172 Chapter 8 and 30 ± 3 days for the controls, and were classified as Stage -4 (peak reproductive age) according to the STRAW+10 criteria.18 The groups differed slightly in age, with a mean age of 37 ± 7·1 years in the migraine group and 31 ± 5·8 years in the control group (p = 0·002). However, in a sensitivity analysis excluding women over 40 years of age, the difference was no longer statistically significant (migraine group: 34 ± 5·5 years, n = 20; control group: 31 ± 5·8 years; p = 0·09). Although 9/28 (32%) also had migraine with aura attacks, the vast majority of their attacks were without aura. Consistent with findings from a prior study,4 all perimenstrual attacks recorded during the study period were attacks without aura. Participants with migraine used their regular acute migraine medications during attacks on an average of 5·5 days per month (SD ± 2·8), which included paracetamol, NSAIDs (such as ibuprofen), or triptans. Table 1. Characteristics of the study population. Migraine Control p-value Participants, n 28 18 Age (mean ± SD) Sensitivity analysis subgroup <45 years Sensitivity analysis subgroup <40 years 37 ± 7·1 35 ± 6·2 (n=24) 34 ± 5·5 (n=20) 31 ± 5·8 31 ± 5·8 31 ± 5·8 0·002 0·03 0·09 Age of menarche (mean ± SD) 13 ± 1 13 ± 1 0·99 BMI (mean ± SD) 23·2 ± 3·2 23·2 ± 2·7 0·84 Cycle length in days (mean ± SD) 28 ± 3 30 ± 3 0·22 Migraine with aura, n (%) 9 (32) Monthly migraine days (mean ± SD) 4·8 ± 2·5 Monthly headache days (mean ± SD) 7·5 ± 3·7 Monthly acute medication days (mean ± SD) 5·5 ± 2·8 Missing data Among the 46 participants, four (9%) completed only one month of measurements. Across all 88 months, with seven intended measurement days per month, 187 of 616 (30%) measurements were missing (equally distributed among both groups). Missing data occurred randomly, 158/187 (85%) due to weekends or holidays, 24/187 (13%) due to participants’ unavailability and 5/187 (3%) due to the coordinating investigator’s unavailability. After averaging corresponding days across two months for participants who completed both, the proportion of missing values decreased to 11% (37/322). The proportion of missing values for our primary outcome measure was 13% across days 8, 10, 12, or 14, with a rate of 11% specifically on day 8.
RkJQdWJsaXNoZXIy MjY0ODMw