Obesity as an Effect Modifier in Sleep-Disordered Breathing and Postcardiac Surgery Atrial Fibrillation.

Abstract

Because the interrelationships of objectively ascertained sleep-disordered breathing (SDB), postcardiac surgery atrial fibrillation (PCSAF), and obesity remain unclear, we aimed to further investigate the interrelationships in a clinic-based cohort. Patients with polysomnography and cardiac surgery (coronary artery bypass surgery and/or valvular surgery) within 3 years, from January 2009 to January 2014, were identified, excluding those with preexisting atrial fibrillation. Logistic models were used to determine the association of SDB (apnea hypopnea index [AHI] per 5-unit increase) and secondary predictors (central sleep apnea [CSA] [central apnea index ≥ 5] and oxygen desaturation index [ODI]) with PCSAF. Models were adjusted for age, sex, race, BMI, and hypertension. Statistical interaction and stratification by median BMI was performed. ORs and 95% CIs are presented. There were 190 patients who comprised the analytic sample (mean age, 60.6 ± 11.4 years; 36.1% women; 80% white; BMI, 33.3 ± 7.5 kg/m²; 93.2% had an AHI ≥ 5; 30% had PCSAF). Unlike unadjusted analyses (OR, 1.06; 95% CI, 1.01-1.1), in the adjusted model, increasing AHI was not significantly associated with increased odds of PCSAF (OR, 1.04; 95% CI, 0.98-1.1). Neither CSA nor ODI was associated with PCSAF. A significant interaction with median BMI was noted (P = .015). Effect modification by median BMI was observed; those with a higher BMI > 32 kg/m² had 15% increased odds of PCSAF (OR, 1.15; 95% CI, 1.05-1.26; P < .003). SDB was significantly associated with PCSAF in unadjusted analyses, but not after taking into account obesity; those with both SDB and obesity may represent a vulnerable subgroup to target to reduce PCSAF and its associated morbidity.