BMI-based obesity cutoffs and excess adiposity in a Caribbean adolescent population of African origin

To evaluate the diagnostic utility of recommended body mass index (BMI)-based classification systems in detecting excess body fat (BF) in an Afro-Caribbean adolescent population.All adolescents aged 12-18 years were invited to participate in the study. A total of 3749 persons participated fully in the study.Males (12.2%) (95% confidence interval (CI): 10.7, 13.9) and females (43.6%) (95% CI: 41.5, 45.7) had excess adiposity. Specificities were high for all cutoff values (range 90.1-99.7%). Sensitivities ranged from 25.7 to 86.4% and from 15.6 to 54.4% for Centre for Disease Control (CDC) and International Obesity Task Force (IOTF) cutoff values, respectively. The mean areas under the receiver-operating characteristic curves ranged from 62.7 to 88.2% in males and 57.7 to 75.1% in females, while positive likelihood ratios ranged from 8.7 to 90.9 for the recommended BMI cutoff values.Overall, the CDC cutoffs had a higher sensitivity than the IOTF cutoffs in screening for excess fat. Specificities were high for cutoff values in both systems and would produce fewer false negatives.