Correlation between invasive and noninvasive oxygen indices in critically ill COVID patients with ARDS: A retrospective analysis

Objective: This study aims to evaluate the correlation between peripheral oxygen saturation/fraction of inspired oxygen (SpO 2 /FiO 2 ) and arterial partial pressure of oxygen (PaO 2 )/FiO 2 ratio in the diagnosis and assessment of severity in acute respiratory distress syndrome (ARDS) patients admitted to intensive care unit (ICU) during the SARS-CoV-2 outbreak. Methods: A retrospective observational study was conducted in patients admitted to COVID-ICU during a 1-year period. Time coinciding values of SpO 2 and PaO 2 were noted from patients’ medical records. An evaluation was made to judge the efficacy of the SpO 2 /FiO 2 ratio (S/F ratio) for the prediction of 100, 200, 300, and >300 mmHg cutoff values of PaO 2 /FiO 2 ratio (P/F ratio) using receiver operating characteristic (ROC) curves. Results: The medical records of 196 patients were evaluated. A strong positive correlation was identified between P/F and S/F ratios in the sample population (correlation coefficient r = 0.716, p < 0.002). The area under the ROC curve of the S/F ratio was 0.720 (95% confidence interval [CI] 0.637–0.804) for mild ARDS. The optimal S/F threshold was found to be ≤411% (sensitivity and specificity 65% and 80%, respectively). Similarly, the S/F threshold was ≤405% (area under the curve [AUC] 0.781, 95% CI 0.713–0.848, p < 0.000) for moderate ARDS (sensitivity 71.8% and specificity 70.8%), and ≤ 395% (AUC 0.762, 95% CI 0.689–0.835, p < 0.000) for severe ARDS (specificity 73.5% and sensitivity 66.7%). Conclusions: The S/F ratio can be used as a reliable, noninvasive alternative tool for the diagnosis of ARDS and assessment of severity. This can be used as an early warning score in patients admitted in low-resource settings with respiratory illness, and an escalation of care can be decided.