© The Author 2015. study question: By investigating a birth cohort with a high ongoing participation rate to derive an unbiased population, what are the parameters and influences upon testicular function for a population not selected with regard to fertility? summaryanswer: While varicocele, cryptorchidism and obesitymayimpact on human testicular function, mostcommondrug exposures and the presence of epididymal cysts appear to have no or minimal adverse impact. what is known already: The majority of previous attempts to develop valid reference populations for spermatogenesis have relied on potentially biased sources such as recruits from infertility clinics, self-selected volunteer sperm donors for research or artificial insemination or once-fertile men seeking vasectomy. It is well known that studies requiring semen analysis have low recruitment rates which consequently question their validity. However, there has been some concern that a surprisingly high proportion of young men may have semen variables that do not meet all the WHO reference range criteria for fertile men, with some studies reporting that up to one half of participants have not meet the reference range for fertile men. Reported median sperm concentrations have ranged from 40 to 60 million sperm/ml. study design, size and duration: TheWestern Australian Pregnancy Cohort (Raine)was established in 1989. At 20-22 years of age, members of the cohort were contacted to attend for a general follow-up, with 753 participating out of the 913 contactable men. Of these, 423 men(56% of participants in the 20-22 years cohort study,46% of contactable men) participated in a testicular function study. Of the 423 men, 404 had a testicular ultrasound, 365 provided at least one semen sample, 287 provided a second semen sample and 384 provided a blood sample. participants/materials, setting, methods: Testicular ultrasound examinations were performed at King Edward Memorial Hospital, Subiaco, Perth, for testicular volume and presence of epididymal cysts and varicoceles. Semen samples were provided and analysed by standard semen assessment and a sperm chromatin structural assay (SCSA) at Fertility Specialists ofWestern Australia, Claremont, Perth. Serum blood samples were provided at the University ofWestern Australia, Crawley, Perth and were analysed for serum luteinizing hormone (LH), follicular stimulating hormone (FSH), inhibin B, testosterone, dihydrotestosterone (DHT), dehydroepiandrosterone (DHEA), estradiol, estrone and the primary metabolites of DHT: 5a-androstane-3a,17b-diol (3a-diol) and 5-a androstane-3-b-17-beta-diol (3b-diol). Serum steroids were measured by liquid chromatography, mass spectrometry and LH, FSH and inhibin B were measured by ELISA assays. main results and the role of chance: Cryptorchidism was associated with a significant reduction in testicular (P = 0.047) and semen (P = 0.027) volume, sperm concentration (P = 0.007) and sperm output (P = 0.003). Varicocele was associated with smaller testis volume (P , 0.001), lower sperm concentration (P = 0.012) and total sperm output (P = 0.030) and lower serum inhibin B levels (P = 0.046). Smoking, alcohol intake, herniorrhaphy, an epididymal cyst, medication and illicit drugs were not associated with any significant semen variables, testicular volume or circulating reproductive hormones. BMI had a significantly negative correlation with semen volume (r = 20.12, P = 0.048), sperm output (r = 20.13, P = 0.02), serum LH (r = 20.16, P = 0.002), inhibin B (r = 20.16, P , 0.001), testosterone (r = 20.23, P , 0.001) and DHT (r = 20.22, P , 0.001) and a positive correlation with 3aD (r = 0.13, P = 0.041) and DHEA (r = 0.11, P = 0.03). Second semen samples compared with the first semen samples in the 287 participants who provided two samples, with no significant bias by Bland-Altman analysis. Testis volume was significantly correlated positively with sperm concentration (r = 0.25, P , 0.001) and sperm output (r = 0.29, P , 0.001) and inhibin B (r = 0.42, P , 0.001), and negatively correlated with serum LH (r = 20.24, P , 0.001) and FSH (r = 20.32, P , 0.001). SCSA was inversely correlated with sperm motility (r = 20.20, P , 0.001) and morphology (r = 20.16, P = 0.005). WHO semen reference criteria were all met by only 52 men (14.4%). Some criteria were not met at first analysis in 15-20% of men, incluDing semen volume (,1.5 ml, 14.8%), total sperm output (,39 million, 18.9%), sperm concentration (,15 million/ml, 17.5%), progressive motility (,32%, 14.4%) and morphologically normal sperm (,4%, 26.4%), while all fiveWHOcriteria were not met in four participants (1.1%). limitations and reasons for caution: This was a large cohort study; however, potential for recruitment bias still exists. Men who did not participate in the testicular evaluation study (n = 282) did not differ from those who did (n = 423) with regard to age, weight, BMI, smoking or circulating reproductive hormones (LH, FSH, inhibin B, T, DHT, E2, E1, DHEA, 3a-diol, 3b-diol), but were significantly shorter (178 versus 180 cm, P = 0.008) and had lower alcohol consumption (P = 0.019) than those who did participate. wider implications of the findings: This study demonstrated the feasibility of establishing a birth cohort to provide a relatively unbiased insight into population-representative sperm output and function and of investigating its determinants from common exposures. While varicocele, cryptorchidism and obesity may impact on human testicular function, most common drug exposures and the presence of epididymal cysts appear to have little adverse impact, and this study suggests that discrepancies from the WHO reference ranges are expected, due to its derivation from non-population-representative fertile populations.