To the Editor:
We have read with exceptional interest the article by Villar et al (1) published in the recent issue of Critical Care Explorations.
The use of corticosteroids in the critically ill patient should be under a precise indication and not, in response to a question, that we cannot yet perform. The scenarios contemplated in the article by Villar et al (1) are acute respiratory distress syndrome (ARDS) from coronavirus disease 2019, ARDS nonviral and dysregulated systemic inflammation (cytokine storm), in which the World Health Organization does not recommend the use of corticosteroids routinely in viral pneumonia, understanding the pros and cons of the administration of corticosteroids (2,3) (Table 1T1).
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 meta-analysis, 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:
ISevents = Pc × IS / 2 + PE × 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 (I2 = 57%) and little variability in effect size between studies (Tau2 = 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.
1. Villar J, Confalonieri M, Pastores SM, et al.Rationale for Prolonged Corticosteroid Treatment in the Acute Respiratory Distress Syndrome Caused by Coronavirus Disease 2019.Crit Care Expl20202e0111
2. Sibila O, Ferrer M, Agustí C, et al.Corticosteroids as adjunctive treatment in community-acquired pneumonia.Minerva Anestesiol20148013361344
3. Tirapegui F, Díaz O, Saldías FClinical efficacy of corticoids treatment in hospitalized adult patients with community-acquired pneumonia.Rev Chil Enferm Respir201834236248
4. Centro Cochrane Iberoamericano: ¿Está justificado el uso de corticoesteroides en el manejo de pacientes con COVID-19? 2020. Manual Cochrane de Revisiones Sistemáticas de Intervenciones, Barcelona, versión 5.1.0Available at: https://es.cochrane.org/es/¿está-justificado-el-uso-de-corticoesteroides-en-el-manejo-de-pacientes-con-covid-19?
. Accessed May 6, 2020
5. Borm GF, Donders ARUpdating meta-analyses leads to larger type I errors than publication bias.J Clin Epidemiol200962825830.e10
6. Thorlund K, Engstrøm J, Wetterslev J, et al.User Manual for Trial Sequential Analysis (TSA).2017Copenhagen, DenmarkCopenhagen Trial Unit, Centre for Clinical Intervention Research1115Available at: http://www.ctu.dk/tsa/files/tsa_manual.pdf
. Accessed May 6, 2020
7. Peter JV, John P, Graham PL, et al.Corticosteroids in the prevention and treatment of acute respiratory distress syndrome (ARDS) in adults: Meta-analysis.BMJ200833610061009
8. Ruan SY, Lin HH, Huang CT, et al.Exploring the heterogeneity of effects of corticosteroids on acute respiratory distress syndrome: A systematic review and meta-analysis.Crit Care201418R63
9. Tang BM, Craig JC, Eslick GD, et al.Use of corticosteroids in acute lung injury and acute respiratory distress syndrome: A systematic review and meta-analysis.Crit Care Med20093715941603
10. Lamontagne F, Briel M, Guyatt GH, et al.Corticosteroid therapy for acute lung injury, acute respiratory distress syndrome, and severe pneumonia: A meta-analysis of randomized controlled trials.J Crit Care201025420435