Abstract
OBJECTIVES: Insulin-like growth factor-1 (IGF-1) levels are affected by nutritional status, yet there is limited research exploring the association between body mass index (BMI) and IGF-1 levels among children. METHODS: This cross-sectional study included 3227 children aged 2-18 years without specific diseases, whose height, weight, and pubertal stages were measured and assessed by pediatricians. BMI standard deviation scores (BMISDS) were used to categorize children as underweight (BMISDS < -2); normal-weight (-2 ≤ BMISDS ≤ 1); overweight (1 < BMISDS ≤ 2); and obese (BMISDS > 2). Children were divided into low-level (<-0.67 SD) and nonlow-level (≥-0.67 SD) groups based on IGF-1 standard deviation scores (IGF-1SDS). The association between IGF-1 and BMI as categorical and continuous variables was explored by Binary logistic regression, the restrictive cubic spline model, and the generalized additive model. Models were adjusted by height and pubertal development. Recursive algorithm and multivariate piecewise linear regression were further utilized to assess the threshold of the smooth curve. RESULTS: IGF-1 levels varied by BMI categories, with the highest levels observed in the overweight group. The proportion of low IGF-1 levels in underweight, normal-weight, overweight, and obese groups was 32.1%, 14.2%, 8.4%, and 6.5%, respectively. The risk odds of low IGF-1 levels in underweight children were 2.86-, 2.20-, and 2.25-fold higher than in children with normal weight before adjustment, after adjustment for height, and after adjustment for height and puberty, respectively. When analyzing the association between BMI and low IGF-1 levels, dose-response analysis demonstrated an inverted J-shaped relationship between BMISDS and low IGF-1 levels. Lower or higher BMISDS increased the odds of low IGF-1 levels, and significance was retained in underweight children but not in obese children. When BMI and IGF-1 levels were used as continuous variables, the relationship between the BMISDS and IGF-1SDS followed a nonlinear inverted U shape. IGF-1SDS increased with the increase of BMISDS (β = 0.174, 95% CI: 0.141 to 0.208, p < 0.01) when BMISDS was less than 1.71 standard deviation (SD) and decreased with the increase of BMISDS (β = -0.358, 95% CI: -0.474 to -0.241, p < 0.01) when BMISDS was greater than 1.71 SD. CONCLUSIONS: The relationship between BMI and IGF-1 levels was found to depend on the type of variable, and extremely low or high BMI values could result in a tendency toward low IGF-1 levels, emphasizing the importance of maintaining a normal BMI range for normal IGF-1 levels.