Acute Decompensated Heart Failure in Patients With and Without COVID-19 — The Experience of a Romanian Center

COVID-19 is an infectious disease caused by a coronavirus similar to SARS-CoV (2002–2004 outbreak) and MERS-CoV (2012 outbreak). The SARS-CoV2 virus produces upper respiratory tract symptoms that may lead to severe acute respiratory distress syndrome and death. Since the first documented case in December 2019, COVID-19 has led to a global health crisis – proof of infectivity and high transmissibility [1].

Patients with concurrent COVID-19 and heart failure (HF) seem to be at a high risk of developing complications. Heart failure patients are more likely to be frail, older, and at an increased risk of infection while COVID-19 in itself gives rise to a systemic inflammatory response that leads to an increase in cardiac output. The combination of factors may generate either exacerbations in patients with preexisting HF or new-onset HF in previously healthy individuals [1]. COVID-19 leads to worse outcomes in patients with advanced heart failure, a fact proven by prolonged hospitalization, higher rates of ICU admission and rehospitalizations, and ultimately a higher mortality rate [2].

In Romania, the impact of the COVID-19 pandemic has not been small. Since the first confirmed case on February 26^th, 2020, there have been over 2,872,849 confirmed cases and up to 65,170 deaths with Bucharest being the highest ranking city in both regards [3,4].

The prevalence of heart failure in Romania is 4.7% of the general population over the age of 35. Patients hospitalized for HF often have a series of associated conditions such as coronary artery disease (CAD), diabetes (T2DM), arterial hypertension (HTN), and chronic kidney disease (CKD) as well as an in-hospital mortality rate of 4.9% [5].

As such, exploring the effects of COVID-19 in cohorts of Romanian heart failure patients may lead to a better understanding of the risks these patients face, as well as offer insights into potential system-wide healthcare solutions looking to lower morbidity and mortality rates. Our goal was to evaluate the HF patients admitted to our department in light of the third wave of the COVID-19 pandemic and characterize the cohort in terms of demographics, associated comorbid conditions, treatment regimens, laboratory assay and echocardiography results as well as outcome (ICU admission, mechanical ventilation, mortality).

We performed a retrospective, unicentric, cohort study on consecutive patients admitted to the Internal Medicine and Cardiology Department between September 1^st and December 1^st, 2021. Enrollment included COVID-19 positive (+) as well as COVID-19 negative (−) patients. Only patients with a heart failure diagnosis, either preexisting or established during hospitalization, were included in the subsequent analysis – patients without a heart failure diagnosis on discharge were excluded. As a result of regulations concerning COVID-19, all patients were admitted in an acute care setting, through the emergency department or as transfers from other hospitals. All patients included in the analysis were admitted because of either acute or acute decompensated HF, COVID-19 (+) status or both; no patients were admitted for routine evaluation.

Demographic data, vaccination status, medical history, treatment records (at-home treatment and treatment at discharge), laboratory assay results at admission as well as echocardiographic data were recorded. Follow-up mortality rates were calculated at 4–7 months after discharge. Vaccination status was deemed complete after two doses of Pfizer / AstraZeneca / Moderna or one dose of Johnson & Johnson. The booster shot was introduced in Romania on September 28, 2021 (during the enrollment period for this study).

Statistical analysis was performed using SPSS Statistics for Windows, Version 20.0. Data were expressed as median and range, after normality testing using the Kolomogorov-Smirnov test. Descriptive data were compared using the chi-square test, while comparisons of differences between groups were performed by using ANOVA or the independent samples t-test. Statistical significance was assumed at the 5% level.

We enrolled a total of 209 consecutive patients; the current analysis only included HF patients (n = 76), regardless of COVID-19 status: 65.3% of patients were COVID-19 (+). As the other 133 patients did not have a preexisting diagnosis of HF, nor were they diagnosed with HF based on symptoms, echocardiographic data, and NTproBNP values in accordance with the 2021 ESC Heart Failure Guidelines [6], they were not included in the analysis. Of the 76 patients evaluated, all had a diagnosis of HF at discharge; however, only 47.4% had a clearly stated diagnosis of heart failure prior to admission (based on medical records available at that time), while the other 52.6% either had not been diagnosed with HF before the current hospitalization or had unavailable medical records in regards to a previous HF diagnosis. Most patients with non-HFpEF (Heart failure with preserved ejection fraction) had a preexisting diagnosis of heart failure (38.9% vs. 2.5%, p < 0.001). About a third of patients had severe HF symptoms (NYHA III–IV): 27.6%. The patient distribution based on functional class was as follows: 60.5% NYHA I, 11.8% NYHA II, 13.2% NYHA III, 14.5% NYHA IV. Most patients with severe HF (NYHA III–IV) were non-HFpEF (i.e., HFmrEF or HFrEF): 80.0% vs. 14.8%, p < 0.001 and had a diagnosis of HF prior to hospitalization: 52.8% vs. 5.0%, p < 0.001.

The median age of the heart failure cohort was 72 years (range: 33–95 years old [y.o.]), and there was a female predominance (59.2%), with women being older than men (median [range]: 75 [51–95] y.o. vs. 68 [33–91] y.o., p = 0.001). There were no differences related to age (71 [33–91] y.o. vs. 75 [55–95] y.o., p = 0.486) or gender (58.8% vs. 60.0% women, p = 0.588) differences between COVID-19 (+) and (−) patients. Furthermore, there were no gender differences in terms of COVID-19 severity: 62.5% of women had severe COVID-19 compared to 42.1% of the men (p = 0.153).

The median length of hospital stay was 13 days, with COVID-19 (+) patients being hospitalized for longer periods of time (16 [1–38] days vs. 8 [1–22] days, p < 0.001).

The overall level of COVID-19 vaccination was relatively low: only 20.7% of all patients were fully vaccinated (3.2% had a booster shot – the booster shot only become available in Romania from September 28, 2021, during the enrollment process), while the remaining majority (78.4%) were either completely unvaccinated (76.2%) or had an incomplete vaccination status (3.2%). In our cohort, there were no differences in terms of vaccination status between COVID-19 (+) and (−) patients (fully vaccinated: 16.7% vs. 29.4%, p = 0.225), probably due to the low number of fully vaccinated patients.

Patient deterioration resulting in ICU admission occurred in 17 patients (23.3%). COVID-19 (+) patients were more likely to require ICU management: 30.6% vs. 8.3%, p = 0.030. There were no gender differences for ICU admission or mechanical ventilation: 26.7% of women required ICU management compared to 36.8% of the men (p = 0.329) while 33.3% of both men and women required mechanical ventilation (p = 0.621).

Furthermore, 16 patients (21.1%) died during hospitalization – mortality was higher among COVID-19 (+) patients: 27.5% vs. 8.0%, p = 0.044; as expected, COVID-19 (+) patients who required ICU management had the highest mortality rates (53.3% vs. 11.8%, p = 0.004). No vaccinated COVID-19 (+) patients required ICU admission (vaccinated vs. unvaccinated: 0.0% vs. 37.8%, p = 0.053), and there were fewer in-hospital deaths (12.5% vs. 28.9%, p = 0.317), though statistical significance was not achieved due probably to the small number of patients.

At follow-up, 4–7 months after hospital discharge, patients who had been COVID-19 (+) had even higher rates of mortality, while for the others mortality rates did not change (41.2% vs. 8.0%, p = 0.002). There were no gender differences in terms of mortality, either inhospital or at follow-up: 30.0% of women compared to 23.8% of men (p = 0.437) had died by the end of hospitalization, while 43.3% of women and 38.1% of men (p = 0.467) died during the follow-up period.

The odds ratio for follow-up mortality in heart failure COVID-19 (+) patients is 8.050 (CI: 1.711–37.881). The relationship between vaccination status, ICU admission, in-hospital mortality, and follow-up mortality is modeled in Figure 1.

Figure 1

In terms of heart failure diagnosis, patients were stratified according to current ESC guideline [6] recommendations into heart failure with preserved ejection fraction (HFpEF, 80.3%), heart failure with reduced ejection fraction (HFrEF, 13.2%), and heart failure with mildly reduced ejection fraction (HFmrEF, 6.6%). Because of the low number of patients, HFrEF and HFmrEF were pooled into non-HFpEF.

In regard to COVID-19 status, fewer patients with non-HFpEF were hospitalized for COVID-19 infection than HFpEF patients: 10.6% vs. 40.0%, p = 0.005 (Figure 2, Table 1). Fewer COVID-19 (+) patients had a low functional class in terms of HF severity (class III or IV NYHA): 13.7% vs. 56.0%, p < 0.001. In other words, patients with HFpEF were more likely to be hospitalized because of a COVID-19 (+) status and the diagnosis of HF established during hospitalization based on echocardiography and NTproBNP values, while non-HFpEF (HFmrEF and HFrEF) patients were more likely to be hospitalized for severe heart failure symptoms (NYHA III–IV) despite a COVID-19 (−) status. Furthermore, none of the COVID-19 (+) non-HFpEF patients required ICU management or mechanical ventilation. NTproBNP values as well as the results of other laboratory assays obtained at admission have been reported comparatively between COVID-19 (+) and (−) patients in Table 2.

Figure 2

The associated conditions most commonly encountered in our HF patients were arterial hypertension (HTN – 78.9%), followed by type 2 diabetes mellitus (T2DM – 38.2%), cancer – either past or current (18.4%), known coronary artery disease (CAD – 17.1%), late-stage chronic kidney disease (CKD – G4 and lower eGRF – 16.7%), atrial fibrillation/flutter (AF – 14.5%), and a history of stroke (11.8%). There were no statistically significant differences in terms of gender and COVID-19 status in the prevalence of associated conditions (Figure 3, Table 1).

Figure 3

The relative risk of developing COVID-19 was as follows: HTN – 1.154 (CI: 0.823 – 1.618), T2DM – 1.234 (CI: 0.867 – 1.757), AF – 1.064 (CI: 0.660 – 1.714), CAD – 1.548 (CI: 0.843 – 2.841), stroke – 0.722 (CI: 0.539 – 0.967), CKD (G2 and lower) – 1.400 (CI: 1.064 – 1.843) and cancer – 1.054 (CI: 0.688 – 1.614). Out of all associated conditions, only COVID-19 (+) patients with a history of stroke were at a higher risk of ICU admission (75.0% vs. 22.0%, p = 0.007) and mechanical ventilation (87.5% vs. 22.5%, p = 0.001), potentially overestimated due to the low absolute number of patients. For other comorbidities, there were no differences in terms of ICU admission rates or need for mechanical ventilation: HTN – ICU: 30.8% hypertensives vs. 30.0% non-hypertensives, p = 0.642, MechV: 33.3% of either hypertensives and non-hypertensives, p = 0.642; T2DM – ICU: 29.4% of diabetics vs. 31.2% of non-diabetics, p = 0.581, MechV: 41.2% vs. 29.0%, p = 0.295); CKD (G2 or lower eGFR) – ICU: 33.3% of patients with CKD vs. 28.6% of patients without CKD, p = 0.516, MechV: 39.4% vs. 23.1%, p = 0.245; CAD – ICU: 33.3% of patients with CAD vs. 30.2% of patients without CAD, p = 0.605, MechV: 50.0% vs. 31.0%, p = 0.312.

In regard to treatment, data were collected from admission and discharge records regarding the use of the following HF medication: beta blockers (BB), angiotensin-converting enzyme inhibitors (ACEi), angiotensin receptor – neprilysin inhitor (ARNi), mineralcorticoid receptor antagonist (MRA), loop diuretics (furosemide) and sodium-glucose cotransporter-2 inhitors (SGLT2-i). Overall, on admission, treatment regimens included: BB – 43.4%, ACEi/ARNi – 27.6%, loop diuretic 19.7%, MRA – 13.2%, SGLT2-i – 5.3% (30% in HFrEF patients).

Our study did not show any difference between patients with and without ACEi/ARB/ARNi before admission in terms of COVID-19 severity (45.5% of patients with treatment vs. 56.2% without had severe COVID-19, p = 0.393), need for mechanical ventilation (15.4% with treatment vs. 40.0% without required MechV, p = 0.101 N.B: the absolute number of patients, 2 vs. 14, was very low and the computation may not be valid) or mortality (in-hospital: 35.7% with treatment vs. 24.3% without died during hospitalization, p = 0.316; follow-up: 42.9% with treatment vs. 40.5% without had died by the date of the follow-up, p = 0.563). Rather, ACEi/ARB/ARNi treatment seems to have a protective effect: fewer patients with ACEi/ARB/ARNi at home were admitted as COVID-19 positive (50.0% vs. 77.1%, p = 0.015) and fewer of the COVID-19 (+) patients required ICU management (7.1% vs. 40.0%, p = 0.023).

At discharge, percentages were slightly higher: BB – 63.3%, ACEi/ARNi – 26.7%, loop diuretic – 30%, MRA – 21.7%, SGLT2-i – 7.9% (40% in HFrEF patients), without reaching statistical significance for any other class except BB (p = 0.011). A comparison in terms of HF treatment between COVID-19 (+) and (−) patients is provided in Table 3.

Up to 42% of patients did not undergo any treatment for heart failure prior to admission (probably explained by the fact that a little over half of the patients diagnosed with HF during hospitalization did not have a preexisting HF diagnosis), while only 31.5% of patients underwent treatment with a combination of at least two classes of HF drugs. Patients with HFpEF took fewer HF drugs compared to non-HFpEF patients both on admission (median [range]: 1 [0–4] vs. 2 [0–4], p = 0.002) and at discharge (1 [0–4] vs. 2 [1–5], p < 0.001). It should be noted that all non-HFpEF patients were started on at least one drug class for the treatment of HF during hospitalization.

Arterial hypertension (HTN) had a high prevalence in our patient cohort – more than two-thirds of patients – and grade 3 HTN was the most common (45.6% of hypertensive patients).

Antihypertensive treatment prior to admission included: ACEi/ARB – 40% (ACEi 28.3%), diuretics – 33.3% (indapamide 16.7%), CCB – 10%, BB – 50%. Consistent with the above-mentioned percentages, suggesting incomplete anti-HTN treatment, hypertensive patients had higher blood pressure values on admission (systolic: 140 [95–210] vs. 120 [83–130] mmHg, p = 0.001; diastolic: 80 [50–110] vs. 65 [60–85] mmHg, p = 0.003; MAP: 96.7 [65–143] vs. 83.3 [69–100] mmHg, p = 0.001). Upon discharge, however, the differences between hypertensive and non-hypertensive patients disappeared (systolic: p = 0.095; diastolic: p = 0.928; MAP: p = 0.400) – treatment distribution: ACEi/ARB – 40.8% (ACEi 26.5%), diuretics – 42.9%, CCB – 22.4%, BB – 65.3%.

Previous studies have reported a male predominance in COVID-19 infections associated with HF – potentially explained by higher levels of ACE2 in male patients [7,8,9]. However, our study found no gender differences in COVID-19 status, severity, rates of ICU admission, mechanical ventilation, nor mortality, either inhospital or at follow-up. This may be explained through the higher prevalence of HFpEF among our patients – since women are more likely to develop HFpEF [10].

The prevalence of cardiovascular conditions cited in previous publications has been highly variable: a meta-analysis evaluating the prevalence of cardiovascular diseases correlated with COVID-19 severity and ICU admission rate has shown a prevalence of 17.1%, 16.4%, 9.7% for HTN, cerebrovascular disease and diabetes, respectively, as well as an increased likelihood of developing severe symptoms that require ICU management for each of these condition [8]. Another study examining critically ill COVID-19 patients has described higher incidences of cardiovascular disease or cardiovascular risk factors: T2DM – 33.3%, CKD – 47.6%, HF – 42.9% [9]. A study performed in the Republic of Moldova showed a prevalence of CAD – 39.1%, T2DM – 13.0%, malignancy – 13.0%, cerebrovascular disease – 8.6%, CKD – 4.3% [11]. Our study cohort only examined patients with heart failure; because HTN, T2DM, CKD, and CAD are commonly associated with HF, it is not surprising that their prevalence was higher than previously cited. The fact that we did not find higher rates of ICU admission or mechanical ventilation among patients with HTN, T2DM, CKD, or CAD may be due to the fact that all patients already had varying degrees of heart failure, advanced age, and numerous cardiovascular risk factors. The fact that we had a higher percentage of COVID-19 (+) patients with a history of stroke who required ICU management or mechanical ventilation may be in part due to the low absolute number of patients with a history of stroke. However, prior large cohort studies have shown that COVID-19 (+) with a prior history of stroke tend to be older, have more comorbidities and develop worse clinical outcomes after COVID-19 [12,13].

Since the beginning of the COVID-19 pandemic, there have been concerns regarding treatment with ACEi/ARB due to their effect in upregulating ACE2, a receptor that also provides a cell entryway for SARS-CoV2. Following numerous studies, both the European Society of Cardiology and the American Heart Association have published papers stating that ACEi/ARB/ARNi treatment is safe even for COVID-19 patients [14,15]. Our study did not show any difference in COVID-19 severity, need for mechanical ventilation, or mortality (in-hospital or at follow-up between patients with and without ACEi/ARB/ARNi prior to admission. Moreover, patients who were undergoing home treatment with ACEi/ARB/ARNi were less likely to be hospitalized with COVID-19 or require ICU management.

Firstly, the small number of patients, as well as some missing laboratory results, made computations untrustworthy occasionally: D-dimer distribution had negative skewedness in favor of COVID-19 (+) patients but a similar median to COVID-19 (−) patients; different ferritin values between groups did not reach statistical significance – data unavailable for most COVID-19 negative patients (Table 2).

Secondly, most COVID-19 (+) patients were referred from other hospitals/departments due to preexisting cardiovascular conditions, while COVID-19 (−) patients were admitted in an acute care setting. Thus, the study group was rather heterogenous: COVID-19 (−) patients were more likely to be symptomatic due to HF (many with NYHA III–IV), while COVID-19 (+) patients were diagnosed incidentally during hospitalization based on echocardiographic abnormalities and elevated NTproBNP, as well as less significant HF symptoms (NYHA I–II).

Additionally, due to the high viral infectivity and transmission rate of COVID-19, the patients included in this analysis generate a selection bias: hospitalized COVID-19 patients and, even more so, COVID-19 (+) patients with preexisting cardiovascular conditions are not representative for the whole COVID-19 (+) population.

Our study evaluated a cohort of patients hospitalized with acute, decompensated HF with or without a concurrent COVID-19 infection. Heart failure with preserved ejection fraction and low symptom severity were common findings among COVID-19 (+) patients. However, COVID-19 positive patients were also hospitalized for longer periods of time, required more ICU care, and had higher death rates both in-hospital and at follow-up.