A Genetic Risk Approach to Differential Diagnosis of Type 1 Diabetes and Risk Stratification in the Spanish Population.

CONTEXT: Accurate type 1 diabetes (T1D) risk assessment can improve diagnosis and preventive care, reducing misdiagnosis and enabling preclinical interventions. Genetic Risk Scores (GRS) offer a promising tool, but population-specific validation is required.Objective. To assess the performance of established vs. a novel Spanish-specific GRS for T1D differential diagnosis and risk stratification.

METHODS

From 2,338 genotyped individuals, 2,215 Spanish participants of European ancestry passed quality control: 816 T1D, 469 type 2 diabetes (T2D), 457 monogenic diabetes (MD), and 473 controls. A novel Spanish-specific GRS (T1D-SP-GRS) was developed and compared with HLA-DRB1 and two European-derived models (T1D-GRS and T1D-GRS2). The best-performing score was evaluated for T1D discrimination vs. T2D and MD, and for association with T1D-onset age.

RESULTS

T1D-GRS2 outperformed T1D-SP-GRS for T1D vs. controls (AUC-ROC: 0.888 [95% CI: 0.867-0.908] vs. 0.837 [0.797-0.877], p=0.001), retaining high discrimination vs. T2D and MD (0.886 [0.865-0.907]). In the combined pediatric-onset T1D and control cohort, the P50 cutoff yielded a sensitivity of 0.80 (95% CI: 0.76-0.84) and specificity of 0.81 (0.77-0.84); P25 prioritized sensitivity (0.97 [0.95-0.99]) while P75 specificity (0.96 [0.93-0.97]). A P90 threshold conferred 26-fold higher T1D odds vs. controls, and 18- and 17-fold vs. T2D and MD (p <0.001). Higher T1D-GRS2 in childhood- vs. adulthood-onset T1D revealed an age-dependent risk profile primarily driven by HLA.

CONCLUSIONS

T1D-GRS2 discriminates T1D from other diabetes etiologies, providing a complementary tool to support diagnostic accuracy. The higher genetic risk associated with a younger age at onset provides key insights for early-stage T1D screening and intervention strategies.