Sleep duration and auditory hallucinations: Genetic correlation and two-sample Mendelian randomization study.

OBJECTIVES

Previous studies have found that sleep problems are associated with psychotic experiences (PEs) across clinical designs, but many confounding factors from uncertain variables have been unclear in observational designs. The aim of the present study was to detect the genetic correlations and causal relationship between sleep-related traits and PEs using the largest current genome-wide association study (GWAS) summary statistics.

METHODS

GWAS results were obtained for positive PEs (N = 116,787-117,794) and sleep-related traits [insomnia complaints (N = 386,533), morningness (N = 345,552), sleep duration (N = 384,317), ease of getting up in the morning (N = 385,949), daytime napping (N = 386,577), daytime sleepiness (N = 386,548), and snoring (N = 359,916)]. Linkage disequilibrium score regression (LDSC) was used to investigate genetic correlations. Mendelian randomization (MR) was conducted on trait pairs with significant genetic associations.

RESULTS

We found that auditory hallucinations were significantly genetically correlated with insomnia complaints (rg = -0.27, p = 1.1 × 10-3), sleep duration (rg = 0.21, p = 9.7 × 10-3), and ease of getting up (rg = 0.31, p = 2 × 10-4). Visual hallucinations and insomnia complaints were highly genetically correlated (rg = 0.36, p = 6.4 × 10-5). Mendelian randomization indicated a unidirectional causal relationship between sleep duration with auditory hallucinations (β = -0.93, p = 7.9 × 10-4 for sleep duration as the exposure). We used large GWAS summary statistics across the LDSC and MR programs to determine that sleep duration as exposure to increase the risk of auditory hallucinations.

CONCLUSIONS

Taken together, these findings suggest that treatment of sleep problems should be considered as a higher priority for future mental health services.