Abstract
STUDY QUESTION
Is paternal physical activity associated with semen quality parameters and with outcomes of infertility treatment?
SUMMARY ANSWER
Among men presenting for infertility treatment, weightlifting and outdoor activities were associated with higher sperm concentrations but not with greater reproductive success.
WHAT IS ALREADY KNOWN
Higher physical activity is related to better semen quality but no studies to date have investigated whether it predicts greater reproductive success.
STUDY DESIGN, SIZE, DURATION
The Environment and Reproductive Health (EARTH) Study is an on-going prospective cohort study which enrolls subfertile couples presenting at Massachusetts General Hospital (2005–2013). In total, 231 men provided 433 semen samples and 163 couples underwent 421 IVF or intrauterine insemination cycles.
PARTICIPANTS/MATERIALS, SETTING, METHODS
Leisure time spent in physical and sedentary activities over the past year was self-reported using a validated questionnaire. We used mixed models to analyze the association of physical and sedentary activities with semen quality and with clinical pregnancy and live birth rates.
MAIN RESULTS AND THE ROLE OF CHANCE
Men in this cohort engaged in a median of 3.2 h/week of moderate-to-vigorous activities. Men in the highest quartile of moderate-to-vigorous activity had 43% (95% confidence interval (CI) 9, 87%) higher sperm concentrations than men in the lowest quartile (P-trend = 0.04). Men in the highest category of outdoor activity (≥1.5 h/week) and weightlifting (≥2 h/week) had 42% (95% CI 10, 84%) and 25% (95% CI −10, 74%) higher sperm concentrations, respectively, compared with men in the lowest category (0 h/week) (P-trend = 0.04 and 0.02). Conversely, men who reported bicycling ≥1.5 h/week had 34% (95% CI 4, 55%) lower sperm concentrations compared with men who reported no bicycling (P-trend = 0.05). Paternal physical and sedentary activities were not related to clinical pregnancy or live birth rates following infertility treatment.
LIMITATIONS, REASONS FOR CAUTION
The generalizability of the findings on live birth rates to populations not undergoing infertility treatment is limited.
WIDER IMPLICATIONS OF THE FINDINGS
Certain types of physical activity, specifically weightlifting and outdoor activities, may improve semen quality but may not lead to improved success of infertility treatments. Further research is needed in other non-clinical populations.
STUDY FUNDING/COMPETING INTEREST(S)
The authors are supported by NIH grants R01-ES009718, ES000002, P30-DK046200, T32-DK007703-16 and ES022955 T32-HD060454. None of the authors has any conflicts of interest to declare.
Keywords: physical activity, semen quality, male fertility, assisted reproduction
Introduction
Physical activity has multiple health benefits (Blair and Morris, 2009) yet its association with male fertility remains unclear. Previous researchers have proposed strenuous exercise as a risk factor for male factor infertility based on findings of reduced semen quality in long-distance runners and endurance cyclists (Arce and De Souza, 1993; Arce et al., 1993; De Souza et al., 1994; Miller et al., 1997; Viru et al., 2001; Gebreegziabher et al., 2004; Vaamonde et al., 2006, 2009; Safarinejad et al., 2009; Wise et al., 2011). However, more recent evidence suggests that exercise might improve semen quality (Vaamonde et al., 2012; Gaskins et al., 2013). Further complicating this picture, certain activities might differentially affect semen quality, an imperfect marker of male fertility. Bicycling, for example, has been consistently associated with lower semen quality parameters (Lucia et al., 1996; Gebreegziabher et al., 2004; Wise et al., 2011). At present, no studies have investigated in the same cohort of men both the effect of different types of physical activity on semen quality parameters and their effect on clinically relevant measures of fertility such as ability to father a pregnancy.
Thus, the objective of our study was 2-fold, first to evaluate the relationship of specific physical and sedentary activities with semen quality parameters and second to evaluate their relationship with reproductive treatment outcomes among a cohort of couples from a fertility center.
Materials and Methods
Study population
Participants were couples enrolled in an ongoing study at the Massachusetts General Hospital (MGH) Fertility Center (Boston, MA, USA) designed to identify determinants of fertility. In brief, all men presenting to the MGH Fertility Center who meet eligibility requirements (e.g. between 18 and 55 years of age at enrollment and no history of vasectomy) are approached by study personnel to participate in the Environment and Reproductive Health (EARTH) study. Approximately 60% of those contacted by the research nurses participate in this study. For the semen quality analysis, men were eligible if they provided at least one semen sample between January 2005 and May 2013 (n = 294). Azoospermic men (n = 4), men with incomplete semen analyses (n = 1) and men missing physical activity data (n = 58) were excluded. Men excluded from the semen quality analysis had similar demographic and reproductive characteristics compared with those included. For the analysis of infertility treatment outcomes we further excluded 44 men whose partner was not enrolled or had not completed an intrauterine insemination (IUI) or assisted reproduction technology (ART) cycle by May 2013 and 24 men who only had cycles initiated prior to physical activity assessment or >2 years after physical activity assessment. Men excluded from the analysis of infertility treatment outcomes had similar demographic (e.g. age, race, body mass index (BMI) and education), reproductive (e.g. history of variocele, previous infertility examination and undescended testes) and female partner characteristics (e.g. female age, race, BMI and infertility diagnosis) to those included. After exclusions, 231 men (433 semen samples) were included in the semen parameters analysis and 163 couples (421 cycles) were included in the infertility treatment outcomes analysis.
At enrollment, height and weight were measured by a research nurse, from which we calculated BMI (kg/m2). Participants also completed a detailed take-home questionnaire focused on lifestyle factors, medical and reproductive history. A trained research nurse abstracted clinical information including infertility diagnosis and treatment protocols from electronic medical records. The Institutional Review Boards of MGH and the Harvard School of Public Health approved the study. All participants provided written informed consent after a research nurse explained study procedures.
Physical and sedentary activity
Time spent in leisure time physical and sedentary activities was assessed using a validated questionnaire (Wolf et al., 1994). Specifically, men reported the average time per week during the preceding year spent on any of the following activities: walking, jogging, running, bicycling, swimming, tennis, squash, weightlifting, aerobics and moderate (e.g. yard work and gardening) and heavy (e.g. digging and chopping) outdoor work. For the questions on walking, jogging and running, men were instructed to include time spent on treadmills and for bicycling men were instructed to include time spent on a stationary bike. Men also reported the average time per week during the preceding year spent sitting at work, while driving and at home. Each activity question had 13 categories for response ranging from ‘never’ to ‘40+ hours per week’. We assigned the duration of activity by using the median value for each category. We calculated total physical and sedentary activity (hours/week) by summing across all physical and sedentary activities. Moderate-to-vigorous activity (hours/week) was calculated by summing time spent in all physical activities except walking. Total metabolic equivalents (MET hours/week) were calculated by multiplying the average MET level of a given activity (Ainsworth et al., 2000) by its reported duration and summing across all physical activities. The validity of this questionnaire was assessed in a different cohort by comparison against four 7-day activity recalls collected over the course of 1 year (Wolf et al., 1994). The Spearman correlation for total physical activity was 0.79 and for total sedentary activity was 0.41.
Semen analysis
Semen samples were obtained on site by masturbation and collected into a sterile plastic container for analysis. Out of 231 men, 48% provided 1 sample, 30% provided 2 samples, 12% provided 3 samples and 10% provided ≥4 samples (a maximum of 6). Men were instructed to abstain from ejaculation for at least 48 h before producing the sample and to report the specific time of abstinence. Thirty-nine men (48 semen samples) did not report their last ejaculation date and were assigned to the most common abstinence time category (48–72 h). Semen samples were liquefied at 37°C for 20 min before analysis. If a sample did not liquefy after 20 min, it was given a minimum of 10 more minutes before starting analysis. Sperm counts and motility were assessed with a computer-aided semen analysis system (Hamilton-Thorne Biosciences, Ceros, Version 14), which shows a high correlation with manual analyses. Sperm morphology was determined using strict criteria and results are expressed as percentage normal spermatozoa (Kruger et al., 1988).
Clinical outcomes
Couples underwent IUI or ART via IVF with either conventional insemination or ICSI, as clinically indicated. We defined clinical pregnancy as the presence of an intrauterine pregnancy confirmed by ultrasound and live birth as the birth of a neonate on or after 24 weeks of gestation.
Statistical analysis
We classified men into quartiles based on time spent in moderate-to-vigorous and sedentary activities and calculated descriptive statistics for demographic, reproductive and lifestyle characteristics according to quartile of activity. We used Fisher's exact test, χ2 test and Kruskal–Wallis test, as appropriate, to test for associations across quartiles. We log-transformed sperm concentration and sperm count to normalize distributions. We used multivariable mixed models with random intercepts to evaluate the relationship of physical and sedentary activities with semen quality parameters and infertility treatment outcomes. These models allow the use of multiple outcome observations per individual while accounting for within-person correlations in outcomes. These models can also generate unbiased estimates in the presence of an unbalanced design when data are not missing completely at random and the lack of balance can be accurately predicted by all measured covariates. We used linear mixed models with autoregressive correlation structure for semen quality parameters and generalized linear mixed models with a binomial outcome distribution and logit link function for the infertility treatment outcomes. Tests for trend were conducted across quartiles using the median activity level in each quartile as a continuous variable in the regression models. All results are presented as population marginal means, adjusted for covariates (Searle et al., 1980).
Confounding was evaluated using a hybrid approach combining prior knowledge using directed acyclic graphs (DAGs) and a statistical approach based on change in point estimates (Weng et al., 2009). A set of variables was determined by a review of the prior literature and a detailed DAG was created identifying variables that should be included in the models. Variables retained in the final semen quality models were abstinence time, age, smoking status, race, education level and BMI. Final variables retained in the infertility treatment outcome models were maternal and paternal age, paternal race, paternal smoking status and maternal and paternal BMI. All models were additionally run without adjusting for paternal BMI as we hypothesized this variable to be a potential mediator as well as confounder of the associations. SAS version 9.2 (SAS Institute, Cary, NC, USA) was used for all statistical analyses. A P < 0.05 was considered significant.
Results
Men in this cohort had mean (SD) age of 36.3 (5.0) years and BMI of 27.2 (4.1) kg/m2. They engaged in a median (interquartile range) of 3.2 (1.1, 10.5) h/week of moderate-to-vigorous activities and 52.5 (32.0, 66.0) h/week of sedentary activities. The majority was Caucasian (86%), never smokers (68%), had a graduate degree (50%) and reported a previous infertility evaluation (77%). Time spent in moderate-to-vigorous activities and sedentary activities was only weakly correlated (rSpearman = −0.08). Men who were more physically active and more sedentary were more likely to have a graduate degree (P-value = 0.02 and<0.001, respectively) (Table I). More sedentary men were also more likely to have a history of smoking (P-value = 0.04). Other baseline demographic characteristics were similar across activity levels. Most physical activity was moderate and heavy outdoor activities (38%), followed by running and jogging (24%), weightlifting (16%), bicycling (16%), swimming laps (3%), playing tennis or squash (2%) and aerobics (1%).
Table I.
Demographic characteristics of male partners by quartile of moderate to vigorous activity and by quartile of sedentary activity.
| Moderate–vigorous activity |
Sedentary activity |
|||
|---|---|---|---|---|
| Quartile | Q1 | Q4 | Q1 | Q4 |
| Range of activity (h/week) | 0–1 | ≥7.1 | 0.9–31 | ≥66 |
| n = 57 | n = 57 | n = 57 | n = 59 | |
| Male characteristics | ||||
| Age, years | 36.0 (4.8) | 37.3 (5.2) | 37.5 (5.3) | 35.1 (5.1) |
| BMI, kg/m2 | 27.8 (4.4) | 26.8 (3.3) | 27.6 (4.0) | 27.0 (3.9) |
| Ever smoker, n (%) | 19 (33.3) | 17 (29.8) | 20 (35.1) | 24 (40.7)* |
| Race, n (%) | ||||
| White | 45 (79.0) | 49 (86.0) | 45 (79.0) | 50 (84.8) |
| Black | 1 (1.8) | 2 (3.5) | 3 (5.3) | 0 (0.0) |
| Asian | 6 (10.5) | 0 (0.0) | 3 (5.3) | 5 (8.5) |
| Other | 5 (8.8) | 6 (10.5) | 6 (10.5) | 4 (6.8) |
| Education level, n (%) | ||||
| Less than college | 17 (29.8) | 11 (19.3)* | 25 (43.9) | 8 (13.6)* |
| College degree | 16 (28.1) | 14 (24.6) | 14 (24.6) | 26 (44.1) |
| Graduate degree | 24 (42.1) | 32 (56.1) | 18 (31.6) | 25 (42.4) |
| Moderate–vigorous activity, h/week | 0.4 (0.4) | 13.9 (10.4)* | 5.8 (8.8) | 4.8 (8.9) |
| Sedentary activity, h/week | 51.9 (31.6) | 47.0 (20.5) | 17.1 (9.1) | 79.0 (11.3)* |
| Previous infertility evaluation, n (%) | 40 (70.2) | 43 (75.4) | 43 (75.4) | 44 (74.6) |
| Abstinence time <48 h, n (%) | 17 (29.8) | 19 (33.3) | 18 (31.6) | 24 (40.7) |
| Any male factor diagnosis,a n (%) | 30 (52.6) | 30 (52.6) | 30 (52.6) | 32 (54.2) |
| Undescended testes, n (%) | 2 (3.6) | 2 (3.5) | 4 (7.0) | 1 (1.7) |
| Varicocele, n (%) | 4 (7.0) | 1 (1.8) | 4 (7.0) | 4 (6.8) |
| Prostatitis, n (%) | 1 (1.8) | 1 (1.8) | 2 (3.5) | 1 (1.7) |
| Epididymitis, n (%) | 0 (0.0) | 2 (3.5) | 2 (3.5) | 0 (0.0) |
| Any reproductive surgery,b n (%) | 13 (22.8) | 20 (35.1) | 16 (28.1) | 14 (23.7) |
| Female characteristicsc | n = 42 | n = 42 | n = 34 | n = 47 |
| Age, years | 35.1 (4.5) | 34.9 (4.5) | 33.7 (7.4) | 34.6 (3.9) |
| BMI, kg/m2 | 25.2 (5.2) | 23.7 (3.6) | 24.3 (4.3) | 24.0 (4.1) |
| Primary infertility diagnosis, n% | ||||
| Male factor | 15 (26.3) | 17 (29.8) | 17 (29.8) | 19 (32.2) |
| Female factor | 22 (38.6) | 16 (28.1) | 15 (26.3) | 17 (28.8) |
| Diminished ovarian reserve | 3 (5.3) | 7 (12.3) | 4 (7.0) | 3 (5.1) |
| Endometriosis | 4 (7.0) | 2 (3.5) | 0 (0.0) | 4 (6.8) |
| Ovulation disorders | 9 (15.8) | 4 (7.0) | 5 (8.8) | 7 (11.9) |
| Tubal | 5 (8.8) | 3 (5.3) | 6 (10.5) | 3 (5.1) |
| Uterine | 1 (1.8) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Unexplained | 20 (35.1) | 24 (42.1) | 25 (43.9) | 23 (39.0) |
| Moderate-vigorous activity, h/week | 2.1 (2.5) | 3.9 (3.5) | 3.1 (3.3) | 2.5 (3.0) |
| Sedentary activity, h/week | 46.3 (19.8) | 43.5 (21.0) | 44.1 (21.5) | 49.0 (23.4) |
For continuous variables, the Kruskal–Wallis test was used to test for associations across quartiles of activity. For categorical variables, χ2 and Fisher's exact tests (when one or more cell counts were ≤5) were used to test the associations between quartiles of activity.
aAt baseline visit, at least one semen quality parameter was below WHO 2011 cut-off values for concentration, motility or morphology.
bHistory of orchidopexy, varicocelectomy, hydrocelectomy, hernia repair, bladder neck surgery or other reproductive surgery.
cOne hundred and sixty-three men had a female partner in the study who had completed at least one cycle.
*P-value from trend across categories <0.05.
Higher levels of moderate-to-vigorous physical activity were related to higher sperm concentrations (Table II). In the multivariable model, men in the second, third and fourth quartiles of physical activity had 10% (95% confidence interval (CI) −16, 44%), 30% (95% CI −1, 70%) and 43% (95% CI 9, 87%) higher sperm concentrations than men in the lowest quartile (P-trend = 0.04). This association was similar when male BMI was excluded from the multivariable model. There was no evidence of effect modification by male BMI (P-interaction = 0.43). Physical activity had a marginal positive association with total sperm count (P-trend = 0.08). Sperm motility (total and progressive) and morphology were unrelated to physical activity.
Table II.
Association between paternal physical and sedentary activities and semen quality parameters.
| Adjusted means (95% CI)a | Quartile of activity |
P for trendb | |||
|---|---|---|---|---|---|
| Q1 | Q2 | Q3 | Q4 | ||
| Moderate-to-vigorous activity (h/week) | |||||
| Number of men | 57 | 59 | 58 | 57 | |
| Median [range] | 0.3 [0, 1.0] | 2.2 [1.1, 3.2] | 5.5 [3.3, 7.0] | 10.7 [7.1, 67.0] | |
| Sperm concentration, ×106 sperm per milliliters | 43.1 (33.1, 56.0) | 47.5 (36.4, 61.9) | 55.8 (42.5, 73.3) | 61.6 (47.0, 80.7) | 0.04 |
| Progressive motility, % A + B | 42.9 (37.0, 48.8) | 41.3 (35.4, 47.3) | 49.6 (43.5, 55.7) | 43.7 (37.7, 49.8) | 0.53 |
| Morphology, % normal | 6.4 (5.5, 7.3) | 6.7 (5.8, 7.7) | 6.8 (5.9, 7.8) | 6.2 (5.2, 7.1) | 0.62 |
| Sperm count, ×106 sperm | 114.4 (88.3, 148.2) | 119.1 (91.6, 154.7) | 143.6 (109.6. 188.0) | 152.5 (116.6, 199.3) | 0.08 |
| Total METS (h/week) | |||||
| Number of men | 57 | 58 | 58 | 58 | |
| Median [range] | 5.4 [0, 11.3] | 20.5 [11.4, 30.4] | 41.2 [30.5, 57.5] | 79.3 [57.6, 354] | |
| Sperm concentration, ×106 sperm/ml | 47.2 (36.2, 61.6) | 55.2 (42.2, 72.2) | 46.6 (35.5, 61.3) | 56.7 (43.3, 74.3) | 0.49 |
| Progressive motility, % A+B | 44.5 (38.6, 50.4) | 45.8 (39.8, 51.8) | 42.1 (36.0, 48.2) | 44.7 (38.7, 50.8) | 0.89 |
| Morphology, % normal | 6.6 (5.7, 7.6) | 6.7 (5.8, 7.6) | 5.9 (4.9, 6.8) | 6.9 (6.0, 7.8) | 0.78 |
| Sperm count, ×106 sperm | 135.8 (104.5, 176.5) | 128.2 (98.4, 167.0) | 118.5 (90.5, 155.2) | 139.5 (106.8, 182.4) | 0.83 |
| Sedentary activity (h/week) | |||||
| Number of men | 57 | 58 | 57 | 59 | |
| Median [range] | 17.5 [0, 31.0] | 42.3 [32.0, 52.0] | 59.5 [53.0, 65.0] | 77 [66.0, 131.0] | |
| Sperm concentration, ×106 sperm/ml | 54.1 (41.5, 70.7) | 51.6 (38.8, 68.5) | 55.4 (41.7, 73.6) | 44.1 (33.7, 57.7) | 0.37 |
| Progressive motility, % A + B | 45.1 (39.2, 51.1) | 42.6 (36.2, 49.0) | 45.4 (39.1, 51.8) | 43.7 (37.7, 49.8) | 0.92 |
| Morphology, % normal | 7.2 (6.2, 8.1) | 6.4 (5.4, 7.4) | 6.2 (5.2, 7.1) | 6.2 (5.3, 7.2) | 0.16 |
| Sperm count, ×106 sperm | 139.4 (107.2, 181.4) | 119.5 (90.3, 158.3) | 145.1 (109.7, 192.0) | 117.9 (90.3, 154.0) | 0.61 |
CI, confidence interval; METs, metabolic equivalents. All analyses were run using linear mixed models with random intercepts and autoregressive correlation structure. Sperm concentration and count were log transformed for normality.
aThe marginal means are presented adjusted for abstinence time (<48, 48–72, ≥72 h), age (continuous), smoking status (ever, never), race (white, other), education (less than college, college, graduate) and BMI (continuous).
bTests for trend across quartiles using the median activity level in each quartile as a continuous variable.
Next we investigated whether specific physical activities were associated with sperm concentration (Fig. 1). Greater time spent in outdoor activities and weightlifting was associated with higher sperm concentration (P-trend = 0.02 and 0.04, respectively). Specifically, men in the highest category of outdoor activity (≥1.5 h/week) and weightlifting (≥2 h/week) had 42% (95% CI 10, 84%) and 25% (95% CI −10, 74%) higher sperm concentrations, respectively, than men with 0 h/week. Bicycling, in contrast, was inversely associated with sperm concentrations (P-trend = 0.05). Men who reported bicycling ≥1.5 h/week had 34% (95% CI 4, 55%) lower sperm concentration compared with men who reported no bicycling. Time spent walking, running, jogging or in other specific physical or sedentary activities was not associated with sperm concentration (Supplementary data, Table SI). Similar results were observed when sperm concentration and total count were classified according to the World Health Organization (WHO) 2010 cutoffs (<15 × 106 sperm/ml and <39 × 106sperm) (Supplementary data, Table SII) (WHO 2010).
Figure 1.
Adjusted mean (95% CI) sperm concentration (million/ml) by type of moderate-to-vigorous physical activity. All analyses were run using log-transformed concentration and linear mixed models with random intercepts and autoregressive correlation structure. The marginal means are presented adjusted for abstinence time (<48, 48–72, ≥72 h), age (continuous), smoking status (ever, never), race (white, other), education (less than college, college, graduate) and BMI (continuous). *P-value compared with lowest category is <0.05.
Finally, we evaluated the relationship of paternal physical and sedentary activities with outcomes of infertility treatment. In the final multivariable model, we did not find an association between time spent in moderate-to-vigorous or sedentary activities and clinical pregnancy or live birth rates following IUI or ART (Table III). For instance, the adjusted difference (95% CI) in live birth rates comparing men in the fourth versus first quartile of moderate-to-vigorous activity was −0.02 (−0.07, 0.11) following IUI and −0.09 (−0.21, 0.06) following ART. Similarly, we did not find any association of the specific physical or sedentary activities with outcomes following infertility treatment (Supplementary data, Table SIII). These associations were similar after adjustment for initial infertility diagnosis and protocol type (data not shown).
Table III.
Association between paternal physical activity and sedentary activity and clinical outcomes in couples from a fertility clinic.
| Adjusted mean rates (95% CI)a | Intrauterine insemination 83 couples, 211 cycles |
Assisted reproduction technologies, 124 couples, 210 cycles |
||
|---|---|---|---|---|
| Clinical pregnancy | Live birth | Clinical pregnancy | Live birth | |
| Moderate-to-vigorous activity (h/week) | ||||
| 0–1.0 | 0.14 (0.07, 0.27) | 0.09 (0.03, 0.21) | 0.45 (0.30, 0.61) | 0.42 (0.27, 0.59) |
| 1.1–3.2 | 0.07 (0.02, 0.20) | 0.06 (0.02, 0.20) | 0.50 (0.35, 0.65) | 0.43 (0.28, 0.59) |
| 3.3–7.0 | 0.11 (0.04, 0.26) | 0.03 (0.01, 0.13) | 0.50 (0.36, 0.63) | 0.50 (0.35, 0.64) |
| >7.0 | 0.07 (0.02, 0.22) | 0.06 (0.02, 0.20) | 0.37 (0.24, 0.51) | 0.33 (0.21, 0.48) |
| P for trendb | 0.46 | 0.52 | 0.28 | 0.36 |
| Sedentary activity (h/week) | ||||
| 0–31 | 0.05 (0.01, 0.19) | 0.05 (0.01, 0.22) | 0.53 (0.36, 0.69) | 0.48 (0.31, 0.66) |
| 32–52 | 0.10 (0.03, 0.27) | 0.07 (0.02, 0.21) | 0.36 (0.24, 0.50) | 0.34 (0.22, 0.49) |
| 53–65 | 0.17 (0.08, 0.34) | 0.11 (0.04, 0.26) | 0.52 (0.38, 0.66) | 0.48 (0.33, 0.64) |
| >65 | 0.08 (0.03, 0.18) | 0.03 (0.01, 0.10) | 0.41 (0.29, 0.54) | 0.40 (0.27, 0.55) |
| P for trend | 0.54 | 0.59 | 0.58 | 0.80 |
All analyses were run using generalized linear mixed models with random intercepts, binomial distribution, logit link function.
aMarginal means are presented adjusted for maternal and paternal age (continuous), paternal race (white, other), paternal smoking status (ever, never), and maternal and paternal BMI (continuous).
bTests for trend across quartiles using the median activity level in each quartile as a continuous variable.
Discussion
Among men presenting for infertility treatment, higher moderate-to-vigorous physical activity, specifically weightlifting and outdoor activities, was associated with higher sperm concentration. We did not find, however, that physical activity was related to clinical pregnancy or live birth rates following IUI or ART. Higher time spent bicycling was associated with lower sperm concentration. Our findings suggest that specific types of physical activity affect semen quality parameters differently. Furthermore, despite the apparent benefit of physical activity on sperm concentration, we did not observe that it translated into increased reproductive success in the context of infertility treatment.
Our results of a positive association between physical activity and semen quality are supported by some (Vaamonde et al., 2009, 2012; Gaskins et al., 2013) but not all human studies (Arce et al., 1993; De Souza et al., 1994; Jensen et al., 1995; Lucia et al., 1996; Hall et al., 1999; Gebreegziabher et al., 2004; Safarinejad et al., 2009). In the largest and most similar study to date, Wise et al. found no association between self-reported physical activity and semen quality among 2261 men attending a fertility clinic (Wise et al., 2011). Important differences in study design could explain the discrepant findings. First, Wise et al. assessed physical activity only among current exercisers. Thus, any men who consistently exercised over the previous year but eliminated their activity before completing the baseline questionnaire would have been misclassified as non-exercisers. Perhaps because of this difference in assessment, 48% of the men in the study by Wise et al. reported no physical activity compared with <2% of our men. In addition, Wise et al. asked men to report their primary type of exercise, but there was no information on the frequency with which men engaged in specific activities.
Our finding that specific types of activity affect semen quality parameters differently is supported in the literature and biologically. Others have found that bicycling was associated with poor semen quality (Lucia et al., 1996; Gebreegziabher et al., 2004; Wise et al., 2011). Possible reasons for this link are mechanical trauma caused by compression of the scrotum on the bicycle saddle and higher scrotal temperatures related to exercise itself or wearing constrictive clothing (Leibovitch and Mor, 2005). We are, however, the first study to report that weightlifting and outdoor activities were beneficial for sperm concentrations. Moderate-intensity weight training induces transient increases in free and total testosterone levels shortly after exertion (Schwab et al., 1993) and improves insulin sensitivity in healthy adults (Black et al., 2010). Higher testosterone levels and insulin sensitivity increase sperm concentrations (Singh et al., 2009; Morgante et al., 2011). The link between outdoor activities and semen quality was unexpected. Our best explanation for this association was that time spent in outdoor activities might be a good proxy for endogenous vitamin D levels. Although still preliminary, recent studies suggest a direct relationship between vitamin D and testosterone levels (Wehr et al., 2010; Pilz et al., 2011). We did not find any association of sedentary activity and semen quality parameters, which is similar to what others have found (Hjollund et al., 2002; Stoy et al., 2004) but in contrast to findings we published last year linking increased television watching with lower sperm concentrations (Gaskins et al., 2013). Further research is needed on the relationship between specific physical and sedentary activities and semen quality to confirm our results and explore the possible mechanisms of action. Specifically, a trial of different types of activity (e.g. weight training) and semen quality parameters where intermediate end-points, such as testosterone levels and insulin sensitivity measures, are measured would help determine if there is a causal relationship and, if so, the potential biological mechanisms.
We did not find an association between physical activity and clinical outcomes of infertility treatment. There are many potential explanations for these findings. First, in the context of infertility treatments in which sperm concentrations are concentrated for IUI and IVF or directly injected into the egg for ICSI, factors related to higher sperm counts might not confer any additional benefit. Secondly, it is likely that men who present to fertility clinics with normal semen quality parameters are more likely to have female partners with reduced fertility (Tielemans et al., 2002). Therefore, the relationship of physical activity with clinical outcomes of infertility treatment could be strongly confounded by determinants of female subfertility. While controlling for female age, BMI and initial infertility diagnosis helped account for some of this bias, the underlying fertility potential of the female partner is difficult to quantify and there may be residual bias. Finally, it is possible that sperm concentrations are unrelated to reproductive success at the relatively high levels we observed in this study. For example, only 12% of our men had a sperm concentration less than the WHO cut-off of 15 × 106 sperm/ml.
One study limitation is we used a single physical activity questionnaire to characterize exposure, which could have led to misclassification during follow-up. While this type of misclassification is likely non-differential, it would tend to attenuate effects to the null and could be one explanation for our lack of significant findings for the clinical outcomes. Due to the observational nature of our study we cannot rule out residual or unmeasured confounding. While we tried to account for many of these potential confounding factors in our adjusted models, it is possible there were other variables we did not measure or account for in our analyses. As such, our findings need to be confirmed by a randomized trial before the causal relationship between exercise and semen quality can be established. Finally, it is possible that men with previous knowledge of their fertility potential might have changed their physical activity level (leading to reverse causation); however, history of previous infertility evaluation was not associated with physical activity level in our cohort. Furthermore, physical activity is not a well-characterized risk factor for male infertility and, therefore, it is unlikely that men with previous fertility troubles were counseled to make changes to their exercise regime. Strengths of our study include its prospective design for IVF outcome, the use of a previously validated activity questionnaire and the use of repeat semen samples which allowed us to account for within-person variability in semen parameters.
In summary, we found that higher physical activity, specifically weightlifting and outdoor activities, was associated with higher sperm concentration, whereas bicycling was related to lower sperm concentration. These changes, however, did not translate into greater reproductive success in the context of infertility treatment. Further research is needed to investigate the association between male physical activity and time to pregnancy in other non-clinical populations.
Supplementary data
Supplementary data are available at http://humrep.oxfordjournals.org/.
Authors' roles
R.H. and J.E.C.: study concept and design. R.H., J.C.P. and C.T.: acquisition of data. A.J.G., M.C.A., P.L.W., R.H., M.W.G. and J.E.C.: analysis and interpretation of data. A.J.G. and J.E.C.: drafting of the manuscript. A.J.G., M.C.A., P.L.W., R.H., M.W.G., J.C.P., C.T. and J.E.C.: critical revision of the manuscript for important intellectual content. A.J.G. and P.L.W.: Statistical analysis.
Funding
The authors are supported by NIH grants R01-ES009718, R01-ES022955, ES000002, P30-DK046200, T32-DK007703-16 and T32-HD060454. The funding sources had no influence on the study design, data analysis, writing of the report or the decision to submit the findings of the present study for publication.
Conflict of interest
None declared.
Supplementary Material
References
- Ainsworth BE, Haskell WL, Whitt MC, Irwin ML, Swartz AM, Strath SJ, O'Brien WL, Bassett DR, Schmitz KH, Emplaincourt PO, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc. 2000;9(Suppl):S498–S504. doi: 10.1097/00005768-200009001-00009. [DOI] [PubMed] [Google Scholar]
- Arce JC, De Souza MJ. Exercise and male factor infertility. Sports Med. 1993;3:146–169. doi: 10.2165/00007256-199315030-00002. [DOI] [PubMed] [Google Scholar]
- Arce JC, De Souza MJ, Pescatello LS, Luciano AA. Subclinical alterations in hormone and semen profile in athletes. Fertil Steril. 1993;2:398–404. [PubMed] [Google Scholar]
- Black LE, Swan PD, Alvar BA. Effects of intensity and volume on insulin sensitivity during acute bouts of resistance training. J Strength Cond Res. 2010;4:1109–1116. doi: 10.1519/JSC.0b013e3181cbab6d. [DOI] [PubMed] [Google Scholar]
- Blair SN, Morris JN. Healthy hearts—and the universal benefits of being physically active: physical activity and health. Ann Epidemiol. 2009;4:253–256. doi: 10.1016/j.annepidem.2009.01.019. [DOI] [PubMed] [Google Scholar]
- De Souza MJ, Arce JC, Pescatello LS, Scherzer HS, Luciano AA. Gonadal hormones and semen quality in male runners. A volume threshold effect of endurance training. Int J Sports Med. 1994;7:383–391. doi: 10.1055/s-2007-1021075. [DOI] [PubMed] [Google Scholar]
- Gaskins AJ, Mendiola J, Afeiche M, Jorgensen N, Swan SH, Chavarro JE. Physical activity and television watching in relation to semen quality in young men. Br J Sports Med. 2013 doi: 10.1136/bjsports-2012-091644. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gebreegziabher Y, Marcos E, McKinon W, Rogers G. Sperm characteristics of endurance trained cyclists. Int J Sports Med. 2004;4:247–251. doi: 10.1055/s-2004-819933. [DOI] [PubMed] [Google Scholar]
- Hall HL, Flynn MG, Carroll KK, Brolinson PG, Shapiro S, Bushman BA. Effects of intensified training and detraining on testicular function. Clin J Sport Med. 1999;4:203–208. doi: 10.1097/00042752-199910000-00004. [DOI] [PubMed] [Google Scholar]
- Hjollund NH, Storgaard L, Ernst E, Bonde JP, Olsen J. Impact of diurnal scrotal temperature on semen quality. Reprod Toxicol. 2002;3:215–221. doi: 10.1016/s0890-6238(02)00025-4. [DOI] [PubMed] [Google Scholar]
- Jensen CE, Wiswedel K, McLoughlin J, van der Spuy Z. Prospective study of hormonal and semen profiles in marathon runners. Fertil Steril. 1995;6:1189–1196. [PubMed] [Google Scholar]
- Kruger TF, Acosta AA, Simmons KF, Swanson RJ, Matta JF, Oehninger S. Predictive value of abnormal sperm morphology in in vitro fertilization. Fertil Steril. 1988;1:112–117. doi: 10.1016/s0015-0282(16)59660-5. [DOI] [PubMed] [Google Scholar]
- Leibovitch I, Mor Y. The vicious cycling: bicycling related urogenital disorders. Eur Urol. 2005;3:277–286. doi: 10.1016/j.eururo.2004.10.024. discussion 286–277. [DOI] [PubMed] [Google Scholar]
- Lucia A, Chicharro JL, Perez M, Serratosa L, Bandres F, Legido JC. Reproductive function in male endurance athletes: sperm analysis and hormonal profile. J Appl Physiol. 1996;6:2627–2636. doi: 10.1152/jappl.1996.81.6.2627. [DOI] [PubMed] [Google Scholar]
- Miller BE, Hackney AC, De Souza MJ. The endurance training on hormone and semen profiles in marathon runners. Fertil Steril. 1997;3:585–586. doi: 10.1016/s0015-0282(97)80096-9. author reply 586–587. [DOI] [PubMed] [Google Scholar]
- Morgante G, Tosti C, Orvieto R, Musacchio MC, Piomboni P, De Leo V. Metformin improves semen characteristics of oligo-terato-asthenozoospermic men with metabolic syndrome. Fertil Steril. 2011;6:2150–2152. doi: 10.1016/j.fertnstert.2010.12.009. [DOI] [PubMed] [Google Scholar]
- Pilz S, Frisch S, Koertke H, Kuhn J, Dreier J, Obermayer-Pietsch B, Wehr E, Zittermann A. Effect of vitamin D supplementation on testosterone levels in men. Horm Metab Res. 2011;3:223–225. doi: 10.1055/s-0030-1269854. [DOI] [PubMed] [Google Scholar]
- Safarinejad MR, Azma K, Kolahi AA. The effects of intensive, long-term treadmill running on reproductive hormones, hypothalamus-pituitary-testis axis, and semen quality: a randomized controlled study. J Endocrinol. 2009;3:259–271. doi: 10.1677/JOE-08-0477. [DOI] [PubMed] [Google Scholar]
- Schwab R, Johnson GO, Housh TJ, Kinder JE, Weir JP. Acute effects of different intensities of weight lifting on serum testosterone. Med Sci Sports Exerc. 1993;12:1381–1385. [PubMed] [Google Scholar]
- Searle SR, Speed FM, Milliken GA. Population marginal means in the linear model: an alternative to least square means. Am Stat. 1980;4:216–221. [Google Scholar]
- Singh S, Malini T, Rengarajan S, Balasubramanian K. Impact of experimental diabetes and insulin replacement on epididymal secretory products and sperm maturation in albino rats. J Cell Biochem. 2009;5:1094–1101. doi: 10.1002/jcb.22337. [DOI] [PubMed] [Google Scholar]
- Stoy J, Hjollund NH, Mortensen JT, Burr H, Bonde JP. Semen quality and sedentary work position. Int J Androl. 2004;1:5–11. doi: 10.1046/j.0105-6263.2003.00428.x. [DOI] [PubMed] [Google Scholar]
- Tielemans E, Burdorf A, te Velde E, Weber R, van Kooij R, Heederik D. Sources of bias in studies among infertility clients. Am J Epidemiol. 2002;1:86–92. doi: 10.1093/aje/kwf011. [DOI] [PubMed] [Google Scholar]
- Vaamonde D, Da Silva ME, Poblador MS, Lancho JL. Reproductive profile of physically active men after exhaustive endurance exercise. Int J Sports Med. 2006;9:680–689. doi: 10.1055/s-2005-872906. [DOI] [PubMed] [Google Scholar]
- Vaamonde D, Da Silva-Grigoletto ME, Garcia-Manso JM, Vaamonde-Lemos R, Swanson RJ, Oehninger SC. Response of semen parameters to three training modalities. Fertil Steril. 2009;6:1941–1946. doi: 10.1016/j.fertnstert.2008.09.010. [DOI] [PubMed] [Google Scholar]
- Vaamonde D, Da Silva-Grigoletto ME, Garcia-Manso JM, Barrera N, Vaamonde-Lemos R. Physically active men show better semen parameters and hormone values than sedentary men. Eur J Appl Physiol. 2012;112:3267–3273. doi: 10.1007/s00421-011-2304-6. [DOI] [PubMed] [Google Scholar]
- Viru AM, Hackney AC, Valja E, Karelson K, Janson T, Viru M. Influence of prolonged continuous exercise on hormone responses to subsequent exercise in humans. Eur J Appl Physiol. 2001;6:578–585. doi: 10.1007/s004210100498. [DOI] [PubMed] [Google Scholar]
- Wehr E, Pilz S, Boehm BO, Marz W, Obermayer-Pietsch B. Association of vitamin D status with serum androgen levels in men. Clin Endocrinol. 2010;2:243–248. doi: 10.1111/j.1365-2265.2009.03777.x. [DOI] [PubMed] [Google Scholar]
- Weng HY, Hsueh YH, Messam LL, Hertz-Picciotto I. Methods of covariate selection: directed acyclic graphs and the change-in-estimate procedure. Am J Epidemiol. 2009;10:1182–1190. doi: 10.1093/aje/kwp035. [DOI] [PubMed] [Google Scholar]
- Wise LA, Cramer DW, Hornstein MD, Ashby RK, Missmer SA. Physical activity and semen quality among men attending an infertility clinic. Fertil Steril. 2011;3:1025–1030. doi: 10.1016/j.fertnstert.2010.11.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wolf AM, Hunter DJ, Colditz GA, Manson JE, Stampfer MJ, Corsano KA, Rosner B, Kriska A, Willett WC. Reproducibility and validity of a self-administered physical activity questionnaire. Int J Epidemiol. 1994;5:991–999. doi: 10.1093/ije/23.5.991. [DOI] [PubMed] [Google Scholar]
- World Health Organization (WHO) WHO laboratory manual for the examination and processing of human sperm. 2010.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.