Biomarkers of insulin resistance and their performance as predictors of treatment response in overweight adults

Insulin Resistance (IR) contributes to the pathogenesis of type 2 diabetes mellitus (T2DM), and is a risk factor for cardiovascular and neurodegenerative diseases. Amino acid and lipid metabolomic biomarkers associate with future T2DM risk in epidemiological cohorts. Whether these biomarkers can accurately detect changes in IR status following treatment is unclear. Herein we evaluated the performance of clinical and metabolomic biomarkers to predict altered IR following lifestyle-based interventions. First, we evaluated the performance of two distinct insulin assay types (high-sensitivity ELISA and Immunoassay) and built cross-sectional clinical and metabolomic IR diagnostic models. These were utilised to stratify IR status in pre-intervention fasting samples, from three independent cohorts (META-PREDICT (MP, n=179), STRRIDE-AT/RT (S-2, n=116) and STRRIDE-PD (S-PD, n=149)). Linear and Bayesian projective prediction strategies were used to evaluate biomarkers for fasting insulin and HOMA2-IR and change in fasting insulin with treatment. Both insulin assays accurately quantified international standard insulin (R 2 >0.99), yet agreement for fasting insulin was less congruent (R 2 =0.65). A mean treatment effect on fasting insulin was only detectable using an ELISA. Clinical-metabolomic models were statistically related to fasting insulin (R 2 0.33-0.39) with modest capacity to classify IR at a clinically relevant HOMA2-IR threshold. Furthermore, no model predicted treatment responses in any cohort. Thus, we demonstrate that the choice of insulin assay is critical when quantifying the influence of treatment on fasting insulin, while none of the clinical-metabolomic biomarkers, established in cross-sectional data, are suitable for monitoring longitudinally changes in IR status.