Multidimensional Pancreatic Islet β-cell Function (PIF) Assessment Improves Predictive Effect of Diabetes Risk Scores

Abstract Aims/hypothesis Comprehensive assessment of pancreatic islet β-cell function (PIF) is crucial for diabetes management. We proposed a multidimensional, relative quantification system for PIF measurement. Methods Our novel approach evaluates PIF using three dimensions: stationary-baseline (PIF-S), load-peak (PIF-L), and accelerated-slope (PIF-A). The system was evaluated in 814 JR Cohort volunteers (195 metabolically healthy, 619 abnormal), 12 Botnia clamp study participants, 3394 type 2 diabetes patients, and 6345 METSIM cohort study participants. Restricted Cubic Spline (RCS) modeling determined ideal values based on human physiological parameters. Each subject's actual values were compared with predicted ideals and converted into percentile indices. Results The Botnia clamp experiment confirmed distinct meaning of three PIF indices. Cluster analysis in metabolically abnormal individuals identified three clusters. Cluster 1, with the highest PIF-A, had the best metabolic profiles and lowest cardiovascular and renal disease risks. Cluster 3, with the highest PIF-S and PIF-L but lowest PIF-A, had the poorest metabolic profiles and highest disease risks. Type 2 diabetes patients with high PIF-S and PIF-L were more prone to complications. Similar patterns were observed in the METSIM cohort, Cluster 1 showing the lowest diabetes risk, with hazard ratios for Clusters 2 and 3 at 2.499 (95% CI 1.932-3.233, P = 3.11E-12) and 3.185 (95% CI 2.353-4.311, P = 6.35E-12), respectively. The novel three-dimensional PIF indices surpass previous indicators in predicting diabetes. Combined with existing diabetes risk scores, novel PIFs also significantly improved their predictive efficiency. Conclusions This novel system offers an effective method for PIF assessment, enhancing diabetes prediction and management by deepening the understanding of diabetes complexity and aiding in precise therapy.