Discussion
In this study, the use of caffeine was associated with greater vasomotor symptom bother in women aged 40 years or older. Interestingly, we found evidence suggesting that caffeine consumption was associated with a decrease in neurocognitive symptom bother, but only in premenopausal women and not in postmenopausal women. Other menopausal symptom domains (including sleep, bowel/bladder function, sexual function, and general symptoms/other) and the total symptom score did not show an association (improvement or worsening) with the use of caffeine.
This is the largest study, to date, to examine the association between caffeine intake and menopausal symptom bother, particularly vasomotor symptoms. Existing data regarding the relationship of caffeine to vasomotor symptoms are conflicting. The positive association between caffeine intake and vasomotor symptoms observed in our study is consistent with the findings of a prior study by Kandiah and Amend involving 196 postmenopausal women in which caffeine intake predicted hot flash severity, but not hot flash frequency. These results differ, however, from the findings of a recent community-based cross-sectional study involving 243 women in which caffeine intake was found to be a strong negative predictor of vasomotor symptom severity. Pimentaetal theorized that neurocognitive benefits of caffeine may have resulted in a decrease in vasomotor symptoms or that increased levels of estrone, which have been associated with high caffeine intake, could explain the observed effect.
Published data have reported inconsistent associations between caffeine intake and estradiol (E2) levels. Several mechanisms have been proposed for the effects of caffeine on E2 levels. E2 and caffeine are both metabolized by the hepatic enzyme CYP1A2; polymorphisms in this enzyme have been linked to variable E2 levels and clearance of caffeine. Some studies reported an inverse association between caffeine intake and E2 levels in premenopausal women, whereas others reported positive or null associations. London et al reported an inverse association between caffeine and E2 levels in 325 perimenopausal women aged between 50 and 60 years. However, a reduction in E2 levels with caffeine intake was not seen in postmenopausal women in a study by Kotsopoulos et al, possibly owing to the sensitivity of the assays or to the already low circulating levels of E2 in postmenopausal women. Another possible mechanism by which caffeine may affect E2 levels is through inhibition of aromatase, the enzyme that converts androgen into estrogen. Furthermore, caffeine may alter circulating levels of sex steroids through a hepatic effect on sex hormoneYbinding globulin (SHBG). Several studies have shown a positive association between caffeine intake and SHBG levels, although effects on E2 levels in these same studies were mixed.
Caffeine may affect other sex steroid hormone levels. Caffeine intake was inversely associated with bioavailable testosterone and positively associated with SHBG and estrone levels in a study of postmenopausal women by Ferrini and Barrett-Connor. Similarly, in a study by Kotsopoulos et al, increasing caffeine intake was positively associated with SHBG levels, and there was a trend toward decreasing free testosterone levels in postmenopausal women.
How researchers should interpret the finding of reduced neurocognitive symptom bother with caffeine use in premenopausal women is unclear. The finding could have occurred by chance, and it is important to note that the individual symptom domains, including the neurocognitive domain, have not been validated. However, caffeine is known to have beneficial effects on arousal, mood, and attention, which may provide a theoretical basis for the finding. Why this association was limited to premenopausal women is also unclear, although this may be related to the observation that premenopausal women demonstrated greater baseline neurocognitive complaints than did postmenopausal women. Future prospective studies gathering more detailed information on caffeine consumption by women in different reproductive stages would be helpful.
Although our study provides valuable information about the possible association between caffeine and menopausal symptom bother in a large group of premenopausal and postmenopausal women, there are some important limitations, most notably related to the subjective reporting of menopausal symptoms on a survey instrument used for clinical assessment. The study was not prospective, creating the possibility of recall bias. Underreporting of caffeine use is a possibility and might confound the findings. The survey instrument did not allow us to obtain detailed information about the type and amount of caffeinated beverages consumed, which could potentially modulate the effects of caffeine. Furthermore, we were not able to control for the use of hormone therapy, as details on hormone therapy use were not collected for the purposes of this study. Finally, because the study was conducted at a subspecialty menopause clinic, our results may not be generalizable to other groups of women.