Tobacco smoke exposure and sleep problems: A secondary dataset analysis of NHANES and Mendelian randomization analyses.

Abstract

Introduction: Smoking and sleep problems are both major public health concerns, but the causal nature of their association remains unclear due to confounding and reverse causality. To address this, observational and Mendelian randomization approaches were combined to investigate the causal relationships between smoking status, cotinine levels, and sleep deprivation and sleep disorder.

Methods: Data from NHANES (2005-2020) were analyzed using weighted logistic regression, subgroup, and threshold analyses to assess associations between smoking status, cotinine levels, and sleep deprivation and sleep disorder. Mendelian randomization was then performed using genetic instruments for smoking phenotypes to evaluate causal relationships.

Results: In NHANES (n=22380), current smokers showed significantly higher risk of sleep disorder (OR=1.68; 95% CI: 1.49-1.90) and sleep deprivation (OR=1.52; 95% CI: 1.36-1.70) compared to never smokers. Serum cotinine was positively associated with both outcomes in a nonlinear manner, with threshold effects observed. In Mendelian randomization, genetic predisposition to current smoking was causally associated with increased risk of sleep deprivation (OR=1.253; 95% CI: 1.015-1.548), while never smoking showed protective causal effects on both outcomes. However, no causal associations were found between cotinine and sleep outcomes.

Conclusions: Observational analyses revealed associations between smoking status, serum cotinine levels, and both insufficient sleep and sleep disorders. However, while Mendelian randomization provided evidence supporting a causal relationship between smoking status and sleep outcomes, it did not establish a causal link between serum cotinine and either insufficient sleep or sleep disorders.