OBJECTIVES: 

The objectives of our systematic review and meta-analyses were to determine the diagnostic accuracy of central venous oxygen saturation (Scvo₂) in estimating mixed venous oxygen saturation (Svo₂) and cardiac index in critically ill patients.

DATA SOURCES: 

A systematic search using MEDLINE, Cochrane Central Register of Controlled Trials, and Embase was completed on May 6, 2024.

STUDY SELECTION: 

Studies of patients in the ICU for whom Scvo₂ and at least one reference standard test was performed (thermodilution and/or Svo₂) were included.

DATA EXTRACTION: 

Individual patient data were used to calculate the pooled intraclass correlation coefficient (ICC) for Svo₂ and Spearman correlation for cardiac index. The Quality Assessment of Diagnostic Accuracy Studies-2 and Grading Recommendations Assessment, Development, and Evaluation tools were used for the risk of bias and certainty of evidence assessments.

DATA SYNTHESIS: 

Of 3427 studies, a total of 18 studies with 1971 patients were identified. We meta-analyzed 16 studies (1335 patients) that used Svo₂ as a reference and three studies (166 patients) that used thermodilution as reference. The ICC for reference Svo₂ was 0.83 (95% CI, 0.75–0.89) with a mean difference of 2.98% toward Scvo₂. The Spearman rank correlation for reference cardiac index is 0.47 (95% CI, 0.46–0.48; p < 0.0001).

CONCLUSIONS: 

There is moderate reliability for Scvo₂ in predicting Svo₂ in critical care patients with variability based on sampling site and presence of sepsis. There is limited evidence on the independent use of Scvo₂ in predicting cardiac index.