During the SARS-CoV-2 pandemic, several corticosteroid regimens have been used in the treatment of the disease, with disparate results according to drug and regimen used. For this reason, we wanted to analyze differences in early mortality derived from the use of different regimens of dexamethasone and methylprednisolone in SARS-CoV-2 infection in critically ill patients requiring admission to an ICU.
MethodObservational, analytical and retrospective study, in an intensive care unit of a third-level university hospital, (March 2020 and June 2021). Adult patients (>18 years old) who were admitted consecutively for proven SARS-CoV-2 infection were included. The association with mortality in ICU at 28 days, different corticosteroid regimens used, was analyzed using a Cox proportional risk regression model.
ResultsData from a cohort of 539 patients were studied. Patient age (RR: 1.06; 95% CI: 1.02–1.10; P=<0.01) showed a significant association with 28-day mortality in the ICU. In the comparison of the different corticosteroid regimens analyzed, taking as a reference those patients who did not receive corticosteroid treatment, the dose of dexamethasone of 6mg/day showed a clear trend towards statistical significance as a protector of mortality at 28 days in the ICU (RR: 0.40, 95% CI: 0.15–1.02, p=0.05). The dose of dexamethasone of 6mg/day and low doses of methylprednisolone show a similar association with survival at 28 days (OR: 1.19; 95% CI: 0.63–2.26).
ConclusionsThe use of corticosteroids has been associated with better mortality outcomes in severe cases of SARS-CoV-2 infection. However, the therapeutic benefits of corticosteroids are not limited to dexamethasone alone.
Durante la pandemia por SARS-CoV-2, varias pautas de corticoides han sido utilizadas en el tratamiento de la enfermedad, con resultados dispares según fármaco y pauta utilizada. Por este motivo hemos querido analizar diferencias en la mortalidad temprana derivadas del empleo de diferentes pautas de dexametasona y metilprednisolona en infección por SARS-CoV-2 en pacientes graves que requieren ingreso en una unidad de cuidados intensivos (UCI).
MétodoEstudio observacional, analítico y retrospectivo llevado a cabo en una unidad de cuidados intensivos de un hospital universitario de tercer nivel, entre marzo de 2020 y junio de 2021. Se incluyeron pacientes adultos (>18 años) que ingresaron de forma consecutiva debido a una infección confirmada por el virus SARS-CoV-2. Se evaluó la asociación entre diferentes pautas de corticoides utilizadas y la mortalidad en la UCI a los 28 días mediante un modelo de regresión de riesgos proporcionales de Cox.
ResultadosCohorte de 539 pacientes donde la edad de los pacientes mostró una asociación significativa con la mortalidad a los 28 días en la UCI (RR: 1,06; IC 95%: 1,02-1,10; p<0,01). Al analizar las diferentes pautas de corticoides utilizadas, se identificó una posible tendencia en la asociación entre la dosis de dexametasona de 6mg/día y una menor mortalidad a los 28 días en la UCI, con un RR de 0,40 (IC 95%: 0,15-1,02; p=0,05). Por otro lado, las dosis de dexametasona mayores a 6mg/día y las dosis bajas de metilprednisolona mostraron una tendencia a reducir la mortalidad, aunque no alcanzaron significación estadística (RR: 0,54; IC 95%: 0,21-1,37; p=0,19 y RR: 0,51; IC 95%: 0,20-1,27; p=0,15, respectivamente).
ConclusionesEl estudio sugiere que la edad y la dosis de dexametasona de 6mg/día podrían ser factores importantes en la mortalidad a los 28 días en pacientes con infección por SARS-CoV-2 en UCI.
Since the onset of the SARS-CoV-2 pandemic, glucocorticoids have been considered a potential treatment for the disease, based on experience of their use in acute respiratory distress syndrome (ARDS),1 influenza infection,2 and other similar coronavirus infections such as severe acute respiratory syndrome (SARS-CoV) or Middle East respiratory syndrome (MERS-CoV).3,4 Their potent anti-inflammatory action has been proposed as the basis for their beneficial effect, especially in the hyperinflammatory state (“cytokine storm”).5
In this context, following the publication of the RECOVERY trial results, it was shown that the use of corticosteroids, specifically dexamethasone at a dose of 6mg per day, resulted in a significant 11% reduction in patient mortality. As a result of these findings, the World Health Organisation (WHO) announced that corticosteroids be considered the gold standard treatment for severe cases of SARS-CoV-2.6,7
However, methylprednisolone has a higher lung tissue-to-plasma ratio in experimental animals compared to dexamethasone, which may indicate a different action on lung injury.8
In light of the need for a better understanding of the effects of corticosteroids in critically ill patients with COVID-19, to provide information on the efficacy and safety of different corticosteroids in the management of critically ill patients with COVID-19, the aim of this study is to analyse the impact of dexamethasone and methylprednisolone therapy on short-term survival (28 days) of patients requiring intensive care unit (ICU) admission due to SARS-CoV-2 infection.
MethodRetrospective, observational, and analytical study which included adult patients admitted to the COVID area with a confirmed diagnosis of SARS-CoV-2 infection by polymerase chain reaction (PCR) in respiratory tract cells from the intensive care unit between March 2020 and June 2021.
Data were obtained from the patient cohort of the COVID patient registry of the intensive care medicine department, approved by the local research ethics committee (2020.250), and with patient/representative consent (written and/or telephonic).
The clinical-demographic variables of the patients were recorded: date of admission to the ICU, age in calendar years at the time of admission to the ICU, sex (dichotomous variable male or female), reason for the main clinical admission being acute respiratory failure (as dichotomous variable yes/no); comorbidities (as dichotomous variables yes/no; hypertension (HTN): diagnosed in the clinical history prior to the current admission; obesity: body mass index previously diagnosed in the clinical history or described as ≥ 30; diabetes mellitus: diagnosed in the clinical history prior to the current admission (no differentiation into type I or II); dyslipidaemia: diagnosed in the clinical history as hypercholesterolaemia and/or hypertriglyceridaemia prior to the current admission; smoker: diagnosed in the clinical history prior to the current admission as active smoker or collected during history taking from patient or relatives; biomarkers analysed (first value) as continuous variables collected in the in the first 24h from admission to ICU: creatine kinase (CK in units/litre [U/L], reference range of normality: 46–171), D-dimer (DD in ng/m, normal reference range: 0–500); determination of arterial oxygen pressure/inspiratory oxygen fraction (P/F) ratio as a continuous variable at ICU admission; therapies required during ICU admission as dichotomous variables yes/no: high-flow nasal cannula (HFNC), mechanical ventilation (MV), anticoagulation dose with prophylactic characteristics according to patient weight, use of antibiotherapy for prophylactic purposes, use of prone position as a therapeutic measure, use of corticosteroids as antiviral treatment, need for vasoactive and/or inotropic drugs due to septic shock, use of continuous renal replacement therapy (CRRT) due to renal failure, and evolutionary variables: days of mechanical ventilation in days as continuous variable, development of pulmonary thromboembolism (PTE) confirmed by computed tomography (CT) as dichotomous variable yes/no, development of ventilator-associated tracheobronchitis and ventilator-associated pneumonia (VAT and VAP) as dichotomous variable yes/no, need for tracheostomy (according to clinical criteria) due to prolongation of MV as dichotomous variable yes/no, ICU stay in days as continuous variable.
Patients were categorised into 4 groups, 1) patients who did not receive corticosteroid treatment during their stay in the ICU, 2) patients treated with dexamethasone 6mg daily for up to 10 days,6 3) patients who received dexamethasone in doses higher than 6mg daily,9 and 4) patients managed with low doses of methylprednisolone (40–80mg daily) for a period of 3 to 5 days.10,11
We performed a descriptive analysis of the sample: for continuous variables, the Shapiro–Wilk test was used as a test of normality. Continuous variables with a normal distribution are described as mean and standard deviation (SD), while continuous variables with a non-normal distribution are expressed as median and their 25–75 percentiles (p25–75). Pearson’s χ2 test or Fisher’s exact test were used to compare between groups to compare proportions, and the Kruskal–Wallis test to compare continuous variables.
A Cox proportional hazards regression model was used in the study of risk factors for 28-day ICU mortality, entering variables with a value of p<.1 in the bivariate analysis against 28-day mortality. Results are shown by the risk ratio (RR) and its 95% confidence interval (95% CI) for Cox regression.
Survival analysis was performed using the Kaplan–Meier method and comparisons were made using the long-rank test.
Statistical software MedCalc®, version 19.5.3 (MedCalc Software Ltd, Ostende, Belgium; https://www.medcalc.org; 2020) was used for calculations and statistical analyses.
ResultsOver the study period, 565 patients were admitted to the ICU with a diagnosis of SARS-CoV-2 infection or high suspicion of SARS-CoV-2. Twenty-six patients were excluded because SARS-CoV-2 infection could not be microbiologically confirmed. The main baseline and clinical-epidemiological characteristics of the remaining 539 patients are shown in Table 1.
Baseline clinical-epidemiological characteristicsof patients admitted with confirmed SARS-CoV-2 pneumonia.
| Variables | n=539 |
|---|---|
| Age (mean [SD]) | 63 (14) |
| Sex | |
| Males (n [%]) | 368 (68.3%) |
| Main comorbidities | |
| HTN (n [%]) | 274 (50.8%) |
| DM (n [%]) | 117 (21.7%) |
| Obesity (n [%]) | 113 (21.0%) |
| Dyslipidaemia (n [%]) | 188 (34.9%) |
| Smoker (n [%]) | 170 (31.5%) |
| SOFA score on admission to ICU (mean [SD]) | 4 (3) |
| PaO2/FiO2 on admission to ICU (median [q25–75]) | 139 (103–180) |
| Laboratory data on admission to ICU | |
| Creatine kinase (median [q25–75]) | 109 (54–302) |
| D-dimer (median [q25–75]) | 917 (553–1883) |
| Treatment used during ICU stay | |
| Tocilizumab and corticosteroid combination immunomodulatory therapy (n [%]) | 395 (73.3%) |
| Use of prone positioning (n [%]) | 233 (57.2%) |
| Use of HNFC on admission to ICU (n [%]) | 292 (58.8%) |
| Need for IMV (n [%]) | 370 (70.1%) |
| Need for vasopressors (n [%]) | 242 (44.9%) |
| Use of ECMO (n [%]) | 5 (.9%) |
| Use of iNO therapy (n [%]) | 32 (5.9%) |
| Need for RRT (n [%]) | 25 (4.6%) |
| Hyperimmune plasma treatment | 180 (33.4%) |
| Treatment with remdesivir | 64 (11.9%) |
| Stay in ICU in days (median [q25–75]) | 9 (5–18) |
| Mortality in ICU at 28 days (n [%]) | 70 (13.1%) |
DM, diabetes mellitus; ECMO, extracorporeal membrane oxygenation; HT, hypertension; HFNC, high-flow nasal cannula; ICU, intensive care unit; IMV, invasive mechanical ventilation; iNO, inhaled nitric oxide; RRT, renal replacement techniques; SD, standard deviation; SOFA, Sequential Organ Failure Assessment score.
As shown in Table 2, age showed a significant difference between the corticosteroid treatment groups (p=.03), and no significant differences were found in the main comorbidities, except for obesity (p<.01). Regarding laboratory data on ICU admission, significant differences were observed in creatine kinase (p<.01) and D-dimer (p<.01) values between treatment groups. With regard to adjuvant antiviral treatments and therapies during ICU admission, significant differences were observed. Plasma use was more frequent in the group treated with dexamethasone at a dose of 6mg/day (p<.01), while the need for invasive mechanical ventilation (IMV) was higher in the group treated with methylprednisolone (p<.01). ICU stay in days showed a significant difference between treatment groups (p<.01). In addition, 28-day ICU mortality was significantly lower in the groups treated with dexamethasone at a dose of 6mg/day (p=.01) and methylprednisolone (p<.01) compared to the group that did not receive corticosteroid treatment.
Main basic clinical differences between the four groups analysed in the study.
| Variables | No corticosteroid use | Dexamethasone 6mg/day | Dexamethasone >6mg/day | Methylprednisolone 40–80mg/day | p-Value* |
|---|---|---|---|---|---|
| N=123 | N=196 | N=91 | N=129 | ||
| Age (median [p25–75]) | 66 (54–73) | 64 (54–71) | 69 (57–74) | 68 (58–74) | .03 |
| Sex | |||||
| Males (n [%]) | 79 (64%) | 129 (66%) | 64 (70%) | 96 (74%) | .27 |
| Main comorbidities | |||||
| HTN (n [%]) | 57 (46%) | 106 (54%) | 39 (43%) | 72 (56%) | .14 |
| DM (n [%]) | 28 (23%) | 47 (24%) | 15 (16%) | 27 (21%) | .53 |
| Obesity (n [%]) | 12 (10%) | 50 (25%) | 14 (15%) | 37 (29%) | <.01 |
| Dyslipidaemia (n [%]) | 40 (32%) | 70 (36%) | 34 (37%) | 44 (34%) | .88 |
| Smoker (n [%]) | 33 (27%) | 63 (32%) | 28 (31%) | 46 (36%) | .50 |
| PaO2/FiO2 on admission to ICU (median [q25–75]) | 112 (89–184) | 146 (117–180) | 138 (99–177) | 145 (116–187) | .02 |
| Laboratory data on admission to ICU | |||||
| Creatine kinase (median [q25–75]) | 308 (92–705) | 101 (56–215) | 72 (43–193) | 81 (47–187) | <.01 |
| D-dimer (median [q25–75]) | 4.545(1.567–19.000) | 743 (522–1.148) | 899 (543–1.734) | 780 (478–1.310) | <.01 |
| Adjuvant antiviral treatments | |||||
| Remdesivir n (%) | 12 (10%) | 28 (14%) | 8 (9%) | 16 (12%) | .47 |
| Plasma | 13 (10%) | 86 (44%) | 30 (33%) | 51 (39%) | <.01 |
| Immunomodulatory treatment with tocilizumab (n [%]) | 14 (11%) | 33 (17%) | 17 (19%) | 32 (25%) | .04 |
| Treatment during admission to ICU | |||||
| Use of HFNC on admission to ICU (n [%]) | 32 (26%) | 137 (70%) | 44 (48%) | 79 (61%) | <.01 |
| Need for IMV (n [%]) | 84 (68%) | 135 (69%) | 76 (83%) | 75 (58%) | <.01 |
| Prone positioning (n [%]) | 61 (49%) | 74 (38%) | 51 (56%) | 47 (36%) | <.01 |
| Need for vasopressors (n [%]) | 56 (45%) | 87 (44%) | 37 (41%) | 62 (48%) | .72 |
| Use of ECMO (n [%]) | 2 (2%) | 3 (1%) | 0 (0%) | 0 (0%) | .32 |
| Use of iNO therapy (n [%]) | 7 (6%) | 15 (8%) | 6 (6%) | 4 (3%) | .39 |
| Need for RRT (n [%]) | 12 (10%) | 7 (3%) | 1 (1%) | 5 (4%) | .02 |
| VAP (n [%]) | 11 (9%) | 40 (20%) | 24 (26%) | 14 (11%) | <.01 |
| VAT (n [%]) | 13 (10%) | 18 (9%) | 12 (13%) | 17 (13%) | .71 |
| Stay in ICU in days (median [p25–75]) | 8 (3-16) | 9 (5-18) | 11 (8-19) | 8 (4-15) | <.01 |
| Mortality in ICY at 28 days (n [%]) | 26 (21%) | 19 (10%) | 13 (14%) | 12 (9%) | .01 |
DM, Diabetes Mellitus; ECMO, Extracorporeal Membrane Oxygenation; HFNC, High-flow nasal cannula; HTN, Hypertension; ICU, Intensive Care Unit; IMV, Invasive mechanical ventilation; iNO, inhaled nitric oxide; RRT, Renal replacement techniques; VAT/VAP, Ventilator-associated tracheobronchitis/ pneumonia.
For the dexamethasone dose of 6mg/day, an RR of .40 (95% CI .15–1.02, p=.05) was found, suggesting an association with lower short-term mortality. For a dexamethasone dose greater than 6mg/day, the RR was .54 (95% CI: .21–1.37; p=.19), indicating a trend towards a beneficial effect, although it was not statistically significant. For the methylprednisolone dose (40–80mg/day), an RR of .51 (95% CI: .20–1.27; p=.15) was obtained, which was not statistically significant either. Regarding patient characteristics, age showed a significant association with short-term mortality in the ICU (RR: 1.06; 95% CI: 1.02–1.10; p<.01). For each year of life, a slight increase in mortality risk was observed. However, the presence of obesity did not show a significant association with mortality (RR: 1.66; 95% CI: .76–3.62; p=.19) (Table 3).
Data from the Cox proportional hazards regression model in the study of risk factors for 28-day ICU mortality, introducing variables with a p-value of <.1 in the bivariate analysis against 28-day morality.
| Variables | RR | 95% CI | p-Value |
|---|---|---|---|
| Dexamethasone dose of 6mg/daya | .40 | .15–1.02 | .05 |
| Dexamethasone dose of >6mg/daya | .54 | .21–1.37 | .19 |
| Methylprednisolone dose of (40–80mg/day)a | .51 | .20–1.27 | .15 |
| Age (per year of life) | 1.06 | 1.02–1.10 | <.01 |
| Presenting obesity | 1.66 | .76–3.62 | .19 |
| Creatine kinase value | 1.00 | .99–1.00 | .15 |
| D-dimer value on admission | 1.00 | 1.00–1.00 | .94 |
| PaO2/FiO2 on admission to ICU | .99 | .99–1.00 | .88 |
| Use of tocilizumab | .64 | .30–1.37 | .25 |
| Use of hyperimmune plasma | .69 | .33–1.42 | .32 |
| Prone positioning therapy | .74 | .37–1.46 | .39 |
| Use of RRT | .85 | .28–2.51 | .77 |
| Onset of VAP | 1.02 | .51–2.03 | .95 |
| Use of HFNC on admission to ICU | .80 | .42–1.52 | .50 |
| Need for IMV during admission | .70 | .28–1.75 | .43 |
95% CI, 95% Confidence Interval; HFNC, High-flow nasal cannula; ICU, Intensive Care Unit; IMV, Invasive mechanical ventilation; RR, Relative risk; RRT, Renal replacement techniques; VAP, Ventilator-associated pneumonia.
In the survival analysis, comparison using Kaplan–Meier curves at 28 days follow-up, showed a difference between groups as seen in Fig. 1 (long-rank test: p≤.01).
Survival analysis (Kaplan–Meier) referring to survival at 28 days in the 4 groups established. Survival curves at 28 days for patients categorised into 4 groups: 1) patients who did not receive corticosteroid treatment during their ICU stay; 2) patients treated with dexamethasone 6mg daily for up to 10 days; 3) patients who received dexamethasone in doses higher than 6mg daily, and 4) patients managed with low doses of methylprednisolone (40–80mg daily) for a period of 3 to 5 days.
In our analysis, 1) patients who received corticosteroid treatment had a longer survival at 28 days compared to patients who did not receive corticosteroid treatment, and 2) both dexamethasone at a dose of 6mg/day and low-dose methylprednisolone appear to be effective in the treatment of severe SARS-CoV-2 disease requiring ICU admission, in terms of 28-day mortality.
However, in our analysis it should be considered that the group of patients not treated with corticosteroids show significantly higher severity criteria (lower PaO2/FiO2 ratio, higher CK, and DD) on admission to the ICU.
There is heterogeneity in the use of glucocorticoids and their effect on mortality in acute lung injury induced by viral diseases.12–16 High doses of glucocorticoids may delay viral clearance and increase mortality and the risk of secondary infections.17,18 However, in patients with severe SARS-CoV-2 disease, short-duration glucocorticoid regimens have proven beneficial and safe, and do not behave as an independent risk factor for prolonged viral RNA shedding.19,20
After publication of the RECOVERY trial,6 dexamethasone became the standard treatment for the most severely ill patients. The study showed that the use of dexamethasone resulted in lower mortality among patients undergoing mechanical ventilation or receiving high concentrations of oxygen, but not among those not receiving respiratory support.
The effect of corticosteroids as a whole was studied using data from a prospective meta-analysis that pooled data from 7 randomised clinical trials (RCTs). In their results, the odds ratio (OR) for association with mortality was .64 (95% CI .50–.82, p<.001) for dexamethasone compared to usual care or placebo (3 trials, 1282 patients, and 527 deaths), while it was .69 (95% CI .43–1.12; p=.13) for hydrocortisone (3 trials, 374 patients, and 94 deaths) and .91 (95% CI .29–2.87; p=.87) for hydrocortisone and methylprednisolone respectively (1 trial, 47 patients and 26 deaths).21
A subsequent comprehensive analysis of 11 RCTs with 8019 hospitalised participants with COVID-19 evaluated the efficacy of systemic corticosteroids together with standard care compared to standard care alone, or standard care plus placebo. We found that systemic corticosteroids together with standard care probably slightly reduce all-cause mortality up to 30 days, although the evidence on mortality up to 120 days was unclear. In addition, systemic corticosteroids appear to slightly increase the likelihood of clinical improvement at 28 days. For serious adverse events, nosocomial infections and invasive mycoses, there was insufficient evidence for conclusive analyses. When comparing different types and doses of corticosteroids, we found that methylprednisolone compared to dexamethasone showed no significant difference in mortality. High-dose dexamethasone (12mg or more) could reduce mortality up to 30 days, but no clear conclusion was found on mortality up to 120 days and clinical improvement at 28 days. Subgroup analyses were performed to explore equity in the effects of corticosteroids. It was observed that patients under 70 years might benefit more from corticosteroids compared to those aged over 70 years.22
The limitations of our study are those inherent to retrospective studies: under-reporting and variability among professionals. In this regard, the lack of any variable (especially analytical tests) or a more specific categorisation of some of them (for example, CK or DD values) that could influence the final results should be considered a limitation of the study. In addition, these data are analysed from a single centre, with the difficulty in external validation of the conclusions that this always entails. Another relevant limitation of this study is the absence of an explicit consideration of the possible impact of the timing of patient admission in relation to mortality outcomes.
In this study, certain variables relevant to corticosteroid treatment in patients hospitalised with COVID-19 were identified and analysed. However, it is possible that other variables not considered in this analysis may have had an impact on the results. Some potential additional variables that could have had an influence include differences in patient comorbidities, disease duration and severity, individual response to corticosteroids, and the presence of other concurrent therapies or medications. In addition, genetic characteristics and access to care may also play a role in the variability of outcomes.
In order to improve the understanding and applicability of these findings, future research should consider including a wider range of relevant variables and conducting more detailed analyses to better understand the impact of corticosteroids in patients with COVID-19.
Based on the data presented and the analysis of the literature, we can conclude that the use of systemic corticosteroids in conjunction with standard care may have a beneficial effect in reducing short-term mortality in patients hospitalised with COVID-19. However, longer-term effects and adverse events require further investigation. In addition, outcomes may vary in different patient subgroups, emphasising the importance of considering equity factors in clinical decision-making.
Key concepts- •
During the COVID pandemic, no standard/uniform corticosteroid regimen was established for the treatment of severely ill patients.
- •
Dexamethasone and methylprednisolone were the most commonly used corticosteroids in different regimens.
- •
Patients who received corticosteroid treatment had longer survival at 28 days compared to patients who did not receive corticosteroid treatment
- •
A possible efficacy of both dexamethasone at a dose of 6mg/day and low-dose methylprednisolone in the treatment of severe SARS-CoV-2 disease requiring ICU admission is suggested in relation to 28-day mortality.
No funding was received for this manuscript.
Conflict of interestsThe authors have no conflict of interests to declare.
