Evaluation of Systemic Sclerosis Primary Heart Involvement and Chronic Heart Failure in the European Scleroderma Trials and Research Cohort

Abstract

Background

Systemic sclerosis (SSc) primary heart involvement (SSc‐pHI) is one of the leading causes of mortality in SSc. We aimed to evaluate risk factors for SSc‐pHI and its progression and the outcomes in the EUSTAR (European Scleroderma Trials and Research) cohort.

Methods

SSc‐pHI was defined according to the World Scleroderma Foundation/Heart Failure Association definition. Data from 5741 patients with SSc in the EUSTAR cohort were analyzed. Additional cardiovascular data were collected from a subcohort of 838 patients with SSc. Lasso regression was used for risk factor analyses. Kaplan–Meier estimator was used for survival analyses. Progression of SSc‐pHI was evaluated by a study definition developed by rheumatology and cardiology experts.

Results

Risk factors for the presence of SSc‐pHI comprised skeletal muscle atrophy (odds ratio [OR], 2.00 [95% CI, 1.00–2.68]), age (OR, 1.91 [95% CI, 1.73–2.03]), male sex (OR, 1.77 [95% CI, 1.42–2.05]), swollen joints (OR, 1.70 [95% CI, 1.47–1.98]), skeletal muscle weakness (OR, 1.38 [95% CI, 1.00–1.85]), and tendon friction rubs (OR, 1.46 [95% CI, 1.00–1.77]) (n=3276). Telangiectasia (OR, 2.10 [95% CI, 1.38–2.72]), intestinal symptoms (OR, 1.70 [95% CI, 1.04–2.42]), age (OR, 1.47 [95% CI, 1.21–1.62]), and antitopoisomerase I antibodies (OR, 1.37 [95% CI, 1.00–1.77]) were associated with an increased risk for new onset of SSc‐pHI (n=1000). Survival rate was significantly lower in patients with SSc‐pHI than in those without (P value <0.0001, n=3768). Patients with SSc‐pHI had a lower survival rate than patients with interstitial lung disease (n=3365). Swollen joints were associated with an increased risk of progressive SSc‐pHI (OR, 2.49 [95% CI, 1.79–3.52]) (n=595). Tendon friction rubs (OR, 1.21 [95% CI, 0.94–1.90]) increased the risk of heart failure with preserved ejection fraction in patients with SSc‐pHI.

Conclusions

We defined progressive SSc‐pHI and identified risk factors for new onset and progression of SSc‐pHI and for SSc‐pHI‐associated heart failure with preserved ejection fraction in the largest cohort with SSc. These findings may guide patient stratification for diagnostic workup and therapy.

Nonstandard Abbreviations and Acronyms

EUSTAR

European Scleroderma Trials and Research

GLS

global longitudinal strain

HFpEF

heart failure with preserved ejection fraction

ILD

interstitial lung disease

PAH

pulmonary arterial hypertension

SSc

systemic sclerosis

SSc‐pHI

systemic sclerosis‐associated primary heart involvement

WSF/HFA

World Scleroderma Foundation/Heart Failure Association

Clinical Perspective.

What Is New?

• Together with cardiology and rheumatology experts we defined progressive systemic sclerosis primary heart involvement (SSc‐pHI) and identified telangiectasia; intestinal symptoms such as diarrhea, bloating, or constipation; age; and antitopoisomerase I antibody positivity as risk factors for incident SSc‐pHI and swollen joints as a risk factor for progressive SSc‐pHI in the largest multinational cohort with SSc.

• We identified a lower survival rate in patients with SSc‐pHI compared with patients with SSc‐associated interstitial lung disease after adjusting for confounders. SSc‐pHI‐associated heart failure with preserved ejection fraction affected approximately one‐third of patients with SSc‐pHI; however, it did not further affect mortality in patients with SSc‐pHI.

What Are the Clinical Implications?

• Our findings increase the awareness for SSc‐pHI and SSc‐pHI‐associated heart failure with preserved ejection fraction in patients with SSc and may guide patient stratification for diagnostic workup and optimal treatment in SSc to improve outcomes in patients with SSc.

Systemic sclerosis (SSc) is a systemic fibrosing autoimmune rheumatic disease. ¹ Among all systemic rheumatic diseases, SSc is associated with the highest disease‐related mortality, and heart involvement is one of the leading causes of mortality in patients with SSc. ² The reported prevalence of clinically manifest primary heart involvement (SSc‐pHI) ranges from 7% to 39%, ³ with clinical findings of arrhythmogenic disturbances and heart failure (HF). Autopsy studies, however, have demonstrated myocardial fibrosis, myocardial inflammation, or contraction band necrosis in up to 80% of patients with SSc. ⁴ , ⁵ Clinically manifest SSc‐pHI is associated with a 5‐year mortality of 75%. ² , ⁵ , ⁶

SSc‐pHI comprises a broad range of clinical entities, whose pathophysiology and clinical course are incompletely understood. Only recently, a literature‐driven consensus‐based definition of SSc‐pHI was published by the World Scleroderma Foundation/Heart Failure Association (WSF/HFA), defining SSc‐pHI as cardiac abnormalities that are predominantly attributable to SSc rather than other causes (such as ischemic heart disease, arterial hypertension, drug toxicity, other cardiomyopathy, primary valvular disease) or SSc complications (such as SSc‐associated pulmonary arterial hypertension [SSc‐PAH], renal involvement, SSc‐associated interstitial lung disease [SSc‐ILD]). ⁷ Treatment of SSc‐pHI is challenging due to the lack of evidence‐based treatment guidelines. Furthermore, the previous lack of a standardized definition of SSc‐pHI further complicates the interpretation, integration, and extrapolation of previously published data. Moreover, most of the previous studies did not strictly differentiate between primary and secondary heart involvement in SSc, and previous clinical, histopathological, and imaging studies reporting on SSc‐pHI also included in particular patients with SSc‐PAH, scleroderma renal crisis, arterial hypertension, or diabetes and as such confounded the results related to SSc‐pHI. ² , ⁸ , ⁹ Recently a consensus of the WSF/HFA on the assessment of SSc‐pHI has been published offering guidance on screening, diagnosis, and follow‐up assessment of SSc‐pHI. ¹⁰ However, further studies on the heterogeneity of SSc‐pHI are needed and these will require a large sample size, which is achievable only by multicenter studies.

The extent, course, and outcome of SSc‐pHI are highly variable. It is currently impossible to predict whether individual patients will develop clinically relevant SSc‐pHI, whether it will be progressive, and how it may respond to immunomodulatory or antifibrotic therapies.

There are few proposed risk factors for new onset of SSc‐pHI; however, these data are based on relatively small numbers of patients and their clinical relevance is further challenged by the aforementioned aspects. ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ Screening for SSc‐pHI with serum cardiac biomarkers has been recommended, although they are not specific for SSc‐pHI and can be influenced by concomitant cardiac diseases, by SSc‐PAH, or by impaired renal function. ⁹ , ¹⁰ , ¹⁸

The data available for progression of SSc‐pHI are even more limited. Even though follow‐up studies of SSc‐pHI are scarce, it is known that some patients with SSc‐pHI remain relatively stable over time, whereas others progress despite treatment. However, a standardized definition of progressive SSc‐pHI is currently lacking and no risk factors for the progression of SSc‐pHI have been described so far.

HF with mildly reduced ejection fraction (EF) and HF with reduced EF are uncommon in SSc. ⁵ However, little is known about the prevalence and outcome of HF with preserved EF (HFpEF) in patients with SSc‐pHI and the current data are confounded by the simultaneous presence of cardiovascular risk factors. ¹⁹ The current possibility of using innovative drugs for HFpEF such as SGLT‐2 (sodium‐glucose cotransporter 2) inhibitors dictates the need to define patients with HFpEF in the context of SSc‐pHI.

Here, we aimed to characterize risk factors, course, and outcome of SSc‐pHI, as defined by the WSF‐HFA, in the EUSTAR (European Scleroderma Trials and Research) cohort, the largest multinational database of patients with SSc and to confirm these results in a EUSTAR subcohort with additional cardiac phenotyping. We further aimed to provide a definition for progressive SSc‐pHI and to identify risk factors for progression of SSc‐pHI. Moreover, we aimed to investigate the presence and the outcome of HFpEF in the context of SSc‐pHI and to identify risk factors for HFpEF in patients with SSc‐pHI.

METHODS

Study Cohort

Our primary study cohort included 17 760 patients with SSc from the EUSTAR database. This cohort will be referred to throughout the article as the EUSTAR cohort. Data collected before the first of July 2021 were analyzed. A subcohort of patients with SSc was phenotyped in more detail by obtaining data beyond that routinely collected in the EUSTAR database, which included echocardiographic, electrocardiographic, cardiac magnetic resonance (CMR) parameters, circulating cardiac biomarkers, and myocardial biopsy data together with supplementary cardiovascular risk factors. This allows for a stricter delineation between cardiac abnormalities in the context of SSc‐pHI and those in the context of secondary SSc heart involvement or non‐SSc‐related diseases. This cohort will be referred to throughout the article as the EUSTAR subcohort. EUSTAR centers with particular expertise in SSc‐pHI were identified and contacted for their collaboration on this project. Eleven EUSTAR centers agreed to collaborate for extended phenotyping and provided further cardiac data from a total of 838 patients and 1809 visits. A table with predefined findings including information regarding antibody profile as well as specific echocardiographic, electrocardiographic, CMR, circulating cardiac biomarker, and myocardial biopsy data together with supplementary cardiovascular risk factors, which are not included in the EUSTAR database, was generated (Table S1). The participating EUSTAR centers completed questionnaires for patients with SSc with clinical suspicion of or confirmed SSc‐pHI as well as for at least an equal number of patients with SSc without suspicion of SSc‐pHI or patients in whom SSc‐pHI was excluded based on CMR or myocardial biopsy. Centers were asked to provide the data for as many follow‐up visits as possible. Patients with SSc with and without SSc‐pHI were matched by the year of their first EUSTAR visit. The additional SSc‐pHI data were matched with the data from the EUSTAR database for each patient by the unique EUSTAR visit ID. Subsequent analyses were performed both in the subcohort of patients with extensive cardiac phenotyping as well as in the whole EUSTAR cohort.

Each EUSTAR center obtained approval from their local ethics committee and all registered patients granted their written informed consent.

The data that support the findings of this study are available from the corresponding author upon reasonable request and after providing the informed written consent of the EUSTAR committee.

Definition of SSc‐pHI

For the definition of SSc‐pHI, a set of inclusion and exclusion criteria was generated based on consensus of rheumatology and cardiology experts (Table S2). Because several of the parameters of the inclusion and exclusion criteria are not routinely collected in the EUSTAR database, only a reduced set of inclusion and exclusion criteria was evaluated in the whole EUSTAR cohort (Table S3). We defined patients with SSc‐pHI as patients fulfilling at least 1 of the inclusion criteria and none of the exclusion criteria, and patients without SSc‐pHI as patients fulfilling none of the inclusion criteria and none of the exclusion criteria.

SSc‐pHI was defined as (1) histologically or CMR‐proven nonischemic myocardial fibrosis or (peri)myocarditis, (2) left ventricular (LV) systolic dysfunction (defined as a LVEF <50% or a global longitudinal strain [GLS] <16% [GLS expressed as a negative number]), (3) LV diastolic dysfunction (defined in echocardiography as E/A ≥2, or E/A <2 and >0.8 and at least 2 out of the following 3 criteria positive: (a) E/e >14; (b) tricuspid regurgitation velocity >2.8 m/s; (c) left atrium volume index >34 mL/m², ²⁰ (4) cardiac conduction defects, or (5) any type of arrhythmias including any number of ectopic beats. Even though valvular disease could be a consequence of SSc, patients with valvulopathies of at least middle grade severity were excluded from our analyses due to the lack of tools to differentiate between valvulopathies in the context of SSc‐pHI and degeneration or infection. Even though pericardial effusion can also be a manifestation of SSc‐pHI, we did not include pericardial effusion in our analyses due to its limited discriminative capacity for SSc‐pHI over secondary heart involvement and other possible causes.

To allow for a strict differentiation between SSc‐pHI and secondary heart involvement or other cardiac diseases, patients with SSc with right heart catheterization‐proven PAH or a pulmonary artery systolic pressure in echocardiography of at least 40 mm Hg, patients with diabetes, coronary heart disease, myocardial infarction, arterial hypertension, preexisting cardiomyopathy, moderate or severe valvulopathies, reduced renal function with a glomerular filtration rate <60 mL/min/1.73 m², or history of scleroderma renal crisis, stem cell transplantation, drug‐induced toxicity, or thorax‐radiotherapy were excluded from further analyses. ²¹ This definition of SSc‐pHI is in line with the recent literature‐driven consensus‐based WSF/HFA definition, which defines SSc‐pHI as all cardiac abnormalities that are primarily attributable to SSc and are not due to other SSc‐complications (such as SSc‐PAH, renal involvement, SSc‐ILD) or non‐SSc‐related diseases (such as ischemic heart disease, arterial hypertension, drug toxicity, other cardiomyopathy, primary valvular disease). ⁷

Myocardial fibrosis in CMR was defined by late gadolinium enhancement, increased native T1 values, or increased extracellular volume fraction as reported by cardiologists/radiologists. ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ (Peri)myocarditis in CMR was defined according to the 2018 revised Lake Louise criteria. ²⁸ Histopathological diagnosis of myocardial fibrosis and myocarditis in the context of SSc‐pHI was performed in accordance with previous publications. ⁸

Definition of Progressive SSc‐pHI

Because no definition of progressive SSc‐pHI was available, a definition of progression of SSc‐pHI was developed as part of our study with rheumatology and cardiology experts using the nominal group technique. The process consisted of 4 steps. Four rheumatologists and 2 cardiologists were asked how they would define the progression of SSc‐pHI. Each of them independently generated a list of statements. Next, the round‐robin method was employed to record all statements from all participants. In the third step, the statements were discussed with all participants and merged. The final statements were then voted by each participant on a scale of agreement ranging from 0% to 100%. The consensus definition included all statements with an agreement of at least 75%.

Progression of SSc‐pHI was defined as at least 1 of the following, acquired over any time frame and not explained by any other causes:

1. Progression of SSc‐pHI in CMR, defined as (1) any quantitative increase in the scar extent or appearance of a new fibrotic area in a follow‐up CMR; (2) new onset of myocardial fibrosis, defined by late gadolinium enhancement, increased native T1 values, or increased extracellular volume fraction, ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ in a follow‐up CMR in a patient with known SSc‐pHI, defined according to the inclusion criteria; and (3) new onset of (peri)myocarditis defined according to the 2018 revised Lake Louise criteria ²⁸ in a follow‐up CMR in a patient with known SSc‐pHI, defined according to the inclusion criteria.

2. ≥5% reduction in the LVEF in echocardiography between 2 successive echocardiograms.

3. ≥5% reduction in the GLS in 2 successive echocardiograms (GLS expressed as a negative number). ²⁹

4. Progression of left ventricular diastolic dysfunction defined as E/A ≥2, or E/A <2 and >0.8 and at least 2 of the following 3 criteria positive: (1) E/e >14; (2) tricuspid regurgitation velocity >2.8 m/s; or (3) left atrium volume index >34 mL/m² to restrictive dysfunction in 2 successive echocardiograms. ²⁰

5. Progression of left ventricular diastolic dysfunction to a systolic dysfunction defined either as a GLS (expressed as a negative number) ≤16% or an LVEF <50% in 2 successive echocardiograms.

6. New onset of diastolic dysfunction (defined as E/A ≥2, or E/A <2 and >0.8 and at least 2 of the following 3 criteria positive: (1) E/e > 14; (2) tricuspid regurgitation velocity >2.8 m/s; or (3) left atrium volume index >34 mL/m²) in a patient with SSc‐pHI without diastolic dysfunction at baseline.

7. New onset of systolic dysfunction defined as an LVEF <50% or a GLS (expressed as a negative number) ≤16% in a patient with SSc‐pHI without systolic dysfunction at baseline.

8. New onset of cardiac arrhythmia or a conduction block at any time during follow‐up in a patient with SSc‐pHI without cardiac arrhythmias or conduction blocks at baseline.

9. Pacemaker/implantable cardioverter‐defibrillator implantation at any time during follow‐up in a patient with SSc‐pHI and cardiac conduction disorders or arrythmias at baseline.

10. New onset of HFpEF in a patient with SSc‐pHI according to the 2021 European Society of Cardiology guidelines. ³⁰

Statistical Analysis

Descriptive statistics for cohort description were presented as absolute numbers and proportion for binary and nominal covariates. For continuous covariates mean±SD were used.

A complete case analysis was performed for the analysis of risk factors. ³¹ Hereby, the set of potential risk factors was chosen as a tradeoff between the amount of missing data and the relevance of each variable. The risk factor analysis was conducted for the 3 outcomes of interest: the presence, the new onset, and the progression of SSc‐pHI. Because all 3 outcomes are binary, a mixed logistic Lasso regression was used for the identification of risk factors, with a random intercept per patient. CIs were calculated using the 2.5th and 97.5th percentiles of the bootstrap distribution based on 200 iterations. The penalty term was chosen by searching along a predefined grid and the value with the lowest Bayesian information criterion was used for the final analysis. The Lasso regression models were adjusted for age and sex. For the survival analysis the distribution of potential confounding variables was balanced by using entropy balancing weights. By the use of entropy balancing weights the Kaplan–Meier estimator is not biased by different distributions of observed confounding variables. In addition to the Kaplan–Meier estimator, weighted Cox regression models were used for survival analyses. ³²

The following potential risk factors for the presence, new onset, or progression of SSc‐pHI and HFpEF for which data from at least 30% of visits were available were included in risk factor analyses: skeletal muscle atrophy, skeletal muscle weakness, swollen joints, tendon friction rubs, ILD, intestinal symptoms, telangiectasia, digital ulcers, antitopoisomerase I antibodies, age, and sex.

“Intestinal symptoms” refers to at least 1 of the following symptoms: diarrhea, constipation, or bloating. Skeletal muscle weakness and atrophy refer to the weakness or atrophy of skeletal muscle as measured by clinical examination.

All statistical analyses were performed using R version 4.2.1. ³³ The mixed logistic Lasso regression was implemented in the R package glmmLasso, ³⁴ and weighted Cox regression was calculated using the survival package ³⁵ with weights implemented in the weightIt package. ³⁶

Andrea‐Hermina Györfi and Tim Filla had access to all data in the study and take full responsibility for their integrity and the data analysis.

RESULTS

Skeletal Muscle Atrophy, Skeletal Muscle Weakness, Swollen Joints, Tendon Friction Rubs, Age, and Male Sex as Risk Factors for the Presence of SSc‐pHI

At the time of data collection, 17 760 patients with SSc had been included in the EUSTAR cohort. Of these patients, 5167 fulfilled 1 or more of the exclusion criteria and were thus not considered for further analyses. The most frequently fulfilled exclusion criteria were arterial hypertension (n=3982) and PAH (n=2044). A total of 6852 patients were excluded due to missing data regarding all inclusion criteria of the study (Figure 1). Of the 5741 remaining patients in the EUSTAR cohort, 2465 patients were excluded from further risk factor analyses due to missing data related to at least 1 of the investigated potential risk factors for SSc‐pHI. Of the remaining 3276 patients, 1531 patients were classified as having SSc‐pHI and 1745 patients were classified as not having SSc‐pHI. A majority (66.8%) of patients with SSc‐pHI had LV diastolic dysfunction, and 5.4% had LV systolic dysfunction; 39.1% of patients with SSc‐pHI had conduction defects, and 18.2% of patients had cardiac arrhythmias. Perimyocarditis was present in 27.8% of patients with SSc‐pHI. Clinical characteristics of patients with and without SSc‐pHI in the EUSTAR cohort are illustrated in Table 1. ³⁷ Skeletal muscle atrophy (odds ratio [OR], 2.00 [95% CI, 1.00–2.68]), age (OR, 1.91 [95% CI, 1.73–2.03]), male sex (OR, 1.77 [95% CI, 1.42–2.05]), swollen joints (OR, 1.70 [95% CI, 1.47–1.98]), skeletal muscle weakness (OR, 1.38 [95% CI, 1.00–1.85]), and tendon friction rubs (OR, 1.46 [95% CI, 1.00–1.77]) increased the risk for having SSc‐pHI among the 3276 patients of the EUSTAR cohort. Analysis of associations with individual clinical manifestations of SSc‐pHI revealed that swollen joints and tendon friction rubs increased the risk for conduction blocks. Male sex was associated with perimyocarditis and LV systolic dysfunction. Skeletal muscle atrophy increased the risk of cardiac arrhythmias, and skeletal muscle weakness was associated with LV systolic dysfunction (Figure 2 and Figure S2).

Figure 1. Schematic representation of the study design.

At the time of data collection (July 1, 2021), 17 760 patients with SSc had been included in the EUSTAR cohort. Of these patients, 5167 were excluded from further analyses as they fulfilled 1 or more of the exclusion criteria. Another 6852 patients were excluded due to missing data regarding all inclusion criteria of the study. The remaining 5741 patients from the EUSTAR cohort were used for further analyses: 3276 patients were used for the analysis of risk factors for the presence of SSc‐pHI, 1000 patients were used for the analysis of risk factors for the development of SSc‐pHI, and 595 patients were used for the analysis of risk factors for the progression of SSc‐pHI. Survival analyses included 3768 patients. EUSTAR indicates European Scleroderma Trials and Research; NA, Not available; and SSc‐pHI, systemic sclerosis primary heart involvement.

Table 1.

Clinical Characteristics of Patients With and Without SSc‐pHI in the EUSTAR Cohort

	No SSc‐pHI	SSc‐pHI
No.	1745	1531
Age, y (mean±SD)	49.76 (13.08)	56.62 (13.44)
Male sex (%)	223 (12.8)	265 (17.3)
Body mass index, kg/m2 (mean±SD)	23.79 (4.43)	24.18 (4.32)
Year of appearance of first non‐Raynaud's phenomenon manifestation (mean±SD)	2010.35 (8.40)	2007.98 (9.00)
Anticentromere antibodies positivity (%)	678 (40.1)	568 (38.1)
Antitopoisomerase I antibodies positivity (%)	619 (35.5)	566 (37.0)
Anti‐RNA‐polymerase III antibodies positivity (%)	105 (8.1)	62 (6.0)
Limited cutaneous systemic sclerosis (%)	1005 (67.6)	822 (65.0)
New York Heart Association stages
I	1181 (72.9)	855 (59.2)
II	379 (23.4)	493 (34.1)
III	54 (3.3)	89 (6.2)
IV	5 (0.3)	7 (0.5)
Cardiac arrhythmias (%)	0 (0.0)	143 (18.2)
Conduction blocks (%)	0 (0.0)	545 (39.1)
Diastolic dysfunction (%)	0 (0.0)	934 (66.8)
Perimyocarditis (%)	0 (0.0)	10 (27.8)
LVEF (mean±SD)	62.07 (5.01)	61.07 (7.81)
Systolic dysfunction (LVEF <50%) (%)	0 (0.0)	73 (5.4)
HF with preserved EF (%)	0	514 (36.2)
HF with mildly reduced EF (%)	0	21 (1.5)
HF with reduced EF (%)	0	22 (1.5)
ILD (%)	608 (40.0)	639 (45.7)
FVC % predicted (mean±SD)	97.99 (19.99)	95.55 (23.14)
Single breath method of measuring the diffusing lung capacity for carbon monoxide percentage predicted (mean±SD)	75.06 (18.89)	70.37 (20.35)
Extensive ILD (at least 20% fibrotic involvement in high‐resolution computed tomography or in indeterminate cases by an FVC <70% according to a previous definition, 37 %)	200 (48.8)	242 (67.4)
Tendon friction rubs (%)	60 (3.4)	91 (5.9)
Subcutaneous calcinosis (%)	94 (8.5)	65 (11.7)
Swollen joints (%)	734 (42.1)	1022 (66.8)
Rheumatoid factor positivity (%)	95 (11.8)	56 (14.3)
Skeletal muscle weakness (%)	171 (9.8)	219 (14.3)
Skeletal muscle atrophy (%)	41 (2.3)	86 (5.6)
Telangiectasia (%)	874 (50.1)	898 (58.7)
Digital ulcers (%)
Current	212 (12.1)	208 (13.6)
Never	1053 (60.3)	906 (59.2)
Previously	480 (27.5)	417 (27.2)
Ramified bushy capillaries (%)
Extensive	36 (8.5)	21 (9.5)
Moderate	76 (17.9)	52 (23.4)
None	202 (47.6)	91 (41.0)
Rare	110 (25.9)	58 (26.1)
Capillary loss (%)
Extensive	65 (14.3)	54 (21.3)
Moderate	114 (25.1)	65 (25.6)
None	182 (40.0)	78 (30.7)
Rare	94 (20.7)	57 (22.4)
Hemorrhages (%)
Extensive	54 (12.2)	23 (9.8)
Moderate	114 (25.7)	65 (27.8)
None	112 (25.3)	67 (28.6)
Rare	163 (36.8)	79 (33.8)
Giant capillaries (%)
Extensive	90 (18.9)	42 (16.7)
Moderate	159 (33.3)	88 (34.9)
None	54 (11.3)	23 (9.1)
Rare	174 (36.5)	99 (39.3)
Intestinal symptoms	349 (20.0)	377 (24.6)
Malabsorption syndrome (%)	7 (1.6)	12 (2.2)
EUSTAR disease activity index (mean±SD)	0.26 (0.78)	0.56 (1.29)
Health assessment questionnaire score (mean±SD)	0.51 (0.65)	0.74 (0.69)
Erythrocyte sedimentation rate (mean±SD)	16.61 (14.85)	20.74 (18.10)
C‐reactive protein elevation (%)	111 (18.2)	205 (22.7)
Patients symptomatic for cardiac disease (%)	8 (1.8)	52 (9.4)
Troponin T elevation (%)	5 (21.7)	11 (52.4)
N‐terminal pro‐B‐type natriuretic peptide, pg/mL (mean±SD)	192.72 (1234.06)	518.76 (4044.61)
Creatine kinase elevation (%)	43 (7.6)	98 (11.2)
Hypocomplementemia (%)	41 (8.9)	66 (9.0)
Current therapy
Calcium channel blockers (%)	540 (30.9)	500 (32.7)
Angiotensin‐converting enzyme inhibitors/angiotensin receptor blockers (%)	45 (2.6)	121 (7.9)
Antiplatelet drugs (%)	304 (17.4)	374 (24.4)
Prostanoids (%)	196 (11.2)	178 (11.6)
Phosphodiesterase‐5 inhibitors (%)	88 (5.0)	69 (4.5)
Endothelin receptor antagonist (%)	145 (8.3)	101 (6.6)
Glucocorticoids (%)	286 (16.4)	331 (21.6)
Nonsteroidal anti‐inflammatory drugss (%)	84 (4.8)	86 (5.6)
Proton pump inhibitors (%)	712 (40.8)	686 (44.8)
Conventional synthetic disease modifying drugs (including hydroxychloroquine, methotrexate, leflunomide, and sulfasalazine) (%)	226 (13.0)	236 (15.4)
Mycophenolate mofetil/mycophenolic acid (%)	183 (10.5)	158 (10.3)
Cyclophosphamide (%)	20 (1.1)	35 (2.3)
Tocilizumab (%)	39 (2.2)	21 (1.4)
Rituximab (%)	37 (2.1)	38 (2.5)
Biologicals (tocilizumab, rituximab, tumor necrosis factor alpha inhibitors, abatacept, and Janus kinase inhibitors, %)	81 (4.6)	64 (4.2)

EF indicates ejection fraction; EUSTAR, European Scleroderma Trials and Research; FVC, forced vital capacity; HF, heart failure; ILD, interstitial lung disease; LVEF, left ventricular ejection fraction; and SSc‐pHI, systemic sclerosis‐associated primary heart involvement.

Figure 2. Risk factor analysis for the presence of SSc‐pHI and its distinct clinical entities in the EUSTAR cohort.

Dot plot illustrating the odds of the presence of muscle atrophy, age, male sex, swollen joints, muscle weakness, tendon friction rubs, telangiectasia, antitopoisomerase I antibodies, intestinal symptoms, interstitial lung disease, and digital ulcers for the presence of SSc‐pHI (black dot), cardiac arrhythmias (blue rhombus), cardiac conduction blocks (yellow rhombus), left ventricular diastolic dysfunction (violet rhombus), perimyocarditis (green rhombus), and left ventricular systolic dysfunction (defined as a left ventricular ejection fraction below 50%) (red rhombus). EUSTAR indicates European Scleroderma Trials and Research; ILD, interstitial lung disease; OR, odds ratio; and SSc‐pHI, systemic sclerosis primary heart involvement.

Next, a subcohort of 838 patients with SSc from 11 EUSTAR centers with detailed cardiac phenotyping was analyzed. Of these patients, 484 were excluded from further analyses as they fulfilled at least 1 exclusion criterion of the study. The most frequent exclusion criteria were chronic kidney disease (n=237), presence of moderate or severe valvulopathies (n=142), arterial hypertension (n=99), presence of coronary heart disease (n=79), PAH (n=57), history of myocardial infarction (n=37), and presence of diabetes (n=20). An additional 27 patients were excluded due to missing information related to all inclusion criteria of the study. A total of 327 patients were included in further analyses; 67 patients were classified as having SSc‐pHI and 92 patients as not having SSc‐pHI. The clinical characteristics of patients with and without SSc‐pHI are illustrated in Table S4. Consistent with our previous results in the whole EUSTAR cohort, male sex (OR, 2.17 [95% CI, 1.00–4.21]), skeletal muscle weakness (OR, 1.98 [95% CI, 1.00–5.50]), and tendon friction rubs (OR, 1.87 [95% CI, 0.88–7.7]) increased the risk for having SSc‐pHI.

Telangiectasia, Intestinal Symptoms, Age, and Antitopoisomerase I Antibodies as Risk Factors for New Onset of SSc‐pHI

Next, we evaluated possible risk factors for new onset of SSc‐pHI in a time period of up to 1.25 years. In the EUSTAR cohort, 118 patients developed SSc‐pHI, and 882 did not develop SSc‐pHI. Telangiectasia (OR, 2.10 [95% CI, 1.38–2.72]), intestinal symptoms (OR, 1.70 [95% CI, 1.04–2.42]), age (OR, 1.47 [95% CI, 1.21–1.62]), and antitopoisomerase I antibody positivity (OR, 1.37 [95% CI, 1.00–1.77]) increased the risk for new onset of SSc‐pHI. Of note, the frequencies of vasoactive or immunosuppressive therapies were similar between the patients who developed and those who did not develop SSc‐pHI (Table 2). The clinical characteristics of patients with new onset of SSc‐pHI versus patients without new onset of SSc‐pHI are summarized in Table 2.

Table 2.

Clinical Characteristics of Patients With and Without New Onset of SSc‐pHI in the EUSTAR Cohort

	Without new onset of SSc‐pHI	With new onset of SSc‐pHI
No.	882	118
Age, y (mean±SD)	48.60 (12.61)	53.82 (11.93)
Male sex (%)	104 (11.8)	15 (12.7)
Body mass index, kg/m2 (mean±SD)	23.81 (4.36)	23.97 (4.18)
Year of appearance of first non‐Raynaud's phenomenon manifestation mean±SD	2008.56 (8.50)	2008.85 (8.48)
Anticentromere antibodies positivity (%)	341 (40.4)	45 (38.5)
Antitopoisomerase I antibodies positivity (%)	315 (35.7)	50 (42.4)
Anti‐RNA‐polymerase III antibodies positivity (%)	46 (7.5)	2 (2.2)
New York Heart Association stages
I	609 (73.5)	80 (71.4)
II	192 (23.2)	27 (24.1)
III	25 (3.0)	5 (4.5)
IV	3 (0.4)	0 (0.0)
Limited cutaneous systemic sclerosis (%)	508 (69.3)	72 (70.6)
ILD (%)	287 (35.9)	43 (38.1)
Extensive ILD (at least 20% fibrotic involvement in high‐resolution computed tomography or in indeterminate cases by an FVC <70% according to a previous definition, 37 %)	84 (52.5)	12 (44.4)
FVC % predicted (mean±SD)	98.70 (19.08)	101.37 (21.62)
Single breath method of measuring the diffusing lung capacity for carbon monoxide percentage predicted (mean±SD)	75.25 (18.37)	73.38 (18.14)
Tendon friction rubs (%)	36 (4.1)	9 (7.6)
Subcutaneous calcinosis (%)	35 (10.7)	7 (14.9)
Swollen joints (%)	582 (66.0)	77 (65.3)
Rheumatoid factor positivity (%)	27 (12.4)	7 (20.0)
Skeletal muscle weakness (%)	89 (10.1)	12 (10.2)
Skeletal muscle atrophy (%)	21 (2.4)	4 (3.4)
Telangiectasia (%)	439 (49.8)	82 (69.5)
Digital ulcers (%)
Current	105 (11.9)	15 (12.7)
Never	506 (57.4)	69 (58.5)
Previously	271 (30.7)	34 (28.8)
Ramified bushy capillaries (%)
Extensive	10 (7.5)	3 (10.3)
Moderate	32 (23.9)	6 (20.7)
None	59 (44.0)	14 (48.3)
Rare	33 (24.6)	6 (20.7)
Capillary loss (%)
Extensive	27 (17.6)	8 (25.0)
Moderate	41 (26.8)	13 (40.6)
None	52 (34.0)	9 (28.1)
Rare	33 (21.6)	2 (6.2)
Hemorrhages (%)
Extensive	21 (14.4)	6 (19.4)
Moderate	46 (31.5)	9 (29.0)
None	32 (21.9)	5 (16.1)
Rare	47 (32.2)	11 (35.5)
Giant capillaries (%)
Extensive	30 (18.9)	9 (26.5)
Moderate	63 (39.6)	9 (26.5)
None	13 (8.2)	7 (20.6)
Rare	53 (33.3)	9 (26.5)
Intestinal symptoms (%)	191 (21.7)	41 (34.7)
EUSTAR disease activity index (mean±SD)	0.46 (1.05)	0.48 (1.01)
Health assessment questionnaire score (mean±SD)	0.47 (0.62)	0.46 (0.54)
Erythrocyte sedimentation rate (mean±SD)	16.90 (16.03)	19.03 (16.49)
C‐reactive protein elevation (%)	103 (19.2)	12 (16.9)
Patient symptomatic for cardiac disease (%)	6 (1.6)	1 (1.6)
Troponin T elevation (%)	4 (21.1)	0 (0.0)
N‐terminal pro‐B‐type natriuretic peptide, pg/mL (mean±SD)	114.94 (191.48)	141.24 (139.38)
Creatine kinase elevation (%)	35 (7.2)	7 (11.1)
Hypocomplementemia (%)	40 (10.4)	5 (8.2)
Cigarette smoking ever (%)	144 (20.7)	23 (25.3)
Current therapy
Calcium channel blockers (%)	237 (26.9)	30 (25.4)
Angiotensin‐converting enzyme inhibitors/angiotensin receptor blockers (%)	14 (1.6)	3 (2.5)
Antiplatelet drugs (%)	149 (16.9)	17 (14.4)
Prostanoids (%)	90 (10.2)	14 (11.9)
Phosphodiesterase‐5 inhibitors (%)	29 (3.3)	7 (5.9)
Endothelin receptor antagonist (%)	69 (7.8)	13 (11.0)
Glucocorticoids (%)	126 (14.3)	17 (14.4)
Nonsteroidal anti‐inflammatory drugs (%)	38 (4.3)	4 (3.4)
Proton pump inhibitors (%)	314 (35.6)	45 (38.1)
Conventional synthetic disease modifying drugs (including hydroxychloroquine, methotrexate, leflunomide, and sulfasalazine) (%)	137 (15.5)	20 (16.9)
Mycophenolate mofetil/mycophenolic acid (%)	56 (6.3)	10 (8.5)
Cyclophosphamide (%)	6 (0.7)	2 (1.7)
Tocilizumab (%)	16 (1.8)	1 (0.8)
Rituximab (%)	16 (1.8)	1 (0.8)
Biologicals (tocilizumab, rituximab, tumor necrosis factor alpha inhibitors, abatacept, and Janus kinase inhibitors, %)	33 (3.7)	5 (4.2)

EUSTAR indicates European Scleroderma Trials and Research; FVC, forced vital capacity; ILD, interstitial lung disease; and SSc‐pHI, systemic sclerosis‐associated primary heart involvement.;

Increased Mortality in Patients With SSc‐pHI

From the whole EUSTAR cohort, 3768 patients (2554 with and 1214 without SSc‐pHI) were included in the survival analyses. SSc‐pHI was associated with an increased hazard ratio (HR) for mortality (HR, 5.06 [95% CI, 3.18–8.18], P value 0.001). The 10‐year survival rate in patients without SSc‐pHI was 94.9% [95% CI, 92.1%–97.8%], whereas the 10‐year survival rate of patients with SSc‐pHI was significantly lower (72.5% [95% CI, 69.4%–75.9%]). The high mortality was confirmed in 273 patients from the further phenotyped EUSTAR subcohort including 109 patients with and 164 patients without SSc‐pHI (HR, 3.46 [95% CI, 1.70–7.06], P alue 0.001). After propensity weighting, patients with SSc‐pHI had a lower survival rate compared with patients without SSc‐pHI in both cohorts (P value <0.0001) (Figure 3). The lowest mortality rate was observed in patients with SSc without SSc‐pHI and SSc‐ILD (n=621 patients), and the highest mortality rate was observed in patients with both SSc‐pHI and SSc‐ILD (n=1122 patients). Patients with SSc‐pHI and without ILD (n=1237 patients) had a lower survival rate compared with patients with ILD and without SSc‐pHI (n=385 patients) in the EUSTAR cohort (Figure 4).

Figure 3. Survival analyses in patients with and without SSc‐pHI in a EUSTAR subcohort and the EUSTAR database.

Kaplan–Meier curves illustrating the survival probabilities of patients with SSc‐pHI (blue curve) and without SSc‐pHI (red curve) in the whole EUSTAR cohort (A) and a EUSTAR sub‐cohort (B). The shaded areas represent the 95% CIs. EUSTAR indicates European Scleroderma Trials and Research; and SSc‐pHI, systemic sclerosis primary heart involvement.

Figure 4. Survival analysis in patients with and without ILD or SSc‐pHI in the EUSTAR cohort.

Kaplan–Meier curves illustrating the survival probabilities of patients without ILD and SSc‐pHI (red curve), patients with SSc‐pHI and without ILD (green curve), patients without SSc‐pHI and with ILD (blue curve), and patients with both ILD and SSc‐pHI (violet curve) in the EUSTAR cohort. Patients meeting the exclusion criteria of the study were not included. The shaded areas represent the 95% CIs. EUSTAR indicates European Scleroderma Trials and Research; ILD, interstitial lung disease; and SSc‐pHI, systemic sclerosis primary heart involvement.

Swollen Joints as Risk Factor for Progression of SSc‐pHI

Of the 1531 patients with SSc‐pHI from the whole EUSTAR cohort, 414 had experienced progression of their SSc‐pHI during the follow‐up period according to the study definition. Of these, 293 patients (70.8%) experienced a decrease in their LVEF of at least 5%, 110 patients (26.6%) progressed from LV diastolic dysfunction to systolic dysfunction with an LVEF below 50%, 111 patients (26.8%) had a new onset of arrhythmia or a conduction disorder, and 4 patients (1%) received a pacemaker/implantable cardioverter defibrillator at any time during their follow‐up. The presence of swollen joints increased the risk of progression of SSc‐pHI (OR, 2.49 [95% CI, 1.79–3.52]), whereas digital ulcers, telangiectasia, intestinal symptoms, skeletal muscle weakness, skeletal muscle atrophy, tendon friction rubs, antitopoisomerase I antibody positivity, age, and male sex did not increase the risk for progressive SSc‐pHI (Figure S1). This effect was still significant when adjusting for CRP (C‐reactive protein) and rheumatoid factor positivity.

Of note, a different cutoff for the decline of the LVEF than the one recommended by the task force for cancer treatments and cardiovascular toxicity of the European Society of Cardiology for cancer treatment associated cardiovascular cytotoxicity was chosen for the definition of progressive SSc‐pHI. ³⁸ Consistent with the clinical practice only 10.2% (n=156) of the patients with SSc‐pHI from the EUSTAR cohort experienced a decline in their LVEF of at least 10% between 2 consecutive echocardiographies, performed at a mean follow‐up period of 1.25 years. However, declines in the LVEF of <10% significantly affected all‐cause mortality. A 5% decline increased the risk of mortality by 16%. A decline in the LVEF of 10% increased the risk of mortality by 34% compared with patients with SSc‐pHI with stable LVEF. The cutoff of 5% for the decline in the LVEF thus was chosen as a tradeoff between the common (inter‐ and intraobserver) variability of LVEF measurements in nonprogressive patients and high sensitivity to detect progression of SSc‐pHI that affects mortality. ³⁹ A risk factors analysis in which a 10% cutoff for the LVEF instead of 5% was chosen showed again that swollen joints were the most relevant risk factor for the progression of SSc‐pHI (OR, 1.74 [95% CI, 1.28–2.67]).

The frequencies of vasoactive and immunomodulatory therapies were similar between patients with and without progressive SSc‐pHI (Table 3). The clinical characteristics of patients with and without progression of SSc‐pHI are illustrated in Table 3.

Table 3.

Clinical Characteristics of Patients With and Without Progressive SSc‐pHI in the EUSTAR Cohort

	Without progressive SSc‐pHI	With progressive SSc‐pHI
No.	181	414
Age, y (mean±SD)	57.64 (10.65)	55.04 (12.04)
Male sex (%)	28 (15.5)	64 (15.5)
Body mass index, kg/m2 (mean±SD)	24.10 (4.03)	23.94 (4.05)
Year of appearance of first non‐Raynaud's phenomenon manifestation, mean±SD	2006.72 (10.18)	2004.85 (8.45)
Anticentromere antibodies positivity (%)	78 (43.6)	147 (35.9)
Antitopoisomerase I antibodies positivity (%)	60 (33.1)	154 (37.2)
Anti‐RNA‐polymerase III antibodies positivity (%)	8 (6.2)	13 (4.5)
Limited cutaneous systemic sclerosis (%)	105 (72.4)	223 (67.2)
New York Heart Association classes
I	111 (62.4)	239 (59.0)
II	51 (28.7)	139 (34.3)
III	16 (9.0)	27 (6.7)
IV	0 (0.0)	0 (0.0)
Cardiac arrhythmias (%)	16 (11.1)	21 (10.7)
Conduction blocks (%)	36 (20.9)	71 (18.6)
Diastolic dysfunction (%)	84 (50.0)	145 (38.2)
Perimyocarditis (%)	1 (12.5)	3 (30.0)
LVEF (mean±SD)	60.62 (5.83)	63.80 (6.07)
Systolic dysfunction (LFEV <50%) (%)	4 (2.2)	7 (1.8)
HF with preserved EF (%)	37 (21.1)	122 (31.4)
HF with mildly reduced EF (%)	1 (0.6)	3 (0.7)
HF with reduced EF (%)	0 (0.0)	3 (0.7)
ILD (%)	79 (46.7)	199 (50.1)
Extensive ILD (at least 20% fibrotic involvement in high‐resolution computed tomography or in indeterminate cases by an FVC <70% according to a previous definition, 37 %)	26 (51.0)	59 (67.8)
FVC percentage predicted (mean±SD)	98.55 (21.51)	98.14 (22.60)
Single breath method of measuring the diffusing lung capacity for carbon monoxide percentage predicted (mean±SD)	71.51 (20.33)	70.44 (18.63)
Tendon friction rubs (%)	9 (5.0)	24 (5.8)
Subcutaneous calcinosis (%)	11 (12.4)	9 (12.0)
Swollen joints (%)	103 (56.9)	345 (83.3)
Rheumatoid factor positivity (%)	12 (19.7)	6 (12.0)
Skeletal muscle weakness (%)	26 (14.4)	55 (13.3)
Skeletal muscle atrophy (%)	6 (3.3)	15 (3.6)
Telangiectasia (%)	121 (66.9)	271 (65.5)
Digital ulcers (%)
Current	22 (12.2)	44 (10.6)
Never	105 (58.0)	229 (55.3)
Previously	54 (29.8)	141 (34.1)
Ramified bushy capillaries (%)
Extensive	5 (11.9)	5 (17.9)
Moderate	16 (38.1)	9 (32.1)
None	16 (38.1)	7 (25.0)
Rare	5 (11.9)	7 (25.0)
Capillary loss (%)
Extensive	13 (26.0)	14 (29.2)
Moderate	14 (28.0)	19 (39.6)
None	14 (28.0)	8 (16.7)
Rare	9 (18.0)	7 (14.6)
Hemorrhages (%)
Extensive	7 (15.2)	3 (8.6)
Moderate	13 (28.3)	11 (31.4)
None	13 (28.3)	10 (28.6)
Rare	13 (28.3)	11 (31.4)
Giant capillaries (%)
Extensive	6 (13.0)	10 (24.4)
Moderate	14 (30.4)	18 (43.9)
None	3 (6.5)	6 (14.6)
Rare	23 (50.0)	7 (17.1)
Intestinal symptoms (%)	53 (29.3)	118 (28.5)
EUSTAR disease activity index (mean±SD)	0.47 (1.12)	0.63 (1.26)
Health assessment questionnaire score (mean±SD)	0.57 (0.51)	0.74 (0.60)
Erythrocyte sedimentation rate (mean±SD)	18.47 (15.87)	19.68 (16.07)
C‐reactive protein elevation (%)	21 (24.7)	56 (17.8)
Patient symptomatic for cardiac disease (%)	3 (3.8)	12 (5.9)
Troponin T elevation (%)	2 (25.0)	5 (50.0)
N‐terminal pro‐B‐type natriuretic peptide, pg/mL (mean±SD)	124.65 (116.82)	296.55 (670.92)
Creatine kinase elevation (%)	8 (9.9)	19 (6.1)
Hypocomplementemia (%)	10 (12.7)	24 (8.7)
Cigarette smoking ever (%)	23 (16.9)	57 (21.0)
Current therapy
Calcium channel blockers (%)	29 (16.0)	49 (11.8)
Angiotensin‐converting enzyme inhibitors/angiotensin receptor blockers (%)	22 (12.2)	44 (10.6)
Antiplatelet drugs (%)	5 (2.8)	11 (2.7)
Prostanoids (%)	0 (0.0)	5 (1.2)
Phosphodiesterase‐5 inhibitors (%)	29 (3.3)	7 (5.9)
Endothelin receptor antagonist (%)	0.0 (0.0)	0.0 (0.0)
Glucocorticoids (%)	58 (32.0)	101 (24.4)
Nonsteroidal anti‐inflammatory drugs (%)	4 (2.2)	7 (1.7)
Proton pump inhibitors (%)	2 (1.1)	12 (2.9)
Conventional synthetic disease modifying drugs (including hydroxychloroquine, methotrexate, leflunomide, and sulfasalazine) (%)	1 (0.6)	3 (0.7)
Mycophenolate mofetil/mycophenolic acid (%)	2 (1.1)	4 (1.0)
Cyclophosphamide (%)	11 (6.1)	11 (2.7)
Tocilizumab (%)	4 (2.2)	6 (1.4)
Rituximab (%)	10 (5.5)	14 (3.4)
Biologicals (tocilizumab, rituximab, tumor necrosis factor alpha inhibitors, abatacept, and Janus kinase inhibitors, %)	15 (8.3)	26 (6.3)

EF indicates ejection fraction; EUSTAR, European Scleroderma Trials and Research; FVC, forced vital capacity; HF, heart failure; ILD, interstitial lung disease; LVEF, left ventricular ejection fraction; and SSc‐pHI, systemic sclerosis‐associated primary heart involvement.

We also observed a higher frequency of swollen joints (86.7%) in patients with progressive SSc‐pHI compared with patients without progressive SSc‐pHI (65.4%) in the EUSTAR subcohort (n=82).

Chronic Heart Failure in the EUSTAR Cohort

One‐third (36.3%) of the patients with SSc‐pHI (= 15.7% of the patients with SSc with or without SSc‐pHI) fulfilled the European Society of Cardiology diagnostic criteria for HFpEF. ⁴⁰ HF with reduced EF was present in 1.5% of the patients with SSc‐pHI (= 0.67% of the patients with SSc with or without SSc‐pHI). ²² HF with mildly reduced EF was present in 1.5% of the patients with SSc‐pHI (= 0.64% of the patients with SSc with or without SSc‐pHI). ⁴ Consistent with the prevalence of HFpEF in the general population, the prevalence of HFpEF was slightly higher in female patients. The mean NT‐proBNP (N‐terminal pro‐B‐type natriuretic peptide) values were pathological in both patients with SSc‐pHI with and without HFpEF; however, the mean value was 6 times higher in patients with HFpEF compared with patients without. Patients with HFpEF had more frequently limited cutaneous involvement or anticentromere antibodies compared with patients without HFpEF. ILD was present in similar proportions between patients with and without HFpEF. However, patients with SSc‐pHI with HFpEF had more frequently extensive ILD, defined by at least 20% fibrotic involvement in high resolution computed tomography, or in indeterminate cases by a forced vital capacity <70%. ³⁷ However, mean forced vital capacity and diffusing lung capacity for carbon monoxide values were similar between the 2 groups. Cardiac arrhythmias and conduction defects were less frequent in patients with SSc‐pHI with HFpEF. The clinical characteristics of the patients with SSc‐pHI with and without HFpEF are illustrated in Table 4.

Table 4.

Clinical Characteristics of Patients With SSc‐pHI With and Without HFpEF in the EUSTAR Cohort

	SSc‐pHI no HFpEF	SSc‐pHI‐ HFpEF
No.	1017	514
Age, y (mean±SD)	56.00 (14.05)	57.83 (12.08)
Male sex (%)	188 (18.5)	77 (15.0)
Body mass index, kg/m2 (mean±SD)	23.67 (4.08)	25.26 (4.62)
Year of appearance of first non‐Raynaud's phenomenon manifestation (mean±SD)	2008.70 (9.14)	2006.59 (8.55)
Anticentromere antibodies positivity (%)	351 (35.8)	217 (42.7)
Antitopoisomerase I antibodies positivity (%)	408 (40.1)	158 (30.7)
Anti‐RNA‐polymerase III antibodies positivity (%)	45 (6.4)	17 (5.3)
Limited cutaneous SSc (%)	531 (61.5)	291 (72.6)
New York Heart Association stages
I	547 (58.8)	308 (59.9)
II	314 (33.8)	179 (34.8)
III	63 (6.8)	26 (5.1)
IV	6 (0.6)	1 (0.2)
Cardiac arrhythmias (%)	137 (20.8)	6 (4.7)
Conduction blocks (%)	493 (52.6)	52 (11.4)
Diastolic dysfunction (%)	420 (47.5)	514 (100)
Perimyocarditis (%)	7 (26.9)	3 (30.0)
LVEF (mean±SD)	59.72 (8.65)	63.25 (5.57)
Systolic dysfunction (LVEF < 50%) (%)	73 (8.8)	0 (0.0)
ILD (%)	134 (45.4)	158 (43.5)
FVC percentage predicted (mean±SD)	93.74 (23.04)	98.74 (23.00)
Single breath method of measuring the diffusing lung capacity for carbon monoxide percentage predicted (mean±SD)	69.17 (20.75)	72.48 (19.49)
Extensive ILD (at least 20% fibrotic involvement in high‐resolution computed tomography or in indeterminate cases by an FVC <70% according to a previous definition, 37 %)	176 (64.5)	66 (76.7)
Tendon friction rubs (%)	59 (5.8)	32 (6.2)
Subcutaneous calcinosis (%)	65 (11.7)	0 (0.0)
Swollen joints (%)	508 (50.0)	514 (100.0)
Rheumatoid factor positivity (%)	53 (13.8)	3 (33.3)
Skeletal muscle weakness (%)	156 (15.3)	63 (12.3)
Skeletal muscle atrophy (%)	65 (6.4)	21 (4.1)
Telangiectasia (%)	604 (59.4)	294 (57.2)
Digital ulcers (%)
Current	159 (15.6)	49 (9.5)
Never	552 (54.3)	354 (68.9)
Previously	306 (30.1)	111 (21.6)
Ramified bushy capillaries (%)
Extensive	19 (8.8)	2 (28.6)
Moderate	47 (21.9)	5 (71.4)
None	91 42.3n	0 (0.0)
Rare	58 (27.0)	0 (0.0)
Capillary loss (%)
Extensive	49 (20.3)	5 (38.5)
Moderate	57 (23.7)	8 (61.5)
None	78 (32.4)	0 (0.0)
Rare	57 (23.7)	0 (0.0)
Hemorrhages (%)
Extensive	21 (9.3)	2 (25.0)
Moderate	59 (26.1)	6 (75.0)
None	67 (29.6)	0 (0.0)
Rare	79 (35.0)	0 (0.0)
Giant capillaries (%)
Extensive	38 (15.9)	4 (30.8)
Moderate	79 (33.1)	9 (69.2)
None	23 (9.6)	0 (0.0)
Rare	99 (41.4)	0 (0.0)
Intestinal symptoms	245 (24.1)	132 (25.7)
Malabsorption syndrome (%)	6 (2.2)	6 (2.1)
EUSTAR disease activity index (mean±SD)	0.43 (1.20)	0.80 (1.43)
Health assessment questionnaire score (mean±SD)	0.74 (0.69)	NaN (NA)
Erythrocyte sedimentation rate (mean±SD)	21.17 (18.86)	20.01 (16.72)
C‐reactive protein elevation (%)	101 (24.0)	104 (21.4)
Patients symptomatic for cardiac disease (%)	31 (11.6)	21 (7.3)
Troponin T elevation (%)	8 (57.1)	3 (42.9)
N‐terminal pro‐B‐type natriuretic peptide, pg/mL (mean±SD)	306.70 (813.25)	1861.79 (10773.93)
Creatine kinase elevation (%)	43 (7.6)	98 (11.2)
Hypocomplementemia (%)	49 (12.0)	49 (10.4)
Current therapy
Calcium channel blockers (%)	342 (33.6)	158 (30.7)
Angiotensin‐converting enzyme inhibitors/angiotensin receptor blockers (%)	78 (7.7)	43 (8.4)
Antiplatelet drugs (%)	241 (23.7)	133 (25.9)
Prostanoids (%)	196 (11.2)	178 (11.6)
Phosphodiesterase‐5 inhibitors (%)	137 (13.5)	41 (8.0)
Endothelin receptor antagonist (%)	84 (8.3)	17 (3.3)
Diuretics (%)	61 (6.0)	26 (5.1)
Glucocorticoids (%)	210 (20.6)	121 (23.5)
Nonsteroidal anti‐inflammatory drugs (%)	51 (5.0)	35 (6.8)
Proton pump inhibitors (%)	461 (45.3)	225 (43.8)
Conventional synthetic disease modifying drugs (including, hydroxychloroquine, methotrexate, leflunomide, and sulfasalazine) (%)	113 (11.1)	89 (17.3)
Mycophenolate mofetil/mycophenolic acid (%)	127 (12.5)	31 (6.0)
Cyclophosphamide (%)	21 (2.1)	14 (2.7)
Tocilizumab (%)	20 (2.0)	1 (0.2)
Rituximab (%)	31 (3.0)	7 (1.4)
Biologicals (tocilizumab, rituximab, tumor necrosis factor alpha inhibitors, abatacept, and Janus kinase inhibitors, %)	56 (5.5)	8 (1.6)

EUSTAR indicates European Scleroderma Trials and Research; FVC, forced vital capacity; HFpEF, heart failure with preserved ejection fraction; ILD, interstitial lung disease; and SSc‐pHI, systemic sclerosis‐associated primary heart involvement.

Tendon friction rubs (OR, 1. 21 [95% CI, 0.94–1.90]) increased the risk for HFpEF in patients with SSc‐pHI, whereas digital ulcers decreased the risk for HFpEF (OR, 0.61 [95% CI, 0.49–0.75]).

Survival analyses included 831 patients with SSc‐pHI and HFpEF, 1723 patients with SSc‐pHI and without HFpEF, and 1214 patients without SSc‐pHI. Interestingly, the survival rate was similar between patients with SSc‐pHI with and without HFpEF after adjusting for age, sex, body mass index, subset of disease, LVEF, ILD, and extent of ILD (Figure 5). None of the patients with HFpEF progressed toward HF with reduced EF during the follow‐up period.

Figure 5. Survival analyses in patients with SSc‐pHI with and without HFpEF and patients with SSc without SSc‐pHI in the EUSTAR cohort.

Kaplan–Meier curves illustrating the survival probabilities of patients with SSc‐pHI with HFpEF (red curve) and without HFpEF (blue curve) as well as patients with SSc without SScpHI (green curve) in the whole EUSTAR cohort. EUSTAR indicates European Scleroderma Trials and Research; HFpEF, heart failure with preserved ejection fraction; SSc, systemic sclerosis; and SSc‐pHI, systemic sclerosis primary heart involvement.

DISCUSSION

Our study provides insights into risk factors for new onset and progression of SSc‐pHI, one of the leading causes of mortality in patients with SSc. This is, to our knowledge, the first study that evaluates the risk factors and outcomes of SSc‐pHI as recently defined by the WSF/HFA using the worldwide largest multicenter, multinational cohort of patients with SSc. ⁷ Furthermore, our study offers a first, preliminary multidisciplinary definition for the progression of SSc‐pHI and identifies risk factors associated with the progression of SSc‐pHI. In addition, we confirmed our results in a EUSTAR subcohort with in‐depth cardiac phenotyping, including CMR (T1 and T2 mapping, extracellular volume, late gadolinium enhancement) and myocardial histopathological data, as well as several echocardiographic parameters such as the GLS, which are not collected in the EUSTAR database.

We found that skeletal muscle atrophy, age, male sex, swollen joints, skeletal muscle weakness, and tendon friction rubs are risk factors for SSc‐pHI, confirming previous findings in a larger, multicenter cohort. Consistent with the general population LV systolic dysfunction was more frequently present in patients of male sex. ⁴¹ , ⁴² , ⁴³ , ⁴⁴ A previous EUSTAR study showed that male sex and the presence of myositis are risk factors for LV systolic dysfunction defined by an LVEF <55%. ¹⁶ An association between myositis‐specific or myositis‐associated antibodies and SSc‐pHI could not be evaluated as testing for these antibodies was not available for most patients. A previous EUSTAR analysis showed that an older age at the diagnosis of SSc is associated with an increased risk of cardiac conduction blocks and LV diastolic dysfunction. ¹² Previous studies have reported an association between tendon friction rubs and cardiac involvement. ⁴⁵ , ⁴⁶ , ⁴⁷ Even though tendon friction rubs are mostly present in patients with diffuse cutaneous SSc and as such in patients who are antitopoisomerase I positive, diffuse cutaneous involvement or the presence of antitopoisomerase I antibodies was not associated with an increased risk for SSc‐pHI in our study. Skeletal muscle weakness, male sex, and tendon friction rubs were further confirmed as risk factors for the presence of SSc‐pHI in a EUSTAR subcohort with extensive cardiac phenotyping. Consistent with the postulated microvascular pathogenesis of SSc‐pHI, we found that telangiectasia increased the risk for new onset of SSc‐pHI in the EUSTAR cohort. ¹⁶ Furthermore, we also observed the presence of intestinal symptoms such as diarrhea, bloating, or constipation to be associated with an increased risk for the development of SSc‐pHI. This effect might be due to malabsorption or intestinal dysbiosis associated with intestinal involvement in SSc. ⁴⁸ Our hypothesis is supported by a recent study which demonstrated that the plasma levels of trimethylamine N‐oxide, a bacterial metabolite produced by dysbiotic microbiota significantly correlated with the LVEF and the NT‐proBNP levels in patients with SSc. ⁴⁹

We also evaluated the survival rates of patients with and without SSc‐pHI in the EUSTAR cohort. Consistent with previous data, we found a significantly lower survival rate of patients with clinically manifest SSc‐pHI. ² , ⁵ , ⁶ Of note, asymptomatic SSc‐pHI was also associated with increased mortality compared with patients without SSc‐pHI, when adjusting for confounders such as age, sex, body mass index, and the presence and extent of SSc‐ILD. ² , ⁵⁰ We further confirmed these results in a EUSTAR subcohort with extensive cardiac characterization. Furthermore, when comparing survival across patients with and without SSc‐pHI and with and without ILD in the EUSTAR cohort, patients with SSc‐pHI and without ILD had a lower survival rate compared with patients with ILD and without SSc‐pHI.

Because no definition of progressive SSc‐pHI was available, we assembled a team of SSc experts and cardiologists to propose a consensus‐definition of progressive SSc‐pHI. Using this definition, we evaluated risk factors for progressive SSc‐pHI in the EUSTAR database. No risk factors for the progression of SSc‐pHI have been described so far. We found that swollen joints were associated with an increased risk for progression of SSc‐pHI, independent of the CRP levels or rheumatoid factor positivity. Previous studies have implicated IL‐1 (interleukin‐1) in the pathogenesis of SSc‐pHI. ⁴⁰ However, whether other arthritogenic cytokines or immune cell subsets might also play a role in the progression of SSc‐pHI is yet unknown. Our data might warrant more frequent evaluations in patients with SSc‐pHI with arthritis. The risk factors for the progression of SSc‐pHI should be considered preliminary as further studies are needed to refine and evaluate the clinical significance of our proposed definition for progressive SSc‐pHI.

The EUSTAR cohort allowed the unique opportunity to evaluate HFpEF due to SSc‐pHI according to the most recent European Society of Cardiology guidelines in a large number of patients with SSc lacking most of the standard cardiovascular risk factors including coronary artery disease, diabetes, arterial hypertension, and significant chronic kidney disease. In contrast with previous results from the Australian Scleroderma Cohort Study (n=225 patients), in the EUSTAR cohort HFpEF was rather common and present in one third of patients with SSc‐pHI and 15.7% of patients with or without SSc‐pHI. Due to our study design, which involved exclusion of patients with confirmed pulmonary hypertension or systolic pulmonary artery pressure >40 mm Hg, as well as most of the classical cardiovascular risk factors, HFpEF in patients with SSc might be even more frequent. However, one must consider that dyspnea in SSc is rather multifactorial and in our cohort, it could be at least partly also due to SSc‐ILD, pleural effusion, pericardial effusion, deconditioning, or involvement of respiratory muscles. Indeed, the presence of extensive ILD was higher in patients with SSc‐pHI with HFpEF compared with patients with SSc‐pHI without HFpEF. Consistent with previous reports, HF with reduced EF and HF with mildly reduced EF were rare in patients with SSc‐pHI in the EUSTAR cohort. ⁵

To our knowledge this is the first study to identify risk factors for HFpEF in patients with SSc‐pHI according to the most recent definitions of HFpEF and SSc‐pHI, respectively. We identified tendon friction rubs as risk factor for HFpEF in patients with SSc‐pHI. Interestingly, the survival rate of patients with SSc‐pHI with HFpEF was similar to that of patients without HFpEF after adjusting for confounders. Diastolic dysfunction has been associated with an increased mortality in patients with SSc; however, even though diastolic dysfunction has been found twice as frequently in patients with SSc‐pHI with HFpEF compared with patients without HFpEF, it did not contribute to an increased overall mortality in patients with HFpEF. ¹⁹ , ⁵¹

Our study has limitations. Although the EUSTAR registry is the largest registry for SSc worldwide, our conclusions would have been strengthened by revalidation of our findings in an independent external validation cohort. Moreover, there are currently no established criteria for the definition of SSc‐pHI and the definition of progressive SSc‐pHI. The criteria used in our study are thus based on interdisciplinary expert consensus but are not validated to date. These limitations need to be kept in mind when interpreting the results of our study. In consequence, follow‐up studies are required to confirm the findings presented herein. Furthermore, in clinical practice, SSc‐pHI frequently coincides with other SSc organ involvements such as SSc‐PAH or scleroderma renal crisis, as well as with other comorbidities such as arterial hypertension, diabetes, or chronic kidney disease. To strictly differentiate SSc‐pHI from secondary heart involvement, patients with SSc‐PAH, scleroderma renal crisis, arterial hypertension, diabetes, or chronic kidney disease of at least third stage were excluded from our analyses. Hyperlipoproteinemia and family history of cardiac disease as established cardiovascular risk factors could not be evaluated, as they are not routinely collected in the EUSTAR database. Even though valvular disease could be a consequence of SSc, patients with valvulopathies of at least middle grade severity were excluded from our analyses due to the lack of tools to differentiate between valvulopathies in the context of SSc‐pHI and congenital, degenerative, or postinfectious valvular disease. Even though pericardial effusion can also be a manifestation of SSc‐pHI, we did not include pericardial effusion in our analyses due to its limited discriminative capacity for SSc‐pHI over secondary heart involvement and other possible causes. Furthermore, cardiovascular death and hospitalization due to HF could not be evaluated in our cohort because they are not routinely collected in the EUSTAR database.

Conclusions

In summary, we identified clinical and demographic risk factors for new onset of SSc‐pHI in the largest, multinational cohort with SSc and showed that patients with SSc‐pHI, even if asymptomatic, have a lower survival rate. Furthermore, the survival rate of patients with SSc‐pHI and without ILD was even lower than that of patients with ILD and without SSc‐pHI. Moreover, we propose a first, preliminary definition of progressive SSc‐pHI and identify risk factors associated with the progression of SSc‐pHI. Our results could help to stratify patients with SSc according to the risk of development or progression of SSc‐pHI to guide diagnostic efforts and treatment initiation; however, these results need further external validation in independent cohorts. Furthermore, we show that HFpEF is frequent in patients with SSc. Due to the multifactorial origin of dyspnea in SSc, HFpEF can easily remain unidentified in patients with SSc. We provide a rationale for increased awareness for HFpEF in patients with SSc to initiate confirmatory diagnostics in a subset of patients and proper treatment including SGLT2 inhibitors.

Sources of Funding

This work was supported by the German Research Foundation (DFG) [DI 1537/20–1, DI 1537/22–1, DI 1537/23–1, SFB CRC1181 (project C01) and SFB TR221 (project number 324392634 (B04)) to J.H.W. Distler], the Wilhelm‐Sander‐Foundation [2013.056.1 to J.H.W. Distler], the German Federal Ministry of Education and Research (BMBF) [MASCARA program, TP 2 (01EC1903A)] and a Career Support Award of Medicine of the Ernst Jung Foundation to J.H.W. Distler. Further support was provided by German Research Foundation [MA 9219/2–1 to A.‐E. Matei], the Else‐Kröner‐Fresenius‐Foundation [2021_EKEA.03 and 2022_EKMS.02 to A.‐E. Matei] and a grant of the research committee of the Medical Faculty of Heinrich‐Heine University of Düsseldorf to A.‐H. Györfi. L.A. Saketkoo has received funding from aTyr, EMD Serono, Horizon, Kinevant, Priovant, Pfizer, and the United States Department of Defense.

Disclosures

Dr Smith is senior clinical investigator of the Research Foundation—Flanders (Belgium) (FWO) [1.8.029.20N]. Dr Saketkoo is consultant or part of the advisory board for/of Argenx, aTyr, Boehringer Ingelheim, Kinevant, Mallinckrodt, Novartis, Scleroderma Foundation, Scleroderma Foundation of Chicago, Steffens Foundation. The remaining authors have no disclosures to report.

Supporting information

Data S1

Tables S1–S4

Figures S1–S2