# Genomic risk prediction of type 2 diabetes in people living with and without HIV

## TL;DR

This study evaluates how well genetic risk scores predict type 2 diabetes in people with and without HIV, finding similar performance across groups.

## Contribution

The study demonstrates the transferability of general population-derived genetic scores to people living with HIV.

## Key findings

- MetaPRS models showed non-significant improvements in T2D risk prediction compared to single-trait PRS and clinical factors.
- Performance of genetic risk scores was similar in people with and without HIV.
- The study highlights the need for refining PRS models in diverse populations and exploring HIV-specific genetic factors.

## Abstract

Type 2 diabetes (T2D) risk prediction remains a challenge, particularly in underrepresented populations, including people living with HIV (PWH) and those of non-European ancestry. We evaluated the performance of two metaPRS (polygenic risk score) models, integrating genetic markers related to inflammation and lipid metabolism, in predicting T2D risk across ancestry groups (African and European), with and without HIV. The metaPRS were generated in a subset from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study (6,034 Black; 11,972 White) and validated in 7,580 (4,120 Black; 3,460 White) PWH from the Centers for AIDS Research of Integrated Clinical Systems (CNICS), as well as an additional 4,152 (2,586 Black; 1,566 White) seronegative participants from REGARDS. Incorporating the metaPRS into models provided non-significant improvements in T2D risk prediction compared to single-trait T2D PRS and clinical risk factors. Performance was similar in PWH and in people without HIV, suggesting that these general population-derived genetic scores are transferable to PWH. Future studies should focus on refining PRS models in diverse populations and exploring genetic factors specific to PWH regarding T2D risk.

The online version contains supplementary material available at 10.1038/s41598-025-31471-7.