We examined the relationships between caffeine, coffee, and tea intake with circulating sex hormones, prolactin and SHBG levels in a large cross-sectional study. Among premenopausal women, caffeine and coffee intake were inversely associated with luteal levels of total and free estradiol. Furthermore, caffeine intake was positively associated with luteal progesterone levels. We also observed a positive association between tea intake and follicular and free estradiol levels. There was a positive association of caffeine and coffee intake with SHBG levels among postmenopausal women. Decaffeinated coffee was inversely associated with DHEAS levels in both groups.
The results from three earlier studies of these associations in premenopausal women have been inconsistent8, 26, 27
. Lucero et al.
reported higher early follicular estradiol levels with increasing daily caffeine and coffee consumption (n=498)8
. Conversely, among 50 premenopausal women, green tea, but not coffee or total caffeine intake, was inversely correlated with follicular estradiol27
. London et al.
reported a significant inverse correlation between caffeine intake and free estradiol among 325 perimenopausal women aged 50–60 years 26
. In our study, premenopausal women with the highest versus lowest quartile of caffeine and coffee intake had 12–15% lower luteal total and free estradiol levels, while follicular estradiol and free estradiol levels were 16–19% higher among women in the highest versus lowest quartile of tea intake. Differences between studies may be due to small sample sizes or poor timing of blood collection during the menstrual cycle in prior studies. With our method of blood collection, we were able to accurately calculate the date of the menstrual cycle for women with timed samples.
Because of the similar associations of coffee and caffeine with luteal estrogen levels, it is probable that caffeine is the component influencing estrogen metabolism. Since there was suggestive evidence for higher testosterone levels with higher intakes of caffeine and coffee, caffeine may be inhibiting CYP19
, or aromatase, the key enzyme mediating the conversion of androgens to estrogens24, 28
. Nevertheless, it is unclear why this would not also influence follicular estrogen levels or sex hormone levels in postmenopausal women where aromatase plays a more critical role in dictating estrogen and testosterone levels. We observed a significant trend for higher circulating progesterone with higher caffeine but not coffee intake. To our knowledge, there are no prior reports of caffeine and progesterone levels in humans; however, one study in male rats showed increased plasma progesterone with intraperitoneal injections of caffeine29
Since we did not observe any effect of coffee or caffeine on follicular estrogen levels, the positive association with tea could be attributed to another component rather than caffeine. Our results require further exploration and should be interpreted with caution given that there are no prior assessments of this association and the number of women with daily tea consumption was small (24% consumed ≥1 cup/day). Nagata et al.
reported an inverse correlation between follicular estradiol levels and green but not black tea among premenopausal women27
. Another study reported that tea intake was not associated with estradiol and SHBG levels8
. We were not able to evaluate the effect of green tea since our FFQ asked about non-herbal tea intake, which consists primarily of black tea in the United States.
Although androgens and estrogens are well-established risk factors for postmenopausal breast cancer30
, the role of these hormones in premenopausal breast cancer is not fully understood. Our findings of lower luteal estrogen levels with higher caffeine intake represents one potential protective mechanism for premenopausal hormonally-dependent cancers. Of the two studies examining estrogens and breast cancer risk in premenopausal women, one reported no association 31
, and the other reported an increased risk of premenopausal breast cancer only with follicular, but not luteal, estradiol levels14
. The elevated levels of follicular estrogens with tea intake suggests that tea may increase the risk of hormone-related cancers, although to our knowledge no epidemiologic data have reported an association between tea intake and premenopausal breast cancer risk. The higher progesterone levels with caffeine intake requires further exploration particularly given the inconsistency in the data regarding a role for this hormone in breast cancer development14, 31
and the possible inverse relationship with ovarian cancer6
. Furthermore, our findings do not explain the increased risk of premenopausal ovarian cancer with caffeine intake we and others have reported5, 32, 33
Among postmenopausal, but not premenopausal, women, high intakes of caffeine or caffeinated coffee were associated with a 13% higher SHBG levels versus low intakes. Four previous studies have reported a similar positive association with caffeine and/or coffee intake and levels of SHBG8, 26, 27, 34
; however, only one study was limited to postmenopausal women34
. Since SHBG is the major carrier of estrogen and testosterone, we expected to see a concomitant inverse association with free levels of these two hormones. We did not observe a strong effect; however, there was suggestive evidence for decreasing free testosterone levels with increasing caffeine intake; testosterone preferentially binds SHBG versus estradiol35
. Similarly, Ferrini et al.
reported an inverse association between caffeine intake and bioavailable testosterone levels, but a positive association with estrone levels34
. The lack of an association with estrogen levels may be attributed to the sensitivity our assays and consequently, inability to detect small changes in circulating levels particularly given the low estrogen, and to a lesser extent androgen, production among postmenopausal women.
Higher SHBG levels have been associated with a lower risk of postmenopausal breast cancer among both users and non- users of PMH7, 12, 36
. Only one study has evaluated the relationship between SHBG concentrations and ovarian cancer risk, observing an inverse association among women diagnosed before age 5537
. In general, the protective effect of caffeine and/or coffee on breast and ovarian cancer appears to be strongest among postmenopausal women4, 5
. Hormonal changes in postmenopausal women include a substantial decrease in estradiol and estrone levels, but only a small change in androgen synthesis by the ovaries and adrenal glands24
. This suggests that the inverse association between caffeine intake and risk of postmenopausal ovarian and breast cancer may be mediated by its effect on hepatic production of SHBG and subsequent reduction in free testosterone. Indeed, high endogenous testosterone levels have been clearly implicated in the etiology of postmenopausal breast cancer7,12
, although a positive association with risk of ovarian cancer is less clear38
Since adipose tissue is an important source of sex hormones in postmenopausal women, we also evaluated whether BMI modified any of these associations. The inverse association between caffeine and SHBG appeared strongest among overweight/obese women. Although not statistically significant, this possible interaction warrants further evaluation since adiposity has consistently been associated with lower SHBG levels.
We did not observe associations between caffeine, coffee, or tea with levels of androgens, estrogens, or prolactin in the postmenopausal women. The null association with prolactin is of interest given that this hormone has been associated with an increased risk of premenopausal10
and postmenopausal breast cancer15
. However, in other studies of postmenopausal women, black tea has been associated with higher plasma levels of estrone and prolactin39, 40
The inverse association between decaffeinated coffee and DHEAS requires further confirmation given the small percentage of women that regularly consumed decaffeinated coffee in our cohort (≥ 2 cups/day: 12% for premenopausal, 31% for postmenopausal women) and more importantly, the biological effect of decaffeinated coffee on adrenal androgens levels has never been explored.
This is the largest study to examine the relationship between endogenous androgens, estrogens, prolactin and SHBG with caffeine, coffee and tea intake. We were able to obtain timed blood samples from a large number of premenopausal women to accurately assess hormone concentrations during both the luteal and follicular phases and we limited our analysis of postmenopausal women to those not using PMH. The major limitation of this study is the inability to establish a temporal relationship between the exposure (i.e. caffeine) and hormones levels, although it is unlikely that endogenous sex hormones would influence coffee consumption. Also, validation studies have shown a high correlation between self-reported coffee intake on the FFQ compared to that from a 28-day diet record (ρ=0.75) 41
In summary, our data suggest that caffeine-mediated changes in circulating levels of luteal estrogens and SHBG represent possible mechanisms by which coffee or caffeine may be associated with pre- and postmenopausal hormonally-related malignancies, respectively. Since hormones are clearly implicated in the etiology of many female cancers, further evaluation of how caffeine-mediated alterations in sex hormones and binding protein levels affect the risk of breast or ovarian cancer are warranted.