In summary, for patients who received dexamethasone:

Benefits in NNT

    2.8% lower risk of death in the overall cohort
    1 in 36 were helped (death prevented)

Harms in NNT

    Not reported
    Not reported

View As:

Source: RECOVERY Collaborative Group, Horby P, Lim WS, et al. Dexamethasone in Hospitalized Patients with Covid-19 - Preliminary Report [published online ahead of print, 2020 Jul 17]. N Engl J Med. 2020.

Study Population: 6425 hospitalized adults with COVID-19 in 176 hospitals in UK

Efficacy Endpoints: All-cause mortality within 28 days, discharge from hospital within 28 days, and requiring invasive mechanical ventilation or extracorporeal membrane oxygenation

Harm Endpoints: Not reported

Narrative: The novel severe acute respiratory syndrome coronavirus (SARS-CoV2) emerged in 2019 resulting in a global pandemic with millions infected and hundreds of thousands dead from coronavirus disease 2019, i.e. ‘COVID-19’.1 COVID-19 can cause lung inflammation and respiratory failure believed to be due to an exaggerated immune response.2 Corticosteroids such as dexamethasone may temper this response.

RECOVERY is a complicated randomized, controlled, open label trial exploring multiple treatment options for COVID-19.3 The results for the dexamethasone arm have been released in a preliminary report. A total of 6425 hospitalized adult patients with suspected or confirmed SARS-CoV-2 infection in the UK were randomized (2:1) to receive usual care or usual care plus dexamethasone 6 mg once daily for up to 10 days. The primary outcome was all-cause mortality. Secondary outcomes included time to discharge and invasive mechanical ventilation or extracorporeal membrane oxygenation. Outcomes were measured at discharge or on day 28 if still in the hospital.3

Administration of dexamethasone was associated with lower mortality overall (relative risk [RR]: 0.83, 95% CI, 0.8-0.9; absolute risk difference [ARD]: 2.8%; NNT: 36). In subgroup analyses this was more pronounced for those receiving invasive mechanical ventilation (RR: 0.6, 95% CI, 0.5-0.8; ARD 12.1%; NNT 8) or requiring oxygen but not ventilation (RR: 0.8, 95% CI, 0.7-0.9; ARD 2.9%; NNT 34). In non-ventilated patients, dexamethasone reduced progression to invasive mechanical ventilation (RR: 0.8, 95% CI, 0.6-0.9; ARD 2.1%; NNT 48). Dexamethasone group mortality was higher in those not on oxygen, but the difference was not statistically significant.3

Caveats: This trial was neither blinded nor placebo-controlled. These limitations introduce significant systemic bias into the findings and limit the validity of the results. There are other limitations as well. Usual care, for instance, was not defined or protocolized which may have led to differential care decisions such as more common withdrawal of measures in the control group or more aggressive care in the treatment group. Furthermore, dexamethasone harms are not detailed in this study even as prior reviews have at times suggested possible increased risks for superinfection, metabolic derangements, and other corticosteroid adverse effects.4

The readers should exercise caution interpreting the results of subgroup analyses from this trial as it is not clear if these analyses were pre-planned or appropriately powered. Moreover, if they were both preplanned and adequately powered the differences would still be considered hypothesis-generating, not definitive.

Although the preliminary results of this study are promising, more research is needed to confirm these findings. Administration of dexamethasone seems reasonable, and hopeful, based on these data. But the overall lack of methodologic rigor, as well as hints of harm in subgroup analysis, suggest these results are unlikely to be stable. Replication studies and subgroup-based investigations will be the only path to accurately answering questions on the utility of dexamethasone in Covid-19. We believe clinical equipoise remains, and have therefore assigned a color recommendation of yellow (unclear if benefits, more data needed). We hope to see large scale further randomized trials as long as the pandemic continues.

Author: Kenneth Lu, MD; Eric Tang, MD
Supervising Editors: Shahriar Zehtabchi, MD

Published/Updated: September 1, 2020

  1. The Title Bar

    The title bar is color-coded with our overall recommendation.

    • Green: Benefits outweigh risks.
    • Yellow: Unclear risk/benefit profile.
    • Red: Benefits do not outweigh risks.
    • Black: Obvious harms, no clear benefits.
  2. Tip content...