Cardiac Structural and Functional Characteristics in Patients with Coronavirus Disease 2019: A Serial Echocardiographic Study

Study Design and Participants

From January 21, 2020 to April 8, 2020, we conducted a prospective observational study to investigate the cardiac structural and functional characteristics by transthoracic echocardiographic examinations in a cohort of patients who were admitted to our ICU for COVID-19 (Trial Registration: NCT04352842). Including criteria were patients who: 1) had been diagnosed with COVID-19 according to the criteria established by WHO interim guidance ^2,11,12 and 2) were admitted to the ICU because of severe or critical conditions. Patients who were < 18 years old or those whose entire stay in hospital lasted for < 48 hours were excluded. The severity of the disease was categorized according to published criteria ^13,14 Criteria for severe cases included any of the following: 1) Respiratory rate ≥30 per minute; 2) blood oxygen saturation (SPO2) ≤ 93% at rest; 3) partial pressure of arterial oxygen to fraction of inspired oxygen ratio <300; or 4) more than 50% of lung infiltrates within 24 to 48 hours. Patients needing mechanical ventilatory support or presenting with septic shock or multiple organ dysfunction or failure constituted the critical cases.

The study protocol conformed to the Ethical Guidelines of the 1975 Declaration of Helsinki. All identification information that would impact personal privacy was removed during data collection. The Hospital Ethics Commission approved the study protocol and waived written informed consent for emerging infectious diseases according to the regulation issued by the National Health Commission of the People’s Republic of China.

Data Collection

All epidemiological, clinical, laboratory, and outcome parameters were prospectively collected with standardized data collection forms from an electronic medical records system. Personal history was confirmed with patients or family members. Two researchers (H.G. and M.Z.) independently reviewed the forms to double check the data collected.

Echocardiography Protocol and Data Analysis

Echocardiography was performed using a GE Vingmed Ultrasound System (GE Healthcare, Horten, Norway) with special protection in the contaminated area. Two-dimensional (2D) and Doppler echocardiographic measurements followed the recommendations of the American Society of Echocardiography ¹⁵. Protections of the echocardiographer and cleaning of the equipment were in accordance with the latest ESC recommendations ¹⁶. Conventional parasternal long-axis view as well as apical 4-chamber, 2-chamber and 3-chamber views were obtained. Images were analyzed using GE Echopac commercial software. All echocardiography determinations were based on the observations of two independent experienced reviewers (J.D. and W.W.) who were blinded to other data.

Events and Follow-up

Clinical outcomes were censored at the time of data cutoff which occurred on April 8, 2020. For patients who were discharged before April 8, follow-up was continued by telephone interview. Time from diagnosis to death was recorded.

Statistics

Data analyses were performed using SPSS 23.0 (SPSS Inc., Chicago, Illinois) and SAS version 9.2 (SAS Institute, Cary, North Carolina). Categorical data are expressed as absolute values and percentages and were compared using chi-square or Fisher exact tests, and continuous data are reported as median (interquartile range [IQR]) and were compared using Kruskal-Wallis/Wilcoxon rank sum tests. Cox proportional hazards regression model was used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) with adjustment for potential confounders for mortality risk. Multivariable Cox regression models were used to determine the independent risk factors for death. Two criteria were considered necessary for a variable to be entered in the multivariable analysis model: (1) a univariate P value for survivors vs. non-survivors comparison ≤0.10; and (2) a plausible association with the risk of death in COVID-19 according to data provided by the literature. Time to events was described by the Kaplan-Meier curves, and groups were compared using the log-rank test. Inferential statistical tests were conducted at a significance level of 0.05. The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agreed to the report as written.

Demographics and Baseline Characteristics

From Jan 21, 2020, to April 8, 2020, a total of 51 patients who were admitted to the ICU and underwent serial echocardiography examinations were prospectively enrolled in this study. All patients had laboratory-confirmed COVID-19 infection. Summary of general information of all study patients is shown in Supplementary Table 1.

The median patient age was 70 years (IQR, 58.0–79.0; range, 25-93 years); and 37 (72.5%) were male. Thirty-three (64.7%) had underlying cardiovascular diseases including hypertension (43.1%), coronary artery disease (23.5%), congestive heart failure (7.8%), and atrial fibrillation (5.9%); and 36 (70.6%) had other comorbidities including diabetes mellitus (31.4%), pulmonary disease (13.7%), renal disease (23.5%), liver disease (11.8%), or cancer (9.8%). Overall, 84.3% had lymphopenia, 33.3% had monopenia, 92.2% had increased D-Dimer levels, and 45.1% had increased fibrin degradation product (FDP) levels. Hyperinflammatory status was observed in most patients as evidenced by increased peak concentrations of high-sensitivity C-reactive protein (hs-CRP) (84.3%) and interleukin-6 (IL-6) (80.4%). Elevated levels of cardiac markers including hs-cTnI and BNP were observed in 62.7% and 86.3% of patients, respectively.

According to the severity of the disease ^13,14, 31 (60.8%) were classified as critically ill and 20 patients (39.2%) were classified as severely ill. Demographic variables including the prevalence of underlying cardiovascular diseases and other comorbidities were similarly distributed between severely ill with critically ill patients (Table 1). However, compared with severely ill patients, critically ill patients had higher blood leukocyte count (P=0.002) and neutrophilic granulocyte percentage (P<0.001), as well as lower lymphocyte percentage (P=0.001), monocyte percentage (P=0.001), and platelet count (P=0.002). Compared with severely ill patients, the levels of D-Dimer (P=0.008) and FDP (P=0.035), inflammation markers including hs-CRP (P<0.001) and IL-6 (P<0.001), and cardiac injury indicators including hs-cTnI (P<0.001) and BNP (P=0.001), were greatly increased in critically ill patients.

Echocardiographic Findings

Forty-two (82.3%) had at least one abnormal left and/or right-heart echocardiographic parameter: 38 patients (74.5%) had at least one abnormal left heart echocardiographic manifestation, and 22 patients (43.1%) had at least one abnormal right heart echocardiographic manifestation. Detailed comparisons of echocardiographic manifestations between severely and critically ill patients have been illustrated in Figure 1 and Table 2.

• Download figure
• Open in new tab

Figure 1

Comparison of cardiac structural and functional parameters between critically ill and severely ill COVID-19 patients. Categorical data were compared using chi-square or Fisher exact tests, and continuous data were compared using Kruskal-Wallis/Wilcoxon rank sum tests. LVEF, left ventricular eject fraction; LVESD, left ventricular end-systolic diameter; LVEDD, left ventricular end-diastolic diameter; IVS, interventricular septal thickness; PASP, pulmonary artery systolic pressure.

• Download figure
• Open in new tab

Figure 2

Comparison of cardiac structural and functional parameters between non-survivors and survivors. Categorical data were compared using chi-square or Fisher exact tests, and continuous data were compared using Kruskal-Wallis/Wilcoxon rank sum tests. LVEF, left ventricular eject fraction; LVESD, left ventricular end-systolic diameter; LVEDD, left ventricular end-diastolic diameter; IVS, interventricular septal thickness; PASP, pulmonary artery systolic pressure.

Cardiac Involvement in Survivors vs. Non-survivors

Twelve (23.5%) patients died. The median age of non-survivors was 75.0 years (IQR, 57.75-80.50 years; range, 25-82 years), and 75.0% were male. As shown in Table 3, the demographic characteristics and the prevalence of chronic diseases were comparable between survivors and non-survivors, with only a tendency of a more frequent cancer history in non-survivors (P=0.078). However, compared with survivors, non-survivors had higher neutrophilic granulocyte percentage (P=0.002), and aspartate aminotransferase (AST) (P=0.022), D-Dimer (P=0.024), FDP (P=0.044), hs-CRP (P=0.024), and IL-6 (P<0.001) levels as well as lower lymphocyte percentage (P=0.014), monocyte percentage (P=0.003), platelet count (P<0.001), and albumin level (P=0.015). Cardiac injury indicators such as hs-cTnI (P=0.039) and BNP (P=0.045) tended to increase in non-survivors (Table 3).

When comparing echocardiographic findings, we did not observe any significant in left-heart parameters except for a thicker IVS in non-survivors (12.00 [IQR 10.00-12.00] mm vs.10.00 [IQR 9.00-11.00] mm; P=0.029). However, non-survivors had more abnormal right-heart parameters including elevated PASP (58.3% vs. 23.1%; P=0.028), RA enlargement (50.0% vs. 17.9%; P=0.036), RV enlargement (58.3% vs. 17.9%, P=0.011), and decreased TAPSE (41.7% vs. 7.7%, P=0.012) (Table 4).

Figure 3 shows the serial changes in cardiac structural and functional parameters in 12 non-survivors. Compared with the echocardiographic findings on admission to the ICU, there was no difference in LVEF (63.50 [IQR 60.75-67.25]% vs. 64.50 [62.00-67.75]%; P=0.130), LVEDD (45.50 [43.00-49.75] mm vs. 48.50 [45.50-50.75] mm; P=0.777), and LVESD (31.00 [28.00-32.75] mm vs. 31.00 [27.50-33.75] mm, P=0.118) in non-survivors on the day before death. However, IVS thickness (9.50 [IQR 9.0-10.75] mm vs.12.0 [IQR 10.25-12.0] mm; P=0.001), PASP (31.00 [IQR 29.25-35.75] mmHg vs. 44.50 [33.50-50.50] mmHg; P=0.009), and RV diameter (35.00 [33.00-36.00] mm vs. 42.0 [33.25-42.75] mm; P=0.004) increased significantly in non-survivors before death.

• Download figure
• Open in new tab

Figure 3

Changes in cardiac structural and functional parameters among 12 non-survivors. LVEF, left ventricular eject fraction; LVESD, left ventricular end-systolic diameter; LVEDD, left ventricular end-diastolic diameter.

Cox proportional hazards regression model was used to estimate the mortality risk for left-heart abnormality or right-heart abnormality with adjustment for potential confounders. Left-heart abnormality was not associated with mortality (HR 1.78 (95% CI 0.39 to 8.16), P_trend=0.456. However, abnormalities of the right heart were associated with increased mortality in patients with COVID-19: HR (95% CI) for mortality was 4.58 (1.24 to 16.96), P_trend⁼0.023, for patients with right-heart abnormality vs. patients without right-heart abnormality. As shown in Figure 4, Kaplan-Meier survival curves were similar between patients with vs. without left-heart abnormalities (P=0.450 by the log-rank test). However, when patients manifested right-heart abnormalities, the survival rate decreased (P=0.012 by the log-rank test).

• Download figure
• Open in new tab

Figure 4

Impact of left-heart or right-heart echocardiographic abnormalities on mortality in COVID-19 patients during the study period. Time to events was described by the Kaplan-Meier curves, and groups were compared using the log-rank test.

Impact of Use of Extracorporeal Membrane Oxygenation (ECMO)

Nine (17.6%) patients received ECMO. We monitored dynamic changes in cardiac structural and functional parameters before and after ECMO (Figure 5). ECMO reduced PASP (34.00 [IQR 30.50-40.50] mmHg vs. 25.00 [21.00-28.00] mmHg; P=0.011), and RA diameter (40.00 [IQR 37.00-50.00] mm vs. 36.00 [IQR 33.50-38.00] mm; P=0.044) and RV (40.50 [33.25-45.75] mm vs. 35.00 [34.00-39.50] mm; P=0.054) in critically ill cases. Of 9 patients receiving ECMO, four survived and five died. The changes in echocardiographic parameters post-ECMO in survivors vs. non-survivors are illustrated in Supplementary Figure 1.

• Download figure
• Open in new tab

Figure 5

Changes in cardiac structural and functional parameters among nine COVID-19 patients treated with extracorporeal membrane oxygenation (ECMO). LVEF, left ventricular eject fraction; LVESD, left ventricular end-systolic diameter; LVEDD, left ventricular end-diastolic diameter.

Predictors of Mortality

In univariate Cox proportional hazards regression analyses, previous malignancy, neutrophilic granulocyte percentage, lymphocyte percentage, platelet count, albumin level, hs-cTnI level, IL-6 level, IVS thickness, RA enlargement, RV enlargement, elevated PASP and decreased TAPSE were significantly associated with mortality (Table 5). Multivariable analyses identified that hs-cTnI (HR: 1.138, 95% CI: 1.029 to 1.258, P=0.012) and IL-6 (HR: 1.095, 95% CI: 1.002 to 1.196, P=0.001) were independently associated with mortality in ICU patients with COVID-19.

Left Heart Abnormalities

Lessons from previous coronavirus and influenza epidemics suggested that the virus would either induce new cardiac pathologies and/or exacerbate underlying cardiovascular diseases ²². It was postulated that the systemic inflammatory storm in COVID-19 might lead to acute myocarditis⁶, stress cardiomyopathy ²³, or acute coronary syndromes ⁷, recently described as the “acute COVID-19 cardiovascular syndrome” ²⁴ The latter should significantly impair LV systolic function, which might be expected to be a key characteristic of cardiac involvement of patients with COVID-19. LVEF has been confirmed as a key determinant of prognosis in cardiac patients ^25,26. Therefore, it was surprising and unexpected to find that in our cohort of severely and critically ill COVID-19 patients, the majority had preserved LVEF; only 15.7% had reduced LVEF that was often associated with pre-existing heart disease. Moreover, LVEF was comparable in critically ill vs. severely ill COVID-19 patients and in COVID-19 survivors vs. non-survivors; serial echocardiography did not reveal LVEF deterioration in non-survivors from ICU admission to death; and left-heart abnormalities were not associated with increased mortality in COVID-19 patients. Thus, our findings did not support the notion that LV systolic dysfunction was the principal form of cardiac involvement associated with COVID-19.

In contrast, we observed abnormal echocardiographic manifestations associated with impaired LV diastolic function in COVID-19 patients, as evidenced by E/A ratio and E/e’ ratio abnormality and IVS hypertrophy. The impaired LV diastolic function could not be attributed to the underlying cardiovascular or other comorbidities as the comorbidities were comparable in critically ill vs. severely ill patients as well as in non-survivors vs. survivors. Moreover, IVS thickness paralleled disease severity and even increased significantly in non-survivors from ICU admission to death. Similarly, a previous study found that subclinical diastolic left ventricular impairment was common in acute severe acute respiratory syndrome (SARS) infection²⁷.

Our findings of impaired LV diastolic function with preserved LV systolic function were in accordance with the findings of several recent autopsy studies in COVID-19 ^8,28. On one hand, no apparent myocardial necrosis/apoptosis was found in myocardium in COVID-19 at autopsy ^8,28. Consistently, although the prevalence was high, the elevation of hs-cTnI level was mostly mild (median 0.07 [IQR 0.02–0.23]) in our cohort. This may explain the preserved LVEF among patients studied. On the other hand, the diastolic LV dysfunction in the current study was similar to that in Clancy’s report in which as many as 60% of patients with severe sepsis had echocardiographic manifestations of diastolic dysfunction ²⁹ Furthermore, several studies have reported that diastolic rather than systolic LV dysfunction was an independent predictor of mortality in patients with severe sepsis ^30,31. Xu et al. reported a patient who died from severe COVID-19 who had mononuclear inflammatory infiltration in the myocardial interstitium, but no substantial damage to the myocytes. Tian et al. reported four patients who died of COVID-19 who had various degrees of cardiac focal edema and interstitial fibrosis ²⁸. Inciardi et al. described cardiac magnetic resonance imaging showing marked ventricular interstitial edema and increased wall thickness in a 53-year-old woman with COVID-19⁶. Taking together, the increased IVS thickness and deteriorating LV compliance observed in our study might be reflective of interstitial inflammatory edema, as has been found in other sepsis studies ²⁹.

Right Heart Abnormalities

About one third of the study patients had right-heart abnormalities as evidenced by elevated PASP, larger RA and RV diameters, and decreased TAPSE. Among them, only 13.6 % had a history of chronic lung disease. Moreover, right-heart dysfunction, but not left-heart dysfunction, was associated with increased mortality in our cohort. The mechanism of right-heart dysfunction in COVID-19 is unclear but may be related to vascular thickening in the lung ³², hypoxemia and pro-inflammatory cytokines that would provoke pulmonary vasoconstriction ², or the possibility that pulmonary embolism as suggested by the extremely high concentration of D-dimer (observed in our cohort and other populations) and its close relationship with mortality ³³. A postmortem study in eight patients who died from SARS showed that four patients had pulmonary thromboembolism ³⁴. Rapid development and progression of right-heart dysfunction has been reported in patients with acute respiratory distress syndrome or acute pulmonary embolism ³⁵. The adverse impact of right-heart impairment on prognosis also was reported in ARDS patients ^36,37

Impact of ECMO

Reports of ECMO utility in COVID-19 patients have been scarce, and its clinical effectiveness and outcomes remained unknown. We monitored the dynamic changes of cardiac structural and functional parameters before and after ECMO implantation. ECMO implantation tended to alleviate elevated PASP and other right-heart dysfunction in critically ill patients. Theoretically, the application of ECMO might reduce pulmonary hypertension by improving oxygenation, correcting hypoxemia and alleviating pulmonary vasoconstriction. Moreover, RV volume may be unloaded with a venoarterial ECMO. However, in non-survivors treated with ECMO, the echocardiographic improvement was temporary; and the right-heart dysfunction deteriorated again before death (Supplementary Figure 1), similar to the post ECMO scenario in children with pulmonary hypertension from pertussis ³⁸. Thus, it seems that the key value of ECMO implantation is as a temporary life support to gain time for therapy, and the treatment of the primacy pulmonary disease is vital for the final prognosis.