Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Apr 1.
Published in final edited form as: J Occup Health Psychol. 2021 Jul 29;27(2):258–265. doi: 10.1037/ocp0000279

The association between work hours, shift work, and job latitude with fecundability: a preconception cohort study.”

Craig James McKinnon 1, Elizabeth Elliott Hatch 2, Olivia R Orta 3, Kenneth J Rothman 4, Michael L Eisenberg 5, Johanna Wefes-Potter 6, Lauren A Wise 7
PMCID: PMC8799775  NIHMSID: NIHMS1729457  PMID: 34323556

Abstract

The role of occupational stress on male fertility is understudied. We examined associations between male occupational stress and fecundability.

We used data from Pregnancy Online Study (PRESTO), a North American preconception cohort study. At baseline (2013–2019), male participants aged ≥21 years completed a baseline questionnaire on employment status, hours worked per week, time of day worked (daytime, evening, nights, and changing or rotating shifts), and job title. We used the O*NET occupational database to rate independence by job title. Female partners were followed via bimonthly follow-up questionnaires for 12 months or until pregnancy. We restricted analyses to 1,818 couples attempting conception for ≤6 cycles at enrollment. We used proportional probabilities regression to estimate fecundability ratios (FRs) and 95% confidence intervals (CIs).

The FR comparing unemployed with employed men was 0.84 (95% CI: 0.62–1.14). Among employed men, FRs and 95% CIs for evening shift work, night shift work, and rotating shift work were 0.89 (95% CI: 0.68–1.17), 0.94 (95% CI: 0.66–1.33), and 0.91 (95% CI: 0.75–1.11) relative to daytime shift work. The FR for any non-daytime shift work was 0.91 (95% CI: 0.78–1.07). Total work hours (long or short) and job independence scores were not appreciably associated with fecundability.

Total hours worked was not associated with time to pregnancy. Working non-daytime shifts and being unemployed were associated with slightly decreased fecundability. However, the variability in these estimates was substantial and the results were reasonably consistent with chance. Little association was observed for other occupation measures.

Keywords: Fertility, Occupational Health, Work Stress

INTRODUCTION

Psychological stress is an important determinant of overall health. Occupational stress is a form of psychological stress related to one’s job and can be divided into work content (e.g. work load, pace, working hours, participation and control) and context stressors (career development, status, pay, role in organization, interpersonal relationships and culture). It may occur when work demands and pressures are not matched to the workers’ knowledge and abilities; workers have little support from colleagues; or workers have little control over work processes or find that their efforts are disproportionate to the job’s rewards (WHO, 2020). Some studies report an association between high levels of occupational stress and lower testosterone levels or reduced fertility (Auger et al., 2001; Giblin, Poland, Moghissi, Ager, & Olson, 1988; Li, Lin, Li, & Cao, 2011; Mcgrady, 1984; Negro-Vilar, 1993), but not all (Fenster et al., 1998; Hjollund, Bonde, Henriksen, Giwercman, & Olsen, 2004; Wesselink et al., 2018).

Previous studies of occupational stress and fertility have mostly focused on men seeking infertility treatment, or have evaluated the effects of selected components of occupational stress on semen characteristics (Auger et al., 2001; Bigelow et al., 1998; Eisenberg, Chen, Ye, & Buck Louis, 2015; Giblin et al., 1988; Gollenberg et al., 2010; Irgens, Msc, & Ulstein, 1999; Janevic et al., 2014; Li et al., 2011; Sheiner, Sheiner, Carel, Potashnik, & Shoham-Vardi, 2002). While there have been two epidemiologic studies of male occupational stressors—a prospective study including job control and demands (Hjollund et al., 2004) and a retrospective study examining night shift work (Zhu, 2003)) and time to pregnancy (TTP)—there was little evidence of an association in either study. However, both studies were conducted outside North America, where occupational environments and policies differ considerably.

Total hours worked, shift work (defined as work outside normal daytime hours), and job latitude are work-content related measure of occupational stress that may affect fecundability (Auger et al., 2001; Sheiner et al., 2002; Tuntiseranee, Olsen, Geater, & Kor-anantakul, 1998). Long work hours can affect hormone balance (Ulhôa, Marqueze, Burgos, & Moreno, 2015) and interrupt sleep duration (NIOSH et al., 2015). Results linking these individual context- and content-related occupational exposures to fecundability are mixed. One retrospective cohort study found evidence of delayed conception among women who worked ≥70 hours per week, and stronger associations in couples where both partners worked ≥70 hours per week relative to both partners working ≤60 hours/week (odds ratio (OR)=2.0, 95% Confidence Interval (CI): 1.1–3.8) for TTP >7.8 months (i.e., 75th percentile for cohort) (Tuntiseranee et al., 1998). Conversely, a cross-sectional study of 1,001 men found that working <40 hours per week was associated with sperm morphological defects (Auger et al., 2001). Some studies have reported an association between shift work and reduced semen quality (Irgens et al., 1999) and male infertility (El-Helaly, Awadalla, Mansour, & El-Biomy, 2010), other studies have found little association between shift work and male fertility (Bigelow et al., 1998; Eisenberg et al., 2015; Tuntiseranee et al., 1998; Zhu, 2003).

Studies of the association between employment status itself and male fertility have also been inconclusive. One cross-sectional study found that exposure to ≥2 recent stressful life events, including job loss or unemployment, was associated with lower sperm concentration and motility, and abnormal sperm morphology (Gollenberg et al., 2010). However, in a longitudinal study, unemployment status was associated with greater fertility (Raymo & Shibata, 2017). A challenge in studying the effect of unemployment on fertility is that pregnancy planning may be put on hold when experiencing employment issues (Kreyenfeld & Andersson, 2014), thus studies that collect data on pregnancy intention are particularly important.

Finally, low job independence, defined here as occupations with high job burden and low decision latitude, which reflect both work-content and work-context related stressors, have been associated with detrimental health outcomes, such as heart disease (Kivimäki et al., 2006), major depressive disorder (Ahola et al., 2006), and fertility in men (Auger et al., 2001; El-Helaly et al., 2010; Sheiner et al., 2002). However, other studies (Bigelow et al., 1998; Hjollund et al., 2004; Janevic et al., 2014; Karasek et al., 1998) indicate little association between occupational stressors and fertility.

In this prospective cohort study of pregnancy planners, we evaluated the extent to which fecundability (i.e., the per-cycle probability of conception among non-contracepting couples) was associated with male occupational stressors, including employment status, hours worked per week, time of day those hours were worked, and a novel measure of occupational independence calculated by the Occupational Information Network (O*NET), a publicly-available database(O*NET, 2018). Finally, we quantified the extent to which selected variables, such as intercourse frequency and sleep duration, mediated the association between by which occupational stressors affected fecundability.

METHODS

Pregnancy Study Online (PRESTO) is an ongoing web-based preconception cohort study. The study methods have been described in detail elsewhere (Wise et al., 2015). Briefly, women aged 21–45 years residing in the U.S. or Canada, who are actively trying to conceive, and not using contraception or fertility treatment are eligible for participation. Female participants complete an online baseline questionnaire with items on demographics, behavioral factors, medical and reproductive history, and medication use. After completion of the baseline questionnaire, female participants are given the option to invite their male partners to participate. Male partners aged ≥21 years are eligible. Male partner participation involves completion of a baseline questionnaire. Female participants complete follow-up questionnaires every 8 weeks for up to 12 months to update their pregnancy status. This study was approved by the Institutional Review Board at Boston Medical Center, and online informed consent was obtained from all participants.

Assessment of exposure.

On the male baseline questionnaire, participants reported their employment status, average number of hours worked each week, the time of day they mainly worked (daytime, evening, nights, and changing or rotating shifts), and their job title. To estimate job independence, we linked self-reported current job title through the Occupational Information Network (O*NET) database, a publicly-available database containing standardized and occupational specific measures of job requirements, skills, tasks, and other occupational assessments designed by the U.S. Department of Labor/Employment and Training Administration (O*NET, 2018). Once a job title is linked to the O*NET system, a measure of job independence on a scale of 0–100 (with 100 indicating highest level of job independence) can be found under the sub-heading of work values and the specific measure of independence, defined as the following: “Occupations that satisfy this work value allow employees to work on their own and make decisions. Corresponding needs are Creativity, Responsibility and Autonomy” (O*NET, 2018). We used additional data from social media (e.g., LinkedIn) to gather more specific information on job titles that were too ambiguous to categorize accurately using O*Net by title alone (such as engineer or teacher) (29.7%). Nearly 16% of job titles remained unclassified due to missing job title (4.3%); a job title that remained too ambiguous to code (such as engineer) (9.5%); or military or other occupations not found in database (2.1%). These titles were marked as missing and job independence score was imputed using multiple imputation (Sterne et al., 2009).

Assessment of outcome.

At baseline, female participants reported their LMP date, usual menstrual cycle length, and the number of conception attempt cycles before study entry. On follow-up questionnaires, they reported their most recent LMP date and whether they had become pregnant since the previous questionnaire. Total discrete menstrual cycles at risk were calculated as follows: menstrual cycles of attempt at study entry + [(LMP from most recent follow-up questionnaire - date of baseline questionnaire completion)/usual menstrual cycle length] +1. Couples contributed observed cycles of attempt time to the analysis from baseline until reported conception, loss to follow-up, withdrawal, initiation of fertility treatment, or 12 cycles, whichever came first.

Assessment of covariates.

On the baseline questionnaire, both partners reported their age, race/ethnicity, education, height, and weight. On the male baseline questionnaire, men reported physical activity, alcohol intake, smoking history, average hours of sleep per night and whether they previously fathered a child. Household income and frequency of intercourse were ascertained from the female baseline questionnaire. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Stress was evaluated via the 10-item version of the Perceived Stress Scale (PSS-10) (Cohen, Kamarck, & Mermelstein, 1983). In addition, each participant completed the Major Depression Inventory (MDI), a 12-item tool that assesses depressive symptoms during the past 2 weeks (range of scores: 0–50), with a higher MDI score indicating more severe symptoms (Bech, 1997). The MDI has been previously validated and has high sensitivity (0.86) and specificity (0.86) compared with clinician-diagnosed major depressive disorder (Bech, Rasmussen, Olsen, Noerholm, & Abildgaard, 2001). Total Metabolic Equivalents of Task (MET) of physical activity were calculated by multiplying the average number of hours per week engaged in various activities by metabolic equivalents estimated from the Compendium of Physical Activities.

Exclusions.

From June 2013 through June 2019, 10,518 eligible women completed the baseline questionnaire. We excluded 118 women whose baseline date of last menstrual period (LMP) was >6 months before study entry and 30 women with missing/implausible LMP data. We then excluded 2,084 women who had been trying to conceive for more than 6 cycles at enrollment, to reduce potential for differential exposure misclassification (i.e., subfertility causing changes in behavior). Of the 8,286 remaining female participants, 4,446 (54%) invited their male partners to participate, and 2,046 males (46%) enrolled. Finally, we excluded 228 male students (11.1%) from the analysis, resulting in 1,818 couples in the study.

Data analysis.

Hours of work per week were categorized into ≤20, 21–34, 35–40, 41–50, 51–60, and >60 hours. Daily hours worked were categorized as daytime shifts only, evening shifts only, night shifts only, and changing or rotating shifts. We divided the O*NET-derived job independence score into quintiles based on the frequency distribution in the analytic cohort, yielding categories of <64 (low independence), 64–68, 69–77, 78–81, and ≥82 (high independence). The independence score and the number of hours worked were also analyzed as continuous variables using restricted cubic splines to allow for non-linearity (Durrleman & Simon, 1989).

The data were formatted into Andersen-Gill structure, with one row per at-risk menstrual cycle under observation to account for variations in attempt time at study entry and reduce bias from left truncation (Therneau & Grambsch, 2000). We used proportional probabilities regression models to estimate fecundability ratios (FRs) and 95% confidence intervals (CIs) for the association between occupational measures and fecundability. The FR is the average per-cycle probability of conception comparing the exposed category with the reference category (FExp / FUnexp). A FR of <1 indicates a longer TTP comparing exposed with unexposed men.

We selected potential confounders a priori based on the available literature and assessment of a causal diagram (Supplemental Figure 1). Results were adjusted for covariates ascertained at baseline, including age (<25, 25–29, 30–34, ≥35 years) race/ethnicity (White/non-Hispanic vs. other), education (high school or less, some college, college graduate, graduate school), total METs, alcohol consumption (0,1–6,7–13, and ≥14 drinks/week), BMI (<25, 25–29, ≥30 kg/m2), cigarette smoking (never, current, and past), perceived stress score (PSS) (<10, 10–19, 20–29, ≥30), and having previously fathered a child (yes vs. no). Additionally, we adjusted for female partner age and education. We ran further models mutually adjusting for hours worked, time of day worked, and independence score.

We stratified analyses by sleep duration (<7 vs. ≥7 hours/night) because it is an important determinant of fecundability in our cohort (Wise et al., 2018) and it is likely associated with some occupational stressors. Additionally we performed mediation analyses (VanderWeele, 2011) to analyze the extent to which intercourse frequency and sleep duration mediated associations between hours worked or non-day shift work and fecundability. We chose intercourse frequency as a plausible mediator as many occupational stressors can decrease the frequency of intercourse, such as shift work (Gaskins, Sundaram, Buck Louis, & Chavarro, 2018) or employment status (Wellings, Palmer, Machiyama, & Slaymaker, 2019). We chose sleep duration as a plausible mediator as longer work hours and non-day shift work may reduce sleep duration, a key factor for optimal testosterone production.

We used multiple imputation to impute missing outcome, exposure, and covariate data (Sterne et al., 2009). We generated five imputation data sets using PROC MI and combined point estimates and standard errors from each data set using PROC MIANALYZE. This method of imputation assumes that the variables are missing at random conditional on measured characteristics included in the imputation model. For the 69 couples without follow-up data, we assigned them one cycle of follow-up and imputed their pregnancy status (pregnant vs. not pregnant). Missingness for covariates ranged from 0.1% (smoking, physical activity) to 15.9% (independence score). We used the weighted copy method to improve convergence of the regression model (Deddens & Petersen, 2008). All statistical analyses were performed using SAS version 9.4.

RESULTS

A total of 1,818 couples contributed 1,152 pregnancies during 7,195 observed menstrual cycles of attempt time. Hours worked per week was positively associated with partner education and inversely associated with intercourse frequency. Daytime shift work was positively associated with higher male education, higher independence scores, and inversely associated with current smoking. Rotating shift work was positively associated with hours worked per week, previously fathering a child, and inversely associated with BMI, Non-Hispanic White race/ethnicity, intercourse frequency, and shorter sleep duration (Table 1). Independence score was positively associated with male age and education, partner education, and inversely associated with BMI, shorter sleep duration, intercourse frequency, and shift work. Unemployment was positively associated with male age, BMI, shorter sleep duration, intercourse frequency, and previously fathering a child (Table 2).

Table 1:

Age Adjusted Baseline Characteristics of 1,818a Men According to Selected Occupational Factors

Characteristic Job Hours Worked per Week1 Shift Worked1
1–20 21–34 35–40 41–50 51–60 >60 Daytime Evening Nights Rotating
Number of men (%) 41 (2.4) 74 (4.2) 778 (44.6) 621 (35.6) 164 (9.4) 66 (3.8) 1412 (81.0) 87 (5.0) 65 (3.7) 180 (10.3)
Male age, years (mean) 34.8 31.1 31.7 31.8 32.7 31.2 32.0 31.8 31.0 31.4
Total MET-hours per week (mean) 35.6 32.8 30.7 32.8 35.5 37.2 32.7 29.8 26.0 32.9
Cycles of attempt time at study entry (mean) 2.3 2.2 1.9 1.8 1.8 2.0 1.8 2.3 2.4 2.0
White, non-Hispanic, % 78.6 82.8 85.9 87.2 91.2 77.6 87.3 76.7 81.3 84.8
Male alcohol ≥14 drinks/week, % 2.3 11.8 10.9 13.1 14.0 12.1 12.4 10.2 3.9 12.4
Male BMI <25, kg/m2 % 31.7 33.2 30.3 36.2 26.3 33.9 33.8 23.4 32.2 25.5
Male BMI ≥30, kg/m2 % 27.5 36.7 29.1 27.1 36.9 18.5 27.1 42.4 37.2 34.8
Male education ≥16 years, % 50.8 47.0 68.0 74.1 64.6 64.3 74.0 40.7 34.4 51.2
Partner’s education ≥16 years, % 47.8 48.8 60.8 62.3 62.3 64.9 62.8 50.9 45.8 54.1
Current smoker, % 29.1 15.2 12.1 11.4 13.0 18.4 11.2 18.0 31.1 14.1
Former smoker, % 23.0 20.0 17.3 16.9 17.8 14.7 17.0 24.0 14.0 18.5
Male sleep <7 hours/day, % 38.7 21.8 34.5 32.2 42.4 61.8 31.3 44.3 60.2 52.4
Male sleep ≥9 hours/day, % 5.7 10.7 2.8 2.7 0.6 1.4 2.4 5.2 5.9 3.9
Intercourse ≤1 times/month, % 8.1 4.3 4.2 3.8 6.3 3.6 4.2 4.9 5.8 4.4
Intercourse ≥4 times/week, % 21.3 18.8 13.8 12.6 12.1 6.6 13.3 17.7 14.5 12.3
PSS-10 score ≥20, % 27.5 22.6 19.8 22.3 22.2 22.0 21.0 26.2 24.2 23.6
MDI ≥25, % 15.8 3.9 5.0 3.6 5.3 3.8 4.0 8.0 10.0 6.7
Previously fathered a child, % 75.3 50.2 44.7 45.9 45.8 33.9 45.2 50.4 57.7 47.4
Job hours worked per week 44.1 41.0 39.2 47.8
Daytime shifts, % 58.1 60.2 85.3 82.7 78.1 57.6
Evening shifts, % 5.1 16.3 5.4 3.3 5.3 0.0
Night shifts, % 22.5 6.8 2.8 3.4 3.1 3.9
Changing or rotating shifts, % 10.1 16.8 6.5 10.7 13.5 38.5
Independence score 72.3 69.0 70.9 72.3 73.5 72.8 72.8 66.4 66.7 67.5
Open in a new tab

Abbreviations: BMI=body mass index, MDI=major depression inventory, METs=metabolic equivalent of task, PSS=perceived stress scale.

1

Excludes 74 unemployed men.

Table 2:

Age Adjusted Baseline Characteristics of 1,818 (1,744 Employed) Men According to Selected Occupational Factors

Characteristic Independence Score1 Unemployed
<64 64–68 69–77 78–81 ≥82 No Yes
Number of men (%) 380 (21.8) 218 (12.5) 548 (31.4) 247 (14.2) 351 (20.1) 1744 (95.9) 74 (4.1)
Male age, years (mean) 31.0 31.9 31.7 32.5 32.7 31.9 33.3
Total MET-hours per week (mean) 28.4 33.5 31.5 35.1 34.7 32.3 31.1
Cycles of attempt time at study entry (mean) 2.0 2.0 1.9 1.6 1.8 1.9 2.0
White, non-Hispanic, % 84.6 89.1 85.4 88.7 85.6 86.2 76.6
Male alcohol ≥14 drinks/week, % 10.1 15.4 10.6 12.0 13.9 12.1 11.2
Male BMI <25, kg/m2 % 26.9 31.6 33.3 29.8 38.3 32.4 30.5
Male BMI ≥30, kg/m2 % 34.9 28.5 28.5 28.6 24.4 28.9 37.5
Male education ≥16 years, % 44.5 65.5 72.6 79.8 83.9 68.7 48.3
Partner’s education ≥16 years, % 54.2 56.8 60.8 65.8 67.9 60.8 58.3
Current smoker, % 20.5 12.6 9.9 9.8 11.5 12.7 14.6
Former smoker, % 18.9 19.4 16.2 13.1 19.3 17.3 22.9
Male sleep <7 hours/day, % 42.0 35.6 34.8 32.7 29.5 35.1 40.4
Male sleep ≥9 hours/day, % 4.8 1.9 2.9 1.1 2.8 2.9 10.8
Intercourse ≤1 times/month, % 2.8 6.2 3.0 6.4 5.3 4.3 4.8
Intercourse ≥4 times/week, % 14.9 15.9 11.8 15.7 12.5 13.5 19.6
PSS-10 score ≥20, % 23.1 19.1 21.4 20.6 21.7 21.6 40.0
MDI ≥25, % 5.2 5.6 5.5 3.5 3.0 4.8 16.9
Previously fathered a child, % 52.2 46.5 43.5 44.0 44.2 46.2 51.5
Job hours worked per week 43.5 44.2 43.7 46.7 44.1
Daytime shifts, % 70.2 83.3 82.6 80.5 88.9
Evening shifts, % 8.5 5.2 4.1 3.5 3.3
Night shifts, % 5.9 3.4 3.6 2.8 1.9
Changing or rotating shifts, % 15.4 8.2 9.8 13.2 5.9
Open in a new tab

Abbreviations: BMI=body mass index, MDI=major depression inventory, METs=metabolic equivalent of task, PSS=perceived stress scale.

1

74 men were excluded due to being unemployed

Hours worked per week showed little association with fecundability in the fully-adjusted model. Relative to 35–40 hours worked per week, the FRs and 95% CIs for <20, 21–34, 41–50, 51–60, and >60 hours per week were 1.26 (0.88–1.83), 0.95 (0.71–1.27), 1.05 (0.93–1.18), 1.04 (0.86–1.26), and 1.15 (0.86–1.55), respectively (Table 3). Spline regression models examining continuous work duration were consistent with the categorical analyses (Supplemental Figure 2).

Table 3.

Fecundability Ratios and 95% Confidence Intervals for Selected Occupational Factors, PRESTO, 2013–2019.

Hours Worked Per Week1 Men Pregnancies Cycles Unadjusted Adjusted 2 Mutually Adjusted 3
 ≤20 hours per week 41 25 139 1.15 (0.80–1.66) 1.26 (0.88–1.83) 1.28 (0.89–1.86)
 21–34 hours per week 74 42 305 0.94 (0.71–1.25) 0.95 (0.71–1.27) 0.96 (0.72–1.29)
 35–40 hours per week 778 488 3,141 Reference Reference Reference
 41–50 hours per week 621 416 2,446 1.10 (0.98–1.23) 1.05 (0.93–1.18) 1.05 (0.94–1.19)
 51–60 hours per week 164 103 622 1.05 (0.87–1.27) 1.04 (0.86–1.26) 1.05 (0.86–1.27)
 >60 hours per week 66 40 223 1.12 (0.84–1.49) 1.15 (0.86–1.55) 1.17 (0.87–1.58)
Shift Work 1
 Daytime 1,412 926 5,472 Reference Reference Reference
 Evening 87 51 376 0.84 (0.64–1.10) 0.89 (0.68–1.17) 0.88 (0.66–1.16)
 Nights 65 34 290 0.78 (0.56–1.11) 0.94 (0.66–1.33) 0.91 (0.63–1.31)
 Changing or rotating shifts 180 103 738 0.87 (0.71–1.06) 0.91 (0.75–1.11) 0.89 (0.73–1.09)
 All non-day shifts 332 188 1,404 0.85 (0.73–0.98) 0.91 (0.78–1.07) 0.89 (0.76–1.04)
Independence Score 1
 <64 (20th Percentile) 380 227 1,504 1.03 (0.85–1.24) 1.13 (0.93–1.36) 1.14 (0.94–1.38)
 64 – 68 (21st-33rd Percentile) 218 131 909 0.98 (0.80–1.20) 1.02 (0.83–1.26) 1.02 (0.82–1.26)
 69 – 77 (34th-66th Percentile) 548 357 2,072 1.14 (0.98–1.33) 1.17 (1.00–1.37) 1.18 (1.01–1.38)
 78 – 81 (67th-80th Percentile) 247 169 1,010 1.07 (0.89–1.28) 1.04 (0.87–1.25) 1.05 (0.88–1.26)
 ≥82 (81st Percentile) 351 230 1,381 Reference Reference Reference
Unemployed
 Yes 74 38 319 0.75 (0.55–1.01) 0.84 (0.62–1.14) NA
 No 1,744 1,114 6,876 Reference Reference NA
Open in a new tab

NA=not applicable

1

Excludes 74 unemployed men.

2

Adjusted for age, race/ethnicity, education, total METs, alcohol consumption, body mass index (BMI), current smoker, past smoker, sleep duration, MDI, PSS-10 score, previously fathered a child, and female partner factors (age and education).

3

Mutually adjusted for other occupational stressors.

FRs (95% CIs) for evening, night, and changing/rotating shifts were 0.89 (95%: 0.68–1.17), 0.94 (95%: 0.66–1.33), and 0.91 (95%: 0.75–1.11), respectively, compared with daytime shifts. A combined measure for all non-daytime shift workers yielded an FR of 0.91 (95%: 0.78–1.07). The FR for comparing unemployed with employed men was 0.84 (95%: 0.62–1.14) (Table 3).

There was no consistent association between job independence score and fecundability. Comparing the lower four quintiles (<64, 64–68, 69–77, and 78–81) with the highest quintile (≥82) yielded FRs (95% CIs) of 1.13 (95%: 0.93–1.36), 1.02 (95%: 0.83–1.26), 1.17 (95%: 1.00–1.37), and 1.04 (95%: 0.87–1.25), respectively (Table 3). Spline regression models that examined continuous job independence showed similar results (Supplemental Figure 3).

Among participants with ≥7 hours of sleep/day, we found slightly stronger inverse associations between evening shift work (FR=0.84, 95% CI: 0.58–1.20) and rotating shift work (FR=0.76, 95% CI: 0.57–1.02) and fecundability, but similar associations for unemployment status, hours worked, and job independence score. Results were similar among those with <7 hours/day of sleep, with the exception of rotating and evening shift work, where we found little association (FR=1.10, 95% CI: 0.84–1.43 and FR=0.94, 95% CI: 0.62–1.41, respectively), although the reduced sample sizes lead to wider confidence intervals and thus less precise estimates.

Mutually adjusting for the three occupational stress-related exposures (i.e. hours worked, shift work, independence score) did not meaningfully change the FRs (Table 3). Finally, there was little evidence of mediation of the associations in this study by sleep duration or intercourse frequency (Supplemental Table 1).

DISCUSSION

In this preconception cohort study, there was little relation between total hours worked per week or O*NET-estimated job independence score and fecundability. Unemployment status and non-daytime shift work were associated with slightly decreased fecundability, but the small magnitude of the association could readily be accounted for by random variation. Results for evening and rotating shift work were weaker among those with shorter sleep durations (<7 hours/day). Finally, there was little evidence that the associations for unemployment status and shift work were mediated by intercourse frequency or sleep duration.

Our null findings for number of hours worked per week tend to agree with previous studies. In a cross-sectional study of 1,496 pregnant women in Thailand (Tuntiseranee et al., 1998), delays in conception were observed when both partners worked long hours (>70 hours per week each), but not when just the male partner worked >40 hours/week; retrospective ascertainment of work hours could have introduced recall bias. In a cross-sectional study of 1,001 male partners of pregnant women, working <40 hours per week was associated with greater sperm morphological defects, but not other semen parameters (Auger et al., 2001). However, abnormal sperm morphology has not been strongly associated with fecundity in previous studies (Bonde et al., 1998; Buck Louis et al., 2014; Irvine, 1998; Loft et al., 2003; Slama et al., 2002).

Our finding of an association between working non-daytime shifts and slightly decreased fecundability agrees with a cross-sectional Norwegian study of 365 men undergoing infertility investigation (Irgens et al., 1999), as well as an Egyptian case-control study comprising 255 infertile and 267 fertile men (El-Helaly et al., 2010). In the Norwegian study, self-reported shift work was associated with increased odds of reduced semen quality, classified as low count or low percent normal morphology (OR=1.46, 95% CI: 0.96–2.40). Additionally, in the Egyptian study, shift work was positively associated with infertility (OR=3.60, 95% CI: 1.12–11.57). Several other studies, however, have reported null results. In a cross-sectional study of pregnant women in Thailand, shift work was not associated with fecundability, as measured by fecundity odds ratios (OR=1.0, 95% CI: 0.6–1.6) (Tuntiseranee et al., 1998). Similarly, a case-control study of 845 males attending a Canadian fertility clinic found little association between shift work and infertility (Bigelow et al., 1998). In a prospective U.S. cohort study of 456 couples, investigators found a negligible association between shift work and semen quality, as measured using standard clinical semen analyses (Eisenberg et al., 2015). One additional cross-sectional study reported evidence that fixed evening and night shifts were associated with longer TTP (OR=0.80 for both), but rotating shifts showed little association (OR=0.99) (Zhu, 2003). The authors hypothesized that their associations may have been mediated by intercourse frequency or pregnancy planning, which were not measured. The present study addressed these issues by restricting to pregnancy planners and by examining the mediating effects of intercourse frequency, though we found intercourse frequency was not a mediator of the association between male shift work and couple fecundability.

The lack of meaningful association between job independence score and fecundability in PRESTO aligns with data from a case-control study of 944 men from a university fertility clinic (Bigelow et al., 1998). In that study, researchers found that while self-reported job strain (reported in categories of “a great deal,” “some,” “hardly any,” “none”) was associated with decreased semen quality (i.e., percent progressive sperm, total motile count, morphology, abnormal heads, and coiled tail defects), it was not associated with clinically-defined infertility (Bigelow et al., 1998). Additionally, PRESTO results agree with two cross-sectional studies of semen parameters (193 and 399 men, respectively), in which neither study found an association between the Karasek measure of job strain and sperm concentration, motility, morphology, or TTP (Hjollund et al., 2004; Janevic et al., 2014).

Limitations of the present study include a single measurement of occupational exposures, assessed at baseline. If participants’ work schedules or jobs changed during follow-up, we would not have captured that change. Given the prospective design of this study, however, we would expect misclassification of occupational measures at baseline to be non-differential with respect to subsequent fecundability. Though several publications have relied on the O*NET-derived measure of job independence to assess its influence on various health outcomes (Crouter, Lanza, Pirretti, Goodman, & Neebe, 2006; Lee et al., 2016; Meyer, Cifuentes, & Warren, 2011), the measure is relatively new and has not been validated. Also there is potential for residual confounding due to uncontrolled (or inadequately measured) confounders, such as working multiple jobs. PRESTO participants were only asked to report on characteristics of their primary job. Another theoretical limitation is selection bias stemming from web-based recruitment. We consider it unlikely that use of the internet would be associated with both exposures and outcome in this study. In our similarly-designed Danish preconception study, birth registry data were used to compare six perinatal associations from the entire nation with our volunteer cohort. Despite differences in the prevalence of demographic and lifestyle covariates, measures of associations were nearly identical (Hatch et al., 2016). This finding is consistent with two studies examining the effects of low participation rates in cohort studies, which found that low participation rates did not meaningfully influence effect estimates (Nilsen et al., 2009; Nohr, Frydenberg, Henriksen, & Olsen, 2006). Finally, given the variety of different methodologies employed as well as diverse study designs, it can be quite difficult to reach clear conclusions from previous literature. In this study, we attempted to compare the results of the associations between occupational stress and previous studies and to synthesize those with the results presented here.

Despite these limitations, our study improves on previous studies of occupational stressors and fecundity in several ways. We enrolled couples during the early stages of conception attempts, and thus were able to study the full fertility spectrum and reduce potential for recall and selection biases, and the loss of accuracy due to left truncation of TTP data (Schisterman, Cole, Ye, & Platt, 2013). We accounted for a number of important covariates not controlled for in prior studies, including physical activity, race/ethnicity, intercourse frequency, and previously fathering a child. We assessed fecundability as opposed to semen parameters, a strength given that semen quality previously has been only weakly associated with fecundability (Bonde et al., 1998; Buck Louis et al., 2014; Irvine, 1998; Loft et al., 2003; Slama et al., 2002). Finally, by linking job title to independence via O*NET, we utilized a novel measure of job independence derived from self-reported work-content and work-context related characteristics for almost 1,000 occupations, continually updated with input from workers within each occupation (O*NET, 2018).

In this preconception cohort study, we found some evidence that unemployment status and non-daytime shiftwork were associated with decreased fecundability. These associations did not appear to be mediated by intercourse frequency or sleep duration.

Supplementary Material

Supplemental Material

Contributor Information

Craig James McKinnon, Boston University School of Public Health, Department of Epidemiology, 715 Albany Street, Boston, MA 02118 USA.

Elizabeth Elliott Hatch, Boston University School of Public Health, Department of Epidemiology, Boston, USA.

Olivia R Orta, Boston University School of Public Health, Department of Epidemiology, Boston, USA.

Kenneth J Rothman, Boston University School of Public Health, Department of Epidemiology, Boston, USA.

Michael L Eisenberg, Department of Urology and Obstetrics & Gynecology, Stanford University School of Medicine, Stanford, USA.

Johanna Wefes-Potter, Boston University School of Public Health, Department of Epidemiology, Boston, USA.

Lauren A Wise, Boston University School of Public Health, Department of Epidemiology, Boston, USA.

REFERENCES

  1. Ahola K, Honkonen T, Kivimaki M, Virtanen M, Isometsa E, Aromaa A, & Lonnqvist J. (2006). Contribution of burnout to the association between job strain and depression: the health 2000 study. Journal of Occupational and Environmental Medicine, 48(10), 1023–1030. 10.1097/01.jom.0000237437.84513.92 [DOI] [PubMed] [Google Scholar]
  2. Auger J, Eustache F, Andersen AG, Irvine DS, Jørgensen N, Skakkebæk NE, … Jouannet P. (2001). Sperm morphological defects related to environment, lifestyle and medical history of 1001 male partners of pregnant women from four European cities. Human Reproduction. 10.1093/humrep/16.12.2710 [DOI] [PubMed] [Google Scholar]
  3. Bech P. (1997). Quality of life instruments in depression. European Psychiatry : The Journal of the Association of European Psychiatrists, 12(4), 194–198. 10.1016/S0924-9338(97)89104-3 [DOI] [PubMed] [Google Scholar]
  4. Bech P, Rasmussen N-A, Olsen LR, Noerholm V, & Abildgaard W. (2001). The sensitivity and specificity of the Major Depression Inventory, using the Present State Examination as the index of diagnostic validity. Journal of Affective Disorders. 10.1016/S0165-0327(00)00309-8 [DOI] [PubMed] [Google Scholar]
  5. Bigelow PL, Canada A, Jarrell J, Young MR, Keefe TJ, & Love EJ (1998). Association of semen quality and occupational factors: comparison of case-control analysis and analysis of continuous variables. In FERTILITY AND STERILITY @ (Vol. 69). [DOI] [PubMed] [Google Scholar]
  6. Bonde JPE, Ernst E, Jensen TK, Hjollund I NH, Kolstad H, Henriksen TB, … Skakkebæk NE (1998). Relation between semen quality and fertility: A population-based study of 430 first-pregnancy planners. Lancet. 10.1016/S0140-6736(97)10514-1 [DOI] [PubMed] [Google Scholar]
  7. Buck Louis GM, Sundaram R, Schisterman EF, Sweeney A, Lynch CD, Kim S, … Chen Z. (2014). Semen quality and time to pregnancy: the Longitudinal Investigation of Fertility and the Environment Study. Fertility and Sterility, 101(2), 453–462. 10.1016/j.fertnstert.2013.10.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cohen S, Kamarck T, & Mermelstein R. (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24(4), 385–396. [PubMed] [Google Scholar]
  9. Crouter AC, Lanza ST, Pirretti A, Goodman WB, & Neebe E. (2006). The O*Net Jobs Classification System: A Primer for Family Researchers. Family Relations, 55(4), 461–472. 10.1111/j.1741-3729.2006.00415.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Deddens JA, & Petersen MR (2008). Approaches for estimating prevalence ratios. Occupational and Environmental Medicine, 65(7), 481,501–506. 10.1136/oem.2007.034777 [DOI] [PubMed] [Google Scholar]
  11. Durrleman S, & Simon R. (1989). Flexible regression models with cubic splines. Statistics in Medicine. 10.1002/sim.4780080504 [DOI] [PubMed] [Google Scholar]
  12. Eisenberg ML, Chen Z, Ye A, & Buck Louis GM (2015). Relationship between physical occupational exposures and health on semen quality: Data from the Longitudinal Investigation of Fertility and the Environment (LIFE) study. Fertility and Sterility. 10.1016/j.fertnstert.2015.02.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. El-Helaly M, Awadalla N, Mansour M, & El-Biomy Y. (2010). Workplace exposures and male infertility - A case-control study. International Journal of Occupational Medicine and Environmental Health. 10.2478/v10001-010-0039-y [DOI] [PubMed] [Google Scholar]
  14. Fenster L, Katz DF, Wyrobek AJ, Pieper C, Rempel DM, Oman D, & Swan SH (1998). Effects of Psychological Stress on Human Semen Quality. The Journal of Urology, 18(2), 1939. 10.1097/00005392-199811000-00100 [DOI] [PubMed] [Google Scholar]
  15. Gaskins AJ, Sundaram R, Buck Louis GM, & Chavarro JE (2018). Predictors of Sexual Intercourse Frequency Among Couples Trying to Conceive. The Journal of Sexual Medicine, 15(4), 519–528. 10.1016/j.jsxm.2018.02.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Giblin PT, Poland ML, Moghissi KS, Ager JW, & Olson JM (1988). Effects of stress and characteristic adaptability on semen quality in healthy men. Fertility and Sterility. 10.1016/S0015-0282(16)59663-0 [DOI] [PubMed] [Google Scholar]
  17. Gollenberg AL, Liu F, Brazil C, Drobnis EZ, Guzick D, Overstreet JW, … Swan SH (2010). Semen quality in fertile men in relation to psychosocial stress. Fertility and Sterility. 10.1016/j.fertnstert.2008.12.018 [DOI] [PubMed] [Google Scholar]
  18. Hatch EE, Hahn KA, Wise LA, Mikkelsen EM, Kumar R, Fox MP, … Rothman KJ (2016). Evaluation of Selection Bias in an Internet-based Study of Pregnancy Planners. Epidemiology (Cambridge, Mass.), 27(1), 98–104. 10.1097/EDE.0000000000000400 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hjollund NHI, Bonde JPE, Henriksen TB, Giwercman A, & Olsen J. (2004). Job strain and male fertility. Epidemiology. 10.1097/01.ede.0000100290.90888.4a [DOI] [PubMed] [Google Scholar]
  20. Irgens K, Msc,;, & Ulstein. (1999). The Effect of Male Occupational Exposure in Infertile Couples in Norway Materials and Methods Study Design Ovid: The Effect of Male Occupational Exposure in Infertile Couples in. Original Articles] Journal of Occupational and Environmental Medicine Issue, 41(12), 1116–1120. Retrieved from http://ovidsp.tx.ovid.com.ezproxy.bu.edu/sp-3.22.1b/ovidweb.cgiof8 [DOI] [PubMed] [Google Scholar]
  21. Irvine DS (1998). Epidemiology and aetiology of male infertility. Human Reproduction (Oxford, England), 13 Suppl 1, 33–44. [DOI] [PubMed] [Google Scholar]
  22. Janevic T, Kahn LG, Landsbergis P, Cirillo PM, Cohn BA, Liu X, & Factor-Litvak P. (2014). Effects of work and life stress on semen quality. Fertility and Sterility. 10.1016/j.fertnstert.2014.04.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Karasek R, Brisson C, Kawakami N, Houtman I, Bongers P, & Amick B. (1998). The Job Content Questionnaire (JCQ): an instrument for internationally comparative assessments of psychosocial job characteristics. Journal of Occupational Health Psychology, 3(4), 322–355. 10.1037/1076-8998.3.4.322 [DOI] [PubMed] [Google Scholar]
  24. Kivimäki M, Head J, Ferrie JE, Shipley MJ, Brunner E, Vahtera J, & Marmot MG (2006). Work stress, weight gain and weight loss: Evidence for bidirectional effects of job strain on body mass index in the Whitehall II study. International Journal of Obesity. 10.1038/sj.ijo.0803229 [DOI] [PubMed] [Google Scholar]
  25. Kreyenfeld M, & Andersson G. (2014). Socioeconomic differences in the unemployment and fertility nexus: Evidence from Denmark and Germany. Advances in Life Course Research, 21, 59–73. 10.1016/j.alcr.2014.01.007 [DOI] [PubMed] [Google Scholar]
  26. Lee LJ, Symanski E, Lupo PJ, Tinker SC, Razzaghi H, Pompeii LA, … Chan W. (2016). Data linkage between the National Birth Defects Prevention Study and the Occupational Information Network (O*NET) to assess workplace physical activity, sedentary behaviors, and emotional stressors during pregnancy. American Journal of Industrial Medicine, 59(2), 137–149. 10.1002/ajim.22548 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Li Y, Lin H, Li Y, & Cao J. (2011). Association between socio-psycho-behavioral factors and male semen quality: Systematic review and meta-analyses. Fertility and Sterility. 10.1016/j.fertnstert.2010.06.031 [DOI] [PubMed] [Google Scholar]
  28. Loft S, Kold-Jensen T, Hjollund NH, Giwercman A, Gyllemborg J, Ernst E, … Bonde JP (2003). Oxidative DNA damage in human sperm influences time to pregnancy. Human Reproduction (Oxford, England), 18(6), 1265–1272. [DOI] [PubMed] [Google Scholar]
  29. Mcgrady AV (1984). Effects of psychological stress on male reproduction: a review. [DOI] [PubMed] [Google Scholar]
  30. Meyer JD, Cifuentes M, & Warren N. (2011). Association of self-rated physical health and incident hypertension with O*NET factors: validation using a representative national survey. Journal of Occupational and Environmental Medicine, 53(2), 139–145. 10.1097/JOM.0b013e318203f220 [DOI] [PubMed] [Google Scholar]
  31. Negro-Vilar A. (1993). Stress and other environmental factors affecting fertility in men and women: Overview. Environmental Health Perspectives. 10.1289/ehp.93101s259 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Nilsen RM, Vollset SE, Gjessing HK, Skjaerven R, Melve KK, Schreuder P, … Magnus P. (2009). Self-selection and bias in a large prospective pregnancy cohort in Norway. Paediatric and Perinatal Epidemiology, 23(6), 597–608. 10.1111/j.1365-3016.2009.01062.x [DOI] [PubMed] [Google Scholar]
  33. NIOSH Caruso, C., Geiger-Brown J, Takahashi M, Trinkoff A, & Nakata A. (2015). NIOSH training for nurses on shift work and long work hours. In DHHS (NIOSH) Publication No. 2015–115. 10.26616/NIOSHPUB2015115 [DOI] [Google Scholar]
  34. Nohr EA, Frydenberg M, Henriksen TB, & Olsen J. (2006). Does low participation in cohort studies induce bias? Epidemiology (Cambridge, Mass.), 17(4), 413–418. 10.1097/01.ede.0000220549.14177.60 [DOI] [PubMed] [Google Scholar]
  35. O*NET. (2018). Occupational Information Network. O*NET Resource Center. Retrieved July 18, 2018, from National Center for O*NET Development; website: https://www.onetonline.org/ [Google Scholar]
  36. Raymo JM, & Shibata A. (2017). Unemployment, Nonstandard Employment, and Fertility: Insights From Japan’s “Lost 20 Years”. Demography, 54(6), 2301–2329. 10.1007/s13524-017-0614-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Schisterman EF, Cole SR, Ye A, & Platt RW (2013). Accuracy loss due to selection bias in cohort studies with left truncation. Paediatric and Perinatal Epidemiology, 27(5), 491–502. 10.1111/ppe.12073 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Sheiner EK, Sheiner E, Carel R, Potashnik G, & Shoham-Vardi I. (2002). Potential association between male infertility and occupational psychological stress. Journal of Occupational and Environmental Medicine. 10.1097/00043764-200212000-00001 [DOI] [PubMed] [Google Scholar]
  39. Slama R, Eustache F, Ducot B, Jensen TK, Jorgensen N, Horte A, … Jouannet P. (2002). Time to pregnancy and semen parameters: a cross-sectional study among fertile couples from four European cities. Human Reproduction (Oxford, England), 17(2), 503–515. [DOI] [PubMed] [Google Scholar]
  40. Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, … Carpenter JR (2009). Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ (Clinical Research Ed.). 10.1136/bmj.b2393 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Therneau T, & Grambsch P. (2000). Modeling Survival Data: Extending the Cox Model. New York, NY: Springer-Verlag. [Google Scholar]
  42. Tuntiseranee P, Olsen J, Geater A, & Kor-anantakul O. (1998). Are long working hours and shiftwork risk factors for subfecundity? A study among couples from southern Thailand. Occupational and Environmental Medicine. 10.1136/oem.55.2.99 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Ulhôa MA, Marqueze EC, Burgos LGA, & Moreno CRC (2015). Shift work and endocrine disorders. International Journal of Endocrinology, 2015, 826249. 10.1155/2015/826249 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. VanderWeele TJ (2011). Causal mediation analysis with survival data. Epidemiology. 10.1097/EDE.0b013e31821db37e [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Wellings K, Palmer MJ, Machiyama K, & Slaymaker E. (2019). Changes in, and factors associated with, frequency of sex in Britain: evidence from three National Surveys of Sexual Attitudes and Lifestyles (Natsal). BMJ (Clinical Research Ed.), 365, l1525. 10.1136/bmj.l1525 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Wesselink AK, Hatch EE, Rothman KJ, Weuve JL, Aschengrau A, Song RJ, & Wise LA (2018). Perceived Stress and Fecundability: A Preconception Cohort Study of North American Couples. American Journal of Epidemiology, 187(12), 2662–2671. 10.1093/aje/kwy186 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. WHO. (2020). Stress at the workplace. Retrieved from World Health Organization website: https://www.who.int/occupational_health/topics/stressatwp/en/
  48. Wise LA, Rothman KJ, Mikkelsen EM, Stanford JB, Wesselink AK, McKinnon C, … Hatch EE (2015). Design and Conduct of an Internet-Based Preconception Cohort Study in North America: Pregnancy Study Online. Paediatric and Perinatal Epidemiology, 29(4), 360–371. 10.1111/ppe.12201 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Wise Rothman, Wesselink K, Mikkelsen A, Sorensen E, McKinnon C, & Hatch E. (2018). Male sleep duration and fecundability in a North American preconception cohort study. Fertility and Sterility, 109(3), 453–459. 10.1016/j.fertnstert.2017.11.037 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Zhu JL (2003). Shift work and subfecundity: a causal link or an artefact? In Occup Environ Med (Vol. 60). Retrieved from http://www.occenvmed.com/cgi/content/full/60/9/e12 [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 Material
NIHMS1729457-supplement-Supplemental_Material.pdf (298.8KB, pdf)

RESOURCES