A clinical prediction model for beta cell monogenetic diabetes in Chinese patients with early-onset type 2 diabetes

Abstract

Aim

Monogenic diabetes is a group of disorders arising from single gene mutations with a clear pathophysiology, most of which present with impaired beta cell function rather than insulin resistance. This study aims to evaluate the ability of TyG index and polygenetic risk score (PRS) to identify multi-type beta cell monogenetic diabetes (beta-cell-MgD) in Chinese early-onset type 2 diabetes (EOD) population.

Methods

A prediction model for beta-cell-MgD was established by logistic regression analysis in Cohort 1 (92 beta-cell-MgD, 512 EOD). Model performance was evaluated by receiver operating characteristic curves (ROC) and validated in an independent case-control sample (Cohort 2, 35 beta-cell-MgD, 50 EOD) and a newly diagnosed drug-naive EOD cohort (Cohort 3, 7 beta-cell-MgD, 176 EOD). PRS was constructed based on Genome-wide genotyping data from participants in Cohort 3. The ability of PRS to identify beta-cell-MgD was tested by ROC.

Results

The TyG-MgD score based on age at diagnosis, BMI and TyG presented a good performance to distinguish beta-cell-MgD (AUC=0.769), and achieving AUCs of 0.966 and 0.754 respectively in validation cohorts. At the optimal cutoff point -16.19, the model achieved a sensitivity of 66.3% and a specificity of 75.39%, allowing one case of beta-cell-MgD identified among every three patients. -16.85 could be used as the screening threshold prioritizing 80% sensitivity (with 59% specificity). Models combining TyG-MgD with East Asian PRS and beta-cell dysfunction-high proinsulin partitioned polygenetic score showed AUCs of 0.842 and 0.834 respectively for indentifying beta-cell-MgD.

Conclusion

We developed a clinical prediction model as a simple screening tool for multi-type beta-cell-MgD, identifying who are most likely to benefit from next genetic sequencing in Chinese population. PRS might be helpful for further screening of MgD.