Abstract Introduction: The Homeostasis Model Assessment of insulin resistance (HOMA-IR) is a validated surrogate measure of insulin sensitivity and beta-cell function for epidemiological studies. Insulin resistance is closely associated with obesity and shares dysmetabolic state with other cardiometabolic risk factors. The aim of this study was to analyze the change in HOMA-IR levels in US youth from 1999 to 2018. Methods: Serial cross-sectional data of fasting insulin and glucose levels measured in youth aged 12 to 18 years enrolled in National Health and Nutrition Examination Surveys representative of US population from 1999 to 2018 were used. Two-year survey cycles were combined to derive 4-year survey periods, and NHANES recommended adjustments were made for glucose and insulin levels for forward conversion to 2017–18 instrument levels. Median HOMA-IR was compared across the survey periods in subgroups by age, gender, race/ethnicity and levels of obesity. Association of HOMA-IR with selected cardiometabolic risk factors was assessed. Change of HOMA-IR levels were analyzed using design adjusted multivariable regression models, association of HOMA-IR with cardiometabolic risk factors was assessed and percentiles were derived. Results: In an analysis of 5,541 youth, the unadjusted median HOMA-IR increased from 1.83 (95% CI 1.77–1.89) in 1999-02 to 2.41 (95% CI 2.29–2.54) in 2007–10 and has remained in this range till 2015–18. Youth with obesity (Body mass index > 95th percentile or waist circumference > 90th percentile), female gender, those of Hispanic and Black origin, and those aged 12–16 years had higher levels of HOMA-IR (p <.001). While adjusting for age category, sex, race/ethnicity, obesity represented by waist circumference z-score with its quadratic term and interaction with gender, HOMA-IR increased by 26% (95% CI 15–37) in 2015–18 as compared to 1999-02. For waist-circumference z-score (WCZ) > 1, the HOMA-IR levels were higher in boys compared to girls. The addition of total body fat from DXA scans in the model, when available, explained the variance related to gender, but not the interaction between gender and WCZ. No change was observed in the higher levels of HOMA-IR by ethnicity when high sensitivity CRP level as a measure of inflammation was added to the model. There was a positive association of HOMA-IR with triglycerides (29 %/unit, 95% CI 25–33), non-HDL cholesterol (8 %/unit, 95% CI 6–10), systolic blood pressure (2% /unit, 95% CI 1–2) and alanine transferase (11 %/unit, 95% CI 8–14) while controlling for age, sex, race/ethnicity, obesity and survey period. In youth with BMI between 5-85th percentile, HOMA-IR decreased with increase in age from 12 to 18 years. Conclusions: HOMA-IR levels in US youth have increased since 1999-02 and may herald future cardiometabolic risks.