Exploring Exclusive Breastfeeding and Childhood Cancer Using Linked Data

Abstract

Søegaard and colleagues 1 investigated whether prolonged exclusive breastfeeding (defined as Ն3 months) was associated with childhood cancer. Their register-based study included all children in the Danish National Child Health Register who were born between 2005 and 2018 and who had information about exclusive breastfeeding, defined as breast milk supplemented only by water or milk formula (maximum of once weekly). Until 2011, the reporting of exclusive breastfeeding in Denmark was voluntary (data available for 21% of births) but became mandatory in 2012 (data available for 62% of births). These data were linked to other population-based registers (Medical Birth Register, Civil Registration System, Population Education Register, Danish National Patient Register, and Danish Cancer Register). Children were followed up until diagnosis, loss to follow-up or emigration, death, age 15 years, or December 31, 2020 (ie, minimum of 2 years of follow-up). Using Cox proportional hazards regression and based on approximately 310 000 births and 1.6 million person-years of follow-up, the authors found that children who were exclusively breastfed for at least 3 months had a decreased risk of hematologic cancers (based on 124 cases), particularly B-cell precursor acute lymphoblastic leukemia (BCP-ALL; 74 cases), compared with those who were breastfed for less than 3 months. No such association was found for central nervous system (CNS) tumors (44 cases) or solid tumors (80 cases). The study was underpowered to investigate rarer cancers, such as acute myeloid leukemia (AML) or Hodgkin lymphoma.The study by Søegaard et al 1 is important, as it is, to my knowledge, the first population-based cohort study to report that prolonged exclusive breastfeeding may be associated with decreased risk of ALL. These findings support the existing literature based on case-control studies for both ALL 2,3 and CNS tumors. 3,4 A 2005 meta-analysis found an association of prolonged breastfeeding for more than 6 months with a decreased risk of ALL (13 studies) and acute nonlymphoblastic leukemia (9 studies) but no association with risk of CNS tumors (4 studies). 3 Similarly, using pooled data from multiple case-control studies, the Childhood Cancer and Leukemia International Consortium recently reported decreased risks of ALL (>10 000 cases) and AML (approximately 1700 cases) associated with breastfeeding for 4 months or more 2 but no association between CNS tumors and at least 6 months of breastfeeding. 4 Childhood cancer research is challenging because of the rarity of disease, leading to sample size issues. The consistency of findings between cohort and case-control studies shows that case-control studies can play an important role in identifying potential risk factors for rare diseases despite the risk of recall bias.The article by Søegaard et al 1 highlights some of the benefits of linking routinely collected population-based data for epidemiologic research, especially when the linkages move beyond the domain of conventional registers (births and ...