The effect of mild hypercapnia on hospital mortality after cardiac arrest may be modified by chronic obstructive pulmonary disease

Background: The main objective was to evaluate the effect of carbon dioxide on hospital mortality in chronic obstructive pulmonary disease (COPD) and non-COPD patients with out-of-hospital cardiac arrest (OHCA). Methods: We conducted a retrospective observational study in OHCA patients from the eICU database (eicu-crd. mit.edu). The main exposure was the partial pressure of arterial carbon dioxide (PaCO2). The proportion of time spent (PTS) within four predefined PaCO2 ranges (hypocapnia: <35 mmHg, normocapnia: 35-45 mmHg, mild hypercapnia: 46-55 mmHg, and severe hypercapnia: >55 mmHg) were calculated respectively. The primary outcome was hospital mortality. Multivariable logistic regression models were performed to assess the independent relationship between PTS within PaCO2 range and hospital mortality, and the interaction between PTS within PaCO2 range and COPD was explored. Results: A total of 1721 OHCA patients were included, of which 272 (15.8%) had COPD. After adjusted for the confounders, the PTS within mild hypercapnia was associated with lower odds ratio for hospital mortality in COPD patients (OR 0.923; 95% CI 0.857-0.992; P = 0.036); however, it was associated with higher odds ratio for hospital mortality in non-COPD patients (OR 1.053; 95% CI 1.012-1.097; P = 0.012; Pinteraction = 0.008). The PTS within normocapnia was not associated with hospital mortality in COPD patients (OR 0.987; 95% CI 0.914-1.067; P = 0.739); however, it was associated with lower odds ratio for hospital mortality in non-COPD patients (OR 0.944; 95% CI 0.916-0.973; P < 0.001; Pinteraction = 0.113). Conclusions: The effect of carbon dioxide on hospital mortality differed between COPD and non-COPD patients. Mild hypercapnia was associated with increased hospital mortality for non-COPD patients but reduced hospital mortality for COPD patients. It would be reasonable to adjust PaCO2 targets in OHCA patients with COPD. (c) 2021 Elsevier Inc. All rights reserved.