Severity and Mortality Associated with Steroid Use among Patients with COVID-19: A Systematic Review and Meta-Analysis

Background There are controversial suggestions about steroid use to treat patients infected with COVID-19. Conclusive evidence regarding the use of steroids to treat COVID-19 is still lacking. This meta-analysis aimed to determine the mortality and severity associated with corticosteroid therapy compared to noncorticosteroid treatment in patients with COVID-19. Methods The information was collected from electronic databases: PubMed, CINAHL, the Cochrane Library, clinicaltrials.gov, and Google scholar through January 30, 2021. Risk ratios (RRs) with 95% confidence intervals (CIs) were performed using random effect models. Endnote citation manager software version X9 for Windows was utilized to collect and organize search outcomes (into relevant and irrelevant studies) and to remove duplicate articles. Results Thirty-two studies were included in the meta-analysis, including 14,659 COVID-19 patients. No significant differences in mortality between the steroid and nonsteroid treatment groups (RR = 0.95; 95% CI: 0.80–1.13; p = 0.57). There was no significant reduction in mortality in critically ill COVID-19 patients treated with corticosteroid (RR = 0.89; 95% CI: 0.62–1.27; p = 0.52). Significant differences were observed in severe disease conditions between the steroid and nonsteroid treatment groups (RR = 1.10; 95% CI, 1.03–1.19, p = 0.007). Conclusion There was no significant difference in all-cause mortality between the steroid and nonsteroid treatment users' of COVID-19 patients. There was no significant reduction of all-cause mortality in critically ill COVID-19 patients treated with corticosteroids.


Background
Coronavirus disease-19  was first identified at the end of 2019 in Wuhan City, China. It rapidly spread in China and other countries throughout the world [1,2]. In March 2020, the World Health Organization characterized the disease as pandemic [3]. As of May 21, 2020, more than five million confirmed cases have been documented and several death cases reported globally. It is affecting 213 countries and territories around the world and 2 international conveyances [4]. Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is identified as the cause of COVID-19 [5].
Currently, there is no drug confirmed by clinical trial to prevent or treat COVID-19. However, more than 300 active clinical treatment trials are under investigation [6]. Drugs that are already marketed for other conditions are being off-label used. For example, antimalarial medications chloroquine and hydroxychloroquine are widely used to treat COVID-19 [7]. A multinational registry analysis debunked the benefit of hydroxychloroquine or chloroquine when used alone or with a macrolide. e finding of the study showed that each of these regimens was associated with decreased in-hospital survival and an increased frequency of ventricular arrhythmias when used for the treatment of COVID-19 [8]. Remdesivir is another drug being used as a promising option for treating COVID-19 based on laboratory experiments [9].
Corticosteroids were widely used to treat severe acute respiratory syndrome coronavirus-1(SARS-CoV-1) and Middle East respiratory syndrome coronavirus (MERS) during their outbreaks and are being used in patients with COVID-2019 [10]. Up to 70% of critically ill patients are receiving systemic corticosteroids. It is identified that patients treated with a corticosteroid had more clinical symptoms, a higher inflammation index, and more abnormalities on chest computed tomography [11]. A study showed that the use of high-dose corticosteroids increases the risk of death in patients with severe COVID-19 [12], indicating several controversial issues about the use of the steroid to treat patients infected with COVID-19. It is suggested that available evidence does not endorse the use of a steroid for COVID-19 patients, which may cause several side effects [13]. However, it is believed that short-term glucocorticoid therapy with small-or medium-dose could be beneficial for patients with severe conditions [14]. According to the World Health Organization (WHO) guidelines recommendation, glucocorticoids should only be used under clinical trial conditions [15]. Conclusive evidence regarding the use of the steroid to treat COVID-19 is still lacking. erefore, this study aims to summarize the current evidence of the severity and mortality associated with steroid therapy for patients with COVID-19 which will support us in making the best decision in the management of the COVID-19.

Study Design.
Analyses were performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [16]. is study was registered in PROSPERO with registration number CRD42020185773.

Search Strategy.
e information was collected from electronic databases: PubMed, CINAHL, the Cochrane Library, clinicaltrial.gov, and Google scholar. ere was no limitation applied to the language. e reference list of all identified articles was searched for additional studies. en, an extensive list of search terms was prepared by the analysis of title, abstract, and keywords of retrieved articles. A fullscale search of databases in PubMed, CINAHL, the Cochrane Central Register of Controlled Trials, clinicaltrial. gov, and Google scholar from inception to January 30, 2021, was performed. No language restriction was imposed during the identification of studies. Flow diagram was used to summarize the number of studies identified, screened, excluded, and finally included in the study.

Study Selection.
Two reviewers independently carried out a literature search and examined relevant studies and sequentially screened their titles and abstracts for eligibility. e full texts of potentially eligible studies were retrieved. Disagreements were solved in a discussion. A screening guide was used to ensure that all review authors reliably apply the selection criteria.

Inclusion and Exclusion
Criteria. Randomized controlled trials (RCTs), observational studies, prospective and retrospective comparative cohort studies, and case-control studies were eligible for the review. e review considered all articles comparing corticosteroid with noncorticosteroid treatment for patients with the diagnosis of COVID-19. Nonhuman studies and studies that did not report mortality and severity data were excluded from the review. e outcome considered was mortality, death within the hospital (all-cause mortality) of COVID 19 patients. e other outcome studied was the number of severe cases in the two groups (corticosteroid versus noncorticosteroid treatment groups).

Methodological Quality Assessment. Selected papers
were assessed by two independent reviewers for methodological validity before inclusion in the review. Observational studies were assessed using Newcastle-Ottawa Scale (NOS), which consists of three domains: (1) subject selection, (2) comparability of the study groups, and (3) assessment of outcome(s). A score of 0-9 was allocated to each study. A standardized Cochrane risk of bias tool was used for the randomized controlled trial (RCT). Any disagreements that arise between the reviewers were resolved through discussion.
2.6. Data Extraction. Two reviewers independently extracted data from the studies using a predesigned format prepared. Data that were extracted include first author, a region of study, included population, study design, sample size, comparator group, patient status, age, gender, interventions, mortality, and severity. Severe cases were characterized as patients admitted to ICU or on invasive mechanical ventilation.

Data Analysis and Statistical Methods.
To perform a meta-analysis, Review Manager 5.4 (Copenhagen: e Cochrane Collaboration, 2014) was used. e outcome variables were calculated using the Mantel Haenszel formula. Risk ratios (RRs) were reported with 95% confidence intervals (CIs) for the variables. e p value was two-tailed, and the statistical significance was set at ≤0.05. Heterogeneity was assessed with the Q-statistic test and the I 2 test. e I 2 statistic measured the percentage of total variation across the studies due to clinical or methodological heterogeneity instead of chance. e significant Q statistics (p < 0.05) indicated heterogeneity across the studies; thus, a random effect model was utilized. Substantial heterogeneity was represented by I 2 for >50%.

Results and Discussion
e search returned a total of 1419 (PubMed: 85, Google Scholar: 1290, clinicaltrials.gov: 24, Cochrane Library: 1, and other sources: 19) citations, of which 40 were duplicates. According to the exclusion criteria, 1292 citations were excluded after the title and abstract screening. Figure 1 shows the study selection processes.

Publication Bias.
To assess the small-study effect and publication bias, the regression-based Egger's test was performed. e funnel-plot analysis showed a symmetrical shape for mortality ( Figure 5), indicating no publication bias; Egger's test indicated nonsignificant small-study effects (p � 0.570). e funnel-plot analysis showed asymmetrical shape for severe cases ( Figure 6), indicating publication bias; Egger's test indicated significant small-study effects for severe cases (p � 0.027).
3.6. Sensitivity Analysis. Sensitivity analysis for betweenstudy heterogeneity and analytical methods was performed to show the robustness of the finding. e analysis showed that the result was stable except for the exclusion of three studies [18,21,33] that changed the conclusion ( Figure 7). e rationale for the use of corticosteroids is to decrease the host inflammatory responses in the lungs, which may lead to acute lung injury and acute respiratory distress syndrome [6]. Because of the high amount of cytokines induced by COVID-19 infection, corticosteroids were used frequently for the treatment of patients with severe illness, for possible benefit by reducing inflammatory-induced lung injury [24]. ere were potential harms and a lack of proven benefit for corticosteroid cautions against their routine use in patients with COVID-19 [6].
In our finding, there were no significant differences in mortality between the corticosteroid and noncorticosteroid treatment groups in critically ill COVID-19 patients. e initial clinical trial results from the United Kingdom (UK) showed that dexamethasone, a corticosteroid, can be lifesaving for patients who are critically ill with COVID-19. For patients on ventilators, the treatment was shown to reduce mortality by about one-third, and for patients requiring only oxygen, mortality was cut by about one-fifth, according to preliminary findings shared with the World Health Organization (WHO) [50]. e most commonly invoked rationale for giving steroids in patients with severe COVID-19 is to modulate the destructive inflammatory immune response that occurs with advancing disease. e guidelines recommend against using corticosteroids to try to modulate the immune system in mechanically ventilated COVID-19 patients without acute Studies included in meta-analysis (n = 32)        Low risk of bias Low risk of bias Low risk of bias a: representativeness of the exposed cohort, b: selection of the nonexposed cohort, c: ascertainment of exposure, d: demonstration that the outcome of interest was not present at the start of the study, e: comparability of cohorts based on the design or analysis, f: assessment of outcome, g: follow-up long enough for outcomes to occur, h: adequacy of follow-up of the cohort; i: random sequence generation (selection bias), j: allocation concealment (selection bias), k: blinding of participants and personnel (performance bias), l: blinding of outcome assessment (detection bias), m: incomplete outcome data (attrition bias), n: selective reporting (reporting bias).
A previous meta-analysis in persons with SARS-CoV-2, SARS-CoV, or MERS-CoV (Middle East respiratory syndrome coronavirus) infection indicated that corticosteroid did not significantly reduce the risk of death [52]. Another meta-analysis that included 5 studies indicated a lack of benefit of corticosteroid therapy on mortality in critically ill patients with COVID-19 [53]. In addition, a review that included 15 studies with coronavirus-infected patients (2 studies with COVID-19 patients) showed that corticosteroid treatment was associated with higher mortality and critical patients were more likely to require corticosteroids therapy [54].
Some of the limitations of this meta-analysis are as follows: the included studies are retrospective cohort studies, there is a high degree of between-study heterogeneity, and mortality may be influenced by other therapeutic options.

Conclusion
No significant differences in mortality between the corticosteroid and noncorticosteroid treatment groups were observed. ere was no significant reduction in mortality in critically ill COVID-19 patients treated with corticosteroids. More randomized clinical trials are needed to further verify this conclusion.

ARDS:
Acute respiratory distress syndrome CI: Confidence interval HLH: Hemophagocytic lymphohistiocytosis ICU: Intensive care unit MERS-CoV: Middle east respiratory syndrome coronavirus NOS: Newcastle-Ottawa scale PRISMA: Preferred reporting items for systematic reviews and meta-analyses RCT: Randomized controlled trial RR: Risk ratio SARS-CoV-2: Severe acute respiratory syndrome coronavirus 2 UK: United Kingdom WHO: World Health Organization.

Data Availability
e data presented are properly cited and can be obtained from already published original research articles, which are available on electronic databases.