Abstract
Background
Few risk factors have been established for childhood germ cell tumors (GCT). Parental infertility and infertility treatment may be associated with GCT development but these risk factors have not been fully investigated.
Methods
A case-control study of childhood GCT was conducted through the Children’s Oncology Group (COG). Cases, under the age of 15 years at diagnosis, were recruited through COG institutions from January 1993 to December 2002. Controls were obtained through random digit dialing. Information about infertility and infertility treatment along with demographic factors was collection through maternal interviews. Subgroups created by gender, age at diagnosis, and tumor location were examined separately. Statistical analysis was performed using multivariate logistic regression models.
Results
Overall, no association between GCT and infertility or its treatment was found. In subgroup analysis, females whose mothers had two or more fetal losses were found to be at increased risk for non-gonadal tumors (Odds ratio (OR) = 3.32, 95% Confidence interval (CI) = 1.12–9.88). Younger maternal age was associated with a lower risk of gonadal GCT in females (OR = 0.52, 95% CI = 0.28–0.96). There was an increased risk of all GCT and gonadal GCT in males born to older mothers (OR = 2.88, 95% CI = 1.13–7.37 and OR = 3.70, 95% CI = 1.12–12.24).
Conclusion
While no association between parental infertility or its treatment and childhood GCT was found overall, possible associations with maternal age and history of recurrent fetal loss were found in subgroups defined by gender.
Keywords: Germ cell tumor, infertility, pediatrics, epidemiology
Introduction
Childhood germ cell tumors (GCT) are a group of heterogeneous diagnoses with a common origin in the primordial germ cell[1]. There is a distinctive U-shaped incidence curve by age with incidence rates of 21.0, 3.9, 2.2, 7.4, and 29.0 per million for age groups <1, 1–4, 5–9, 10–14, and 15–19 years, respectively, in the United States from 1992–2004[1]. In addition to the wide variance with age, there is a difference in the location of the tumors by sex. In females below the age of 4 years, the most common type of GCT is non-gonadal GCT, whereas in boys testicular GCT and non-gonadal GCT have approximately equal incidence rates in infancy and, after that, gonadal GCT predominates until age four[1]. Incidence is low for both males and females until around puberty when the incidence of ovarian GCT in females and testicular GCT in males rises dramatically [1].
The literature on the etiology of GCT in children under the age of 15 years is sparse, and associations with potential risk factors have been noted only inconsistently [2–12]. Parental infertility and infertility treatment have not been studied previously in pediatric GCT; however, there are reasons that an association would be plausible. For example, superovulation has been shown to disrupt acquisition and maintenance of genomic imprinting in developing embryos [13]. This abnormal imprinting could potentially lead to tumor development.
There are also factors associated with infertility or infertility treatment that could increase risk of GCT. Congenital abnormalities of the genital organs have been found to be associated with childhood GCT [5, 6, 8, 11] and studies have indicated increased risk of these abnormalities following intracytoplasmic sperm injection (ICSI) [14, 15]. Another study noted an increased risk of cryptorchidism after intrauterine insemination, although there was no increase after in vitro fertilization or ICSI [16]. An increase in congenital abnormalities could suggest an increase in cancer. Risk of testicular GCT could also be increased through transmission of infertility to children born through the use of ICSI procedures [17–19]. Male factor infertility has been shown to increase risk for adult GCT [20–22] and if male children of infertile males are more likely to be infertile themselves, their risk for testicular GCT could also increase. However, it is not clear that the link between GCT and male infertility could be extended to childhood GCT, given their histological and cytogenetic differences that imply distinct etiologies. Thus more research is needed into parental infertility and risk of childhood GCT.
We used data from a Children’s Oncology Group (COG) case-control study to evaluate the potential association between parental infertility and its treatment and GCT in children.
Materials and Methods
The methods for this study have been detailed previously [3, 4]. Briefly, a total of 84 member institutions of the COG provided cases for this study. Cases were eligible if they had a newly diagnosed GCT under the age of 15 years from January 1993 to December of 2002. Specific GCT diagnoses included germinoma, embryonal carcinoma, yolk sac tumor, choriocarcinoma, immature teratoma, and mixed GCT. Patients with tumors located in the brain were excluded. Cases were required to have a biological mother available who spoke English and have a telephone in their residence. The treating physician gave approval for study coordinators to contact the case’s mother.
Controls were selected through random digit dialing and were frequency matched to cases based on sex, age, and telephone exchange. Using methods similar to those of Waksberg [23, 24], potential control phone numbers were generated from case phone numbers. The area code and exchange of the case phone number were retained and the last four digits were randomly selected in order to obtain a control number. If a residential number was reached, screening questions were asked to determine whether or not a child under the age of 15 was present in the household. Case-control ratios were different based on sex; for males the ratio was 1:2 and females 1:1 since GCT in children is more common in females.
Approval for this study was granted by the institutional review boards at each participating institution as well as from the University of Minnesota.
Information on potential exposures was collected for cases and controls through maternal interview. The interview included questions about pregnancy history, maternal exposures during pregnancy with the index child, family history of cancer and other diseases, and information about the medical history of the mother. In the maternal questionnaire, several questions about infertility and infertility treatment were asked including length of time to index pregnancy, history of infertility (more than one year of trying without becoming pregnant), visit to a doctor by mother or index biological father due to non-pregnancy, specific infertility treatment, use of ovulating stimulating drugs in the year before or during pregnancy, history of multiple birth, and history of recurrent pregnancy loss.
Unconditional logistic regression was used to explore the association between infertility and infertility treatment and GCT. Exposures were included in the logistic analysis if at least four cases and controls were represented in all exposure categories. Matching factors sex and year of birth were included in all analyses. In addition, after exploration of potential confounding variables, the following variables were included in all models: maternal age, maternal race, household income, maternal education, and child’s gestational age. Results are reported as odds ratios (OR) and 95% confidence intervals (CI). Subgroup analysis was performed for age at diagnosis (<5, 5+) and tumor location (non-gonadal, gonadal), and gender. A sensitivity analysis was performed in subgroups limited to children with white, non-Hispanic mothers to assess the influence of race/ethnicity on results. Gender was analyzed in subgroups rather than through interactions since the etiology of GCT may differ by gender. All analysis was performed using SAS 9.1 (SAS institute, Inc, Cary, NC).
An additional method for assessing infertility was explored based on latent class analysis (LCA) [25]. The variables used in the latent class analysis were maternal age, history of infertility, use of female hormones before or during pregnancy, history of multiple birth, and history of recurrent pregnancy loss. While each variable individually has the potential for misclassification of the true exposure of interest, using all variables together may mitigate this misclassification. Models with and without maternal age were explored to determine if the effect of infertility was only through maternal age or if there was an independent risk factor for infertility apart from age. LCA was conducted using M-Plus software [26]. Predicted class membership was dichotomized and used as a predictor in a logistic regression model along with potential confounders.
Results
Out of 344 eligible cases, 278 completed the maternal interview. Controls were obtained from a total of 17,292 randomly selected phone numbers with 5,912 residential numbers identified. There were 634 households with confirmed eligible children, of which 422 completed maternal interviews (67%).
Demographic characteristics of cases and controls are presented in table 1. Cases were more likely to be non-white and have a high school education or less. Cases and controls were similar with regard to gestational age and birth weight.
Table 1.
Demographic characteristics
| Overall | Males | Females | ||||
|---|---|---|---|---|---|---|
| Controls n (%) |
Cases n (%) |
Controls n (%) |
Cases n (%) |
Controls n (%) |
Cases n (%) |
|
| Mother’s race | ||||||
| White | 356 (84.4) | 213 (76.9) | 152 (84.0) | 70 (84.3) | 203 (84.6) | 143 (73.7) |
| Non-white | 66 (15.6) | 64 (23.1) | 29 (16.0) | 13 (15.7) | 37 (15.4) | 51 (26.3) |
| Mother’s education | ||||||
| ≤ High school | 166 (39.3) | 138 (49.8) | 66 (36.6) | 39 (47.0) | 100 (41.7) | 99 (51.0) |
| Some post HS | 97 (23.0) | 53 (19.1) | 40 (22.1) | 16 (19.3) | 57 (23.8) | 37 (19.1) |
| College graduate | 159 (37.7) | 86 (31.0) | 75 (41.4) | 28 (33.7) | 83 (34.6) | 58 (29.9) |
| Household income | ||||||
| ≤ $30,000 | 192 (45.9) | 145 (52.9) | 81 (44.8) | 40 (50.0) | 111 (47.0) | 105 (54.1) |
| $30,001–$50,000 | 124 (29.7) | 63 (23.0) | 53 (29.3) | 21 (26.3) | 70 (29.7) | 42 (21.6) |
| > $50,000 | 102 (24.4) | 66 (24.1) | 47 (26.0) | 19 (23.8) | 55 (23.3) | 47 (24.2) |
| Index child’s gestational age | ||||||
| < 37 weeks | 44 (10.4) | 35 (12.6) | 16 (8.8) | 10 (12.0) | 27 (11.2) | 25 (12.8) |
| 37–41 weeks | 343 (81.1) | 220 (79.1) | 150 (82.9) | 67 (80.7) | 193 (80.1) | 153 (78.5) |
| ≥ 42 weeks | 36 (8.5) | 23 (8.3) | 15 (8.3) | 6 (7.2) | 21 (8.7) | 17 (8.7) |
| Index child’s birth weight Mean (SD) | 3373.72 (590.31) | 3354.27 (680.58) | 3462.77 (607.81) | 3499.00 (644.45) | 3309.19 (569.33) | 3292.67 (687.78) |
| Index child’s gender | ||||||
| Male | 181 (42.9) | 83 (29.9) | ||||
| Female | 241 (57.1) | 195 (70.1) | ||||
| Index child’s age at diagnosis | ||||||
| <1 years | 57 (20.5) | 22 (26.5) | 35 (17.9) | |||
| 1–4 years | 91 (32.7) | 42 (50.6) | 49 (25.1) | |||
| 5–9 years | 41 (14.7) | 4 (4.8) | 37 (19.0) | |||
| 10–15 years | 89 (32.0) | 15 (18.1) | 74 (37.9) | |||
| Type of tumor | ||||||
| Gonadal | 144 (53.1) | 47 (58.0) | 97 (51.1) | |||
| Non-gonadal | 120 (44.3) | 31 (38.3) | 89 (46.8) | |||
| Metastatic | 7 (2.6) | 3 (3.7) | 4 (2.1) | |||
Overall results suggested little relationship between infertility or infertility treatment and GCT (table 2). Gonadal and non-gonadal subgroup results were fairly concordant with the overall results. However, analysis for each gender separately indicated some potential differences between cases and controls (tables 3 and 4). For females, there was an increased risk of non-gonadal GCT in children born to women with two or more fetal losses (OR = 3.32, 95% CI = 1.12–9.88, p = 0.03). Maternal age <27 was found to be significantly associated with a lower risk of gonadal GCT in females (OR = 0.52, 95% CI = 0.28–0.96, p = 0.04). Finally, in males there appeared to be an increased risk of GCT in children born to mothers aged 35 years or greater both overall (OR = 2.88, 95% CI = 1.13–7.37, p = 0.03) and in gonadal tumors (OR = 3.70, 95% CI = 1.12–12.24, p = 0.03). Findings were similar in analysis limited to children of white, non-Hispanic mothers (data not shown).
Table 2.
Association between childhood GCT and parental infertility and infertility treatment
| Controls n (%) |
Cases n (%) |
ORa | 95% CI | p-value | Gonadal n (%) |
ORa | 95% CI | p-value | Non- gonadal n (%) |
ORa | 95% CI | p-value | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Prior fetal loss | |||||||||||||
| None | 332 (78.5) | 227 (81.7) | Ref | 119 (82.6) | Ref | 97 (80.8) | Ref | ||||||
| One | 74 (17.5) | 34 (12.2) | 0.65 | 0.41–1.03 | 0.06 | 19 (13.2) | 0.65 | 0.37–1.16 | 0.15 | 13 (10.8) | 0.57 | 0.29–1.12 | 0.10 |
| Two or more | 17 (4.0) | 17 (6.1) | 1.30 | 0.63–2.71 | 0.48 | 6 (4.2) | 0.96 | 0.35–2.63 | 0.93 | 10 (8.3) | 1.51 | 0.62–3.68 | 0.37 |
| History of multiple birth prior to index | |||||||||||||
| None | 410 (96.9) | 269 (96.8) | Ref | 139 (96.5) | Ref | 117 (97.5) | Ref | ||||||
| At least one multiple birth | 13 (3.1) | 9 (3.2) | 1.08 | 0.43–2.67 | 0.87 | 5 (3.5) | 1.29 | 0.43–3.91 | 0.65 | 3 (2.5) | c | ||
| Maternal age | |||||||||||||
| < 27 | 187 (44.2) | 126 (45.3) | 0.84 | 0.59–1.21 | 0.36 | 64 (44.4) | 0.79 | 0.50–1.25 | 0.32 | 54 (45.0) | 0.91 | 0.55–1.50 | 0.71 |
| 27–34 | 197 (46.6) | 123 (44.2) | Ref | 66 (45.8) | Ref | 54 (45.0) | Ref | ||||||
| ≥35 | 39 (9.2) | 29 (10.4) | 1.23 | 0.70–2.15 | 0.47 | 14 (9.7) | 1.23 | 0.61–2.49 | 0.57 | 12 (10.0) | 1.05 | 0.49–2.29 | 0.90 |
| Use of OSd drugs (prior to or before knowledge of pregnancy) | |||||||||||||
| No | 410 (97.4) | 273 (98.2) | Ref | 139 (96.5) | Ref | 120 (100.0) | Ref | ||||||
| Yes | 11 (2.6) | 5 (1.8) | 0.84 | 0.28–2.58 | 0.77 | 5 (3.5) | 1.48 | 0.48–4.63 | 0.50 | 0 (0.0) | c | ||
| No pregnancy after a year or more of trying | |||||||||||||
| No | 334 (79.1) | 221 (79.5) | Ref | 112 (77.8) | Ref | 96 (80.0) | Ref | ||||||
| Yes | 88 (20.9) | 57 (20.5) | 0.95 | 0.64–1.41 | 0.78 | 32 (22.2) | 1.01 | 0.62–1.63 | 0.97 | 24 (20.0) | 0.91 | 0.52–1.57 | 0.72 |
| Time to index pregnancy | |||||||||||||
| Not trying | 191 (45.3) | 134 (48.4) | 0.93 | 0.65–1.32 | 0.67 | 68 (47.2) | 0.97 | 0.63–1.51 | 0.90 | 58 (48.7) | 0.92 | 0.56–1.52 | 0.75 |
| < 1 year of trying | 183 (43.4) | 119 (43.0) | Ref | 61 (42.4) | Ref | 52 (43.7) | Ref | ||||||
| ≥ 1 year of trying | 48 (11.4) | 24 (8.7) | 0.71 | 0.40–1.26 | 0.24 | 15 (10.4) | 0.86 | 0.44–1.72 | 0.67 | 9 (7.6) | 0.65 | 0.29–1.49 | 0.31 |
| Latent class infertilityb (with maternal age) | |||||||||||||
| No | 390 (92.2) | 254 (91.4) | Ref | 130 (90.3) | Ref | 111 (92.5) | Ref | ||||||
| Yes | 33 (7.8) | 24 (8.6) | 1.05 | 0.56–1.98 | 0.87 | 14 (9.7) | 1.32 | 0.62–2.79 | 0.47 | 9 (7.5) | 0.78 | 0.32–1.89 | 0.57 |
| Latent class infertilityb (without maternal age) | |||||||||||||
| No | 408 (96.5) | 267 (96.0) | Ref | 138 (95.8) | Ref | 115 (95.8) | Ref | ||||||
| Yes | 15 (3.5) | 11 (4.0) | 1.18 | 0.51–2.73 | 0.70 | 6 (4.2) | 1.20 | 0.43–3.32 | 0.73 | 5 (4.2) | 1.23 | 0.39–3.86 | 0.72 |
Logistic regression models were adjusted for child’s age, sex, and gestational age as well as maternal age, race, education, and household income
Infertility predicted from the latent class model
Sample size too small for logistic regression analysis
Ovarian stimulating
Table 3.
Association between childhood GCT and parental infertility and infertility treatment (Females only)
| Controls n (%) |
Cases n (%) |
ORa | 95% CI | p-value | Gonadal n (%) |
ORa | 95% CI | p-value | Non- gonadal n (%) |
ORa | 95% CI | p-value | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Prior fetal loss | |||||||||||||
| None | 185 (76.8) | 157 (80.5) | Ref | 80 (82.5) | Ref | 69 (77.5) | Ref | ||||||
| One | 48 (19.9) | 23 (11.8) | 0.59 | 0.34–1.03 | 0.07 | 12 (12.4) | 0.48 | 0.22–1.02 | 0.06 | 10 (11.2) | 0.61 | 0.28–1.36 | 0.23 |
| Two or more | 8 (3.3) | 15 (7.7) | 2.12 | 0.84–5.36 | 0.11 | 5 (5.2) | 0.77 | 0.21–2.87 | 0.70 | 10 (11.2) | 3.32 | 1.12–9.88 | 0.03 |
| History of multiple birth prior to index | |||||||||||||
| None | 235 (97.5) | 187 (95.9) | Ref | 93 (95.9) | Ref | 86 (96.6) | Ref | ||||||
| At least one multiple birth | 6 (2.5) | 8 (4.1) | 1.65 | 0.53–5.15 | 0.39 | 4 (4.1) | 1.55 | 0.36–6.61 | 0.56 | 3 (3.4) | c | ||
| Maternal age | |||||||||||||
| < 27 | 119 (49.4) | 90 (46.2) | 0.67 | 0.43–1.05 | 0.08 | 44 (45.4) | 0.52 | 0.28–0.96 | 0.04 | 41 (46.1) | 0.87 | 0.48–1.58 | 0.64 |
| 27–34 | 97 (40.2) | 88 (45.1) | Ref | 46 (47.4) | Ref | 39 (43.8) | Ref | ||||||
| ≥35 | 25 (10.4) | 17 (8.7) | 0.79 | 0.39–1.61 | 0.52 | 7 (7.2) | 0.63 | 0.22–1.76 | 0.38 | 9 (10.1) | 0.87 | 0.35–2.20 | 0.77 |
| Use of OSd drugs (prior to or before knowledge of pregnancy) | |||||||||||||
| No | 234 (97.9) | 191 (97.9) | Ref | 93 (95.9) | Ref | 89 (100.0) | Ref | ||||||
| Yes | 5 (2.1) | 4 (2.1) | 0.99 | 0.25–3.94 | 0.99 | 4 (4.1) | 1.79 | 0.42–7.66 | 0.43 | 0 (0.0) | c | ||
| No pregnancy after a year or more of trying | |||||||||||||
| No | 184 (76.3) | 155 (79.5) | Ref | 76 (78.4) | Ref | 70 (78.7) | Ref | ||||||
| Yes | 57 (23.7) | 40 (20.5) | 0.85 | 0.53–1.37 | 0.50 | 21 (21.6) | 1.02 | 0.53–1.95 | 0.95 | 19 (21.3) | 0.86 | 0.45–1.64 | 0.65 |
| Time to index pregnancy | |||||||||||||
| Not trying | 119 (49.4) | 94 (48.5) | 0.75 | 0.48–1.20 | 0.23 | 50 (51.5) | 0.88 | 0.48–1.62 | 0.68 | 40 (45.5) | 0.71 | 0.37–1.33 | 0.28 |
| < 1 year of trying | 95 (39.4) | 82 (42.3) | Ref | 37 (38.1) | Ref | 40 (45.5) | Ref | ||||||
| ≥ 1 year of trying | 27 (11.2) | 18 (9.3) | 0.77 | 0.39–1.53 | 0.46 | 10 (10.3) | 1.21 | 0.48–3.03 | 0.69 | 8 (9.1) | 0.67 | 0.27–1.70 | 0.40 |
| Latent class infertilityb (with maternal age) | |||||||||||||
| No | 222 (92.1) | 177 (90.8) | Ref | 88 (90.7) | Ref | 80 (89.9) | Ref | ||||||
| Yes | 19 (7.9) | 18 (9.2) | 1.23 | 0.57–2.62 | 0.60 | 9 (9.3) | 1.52 | 0.55–4.25 | 0.42 | 9 (10.1) | 1.19 | 0.45–3.19 | 0.71 |
| Latent class infertilityb (without maternal age) | |||||||||||||
| No | 235 (97.5) | 185 (94.9) | Ref | 92 (94.8) | Ref | 84 (94.4) | Ref | ||||||
| Yes | 6 (2.5) | 10 (5.1) | 2.35 | 0.80–6.92 | 0.12 | 5 (5.2) | 1.75 | 0.47–6.52 | 0.41 | 5 (5.6) | 3.06 | 0.77–12.10 | 0.11 |
Logistic regression models were adjusted for child’s age, sex, and gestational age as well as maternal age, race, education, and household income
Infertility predicted from the latent class model
Sample size too small for logistic regression analysis
Ovarian stimulating
Table 4.
Association between childhood GCT and parental infertility and infertility treatment (Males only)
| Controls n (%) |
Cases n (%) |
ORa | 95% CI | p-value | Gonadal n (%) |
ORa | 95% CI | p-value | Non-gonadalb n (%) |
|
|---|---|---|---|---|---|---|---|---|---|---|
| Prior fetal loss | ||||||||||
| None | 146 (80.7) | 70 (84.3) | Ref | 39 (83.0) | Ref | 28 (90.3) | ||||
| One | 26 (14.4) | 11 (13.3) | 0.71 | 0.30–1.70 | 0.44 | 7 (14.9) | 1.06 | 0.37–3.09 | 0.91 | 3 (9.7) |
| Two or more | 9 (5.0) | 2 (2.4) | d | 1 (2.1) | d | 0 (0.0) | ||||
| History of multiple birth prior to index | ||||||||||
| None | 174 (96.1) | 82 (98.8) | Ref | 46 (97.9) | Ref | 31 (100.0) | ||||
| At least one multiple birth | 7 (3.9) | 1 (1.2) | d | 1 (2.1) | d | 0 (0.0) | ||||
| Maternal age | ||||||||||
| < 27 | 68 (37.6) | 36 (43.4) | 1.37 | 0.70–2.67 | 0.36 | 20 (42.6) | 1.55 | 0.66–3.68 | 0.32 | 13 (41.9) |
| 27–34 | 99 (54.7) | 35 (42.2) | Ref | 20 (42.6) | Ref | 15 (48.4) | ||||
| ≥35 | 14 (7.7) | 12 (14.5) | 2.88 | 1.13–7.37 | 0.03 | 7 (14.9) | 3.70 | 1.12–12.24 | 0.03 | 3 (9.7) |
| Use of OSd drugs (prior to or before knowledge of pregnancy) | ||||||||||
| No | 175 (96.7) | 82 (98.8) | Ref | 46 (97.9) | Ref | 31 (100.0) | ||||
| Yes | 6 (3.3) | 1 (1.2) | d | 1 (2.1) | d | 0 (0.0) | ||||
| No pregnancy after a year or more of trying | ||||||||||
| No | 149 (82.8) | 66 (79.5) | Ref | 36 (76.6) | Ref | 26 (83.9) | ||||
| Yes | 31 (17.2) | 17 (20.5) | 1.30 | 0.61–2.75 | 0.50 | 11 (23.4) | 1.80 | 0.70–4.64 | 0.23 | 5 (16.1) |
| Time to index pregnancy | ||||||||||
| Not trying | 71 (39.4) | 40 (48.2) | 1.17 | 0.64–2.14 | 0.62 | 18 (38.3) | 0.88 | 0.41–1.90 | 0.75 | 18 (58.1) |
| < 1 year of trying | 88 (48.9) | 37 (44.6) | Ref | 24 (51.1) | Ref | 12 (38.7) | ||||
| ≥ 1 year of trying | 21 (11.7) | 6 (7.2) | 0.56 | 0.17–1.81 | 0.33 | 5 (10.6) | 1.09 | 0.27–4.36 | 0.90 | 1 (3.2) |
| Latent class infertilityb (with maternal age) | ||||||||||
| No | 167 (92.3) | 77 (92.8) | Ref | 42 (89.4) | Ref | 31 (100.0) | ||||
| Yes | 14 (7.7) | 6 (7.2) | 0.72 | 0.21–2.48 | 0.60 | 5 (10.6) | 1.68 | 0.39–7.29 | 0.49 | 0 (0.0) |
| Latent class infertilityb (without maternal age) | ||||||||||
| No | 172 (95.0) | 82 (98.8) | Ref | 46 (97.9) | Ref | 31 (100.0) | ||||
| Yes | 9 (5.0) | 1 (1.2) | d | 1 (2.1) | d | 0 (0.0) |
Logistic regression models were adjusted for child’s age, sex, and gestational age as well as maternal age, race, education, and household income
Too few cases of non-gonadal for logistic regression analysis
Infertility predicted from the latent class model
Sample size too small for logistic regression analysis
Ovarian stimulating
No notable associations were observed in analysis stratified by age at diagnosis (data not shown).
Discussion
While parental infertility or treatment was not found to be a risk factor for childhood GCT overall, there were some associations in clinically relevant subgroups. Specifically, we found that history of recurrent pregnancy loss elevated risk of non-gonadal GCT in females. Although the American Society for Reproductive Medicine defines recurrent pregnancy loss as a separate disease than infertility [27], other evidence shows that repeated pregnancy loss may be related to infertility, particularly in the male [28]. Still, there are many potential causes of recurrent pregnancy loss including a large percentage with no known cause [29]; thus possible underlying reasons for this association would be difficult to determine. However, the association found in this study could represent a novel risk factor for non-gonadal GCT in females which may warrant further examination.
The associations with maternal age in female and male gonadal GCT are in opposite directions with young and old maternal age protective in females and detrimental in males. This type of qualitative interaction is more likely to be due to chance than represent a real biological phenomenon since it seems unlikely that the effect of maternal age would be opposite for male and female gonadal GCT. As others have noted, it is usually biologically implausible for a single exposure to both cause and protect from the same disease in different subgroups, but is more likely a statistical anomaly [30]. In addition there is little evidence for a maternal age effect in GCT both in childhood [31] and in adolescent and adult testicular cancer [32].
Very few studies of pediatric GCT have been conducted. Of those that have been carried out, this study includes the largest number of cases. The relatively large size allowed us to examine subgroups within GCT which potentially have different etiologies and is the study’s greatest strength.
Several limitations also apply to this study. First, even though the number of GCT case was relatively large it was still small in absolute number, particularly in subgroup analysis, leaving us with limited ability to detect moderate associations. Many associations were examined in this study as well, which could lead to false positive findings due to chance alone. There is also the potential for selection bias to influence the study results. Our controls were more likely to be white and have higher income and education compared to cases which may indicate some level of selection bias. However, since these factors are antecedents to both the exposure and disease, adjusting for them in the model may be expected to reduce selection bias [33].
Selection of cases through COG may also have introduced bias, as the proportion of children with cancer seen at COG institutions decreases as the age of the child increases [34]. However, most cases of childhood GCT occur before the age of 5 so that COG institutions would be likely to see a majority of cases; in the current study over half of the cases in this study were diagnosed at 5 years of age or younger. In addition, in this study, a higher proportion of girls with GCT participated as well as a higher percentage of children who were white [3]. This might indicate some differential selection in the cases; however, since controls are also expected to have a higher proportion of children with white race, this could, in fact, mitigate the impact of selection bias. In any case, COG is the only organization which could practicably facilitate case recruitment for such a rare disease in North America.
Another limitation is the potential for recall bias since the exposures of interest could have taken place many years prior to the study. Other studies have examined the possibility of recall bias relating to infertility and infertility treatment. History of infertility was found to be self-reported less often than on medical records as well as slightly more often by cases than controls in one study [35]. Another study showed little difference in the sensitivity of reporting previous spontaneous abortion in cases and controls in a study on sudden infant death syndrome [36]. In this study, the only significant associations were found for maternal age and history of recurrent fetal loss both of which appear to be less subject to recall bias.
While we did not find any association with infertility or infertility treatment and pediatric GCTs, we did find some possible associations with maternal age and recurrent pregnancy loss in sex and location specific subgroups. These associations are likely due to chance but could be considered in future studies.
Acknowledgments
This work was supported by the National Institutes of Health [R01CA067263, U10CA98413, U10CA13539 and U10CA98543]; and the Children’s Cancer Research Fund (Minneapolis, Minnesota).
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of interest
None.
References
- 1.Bernstein L, Smith MA, Liu L, Deapen D, Friedman DL. Germ cell, trophoblastic and other gonadal neoplasms. In: Ries L, Smith M, Gurney J, Linet M, Tamra T, Young J, et al., editors. Cancer Incidence and Survival among Children and Adolescents: United States SEER Program 1975–1995. Bethesda, MD: National Cancer Institute, SEER Program; 1999. pp. 125–138. NIH Pub. No. 99-4649. [Google Scholar]
- 2.Chen Z, Robison L, Giller R, Krailo M, Davis M, Davies S, et al. Environmental exposure to residential pesticides, chemicals, dusts, fumes, and metals, and risk of childhood germ cell tumors. Int J Hyg Environ Health. 2006;209(1):31–40. doi: 10.1016/j.ijheh.2005.08.001. [DOI] [PubMed] [Google Scholar]
- 3.Chen Z, Robison L, Giller R, Krailo M, Davis M, Gardner K, et al. Risk of childhood germ cell tumors in association with parental smoking and drinking. Cancer. 2005;103(5):1064–1071. doi: 10.1002/cncr.20894. [DOI] [PubMed] [Google Scholar]
- 4.Chen Z, Stewart PA, Davies S, Giller R, Krailo M, Davis M, et al. Parental occupational exposure to pesticides and childhood germ-cell tumors. Am J Epidemiol. 2005;162(9):858–867. doi: 10.1093/aje/kwi294. [DOI] [PubMed] [Google Scholar]
- 5.Johnston HE, Mann JR, Williams J, Waterhouse JA, Birch JM, Cartwright RA, et al. The Inter-Regional, Epidemiological Study of Childhood Cancer (IRESCC): case-control study in children with germ cell tumours. Carcinogenesis. 1986;7(5):717–722. doi: 10.1093/carcin/7.5.717. [DOI] [PubMed] [Google Scholar]
- 6.Shankar S, Davies S, Giller R, Krailo M, Davis M, Gardner K, et al. In utero exposure to female hormones and germ cell tumors in children. Cancer. 2006;106(5):1169–1177. doi: 10.1002/cncr.21670. [DOI] [PubMed] [Google Scholar]
- 7.Shu XO, Nesbit ME, Buckley JD, Krailo MD, Robinson LL. An exploratory analysis of risk factors for childhood malignant germ-cell tumors: report from the Childrens Cancer Group (Canada, United States) Cancer Causes Control. 1995;6(3):187–198. doi: 10.1007/BF00051790. [DOI] [PubMed] [Google Scholar]
- 8.Swerdlow AJ, Stiller CA, Wilson LM. Prenatal factors in the aetiology of testicular cancer: an epidemiological study of childhood testicular cancer deaths in Great Britain, 1953-73. J Epidemiol Community Health. 1982;36(2):96–101. doi: 10.1136/jech.36.2.96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Walker AH, Ross RK, Haile RW, Henderson BE. Hormonal factors and risk of ovarian germ cell cancer in young women. Br J Cancer. 1988;57(4):418–422. doi: 10.1038/bjc.1988.95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Johnson KJ, Poynter JN, Ross JA, Robison LL, Shu XO. Pediatric germ cell tumors and maternal vitamin supplementation: a Children's Oncology Group study. Cancer Epidemiol Biomarkers Prev. 2009;18(10):2661–2664. doi: 10.1158/1055-9965.EPI-09-0462. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Johnson KJ, Ross JA, Poynter JN, Linabery AM, Robison LL, Shu XO. Paediatric germ cell tumours and congenital abnormalities: a Children's Oncology Group study. Br J Cancer. 2009;101(3):518–521. doi: 10.1038/sj.bjc.6605169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Poynter JN, Radzom AH, Spector LG, Puumala S, Robison LL, Chen Z, et al. Family history of cancer and malignant germ cell tumors in children: A report from the Children's Oncology Group. Cancer Causes Control. 2009 doi: 10.1007/s10552-009-9448-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Market-Velker BA, Zhang L, Magri LS, Bonvissuto AC, Mann MR. Dual Effects of Superovulation: Loss of maternal and paternal imprinted methylation in a dose-dependent manner. Hum Mol Genet. 2009 doi: 10.1093/hmg/ddp465. [DOI] [PubMed] [Google Scholar]
- 14.Ericson A, Kallen B. Congenital malformations in infants born after IVF: a population-based study. Hum Reprod. 2001;16(3):504–509. doi: 10.1093/humrep/16.3.504. [DOI] [PubMed] [Google Scholar]
- 15.Zhu JL, Basso O, Obel C, Bille C, Olsen J. Infertility, infertility treatment, and congenital malformations: Danish national birth cohort. Bmj. 2006;333(7570):679. doi: 10.1136/bmj.38919.495718.AE. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Damgaard IN, Jensen TK, Petersen JH, Skakkebaek NE, Toppari J, Main KM. Risk factors for congenital cryptorchidism in a prospective birth cohort study. PLoS ONE. 2008;3(8):e3051. doi: 10.1371/journal.pone.0003051. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Feng C, Wang LQ, Dong MY, Huang HF. Assisted reproductive technology may increase clinical mutation detection in male offspring. Fertil Steril. 2008;90(1):92–96. doi: 10.1016/j.fertnstert.2007.06.004. [DOI] [PubMed] [Google Scholar]
- 18.Cram DS, Ma K, Bhasin S, Arias J, Pandjaitan M, Chu B, et al. Y chromosome analysis of infertile men and their sons conceived through intracytoplasmic sperm injection: vertical transmission of deletions and rarity of de novo deletions. Fertil Steril. 2000;74(5):909–915. doi: 10.1016/s0015-0282(00)01568-5. [DOI] [PubMed] [Google Scholar]
- 19.Page DC, Silber S, Brown LG. Men with infertility caused by AZFc deletion can produce sons by intracytoplasmic sperm injection, but are likely to transmit the deletion and infertility. Hum Reprod. 1999;14(7):1722–1726. doi: 10.1093/humrep/14.7.1722. [DOI] [PubMed] [Google Scholar]
- 20.Walsh TJ, Croughan MS, Schembri M, Chan JM, Turek PJ. Increased risk of testicular germ cell cancer among infertile men. Arch Intern Med. 2009;169(4):351–356. doi: 10.1001/archinternmed.2008.562. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Skakkebaek NE, Holm M, Hoei-Hansen C, Jorgensen N, Rajpert-De Meyts E. Association between testicular dysgenesis syndrome (TDS) and testicular neoplasia: evidence from 20 adult patients with signs of maldevelopment of the testis. Apmis. 2003;111(1):1–9. doi: 10.1034/j.1600-0463.2003.11101031.x. discussion 9–11. [DOI] [PubMed] [Google Scholar]
- 22.Peng X, Zeng X, Peng S, Deng D, Zhang J. The association risk of male subfertility and testicular cancer: a systematic review. PLoS One. 2009;4(5):e5591. doi: 10.1371/journal.pone.0005591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Waksberg JS. Sampling methods for random digit dialing. J Amer Statistical Assoc. 1978;73(361):40–46. [Google Scholar]
- 24.Robison LL, Daigle A. Control selection using random digit dialing for cases of childhood cancer. Am J Epidemiol. 1984;120(1):164–166. doi: 10.1093/oxfordjournals.aje.a113866. [DOI] [PubMed] [Google Scholar]
- 25.Formann AK, Kohlmann T. Latent class analysis in medical research. Stat Methods Med Res. 1996;5(2):179–211. doi: 10.1177/096228029600500205. [DOI] [PubMed] [Google Scholar]
- 26.Muthen L, Muthen B. Mplus user's guide. Los Angeles: Muthen & Muthen; 1998–2004. [Google Scholar]
- 27.Definitions of infertility and recurrent pregnancy loss. Fertil Steril. 2008;90(5 Suppl):S60. doi: 10.1016/j.fertnstert.2008.08.065. [DOI] [PubMed] [Google Scholar]
- 28.Puscheck EE, Jeyendran RS. The impact of male factor on recurrent pregnancy loss. Curr Opin Obstet Gynecol. 2007;19(3):222–228. doi: 10.1097/GCO.0b013e32813e3ff0. [DOI] [PubMed] [Google Scholar]
- 29.Ford HB, Schust DJ. Recurrent pregnancy loss: etiology, diagnosis, and therapy. Rev Obstet Gynecol. 2009;2(2):76–83. [PMC free article] [PubMed] [Google Scholar]
- 30.Weiss NS. Subgroup-specific associations in the face of overall null results: should we rush in or fear to tread? Cancer Epidemiol Biomarkers Prev. 2008;17(6):1297–1299. doi: 10.1158/1055-9965.EPI-08-0144. [DOI] [PubMed] [Google Scholar]
- 31.Johnson KJ, Carozza SE, Chow EJ, Fox EE, Horel S, McLaughlin CC, et al. Parental age and risk of childhood cancer: a pooled analysis. Epidemiology. 2009;20(4):475–483. doi: 10.1097/EDE.0b013e3181a5a332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Cook MB, Akre O, Forman D, Madigan MP, Richiardi L, McGlynn KA. A systematic review and meta-analysis of perinatal variables in relation to the risk of testicular cancer--experiences of the mother. Int J Epidemiol. 2009 doi: 10.1093/ije/dyp287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Greenland S. Basic methods for sensitivity analysis of biases. Int J Epidemiol. 1996;25(6):1107–1116. [PubMed] [Google Scholar]
- 34.Ross JA, Severson RK, Pollock BH, Robison LL. Childhood cancer in the United States. A geographical analysis of cases from the Pediatric Cooperative Clinical Trials groups. Cancer. 1996;77(1):201–207. doi: 10.1002/(SICI)1097-0142(19960101)77:1<201::AID-CNCR32>3.0.CO;2-7. [DOI] [PubMed] [Google Scholar]
- 35.Werler MM, Pober BR, Nelson K, Holmes LB. Reporting accuracy among mothers of malformed and nonmalformed infants. Am J Epidemiol. 1989;129(2):415–421. doi: 10.1093/oxfordjournals.aje.a115145. [DOI] [PubMed] [Google Scholar]
- 36.Drews CD, Kraus JF, Greenland S. Recall bias in a case-control study of sudden infant death syndrome. Int J Epidemiol. 1990;19(2):405–411. doi: 10.1093/ije/19.2.405. [DOI] [PubMed] [Google Scholar]