Background
The causes of childhood cancer remain poorly understood, and prenatal environmental factors including maternal occupational exposures may contribute to susceptibility. Childhood cancers likely originate from influences acting before birth, and maternal occupational exposures have been associated with childhood cancer in some studies (e.g., benzene exposure in relation to childhood lymphoma and pesticide exposure in relation to childhood leukaemia). The limited evidence suggests a need for large-scale studies with detailed occupational histories and consistent exposure assessment. 

Objectives
To investigate the associations between maternal occupational exposures (pre-pregnancy and during pregnancy) and childhood leukaemia, central nervous system (CNS) tumours, and lymphoma in the offspring. 

Methods
We conducted a nationwide, register-linked cohort study in Denmark using the data DOCX-Generation (approx. 2.5 million recorded live births from 1977). Each child contributed one record and was followed from date of birth until the earliest of first cancer diagnosis, 19th birthday, date of emigration, date of death, or end of follow-up 31 December 2022. Outcomes were grouped a priori into leukaemia, CNS tumours, and lymphoma; only the first diagnosis up to age 19 years was considered. Maternal occupational exposures were assigned at the occupation level (ISCO-88) with general population job-exposure matrices (JEMs) for organic compounds, heavy metals, pesticides, and extremely low frequency magnetic fields (ELF-MF). Exposure before pregnancy onset and during pregnancy were evaluated separately. We used Cox regression to estimate hazard ratios (HR) with 95% confidence intervals (95% CI), adjusting for the pre-specified confounders including maternal age and family socioeconomic status. 

Results
Maternal occupational exposures yielded signals both before and during pregnancy. In the pre-pregnancy window, maternal exposure to ELF-MF was associated with a higher risk of CNS tumours (HR 1.19; 95% CI 1.02-1.40), and pesticides were associated with a higher risk of lymphoma (HR 1.55; 95% CI 1.02-2.38). Maternal exposure during pregnancy, the association between ELF-MF and CNS tumours persisted at a similar magnitude (HR 1.19; 95% CI 1.01-1.40), while for lymphoma the estimate for pesticides shifted below unity (HR 0.84; 95% CI 0.71-0.99). 

Conclusions/Implications
The increase of CNS tumour risk with maternal ELF-MF across both exposure windows suggests a stable association in these data, noting that our exposure measure is probability-based at the occupation level. The contrast for a positive association between pesticides and lymphoma before pregnancy, but a decreased risk with exposure during pregnancy may reflect changes in work patterns once pregnancy is recognized and/or JEM-related exposure misclassification during pregnancy, rather than an actual decreased risk. To further evaluate long-term exposure patterns, we are currently also investigating cumulative maternal exposure across multiple years before pregnancy, to capture exposure duration and cumulative dose and to align exposure assessment with plausible biological windows for effects relevant to maternal occupational exposures.