key: cord-1055582-f73wpmye
authors: Cano, Edison J.; Fuentes, Xavier Fonseca; Campioli, Cristina Corsini; O’Horo, John C.; Saleh, Omar Abu; Odeyemi, Yewande; Yadav, Hemang; Temesgen, Zelalem
title: “Impact of Corticosteroids in COVID-19 Outcomes: Systematic Review and Meta-Analysis”
date: 2020-10-28
journal: Chest
DOI: 10.1016/j.chest.2020.10.054
sha: 2d5aea05f14717c018504c8ec233eb7a7c60a613
doc_id: 1055582
cord_uid: f73wpmye

Background Since its appearance in late 2019, infections caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have created unprecedented challenges for health systems worldwide. Multiple therapeutic options have been explored including corticosteroids (CS); preliminary results of CS in coronavirus disease 2019 (COVID-19) are encouraging, however, the role of CS is still controversial. Research Question What is the impact of CS in mortality, ICU admission, mechanical ventilation and viral shedding in COVID-19 cases? Study Design and Methods We conducted a systematic review of literature on CS and COVID-19 in major databases (PubMed, MEDLINE, and EMBASE) of published literature until July 22, 2020, that report outcomes of interest in COVID-19 patients receiving CS with a comparative group. Results A total of 73 studies with 21,350 COVID-19 cases were identified. CS use was widely reported in mechanically ventilated (35.3%), ICU (51.3%) and severe COVID-19 cases (40%). CS showed mortality benefit in severelly ill COVID-19 cases (OR 0.65, 95%CI 0.51-0.83, P=0.0006), however, no beneficial or harmful effects were noted amongst high- or low-dose CS regimens. Emerging evidence shows that low-dose CS do not have a significant impact in the duration of SARS-CoV-2 viral shedding. The analysis was limited by highly heterogeneous literature for high- and low-dose CS regimens. Interpretation Our results show evidence of mortality benefit in severely-ill COVID-19 treated with CS. CS are widely used in COVID-19 cases worldwide and a rapidly developing global pandemic warrants further high-quality clinical trials to define the most beneficial timing and dosing for CS.

The inclusion criteria were (1) peer-reviewed publications on COVID-19 only, (2) retrospective or prospective studies with more than 3 cases, (3) reporting the outcomes of interest for adult patients receiving corticosteroids in (4) all languages available. We excluded studies (1) without a comparison group to better characterize the effect of corticosteroids; (2) also excluded studies on special populations such as pregnant or pediatric patients as COVID-19 presentation and management are different in these populations; and (3) excluded studies on organ transplant recipients, inflammatory or rheumatologic patients that reported chronic corticosteroid use.

A total of 945 studies were identified after removing duplicates, 774 were excluded after initial screening.

Two investigators (EJC, CCC) independently reviewed the identified abstracts and selected articles for full review.

Discordances were resolved by a third investigator (XF). The excluded studies comprised reviews (n=275), short communications or letters (n=229), case reports with less than 4 cases (n=134), literature on pregnant or children (n=49), guidelines or society recommendations (n=47), studies not reporting outcomes on COVID-19 (n=27), and preclinical data (n=13). A total of 171 full-text studies were analyzed for eligibility and 73 peer-reviewed articles were included for qualitative and quantitative analysis ( Figure 2 ). Data extracted for each study included study design, median/mean age, country, region or hospital to assess possible population overlap, sample size, patients receiving corticosteroids, corticosteroid dose and duration, other reported therapies, whether they reported outcomes on special populations, and outcomes of interest. Quantitative meta-analysis was performed for mortality outcomes while other clinically relevant endpoints such as severity of COVID-19, ICU admission, need for mechanical ventilation, viral clearance and other adverse events were summarized in a qualitative fashion. We labeled as low-dose corticosteroids any reported dose of methylprednisolone (MP) ≤200mg daily or ≤2mg/kg/day or equivalent in other corticosteroids.

J o u r n a l P r e -p r o o f Risk of bias was determined using ROBINS-I ("Risk Of Bias In Non-randomized Studies -of Interventions") tool for non-randomized studies, 26 and version 2 of the Cochrane risk-of-bias tool for randomized trials (RoB-2). 27 Studies from same hospital or region were noted to perform sensitivity analysis in the likelihood of population overlap.

Summary risk ratios (RR) and their 95% confidence intervals (CI) were calculated using the DerSimonian and Laird random-effects model, and fixed-effect model for specific populations as deemed appropriate. 28 Heterogeneity was assessed with an I 2 statistic, where 0% indicates no heterogeneity, and 100% indicates the highest level of heterogeneity. 29 Sensitivity and subgroup analyses were performed to analyze sources of heterogeneity. Data analysis was performed using Review Manager (RevMan, [Computer program]; Version 5.4, The Cochrane Collaboration, 2020). This meta-analysis used de-identified publicly-available published data and required no Ethics Committee approval.

A total of 73 peer-reviewed articles were included for qualitative (n=55) and quantitative analysis (n=33).

Variables extracted for each publication are listed in Table 1 . [17] [18] [19] 22 ,23,30-98 All 73 studies included outcomes on COVID-19 patients receiving CS and a comparison group that did not receive CS. We were able to find 4 studies that reported outcomes of propensity-score matched populations 23,40,47,90 and one randomized clinical trial. 50 The remaining studies had limited information about baseline characteristics of their population receiving corticosteroids.

A total of 21,350 COVID-19 cases were included in the 73 studies, 4,618 (21.6%) cases received CS. The median/mean age of patients in these studies ranged from 39 (IQR 32-54) to 88 years (IQR 86.6-90). The use of CS across studies was highly variable and ranged from 1% to 97%, with a median corticosteroid use of 35.5% across studies. Most studies were generated in China (n=55, 75. Adjunctive therapies reported concomitantly with CS are reported in Table 1 , and include antibiotics, antivirals, tocilizumab, immunomodulators, traditional Chinese medicine, intravenous immunoglobulin, and convalescent plasma. Limited information was available on medication overlap for the majority of studies and qualitative synthesis was not performed.

Nineteen studies (26%) reported corticosteroid use with significant variability ranging from 1% to 100% across studies. CS use was reported in 396 of 987 (40%) severe COVID-19 cases. These numbers were interpreted as baseline characteristics rather than outcomes due to the lack of information of baseline characteristics across studies. 

Thirteen studies reported viral clearance in 1,482 COVID-19 cases receiving CS vs. no CS. The nucleic acid test results and timing were not standardized and the way to report viral clearance significantly varied among studies. Three studies that did not report CS dose concluded that patients treated with CS might have prolonged viral shedding. 30, 63, 73 Seven studies reported low-dose CS and viral clearance in 604 COVID-19 cases. Findings are summarized in Table 2 

Thirty-three of 73 studies reported mortality outcomes in patients receiving CS with a comparison group not receiving CS. One study was excluded (Ding Q, et al.) 43 due to critical risk of bias as it described outcomes in patients with COVID-19 and influenza co-infection.

We identified 32 studies comparing glucocorticoids to not administering glucocorticoids in COVID-19 cases (Figure 3) , heterogeneity was too high (I²=90%) to meaningfully combine for meta-analysis in this set with statistically significant heterogeneity (P<0.00001) with an overall detrimental effect of CS in mortality of (OR 2.30; 95%CI 1.45-3.63; P=0.0004).

We identified eight studies reporting mortality outcomes exclusively in severely ill COVID-19 (ARDS, mechanically ventilated, or critically ill) in patients receiving CS versus those who didn't ( Figure 4 ). There was low heterogeneity (I²=29%, heterogeneity P=0.19), with favorable odds of mortality (fixed-effect model) among those receiving CS, achieving statistical significance (OR 0.65, 95%CI 0.51-0.83, P=0.0006).

We also identified two studies which used high-dose corticosteroid protocols, and 15 studies specifying low dose regimens. Amongst those studies reporting higher doses ( Figure 5 ), there was low heterogeneity (I 2 =0%), but the odds of mortality (random-effects model) among those receiving high-dose CS did not achieve statistical significance (OR 0.57, 95%CI 0.27-1.23, P=0.16).

Low-dose CS were assorted with moderate heterogeneity (I 2 =60%), also with a non-significant odds (random-effects model) for mortality (OR 1.13, 95%CI 0.71-1.8, P=0.61) ( Figure 6 ). Because of concern of possible overlap of some study populations, several iterations of sensitivity analyses were performed, serially J o u r n a l P r e -p r o o f removing studies in which the same patient may have been reported more than once. None of these resulted in a meaningful change in heterogeneity metrics, nor odds of benefit or harm reaching statistical significance

In this systematic review and meta-analysis, we identified 73 comparative studies describing the experience of CS in COVID-19, which represents a considerable number of publications for a relatively new disease. There is also significant potential population overlap in studies generated in China that should be considered in future syntheses.

Overall, 21.6% of COVID-19 cases received CS in our analysis, highlighting the wide use of CS despite the lack of well-established indications or high-quality literature in favor or against CS. Almost half of studies reported dose or timing of CS, being low-dose methylprednisolone the most common approach. CS were widely used in mechanically ventilated (35.3%), ICU (51.3%) and severe COVID-19 cases (40%), which could potentially reflect a general practice rather than the impact of CS in severity of disease pending high-quality studies. There is also emerging evidence in our synthesis showing that low-dose CS do not have significant impact in duration of SARS-CoV-2 viral shedding, in contrast with data from SARS and MERS. 16 Severely ill COVID-19 cases showed a statistically significant mortality benefit from CS (OR 0.65, 95%CI 0.51-0.83, P=0.0006) in our analysis. No beneficial or harmful effect was noted amongst high-or low-dose CS.

Overall mortality of COVID-19 cases receiving CS was higher than cases not receiving CS with the caveat that the population studied was too heterogeneous, possibly due to selection bias among studies, with CS administered to cases with grave prognosis at baseline. The vast majority of studies did not report baseline characteristics of the group receiving CS.

Side effects in COVID-19 cases receiving CS included superinfection, CAPA, and hyperglycemia, however literature on side effects is lacking. Well-known corticosteroid side effects such as hyperglycemia and superimposed infections have been reported in coronavirus diseases. 99, 100 However the largest meta-analysis on low-dose corticosteroid use in patients with sepsis did not show an increased risk for superinfection (n=5356, J o u r n a l P r e -p r o o f While the role of CS in COVID-19 remains unclear, there is evidence suggesting benefits of CS in ARDS.

A meta-analysis published in 2018 in patients with ARDS receiving CS (n=494 for hydrocortisone and n=272 for methylprednisolone), showed reduced time to extubation, duration of hospitalization, and mortality; with an increase in ventilation-free days and ICU-free days. 102 The proposed doses for methylprednisolone in this setting are 1-2 mg/Kg bolus followed by the same daily dose at an infusion rate of 10 ml/hour daily with a gradual taper. 103,104 Based on similar information, the Society of Critical Care Medicine/the European Society of Intensive Care Medicine guidelines also recommended the early use of CS in moderate to severe ARDS. 105 However, the quality supporting these findings has been questioned. 106

Our qualitative synthesis was limited by the detail of reported patient's characteristics amongst studies.

There was also a lack of details in dosing, indication and timing of CS across studies. Potential for population overlap was also noted in the majority of studies generated in China, although this was mitigated by sensitivity analysis in the quantitative synthesis, it is difficult to assess the impact of the overlap in the qualitative synthesis.

Our qualitative synthesis was limited by the heterogeneity of studies included in high-and low-dose CS.

The international recommendations for corticosteroid use in COVID-19 are summarized in Table 3 

• Corticosteroids use is widely reported in COVID-19 cases worldwide, although their impact on clinically relevant outcomes in specific populations remains unclear

• Our study findings show mortality benefit for severely ill COVID-19 patients receiving corticosteroids 

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Asymptomatic COVID-19 infection management: the key to stop COVID-19

First Case of 2019 Novel Coronavirus in the United States

Emerging infectious diseases: Focus on infection control issues for novel coronaviruses (Severe Acute Respiratory Syndrome-CoV and Middle East Respiratory Syndrome-CoV), hemorrhagic fever viruses (Lassa and Ebola), and highly pathogenic avian influenza viruses, A(H5N1) and A(H7N9)

SARS-Cov-2 infection: Response of human immune system and possible implications for the rapid test and treatment

Pathogenic human coronavirus infections: causes and consequences of cytokine storm and immunopathology

The trinity of COVID-19: immunity

Corticosteroid treatment of severe acute respiratory syndrome in Hong Kong

Severe acute respiratory syndrome: report of treatment and outcome after a major outbreak

The effect of corticosteroid treatment on patients with coronavirus infection: a systematic review and meta-analysis

Clinical Characteristics of 138 Hospitalized Patients With 2019 Novel Coronavirus-Infected Pneumonia in Wuhan

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Impact of corticosteroid therapy on outcomes of persons with SARS-CoV-2, SARS-CoV, or MERS-CoV infection: a systematic review and meta-analysis

Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study

Clinical course and risk factors for mortality of adult inpatients with COVID-19

Clinical Features and Short-term Outcomes of 102 Patients with Coronavirus Disease

Clinical characteristics of Coronavirus Disease

A report of clinical diagnosis and treatment of nine cases of coronavirus disease 2019

Clinical characteristics and outcomes of older patients with coronavirus disease 2019 (COVID-19) in Wuhan, China (2019): a single-centered, retrospective study

Clinical characteristics and treatment of critically ill patients with COVID-19 in Hebei

Corticosteroids in patients with COVID-19: what about the control group?

Clinical characteristics of coronavirus disease 2019 in patients aged 80 years and older

Clinical characteristics of fatal and recovered cases of coronavirus disease 2019 in Wuhan, China: a retrospective study

The clinical characteristics of pneumonia patients coinfected with 70

Effectiveness of corticoid pulses in patients with cytokine storm syndrome induced by SARS-CoV-2 infection

Epidemiological and clinical features of 2019 novel coronavirus diseases (COVID-19) in the South of Iran

Epidemiology and clinical course of COVID-19 in Shanghai

Clinical Characteristics and Factors Associated With Long-Term Viral Excretion in Patients With Severe Acute Respiratory Syndrome Coronavirus 2 Infection: a Single-Center 28-Day Study

Clinical features of patients with coronavirus disease 2019 from a designated hospital in Beijing

The clinical value of two combination regimens in the 94

The clinical characteristics and outcomes of patients with diabetes and secondary hyperglycaemia with coronavirus disease 2019: A single-centre, retrospective, observational study in Wuhan

A Retrospective Analysis of the Clinical and Epidemiological Characteristics of COVID-19 Patients in Henan Provincial People's Hospital

Clinical characteristics of patients with 2019 coronavirus disease in a non-Wuhan area of Hubei Province, China: a retrospective study

Risk-adapted Treatment Strategy For COVID-19 Patients

Clinical characteristics of 34 COVID-19 patients admitted to intensive care unit in Hangzhou

Clinical evidence does not support corticosteroid treatment for 2019-nCoV lung injury. The Lancet

SARS: systematic review of treatment effects

Corticosteroids for treating sepsis in children and adults

The clinical practice guideline for the management of ARDS in Japan. Care Medicine (SCCM) and European Society of Intensive Care Medicine (ESICM) 2017

Guidelines on the management of acute respiratory distress syndrome

Chinese COVID-19 guidelines

COVID-19: Interim Guidance on Management Pending Empirical Evidence

Infectious Diseases Society of America Guidelines on the Treatment and Management of Patients with COVID-19

Surviving Sepsis Campaign: Guidelines on the Management of Critically Ill Adults with Coronavirus Disease 2019 (COVID-19). Crit. Care Med. 2020. 112. Corticosteroids for COVID-19

J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f