RESUMO
OBJECTIVE: The arterial to end-tidal carbon dioxide tension difference (CO2 gap) correlates with physiologic dead space. The prognostic value of increased CO2 gap in trauma and respiratory distress patients is documented. Transpulmonary arteriovenous shunting is identified as a predictor of mortality in non-pulmonary sepsis. We set out to investigate the prognostic value of the CO2 gap in a pilot study of patients with suspected sepsis from non-respiratory causes. METHODS: Patients presenting to tertiary Australian ED with suspected sepsis (n = 215) underwent near-simultaneous end-tidal carbon dioxide and partial pressure of carbon dioxide measurements. We investigated the correlation of CO2 gap levels with the primary outcome of in-hospital mortality (IHM) and secondary outcomes of sepsis (ΔSOFA ≥2) and IHM and/or intensive care unit stay ≥72 h (IHM/ICU72h) in patients with sepsis because of non-respiratory causes. RESULTS: Among patients included in the analysis (n = 165), the CO2 gap showed modest positive correlation with qSOFA (ρ = 0.39) and weak positive correlation with SOFA scores (ρ = 0.29) (both P < 0.01). The CO2 gap had modest predictive value for primary outcome (IHM), area under receiver operating curve (AUROC 0.85, 95% confidence interval [CI] 0.78-0.90) and IHM/ICU72h outcome (AUROC 0.80, 95% CI 0.73-0.86), but lower predictive value for sepsis outcome (AUROC 0.64, 95% CI 0.55-0.71) (all P < 0.001). We report modest test performance for primary outcome at CO2 gap ≥5 and ≥10 mmHg cut-offs. CONCLUSION: In this pilot study of patients with suspected sepsis from non-respiratory causes, an increased CO2 gap demonstrates value in risk stratification and needs to be further evaluated and compared to other existent biomarkers.