Abstract
Objective:
To prospectively evaluate the association between female sleep patterns, shift work, and fecundability.
Design:
Pregnancy Study Online (PRESTO) is a web-based preconception cohort study
Setting:
North American couples during the preconception period
Patients:
Female participants aged 21-45 years attempting pregnancy
Interventions:
Not applicable
Main Outcome Measures:
At baseline, participants reported their average sleep duration per 24-hour period in the previous month, the frequency of trouble sleeping within the last 2 weeks (as measured by the Major Depression Inventory), and shift work patterns. To ascertain pregnancy status, follow-up questionnaires were completed every 8 weeks for up to 12 months or until conception. Analyses were restricted to 6,873 women attempting pregnancy for ≤6 months at enrollment from June 2013 through September 2018. We used proportional probabilities regression models to estimate fecundability ratios (FRs) and 95% confidence intervals (CIs), adjusting for potential confounders.
Results:
Relative to 8 hours of sleep/day, FRs for <6, 6, 7, and ≥9 hours of sleep/day were 0.89 (CI: 0.75-1.06), 0.95 (CI: 0.86-1.04), 0.99 (CI: 0.92-1.06), and 0.96 (CI: 0.84-1.10), respectively. Compared with no trouble sleeping, FRs for trouble sleeping <50% of the time or trouble sleeping >50% of the time were 0.93 (CI: 0.88-1.00) and 0.87 (CI: 0.79-0.95), respectively. Results were slightly stronger among women with higher depressive symptoms and perceived stress levels. There was no association between shift work and fecundability.
Conclusion:
Trouble sleeping at night was associated with modestly reduced fecundability. A weaker inverse association was observed between shorter sleep duration and fecundability.
Keywords: sleep, fertility, time-to-pregnancy, preconception, cohort studies
Capsule:
We evaluated prospectively the association between sleep patterns and fecundability among pregnancy planners. Trouble sleeping and short sleep duration were modestly associated with reduced fecundability.
INTRODUCTION
In the United States (US), approximately 15% of couples experience infertility, defined as the inability to conceive within 12 months of unprotected intercourse (1). Despite the large financial and emotional costs associated with infertility and its treatments, few modifiable risk factors for infertility have been identified.
The prevalence of sleep deprivation has been increasing in the US. Since 1985, the frequency of adults sleeping fewer than 6 hours per night increased by 31% (2). Nationally representative data from the US and Canada indicate that about 30-32% of premenopausal women sleep fewer than 7 hours per night on average (3, 4); 24% of US women report having problems staying asleep (4). Sleep disturbances are more prevalent among women than men, with women being more likely to have difficulty falling asleep and to have poorer sleep quality than men, potentially due to sleep disturbances often coinciding with fluctuations in reproductive hormones (5). Suboptimal sleep duration has been associated with increased risks of heart disease (6), obesity (7), diabetes (8), and all-cause mortality (9), often displaying U-shaped associations, with 7-8 hours having the lowest risk.
The extent to which sleep influences female fertility has not been well studied. Rodent studies have shown that disruption of circadian rhythms adversely affects fertility, as evidenced by a higher proportion of irregular or absent estrus cycles (10, 11). Less is known about the effect of sleep on fertility among women. In a Taiwanese prospective cohort study, women with medically-diagnosed non-apnea sleep disorders had a 3.7-fold increased risk of infertility compared with women without sleep disorders (12). Other studies have been either cross-sectional or retrospective in design, or have been conducted in selective populations such as couples attending fertility clinics and women with polycystic ovary syndrome (PCOS). A cross-sectional study among women with PCOS reported an association between short sleep duration (<6 hours per night) and abnormal menstrual cycle length, but found no association with androgenic and ovarian measures (13). In an infertility clinic study population, disturbed sleep was associated with diminished ovarian reserve as measured by follicle stimulating hormone (FSH), estradiol, inhibin B, and Mullerian-inhibiting substance, although temporality of the association was unclear (14). In a study of women undergoing in vitro fertilization (IVF), the number of oocytes retrieved was positively associated with actigraph-measured sleep time (15).
Shift work and work at night are often examined to study the effect of disrupted circadian rhythms on reproductive health (11). Shift work is defined as any work that takes place outside of the traditional work hours of 9:00 a.m.- 5:00 p.m., including evening or night (defined as work starting from around 10 p.m. to 2 a.m.), early morning, or rotating shifts (16). In the US, around 12% of employed females work a shift schedule (11). Shift work has been associated with depression (18), cardiovascular disease (17), and breast cancer (11).
Of the four retrospective cohort studies that evaluated shift work and subfecundity among pregnant women or women with a prior pregnancy, three reported a slight positive association between shift work and subfecundity (19, 20) and a fourth reported little association (21). A fifth study, a prospective cohort study of U.S. nurses, reported slightly longer time to pregnancy (TTP) among women who worked rotating shifts (22). Of these five studies, three investigated the association between night work and TTP: two found slightly reduced fecundity among night workers (20, 21) while the other found slightly improved fecundity among night workers (22).
Given the scarcity of research on female sleep patterns and fertility, and the overall trend towards reduced sleep duration among reproductive-aged women nationally, we investigated the effects of duration and quality of sleep on fecundability among female participants in a North American preconception cohort study. We also examined the extent to which shift work was associated with fecundability.
MATERIALS AND METHODS
Study population:
Pregnancy Study Online (PRESTO) is an ongoing web-based preconception cohort study of North American pregnancy planners. Study methods have been described in detail elsewhere (23). Briefly, women aged 21-45 years living in Canada or the US, and not using contraception or fertility treatment are eligible for participation. At baseline, female participants complete a web-based questionnaire on demographics, lifestyle, and medical and reproductive histories. To update pregnancy status over time, female participants complete follow-up questionnaires every 8 weeks or until reported conception, initiation of fertility treatment, cessation of pregnancy attempt, withdrawal, loss to follow-up, or 12 months, whichever came first. Male participation was optional and involves completion of an online baseline questionnaire similar to the female questionnaire. This study was approved by the Institutional Review Board at Boston University Medical Center; online informed consent was obtained from all participants.
From June 2013 until September 2018, 8,772 eligible women completed the baseline questionnaire. We excluded 102 women whose baseline date of last menstrual period (LMP) was >6 months before study entry, and 35 women with insufficient or missing LMP data. We additionally excluded 1,762 women attempting conception for >6 cycles at study entry, for a final analytic sample of 6,873. Of these 6,873 women, 54% invited their male partners to participate and 47% of these men completed the survey, yielding 1,743 women with available partner data.
Assessment of sleep exposure
On the female baseline questionnaire, participants were asked “In the last month, on average how many hours of sleep did you get each night?”, with response options of <5, 5, 6, 7, 8, 9, or 10 or more hours. Average sleep duration during the last month was categorized as follows: <6, 6, 7, 8 (reference), or ≥9 hours. To obtain data on sleep quality, a question on the Major Depression Inventory (MDI) asked, within the past two weeks: “Have you had trouble sleeping at night?” with response options of “all of the time,” “most of the time,” “slightly more than half of the time,” “slightly less than half of the time,” “some of the time,” or “at no time.” This variable was categorized into 3 groups: “at no time” (reference), “less than half the time”, or “more than half the time.”
To ascertain shift work, we asked "In the past month, did you work rotating shifts (hours that shift in a predictable way from day-to-day, week-to-week, or month-to-month)?" We assessed night work by asking "How many times in the past month did you work at night (defined as working sometime between midnight and 2 AM)?" Night work and shift work were both categorized as “none,” “1-4 times per month,” or “≥5 times per month.” Shift work analyses were restricted to currently employed participants.
Assessment of TTP
We estimated TTP using data from the baseline and follow-up questionnaires. On the baseline questionnaire, women reported their LMP, usual menstrual cycle length, and the number of cycles they attempted pregnancy at study entry. On each follow-up questionnaire, participants reported their most recent LMP date and whether they had conceived since the prior questionnaire. Among women who reported irregular cycles, cycle length was estimated based on the LMP date at baseline and the consecutive LMP dates reported at follow-up. TTP was estimated based on the total discrete cycles at risk, calculated as: cycles of attempt at study entry + [(LMP date from most recent follow-up questionnaire - date of baseline questionnaire completion)/usual cycle length] +1.
Assessment of covariates
On the baseline questionnaire, we collected covariate data, including age, height, weight, race/ethnicity, household income, employment status, education, marital status, smoking, marijuana use, alcohol and caffeine consumption, intercourse frequency, contraception history, reproductive history, stress via the 10-item version of the perceived stress scale (PSS-10) (24), history of anxiety, depressive symptoms via the MDI, and a history of other physician-diagnosed medical conditions (e.g., major depressive disorder and anxiety/panic disorder). Body mass index (BMI) was calculated as weight (kilograms) divided by height (meters) squared.
Data analysis
We used proportional probabilities regression models to estimate fecundability ratios (FRs) and 95% confidence intervals (CIs) for the association between selected sleep variables and fecundability. The FR is the average per-cycle probability of conception within each exposure category in comparison to the reference group. A FR of <1 indicates a longer TTP among exposed relative to unexposed women. For example, a FR of 0.85 indicates that, in any given cycle of attempt time, exposed women had a 15% reduced probability of conception than unexposed women. The proportional probabilities model includes indicator variables for each cycle at risk, thereby accounting for the baseline decline in fecundability over time (25). The Anderson-Gill data structure outputs a single menstrual cycle per observation and accounts for left truncation from delayed entry into the study. To allow for non-linear associations, we fit restricted cubic splines, transforming the categorical sleep duration variable into a continuous variable and assigning values of 4 to “<5 hours” and 10 to “10 or more hours.”
Potential confounders were determined a priori based on the literature and an assessment of a directed acyclic graph. Final models were adjusted for female age (<25, 25-29, 30-34, ≥35 years), race/ethnicity (White non-Hispanic vs other race/ethnicity), BMI (<18.5, 18.5-24.9, 25-29.9, ≥30 kg/m2), income (<$50,000, $50,000-99,000, ≥$100,000 US dollars/year), education (<12, 12, 13-15, 16, ≥17 years), prior birth (yes vs no), use of hormonal contraceptives as last method of contraception (yes vs no), current smoking (yes vs no), hours working per week (<20, 20-40, ≥41 hours), current unemployment (yes vs no), current caffeine consumption (<100, 100-199, 200-299, ≥300 mg/day), and history of infertility (yes, no, never attempted to conceive). Final models were mutually adjusted for sleep duration and sleep quality. Marijuana use, intercourse frequency, marital status, current levels of depressive symptoms, current levels of perceived stress, history of anxiety or depression, and other medical diagnoses (PCOS, thyroid disease, and diabetes) were assessed as potential confounders but were omitted from final models because they had no appreciable effect on the exposure-outcome association (26).
To assess effect measure modification by comorbid mental health conditions, we stratified by history of anxiety or depression; current depressive symptoms (MDI <15 vs ≥15); and current perceived stress (PSS-10 score: <20 and ≥20). Because we hypothesized that parous women have irregular sleeping schedules, we conducted additional analyses restricted to nulliparous women. Lastly, to address concerns that sleep habits were ascertained only at baseline and that women with longer attempt times at study entry may have changed their behaviors in response to subfertility, we stratified by attempt time at study entry (<3 vs 3-6 cycles). We also stratified shift work analyses by age (<30 vs ≥30 years).
Previous findings from the PRESTO cohort showed inverse associations between short sleep duration and fecundability among males (27). Therefore, we performed sub-analyses restricted to couples with complete male and female data (N=1,743), in which we further adjusted for male sleep duration (<6, 6, 7, 8, ≥9 hours/night).
We used multiple imputation to impute missing data on exposures, covariates, and pregnancy status (28). We generated five imputed datasets using a Markov chain Monte Carlo method with over 200 covariates to predict missing values. Each imputed dataset was analyzed separately and pooled to account for between- and within-imputation variation (29). To reduce selection bias from differential loss to follow-up, we assigned one cycle of follow-up for the women with no follow-up data (N=948) and then imputed their pregnancy status (29). Fewer than 0.5% of women were missing data on sleep duration and trouble sleeping; 4% were missing data on night work. Missingness for covariates ranged from <0.1% (prior pregnancy, caffeine use, and history of anxiety) to 3% (income). There were no missing values for age.
RESULTS
Overall, 6,873 female participants contributed 3,933 pregnancies and 26,339 menstrual cycles of attempt time. Among participants who conceived, 3,657 pregnancies were self-reported or found in birth registries; 276 pregnancies were imputed. Of the participants who did not conceive, 359 were actively contributing follow-up, 168 stopped trying to conceive, 522 initiated fertility treatment, 947 were lost to follow-up, and 944 completed 12 cycles of attempt time. The median follow-up time among all women was 3 cycles (interquartile range: 1-6 cycles). The unadjusted time to pregnancy among women who conceived at the 25th, 50th, and 75th percentiles were 2, 4, and 6 cycles, respectively. Among 1,743 couples with complete questionnaires from both partners, 1,098 couples conceived and contributed 6,930 menstrual cycles of attempt time. At baseline, 6% of participants reported <6 hours of sleep/night and 6% reported ≥9 hours of sleep/night; 33% reported having no trouble sleeping, 47% reported having trouble sleeping less than half the time, and 20% reported having trouble sleeping more than half the time. The cumulative proportion of pregnancies at 12 months for women with trouble sleeping more than half of the time, trouble sleeping less than half of the time, and no trouble was 64%, 70%, and 76%, respectively. The cumulative proportion of pregnancies at 12 months for women who sleep <6, 6, 7, 8, and ≥9 hours was 62%, 67%, 73%, 73%, and 66%, respectively.
At baseline (Table 1), a U-shaped association was observed between sleep duration and PSS-10 score, depression (both physician diagnosis and MDI score), and current smoking status. Those with short and long sleep durations were more likely to have trouble sleeping at night. Individuals at the extremes of sleep duration had lower education and income, worked less per week, and drank less alcohol. Gravidity and parity were inversely associated with sleep duration and positively associated with trouble sleeping. Trouble sleeping was positively associated with BMI, depression diagnosis, PSS-10 score, and current smoking, and inversely associated with income, education, and sleep duration.
Table 1.
Baseline characteristics* of women based on average hours of sleep per 24-hour period and trouble sleeping
| Average sleep duration during past month (hours/24-hour period) | Trouble sleeping at night during past two weeks | |||||||
|---|---|---|---|---|---|---|---|---|
| <6 | 6 | 7 | 8 | ≥9 | No | <50% of time | >50% of time | |
| Number of women (n) | 398 | 1,318 | 2,938 | 1,783 | 436 | 2,260 | 3,220 | 1,393 |
| Female age, years (mean) | 29.9 | 29.9 | 30.0 | 29.8 | 29.7 | 29.9 | 30.0 | 29.8 |
| Female BMI, kg/m2 (mean) | 31.2 | 29.2 | 27.3 | 26.8 | 27.7 | 26.9 | 27.5 | 29.9 |
| Time trying at study entry, cycles (mean) | 2.5 | 2.2 | 2.1 | 1.9 | 2.1 | 2.0 | 2.0 | 2.4 |
| Household income, USD (%) | ||||||||
| <50K | 42.2 | 25.8 | 16.4 | 17.8 | 25.3 | 15.4 | 18.6 | 33.3 |
| 50K-99K | 38.3 | 39.7 | 38.0 | 39.2 | 41.9 | 38.1 | 40.0 | 38.4 |
| ≥100K | 19.5 | 34.5 | 45.7 | 43.0 | 32.9 | 46.5 | 41.4 | 28.4 |
| White, non-Hispanic (%) | 70.9 | 79.2 | 85.5 | 86.8 | 84.2 | 84.5 | 84.6 | 80.9 |
| Married (%) | 82.4 | 87.6 | 92.0 | 92.1 | 89.7 | 92.4 | 91.4 | 85.4 |
| Education≥ college degree (%) | 47.9 | 64.4 | 77.6 | 77.9 | 66.7 | 79.3 | 74.7 | 57.7 |
| Prior birth (%) | 55.5 | 38.8 | 29.6 | 25.0 | 21.4 | 29.8 | 29.5 | 37.3 |
| Prior pregnancy (%) | 72.6 | 56.0 | 48.5 | 44.0 | 45.9 | 47.7 | 48.1 | 57.9 |
| Hormonal contraception as last form of birth control (%) | 39.1 | 39.5 | 39.5 | 38.4 | 35.8 | 38.6 | 39.1 | 38.5 |
| Perceived Stress Score (mean) | 19.3 | 17.3 | 16.0 | 15.1 | 16.2 | 14.1 | 16.1 | 19.9 |
| Ever diagnosed with depression (%) | 33.7 | 26.6 | 22.8 | 22.9 | 31.1 | 18.5 | 23.0 | 37.9 |
| Major Depression Inventory score (mean) | 17.7 | 12.7 | 10.1 | 9.0 | 11.4 | 5.8 | 10.0 | 20.7 |
| Intercourse, times per week (%) | ||||||||
| ≤1 | 37.6 | 41.4 | 39.2 | 37.3 | 36.6 | 38.4 | 39.1 | 37.9 |
| 2-3 | 35.6 | 43.1 | 46.1 | 45.8 | 43.3 | 45.5 | 45.6 | 42.0 |
| ≥4 | 26.8 | 15.5 | 14.7 | 16.9 | 20.1 | 16.2 | 15.3 | 20.1 |
| Work duration, hours per week (%) | ||||||||
| ≤20 | 9.4 | 7.5 | 8.4 | 10.4 | 17.3 | 9.0 | 9.1 | 11.0 |
| 21-40 | 71.0 | 72.2 | 72.7 | 76.0 | 73.5 | 73.8 | 73.0 | 73.0 |
| ≥41 | 22.6 | 23.8 | 22.6 | 18.0 | 15.6 | 20.7 | 21.7 | 20.9 |
| Worked rotating shifts (%) | 23.0 | 18.1 | 15.2 | 13.7 | 17.6 | 14.3 | 15.6 | 19.1 |
| Current smoker (%) | 16.8 | 9.0 | 5.0 | 4.9 | 9.1 | 4.3 | 5.8 | 12.7 |
| No alcohol use (%) | 40.2 | 26.9 | 23.7 | 25.2 | 33.1 | 25.3 | 24.6 | 32.2 |
| Marijuana use (%) | 16.2 | 12.5 | 12.7 | 11.1 | 17.5 | 11.3 | 12.6 | 15.4 |
| No caffeine drinks (%) | 19.3 | 15.3 | 10.8 | 13.8 | 17.0 | 13.2 | 12.5 | 15.1 |
| Trouble sleeping at night (MDI item)(%) | ||||||||
| No trouble | 13.4 | 22.5 | 32.3 | 44.7 | 39.6 | -- | -- | -- |
| <50% of the time | 28.1 | 47.5 | 51.5 | 43.1 | 42.1 | -- | -- | -- |
| >50% of the time | 58.5 | 30.0 | 16.2 | 12.2 | 18.3 | -- | -- | -- |
| Total sleep duration (average hours/24-hour period) (%) | ||||||||
| <7 | -- | -- | -- | -- | -- | 15.4 | 22.9 | 45.4 |
| 7-8 | -- | -- | -- | -- | -- | 77.1 | 71.2 | 48.9 |
| ≥9 | -- | -- | -- | -- | -- | 7.5 | 5.9 | 5.7 |
All characteristics, except for age, are age standardized to cohort at baseline
Relative to 8 hours of sleep, <6 hours of sleep was associated with a slight reduction in fecundability (FR=0.89, 95% CI: 0.75, 1.06) (Table 2). No appreciable associations were observed for the intermediate and highest categories of sleep duration. We observed a dose-response association between trouble sleeping and fecundability: compared with no trouble sleeping, FRs were 0.93 for trouble sleeping less than half of the time (95% CI: 0.88-1.00) and 0.87 for trouble sleeping more than half of the time (95% CI: 0.79-0.95). The restricted cubic spline displaying the association between female sleep duration and fecundability was consistent with the categorical analysis (Supplemental Figure 1). Sub-analyses among couples where sleep information was available on both partners yielded similar results. After additional adjustment for male sleep duration, the effect estimate for short sleep duration among females was attenuated.
Table 2.
Association between sleep duration and trouble sleeping and fecundability
| No. of pregnancies |
No. of cycles |
Unadjusted FR |
95% CI | Adjusted FRa |
95% CI | |||
|---|---|---|---|---|---|---|---|---|
| Total sleep duration (average hours/24-hour period)* | ||||||||
| <6 | 180 | 1,502 | 0.80 | 0.67-0.94 | 0.89 | 0.75-1.06 | ||
| 6 | 699 | 5,049 | 0.89 | 0.81-0.98 | 0.95 | 0.86-1.04 | ||
| 7 | 1,727 | 11,258 | 0.98 | 0.91-1.05 | 0.99 | 0.92-1.06 | ||
| 8 | 1,093 | 6,799 | 1.00 | Reference | 1.00 | Reference | ||
| ≥9 | 234 | 1,731 | 0.91 | 0.80-1.03 | 0.96 | 0.84-1.10 | ||
| Trouble sleeping at night (MDI item)** | ||||||||
| No trouble | 1,411 | 8,411 | 1.00 | Reference | 1.00 | Reference | ||
| <50% of time | 1,853 | 12,516 | 0.91 | 0.86-0.97 | 0.93 | 0.88-1.00 | ||
| <50% of time | 669 | 5,412 | 0.80 | 0.73-0.87 | 0.87 | 0.79-0.95 | ||
| Restricted to women with partner data available | ||||||||
| No. of pregnancies |
No. of cycles |
Unadjusted FR |
95% CI | Adjusted FRa |
95% CI | Adjusted FRb |
95% CI | |
| Total sleep duration (average hours/24-hour period)* | ||||||||
| <6 | 35 | 292 | 0.78 | 0.56-1.08 | 1.00 | 0.71-1.41 | 1.01 | 0.71-1.42 |
| 6 | 172 | 1,304 | 0.80 | 0.68-0.95 | 0.85 | 0.72-1.02 | 0.86 | 0.72-1.03 |
| 7 | 488 | 2,931 | 0.95 | 0.84-1.08 | 0.94 | 0.83-1.07 | 0.94 | 0.83-1.07 |
| 8 | 328 | 1,853 | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| ≥9 | 75 | 550 | 0.82 | 0.65-0.98 | 0.86 | 0.67-1.09 | 0.86 | 0.67-1.09 |
| Trouble sleeping at night (MDI item)** | ||||||||
| No trouble | 417 | 2,304 | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| <50% of time | 517 | 3,336 | 0.89 | 0.79-1.00 | 0.92 | 0.82-1.04 | 0.93 | 0.82-1.04 |
| <50% of time | 164 | 1,290 | 0.76 | 0.64-0.90 | 0.83 | 0.69-0.99 | 0.83 | 0.70-1.00 |
Adjusted for female age, BMI, income, non-Hispanic white, prior birth, prior form of birth control hormonal, current smoker, hours working, history of infertility, unemployment, and caffeine consumption.
Adjusted for covariates in model a with additional adjustments for male sleep per 24-hour period
Additional adjustment for trouble sleeping
Additional adjustment for hours of sleep per 24-hour period
Women with histories of anxiety or depression, MDI scores ≥15, or PSS scores ≥20 had a similar distribution of sleep duration compared with unaffected women. However, women with high MDI (≥15) and PSS (≥20) scores tended to report greater trouble sleeping than those with lower scores. Short sleep duration showed a stronger inverse association among women without a history of anxiety or depression than among women with a history of anxiety or depression, although results were imprecise. For trouble sleeping at night, FRs among those without a history of anxiety or depression were also stronger than among women with a history of these disorders. Stronger effects were observed for trouble sleeping >50% of the time among women with MDI scores ≥15 (compared with MDI scores <15) and among women with PSS scores ≥20 (compared with PSS scores <20).
Similar results were found among nulliparous women for sleep duration and trouble sleeping relative to the full cohort although the FR for the lowest category of sleep duration was attenuated (Supplemental Table 2). Results were slightly stronger among women with longer attempt times (Supplemental Table 3).
When stratifying by trouble sleeping at night, we observed a U-shaped association for hours of sleep/night with those at the extremes having reduced fecundability (Supplemental Table 1). For those with trouble sleeping >50% of the time, those with ≥9 vs 8 hours of sleep had reduced fecundability (FR=0.76, 95% CI: 0.52-1.11).
In comparison with those not working at night or on rotating shifts, there was little association for those who worked night shifts or rotating shifts during the past month (Table 4). We found no evidence of effect measure modification of these associations by age (<30 vs ≥30 years; data not shown).
Table 4.
Association between night work and shift work and fecundability, restricted to employed participants
| No. of pregnancies |
No. of cycles |
Unadjusted | 95% CI | Adjusteda | 95% CI | Adjustedb | 95% CI | |
|---|---|---|---|---|---|---|---|---|
| No night or shift work | 2,623 | 16,966 | 1.00 | Reference | 1.00 | Reference | 1.00 | Reference |
| Night worker only | 233 | 1,650 | 0.93 | 0.82-1.05 | 0.95 | 0.84-1.08 | 0.96 | 0.85-1.09 |
| 1-4 times/month | 142 | 976 | 0.95 | 0.81-1.10 | 0.95 | 0.81-1.10 | 0.94 | 0.80-1.10 |
| ≥5 times/month | 91 | 674 | 0.91 | 0.74-1.10 | 0.97 | 0.80-1.18 | 0.98 | 0.81-1.19 |
| Shift worker only | 344 | 2,568 | 0.93 | 0.84-1.04 | 0.95 | 0.86-1.06 | 0.96 | 0.86-1.07 |
| Night and shift worker | 205 | 1,367 | 0.97 | 0.85-1.11 | 0.98 | 0.86-1.13 | 0.99 | 0.87-1.14 |
| 1-4 times/month | 92 | 590 | 0.97 | 0.79-1.20 | 0.99 | 0.80-1.23 | 0.99 | 0.80-1.22 |
| ≥5 times/month | 113 | 777 | 0.97 | 0.81-1.16 | 0.98 | 0.82-1.18 | 0.99 | 0.82-1.19 |
Adjusted for female age, BMI, income, non-Hispanic white; prior birth, prior form of birth control hormonal, current smoker, hours working per week, history of infertility, and caffeine consumption.
Adjusted for covariates in model a with additional adjustments for trouble sleeping and hours of sleep per 24-hour period
DISCUSSION
In this North American preconception cohort study of female pregnancy planners, trouble sleeping at night was associated with modestly reduced fecundability in a dose-response manner. Short sleep duration (<6 hours/night) was weakly associated with reduced fecundability. Little association was observed between shift work and fecundability. Associations were similar when stratifying by factors associated with sleep disturbances such as physician-diagnosed depression and anxiety, depressive symptoms, and perceived stress levels, although the findings for trouble sleeping at night were slightly stronger among those with high depressive symptoms and perceived stress levels. For trouble sleeping, findings persisted among those trying <3 cycles at study entry, among whom reverse causation is less likely (i.e., subfertility causing disturbances in sleep). When restricting to individuals with no trouble sleeping, we observed a stronger albeit imprecise association with short sleep duration. Results for short female sleep duration were attenuated with additional adjustment for male sleep duration, which was associated with reduced fecundity in this cohort (27).
To our knowledge, no previous study has examined the role of sleep patterns on fecundability among women with no known fertility problems. The present study builds on research conducted among populations with known fertility impairments such as PCOS or diagnosed infertility. While these studies differ methodologically, our findings agree with cross-sectional studies showing associations between infertility and impaired sleep, including hours of sleep, quality of sleep, and sleep disorders (12-15, 30).
We found no reduction in fecundability for women who worked in rotating shifts or at night. Our findings disagree with a meta-analysis of previous cohort studies, which found that shift workers had a slightly higher odds of subfecundity (OR=1.12, 95% CI: 0.86-1.44) compared with women working normal hours or unemployed women (31).
In addition to being the first prospective cohort study to examine the association between sleep duration, sleep quality, and TTP, PRESTO included a large, geographically diverse cohort of women. All couples were enrolled during the preconception period, with over two-thirds of couples enrolling during their first three cycles of attempt time. Additionally, data were collected on a wide range of confounders, including those from male partners.
Limitations include potential misclassification of sleep variables due to self-reported data ascertainment. However, most studies have relied on self-reported measures of sleep. Self-reported sleep duration has been moderately correlated with actigraph-measured sleep (32, 33), with correlations ranging from 0.43 (34) to 0.47 (33) and within-subject year-to-year correlation of sleep duration, as measured by wrist actigraphy, is high (r=0.76). In PRESTO, sleep measures were reported before the occurrence of subfertility; thus, misclassification is likely to be non-differential, attenuating FRs for extreme categories toward the null. Although the distribution of sleep duration was compatible with national data (4), PRESTO ascertained sleep data only at baseline, averaging over the month before the baseline survey was collected. Thus, we were unable to assess whether sleep patterns changed over time. Additionally, we had small numbers in the extreme categories of sleep duration resulting in imprecise estimates. Although many factors were considered and adjusted for in final models, there is potential for unmeasured confounding. Lastly, we were unable to evaluate the effects of sleep among individuals with very high MDI scores and levels of stress, owing to a small number of women within these categories.
In animal studies, circadian rhythms assist with regulation of the estrus cycle, luteinizing hormone (LH) surge, and ovulation (35); comparably, throughout a woman’s reproductive lifetime, fluctuations in ovarian hormones may affect sleep quality via these rhythms (36). Potential biologic mechanisms underlying the association between suboptimal sleep patterns and reduced fecundity involve dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, which is associated with infertility (37). One hypothesis is that the HPA axis may act independently on fecundity via stress levels. A second hypothesis is that sleep dysregulation influences fecundity through chronic HPA axis activation, leading to the alteration of hormones important for reproduction such as thyroid stimulating hormone, LH, FSH, testosterone, anti-Mullerian hormone, and prolactin. The interaction between stress and sleep may provide a biologic rationale for slightly stronger results for trouble sleeping among women with higher MDI scores and higher levels of perceived stress. Stress may temporarily raise melatonin levels, an important sleep regulating hormone. Melatonin has been found to have mixed effects on fecundity, affecting reproductive hormones differently depending on the cycle phase in which sleep disruption occurs (38, 39).
CONCLUSION
We observed modestly reduced fecundability among North American women who reported disturbed sleep as measured by the Major Depression Inventory and a weak inverse association between shorter sleep durations and fecundability. Little association was seen between shift work and fecundability. These results persisted after accounting for mental health status and other factors associated with sleep disturbances.
Supplementary Material
Supplemental Figure 1. Restricted cubic spline of the association between total sleep duration and fecundability among 6,873 female PRESTO participants, 2013-2018. Reference value for spline is 8 hours of sleep per 24-hour period, with knot points at 6, 7, 8, and 9 hours per 24-hour period. Adjusted for female age, BMI, income, non-Hispanic white, prior birth, prior form of birth control hormonal, current smoker, hours working per week, history of infertility, unemployment, and caffeine consumption.
Table 3.
Association between sleep duration and trouble sleeping and fecundability, stratified by history of and current mental health measures
| No. of pregnancies |
No. of cycles |
Unadjusted FR |
95% CI | Adjusted FRa |
95% CI | |
|---|---|---|---|---|---|---|
| No history of anxiety or depression | ||||||
| Total sleep duration (average hours/24-hour period)* | ||||||
| <6 | 102 | 929 | 0.76 | 0.62-0.94 | 0.87 | 0.69-1.11 |
| 6 | 447 | 3,314 | 0.88 | 0.78-0.99 | 0.90 | 0.79-1.01 |
| 7 | 1,142 | 7,514 | 0.99 | 0.90-1.08 | 0.97 | 0.88-1.06 |
| 8 | 744 | 4,559 | 1.00 | Reference | 1.00 | Reference |
| ≥9 | 131 | 943 | 0.94 | 0.79-1.11 | 0.97 | 0.80-1.17 |
| Trouble sleeping at night (MDI item)** | ||||||
| None | 1,015 | 6,093 | 1.00 | Reference | 1.00 | Reference |
| <50% of time | 1,222 | 8,418 | 0.90 | 0.83-0.97 | 0.92 | 0.85-0.99 |
| >50% of time | 329 | 2,748 | 0.79 | 0.70-0.89 | 0.86 | 0.76-0.97 |
| History of anxiety or depression | ||||||
| Total sleep duration (average hours/24-hour period)* | ||||||
| <6 | 78 | 573 | 0.83 | 0.64-1.08 | 0.98 | 0.74-1.30 |
| 6 | 252 | 1,735 | 0.90 | 0.77-1.05 | 0.97 | 0.82-1.14 |
| 7 | 585 | 3,744 | 0.96 | 0.85-1.09 | 0.97 | 0.86-1.10 |
| 8 | 349 | 2,240 | 1.00 | Reference | 1.00 | Reference |
| ≥9 | 103 | 788 | 0.87 | 0.71-1.06 | 0.96 | 0.78-1.18 |
| Trouble sleeping at night (MDI item)** | ||||||
| None | 396 | 2,318 | 1.00 | Reference | 1.00 | Reference |
| <50% of time | 631 | 4,098 | 0.95 | 0.85-1.06 | 0.97 | 0.86-1.09 |
| >50% of time | 340 | 2,664 | 0.81 | 0.71-0.92 | 0.88 | 0.76-1.02 |
| Current Major Depression Inventory score <15 | ||||||
| Total sleep duration (average hours/24-hour period)* | ||||||
| <6 | 94 | 700 | 0.90 | 0.74-1.11 | 0.99 | 0.80-1.22 |
| 6 | 483 | 3,381 | 0.90 | 0.81-0.99 | 0.94 | 0.84-1.04 |
| 7 | 1,410 | 8,860 | 0.99 | 0.92-1.07 | 0.99 | 0.92-1.07 |
| 8 | 916 | 5,534 | 1.00 | Reference | 1.00 | Reference |
| ≥9 | 180 | 1,250 | 0.93 | 0.81-1.08 | 0.98 | 0.85-1.14 |
| Trouble sleeping at night (MDI item)** | ||||||
| None | 1,322 | 7,925 | 1.00 | Reference | 1.00 | Reference |
| <50% of time | 1,552 | 10,201 | 0.94 | 0.87-1.00 | 0.96 | 0.90-1.02 |
| >50% of time | 209 | 1,599 | 0.85 | 0.74-0.97 | 0.91 | 0.80-1.04 |
| Current Major Depression Inventory score≥15 | ||||||
| Total sleep duration (average hours/24-hour period)* | ||||||
| <6 | 86 | 802 | 0.75 | 0.58-0.98 | 0.83 | 0.63-1.11 |
| 6 | 216 | 1,668 | 0.91 | 0.75-1.11 | 0.96 | 0.79-1.17 |
| 7 | 317 | 2,398 | 0.96 | 0.81-1.14 | 0.99 | 0.83-1.17 |
| 8 | 177 | 1,265 | 1.00 | Reference | 1.00 | Reference |
| ≥9 | 54 | 481 | 0.85 | 0.64-1.13 | 0.88 | 0.66-1.17 |
| Trouble sleeping at night (MDI item)** | ||||||
| None | 89 | 486 | 1.00 | Reference | 1.00 | Reference |
| <50% of time | 301 | 2,315 | 0.78 | 0.63-0.97 | 0.77 | 0.62-0.96 |
| >50% of time | 460 | 3,813 | 0.73 | 0.59-0.89 | 0.72 | 0.58-0.90 |
| Perceived stress score < 20 | ||||||
| Total sleep duration (average hours/24-hour period)* | ||||||
| <6 | 114 | 901 | 0.83 | 0.68-1.02 | 0.97 | 0.80-1.19 |
| 6 | 516 | 3,683 | 0.89 | 0.80-0.99 | 0.94 | 0.85-1.05 |
| 7 | 1,421 | 9,003 | 0.99 | 0.92-1.07 | 0.99 | 0.92-1.07 |
| 8 | 930 | 5,672 | 1.00 | Reference | 1.00 | Reference |
| ≥9 | 179 | 1,319 | 0.89 | 0.77-1.03 | 0.94 | 0.81-1.09 |
| Trouble sleeping at night (MDI item)** | ||||||
| None | 1,254 | 7,461 | 1.00 | Reference | 1.00 | Reference |
| <50% of time | 1,521 | 10,210 | 0.91 | 0.85-0.98 | 0.92 | 0.82-1.02 |
| >50% of time | 385 | 2,907 | 0.84 | 0.76-0.94 | 0.94 | 0.88-1.00 |
| Perceived stress score ≥ 20 | ||||||
| Total sleep duration (average hours/24-hour period)* | ||||||
| <6 | 66 | 601 | 0.77 | 0.58-1.01 | 0.87 | 0.65-1.17 |
| 6 | 183 | 1,366 | 0.93 | 0.76-1.13 | 1.01 | 0.82-1.24 |
| 7 | 306 | 2,255 | 0.95 | 0.80-1.14 | 0.98 | 0.82-1.17 |
| 8 | 163 | 1,127 | 1.00 | Reference | 1.00 | Reference |
| ≥9 | 55 | 412 | 1.00 | 0.75-1.34 | 1.08 | 0.80-1.45 |
| Trouble sleeping at night (MDI item)** | ||||||
| None | 157 | 950 | 1.00 | Reference | 1.00 | Reference |
| <50% of time | 332 | 2,306 | 0.92 | 0.77-1.09 | 0.89 | 0.75-1.06 |
| >50% of time | 284 | 2,505 | 0.74 | 0.62-0.89 | 0.73 | 0.60-0.89 |
Adjusted for female age, BMI, income, non-Hispanic white, prior birth, prior form of birth control hormonal, current smoker, trouble sleeping, history of infertility, unemployment, and caffeine consumption.
Additional adjustment for trouble sleeping
Additional adjustment for hours of sleep per 24-hour period
Acknowledgements:
This research was supported by NICHD (R21-HD072326, R01- HD086742). We acknowledge the contributions of PRESTO participants and staff. We thank Mr. Michael Bairos for technical support in developing the study’s web-based infrastructure. All authors made significant contributions to the manuscript in accordance with the Vancouver group guidelines. None of the authors are affiliated with any organization having direct or indirect financial interest in the subject matter discussed in the transcript.
Footnotes
Conflict of Interest Disclosure: Drs. Hatch, Wesselink, Rothman, Wise and Ms. Willis report grants from National Institutes of Health during the conduct of the study. Dr. Mikkelsen has nothing to disclose.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Thoma ME, McLain AC, Louis JF, King RB, Trumble AC, Sundaram R, et al. Prevalence of infertility in the United States as estimated by the current duration approach and a traditional constructed approach. Fertil Steril 2013; 99:1324–31.e1321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ford ES, Cunningham TJ, Croft JB. Trends in Self-Reported Sleep Duration among US Adults from 1985 to 2012. Sleep 2015; 38:829–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chaput J-P, Wong S, Michaud I. Duration and quality of sleep among Canadians aged 18 to 79.In: Health Reports 2017. [PubMed] [Google Scholar]
- 4.Vahratian A. Sleep Duration and Quality Among Women Aged 40–59, by Menopausal Status. In: NCHS Data Brief 2017. [PubMed] [Google Scholar]
- 5.Abbott SM, Attarian H, Zee PC. Sleep disorders in perinatal women. Best Pract Res Clin Obstet Gynaecol 2014; 28:159–68. [DOI] [PubMed] [Google Scholar]
- 6.Cappuccio FP, Cooper D, D'Elia L, Strazzullo P, Miller MA. Sleep duration predicts cardiovascular outcomes: a systematic review and meta-analysis of prospective studies. Eur Heart J 2011; 32:1484–92. [DOI] [PubMed] [Google Scholar]
- 7.Knutson KL. Does inadequate sleep play a role in vulnerability to obesity? Am J Hum Biol 2012; 24:361–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Knutson KL, Van Cauter E. Associations between sleep loss and increased risk of obesity and diabetes. Ann N Y Acad Sci 2008; 1129:287–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cappuccio FP, D'Elia L, Strazzullo P, Miller MA. Sleep duration and all-cause mortality: a systematic review and meta-analysis of prospective studies. Sleep 2010; 33:585–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kennaway DJ, Boden MJ, Varcoe TJ. Circadian rhythms and fertility. Mol Cell Endocrinol 2012; 349:56–61. [DOI] [PubMed] [Google Scholar]
- 11.Mahoney MM. Shift work, jet lag, and female reproduction. Int J Endocrinol 2010; 2010:813764. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wang ID, Liu YL, Peng CK, Chung CH, Chang SY, Tsao CH, et al. Non-Apnea Sleep Disorder Increases the Risk of Subsequent Female Infertility-A Nationwide Population-Based Cohort Study. Sleep 2018; 41. [DOI] [PubMed] [Google Scholar]
- 13.Lim AJ, Huang Z, Chua SE, Kramer MS, Yong EL. Sleep Duration, Exercise, Shift Work and Polycystic Ovarian Syndrome-Related Outcomes in a Healthy Population: A Cross-Sectional Study. PLoS One 2016; 11:e0167048. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Pal L, Bevilacqua K, Zeitlian G, Shu J, Santoro N. Implications of diminished ovarian reserve (DOR) extend well beyond reproductive concerns. Menopause 2008; 15:1086–94. [DOI] [PubMed] [Google Scholar]
- 15.Goldstein CA, Lanham MS, Smith YR, O'Brien LM. Sleep in women undergoing in vitro fertilization: a pilot study. Sleep Med 2017; 32:105–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Grosswald B The Effects of Shiftwork on Family Satisfaction. Families in Society: The Journal of Contemporary Social Services 2004: 413–23. [Google Scholar]
- 17.Costa G Shift work and health: current problems and preventive actions. Saf Health Work 2010; 1:112–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Cole RJ, Loving RT, Kripke DF. Psychiatric aspects of shiftwork. Occup Med 1990; 5:301–14. [PubMed] [Google Scholar]
- 19.Bisanti L, Olsen J, Basso O, Thonneau P, Karmaus W. Shift work and subfecundity: a European multicenter study. European Study Group on Infertility and Subfecundity. J Occup Environ Med 1996; 38:352–58. [DOI] [PubMed] [Google Scholar]
- 20.Tuntiseranee P, Olsen J, Geater A, Kor-anantakul O. Are long working hours and shiftwork risk factors for subfecundity? A study among couples from southern Thailand. Occup Environ Med 1998; 55:99–105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Zhu JL, Hjollund NH, Boggild H, Olsen J. Shift work and subfecundity: a causal link or an artefact? Occup Environ Med 2003; 60:E12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Gaskins AJ, Rich-Edwards JW, Lawson CC, Schernhammer ES, Missmer SA, Chavarro JE. Work schedule and physical factors in relation to fecundity in nurses. Occup Environ Med 2015; 72:777–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Wise LA, Rothman KJ, Mikkelsen EM, Stanford JB, Wesselink AK, McKinnon C, et al. Design and Conduct of an Internet-Based Preconception Cohort Study in North America: Pregnancy Study Online. Paediatr Perinat Epidemiol 2015; 29:360–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J Health Soc Behav 1983; 24:385–96. [PubMed] [Google Scholar]
- 25.Weinberg CR, Wilcox AJ, Baird DD. Reduced fecundability in women with prenatal exposure to cigarette smoking. Am J Epidemiol 1989; 129:1072–78. [DOI] [PubMed] [Google Scholar]
- 26.Greenland S, Daniel R, Pearce N. Outcome modelling strategies in epidemiology: traditional methods and basic alternatives. Int J Epidemiol 2016; 45:565–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Wise LA, Rothman KJ, Wesselink AK, Mikkelsen EM, Soresen HT, McKinnon CJ, et al. Male sleep duration and fecundability in a North American preconception cohort study. Fertil Steril 2018; 109:453–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Zhou XH, Eckert GJ, Tierney WM. Multiple imputation in public health research. Stat Med 2001; 20:1541–49. [DOI] [PubMed] [Google Scholar]
- 29.Sterne JA, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, et al. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 2009; 338:b2393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lin JL, Lin YH, Chueh KH. Somatic symptoms, psychological distress and sleep disturbance among infertile women with intrauterine insemination treatment. J Clin Nurs 2014; 23:1677–84. [DOI] [PubMed] [Google Scholar]
- 31.Stocker LJ, Macklon NS, Cheong YC, Bewley SJ. Influence of shift work on early reproductive outcomes: a systematic review and meta-analysis. Obstet Gynecol 2014; 124:99–110. [DOI] [PubMed] [Google Scholar]
- 32.Lemola S, Ledermann T, Friedman EM. Variability of sleep duration is related to subjective sleep quality and subjective well-being: an actigraphy study. PLoS One 2013; 8:e71292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Lauderdale DS, Knutson KL, Yan LL, Liu K, Rathouz PJ. Self-reported and measured sleep duration: how similar are they? Epidemiology 2008; 19:838–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Cespedes EM, Hu FB, Redline S, Rosner B, Alcantara C, Cai J, et al. Comparison of Self-Reported Sleep Duration With Actigraphy: Results From the Hispanic Community Health Study/Study of Latinos Sueño Ancillary Study. Am J Epidemiol 2016; 183:561–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Boden MJ, Varcoe TJ, Kennaway DJ. Circadian regulation of reproduction: from gamete to offspring. Prog Biophys Mol Biol 2013; 113:387–97. [DOI] [PubMed] [Google Scholar]
- 36.Lee KA, Baker FC, Newton KM, Ancoli-Israel S. The Influence of reproductive status and age on women's sleep. J Womens Health (Larchmt) 2008; 17:1209–14. [DOI] [PubMed] [Google Scholar]
- 37.Kloss JD, Perlis ML, Zamzow JA, Culnan EJ, Gracia CR. Sleep, sleep disturbance, and fertility in women. Sleep Med Rev 2015; 22:78–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Tamura H, Takasaki A, Miwa I, Taniguchi K, Maekawa R, Asada H, et al. Oxidative stress impairs oocyte quality and melatonin protects oocytes from free radical damage and improves fertilization rate. J Pineal Res 2008; 44:280–7. [DOI] [PubMed] [Google Scholar]
- 39.Eryilmaz OG, Devran A, Sarikaya E, Aksakal FN, Mollamahmutoğlu L, Cicek N. Melatonin improves the oocyte and the embryo in IVF patients with sleep disturbances, but does not improve the sleeping problems. J Assist Reprod Genet 2011; 28:815–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure 1. Restricted cubic spline of the association between total sleep duration and fecundability among 6,873 female PRESTO participants, 2013-2018. Reference value for spline is 8 hours of sleep per 24-hour period, with knot points at 6, 7, 8, and 9 hours per 24-hour period. Adjusted for female age, BMI, income, non-Hispanic white, prior birth, prior form of birth control hormonal, current smoker, hours working per week, history of infertility, unemployment, and caffeine consumption.