Potential causal associations between adiposity and hydration status in children and adolescents: evidence from negative control and Mendelian randomization analyses
Abstract
This study examined the potential causal association between body mass index (BMI), waist circumference (WC), and hydration status among children and adolescents. Participants aged 6–18 years were recruited from 10 cities in Henan Province, China, using a three-stage cluster random sampling strategy. BMI Z -scores and WC Z -scores were used as primary exposure variables. Hydration status was assessed using urine specific gravity (USG). Restricted cubic spline models were used to assess potential dose–response relationships between BMI Z -scores, WC Z -scores, and log-transformed USG. Logistic regression models were applied to estimate associations between adiposity indicators and hydration status. Parental BMI was used as negative control outcomes to assess potential residual confounding arising from shared familial factors. Bidirectional Mendelian randomization (MR) analyses were performed to investigate the potential causal effects. A total of 11,935 children and adolescents (mean age 9.98 ± 2.17 years; 47.6% girls) were included. Both BMI Z -scores and WC Z -scores showed nonlinear associations with log-transformed USG. Higher BMI Z -scores (OR = 1.248, 95% CI: 1.011–1.541) and WC Z -scores (OR = 1.211, 95% CI: 1.109–1.322) were associated with increased risk of inadequate hydration. Negative control analyses showed no significant associations. MR analyses provided genetic evidence supporting potential causal effects of BMI ( β = 0.073, 95% CI: 0.035–0.112) and WC ( β = 0.097, 95% CI: 0.036–0.158) on USG, whereas reverse MR analyses, which were based on a limited number of available genetic instruments, found no evidence of reverse effects. Higher standardized BMI and WC were associated with an increased risk of inadequate hydration, with negative control and MR analyses providing supportive evidence for potential causality, while reverse causality remains uncertain.