Abstract
Background: The oxygenation index (OI) and oxygen saturation index (OSI) are proven mortality predictors in pediatric and adult patients, traditionally using mean airway pressure (P(mean)). We introduce novel indices, replacing P(mean) with DP (ΔP(insp)), MP(dyn), and MP(tot), assessing their potential for predicting COVID-19 acute respiratory distress syndrome (ARDS) mortality, comparing them to traditional indices. Methods: We studied 361 adult COVID-19 ARDS patients for 7 days, collecting ΔP(insp), MP(dyn), and MP(tot), OI-ΔP(insp), OI-MP(dyn), OI-MP(tot), OSI-ΔP(insp), OSI-MP(dyn), and OSI-MP(tot). We compared these in surviving and non-surviving patients over the first 7 intensive care unit (ICU) days using Mann-Whitney U test. Logistic regression receiver operating characteristic (ROC) analysis assessed AUC and CI values for ICU mortality on day three. We determined cut-off values using Youden's method and conducted multivariate Cox regression on parameter limits. Results: All indices showed significant differences between surviving and non-surviving patients on the third day of ICU care. The AUC values of OI-ΔP(insp) were significantly higher than those of P/F and OI-P(mean) (P values .0002 and <.0001, respectively). Similarly, AUC and CI values of OSI-ΔP(insp) and OSI-MP(dyn) were significantly higher than those of SpO(2)/FiO(2) and OSI-P(mean) values (OSI-ΔP(insp): P < .0001, OSI-MP(dyn): P values .047 and .028, respectively). OI-ΔP(insp), OSI-ΔP(insp), OI-MP(dyn), OSI-MP(dyn), OI-MP(tot), and OSI-MP(tot) had AUC values of 0.72, 0.71, 0.69, 0.68, 0.66, and 0.64, respectively, with cut-off values associated with hazard ratios and P values of 7.06 (HR = 1.84, P = .002), 8.04 (HR = 2.00, P ≤ .0001), 7.12 (HR = 1.68, P = .001), 5.76 (HR = 1.70, P ≤ .0001), 10.43 (HR = 1.52, P = .006), and 10.68 (HR = 1.66, P = .001), respectively. Conclusions: Critical values of all indices were associated to higher ICU mortality rates and extended mechanical ventilation durations. The OI-ΔP(insp), OSI-ΔP(insp), and OSI-MP(dyn) indices displayed the strongest predictive capabilities for ICU mortality. These novel indices offer valuable insights for intensivists in the clinical management and decision-making process for ARDS patients.