PaCO2 Levels at Admission Influence the Prognosis of Sepsis Patients: A Nonlinear Relationship

Background: Sepsis is a life-threatening organ dysfunction caused by a maladjusted host response to infection (Sequential Organ Failure Assessment [SOFA] ≥ 2). PaCO2 is a blood gas that is commonly tested in hospitalized patients. Few studies have investigated how abnormal PaCO2 levels influence the prognosis of sepsis patients. Methods: A retrospective study was applied to data on sepsis patients extracted from the MIMIC-III public database and the results of their first PaCO2 examination after admission. The outcome measure was all-cause death within 90 days of follow-up. The patients were divided into five quintile groups, and the relationship between PaCO2 and all-cause death in sepsis patients was studied using restricted cubic splines (RCSs). Results: Our study included 4898 sepsis patients, the hypocapnia (PaCO2 < 31 mmHg) and hypercapnia (PaCO2 ≥ 49 mmHg) groups were associated with the risk of death in sepsis patients using quintile grouping and the multivariate Cox model, with HRs of 1.12 (95% CI = 1.03–1.29) and 1.25 (95% CI = 1.10–1.41), respectively, when compared with the control group (37 ≤ PaCO2 < 41 mmHg). The relationship between PaCO2 and all-cause death in sepsis patients was U-shaped. Conclusion: The presence of hypercapnia and hypocapnia at the time of hospital admission will have adverse effects on all-cause death in sepsis patients.