Predictors of length of hospital stay in patients with acute COVID-19

Abstract

The aim of the study was to identify the predictors of length of hospital stay in patients with acute COVID-19, based on the pathophysiological particularities of SARS. Materials and methods: The study was conducted from December 2020 to May 2021 on 103 randomized patients (59 men, mean age 62±13 years, body mass index 30.3±5.5 kg/m2 ) with moderate to severe acute COVID-19 infection who were hospitalized for emergency non-invasive oxygen therapy. Log-regression models were used to assess the suitability of some functional spirometric variables and/or SpO2/FiO2 surrogate oxygenation index to predict the duration of inpatient treatment from the day of the examination (≤ 7 vs. > 7 days). Results: The analysis of the receiver operating characteristic curves showed that the relative indicator of the “ventilatory reserve” (the ratio of maximum voluntary ventilation to minute ventilation at rest, VR= MVV / MV) has sufficient sensitivity (82%), specificity (69%) and the area under the curve (AUC=0.7), although the SpO2/FiO2 ratio has a better predictive capacity (78%, 84% and 0.8 accordingly). The model combining these two integral indicators of gas exchange and the respiratory muscles reserve showed the best sensitivity (89%), specificity (84%) and area under the curve (0.9). Conclusion: The proposed model for determining of the hypoxia vector by assessing the severity of ventilation-perfusion dissociation with indicators of oxygenation and ventilation showed that the MVV/MV ratio and SpO2/FiO2 ratio can be used alone and especially in combination, as effective outcome predictors of the acute phase of the infectious process caused by SARS-CoV-2.