American Journal of Epidemiology Advance Access originally published online on June 14, 2006
American Journal of Epidemiology 2006 164(3):222-231; doi:10.1093/aje/kwj174
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Original Contribution |
Parental Heat Exposure and Risk of Childhood Brain Tumor: A Children's Oncology Group Study
1 Division of Oncology, Children's Hospital of Philadelphia, Philadelphia, PA
2 Department of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA
3 Department of Epidemiology and Cancer Control, St. Jude Children's Research Hospital, Memphis, TN
4 Division of Human Genetics, Children's Hospital of Philadelphia, Philadelphia, PA
5 Department of Neurosurgery, Children's Hospital of Pittsburgh, University of Pittsburgh School of Medicine, Pittsburgh, PA
6 Departments of Pathology and Neurology, Children's Hospital of Philadelphia, Philadelphia, PA
Correspondence to Dr. Greta R. Bunin, Division of Oncology, Room 1472, Children's Hospital of Philadelphia, 3535 Market Street, Philadelphia, PA 19104 (e-mail: bunin{at}email.chop.edu).
Received for publication November 21, 2005. Accepted for publication February 8, 2006.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Medulloblastoma (MB) and primitive neuroectodermal tumor (PNET) are histologically similar brain tumors that occur mostly in children. As part of a comprehensive case-control study of MB/PNET, this study explored parental exposure to heat and electromagnetic fields as potential risk factors. Parents of 318 cases (<6 years of age at diagnosis in 1991–1997 and registered with the Children's Cancer Group) and 318 controls selected by random digit dialing were interviewed. In univariate analyses, moderately strong associations were observed for mother's sauna use close to conception (odds ratio = 3.8, 95% confidence interval (CI): 1.0, 13.7) or in the first trimester (odds ratio = 3.6, 95% CI: 0.7, 17.3) and for father's exposure in the 3 months before the pregnancy to sauna (odds ratio = 2.4, 95% CI: 1.3, 4.5), electric blanket (odds ratio = 2.0, 95% CI: 0.9, 4.3), or any heat source (for higher exposure: odds ratio = 2.5, 95% CI: 1.4, 4.6). In multivariate models, father's sauna use and father's exposure to any heat source were associated with MB/PNET in a dose-response fashion (for high exposure: odds ratio = 3.4, 95% CI: 1.2, 9.7, and odds ratio = 2.1, 95% CI: 1.1, 4.3, respectively). This new observation regarding paternal exposure to heat just prior to the index pregnancy deserves consideration in future animal and human studies of MB/PNET.
brain neoplasms; case-control studies; child, preschool; heat; infant; maternal exposure; medulloblastoma; paternal exposure
Abbreviations: CI, confidence interval; MB, medulloblastoma; PNET, primitive neuroectodermal tumor
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Medulloblastoma (MB) and primitive neuroectodermal tumor (PNET) are histologically similar tumors that occur in various parts of the brain (1
). Medulloblastoma is a cerebellar tumor that accounts for about 80 percent of the combined group of MB/PNET (2). It is the second most common brain tumor in children, accounting for about 20 percent of cases (3), and is rare in adults. Few epidemiologic studies have investigated MB/PNET specifically.
Only ionizing radiation and some genetic conditions are known to increase the risk of brain tumors in childhood (3). For a few other exposures, the data cannot be considered conclusive but show some consistency. Mothers of children with brain tumors report more frequent eating of cured meats during pregnancy compared with mothers of controls (4–6), although confounding by other aspects of diet has not been ruled out as an explanation of the findings. The association with cured meats seems to be specific for astrocytic tumors and has generally not been observed for MB/PNET (7, 8). Living on a farm and exposure to farm animals have been associated with childhood brain tumors, usually with MB/PNET (9–11). The epidemiologic data on these two risk factors, as well as the incidence patterns by age and sex, suggest that MB/PNET may differ in etiology from the more common astrocytic tumors.
In animals, elevated temperatures during pregnancy can result in embryonic death, spontaneous abortion, growth retardation, and congenital anomalies (12). The developing central nervous system is especially susceptible to excessive heat, which has been linked to neural tube defects in both animals and humans (12). To our knowledge, no animal studies of the effect of hyperthermia during pregnancy on the risk of cancer in offspring have been reported. As some agents are both teratogenic and carcinogenic (13–15), we included maternal heat exposure in our study of MB/PNET.
If paternal exposure to heat played an etiologic role in childhood cancer, an effect on sperm would be most likely. Sperm production requires temperatures 3–4 degrees lower than normal body temperature (16). Heat is known to decrease sperm count and quality (17). In animals, the effect of paternal heat exposure extends to the embryo. The mating of heat-stressed males to normal females results in an increase in nondeveloping and abnormal embryos and in a decrease in embryo survival (18, 19).
Electromagnetic fields have been studied in relation to childhood brain tumors with mostly negative results for exposure from power lines (20) and other sources, such as electric blankets and electrically heated water beds (21–23). The only study to report results for MB/PNET observed no association with either source of electromagnetic field (22).
Because relevant data are limited, we included selected sources of parental exposure to heat and electromagnetic fields as exploratory questions in a case-control study of MB/PNET of the Children's Oncology Group and report the results here.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Study population
The study (Children's Cancer Group protocol E21) aimed to investigate maternal diet during pregnancy in the etiology of MB/PNET, with examination of other hypothesized risk factors as secondary goals. Eligible patients were those diagnosed between July 1991 and November 1997 with a MB or PNET of the brain before the age of 6 years and registered with the member institutions of the Children's Cancer Group, one of the pediatric oncology cooperative groups that later merged to form the Children's Oncology Group. At the time of the study, the Children's Cancer Group consisted of over 100 medical centers that were estimated to diagnose and/or treat about half of all cancers in US and Canadian children under the age of 15 years (24
). The institutional review boards of all Children's Cancer Group institutions that enrolled patients approved the study. All participants provided consent for the telephone interview.
Participating institutions registered 558 potentially eligible cases with the Children's Cancer Group. Cases were excluded for the following reasons: physician did not give consent to contact family (n = 35), not PNET on pathology review (n = 28), biologic mother not available (n = 17), language barrier (n = 16), no phone in household (n = 13), residence outside North America (n = 6), other cancer (n = 3), and no response from institution (n = 2). The institutional review boards at many participating institutions required that the parents give their consent before they were invited to participate by the principal investigator. The institution could not locate the families of seven cases to obtain their consent, and parents of 39 cases either did not respond to the letter sent by the institution or actively refused through the institution. When we attempted to contact the families of cases, 41 could not be located and 27 refused. We were unable to find controls for six cases. A total of 318 case mothers (73 percent of eligible cases) for whom controls were interviewed comprise the final sample.
We estimated misclassification of diagnosis by conducting pathology review of the tumors for which slides and/or blocks were available. This pathology review was done independently of determination of eligibility on other criteria and contact of the parents. On the basis of this review, which is reported in more detail elsewhere (8), we estimate that 17 (11 percent) of the 154 nonreviewed tumors were not MB/PNET; the diagnosis of the other 164 cases was confirmed by the reviewing pathologist (L. B. R-A.). The estimated 17 non-MB/PNET tumors comprise 5 percent of the final sample of 318 and, thus, we estimate that 5 percent were not MB/PNET.
Control children were selected by random digit dialing. We matched on area code, as a convenient way of matching on geographic area, a possible confounder for many exposures including use of the selected electromagnetic fields and heat sources. We matched on race (non-Hispanic White, non-Hispanic Black, other), as exposures and incidence may vary by race. We also matched on date of birth (within 6 months for cases with age at diagnosis of <1 year, within 1 year for all others), so that parents of both cases and controls would be recalling their exposures over a similar time interval. The results from random digit dialing have been described previously (8). The response rates for the random digit dialing screening call and the study questionnaire were 67 percent and 73 percent, respectively.
We interviewed the mothers of 318 cases and 318 controls. Of these, we obtained information on the fathers of 272 cases (86 percent of the 318 cases) and 258 controls (81 percent of the controls). The mother completed a proxy interview for 28 (10 percent) of the fathers of cases and 63 (24 percent) of the fathers of controls.
Trained interviewers conducted structured telephone interviews with mothers and fathers of the cases and controls. The mother was asked about the timing and frequency of exposure to several sources of excessive heat (sauna, hot tubs, fever), a source of electromagnetic fields (electrically heated water beds), and a source of both heat and electromagnetic fields (electric blankets) in the month before and during pregnancy. If recall bias occurred for the personal habits in this section of the questionnaire, we might expect to see associations with all habits including those that do not expose the individual to excessive heat. To help assess the likelihood of recall bias, we asked about hot baths, which are unlikely to expose the bather to excessive heat. Fathers were asked about sauna, hot tub, and electric blanket use, as well as hot baths and water beds, in the 3 months before the pregnancy began. Fathers were not asked about fever; the fever questions were in a section of the mother's questionnaire on selected illnesses and medications during pregnancy that was not included in the father's questionnaire. The 3 months prior to the pregnancy were chosen because spermatogenesis takes 10 weeks (16). For sauna, hot tub, hot baths, and electric blanket use, parents indicating that they had the exposure were asked about its frequency (most days, 1–3 times per week, 1–3 times per month, <1 time per month, once or twice only). Electric blanket users were asked about their pattern of use, that is, whether or not they turned it off before getting into bed, and if they left it on, whether it was on for less than an hour, 1–3 hours, or throughout the night. Mothers were asked in which part of pregnancy the use occurred (month before or first, second, third trimester). Mothers were asked the duration in days of fevers.
Interviews of cases took place a median of 1.6 years after the child's diagnosis (range: 0.1–6.2 years). For cases and controls, the mean length was 58 (range: 25–135) minutes for the mother's interview and 40 (range: 10–120) minutes for the father's interview.
Statistical analysis
Demographic characteristics and possible confounders were compared by chi-squared tests for categorical variables and t tests for continuous variables.
Logistic regression was used to calculate odds ratios for MB/PNET and sources of heat and electromagnetic fields. Analyses for mother's exposure and father's exposure for each source were performed as were analyses of heat from any of the sources and electromagnetic fields from electric blankets or water beds. Analyses of maternal exposure were also done for each time period of pregnancy. Fathers were considered exposed to excessive heat if they used a sauna, hot tub, or an electric blanket. Mothers were considered exposed to excessive heat if they used a sauna, hot tub, or electric blanket, or if they reported having a fever over 100°F (37.78°C). Electromagnetic field exposure was defined as using an electric blanket that was on when the parent got into bed or using an electrically heated water bed. Analyses were also performed by level of exposure using indicator variables. For sauna, hot tub, and hot baths, use three times per month or less was considered low exposure, and weekly use was considered high exposure. Electric blanket use was considered high if the heat was on when the parent got into bed. Mothers were considered to have high exposure to fever if the fever lasted more than 1 day. A parent was considered to have high heat exposure from all sources combined if he/she had high use from at least one source.
Levels of exposure to electric water beds were not defined; information on the frequency and pattern of use was not collected as these beds must be kept heated if they are slept on. Associations with levels of exposure to electromagnetic fields were not evaluated, as few parents had "low" electric blanket exposure, and the level of exposure could not be defined for electric water beds. Tests for trend were performed by using the level of exposure, that is, 0, 1, or 2, as a continuous variable in logistic regression.
Unconditional logistic regression was used to compute odds ratios adjusted for possible confounders. Because the study design was matched, the matching factors (date of birth, race, and area code) were considered possible confounders. We matched on area code to account for broad geographic differences in exposures; for the analysis, we divided the US states and Canadian provinces into 14 geographic regions. The father's and mother's demographic and interview characteristics were highly correlated; we chose to use the father's values. Controls were consistently older than cases and interviewed later because an individually matched control could not be sought until the case was interviewed. The inclusion of date of child's birth (matching factor), date of father's interview, and age of child at interview in the same model created perfect multicollinearity; that is, any two of the variables determined the third. We chose to include the date of interview and the child's age at interview in the models. Other possible covariates were the father's interview by proxy, exposure of the other parent to the same heat or electromagnetic field source, income level, father's educational level, paternal smoking, residential mobility (number of moves), household size, father's age at the child's birth, and father's marital status. The final models included the father's interview by proxy, the other parent's use of same heat or electromagnetic field source, father's race, date of interview, child's age at interview, income level, and smoking status (yes/no), because these altered one or more coefficients of the heat or electromagnetic field factor by at least 15 percent, differed significantly or nearly so between cases and controls, or were matching factors. For analyses of the father's level of exposure to individual sources except for hot baths, the covariate used for the mother's exposure was any use of the source because of the small number of exposed mothers. Geographic area did not change the results appreciably and was excluded. Although the results of unconditional logistic regression are presented here, the results were not substantially different when we performed conditional logistic regression on matched pairs (a total of 250 matched pairs with data on mothers and fathers).
Most analyses were performed using SPSS statistical software (25). SAS software was used for conditional logistic regression (26).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Demographic characteristics and potential confounders
Parents of cases and controls were similar in race (a matching factor), education, residential mobility, and age at child's birth (table 1). There was a modest trend for families of cases to have higher household incomes compared with the families of controls, based on either parent's report of income level (table 1). As a result of the study methods, which specified that mothers of cases be interviewed before a control was sought, controls were older at interview than were cases, and interviews of controls occurred later in the study (table 1). Mothers of cases were less likely to have smoked compared with mothers of controls. More fathers of controls than cases were interviewed by proxy.
|
Maternal exposure to sources of heat and electromagnetic fields
All but one of the risk estimates for sauna, hot tub, and electric blanket were greater than 1.0 for any use, use in the month before the pregnancy, and use in the first trimester (table 2). Only sauna use was associated with a substantially elevated risk, an odds ratio of greater than 3.5 that reached or approached statistical significance. Fever and electric water bed use were not associated with MB/PNET, and hot baths were inversely associated.
|
When heat from any of the sources (sauna, hot tub, electric blanket, fever) was considered, modestly elevated odds ratios were observed for the month before and first trimester. Electromagnetic field exposure from water beds and/or electric blankets showed no association with MB/PNET.
The numbers of mothers who reported use of a sauna, hot tub, or electric blanket, or who reported a fever, were small for exposure at any time during the pregnancy and particularly small for specific time periods, limiting analyses by frequency and timing. Only for hot baths was the number of exposed mothers sufficient to permit analysis by frequency of exposure. The odds ratios for less and more frequent baths were 0.8 (95 percent confidence interval (CI): 0.3, 1.8; p = 0.57) and 0.7 (95 percent CI: 0.5, 1.0; p = 0.03), respectively, with a significant test for trend (p = 0.02) (data not shown).
The results were not appreciably different when we restricted the analysis to the wives/partners of the fathers for whom we had interview data rather than including all interviewed mothers.
Paternal exposure to sources of excessive heat and electromagnetic fields
We observed associations between MB/PNET and each of the three sources of excessive heat studied (table 3). The odds ratios were 1.5 for hot tubs, 2.0 for electric blankets, and 2.4 for sauna, with confidence intervals that excluded or nearly excluded 1.0. The analyses by frequency of exposure to sauna resulted in similar risk estimates for low and high exposure, although the test for trend was significant (p = 0.009). No dose response was seen for hot tub use. For electric blankets, there was insufficient variability to assess level of use. The odds ratio for water bed use was not elevated. For hot baths, the odds ratio for less frequent, but not more frequent, use was significantly elevated, that is, was not indicative of increasing risk with greater exposure.
|
For heat exposure from any of the sources, we observed a significantly elevated odds ratio of 1.9 (95 percent CI: 1.3, 2.9; p = 0.002) and a dose-response relation (ptrend = 0.001), with an odds ratio of 2.5 (95 percent CI: 1.4, 4.6; p = 0.003) for the higher level of exposure. To further explore the association with paternal heat exposure, we categorized exposure into four levels based on the number of times any of the heat sources was reported used in the 3 months prior to the pregnancy. The odds ratios for the three categories above the referent category (no exposure) were 1.8 (95 percent CI: 1.1, 3.0; p = 0.03), 2.1 (95 percent CI: 1.0, 4.5; p = 0.05), and 2.0 (95 percent CI: 0.9, 4.6; p = 0.11) for fewer than 10 times, 10–49 times, and 50 or more times, respectively. For exposure to either source of electromagnetic fields (water beds and/or electric blankets), no association was observed.
Analyses adjusted for possible confounders
As maternal and paternal exposure to the lifestyle factors examined may be correlated and as demographic and other factors may confound the observed associations, we performed multivariate analyses to adjust for use by the other parent and for other possible confounders (tables 4 and 5).
|
|
In the adjusted analysis, the risk estimate for maternal sauna use was 2.0 with a large confidence interval. For paternal sauna use, a dose-response relation was observed; the odds ratios for less and more frequent use were 1.4 (95 percent CI: 0.6, 3.5; p = 0.48) and 3.4 (95 percent CI: 1.2, 9.7; p = 0.02), respectively (ptrend = 0.02).
For hot baths of the father, the pattern seen in the crude analysis of an elevated odds ratio for less frequent but not more frequent baths remained. Maternal hot baths were inversely associated with MB/PNET (ptrend = 0.01). No noteworthy associations were observed for maternal or paternal use of hot tubs, electric blankets, or waterbeds or for maternal fever.
In the adjusted analysis, the relations of paternal and maternal exposure to heat from any of the sources with MB/PNET were attenuated compared with the univariate results. However, the association with paternal exposure remained, with the effect mostly seen at the higher level of exposure (for higher exposure: odds ratio = 2.1, 95 percent CI: 1.1, 4.3; p = 0.03, ptrend = 0.001). The small numbers of fathers with higher exposure precluded detailed exploration in multivariate models with more levels of exposure.
Use of either source of electromagnetic field (water beds and/or electric blankets) by the mother or father was not associated with disease.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
We observed associations between paternal heat exposure in the 3 months before the pregnancy and MB/PNET. The associations with sauna use and heat exposure from any of the sources were the most robust and showed some indication of a dose-response relation. In contrast, the association with maternal sauna use was much attenuated and imprecise after adjustment for possible confounders, although the magnitude of the adjusted odds ratio was similar to that for paternal sauna use, both being about 2.0. No association with risk was seen for the electromagnetic field from electric blankets or water beds.
If paternal heat exposure were, in fact, causally related to MB/PNET, it would presumably act through a genetic or epigenetic effect on sperm. For example, heat exposure might result in mutation in genes predisposing to MB/PNET, which would be passed to the child as a new germline mutation. Germline mutations in several genes, patched-2 (PTCH2), adenomatous polyposis of the colon (APC), ataxia telangiectasia mutated (ATM), and human suppressor of fused homolog (SUFU), are associated with MB or PNET (27–29). In at least some instances, these mutations are present in the child but not in the parents; that is, they occur as new germline mutations. Thus, a role for new germline mutation in the etiology of MB/PNET is established but is thought to explain only a very small proportion of cases (3). Whether heat exposure could result in new germline mutation is not known. Heat can cause chromosomal aberrations in vitro and can inhibit repair of double strand breaks in DNA (30). However, whether these experiments exposing somatic cells to severe hyperthermia have any relevance to sperm and their precursors exposed to milder conditions is not known. Another theoretical mechanism is that heat induces an epigenetic effect, such as a change in imprinting or methylation, which increases the risk of MB/PNET. Imprinting errors are not known to be associated with MB/PNET, but altered methylation patterns of several genes—HIC1, CASP8, and RASSF1A—have been observed in MB (31). It is possible that altered methylation early in embryogenesis could result from damage to sperm and predispose the child to MB/PNET. The animal data provide evidence that heat exposure of the male can damage the survival and development of embryos, but germline mutation has not been studied, to our knowledge. In summary, the idea that paternal heat exposure before a child's conception could increase the risk of MB/PNET must be considered speculative, as only limited animal and laboratory evidence supports it.
We observed no association between MB/PNET and maternal or paternal exposure to electromagnetic fields from electric blankets and electrically heated water beds. Although two early studies observed odds ratios of about 2.0 associated with higher electromagnetic field exposure from the electric power distribution system of residences, larger studies and those using more direct measures of exposure did not replicate this finding (refer to the review by Gurney and van Wijngaarden (20)). Magnetic fields from electric appliances are stronger than those from residential power lines. Electric blankets and water beds can substantially contribute to an individual's exposure as, unlike many other appliances, they are used close to the body and for long periods of time. Of the three relevant studies, two showed no association (22, 23). In the third study, a significantly increased risk was observed, but the sample size was small, and the case group was heterogeneous in terms of histology (21). The one study to report results for MB/PNET observed no increase in risk (22). Our results add to the overall negative findings related to electromagnetic fields and childhood brain cancer, as well as MB/PNET specifically.
As this is the first study to examine exposure to these sources of heat, other studies need to replicate these findings to better assess the possibility of chance, confounding, and bias as explanations. The less than optimal participation rates, especially for controls, might have resulted in selection bias, if participants and nonparticipants differed in their exposure to the sources of heat studied. One of the limitations of random digit dialing as a source of controls is the lack of data on nonresponders. We accounted for possible selection bias based on socioeconomic status by adjusting for education and household income. However, we cannot rule out the possibility of selection bias on other characteristics. Another concern is inaccurate recall of exposures on average 5 years in the past. If misclassification due to inaccurate recall is nondifferential, the odds ratios will generally be closer to 1.0 (32). If parents of cases and controls differed in their ability to recall these exposures, the results could be biased in either direction. We know of no data on the recall accuracy of the relevant exposures during a pregnancy several years in the past.
The small numbers of exposed mothers for most of the heat sources limited analyses by time period of pregnancy and level of exposure. Another limitation is our definition of fever as 100°F compared with the definition of 102°F (38.89°C) used in some studies of neural tube defects and other congenital anomalies (33). If only high fevers increased risk, including lower degrees of fever, that is, an irrelevant exposure, along with the exposure of interest, would dilute the observed effect and result in odds ratios biased toward the null. It is likely that most of the fevers reported in this study were below 102°F, as lower fevers are more common in adults than higher fevers. Our results, then, do not answer the question about high fevers. However, our results for maternal heat exposure from any source did not change appreciably when we excluded fever as one of the heat sources. In other words, although our definition of fever might have obscured an observed association, including fever as a source of heat did not obscure an association with the other heat sources.
In summary, we report an association between paternal heat exposure during the 3 months prior to the index child's conception and MB/PNET development in the child. This novel observation is worth investigating further, as so little is known about the etiology of childhood brain tumors.
|
ACKNOWLEDGMENTS |
|---|
This work was supported in part by grant CA60951 from the National Cancer Institute.
The authors thank Kathy Walsh, Anne Goldblatt, the late Jean Rodwell, Mary Rewinski, Christine Plourde, and Sallie McLaughlin for their hard work and dedication during the data collection phase of the study.
Conflict of interest: none declared.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
- Rorke LB, Gilles FH, Davis RL, et al. Revision of the World Health Organization classifications of brain tumors for childhood brain tumors. Cancer 1985;56:1869–986.[CrossRef][Web of Science][Medline]
- Surveillance, Epidemiology, and End Results (SEER) Program public-use data (1973–2001), released April 2004, based on the November 2003 submission. Bethesda, MD: Cancer Statistics Branch, Surveillance Research Program, Division of Cancer Control and Population Sciences, National Cancer Institute, 2004. (www.seer.cancer.gov).
- Gurney JG, Smith MA, Bunin GR. CNS and miscellaneous intracranial and intraspinal neoplasms. In: Ries LAG, Smith MA, Gurney JG, et al, eds. Cancer incidence and survival among children and adolescents: United States SEER Program 1975–1995. Bethesda, MD: National Cancer Institute, 1999:51–64. (NIH publication no. 99-4649).
- Preston-Martin S, Yu MC, Benton B, et al. N-Nitroso compounds and childhood brain tumors: a case-control study. Cancer Res 1982;42:5240–5.
[Abstract/Free Full Text] - Preston-Martin S, Pogoda JM, Mueller BA, et al. Maternal consumption of cured meats and vitamins in relation to pediatric brain tumors. Cancer Epidemiol Biomarkers Prev 1996;5:599–605.[Abstract]
- Bunin GR, Kuijten RR, Boesel CP, et al. Maternal diet and risk of astrocytic glioma in children: a report from the Children's Cancer Group. Cancer Causes Control 1994;5:177–87.[CrossRef][Web of Science][Medline]
- Bunin GR, Kuijten RR, Buckley JD, et al. Relation between maternal diet and subsequent primitive neuroectodermal brain tumors in young children. N Engl J Med 1993;329:536–41.
[Abstract/Free Full Text] - Bunin GR, Kushi LH, Gallagher PR, et al. Maternal diet during pregnancy and its association with medulloblastoma in children: a Children's Oncology Group study. Cancer Causes Control 2005;16:877–91.[CrossRef][Web of Science][Medline]
- Efird JT, Holly EA, Preston-Martin S, et al. Farm-related exposures and childhood brain tumours in seven countries: results from the SEARCH International Brain Tumour Study. Paediatr Perinat Epidemiol 2003;17:201–11.[CrossRef][Web of Science][Medline]
- Kristensen P, Andersen A, Irgens LM, et al. Cancer in offspring of parents engaged in agricultural activities in Norway: incidence and risk factors in the farm environment. Int J Cancer 1996;65:39–50.[CrossRef][Web of Science][Medline]
- Bunin GR, Buckley JD, Boesel CP, et al. Risk factors for astrocytic glioma and primitive neuroectodermal tumor of brain in young children: a report from the Children's Cancer Group. Cancer Epidemiol Biomarkers Prev 1994;3:197–204.[Abstract]
- Edwards MJ, Saunders RD, Shiota K. Effects of heat on embryos and foetuses. Int J Hyperthermia 2003;19:295–324.[CrossRef][Web of Science][Medline]
- Mandal PK. Dioxin: a review of its environmental effects and its aryl hydrocarbon receptor biology. J Comp Physiol [B] 2005;175:221–30.[CrossRef][Medline]
- Prater MR, Zimmerman KL, Ward DL, et al. Reduced birth defects caused by maternal immune stimulation in methylnitrosourea-exposed mice: association with placental improvement. Birth Defects Res A Clin Mol Teratol 2004;70:862–9.[CrossRef][Web of Science][Medline]
- Lijinsky W. Chemistry and biology of N-nitroso compounds. New York, NY: Cambridge University Press, 1992.
- Braunstein GD. Testes. In: Greenspan FS, Gardner DG, eds. Basic & clinical endocrinology. 7th ed. New York, NY: Lange Medical Books/McGraw-Hill, 2004:478–510.
- Procope B. Effect of repeated increase of body temperature on human sperm cells. Int J Androl 1965;10:333–9.
- Zhu B, Walker SK, Oakey H, et al. Effect of paternal heat stress on the development in vitro of preimplantation embryos in the mouse. Andrologia 2004;36:384–94.[CrossRef][Web of Science][Medline]
- Mieusset R, Quintana Casares P, Sanchez Partida LG, et al. Effects of heating the testes and epididymides of rams by scrotal insulation on fertility and embryonic mortality in ewes inseminated with frozen semen. J Reprod Fertil 1992;94:337–43.
[Abstract/Free Full Text] - Gurney JG, van Wijngaarden E. Extremely low frequency electromagnetic fields (EMF) and brain cancer in adults and children: review and comment. Neuro-oncol 1999;1:212–20.[Abstract]
- Savitz DA, John EM, Kleckner RC. Magnetic field exposure from electric appliances and childhood cancer. Am J Epidemiol 1990;131:763–73.
[Abstract/Free Full Text] - Preston-Martin S, Gurney JG, Pogoda JM, et al. Brain tumor risk in children in relation to use of electric blankets and water bed heaters. Results from the United States West Coast Childhood Brain Tumor Study. Am J Epidemiol 1996;143:1116–22.
[Abstract/Free Full Text] - Gurney JG, Mueller BA, Davis S, et al. Childhood brain tumor occurrence in relation to residential power line configurations, electric heating sources, and electric appliance use. Am J Epidemiol 1996;143:120–8.
[Abstract/Free Full Text] - Ross JA, Severson RK, Robison LL, et al. Pediatric cancer in the United States: a preliminary report of a collaborative study of the Childrens Cancer Group and the Pediatric Oncology Group. Cancer 1993;71(suppl):3415–21.[CrossRef][Web of Science][Medline]
- SPSS base 10.0 for Windows user's guide. Chicago, IL: SPSS, Inc, 1999.
- SAS/STAT user's guide, version 8. Cary, NC: SAS Institute, Inc, 1999.
- Online Mendelian Inheritance in Man (OMIM). Center for Medical Genetics, Johns Hopkins University (Baltimore, MD) and National Center for Biotechnology Information, National Library of Medicine (Bethesda, MD), 1999. (http://www.ncbi.nlm.nih.gov/omim).
- Taylor MD, Liu L, Raffel C, et al. Mutations in SUFU predispose to medulloblastoma. Nat Genet 2002;31:306–10.[CrossRef][Web of Science][Medline]
- Sevenet N, Sheridan E, Amram D, et al. Constitutional mutations of the hSNF5/INI1 gene predispose to a variety of cancers. Am J Hum Genet 1999;65:1342–8.[CrossRef][Web of Science][Medline]
- Dewhirst MW, Lora-Michiels M, Viglianti BL, et al. Carcinogenic effects of hyperthermia. Int J Hyperthermia 2003;19:236–51.[CrossRef][Web of Science][Medline]
- Lindsey JC, Lusher ME, Anderton JA, et al. Identification of tumour-specific epigenetic events in medulloblastoma development by hypermethylation profiling. Carcinogenesis 2004;25:661–8.
[Abstract/Free Full Text] - Barron BA. The effects of misclassification on the estimation of relative risk. Biometrics 1977;33:414–18.[CrossRef][Web of Science][Medline]
- Miller P, Smith DW, Shephard TH. Maternal hyperthermia as a possible cause of anencephaly. Lancet 1978;1:519–21.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||