Data-driven Cluster Analysis Reveals Increased Risk for Severe Insulin-deficient Diabetes in Black/African Americans

Area of Science:

• Endocrinology and Metabolism
• Population Health
• Genetics and Genomics

Background:

• Diabetes mellitus is a complex, heterogeneous disease with emerging subtypes.
• Previous research has understudied the implications of these subtypes across diverse populations.
• Understanding diabetes heterogeneity is crucial for addressing health disparities.

Purpose of the Study:

• To apply cluster analysis to a diverse diabetes cohort in the US Deep South.
• To investigate the distribution of diabetes subtypes based on key clinical parameters.
• To examine the influence of racial background on diabetes subtype prevalence and characteristics.

Main Methods:

• Retrospective hierarchical cluster analysis of electronic health records from 89,875 patients.
• Inclusion of adult patients with diabetes codes and data for glutamic acid decarboxylase autoantibody, hemoglobin A1c, body mass index, diagnosis age, HOMA2-B, and HOMA2-IR.
• Analysis of a cohort with approximately 42% Black/African American representation.

Main Results:

• Significant differences in diabetes cluster distribution were observed between Black/African Americans and Whites (P < .001).
• Black/African Americans exhibited a higher likelihood of severe insulin-deficient diabetes (OR, 1.83; P < .001) with increased metabolic perturbations and complication risks (OR, 1.42; P = .020).
• Black/African Americans showed more severe impairment of beta-cell function (P < .001) without increased obesity or insulin resistance.

Conclusions:

• Racial background significantly influences diabetes cluster distribution.
• Black/African Americans are disproportionately affected by severe insulin-deficient diabetes, suggesting underlying biological or environmental factors.
• These findings may help explain outcome disparities and inform tailored treatment strategies for diverse diabetes populations.