<b><i>Background:</i></b> Chronic obstructive pulmonary disease (COPD) and sleep apnea are common conditions and often coexist. The proper diagnosis of sleep apnea may affect the management and outcome of patients with COPD. <b><i>Objective:</i></b> To determine the accuracy of home nocturnal oximetry to distinguish between nocturnal oxygen desaturation related to COPD alone or to sleep apnea in patients with moderate-to-severe COPD who have significant nocturnal hypoxemia with cyclical changes in saturation. <b><i>Methods:</i></b> This study involved a comparison of home nocturnal oximetry and laboratory-based polysomnography (PSG) in patients with moderate-to-severe COPD considered for inclusion in a trial of nocturnal oxygen therapy. All of the patients had significant nocturnal oxygen desaturation (defined as ≥30% of the recording time with a transcutaneous arterial oxygen saturation &#x3c;90%) with cyclical changes in saturation suggestive of sleep apnea. <b><i>Results:</i></b> PSG was obtained in 90 patients; 45 patients (mean age = 68 years, SD = 8; 71% men; mean forced expiratory volume in 1 s [FEV₁] = 50.6% predicted value, SD = 18.6%; data from 41 patients) fulfilled the criteria for sleep apnea (mean apnea-hypopnea index = 32.6 events/h, SD = 19.9) and 45 patients (mean age = 69 years, SD = 8; 87% men; mean FEV₁ predicted value 44.6%, SD = 15%) did not (mean apnea-hypopnea index = 5.5 events/h, SD = 3.8). None of the patients’ characteristics (including demographic, anthropometric, and physiologic measures) predicted the diagnosis of sleep apnea according to PSG results. <b><i>Conclusion:</i></b> The diagnosis of sleep apnea in patients with moderate to severe COPD cannot rely on nocturnal oximetry alone, even when typical cyclical changes in saturation are seen on oximetry tracing. When suspecting an overlap syndrome, a full-night, in-laboratory PSG should be obtained.