Sleep Patterns, Plasma Metabolome, and Risk of Incident Type 2 Diabetes Mellitus

Abstract

Context: A healthy sleep pattern has been related to a lower risk of type 2 diabetes mellitus (T2DM).

Objective: We aimed to identify the metabolomic signature for the healthy sleep pattern and assess its potential causality with T2DM.

Methods: This study included 78 659 participants with complete phenotypic data (sleep information and metabolomic measurements) from the UK Biobank study. Elastic net regularized regression was applied to calculate a metabolomic signature reflecting overall sleep patterns. We also performed genome-wide association analysis of the metabolomic signature and one-sample mendelian randomization (MR) with T2DM risk.

Results: During a median of 8.8 years of follow-up, we documented 1489 incident T2DM cases. Compared with individuals who had an unhealthy sleep pattern, those with a healthy sleep pattern had a 49% lower risk of T2DM (multivariable-adjusted hazard ratio [HR], 0.51; 95% CI, 0.40-0.63). We further constructed a metabolomic signature using elastic net regularized regressions that comprised 153 metabolites, and robustly correlated with sleep pattern (r = 0.19; P = 3×10-325). In multivariable Cox regressions, the metabolomic signature showed a statistically significant inverse association with T2DM risk (HR per SD increment in the signature, 0.56; 95% CI, 0.52-0.60). Additionally, MR analyses indicated a significant causal relation between the genetically predicted metabolomic signature and incident T2DM (P for trend < .001).

Conclusion: In this large prospective study, we identified a metabolomic signature for the healthy sleep pattern, and such a signature showed a potential causality with T2DM risk independent of traditional risk factors.

Keywords: causality; genetics; metabolome; sleep; type 2 diabetes mellitus.