Serum caffeine and paraxanthine concentrations and menstrual cycle function: correlations with beverage intakes and associations with race, reproductive hormones, and anovulation in the BioCycle Study

Abstract

Background: Clinicians often recommend limiting caffeine intake while attempting to conceive; however, few studies have evaluated the associations between caffeine exposure and menstrual cycle function, and we are aware of no previous studies assessing biological dose via well-timed serum measurements.

Objectives: We assessed the relation between caffeine and its metabolites and reproductive hormones in a healthy premenopausal cohort and evaluated potential effect modification by race.

Design: Participants (n = 259) were followed for ≤2 menstrual cycles and provided fasting blood specimens ≤8 times/cycle. Linear mixed models were used to estimate associations between serum caffeine biomarkers and geometric mean reproductive hormones, whereas Poisson regression was used to assess risk of sporadic anovulation.

Results: The highest compared with the lowest serum caffeine tertile was associated with lower total testosterone [27.9 ng/dL (95% CI: 26.7, 29.0 ng/dL) compared with 29.1 ng/dL (95% CI: 27.9, 30.3 ng/dL), respectively] and free testosterone [0.178 ng/mL (95% CI: 0.171, 0.185 ng/dL) compared with 0.186 ng/mL (95% CI: 0.179, 0.194 ng/dL), respectively] after adjustment for age, race, percentage of body fat, daily vigorous exercise, perceived stress, depression, dietary factors, and alcohol intake. The highest tertiles compared with the lowest tertiles of caffeine and paraxanthine were also associated with reduced risk of anovulation [adjusted RRs (aRRs): 0.39 (95% CI: 0.18, 0.87) and 0.40 (95% CI: 0.18, 0.87), respectively]. Additional adjustment for self-reported coffee intake did not alter the reproductive hormone findings and only slightly attenuated the results for serum caffeine and paraxanthine and anovulation. Although reductions in the concentrations of total testosterone and free testosterone and decreased risk of anovulation were greatest in Asian women, there was no indication of effect modification by race.

Conclusion: Caffeine intake, irrespective of the beverage source, may be associated with reduced testosterone and improved menstrual cycle function in healthy premenopausal women.

Keywords: 1,7-dimethylxanthine; anovulation; biological markers; caffeine; menstrual cycle.

FIGURE 1

Mean ± SD variations in fasting serum caffeine (A), paraxanthine (B), and theobromine (C) concentrations across the menstrual cycle (n = 259 women). Comparisons were made with the use of linear mixed models to account for repeated measures within women (both across the cycle and between cycles). Pairwise comparisons were made between menses, early follicular, midfollicular, LH/FSH surge, ovulatory, early luteal, midluteal, and late-luteal visits with the use of Tukey’s method to account for multiple comparisons. Values that do not share a common lowercase letter were significantly different at P < 0.05 on the basis of Tukey’s test. Overall, there were ≥3650 measurements (90%) for each reproductive hormone collected from ≤16 clinic visits across 2 menstrual cycles for the 259 women. Note: 1.00 μmol serum caffeine/L = 194.2 ng/mL; 1.00 μmol serum paraxanthine/L = 180.2 ng/mL; and 1.00 μmol serum theobromine/L = 180.2 ng/mL. FSH, follicle-stimulating hormone; LH, luteinizing hormone.

FIGURE 2

Adjusted RRs (95% CIs) of anovulation according to serum caffeine (A), paraxanthine (B), and theobromine (C) tertiles (n = 259 women). Analyses were performed with the use of generalized linear mixed models and adjusted for age, race, percentage of body fat, daily vigorous exercise, perceived stress, depression, Mediterranean diet score, and total energy and alcohol intakes (all continuous except for race, which was categorized as white, black, Asian, or other). A total of 259 women were followed for ≤1 (n = 9) or 2 (n = 250) menstrual cycles. Anovulation was defined as any cycle with a peak progesterone concentration ≤5 ng/mL and no observed serum luteinizing hormone peak at the midluteal or late-luteal phase visits (n = 42 of 509 cycles; 8.3%); 28 women had 1 anovulatory cycle, and 7 women had 2 anovulatory cycles. T, tertile.