We conducted a historical cohort study to examine the relationship between survival and use of inhaled corticosteroids (ICS) and/or long-acting beta agonists (LABA) in patients with chronic obstructive pulmonary disease (COPD). All COPD patients aged 40 years who were enrolled in one of two regional managed care organizations during 1995-2000, and who had 90 days use of an ICS and/or LABA (N=1288) or of a short-acting bronchodilator (N=397), were identified. Of patients treated with ICS and/or LABA, 14.4% died during the follow-up period, as compared to 28.2% of comparison patients (P<0.01). In a Cox proportional hazards model that controlled for age, sex, comorbidities, COPD severity, and asthma status, a reduced risk of death was found for ICS treatment (HR 0.59 [95% CI 0.46-0.78]), LABA (HR 0.55 [0.34-0.89]), and ICS plus LABA treatment (HR 0.34 [0.21-0.56]). A second model that excluded any patient who also had an ICD-9 code for asthma (N=840) still found improved survival among those using the combination of ICS plus LABA (HR 0.35 [CI 0.17-0.71]). Additional analyses that varied the exposure criteria also found a consistent treatment benefit. Inclusion of ICS or bronchodilator treatment during the follow-up period as a time-dependent function appears to negate the survival benefit; however, the underlying assumptions for valid time-dependent modeling are clearly violated in this situation. In conclusion, we found that COPD patients who used ICS alone or in combination with LABA had substantially improved survival even after adjustment for asthma and other confounding factors.