76. Effect of Early Administration of Systemic Corticosteroids on Outcomes in Patients with COVID-19 Pneumonia

Abstract

Background Systemic corticosteroids (steroids) have been empirically used in acute respiratory distress syndrome, an entity also present in coronavirus disease 19 (COVID-19). Early steroids administration could accelerate resolution of symptoms and reduce intensive care unit (ICU) stay in these patients, but practice varies widely as evidence is scant. Methods We reviewed the records of 498 adults admitted to Stony Brook University Hospital, NY, from 3/1 to 4/15, 2020 with COVID-19 requiring high-flow O2 (non-rebreather mask, Venturi mask with FiO2 >50%, or high-flow nasal cannula). We excluded those (N=29) who received mechanical ventilation (MV) or died within 24h of admission. We followed patients until death or discharge. We compared outcomes between patients who received early steroids (i.e. prior to MV) and those who did not. We used adjusted Cox models to evaluate the composite of death or need for MV. We also evaluated healthcare resources utilization. Results Of 469 patients, 175 (37.3%) received steroids while on high flow O2. Table 1 summarizes the baseline characteristics. Patients who received steroids were more likely to have asthma, had slightly longer duration of symptoms, lower O2 saturation, higher NT-proBNP and lower IL-6 levels at baseline. In total, 228 patients (48.6%) reached the composite endpoint (123 died and 105 received MV). By 7 days, 32.5% of patients who received steroids died or were intubated vs. 44.8% of those who did not (log-rank P=0.008), Figure 1. In models adjusted for race, age, sex, comorbidities, baseline O2 saturation and procalcitonin, steroids reduced risk for death or MV by 44% (hazard ratio [HR] 0.56; 95%CI 0.42–0.76; P< 0.001). The effect was time-dependent with initial HR 0.34 (95%CI 0.21–0.56; P< 0.001) and daily attenuation by 10.2% (95%CI 1.7%–19.4%; P=0.017). Mortality at 7 and 14 days did not differ between groups (8.1% vs. 8.3% and 19.1% vs. 21.0%, respectively, log-rank P=0.75). Among discharged patients, length of hospital stay was longer, but ICU stay was shorter with steroids, Table 2. Patient Characteristics According to Use of Early Steroids Healthcare Resources Utilization According to Use of Early Steroids Among Discharged Patients Conclusion Early administration of steroids reduced primarily the need for MV in our high-risk COVID-19 patients, with shorter ICU utilization, at the expense of longer hospital stay. Further studies are needed to optimize the use of steroids in these patients. Disclosures All Authors: No reported disclosures