Association Between Use of Marijuana and Male Reproductive Hormones and Semen Quality

A Study Among 1,215 Healthy Young Men

Tina Djernis Gundersen; Niels Jørgensen; Anna-Maria Andersson; Anne Kirstine Bang; Loa Nordkap; Niels E. Skakkebæk; Lærke Priskorn; Anders Juul; Tina Kold Jensen


Am J Epidemiol. 2015;182(6):473-481. 

In This Article



Because of the military draft in Denmark, all men aged 18 years, except those suffering from severe chronic disease, are required to undergo a compulsory physical examination to determine their fitness for military service. Some men postpone their examination until completion of their education. Since 1996, trained staffs from the Department of Growth and Reproduction at Copenhagen University Hospital (Rigshospitalet, Copenhagen, Denmark) have approached draftees when they appeared for their compulsory physical examination and invited them to participate in a study of semen quality. Men recruited from January 2008 to June 2012 were included in the present study because they were asked detailed questions about marijuana and/or recreational drug use. A total of 1,221 men, representing 30% of those who were invited, agreed to take part in the study and completed a questionnaire, delivered a semen sample, had a blood sample drawn, and underwent a physical examination. They received kr500 (approximately US $85) in compensation for their time.

Participants did not differ from nonparticipants with regard to age, but they were generally better educated than nonparticipants (data not shown). Study approval was obtained from the local ethics committee. A detailed description of the study has been published previously.[9,10]

Physical Examination

A physician examined the participants for the Tanner stage of pubic hair and genital development, testicular volumes by use of a Prader wooden orchidometer (Pharmacia & Upjohn Company, LLC, London, United Kingdom), location of the testes in the scrotum, and the consistency of the testes and epididymis. The possible presence of variocele (grades 1–3), hydrocele, or any particular malformation was recorded. Height and weight of the men were measured, and body mass index (BMI) was calculated as weight in kilograms divided by squared height in meters.


Before the examination all the men filled in a questionnaire on previous and/or current genital diseases, such as inguinal hernia, variocele, epididymitis, chlamydia, and gonorrhoea, and whether they had had surgery for testicular torsion. They reported whether they had been responsible for a pregnancy and if they were born with both testes in the scrotum, as well as whether they had had fever above 38°C (100.4°F) in the previous 3 months. Fever was not reported by 6.2% of the men, but they were categorized as not having had fever because their semen quality resembled that of those without fever. Self-reported genital diseases were divided into 2 categories: "self-reported genital conditions" (including torsion of testis, epididymitis, and inguinal hernia) and "sexually transmitted diseases" (including chlamydia and gonorrhea).

They were asked to estimate their fitness by self-reported physical activity converted to watts per week using the method of Craig et al..[11] Men were asked about current tobacco-smoking habits, and daily caffeine intake was estimated on the basis of their reported intake of caffeine-containing beverages the week prior to the visit. They completed a diary reporting their daily intake of red and white wine, beer, strong alcoholic drinks, alcopops, and others alcoholic drinks during the week prior to participation; these were categorized into >20 or ≤20 units per week, the former being the advised maximum intake for men by the Danish Health and Medicines Authority.[12] In addition, they were asked about television watching, sleeping habits, and stress 3 months before (categorized as described by Craig et al.[11] and Jensen et al.[13]), and they answered a food frequency questionnaire from which their intake of 681 macronutrients was calculated.[12–14] Their mothers answered questions about their tobacco-smoking habits during pregnancy, and the educational level of mothers was coded as below 9 years, 9–10 years, or above 10 years of schooling.

The men were asked, "Have you smoked hashish (pot, marijuana) within the last 3 months?" and "Have you taken recreational drugs other than marijuana (e.g., amphetamine, ecstasy, cocaine) within the past 3 months?" Answer categories were daily, several times per week, at least once per week, less than once per week, and never.

Blood Sample

Blood samples were drawn from a cubital vein and centrifuged, and the serum was separated and frozen. Serum levels of follicle-stimulating hormone, luteinizing hormone, and sex hormone-binding globulin were determined by using a time-resolved immunofluorometric assay (Delfia; Wallac Oy, Turku, Finland). Testosterone and estradiol levels were determined by using time-resolved fluoroimmunoassays (Delfia; Wallac Oy). The inhibin B level was determined by means of a specific 2-sided enzyme immunometric assay (Inhibin B Gen II; Beckman Coulter, Ltd., High Wycombe, United Kingdom). The hormones were all measured within same time period and in the same assay batches. Free testosterone was calculated on the basis of the measured serum concentrations of total testosterone and sex hormone-binding globulin by using the method of Vermeulen et al.[15] and a fixed albumin concentration of 43.8 g/L.[15]

Semen Analysis

Participants of the study provided a semen sample by means of masturbation in a room close to the semen laboratory. The period of ejaculation abstinence (time since last ejaculation) and the month of delivery of the sample (divided into October–March or April–September intervals) were recorded. The semen samples were analyzed for volume, sperm concentration, total sperm count, percentage of motile spermatozoa, and percentage of morphologically normal spermatozoa as described by Jørgensen et al.[16] in accordance with the most recent World Health Organization guidelines.[17] Since 1996, our laboratory has led a quality-control program for the assessment of sperm concentration; the laboratory has kept the interlaboratory difference unchanged,[14] and the variation between technicians was less than 10%. The sperm morphology has currently been assessed by experienced technicians according to strict criteria for the first 838 men.

Statistical Methods

Outcome variables for semen quality were semen volume, sperm concentration, total sperm count, and percentages of motile and morphologically normal spermatozoa. Outcome variables for hormone analyses were testosterone, luteinizing hormone, follicle-stimulating hormone, sex hormone-binding globulin, and calculated free testosterone. Exposure variables included frequency of smoking marijuana during the past 3 months, specified as none (reference value), ≤1 time per week, and >1 time per week. In addition, a variable including the combined use of marijuana and other recreational drugs was created as marijuana use ≤1 time per week and no use of other recreational drugs (reference value), marijuana use >1 time per week and no use of other recreational drugs, and marijuana use >1 time per week and use of other recreational drugs.

First, semen quality and reproductive hormone levels were compared among men with and without marijuana use. Then variables from the physical examination and questionnaire were compared for men with and without marijuana smoking by use of the χ2 or t test to identify potential confounders. The data were then examined by multiple linear regression. Because of the nonnormal (skewed) distributions of semen quality and serum reproductive hormones, all variables except for motile sperm and morphologically normal forms were transformed by a natural logarithmic scale and back-transformed to obtain the expected percentage change for marijuana use. The variables for motile sperm and morphologically normal forms were analyzed untransformed because of the normal distribution to obtain the change in percentage points. To select confounders, we initially included those associated with semen parameters, hormone levels, or use of marijuana. Potential confounders were then stepwise excluded, on the basis of their changing the estimates by more than 10%. One set of confounders was used for all semen quality analyses: hours of abstinence (subdivided into <48 hours, 48–95 hours, and >95 hours), tobacco smoking, alcohol consumption, self-reported sexually transmitted diseases, and use of other recreational drugs. There was 1 set of confounders for all hormone analyses: BMI levels (<20, 20–24.99, and ≥25), tobacco smoking, hour of day of blood sampling, and use of other recreational drugs. In the analyses of the combined association of marijuana and other recreational drug use, use of recreational drugs was not included as a confounder.

We evaluated a trend in the association of marijuana use by inserting the categorical marijuana variable (never, less than once per week, at least once per week, several times per week, and daily) into the model assuming the association to be linear. Finally, we compared marijuana use for men with semen concentration below World Health Organization reference values[17] (15 and 20 million/mL) by binary logistic regression analyses adjusting for the same set of confounders.

The maternal educational level was left out in the final analysis as answers were missing in 12% of the cases. However, we repeated all of the analyses with inclusion of maternal education. For the semen quality analyses, we tested for interaction with BMI, alcohol intake, and smoking, which was not significant. However, we stratified the analyses on BMI, alcohol intake, and tobacco smoking in order to evaluate if these had any modifying effect.

The fit of regression models was evaluated by testing the residuals for normality and by inspecting the residual plots. All of the analyses were conducted as general linear models in SPSS, version 19, statistical software (SPSS, Inc., Chicago, Illinois), and the results are presented with 95% confidence intervals.