Predicting survival of patients with COVID-19 admitted at Mulago National Referral Hospital using the chest x-ray (BRIXIA) score

Abstract

Introduction: By 31st October, 126,000 cases of Covid-19 pneumonia and 3,215 deaths had been reported in Uganda. Chest x-ray score (Brixia score) plays an important role in the diagnosis, treatment and monitoring of patients with covid-19 pneumonia. However, its performance in the Ugandan setting has not been fully studied. Therefore, this study aimed to describe chest- X ray findings, determine the Brixia score and its role in predicting survival among of patients with COVID-19 admitted at Mulago National Referral Hospital (MNRH). General objective: To determine the survival of patients with COVID-19 admitted at MNRH using Brixia score at 40 days following admission. Methods: This was a retrospective cohort study that involved reviewing records of patients admitted with Covid-19 at MNRH from March 2020 to July 2021. Records of 212 patients were randomly sampled. The Chest X-ray findings and Brixia scores were summarized using proportions, mean and standard deviation. The outcome of patients was determined using survival analysis. The predictors of survival after 40 days of admission were determined using proportional cox regression with Hazard ratios at P-value <0.05 and 95% confidence interval. Results: Data of 212 patients were analysed. Majority were at least 40 years old, 77.8% (165/212) and the mean age was 51.9 years (SD+/-14.7years). More were of male gender, 58.7% (121/206) and the most common co-morbidity was hypertension, 43.7% (90/206). The most common presenting symptoms at admission were cough 82.1% (174/212) and difficulty in breathing, 72.2% (153/212). Majority of the patients had severe WHO clinical disease, 69.3% (147/212) and the most common Chest X-ray finding was GGO, 71.7% (152/212) which was mostly distributed in the middle and lower zones, 66.0% (140/212). The mean Brixia score was 11.5 (SD+/-6.0) and there was a statistically significant positive correlation between Brixia score and the WHO clinical severity groups, p-value<0.001. In the multivariate model, patients with cough or throat symptoms were 0.64 (aHR= 0.36) less likely to die compared to those with no cough or throat symptoms, p-value =<0.001. Patients with Brixia score >9 were 6.03 times more likely to die compared to those with Brixia ≤9, (aHR=6.03), p-value= 0.001. Conclusion: At MNRH, most patients admitted with Covid-19 had GGO on Chest- X ray especially in middle and lower lung zones. The mean Brixia score was high and directly correlated with WHO clinical grading. Presence of cough or throat symptoms, and low Brixia score at admission increased survival of patients. Therefore, Brixia score may be used in the management of Covid-19 in a low resource setting and responsible stakeholders should make x-rays readily available and affordable in all facilities deemed to manage Covid-19 patients.