OR08-05 Diagnostic Tests to Discriminate Selective and Global Insulin Resistance

Abstract M. Brush: None. S. Auh: None. A.L. Ramanathan: None. A. Prakash: None. M. Walter: None. R. Muniyappa: None. R.K. Semple: Consulting Fee; Self; AstraZeneca, Alynylam. Grant Recipient; Self; Wellcome Trust. Speaker; Self; Novartis Pharmaceuticals, Eli Lilly & Company, Amryt. R.J. Brown: None. Severe insulin resistance (IR) can be classified into selective and non-selective IR. In selective IR, glucoregulatory insulin signaling is attenuated in only some insulin-responsive tissues and/or some signal transduction pathways. This causes compensatory hyperinsulinemia, which increases signaling in the relatively unimpaired tissues or pathways. An important example is hepatic lipogenesis, which is overstimulated in “common” obesity-related IR. Selective IR is also seen in the rare condition of lipodystrophy, in which patients lack some or all subcutaneous fat. In contrast, in global, or non-selective IR (usually caused by insulin receptor pathogenic variants), all insulin signal transduction is blocked at the receptor. Impaired receptor signaling affects all pathways and leads to low triglycerides with hyperglycemia. Despite these biochemical differences, these conditions have similar clinical presentations and can require genetic testing to reach a definitive diagnosis. We created accurate and accessible diagnostic tests using IR biomarkers to aid clinical discrimination of selective vs non-selective insulin resistance. We assessed fasting serum insulin (mcU/mL), HDL (mg/dL), triglycerides (mg/dL), adiponectin (ng/mL), and uric acid (mg/dL) from 182 patients with rare and common forms of non-selective (n=59) and selective IR (n=123). Patients with selective IR included those with a diagnosis of partial (n=26) and generalized lipodystrophy (n=14), and obesity (n=83). Patients with non-selective IR included those with pathogenic variants of the insulin receptor gene (n=40) or autoantibodies to the insulin receptor (n=19). The Receiver Operating Characteristic (ROC) area under the curve (AUC), sensitivity, and specificity of each combination of biomarkers were assessed in development and validation cohorts. We identified two combinations of biomarkers with high sensitivity and specificity. Test one, (insulin*HDL)÷(triglycerides*uric acid) had an AUC of 0.988 in the development cohort and 0.997 in the validation cohort. A cut-off of ≥8.3 predicted non-selective IR with a sensitivity of 100% in both the development and validation cohorts and a specificity of 94% and 89% in the development and validation cohorts respectively. Test two, (insulin*HDL*adiponectin)÷ (triglycerides*uric acid), had an AUC of 0.983 in the development cohort and 1.0 in the validation cohort. A cut-off of ≥51682 predicted non-selective IR in the validation cohort with 93% and 100% sensitivity in the development and validation cohorts respectively and 94% specificity in both cohorts. Both tests use easily measured biomarkers to discriminate selective and non-selective IR with high diagnostic accuracy. These accurate and accessible tests could be used in clinical settings to aid diagnostic classification of severe IR. Saturday, June 1, 2024