Glucocorticoid therapy in patients with COVID-19 and concurrent heart failure.

Dear Director:

We read with interest the multicentre, retrospective study performed by Salinas-Botrán et al. to identify the risk factors associated with in-hospital mortality among patients with heart failure hospitalized due to coronavirus disease 2019 (COVID-19).

It was reported from their multivariate analysis that age (adjusted odds ratio [AOR]: 1.03; 95% confidence interval [95% CI] 1.02–1.05), severe dependence (AOR: 1.62; 95% CI 1.19–2.20), baseline tachycardia (AOR: 1.01; 95% CI 1.00–1.01), baseline C-reactive protein level (AOR: 1.004; 95% CI 1.002–1.004), baseline lactate dehydrogenase level (AOR: 1.001; 95% CI: 1.001–1.002), and baseline serum creatinine level (AOR: 1.35; 95% CI 1.18–1.54) were independently associated with in-hospital mortality in their cohort of patients with heart failure hospitalized due to COVID-19. In fact, these identified risk factors of mortality are common in patients with COVID-19, including those without heart failure2, 3.

Nevertheless, based on their findings, it appears that the use of glucocorticoids, which was not incorporated into their multivariate analysis, could also be associated with in-hospital mortality in their cohort of patients. The study reported that the deceased patients had a significantly higher rate of glucocorticoid use than the patients who stayed alive during hospitalization (47.4% vs. 41.7%; p = .015). While this may be due to confounding bias, in which the use of glucocorticoids could have selected patients with higher disease severity, we took notice that the deceased patients had a significantly higher rate of development of acute decompensated heart failure than the patients who stayed alive during hospitalization (35.7% vs. 28.6%; p < .001).

Apart from their anti-inflammatory activity, glucocorticoids, especially hydrocortisone, prednisone, and prednisolone, can produce an appreciable mineralocorticoid effect, subsequently leading to fluid retention. This may be clinically insignificant in otherwise normal subjects (without heart failure) due to the phenomenon of mineralocorticoid escape that prevents progressive fluid overload. Still, patients with underlying heart disease, particularly those with congestive heart failure, may not be able to tolerate the mineralocorticoid effect of glucocorticoids, which can worsen their pre-existing fluid overload and precipitate acute decompensation of heart failure, as well as subsequent morbidity and mortality. Indeed, a recent study (n = 1155) reported that the use of glucocorticoids was associated with higher rates of in-hospital death, acute decompensated heart failure, need for invasive and non-invasive mechanical ventilation, and in-hospital complications, in patients with heart failure hospitalized for COVID-19. The findings contrast with the widely recognized mortality benefits of glucocorticoid therapy in patients with severe course of COVID-19.

Therefore, pending more investigations, we believe that caution should be exercised in the administration of glucocorticoids in patients with heart failure hospitalized for COVID-19; glucocorticoids with appreciable mineralocorticoid effect such as hydrocortisone should be avoided, while dexamethasone and methylprednisolone with no clinically important mineralocorticoid activity should be preferred when clinically indicated. Indeed, hydrocortisone can also have lower potency compared to dexamethasone in terms of anti-inflammatory activities. In addition, the short-term use of glucocorticoids with minimal mineralocorticoid action, when added to maximum diuretic therapy, can potentiate renal responsiveness to diuretic therapy in patients with congestive heart failure. Alternatively, if glucocorticoids are deemed inappropriate, interleukin-6 antagonists can be administered.

We look forward to the authors’ reply to report the types of glucocorticoids administered to their cohort of patients with heart failure hospitalized due to COVID-19. In addition, if feasible, the authors should incorporate the use of different types of glucocorticoids in their multivariate analysis to determine if the use of glucocorticoids was associated with in-hospital mortality in their cohort of patients.