1 Ingram School of Nursing, Faculty of Medicine and Health Sciences, Mcgill University, 680 Sherbrooke West, Montreal, QC, H3A 2M7, Canada. andraea.vanhulst@mcgill.ca.
2 Ingram School of Nursing, Faculty of Medicine and Health Sciences, Mcgill University, 680 Sherbrooke West, Montreal, QC, H3A 2M7, Canada.
3 Department of Mathematics and Statistics, Faculty of Arts and Science, Concordia University, Montreal, Canada.
4 Department of Pediatrics, London Health Sciences Centre, Children's Hospital of Western Ontario, Western University, London, Canada.
5 Department of Pediatrics, Faculty of Medicine & Dentistry, College of Health Sciences, University of Alberta, Edmonton, Canada.
6 School of Health, Concordia University, Montreal, Canada.

The World Health Organization recommends using + 2 SD of body mass index z-score (zBMI) to define overweight/obesity (OWO) in children ages 2 to 5 years whereas + 1 SD is used as cut-point from 5 years onwards. Empirical evidence for using different cut-points across childhood is lacking. Our objective was to compare the ability of OWO in early childhood defined using zBMI cut-points at + 2 SD and + 1 SD to predict obesity at 10 years. Data from a prospective birth cohort (QLSCD) were analyzed. At ages 2.5, 3.5, and 4.5 years, children were classified as OWO based on + 2 SD and + 1 SD zBMI cut-points. At 10 years, obesity was assessed (zBMI and waist circumference). Associations between OWO (vs non-OWO) and later obesity were estimated using multivariable linear regressions. Outcome predictions for each cut-point were compared using partial eta-squared values. The sample included 1092 children (53% female). OWO in early childhood was 2-3 times more prevalent when using + 1 SD vs + 2 SD cut-points. In relation to later obesity, partial eta-squared values for both cut-points of OWO were in the small to medium effect size range (ranging from 3 to 15%), suggesting that OWO regardless of cut-point contributed only modestly to obesity measured at 10 years. However, across all time points, eta-squared values were slightly higher for OWO defined at + 1 SD vs + 2 SD, indicating a higher proportion of variance in outcomes being accounted for at zBMI + 1 SD. Conclusion: In children 2 to 5 years old, both definitions of OWO had small to modest effect sizes in relation to obesity in childhood albeit with a marginally superior predictive ability of the + 1 SD over the + 2 SD cut-point across early childhood. From a clinical perspective, using a single cut-point from early childhood onwards may be more practical to monitor growth and weight gain over time and identify children at risk of persistent obesity. What is Known: • The World Health Organization recommends using zBMI cut-points at + 2 SD for children ages 2-5 years, and + 1 SD from 5 years onwards to define overweight/obesity • Research is needed to determine which zBMI cut-point (+ 2 SD or + 1 SD) in children under 5 years best predicts subsequent obesity What is New: • Both definitions of overweight/obesity in early childhood contributed modestly to obesity at 10 years, with + 1 SD being marginally more effective than + 2 SD • Using a single cut-point at + 1 SD across childhood may be more practical for monitoring growth, weight gain, and identifying children at risk of persistent obesity.