Background Cardiovascular (CV) involvement in patients with systemic lupus erythematosus (SLE) is presumably subclinical for the major part of its evolution. We evaluated the associations between high-sensitive troponin T (hs-TropT), a sensitive marker of myocardial injury, and CV involvement using cardiac magnetic resonance (CMR).
Methods and results This is a two-centre (London and Frankfurt) CMR imaging study at 3.0 Tesla of consecutive 92 patients with SLE free of cardiac symptoms, undergoing screening for cardiac involvement. Venous samples were drawn and analysed post-hoc for cardiac biomarkers, including hs-TropT, high-sensitive C reactive protein and N-terminal pro brain natriuretic peptide. Compared with age-matched/gender-matched non-SLE controls (n=78), patients had significantly raised cardiac biomarker levels, native T1 and T2, aortic and ventricular stiffness, and reduced global longitudinal strain (p<0.01). In SLE, hs-TropT was significantly and independently associated with native T2, followed by the models including native T1 and aortic stiffness (Χ2 0.462, p<0.01). There were no relationships between hs-TropT and age, gender, CV risk factors, duration of systemic disease, cardiac structure or function, or late gadolinium enhancement.
Conclusions Patients with SLE have a high prevalence of subclinical myocardial injury as demonstrated by raised high-sensitive troponin levels. CMR with T2 mapping reveals myocardial oedema as the strongest predictor of hs-TropT release, underscoring the inflammatory interstitial remodelling as the main mechanism of injury. Patients without active myocardial inflammation demonstrate diffuse interstitial remodelling and increased vascular stiffness. These findings substantiate the role of CMR in screening of subclinical cardiac involvement.
Trial registration numer NCT02407197; Results.
- systemic lupus erythematosus
- cardiovascular disease
- magnetic resonance imaging
Statistics from Altmetric.com
Systemic lupus erythematosus (SLE) is a chronic systemic inflammatory condition associated with complex cardiovascular (CV) involvement with high morbidity and CV mortality.1–3 It is recognised that the first decade since the diagnosis of systemic disease represents a period of intense systemic disease activity. This includes CV involvement, which in contrast to systemic disease progresses silently for the major part of its evolution. Only a small proportion of patients eventually proceed into manifest cardiac disease, most commonly congestive heart failure, which is poorly reversible and difficult to treat.4–8 Underscored by the pathophysiological pathways of inflammation and remodelling, CV involvement is phenotypically characterised by non-ischaemic cardiomyopathy, and coronary and aortic vasculitis.2 9–14 A number of previous studies using cardiovascular magnetic resonance (CMR) reported significantly elevated non-invasive quantifiable myocardial tissue measures of remodelling and inflammation T1 and T2 mapping, and myocardial scar by late gadolinium enhancement (LGE), respectively (reviewed in refs 9 10 15–21). Studies further demonstrated considerable vascular involvement in SLE, including carotid, aortic and coronary vessel wall vasculitis, and high frequency of myocardial microvascular disease and myocardial fibrosis.2 6 7 11 22–29 Despite the above knowledge, the relationship with the biomarkers of myocardial injury (high-sensitive troponin, hs-TropT26 27), inflammation (high-sensitive C reactive protein, hs-CRP) and remodelling (N-terminal pro brain natriuretic peptide, NT-proBNP) remains unclear.28 Furthermore, the pathophysiological relationship between these markers and changes observed in subclinical CV involvement is not ascertained.
Consecutive 92 patients with an established diagnosis of SLE as per the American College of Rheumatology revised classification criteria,29 without previously known or symptomatic cardiac disease, were referred for screening with CMR for CV involvement from the local rheumatology departments (London, n=56; Frankfurt, n=36). Seventy-eight (non-SLE) subjects, matched for age, gender and CV risk factor profile, with low-pretest likelihood for left ventricular (LV) cardiomyopathy, no clinical or serological evidence for systemic inflammation, taking no anti-inflammatory medication, and no history of cardiac events or known coronary artery disease (CAD), served as controls. The general contraindication to contrast-enhanced CMR was observed as per local rules (known allergy to contrast agents, pregnancy, cochlear implants, cerebral aneurysm clips, non-CMR compatible pacemakers; no subject was excluded due to these contraindications in the present study). Patient characteristics were recorded using a standardised questionnaire for all subjects, including symptoms, age, gender, body mass index, renal function, CV risk factors and immunosuppressive therapy (table 1).
Cardiovascular magnetic resonance
All subjects underwent a routine clinical scan protocol using a 3 Tesla clinical scanner (Achieva, Philips Healthcare, Best, The Netherlands, and Skyra, Siemens Healthineers, Erlangen), CMR protocol (figure 1). After standardised patient-specific planning, myocardial T1 and T2 mapping was performed in a single mid-ventricular short-axis (SAX) slice. T1 mapping was performed using modified Look-Locker imaging.30 31 For T2 mapping, a hybrid gradient and spin echo (GraSE) sequence was used at the London site,32 whereas in Frankfurt T2 mapping fast low angle shot (FLASH) sequence was employed.33 34 All sequence types and parameters have been validated and reported previously.12 18 20 Sequence-specific normal ranges were employed (native T1: 3.0 T; mean of the normal range 1052±23 ms; ie, upper limit of normal range: 1098 ms at 3.0 T35; native T2: GraSE sequence: 45±4 ms32 36; T2-FLASH sequence 35±4 ms33 34 36). Myocardial perfusion imaging with adenosine infusion (140 μg/kg/min) and administration of 0.1 mmol/kg body weight gadobutrol (Gadovist, Bayer AG, Leverkusen, Germany) was performed,37 38 followed by volumetric cavity assessment by whole-heart coverage of SAX slices. Central aortic pulse wave velocity (PWV), a measure of aortic stiffness, was obtained with an inplane sagittal oblique acquisition of the ascending and descending aorta, using a retrospectively gated, free-breathing, phase-contrast gradient echo sequence, as reported previously.39 40 Representative acquisitions are included in figure 1.
Cardiac volume and function were quantified using commercially available software (Medis, Leiden) following standardised postprocessing recommendations.41 LGE was characterised as present or absent, and ischaemic or non-ischaemic in type, based on the predominant pattern.38 41 Quantitative tissue characterisation and myocardial deformation analysis were performed by the core-lab (Goethe CVI, Frankfurt), blind to the underlying subject group allocation and the time-point of the examination. The rates of T1 and T2 relaxation were measured in the septal myocardium of the mid-ventricular SAX using the ConSept approach, as previously described.31 42 Areas of LGE were excluded from mapping the region of interest (ROIs) to allow T1 and T2 measurements to avoid areas of replacement scar.42 43 Global longitudinal and circumferential strain was measured within a feature tracking application using feature tracking two-dimensional prototype software (TomTec, Munich, Germany), as previously described.44 Longitudinal deformation was obtained in three long-axis views, whereas circumferential included three SAX slices. Both are expressed as an absolute global peak systolic strain per direction.9 All patients underwent venous blood sampling at the time of the CMR study. Plasma samples were frozen at −80°C and analysed subsequently using standardised commercially available test kit analysis for hs-TropT, hs-CRP and NT-proBNP (Elecsys 2010, Roche, Basel, Switzerland). Analytical validation and limits of detection of the hs-TropT test were used to define normal/abnormal (using a cut-off value of 99th percentile of 13.9 ng/L, as previously reported).45
Statistical analysis was performed using SPSS V.24. Normality of distributions was tested using Shapiro-Wilk test. Categorical data are expressed as counts (percentages), and continuous variables as mean±SD or median (range), as appropriate. Comparisons between groups were performed using Student’s t-test or one-way analysis of variance for normally distributed variables, and χ2 and Mann-Whitney test for non-normally distributed variables. Fisher’s exact tests were conducted for proportions. The associations were analysed by univariate and multivariate linear regression analyses. Collinearity diagnostics was used to examine the variance inflation factor analysis. Interobserver and intraobserver reproducibility and agreement of postprocessing approaches have been reported previously.36 42 46 All tests were two-tailed and p values <0.05 were considered statistically significant.
Subject characteristics, results of blood markers and CMR measurements are summarised in table 1. Patients were well matched to controls for age, gender, blood pressure, heart rate and CV risk profile. The average duration of systemic disease was 7±5 years. Patients with SLE took a variety of disease-modifying antirheumatic medication, including prednisolone (39, 42%), hydroxychloroquine (37, 40%), mycophenolate (25, 27%) or others (11, 12%). Twenty-nine patients with SLE (32%) were taking anticoagulation treatment in the context of antiphospholipid syndrome.
Compared with controls, LV end-systolic volume (LV-ESV) and mass were increased and LV ejection fraction reduced in patients with SLE (p<0.05 for all), although the means remained within the normal range.47 48 Myocardial native T1 and T2 and central aortic PWV were significantly higher in patients with SLE, whereas the global longitudinal strain (GLS) was reduced (p<0.001 for all). LGE was present in 28 (30%) patients with SLE, of whom a minority (6, 7%) were ischaemic in type.41 Pericardial enhancement and effusion were present in 9 (13%) and 7 (10%) patients (figure 1). Myocardial perfusion imaging in patients with SLE revealed no subjects with regional hypoperfusion due to obstructive epicardial CAD,40 whereas microvascular disease pattern was observed in 24 (31%).11
Hs-TropT levels were detectable in 81% (n=139) of all subjects. A total of 50 (64%) patients with SLE had hs-TropT above the 99th percentile,45 49 although only 4 (8%) above the clinically relevant threshold >30 ng/L.49 Compared with controls, patients with SLE had mildly but significantly raised hs-CRP and NT-proBNP and reduced estimated glomerular filtration rate levels (p<0.01; table 1). In patients with evidence of myocardial oedema defined as raised native T2 (defined as >2 SD above the mean of the sequence-specific normal range, n=48 (61%)),32 33 45 there was also a significant rise in hs-TropT, hs-CRP and NT-proBNP (figure 2).
Analyses of the relationships between hs-TropT, serological markers and CMR measures of CV involvement are presented in table 2 and figures 3 and 4. In controls, hs-TropT levels were mildly associated with native T1 and LVmass (<0.05). In patients with SLE, hs-TropT was associated with native T1 and T2, GLS, E/e’, PWV, hs-CRP and NT-proBNP (p<0.01 for all). Furthermore, hs-CRP was associated with PWV (r=0.46), native T1 (r=0.39) and T2 (r=0.56) (p<0.01 for all), but not E/e’. NT-proBNP was moderately associated with PWV (r=0.35), native T1 (r=0.34), native T2 (0.29) and strongly associated with E/e’ (r=0.79) and left atrium (LA) area (r=0.60) (p<0.01 for all). In patients with SLE, there was no association with LVmass, left ventricular ejection fraction or left ventricular end-diastolic volume for any of the serological markers. The results of multivariate linear regression analysis (stepwise) of the relationships with hs-TropT are presented in table 2. On the contrary, in patients with SLE, hs-TropT was independently associated with native T2 (χ2 0.391, p<0.01), followed by the models which also included native T1 and PWV (χ2 0.462, p<0.01). The relationships between diffuse myocardial remodelling by native T1 and markers of aortic and ventricular stiffness and deformation were significant and accentuated in the subgroups of patients with evidence of myocardial inflammation (figure 3). The results of patients (n=3) who underwent endomyocardial biopsy (EMB) following CMR findings are included in table 3 and figure 5.
Patients with SLE and no known or symptomatic cardiac disease have a high prevalence of subclinical myocardial injury, as evidenced by detectable hs-TropT release. The extent of subclinical myocardial injury is independently associated with the relevant components of CV involvement, including myocardial inflammation (by native T2), interstitial remodelling (by native T1), aortic (by PWV) and ventricular stiffness (by E/e’), respectively. There are no relationships between hs-TropT and changes of cardiac function or structure. The role of inflammation as the pathophysiological driver of myocardial injury is further reiterated by a considerably greater rise of hs-TropT (as well as hs-CRP and NT-proBNP) in the subgroup of patients with significantly raised native T2, indicating the presence of oedema (33,34). These findings substantiate the role of CMR with native T1 and T2 mapping and PWV in screening for the presence of subclinical cardiac involvement.
The present observations importantly complement the insights of previous studies in deciphering the natural course of inflammatory cardiomyopathy in patients with SLE.15 18 The significantly and prevalently raised troponin levels in patients with SLE provide a sensitive validation of the subclinical myocardial injury.26 We and others have previously shown raised troponin levels in clinically manifest lupus myocarditis (10,16,52-54). In the present study, the post-hoc analysis of systematically collected research blood samples in patients with no known CV involvement reveals detectable troponin release in a majority of patients with SLE, which are on average significantly higher than in non-SLE controls.28 50 Different from the presentation of an acute coronary syndrome, troponin levels in most patients with SLE remain below the acute coronary syndrome (ACS) threshold (>30 ng/L),45 49 indicating a much smaller amount of the total myocardial necrosis in SLE myocardial injury, compared with an acute atherosclerotic coronary artery occlusion. This important difference in the underlying pathophysiology—the non-coronary mechanism of myocardial injury in SLE—is further apparent through the absence of a typical ischaemic pattern of LGE, corresponding to a coronary vascular territory. On the contrary, the pattern of myocardial scar (if observed) is non-ischaemic and intramyocardial (41,43) (figure 1). Furthermore, tissue characterisation with native T2 mapping, a sensitive marker of myocardial oedema, is the strongest predictor of hs-TropT release, demonstrating the underlying myocardial inflammation as the main mechanism of injury. hs-TropT levels and native T1 both remain significantly elevated (compared with controls) in patients without significant myocardial oedema (normal native T2), although commonly beyond the abnormal threshold (hs-TropT—13.9 ng/L; native T1—1098 ms; please see the Methods section), indicating the sustained diffuse process of interstitial remodelling and fibrosis.43 46 These observations correspond with the finding of hs-CRP rise in the group with oedema (raised native T2) but not without, despite it being commonly subdued in SLE.51–53 A number of previous observations on NT-proBNP levels in patients with SLE, such as associations with echocardiographic LA diameter and aortic stiffness,54 but not LV structure or function, were also replicated in our study.55 We expand these observations by demonstrating systematic associations between NT-proBNP with aortic and ventricular stiffness (by PWV and E/e’), and interstitial and inflammatory remodelling (by native T1 and T2). In non-ischaemic cardiomyopathy, myocardial native T1 correlates with the degree of aortic and ventricular stiffness38 56 57; similar associations in the present cohort of patients with SLE relate the communality of vascular and myocardial remodelling processes. Our findings also revealed detectable structural LV enlargement (LV-ESV and LVmass) and reduction in global LV and RV function, although not reaching the clinically relevant thresholds nor showing significant association with hs-TropT. Thus, the absence of significant associations between hs-TropT and markers of cardiac structure and function exposes the limited value of methods, such as echocardiography, in detecting early abnormalities or assessment of the subclinical disease. Our observations first communicate the link between the pathophysiological mechanism of myocardial inflammation and CV involvement based on the common direction of change in the serological and imaging measures. Second, native T1 and T2 are quantitative measures of tissue abnormalities, allowing to stratify the degree of inflammatory and interstitial myocardial involvement. Third, whereas significant hs-TropT levels correspond well with the presence of myocardial oedema, ‘biochemically low-grade’ troponin release (<13.9 ng/L) in patients with diffuse fibrosis but no active inflammation may lead to falsely misclassify this subgroup as having no CV involvement. Native T1 can unravel myocardial abnormalities that develop beyond the analytical limits of detection by the hs-TropT assay (45). Whereas it might be possible to adjust the troponin thresholds to the context of inflammatory injury, further studies are required to establish the sensitivity/specificity ranges, prognostic associations and its ability to change treatment. Troponin identifies patients with sufficiently active inflammation leading to cell necrosis and its release. Also, raised troponin levels may frequently persist over several weeks, obscuring the insight into changes in disease activity or recovery. On the contrary, imaging measures respond rapidly to altered disease dynamics and/or to anti-inflammatory treatment, providing an objective and immediate clinical assessment.10 56 While not used routinely to support the diagnosis of myocardial involvement in SLE, the results from the few patients who underwent EMB following CMR results corroborate our findings with histological evidence of chronic autoimmune macrophage-rich diffuse interstitial myocardial inflammation and interstitial and perivascular fibrosis.4
A few limitations apply. This is a two-centre multivendor observational study with patients on stable treatment, as guided by their systemic and not cardiac disease. Whereas an effect of reduced inflammation due to treatment could not have been fully avoided, we strived for consecutive broad inclusion of subjects with SLE, allowing a balanced and realistic representation of a typical cohort. Standardisation of data acquisition and postprocessing for application of T1 and T2 mapping methods were achieved by unifying the T1 mapping imaging parameters, performing the on-the-fly image quality control and by using centralised postprocessing.35 Whereas the T2 mapping sequence was different at the two sites, the acquisitions show similar precision and accuracy (ie, sensitivity to inflammatory substrate).57
In summary, patients with SLE without known or symptomatic cardiac disease have a high prevalence of subclinical myocardial injury as evidenced by a significant hs-TropT elevation. This is paralleled by a significant rise in hs-CRP and NT-proBNP, as well as of the non-invasive imaging measures of CV inflammatory involvement. We demonstrate a strong and predictive relationship between hs-TropT and non-invasive tissue measures of diffuse myocardial fibrosis and inflammation by native T1 and T2, aortic by PWV, and ventricular stiffness by E/e’, respectively, but not changes in cardiac function or structure. Taken together, these findings reveal a significant subclinical CV involvement in patients with SLE, which is amenable to detection with non-invasive tissue characterisation measures and is directly related to CV inflammatory injury.
We would like to acknowledge the support of cardiac radiographers, local cardiology and rheumatology departments at participating centres, and MR clinical scientists, Dr Christian Stehning (Philips), Dr Andreas Greiser (Siemens) and Dr Ralph Strecker (Siemens). We would also like to acknowledge Banher Sandhu (London), Catia Canepa and Franziska Weiss (Frankfurt) for their contribution to clinical data collection.
Handling editor Josef S Smolen
Contributors All authors have made relevant contributions to this manuscript as outlined by the International Committee of Medical Journal Editors. Substantial contributions to the conception or design of the work; or the acquisition, analysis or interpretation of data for the work: LW, RH, AB, UD, HB, SS, DPD’C, GC-W, MM, KS, CA, KK, TJV, AMZ, EN, VOP. Drafting the work or revising it critically for important intellectual content: LW, RH, AB, UD, HB, EN, VOP. Final approval of the version to be published: LW, RH, AB, UD, HB, SS, DPD’C, GC-W, MM, KS, CA, KK, TJV, AMZ, EN, VOP. Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: LW, EN, VOP.
Funding Department of Health via the National Institute for Health Research (NIHR) comprehensive Biomedical Research Centre (BRC) award to Guy’s and St Thomas’ NHS Foundation Trust in partnership with King’s College London and King’s College Hospital NHS Foundation Trust. Spanish Cardiology Society Fellowship to RH. German Ministry of Education and Research via the German Centre for Cardiovascular Research (DZHK) to VOP, AMZ and EN.
Competing interests None declared.
Patient consent Obtained.
Ethics approval The study protocol was reviewed and approved by the respective local ethics committee (Guy’s and St Thomas' Ethics Committee (London site), and Ethikkommission, Goethe University Frankfurt, Germany). All procedures were carried out in accordance with the Declaration of Helsinki.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.