Age-period-cohort effects on incidence trends of childhood leukemia from four population-based cancer registries in Colombia

Background: Childhood leukemia (CL) is the most prevalent form of pediatric cancer on a global scale. However, there is a limited understanding of the dynamics of CL incidence in South America, with a specific knowledge gap in Colombia. This study aimed to identify trends in CL incidence and to analyze the effects of age, period, and birth cohort on the risk of leukemia incidence in this population. Methods: Information on all newly diagnosed leukemia cases (in general and by subtype) among residents aged 0–18 years and living in the serving areas of population-based cancer registries of Cali (2008–2017), Bucaramanga (2000–2017), Manizales (2003–2017), and Pasto (1998–2018). Estimated annual percent changes (EAPC) in incidence over time and potential changes in the slope of these EAPCs were calculated using joinpoint regression models. The effects of age, period, and cohort in CL incidence trends were evaluated using age-period-cohort models addressing the identifiability issue through the application of double differences. Results: A total of 966 childhood leukemia cases were identified. The average standardized incidence rate (ASIR) of leukemia was calculated and expressed per 100,000 person-years – observing ASIR of 4.46 in Cali, 7.27 in Bucaramanga, 3.89 in Manizales and 4.06 in Pasto. Concerning CL trends there were no statistically significant changes in EAPC throughout the different periods, however, when analyzed by leukemia subtype, statistically significant changes were observed in the EAPC for both ALL and AML. Analysis of age-period-cohort models revealed that age-related factors significantly underpin the incidence trends of childhood leukemia in these four Colombian cities. Conclusions: This study offers valuable insights into the incidence trends of childhood leukemia in four major Colombian cities. The analysis revealed stable overall CL incidence rates across varying periods, predominantly influenced by age-related factors and the absence of cohort and period effects. This information is useful for surveillance and planning purposes for CL diagnosis and treatment in Colombia.