Marijuana Smoking and Markers of Testicular Function Among men From a Fertility Centre

Feiby L. Nassan; Mariel Arvizu; Lidia Mínguez-Alarcón; Paige L. Williams; Jill Attaman; John Petrozza; Russ Hauser; Jorge Chavarro; for the EARTH Study Team


Hum Reprod. 2019;34(4):715-723. 

In This Article


Contrary to our hypothesis, we observed that men who had ever smoked marijuana had higher sperm concentration and total sperm count, lower prevalence of sperm parameters below the WHO reference values, and lower FSH concentrations than men who had never smoked marijuana. These findings were robust after conducting several sensitivity analyses and considering different metrics of marijuana smoking. Specifically, more intense use was associated with significantly higher concentrations of testosterone, SHBG and inhibin-B, and later initiation of marijuana had an association with lower sperm count of marginal statistical significance. These results are consistent with a direct pro-spermatogenic testicular effect and secondary compensation in FSH secretion. On the other hand, the associations of marijuana smoking with sperm count and FSH concentrations were stronger for past smokers than for current smokers even though these two groups did not differ significantly from each other. Furthermore, longer duration since last use of marijuana was related to higher sperm count. These other results raise the possibility that our findings are not explained by a true underlying biologic mechanism but are instead spurious associations.

Let us first consider the possibility that the observed relations are spurious. While we considered a large number of potential confounders, residual confounding must still be considered. Our analysis suggests that, in order to account for the observed relations, an unmeasured confounder would have to be positively related to marijuana smoking and simultaneously positively related to semen quality by ≥2.08 risk ratio. In other words, for an unmeasured confounder to explain the observed associations, it would have to have a relation with marijuana smoking of greater magnitude than the association between marijuana and tobacco smoking (RR = 1.6) (one of the potential confounders most strongly related to marijuana smoking in our data) and a similarly strong positive association with semen quality, independently of all measured confounders. This seems unlikely. Selection bias does not seem likely either. Although we observed some differences in BMI, race and primary infertility diagnosis between men included and excluded from analysis, we did not observe systematic differences in terms of frequency of cocaine use, sperm concentration or sperm count between included and excluded men from analysis. The close match in the frequency of marijuana smoking in this population and the general USA population (Azofeifa et al., 2016), as well and the lack of difference in semen quality between men who joined this study and men from the same clinic who did not join the study (Hauser et al., 2005) also argues against selection bias possibility.

Another possibility is that the assumed causal structure is incorrect and the association reflects reverse causation. Specifically, we had assumed that marijuana use would have a negative effect on the testis impairing spermatogenesis and, secondarily, affecting concentrations of reproductive hormones (Supplementary Figure S2A). In an equally plausible alternate causal structure (Supplementary Figure 2B), men with higher circulating testosterone concentrations are more likely to engage in risk-seeking behaviours (Campbell et al., 2010), including marijuana and cocaine use, and testosterone is positively related to sperm count to the extent that testosterone reflects the normal gonadotropic activity to maintain intra-testicular testosterone concentrations and sustain spermatogenesis (Walker, 2011). These two causal structures are difficult to differentiate with only the available data. If anything, the lack of substantial change in estimates of the relation between marijuana and semen quality after adjusting for testosterone concentrations in addition to the opposite relations of cocaine and marijuana in our data argue more strongly for the first causal structure (Supplementary Figure S2A). In the absence of randomised trials of marijuana use, new studies with detailed information on within-person changes in marijuana use over time will be necessary to identify the correct causal structure.

Supplementary Figure S2.

Association between marijuana use and spermatogenesis. A: A priori assumed causal structure of the association between marijuana use and spermatogenesis. B: Alternative causal structure of the association between marijuana use and spermatogenesis. C: Proposed underlying association of marijuana use and spermatogenesis.

Our results are also consistent with a true biological association whereby the effect of marijuana smoking on testicular function, both in terms of spermatogenesis and hormone production, is dose dependent and non-linear. Specifically, and similar to the relation between alcohol intake and cardiovascular disease risk (Chiuve et al., 2006), we hypothesise that moderate use of marijuana may be related to improved testicular function but this relation reverses at higher doses, resulting in adverse effects (Supplementary Figure 2C). This possible scenario is consistent not only with our results but also with past data in humans and experimental models. If this hypothesis is correct, the apparent discrepancy in the association between marijuana use and sperm counts between this study and the report among young Danish men (Gundersen et al., 2015) could be explained by the differences in intensity of marijuana use between populations. Gundersen et al. (2015) reported that among 1215 healthy young men, men in the highest frequency of marijuana use had a 28% (95% CI: −48, −1) lower sperm concentration than non-users. Similar deleterious effects at high levels of exposure have been documented by others studying men with a history of drug abuse (Hembree et al., 1978; Issidorides, 1978; Singer et al., 1986; el-Gothamy and el-Samahy, 1992; Vescovi et al., 1992), although a positive correlation between marijuana use and percentage of motile sperm has also been reported (Close et al., 1990). Similarly, animal models have shown disruption of spermatogenesis associated with marijuana exposure (du Plessis et al., 2015; Alagbonsi et al., 2016; Di Giacomo et al., 2016). However, cannabinoid receptor 1 (CB1) knockout mice have a reproductive phenotype that strongly suggests an important effect of endocannabinoids, and potentially exogenous cannabinoids, on testicular function including decreased testicular production of testosterone, low numbers of Leydig cells in adulthood and abnormal spermatogenesis (Cacciola et al., 2008, 2013). CB1 receptors are found in the testis, vas deferens, and human sperm cells, and anterior pituitary, and activation of CB1 in spermatozoa by THC is different at low doses (hyper-activation) and high doses (inactivation) (Rossato et al., 2005). Clearly, additional research is needed to evaluate whether the effects of marijuana smoking on testicular function are dose dependent as suggested.

The most important limitation of the study is the possibility of underreporting of marijuana use given its status as an illegal drug during most of the study, its social stigma and potential effects on insurance coverage for infertility services of disclosing this information. In addition, we did not have information about other forms of marijuana use other than marijuana smoking. However, it has been shown that the self-report of marijuana was highly correlated with the blood and urinary cannabinoids levels (Fried, 1980; Greenland et al., 1982). Also, our results may not generalisable to men in the general population because men in the current study were enroled from a fertility centre. Strengths of our study include its prospective design with multiple semen samples in a large proportion of men and our ability to adjust for a wide range of potential confounders. We had data for many reproductive outcomes including semen parameters, sperm DNA integrity and serum reproductive hormones, which allow for more comprehensive assessments of testicular function.

In conclusion, marijuana smokers had higher sperm concentration and sperm count, lower prevalence of sperm parameters below the WHO reference values, and lower FSH concentrations than never marijuana smokers. These findings are not consistent with a deleterious role of marijuana smoking on testicular function as initially hypothesised. The findings are equally also consistent, however, with a non-causal interpretation. Whether these findings are reflective of the previously described role of cannabinoids in spermatogenesis and dose-dependent effects of the activation of endocannabinoid receptors on testicular function or are, instead, reflective of a spurious association, requires further work.