotension in a population not preselected for hypertension and receiving a fixed-dose triple antihypertensive combination. Exclusion criteria were any contraindication to the medications, abnormal creatinine or electrolytes, and cluster headache or any chronic headaches (headache occurring on more than 15 days per month) (see full list of eligibility criteria in Supplementary Table 1).

The study protocol was approved by the Research Ethics and Governance Office of Sydney Local Health District (Protocol No X15-0410, HREC/15/RPAH/554). Informed consent was obtained from participants.

Using a central, computer-based randomization system, participants were randomly assigned to: (1) low-dose BP lowering combination - a single encapsulated pill containing three BP-lowering drugs, at low-dose (telmisartan 20 mg, amlodipine 2.5 mg and indapamide 1.25 mg).; (2) propranolol at a dose of 160 mg/day; or (3) placebo. The placebo capsule appeared identical to, and was of similar weight to, the active drugs. An additional factorial randomization was implemented with participants also allocated to simvastatin 20 mg, low-dose combination cholesterol lowering (rosuvastatin 10 mg plus ezetimibe 10 mg) or placebo (Figure 1). The results for the cholesterol lowering arm will be reported separately. Trial medicines were prepared and packaged at a manufacturing facility licensed with a Certificate of Good Manufacturing Practice by the Therapeutic Goods Administration of Australia.

Recruitment was through general advertisement in hospital clinics, primary care newsletters, health appointment booking sites and a charitable headache organization in Australia. Migraine diagnosis was based on International Classification of Headache Disorders, third edition (ICHD-3) (4). The participants underwent a 4-week screening period, 12 weeks of study medication and another 4 weeks of post study-drug observational period (Figure 1). Visits conducted during the 12 weeks were face-to-face. Post-study follow-up involved a phone call. Biological sex was inquired during the initial study visit. Office BP was measured when the participants were rested in the seated position for 5 min, an appropriate cuff size was selected, and 3 measurements of BP were recorded using an Omron HEM907 monitor or equivalent. The second and third readings were averaged for study analysis. Participants completed paper or electronic based headache diaries at any time a headache occurred after screening. The diary was used to track the number and duration of migraine headache days, acute symptomatic treatment of migraine headache and effectiveness of treatment. All serious adverse events were recorded. Participants and study assessors were blinded to study drug allocation.

For this pilot trial, testing for trial recruitment and data collection procedures; participant acceptability; and SBP, DBP and LDL differences between the groups were mainly assessed.

The primary outcome was the reduction in mean monthly headache days at 12 weeks when compared to baseline. Secondary outcomes were change from baseline in mean monthly hours with migraine; proportion of subjects with at least a 50% reduction from baseline in monthly headache days; change from baseline in monthly headache medication treatment days; change in SBP and DBP; and adverse events (AEs).

All analyses of study outcomes were conducted according to the principle of intention-to-treat. Mean (standard deviation) and number (%) are reported. The feasibility of the study was assessed as the proportion of patients randomized out of those screened. The pilot was not powered for headache outcomes but rather designed to evaluate feasibility, tolerability, and variance estimates to inform the design of a future, adequately powered efficacy trial. As such, observed treatment effects are considered exploratory and hypothesis-generating. The analyses of continuous outcomes such as change in headache frequency and SBP at 12 weeks were performed using ANCOVA including the treatment arm and baseline values as a covariate by reporting mean change (95% confidence interval). Relative risk (95% confidence interval) was reported for comparison of dichotomous variables such as proportion of participants with at least a 50% reduction in monthly headache. Pre-defined subgroup analyses included headache type, baseline BP, sex and age.

Between March 2017 and June 2018, 378 participants were screened, and 30 (8% feasibility) were eligible and randomly allocated to the treatment groups (Figure 2). For the 334 not eligible, 26% had low BP, 23% were unwilling to participate, 16% did not meet episodic migraine criteria and 12% were taking medications that were prohibited during the study. Reasons for unwillingness to participate after initial interest were issues with participating in a drug trial and trial burden. There were 14 participants not randomized after being enrolled because they did not attend randomization visits, were ineligible for randomization due to low BP or deranged blood tests and not completing screening visits. One participant discontinued the study in the first 4 weeks of the trial for personal reasons.

Baseline characteristics are shown on Supplementary Table 2. The mean age (SD) of the participants was 49 (11.5), 83% were female, 90% Caucasians and 10% Asians and 50% worked full time. The mean body mass index was 27 kg/m², the mean (SD) SBP was 131 (12) mmHg and the mean diastolic BP was 87 (8) mmHg. 16% reported a history of depression. Participants had an average of 67 h of migraine per month (median 67, interquartile range 15–80); 47% reported having aura symptoms, all participants experienced photophobia and over 80% were nauseated. Triptans were the most used abortive therapy and 23% of the participants were on prophylactic medications.

The low-dose BP treatment combination had on average 0.6 fewer monthly migraine days with wide confidence intervals (95% confidence interval [CI] −2.9 to 1.8) while in the propranolol group there were −1.1 fewer days (95% CI −3.8 to 1.6) (Supplementary Table 2). The effects on BP compared to placebo at 12 weeks had wide confidence intervals, e.g., −7 (−19 to 5) and −8 (−17 to 0) mm Hg systolic reductions for low-dose combination and propranolol, respectively, (Table 1). There was no significant difference in any other outcomes. There were no significant differences in outcomes in any of the pre-defined subgroups, including by age or sex (Supplementary Table 3).

AEs are shown in Supplementary Table 4. There were no reported serious AEs. The two most common AEs reported between week 4 and week 12, apart from headache, were feeling faint/dizzy and insomnia. The former was seen in 40% participants of all participants but mostly in the low-dose BP lowering combination arm (78%) and the placebo arm (64%). There were a similar percentage of participants reporting insomnia in all groups (between 33 and 55%). There were no dropouts due to adverse events.