Association of Cumulative Smoking Exposure with REM Sleep Alterations in Obstructive Sleep Apnea: A Cross-Sectional...

Objective: The association of smoking with sleep apnea is often based on subjective data. This study quantified the effects of smoking on sleep architecture using exhaled carbon monoxide (eCO) and polysomnography (PSG). Methods: A total of 183 patients with suspected obstructive sleep apnea (OSA) were included in this prospective study. Following full-night PSG, eCO was measured within 10 min. Data were analyzed using the Generalized Linear Model (GLM). Results: Although initial unadjusted analyses showed an inverse correlation between eCO levels and central apnea count, GLM revealed that male gender was the only independent predictor for central apnea, negating the effect of eCO. GLM analyses, adjusted for age, gender, BMI, and alcohol and drug use, revealed that cumulative smoking load (pack-years) was independently associated after multivariable adjustment with reduced REM sleep duration (B = -0.345, 95% CI [-0.571; -0.119], p = 0.003) and REM sleep percentage (B = -0.099, 95% CI [-0.158; -0.040], p = 0.001). Similarly, smoking duration (years) significantly predicted decreased REM sleep duration (B = -0.426, 95% CI [-0.724; -0.128], p = 0.005) and REM percentage (B = -0.119, 95% CI [-0.197; -0.041], p = 0.003). Formal interaction analyses did not detect a statistically significant interaction with body mass index (BMI) (p > 0.05 for all interaction terms). Conclusions: In OSA, smoking is independently associated with alterations in REM sleep architecture rather than respiratory events. Cumulative smoking load and smoking duration are independently associated with alterations in REM sleep after adjusting for any other major clinical comorbidities.