Target Groups for a Short Dexamethasone Course among Critically Ill COVID-19 Patients

Introduction Corticosteroids are one of the most promising therapeutic agents for critically ill patients with coronavirus disease 2019 (COVID-19). Despite emerging data, assessed populations and regimens vary, and there are patient subgroups whose response to steroids remains unclear. We aimed to evaluate the outcomes of COVID-19 patients admitted to the intensive care unit (ICU) and treated with a short dexamethasone course to determine which patient categories derive the highest benefit. Methods A retrospective cohort study was conducted using a prospectively collected single-center ICU database (April 1–October 1, 2020). Adult COVID-19 patients were assigned to dexamethasone (12 mg × 3 days) and usual care groups. Patient, management, and outcome data were extracted. The primary outcome was the 28-day ICU mortality. Subgroup analysis was performed to assess the impact of dexamethasone on mortality in patients with invasive mechanical ventilation (IMV). Results Of 233 patients, 220 (median age: 65 years, 38% female) were included: 83 patients received dexamethasone and 137 received usual care. Overall, 28 (33.7%) and 54 (39.4%) patients in the dexamethasone and usual care groups, respectively, died within 28 days since ICU admission (rate ratio (RR) 0.86; 95% confidence interval (95% CI): 0.59–1.23; p=0.405). In the IMV cohort, dexamethasone did not decrease the 28-day mortality compared with usual care (47.5% vs. 62.0%; RR 0.78; 95% CI: 0.57–1.09; p=0.107). A subgroup analysis revealed significantly lower 28-day mortality in IMV patients <65 years receiving dexamethasone vs. usual care (22.6% vs. 48.5%; RR 0.47; 95% CI: 0.22–0.98; p=0.043), which was not seen in IMV patients ≥65 years (75.0% vs. 71.1%; RR 1.06; 95% CI: 0.79–1.42; p=0.719). Patients ≥65 years experienced hyperglycemia, bacterial infection, and septic shock significantly more often than younger patients who received dexamethasone (p=0.002, p=0.025, and p < 0.001, respectively). Conclusions A 3-day dexamethasone course is not associated with lower 28-day mortality in critically ill COVID-19 patients, either in the entire ICU cohort or in the IMV. Dexamethasone may significantly reduce the 28-day mortality in IMV patients <65 years, but not in the older IMV subgroup. Dexamethasone administration in patients ≥65 years is associated with a significantly higher rate of adverse events than that in younger patients.


Introduction
e global coronavirus disease 2019 (COVID-19) pandemic continues despite widespread prevention measures and the recent introduction of effective vaccines [1][2][3][4][5][6][7]. Over 300,000 new COVID-19 cases are reported worldwide every day, and up to 18% of patients develop severe disease requiring hospitalization [8,9]. About 15-35% of hospitalized patients are treated in the intensive care unit (ICU), and up to 91% of them require invasive mechanical ventilation (IMV) [7,10]. Current data show that the hospital mortality may be up to 20%, and 40% among ICU patients; however, due to insufficient testing for COVID-19, the actual mortality rates might be even higher [7,11]. us, there is an urgent need to identify effective drugs for these patients.
Currently, corticosteroids are one of the most vigorously studied therapeutics showing overall promising results in patients with COVID-19 [12]. e growing interest for these anti-inflammatory drugs is based on their ability to accelerate resolution of acute respiratory distress syndrome (ARDS), which also presents in severe COVID-19 [7,[13][14][15]. e authors of a retrospective cohort study comprising 201 COVID-19 patients reported that treatment with methylprednisolone was associated with a lower risk of death in patients who developed ARDS [16]. On the contrary, a randomized controlled trial from Brazil including 393 COVID-19 patients did not find a difference in the 28-day mortality between methylprednisolone and placebo [17]. Dexamethasone, a well-known and widely available corticosteroid, was also actively investigated in several studies [18], one of which showed lower 28-day mortality among IMV patients and those on oxygen support than patients who received usual care [11]. However, the experimental arm in this trial received dexamethasone for up to 10 days that could increase the risk of drug-related side effects [19][20][21][22]. ereby, an early 3-day course of methylprednisolone was suggested for use, as it showed improved clinical outcomes in patients with moderate-to-severe COVID-19 [23].
Despite continuously emerging data about the use of steroids in COVID-19, the assessed populations and reported regimens vary significantly, and there are still clinical groups of patients whose response to this therapy remains unclear.
erefore, our study focused on critically ill COVID-19 patients admitted to the ICU and aimed to evaluate their outcomes following treatment with a short 3day dexamethasone course as well as determine a subgroup of ICU patients who will benefit the most from this approach.

Study Design.
We conducted a retrospective cohort study to evaluate the impact of a short 3-day dexamethasone course on outcomes of critically ill COVID-19 patients admitted to the ICU.

Setting.
is study was designed and conducted in the ICU of Clinical Hospital #1 MEDSI (Otradnoe, Krasnogorsky Rayon, Moscow Oblast, Russia), a medical center completely reorganized in March 2020 for the management of COVID-19 patients and launched on April 1, 2020. All patients were admitted to the hospital wards via the emergency department or directly to the ICU depending on their condition. e principal indication for ICU admission was respiratory failure defined as peripheral oxygen saturation (SpO 2 ) ≤90% with oxygen flow up to 10 L/min and respiratory rate ≥30/min. e usual care in the ICU included, if indicated, oxygen therapy; noninvasive respiratory support; IMV; administration of antibiotics, anticoagulant and vasopressor agents, and enteral and parenteral nutrition; as well as renal-replacement therapy. Some of the patients received hydroxychloroquine or anti-interleukin-6 receptor monoclonal antibodies (i.e., tocilizumab or sarilumab) until they were found ineffective to reduce mortality [24,25].

Ethics.
e Ethical Committee of Clinical Hospital #1 MEDSI determined that this project met the definition of human subject research and approved its conduction (protocol no. 82 dated December 28, 2020).

Data Sources, Participants, and Intervention.
A prospectively collected ICU database was reviewed from April 1 to October 1, 2020, and the follow-up continued until December 1, 2020. Male and nonpregnant female patients ≥18 years of age admitted to the ICU were included if they had a laboratory-confirmed (viral RNA detected from nasopharyngeal or oropharyngeal swab samples via reverse transcriptase-polymerase chain reaction (RT-PCR)) or clinically suspected SARS-CoV-2 infection (fever, bilateral ground-glass opacification on chest computed tomography (CT), absence of catarrhal phenomena typical for the influenza virus, negative test for influenza, and disparity of an extent of respiratory distress and percentage of lung injury on chest CT). Patients with another primary diagnosis along with those transferred to another facility and lost for follow-up were excluded. Eligible patients were assigned to two groups based on the administration of a short 3-day dexamethasone course, and their characteristics, management, and outcomes were compared between the groups. e 3day dexamethasone therapy comprised a 12 mg intravenous dose on day 1 of ICU admission followed by the same dose for 2 more days. e dose and duration of dexamethasone were chosen based on data about its effectiveness in ARDS and the risk of prolonged viral clearance in higher doses of steroids [15,20,26]. Although the decision to administer corticosteroids was primarily based on the physician's clinical judgment, working in circumstances of an unstudied disease pandemic and lack of evidence-based approaches certainly influenced these decisions. In our department, steroids were used more routinely at the beginning due to available data on non-COVID-19 ARDS, [14,15] which is reflected in the 30% rate of dexamethasone administration in April 2020 and even 70% rate in May 2020. en, we made a pause in using it so routinely and thus only 10% of patients received dexamethasone in June and 25% in July. When the RECOVERY trial was released [11], the rate of steroid administration increased back again in August and September 2020. No patients received dexamethasone or other steroids before the ICU admission.

Assessed Variables.
We retrieved the following patient characteristics: age, sex, body mass index (BMI, weight in kilograms divided by the square of height in meters), past and current medical history, vital signs, laboratory test results, percentage of lung tissue injury according to chest CT, and administered medications recorded at the time of ICU admission. Data regarding oxygen therapy (i.e., high-flow nasal cannula and noninvasive respiratory support), IMV, adverse events, and the length of ICU and hospital stay along with ICU readmission were also extracted. Registered bleeding referred to hemorrhage from the gastrointestinal, respiratory, and genitourinary tracts. Occurred thromboembolic events included deep vein thrombosis, splanchnic vein thrombosis, and pulmonary embolism. Hyperglycemia was defined as a random blood glucose level ≥10.0 mmol/L (≥180 mg/dL) [21,27]. Bacterial infection was identified by present clinical signs of systemic inflammatory response and confirmed by positive bacterial culture. Septic shock was diagnosed according to the 2016 SCCM/ESICM task force criteria in patients with a confirmed bacterial infection requiring vasopressor agents for maintaining mean blood pressure above 65 mm Hg despite fluid resuscitation [28].

Outcomes.
e primary outcome was 28-day ICU mortality, which was defined as an all-cause death within 28 days since ICU admission. Secondary outcomes included 28day mortality in the IMV cohort, extubation or decannulation rate, duration of IMV, lengths of ICU and hospital stay, and the final outcome, which was either discharge from the hospital or death from any cause.

Statistical
Analysis. Statistical analysis was conducted using the IBM SPSS Statistics for Windows software (Version 23.0; IBM Corporation, Armonk, NY). Categorical variables are reported as proportions, and continuous variables are presented as medians with interquartile range (IQR). Fisher's exact test and chisquare analysis were used for categorical variables, and the unpaired Mann-Whitney test was used for continuous variables. e primary outcome of 28-day ICU mortality was estimated using the Kaplan-Meier method. By the time of data cut-off on December 1, 2020, the outcome (discharged or deceased) of all patients was known, and discharged patients' data were censored on day 29. e impact of a short dexamethasone course on the 28-day ICU mortality was presented as a rate ratio (RR) with a 95% confidence interval (95% CI). Multivariable logistic regression analysis was performed with 28-day mortality as a dependent variable to evaluate its association with dexamethasone adjusting for other factors. All predictors were included in the analysis, and a stepwise elimination was used to build a final model. e results were presented as an odds ratio (OR) with 95% CI. For all analyses, p < 0.05 was considered to indicate statistical significance. Subgroup analysis was performed to assess the impact of dexamethasone on mortality in IMV patients.

Participants.
Of 233 patients in the database, 13 were excluded owing to having non-COVID-19 pneumonia or another primary pathology (urgent surgery, cerebral artery aneurysm rupture, meningitis, etc.); transfer to another facility; and follow-up loss. Finally, 220 patients were included in the study: 83 in the dexamethasone group and 137 in the usual care group.

ICU Management.
e median start of the studied dexamethasone course was on the 10 th (IQR: 8-13) day since the first day of fever along with a median cumulative dose of 36 (IQR: 28-40) mg. e groups were well balanced with respect to other medications and interventions (Table 2). During hospitalization, significantly more patients in the dexamethasone group received sarilumab than in the usual care group (12 (14%) vs. 8 (6%), p � 0.031). In the ICU, the need for vasopressor agents was higher in the dexamethasone (65%) group than the usual care group (48%) (p � 0.015). In the dexamethasone group, noninvasive respiratory support was used more frequently (16% vs. 2%) (p < 0.001).

Adverse Events.
Bleeding and nonventricular arrhythmia occurred more often in the dexamethasone group than in the usual care group (27% vs. 12%, p � 0.005, and 31% vs. 19%, p � 0.037; respectively) ( Table 3). On the contrary, the rates of ventricular arrhythmia, Clostridioides difficile infection, and thromboembolic events did not differ between the two groups. Hyperglycemia occurred more frequently in the dexamethasone group than in the usual care group (40% vs. 23%, p � 0.010). e bacterial infection rate was 71% in the dexamethasone group and 61% in the usual care group (p � 0.141). Septic shock developed in 39 (47%) dexamethasone patients vs. 44 (32%) usual care patients (p � 0.027p).
After the follow-up was completed, the overall ICU mortality was 41.4% and overall mortality in IMV patients was 61.5%.

Multivariable Logistic Regression Analysis.
All variables included in the multivariable logistic regression models are presented in Table 4. e final model for all patients included 168 patients and 64 events and dexamethasone use showed to be associated with lower 28-day ICU mortality (OR 0.41; 95% CI: 0.16-1.00, p � 0.049), whereas age, creatinine level, and the necessity of IMV negatively affected mortality adjusting for other factors. e final model for IMV patients included 104 patients and 55 events, and dexamethasone did not show any association with 28-day ICU mortality. In IMV patients <65 years, there were 52 patients and 17 events in the final model, and dexamethasone administration was associated with lower 28-day ICU mortality (OR 0.17; 95% CI: 0.03-0.84, p � 0.030) in this subgroup adjusting for other predictors.

Discussion
Systemic corticosteroids are currently one of the most investigated and discussed medications for COVID-19 management. However, even this approach does not ensure reduced mortality in all COVID-19 patients; therefore, it is important to determine the categories of patients who stand to benefit most from this treatment. e present study was conducted with this particular aim and a focus on critically ill patients with COVID-19 admitted to the ICU.
Data regarding the use of steroids in COVID-19 patients are inconsistent. In a multicenter CoDEX trial comprising 299 ICU patients, the use of dexamethasone did not reduce mortality in COVID-19-induced moderate-to-severe ARDS compared with the control group but significantly increased the number of ventilator-free days for the 28-day period [29]. In the MetCOVID trial, no reduction in 28-day mortality was demonstrated in hospitalized COVID-19 patients treated with methylprednisolone (0.5 mg/kg twice daily for 5 days); however, the authors noted a lower mortality in patients over 60 years [17]. By contrast, in the   RECOVERY trial, 6 mg of dexamethasone for 10 days led to lower 28-day mortality in hospitalized COVID-19 patients with the greatest reduction by 36% seen in the most critically ill patients receiving IMV [11]. In the present study, a short 3-day dexamethasone course was not associated with lower 28-day mortality in the entire ICU cohort compared to usual care. However, dexamethasone led to a significant reduction in the 28-day ICU mortality in IMV patients <65 years of age. is result is consistent with data from the RECOVERY trial (benefit of steroids is seen in patients <70 years) and a meta-analysis conducted by the WHO Rapid Evidence Appraisal for COVID-19 erapies (REACT) Working Group [18].
In our study, we also saw no decrease in mortality in the entire IMV subgroup; however, it does not contradict the RECOVERY findings. We believe it can be explained by the difference in IMV patients' age: 59 years (mean) in the RECOVERY trial vs. 65 years (median) in present study. When we evaluated the effect of dexamethasone therapy solely in a group of IMV patients <65 years with a median age of 55 years, we obtained a significant benefit and even higher (53%) mortality reduction. It should be noted that our study was carried out without limiting the patient inclusion due to strict criteria, which is typical for prospective randomized trials. For instance, in the RECOVERY trial, physicians were able to exclude severely comorbid patients whom they determined to be unsuited candidates for treatment with dexamethasone. As a result, in the group of IMV patients, the number of comorbidities was much lower than in patients who did not require oxygen upon admission.
e exclusion of patients with severe comorbid pathology might have contributed to an increase in the steroids' effectiveness in the RECOVERY trial; therefore in this study, we aimed to include as many ICU patients as possible in order to best reflect actual clinical practice. e most important concern of corticosteroid use in COVID-19 patients is a risk of adverse events, particularly prolonged viral clearance [30,31] and secondary bacterial and fungal infections [32,33] previously reported in patients with ARDS of other viral etiology. In our study, patients ≥65 years who received dexamethasone experienced hyperglycemia, bacterial infection, septic shock, and hypotension requiring vasopressor agents significantly more often than younger patients (Figure 3, Tables S5 and S6 in Supplementary Materials). One of the aims was to determine whether these differences were due to either the patients' age and their higher comorbidity rate or the result of dexamethasone use. In this respect, we assessed the outcomes of patients who did not receive dexamethasone (Tables S7 and S8 in Supplementary Material). ese complication rates were similar in patients <65 years regardless of steroid use. Moreover, there were no differences in complication rates between the patients <65 and those ≥65 years who did not receive dexamethasone despite more comorbidities in the older cohort. us, we suppose that the use of dexamethasone in critically ill ICU patients ≥65 years is associated with an increased number of adverse events that probably explains the absence of a positive impact of dexamethasone on the mortality in this subgroup of COVID-19 patients.
is study has some limitations due to its single-center retrospective design. e decision to use steroids was based   primarily on the physician's clinical judgment, which could have led to the use of dexamethasone predominantly in patients with more severe disease. Indeed, variables such as the rate of hypoxia, percentage of lung tissue injury on chest CT, levels of C-reactive protein and ferritin, and the use of noninvasive respiratory support were significantly higher in the dexamethasone group than in the usual care group, which likely explains the higher rates of intubation and IMV in the former group. ese confounding factors might have reduced the value of dexamethasone efficacy in patients requiring any level of respiratory support.

Conclusions
A short 3-day dexamethasone course is not associated with lower 28-day mortality in critically ill COVID-19 patients, either in the entire ICU cohort or in those requiring IMV. However, this therapeutic approach may reduce the 28-day ICU mortality rate in younger IMV patients, but not in the older IMV subgroup. In patients aged ≥65 years, the administration of dexamethasone is associated with a significantly higher rate of adverse events. Further randomized studies are needed to determine the most beneficial regimens of systemic steroid use in elderly patients with severe COVID-19.    Peripheral oxygen saturation qSOFA score:

Abbreviations
Quick sequential organ failure assessment score WHO: World Health Organization.
Data Availability e data sets used and analyzed during the current study are available from the corresponding author on reasonable request. e participant data were de-identified and are kept confidential.

Ethical Approval
e Ethical Committee of Clinical Hospital #1 MEDSI determined that this project met the definition of human subject research and approved its conduction without written informed consent due to the current epidemiologic situation and the exclusive use of electronic medical documentation (protocol no. 82 dated December 28, 2020).

Conflicts of Interest
Dr. Armen Oganesyan reports working as a coinvestigator with Novartis and R-pharm. Dr. Ruslan Menzulin reports working as a coinvestigator with R-pharm. Dr. Kirill Zykov is an investigator for, has received research support from, and is a consultant and/or speaker for Astra-Zeneca, Bayer, Boehringer Ingelheim, KRKA, MSD, Novartis, Solopharm, Stada, Takeda, and ermo Fisher Scientific. Other authors declare no conflicts of interest.

Authors' Contributions
AO, RM, and AN developed the study conception and design. AO and RM performed the literature review, collected and interpreted the data, and drafted the first version of the manuscript. YS contributed to data acquisition, literature review, writing, and revision of the manuscript. AN performed the statistical analysis, visualization, and contributed to manuscript drafting, its critical revision for content, and English polishing. KZ administered the project and contributed to the manuscript revision. All authors contributed to the review and amendments of the manuscript for important intellectual content and approved the final version for submission.

Supplementary Materials
Table S1: characteristics of IMV patients <65 years at ICU admission. Table S2: management and outcomes of IMV patients <65 years. Table S3: characteristics of IMV patients ≥65 years at ICU admission. Table S4: management and outcomes of IMV patients ≥65 years. Table S5: characteristics of a dexamethasone cohort at ICU admission. Table S6: management and outcomes of a dexamethasone cohort. Table S7: characteristics of a usual care cohort at ICU admission.