The presence of abnormal septal motion on echocardiography is a predictor of abnormal cardiac magnetic resonance in systemic sclerosis

Abstract

Objectives

We aimed to perform a comprehensive analysis of the ECG, two-dimensional echocardiography (2DE) and cardiac MRI (CMR) findings in patients with systemic sclerosis (SSc), and also to investigate correlations between CMR findings and some ECG and echocardiography (ECHO) results.

Methods

We retrospectively analysed data from patients with SSc who were regularly seen at our outpatient referral centre, all assessed with ECG, Doppler ECHO and CMR.

Results

Ninety-three patients were included; mean (s.d.) age of 48.5 (10.3) years, 86% female, 52% diffuse SSc. Eighty-four (90%) of the patients had sinus rhythm. The most common ECG finding was the left anterior fascicular block, recorded in 26 patients (28%). The abnormal septal motion (ASM) was found in 43 (46%) patients on ECHO. Myocardial involvement (inflammation or fibrosis), as assessed by multiparametric CMR, was present in >50% of our patients. The age- and sex-adjusted model showed that ASM on ECHO increased significantly the odds of increased extracellular volume [odds ratio (OR) 4.43, 95% CI 1.73, 11.38], increased T1 Relaxation time (OR 2.67, 95% CI 1.09, 6.54), increased T2 Relaxation time (OR 2.56, 95% CI 1.05, 6.22), increased signal intensity ratio in T2-weighted imaging (OR 2.56, 95% CI 1.05, 6.22), presence of late gadolinium enhancement (OR 3.85, 95% CI 1.52, 9.76) and mid-wall fibrosis (OR 3.64, 95% CI 1.48, 8.96).

Conclusion

This study indicates that the presence of ASM on ECHO is a predictor of abnormal CMR in SSc patients, and a precise assessment of ASM may serve as an important point for selecting the patients that should be evaluated by CMR for early detection of myocardial involvement.

Introduction

SSc is a complexCTD characterized by fibrosis of the skin and internal organs [1]. One of the feared complications of the disease is cardiac involvement, which is often clinically asymptomatic and is considered a poor prognostic factor. Among SSc-associated deaths, cardiac complications are one of the leading causes [2, 3].

Clinically evident primary cardiac involvement (i.e. not secondary to pulmonary arterial hypertension or ischaemic heart disease) includes inflammatory myocarditis, systolic and diastolic ventricular dysfunction due to myocardial fibrosis, arrhythmias, conduction abnormalities and pericardial disease [4]. The role of myocardial inflammation and myocardial fibrosis in the pathogenesis of primary SSc heart disease has recently been emphasized [5]. An association between cardiac MRI (CMR) indices of myocardial fibrosis and adverse cardiovascular outcomes in patients with SSc has also been observed [6]. Diffuse myocardial fibrosis has been reported in up to 80% of cases in autopsy studies [7].

Therefore, early and accurate detection of myocardial involvement using a non-invasive screening modality is of great clinical importance in patients with SSc. The diagnostic gold standard for myocardial involvement remains endomyocardial biopsy [8]. However, this test is an invasive procedure with sampling errors and low diagnostic sensitivity, and is unlikely to be available outside major cardiac centres [8].

In clinical practice, non-invasive and readily available diagnostic methods include ECG and echocardiography (ECHO). Two-dimensional echocardiography (2DE) is the mainstay of serial monitoring of SSc patients. However, some studies have reported that subtle changes in myocardial contractility can be underestimated using conventional echocardiographic techniques, and more sensitive modalities are required to determine subclinical myocardial involvement [9].

CMR is an accurate and sensitive method to study heart structure [10]. With a comprehensive protocol including late gadolinium enhancement (LGE) and quantitative T1 and T2 mapping techniques, it allows functional assessment of the cardiac system, tissues characterization and detection of various mechanisms of cardiac involvement, including inflammatory, microvascular and fibrotic processes [11, 12]. CMR overcomes the shortcomings of ECHO, leading to earlier diagnosis of silent cardiac abnormalities in SSc. However, it is not always available and, therefore, not yet feasible as a universal diagnostic modality for the assessment of SSc heart involvement. Currently, we use CMR in some instances based on ECG and echocardiographic findings.

In this study, we investigated the ECG, 2DE and CMR findings of our SSc patients to perform a comprehensive analysis of these data and to assess correlations between the cardiac MRI findings and some ECG and ECHO results.

Methods

We retrospectively analysed data from SSc patients who were regularly seen at our outpatient referral centre, Iran Rheumatology Center, between 2017 and 2021. All patients enrolled in this observational cohort study fulfilled the criteria of the ACR/EULAR 2013.

According to our centre’s clinical protocol, all patients, regardless of clinical symptoms, underwent ECG and Doppler ECHO annually. When needed, CMR was performed. In our clinic, the indications for magnetic resonance heart assessment include the following:

• The presence of regional wall motion abnormalities on ECHO

• Left ventricular ejection fraction (LVEF) ≤50% on ECHO

• LV diastolic dysfunction GII or GIII on ECHO

• The presence of pericardial effusion on ECHO

• Frequent premature ventricular contractions (PVCs) (more than five PVCs per minute on the routine ECG or >20% of all QRS complexes on standard 24-h Holter monitoring)

• Post-capillary pulmonary hypertension on right heart catheterization (RHC)

• A positive fluid challenge response on fluid challenge test during RHC (some recent cases)

• Cases with abnormal myocardial perfusion imaging (SPECT protocols) but with a normal coronary angiography

• The patients with abnormal right ventricular (RV) systolic function on ECHO but with a normal pulmonary arterial pressure on RHC

The patient population consisted of 93 SSc patients, with the limited or diffuse disease according to LeRoy’s classification, between the ages of 27 and 73 years. All patients were assessed by CMR. These are among the 700 SSc patients who were regularly seen at our outpatient referral centre for several years.

Clinical characteristics, including age, gender, disease subset (diffuse, limited), medications, cardiovascular risk factors, thyroid dysfunction and anaemia, were recorded. Autoimmune antibody test results (anti-Scl-70, anti-centromere) were extracted from our centre’s database.

The study was approved by the Research Ethics Committees of the Evidence-Based Medicine Research Center, Tehran University of Medical Sciences (approval ID: IR.TUMS.EMRI.REC.1400.104).

All patients underwent standard 12-lead ECG (ECG amplifier sensitivity of 10 mm/mV, at a speed of 25 mm/s). The traces were carefully examined for supraventricular or ventricular arrhythmias, conduction disorders, ST-T wave abnormality, Septal Q-wave pattern, signs of LV hypertrophy and right atrial enlargement.

The patients first underwent an ECHO and then CMR was performed. Therefore, the ECHO interpreter was blinded to the clinical data, but the CMR interpreters had access to the ECHO findings. The time interval between CMR and ECHO was on average 4–6 weeks.

Echocardiography

The standard echocardiographic studies and analyses were performed by an experienced senior cardiologist using an US system (Samsung Medison, EKO7system, Samsung, Korea) with a 2–4 MHz transducer.

Echocardiographic parameters recorded include left ventricle end-diastolic diameter (LVEDD), left ventricle end-systolic diameter (LVESD), interventricular septum (IVS), posterior wall (PW) thicknesses, right ventricle end-diastolic diameter (RVEDD) and right ventricle wall thickness, systolic and diastolic function of the LV and systolic function of the right ventricle, the presence of pericardial effusion (small, moderate, large), and the presence of valve disease (stenosis and regurgitations). We specified the patients with systolic peroxidase–antiperoxidase >35 mm.

All measurements were performed according to the recommendations of the American Society of Echocardiography and the European Association of Echocardiography (Supplementary Data S1, available at Rheumatology online).

M-mode ECHO was interrogated on the parasternal views to analyse septal motion. Abnormal septal motion (ASM) has been characterized as an atypical jerky movement of the basal and mid-interventricular septum, occurring with every cardiac cycle, without a relationship to the respiratory cycle. If multiple ECHO were performed for the same patient within this specified time frame, the last study performed before CMR was analysed.

CMR protocol and analysis

The CMR studies were performed in two CMR centres using a 1.5 T GE Signa Explorer system (Florence, South Carolina, USA) and 1.5 T Siemens Sola (Erlangen, Germany). All examinations were interpreted by two cardiologists experienced in cardiac MRI.

The acquisition sequences include:

• Cine steady-state free precession imaging in short- and long-axis views for quantifying cardiac volumes and functions.

• The Modified Look-Locker Inversion Recovery (MOLLI) sequence to perform T1 mapping was used. Short-axis slices were acquired at the basal and mid-ventricular levels. In addition to T1 mapping, measurement of the extracellular volume (ECV) using the same MOLLI sequence was done. ECV is a marker of fibrosis and myocardial remodelling. To calculate ECV, pre- and post-contrast T1 mapping was performed (15–20 min after the contrast injection to allow for equilibration of the contrast agent between the blood pool and the myocardium).

• A T2 mapping sequence (multi-echo fast spin echo (FSE) sequence with 12 echoes spaced evenly between 0 and 72 ms) to assess the T2 Relaxation time of the myocardium. Short-axis slices were acquired at the basal and mid-ventricular levels to cover the whole LV.

• LGE imaging was acquired in 2D phase-sensitive inversion recovery reconstruction and with 2D single shot technique in patients with difficulty breath-holding, on average 10–15 min after injection of 0.2 mmol/kg of gadolinium-containing contrast agents.

Post-processing analysis was performed using Cvi42 workstation (Circle Cardiovascular Imaging, Calgary, Alberta, Canada) (Supplementary Data S2, available at Rheumatology online).

The presence and extent of scar tissue were assessed by visually analysing the signal intensity of the myocardium on the LGE images. Myocardial fibrosis was assessed in terms of distribution (standard LV 17segmentation), pattern (focal, multifocal, linear and diffuse) and localization of fibrosis (subepicardial, mid-wall, subendocardial, insertion point).

A normal LVEF was defined as LVEF ≥55%.T1- and T2-based criterion was defined as pathological if: LGE >0% of LV mass; native T1 mapping >1050 ms; ECV >27%; T2 ratio >2 and T2 mapping >52 ms.

Statistical analysis

All statistical analysis was performed using the SPSS version 18. Continuous and categorical data are expressed as mean with s.d. and number (%), respectively. Comparison of continuous and categorical findings of CMR with ASM was assessed using a t-test and χ² test, respectively. Pearson correlation test was used to assess the correlation between LVEF by two modalities (ECHO and CMR).

A crude and adjusted odds ratio (OR) by logistic regression test was used to assess the association between ASM as independent variable with CMR findings as dependent variables.

Variables with a P-value <0.1 in the crude model were included in the adjusted model. The adjustment was performed for age and sex. Results of the logistic regression model were reported as OR with a 95% CI. A P-value <0.05 was considered statistically significant. All P-values were corrected using the Benjamini–Hochberg correction method to control the false discovery rate due to multiple comparisons problem [13].

Results

Characteristics of the study population

The clinical characteristics of the patients are shown in Table 1. We studied 93 patients with SSc (diffuse, 48; limited, 45). The mean age was 48.5 years. Most of the patients in the cohort were female (86%). Anti-Scl-70 antibodies and ACA were positive in 29 (31%) and 27 (29%) patients, respectively. Twenty-six (28%) patients received a β-blocker drug, and 42 (45%) an angiotensin-converting enzyme inhibitor/angiotensin-receptor blocker. The proportion of patients using calcium-channel blockers was 82%, and pulmonary arterial hypertension medication was 30%. Some of the patients on pulmonary arterial hypertension medication had pre-capillary pulmonary hypertension in RHC; some of them were on these medications for digital ulcers, and the others had been treated with these drugs for RP.

Cardiovascular risk factors such as diabetes mellitus, systemic arterial hypertension and dyslipidaemia were assessed. Twenty-three (25%) patients had systemic arterial hypertension and five (5%) had diabetes. Twenty-two (24%) patients were anaemic and 11 (12%) had thyroid dysfunction.

ECG findings

Resting ECG findings are shown in Table 2. Eighty-four (90%) of the patients had sinus rhythm. Nine of the patients (10%) had supraventricular arrhythmias, of whom three patients had atrial fibrillation and six patients had atrial flutter. Sixteen (17%) patients had premature supraventricular complexes, and 23 (25%) patients had premature ventricular complexes (Isolated, Bigemini, Trigemini, Couplets, Nonsustained VT and Polymorphic PVCs). Conduction abnormalities including Incomplete right bundle-branch block (RBBB), Complete RBBB, Incomplete left bundle-branch block (LBBB), Complete LBBB, non-specific intraventricular conduction delay and left anterior fascicular block were assessed (see Table 2 for details). Eighteen (19%) patients had a septal Q-wave pattern, and 19 (20%) had ST-T wave abnormalities.

Echocardiographic findings

The mean LVEDD was 4.52 ± 0.41 mm, LVESD was 3.1 ± 0.43 mm, IVS was 0.91 ± 0.14 mm and PW was 0.91 ± 0.14 mm. The mean RVEDD was 2.8 ± 0.27 mm, and the RV wall thickness was3.7 ± 0.74 mm. The mean LVEF was 49.82 ± 6.2%. Abnormal RV systolic function was found in 32 (34%) patients. Twenty-nine patients (31%) were diagnosed with LV diastolic dysfunction. ASM was found in 43 (46%) patients. Fifty-six patients presented regional wall motional abnormalities. Systolic pulmonary artery pressure more than 35 mm Hg was observed in 43 patients. Eighteen (19%) patients had pericardial effusion. No large pericardial effusion was found. Eighteen (19%) patients had valve disease (Table 2).

CMR abnormalities

The main MRI abnormalities are shown in Table 3.

Myocardial inflammation and fibrosis

Increased signal intensity on T2-weighted sequences was found in 49 patients (53%), and on T2 mapping, increased T2 Relaxation time was found in 49 patients (53%). Overall, 56% of SSc patients demonstrated increased ECV, and 60% demonstrated increased T1 Relaxation time. Myocardial delayed contrast enhancement was detected in 54 patients (58%). Fibrosis was mid-wall in 42 patients (45%), subepicardial in 30 patients (32%) and subendocardial in 9 patients (10%). It was mainly multifocal (26%).

LV/RV dimension and function

The mean (s.d.) indexed LV end-diastolic volume was 73.46 ± 15.9 ml/m², and the mean (s.d.) indexed LV end-systolic volume was36.8 ± 12.2 ml/m². LV dilatation was found in 13 patients (14%) and RV dilatation in 13 patients (14%). On cine-MRI, 48 patients (52%) had an impaired LVEF, and 43 (46%) had an impaired RV ejection fraction. The mean (s.d.) LVEF was 51.4 ± 8.1%. Pericardial effusion was observed in 33 patients (36%).

Correlation between ejection fraction obtained in ECHO and CMR findings

The LVEF was comparable between the two modalities. The mean (s.d.) LVEF obtained by ECHO was 49.8 (6.2)%, and that obtained by MRI was 51.4 (8.1)%. There was a significant linear correlation between LVEFs obtained by these two investigations (r = 0.72, P ˂ 0.001) (see Fig. 1).

Correlation between CMR abnormalities and PVCs in ECG

According to univariable and multivariable logistic regression, CMR findings were not significantly associated with PVCs in ECG (P > 0.05 for all variables).

Association between CMR abnormalities and ASM in ECHO

When SSc patients were stratified according to with or without ASM, the χ² test shows that the presence of increased ECV, increased T1 Relaxation time and LGE were associated significantly with the presence of ASM in ECHO (Table 4). Thirty-one (72%) patients with ASM had an increased ECV, but 21 (42%) patients without ASM had an increased ECV (P = 0.01). LGE was present in 32 (74%) patients with ASM while it was present in 22 (44%) patients without ASM (P = 0.01). Moreover, patients with ASM had significantly more mid-wall myocardial fibrosis within the LV (61% vs 32% in another group; P = 0.02). Furthermore, 27 patients with ASM had increased T2 Relaxation time on T2 mapping, but 22 patients without ASM had increased T2 Relaxation time (P = 0.07). Also, an increased signal intensity ratio in T2-weighted imaging was observed in 27 patients with ASM, but 22 patients without ASM had an increased signal intensity ratio (P = 0.07). The other CMR indexes showed no significant differences between the two groups.

Table 5 shows the crude, and age- and sex-adjusted association between ASM and CMR findings, which had a P-value <0.1 in Table 4. Adjusted model shows that ASM on ECHO increased significantly the odds of increased ECV (OR 4.43, 95% CI 1.73, 11.38), increased T1 Relaxation time (OR 2.67, 95% CI 1.09, 6.54), increased T2 Relaxation time (OR 2.56, 95% CI 1.05, 6.22), increased signal intensity ratio in T2-weighted imaging (OR 2.56, 95% CI 1.05, 6.22), presence of LGE (OR 3.85, 95% CI 1.52, 9.76) and mid-wall fibrosis (OR 3.64, 95% CI 1.48, 8.96).

Discussion

The present study assessed ECG, ECHO and CMR findings in SSc patients. We also investigated the relationship between some ECG and ECHO items and CMR data. The main findings of the study are as follows.

Eighty-four patients had sinus rhythm. Nine (10%) patients presented supraventricular arrhythmias. The most common ECG finding was the left anterior fascicular block recorded in 26 patients (28%), followed by premature ventricular complexes. As in the previous study, the most common reported standard ECG abnormality was the presence of ST-T non-specific changes [14]. RBBB (complete or incomplete) was present in 24 of our patients (26%), and this is a considerable point among our findings, as the study by Draeger et al. on 265 SSc patients showed that a complete RBBB on ECG predicted a higher risk of mortality [15]. Ventricular ectopic beats (VEBs) were detected in 23 patients (25%); 16 of our patients (17%) had supraventricular ectopic beats. Septal Q-wave patterns were found in 18 patients (19%), although more than in a previous study [16]. The difference between those results and our findings may stem from differences in interpretations.

Despite well-described limitations, 2DE is the mainstay of serial monitoring of SSc patients. Using 2DE, impaired RV function was found in 34% of our patients. This is clinically important because it is a predictor of mortality in SSc [17]. Fifty-six patients presented RWMA. Previous studies using Doppler ECHO have also shown regional contractile alterations in SSc patients [16]. We identified ASM (an atypical jerky movement of the basal and mid-interventricular septum, occurring with every cardiac cycle, without a relationship to the respiratory cycle) in 43 (46%) patients. About 19% of our patients had moderate or severe valvular heart disease. Zairi et al. showed first-grade aortic regurgitation in one patient and high-grade regurgitation that required surgical management in none of the patients [18].

Using CMR, we enable to analyse the different patterns of cardiac involvement in SSc. However, the routine use of CMR for the evaluation of myocardial involvement in SSc is not practical because there are some limitations, including their availability and financial constraints [10]. In our cohort, CMR findings included focal areas of fibrosis (as detected by LGE) and diffuse myocardial fibrosis (as detected by T1-weighted imaging and ECV quantification). Using delayed-enhancement MRI (DE-MRI), we found myocardial fibrosis in 54 patients (58%). In comparison with the present study, in the study of Hachulla et al. based on delayed contrast enhancement, a prevalence of myocardial fibrosis of 21% was reported [10]. Another important innovation of our study was an increased ECV, a marker indicative of diffuse myocardial fibrosis in 56% of patients because, according to the study of Dumitru et al., higher ECV has a trend for association with cardiovascular outcomes [6]. Like some other studies, the most common fibrosis location in our patient population was midwall in the LV. Our study’s most common form of fibrosis was multifocal, different from the generally described linear pattern. Increased signal intensity on T2-weighted images and increased T2 Relaxation time (T2 mapping) are indicators of soft-tissue oedema, and they were present in 53% of our patients. We observed that LVEFs were reduced in 52% of patients on CMR, and RV ejection fractions were altered in 46% of patients. Hachulla et al. showed that 12 of 52 patients (23%) had an impaired LVEF and reduced RV ejection fraction was present in 11 patients (21%) [10]. As previously suggested, the alteration of LV and RV ejection fractions is most likely a direct consequence of myocardial fibrosis [19]. In our retrospective cohort study, pericardial effusion was observed in 33 patients (36%), whereas in another study, 10 patients (19%) had a moderate pericardial effusion [10].

Another finding of this study is the absence of a correlation between ventricular arrhythmias and the presence of fibrosis and inflammation on CMR. In contrast to our study, The Scleroderma Arrhythmia Clinical Utility Study on 150 patients reported the predictive value of replacement fibrosis (%LGE) and increased T2 ratio for the development of ventricular rhythm disturbances [20].

In SSc patients, myocardial fibrosis correlates with the number of VEBs on 24-h ECG Holter. The presence of frequent VEBs is associated with an increase in the risk of sudden cardiac death [14, 20]. De Luca et al. reported the correlation between the number of PVCs/24 h with poor prognosis in SSc [14]. Interestingly, we found cardiac MRI findings were not significantly associated with PVCs in ECG (P > 0.05 for all variables). Since 24-h Holter is not routinely performed as part of our protocol, we evaluate PVCs in resting 12-lead ECGs, not in 24-h Holter ECG recordings, and PVCs may not be fully detected by resting ECG and might have been missed.

We evaluated the crude, and age- and sex-adjusted association between ASM and CMR findings, which had a P-value of <0.1. The adjusted model showed the presence and localization of myocardial fibrosis (mid-wall fibrosis), increased ECV, increased T1 Relaxation time, increased T2 Relaxation time and increased signal intensity ratio in T2-weighted imaging were associated with the presence of ASM; therefore, this study showed an association of ASM with the CMR measures of myocardial fibrosis and inflammation. Previous studies showed that LV regional dysfunction is related to myocardial abnormalities on CMR in SSc patients [21].

The main aim of this study was to investigate the association of ASM on ECHO (due to primary reasons such as myocardial involvement or secondary reasons such as pulmonary hypertension) with CMR findings. In this study, we did not intend to separate the various causes of ASM and their association with CMR findings. In this study we have shown that whenever abnormal septal motion is observed on ECHO, there are abnormal CMR findings in a significant percentage of these patients. In all cases with ASM, myocardial involvement, to varying extents, may be present. This is, of course, an issue that should be explored in future studies.

LV end-systolic volumes and LV end-diastolic volumes of the patients with ASM were not statistically significantly different compared with patients without ASM, which may indicate that progressive myocardial function impairment due to SSc has not happened. Based on our findings, it can be hypothesized that, probably before the development of overt cardiac involvement, there is interventricular septum dyssynchrony. Therefore, septal contractile dysfunction may precede the diffuse myocardial involvement in SSc, which needs further investigation by prospective studies.

It has been reported that the degree of both myocardial fibrosis and inflammation is associated with a poor 22-month outcome in SSc patients [22]. Also, studies showed that early therapeutic intervention of acute myocardial inflammatory process has the potential to prevent late cardiac remodelling and improvement of patients’ outcomes [5]. Myocardial involvement (inflammation or fibrosis), as assessed by multiparametric CMR, including T2-weighted imaging, native T1, quantitative LGE and ECV measurement, was present in >50% of our patients.

As a result, regarding the high percentage of myocardial involvement in CMR, our findings have clinical importance, but whether detecting these early impairments in patients with SSc correlates with outcomes in these patients requires further investigation; in addition, these changes in SSc patients during the follow-up could be important points for therapeutic trials.

The results of our retrospective analysis show the importance of comprehensive screening and assessment to recognize myocardial involvement in SSc. More extensive cohorts with long-term follow-ups are warranted to validate the significance of the initial findings of this cohort. Our preliminary data can be a starting point for these studies to determine the prognostic importance of myocardial involvement in SSc.

Limitations

Our study has some limitations. The main limitation is its modest size, the other is its retrospective design. In addition, the presence of myocardial inflammation and fibrosis in our patients was not assessed by myocardial biopsy, which is the gold standard. However, myocardial biopsy is too invasive to be considered a routine diagnostic procedure.

Also, some patients had a follow-up CMR, and we have utilized the first one; the results may be different from the subsequent CMRs.

Since data for two commonly used biochemical parameters, troponin and proBNP, were unavailable for all 93 patients, we did not mention them in this study. Finally, we did not evaluate the prognostic impact of our results.

In summary, this retrospective cohort study indicates that the presence of ASM in ECHO is a predictor of abnormal CMR in SSc patients, and it could provide a basis for more refined screening algorithms for at-risk patients. Therefore, a precise assessment of ASM may serve as an important point for selecting the patients that should be evaluated by CMR for early detection of myocardial involvement. Finally, the study provides valuable insights into the CMR findings in these patients.

Supplementary material

Supplementary material is available at Rheumatology online.

Data availability

As this project has been designed in several phases and is still ongoing, we cannot currently share any data from the study publicly. The data will be shared when a reasonable request is made to the corresponding author.

Funding

No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article.

Disclosure statement: The authors have declared no conflicts of interest.

References