The study examines the best cutoff points of the Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) to identify persons at risk of metabolic syndrome (MetS) among a Mexican adult population. Because there is current uncertainty over the thresholds, the study compares two approaches: percentile distribution and receiver operating characteristic (ROC) curve analysis.
Methodology:
Cross-sectional analysis was performed in 765 adults recruited from a hospital outpatient clinic in Mexico City. The participants were stratified into two groups: a reference group (normal weight and normal fasting glucose) and a MetS group (overweight/obese subjects stratified by the presence or absence of MetS). HOMA-IR values were compared using the 75th percentile in the reference group and ROC curve analysis in the MetS group. Optimal cutoffs were estimated using the Youden index.
Findings:
The reference group consisted of 218 participants and MetS group consisted of 547 participants. The 75th percentile and 90th percentile HOMA-IR were 2.72 and 3.71, respectively. ROC analysis indicated higher cutoff values in the diagnosis of MetS, and the most appropriate thresholds were 4.21 (AHA/NHLBI definition) and 4.23 (WHO definition). The percentile approach allowed the identification of participants at risk at an earlier point before the development of clinical features.
Conclusion:
Results identify heterogeneity in HOMA-IR cutoff points by methodology and the requirement of population-specific reference points. The percentile-based approach is valuable for early detection of insulin resistance, helpful for preventive treatment in the preclinical phase. The present study once again emphasizes the value of customized cutoff points in risk stratification and clinical decision-making.
