Semen quality and sedentary work position

Transcription

1 international journal of andrology, 27:5 11 (2004) Semen quality and sedentary work position JULIE STØY,* NIELS HENRIK I. HJØLLUND,* JENS TØLBØLL MORTENSEN, HERMANN BURRà and JENS PETER BONDE* *Department of Occupational Medicine, Aarhus University Hospital, Aarhus, Department of Occupational Medicine, Skive, and ànational Institute of Occupational Health, Copenhagen, Denmark Summary Increased scrotal temperature can, in experimental settings, markedly disturb the production of semen. Sedentary work position may increase the temperature of the scrotum, but previous studies have failed to determine whether changes in scrotal temperature caused by sedentary work actually do affect semen quality. This study was carried out to elucidate the possible harmful effects of sedentary work on sperm count and other semen characteristics. In a semen sample was obtained from 3119 men who attended an infertility workup in one of four Danish fertility centres. A total of 2517 men returned a postal questionnaire with information on life style, leisure time activities, occupational history and job duties. Information on job specific work position was obtained from The Danish Work Environment Cohort study 1990 (DWECS). In this analysis DWECS data for a total of 1747 men was included from men aged years with >30 h of work per week. For all job titles represented in the DWECS, the mean proportion of sedentary work was estimated. The sperm cell concentration was 30.6 million/ml among men in the quintile with lowest job specific sedentary work compared with 40.5 million/ml in the highest quintile. The difference was, however, not statistically significant. Stratification on infertility period, educational level of the man, fertility centre, and fertility-related disease of the spouse did not influence the results. The analyses do not suggest that sedentary work is a risk factor for abnormal semen characteristics. Keywords: occupation, sedentary work, semen characteristics, sperm count Introduction Recent findings suggest that sperm count has been declining over the last decades (Carlsen et al., 1992; Auger et al., 1995) but the causes of the observed decline are mainly unsettled (Mieusset & Bujan, 1995). Clearly, a substantial and persistent increase in scrotal temperature can markedly disturb the production of semen (Mieusset et al., 1987, 1991; Loughlin et al., 1991; Shafik, 1991; Figa-Talamanca et al., 1992; Thonneau et al., 1996; Jannes et al., 1998). Sedentary work, that is work in sitting posture, lead to an increase in Correspondence: Julie Støy, Department of Occupational Medicine, Aarhus University Hospital, Noerrebrogade 44, DK-8000 Aarhus C, Denmark. stoey@stofanet.dk scrotal temperature (Bujan et al., 2000; Hjollund et al., 2000, 2002a) and the prevalence of sedentary work has been increasing in the last decade. It is therefore, of interest to elucidate if parts of the apparent decline in sperm count possibly can be explained by the increasing amount of sedentary work. The impact of the relatively moderate elevations of scrotal temperature as encountered in sedentary work position is, however, not settled (Sas & Szollosi, 1979; Buiatti et al., 1984; Oldereid et al., 1992; Figa-Talamanca et al., 1996; Thonneau et al., 1996; Bigelow et al., 1998; Hjollund et al., 2000). Previous studies of the association between sperm count, semen characteristics and sedentary work suffer from Ó 2004 Blackwell Publishing Ltd.

2 6 J. Støy et al. methodological problems: recall bias (when obtaining exposure information from couples with known fertility problems), selection bias (when recruiting from an infertility clientele), small sample sizes and poorly characterized exposure to sedentary work. This study was carried out to address this issue in an extensive data material with independent exposure information and a sufficient contrast in the exposure to sedentary work position in the data material. Subjects and methods Study population and data collection Semen samples were collected from 3119 men who attended an infertility work up from 1 January 1981 to 31 December 1983 in four fertility centres in Denmark. During 1984, one of the authors sent a postal questionnaire to the men from three of the four fertility clinics, whereas the men from the fourth fertility clinic already filled in a similar questionnaire at the time of the infertility work up. A total of 2517 men (81%) returned the questionnaire with information on current occupation, employment history, lifestyle characteristics, medical history and sociodemographic characteristics. Questions on the current occupation emphasized job duties and exposures to hazardous chemicals and physical agents. The job titles of the men were classified according to the five-figure Danish occupational classification from 1968 (DISCO-68) (Arbejdsdirektoratet, 1986). Additional information on the study population is provided elsewhere (Mortensen, 1986, 1988). Laboratory methods Trained laboratory technicians performed the semen analyses in the four fertility centres, while physicians did the morphological diagnosis. As in usual clinical practice, the participants were asked to be sexually abstinent 3 4 days prior to the semen analysis. Semen sample examinations were part of the infertility workup routine and no inter- or intralaboratory quality assurance was performed. Exposure assessment The questionnaire did not include items about duration of sedentary work. External data were extracted from The Danish Work Environment Cohort study (DWECS), which is conducted by The National Institute of Occupational Health in Copenhagen, Denmark, every fifth year (Nord- Larsen et al., 1992; Borg et al., 2000). The purpose of the cohort study is to describe the working environment and health among Danish wage earners. The wage earners are randomly selected from the central population register of Statistics Denmark. The cross-sectional data was used from 1990 where a total of 5940 telephone interviews of wage earners were conducted. Only information on men aged years with more than 30 h of work per week from the DWECS were extracted, a total of 1747 men. The following variables were obtained: DISCO-68, working hours per week and three questions on the proportion of a working day spent sitting, walking or standing. There were six response levels for each of the questions on work position. A proportion was assigned for each of the six response levels: Ôalmost all the timeõ was assigned a value of 1.00, Ô75% of the timeõ a value of 0.75, Ô50% of the timeõ a value of 0.50, Ô25% of the timeõ a value of 0.25, Ôalmost noneõ a value of 0.075, and ÔnoneÕ a value of The mean proportion of sedentary work was calculated and the corresponding standard deviation for all job titles represented in the data from the DWECS. The job titles were then grouped into quintiles with respect to the mean proportion of sedentary work. The first quintile contained the job titles with the lowest proportion of sedentary work and the fifth quintile contained the job titles with the highest proportion of sedentary work. The standard deviations were highly dependent of the proportion of sedentary work (Fig. 1). To minimize misclassification, only job titles with standard deviations below the median of standard deviations in the quintile in question were included in the analyses. The choice of this threshold of standard deviations was a trade off between statistical power and a sufficient exposure contrast within the data material. Some job titles in the participant database could not be assigned a proportion of sedentary work from the DWECS database, because some job titles were not represented in both databases. To increase the number of job titles represented in both databases, the last two digits of the DISCO-68 was ignored if a chi-square test did not reveal unequally distributed proportions of sedentary work position in the five-figure DISCO-68 job titles (p > 0.10). Of the 2517 men in the participant database, 961 men were included in the analyses; 187 men were excluded because of a missing DISCO-68, 194 because they were represented by less than two observations in the DWECS database and 1175 because they held a job title where the corresponding mean proportion of sedentary work had a standard deviation above the median of standard deviations in the quintile. Analysis and statistical methods The relation between sperm cell concentration and jobspecific proportion of sedentary work was analysed in linear models, where sperm-cell concentration was transformed to the logarithm to obtain normality of residuals and homogeneity of residual variances (SAS PROC GLM) (SAS Institute, 1989). When this transformation is applied, the estimates can be interpreted as percentage change per unit of the independent variable. To enable inclusion of azoospermic men, the value of 1 million/ml was added to the sperm density of all men. The analyses was repeated after exclusion of azoospermic men. Homogeneity and conditional normal distribution of residuals were examined in a residual plot and

3 Semen quality and sedentary work position 7 Figure 1. Plot of the mean proportion of sedentary work vs. the corresponding standard deviation for all male job titles in The Danish Work Environment Cohort study. by test of the Shapiro Wilk statistic. In other analyses, the participants were classified in relation to abnormal semen characteristics, defined as a sperm concentration <20 million/ml or <50% of sperm cells with normal morphology or <50% of sperm cells with normal motility. The characteristics of semen were considered abnormal if any of these criteria were fulfilled. The odds ratio of abnormal semen characteristics in the five quintiles of sedentary work and with sedentary work entered as a continuous variable was calculated by logistic regression (SAS PROC LOGISTIC) (SAS Institute, 1989). The following potential confounders in the statistical models were included: age (years), current smoker (yes/no), high alcohol intake (+20 U/week), occupation with suspected negative-reproductive effects (yes/no; exposure in the working environment to radiant heat, micro-waves, X-ray, welding, soldering, organic solvents, cutting-oils, lubricants, dyes, lead or pesticides), educational level (unskilled worker, skilled worker, academic degree, others) genital disorders (yes/no; one or both testicles not in the scrotum, treatment for non-descended testicle, genital infections, testis invasive mumps, operation performed on scrotum, penis or testicles or trauma to the testicles), fever prior to the semen analysis (yes/no) and fertility centre (four levels). Results Characteristics of the 2517 questionnaire respondents in the participant database are described in Table 1. Sperm motility and sperm morphology were unequally distributed in the four fertility centres. The results from the linear regression are shown in Table 2. The concentration of sperm cells was higher in all quintiles compared with the lowest quintile of the proportion of sedentary work, but not in a dose-response manner. None of these findings were, however, statistically significant. The log-transformation made the model sensitive to the distribution of very low sperm densities, but the use of an alternative transformation or exclusion of azoospermic men did not change the results substantially. The analyses with the logarithm of total sperm count as the dependent variable was also performed. These results were, however, not substantially different. The odds ratios for abnormal semen characteristics in the five quintiles of sedentary work are shown in Table 3. The risk of abnormal semen quality was lower in all quintiles compared with the lowest quintile of the proportion of sedentary work, but none of the findings were statistically significant. In couples where the spouse had a well-defined cause of infertility, the distribution of semen characteristics may be closer to the population values, but analyses stratified on the fertility status of the spouse only changed the results marginally. In addition, the stratification for infertility period, fertility centre and educational level did not substantially change the results. In sub-analyses the aim was to increase the precision of the estimates of the proportions of sedentary work by only including the observations with standard deviations below the 25 percentile of standard deviations in the quintile. The results were similar to the results in the main analyses (results not shown). In other sub-analyses observations with standard deviations between 0 and 95 percentile of standard deviations in the quintile were included to increase the statistical power of the analyses. The results were, however, not substantially different from the main analyses (results not shown). Discussion The results do not suggest a harmful effect of sedentary work on sperm density or other semen characteristics. The strength of this study was the relatively large sample and unbiased data on sedentary work position. One of the

4 8 J. Støy et al. Characteristic Aalborg (n ¼ 237) Aarhus (n ¼ 302) Odense (n ¼ 281) Sønderborg (n ¼ 141) Table 1. Characteristics of the study population by fertility centre Male age (years) and above Occupational group (by 1. digit of DISCO-68) a 0/1 (research, education, health) (management, clerical work) (trade, service trade) (agriculture, extraction of raw materials, industry, transportation, industry trade) Duration of infertility period (months) and above Unknown Sperm density (million/ml) < and above Sperm motility (percent motile sperm) < and above Sperm morphology (percent sperm with normal morphology) < and above a Danish occupational code from limitations was, however, the precision of the estimates of the proportions of sedentary work. In order to validate the information on job specific work position, the three proportions of work position was summed: sitting, walking and standing (Fig. 2). The expected value of the sum of proportions is 1. The most remarkable departure from this value was over reporting of walking and standing work position, while there were no indications of over- or underreporting of sedentary work. In order to increase precision of the estimates of sedentary work, job titles with standard deviations of sedentary work below the median of the standard deviation in the quintile were alone included. The choice of this threshold of standard deviations was an arbitrary trade off to ensure a sufficient number of job titles in the analyses and at the same time to promote the precision of the exposure estimate. As demonstrated, the results were, however, not very sensitive to the chosen threshold. Sedentary position at work was alone considered. The results may, thus, be confounded by the leisure time exposure to sedentary position. Predominant sedentary work position may combine with an active leisure time lifestyle,

5 Semen quality and sedentary work position 9 Table 2. Sperm concentration by level of sedentary work Proportion of sedentary work n Median sperm cell concentration (million/ml) Percentage difference in sperm cell concentration between men in first quintile of sedentary work and men in the other quintiles Crude (95% CI) Adjusted (95% CI) First quintile ( ) Reference Reference Second quintile ( ) () ) 11.5 () ) Third quintile ( ) () ) 17.9 () ) Fourth quintile ( ) () ) 31.9 () ) Fifth quintile ( ) () ) 18.7 () ) Continuous ( ) () ) 33.3 () ) CI, confidence interval. Table 3. Odds ratios for abnormal semen quality by level of sedentary work Proportion of sedentary work Abnormal semen quality (n) Normal semen quality (n) Odds ratio and 95% confidence interval Crude Adjusted First quintile ( ) Second quintile ( ) ( ) 0.60 ( ) Third quintile ( ) ( ) 0.63 ( ) Fourth quintile ( ) ( ) 0.56 ( ) Fifth quintile ( ) ( ) 0.65 ( ) Continuous ( ) ( ) 0.70 ( ) Figure 2. Distribution of summarized proportions of work positions in The Danish Work Environment Cohort study. whereas predominant standing work position may lead to more sedentary leisure time activities. The 24-h exposure to sedentary position could therefore add up to the same total for men with low and high proportions of sedentary work. To elucidate this potential problem, additional data from the DWECS 2000 was obtained (Burr, 2002). A cross-tabulation was made of the response levels in the question of sedentary work (six levels) and the leisure time activity (four levels) (Table 4). Statistically significant association was not found between the proportions of sedentary work and leisure time exercise. On the contrary, a weak association between no sedentary work and a high level of exercise in the leisure time was found (Pearson chi-square, 1 d.f., p ¼ 0.053). Thus, the findings seem not to be explained by a similar 24-h exposure to sedentary position among men with high and low proportions of sedentary work. Selection bias may seriously distort results in studies in sub-fertile populations. If the duration of infertility varies with the level of sedentary work: men with high proportions of sedentary work, who on average have a higher educational level, may only accept a shorter delay in conception before they attend an infertility work up than less-educated men with a low proportion of sedentary work (Collins et al., 1994). Men with long infertility periods who finally attend the fertility clinics are consequently less fertile than the men who attended the fertility clinic after a short infertility period. However, no differences was found in duration of infertility before referral in the five levels of sedentary work.

6 10 J. Støy et al. Table 4. Level of physical activity in the leisure time by level of sedentary work Time spend in sedentary position during a work day Almost no exercise in the leisure time 2 4 h/week of moderate exercise in the leisure time >4 h/week of moderate exercise in the leisure time >4 h/week of strenuous exercise in the leisure time Almost all the time % of the time % of the time % of the time Almost none None The mean sperm cell concentration for responders of the questionnaire was 69.2 million/ml and for non-responders 64.2 million/ml. Thus, the mean sperm cell concentration was not significantly different for responders and nonresponders, therefore, dropouts are not likely to bias the analyses. Sexual abstinence is an important determinant of sperm density (Raziel et al., 2001). Participants were asked to be sexually abstinent 3 4 days before the semen analyses, but there was no information on the actual duration of sexual abstinence prior to the semen analysis, and it might be unevenly distributed among those with low and high level of sedentary work. To elucidate the association between sedentary work and duration of sexual abstinence, additional data was obtained from The Danish First Pregnancy Planner study (Bonde et al., 1998). In this study there was information on both predominant work position and duration of sexual abstinence and no association between work position and sexual abstinence was found (p ¼ 0.69). Even if the period of sexual abstinence differed between the fertility centres this potential confounder is adjusted for when adjusting for fertility centre. In the majority of previous studies of sedentary work and semen quality, the results suggested no association between abnormal semen quality and sedentary work. Thus, the study confirms the findings of others (Buiatti et al., 1984; Oldereid et al., 1992; Bigelow et al., 1998; Hjollund et al., 2002b). In conclusion, no differences was found in sperm density or other semen characteristics between men in jobs with low proportions of sedentary work compared with men in jobs with high proportions of sedentary work. Acknowledgements The study was supported by the Danish Medical Research Council. Journal number Arbejdsdirektoratet (1986) Dansk Fagkode, 2 edn. Arbejdsdirektoratet, København. Auger, J., Kunstmann, J. M., Czyglik, F. & Jouannet, P. (1995) Decline in semen quality among fertile men in Paris during the past 20 years. The New England Journal of Medicine 332, Bigelow, P. L., Jarrell, J., Young, M. R., Keefe, T. J. & Love, E. J. (1998) Association of semen quality and occupational factors: comparison of case-control analysis and analysis of continuous variables. Fertility and Sterility 69, Bonde, J. P., Hjollund, N. H., Jensen, T. K., Ernst, E., Kolstad, H., Henriksen, T. B., Giwercman, A., Skakkebaek, N. E., Andersson, A. M. & Olsen, J. (1998) A follow-up study of environmental and biologic determinants of fertility among 430 Danish first-pregnancy planners: design and methods. Reproductive Toxicology 12, Borg, V., Kristensen, T. S. & Burr, H. (2000) Work environment and changes in self-rated health: a five-year follow-up study. Stress Medicine 16, Buiatti, E., Barchielli, A., Geddes, M., Nastasi, L., Kriebel, D., Franchini, M. & Scarselli, G. (1984) Risk factors in male infertility: a case-control study. Archives of Environmental Health 39, References Bujan, L., Daudin, M., Charlet, J. P., Thonneau, P. & Mieusset, R. (2000) Increase in scrotal temperature in car drivers. Human Reproduction 15, Burr, H. (2002) Arbejdsmiljøi Danmark 2000: en kortlægning af lønmodtageres og selvstændiges arbejdsmiljøog helbred. Arbejdsmiljørådets service center, Ramsingsvej, Valby, Denmark. Carlsen, E., Giwercman, A., Keiding, N. & Skakkebaek, N. E. (1992) Evidence for decreasing quality of semen during past 50 years. British Medical Journal 305, Collins, J. A., Burrows, E. A. & Willan, A. R. (1994) Occupation and the clinical characteristics of infertile couples. Canadian Journal of Public Health 85, Figa-Talamanca, I., Dell Orco, V., Pupi, A., Dondero, F., Gandini, L., Lenzi, A., Lombardo, F., Scavalli, P. & Mancini, G. (1992) Fertility and semen quality of workers exposed to high temperatures in the ceramics industry. Reproductive Toxicology 6, Figa-Talamanca, I., Cini, C., Varricchio, G. C., Dondero, F., Gandini, L., Lenzi, A., Lombardo, F., Angelucci, L., Di Grezia, R. & Patacchioli, F. R. (1996) Effects of prolonged autovehicle driving on male reproduction function: a study among taxi drivers. American Journal of Industrial Medicine 30,

7 Semen quality and sedentary work position 11 Hjollund, N. H., Bonde, J. P., Jensen, T. K. & Olsen, J. (2000) Diurnal scrotal skin temperature and semen quality. The Danish First Pregnancy Planner Study Team. International Journal of Andrology 23, Hjollund, N. H., Storgaard, L., Ernst, E., Bonde, J. P. & Olsen, J. (2002a) The relation between daily activities and scrotal temperature. Reproductive Toxicology 16, Hjollund, N. H., Storgaard, L., Ernst, E., Bonde, J. P. & Olsen, J. (2002b) Impact of diurnal scrotal temperature on semen quality. Reproductive Toxicology 16, Jannes, P., Spiessens, C., Van der Auwera, I., D Hooghe, T., Verhoeven, G. & Vanderschueren, D. (1998) Male subfertility induced by acute scrotal heating affects embryo quality in normal female mice. Human Reproduction 13, Loughlin, K. R., Manson, K., Foreman, R., Schwartz, B. & Heuttner, P. (1991) The effect of intermittent scrotal hyperthermia on the Sprague Dawley rat testicle. Advances in Experimental Medicine and Biology 286, Mieusset, R. & Bujan, L. (1995) Testicular heating and its possible contributions to male infertility: a review. International Journal of Andrology 18, Mieusset, R., Bujan, L., Mansat, A., Pontonnier, F. & Grandjean, H. (1987) Hyperthermia and human spermatogenesis: enhancement of the inhibitory effect obtained by Ôartificial cryptorchidismõ. International Journal of Andrology 10, Mieusset, R., Quintana Casares, P. I., Sanchez-Partida, L. G., Sowerbutts, S. F., Zupp, J. L. & Setchell, B. P. (1991) The effects of moderate heating of the testes and epididymides of rams by scrotal insulation on body temperature, respiratory rate, spermatozoa output and motility, and on fertility and embryonic survival in ewes inseminated with frozen semen. Annals of the New York Academy of Sciences 637, Mortensen, J. T. (1986) Erhvervsbetingede sædcellepåvirkninger. Arbejdsmiljørådets service center, Ramsingsvej, Valby, Denmark. Mortensen J. T. (1988) Risk for reduced sperm quality among metal workers, with special reference to welders. Scandinavian Journal of Work, Environment & Health 14, Nord-Larsen, M., Ørhede, E., Nielsen, J. & Burr, H. (1992) Lønmodtagernes arbejdsmiljø , report. Oldereid, N. B., Rui, H. & Purvis, K. (1992) Life styles of men in barren couples and their relationship to sperm quality. International Journal of Fertility 37, Raziel, A., Friedler, S., Schachter, M., Kaufman, S., Omanski, A., Soffer, Y. & Ron-El, R. (2001) Influence of a short or long abstinence period on semen parameters in the ejaculate of patients with nonobstructive azoospermia. Fertility and Sterility 76, SAS Institute (1989) SAS/STAT User s Guide, Version 6, Volume 2, 4th edn. SAS Institute Inc, Cary, NC, USA. Sas, M. & Szollosi, J. (1979) Impaired spermiogenesis as a common finding among professional drivers. Archives of Andrology 3, Shafik, A. (1991) Testicular suspension: effect on testicular function. Andrologia 23, Thonneau, P., Ducot, B., Bujan, L., Mieusset, R. & Spira, A. (1996) Heat exposure as a hazard to male fertility. Lancet 347, Received 21 January 2003; revised 12 May 2003; accepted 12 May 2003