Abstract
Childhood obesity is a growing global health concern, contributing to numerous non-communicable diseases and long-term health complications. The prevalence of obesity in children and adolescents continues to rise, driven by complex interactions among various factors. The key risk factors include both environmental and genetic influences. Environmental factors include family elements like household conditions and lifestyle, while genetic factors refer to inherited predispositions. More recently, epigenetic factors have gained attention, focusing on chemical modifications such as DNA methylation that are influenced by the prenatal and early-life environment and may contribute to obesity risk. Unlike obesity in adults, the risk factors for obesity in children are largely dependent on their family environments rather than individual behaviors. For effective intervention, it is important to identify at-risk children and their families as early as possible after birth. Despite advances in machine learning, polygenic risk scores, and epigenomic markers-which show promise as being more accurate and comprehensive prediction methods-no risk prediction models are currently in clinical use. Achieving predictions with higher accuracy, external validation, and consideration of population-specific factors (e.g., ethnic variability) while avoiding bias or stigma in targeted interventions is needed for effective childhood obesity prevention. Herein, we summarize environmental, genetic, and epigenetic risk factors for childhood obesity and review the unique situations and regional factors in Japan, which are the focus of our study. Furthermore, we introduce the major advances in risk prediction models for childhood obesity.