Study objectives: To test the clinical observations that patients with chronic obstructive pulmonary disease (COPD) have impaired sleep quality without excessive daytime sleepiness (EDS), and to analyse the aetiological factors.Participants: Fifteen non-diabetic postmenopausal women with moderate to severe COPD and 20 community dwelling age-matched control women.Measurements and results: Patients completed questionnaires, had a polysomnography and blood tests. Controls filled in the questionnaires. In the Basic Nordic Sleep Questionnaire, the average (±sd) scores for sleepiness (9.9±3.0 in patients vs. 7.6±3.2 in controls, P=0.025, test range 4–20) and insomnia (18.3±3.4 vs. 16.6±4.4, P=123, test range 7–35) were low. Although 53% had a good night's sleep seldom or never and 70% slept restlessly, only 33% felt tired in the mornings. Controls reported better sleep quality, less tiredness and sleepiness. With polysomnography, the total sleep time was 4h 41min ±1h 20min in patients. Sleep was fragmented, the proportion of stage 1 sleep high and rapid eye movement (REM) latency delayed. Sleepiness correlated with fasting serum insulin levels (r=0.59, P=0.027) and body movements (r=0.52, P=0.047). In stepwise linear regression analyses, sleepiness was positively associated with insulin levels (P=0.025) but not with body movements. Insulin explained 38.0% of the variance in the sleepiness score, when adjusted for body mass index (BMI).Conclusions: Despite short and fragmented sleep, non-diabetic patients with COPD did not have marked EDS. An association between fasting insulin and sleepiness suggests that insulin resistance is involved in EDS.