Prognostic effects of arterial carbon dioxide levels in patients hospitalized into the cardiac intensive care unit for acute heart failure

AIMS: Although both hypercapnia and hypocapnia are common in acute heart failure (AHF) patients, routine assessment of arterial blood gas is not recommended. Additionally, no association between hypercapnia and increased mortality has been found, and the prognostic value of hypocapnia in AHF patients remains to be elucidated. In this observational study, we aimed to investigate the relationship between partial pressure of arterial carbon dioxide (PaCO2), especially low PaCO2, and long-term mortality in AHF patients. METHODS AND RESULTS: Acute heart failure patients hospitalized in the cardiac intensive care unit of our institution between 2007 and 2011 were screened. All eligible patients were divided into two groups based on the inflection point (i.e. 31.0 mmHg) of the 3-knot cubic spline curve of the hazard ratio (HR), with a PaCO2 of 40 mmHg as a reference. The association between PaCO2 levels and all-cause mortality was assessed using Cox proportional hazards regression models. Among 435 patients with a median follow-up of 1.8 years, 115 (26.4%) died. Adjusted analysis with relevant variables as confounders indicated that PaCO2 <31 mmHg was significantly associated with increased all-cause mortality [HR 1.71, 95% confidence interval (CI) 1.05-2.79; P = 0.032]. When PaCO2 was considered as a continuous variable, the lower was the log-transformed PaCO2, the greater was the increased risk of mortality (HR 0.71, 95% CI 0.52-0.96; P = 0.024). CONCLUSIONS: In AHF patients, lower PaCO2 at admission was associated with increased long-term mortality risk.