Abstract
Background & aims: Body mass index (BMI) is used to identify high-risk groups in childhood and to target early interventions to prevent later metabolic disorders. This is a priority area for research among children born very preterm (VPT) given growing concern with their long-term risks of adverse metabolic outcomes. Two principal international BMI classifications exist for underweight, overweight and obesity (OWOB) in children (International Obesity Task Force (IOTF) and World Health Organization (WHO)), but how the choice of reference affects results of research in children born VPT is unknown. Our objective was to compare the prevalence and risk factors for underweight and OWOB using these two references among five-year-old children born VPT.
Methods: Data comes from a population-based cohort of children born VPT in 11 European countries in 2011-12 with information from medical records during the neonatal hospitalization and parental questionnaires at age five. BMI at five years of age was classified into underweight and OWOB using IOTF and WHO references (n=2,654 children). Conversion algorithms were also applied to prevalence estimates. Associations with sociodemographic, perinatal and neonatal characteristics were assessed using multinomial logistic regression with multiple imputation and inverse probability weighting to account for missing data and attrition.
Results: After applying a cutoff of -1SD to define mild underweight for WHO – since this category exists only for IOTF – the estimated prevalence of total underweight was similar between IOTF and WHO references (27.8% vs 27.0%, respectively). IOTF classified a higher proportion of children as having severe underweight (4.1% vs 1.6%). For OWOB, prevalence estimates were lower using IOTF, particularly among boys (8.6% vs 14.1%). The algorithms provided good conversion (1-2% absolute difference) of prevalence between references for underweight and obesity overall, and overweight for girls, but had a larger error for overweight in boys. Risk factors were similar for underweight and OWOB for both references, with the exception of sex and maternal country of birth (OWOB significantly associated using IOTF but not WHO references).
Conclusion: Clinicians and researchers should be aware of the difference in prevalence of suboptimal BMI when interpreting findings from studies using different classifications and IOTF should be used with caution for investigating sex differences. Given the high prevalence of mild underweight in children born VPT, the cutoff of -1SD should be used with the WHO references in studies of BMI in this population.
Keywords: Body mass index; IOTF growth charts; Overweight and obesity; Underweight; Very preterm; WHO growth charts.