Corticosteroid Treatment in Critically Ill Patients with COVID-19: A Retrospective Cohort Study

Objectives : The role of corticosteroids in the treatment of COVID-19 is controversial. In this paper, we intend to study the application value of corticosteroid in critically ill patients with COVID-19. Methods : We collected data from 120 patients who were diagnosed with COVID-19 in the Wuhan Union Hospital. In the first part of this study, the outcome of patients given corticosteroids was compared with that of patients not given corticosteroids. The second part of the study was a matched (1:1) case–control study. After matching the baseline characteristics between the two groups, the effect of corticosteroid use on the outcome was analyzed again. Results : Analysis of data without adjusting differences in baseline characteristics indicated that the proportion of mechanical ventilation and the mortality was higher in the corticosteroid treatment group in severely ill patients. The length of hospitalization and viral shedding have no significant difference between the two groups. After adjusting the difference between the corticosteroid and non-corticosteroid treatment group, analysis revealed that the use of corticosteroids had no effect on the outcome. Conclusions : Among the critically ill patients, the use of corticosteroids has no effect on length of hospitalization, viral shedding or mortality rates. These results indicated that corticosteroid treatment


Introduction
In December, 2019, a series of pneumonia cases of SARS-CoV-2 emerged in Wuhan, Hubei, China, which spreads rapidly and the clinical symptoms are diverse. Fever, dry cough and anorexia are the main symptoms. Mild patients may only show low fever and fatigue, and severe patients may have acute respiratory distress syndrome, septic shock, multiple organ failure or even death in a short duration of time. From SARS, H1NI to MERS, the application of corticosteroids in viral pneumonia has long been controversial. Among them, Clark D Russal and colleagues believe that corticosteroids are not suitable for 2019n-Cov lung injury based on existing clinical evidence 1  In the treatment of this disease, corticosteroids was usually empirically applicated to critically ill patients, which was based on the situation of patient, besides the basically routine treatment.
However, few study have been done on the treatment effect of corticosteroids on patients with COVID-19. Therefore, the aim of this study is to propose some valuable suggestions for the application of corticosteroids in clinical work.

Study population
This single-center, retrospective study includes 120 patients who were confirmed with COVID-19 in the Wuhan Union Hospital after January 8, 2020 and were discharged or had died before April 10, 2020. The diagnostic criteria and clinical classification criteria of COVID-19 refer to the provisional WHO guidelines. This study was approved by the institutional review board of Wuhan Union Hospital.
All patients included in this study were treated with standard drugs in the isolation ward of Wuhan Union Hospital. Methylprednisolone succinate is the most common used corticosteroid. The dosage and duration of the hormone vary from person to person. All patients were given antiviral treatment, regularly monitored the patient's vital signs, laboratory and imaging changes during hospitalization, and varied degrees of oxygen therapy support according to the patient's severity.

Data collection
The following information was collected from the electronic case for included patients: epidemiology, demographic characteristics, clinical, impactology (including lung CT or chest radiography results), laboratory (blood routine, biochemistry, coagulation function, myocardial enzymes, inflammation indicators), medication and other interventions, prognostic indicator.
The main outcome event was the mortality during hospitalization. The secondary outcome event is the length of virus shedding which is defined as the event from the onset to two consecutive tests that turn negative and no more positive tests occur, and the length of hospitalization of surviving patients.

Statistic analysis
Results were analyzed with SAS software version 9.4. Categorical variables were compared by Fisher exact test or chi-square test, and continuous variables were compared by Mann-Whitney U test. All tests were tow-tailed and P <0.05 is considered statistically significant.
To reduce the effect of potential confounding in this study, we performed rigorous adjustment for differences in baseline characteristics by propensity score analysis, and then compared the outcomes between corticosteroid treatment group and non-corticosteroid treatment group.

Results
In this study, 71 patients used corticosteroids and 49 patients did not among the 120 critically ill patients. Among them, the levels of lactate dehydrogenase and C-reactive protein were higher in the corticosteroid treatment group, while the lymphocyte count was lower. Furthermore, the proportion of mechanical ventilation and mortality in the corticosteroid treatment group were higher. There were no significant differences between the two groups among age, gender, symptoms, length of hospitalization and viral shedding. (Table 1) Table 1 shows that the mortality rate in the corticosteroid treatment group was significantly higher than that in the non-corticosteroid treatment group, and it was statistically significant. However, this does not necessarily mean that adjuvant corticosteroid therapy will increase the mortality rate of patients, and this phenomenon may be partially attributed to our clinical medication "habit" of corticosteroids to critically ill patients. As can be seen in the table, the baseline characteristics of the two groups of patients were obviously unevenly distributed. Therefore, we matched the baseline characteristics between the two groups and analyzed the effects of corticosteroid on mortality, length of hospitalization and viral shedding.
There were 66 patients in 33 pairs with and without steroid treatment who were successfully matched for propensity for steroid treatment. Table 2 shows the general characteristics and outcomes of the corticosteroid treatment group and the non-corticosteroid treatment group after matching the baseline characteristics. The mortality rates during hospitalization of the corticosteroid treatment group was 6.1%, the non-corticosteroid treatment group was 9.1%. The hospitalization stay and viral shedding of the corticosteroid treatment group were longer, while none of the above-mentioned outcome events reached statistical differences. These results indicated that corticosteroid treatment did not have affect on the hospitalization stay, SARS-CoV-2 viral shedding or mortality rates of critically ill patients.

Discussion
Corticosteroids are widely utilized in the treatment of severe pneumonia, including the Middle East respiratory syndrome and severe acute respiratory syndrome. However, this therapy has been controversial so far. This study revealed that the use of corticosteroids has no effect on the hospitalization stay, SARS-CoV-2 viral shedding or mortality rates of critically ill patients. Our study also has the following limitations: ① This experiment is a retrospective study, and due to the limitation of sample size and the partial lack of patient data, it is impossible for us to match all baseline characteristics between groups; ②The time for patient nucleic acid testing is determined by the doctor-in-charge, therefore, the length of virus shedding may be limited by the frequency of specimen collection, and at the same time is also limited by the low positive rate of the detection method; ③ Our study only focused on the patients' short-term partial outcomes, superinfections, complications, and long-term side effects of corticosteroids may require further research to ensure the safety of patients receiving corticosteroid therapy.
The current application of corticosteroids in critically ill patients with COVID-19 is inconclusive. In conclusion, our analysis of 120 critically ill patients with COVID-19 in Wuhan Union Hospital showed that among the critically ill patients, the use of corticosteroids has no effect on length of hospitalization, SARS-CoV-2 viral shedding or mortality rates.

Ethics approval and consent to participate
This study was approved by the institutional review board of Wuhan Union Hospital [0124-1] and the patients consented to publication. This study does not contain any personal information that could lead to the identification of the patients.

Competing interests
The authors declare no competing interests.  Tables   Table 1 General characteristics of the patients between corticosteroid treatment and no corticosteroid treatment groups in 120 critically ill patients Data reported as n (%) or median The P value of the data comparison between the corticosteroid treatment group and the no corticosteroid treatment group co from the chi-square test Fisher exact test and the Mann-Whitney U test All the above patients' lung imaging is bilateral infiltration, and all are given antiviral treatment Data reported as n (%) or median The P value of the data comparison between the corticosteroid treatment group and the no corticosteroid treatment group co from the chi-square test Fisher exact test and the Mann-Whitney U test All the above patients' lung imaging is bilateral infiltration, and all are given antiviral treatment