Abstract
Evaluation of: Schmidt LS, Kamper-Jorgensen M, Schmeigelow K et al. Infectious exposure in the first years of life and risk of central nervous system tumours in children: analysis of birth order, childcare attendance and seasonality of birth. Br. J. Cancer 102(11), 1670–1675 (2010).
Early infectious exposures have been implicated in the etiology of childhood CNS tumors. The article evaluated here assesses whether infectious exposure in the first years of life is involved by analyzing birth order, childcare attendance and seasonality of birth. The conclusion is that exposure to infectious disease in early childhood does not play an important role in the etiology of pediatric CNS tumors. It is noted that the measures used are only proxies for infectious exposures and, as such, may not accurately reflect the underlying infectious burden. Furthermore, this study is from the Nordic countries and from a later period than previous studies. Differences in patterns of exposure to infection may vary geographically between countries and may have changed over time. However, the authors have not excluded the possibility that a specific infectious agent is involved, and it is agreed here that future efforts should include a focus on a search for such an agent or agents, as well as formulating plausible, diagnostic-specific, mechanistic hypotheses.
It has been suggested that infectious exposures may play a role in the etiology of CNS tumors in children Citation[1]. Some epidemiological studies have suggested that infectious exposures occurring during pregnancy or around the time of birth may be involved. These exposures included measles and influenza Citation[2–5]. Evidence of viral genomic inclusions within certain childhood CNS tumor cells has also been reported Citation[6,7]. Descriptive epidemiological studies, including an analysis of space–time clustering and seasonal variation, have also suggested an infectious etiology for childhood CNS tumors Citation[8,9].
The study reviewed here sought to determine whether infectious exposure in the first years of life is related to the etiology of CNS tumors in children by analyzing birth order, childcare attendance and seasonality of birth Citation[10].
Methods & results
The authors undertook a case–control study in a cohort of the Nordic childhood population, which totals 4.4 million children aged 0–14 years. The aim of their investigation was to examine infectious disease in relation to the risk of developing a CNS tumor. They used birth order and season of birth as ‘proxies’ for infectious exposures. Using a subset of the data, from Denmark, they analyzed the effect of early childcare attendance on the subsequent risk of a CNS tumor.
The case–control study included 3983 children aged 0–14 years who had been diagnosed with a primary CNS tumor during the period 1985–2006 and who lived in Denmark, Norway, Sweden or Finland at the time of diagnosis. Patients were identified from both national and specialized cancer registries. CNS tumors were classified according to the main group III of the third edition of the International Classification of Childhood Cancer Citation[11]. This includes both benign and malignant CNS tumors, but excludes germ cell tumors and lymphomas located in the CNS.
For the birth order analyses, each case was matched by age (birth month, year), sex and country to five controls, obtained at random from the national Nordic childhood population registries. These registries allowed the identification of all maternal siblings, and hence birth order was obtained. Birth order was obtained for 3600 cases and 17,848 matching controls.
For the analysis of seasonality in birth, all children born between 1 January 1985 and 31 December 2006 in the Nordic countries and who were registered in a national birth registry were included. The analysis was restricted to 2771 CNS tumor cases who were born in the same period.
For the analysis of childcare attendance in the first 2 years of life using Danish data, information was included on children born between 1 January 1989 and 17 November 2002 and diagnosed with a CNS tumor during 1989–2006. To avoid possible confounding, children aged less than 2 years at the date of diagnosis were excluded. A total of 351 children were eligible and were matched to 3395 controls.
Conditional logistic regression was used to analyze birth order and childcare attendance. Walter and Elwood’s test was used to analyze sinusoidal variation Citation[12].
The study found no association between birth order and risk of CNS tumors analyzed together or separately by histological sub-group. The odds ratio (OR) for second born or later, compared with first born, was 1.03 (95% CI: 0.96–1.10). Childcare attendance compared with care in the home showed a modestly raised OR (1.29; 95% CI: 0.90–1.86) for all CNS tumors. The highest risk was found for children attending a crèche, and the strongest association was for embryonal CNS tumors. There was no effect of age at enrolment or duration in childcare. The seasonal variation in birth was no different to that of the underlying population. The investigators concluded that the results do not support the hypothesis that exposure to infectious disease in early childhood plays an important role in the etiology of pediatric CNS tumors.
Discussion & significance
The authors of this important article have conducted a well-designed and analyzed study of CNS tumors in the Nordic countries. Their conclusions of no apparent association with early childhood exposures are certainly supported by their data. However, it must be acknowledged that there are some limitations to their conclusions.
First of all, the authors have chosen three proxy measures to represent early childhood exposure to infections: birth order, childcare attendance and seasonal variation. Since the measures used are only proxies for infectious exposures they may not accurately reflect the underlying infectious burden. Indeed, the authors acknowledge that birth order may not be an accurate surrogate measure for all infections, since the risk of infectious disease will depend on the specific virus or bacterium involved Citation[13]. Furthermore, the authors recognize that birth order may not truly reflect exposure to childhood infections in the Nordic countries, since a larger proportion of children use childcare facilities (an important place of social mixing, and hence exposure to infections) Citation[14]. Also, high rates of divorce and new family arrangements suggest the unreliability of birth order in this setting.
Second, the time period of the study is markedly later than previously published studies. Patterns of exposure to infection may have changed over time. For example, immunization programs have become more widespread in recent years. Third, the study was conducted in the Nordic countries. Differences in patterns of exposure to infection may vary geographically between countries. This will be most marked, not only between resource-rich countries and those with more limited resources, but also between countries that have different amounts of variation in wealth distribution. The Nordic countries present a different pattern of socio-economic distribution compared with the UK or the USA Citation[101]. This will inevitably influence the pattern of exposure to early childhood infections.
Expert commentary & five-year view
The evidence concerning the putative role of early infectious exposures in the etiology of childhood CNS tumors is far from clear. Whilst some descriptive and epidemiological studies provide support for such a role Citation[1–5,8,9], the present study, which is the subject of this evaluation, does not Citation[10]. To gain greater understanding of these apparently contradictory findings we need to make comparisons with other malignancies for which an infectious etiology has been proposed. In particular, a role for early childhood exposures has been suggested for childhood acute lymphoblastic leukemia (ALL) and is supported by both descriptive and epidemiological studies. For childhood ALL, it should be noted that there is inconsistency between studies, although overall the evidence is supportive of such an infectious etiology. Furthermore, no specific infectious agent has been identified for childhood ALL Citation[15]. Three mechanistic hypotheses have been put forward Citation[16–19].
There is a lack of similar mechanistic hypotheses for childhood CNS tumors. The process leading to the onset of a childhood CNS tumor (or any other childhood malignancy) is likely to involve at least two events Citation[20]. The initial ‘event’ may be germline or somatic (due to either endogenous or environmental factors). By contrast, the final event is likely to always involve an environmental agent. For childhood ALL, the ‘lag time’ between the final event and the onset of leukaemia is likely to be relatively short. By contrast, for childhood CNS tumors, lag times from the final event that precipitated the onset of a tumor to clinical diagnosis are likely to be quite heterogeneous and much more prolonged. This makes it much more difficult and complex to disentangle the descriptive and epidemiological evidence for childhood CNS tumors compared with childhood ALL.
One major objective of future research should be to develop good plausible mechanistic hypotheses for childhood CNS tumors. These should be very specific for diagnostic subgroups. A second objective for future research should be to search for specific infectious agents that may be implicated for specific childhood CNS tumors. This should involve international collaborative laboratory and epidemiological studies. A final and important objective should be to examine the role of gene–environment interactions in the etiology of childhood CNS tumors. In this context, the environmental exposure would be one or more infectious agents.
Key issues
Key conclusions
• The study found no association between birth order and the risk of CNS tumors.
• The seasonal variation at birth was similar to the underlying population.
• The investigators concluded that there was no support for the involvement of early childhood infectious exposures in pediatric CNS tumors.
Unresolved issues
Issues that still need to be explored include:
• Assessment of the adequacy of the proxy measures used to represent early childhood exposures to infections.
• Differences in patterns of early childhood exposure to infections between the Nordic countries and other countries.
Future work
Future research should focus on:
• Development of plausible diagnostic specific mechanistic hypotheses for childhood CNS tumors.
• Detailed studies of specific putative infectious agents.
• Examination of the role of genetic predisposition to infectious agents.
Financial & competing interests disclosure
The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this key paper evaluation manuscript.
References
- Barbanti-Brodano G, Martini F, De Mattei M et al. BK and JC human polyomaviruses and simian virus 40: natural history of infection in humans, experimental oncogenicity and association with human tumors. Adv. Viral Res.50, 66–99 (1998).
- Linet MS, Gridley G, Cnattingius S et al. Maternal and perinatal risk factors for childhood brain tumors (Sweden). Cancer Causes Control7, 437–448 (1996).
- Linos A, Kardara M, Kosmidis H et al. Reported influenza in pregnancy and childhood tumour. Eur. J. Epidemiol.14, 471–475 (1998).
- Fear NT, Roman E, Ansell P et al. Malignant neoplasms of the brain during childhood: the role of prenatal and neonatal factors (United Kingdom). Cancer Causes Control12, 443–449 (2001).
- Dickinson HO, Nyari TA, Parker L. Childhood solid tumours in relation to infections in the community in Cumbria during pregnancy and around the time of birth. Br. J. Cancer87, 746–750 (2002).
- Krynska B, Del Valle L, Croul S et al. Detection of human neurotropic JC virus DNA sequence and expression of the viral oncogenic protein in pediatric medulloblastomas. Proc. Natl Acad. Sci. USA96, 11519–11524 (1999).
- Kim JY, Koralnik IJ, LeFave M et al. Medulloblastomas and primitive neuroectodermal tumors rarely contain polyomavirus DNA sequences. Neuro. Oncol.4, 165–170 (2002).
- McNally RJ, Cairns DP, Eden OB et al. An infectious aetiology for childhood brain tumours? Evidence from space–time clustering and seasonality analyses. Br. J. Cancer86, 1070–1077 (2002).
- McNally RJ, Kelsey AM, Eden OB et al. [Evidence for an infectious component to the aetiology of childhood brain tumours]. Salud Ciencia13, 26–28 (2005).
- Schmidt LS, Kamper-Jorgensen M, Schmiegelow K et al. Infectious exposure in the first years of life and risk of central nervous system tumours in children: analysis of birth order, childcare attendance and seasonality of birth. Br. J. Cancer102, 1670–1675 (2010).
- Steliarova-Foucher E, Stiller C, Lacour B et al. International Classification of Childhood Cancer, third edition. Cancer103, 1457–1467 (2005).
- Walter SD, Elwood JM. A test for seasonality of events with a variable population at risk. Br. J. Prev. Soc. Med.29, 18–21 (1975).
- Law GR. Host, family and community proxies for infections potentially associated with leukaemia. Radiat. Prot. Dosimetry132, 267–272 (2008).
- Kamper-Jorgensen M, Wohlfahrt J, Simonsen J et al. The childcare database: a valuable register linkage. Scand. J. Public Health35, 323–329 (2007).
- McNally RJ, Eden OB. An infectious aetiology for childhood acute leukaemia: a review of the evidence. Br. J. Haematol.127, 243–263 (2004).
- Greaves MF. Speculations on the cause of childhood acute lymphoblastic leukemia. Leukemia2, 120–125 (1988).
- Kinlen L. Evidence for an infective cause of childhood leukaemia: comparison of a Scottish New Town with nuclear reprocessing sites in Britain. Lancet2, 1323–1327 (1988).
- Kinlen L. Epidemiological evidence for an infective basis in childhood leukaemia. Br. J. Cancer71, 1–5 (1995).
- Smith M. Considerations on a possible viral etiolgy for B-precursor acute lymphoblastic leukemia of childhood. J. Immunother.20, 89–100 (1997).
- Knudson AG. A two-mutation model for human cancer. Adv. Viral Oncol.7, 1–17 (1987).
Website
- Forster M, d’Ercole MM. Income distribution and poverty in OECD countries in the second half of the 1990s www.oecd.org/dataoecd/48/9/34483698.pdf