This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pang, D. Articles by Birch, J. M. Search for Related Content PubMed PubMed Citation Articles by Pang, D. Articles by Birch, J. M.

Cancer Incidence and Mortality among the Parents of a Population-based Series of 2604 Children with Cancer¹

University of Manchester, Cancer Research UK, Paediatric & Familial Cancer Research Group, Royal Manchester Children’s Hospital, Manchester M27 4HA [D. P., R. M., J. M. B.], and Central Manchester and Manchester Children’s University Hospitals National Health Service Trust, Department of Histopathology, Royal Manchester Children’s Hospital, Manchester M27 4HA [A. K.], United Kingdom

	Abstract

Top Abstract Introduction Materials and Methods Results Discussion References

 
The purpose is to determine cancer incidence and mortality from all causes among parents of children with solid tumors and to formulate hypotheses about heritable risks and associations with parental environmental exposures. A total of 2604 children ages < 15 years included in the Manchester Children’s Tumor Registry, 1954–1996, was eligible. Parents of index children were traced and followed up to December 31, 2000, through the United Kingdom National Health Service Central Register. Data on cancer diagnosis and all causes of deaths were obtained. Standardized incidence (SIR) and standardized mortality (SMR) ratios, Ps, and 95% confidence intervals were calculated from serial age and sex-specific cancer incidence and mortality data for England and Wales. There was a significant excess of cancers in parents overall (observed number = 481, expected number = 428.9, SIR = 1.1; P < 0.05) because of central nervous system tumors (SIR = 1.7; P < 0.05), carcinoma breast (SIR = 1.3; P < 0.05), bone and soft tissue sarcoma (SIR = 2.9; P < 0.01), and retinoblastoma (SIR = 62.1; P < 0.001). Diagnoses in the index children associated with excess risk of cancer in their parents were retinoblastoma (SIR = 1.8; P < 0.01) and gonadal germ cell tumors (SIR = 2.3; P < 0.01). There was no excess risk of death from all causes considered together (observed number = 791, expected number = 814.2) nor from any specific cause. These results indicate that there is a small excess of incident cancers among parents of children with solid tumors, probably mainly because of known genes, but mortality is similar to the general population. The study is very reassuring for families of children with cancer.

	Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

 
Genetic epidemiological studies show that certain cancers can manifest a tendency to aggregate in families, and it is known that mutations in certain cancer-associated genes predispose to the development of both specific childhood and adult onset cancers (1) . Such an inherited susceptibility to tumors has been well documented for bilateral retinoblastoma and for Li-Fraumeni syndrome, which are linked to germ-line mutations in the tumor suppressor genes Rb1 and TP53, respectively (2 , 3) . The possibility that other unknown genes may also be involved exists. One way to assess this is to study cancer incidence in parents of children with cancer. A study covering the period 1943–1985 linked birth registry data to the Danish Cancer Registry (4) . This study found no overall excess of cancers in parents. To date, this is the only comprehensive study of cancer in parents of children with cancer, but although population based with 4000 children included, there were some weaknesses. Notably, the cancer incidence data on the parents were coded in ICD³ 7 (5) , and a broad, nonspecific classification into 14 site-based cancer groups was used. The childhood cancer data were based on original registration documents dating back to 1943, and many diagnoses would therefore have been inconsistent with current classifications. Given these limitations, specific associations between childhood and adult cancers may not have been recognized.

We have carried out a study of cancer incidence and mortality in the parents of children with solid tumors by linking the parents of children with cancer included in a specialized, comprehensive pediatric tumor registry with the United Kingdom national central registration system, which provides continuous notification of cancer registrations and causes of death, including noncancer causes.

We formulated three prior hypotheses. First, there may be excesses of specific cancers in parents of children with certain tumors because of mutations in the TP53, NF1, NF2, patched, and Rb1 genes. Second, there may be excesses of carcinoma of the lung, leukemia, and non-Hodgkin’s lymphoma in the parents as a result of exposure to cigarette smoking, hydrocarbons, and agrochemicals. Third, the pattern of mortality from noncancer causes among the parents of children with cancer may provide indications of lifestyle factors that could increase or decrease the risk of childhood cancer. To test these hypotheses, we have used comprehensive population-based cancer incidence and mortality data to examine patterns of cancer, mortality from all causes, and specific causes of death in parents of children with solid tumors and compared these with the expected patterns in the general population.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

 
Study Subjects.
Solid tumors, including lymphomas, were selected for study because previous analyses suggest there may be a higher degree of familial clustering among these diagnostic groups than among leukemias (1 , 4) . The parents of 2604 children with histologically confirmed solid tumors diagnosed younger than 15 years of age, who were included in the MCTR between January 1, 1954, and December 31, 1996, were eligible.

The MCTR is described in detail elsewhere (6) . A summary is provided below. The MCTR is population based and includes all cases of malignant disease in children resident in a geographically defined area of Northwest England at the time of diagnosis of their cancer. The MCTR has near complete ascertainment of cases, and reporting practice has not changed over time. A computerized record-keeping system maintains detailed clinical and pathological records on each case. In addition, histopathological slides (stained and unstained) were available on virtually all cases. These were periodically re-reviewed by a panel of pathologists and diagnoses updated to take account of developments in histopathological classification. Frozen blood and tumor samples are also stored where possible to facilitate detailed studies in line with advances in medical technology, in particular, genetic techniques.

Tumors in all case children were individually coded according to the ICD-O (7) on the basis of diagnostic slides and clinical records. This procedure avoids possible misclassification of cases that might arise as a result of applying conversion codes from earlier coding systems to ICD-O. All cases were categorized on the basis of morphology and topography codes into broad groupings as follows: HD; NHL; astrocytoma; ependymoma; PNETs; other CNS tumors; Rb; neuroblastoma; peripheral PNETs; WT; HB; osteosarcoma and chondrosarcoma; RMS; other bone and soft tissue sarcoma; gonadal GCTs; nongonadal GCTs; ACC; malignant neoplasms of skin; carcinoma (excluding of skin); other rare specified tumors; and unspecified neoplasms. The morphology and topography code allocations defining these groups and algorithms for selecting the groups are available online.⁴

After approval from the appropriate research ethics committees, parents of the index children were traced and followed-up to December 31, 2000, through the United Kingdom NHSCR. Each traced individual was then flagged at NHSCR to provide continuous notification of details of cancer registrations and death notifications. Information required for successful flagging includes full names (including former names), dates of birth, current and previous residential addresses, and National Health Service number. It is sometimes possible to trace and flag with less than the optimum information. Information on parents was extracted from the MCTR records, other medical records, interview data, and birth records. A database was set up for sending data to NHSCR and for recording the quarterly flagging returns, including dates of diagnosis, morphology and topography codes for cancer registrations, and dates and causes of death. Cancer diagnoses and underlying causes of deaths in parents were coded according to the contemporaneous revision of ICD-O and the ICD, respectively. The United Kingdom national cancer registration scheme is described by Quinn et al. (8) .

Statistical Methods.
Expected numbers of cancers and deaths were calculated from serial age and sex-specific cancer incidence and mortality data for England and Wales applied to the similarly defined arrays of person-years at risk generated by the data. All population data were supplied by the Office for National Statistics, London, United Kingdom (9) . For cancer incidence, parents entered the person-years at risk at their date of birth and left the person-years at risk on the date of cancer diagnosis, date last known to be alive, date of embarkation from the United Kingdom, or the closing date of the study, whichever was the earlier. Cancer risk was calculated from the parents’ dates of birth and not birth of the index child because a proportion of children survived malignancy even in the prechemotherapy era after treatment by surgery alone, and we wanted to allow for the possibility that some parents may have survived cancer in childhood (10) . For mortality (morbidity), parents entered the person-years at risk at birth of the index child and left the person-years at risk on the date of death (diagnosis), date last known alive, date of embarkation, or the closing date of the study, whichever was the earlier. All subjects were censored on reaching their 85^th birthday, i.e., they made no additional contributions to expected or observed numbers of cancers or deaths past this age. This is because the general population count is supplied by combining all those ages ≥ 85 years. If the age structure of this open-ended age group is different from that of parents, the expected and observed numbers would not be comparable. It is also thought that diagnoses and deaths were recorded with less accuracy for those over age 85 years (11) , and this would make comparisons of the observed with expected numbers for a specific cancer unreliable. Parental cancers were classified into diagnostic groups using a morphology-based scheme developed for analyzing familial cancer patterns associated with childhood cancers (12) . Two-sided Ps and 95% CIs of SMRs/ SIRs were computed on the assumption the observed numbers followed a Poisson distribution. All analyses have been carried out using Stata (13) . Separate analyses were performed for each diagnostic group of childhood cancer, each diagnostic group in parents, or both.

A proportion of certain childhood cancers is known to be associated with germ-line mutations in specific genes, which can also predispose to adult-onset cancers. These genes include TP53, NF1, NF2, patched, and Rb1. The main childhood-onset cancers associated with mutations in these genes include CNS tumors, soft tissue sarcoma, osteosarcoma, chondrosarcoma, ACC, and retinoblastoma. The most frequent adult-onset cancers associated with such mutations are carcinoma of the breast, bone and soft tissue sarcoma, and CNS tumors (1 , 2 , 12) . To test the extent to which these known genes might be responsible for the observed parental cancers, the parents of children with CNS tumors, bone and soft tissue sarcoma, ACC, and retinoblastoma were removed from the cohort, and the analyses repeated for parents of the remaining diagnostic groups.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

 
Table 1 shows numbers of children eligible for the study, and the numbers of their parents traced. A total of 2604 children with solid tumors were eligible; 2288 mothers and 2204 fathers were successfully traced and flagged at NHSCR. Success rates of follow-up were 88 and 85% for mothers and fathers, respectively. Reasons for nontracing were adoption, emigration, in the armed services, or insufficient information.

Table 5 shows the SMRs by cause of death. Overall, the O of deaths from all causes is close to the expected (O = 791, E = 814.2, SMR = 0.97; 95% CI, 0.91–1.04). There was no significant excess mortality from any specific cancer or all cancers (O = 269, E = 268.9, SMR = 1.0; 95% CI, 0.9–1.1). Similarly, there was no excess mortality from noncancer causes of death (O = 522, E = 545.3, SMR = 0.96; 95% CI, 0.88–1.04). However, there were significant deficits of deaths from diseases of the nervous system, SMR = 0.4 (95% CI, 0.2–0.9) and suicide SMR = 0.3 (95 CI, 0.1–0.8).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This is unlikely to be explained by chance because of selection in the ascertainment procedure (14) because in this population-based study, a near complete ascertainment of children with solid tumors in a well-defined geographical area was achieved. In addition, the observed excesses were associated with specific diagnostic groups. Neither can these excesses be explained by the healthy parent effect, by analogy with the healthy worker effect (15 , 16) , whereby the cohort of parents may experience a lower morbidity than the general population because these individuals must be reproductively fit (17) . If anything, this would bias the SMR/SIR in the opposite direction. Furthermore, there is no reporting bias because all data were obtained from population registries. This may represent a new multiple cancer syndrome, and the most biologically feasible interpretation would be inherited mutations in an unknown gene transmitted from a carrier parent. However, numbers of cases in this group are small, and a chance association cannot be ruled out. Confirmation of this finding in an independent dataset is required.

Li (18) pointed out that inherited cancer genes are the most potent oncogenic influence in humans, exceeding the effects of such environmental agents as ionizing radiation, tobacco, and occupational carcinogens. Although mutations in known genes, including TP53, NF1, NF2, patched, and Rb1, may account for a substantial part of the excess cancer risk, other unknown genes may also be contributing to the development of cancers in parents of children with certain diagnoses other than GCTs. It has been demonstrated recently that germ-line mutations to TP53 predispose to a specific spectrum of mainly rare cancers (12) . Such mutations are unlikely to account for the excess of CNS tumors in parents of children with types of cancers not known to be associated with germ-line TP53 mutations. Other genes associated with increased risk of specific childhood cancers, e.g., WT1 are not known to predispose to adult-onset cancers. It is possible, therefore, that rare predisposing mutations in an unknown gene or genes may be present in a few families.

Certain parental environmental exposures may increase the risk of cancer in offspring. Our second hypothesis proposed that parental exposures to such environmental agents might result in increased cancer risks in the parents as well as their offspring. Three such exposures that have been most consistently linked with childhood cancer risk are parental smoking behavior and occupational exposures to hydrocarbons (especially vehicle related) and agrochemicals (especially pesticides; Refs. 19, 20, 21, 22, 23 ). Such exposures may also lead to increased risk of lung cancer, leukemia, and lymphoma in the parents themselves (24, 25, 26) . We therefore hypothesized that there may be an excess of specific cancers in the parents of children with cancer because of the relevant exposures. Such excesses of cancers in parents would support an etiological role for the corresponding environmental agents in childhood cancer.

However, this study does not provide evidence that parental tobacco use and parental occupational exposures to hydrocarbons and agrochemicals are risk factors for childhood solid tumors because we did not find excess carcinoma of the lung in parents and similarly, we failed to detect excesses of leukemia and lymphoma. We cannot reject the possibility of links between such parental exposures and risk of childhood cancer, but the present results do not support our prior hypothesis.

Our third hypothesis was that the pattern of mortality from noncancer causes among parents of children with cancer might indicate lifestyle factors that could influence childhood cancer risk. However, with respect to mortality, this study does not suggest that families of children with cancer experience greater mortality than the rest of the population. The observed number of deaths overall is close to that expected in the general population, and there were deficits of deaths from nervous system diseases and suicide. Although having a child with cancer or the death of a child because of cancer is a severe adverse life event that could increase the likelihood of suicide (27) , the deficit of deaths from suicide suggests that this is not the case. However, in the general population, reasons for suicide include other life events such as divorce, separation, and unemployment. We do not have information about such life events in our cohort.

This study has a number of advantages over previous studies. Data on childhood cases were population based with near complete ascertainment in a geographically defined area. Although the study covers nearly half a century, histopathological material was available for review, and cases were classified according to current criteria. Adult onset cancers in parents were also derived from population-based data, and a good follow-up rate was achieved. Furthermore, in the United Kingdom national cancer registration system, cancers are classified by morphology as well as site. Information for both the observed and expected cancers in the parents was derived from the same source. Thus, ascertainment bias and information bias were unlikely. In addition, uniquely, the study has examined deaths from all causes, including noncancer causes and, therefore, provided an opportunity to explore shared environmental and lifestyle hypotheses and to assess the general health of the parental cohort in terms of general mortality rate.

In conclusion, we found a small excess cancer risk in parents of children with solid tumors, including novel associations in certain subgroups. Additional pedigree studies and in molecular studies among selected members of the families may clarify this aspect. This study lends additional credence to the hypothesis that familial factors play a role in the etiology of childhood tumors, but high penetrance mutations in cancer-associated genes probably account for a small percentage of cases only. Overall, the parents of children with cancer do not experience poorer health than the general population in terms of mortality from any specific cause. These results are very reassuring for families of children with cancer.

	Acknowledgments

 
We thank the staff of the National Health Service Central Registers for England and Wales and Scotland for flagging and tracing of parents by. We also thank Ewa Dale, Joy Hogg, and Terry Mather for their hard work on data preparation. We thank the families who took part in the study.

	Footnotes

 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

¹ The project is supported by Cancer Research UK.

² To whom requests for reprints should be addressed, at Cancer Research UK, Paediatric and Familial Cancer Research Group, Royal Manchester Children’s Hospital, Stancliffe, Hospital Road, Manchester M27 4HA, United Kingdom. E-mail: jillian.m.birch{at}man.ac.uk

³ The abbreviations used are: ICD, International Classification of Diseases; MCTR, Manchester Children’s Tumor Registry; NHSCR, National Health Service Central Register; HD, Hodgkin’s disease; NHL, non-Hodgkin’s lymphoma; PNET, primitive neuroectodermal tumor; CNS, central nervous system; Rb, retinoblastoma; ACC, adrenocortical carcinoma; WT, Wilms’ tumor; HB, hepatoblastoma; RMS, rhabdomyosarcoma; CI, confidence interval; SMR, standardized mortality ratio; SIR, standardized incidence ratio; O, observed number; E, expected number.

⁴ Internet address: http://www.biomed2.man.ac.uk/crcpfcrg/CRUKPFCRG/PFCRG.htm.

Received 10/18/02; revised 2/25/03; accepted 3/ 6/03.

	References

Top Abstract Introduction Materials and Methods Results Discussion References

 

1. Birch J. M. Genes and cancer. Arch. Dis. Child., 80: 1-3, 1999.[Free Full Text]
2. Blanquet V., Turleau C., Gross-Morand M. S., Senamaud-Beaufort C., Doz F., Besmond C. Spectrum of germline mutations in the Rb1 gene: a study of 232 patients with hereditary and non-hereditary retinoblastoma. Hum. Mol. Genet., 4: 383-388, 1995.[Abstract/Free Full Text]
3. Birch J. M. The Li-Fraumeni syndrome and the role of TP53 mutations in predisposition to cancer Eeles R. A. Ponder B. A. J. Easton D. F. Horwich A. eds. . Genetic Predisposition to Cancer, Ed. 2 Edward Arnold Ltd. London 2003.
4. Olsen J. H., Boice J. D., Seersholm N., Bautz A., Fraumeni J. F. Cancer in the parents of children with cancer. N. Engl. J. Med., 333: 1594-1599, 1995.[Abstract/Free Full Text]
5. World Health Organization. . International statistical classification of diseases, injuries and causes of death., Seventh revision. WHO Geneva 1955.
6. McNally R. J. Q., Kelsey A. M., Cairns D. P., Taylor G. M., Eden O. B., Birch J. M. Temporal increases in the incidence of childhood solid tumours seen in Northwest England (1954–1998): are likely to be real. Cancer (Phila.), 92: 1967-1976, 2001.[Medline]
7. World Health Organization . International classification of diseases for oncology, Ed. 2. WHO Geneva 1990.
8. Quinn M., Babb P., Brock A., Kirby L., Jones J. Cancer trends in England and Wales 1950–1999. Studies on medical and population subjects no. 66., The Stationery Office London 2001.
9. Quinn M. J., Babb P. J., Jones J., Baker A., Ault C. Cancer 1971–1997. Registrations of cancer cases and deaths in England and Wales by sex, age, year, health region and type of cancer (CD-ROM)., Office for National Statistics London 1999.
10. Birch J. M., Marsden H. B., Jones P. H., Pearson D., Blair V. Improvements in survival from childhood cancer: results of a population based survey over 30 years. Br. Med. J., 296: 1372-1376, 1988.
11. Swerdlow A., dos Santos Silva I., Doll R. Cancer Incidence and mortality in England and Wales: trends and risk factors. Materials and methods, 4-7, Oxford University Press Oxford 2001.
12. Birch J. M., Alston R. D., McNally R. J. Q., Evans D. G. R., Kelsey A. M., Harris M., Eden O. B., Varley J. M. Relative frequency and morphology of cancers in carriers of germline TP53 mutation. Oncogene, 20: 4621-4628, 2001.[Medline]
13. StataCorp. . Stata statistical software: release 6.0., Stata Corporation College Station, TX 1999.
14. Li F. P. Familial aggregation Schottenfeld D. Fraumeni J. F. eds. . Cancer Epidemiology and Prevention, 546-558, Oxford University Press Oxford 1996.
15. McMichael A. J. Standardized mortality ratios and "healthy worker effect": scratching beneath the surface. J. Occup. Med., 18: 165-168, 1976.[Medline]
16. Carpenter L. M. Some observations on the healthy worker effect. Br. J. Ind. Med., 44: 289-291, 1987.[Medline]
17. Sanders B. M., Jay M., Draper G. J., Roberts E. M. Non-ocular cancer in relatives of retinoblastoma patients. Br. J. Cancer, 60: 358-365, 1989.[Medline]
18. Li F. P. Cancer epidemiology and prevention. Sci. Am. Med., 10: 12:I:28 1987.
19. Sorahan T., McKinney P. A., Mann J. R., Lancashire R. J., Stiller C. A., Birch J. M., Dodd H. E., Cartwright R. A. Childhood cancer and parental use of tobacco: findings from the inter-regional epidemiological study of childhood cancer (IRESCC). Br. J. Cancer, 84: 141-146, 2001.[Medline]
20. Feychting M., Plato N., Nise G., Ahlbom A. Paternal occupational exposures and childhood cancer. Environ. Health Perspect., 109: 193-196, 2001.[Medline]
21. Zahm S. H., Ward M. H. Pesticides and childhood cancer. Environ. Health Perspect., 106(Suppl.3): 893-908, 1998.
22. Colt J. S., Blair A. Parental occupational exposures and risk of childhood cancer. Environ. Health Perspect., 106(Suppl.3): 909-925, 1998.
23. Little J. Exposures to Chemicals and Dusts. IARC Scientific Publications No. 149. Epidemiology of Childhood Cancer., 6: 178 1999.
24. Blot W. J., Fraumeni J. F., Jr. Cancers of the lung and pleura. Cancer Epidemiol. Prev., 31: 637-665, 1996.
25. Linet M. S., Cartwright R. A. The leukaemias. Cancer Epidemiol. Prev., 40: 841-892, 1996.
26. Scherr P. A., Mueller N. E. Non-Hodgkin’s lymphomas. Cancer Epidemiol. Prev., 42: 920-945, 1996.
27. Cavanagh J. T. O., Owens D. G. C., Johnstone E. C. Life events in suicide and undetermined death in south-east Scotland: a case-control study using the method of psychological autopsy. Soc. Psychiatr. Epidemiol., 34: 645-650, 1999.

This article has been cited by other articles:

				E. Y. Chuang, X. Chen, M.-H. Tsai, H. Yan, C.-Y. Li, J. B. Mitchell, H. Nagasawa, P. F. Wilson, Y. Peng, M. M. Fitzek, et al. Abnormal gene expression profiles in unaffected parents of patients with hereditary-type retinoblastoma. Cancer Res., April 1, 2006; 66(7): 3428 - 3433. [Abstract] [Full Text] [PDF]

				D. LICHTERMANN Cancer Risk in Parents of Patients With Schizophrenia Am J Psychiatry, May 1, 2005; 162(5): 1024 - 1024. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Pang, D. Articles by Birch, J. M. Search for Related Content PubMed PubMed Citation Articles by Pang, D. Articles by Birch, J. M.