May 18, 2022
By Aldon Li, MD, FIDSA
The mortality benefit with the use of dexamethasone in hospitalized COVID-19 patients requiring supplemental oxygen was first demonstrated in the preliminary release of data from the RECOVERY trial in June 2020, leading to subsequent guidelines from IDSA and the National Institutes of Health (NIH) endorsing the use of dexamethasone 6 mg for 10 days (or until discharge) in hypoxic, hospitalized patients. However, NIH recognizes there is insufficient evidence to recommend either for or against the use of dexamethasone in patients who are stable for discharge, but still requiring oxygen supplementation.
A recent study in JAMA Network Open conducted a retrospective review of hospitalized adults with COVID-19 who were admitted prior to the widespread availability of remdesivir and use of steroids in Southern California between May 2020 and September 2020, comparing all-cause readmission or mortality within 14 days from discharge in 1) the group who continued dexamethasone on discharge to 2) the group who stopped dexamethasone on discharge.
The authors found a heterogenous cohort of 1,164 patients (59.5% continued dexamethasone on discharge with some patients receiving a total duration, inpatient + outpatient, of > 10 days) with differing demographics (age, sex, race) and clinical characteristics (51% vs. 59% received remdesivir in the continue vs. stop group, 91% vs. 88% received supplemental oxygen while hospitalized in the continue vs. stop group, 23% vs. 19% had O2 saturation < 94% on discharge in the continue vs. stop group).
Given the heterogeneity between the continue and stop groups, the authors used propensity scores to balance the groups in an attempt to adjust for confounding variables that could affect readmission or mortality. The adjusted odds ratio for the continue group had a point estimate of 0.87, suggesting that continuing dexamethasone at discharge resulted in 13% lower odds of all-cause readmission or mortality within 14 days from discharge. However, given the wide 95% confidence interval of 0.58-1.30, the resulting lower odds may be due to chance rather than a true decrease in risk.
Because of the high probability of random variation influencing the results of this study, following the guidance of the original RECOVERY trial and stopping dexamethasone on discharge even if the patient received less than a 10-day course may be the best option until further studies are available.