170 Chapter 8 control samples were included in each assay batch and external quality was assured via appropriate programs. The intra- and inter-assay coefficients of variation were <3% and <6·2% for FSH, <3·5% and <6·4% for LH, <3% and <5% for testosterone, <8% and <11% for androstenedione, and <4% and <5% for SHBG. For the remaining hormones, inter-assay performance was quantified as follows. For 17-hydroxyprogesterone, the standard deviation (SD) was 0·1 nM, the coefficient of variation (CV) was 2·2%, and the mean concentration was 4·9 nM. DHEA exhibited an SD of 3·4 nM, a CV of 3·3%, and a mean of 103 nM. For DHEA-S, the SD was 0·4 µM, the CV was 4·0%, and the mean was 9·9 µM. Dihydrotestosterone showed an SD of 0·1 nM, a CV of 4·9%, and a mean of 2·1 nM. Outcomes The primary outcome was the peak during the mid-luteal phase (day 8) and associated decline in estradiol (pmol/L) in plasma between women with migraine and healthy controls. The decline was calculated using the difference in plasma levels for each participant between days 8–14. Secondary outcomes included a comparative analysis between women with migraine and healthy controls for all plasma levels of LH (U/L), FSH (U/L), sex hormone-binding globulin (SHBG, nmol/L), 17-hydroxyprogesterone (17-OHP, nmol/L), androstenedione (nmol/L), dehydroepiandrosterone (DHEA, nmol/L), dehydroepiandrosterone sulfate (DHEA-S, nmol/L), dihydrotestosterone (DHT, nmol/L), progesterone (nmol/L), and testosterone (nmol/L) (eFigure 1). Additionally, the decline in progesterone was compared for days 8–14 and the decline in LH for days 1–3. Ratios of hormones were calculated and compared at each time point for the following hormone pairs: LH:FSH, FSH:estradiol, LH:progesterone, testosterone:estradiol, estradiol:progesterone, and progesterone:estradiol. Statistical analysis The sample size was determined based on available data, as the inclusion target of n=33 per group was not reached within the study period. Data cleaning and analysis were conducted using R version 4·4·0. Baseline characteristics were summarized as means with standard deviations or frequencies with proportions. Data was visualized using principle component analysis and scatter plots to identify potential outliers. For skewed data, a log₂ transformation was applied, which was applicable to all hormones. Measurements were standardized relative to the LH surge (defined as day 0) to maintain consistency across individuals. For participants who completed two months of measurements, mean values for each measurement day were calculated.
RkJQdWJsaXNoZXIy MjY0ODMw