The new definition of obesity: an analysis of a population-based survey in an Andean country

Background: Traditional obesity classification based on body mass index (BMI) fails to capture body fat distribution or clinical dysfunction, and may therefore fail to identify people at highest cardiometabolic risk. The Lancet Diabetes & Endocrinology Commission recently proposed a new framework distinguishing preclinical from clinical obesity based on excessive adiposity and clinical dysfunction. We aimed to estimate the prevalence of clinical and preclinical obesity in Peruvian adults using the Commission's criteria and adapted regional cutoffs, and to generate ethnicity-specific reference curves for waist circumference (WC) and waist-to-height ratio (WHtR). Methods: This cross-sectional analysis used nationally representative data from 2021 to 2023 Peruvian Demographic and Health Surveys (ENDES), including 84,622 adults aged ≥20 years. Clinical obesity was defined as excess body fat (BMI, WC, or WHtR) plus diabetes or hypertension diagnosis. Preclinical obesity was defined as excess body fat without clinical dysfunction. Age-adjusted prevalence estimates were calculated using four anthropometric criteria and stratified by sex. Ethnicity-specific WC and WHtR reference curves were generated using GAMLSS models, stratified by age and sex. Findings: A total of 84,622 participants were included in the study. Among these participants, mean age was 44.1 (range: 20-97) years, and 48,300 participants (51.7%) were female. Clinical obesity age-adjusted prevalence ranged from 15.7% to 22.1%, and preclinical obesity from 28.7% to 53.8%, depending on cutoffs used. Up to 13.5% of individuals with normal BMI and 21% of those overweight met criteria for clinical obesity. Women showed the highest prevalence estimates of preclinical obesity, ranging from 33.4% to 65.8%, whereas men reached their highest prevalence (41.3%) when the International Diabetes Federation (IDF) cutoffs were applied. In the case of clinical obesity, women had higher prevalence estimates of clinical obesity when applying the Lancet Commission approach (18.7%) and the Peruvian national guidelines (21.4%). Men showed higher prevalence estimates when using the cutoffs proposed by the Latin American Consortium of Studies in Obesity (LASO) (16.8%) and the IDF (22.8%). Reference curves showed that Quechua-Aymara individuals had lower WC and WHtR values compared to Afro-Peruvian and other groups at the 97th percentile, in both men and women. Interpretation: Reliance on BMI alone underestimates a large proportion of clinically relevant cases. Incorporating WC-measurements and clinical dysfunction into diagnostic frameworks could improve identification, prevention, and policy responses to obesity in Peru and similar settings. Funding: The authors received no financial support.