Physical Activity, Television Watching and Semen Quality
Participants were part of the Rochester Young Men's Study (RYMS), a cross-sectional study conducted during 2009–2010 at the University of Rochester (Rochester, New York, USA). Men were recruited into RYMS through flyers and newspapers at college campuses in the Rochester area. Subjects were eligible if they were born in the USA after 31 December 1987, able to read and speak English and able to contact their mother and ask her to complete a questionnaire. A total of 389 potential participants contacted the study. Of these, 305 (78.4%) met all eligibility criteria and 222 men participated in the study. Our analysis only includes men with complete information on both physical activity and TV watching (n=189). The University of Rochester Research Subjects Review Board approved the study and written informed consent was obtained from all subjects before their participation.
Men were asked to report the number of hours they spent in a normal week over the past 3 months engaged in vigorous, moderate or mild exercise. Our main exposure of physical activity was hours per week of moderate-to-vigorous activity defined as the sum of those two categories. This type of activity corresponded to any exercise that made you somewhat to very winded or sweaty. Alternate measures of physical activity were also calculated including total metabolic equivalents (METs) and moderate-to-vigorous METs. Mild (<3 METs), moderate (3–6 METS) and vigorous (>6 METs) activities were given an average MET level of 2, 4.5 and 6, respectively, to calculate the total METs per person. TV watching was assessed in the same questionnaire by asking men to select the category of TV watching time per workday or weekend day that corresponded to their average habits over the past 3 months. Categories for response were 'none/almost none', '1–3 h/day', '4–6 h/day', '7–9 h/day' and 'over 10 h/day'. The median value for each category was used to assign TV watching time, and a weighted average of weekend and workdays was taken to give the average amount of TV watching time per week.
Semen samples were collected by masturbation at the clinic where, upon arrival, men were asked to report the time of their previous ejaculation. The men had been asked to abstain from ejaculation for at least 48 h before sample collection; however, they were not excluded if this was not the case (n=26). Abstinence times >240 h (n=7) were truncated at 240 h. Sample processing was initiated within 30 min of collection. Ejaculate volumes were estimated by specimen weight, assuming a semen density of 1.0 g/ml. Sperm concentration was evaluated by a haemocytometer (Improved Neubauer; Hauser Scientific Inc., Horsham, Pennsylvania, USA). Two chambers of the haemocytometer were counted, and the average was used in this analysis. Motility was analysed using WHO 1999 criteria and was classified as both progressive (A+B) and total (A+B+C). Smears for morphology were air-dried, fixed and shipped to the University Department of Growth and Reproduction at the Rigshospitalet (Copenhagen, Denmark). The slides were Papanicolaou stained and assessed using strict criteria. To increase the consistency and comparability of methods, six sets of duplicate semen samples were sent over the course of the study from the University of Copenhagen's Department of Growth and Reproduction to the Andrology Laboratory (University of Rochester), which is Clinical Laboratory Improvement Amendments certified.
A physical examination of each participant was performed on the same day as semen sampling. Assessments included weight, height, testis size by palpation using the Prader's orchidometer (Andrology Australia, Clayton, Victoria, Australia) and the presence of varicocele or other genital abnormalities. Men also completed questionnaires concerning demographics, medical and reproductive history, psychological stress, medication use (antibiotics, antidepressants and hormones) and smoking habits. Substantial psychosocial stress was defined as indicating a positive response to two or more questions out of six questions on stressful life events. Diet was assessed using a validated questionnaire. Diet quality was summarised by two previously described dietary patterns: a Prudent pattern (characterised by high intakes of fish, chicken, fruit, cruciferous vegetables, tomatoes, leafy green vegetables, legumes and whole grains) and Western pattern (characterised by high intakes of red and processed meat, butter, high-fat dairy, refined grains, pizza, snacks, high-energy drinks, mayonnaise and sweets). All covariates were 100% complete.
Men were classified into quartiles according to their average moderate-to-vigorous physical activity and TV watching per week. Descriptive statistics were calculated for demographic characteristics across quartiles of activity. Multivariable linear regression was used to evaluate the associations between quartile of activity and sperm parameters. Sperm concentration and sperm count were log-transformed to normalise distributions. The association between activity and sperm parameters was also evaluated as a continuous linear and quadratic variable. Tests for non-linearity used the likelihood ratio test, comparing the model without any activity term with the model with the linear and quadratic term. Tests for trend were conducted across quartiles using a variable with the median physical activity and TV watching level in each quartile as a continuous variable in the linear regression models. All results are presented as adjusted means for the median level of each covariate. For sperm concentration and count (which were log-transformed for linear regression), adjusted means were obtained by exponentiating ('back-transforming') the estimated β coefficients for the median level of each covariate.
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. A set of variables was determined by a review of the prior literature, and a detailed DAG was created to identify variables that should be included in the models. An exploratory confounding evaluation was also used with covariates being included in the model if they changed the exposure coefficient by more than 15% and were significant at the p=0.10 level. Variables retained in the final multivariable models were abstinence time (h), race (white/other), smoking status (current/former or never), body mass index (BMI) (kg/m), recruitment period (2009/2010), total energy intake (kcal), TV watching (for physical activity analyses) and moderate-to-vigorous exercise (for TV watching analyses). Motility analyses were additionally adjusted for time from semen collection to start of semen analysis.
Effect modification by BMI (<25 kg/m and ≥25 kg/m), smoking status (current and never/former smokers), physical activity (<8.25 h/week (median) and ≥8.25 h/week moderate-to-vigorous activity) and TV watching (<14 h/week (median) and ≥14 h/week) was tested using cross-product terms in the final multivariate model. SAS V.9.2 (SAS Institute, Cary, North Carolina, USA) was used for all statistical analyses.
Methods
Study Population
Participants were part of the Rochester Young Men's Study (RYMS), a cross-sectional study conducted during 2009–2010 at the University of Rochester (Rochester, New York, USA). Men were recruited into RYMS through flyers and newspapers at college campuses in the Rochester area. Subjects were eligible if they were born in the USA after 31 December 1987, able to read and speak English and able to contact their mother and ask her to complete a questionnaire. A total of 389 potential participants contacted the study. Of these, 305 (78.4%) met all eligibility criteria and 222 men participated in the study. Our analysis only includes men with complete information on both physical activity and TV watching (n=189). The University of Rochester Research Subjects Review Board approved the study and written informed consent was obtained from all subjects before their participation.
Physical Activity and TV Watching
Men were asked to report the number of hours they spent in a normal week over the past 3 months engaged in vigorous, moderate or mild exercise. Our main exposure of physical activity was hours per week of moderate-to-vigorous activity defined as the sum of those two categories. This type of activity corresponded to any exercise that made you somewhat to very winded or sweaty. Alternate measures of physical activity were also calculated including total metabolic equivalents (METs) and moderate-to-vigorous METs. Mild (<3 METs), moderate (3–6 METS) and vigorous (>6 METs) activities were given an average MET level of 2, 4.5 and 6, respectively, to calculate the total METs per person. TV watching was assessed in the same questionnaire by asking men to select the category of TV watching time per workday or weekend day that corresponded to their average habits over the past 3 months. Categories for response were 'none/almost none', '1–3 h/day', '4–6 h/day', '7–9 h/day' and 'over 10 h/day'. The median value for each category was used to assign TV watching time, and a weighted average of weekend and workdays was taken to give the average amount of TV watching time per week.
Semen Collection and Analysis
Semen samples were collected by masturbation at the clinic where, upon arrival, men were asked to report the time of their previous ejaculation. The men had been asked to abstain from ejaculation for at least 48 h before sample collection; however, they were not excluded if this was not the case (n=26). Abstinence times >240 h (n=7) were truncated at 240 h. Sample processing was initiated within 30 min of collection. Ejaculate volumes were estimated by specimen weight, assuming a semen density of 1.0 g/ml. Sperm concentration was evaluated by a haemocytometer (Improved Neubauer; Hauser Scientific Inc., Horsham, Pennsylvania, USA). Two chambers of the haemocytometer were counted, and the average was used in this analysis. Motility was analysed using WHO 1999 criteria and was classified as both progressive (A+B) and total (A+B+C). Smears for morphology were air-dried, fixed and shipped to the University Department of Growth and Reproduction at the Rigshospitalet (Copenhagen, Denmark). The slides were Papanicolaou stained and assessed using strict criteria. To increase the consistency and comparability of methods, six sets of duplicate semen samples were sent over the course of the study from the University of Copenhagen's Department of Growth and Reproduction to the Andrology Laboratory (University of Rochester), which is Clinical Laboratory Improvement Amendments certified.
Covariate Assessment
A physical examination of each participant was performed on the same day as semen sampling. Assessments included weight, height, testis size by palpation using the Prader's orchidometer (Andrology Australia, Clayton, Victoria, Australia) and the presence of varicocele or other genital abnormalities. Men also completed questionnaires concerning demographics, medical and reproductive history, psychological stress, medication use (antibiotics, antidepressants and hormones) and smoking habits. Substantial psychosocial stress was defined as indicating a positive response to two or more questions out of six questions on stressful life events. Diet was assessed using a validated questionnaire. Diet quality was summarised by two previously described dietary patterns: a Prudent pattern (characterised by high intakes of fish, chicken, fruit, cruciferous vegetables, tomatoes, leafy green vegetables, legumes and whole grains) and Western pattern (characterised by high intakes of red and processed meat, butter, high-fat dairy, refined grains, pizza, snacks, high-energy drinks, mayonnaise and sweets). All covariates were 100% complete.
Statistical Analysis
Men were classified into quartiles according to their average moderate-to-vigorous physical activity and TV watching per week. Descriptive statistics were calculated for demographic characteristics across quartiles of activity. Multivariable linear regression was used to evaluate the associations between quartile of activity and sperm parameters. Sperm concentration and sperm count were log-transformed to normalise distributions. The association between activity and sperm parameters was also evaluated as a continuous linear and quadratic variable. Tests for non-linearity used the likelihood ratio test, comparing the model without any activity term with the model with the linear and quadratic term. Tests for trend were conducted across quartiles using a variable with the median physical activity and TV watching level in each quartile as a continuous variable in the linear regression models. All results are presented as adjusted means for the median level of each covariate. For sperm concentration and count (which were log-transformed for linear regression), adjusted means were obtained by exponentiating ('back-transforming') the estimated β coefficients for the median level of each covariate.
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. A set of variables was determined by a review of the prior literature, and a detailed DAG was created to identify variables that should be included in the models. An exploratory confounding evaluation was also used with covariates being included in the model if they changed the exposure coefficient by more than 15% and were significant at the p=0.10 level. Variables retained in the final multivariable models were abstinence time (h), race (white/other), smoking status (current/former or never), body mass index (BMI) (kg/m), recruitment period (2009/2010), total energy intake (kcal), TV watching (for physical activity analyses) and moderate-to-vigorous exercise (for TV watching analyses). Motility analyses were additionally adjusted for time from semen collection to start of semen analysis.
Effect modification by BMI (<25 kg/m and ≥25 kg/m), smoking status (current and never/former smokers), physical activity (<8.25 h/week (median) and ≥8.25 h/week moderate-to-vigorous activity) and TV watching (<14 h/week (median) and ≥14 h/week) was tested using cross-product terms in the final multivariate model. SAS V.9.2 (SAS Institute, Cary, North Carolina, USA) was used for all statistical analyses.
Source...