Study to identify predictor of hypoxia in COVID-19 infection: A single-center, retrospective study

Background: Given the limited availability of critical care resources in our country, it is important to identify reliable predictors of hypoxia in patients with COVID-19 infection. We thus sought to compare differential predictive values of demographic, clinical, and laboratory measures and identify predictor for hypoxia in COVID-19 infection. Materials and Methods: This is single-center retrospective analysis of patient admitted at AIIMS Patna between 15th June and 15th Aug. All the patients who had oxygen saturation less than 94% grouped under hypoxic group while ≥94% in non-hypoxic group at room air. Statistical analysis: Logistic regression model used to determine the predictor of hypoxia in COVID-19 infection. Results: Total 73 were used for analysis. Study patients had a mean age of 55.05 ± 12.7 year, of whom 78.08% were male (57/73). 39 (53.42%) patients were found hypoxic at time of admission while 34 (46.56%) were non-hypoxic. Presence of dyspnoea significantly found more frequently in hypoxic patients (P = 0.0003). Patients with O2 saturation of less than 94% have more likely to have diabetes (P = 0.002) and hypertension (P = 0.02). Analysis of laboratory variables showed that leucocytosis (P = 0.007) and neutrophilia (P = 0.01) were significantly higher in hypoxic group compare to non-hypoxic group. Univariate regression model showed patient with any one comorbidities, diabetes, or hypertension was found as strong risk factor for hypoxia after COVID-19 infection (P < 0.05). Conclusion: This is the first study to identify predictor of hypoxia in COVID-19 infection. Patient with any one comorbidities, diabetes, or hypertension was found strong risk factor for hypoxia after COVID-19 infection. Among lab variable, leucocytosis, neutrophilia, lymphocytopenia, and CRP (>27.5 mg/L) were found as predictor of hypoxia in COVID infection.

Abstract licence: CC BY-NC-SA

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.