Introduction

The efficacy of low-dose dexamethasone in patients with COVID-19 with a regimen of 6 mg daily for ten days has been established since the early days of the pandemic (RECOVERY Collaborative Group et al. 2021). However, despite the positive findings (e.g., in RECOVERY trial), questions have been raised about whether using a higher dose of dexamethasone in patients with COVID-19 could be more clinically effective in containing the excessive inflammatory responses, especially in severely or critically ill patients, in order to improve their prognosis. Therefore, we conducted a systematic review with meta-analysis to estimate the risk of mortality using intermediate- to high-dose dexamethasone compared with low-dose dexamethasone in this population of patients.

Methods

We performed a comprehensive literature search in six electronic databases: PubMed, Scopus, and Web of Science for published studies, and medRxiv, Research Square, and SSRN for preprints from inception to 20 March 2023. The search strategy in the electronic databases was built based on a combination of the following keywords and their Medical Subject Headings (MeSH) terms (if applicable): “COVID-19”, “SARS-CoV-2”, “dexamethasone”, “corticosteroid”, “steroid”, and “glucocorticoid”. We also hand-searched the bibliographies of included articles and relevant reviews to retrieve additional relevant records. We limited the search to human and adult studies with no restrictions on publication date or publication status.

Two investigators (CSK and SSH) independently screened all retrieved references for inclusion based on the study title and abstract. In addition, the two investigators (CSK and SSH) retrieved and reviewed the full text of articles deemed possibly eligible for inclusion. We resolved disagreement during the review process through a discussion with a third investigator (DSR) and by consensus.

We restricted our search to randomized clinical trials that assessed mortality outcomes with intermediate- to high-dose dexamethasone against low-dose dexamethasone in adult participants (≥ 18 years) infected with SARS-CoV-2 and requiring any form of respiratory support (e.g., oxygen therapy, non-invasive ventilation, and invasive ventilation). A low-dose or an intermediate- to high-dose dexamethasone was classified based on the a priori–defined cutoff of 10 mg per day. We excluded non-randomized trials, single-arm trials, observational studies, case reports, reviews, conference abstracts, animal studies, and non-English language publications.

Two investigators (CSK and DSR) extracted data independently from each included trial using a standardized data collection form. Data extracted included characteristics of the included studies (first author’s surname, year of publication, country where the trial was performed, and trial design), details of the population enrolled (mean/median age and illness severity), details of the study interventions (dose, frequency, and duration), mortality outcomes, and adverse events. We resolved discrepancies in the data extracted by the two investigators through discussion or, if necessary, by involving a third investigator (SSH) for adjudication.

The risk of bias for each of the included trials was assessed independently by two investigators, CSK and SSH, using the Cochrane Risk of Bias Tool version 2 (Sterne et al. 2019), which is structured into a fixed set of bias domains. Disagreements were resolved by discussion with a third investigator (DSR) and by consensus. We adjudicated the overall risk of bias as ‘low’ only if all domains were assessed as low risk of bias.

The outcome of interest was all-cause mortality, assessed at two time points: 28 days and 60 days and above.

A random-effects model was used to estimate the pooled odds ratio for the development of outcomes of interest with the use of intermediate- to high-dose dexamethasone relative to the use of low-dose dexamethasone, at 95% confidence intervals. To evaluate the robustness of the pooled estimate, an inverse variance heterogeneity (IVhet) model was also fitted. In addition, the heterogeneity between studies was examined using the I2 statistics and the χ2 test, with substantial heterogeneity predetermined at 50% and p < 0.10, respectively. Finally, publication bias was assessed visually using Begg’s funnel plot. All analyses were performed using Meta XL, version 5.3 (EpiGear International, Queensland, Australia).

Results

The systematic literature search conducted on electronic databases retrieved a total of 1,935 records. After deduplication, 1,527 records were identified, and from these, 11 records were evaluated for eligibility. Three studies were excluded due to their observational study designs. Ultimately, our systematic review included eight randomized controlled trials (Bouadma et al. 2022; Granholm et al. 2022; RECOVERY Collaborative Group et al. 2023; Maskin et al. 2022; Munch et al. 2021; Rabascall et al. 2022; Taboada et al. 2021; Toroghi et al. 2022; Wu et al. 2022), comprising 1,800 patients randomized to receive intermediate- to high-dose dexamethasone and 1715 patients randomized to receive low-dose dexamethasone.

The characteristics of the included randomized trials are shown in Table 1. The regimen of intermediate- to high-dose dexamethasone in the intervention group differed across the included trials: five trials (Bouadma et al. 2022; RECOVERY Collaborative Group 2023; Maskin et al. 2022; Taboada et al. 2021; Wu et al. 2022) administered dexamethasone using dose-tapering strategies (16 to 20 mg daily for five days, followed by 8 to 10 mg daily for additional five days), one trial (Rabascall et al. 2022) administered dexamethasone using weight-based dosing strategy (0.2 mg/kg daily for 10 days), while the remaining two trials (Granholm et al. 2022; Munch et al. 2021; Toroghi et al. 2022) administered dexamethasone with doses of 12 to 24 mg daily for ten days. The regimens of low-dose dexamethasone in the comparative group across the included trials (Bouadma et al. 2022; Granholm et al. 2022; RECOVERY Collaborative Group. 2023; Maskin et al. 2022; Munch et al. 2021; Rabascall et al. 2022; Taboada et al. 2021; Toroghi et al. 2022; Wu et al. 2022) were 6–8 mg daily for ten days.

Table 1 Characteristic of included studies
Full size table

We adjudicated two trials (Bouadma et al. 2022; Granholm et al. 2022; Munch et al. 2021) as having a low risk of bias in all domains and, thus, an overall low risk of bias. The remaining trials were rated as having an overall some concerns of bias (some concerns of bias in at least one domain): the five trials reported by Maskin et al. (2022), Taboada et al. (2021), Rabascall et al. (2022), RECOVERY Collaborative Group. (2023) and Wu et al. (2022), respectively, had some concerns in the domain of deviations from intervention due to open-label trial design, while the trial reported by Toroghi et al. (2022) had some concerns in the domain of deviations from intervention due to single-blind trial design.

The meta-analysis of six trials (RECOVERY Collaborative Group. 2023; Maskin et al. 2022; Munch et al. 2021; Rabascall et al. 2022; Taboada et al. 2021; Wu et al. 2022) using a random-effects model revealed no significant difference in the risk of 28-day all-cause mortality between intermediate- tohigh-dose dexamethasone and low-dose dexamethasone. The estimated effect of low-dose dexamethasone on 28-day all-cause mortality (Fig. 1; pooled odds ratio = 1.16, 95% confidence interval 0.77–1.74) indicates mortality benefits compared to intermediate- to high-dose dexamethasone, but is with limited evidence against our model hypothesis of ‘no significant difference’ at the current sample size. The IVhet model also produced same mortality estimate. Notably, one large trial reported by RECOVERY Collaborative Group (2023) demonstrated a statistically significant higher risk of 28-day all-cause mortality with intermediate- to high-dose dexamethasone compared to low-dose dexamethasone.

Fig. 1
figure 1

Pooled odds ratio for all-cause mortality with the administration of intermediate- to high-dose dexamethasone among patients with COVID-19 requiring respiratory support relative to low-dose dexamethasone

Full size image

Similarly, the meta-analysis of five trials (Bouadma et al. 2022; Granholm et al. 2022; Maskin et al. 2022; Taboada et al. 2021; Toroghi et al. 2022) using a random-effects model revealed no significant difference between intermediate- to high-dose dexamethasone and low-dose dexamethasone in terms of 60-day all-cause mortality. The estimated effect with intermediate- to high-dose dexamethasone on the 60-day all-cause mortality (Fig. 1; pooled odds ratio = 0.96, 95% confidence interval 0.74–1.26) suggests mortality benefits compared to low-dose dexamethasone, but is with limited evidence against our model hypothesis of ‘no significant difference’ at the current sample size. The IVhet model also produced non-significant evidence of mortality benefits (pooled odds ratio = 0.92, 95% confidence interval 0.69 to 1.21).

Funnel plots revealed a gross asymmetry to either side, suggesting possible publication bias (Fig. 2). Additionally, the studies were scattered asymmetrically around the summary effect.

Fig. 2
figure 2

Funnel plot with 95% confidence interval on the odds ratio of included studies in the meta-analysis (n = 9)

Full size image

Discussion

In this systematic review and meta-analysis, we found no significant mortality benefits with the use of intermediate- to high-dose dexamethasone in patients with COVID-19 requiring respiratory support compared with low-dose dexamethasone. Our systematic review and meta-analysis provide the most up-to-date evidence (when writing this manuscript) of administering intermediate- to high-dose dexamethasone compared with low dose dexamethasone on mortality outcomes in this population of patients. The methodologic strengths of our systematic review and meta-analysis include a focused research question with a defined population, intervention, comparator, and outcome.

The findings of our review have important implications for clinical practice, especially in the context where clinicians intend to trial a higher than the recommended dose of dexamethasone (6 mg daily for ten days) in patients with COVID-19 who are severely or critically ill, recognizing the positive effects of dexamethasone. In this population of patients, additional anti-inflammatory therapies may be more clinically effective in reducing the risk of mortality than escalating the dose of dexamethasone. For example, available data suggest that both baricitinib and tocilizumab provide a mortality benefit for patients with COVID-19 who experience a severe course of the disease, even if they are already receiving low-dose dexamethasone (Abani et al. 2022; Domingo et al. 2021).

The use of higher doses of dexamethasone is not without safety concerns. One of the dose-related complications with dexamethasone is the development of hyperglycemia, which could potentially offset its mortality benefits in patients with COVID-19 (Kow et al. 2021). The trial reported by Toroghi et al. (2022) observed a higher proportion of patients in the high-dose group (47.8%) and intermediate-dose group (37.5%) developed hyperglycemia compared to the low-dose group (29.8%). In addition, using a higher dose of dexamethasone may predispose patients to a higher risk of secondary infections. In the same trial (Toroghi et al. 2022), a higher proportion of patients in the high-dose group (8.7%) and intermediate-dose group (2.5%) developed secondary infections compared to the low-dose group (2.1%).

Our systematic review and meta-analysis are limited by the lack of information from the included studies on the rate of SARS-CoV-2 vaccination and the predominant SARS-CoV-2 variants during the study period, which can influence the risk of mortality among patients with COVID-19. Nevertheless, the lack of information mentioned above is understandable since the variants of concerns were not widely circulated, sequencing for variants of concerns was not widely available, and COVID-19 vaccination was not publicly available during the study period for most of the included studies.

Overall, our systematic review and meta-analysis of currently existing data from published randomized clinical trials suggest that administering intermediate- to high-dose dexamethasone in patients with COVID-19 requiring respiratory support has similar mortality benefits compared with low-dose dexamethasone. Nevertheless, one large trial (RECOVERY Collaborative Group. (2023)) indicated 28-day mortality benefits with low-dose dexamethasone compared with intermediate- to high-dose dexamethasone. Higher doses of dexamethasone may be associated with safety concerns and does not show apparent superiority over low-dose dexamethasone. Prioritizing low-dose dexamethasone in this population of patients would be beneficial.