Left ventricular segmental strain based on speckle tracking echocardiography versus indications of immune inflammation in patients after COVID-19 pneumonia

Introduction

SARS-CoV-2 induces a massive release of proinflammatory cytokines in COVID-19-associated macrophage activation syndrome resulting in microvascular thrombosis in the pulmonary arterial network [1]. It is essential to know whether the immunothrombotic response underlies the long-term consequences of COVID-19 [2,3]. Long-term excessive release of proinflammatory cytokines leads to endothelial dysfunction, damage to cardiomyocytes (CMC), and collagen accumulation with subsequent myocardial remodeling and congestive heart failure (CHF) [4].

Myocardial damage in COVID-19 was confirmed by Lindner D. et al. during autopsy: virus particles were found in the heart tissue of 61.5% of patients [5]. In 30-50% of patients, three months after COVID-19, myocarditis-like accumulation of gadolinium was detected by MRI in the middle and subepicardial layers of the left ventricular (LV) myocardium [6,7]. Speckle tracking echocardiography (STE) revealed a predominantly subepicardial (rather than subendocardial) decrease in LV contractility [8,9].

It is worth noting that STE method, according to Shah S.J. et al., can serve as a digital biopsy capable of evaluating the function of CMC [10]. At the same time, various authors reported the lesions of LS based on STE data, with various options for the predominant localization of the lesion [11,12].

Objective: To examine the association between echocardiographic parameters and laboratory indicators of immune inflammation in patients recovered from COVID-19 pneumonia depending on segmental longitudinal strain of the left ventricle identified by the speckle tracking echocardiography.

Material and methods

Subjects

Ethical approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

Our study was enumerated in the international register of clinical trials of the USA National Institute of Health (ClinicalTrials.gov Identifier: NCT04501822). The study protocol was approved by the local ethics committee (No.159 of 23 July 2020). Informed consent was obtained from every patient included in the study.

The inclusion criteria for our study were: pneumonia associated with COVID-19 officially diagnosed at the regional clinical hospital and the patient’s willingness to participate in outpatient monitoring. The exclusion criteria were as follows: chronic diseases in the acute stage, tuberculosis and other diseases accompanied by pulmonary fibrosis, valvular heart disease, cancer, HIV, and chronic hepatitis.

Patients who experienced COVID-19 pneumonia were examined three months after two consecutive negative results of polymerase chain reaction test. STE method with high quality of imaging was applied to 273 of 380 patients (71.8%) included in the Prospective Register of Survivors of Pneumonia Associated with COVID-19 (certificate of state registration No. 2021622535). Subsequently, patients with confirmed coronary artery disease (CAD) [13], arterial hypertension (AH) and increased LV myocardial mass were excluded [14]. Consequently, the study involved 216 patients (56.8%) (51.1% men, mean age of 50.1±11.1 years).

Analyzing echocardiography

Echocardiography was performed using the expert-class ultrasound system Vivid S70 and matrix sensor M5Sc-D (1.5-4.6 MHz); the data were saved in the DICOM format. Myocardial LS was assessed by STE in accordance with the current recommendations [15] using the 17-segment LV model to visually assess the nature of its lesion (Figure 1). Patients with a lesion of segmental LS (≥3 LV segments) constituted Group I (n=108); study subjects without LS impairment formed Group II (n=108) [16]. Lesions of ≥4 segments of the same LV level were considered diffuse LS impairment (Figure 1). A decrease in LS ≥3 segments corresponding to the blood supply basins of the anterior, circumflex, or right coronary arteries were considered a regional lesion of the left ventricle (Figure 2).

Figure 1. Diffuse lesion of predominantly basal level of the LV (LS at the basal level =15.3%; LS at the mid-ventricular level =17.6%; LS at the apical level =17.8%; GLS =16.5%).

LV, left ventricle; LS, longitudinal strain; GLS, global longitudinal strain.

Figure 2. Regional lesion of the apical anterior and apical septal segments of the LV (LS at the basal level =16.2%; LS at the mid-ventricular level =15.2%; LS at the apical level =12.8%; GLS =15.5%)

LV, left ventricle; LS, longitudinal strain; GLS, global longitudinal strain. 

Laboratory data

Laboratory studies included a complete blood count and blood biochemistry test. We determined the biochemical markers of inflammation. High-sensitivity C-reactive protein (hsCRP, reference value range: ​​0-3.0 mg/L) was identified by immunoturbidimetric method with the C-reactive protein hs analytical kit (BioSystem, Spain) on a semi-automatic analyzer of the open type (Clima MC-15, Spain). Interleukin (IL) 6 (IL-6, reference value range of ​​≥9.7 pg/mL), IL-1 (reference value range of 0-5 pg/mL), IL-8 (reference value range of 0-62 pg/mL), tumor necrosis factor α (TNF-α, reference value range of ​​0-8.11 pg/mL), homocysteine ​​(​​reference value range of 5.0-15.0 µmol/L), and N-terminal pro-brain natriuretic peptide (NT-proBNP) were measured by competitive method (solid-phase chemiluminescent enzyme immunoassay) on the IMMULITE 2000 analyzer (Siemens Diagnostics, USA).

Statistical data processing

Statistical data processing was carried out using the Statistical Package for the Social Sciences – IBM SPSS Statistics 26. The normality of distributions was checked by the Kolmogorov-Smirnov criterion. Pearson’s χ2 test was applied to compare qualitative variables. We employed Student’s t-test to compare the quantitative variables in case of their normal distribution; the results are presented as M±SD (М: mean; SD: standard deviation). When analyzing quantitative variables with a distribution other than normal, the Mann-Whitney U test was used; results are presented as median with an interquartile range of the 25th and 75th percentiles. Spearman’s rank-order correlation analysis was carried out to determine the severity of relationships. Logistic regression analysis was used to identify an independent relationship. When assessing the diagnostic significance of the indicators obtained in the regression, ROC analysis was employed. The statistical significance of differences in variables was assumed at p <0.05.

Results

Clinical, functional and laboratory characteristics of patients are presented in Tables 1-4.

Table 1. Clinical characteristics of patients three months after COVID-19 pneumonia

Table 2. Characterization of conventional echocardiographic parameters in patients three months after COVID-19 pneumonia

Table 3. LV longitudinal strain in patients three months after COVID-19 pneumonia

Table 4. Laboratory indicators in patients three months after COVID-19 pneumonia

The groups were comparable in terms of the main clinical and functional characteristics three months after COVID-19 pneumonia, with the exception of gender, AH, and body mass index (BMI), which were statistically significantly more common in Group I. There were no significant differences in the severity of lung damage based on the chest computed tomography during hospitalization and at the control point. The groups were also comparable in terms of post COVID-19 syndrome symptoms. However, the scores of the diagnostic algorithm, The Heart Failure Association preserved ejection fraction (HFA-PEFF), were significantly higher in Group I, compared with Group II (Table 1).

LV end diastolic volume and LV myocardial mass index had statistically significantly higher values in Group I patients vs. Group II. There was also a significant reduction in the early diastolic annular velocity according to tissue Doppler imaging (TDI), along with the early diastolic velocity of the transmitral flow and diastolic flow velocity in the pulmonary veins according to pulsed wave Doppler in Group I, as compared to Group II (Table 2, Figure 3).

Figure 3. A patient with symptoms of diastolic dysfunction (transmitral flow: peak Е=57 cm/s, peak A=86 cm/s; TDI e': IVS=9 cm/s and lateral wall of the LV=8 cm/s) and impaired contractility of the LV (LS at the basal level =14.0%; preservation of the GLS=19.2%; LS at the apical segments =25.8%).

GLS, global longitudinal strain; IVS, interventricular septum; LV, left ventrile; TDI, tissue Doppler imaging. 

Analysis of LV contractile function yielded no significant differences in LV ejection fraction, but detected lesions of both GLS (global longitudinal strain) and LS of all LV levels in Group I, compared with Group II (Tables 2,3). Analysis of the segmental LS revealed impairment of contractility of most LV segments in Group I (Table 3). Among Group I patients, we discovered diffuse lesion of LS (n=41, 38.0% of patients) and regional lesion of LV (n=67, 62.0% of patients). 

Analysis of the laboratory indicators yielded statistically significantly higher concentrations of TNF–α, IL-6, triglycerides, hsCRP and some other indicators in Group I vs. Group II (Table 4).

In Group II, based on Spearman’s correlation analysis, we revealed weak correlations between the concentration of ferritin and LS at the basal LV level (r=-0.250, p=0.033), the concentration of ferritin and LS at the mid-ventricular level of the LV (r=-0.262, p=0.025), the concentration of hsCRP and LS at the mid-ventricular level of the LV (r=-0.197, p=0.043), and the concentration of CRP and LS at the apical level of the LV (r=-0.192, p=0.049).

According to logistic regression data, of all baseline variables describing LV systolic and diastolic functions, laboratory biomarkers, and clinical characteristics, solely the complex of LS at the basal LV level (OR 3.028; 95% CI 1.909-4.802; р<0.001), LS at the apical LV level (OR 1.287; 95% CI 1.099-1.507; р=0.002) and TDI (peak e') of the LV lateral wall (OR 0.774; 95% CI 0.657-0.911; р=0.002) had an independent association with diffuse and regional lesions of LS. In ROC analysis, the sensitivity and specificity of this model were 86.9% and 89.8%, respectively. The area under the curve (AUC) was 0.934 (p<0.001), which corresponded to the excellent quality of the predictive model (Figure 4).

Figure 4. Sensitivity and specificity of the independent association of the baseline complex of variables with diffuse and regional LS lesions.

LS, longitudinal strain.

Discussion

The COVID-19 pandemic has generated interest in the concept of immunothrombotic response, as the disease is a relevant model for demonstrating the mutually potentiating action of the immune system and the coagulation system [3].

In our study, among patients with LS impairment assessed by STE, diffuse and regional LV lesions were identified. In patients who underwent COVID-19, a predominant lesion at the basal level, along with posterior and lateral walls, of the LV was revealed by STE [11,12], which is consistent with our results (Figure 1). Moreover, this localization corresponds to the lesion of the LV in myocarditis according to MRI [7]. On the other hand, microvascular thrombi in coronary arteries and infection of endothelial cells have been confirmed in SARS-CoV-2 [17]. Thus, diffuse and regional LV lesions assessed by STE three months after COVID-19 pneumonia may indicate chronic inflammatory cardiomyopathy and coronary microvascular dysfunction, respectively [18,19] (Figure 1, 2).

It is assumed that the proinflammatory cytokine (IL-6)) makes a significant contribution to immune dysregulation [20]. In the absence of regulation, neutrophil extracellular traps (NETs) can contribute to the intensification of inflammation and microvascular thrombosis. It is important to note that a weak correlation of NETs features with CRP concentration was discovered [21]. At the same time, elevated levels of hsCRP and ferritin were used as pilot prognostic criteria for the cytokine storm associated with COVID-19. Elevated triglyceride levels in these patients were also noted [22]. In addition, a reduction in GLS was associated with high levels of IL-6, hsCRP, and ferritin 30-45 days after COVID-19. Hence, a longer follow-up monitoring duration of patients with latent LV dysfunction was proposed [23].

According to our data, in patients with visually noticeable lesions of the LV based on STE, three months after pneumonia associated with COVID-19, the level of laboratory parameters involved in the implementation of the immune-inflammatory response was higher, but still within the normal range (with the exception of CRP). As is known, hsCRP is synthesized in response to the release of IL-6 and TNF-α [24]. It should also be pointed out that elevated levels of cytokines, TNF-α and IL-6, lead to inadequate CMC contact due to displacement of the desmosomal protein, plakoglobin, from the membrane. This contributes to cell membrane damage and fibrofatty replacement of CMC [25,26]. Therefore, it is important state that we have established a decrease in LS of most LV segments without visual markers of their dysfunction in patients with diffuse and regional LV lesions.

SARS-CoV-2 enters the CMC by binding its own spike protein to the membrane protein of angiotensin-converting enzyme 2 (ACE2) [27]. Moreover, an association was found between the immune response and myocardial inflammation in patients with COVID-19 and overexpression of ACE2 in CMC in CHF [28]. It should be noted that we detected AH and diastolic dysfunction more often in patients with impaired LV contractility. Besides, nearly 60% of patients have post COVID-19 syndrome, which includes shortness of breath and chest pain [29]. In addition, HFA-PEFF diagnostic algorithm scores were higher in patients with impaired LV contractility. Therefore, it seems relevant to suggest that such patients have CHF with preserved LV ejection fraction and to conduct a diastolic stress test [30].

Limitations

First, our sample size, while relatively larger than in previous studies, was still somewhat small and subject to the limitations of small groups. Second, all patients were confirmed to have COVID-19. We did not have follow-up data from patients who have not had COVID-19. Finally, the division into groups used by us was not previously described. Accordingly, we can only speculate on a probably cause for the association between echocardiographic parameters and symptoms of immune inflammation three months after COVID-19 associated pneumonia (chronic inflammatory cardiomyopathy or coronary microvascular dysfunction).

Conclusion

Based on speckle tracking echocardiography data, we have identified a relationship of diffuse and regional lesions of the left ventricle with features of systolic and diastolic dysfunction of the left ventricle and markers of immune inflammation in patients three months after COVID-19 pneumonia.

Conflict of Interest

The authors declare that they have no conflicts of interest regarding the presented article. The authors are accountable for all aspects of the study in terms of ensuring the accuracy and integrity of any part of the research.

Funding

No external funding was provided.