Approaches to Converting Sp o2 /F io2 Ratio to Pa o2 /F io2 Ratio for Assessment of Respiratory Failure in Critically Ill Patients: A Systematic Review.

Abstract

The Pa o2 /F io2 (PF) ratio is widely used as an assessment of respiratory failure in guiding ventilation strategies and prognostication in critically ill patients. However, given that it mandates invasive arterial access, the Sp o2 /F io2 (SF) ratio has been suggested as a noninvasive and readily accessible alternative. What are the best ways to convert SF and PF ratios in critically ill patients, in terms of their diagnostic/prognostic accuracy and clinical utility? We comprehensively searched databases (MEDLINE, Embase, Web of Science, Cochrane library) to identify relevant studies. Any observational studies that compared the SF to PF ratio in critically ill patients. We assessed individual study risk of bias (ROB) using the revised QUADAS II tool. We included 45 observational studies, ranging from 61 to 141,000 measurements. SF to PF imputation was less accurate when the Sp o2 was equal to or greater than 97%. Otherwise, all studies were able to establish strong correlational relationships between SF and PF ratios, but there was no clear best equation. Based on ease of use, size, generalizability and methodology, we were able to prioritize four equations (one linear, one logarithmic linear, and two nonlinear). All four equations showed strong correlation between SF and PF ratios, with the linear equation being easiest to apply. The SF ratio also correlated well with clinical outcomes when compared with the PF ratio, both as an individual value and as part of a comprehensive score, with more discriminating performance in some cases. SF and PF ratios demonstrate good correlation, and may have similar prognostic value. Although there is no clear optimal method to convert SF to PF ratios, linear equations show acceptable correlation and are most easily applied at the bedside.