Advertisement

Body mass index in relation to semen quality and reproductive hormonesamong 1,558 Danish men

      Objective

      To examine the relationship between body mass index (BMI) and semen quality among young men from the general population.

      Design

      Cross-sectional study.

      Setting

      Danish young men were approached when they attended a compulsory physical examination to determine their fitness for military service.

      Patient(s)

      From 1996–1998, 1,558 (19%) young men (mean age 19 years) volunteered.

      Main outcome measure(s)

      Semen volume (in milliliters), sperm concentration (in million per milliliter), percentage of motile spermatozoa, percentage of spermatozoa with normal morphology, total sperm count (in million), and testis size (in milliliters). In addition, serum reproductive hormones were measured.

      Result(s)

      Serum T, sex hormone-binding globulin (SHBG), and inhibin B all decreased with increasing BMI, whereas free androgen index and E2 increased with increasing BMI. Serum FSH was higher among slim men. After control for confounders, men with a BMI <20 kg/m2 had a reduction in sperm concentration and total sperm count of 28.1% (95% confidence interval [CI] 8.3%–47.9%) and 36.4% (95% CI 14.6%–58.3%), respectively, and men with a BMI >25 kg/m2 had a reduction in sperm concentration and total sperm count of 21.6% (95% CI 4.0%–39.4%) and 23.9% (95% CI 4.7%–43.2%), respectively, compared to men with BMI between 20–25 kg/m2. Percentages of normal spermatozoa were reduced, although not significantly, among men with high or low BMI. Semen volume and percentage of motile spermatozoa were not affected by BMI.

      Conclusion(s)

      High or low BMI was associated with reduced semen quality. It remains to be seen whether the increasing occurrence of obesity in the Western world may contribute to an epidemic of poor semen quality registered in some of the same countries. If so, some cases of subfertility may be preventable.

      Key words

      Recent studies have shown that low sperm count is a relatively common symptom among young Danish men (
      • Andersen A.G.
      • Jensen T.K.
      • Carlsen E.
      • Jorgensen N.
      • Andersson A.M.
      • Krarup T
      • et al.
      High frequency of sub-optimal semen quality in an unselected population of young men.
      ) who are also at increased risk of other reproductive health problems, including high and increasing risk of testicular cancer (
      • Adami H.-O.
      • Bergström R.
      • Möhner M.
      • Zatonski W.
      • Storm H.
      • Ekbom A
      • et al.
      Testicular cancer in nine Northern European countries.
      ,
      • Bergström R.
      • Adami H.-O.
      • Möhner M.
      • Zatonski W.
      • Storm H.
      • Ekbom A
      • et al.
      Increase in testicular cancer incidence in six European countries: a birth cohort phenomenon.
      ). The etiology of the adverse trends in male reproductive health problems is not known, although environmental factors must be suspected to contribute due to the rapid changes in frequencies of these reproductive problems (
      • Sharpe R.M.
      • Skakkebæk N.E.
      Are oestrogens involved in falling sperm counts and disorders of the male reproductive tract?.
      ).
      An increase in average body mass index (BMI) resulting in increased prevalence of obesity is another recent observation in the Western world (
      • Mokdad A.H.
      • Bowman B.A.
      • Ford E.S.
      • Vinicor F.
      • Marks J.S.
      • Koplan J.P.
      The continuing epidemics of obesity and diabetes in the United States.
      ). Obesity has previously been related to reduced female fertility with prolonged waiting time to pregnancy among women with a BMI >25 kg/m2 (
      • Zaadstra B.M.
      • Seidell J.C.
      • Van Noord P.A.H.
      • te Velde E.R.
      • Habbema J.D.F.
      • Vrieswijk B
      • et al.
      Fat and female fecundity: prospective study of effect of body fat distribution on conception rates.
      ). There are, however, to our knowledge no studies investigating the relationship between BMI and male fertility and semen quality.
      We therefore decided to study the association between BMI and semen quality in young men from the general population.

      Materials and methods

       Populations

      Due to the military drafting system in Denmark, all 18-year-old men are required to attend a compulsory physical examination to determine their fitness for military service. Trained staff personally approached all young men from the general population when they presented for this compulsory physical examination in two cities in Denmark (Copenhagen and Aalborg) from June 1996 to March 1998. Men who suffered from chronic diseases (corresponding to 10%–14% of the population) were not summoned to the military board and therefore not approached in this study.
      The trained staff informed the men about the study and handed out written information. The men could give informed consent for participation and book an appointment immediately or return the consent by mail. For a young man to participate in the study, he and his mother had to be born and raised in Denmark. The participants received economic compensation ($80). The mean participation rate in the centers was 19%. Ethical approval was obtained from the local ethical committee, which is common practice for clinical studies in Denmark. Therefore, no Institutional Review Board approval was obtained. The procedures were in accordance with the Helsinki Declaration of 1975, as revised in 1983. A detailed description of the study has been given elsewhere (
      • Andersen A.G.
      • Jensen T.K.
      • Carlsen E.
      • Jorgensen N.
      • Andersson A.M.
      • Krarup T
      • et al.
      High frequency of sub-optimal semen quality in an unselected population of young men.
      ).

       Semen analysis

      Each man provided a semen sample by masturbation into a wide mouthed plastic container in a room close to the semen laboratory. The period of abstinence was recorded and the semen sample was analyzed according to the World Health Organization 1992 guidelines (
      World Health Organization
      WHO laboratory manual for the examination of human semen and sperm-cervical mucus interaction.
      ) modified in accordance with Jørgensen et al. (
      • Jørgensen N.
      • Auger J.
      • Giwercman A.
      • Irvine D.S.
      • Jensen T.K.
      • Jouannet P
      • et al.
      Semen analysis performed by different laboratory teams: an intervariation study.
      ). From all semen samples a smear was stained and preserved, and the same physician assessed all morphology slides by the use of strict criteria method (
      • Jensen T.K.
      • Scheike T.
      • Keiding N.
      • Schaumburg I.
      • Grandjean P.
      Fecundability in relation to body mass and menstrual cycle patterns.
      ). The investigator, who was blinded, evaluated 90% of the morphology smears. The remaining smears were either damaged by transport or were fixated insufficiently.
      A previous study found interobserver variability in semen analysis between different laboratories, especially concerning motility assessments (
      • Jørgensen N.
      • Auger J.
      • Giwercman A.
      • Irvine D.S.
      • Jensen T.K.
      • Jouannet P
      • et al.
      Semen analysis performed by different laboratory teams: an intervariation study.
      ). In the present study only three technicians from our laboratory were involved in the semen analyses. Our technicians performed a monthly intralaboratory control by examining the same semen samples blindly and the between-technician variation in sperm concentration was <10% throughout the study period. The following semen variables were used as outcome variables: semen volume (in milliliters), sperm concentration (in million per milliliter), percentage of motile spermatozoa, percentage of spermatozoa with normal morphology, and the calculated total sperm count (concentration × volume, in million).

       Hormone analysis

      A blood sample was withdrawn from a cubital vein of each participant, centrifuged, and the serum was separated and frozen. All samples were analyzed at Department of Growth and Reproduction, Rigshospitalet, Denmark. Serum levels of FSH, LH, and sex hormone-binding globulin (SHBG) were determined using a time-resolved immunofluorometric assay (Delfia, Wallac, Turku, Finland). Testosterone levels were determined using a time-resolved flouroimmunoassay (Delfia), E2 by RIA (Pantex, Santa Monica, CA) and inhibin B by a specific two-sided enzyme immunometric assay (Serotec,UK). Intra- and interassay coefficients of variation (CV) for measurements of both FSH and LH were 3% and 4.5%, respectively. The CV for T and SHBG were <8% and <5%, respectively. The intra- and interassay CV for E2 and inhibin B were 7.5% and 13%, and 15% and 18%, respectively. The free T index was calculated as (total T × 100)/SHBG.

       Physical examination

      All physical examinations were performed by three physicians (two in Copenhagen and one in Aalborg). Tanner stage of pubic hair and genital development, testicular volumes (determined by use of a Prader orchidometer and mean of both testes was calculated), and the possible presence of a varicocele, a hydrocele, the location of testis in scrotum, and the consistency of the testis and epididymis were recorded. Weight was measured in kilograms using only one weighing scale in each center, height was measured in centimeters, and the BMI was calculated as weight in kilograms divided by the squared height in meters. The BMI was categorized as <20 kg/m2, 20–25 kg/m2 and >25 kg/m2, as normally described as under-, normal, and overweight in the literature (
      • Cnattingius S.
      • Bergstrom R.
      • Lipworth L.
      • Kramer M.S.
      Prepregnancy weight and the risk of adverse pregnancy outcomes.
      ).

       Questionnaire

      All participants completed a questionnaire that was handed out to them at the day of the military examination. It was returned to the physician at the time of physical examination. The questionnaire included information on previous or current diseases and genital diseases such as cryptorchidism, inguinal hernia, varicocele, epididymitis, gonorrhea, chlamydia, and operation for torsion of the testis.
      The men were asked if they were still in school and if not, what was their highest education and current work situation. They reported on smoking and alcohol intake during the week before completion of the questionnaire. Smoking habits were reported as the average number of cigarettes, cigars, or pipes smoked per day. The total weekly alcohol intake (number of drinks) was calculated by summarizing the beer, wine, and liquor intake. In addition, they were asked whether their mothers' smoked during pregnancy with them and if their parents smoked at home (answer categories; none smoked; one smoked; both smoked).
      Diseases in the reproductive organs (found at the physical examination or from questionnaires) found to affect any of the outcome variables separately were transformed into one variable for each outcome (present or not present).

       Statistics

      The BMI was categorized as <20 kg/m2, between 20–25 kg/m2 (reference), and >25 kg/m2 or in categories of one from below 19 kg/m2 to >29 kg/m2 (27–28.99 kg/m2 combined because of small sample size) with BMI from 22–22.99 kg/m2 as reference. Furthermore, it was entered as a continuous variable including a squared and a third potency BMI.
      The distribution of information obtained from the questionnaire and physical examination was compared for different BMI groups to determine possible confounders. Finally, a multiple linear regression analysis was performed taking into account confounders found to affect outcome variables and be differently distributed among men in different BMI groups. The BMI was entered into the model as dummy variables or as a continuous variable. Normally distributed outcome variables were entered directly as continuous variables in the linear multiple regression analysis, whereas sperm concentration and total sperm count were transformed by use of the natural logarithm to obtain normality. Confounders were excluded stepwise if they were not statistically significant at the 10% level. The results are presented with 95% confidence intervals (95% CI). The fit of the regression models was evaluated by testing the residuals for normality and by inspecting the residual plots.
      Finally, the distribution of BMI was tested by entering BMI as a continuous variable and entering squared and a third potency BMI in the model.

      Results

      A total of 1,558 men participated. Mean testis size was 19.8 mL (standard deviation 4.6 mL), mean semen volume was 3.2 mL (SD 1.5 mL), median sperm concentration and total sperm count were 44 million/mL (25–75 percentile, 21–79 million/mL) and 128 million/mL (55–246 million/mL), respectively, whereas mean percentages of motile spermatozoa and normal forms were 65.2% (12.7%) and 7.2% (4.8%), respectively. Mean serum LH, T, SHBG, inhibin B, and E2 were, respectively, 3.91 IU/L (1.59), 24.4 nmol/L (6.6), 29.8 nmol/L (11.3), 214 pg/mL (80), 85.8 pmol/L (28.7), and median FSH and free androgen index were 2.90 IU/L (2.10–4.22) and 83.6 (67.5–104.7). Median BMI was 22.4 kg/m2 (20.4 –24.3 kg/m2).
      Sperm concentration, total sperm count, the percentages of normal forms and motile spermatozoa were reduced if a varicocele was found at the physical examination, the testes were hard or soft, or the man reported being treated for cryptorchidism and this information was transformed into one variable (present or not). Testis size was affected if the man reported having had cryptorchidism, whereas semen volume was not significantly affected by any of these conditions.
      Men with BMI <20 kg/m2 or >25 kg/m2 had lower sperm concentration, total sperm count, and fewer normal forms (Table 1, Fig. 1). Among men with high or low BMI, 24.4% and 29.0% had sperm count <20 million/mL (lower WHO limit) compared to 21.7% of men with BMI between 20–25 kg/m2. Testis size and semen volume were smaller and percentages of motile spermatozoa were lower among men with low BMI (Table 1), but not among men with high BMI compared to men with a normal BMI. Men with high or low BMI had longer periods of abstinence and more often delivered their samples from April to September. They had more cases of diseases in reproductive organs, were more often not in school or had been exposed to smoking in utero or childhood (Table 1) to a higher extent than men with normal BMI. More men with BMI <20 kg/m2 smoked and the slim men were younger. The men with BMI >25 kg/m2 were older and less often smoked.
      TABLE 1The median sperm concentration and total sperm count (25–75 percentiles), mean percent of motile and morphological normal sperm, semen volume and testis size (standard deviation), reproductive hormones, and information obtained from questionnaires among 1,558 Danish military conscripts with different BMI.
      VariableNumber (total 1,558)%BMI (kg/m2)
      <20 n = 21720–25 n = 1,042⩾25 n = 299
      Mean testis size (mL)
      Computed as yes if condition found in either left or right testis.
      1,54799.318.2 (4.4)20.0 (4.4)20.1 (5.0)
      Mean semen volume (mL)1,558100.03.0 (1.5)3.2 (1.4)3.2 (1.6)
      Median sperm concentration (million/mL)1,558100.040.0 (17–75)46.0 (23–84)39.0 (20–69)
      Median total sperm count (million)1,558100.0105 (47–240)138.0 (59–259)116.0 (46–213)
      Mean motile sperm (%)1,54599.263.7 (14.5)65.4 (12.4)65.5 (12.5)
      Mean morphologically normal forms (%)1,39789.76.8 (4.4)7.4 (5.0)7.1 (4.6)
      Period of abstinence (hours)
       0–951,24780.575.581.879.5
       96+30319.524.518.220.5
      Season
       April–September18311.715.710.812.0
       October–March1,37588.384.389.288.0
      Diseases in reproductive organs
      Varicocele or abnormally hard or soft testes found at physical examination or questionnaire information about cryptorchidism.
       No1,31686.080.487.485.2
       Yes21414.019.612.614.8
      Reproductive hormones
       Median FSH (IU/L)1,55099.53.14 (2.19–4.96)2.89 (2.11–4.15)2.83 (2.00–4.13)
       Mean LH (IU/L)1,55099.53.78 (1.60)3.92 (1.60)3.98 (1.55)
       Mean T (nmol/L)1,55099.525.2 (6.6)24.8 (6.6)22.1 (5.8)
       Mean SHBG (nmol/L)1,55099.534.7 (11.8)30.5 (10.9)23.9 (10.0)
       Median free androgen index1,55099.573.3 (59.9–95.4)83.0 (67.3–102.5)95.5 (76.6–122.1)
       Mean inhibin B (pg/mL)1,54999.4219 (76)217 (81)202 (75)
       Mean E2 (pmol/L)95461.279.7 (28.4)85.2 (28.6)92.7 (27.9)
      Questionnaire information
      City
       Copenhagen1,43291.993.592.588.6
       Aalborg1268.16.57.511.4
      Age (years)
       <201,17076.084.575.571.5
       20+37024.015.524.528.5
      Still in school
       No60939.443.736.845.8
       Yes93560.656.363.254.2
      Alcohol intake (U/week)
       0–191,06973.274.473.172.7
       20+39126.825.626.927.3
      Smoking
       No91260.445.661.965.5
       Yes59939.654.438.134.5
      Exposures in utero and childhood
      Exposed to smoking in utero
       No85559.856.761.456.8
       Yes57440.243.338.643.2
      Exposed to smoking in childhood
       No37324.317.328.116.3
       One parent smoked56937.141.135.340.5
       Two parents smoked59238.641.636.643.2
      a Computed as yes if condition found in either left or right testis.
      b Varicocele or abnormally hard or soft testes found at physical examination or questionnaire information about cryptorchidism.
      Figure thumbnail gr1
      FIGURE 1The relationship between BMI and sperm concentration (A) and total sperm count (B).
      Serum reproductive hormones were related to BMI (Table 1). Serum T, SHBG, and inhibin B all decreased significantly with increasing BMI, whereas free androgen index and E2 increased significantly with increasing BMI. Serum FSH and SHBG were significantly higher among slim men, whereas E2 and free androgen index significantly decreased. These results were not adjusted as all the men were examined before lunch time.
      Serum reproductive hormones were also related to semen quality. A significant correlation between serum inhibin B, LH, FSH, and all semen parameters was found. Free androgen index and serum T correlated only with percentages of motile spermatozoa, whereas serum E2 did not correlate with any semen parameters.
      Linear regression was performed with the semen quality and testis size as dependent variables and BMI categorized and inserted as dummy variables as explanatory variables. Before adjustment, men with a BMI <20 kg/m2 had smaller testis size (Table 2). Men with high or low BMI had lower concentration and lower total sperm count (Table 2).
      TABLE 2Semen quality and testis size among men with BMI >25 or <20 kg/m2 compared to men with BMI between 20–25 kg/m2.
      Body mass index (kg/m2)Reduction in testes size (mL)Reduction in semen volume (mL)Reduction in sperm concentration (in %)Reduction in total sperm count (in %)Reduction in percentage of motile spermatozoaReduction in percentage of normal forms
      Unadjusted:
       <20−1.8 (−2.5;−1.2)−0.21 (−0.43;0.01)−25.3 (−46.4;−6.0)−34.1 (−55.4;−12.8)−1.67 (−3.54;0.20)−0.51 (−1.25;0.24)
       20–24.99ReferenceReferenceReferenceReferenceReferenceReference
       ≥250.1 (−0.5;0.7)0.003 (−0.19;0.19)−21.1 (−38.1;−4.1)−24.9 (−43.7;−6.2)0.12 (−1.52;1.77)−0.27 (−0.93;0.39)
      Adjusted:
       <20−2.0 (−2.7;−1.3)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      −0.24 (−0.45;−0.02)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −28.1 (−47.9;−8.3)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −36.4 (−58.3;−14.6)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −1.68 (−3.65;0.29)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.26 (−1.00;0.49)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       20–24.99ReferenceReferenceReferenceReferenceReferenceReference
       ≥250.1 (−0.5;0.7)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      −0.003 (−0.19;0.18)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −21.6 (−39.4;−4.0)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −23.9 (−43.2;−4.7)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      0.05 (−1.78;1.67)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.37 (−1.03;0.29)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
      Adjusted:
       <19−3.9 (−5.0;−2.8)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      −0.19 (−0.53;0.16)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −28.8 (−61.1;3.5)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −31.3 (−66.9;4.2)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −0.36 (−3.57;2.84)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.43 (−1.64;0.79)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       19–19.99−2.4 (−3.4;−1.5)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      −0.27 (−0.57;0.04)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −30.8 (−59.2;−2.3)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −40.2 (−71.6;−8.8)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −1.79 (−4.61;1.02)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.23 (−1.31;0.86)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       20–20.99−2.2 (−3.0;−1.4)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      0.04 (−0.23;0.31)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −14.2 (−39.1;10.7)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −10.9 (−38.3;16.5)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −0.40 (−2.84;2.04)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.13 (−1.07;0.81)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       21–21.99−1.4 (−2.2;−0.6)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      0.007 (−0.24;0.26)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      5.1 (−18.1;28.3)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      6.6 (−18.8;32.0)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      0.69 (−1.59;2.97)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      0.28 (−0.60;1.16)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       22–22.99ReferenceReferenceReferenceReferenceReferenceReference
       23–23.99−0.8 (−1.6;0.08)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      0.003 (−0.26;0.27)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      0.4 (−24.8;25.7)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      5.8 (−21.8;33.5)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      1.50 (−0.98;3.98)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.02 (−0.97;0.92)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       24–24.99−0.7 (−1.7;0.2)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      −0.08 (−0.37;0.23)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −3.3 (−31.4;24.8)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −4.0 (−34.8;26.8)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      1.07 (−1.68;3.82)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.69 (−1.77;0.39)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       25–25.99−1.4 (−2.5;−0.34)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      −0.0003 (−0.34;0.34)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −36.1 (−68.6;−3.7)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −36.5 (−71.6;−1.0)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −2.18 (−5.34;0.99)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.67 (−1.91;0.58)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       26–26.990.3 (−1.0;1.5)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      −0.03 (−0.42;0.36)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −19.0 (−55.8;17.8)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −18.6 (−58.9;21.7)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      1.79 (−1.83;5.41)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      0.02 (−1.32;1.36)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       27–28.99−1.4 (−2.6;−0.1)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      0.06 (−0.33;0.44)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −36.3 (−72.8;0.3)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −38.5 (−78.6;1.5)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      1.18 (−2.40;4.75)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.89 (−2.31;0.53)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
       ≥29−1.0 (−2.2;0.2)
      Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      −0.06 (−0.44;0.31)
      Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      −2.0 (−37.0;33.0)
      Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      −2.4 (−41.1;36.3)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      1.45 (−1.98;4.89)
      Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      −0.38 (−1.69;0.94)
      Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
      Note: Both unadjusted and adjusted results from linear regression analyses are shown with 95% confidence intervals (95% CI) (a reduction in sperm concentration of 17.1 and in motility of −1.68% means that men with BMI <20 kg/m2 had 17.1% lower concentration and 1.68% point fewer motile spermatozoa, respectively, compared to men with BMI between 20–25 kg/m2).
      a Adjusted for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism.
      b Adjusted for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm).
      c Adjusted for diseases in reproductive organs, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      d Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and period of abstinence (transformed by the use of the natural logarithm).
      e Adjusted for diseases in reproductive organs, age, in utero exposure to smoking, and time from ejaculation until evaluation (transformed by the use of the natural logarithm).
      f Adjusted for diseases in reproductive organs and age. Mean percentage of normal form was 7.2
      After control for diseases in reproductive organs, in utero exposure to smoking, period of abstinence (transformed by the use of the natural logarithm), men with a BMI <20 kg/m2 had a reduction in sperm concentration and total sperm count of 28.1% (8.3%, 47.9%) and 36.4% (14.6%, 58.3%), respectively (the reduction in total sperm count was also adjusted for age). Men with a BMI >25 kg/m2 had a reduction in sperm concentration and total sperm count of 21.6% (4.0%; 39.4%) and 23.9% (4.7%; 43.2%), respectively, compared to men with BMI between 20–25 kg/m2. Men with a BMI <20 kg/m2 had 2.0 mL (1.3–2.7 mL) smaller testis size after control for center of investigation, in utero exposure to smoking, still in school, and cryptorchidism and 0.24 mL (0.02–0.45 mL) smaller semen volume after control for center of investigation, age, and period of abstinence (transformed by the use of the natural logarithm). Percentage of motile and morphological normal spermatozoa were not affected by BMI.
      The BMI was entered in categories of one from <19 kg/m2 to >29 kg/m2 (27–28.99 kg/m2 combined because of small sample size) with BMI from 22–22.99 kg/m2 as reference (Table 2). Testis size, sperm morphology, concentration, and total sperm count showed an inverted U-shaped distribution with lower values for lower and higher BMIs, whereas no U shape was found for volume and motility.
      Furthermore, BMI entered the model as a continuous variable including BMI squared and raised to the third power. All three BMIs were significant at the 10% level for sperm concentration, total sperm count, and testis size suggesting that the U-shaped relationship is a good model for describing the relationship between these semen parameters and BMI.
      The analyses were repeated for men without diseases in reproductive organs and for men with a period of abstinence >96 hours, and the same associations were found.

      Discussion

      We found a significant association between sperm count and BMI, as overweight as well as slim men had lower sperm concentrations and also lower total sperm counts compared to men with ideal weights (BMI between 20–25 kg/m2). Previous studies have shown similar associations between female fecundity and BMI with longer waiting time to pregnancy, increased risk of irregular menstrual cycles, preterm delivery, and stillbirths (
      • Jensen T.K.
      • Scheike T.
      • Keiding N.
      • Schaumburg I.
      • Grandjean P.
      Fecundability in relation to body mass and menstrual cycle patterns.
      ,
      • Cnattingius S.
      • Bergstrom R.
      • Lipworth L.
      • Kramer M.S.
      Prepregnancy weight and the risk of adverse pregnancy outcomes.
      ,
      • Lake J.K.
      • Power C.
      • Cole T.J.
      Women's reproductive health: the role of body mass index in early and adult life.
      ,
      • Hartz A.J.
      • Barboriak P.N.
      • Wong A.
      • Katayama K.P.
      • Rimm A.A.
      The association of obesity with infertility and related menstrual abnormalities in women.
      ,
      • Grodstein F.
      • Goldman M.B.
      • Cramer D.W.
      Body mass index and ovulatory infertility.
      ,
      • Pasquali R.
      • Pelusi C.
      • Genghini S.
      • Cacciari M.
      • Gambineri A.
      Obesity and reproductive disorders in women.
      ) in women with less ideal weights.
      It is also well-documented that BMI is correlated to other human health risks. Many studies have shown that both high and low BMI is associated with increased mortality and morbidity (
      • Singh P.N.
      • Lindsted K.D.
      Body mass and 26-year risk of mortality from specific diseases among women who never smoked.
      ,
      • Troiano R.P.
      • Frongillo E.A.
      • Sobal J.
      • Levitsky D.A.
      The relationship between body weight and mortality: a quantitative analysis of combined information from existing studies.
      ). Low BMI can result from a “healthy lifestyle” but may also be due to many chronic diseases. Although the participants were recruited from the general population, participants with slight malnutrition or subclinical adverse conditions, which may affect their reproductive health, may be over-represented in the group of underweight men. In contrast, it is difficult to disregard obesity as a causal factor for the reduced semen quality in the group of men with BMI >25 kg/m2. High BMI is not only associated with increased risk of cardiovascular disease (
      • Must A.
      • Jacques P.F.
      • Dallal G.E.
      • Bajema C.J.
      • Dietz W.H.
      Long-term morbidity and mortality of overweight adolescents—a follow-up of the Harvard Growth Study of 1922 to 1935.
      ) and type 2 diabetes (
      • Hu F.B.
      • Manson J.E.
      • Stampfer M.J.
      • Colditz G.
      • Liu S.
      • Solomon C.G
      • et al.
      Diet, lifestyle, and the risk of type 2 diabetes mellitus in women.
      ), but also with other endocrine symptoms including changes in reproductive hormones (discussed later).
      High or low BMI was associated with a significantly reduced sperm number. The BMI was, however, not associated with poor sperm motility and morphology. Several studies have shown that sperm concentration is correlated to waiting time to pregnancy; among partners of pregnant women, the probability of achieving a pregnancy increased with increasing sperm concentration up to 55 million/mL (
      • Slama R.
      • Eustache F.
      • Ducot B.
      • Jensen T.K.
      • Jorgensen N.
      • Horte A
      • et al.
      Time to pregnancy and semen parameters: a cross-sectional study among fertile couples from four European cities.
      ) and among couples trying to conceive for the first time, the probability of succeeding increased with sperm count up to 40 million/mL (
      • Bonde J.P.
      • Ernst E.
      • Jensen T.K.
      • Hjollund N.H.I.
      • Kolstad H.
      • Henriksen T.B
      • et al.
      The relation between semen quality and fertility.
      ). A substantial number of obese men may have less than 40 million spermatozoa/mL resulting in less ability to conceive and perhaps, subfertility. The current epidemic of obesity, which is of concern due to its risks of increases in the connected lifestyle diseases, may therefore also contribute to prevalence of subfertility among populations in the Western world.
      The pattern in reproductive hormones differed between the different BMI groups, and it could be speculated that hormonal imbalance may be involved in the decreased semen quality in men with BMI <20 kg/m2 or >25 kg/m2, although the mechanisms behind the decreased semen quality in the slim and overweight men may not be the same. In accordance with several previous studies of reproductive hormones in relation to BMI (
      • Giagulli V.A.
      • Kaufman J.M.
      • Vermeulen A.
      Pathogenesis of the decreased androgen levels in obese men.
      ,
      • Gyllenborg J.
      • Rasmussen S.L.
      • Borch-Johnsen K.
      • Heitmann B.L.
      • Skakkebaek N.E., Juul A.
      Cardiovascular risk factors in men: the role of gonadal steroids and sex hormone-binding globulin.
      ,
      • Vermeulen A.
      • Kaufman J.M.
      • Giagulli V.A.
      Influence of some biological indexes on sex hormone-binding globulin and androgen levels in aging or obese males.
      ), we observed significantly decreased SHBG and T levels in men with BMI ≥25 kg/m2. Especially SHBG levels were strongly negatively associated with BMI. The regulation of SHBG is multifactorial, but there is now much evidence that insulin may have an inhibitory effect on SHBG synthesis (
      • Hautanen A.
      Synthesis and regulation of sex hormone-binding globulin in obesity.
      ,
      • Gascon F.
      • Valle M.
      • Martos R.
      • Ruz F.J.
      • Rios R.
      • Montilla P
      • et al.
      Sex hormone-binding globulin as a marker for hyperinsulinemia and/or insulin resistance in obese children.
      ). As total T and SHBG decreased concurrently with increasing BMI, the free androgen index levels were, however, not decreased but actually higher in overweight men compared to men with normal weight. The fact that both total T and SHBG (but not free T) were decreased in obese men has led to the suggestion that the decrease in total T levels is a consequence of the (insulin-induced) decrease in SHBG to sustain a free T homeostasis (
      • Raman J.D.
      • Schlegel P.N.
      Aromatase inhibitors for male infertility.
      ). However, although circulating levels of bioavailable T are sustained in obese men, it might be speculated that a decreased total T production nevertheless could result in decreased intratesticular T levels, which might affect the function of the seminiferous epithelium.
      The overweight men also had increased E2 levels and thus a reduced T/E2 ratio. Low serum T/E2 ratios are often seen in infertile men and it has been suggested that the combination of low serum T and increased E2 may adversely affect spermatogenesis. Treatment of infertile men showing low T/E2 ratios with aromatase inhibitors has led to normalization of the hormonal profile and in some oligospermic men to improvement of sperm quality (
      • Pavlovich C.P.
      • King P.
      • Goldstein M.
      • Schlegel P.N.
      Evidence of a treatable endocrinopathy in infertile men.
      ,
      • Raman J.D.
      • Schlegel P.N.
      Aromatase inhibitors for male infertility.
      ). However, no difference between the three BMI groups regarding the free androgen index/E2 ratio was observed, thus it seems unlikely that the ratio between circulating bioavailable T and E2 could be causing the decreased sperm concentration. Reproductive hormone levels have been shown to normalize after weight loss (
      • Kaukua J.
      • Pekkarinen T.
      • Sane T.
      • Mustajoki P.
      Sex hormones and sexual function in obese men losing weight.
      ,
      • Vermeulen A.
      • Kaufman J.M.
      • Giagulli V.A.
      Influence of some biological indexes on sex hormone-binding globulin and androgen levels in aging or obese males.
      ). It remains, however, to be seen whether weight loss may also improve semen quality.
      Other factors are likely to be involved in the decreased spermatogenesis in underweight men, as we found increased SHBG levels and consequently significantly lower free androgen index levels than men with a BMI ≥20 kg/m2. They also had significantly lower E2 levels. However, a link between this hormonal pattern and decreased spermatogenesis is not obvious.
      Our participation rate was 20%, which is higher or at the same level as similar studies of semen quality among men from the general population (
      • Swan S.H.
      • Brazil C.
      • Drobnis E.Z.
      • Liu F.
      • Kruse R.L.
      • Hatch M
      • et al.
      Geographic differences in semen quality of fertile U.S. males.
      ,
      • Jorgensen N.
      • Andersen A.G.
      • Eustache F.
      • Irvine D.S.
      • Suominen J.
      • Petersen J.H
      • et al.
      Regional differences in semen quality in Europe.
      ,
      • Jorgensen N.
      • Carlsen E.
      • Nermoen I.
      • Punab M.
      • Suominen J.
      • Andersen A.G
      • et al.
      East–West gradient in semen quality in the Nordic-Baltic area: a study of men from the general population in Denmark, Norway, Estonia and Finland.
      ). We were concerned whether the participants were selected in some way and therefore conducted a study where reproductive hormones among participants and nonparticipants were compared (
      • Andersen A.G.
      • Jensen T.K.
      • Carlsen E.
      • Jorgensen N.
      • Andersson A.M.
      • Krarup T
      • et al.
      High frequency of sub-optimal semen quality in an unselected population of young men.
      ). We found no significant difference with regard to reproductive hormones in the two groups, indicating that our participants represented the general population with respect to reproductive health. In addition, this study compared semen quality in groups of men with different BMI, and it is therefore of less importance whether the groups of men in fact represented the general population.
      Height and weight were measured by instruments installed in each center. Small weight differences between centers may therefore be due to measurement errors. The mean weight was slightly higher in the center in Aalborg compared to Copenhagen, but no differences in height between centers were observed. However, the relation between BMI and semen quality was found in each center separately and we adjusted for center differences in the multiple regression analyses.
      The BMI is a measure of weight in relation to height and does not directly reflect the percentage of body fat. Some of the men with a high BMI may in fact not be obese but have a large muscle mass, which may not be associated with reduced semen quality. However, the lower serum T level among the men with high BMI shows that the high BMI is probably due to high body fat level and not muscles. If, however, some men have been misclassified as obese although they have large muscle mass this will be a nondifferential misclassification and tend to underestimate a true effect. It can therefore not explain our findings.
      Differences in the distribution of confounders as, for example, period of abstinence, diseases in reproductive organs, season, or smoking may influence our finding. We did, however, take into account these factors in the regression analyses and they are unlikely to explain our findings.
      In conclusion, body mass within the ideal “normal” range was associated with higher sperm concentration, higher total sperm count, and a lower percentage of abnormal spermatozoa. Sperm motility and semen volume was not affected by BMI. It remains to be seen whether the increasing occurrence of obesity in the Western world may contribute to an explanation for the recently reported subnormal sperm counts registered in some of the same countries where obesity has increased. If so, some cases of subfertility may be preventable.

      References

        • Andersen A.G.
        • Jensen T.K.
        • Carlsen E.
        • Jorgensen N.
        • Andersson A.M.
        • Krarup T
        • et al.
        High frequency of sub-optimal semen quality in an unselected population of young men.
        Hum Reprod. 2000; 15: 366-372
        • Adami H.-O.
        • Bergström R.
        • Möhner M.
        • Zatonski W.
        • Storm H.
        • Ekbom A
        • et al.
        Testicular cancer in nine Northern European countries.
        Int J Cancer. 1994; 59: 33-38
        • Bergström R.
        • Adami H.-O.
        • Möhner M.
        • Zatonski W.
        • Storm H.
        • Ekbom A
        • et al.
        Increase in testicular cancer incidence in six European countries: a birth cohort phenomenon.
        J Natl Cancer Inst. 1996; 88: 727-733
        • Sharpe R.M.
        • Skakkebæk N.E.
        Are oestrogens involved in falling sperm counts and disorders of the male reproductive tract?.
        Lancet. 1993; 341: 1392-1395
        • Mokdad A.H.
        • Bowman B.A.
        • Ford E.S.
        • Vinicor F.
        • Marks J.S.
        • Koplan J.P.
        The continuing epidemics of obesity and diabetes in the United States.
        JAMA. 2001; 286: 1195-1200
        • Zaadstra B.M.
        • Seidell J.C.
        • Van Noord P.A.H.
        • te Velde E.R.
        • Habbema J.D.F.
        • Vrieswijk B
        • et al.
        Fat and female fecundity: prospective study of effect of body fat distribution on conception rates.
        Br Med J. 1993; 306: 484-487
        • World Health Organization
        WHO laboratory manual for the examination of human semen and sperm-cervical mucus interaction.
        3 ed. Cambridge University Press, Cambridge1992
        • Jørgensen N.
        • Auger J.
        • Giwercman A.
        • Irvine D.S.
        • Jensen T.K.
        • Jouannet P
        • et al.
        Semen analysis performed by different laboratory teams: an intervariation study.
        Int J Androl. 1997; 20: 201-208
        • Jensen T.K.
        • Scheike T.
        • Keiding N.
        • Schaumburg I.
        • Grandjean P.
        Fecundability in relation to body mass and menstrual cycle patterns.
        Epidemiology. 1999; 10: 422-428
        • Cnattingius S.
        • Bergstrom R.
        • Lipworth L.
        • Kramer M.S.
        Prepregnancy weight and the risk of adverse pregnancy outcomes.
        N Engl J Med. 1998; 338: 147
        • Lake J.K.
        • Power C.
        • Cole T.J.
        Women's reproductive health: the role of body mass index in early and adult life.
        Int J Obestet. 1997; 21: 432-438
        • Hartz A.J.
        • Barboriak P.N.
        • Wong A.
        • Katayama K.P.
        • Rimm A.A.
        The association of obesity with infertility and related menstrual abnormalities in women.
        Int J Obestet. 1979; 3: 57-73
        • Grodstein F.
        • Goldman M.B.
        • Cramer D.W.
        Body mass index and ovulatory infertility.
        Epidemiology. 1994; 5: 247-250
        • Pasquali R.
        • Pelusi C.
        • Genghini S.
        • Cacciari M.
        • Gambineri A.
        Obesity and reproductive disorders in women.
        Hum Reprod Update. 2003; 9: 359-372
        • Singh P.N.
        • Lindsted K.D.
        Body mass and 26-year risk of mortality from specific diseases among women who never smoked.
        Epidemiology. 1998; 9: 246-254
        • Troiano R.P.
        • Frongillo E.A.
        • Sobal J.
        • Levitsky D.A.
        The relationship between body weight and mortality: a quantitative analysis of combined information from existing studies.
        Int J Obes. 1996; 20: 63-75
        • Must A.
        • Jacques P.F.
        • Dallal G.E.
        • Bajema C.J.
        • Dietz W.H.
        Long-term morbidity and mortality of overweight adolescents—a follow-up of the Harvard Growth Study of 1922 to 1935.
        N Engl J Med. 1992; 327: 1350-1355
        • Hu F.B.
        • Manson J.E.
        • Stampfer M.J.
        • Colditz G.
        • Liu S.
        • Solomon C.G
        • et al.
        Diet, lifestyle, and the risk of type 2 diabetes mellitus in women.
        N Engl J Med. 2001; 345: 790
        • Dunphy B.C.
        • Neal L.M.
        • Cooke I.D.
        The clinical value of conventional semen analysis.
        Fertil Steril. 1989; 51: 324-329
        • Bonde J.P.
        • Ernst E.
        • Jensen T.K.
        • Hjollund N.H.I.
        • Kolstad H.
        • Henriksen T.B
        • et al.
        The relation between semen quality and fertility.
        A population based study of 430 first-pregnancy planners. Lancet. 1998; 352: 1172-1177
        • Bostofte E.
        • Serup J.
        • Rebbe H.
        Relation between sperm count and semen volume, and pregnancies obtained during a twenty-year follow-up period.
        Int J Androl. 1982; 5: 267-275
        • Slama R.
        • Eustache F.
        • Ducot B.
        • Jensen T.K.
        • Jorgensen N.
        • Horte A
        • et al.
        Time to pregnancy and semen parameters: a cross-sectional study among fertile couples from four European cities.
        Hum Reprod. 2002; 17: 503-515
        • Giagulli V.A.
        • Kaufman J.M.
        • Vermeulen A.
        Pathogenesis of the decreased androgen levels in obese men.
        J Clin Endocrinol Metab. 1994; 79: 997-991000
        • Gyllenborg J.
        • Rasmussen S.L.
        • Borch-Johnsen K.
        • Heitmann B.L.
        • Skakkebaek N.E., Juul A.
        Cardiovascular risk factors in men: the role of gonadal steroids and sex hormone-binding globulin.
        Metab Clin Exper. 2001; 50: 882-888
        • Vermeulen A.
        • Kaufman J.M.
        • Giagulli V.A.
        Influence of some biological indexes on sex hormone-binding globulin and androgen levels in aging or obese males.
        J Clin Endocrinol Metab. 1996; 81: 1821-1826
        • Hautanen A.
        Synthesis and regulation of sex hormone-binding globulin in obesity.
        Int J Obestet. 2000; 24: S64-S70
        • Gascon F.
        • Valle M.
        • Martos R.
        • Ruz F.J.
        • Rios R.
        • Montilla P
        • et al.
        Sex hormone-binding globulin as a marker for hyperinsulinemia and/or insulin resistance in obese children.
        Eur J Endocrinol. 2000; 143: 85-89
        • Pavlovich C.P.
        • King P.
        • Goldstein M.
        • Schlegel P.N.
        Evidence of a treatable endocrinopathy in infertile men.
        J Urol. 2001; 165: 837-841
        • Raman J.D.
        • Schlegel P.N.
        Aromatase inhibitors for male infertility.
        J Urol. 2002; 167: 624-629
        • Kaukua J.
        • Pekkarinen T.
        • Sane T.
        • Mustajoki P.
        Sex hormones and sexual function in obese men losing weight.
        Obesity Res. 2003; 11: 689-694
        • Vermeulen A.
        • Kaufman J.M.
        • Giagulli V.A.
        Influence of some biological indexes on sex hormone-binding globulin and androgen levels in aging or obese males.
        J Clin Endocrinol Metab. 1996; 81: 1821-1826
        • Swan S.H.
        • Brazil C.
        • Drobnis E.Z.
        • Liu F.
        • Kruse R.L.
        • Hatch M
        • et al.
        Geographic differences in semen quality of fertile U.S. males.
        Environ Health Perspect. 2003; 111: 414-420
        • Jorgensen N.
        • Andersen A.G.
        • Eustache F.
        • Irvine D.S.
        • Suominen J.
        • Petersen J.H
        • et al.
        Regional differences in semen quality in Europe.
        Hum Reprod. 2001; 16: 1012-1019
        • Jorgensen N.
        • Carlsen E.
        • Nermoen I.
        • Punab M.
        • Suominen J.
        • Andersen A.G
        • et al.
        East–West gradient in semen quality in the Nordic-Baltic area: a study of men from the general population in Denmark, Norway, Estonia and Finland.
        Hum Reprod. 2002; 17: 2199-2208