WHR performed as the best predictor of type 2 diabetes with an optimal cutoff of >0.95 for both sexes, while BMI and BAI were the least accurate with AUCs less than 60% for both sexes.
Key Findings
Results
Fifteen percent of study participants had type 2 diabetes.
Total sample size was 10,663 participants
Participants were Iranian adults between 40 and 70 years from the Kharameh cohort study
Cross-sectional design using baseline data from the cohort
Diabetic group had significantly higher anthropometric indices than the non-diabetic group
Results
WHR was the best-performing anthropometric index for predicting type 2 diabetes.
Optimal cutoff point was WHR >0.95 for both sexes
WHR outperformed all other indices including BMI, WHtR, ABSI, BRI, and BAI
ROC curves were used to determine sensitivity and specificity of each index
The finding underscores the role of body fat distribution, particularly central or abdominal obesity, in predicting diabetes risk
Results
BMI and BAI were the least accurate anthropometric indices for predicting type 2 diabetes.
Both BMI and BAI had AUCs of less than 60% for both sexes
These two traditional and novel indices respectively showed poor discriminatory ability
Six indices total were evaluated: BMI, WHR, WHtR, ABSI, BRI, and BAI
Results
All anthropometric indices were significantly associated with an increased risk of type 2 diabetes.
Statistical analysis was performed using independent t-tests and logistic regression
Associations were examined separately in men and women
Anthropometric indices were significantly higher in the diabetic group than in the non-diabetic group
Gender was examined as a modifier of the association between anthropometric indices and diabetes
Results
Novel anthropometric indices did not consistently outperform traditional measures for diabetes prediction in this population.
Novel indices evaluated included ABSI, BRI, and BAI
Traditional indices evaluated included BMI and WHR
WHR, a traditional index, performed best overall
BAI, a novel index, performed among the worst with AUC <60% for both sexes
The study aimed to evaluate predictive power of novel indices compared with traditional ones
Baberi F, Rezaeianzadeh S, Rezaianzadeh A, Hamedi A. (2026). Novel anthropometric indices for predicting diabetes mellitus: A population-based study.. Primary care diabetes. https://doi.org/10.1016/j.pcd.2025.10.006