Latent class analysis applied to 1,034 participants (RESET cohort) using 14 cardiometabolic variables—six conventional metabolic syndrome (MetS) measures plus eight additional markers—identified five distinct phenogroups. Two are clinically counterintuitive: Lean-Insulin Resistant (Lean-IR, 13.5%) and Obese-Insulin Sensitive (Obese-IS, 20.4%), both showing low MetS prevalence (21–31%) yet carrying hazard ratios of 1.86 and 1.85 respectively for myocardial infarction or stroke in 344,817 UK Biobank participants. The highest-risk Obese-IR group (75.9% MetS prevalence) showed an HR of 2.75. A four-variable decision tree—visceral adiposity, insulin resistance, systolic blood pressure, and HbA1c—replicated phenogroup assignments across three external cohorts spanning Asian and European populations.
This finding challenges a foundational assumption of cardiometabolic medicine: that metabolic syndrome status adequately stratifies cardiovascular risk. Lean individuals with insulin resistance—often dismissed clinically because their BMI appears normal—carried nearly identical event risk to obese-insulin-sensitive patients, and meaningfully higher risk than those with isolated hypertension. The liver fat data (9.0% vs. 2.8% in Lean-IR vs. Metabolically Preserved) suggests ectopic fat deposition as a plausible mechanism underlying this hidden risk. Proteomic signatures further differentiate groups, opening biomarker discovery opportunities. Limitations include the observational design precluding causal inference, modest discovery cohort size, and potential residual confounding. As a preprint not yet peer-reviewed, these findings warrant scrutiny before influencing clinical guidelines—but the large UK Biobank validation substantially strengthens credibility. If confirmed, routine insulin resistance and visceral fat screening could meaningfully improve cardiovascular risk stratification beyond current BMI-centric or MetS frameworks.
