Determinants of Hypercapnia in Obese Patients With Obstructive Sleep Apnea : A Systematic Review and Metaanalysis of Cohort Studies

BackgroundInconsistent information exists about factors associated with daytime hypercapnia in obese patients with obstructive sleep apnea (OSA). We systematically evaluated these factors in this population.MethodsWe included studies evaluating the association between clinical and physiologic variables and daytime hypercapnia (Paco2, ≥ 45 mm Hg) in obese patients (body mass index [BMI], ≥ 30 kg/m2) with OSA (apnea-hypopnea index [AHI], ≥ 5) and with a < 15% prevalence of COPD. Two investigators conducted independent literature searches using Medline, Web of Science, and Scopus until July 31, 2008. The association between individual factors and hypercapnia was expressed as the mean difference (MD). Random effects models were used to account for heterogeneity.ResultsFifteen studies (n = 4,250) fulfilled the selection criteria. Daytime hypercapnia was present in 788 patients (19%). Age and gender were not associated with hypercapnia. Patients with hypercapnia had higher BMI (MD, 3.1 kg/m2; 95% confidence interval [CI], 1.9 to 4.4) and AHI (MD, 12.5; 95% CI, 6.6 to 18.4) than eucapnic patients. Patients with hypercapnia had lower percent predicted FEV1 (MD, −11.2; 95% CI, −15.7 to −6.8), lower percent predicted vital capacity (MD, −8.1; 95% CI, −11.3 to −4.9), and lower percent predicted total lung capacity (MD, −6.4; 95% CI, −10.0 to −2.7). FEV1/FVC percent predicted was not different between hypercapnic and eucapnic patients (MD, −1.7; 95% CI, −4.1 to 0.8), but mean overnight pulse oximetric saturation was significantly lower in hypercapnic patients (MD, −4.9; 95% CI, −7.0 to −2.7).ConclusionsIn obese patients with OSA and mostly without COPD, daytime hypercapnia was associated with severity of OSA, higher BMI levels, and degree of restrictive chest wall mechanics. A high index of suspicion should be maintained in patients with these factors, as early recognition and appropriate treatment can improve outcomes.