We retrospectively evaluated data from 213 consecutive patients; 152 were affected by obstructive sleep apnea (OSA), 29 had OSA associated with chronic obstructive pulmonary disease (COPD), also known as overlap syndrome, and 32 had COPD. Patients with obesity-hypoventilation syndrome were not included. The aims of the study were to evaluate the anthropometric, pulmonary, and polysomnographic characteristics of patients affected by overlap syndrome compared to "simple" OSA and to COPD subjects and to analyze the determinants of hypercapnia in overlap syndrome. In the comparison between overlap and OSA patients, the overlap group had a significantly higher PaCO2 (44.59 vs. 39.22 mm Hg; p < 0.01), in the presence of a similar AHI (40.46 vs. 41.59/h). Comparing overlap to COPD patients, overlap showed a significantly higher PaCO2 value (44.59 vs. 39.63 mm Hg; p < 0.005) and had significantly less severe obstructive impairment (FEV 162.93 vs. 47.31%; FEV1/FVC ratio 66.71 vs. 59.25%; p < 0.005). Anthropometric, pulmonary function, and polysomnographic data did not differ between normo- and hypercapnic overlap patients. The best model (stepwise multiple regression analysis) for predicting PaCO2 in overlap patients showed r2 value 0.65: PaO2 contributed to 38%, FEV1 to 15%, and weight to 12%. In conclusion, the occurrence of hypercapnia in overlap patients is only partially explained by the combination of overweight and reduced respiratory function, supporting the hypothesis of a multifactorial genesis.