Tel.: +972-2-6777452, Fax: +972-2-6431086
Risk of germ cell testicular cancer according to origin: A migrant cohort study in 1,100,000 Israeli men
Article first published online: 28 SEP 2012
Copyright © 2012 UICC
International Journal of Cancer
Volume 132, Issue 8, pages 1878–1885, 15 April 2013
How to Cite
Levine, H., Afek, A., Shamiss, A., Derazne, E., Tzur, D., Zavdy, O., Barchana, M. and Kark, J. D. (2013), Risk of germ cell testicular cancer according to origin: A migrant cohort study in 1,100,000 Israeli men. Int. J. Cancer, 132: 1878–1885. doi: 10.1002/ijc.27825
- Issue published online: 13 FEB 2013
- Article first published online: 28 SEP 2012
- Accepted manuscript online: 7 SEP 2012 06:56AM EST
- Manuscript Accepted: 23 AUG 2012
- Manuscript Received: 15 MAY 2012
- Israel Cancer Research Fund
- testicular cancer;
- young men;
- environmental health;
- male reproduction
Testicular cancer incidence is highest among men of northern European ancestry and lowest among men of Asian/African descent. We conducted a large-scale migrant cohort study to assess origin and migrant generation as predictors of testicular germ cell tumors (TGCTs), controlling for possible confounders. Data on 1,092,373 Jewish Israeli males, who underwent a general health examination prior to compulsory military service at ages 16–19 between the years 1967–2005, were linked to Israel National Cancer Registry to obtain incident TGCTs up to 2006. Cox proportional hazards was used to model time to event. Overall, 1,001 incident cases (534 seminoma and 467 nonseminoma) were detected during 19.2 million person-years of follow-up. Origin was a strong independent predictor of TGCTs with remarkably low incidence for North African-born (HR = 0.10, 95% CI: 0.04–0.21) and Asian-born (HR = 0.35, 0.20–0.62), while intermediate for Israeli-born of North African origin (HR = 0.48, 0.40–0.58) and Asian origin (HR = 0.56, 0.47–0.66), compared to European origin. A comparison of Israeli born of North African and Asian origin with North African and Asian-born yielded a HR of 2.31 (1.36–3.93). Significant risk factors controlled for were year of birth, years of education and height. Findings persisted when analyses were stratified by histologic subtypes of TGCTs. The findings of lower rates of TGCTs among men born in North Africa and Asia compared to European ancestry, but a steep increase in next generation migrants, particularly among the Israeli-born migrants from North Africa, provide clues to direct further research on the role of modern lifestyle and environment in the etiology of TGCTs.
Testicular germ cell tumors (TGCTs) are the most frequently occurring cancer in developed countries among young men aged 15–35 years.1 Over the last decades, the incidence of TGCTs has increased in many countries for unknown reasons.2 In Israel, too, a steep rise in the incidence of TGCTs was noted between 1992 and 2002.3 TGCTs are commonly grouped into two main types, roughly half each: seminomas and nonseminomas.
The most consistent risk factors are a previous history of TGCTs or carcinoma in situ, a family history of TGCTs and cryptorchidism.4, 5 Meta-analytic synthesis of the available evidence suggests that inguinal hernia, twining, maternal bleeding, low gestational age, low birth order and small sibship size may also be risk factors for TGCTs.6, 7 A recent systematic review and meta-analysis found evidence for a positive monotonic association of adult height with TGCTs.8 Early studies indicated that higher socioeconomic position (SEP) was associated with an increased risk,9 whereas more recent reports have found little association,10 inferring that decreasing variation between socioeconomic strata may have limited the role of SEP as a predictor.11
Globally, testicular cancer incidence is highest among men of northern European ancestry and lowest among men of Asian and African descent. However, it is not known if this difference is due to genetic or environmental factors. Rates among European men, especially Scandinavians, are five to ten times higher than men of African and Asian descent.12 Rates in the Israeli Jewish population, despite the steep increase over the last decades,3 are still within the lower range of western countries such as Scandinavia, Germany and the United States.13 In countries with multiethnic populations, the incidence of TGCTs among white men is higher than the incidence among men of other backgrounds.14 In addition to having the highest absolute rates, men of European ancestry have experienced the greatest increase in incidence over the last decades, but this incline may have reached its peak in some countries.15 In almost all populations studied, the increase in incidence has been found to be consistent with a birth-cohort effect.16 Several migrant studies have reported on differing TGCT incidence by origin and generation of immigration to countries such as Sweden and Denmark.17, 18 Among immigrants to Sweden, the first generation immigrants displayed incidence rates similar to their country of origin (higher or lower), but this difference tended to disappear in the next generation born in Sweden.19
Vast immigration to Israel took place since the establishment of the state, mainly of Jews from Europe, North Africa and Western Asia. At the same time, Israel underwent transition from a developing country into a modern, western life-style, industrial state.20 This unique setting in Israel and our extensive dataset allowed us to conduct a large-scale migrant cohort study of adolescents whose objective was to determine the associations between origin and risk of TGCTs, and examine the extent to which these associations are explained by other risk factors such as birth cohort, height and measures of SEP.
Material and Methods
Study design, setting and population
The study was designed as a historical prospective migrant cohort of adolescents intended to assess the association between origin and the risk of developing TGCTs in young adults, adjusting for socio-demographic and anthropometric determinants measured in adolescence.
Israel is one of the few Western countries where military service is mandatory. Israeli Jewish adolescents are obligated to present themselves for a medical board examination, predominantly at age 17, before military service.
Consequently, use of these data provides a generally representative sample of the young Jewish adult population, particularly of males. Our study population included those examined since 1967, the first year with full computerized records on recruits, until 2005. We included Jewish males aged 16–19 at examination. We excluded men with a cancer diagnosis prior to examination (since 1960), and those of Ethiopian origin who differ significantly from the rest in their phenotypic and genetic characteristics (10,475 born in Ethiopia, 1,596 Israeli born of Ethiopian origin; 0 cases), those with missing values for weight or height (30,548 participants, 17 cases) or country of birth/paternal country of birth for Israeli born (6,369 participants, 7 cases) and 22 subjects without testicles. The study protocol was approved by the Ethics Committee of the Israel Defense Force Medical.
Data sources and variables
Risk factor data
Socio-demographic, anthropometric and clinical data were obtained at the recruitment centers during the obligatory call-up. Origin was defined by country of birth for immigrants to Israel (first generation), and by paternal country of birth for the Israeli born, or grandfather's country of birth if the father was Israeli-born (second and third generations, between which we are unable to distinguish in this data set). Israeli origin was defined for the relatively uncommon fourth generation (or more) Israelis, who are a mixture of European, North African and Asian ancestries. European origin also included countries of emigration from Europe (the Americas, Australia and Southern Africa). Consequently, seven categories were formed: European origin, European born, Asian origin, Asian born, North African origin, North African born and Israeli origin.
Height and weight were measured by trained medical personnel using a stadiometer and a beam balance. Subjects were barefoot and wore only a shirt and underwear. Height and year of birth were treated as continuous variables, while BMI was grouped as <25, 25.0–29.9 and ≥30 kg/m2. Years of schooling prior to military service was grouped into ≤9, 10, 11 and ≥12 years of education.
Residential SEP was based on the village/town/city of residence, according to a national classification of 10 clusters by geographical units, which we categorized as low (1–4), medium (5–7) and high (8–10).
Cases of TGCTs in our cohort were identified by linkage to the Israel National Cancer Registry (INCR) by way of the personal identification number given to all Israeli citizens at birth or immigration. The INCR, a population-based registry in operation since 1960, meets internationally accepted requirements for the coding system (ICDO-Version 3) and completeness of data. Reporting is mandatory by Israeli law since 1982. Completeness of the registry is considered to be about 95% for all types of solid cancer and has maintained consistently high coverage since inception of the registry in the 1960s.21
Only histologically verified TGCTs, registered between 1967 and 2006, the last year with data at the time of linkage, were included. Only primary tumors, ICD-O topography code C62.9 (Testis NOS), C62.0 (undescended testis) and C62.1 (descended testis) were included. Three cases of seminoma and eight cases of nonseminoma not found in the testis were excluded, as were 61 cases of testicular non GCTs and two cases of spermatocytic seminoma as these tumors are thought to have different etiology.22 Diagnoses of TGCTs were grouped as seminoma (morphologic codes 9060–9064, excluding spermatocytic seminoma) or nonseminoma (including embryonal carcinoma, codes 9070–9073; malignant teratoma [including mixed tumors], codes 9080–9085 and 9102; choriocarcinoma, codes 9100–9101; nonseminomatous GCT, code 9065), as in other studies.23 One case of TCGT was excluded as the follow-up was curtailed at age 54 to provide a well-defined grouping.24
Analyses were performed with IBM SPSS version 18. Means of age at diagnosis for seminoma and nonseminoma were compared using a t test. Cox proportional hazards regression models were used to assess the association between origin and time to TGCT diagnosis adjusting for suspected confounders. For the origin and education variables, the largest groups were chosen as the reference category: Israeli born of European origin, and ≥12 years of education. As birth cohort was associated with the incidence of TGCTs, we introduced year of birth into all single predictor (“univariate”) models as well as the multivariable analyses. In an additional analysis rates within the same origin group were compared directly between immigrant generations, adjusted for the covariates. Multivariable analyses were carried with all variables that were statistically significant in the univariate analyses. A stepwise forward likelihood ratio procedure was used with p to enter <0.05 and p to exit ≥0.10 to include the covariates into the model. In addition, models were run separately for seminoma and nonseminoma. About 95% confidence intervals (CI) were computed for the hazard ratios (HRs). Inspection of log minus log plots for each variable verified the assumption of proportionality of the hazards.
Altogether, 1,092,373 Jewish Israeli males were eligible for the study and were followed for an average of 18 years, constituting 19.2 million person-years. The mean age at examination was 17.4 years; 84.1% were examined at age 17, 9.2% at age 16, 5.3% at age 18 and 1.3% at age 19.
During the follow-up 1001 TGCTs were diagnosed: 534 seminoma and 467 nonseminoma. Cancer morphology is detailed in Table 1. The mean age at diagnosis was 30.9 (SD = 7.6). A significant older mean age at diagnosis was noted for seminoma (33.4, SD = 7.5) compared to nonseminoma (28.0, SD = 6.7), (p < 0.0001), with a mean difference of 5.4 (95% CI 4.5–6.3) years.
Year of birth was strongly associated with TGCT (adjusted HR 1.09; 95% CI 1.08–1.10) and is controlled for in all analyses (Table 2). The mean height was 173.6 centimeters (SD = 6.8), and mean BMI was 21.6 (SD = 3.2). A little over half had completed high school and 21.9% were classified as living in high residential SEP areas. The most common origin was European (Israeli-born) which comprised 32.1% of our study population, followed by Asian origin (24.3%), North African origin (21.9%), European born (11.0%), Israeli origin i.e., 4th generation (5.1%), North African born (3.2%) and Asian born (2.5%).
Height, BMI, years of education and residential SEP differed by origin and were (except for BMI) significantly associated with risk of TGCTs in the univariate analysis, thus were considered in the analysis as possible confounders (Table 2). In the multivariable model, height remained positively associated with TGCT incidence (adjusted HR 1.02; 95% CI 1.01–1.03 per centimeter increment in height) as was increasing years of education.
Origin was strongly associated with TGCTs in the univariate model (Table 2). In a forward stepwise procedure, four variables were retained: origin, year of birth (more recent years at higher risk), height (taller adolescents at higher risk) and years of education (more educated at higher risk). BMI and residential SEP were discarded. The coefficients for origin showed slight attenuation after adjustment in the multivariable model (Table 2). Remarkably low rates were seen for first generation North African born (adjusted HR 0.10; 95% CI 0.04–0.21) and Asian born (adjusted HR 0.35; 95% CI 0.20–0.62) compared to Israeli born of European origin (Fig. 1). This protective effect was partly attenuated in the Israeli-born of North African and Asian origin, with these second and third generation immigrants showing intermediate HRs compared to European origin (0.48; 95% CI 0.40–0.58 and 0.56; 95% CI 0.47–0.66, respectively). For the fourth generation Israeli origin group, the hazard was lower than for European origin. To assess the difference in the hazard of TGCTs by generation we directly compared North African and Asian born to North African and Asian origin, adjusting for year of birth, height and years of education. The hazard was significantly higher in the next generation of North African immigrants (adjusted HR 4.00; 95% CI 1.52–10.49, p = 0.005), whereas a similar comparison for Asia showed an adjusted HR of 1.42 (95% CI 0.76–2.64, p = 0.27). In a pooled analysis of North Africa and Asia (in light of the small numbers of cases among those born abroad) the adjusted HR was 2.31 (95% CI 1.36–3.93, p = 0.002). It is noteworthy that among the European born the incidence was not significantly different from that of the Israeli born of European origin.
Overall, the association of origin was similar for both seminoma and non-seminoma when analyzed separately, adjusted for year of birth, height and years of education (Table 3). Height and year of birth were predictive for both, whereas education was associated with seminoma but not nonseminoma.
To compare the association with year of birth between the origins, we calculated adjusted HRs for year of birth, stratified by origin, controlling for height and years of education. For each origin, year of birth was significantly associated with TGCTs. The steepest rate of change was observed for the North African-born (HR = 1.26, 1.11–1.43), intermediate for Asian-born (HR = 1.14, 1.01–1.29) and lowest for European-born (HR = 1.06, 1.03–1.09). Among the Israeli born, each of the above paternal origins showed similar HRs for year of birth (HRs between 1.09 and 1.11).
In this cohort of over 1 million Israeli male adolescents, followed prospectively for nearly four decades, 1,001 incidence cases TGCT were observed. After adjusting for various possible confounders, the association of origin with TGCTs persisted strongly, with the incidence being lowest for North African and Asian born (first generation), intermediate for Israeli-born of North African and Asian origin and highest for European born or origin. The incidence among Israeli born of North African and Asian origin was more than double that of North African and Asian born. The incidence increased by year of birth for all origin groups. This appeared to be steepest for the North African born who differed significantly from the European born in their rate of increase. Height was found to be a strong risk factor, as has been previously reported.8 The rate of TGCTs increased with increasing years of schooling, more evident for seminoma. Overall, risk factors were similar for seminoma and non-seminoma.
Globally, the incidence of TGCTs among populations of European ancestry is substantially higher than those of African or Asian descent.4 The incidence of TGCTs has been increasing in men of Northern European ancestry since World War II.25 Although the determinants of the increase remain unknown, there is clear evidence that this is consistent with a birth cohort phenomenon.25, 26 An increase in testicular cancer incidence during 1992–2002 was reported in all population groups in Israel, more notable among Jews born in Asia or North Africa and evident also in Arabs.3 We found a significant increase during the 4 decades of the study in all origin groups, steepest for North African-born. Our findings, coupled with the increasing incidence among Arab men in Israel may have implications particularly for the predominantly Arab North African populations that currently display extremely low rates of TCGTs. They may expect sharp increases as the effects of globalization and westernization spread.
An earlier study in Israel investigated the risk of cancer in three generations of young Israelis, by continent of birth and origin.21 This analysis included 298 cases of germ-cell tumors and carcinoma of the testis, aged under age 30 at diagnosis. Differing rates were found among first generation migrants in comparison to Israeli-born whose father was born in Israel (a small proportion of the population composed of mixed origins). Relative risks adjusted for age and time-period were higher for European-born, nonsignificantly lower for Asian-born and significantly lower for North African born. Differences were attenuated for second-generation immigrants. Results were similar for maternal and paternal origin. That study was unable to adjust for important determinants of TGCTs that may be confounders such as height was limited to cancer diagnosed before the age of 30 and was more limited in power. In our cohort, we could explore for the first time the extent to which these associations were confounded or mediated by other risk factors such as birth cohort, height and socio-demographic characteristics. Furthermore, with more than threefold the number of incidence cases power was enhanced, estimates are more stable, direct within-ethnic group comparisons were possible, and we were able to restrict the analysis to TGCTs as well as to analyze seminoma and nonseminoma as separate outcomes. We found a similar pattern to the previous study by origin, after adjusting for possible confounders, as the incidence was very low for adolescents who had themselves emigrated from North Africa and was low for those who had emigrated from Asia, compared to those of European origin. This protective effect was attenuated, although still evident, for those whose father (or grandfather) had emigrated from North Africa or Asia. This loss of protection for Israeli-born sons of immigrants from North Africa compared to those born in North Africa. These findings suggest that among those who emigrated from a non-western life-style, low TGCT incidence country, exposure of the parents, including change in the intrauterine environment, or of the offspring himself to the Israeli context of a medium TGCT incidence country, leads to higher TGCTs risk. The lower risk, though to a lesser extent, found for the Israeli origin group compared to European origin might be explained by the composition of the Israeli origin group which includes an unknown proportion of those originally from low incidence countries of North African or Asian ancestry. The Israeli origin group was characterized by lower years of education and lower residential SEP compared to European born or Israeli-born of European origin, pointing to the ethnic admixture of this group. Several studies have reported on TGCT incidence among immigrants to countries such as Sweden and Denmark.17, 18 Among immigrants to Denmark, a high incidence country, rates among second generation immigrants were similar to the native born (SIR 0.9, 95% CI 0.7–1.1), whereas rates among their first generation immigrant fathers were lower (SIR 0.4, 95% CI 0.4–0.5). In contrast, in our study the difference persisted, albeit attenuated, in Israeli born males of North African and Asian descent. Similarly to Denmark, among immigrants to Sweden, the first generation immigrants displayed incidence rates similar to their country of origin (higher or lower) but this difference tended to disappear in the next generation born in Sweden.27 Changes in that study were most consistent for seminoma. In our study, while seminoma presented at an older age, the associations of origin and other risk factors for TGCTs were generally similar for seminona and nonseminoma, as reported in most studies.28 These patterns of TGCT incidence among sons of immigrants compared to immigrants suggest early life environmental exposures, possibly intrauterine.
The marked geographic variations in risk and the rapid increases in the incidence of TGCTs suggest that environmental factors play an important role in etiology.29 Recent publications of genome-wide association studies of TGCTs support a genetic contribution to TGCTs risk,30, 31 whereas the secular trends persuasively argue for the role of strong environmental determinants that are changing across time. Higher risk for TGCTs found for both monozygotic and dizygotic twins points to the importance of shared in utero exposure. The fact that concordance among monozygotic twins was not greater than among the dizygotic suggests a smaller role for a genetic contribution, although lack of statistical power may have played a role.32, 33 Nonetheless, no specific environmental exposure has yet been found to explain the trend of increasing incidence, although exposure to persistent organochlorine pesticides was reported to be associated with higher risk of TGCTs.34 Therefore, there is need for a continued study of early life environmental exposures. Prenatal exposures have been suggested to be instrumental in testicular cancer development, but despite numerous epidemiological studies, no exposure during this time window has been consistently associated with risk, and novel hypotheses are needed. Richiardi et al. proposed that “future etiological studies on testicular cancer should take postnatal exposures during puberty into account and, whenever possible, investigate both main effects and interactions between perinatal factors, genetic factors and postnatal factors.” We suggest adding investigation of factors which could lead to epigenetic changes, such as parental exposures, and in particular intrauterine exposures, in accordance with recent findings which point to a role of epigenetics in TGCTs.35
A recent meta-analysis has found consistent evidence for a positive association of height measured at age 18 or older with TGCTs.8 Previous studies included in the meta-analysis were predominantly of case-control design, except for one cohort study36 and in most of the studies self-reported anthropometric measures were used. Our large cohort further substantiates the evidence that height is a strong predictor of TGCTs. Growth within the first 2 years of life is largely predictive of secular trends in adult height and adult height can be considered as a surrogate marker of childhood nutrition.37 An inference is that environmental exposures, which account for ∼20% of the variability in adult height in most modern, developed countries, are mainly active within a short time-window during early postnatal development. It has been speculated that the trend of increasing adult height and the increasing TGCTs incidence are biologically interrelated with improved nutrition in early life as the missing link.38 However, as adjusting for adult height in our study did not materially alter the HRs for origin or birth cohort, it is clear that additional factors to those affecting height are important for explanation of the origin and time trend associations. We noted a nonlinear association with overweight and obesity that was not statistically significant.
Early studies indicated that higher SEP was associated with an increased risk, while more current studies have found little association.4 Our finding of an independent association (after controlling for origin) of TGCTs with years of education, a valid lifetime marker of SEP, suggests that higher SEP in Israel, measured across four decades of adolescent birth cohorts, is associated with increased incidence, at least for seminoma. Adjustment for years of education (or residential SEP) as a marker of SEP had little effect on the origin-TGCTs association, which indicates that those characteristics associated with our measures of SEP did not confound the association materially.
Strengths of our migrant cohort include its prospective population-based design reflecting mandatory health examinations among 17-year olds in our population, large sample size, long follow-up, high degree of completeness of the cancer registry data throughout study period, and the ability to adjust for possible confounders such as anthropometry, birth cohort and SEP.
This analysis also has certain limitations. We did not have information on possible perinatal risk factors, such as maternal bleeding during pregnancy, birth order, sibship size, birth weight, gestational age and twinning.6, 7 We had data on country of birth and paternal country of birth, but could not distinguish between second and third migrant generations and did not have data on maternal origin. However, maternal and paternal origins in our population were previously shown to be highly associated.39 A previous smaller study in Israel found similar associations of testicular cancer with maternal and paternal origin.21 Despite our large sample size and number of cases, we had relatively small number of cases for Asian-born and North African-born immigrants, thus estimates for these groups may be less stable. There was only one examination of height and weight, conducted during late adolescence. Changes in final adult height after age 17 are expected to be modest. However, significant changes in BMI (due to weight changes) between mean age 17 to mean age 30 were previously reported in a sub-sample of our cohort.40 Indeed, differences in the timing and method of weight measurement were given as possible explanations for heterogeneity in the weight and TGCTs meta-analyses.8 We had shorter follow-up periods for the more recent birth cohorts; therefore our analysis better represents risks for TGCTs manifesting at a young age, rather than lifetime risk. This could be more of an issue for seminoma, as suggested by the slightly younger mean age at diagnosis in our cohort (33.4, SD = 7.5) compared to most other studies.5 However, as TGCTs are mostly diagnosed during young adulthood, predictors are expected to be similar in magnitude. This assumption can in the future be verified by continued follow-up of our cohort in the coming years.
In conclusion, we found far lower rates of TGCTs among North African and Asian born men than among those of European origin, with a substantial increase in the Israeli-born of North African and Asian origin to about half the rates of the European origin in the second and third generations of migrants i.e., reflecting a persistent though declining protective effect. These findings provide helpful clues for further research on the role of modern lifestyle and environment in the etiology of TGCTs. Furthermore, our findings, coupled with the increasing incidence among Arab men in Israel may have implications particularly for the predominantly Arab North African populations that currently display extremely low rates of TCGTs. They may expect sharp increases as the effects of globalization and westernization spread.
- 1Wein AJ, Kavoussi LR, Novick AC, et al., eds. Campbell–Walsh urology, 9th edn., Philadelphia: Saunders Elsevier, 2007, 2205–306.
- 3Epidemiological characteristics and trends of testicular cancer in Israel 1992–2002. Harefuah 2007; 146: 515–9, 575., , , et al.
- 12Parkin DM, Whelan SL, Ferlay J, et al., eds. Cancer incidence in five continents, vol. 8. IARC Scientific Publication No 155. Lyon, France: IARC Press, 2002.
- 20Rosen B, Merkur S, eds. Israel: health system review. Health Syst Trans 2009; 11: 1–226.
- 21The risk of cancer in 3 generations of young Israelis: a study of migrants and their descendants. IARC Technical Report 27, International Agency for Research on Cancer. Lyon, France: IARC Press, 1998., .