Dexamethasone treatment for COVID-19 is related to increased mortality in hematologic malignancy patients: results from the EPICOVIDEHA registry

The optimal treatment strategies for hematological malignancy patients with COVID19 are still unclear with respect to the selection and timing of anti-viral as well as anti-inflammatory therapies. Most COVID-19 management recommendations have been adapted from the ones used in immunocompetent patients.^1,2 However, immunosuppressed patients often have substantial alterations in their adaptive and innate immunity that affect the pathophysiology of SARS-CoV-2 infection and often have reduced anti-viral immunity as well as dysfunctional inflammatory response. As a result, we hypothesize that these patients mainly benefit more from antiviral treatment, whereas dexamethasone may perpetuate the intrinsic immunosuppression and be even detrimental. Our study demonstrates that dexamethasone treatment for SARS-CoV-2 infection is related to increased mortality in hematological malignancy patients, even during the omicron wave with most patients being fully vaccinated. Data included were exported from the EPICOVIDEHA registry (clinicaltrials gov. Identifier: NCT04733729). The corresponding local ethics committee of each participating institution has approved the EPICOVIDEHA study when applicable. The local Institutional Review Board and Ethics Committee of the Fondazione Policlinico Universitario Agostino Gemelli—IRCCS, Università Cattolica del Sacro Cuore of Rome, Italy, approved the multicenter, non-interventional EPICOVIDEHA study (study ID: 3226). Both hospitalized and non-hospitalized patients were eligible for inclusion. Each patient was reviewed for validity following the inclusion criteria: i) patient >18 years old, ii) hematological malignancies with activity during the 5 years before COVID-19, iii) confirmed diagnosis for COVID-19 and iv) COVID-19 treatment information. Mortality rate was reported at 90 days after COVID-19 diagnosis. Classification of COVID-19 role in patient’s death was made by the reporting physician. Patients in the study population were classified as following: i) “dexamethasone only” group, for patients treated with dexamethasone exclusively, ii) “dexamethasone plus antivirals” group, for patients having received dexamethasone in addition to antivirals, and iii) in the “antiviral strategy group”, with patients treated with antivirals exclusively. With regard to antivirals regimens, in both antiviral strategy group and the dexamethasone plus antivirals group, antivirals were used in monotherapy or in combination with monoclonal antibodies and convalescent plasma. Differences between treatment groups were assessed by χ² or Fisher’s exact test. Factors associated with mortality were analyzed by Cox regression. Given the lack of randomization of therapies, a propensity score of receiving dexamethasone was estimated using a backward stepwise logistic regression model that included variables with P values ≤0.05 in the univariable analysis: age, renal dysfunction, smoking history, status of the malignancy, lymphopenia, previous COVID-19 vaccination, season of COVID-19 diagnosis and COVID-19 severity. The propensity score for receiving dexamethasone was then used as a covariable in a multivariable analysis to adjust for potential confounding factors associated with initial anti-COVID-19 treatment. The goodness of fit of the final multivariable model was assessed by the Hosmer-Lemeshow test and the area under the receiver operating characteristic curve (AUC). Sensitivity analyses were performed by repeating the propensity score approach with different methods, including 1:1 matching with replacement and a calliper of 0.25, as well as quintile stratification. A P value <0.05 was considered statistically significant. Statistical analysis was run with SPSS v25.0 (IBM Corp. Chicago, IL, USA). A total of 5,962 patients with COVID-19 and hematological malignancies were enrolled in EPICOVIDEHA registry. Finally, 2,267 patients were included in the analysis, of whom 500 (22.1%) patients were assigned to the dexamethasone only group, 470 (20.7%) to the dexamethasone plus antivirals group and 1,297 (57.2%) to the antiviral strategy group (Table 1; Online Supplementary Table S3; Online Supplementary Figure S1). Anti-SARS-CoV-2 strategies were administered based on internal criteria of the respective treatment team (Online Supplementary Tables S1, S2). Overall, day-90 mortality was 20.5% (464 patients), 9.8% (223 patients) exclusively related to COVID-19, 6.0% (137) related to both hematological malignancies and SARS-CoV-2 infection and 1.6% (36 patients) not related to the COVID-19 episode. Figure 1A-C detailed the survival probability curves for the three treatment groups of the study, regardless of the pandemic waves and in those patients with omicron infection. Figure 1B detailed the survival probability curves for the groups according to the need of hospital admission, intensive care unit (ICU) admission or outpatients’ care. In the dexamethasone only group, 138 patients (27.6%) died at the end of follow-up versus 86 patients (18.3%) in the dexamethasone plus antivirals group (P>0.001) and 55 patients (4.2%) in the antiviral strategy group (P<0.001). The independents factors associated to mortality were age, chronic liver disease, absence of neutropenia, active hematological malignancy, less than three vaccine doses, need of hospital and ICU admission (Table 2). The dexamethasone only group was an independent factor related to mortality (absolute hazard ratio [aHR]=0.562, 95% confidence interval [CI]: 0.418-0.754 in the antiviral strategy group; aHR=0.284, 95% CI: 0.191-0.422 in the dexamethasone plus antivirals group, P<0.001). This finding remained by incorporating the propensity score for receiving dexamethasone into the model. The goodness of fit was assessed by the Hosmer-Lemeshow test (P=0.099), and the discriminatory power of the score, as evaluated by the area under the curve (AUC), was 0.77 (95% CI: 0.75-0.79). The consistency of this result was confirmed by repeating the propensity score analyses by 1:1 matching with replacement and a calliper of 0.25, and by quintile stratification. The main finding of this study is that the use of dexamethasone treatment for COVID-19 was associated with the worse outcomes in patients suffering from hematological malignancies, especially when antiviral strategies were not concomitantly applied (Figure 1A-C).

Figure 1.Survival curves for the three groups of patients with different treatment strategies. (A) ALL patients. (B) Non-hospitaLized (home) patients and hospitaLized patients (non-intensive care unit [non-ICU] admitted patients and ICU-admitted patients. (C) SARS-CoV-2 Omicron variant-infected patients.

Table 1.Clinical characteristic by treatment group.

It is increasingly acknowledged that patients with COVID-19 can present with different clinical phenotypes depending on the pathophysiology complicating the infection.^3-5 Low cycle threshold values of the real-time reverse transcription polymerase chain reaction (rRT-PCR) can guide us about the fact that our patients have a high viral load. Conversely, acute elevations in C-reactive protein, ferritin, or lactate dehydrogenase (LDH) values may indicate a hyper-inflammatory syndrome. Personalizing the treatment that patients receive based on the respective clinical phenotype of COVID-19 is essential to improve the prognosis.^4,6 In this scenario, hematological patients may be different compared with immunocompetent general population. First, the process of immune-mediated viral clearance is often distorted in immunosuppressed patients leading to insufficient viral control, which may end up in long-term persistent positive PCR. Thus, the day from the onset of symptoms may not give us optimal information about the need for antivirals. Secondly, hematologic patients with malignancies have commonly pre-existing elevations in LDH and ferritin. It is therefore important not to analyze the absolute value of these markers in COVID-19 but also to consider the longer-time evolution of these inflammatory biomarkers prior and during the infection.

Since the beginning of the pandemic, hematological patients have had an increased mortality when compared to the general population.⁷, ⁸ Most factors associated with mortality identified in our study are well known.⁹ Our study helps to identify that a delay in antiviral treatment until the patient manifests severe illness and the use of dexamethasone are related to increased mortality in hematological patients with malignancies. Importantly, most patients included in this study presented COVID-19 during the predominance of SARS-CoV-2 Omicron variant.

Dexamethasone was the first drug reported as a treat pared with patients receiving usual care.¹¹ No severe immunosuppressant patients were included. Data validating these results in immunocompromised patients has been never reported. Our study provides clear real-life evidence against the general use of dexamethasone in this population, specially without antivirals, regardless of SARS-CoV-2 variant predominance. Interesting, mortality in ICU patients is very high and seems not to be influenced by detailed treatment strategies. Dexamethasone potentially diminishes type I interferon (INF) response, an endogenous cytokine essential to avoid escape of SARS-CoV-2 and it may also increase SARS-CoV-2 viral load and prolongs SARS-CoV-2 viral shedding.¹² Our study contributes additional evidence to previously documented results supporting the improving outcomes related with the use of early antiviral strategies in patients with hematological malignancies and COVID-19.^13-15 The strengths of this study are the large number of patients included, the multicenter approach and the extensive data gathered. However, there are some limitations, this study was non-randomized and non-interventional, with treatment decisions made by attending physicians. The absence of randomization introduces the potential for selection bias. Nevertheless, we employed propensity score methodology to mitigate the impact of these limitations. The retrospective design of the study may inherently result in lower data quality. Additionally, the analysis spanned a dynamic period, making it impossible to completely rule out the presence of a calendar effect on certain aspects, such as the evolving medical expertise in COVID-19. Data on the cycling time (Ct) of rRT-PCR or other surrogate viral marker (subgenomic RNA) are not available. Finally, we report the limitation of missing information on the day of starting different treatments after symptoms onset.

Table 2.Factors related with mortality in univariate and multivariate analyses.

In conclusion, this real-life large multicenter study showed the potential worse effect of dexamethasone treatment for COVID-19 in hematological patients with malignancies, even in the omicron era with most vaccinated patients. General treatment recommendations for patients with COVID-19 can be used with caution in patients with immunosuppression. New studies to provide high quality recommendations and treatment guidelines addressed to solve the specific problems of COVID-19 in patients with hematological malignancies are needed.

Footnotes

• Received December 12, 2023
• Accepted March 22, 2024

Correspondence

*TFA and JS-G contributed equally as first authors.

^#CG-V and LP contributed equally as senior authors.

C. GARCIA-VIDAL - cgarciav@clinic.cat

J. SALMANTON-GARCIA - jon.salmanton-garcia@uk-koeln.de

Disclosures

CG-V has received honoraria for talks on behalf of Gilead Science, MSD, Pfizer, Janssen, Novartis, Basilea, GSK, Shionogi, AbbVie and Advanz Pharma; and a grant support from Gilead Science, Pfizer, GSK, MSD and Pharmamar.

Funding

Acknowledgments

Acknowledged are the following collaborative group members: M. Mladenović, C. Flasshove, B. Mišković, J-M Ribera-Santa Susana, M. Hoenigl, J. Prattes, M. Mikulska, A. Cuccaro, E. Bekirova, J. Batinić, N. De Jonge, T Adžić-Vukičević, L. Drgoňa, H. M. Orth, F. Reizine, M. Piedimonte, J. Schubert, A. Soto-Silva, J. Loureiro-Amigo, L. Serrano, L. Lorenzo De La Peña, A. Guidetti, I. Ormazabal-Vélez, S. Malak, M. Calbacho, N. Fernández, R. F Duarte, E. De Kort, G. Cengiz Seval, L. Verga, R. Bergantim, M-J. Jiménez-Lorenzo, J. Maertens, N. Khanna, M. Egger, O-F. Coronel-Ayala, P. Zdziarski, A. Busca, E. Busch, C. B. Poulsen, F. Danion, T. Cushion, S. Pinzón, Y. Gonzaga, A. Kulasekararaj, H. Zarrinfar, B. Hoell-Neugebauer, C. S. Kho, R. Duléry, M. Kolditz, M. Fung and A. D. Tanase.