Early Versus Late Use of Dexamethasone in Critically Ill Patients With COVID-19: A Multicenter, Cohort Study

Rationale: It is unclear whether the timing of dexamethasone initiation is associated with positive outcomes. Objectives: To evaluate the appropriate timing of systemic dexamethasone initiation in critically ill patients with COVID-19. Methods: A multicenter, non-interventional cohort study including adults with COVID-19 patients admitted to intensive care units (ICUs) received systemic dexamethasone between March 2020 and January 2021. Patients were divided into two groups based on the timing for dexamethasone use (early vs. late). Early use is dened as the new initiation of dexamethasone within 24 hours of ICU admission. Propensity score matching was used based on the patient’s SOFA score, MV within 24 hours of ICU admission, proning status, and tocilizumab use during ICU stay. Results: A total of 480 patients were included in the study; dexamethasone was initiated early within 24 hours of ICU admission in 367 patients. Among 202 patients matched using propensity score, 101 had received dexamethasone after 24 hours of ICU admission (1:1 ratio). The 30-day mortality (OR [95%CI]: 1.82[1.04, 3.19], P=0.04) and in-hospital mortality (OR [95%CI]: 1.80[1.03, 3.15], P=0.04). Among the non-mechanically ventilated patients, late use of dexamethasone was associated with higher odds of developing respiratory failure that required MV (OR [95%CI]: 3.8 [1.41, 10.3], P=0.008) Conclusion: Early use of dexamethasone within 24 hours of ICU admission in critically ill patients with COVID-19 was associated with mortality benets. Moreover, dexamethasone's early use might be considered a proactive measure in non-mechanically ventilated critically ill patients with COVID-19, to prevent further complications. Our study aimed to assess the time of dexamethasone initiation on the clinical outcomes of critically ill patients with COVID-19. A total of 480 patients were included in the analysis; dexamethasone was initiated early in most of the included patients (76.4%). This study was started in March 2020 before releasing RECOVERY Collaborative group study results about dexamethasone use.(13) During that time, clinicians were hesitant to start dexamethasone early in all patients due to conicting evidence about steroids. After propensity score matching using the SOFA score, MV status within 24 hours of ICU admission, proning position status, and tocilizumab use during ICU stay, we found that dexamethasone early initiation was associated with mortality benets and shorter hospital LOS among the survived patients. Additionally, early initiation was associated with lower odds of respiratory failure that requires MV during ICU stay.


Introduction
In 2019, a newly discovered severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was rst reported in Wuhan, China.(1) Since the outbreak of this virus, more than 180,000,000 con rmed cases have been reported worldwide.(2) More than 4 million deaths occurred among the con rmed SARS-CoV-2 cases, with overall case mortality of almost 2% globally.(1) (1)The clinical presentation of COVID-19 ranges in severity from asymptomatic mild illness to severe pneumonia, leading to acute respiratory distress syndrome (ARDS) and associated with a high mortality rate.(3) Severe respiratory symptoms may increase the risk of hospitalization and intensive care unit (ICU) admission. (4) The ICU mortality rate is increased in patients admitted to the ICU and have severe acute lung injury ranging from 26% to 40%. (5,6) In critically ill patients with severe COVID-19, the hyperactivation of the systematic in ammatory system causes a state known as "cytokine release syndrome" (CRS). (7) This state may lead to multiple complications such as acute respiratory distress syndrome (ARDS), septic shock, and acute kidney injury (AKI), disseminated intravascular coagulation (DIC), increasing the risk of mortality in those patients. (1,8,9) Currently, several variants of SARS-CoV-2 that cause COVID-19 are discovered, yet there are limited treatment options speci c for COVID-19. The mainstay for the treatment of patients with moderate to severe COVID-19 is anti-in ammatory/antirheumatic medications, immune-based therapy, antiviral agents, and convalescent plasma. (10) Given the fact that COVID-19 patients can develop a systemic in ammatory response that can lead to lung injury and multisystem organ dysfunction, CS anti-in ammatory effect can serve as a potential therapeutic option. (11,12) The RECOVERY group multicenter, randomized, open-label trial conducted in the United Kingdom showed a signi cant mortality reduction at 28-days in hospitalized patients who received dexamethasone for up to 10 days compared to patients who received the standard of care. (13) This bene t was observed in patients who were mechanically ventilated or required supplemental oxygen at enrollment.(13) Moreover, a systemic review and meta-analysis including 20,197 patients with COVID-19 reported a signi cant reduction in mortality and ventilator-free days, the number of patients requiring mechanical ventilation for respiratory failure, and the mechanical ventilator timing. (14) Although most previous studies showed favorable clinical outcomes and mortality bene ts in patients who were initiated CS "early" in the COVID-19 treatment. (13,(15)(16)(17) the optimal timing for starting CS in critically ill patients with COVID19 is still being investigated, (18) especially dexamethasones in COVID19 critically ill patients. Since COVID-19 related lung injury and its associated hyperin ammatory and overreacting immune response occur early in ARDS presentation. We hypothesized that early initiation of dexamethasone could attenuate the in ammatory process early, leading to survival bene ts and reduction in further complications. Therefore, this study aims to evaluate the appropriate timing of systemic dexamethasone initiation in critically ill patients with COVID19 and its clinical outcomes.

Study design
This is a multicenter, non-interventional cohort study including critically ill patients aged >18-years with COVID-19 who received dexamethasone and were admitted to the ICU from March 01, 2020, until January 31, 2021. The retrospective component included de-identi ed data of COVID-19 PCR positive patients admitted before the date of IRB approval (March-June 2020). The prospective component was conducted between July 1, 2020 and January 31, 2021.The study was conducted at four hospitals in Saudi Arabia: King Abdulaziz Medical City (Riyadh), King Abdulaziz University Hospital (Jeddah), King Abdullah bin Abdulaziz University Hospital (Riyadh), and King Khalid Hospital (Hail). The primary site for this multicenter prospective study was King Abdulaziz Medical City (Riyadh).
Patients were diagnosed with COVID-19 using Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) nasopharyngeal and or throat swabs. Patients were considered eligible for dexamethasone according to the Suadi Ministry of Health (MOH) protocol near June 2020 for patients with con rmed COVID-19.(19) Eligible patients were then classi ed into two groups based on the timing of dexamethasone initiation during ICU stay to either "early" or "late." The de nition of early versus late CS initiation is debatable in critically ill patients in general and, more speci cally, in COVID-19 patients. Patients who were initiated on dexamethasone 24 hours within ICU admission were considered to have "early" CS initiation. In contrast, patients who received dexamethasone beyond that time were categorized to have "late" CS initiation. The decision of starting early versus late CS was based on the physicians' clinical judgment. The study was approved by King Abdullah International Medical Research Center (KAIMRC) in July 2020 (Ref.# RC20/430/R).

Patient Cohort
We aimed to enroll as many patients as possible, with no prede ned sample size. All included patients have received a dose of 6 mg IV once daily of dexamethasone based on MOH protocol. (13,19) Patients were excluded if the dexamethasone was initiated prior to ICU admission, use of dexamethasone interchangeably with methylprednisolone and vice versa, the ICU length of stay (LOS) ≤ one day or death within 24 hours of ICU admission, and/or labeled as "Do-Not-Resuscitate" code status within 24 hours of ICU admission ( Figure 1).

Data collection
Study data were collected and managed using Research Electronic Data Capture (REDCap®) software electronic data capture tools hosted by King Abdullah International Medical Research Center (KAIMRC) 31,32 . We collected patients demographic data (See additional le 1), comorbidities, vital signs and laboratory tests, Acute Physiology and Chronic Health Evaluation II (APACHE II), Sequential Organ Failure Assessment (SOFA), and Nutrition Risk in Critically ill (NUTRIC) scores, Glasgow Coma Score (GCS), acute kidney injury, uid balance, proning position status, the needs for mechanical ventilation (MV) and MV parameters (e.g., PaO 2 /FiO 2 ratio, FiO 2 requirement) within 24 hours of ICU admission. Also, renal pro le, liver function tests (LFTs), coagulation pro le (i.e., INR, aPTT, brinogen), and in ammatory markers (CRP, procalcitonin) within 24 hours of ICU admission were collected. Tocilizumab use was recorded for the eligible patients. All patients were followed until they were discharged from the hospital or died during in-hospital stay whichever occurred rst.

Outcomes
The primary endpoint was the in-hospital mortality in critically ill patients with COVID-19 in relation to the timing of dexamethasone initiation. The secondary endpoints were 30-day mortality, hospital LOS, ICU LOS, ventilator-free days, and, ICU-related complication (s) during ICU stay (i.e., AKI, liver injury, hospitalacquired pneumonia, secondary fungal infection, respiratory failure requires MV, and thrombosis or infarction).

Outcome de nition (s)
Primary outcome: In-hospital mortality is de ned as death occurring during the hospital stay.

Secondary outcomes :
Ventilator-free days (VFDs) at 30 days were calculated as the following: if the patients die within 30 days of MV, the VFDs = 0, VFDs = 30 − days after MV initiation (if successfully liberated from MV) , and VFDs = 0 if the patient is on MV for >30 days.
ICU complications: a. Acute kidney injury (AKI) was de ned as a sudden decrease of renal function within 48 hours, de ned by an increase in absolute SCr of at least 26.5 μmol/L (0.3 mg/dL) or by a percentage increase in SCr ≥ 50% (1.5× baseline value) during ICU stay . (20) b. Acute liver injury was de ned as alanine aminotransferase (ALT) exceeding three times the upper limit of normal or double in patients with elevated baseline ALT during the ICU stay.
c. Secondary fungal infection was identi ed through the blood, urine, wound, drainage, cerebrospinal uid (CSF), and/or respiratory cultures. The fungal growth is considered signi cant if the growth is ≥ of 100,000 CFU/ml in sputum or endotracheal aspiration shows; bronchoalveolar lavage (BAL) shows the growth ≥of 10,000 CFU of single organism/ml for protected specimen brushes (PSBs), and ≥100,000 CFU of single organism/ml for BAL uid. Additionally, urinary cultures were considered signi cant if showing a growth of ≥100,000 CFU/ml of no more than two species of microorganisms. Cultures were excluded if the laboratory reported them as a "contaminant sample" (21,22) d. Respiratory failure was de ned as either low arterial carbon dioxide tension (PaCO 2 ) or hypoxemic respiratory failure (PaO 2 < 60 mm Hg with a normal or hypercapnic respiratory failure (PaCO 2 > 50 mm Hg) that requires mechanical ventilation. e. Arterial/venous thrombosis was de ned using the International Classi cation of Diseases, Tenth Revision, Clinical Modi cation (ICD10-CM) code (i.e., myocardial infarction (MI), ischemic stroke, pulmonary embolism, deep vein thrombosis).(23)

Statistical analysis
We presented categorical variables as number (percentage), numerical variables (continuous variables) as mean and standard deviation (SD), or median and lower quartile (Q1) and upper quartile (Q3), as appropriate. The normality assumptions were assessed for all numerical variables using a statistical test (i.e., Shapiro-Wilk test) and graphical representation (i.e., histograms and Q-Q plots). We assessed model t using the Hosmer-Lemeshow goodness-of-t test. No imputation was made for missing data as the cohort of patients in our study was not derived from random selection.
We compared categorical variables using the Chi-square or Fisher exact test. We compared the normally distributed continuous variables using student t-test and other non-normally distributed continuous variables with the Mann-Whitney U test. Baseline characteristics, baseline severity, and outcome variables were compared between the two groups. Multivariate logistic and generalized linear regression were used to determine the relationship between the timing of initiation with different outcomes considered in this study. The odds ratios (OR) and incidence rate ratio (IRR) with the 95% con dence intervals (CI) were reported for the associations. We considered a P value of < 0.05 statistically signi cant, and we used SAS version 9.4 for all statistical analyses.
Propensity score matching Procedure (Proc PS match) (SAS, Cary, NC) with a 1:1 ratio were used to match patients in the early group to late group based on the patient's SOFA score, MV within 24 hours of ICU admission, proning status, and tocilizumab Use during ICU stay 4 . A greedy nearest neighbor matching method was used in patients who received early dexamethasone (control) is matched with one patient in the late group; this eventually produces the smallest within-pair difference among all available pairs with treated patients. These patients are matched only if the difference in the logits of the propensity scores for pairs of patients from the two groups is less than or equal to 0.5 times the pooled estimate of the standard deviation.

Demographic and Clinical Characteristics
A total of 790 critically ill patients with COVID-19 who received dexamethasone were screened; 480 patients were included during the study period. Out of the 480 patients, 76.4 % (367 patients) received dexamethasone early within 24 hours of ICU admission. A total of 202 patients were included after propensity score matching (1:1 ratio) based on the selected criteria ( Figure 1). The majority of the patients were men (70.8%), and the mean age of the patients was 62.9±15.4 years. The most common comorbidities were diabetes mellitus (60.9%), hypertension (61.4%), and dyslipidemia (19.8%); comorbidities were not signi cantly different between the two groups as presented in Table E1 available in the online supplement.
There was no signi cant difference between the two groups' baseline severity scores (i.e., APACHE II, SOFA, and NUTRIC scores), the needs for MV within 24 hours, baseline Oxygenation Index (OI), PaO2/FiO2 ratio, best Glasgow coma score (GCS), blood glucose level, lactic acid, platelets count, CRP, Creatine phosphokinase(CPK), ferritin, procalcitonin, D-dimer, serum creatinine and AKI within 24 hours of ICU admission after using propensity score. The median APACHE II score was 13, while the median SOFA score was 4. Additionally, the difference in tocilizumab use during ICU stay was not signi cant between the two groups, as shown in (Table E1).

Mortality and length of stay
The ICU mortality within 30 days in early groups was 42.6% compared with 57. 4    ∆ Denominator of the percentage is the total number of patientŝ^ Chi-square test is used to calculate the P-value.
Wilcoxon rank sum test is used to calculate the P-value.
$ Propensity score matched used based on patient's SOFA score, Mechanical Ventilation within 24 hours of ICU admission, Proning position, and Tocilizumab Use during ICU stay.

Discussion
Our study aimed to assess the time of dexamethasone initiation on the clinical outcomes of critically ill patients with COVID-19. A total of 480 patients were included in the analysis; dexamethasone was initiated early in most of the included patients (76.4%). This study was started in March 2020 before releasing RECOVERY Collaborative group study results about dexamethasone use. (13) During that time, clinicians were hesitant to start dexamethasone early in all patients due to con icting evidence about steroids. After propensity score matching using the SOFA score, MV status within 24 hours of ICU admission, proning position status, and tocilizumab use during ICU stay, we found that dexamethasone early initiation was associated with mortality bene ts and shorter hospital LOS among the survived patients. Additionally, early initiation was associated with lower odds of respiratory failure that requires MV during ICU stay.
In critically ill patients with COVID-19, the dysregulated in ammatory immune response observed can be counteracted by the use of CS to down-regulate the in ammatory immune response and accelerate disease resolution. (24,25) Although the World Health Organization (WHO) initially did not recommend using CS for COVID-19 treatment, as of September 02, 2020, the WHO and the National Institute of Health In our study, early dexamethasone initiation was associated with a signi cant reduction in 30-day mortality and in-hospital mortality. The majority of these patients received MV within 24 hours of ICU admission. The result of our study is consistent with the RECOVERY trial nding showing that the use of dexamethasone was associated with lower 28-day mortality among patients who received respiratory support. (13) However, in the RECOVERY trial, the median time for dexamethasone initiation since symptoms onset was 8 (5-13) days. (13) Suggesting that the bene t of dexamethasone in patients with COVID-19 is noticeable receiving dexamethasone more than seven days post-symptoms onset.(13) Also, the RECOVERY trial did not investigate if early dexamethasone initiation is different from just initiating dexamethasone at any given point in terms of hospital length of stay (LOS), ICU LOS, and MV duration.
Several meta-analyses and systemic reviews reported a signi cant reduction in the mortality rates in patients with COVID-19 receiving CS compared to standard care only. (14,29) However, none of the studies included in the meta-analyses assessed the mortality bene t of early or late CS initiation. (14,29) On the other hand, an observational retrospective study including 615 patients with COVID-19 found that starting CS at > 72 hours from admission associated with signi cant mortality reduction (HR 0.56, 95% CI 0.38-0.82; p = 0.003) compared to that early start of CS (within 24 hours). (30) The mortality bene t was evident in patients who were initiated on CS at > 7 days of symptom onset. (30) It is noteworthy that most of the study patients received methylprednisolone (87%). (30).The present study found shorter hospital LOS and longer VFD in the early group. Similarly, the study by Monedero et al. found that early CS use in critically ill COVID19 patients was associated with shorter MV duration and less ICU LOS compared to delayed or no use of CS. (16) Besides the mortality bene ts that were observed in these studies, limited studies investigated the impact of late CS initiation on respiratory failure that required MV. Our study found that early initiation of dexamethasone was associated with a lower odds of respiratory failure that required MV. That could be related to dexamethasone's prolonged and potent effect, which can mediate downregulation of systemic and pulmonary in ammation, restore homeostasis, and enhance disease resolution. Even though the NIH guideline recommends using CS in severely ill COVID-19 patients who require MV based on the evidence of the RECOVERY trial results (13,26), our study demonstrated lower odds of developing respiratory failure in non-MV patients. Monedero et al. reported a lower MV rate in the early steroids group in parallel to our ndings. (16) However, in that study, the cutoff of early initiation of CS was at 48 hours from ICU admission while including patients who were started on CS before ICU admission. (16) We believe that our multicenter prospective cohort study is one of the rst studies to highlight the appropriate time of dexamethasone initiation and its effects on the clinical outcomes of critically ill patients with COVID-19. It is prospective design allows to prospectively explore the association between the time of dexamethasone therapy initiation in COVID19 patients with ICU mortality. Additionally, it had a prede ned cutoff margin of early vs. late initiation time, and it assessed several important clinical outcomes in the nal analysis. Nevertheless, we also determined some limitations to our study. First, the observational nature of the study design limits the exclusion of missing data of some variables. Second, despite propensity score matching, some residual confounding factors are still possible. Lastly, there was a dynamic change in the national and international COVID-19 management guidelines as more evidence emerged, affecting the general practice.

Conclusion
Early use of dexamethasone within 24 hours of ICU admission in COVID-19 critically ill patients was associated with reduced mortality. Moreover, this approach might be considered a proactive measure in the non-MV critically ill patients with COVID-19 to prevent further complications. This study reinforces previous reports' ndings of the mortality bene t of CS in critically ill patients with COVID-19. Further randomized clinical and interventional studies are needed to con rm the optimal timing for CS initiation, and its bene t in non-MV COVID-19 critically ill patients.