# Developing a Polygenic Risk Score for Weight Gain in People Treated for Psychosis—Application in a Real‐World Setting

## TL;DR

This study develops a genetic score to predict weight gain in women treated for psychosis, finding it explains 59% of the variation in daily weight gain.

## Contribution

A novel polygenic risk score for antipsychotic-induced weight gain in women with first-episode psychosis is developed and validated.

## Key findings

- The additive PRS significantly predicted weight gain in females (adjusted r² = 0.59, p < 0.05).
- High genetic risk groups showed greater weight gain per day (p = 0.018), with significant gender interactions.
- The PRS did not predict BMI percentage change in the study population.

## Abstract

Genetic factors are thought to play an important role in antipsychotic‐induced weight gain (AIWG). Polygenic risk scores (PRS) could provide a measure of genetic predisposition to antipsychotic drug induced weight gain (AIWG).We conducted a study to examine how a PRS, generated using SNPs, identified from a recent meta‐analysis, related to weight‐change over time in people with first episode‐psychosis.

The PRS included SNPs in six different genes, identified as having significant associations (p < 0.05) with AIWG. These were HTR2C rs3813929; MTHFR rs1801133; ADRA2A rs1800544; MC4R rs489693; LEPR rs1137101 and CNR1 rs1049353. An additive PRS and a risk allele based weighted PRS were created based on risk allele counts and presence or absence of risk alleles respectively. The additive PRS was also used to create low/high genetic risk groups for analysis. The association between PRS and weight gain per day (WGPD) in grams/day as well as BMI percentage change (=> 7%) was investigated using regression models.

In multiple regression analysis, the additive PRS significantly predicted AIWG in females (adjusted r
2 = 0.59, B: unstandardised regression coefficient = 24.4 g/day p < 0.05), but not in males. ANCOVA showed that high genetic risk groups had greater WGPD (p = 0.018), with significant PRS gender interactions driven by markedly higher WGPD in high‐risk females (p = 0.039). None of the models tested were associated with BMI percentage change.

We report a PRS that is predictive of weight gain in women treated for first episode psychosis, accounting for 59% of the variance daily weight‐gain over time. Validation in an independent cohort is required, as is determining whether it is feasible to apply the PRS prospectively.