FormalPara Key Points

Heart failure patients with a moderate or severe Charlson Comorbidity Index (CCI) score were more likely than those with a mild score to receive digoxin.

The hypothesis of equivalent protection by digoxin treatment or standard of care of heart failure patients in terms of susceptibility to COVID-19 infection appears to be supported by the data.

In the Military Health System (MHS, with ca. 9.6 million members), it cannot be excluded that digoxin and standard of care protect heart failure patients equivalently from COVID-19.

1 Introduction

Cardiac glycosides such as digoxin, digitoxin, and ouabain are still used around the world to treat patients with chronic heart failure with reduced ejection fraction (HFrEF) and/or atrial fibrillation (AF) [1,2,3,4,5,6]. However, in the United States (US), only digoxin is licensed, and it has been increasingly replaced by a standard-of-care group of drugs which include renin-angiotensin aldosterone inhibitors (ARNI), β-blockers, mineralocorticoid receptor antagonists (MRA), and the SGLT2 inhibitor. However, a recent analysis using the Medical Expenditure Panel Survey (MEPS) showed that an increasing proportion of digoxin users in the US population at large are adults living at or below the federal poverty level [7]. Consistently, the yearly median out-of-pocket cost for standard of care is US$2217.00, whereas that for digoxin is only US$60.00. Other digoxin users in this category are those who self-identify as either Asian or Hispanic [7], and those with comorbidities which are inconsistent with standard of care [8, 9]. By contrast, the MEPS analysis also showed that digoxin use had declined among females and self-identified white adults [7].

Historically, because of the comorbidity problem, digoxin has been given, by design, both here in the US and elsewhere in the world, to the sickest patients who, by definition, have a higher mortality risk [6]. This strategy has caused previous clinical trials of digoxin to be subject to prescription bias, because digoxin was, by design, associated with lower overall survival [1, 6, 10]. However, current reexaminations of these older clinical trial data have led to the suggestion that use of digoxin may lead to overall survival outcomes “equal to or slightly better” than standard of care [3, 6, 11]. Consequently, digoxin may be experiencing a recurrence of interest as reexamination of old clinical trial data has revealed that negative conclusions for both digoxin and digitoxin needed to be reconsidered [11,12,13,14,15].

However, because the COVID-19 pandemic has impacted profoundly on those with concomitant cardiovascular comorbidities [16], we considered whether digoxin-treated patients might be more susceptible to COVID-19 than those treated with standard of care. The reason for this consideration was that increased comorbidities, as defined by the Charlson Comorbidity Index (CCI), have been widely seen to be associated with worse outcomes for COVID-19 [17,18,19,20]. By contrast, it has been reported that digitoxin and ouabain, and with lesser potency digoxin, interfere with penetration of the SARS-CoV-2 virus into human lung cells [21] and green monkey kidney cells [22]. We therefore hypothesized that treatment with digoxin rather than standard of care might equivalently protect heart failure patients with regard to diagnosis of COVID-19, hospitalization and death.

2 Methods

To test this hypothesis, we conducted a cross-sectional study using the US Military Health System (MHS) Data Repository (MDR) to identify all MHS TRICARE Prime and Plus beneficiaries ages 18–64 years with a heart failure (HF) diagnosis during the COVID-19 pandemic (April 1 2020 to August 31 2021). The MDR contains administrative and clinically comprehensive data for 9.6 million beneficiaries, including active duty personnel, retirees, and their family members. In the MHS, all patients receive care through universal health coverage irrespective of rank, gender, or ethnicity. Beneficiaries aged 65 years and older were excluded from analyses due to TRICARE becoming secondary payer to Medicare. Beneficiaries associated with—either as personnel or dependents—the National Guard or Reserves were also excluded due to their inconsistent access to care.

We subsequently identified all persons in our study population diagnosed with COVID-19 or COVID-19-related conditions, and who received digoxin, β-blockers, ACE inhibitors, and angiotensin receptor blockers during the study period, as well as those who were hospitalized or died during the study period. The CCI score was retrospectively calculated by the authors for each patient at the time of their HF diagnosis using International Classification of Diseases, 10th revision (ICD-10), diagnostic codes and used to account for comorbidities and increased risk of disease. The full list of diagnostic codes used to identify heart failure and COVID-19 or related conditions can be found in the electronic supplementary material.

Study analyses included descriptive statistics on patient demographic and clinical characteristics; chi-square tests for differences in demographics and clinical characteristics between those with and without COVID-19; and unadjusted and adjusted logistic regressions, with multiple imputations for missing race, which were used to assess any demographic and clinical associations of COVID-19 diagnosis in HF patients. Twenty iterations and all patient demographics, CCI score categories, and receipt of digoxin status were used in the imputation of race. Parameter estimates and Akaike information criterion (AIC) fit statistics from full and complete case analysis models were compared with imputed results to assess the quality and fit of regression analysis. All patient demographics, CCI scores, and receipt of digoxin were used as adjustment factors in the adjusted logistic regression models. Additional subset analyses stratified by patient demographics and clinical characteristics were performed to determine impact of digoxin on patients with heart failure. All analyses were performed using SAS 9.4. This research was reviewed and determined exempt from human subjects oversight by the Uniformed Services University of the Health Sciences Institutional Review Board.

3 Results

We identified 14,044 beneficiaries with HFrEF during the study period. Of these, 10.6% were diagnosed with COVID-19; 3.5% were treated with digoxin; 27.1% were hospitalized; and 1.9% died during the study period. Among the subset diagnosed with COVID-19, 3.4% were treated with digoxin; 44.8% were hospitalized; and 3.3% died (Table 1). In addition, COVID-19 patients had systematically higher values of comorbidity above mild scores compared with those without COVID-19. Chi-square tests found differences across demographic and clinical characteristics, excluding gender, service, receipt of digoxin, ACE inhibitors and receptors, and other systemic and connective-tissue–related conditions (Table 1). However, these differences in COVID-19 diagnosis were statistically insignificant (Table 2).

Table 1 Demographics and clinical characteristics of beneficiaries with heart failure during the period April 2020 to August 2021
Table 2 Odds of COVID-19 diagnosis among MHS beneficiaries with heart failure, April 2020 to August 2021

Nonetheless, multivariate logistic regressions with imputed race revealed statistically significant associations with age group, officer ranks, and mild CCI score category. For example, all age groups under the referent of 55–64 years were more likely to have COVID-19, with ages 18–24 years having the highest odds (OR 1.63; 95% CI 1.20–2.20). Those associated with Junior (OR 0.69, 95% CI 0.51–0.92) and Senior Officer ranks (OR 0.76, 95% CI 0.63–0.93) were less likely to have COVID-19 compared with those associated with a Senior Enlisted rank. However, patients with a mild CCI score (OR 0.57; 95% CI 0.47–0.71) were also less likely to receive digoxin when compared with patients with no score.

Descriptive and inferential statistics of data from patients prescribed digoxin can be found in Table 3. Use of Chi-square tests revealed statistically significant differences across age groups, beneficiary status, rank, and CCI score. In addition, multivariate logistic regression results found statistically significant associations among beneficiary status and CCI score only. Dependents of active duty personnel with HF were less likely to receive digoxin (OR 0.50, 95% CI 0.26–0.95) compared with retirees, and those with a moderate (OR 1.70, 95% CI 1.38–2.09) or severe CCI score (OR 1.60, 95% CI 1.18–2.18) were more likely than those with a mild score to receive digoxin.

Table 3 Demographic and clinical comparisons of beneficiaries with heart failure and digoxin treatment status

4 Discussion

As hypothesized, these data indicate that digoxin, given to the HF patients with the most comorbidities, and standard of care given to those with the least comorbidities, equivalently protected against COVID-19. The hypothesis of equivalent protection by digoxin treatment of HF patients in terms of susceptibility to COVID-19 infection thus appears to be supported by the data. Whether the same can be said for digitoxin, which is a more potent inhibitor of viral penetration [21], may soon be known when analysis is completed for the DIGIT-HF trial [1]. We are also aware that ouabain, the most potent inhibitor of viral penetration, may eventually have a role to play in viral defense. However, to the best of our knowledge, no trial is ongoing.

An additional inference from this analysis, at least as it might inform regarding current digoxin treatment in the US population at large, is that the socioeconomic relegation of digoxin to the poor, to certain ethic groups, and to those with relevant comorbidities [7] does not necessarily carry with it a poor outcome for either HF treatment or resistance to COVID-19. However, an important caveat is that optimal care, as provided by the MHS to all patients, could be contributing to the equivalent outcome.

Indeed, in contrast to the US population at large, the MHS population of the youngest patients, as marked by their representation as service members and their dependents, and independent of ethnicity, race or gender, are less frequently prescribed digoxin than the older patients/beneficiaries (see Table 2). The older beneficiaries have larger numbers of comorbidities than the younger active duty service members and their families, and are therefore more likely to be prescribed digoxin. Thus, the rationale for who is prescribed digitoxin to treat HF in the US population at large is the inverse of the MHS population. In addition, we noted that HF patients treated with digoxin were also more likely to be prescribed β-blocker drugs (see Table 3).

There are limitations to this study. It is a limitation that there is a smaller number of digoxin-treated patients compared with those receiving standard of care. However, this is also generally the case for the comparison study of the US population at large [7]. Furthermore, it is a limitation that the MHS patients were all in the age range of 18–64 years. However, the racial/ethnic/gender demographics of our two groups were comparable, thus providing further support for our findings, regardless of the limitations [7]. It is also a limitation that it cannot be excluded that digoxin-dependent blockade of SARS-CoV-2 penetration into target cells [21,22,23] may have contributed to successful resistance to COVID-19 infection by HF patients receiving digoxin. However, further tests of this possibility will have to be left for future studies. Furthermore, the use of administrative claims data can lead to underestimation of prevalence and comorbidities, and there is the potential for unseen confounders or biases; both points are additional limitations of the study. Lastly, this cross-sectional study can only provide observed associations; other criteria such as experimental study designs, temporality, and repeatability from further studies are needed to prove or disprove causation.

5 Conclusion

The hypothesis of equivalent protection by digoxin treatment of HF patients in terms of susceptibility to COVID-19 infection appears to be supported by the data. In the MHS (which has ca. 9.5 million members), it cannot be excluded that digoxin and standard of care protect HF patients equivalently from COVID-19.