Introduction

The ongoing pandemic virus disease 2019 (COVID-19) has presented an unprecedented challenge for the medical communities in an effort to develop suitable therapeutic management for viral infection in order to provide optimal medical care. Data from China showed that, although respiratory failure is the dominant and the worst clinical manifestation of COVID-19, a large number of patients would present with preexisting cardiovascular disease (CVD) or develop new-onset cardiac dysfunction during the course of the viral infection.^1–4 Several studies have noted that cardiac injury and cardiovascular (CV) involvement during COVID-19 infection were associated with more severe outcomes.^5,6

In addition, there are limited data regarding the cardiac safety of hydroxychloroquine (HCQ), often in combination with macrolides and antivirals as treatment medications in COVID-19 patients. In fact, the use of these drugs has been associated with QT prolongation and then with potential risk of ventricular tachyarrhythmia.^7–12 Therefore, the degree of cardiac involvement of COVID-19 is variable and requires the development of standardized correct management. ECG changes during in-hospital stay may reflect cardiac expression of illness.

This study is aimed at detecting the development of CV manifestations, with particular regard to the electrocardiographic characteristics at baseline and their changes after medical therapy in a cohort of noncritically ill patients affected by COVID-19.

Methods

Population

In this observational study, we consecutively enrolled 60 patients hospitalized for COVID-19 pneumonia at ‘Tor Vergata Hospital’ of Rome, Italy, from 16 March to 11 May. SARS-CoV2 infection was diagnosed according to the WHO guidance, after positive results of real-time reverse transcription polymerase chain reaction assay of nasal and pharyngeal swabs.¹³ All patients included had noncritical illness, whose management did not require intensive care throughout the whole hospitalization. All patients were in oral therapy with HCQ (200 mg, twice daily) often in combination with azithromycin or clarithromycin (500 mg, once daily) and antivirals (Lopinavir/Ritonavir 400/100 mg, twice daily). In some patients, the clinical course of disease required administration of Tocilizumab (TCZ), a humanized monoclonal antibody against interleukin-6 receptor. In addition, all patients received prophylactic low molecular weight heparin (50–70 IU/kg, with dose adjustment required for creatinine clearance < 30 ml/min). We excluded patients with severe hydro-electrolytic imbalances (defined as serum potassium levels >5 mEq/l; serum sodium levels <130 mEq/l and greater than 150 mEq/l; serum calcium levels <8 mg/dl and >10 mg/dl).

Data collection

At baseline, the following parameters were collected: date of birth, gender, clinical and medication history, anthropometric data, CV risk factors and current pathologies as hypertension, dyslipidaemia, diabetes, coronary artery disease (CAD), atrial fibrillation (AF) or flutter (AFL), obstructive pulmonary disease (COPD), chronic kidney disease (CKD), anaemia, cerebrovascular disease. ECGs were done according to routine practice. All patients had a basal 12-lead ECG, acquired before the beginning of therapy, and a control 12-lead ECG, acquired 3 and 7 days after the beginning of therapy. We evaluated heart rate (HR) and rhythm, PR and QRS duration, QT and HR corrected QT (QTc) interval duration according to Bazett's formula correction, and any repolarization abnormalities. Electrocardiograms were manually evaluated by two cardiologists to calculate all parameters. Laboratory testing including haemoglobin, creatinine, serum potassium, serum magnesium, serum sodium, serum calcium, arterial blood PH, leucocytes, procalcitonin (PCT), C-reactive protein (CRP), tumor necrosis factor alfa (TNF-alfa), interleukin 6 (IL-6), lactate dehydrogenase (LDH), D-dimer, high-sensitive Troponin I, serum creatinine level at maximum QTc, potassium at maximum QTC; were performed according to routine practice. A standard transthoracic echocardiography was performed, focusing on biventricular function, presence of pericardial diseases and valve dysfunction to limit the exposure time of the imager. Afterwards, the following in-hospital cardiac adverse events were assessed: acute coronary syndromes (ACS), myocarditis, pericarditis, pericardial effusion, supraventricular tachyarrhythmia, ventricular tachyarrhythmia and pulmonary embolism. Diagnosis was performed by combining clinical, laboratory, and instrumental data, using ECGs, transthoracic echocardiography, chest X-ray and CT pulmonary angiography.

Statistical analysis

Continuous variables were expressed as mean (SD) or median [interquartile range (IQR)] values when they did not show a normal distribution. Categorical variables were presented as absolute numbers (percentage). Two independent sample testing for continuous variables were performed using Student's t-test if samples had normal distributions, or Mann–Whitney U test if samples did not have normal distributions. Paired samples t-test or Wilcoxon signed rank test were used for paired samples, as appropriate. Categorical variables were compared with Chi-square for independent samples and McNemar test for paired samples. A P value <0.05 was considered to be statistically significant. Statistical analysis was performed using SPSS software, version 26 (SPSS Inc, Chicago, IL, USA).

Result

Study population

Sixty patients were included in this observation study. Baseline demographic characteristics, clinical features and previous therapy of the overall population are summarized in Table 1. Study population had a mean age of 64.12 ± 14.27 years. Patients were predominantly men (61.7%).

Considerable numbers of patients had the common CV risk factors. Specifically, 35 (58.3%) patients had hypertension; 28 (46.6%) were on ACE inhibitors, ARB or ARNI therapy; 11 (18,3%) patients had dyslipidaemia; and 10 (16.7%) were on statins. In addition, 11 (18.3%) patients had diabetes, 23 (38.3%) were current smokers, 4 (6.7%) patients had CKD, and 8 (13.3%) patients had history of CAD. Moreover, 15% of patients had COPD, 26.7% had anaemia, and the median of haemoglobin was 13.3 (11.82–14.45) g/dl. Laboratory tests and COVID-19 related medications are summarized in Table 2. All patients received oral HCQ therapy, alone or in combination with other drugs. Specifically, 33 (55%) patients received triple therapy with HCQ, antivirals and macrolides, 12 (20%) patients received dual therapy with HCQ and antivirals, 12 (20%) patients received dual therapy with HCQ and macrolides, and 3 (5%) patients received single therapy with HCQ. Therapy with TCZ was used in 13 patients, of whom 8 were on triple therapy (HCQ, antiviral and macrolide) and 5 patients were on HCQ and macrolide therapy. Adverse cardiac events of the general population are summarized in Table 2.

ECG changes

We analysed electrographic changes evaluating ECGs acquired at admission and at least 3 days after the beginning of therapy (Table 3). ECGs under treatment showed a lower HR (71.41 ± 11.56 vs 80.1 ± 25,1 beats/min, P = 0.012), a longer QRS duration (100.35 ± 17.61 vs 96.75 ± 17.14, P = 0.010) and a longer QTc interval (449.25 ± 31.93 vs 419.9 ± 33.41, P = 0.000) than ECGs performed before therapy. There was no significant change in PR interval (170.46 ± 27.67 vs 166.94 ± 32.04, P = 0.073). We also analysed ECG changes after 7–10 days of treatment. Similar to 3-day analysis, we recorded lower HR (69.4 ± 8.06 vs 80.1 ± 25.1 beats/min, P = 0.01), longer QRS duration (102.46 ± 15.08 vs 96.75 ± 17.14, P = 0.000) and longer QTc interval (452.15 ± 31.93 vs 419.9 ± 33.41, P = 0.000). After therapy, 18 patients had a QTc ≥ 470 ms and 6 patients had QTc> 500 ms. Interestingly, we observed a case of right bundle branch block at hospital admission which disappeared after treatment. Moreover, we detected the new onset of a left anterior fascicular block (LAFB) during in-hospital stay. Ultimately, new onsets of ventricular repolarization anomalies were found in eight patients, apparently without any clinical correlation.

Cardiac events during in-hospital stay

Our study population was divided into two groups: subjects who had no cardiac adverse events during hospital stay (Group A, N = 45), and subjects with CV involvement (Group B, N = 15). Within the latter group, six (10%) patients had pericardial effusion, one (1,7%) pericarditis without pericardial effusion, two (3.3%) patients had ACS, six (10%) had supraventricular tachyarrhythmia and two (3.3%) patients had pulmonary embolism. The main CV events recorded occurred after the sixth day after the beginning of therapy. Neither myocarditis nor ventricular tachyarrhythmia or sudden cardiac death was found. General characteristics, laboratory tests, and COVID-19 related medications of groups are summarized in Table 4. In Group B the percentage of males was lower (33.3% vs 71.1%, P = 0.009) that Group A. There was no significant difference in common CV risk factors (Table 4). As expected, patients of Group B had a statistically significant lower median ejection fraction (EF) than patients of Group B (55.33 ± 4.81 vs 58.67 ± 4.45, P = 0.017) and higher hs Troponin I (P = 0.004). A significantly difference was also observed for the average levels of haemoglobin at admission and at its minimum values, which were both lower in Group B patients (P = 0.020; P = 0.048). We also noticed a significantly lower percentage of lymphocytes at baseline in Group B than Group A (P = 0.016). Interestingly, we observed significantly higher levels of IL-6 in Group B patients (P = 0.013). Moreover, more patients belonging to Group B were treated with TCZ therapy than Group A patients (P = 0.047). On the contrary, more Group A patients received combined therapy with HCQ and antivirals (P = 0.025).

A standard transthoracic echocardiography was performed in all patients with CV events. In particular, we noticed reduced EF in patients with ACS, presence of pericardial effusion in six patients and right section enlargement in patients with pulmonary embolism. No severe valve defects were found.

Regarding the ECG changes in patients with CV complications, we found reversal T waves in patients with NSTEMI, S1Q3T3 pattern in two patients with pulmonary embolism and reduction of voltage in one patient with pericardial effusion. We also performed a subanalysis by screening the population according to QTc prolongation (corrected QT interval > 470 ms and corrected QT interval < 470 ms) in order to analyse separately CV complications between groups. Of 60 patients, 24 (40%) had ECGs with new QTc interval prolongation to 470 ms. Compared with those without QT prolongation, these patients did not have a higher risk of developing CV complications. In fact, they did not show a higher incidence of pericardial effusion, pericarditis, ACS and PE (Table 5). Of note, no ventricular tachycardia was found in either group. In addition, the onset of supraventricular tachycardia showed no significant difference between patients with and without QT prolongation (8.3% vs 11.1%, P = 0.725). In order to correlate QTc prolongation with the treatment regimen, we have screened samples on the basis of ΔQTc>60 ms or ΔQTc<60 ms. We noted that the majority of patients with ΔQTc>60 were on triple therapy with HCQ, antivirals and macrolides or dual therapy with HCQ and macrolides.

Discussion

The SARS-CoV2 pandemic has led to a health emergency due to a high rate of hospitalization for pneumonia and respiratory failure in infected patients, especially among elderly and those with comorbidity.^1,2,14 Although the heart is not the primary target organ of the virus, cardiac involvement has been noted in many patients.^5,6 Cardiac manifestations may be related to increased cardiac stress due to respiratory failure and hypoxemia, direct myocardial infection by SARS-CoV2, indirect injury from the systemic inflammatory response,¹⁴ or to an effect of drug therapy used to treat the infection.^15–17 In addition to clinically manifest CV events, changes in ECG were noted during hospitalization.¹⁸ The major complications were observed in critical patients admitted to intensive care units.⁵

Our aim is to detect any electrocardiographic disorder at admission and their change during hospitalization in noncritically COVID-19 patients after therapy currently used against the infection. We have also observed the CV events that occurred during hospitalization and we have correlated them to demographic characteristics, laboratory parameters and therapy used.

Our study enrolled 60 consecutive patients hospitalized for respiratory failure at the Pneumology Department of ‘Policlinico Tor Vergata’ from March to May 2020. As already highlighted in previous studies, patients hospitalized for COVID-19 have a high incidence of CV risk factors, in particular systemic hypertension.^1–4 In our analysis patients with hypertension represent 58% of the total. Previous analyses showed that the male sex is related to a worse prognosis in Sars-CoV-2 infection.¹⁹ In our study population 61% of patients were male, but it is unclear whether this is related to a higher susceptibility of the male sex to the infection or to a higher rate of hospitalization for a worse clinical course than the female sex.

The main electrocardiographic parameters were collected and compared with routine 12-lead ECG performed at the time of admission and at least 3 days after the start of therapy for COVID-19. The comparison showed a significant reduction in HR in ECGs recorded under therapy. We think that this may be related to the treatment of the respiratory disease, even though it has been described that HCQ acts as a bradycardic agent in sinoatrial node cells in atrial preparation, and in vivo.²⁰

Significant QTc prolongation is mainly related to therapy with HCQ and macrolide.²¹ At the time of admission, no patient was on antiarrhythmics or other medications associated with QT prolongation. Notably, no patient showed ventricular tachyarrhythmias or sudden cardiac death. This may be due to the limited sample size and the close clinical and instrumental monitoring performed in cases of QTc prolongation beyond certain values. Only six patients showed a QTc >500 ms under therapy, which resulted in drug withdrawal and close control by serial ECG and telemetry monitoring.

In the case of QTc >470 ms, the reduction in drug dosage was evaluated, ECGs were performed more frequently and any electrolyte imbalances were corrected.

We noticed a slight but significant increase in the duration of the QRS interval after starting treatment against SARS-CoV2. This increase is not due to specific delays in intra-ventricular conduction. In fact, as discussed previously, only one patient developed LAFB after therapy. Since there is no known association between the drugs used and change in QRS duration, the explanation for this finding is unclear.

Patients who presented CV events during hospitalization include 25% of the study population. Previous studies show a lower prevalence of CV events in the general population of COVID-19 affected patients (7–12%).^22,23 In these studies, cardiac involvement was expressed only by the finding of direct myocardial damage. In addition, the prevalence seemed to vary considerably across different geographic locations.²³ The ACS registered in our population was NSTE-ACS, with a low GRACE score, treated with medical therapy.

Among the supraventricular tachyarrhythmias we found three patients with AF, 2 with atrial flutter, and one with paroxysmal supraventricular tachycardia. Comparison between groups with (Group B) and without (Group A) development of CV events showed that female gender, low haemoglobin values at admission and lower EF are significantly associated with an increased risk of developing cardiac involvement. We also noted that Group B patients showed a major drop in haemoglobin values during hospitalization compared with Group A patients, although this result is not significant. The correlation between the female gender and the increased CV risk in COVID-19 patients represents interesting data, but it still remains unexplained and requires studies with larger samples and more specific analyses to identify any other variables associated with this finding. It is already known that anaemia has harmful effects in patients with CV risk factors, coronary heart disease, heart failure (HF) and pulmonary hypertension.^24,25 In addition, a reduced ejection fraction and low left ventricular efficiency contribute to increased CV risk.²⁶

The COVID-19 pathophysiology is characterized at first by direct damage determined by the pathogen, followed by a second phase headed by immune host response, which may appear ‘exaggerated’ and lead to the major complications of the disease.^27,28 Interestingly, lower lymphocyte counts and higher IL-6 values are significantly more expressed in Group B. These findings appear consistent with more aggressive immune responses in patients with CV events. Moreover, this is confirmed by an increased use of TCZ therapy among patients who developed CV events during hospitalization.

Lastly we also found from the subanalysis according to QTc prolongation that patients with QT >470 ms did not show a higher risk of developing CV complications. Therefore, although COVID-19 related medications increased the QT interval compared to baseline, in this population the risk of CV involvement did not change.

Study limitations

Our study has some limitations. First, the study analysed a relatively small number of participants. Second, it was an observational single center study. Thus, the results should be confirmed and validated in a different and more representative population. Third, not all possible arrhythmic manifestations were detected since we included noncritically patients and continuous telemetric monitoring was performed only when required.

Conclusions

Patients admitted for SARS-CoV2 infection and treated with anti-COVID-19 drug therapy develop ECG changes such as a reduction in HR, a slight increase in QRS duration and a QTc interval prolongation. In our cohort, one in four patients developed nonfatal CV events. Female gender, lower haemoglobin values, lower lymphocyte count, higher plasma levels of IL-6 and the use of TCZ therapy were more represented in the group of patients with cardiac involvement. Patients with QT prolongation did not show a higher risk of developing CV complications.

Acknowledgements

The authors want to thank all the medical staff of the Division of Respiratory Medicine – Tor Vergata Hospital – Rome, Italy, for their help in collecting data and for their effort and commitment in fighting against the SARS-CoV2 pandemic.

Conflicts of interest

There are no conflicts of interest.