Childhood cancer incidence by migrant background in Sweden (1991-2021): a nationwide cohort study.

Background: Studies indicate that incidence of certain childhood cancers is lower in low-compared to high-income countries. By examining incidence in migrant populations in Sweden, we aimed to disentangle whether this is due to methodological factors, such as underreporting, or to genetic or environmental aetiology. Methods: In this nationwide, population-based cohort study, we grouped all children born in Sweden (1990-2019) as: Swedish background, 2nd generation, 2.5 generation (one foreign-born, one Swedish-born parent), or 3rd generation. Age-standardized incidence rates (ASR) for first childhood cancer identified in the National Cancer Register (ages 1-19) were calculated using the 2013 European standard population and are reported per 100,000 person-years [95% confidence intervals]. Incidence rate ratios (IRR) were calculated using Poisson regression adjusted for year of birth and sex. Using the same methods, we examined childhood cancer incidence by parental lineages separately, parental length of stay before birth, and World Bank income classification of parental country of birth. Findings: During 39,862,393 person-years of follow-up, 6797 children were diagnosed with cancer. The ASR of childhood cancer in Swedish background children was 17.18 [16.63-17.74], this was slightly higher in boys (18.20 [17.42-19.01]) than girls (16.09 [15.34-16.88]). The overall results were largely similar across migrant generations and parental length of stay. Lymphoma incidence was higher in 2nd generation (ASR 3.09 [2.53-3.78]) compared to Swedish background children (2.22 [2.02-2.44]). Children with migrant mothers from high-income countries had a higher incidence of leukaemia (ASR 5.88 [4.78-7.24]) compared to Swedish background children (4.31 [4.09-4.55]), whereas central nervous system tumour and leukaemia incidence was lower in children with mothers from low-income countries (2.64 [1.79-3.88], 3.66 [2.76-4.85], respectively) compared to Swedish background children (4.49 [4.24-4.75], (4.31 [4.09-4.55]). The pattern was similar for fathers from low-income countries and for IRRs. Interpretation: Finding lower incidence of CNS tumours and leukaemia among children born in Sweden to mothers from low-income countries challenges the hypothesis that low incidence of some childhood cancers in low-income countries is solely due to underreporting, but rather suggest that genetic or environmental factors may underlie these observations. Funding: Barncancerfonden (PR2023-0060).