Prognostic effects of arterial partial pressure of carbon dioxide in critical patients with acute heart failure: a retrospective cohort study

Abstract

Abstract Background: The association of the arterial partial pressure of carbon dioxide (PaCO2) with poor outcomes in acute heart failure (AHF) patients has been observed. Nevertheless, the existing studies are controversial and limited by small sample sizes. We assessed whether PaCO2 is associated with intensive care unit (ICU) 60-day mortality in AHF patients. Methods: We extracted and analyzed data from a quality public database, the Medical Information Mart for Intensive Care IV database. We used a smooth curve fitting method to assess a potential curvilinear relationship. Multivariate regression analyses incorporating a two-stage linear regression model were used to investigate the associations of PaCO2 with prognosis. Results: Overall, 4,271 participants were included, comprising 825 ICU 60-day non-survivors and 3,446 ICU 60-day survivors. Smooth curve fitting revealed a U-shaped relationship between PaCO2 and prognosis, with an inflection point of PaCO2 at 37 mmHg that was used to establish a two-stage linear regression model. When PaCO2 was <37 mmHg, the multivariate regression analysis showed that for every 5-mmHg increase in the PaCO2, the ICU 60-day mortality decreased by 16% (p=0.0274). There was no significant association between PaCO2 and ICU 60-day mortality (p=0.4685) when PaCO2 was ≥37 mmHg. Conclusions: These findings suggest that low PaCO2, but not high PaCO2, at ICU admission might help determine the short-term prognosis in critically AHF patients.