Cancer is a leading cause of death for children and adolescents worldwide, where Scandinavian countries together with North America and other European countrieshave the highest cancer incidence among children.¹ The causes of most childhood cancers are unknown. Family history of cancer, certain genetic conditions, and some environmental factors, such as exposure to radiation, account for only a small percentage of the cases.² Hereditary syndromes such as Li-Fraumeni syndrome, an autosomal dominant disorder caused by germline TP53 mutations, are thought to be a rare cause of paediatric cancers.³

Childhood cancer risk factors may be different in distinct populations, and studies carried out in migrant populations might contribute to split the role of genetic and environmental factors in cancer development. In the present study we focus on leukaemia and tumours in the central nervous system (CNS), the most common cancers in children accounting for around 60% of all cancer cases.⁴ Leukaemia can occur at any age but it is most commonly diagnosed in children <4 years old.⁵ Astrocytomas and embryonal tumoursare the most common types of childhood CNS tumours, followed byependymomas.⁶ Younger children have a higher incidence of tumours of embryonal origin, such as medulloblastoma or atypical teratoid/rhabdoid tumour, whereas older children tend to have tumours of glial origin such as astrocytomas.⁷

Over one million Scandinavians emigrated to America throughout the 19th century; some weremembers of the Church of Jesus Christ of the Latter-Day Saints (LDS) from the Scandinavian Mission.^8,9,10,11 They had many of the same motives as other migrants, but in addition they were directly encouraged to settle in Utah where most of the population still practices the LDS religion. This population is well-known for their lifestyle characteristics including large family size, a healthy lifestyle, and proscriptions against alcoholic beverages and tobacco.¹²

The Utah Population Database (UPDB) and the Norwegian National Population Register (DSF) are well established registries that allowfor population-based analyses of the familial nature of diseases. We were able to conduct a unique comparison of incidence and family history of cancer in Utah and in Norway,considering the influence of Scandinavian (Norwegian, Swedish and Danish) ancestry on the development and aetiology of childhood cancers.

The goals of the present study were:

i) to explore childhood CNS tumours and leukaemia incidence ratesin Norway and Utah and across Scandinavian ancestry in Utah; ii) to quantify the potentially increased risk of childhood CNS tumours and leukaemia among those with a family historyof leukaemia, CNS tumours, solid tumours and all cancers, taking intoconsideration Scandinavian ancestry in Utah;iii) to quantify the effect of family history of Li-Fraumenitumour spectrum on the risk of childhood CNS tumours and leukaemia, taking Scandinavian ancestry in Utah into account.

During the 49-year study period, a total of 1,309,797 children with genealogy data as previously described were born in Utah, and 3,835,376 children were born in Norway. Overall, 71.4% of the Utah children had ancestral roots from Scandinavia. Total numbers of cancer cases are presented in Table 1.

To our knowledge, this study is the first comparative study to examine incidence and the familial risk of childhood CNS tumours and leukaemia among children with Scandinavian ancestry born in Utah compared with children born in a Scandinavian country (Norway).

In this study the incidence rate of CNS tumours showed a significant increase with consecutive birth cohorts in both populations. Childhood leukaemia (0–14 years) incidence rates showed more stable birth cohort trendsduring the study period, with only a slight increase in incidence rates among children born in the last periods in Norway.A family history of CNS tumours was associated with an elevated risk of childhood CNS tumours. The HR’s of CNS tumours was also elevated when family members were diagnosed with a tumour belonging to the Li-Fraumeni tumour spectrum.

The results from the present study are in accordance with previous observations of increasing incidence of childhood cancer among Scandinavian children, especially in Denmark and Norway.^28,29,30,31 However, several studies show geographical and ethnicity differences of incidence ofcancer.^32,33 Our results show that the state of Utah has incidence rates of childhood CNS tumours and leukaemia similar to those found in Scandinavia.The high incidence rates reported here probably reflect the completeness of the registries, due to complete coverage in both populations. Incidence rates of CNS tumours in Utah and USA in the general population, regardless of age, show a stable trend from 1975–2016 (Appendix 4). By contrast,in Norway an increased incidence of CNS tumours was noted. However, a decline in the incidence rates has been observed in recent years. Incidence of leukaemia in the general population was relativelystable in the three populations from 1975 to 2016. Information on incidence rates among Utah children is scarce. The National Cancer Institute reports childhood cancer rates in a population consisting of 18 areas, including Utah, of 4.6/100,000 person-years for childhood CNS tumours and 5.4/100,000 person-years for leukaemia in the period 2011–2015, almost similar to our results.³⁴

Childhood cancers, like adult cancers, may be the result of a mixture of genetic, environmental,and behavioural causes, not just one factor by itself. Norwegians descendants in Utah will share genetic makeup, but have grown up in different environments compared with their ancestors from Norway. Cancer incidence in Utah and Norway may be similar because of genetic similarities. However, the role of other factors influencing the aetiology cannot be ruled out.

Geographical and ethnic variations in incidence rates of CNS tumours and leukaemia among children have been reported around the world.^35,36 In USA, incidence rate of childhood cancer is higheramong white children (18.4/100,000 person-years)thanamong African-Americanchildren (13.3/100,000 person-years).³⁶ The incidence rates of CNS tumours and leukaemia found in our study are among the highest in the world and are similar to those observed in other parts of Europe and Canada.The stable incidence rate of leukaemia observed in Utah in Figure 2 might indicate that the prevalence of risk factors involved in the aetiology of the disease have been constant over time.An alternative explanation might be that the number of susceptible individuals in Utah and prevalence of risk factors are at their maximum.

The incidence of leukaemia diagnosed in children before 5 years isoverall higher than the incidence observed in older children. In England and Wales studies report an incidence of 7.9/100,000 person-years. This rate is similar to our study.³⁷ Lower incidence rates of paediatric CNS tumours are reported in some Southern and Eastern European countries with rates ranging from 3.1 to 5.0/100,000 person-years.³⁸

We observed a change over time in the incidence of CNStumours, which may be attributable to the development of radiological imaging techniques such as CT and MRIfor the evaluation of brain tumours.³⁹ In 1971 a new era of imaging based diagnostics began with the creation of the first CT scanner. Later in 1977 the nuclear magnetic resonance technique (MRI) appears. In 1980, the first papers reporting CT scanner as the most accurate diagnostic test in children with brain tumours were published.^40,41 In Norway,the total number ofradiological examinationof CNS tumoursincreased by 16% from 1993 to 2002.⁴² Despite the increased incidence of CNS tumoursobserved in the present study, we also noted that the increase in the incidence of CNS tumoursin the youngest age group flattened out among children born after 1999, which may indicate that the effect of using these diagnostic techniques on the registration of new cases have ceased.The incidence of childhood leukaemia (0–14 years) was more stable in both populations. However, a significantly increased trend was observed in Norway.A similar increase is observed in data published by Cancer statistics for the Nordic countries(NORDCAN).⁴³ The increase in leukaemia rates may also reflect improved in diagnostics, classification and accuracy of reporting in recent years.

Our findings also confirm and strengthen the evidence of common familial risks for childhood CNS tumours and leukaemia, and are consistent with a wide range of studies^44,45,46,47 The biological mechanismsbehind the associations found in the present study might be explained by genetic and environmental factors or interaction between them. However, studies conducted in first-degree relatives suffer from inability to discern between possible environmental effects or shared genetic susceptibility.

Environmental and inherent risk factors consistently associated with childhood cancer are high-dose radiation, prior chemotherapy, birth weight, parental age, birth defects and family cancer syndromes.⁴⁸ Our findings of elevated risk of CNS tumours among children with a family history of Li-Fraumeni tumours support these remarks.We found a strong familial aggregation of CNS cancer and leukaemia which might indicate that these cancers share a heritable aetiology.Zhang et al, 2015⁴⁹ found that 8.5% of the children and young adults included in their study had a pathogenic or probably pathogenic germline mutationconferring cancer risk. Only 40% of these children had a family history of cancer.They also point out that the prevalence of these mutations might be underestimated.

The majority of childhood CNS tumours occur before the age of five, suggesting that prenatal as well as postnatal exposures must be considered as potential etiologic factors.⁵⁰ Leukaemia can occur at any age, but it is most commonly diagnosed in children <4 years old.⁵¹ Environmental factors associated with elevated risk of childhood cancer with studies showing mixed results are background radiation, non-ionising electromagnetic fields, exposures to chemicals,pesticides, infections, parental exposures, such as parental smoking, alcohol consumption and occupational exposures.⁵² Some of the chemical components of tobacco smoke have been demonstrated to be carcinogenic in animals and in humans, such as polycyclic aromatic hydrocarbons and nitrosamines. Both maternal and paternal smoking during pregnancy and before pregnancy are exposures that have been explored in numerous epidemiologic studies.⁵³

Utah has the lowest proportion of smokers of any U.S. state (8.9%), also low compared to Norway (12%).^54,55 In addition,Utah is a population that excludes alcoholic beverages, coffee, tea, and other addictive substances. Similar incidence trends of CNStumours and leukaemia were observed in Utah and Norway. If some of these environmental factors are involved in the development of childhood cancer due to germline mutations and epigenetics, we would probably have expected differences in incidence rates between Utah and Norway.

Perinatal and reproductive factors have been consistently associated with both the risk of childhood CNS and leukaemia cases.^56,57 There is also a growing evidence that for the majority of childhood acute leukaemia the first genetic event occurs prenatally in the foetus.⁵⁸

Parental age and parity have been associated with risk of childhood cancer in early studies, probably because of the possibility of associations with infections in early life,^37,38 and because maternal age might be associated with chromosomal aberrations of female germ cells.⁵⁹ Older maternal age has been associated with an increased risk of leukaemia, especially in the age group 0–4 years.⁶⁰ In the Utah population Mormon women are more likely to have more pregnancies with two children for every one mother compared to1.5 children among Norwegian women, according to statistics for 2017.^61,62 Utah women are also more likely to be younger at birth. The mean age of mothers at all births in Utah is 23.3 years in 2000,⁶³ compared to 29.6 among women in Norway at the same year.⁶⁴

We found a significant increase in incidence of CNS tumourswith consecutive birth cohorts in both populations, regardless of probable life style variations among generations of people living in Norway and Utah.

CNS tumours and leukaemia aggregatin families. Children with a family history of Li-Fraumeni tumours showed anincreasedrisk of developing these cancers.Given the importance of this syndrome in the risk of developing cancer at early age, more research including this risk population is needed.Our research findings may be useful for prioritizing childhood cancer research and continue monitoring changes in incidence of childhood cancers.