Increased Risk of Testicular Germ Cell Cancer Among Infertile Men FREE

Testicular germ cell cancer is the most common cancer among young men in industrialized countries. During the last 3 to 5 decades, there has been a notable and continued increase in the incidence of testicular germ cell cancers in men. Among US men, incidence has increased from 3.8 to 6.8 cases per 100 000 person-years in whites from 1975 to 2002.¹ More profound increases have been seen in Scandinavian countries such as Norway, where incidence has increased from 3.5 to 10 cases per 100 000 person-years from 1960 to 2000.² During the same period, there is evidence of a decline in semen quality and fertility in industrialized nations.^3- 4 However, it is unclear whether these 2 trends are independent or related to one another.

Several epidemiologic studies^5- 11 have investigated the association between male infertility and testicular cancer, with heterogeneous findings. Those studies that used case-control methods have inconsistent findings likely due to the use of surrogate markers for infertility, such as paternity and number of offspring, in the absence of an infertility evaluation. Two cohort studies have demonstrated increased risk for testicular cancer among infertile men. In a large Danish cohort, Jacobsen et al⁷ reported that men evaluated for infertility were 1.6 times more likely to develop subsequent testicular cancer (standardized incidence ratio [SIR], 1.6; 95% confidence interval [CI], 1.3-1.9). Although compelling, it is unclear whether this association can be generalized to US men given that testicular cancer has nearly twice the incidence in Denmark. In another study of a single US institution cohort of men referred for infertility treatment, Raman et al⁶ reported an 18-fold increased risk of testicular cancer (SIR, 18.3; 95% CI, 18.0-18.8). However, this finding represented the occurrence of only 8 cancer cases and did not exclude cases that occurred before the diagnosis of infertility. Thus, it is not clear from the existing literature that the associations between male infertility and testicular cancer drawn from European cohorts are valid for US men or that the true risk of testicular cancer among infertile US men is currently well described.

On the basis of the feasibility of this association, the difficulty in generalizing data from a non-US cohort, and the limited and discrepant data available from US studies, we evaluated the association between infertility and testicular cancer using a multi-institutional cohort of men seeking infertility care in the United States. To the best of our knowledge, this is the largest US investigation of its kind to examine this issue.

After institutional review board approval, men were identified from an initial cohort of 51 461 couples with fertility problems who sought evaluation for infertility in California. This initial cohort consisted of couples in which both partners were older than 18 and who received care from 40 different physicians at 15 infertility centers between January 1, 1967, and January 1, 1998. Eleven fertility centers were located in northern California and 4 in southern California. Four of the fertility centers were part of the Kaiser Permanente Medical Group. All ethnic and language groups were included, and couples were only excluded if (1) the woman was not attempting pregnancy or (2) the woman was not a US resident. Identifying information was obtained from the medical records of all study participants from which 42 274 male partners were identified. Because couples sought care from fertility centers staffed by reproductive endocrinologists, the primary focus of care was the female partner; thus, detailed information was not available for all male partners. Complete demographic information, including date of birth, was available for 22 562 men, who represent the final study cohort (Figure).

Men were considered to have male factor infertility based on a clinical presentation with abnormal semen analysis variables as defined by World Health Organization criteria.¹² The presence or absence of male factor infertility was determined by the treating physician and coded as a dichotomous variable (male factor: yes/no). The presence or absence of male factor infertility was known for 19 106 men (84.7%), of whom 4549 (23.8%) had male factor infertility and 14 557 (76.2%) did not.

All male members of couples evaluated for infertility were linked to the California Cancer Registry (CCR). The CCR combines all registries of the Surveillance Epidemiology and End Results (SEER) program of the National Cancer Institute for the state of California and contains information on all cases of histologically confirmed cancer from January 1, 1988, to December 31, 2004. The CCR provides data on date of diagnosis, age at diagnosis, vital status, tumor behavior (International Statistical Classification of Diseases for Oncology, Second Edition [ICDO-2] coding), histologic type (ICDO-2 coding), and history of cryptorchidism. All testicular tumors were identified by ICDO-2 site codes (testis: C620-C621 and C629) and categorized by ICDO-2 histology codes (seminoma: 9060-9063, embryonal: 9070-9073, teratocarcinoma: 9081, mixed germ cell tumor: 9085, and carcinoma not otherwise specified: 8010).

Automated, probabilistic matching was performed using Social Security number, first name, middle name, last name, date of birth, and address. All matches were reviewed by hand and consensus obtained by 3 investigators (T.J.W., M.S., and M.S.C.). A match was considered definite if there was complete agreement on Social Security number, first name, last name, and date of birth. All cases of testicular cancer were considered definite matches.

Cancer cases occurring before 1988 could not be identified by the CCR; therefore, our analysis was left truncated starting at 1987. By this method, men evaluated for infertility before cancer registry availability do not accumulate “at risk time” until 1987 but remain as members of the cohort in whom cancer cases can be identified. All men were aged accordingly from the time of infertility evaluation to the time of cancer diagnosis or the time at which final cancer registry accrual occurred (December 31, 2004). Men diagnosed as having cancer before the date of infertility evaluation or within 1 year after infertility evaluation were excluded from analysis given our inability to determine whether the treatment of cancer or the presence of an occult cancer led to infertility. Two additional analyses were performed to evaluate the potential for reverse causality. First, we calculated cancer risk stratified by time since infertility evaluation. Second, we performed sensitivity analysis whereby cases of testicular cancer were excluded by increasing intervals from the time of infertility evaluation up to 5 years.

The rate of testicular cancer among men in our cohort was compared with the rate among men in the general California population. We calculated the expected number of cancer cases in the cohort by multiplying the number of years at risk by the 5-year age strata and year-specific primary testicular cancer rates from all SEER registries in California. The SIRs were calculated by dividing the number of observed cancer cases in the cohort by the number of expected cases. A man's at-risk time was calculated from the start of the couple's infertility evaluation. Analyses were performed for the entire cohort of men seeking infertility care and for subgroups of men with and without male factor infertility.

We analyzed the risk for testicular cancer in men with and without male factor infertility using a Cox proportional hazards regression model, controlling for age, duration of treatment, and site of treatment. We used SIRs and hazard ratios and their 95% CIs to estimate the association between infertility and testicular cancer. P ≤ .005 was considered statistically significant, and all tests were 2-sided. All statistical analyses were performed using a commercially available software program (SAS, version 9.1; SAS Institute Inc, Cary, North Carolina).

Of the 51 461 couples who sought infertility care and were enrolled in the primary study, the male partner was identified in 42 274 cases and, of these, 22 562 had complete data for analysis. Table 1 presents the characteristics of men with and without male factor infertility. Men with and without male factor infertility were similar. Men with male factor infertility were slightly older at the time of infertility evaluation (mean age, 38.1 vs 36.4 years) and had longer duration of infertility care (mean duration, 1.7 vs 1.5 years) than men without male factor infertility but had the same duration of follow-up (mean duration, 11.4 years). A larger proportion of men with male factor infertility developed testicular cancer compared with those without male factor infertility (0.3% vs 0.1%).

Sixty-seven cases of testicular cancer occurred among cohort members; however, only 34 cases occurred at least 1 year after the start of infertility evaluation. Of the 34 men who developed testicular cancer, 85.3% had seminomas, 11.8% had nonseminomas, and 2.9% had unspecified disease. Greater than 85.3% of tumors were pathologically confined to the testicle (Table 2). Among all men who developed testicular cancer, 1 (2.9%) had a history of cryptorchidism, and the mean (SD) time from infertility evaluation to cancer diagnosis was 5.6 (4.0) years.

Overall, the infertility cohort members demonstrated a trend for increased risk of testicular germ cell cancer compared with men from the general California population (Table 3). Thirty-four men developed cancer in the cohort compared with an expected 25 (SIR, 1.3; 95% CI, 0.9-1.9). Among men with male factor infertility, risk of testicular cancer was markedly and significantly higher, whereby 13 cases were observed compared with an expected 5 (SIR, 2.8; 95% CI, 1.5-4.8). Among men without male factor infertility, no significant evidence was found of increased testicular cancer risk (SIR, 1.1; 95% CI, 0.6-1.7).

Table 4 gives the SIRs for testicular germ cell cancer stratified by time between infertility evaluation and cancer diagnosis. Among all men, risk for cancer was highest for the 12- to 15-year interval (SIR, 2.0) but similar across all intervals, with the exception of the 8- to 11-year interval (SIR, 0.5). In the subanalysis of men with male factor infertility, risk was highest for the 1- to 3-year interval (SIR, 4.0); however, this analysis is difficult to interpret given the small number of cases in each stratum. No trends in the SIRs during the 4 periods were identified. Table 5 gives the SIR for testicular cancer with exclusion of cases of varying intervals since infertility evaluation was performed. Cancer risk was elevated and stable with cases excluded from 1 to 5 years from the time of infertility evaluation. The SIRs were highest when cases were included from the 0- to 1-year interval from infertility evaluation, confirming the importance of their exclusion in the overall analysis.

The risk of testicular cancer in men with male factor infertility relative to those without male factor infertility was modeled using multivariable Cox proportional hazards regression models. When controlling for age, duration of infertility treatment, and infertility treatment center, men with male factor infertility had 2.8 times the risk of cancer compared with those without male factor infertility (hazard ratio, 2.8; 95% CI, 1.3-6.0). In this same model, for each additional year of receiving infertility treatment, men had 1.2 times the risk of cancer (hazard ratio, 1.2; 95% CI, 1.1-1.4).

Our cohort study, based on more than 22 000 men undergoing evaluation for infertility, observed an association between infertility and the subsequent development of testicular cancer. Male partners of infertile couples were 1.3 times more likely to develop a testicular germ cell cancer than the California population, regardless of male fertility status. Among men with known male factor infertility, the risk of subsequent testicular cancer was more than twice that, such that these men were 2.8 times more likely to develop testicular cancer relative to the general population. When men without male factor infertility were examined, there was no excess risk of testicular cancer. In multivariable analyses, men with male factor infertility were 2.8 times more likely to develop cancer than those in the cohort without male factor infertility, after adjusting for age, duration of infertility, and treatment facility.

Given our understanding of the natural history and presentation of testicular cancer, it is unlikely that these findings represent a screening phenomenon in which men exposed to infertility care are more likely to be diagnosed as having an indolent testicular tumor as a result of diagnostic testing. In fact, testicular germ cell tumors in adults are almost exclusively diagnosed fairly rapidly by a simple physical examination that reveals a nodule or painless scrotal swelling.¹³

These data confirm the findings by Jacobsen and colleagues,⁷ who reported an increased cancer risk among Danish men undergoing semen analysis testing. Although similar in design, this study differs from the Danish cohort in that all men in this analysis were infertility partners rather than men presenting for semen analyses for a variety of reasons. Despite differences in the study populations, the US and Danish studies demonstrate a remarkably similar risk of developing testicular cancer. Although both studies observed an increased risk of cancer in men with male factor infertility, men in our cohort were dichotomized to the presence or absence of these factors but in the absence of semen analysis data; thus, we cannot comment on specific semen analysis variables that may place men at highest risk. On the contrary, our data do not confirm the nearly 20-fold increased risk of testicular cancer among infertile men reported by Raman and colleagues.⁶ Although provocative, this markedly increased risk likely reflects reverse causality, in which the analysis included all prevalent and incident tumors and not just those occurring after infertility. Despite the increased relative risk of testicular cancer among infertile men, the absolute risk of cancer remains low, even among men with male factor infertility.

Similar to both previous cohort studies, seminoma was the predominant cancer histologic feature observed in affected men. Furthermore, most cancers had favorable histologic features (seminoma) and were pathologically confined to the testicle at the time of diagnosis. Although it is possible that infertility is associated with this particular histologic feature, it is more likely that the age distribution of men in our cohort makes the development of seminoma (the most common testicular tumor in older men) more probable. Furthermore, it is possible that having more aggressive nonseminomatous germ cell tumors during or before the reproductive years effectively excluded such patients from our study cohort.¹⁴

These results suggest that male factor infertility may be a risk factor for the subsequent development of testicular germ cell cancer; however, potential sources of bias merit mention. The increased risk of testicular cancer among all cohort members may be the result of unmeasured confounding, whereby some additional factor, common to all members of the cohort, placed men at increased testicular cancer risk. Specifically, higher socioeconomic status has been shown to be a potential risk factor for the development of testicular cancer.¹⁵ Although this variable was not measured, men of higher socioeconomic status may be more likely to seek infertility evaluation and more likely to develop testicular cancer, thus increasing the SIR for all cohort members. Cryptorchidism is a commonly cited risk factor for infertility and testicular cancer.¹⁶ Because we do not know the prevalence of cryptorchidism among cohort members without cancer, we cannot assess the potential for confounding by this variable. By cancer registry data, only 1 man (3%) with cancer had a history of cryptorchidism, suggesting that infertility may be associated with testicular cancer, independent of cryptorchidism. However, this too must be interpreted with caution because a history of cryptorchidism may not be reliably identified by cancer registrars.

In interpreting these data, we considered the postulate that male factor infertility or its treatment could cause testicular cancer. However, this theory is highly improbable given that in many cases infertility treatment involves the use of assisted reproductive technologies rather than specific medical or surgical treatment of the male partner.¹⁷ A more plausible explanation is that a common exposure underlies infertility and testicular cancer. Prior work by us and others^18- 20 suggests that certain severe forms of male infertility are associated with faulty DNA repair. Faulty DNA repair has classically been associated with tumorigenesis, in human and animal models, and could underlie the association of infertility and testicular cancer.^21- 22 In addition to this biological basis, the testicular dysgenesis syndrome, as hypothesized by Skakkebæk and colleagues,²³ is a theoretical construct that attempts to relate environmental modulators, genetics, and infertility in the development of testicular cancer. Thus, the association between infertility and testicular cancer has been shown to be biologically and clinically feasible. More importantly, the association between male infertility and testicular germ cell cancer should stimulate further research that focuses on the etiology of poor germ cell health in these populations.

Correspondence: Thomas J. Walsh, MD, MS, Department of Urology, University of Washington, School of Medicine, 1959 NE Pacific St, Seattle, WA 98195-6510 (walsht@u.washington.edu).

Accepted for Publication: August 14, 2008.

Author Contributions:Study concept and design: Walsh, Croughan, Chan, and Turek. Acquisition of data: Walsh and Croughan. Analysis and interpretation of data: Walsh, Croughan, Schembri, and Chan. Drafting of the manuscript: Walsh, Croughan, Chan, and Turek. Critical revision of the manuscript for important intellectual content: Walsh, Croughan, Schembri, Chan, and Turek. Statistical analysis: Walsh, Croughan, Schembri, and Chan. Obtained funding: Walsh, Croughan, and Turek. Administrative, technical, and material support: Walsh, Schembri, and Turek. Study supervision: Walsh, Croughan, and Turek.

Financial Disclosure: None reported.

Funding/Support: This study was funded by grants K12 HD053943012 (Dr Walsh) and 1 RO1 CA69619 (Dr Croughan) from the National Institute of Child Health and Human Development and the California Urology Foundation (Dr Turek).

Disclaimer: The ideas and opinions expressed herein are those of the authors, and endorsement by the state of California, California Department of Public Health, the National Cancer Institute, and the Centers for Disease Control and Prevention or their contractors and subcontractors neither is intended nor should be inferred.

Additional Contributions: The collection of cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute's SEER program under contract N01-PC-35136 awarded to the Northern California Cancer Center, contract N01-PC-35139 awarded to the University of Southern California, and contract N02-PC-15105 awarded to the Public Health Institute; and the Centers for Disease Control and Prevention's National Program of Cancer Registries under agreement U55/CCR921930-02 awarded to the Public Health Institute.