Leptin as a predictor of anthropometric cutoff points for obesity

This study was conducted to find the association between leptin and adiposity indices. Secondly, to identify optimal threshold of various anthropometric indices for obesity, as assessed by 75th percentile of leptin levels, within a clinic sample of non-diabetic and diabetic Pakistani adults. Fasting serum leptin levels were compared with anthropometric markers of obesity in 164 diabetic and non-diabetic subjects (90 male, 74 female), aged 35 to 65 years. Obesity was defined by body mass index (BMI) of 25 kg/m2 in either sex. The cutoff point of leptin was taken as the 75th percentile in non-obese subjects. Diagnostic accuracy for detecting excess fatness was evaluated through receiver operating characteristics (ROC) analyses with leptin taken as reference test against anthropometric indices as test variables. The 75th percentile of leptin in male and female was 7.0ng/mL and 17.9ng/mL, respectively. Leptin levels were significantly higher in females (p<0.001) and had strong positive correlation (p<0.001) with most anthropometric indices of obesity in both sexes; hip circumference (HC) being most prominent among these. Largest area under ROC curve (AUC) was between WC and leptin (AUC=0.844; CI=0.764, 0.925) in males and BMI and leptin (AUC=0.832; CI=0.740, 0.923) in females. The optimum thresholds for obesity indices in our study were: BMI, WC and HC as 25 kg/m2, 96.25cm, 99.25cm for males; 27 kg/m², 95.50cm, 105.5cm for females, respectively. Leptin can be considered as a potential marker of obesity and may be used to identify obesity cutoffs in future demographic surveys. Longitudinal studies are required that include leptin in coronary artery disease risk assessment models.