key: cord-0832217-xc3prue7
authors: Escarramán-Martínez, Diego; Guerrero Gutiérrez, Manuel Alberto; Pérez-Nieto, Orlando Rubén; Zamarrón López, Eder Iván
title: Evidence of Steroids in Patients With Acute Respiratory Distress Syndrome in Coronavirus Disease 2019
date: 2020-05-11
journal: Crit Care Explor
DOI: 10.1097/cce.0000000000000124
sha: f84dffad0dddbb154b8175072526233b12f319b7
doc_id: 832217
cord_uid: xc3prue7

nan

.

When all available evidence is included, systematic reviews and meta-analyses are considered as the best quality evidence available (4) . In the application of some statistical analyses such as meta-analyses, as additional results accumulate (update of studies), increases the probability of observing false positive results (error type 1) or false negative results (error type 2) causing a phenomenon called multiplicity secondary to repeated significance tests (5) . The trial sequential analysis (TSA) it is a methodology that combines an information size calculation (cumulative number of patients, number of observations of the event of interest in the included studies or impact of the multiplicity), with an adjusted statistical significance threshold (monitoring limits or test penalty) of a metaanalysis, in order to avoid multiplicity secondary to repeated significance tests (6) .

Thirty-two studies from four meta-analyses (7-10) and the study by Villar et al (1) that compared mortality with the use of corticosteroids in ARDS were taken into account (Fig. 1) , for the construction of a single meta-analysis, using a random-effects model with the Biggerstaff-Tweedie method. Subsequently, based on the results, a TSA was constructed with a statistical significance of 95%, a probability of type 1 error (α) of 5%, a probability of type 2 error (β) of 20%, and a statistical power of 80% (1-α). For the size of the information, the required numbers of events for conclusive and reliable information was calculated with a test of bilateral significance according to the formula: IS events = P c × IS / 2 + P E × IS / 2.

PC is the expected proportion in the control group (no steroid), PE is the expected proportion in the experimental group (steroid), and IS is the information size in each group. Due to the existence of trials reporting zero events in both the experimental and control groups, an empirical continuity correction was applied in the zero event trials.

Thirty-two clinical trials included with a total of 2,749 patients with the naked eye it could be inferred that if there is a possible association in the reduction of mortality with the use of steroids (risk ratio, 0.93; 95% CI, 0.78-1.11), however, when analyzing the CI it is observed that it is short and touches the null value, which translates into an inconclusive association and despite the fact that more studies are carried out, it was not possible to improve the clinical significance. In terms of heterogeneity, there is a high proportion of variability observed in steroid use that is due to heterogeneity and not random (I 2 = 57%) and little variability in effect size between studies (Tau 2 = 0.11) (Fig. 1) . For better evidence, a TSA was constructed with the TSA Viewer software Version 0.9.5.10 Beta from the Copenhagen Trial Unit with an adjusted information size of 17,027 patients based on the result of ISevents, the cumulative curve Z does not cross statistical limits of significance (Fig. 2) creating false positive results. Therefore, with all the available evidence, it is concluded that there is no reason that justifies the use of steroids in ARDS.

Dr. Escarramán-Martínez designed the article. Dr. Guerrero Gutiérrez redacted the article. All the authors read and approved the final version of the article.

Rationale for Prolonged Corticosteroid Treatment in the Acute Respiratory Distress Syndrome Caused by Coronavirus Disease

Corticosteroids as adjunctive treatment in community-acquired pneumonia

Clinical efficacy of corticoids treatment in hospitalized adult patients with community-acquired pneumonia

Centro Cochrane Iberoamericano: ¿Está justificado el uso de corticoesteroides en el manejo de pacientes con COVID-19? 2020

Updating meta-analyses leads to larger type I errors than publication bias

Copenhagen Trial Figure 1. Forest plot. Meta-analysis of the effect of corticosteroids on mortality in patients with acute respiratory distress syndrome. Random-effects model of 32 studies with 2

Corticosteroids in the prevention and treatment of acute respiratory distress syndrome (ARDS) in adults: Metaanalysis

Exploring the heterogeneity of effects of corticosteroids on acute respiratory distress syndrome: A systematic review and meta-analysis

Use of corticosteroids in acute lung injury and acute respiratory distress syndrome: A systematic review and meta-analysis

Corticosteroid therapy for acute lung injury, acute respiratory distress syndrome, and severe pneumonia: A meta-analysis of randomized controlled trials

Trial sequential analysis of the meta-analysis. The Z value is the test statistic and |Z| = 1.96 corresponds to a p = 0.05; the higher the Z value, the lower the p value. The size of the information required to accept or reject the reduction in the relative risk of mortality with the use of corticosteroids found in the meta-analysis of the random-effects model was calculated for 17,027 patients using the diversity (D2) of 64% found, significance 95% statistic and 80% power

The authors have disclosed that they do not have any potential conflicts of interest.