|Home | About | Journals | Submit | Contact Us | Français|
Does use of commonly used over-the-counter (OTC) pain medication affect reproductive hormones and ovulatory function in premenopausal women?
Few associations were found between analgesic medication use and reproductive hormones, but use during the follicular phase was associated with decreased odds of sporadic anovulation after adjusting for potential confounders.
Analgesic medications are the most commonly used OTC drugs among women, but their potential effects on reproductive function are unclear.
The BioCycle Study was a prospective, observational cohort study (2005–2007) which followed 259 women for one (n = 9) or two (n = 250) menstrual cycles.
Two hundred and fifty-nine healthy, premenopausal women not using hormonal contraception and living in western New York state. Study visits took place at the University at Buffalo.
During study participation, 68% (n = 175) of women indicated OTC analgesic use. Among users, 45% used ibuprofen, 33% acetaminophen, 10% aspirin and 10% naproxen. Analgesic use during the follicular phase was associated with decreased odds of sporadic anovulation after adjusting for age, race, body mass index, perceived stress level and alcohol consumption (OR 0.36 [0.17, 0.75]). Results remained unchanged after controlling for potential confounding by indication by adjusting for ‘healthy’ cycle indicators such as amount of blood loss and menstrual pain during the preceding menstruation. Moreover, luteal progesterone was higher (% difference = 14.0, −1.6–32.1, P = 0.08 adjusted) in cycles with follicular phase analgesic use, but no associations were observed with estradiol, LH or FSH.
Self-report daily diaries are not validated measures of medication usage, which could lead to some classification error of medication use. We were also limited in our evaluation of aspirin and naproxen which were used by few women.
The observed associations between follicular phase analgesic use and higher progesterone and a lower probability of sporadic anovulation indicate that OTC pain medication use is likely not harmful to reproduction function, and certain medications possibly improve ovulatory function.
This work was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health (contract # HHSN275200403394C). The authors have no conflicts of interest to disclose.
Human reproduction is an inefficient, complex process that requires a series of intricately timed events to achieve pregnancy. Ovulation and implantation are critical events that are affected by endogenous and exogenous factors and, for those trying to conceive, maximizing the likelihood of both events can lead to greater reproductive success. It has been hypothesized that aspirin use could improve fertility outcomes, possibly due to increased ovarian and/or endometrial vascular perfusion (Schisterman et al., 2014). Yet, other medications used for pain relief (i.e. analgesics), have been reported to inhibit ovulatory function in human and animal studies (Priddy et al., 1990; Miyazaki et al., 1991; Kranzfelder et al., 1992; Athanasiou et al., 1996).
Analgesics are among the most commonly used medications and proper use is considered safe and effective (Hancock et al., 1992; Eggen, 1993; Furu et al., 1997; Kaufman et al., 2002). According to a 2006 US survey of ambulatory adults, OTC analgesics are the most frequently used individual products among U.S. Food and Drug Administration-regulated medications, with 17–23% of the population using such medications in a given week (Kaufman et al., 2002). Furthermore, women reported greater OTC analgesic use than men (Kaufman et al., 2002; Koushede et al., 2011) but analgesic use across the menstrual cycle and the effects of OTC analgesics on hormones and ovulatory function in premenopausal women have not been investigated. Given that OTC analgesics are readily available, commonly used to treat dysmenorrhea, and little is understood about their potential effects on reproductive function, we sought to investigate both the acute and chronic effects of daily-measured OTC analgesic use on reproductive hormones and ovulatory function in healthy, eumenorrheic, premenopausal women with carefully timed, repeated measures of blood hormone concentrations across the menstrual cycle.
The BioCycle Study (2005–2007) included 259 healthy, regularly menstruating women in Western New York State, 18–44 years of age. In this prospective cohort study, women participated for one (n = 9 women) or two (n = 250) menstrual cycles. Exclusion criteria included current use of oral contraceptives, vitamin/mineral supplements, or certain prescription medications; pregnancy, breastfeeding, or trying to conceive in the preceding 6 months; self-reported body mass index (BMI) of <18 or >35 kg/m2 at screening; laparoscopy confirmed endometriosis; and diagnosis of chronic conditions, including menstrual or ovulatory disorders. The University at Buffalo Health Sciences Institutional Review Board (IRB) approved the study and served as the IRB designated by the National Institutes of Health under a reliance agreement. All participants provided written informed consent. Further details of the study design and inclusion and exclusion criteria were previously described (Wactawski-Wende et al., 2009).
Participants provided fasting blood samples up to eight times per cycle during the following menstrual cycle phases: menses; mid and late follicular phase; luteinizing hormone (LH) surge; expected ovulation; and the early, mid, and late luteal phase. Visits were timed using fertility monitors (Clearblue Easy Fertility Monitor™; Inverness Medical, Waltham, Massachusetts) (Howards et al., 2009). Nearly all participants (94%) completed ≥7 clinic visits (including blood collection) per cycle, and 100% completed ≥5 clinic visits per cycle.
Biospecimen protocols were designed to minimize variability (Wactawski-Wende et al., 2009). All samples were processed and frozen at −80°C within 90 min of phlebotomy, and analytes were measured in participant-specific batches within a single run. Estradiol, LH, follicle-stimulating hormone (FSH) and progesterone were measured in fasting serum samples using solid-phase competitive chemiluminescent enzymatic immunoassays (DPC Immulite 2000 analyzer, Siemens Medical Solutions Diagnostics, Deerfield, IL, USA) at the Kaleida Health Center for Laboratory Medicine (Buffalo, NY, USA). The inter-assay coefficients of variation for these tests reported by the laboratory were ≤10% for estradiol, ≤5% for LH and FSH, ≤14% for progesterone.
Sporadic anovulatory cycles were defined as cycles with peak serum progesterone concentrations ≤5 ng/ml and no observed serum LH peak during the mid or late luteal phase visit to ensure that progesterone was assessed during the luteal phase (Lynch et al., 2014).
Participants recorded daily medication intake (type [e.g. name of medication], dose, frequency per day). Eighty-nine percent of the participants completed >75% of their daily diaries. Each medication reported in the daily diary was grouped into one of six major categories based on indication: pain, allergy, cold and cough, gastrointestinal, antibiotics, musculoskeletal and central nervous system. Each medication was subsequently identified via the Micromedex online database or manufacturer's website and further categorized by its primary active ingredient (e.g. acetaminophen, ibuprofen, aspirin, etc.) and dose. Combination medications were assigned to a category based on the principal ingredient in the product. This analysis focuses on the active ingredients in the category of pain (i.e. analgesic) medications. Total dosage of medication each day was determined for each woman. If the total doses of medications reported were greater than the maximum recommended daily dose as defined by the label, this was defined as an overdose.
At study enrollment, age, race, smoking, reproductive history and perceived stress were obtained by using questionnaires (Wactawski-Wende et al., 2009). Perceived stress level was also determined at baseline using the Cohen Perceived Stress Scale (PSS) and was subsequently categorized as: low stress (<14), moderate stress (14–27) and high stress (28–42) (Cohen et al., 1983). Participants also prospectively recorded daily stress levels (not stressful , a little stressful , very stressful ) over the course of the two menstrual cycles. Alcohol intake was also recorded daily and was averaged over the period of observation as follows: low (0–0.5 drinks/day), moderate (0.5–1 drinks/day) or high (≥ 1 drinks/ day). Information about menstrual blood loss was assessed using a detailed menstrual flow questionnaire (Wyatt et al., 2001) where participants characterized their blood loss as low, medium or heavy (Dasharathy et al., 2012). Previous week menstrual pain was assessed at each study visit using a questionnaire that captured the severity (none, mild, moderate and severe) of the following symptoms: swelling of hands or feet, breast tenderness or fullness, lower abdominal cramping, generalized aches and pains, lower backache, and headache. Each cycle was then assigned a menstrual pain score, calculated by taking the average of pain scores from assessments during menses and mid-follicular phase.
Descriptive statistics for continuous and categorical covariates were compared between users and non-users of any OTC analgesic, and for each of the four most commonly reported analgesic active ingredients: acetaminophen, aspirin, ibuprofen, and naproxen. Chi-square or Fisher exact tests were used to evaluate significance for categorical variables and Student's t-test for continuous variables.
Timing of medication use during the menstrual cycle was determined using a standardized 28-day cycle to evaluate in what phase the majority of analgesic use occurred. 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., 2011). If the cycle was classified as anovulatory, cycle day 14 was assigned as the estimated day of expected ovulation for comparison purposes. Average frequency of analgesic use was calculated and compared between early cycle (standardized Day 1 [start of menses] to standardized Day 5 [average bleeding length]), which included menses), mid cycle (standardized Day 6 to standardized Day 13 [day of LH surge]) and late cycle (standardized Day 14 [estimated day of ovulation] to standardized Day 28 [day prior to start of next menses]) using linear mixed models to account for repeated cycles within women. Pair-wise comparisons were made between early, mid and late cycle phases using the Tukey method to account for multiple comparisons.
All hormones, as well as blood pressure, pulse, cycle length, menses length and blood loss, were log-transformed for normality. Linear mixed models were used to estimate the associations between follicular phase (start of menses to estimated day of ovulation) analgesic use (yes/no) and estradiol and FSH concentrations across the menstrual cycle, as well as mid-cycle LH, and luteal progesterone concentrations, corresponding to the cycle phases with the most hormonal variability. We also assessed the effects of time-varying analgesic use on reproductive hormone concentrations (estradiol and FSH across the cycle, mid-cycle LH, and luteal progesterone) by averaging reported daily analgesic use for the 5 days before each clinic visit. Time-varying estimates were calculated using weighted models to control for factors such as reproductive hormones which could be both causes and consequences of short-term medication use (Robins et al., 2000). We constructed inverse probability weights using concurrent estradiol, LH, FSH, progesterone, blood loss and menstrual pain variables. Generalized linear models were used to estimate the effect of follicular phase analgesic use on the odds of sporadic anovulation while accounting for multiple cycles per woman.
Models were adjusted for potential confounders including age (continuous), race (white/black/other), BMI (continuous), smoking status (nonsmoker or current smoker), perceived stress (low/medium/high) and alcohol intake (low/moderate/high) (note that in the time-varying models perceived stress and alcohol intake were handled as time-varying confounders). Additional models were adjusted for ‘healthy cycle indicators’: blood loss during menstruation and menstrual pain score. These potential confounders were considered indicators of both analgesic medication use and ovulation. We hypothesized that greater blood loss during menstruation may reflect greater prior-cycle endometrial growth and therefore increase the likelihood of ovulation (Dasharathy et al., 2012). Similarly, cycles with more menstrual pain, a potential indicator for analgesic use, may also be associated with ovulatory status (Dawood, 1985). All analyses were carried out using SAS version 9.3 (SAS Institute, Inc., Cary, NC, USA).
Participants were on average relatively young (mean age, 27.3 years), of healthy weight (mean BMI, 24.1), educated (87% had post-secondary education), and primarily non-smokers (96%) (Table I). Compared with the non-users, women who had taken any OTC analgesic during the study period were older (P = 0.04), and were more commonly white (P = 0.02), with reported high category of alcohol consumption (P < 0.01) and moderate stress level (P = 0.03, Table I).
More than half (68%) of the study participants reported any OTC analgesic during the course of the study. The most commonly used OTC analgesic active ingredients were ibuprofen (45% of women, 513 person-days reported), acetaminophen (33% of women, 227 person-days reported), naproxen (10% of women, 83 person-days reported) and aspirin (10% of women, 71 person-days reported). Of those who reported consuming these active ingredients, the median number of days (intraquartile range) consumed was 3 (1, 6) days for ibuprofen, 2 (1, 3) days for acetaminophen, 2 (1, 5) days for naproxen and 3 (1, 4) days for aspirin. The maximum recommended daily dose, as defined by the label, of each OTC analgesic evaluated in this study was 1200 mg for ibuprofen, 4000 mg of acetaminophen, 660 mg of naproxen and 4000 mg of aspirin. Consumption over the maximum recommended daily dose of an active ingredient was reported on 65 days for ibuprofen (including four occurrences over the prescription dose level of 3200 mg/day), 15 days for acetaminophen and 12 days for naproxen. There were no occurrences of consuming over the recommended daily dose in aspirin in the 71 days it was consumed. A total of 54 women, during 57 cycles, reported taking more than one analgesic. The most common combination was acetaminophen and ibuprofen, which accounted for 21 (38.9%) of the women in the study with use of multiple medications. Women reporting use of multiple analgesic active ingredients contributed to each active ingredient category.
Over-the-counter analgesics were taken primarily during menstruation, with a steep decline after menstruation (Fig. 1). Analgesic users had significantly greater blood loss during menstruation (P = 0.025), higher average luteal progesterone (P = 0.017) and heavier menstrual flow (P = 0.0048) (Table II) than non-users.
Of the 509 cycles studied, 42 (8.3%) were anovulatory and 31 women contributed to these 42 cycles. Based on unadjusted results, OTC analgesic users had a lower percentage of anovulatory cycles compared with non-users (4 versus 14%, P = 0.009). Follicular phase OTC analgesic was associated with a decreased odds of an anovulatory cycle (OR [95% CI] = 0.32 [0.16, 0.65]). This relationship remained after adjusting for age, race, BMI, smoking, perceived stress, and alcohol intake (OR = 0.36 [0.17, 0.75]), and upon adjusting for potential confounding-by-indication variables (blood loss amount and menstrual pain score) (OR = 0.33 [0.14, 0.76]; Table III).
OTC analgesic use during the follicular phase of the menstrual cycle was associated with higher luteal progesterone concentrations (% difference = 14.0 [−1.6, 32.1]; P = 0.08 adjusted), but analgesic use was not associated with differences in estradiol, LH or FSH concentrations (Table IV). Use of naproxen during the follicular phase was associated with lower periovulatory LH whereas no other medication type was associated with LH. Interestingly, when naproxen users were excluded from the any analgesic use analysis as a sensitivity analysis, we observed a significant increase in periovulatory LH concentrations. When analgesic use was modeled as a time-varying variable to assess short-term effects of recent medication use, we observed lower estradiol (% difference = −15.5 [−20.7, −9.9]; P = < 0.001) and higher FSH (% difference = 10.0 [4.8, 15.5]; P = 0.001) among women who used analgesics during the 5 days prior to the blood draw compared with women who did not use analgesics during this time period (Table V). Mid-cycle LH and luteal progesterone were not significantly associated with time-varying analgesic use. Results were similar after adjustment for concurrent reproductive hormones using marginal structural models.
Over-the-counter analgesic use was highest during menses and was associated with reduced odds of having an anovulatory menstrual cycle among healthy women with regular cycles and not on oral contraceptives. The women who took analgesics during the follicular phase of their menstrual cycle also had significantly higher luteal progesterone than those who did not, but LH, FSH and estradiol levels were not significantly different between users and non-users. However, we did observe that analgesics taken during the 5 days prior to blood sample collection were associated with lower estradiol and higher FSH concentrations. These findings support that OTC analgesic use at the reported doses and frequency of use does not adversely affect reproductive function in normally cycling women who are not using oral contraception, and that analgesic use may be associated with ovulatory function. Moreover, a protective effect of aspirin or other OTC analgesics on ovulation lends support to previous studies reporting improved fertility outcomes associated with aspirin use (Empson et al., 2002; Farquharson et al., 2002; Schisterman et al., 2014).
Our findings are not inconsistent with the limited existing literature regarding OTC analgesic use and ovulation and reproductive hormones. Though animal studies consistently have shown that non-steroidal anti-inflammatory drugs (NSAIDs) are associated with inhibition of ovulation, studies among women are less clear (Gaytan et al., 2006). In a randomized crossover trial of ibuprofen use (800 mg, three times per day for 10 days beginning during the follicular phase), no associations with reproductive hormones were observed, though significant delays in the timing of ovulation were found, supporting the theory that ovulation may occur through inflammatory-related processes that lead to the targeted rupture of a follicle and that anti-inflammatory agents can interfere with that process (Uhler et al., 2001). Though this previous study benefited by assessing ovulation by ultrasound, the study utilized high doses of ibuprofen for 10 consecutive days among a small group of women (n = 12), which are important distinctions from the present observational study. Similarly, a retrospective study of over 1800 natural IVF cycles reported short-term low-dose NSAID use diminished the rate of unwanted premature ovulations (Kawachiya et al., 2012). It is also important to note that the present study reveals the vast majority of analgesic use in healthy young women occurs during the early follicular phase, highlighting the relevance of studying the effects of this common timing of exposure. Inconsistencies in the literature regarding the effects of NSAID use on ovulatory function may relate to the timing of exposure, particularly when comparing observational and intervention studies, or due to differences in the populations being studied.
With regard to hormone levels, previous work indicates that circulating levels of hormones (e.g. progesterone, LH, estradiol, FSH) are unchanged in the presence of analgesic use (Bauer et al., 2013) even though ovulation was delayed (Uhler et al., 2001). However, findings in other studies regarding estrogen levels are conflicting, reporting both higher and lower levels of estrogen with OTC analgesic use (Cramer et al., 1998; Gates et al., 2010).
Though the specific mechanisms by which OTC analgesics may affect ovulation are incompletely understood, it is plausible that increased blood flow via increased ovarian and/or endometrial vascular perfusion, associated with aspirin treatment may play an important role in improving fertility outcomes (Rubinstein et al., 1999). Indeed, studies have demonstrated interdependence of blood flow to reproductive organs and reproductive hormones which may have subsequent effects on ovulation (Battaglia et al., 1990), and aspirin treatment has been shown to increase uterine and ovarian blood flow velocity (Rubinstein et al., 1999). Aspirin and other analgesics are known to inhibit two forms of the cyclooxygenase (Cox) enzyme, Cox-1 and Cox-2, which can lead to increased blood flow. Specifically, these enzymes catalyze the conversion of arachidonic acid to eicosanoids, causing vasodilation and also decreased platelet aggregation. At low doses (typically 70–150 mg), aspirin effectively inhibits platelet production of Thromboxane A2 (TXA2) with little effect on endothelial Prostacyclin (PGI2), resulting in a net increase in the PGI2:TXA2 ratio to decrease thrombosis and increase blood flow (FitzGerald et al., 1983; Patrono et al., 2001; Vane and Botting, 2003). Such mechanisms may have contributed to the protective effect of analgesic use on ovulation observed here. Indeed, any effects of analgesic use on ovulatory function are unlikely to be directly due to altered gonadotrophin or steroid hormone concentrations, as we observed only short-term associations between recent OTC analgesic use and lower estradiol, and higher FSH concentrations, but no sustained impact on hormonal patterns. Further studies are needed to determine whether analgesics, and which analgesic types in particular, may be working through hormonal or inflammatory pathways to influence ovulatory function.
Our results of improved ovulatory function among women reporting follicular phase OTC analgesic use were robust to adjustment by various potential confounding factors. As previously stated, several characteristics (age, BMI, race, stress and alcohol) were considered potential confounders because of their known or possible associations with ovulation, reproductive hormones, and analgesic use. Where appropriate, we took the time-varying nature of stress and alcohol consumption into account in our adjusted models. Additionally, we attempted to address potential confounding by indication. Specifically, it is possible that women who have a more painful menses are also more likely to regularly ovulate and use OTC analgesics. To attempt to address this concern, measures of blood loss and menstrual pain were included as cycle-specific markers of a robust/painful menses, a proxy for this healthy cycle characteristic which may be associated with both subsequent analgesic use and subsequent ovulation. Interestingly, the primary findings of a protective effect on ovulation and differences in hormones remained, even after adjusting for these potential confounders. While we were able to account for these healthy cycle indicators, we cannot rule out the possibility of residual confounding.
Strengths of this study include the comprehensive observation and monitoring of a large number of women through two menstrual cycles. No previous study has assessed daily analgesic use and multiple longitudinal measures of reproductive hormone levels and ovulatory status throughout more than one menstrual cycle. Clinic visits timed with fertility monitors provided significant improvement in individualized monitoring of cycles. By assessing multiple participant characteristics in a standardized manner, the ability to adjust for confounding in the study was increased. Self-report daily diaries are not validated measures of medication usage, which could lead to some classification error of medication use status and level of active ingredient consumption. Most of the analgesic use in the BioCycle Study occurred in the early follicular phase, thereby limiting our ability to assess whether the effects of analgesic use varied depending on its timing throughout the cycle (e.g. effects of periovulatory or chronic use). We were also limited in our evaluation of aspirin and naproxen, which were used by few women. Though there were limited reports of naproxen use, our results suggest that there may be differences in the hormonal effects of naproxen, and when we excluded naproxen users from the analysis of any analgesic use we observed a significant increase in periovulatory LH concentrations. Though the mechanisms of action are very similar between naproxen and ibuprofen (both are non-selective COX inhibitors), naproxen has a much longer duration than ibuprofen (around 12 h compared with 4–6 h) which may partially explain some of these differences.
Overall, we observed that OTC analgesic use in healthy, premenopausal women is frequent and varies over the menstrual cycle with increased use during menses. OTC analgesic use was significantly associated with higher luteal levels of progesterone overall, and use of OTC analgesics in the preceding 5 days was associated with reduced estradiol, and increased FSH compared with women who did not use analgesics during this time period. After adjusting for confounders, we also observed a protective effect on ovulation with OTC analgesic use. Overall, these findings indicate that typical use of analgesic medication in normally cycling, premenopausal women is likely not harmful to reproduction function, and use of certain medications may even improve ovulatory function. However, specific mechanisms of these potentially beneficial effects require further study.
R.A.M. analyzed data and wrote the manuscript; S.L.M. participated in data analysis and interpretation, and helped draft and critically review the manuscript; K.C.S. participated in data analysis and interpretation and helped draft and critically review the manuscript; K.A.A. participated in data analysis and interpretation and helped draft and critically review the manuscript; L.A.S. participated in data analysis and interpretation and helped draft and critically review the manuscript; N.J.P. participated in data analysis and interpretation and critically reviewed the manuscript; A.C.F. participated in data analysis and interpretation and critically reviewed the manuscript; D.M. participated in interpretation of the data and critically reviewed the manuscript; S.M.Z. participated in interpretation of the data and critically reviewed the manuscript; J.W.-W. aided in design of study, directed site data collection, and critically reviewed manuscript; E.F.S. designed study, directed data collection, analysis and interpretation, and critically reviewed manuscript.
This work was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health (contract # HHSN275200403394C).