The publisher's final edited version of this article is available at Horm Behav See other articles in PMC that cite the published article. Women were not attempting pregnancy nor using hormonal contraceptives. History of ever having been sexually active was assessed at baseline and frequency of sexual activity, defined as vaginal-penile intercourse, was self-reported daily throughout the study. Sporadic anovulation was identified using peak progesterone concentration.
Linear mixed models were used to estimate associations between sexual activity and reproductive hormone concentrations and generalized linear models were used to estimate associations with sporadic anovulation. Models were adjusted for age, race, body mass index, perceived stress, and alcohol consumption and accounted for repeated measures within women.
Findings from our study suggest that ever having been sexually active is associated with improved reproductive function, even after controlling for factors such as age. However, given the potential evolutionary benefits of sexual activity influencing ovulatory function and consequent pregnancy success Wilcox et al. Female- versus male-initiated sexual activity has been shown to be greater around the time of ovulation Adams et al.
A recent review concluded that although the relationship between sexual activity and menstrual cycle function has been studied, several conflicting results and methodological differences make it difficult to draw definitive conclusions Brown et al. Previous studies have evaluated sexual activity patterns across the menstrual cycle, but have been limited in their examination of sexual activity and ovulatory function, have not included a comparison of reproductive function between sexually active women and women reporting no history of sexual activity, nor have evaluated the effect of reproductive function during one cycle on sexual activity in a subsequent cycle.
Therefore, we investigated the association between sexual activity and reproduction function, examining both the effect of sexual intercourse on reproductive function and the effect of reproductive function on sexual intercourse using longitudinally collected data.
Our hypotheses were that sexually active women would have higher reproductive hormone levels and be less likely to experience anovulatory cycles compared with sexually inactive women and that reproductive hormones would be associated with sexual activity patterns. We investigated these hypotheses in a cohort of healthy premenopausal women, both with or without a history of sexual activity, who were not attempting pregnancy nor using hormonal contraceptives.
Materials and Methods Study Population The BioCycle Study was a prospective cohort study that included healthy, regularly menstruating women aged 18 to 44 years from western New York State during — and followed them for up to two menstrual cycles. Details of the study population, materials and methods have been previously described Wactawski-Wende et al.
Briefly, exclusion criteria included use of oral contraceptives within the past 3 months; a history of pregnancy, breastfeeding, or attempting a pregnancy within the past 6 months; and any recent history of infection or diagnosis of a chronic medical condition, including menstrual and ovulatory disorders, or psychiatric condition, including premenstrual dysphoric disorder. All participants provided written informed consent.
The main findings of the study concerning reproductive hormones and oxidative stress have been previously published Schisterman et al.
Fasting blood samples were scheduled to be collected in the morning between 7 and 8: Blood collection and handling protocols were designed to minimize variability Wactawski-Wende et al. All hormone measurements were log-transformed for normality before statistical analysis and then transformed by exponentiation for table display. In addition, LH and progesterone measurements were restricted in the analysis to mid-cycle three days around the LH surge and luteal phase early, mid and late , respectively, as these are the phases with the greatest biological variance for these hormones.
For certain analyses, progesterone measurements from mid-cycle visits were also analyzed for comparison. These cycles were considered to reflect sporadic rather than chronic anovulation, as study participants were healthy women without reported gynecological or menstrual disorders. Sexual Activity and Covariate Assessment Participant characteristics, such as race, age, and smoking status, were assessed during a baseline visit via questionnaire. Trained study staff measured height and weight, from which BMI was calculated.
In addition, prior sexual history was ascertained by the question: Among sexually active women, contraceptive use history was obtained. Alcohol intake was also assessed via the daily diary and was averaged over the study period and subsequently categorized as: The daily diary also captured information on medication use, hours of sleep, and minutes of vigorous physical activity. Statistical Analysis Baseline characteristics, sexual history, and contraceptive use history were compared among participants grouped into four sexual activity categories: In addition, pair-wise comparisons were performed between sexual activity categories on mean reproductive hormone concentrations, with the Tukey method used to account for multiple comparisons.
Variation in reported daily sexual activity across the menstrual cycle among women sexually active during the study was assessed using linear mixed models to account for repeated measures within women. Days were aligned in relation to the day of ovulation, which was estimated based on dates and levels of LH peak from the fertility monitor compared with the observed LH maximum value in serum and the first day of progesterone rise Mumford et al.
If the cycle was classified as anovulatory, cycle day 14 was assigned as the estimated day of ovulation for comparison purposes. Pair-wise comparisons were made between cycle days using the Tukey method. In addition, average frequency of sexual activity was compared between Mid-Cycle 7 days prior through 2 days after estimated date of ovulation and Early Cycle 8 or more days prior to ovulation, including menses and Late Cycle 3 or more days after ovulation , respectively.
Linear mixed models were also used to estimate the association between sexual activity category and reproductive hormone concentrations. A first-order autoregressive moving-average correlation structure was specified for the correlation between hormone measurements within women. For these analyses, the four level sexual activity category was examined as well as a collapsed two level variable sexually active vs. In addition, we restricted the analyses to sexually active women to examine the association between average frequency of sexual activity and hormone concentrations.
Generalized linear models were used to estimate odds ratios OR for the association between categorized sexual activity both four level and two level categories and the likelihood of experiencing an anovulatory cycle during the study, while accounting for the repeated measurements within women.
These models were also run after restricting the analyses to sexually active women. In terms of the direction of effect, we took advantage of the longitudinal nature of our data to examine both the effect of sexual intercourse on reproductive function and, conversely, the effect of reproductive function on sexual intercourse.
For the effect of sexual intercourse on reproductive function, we used linear mixed models described above to evaluate the association between previous-day sexual intercourse any vs. We also examined sexual intercourse prior to the day of expected ovulation any vs.
For the effect of reproductive function on sexual intercourse, we evaluated associations between previous-day hormone levels and report of next-day sexual intercourse any vs.
In addition, we examined anovulation in the first study cycle in relation to average frequency of sexual intercourse in the second study cycle. And lastly, we examined the association between hormone concentrations and same-day sexual intercourse. As blood draws were scheduled first thing in the morning, this analysis likely reflects the effect of reproductive function on sexual intercourse because sexual intercourse reported for that day presumably occurred later in the day or evening.
Where appropriate we also adjusted for the effect of time-varying confounders using inverse probability weighting Robins et al. All analyses were carried out using SAS version 9.
Mid-cycle progesterone did not differ among sexual activity categories: