Reference intervals for fasting insulin and insulin-related indices in healthy adults: a cross-sectional study in Gandaki Province, Nepal

Background Accurate assessment of insulin resistance, sensitivity and β-cell function is essential for early detection and management of metabolic disorders. However, reference intervals (RIs) commonly used in Nepal have been adapted from Western populations, which may not accurately reflect local physiological characteristics. Thus, this study aimed to establish population-specific RIs for fasting insulin and key insulin-related indices using a direct priori method in healthy adults from Gandaki Province, Nepal. Method This cross-sectional study recruited 135 healthy adults (20–69 years, body mass index 18.5–24.9 kg/m²) representing different districts of Gandaki Province, Nepal. Fasting blood samples were analysed for glucose, insulin and lipids using standardised assays. Insulin was measured using the chemiluminescence immunoassay method. Nineteen different insulin-derived indices (Homeostasis Model Assessment 1 of Insulin Resistance (HOMA1-IR), Homeostasis Model Assessment 2 of Insulin Resistance (HOMA2-IR), Homeostasis Model Assessment for Triglycerides, Fasting Insulin to Glucose Ratio, Fasting Insulin Resistance Index, Metabolic Score for Insulin Resistance (METS-IR), InsuTAG, HOMA1-%S, HOMA2-%S, Quantitative Insulin Sensitivity Check Index (QUICKI), McAuley, Bennett, Raynaud, Glucose-to-Insulin Ratio, Fasting Insulin Sensitivity Index, Single Point Insulin Sensitivity Estimator (SPISE), reciprocal insulin, HOMA1-%B and HOMA2-%B) were calculated. Non-parametric 95% double-sided RIs (2.5th–97.5th percentiles) were established following outlier removal per Clinical and Laboratory Standards Institute-International Federation of Clinical Chemistry and Laboratory Medicine EP28-A3c guidelines. Results The RI for fasting insulin was 2.63–14.56 µIU/mL (median 7.69 µIU/mL). Among the 19 mathematically correlated insulin-derived indices which are calculated from core measurements (fasting serum insulin and glucose), consistent patterns emerged across functional categories. Insulin resistance indices (HOMA1-IR: 0.56–3.50; HOMA2-IR: 0.30–1.70; METS-IR: 25.14–38.94) exhibited concordant right-skewed distributions with elevated upper limits. Conversely, insulin sensitivity indices (QUICKI: 0.32–0.42; HOMA2-%S: 58.83–233.20; SPISE: 5.75–10.86) demonstrated inverse, left-skewed patterns. Beta-cell function indices (HOMA1-%B: 0.54–322.21; HOMA2-%β: 40.74–159.52) also exhibited right skewed characteristics and revealed wide interindividual variability, reflecting preserved pancreatic reserve despite varying insulin resistance. Composite indices incorporating lipid parameters showed broader ranges, capturing additional metabolic heterogeneity. Conclusion This is the first study to define the RIs of fasting insulin and a spectrum of insulin derived indices in a Nepalese population. These findings offer a valuable framework for early detection and management of metabolic disorders in South Asian populations.