Abstract
Objective To explore the effect of left ventricular (LV) diastolic dysfunction (LVDD) in systemic sclerosis (SSc)-associated interstitial lung disease (ILD), and to investigate SSc-specific associations and clinical correlates of LVDD.
Methods There were 102 Australian Scleroderma Cohort Study participants with definite SSc and radiographic ILD included. Diastolic function was classified as normal, indeterminate, or abnormal according to 2016 American Society of Echocardiography/European Association of Cardiovascular Imaging guidelines for assessment of LV diastolic function. Associations between clinical features and patient- and physician-reported dyspnea were evaluated using logistic regression. Survival analyses were performed using Kaplan-Meier survival estimates and Cox regression modeling.
Results LVDD was identified in 26% of participants, whereas 19% had indeterminate and 55% had normal diastolic function. Those with ILD and LVDD had increased mortality (hazard ratio 2.4, 95% CI 1.0-5.7; P = 0.05). After adjusting for age and sex, those with ILD and LVDD were more likely to have severe dyspnea on the Borg Dyspnoea Scale (odds ratio [OR] 2.6, 95% CI 1.0-6.6; P = 0.05) and numerically more likely to record World Health Organization Function Class II or higher dyspnea (OR 4.2, 95% CI 0.9-20.0; P = 0.08). Older age (95% CI 1.0-6.4; P = 0.05), hypertension (OR 5.0, 95% CI 1.8-13.8; P < 0.01), and ischemic heart disease (OR 4.8, 95% CI 1.5-15.7; P < 0.01) were all associated with LVDD, as was proximal muscle atrophy (OR 5.0, 95% CI 1.9-13.6; P < 0.01) and multimorbidity (Charlson Comorbidity Index scores ≥ 4, OR 3.0, 95% CI 1.1-8.7; P = 0.04).
Conclusion LVDD in SSc-ILD is more strongly associated with traditional LVDD risk factors than SSc-specific factors. LVDD is associated with worse dyspnea and survival in those with SSc-ILD.
Systemic sclerosis (SSc) heart involvement (SHI) is an underrecognized complication of SSc despite being a major contributor to mortality.1 Diffuse myocardial fibrosis is the histopathological hallmark of SHI2 and may contribute to excess frequency of diastolic dysfunction in individuals with SSc.3 Left ventricular (LV) diastolic dysfunction (LVDD) has been reported in 20% to 30% of those with SSc,4-6 including exercise-induced LVDD where individuals with normal resting hemodynamics may demonstrate abnormal diastology on exertion.5,7 Exercise intolerance is a cardinal symptom of LVDD,8 and LVDD has been associated with reduced exercise capacity in SSc.9 LVDD in SSc has also been associated with increased mortality.4,6,10
The precise pathophysiology of LVDD in SSc is unknown. Myocardial fibrosis predisposing to LVDD may be associated with recurrent myocardial ischemia and reperfusion injury because of abnormal small-vessel vasoreactivity.11 Risk of LVDD is increased in those with fibrotic SSc manifestations including interstitial lung disease (SSc-ILD),12,13 as well as being associated with traditional LVDD risk factors including older age, smoking, and hypertension (HTN).13
Identifying LVDD in those with ILD may be challenging, because of the overlapping symptoms of dyspnea, exercise intolerance, and reduced physical function. However, failure to recognize LVDD as contributing to an individual’s symptoms may mean progressive physical disability is erroneously attributed to ILD, and opportunities to optimize management may be missed.
Given we have previously identified an association between the presence of ILD and LVDD in our cohort,12 we aimed to investigate the clinical correlates, associations, and prognosis of LVDD in SSc-ILD specifically, and quantify the contribution of LVDD to breathlessness in people with SSc-ILD.
METHODS
Participants enrolled in the prospective Australian Scleroderma Cohort Study (ASCS) at St. Vincent’s Hospital, Melbourne (SVHM), between 2007 and October 2022 were eligible for inclusion. Individuals diagnosed with radiographic ILD and meeting American College of Rheumatology/European Alliance of Associations for Rheumatology criteria14 for SSc were included. Participants were required to have an echocardiogram where at least 3 variables of diastolic function could be assessed in accordance with the American Society of Echocardiography 2016 Assessment of LVDD Guidelines.15 Participants with pulmonary arterial hypertension (PAH) diagnosed by right heart catheter studies according to either revised16 or previous classification criteria were excluded. Participants with significant mitral valve disease or those experiencing atrial fibrillation on the echocardiogram used to define diastolic function were also excluded. Ethics approval for the ASCS was provided by SVHM Human Research Ethics Committee (HREC-A 020/07). Written informed consent was obtained from all participants.
All participants underwent annual transthoracic echocardiography (TTE) and pulmonary function tests (PFTs), including percent predicted forced vital capacity (FVC) and diffusing lung capacity for carbon monoxide (DLCO; corrected for hemoglobin). ILD was diagnosed using high-resolution computed tomography (HRCT) of the chest performed at physician discretion in response to the presence of clinical features of ILD or abnormal PFTs. Severity of ILD was defined by the extent of radiological involvement as either limited (< 20% HRCT involvement) or extensive (> 30% HRCT involvement).17 Where HRCT extent was 20% to 30%, percent predicted FVC of < 70% was used to classify patients as extensive ILD, or FVC ≥ 70% as limited ILD.
TTE. All TTEs were performed at SVHM using GE Vingmed Vivid E9 Ultrasound (GE, Vingmed Ultrasound) and software analyses using Echopac PC v204 analysis software (GE). Standard parasternal, apical, and subcostal windows were used to image cardiac chambers, as well as color, pulsed, and continuous wave Doppler assessment. The modified biplane Simpson method was used to calculate LV ejection fraction (LVEF), with normal systolic function defined as LVEF ≥ 50%. LV hypertrophy (LVH) was determined using linear 2-D measurements using the cube formula.18 Left atrial (LA) volume index (LAVI) was calculated as the average of the volumetric measurement of the LA indexed to body surface area (BSA; m2). Where LAVI was not available, LA enlargement was defined as LA area indexed to BSA > 11.8 cm2/m2.19 E/A ratio was calculated to evaluate diastolic function using pulsed-waved Doppler imaging of mitral valve inflow, measuring peak early diastolic velocity (E) and peak late diastolic velocity (A). Tissue Doppler imaging was used to measure septal and/or lateral e′ velocity (peak early diastolic velocity at mitral anulus) and calculate average E/e′ ratio (ratio of early diastolic mitral inflow velocity to early diastolic mitral annulus velocity). Tricuspid regurgitant (TR) velocity was calculated by continuous wave Doppler and used to measure right ventricular systolic pressure (RVSP) using the modified Bernoulli equation and adding an estimated right atrial pressure.
LVDD was defined according to the 2016 Assessment of LVDD Guidelines15 using 4 criteria: LAVI > 34 mL/m2, average E/e′ ratio > 14 (if unavailable, septal E/e′ > 16 or lateral E/e′ > 13), annular velocities (septal e′ < 7 cm/s or lateral e′ < 10 cm/s), and peak TR velocity > 2.8 m/s. At least 3 criteria needed to be available for analysis to establish either normal or abnormal diastolic function. LVDD was considered present when > 50% of criteria were abnormal, indeterminate diastolic function if 50% of available diastolic function variables were abnormal, and normal diastolic function if < 50% were abnormal.15 LVDD was graded according to 2016 recommendations15 as grade I, II, or III using E/A ratio and the number of abnormal diastolic variables, with grade considered undetermined if insufficient variables were available for analysis. We defined those with definite LVDD as having ILD-LVDD, and those without definite LVDD (ie, with normal or indeterminate diastolic function) as having ILD only.
Data collection. Demographic and disease data were collected annually. Disease onset/duration were defined from the first non-Raynaud SSc manifestation. Disease features were considered present if they were ever reported from SSc diagnosis. LeRoy criteria were used to determine cutaneous SSc subtype (diffuse cutaneous [dcSSc] or limited cutaneous).20 Current medications, the presence of comorbid angina, diabetes, dyslipidemia, HTN, and smoking were recorded from patient-reported history and medical record review at each study visit. Ischemic heart disease (IHD) was defined by the presence of abnormal coronary angiography or physician-reported ischemic chest pain and/or patient-reported IHD. BMI was calculated as weight (kg) divided by height (m2) at each study visit. Scleroderma renal crisis was diagnosed in the concurrent presence of 2 of either of the following: new-onset HTN without alternate cause, unexplained rise in serum creatinine, or microangiopathic hemolytic anemia. Medsger severity scores (MSSs) were calculated21 at each visit to assess the overall burden of SSc, including organ-specific and composite scores.22 To measure multimorbidity, we calculated a modified Charlson Comorbidity Index (CCI) score23 (included items provided in Supplementary Table S1; available from the authors upon request). Hemiplegia, HIV/AIDS, and dementia were excluded from the calculation of the CCI as these data are not collected in the ASCS. The maximum possible CCI score in our cohort was 19. Scores ≥ 4 were defined as multimorbidity.24
At each visit, the physician-reported World Health Organization (WHO) Functional Class was collected. Patients were asked to rate their level of breathlessness using the Borg Dyspnoea Scale, on a numerical rating scale from 0 to 10, where 0 indicated no breathlessness and 10 indicated maximal breathlessness. For analysis, this was dichotomized into scores of < 5 of 10 (no-to-moderate dyspnea) or ≥ 5 of 10 (severe dyspnea).
Statistical analysis. Characteristics of study participants are presented as mean (SD) for normally distributed continuous variables, median (IQR) for nonnormally distributed continuous variables, and as n (%) for categorical variables. Comparisons between groups were performed using 2-sample t test for normally distributed continuous variables, the Wilcoxon signed-rank test for nonnormally distributed continuous variables, and the chi-square test for categorical variables. Logistic regression was used to calculate odds ratios (ORs).
Survival analysis was performed using the endpoint of all-cause mortality. Kaplan-Meier survival curves and the Wilcoxon signed-rank test were used to estimate survival from SSc onset. A multivariable Cox proportional hazards regression analysis was used to determine multivariate associations of mortality. Variables were included in multivariate analyses if they were either clinically relevant or statistically significant on univariate analysis (P < 0.05) and did not violate the proportional hazards assumption. The results are reported as hazard ratios (HRs) with accompanying 95% CIs. Analysis was performed using Stata 17.0 (StataCorp).
RESULTS
One hundred two participants with ILD were included, with a median age at SSc onset of 46.7 years (IQR 35.3-56.2). Extensive ILD was present in 40 participants (41.2%). Twenty-seven (26.5%) participants had definite LVDD, 19 (18.6%) had indeterminate LV diastolic function, and 56 (54.9%) had normal LV diastolic function. Of those with definite LVDD, severity was grade I in 14.8%, grade II in 59.3%, and grade III in 14.8%. Grade was undetermined in 11.1% of participants. Eighty-one participants (79.4%) had all 4 criteria for diastolic function assessment available, whereas 21 participants (20.6%) had 3 of 4 criteria available for assessment. All participants with indeterminate diastolic function had 4 criteria available, whereas 19 (70.4%) of those with definite diastolic function and 34 (60.7%) of those with normal diastolic function had 4 criteria available. Median composite MSS score at baseline was 6 (Table 1).
Characteristics and SSc disease features of the study population (N = 102).
Clinical phenotype and disease features. There was no difference in age at SSc onset or sex between those with ILD and definite LVDD (ILD-LVDD) and those with ILD and normal or indeterminate diastolic function (ILD-only group). Multimorbidity was more common in the ILD-LVDD group (modified CCI ≥ 4; P = 0.02); this group also had higher SSc severity as measured by MSSs (P = 0.03), driven by more severe MSS heart scores in those with ILD-LVDD (P < 0.01).
There were no differences in demographic features, autoantibodies, SSc subtype, duration of disease, and BMI between groups (Table 1). The ILD-LVDD group had a higher modified Rodnan skin score (P = 0.07) and were more likely to have muscle atrophy (P < 0.01) and lower gastrointestinal (GI) symptoms (P < 0.01) than those with ILD alone. Peak erythrocyte sedimentation rate (P < 0.01) and C-reactive protein (P = 0.09) were higher in those with ILD-LVDD.
Those with ILD-LVDD were numerically more likely to have extensive ILD (53.9% vs 36.6%), although this did not reach statistical significance (P = 0.13). Whereas both groups had similar percent predicted FVC at baseline (P = 0.95), baseline percent predicted DLCO was marginally lower in those with ILD-LVDD (P = 0.07). Those with ILD only were more likely to receive immunosuppression, although this did not meet statistical significance (83% vs 67%; P = 0.08), with no difference in prednisolone exposure between groups (P = 0.88; Table 2).
Cardiopulmonary disease features, comorbidities, and treatments.
IHD (P < 0.01) and other vascular disease/risk factors including HTN (P < 0.01), peripheral vascular disease (P < 0.01), and cerebrovascular disease (P = 0.03) were more common in those with ILD-LVDD (Table 2). In keeping with this higher prevalence of cardiovascular (CV) disease, those with ILD-LVDD were more likely to have received multiple antihypertensive, antiplatelet, diuretic, and anticoagulant medications (all P < 0.05), although frequency of calcium channel antagonist use was similar (P = 0.47). On echocardiography, those with ILD-LVDD were more likely to have LVH (P = 0.02), left-sided valvular lesions (P < 0.01), and a higher baseline RVSP (P < 0.01).
Associations of LVDD. Multivariable logistic regression was performed adjusting for age and sex to identify associations of LVDD (Table 3). Participants older than the cohort median at SSc onset (46.7 yrs) were 2.5 times more likely to have LVDD (95% CI 1.0-6.4; P = 0.05). There was no association between LVDD and dcSSc or sex. Participants with IHD (OR 4.8, 95% CI 1.5-15.7; P < 0.01), HTN (OR 5.0, 95% CI 1.8-13.8; P < 0.01), and multimorbidity (OR 3.0, 95% CI 1.1-8.7; P = 0.04) were more likely to have LVDD. Those with proximal muscle atrophy (OR 5.0, 95% CI 1.9-13.6; P < 0.01) were also more likely to have LVDD.
Associations of ILD-LVDD using logistic regression.
Survival. Those with ILD-LVDD had significantly worse survival than those with ILD only (P < 0.01; Figure). In a univariable hazard model, ILD-LVDD was associated with an increased risk of death (HR 2.4, 95% CI 1.0-5.7; P = 0.05; Table 4). This effect was attenuated in a multivariable model including ILD-LVDD (HR 1.1, 95% CI 0.4-2.9; P = 0.85), age at SSc onset, sex, dcSSc, and extensive ILD. Extensive ILD (HR 3.3, 95% CI 1.2-9.1; P = 0.03) and older age at SSc onset (HR 1.1, 95% CI 1.0-1.1; P < 0.01) were also associated with increased risk of mortality.
Kaplan-Meier survival estimates in those with ILD-LVDD, compared to those with ILD only. ILD: interstitial lung disease; LVDD: left ventricular diastolic dysfunction; SSc: systemic sclerosis.
Cox proportional hazard model for survival from SSc onset to death (all-cause mortality).
Comparison of functional status and dyspnea between groups. Those with ILD-LVDD had more severe dyspnea in both univariable analyses (OR 2.5, 95% CI 1.0-6.3; P < 0.05) and after adjusting for age and sex (OR 2.6, 95% CI 1.0-6.6; P = 0.05). Those with ILD-LVDD were more likely to record a WHO Functional Class II dyspnea or above in both univariable analyses (OR 4.2, 95% CI 0.9-20.0; P = 0.08) and after adjusting for age and sex (OR 4.0, 95% CI 0.8-19.3; P = 0.08), although this did not meet statistical significance.
DISCUSSION
Twenty-six percent of 102 individuals with SSc-ILD had definite LVDD, with a further 19% having indeterminate LV diastolic function and 55% having normal diastolic function. Both survival and breathlessness were worse in those with ILD-LVDD. LVDD in the setting of ILD was associated with increasing age, CV disease, and multimorbidity. These data suggest that in people with SSc and ILD, classical risk factors are a significant contributor to the development of LVDD. These include older age, left heart disease, and traditional CV risk factors such as HTN. This is in addition to SSc-specific features that may further increase the risk of LVDD such as skeletal muscle involvement, chronic inflammation, and GI involvement. The strong association with muscle atrophy suggests LVDD may also be associated with myopathy or physical frailty in SSc-ILD. However, regardless of the mechanism, once present, LVDD is associated with poorer survival and more breathlessness. This highlights the importance of considering extrapulmonary contributors to dyspnea in those with SSc-ILD.
We identified a 2-fold increase in mortality in those with ILD and LVDD compared to ILD only, although this effect was attenuated after adjusting for ILD severity, age, sex, and dcSSc. This is consistent with other studies reporting higher mortality in general SSc cohorts with LVDD,4,6,10 with up to a 3-fold increased risk of death.6 There are several reasons for an attenuated signal in our cohort. In SSc-ILD, there may be ILD-related early contributors to mortality before LVDD develops, as LVDD may be a later manifestation as participants age. Further, individuals in our study were recruited approximately 6 years after SSc onset, suggesting a degree of “survivor bias” as those with more aggressive disease and early mortality less frequently survive to recruitment, which may underestimate survival differences between groups. It is also possible that if myocardial fibrosis is the underlying pathogenic mechanism of diastolic dysfunction in SSc, then treating fibroinflammatory complications of SSc (including SSc-ILD) may modify the course of LVDD. However, this hypothesis remains untested in SSc.
Cardiopulmonary disease is closely associated with health-related quality of life in SSc-ILD.25 In particular, LVDD in SSc is correlated with higher levels of dyspnea12 and poorer exercise tolerance.9 In our cohort, those with ILD-LVDD had more dyspnea than those with ILD only. Although both groups exhibited high levels of dyspnea, individuals with LVDD were more likely to report severe limitation of activity. This supports the idea that LVDD, even in the absence of clinical heart failure, contributes to exercise impairment and poor physical function in SSc.5,7 Given that echocardiography is already recommended annually in SSc cohorts to screen for PAH, this study underscores the importance of recognizing LVDD on echocardiography in SSc-ILD as an opportunity to identify a group at risk of increased breathlessness in the absence of ILD progression as well as a potentially poorer prognosis.
We identified a high frequency of important comorbidities in participants with ILD-LVDD. Those with ILD-LVDD more commonly were older and frequently had left heart disease that may contribute to the development of LVDD, and independently contribute to breathlessness and functional decline. We identified a strong association between multimorbidity indices and LVDD. This study lacked the power to fully control for all important comorbidities to establish LVDD as a direct cause of the functional limitation observed. Alternatively, the presence of ILD-LVDD may reflect the significant multimorbidity present in this cohort, thus being a secondary phenomenon of other cardiopulmonary diseases.
In applying a stringent definition of LVDD,15 our cohort’s 26% prevalence of definite LVDD is similar to estimates in general SSc populations of 20% to 30%.4-6 We did, however, note an unusual pattern of LVDD grading, with significantly more cases having grade II LVDD compared to grade I and III, whereas studies in the non-SSc population tend to have more grade I LVDD than the more severe grades. It is possible that this reflects a different pathological process in SSc than in the general population and we have seen this in our previous research.12 The frequency of LVDD may be increased in SSc-ILD12 and in general cohorts with pulmonary fibrosis26 and chronic obstructive pulmonary disease.27,28 However, in SSc, myocardial fibrosis, which is detected on cardiac magnetic resonance imaging in 53% to 94% of people without overt cardiac disease,29,30 may reduce LV compliance and thus predispose to LVDD.31 This myocardial fibrosis may be caused by recurrent microvascular ischemia and reperfusion injury because of abnormal vasoreactivity.11 Coronary vasospasm evidenced by reversible perfusion defects and regional wall motion abnormalities on single photon emission computed tomography in response to peripheral cold exposure has been identified in SSc,32,33 as have exercise-induced perfusion defects in SSc cohorts with normal coronary angiography.34,35 Fibrinoid necrosis of intramural coronary arteries has been reported at autopsy in addition to concentric myocardial fibrosis,36 suggesting that microvascular perfusion abnormalities may be compounded by structural changes in the coronary microvasculature.11 Further, while data are conflicting about an independent association between LVDD and coronary artery disease after adjusting for shared risk factors,37,38 peripheral endothelial dysfunction may be a risk factor for heart failure with preserved ejection fraction in the general population,39-41 and confer poorer prognosis.39,41 This supports an association between microvascular dysfunction and LVDD in both the general population and in SSc. It is also possible that higher frequencies of fibrotic disease occur in those with SSc and LVDD than without, supported by other data demonstrating a higher frequency of LVDD in those with ILD than those without ILD12,13; this association warrants further exploration.
This study has limitations. This study is of modest size with only 102 participants, meaning these analyses should be validated in a larger cohort. Further, this sample size limits the power of certain analyses, especially those exploring less common SSc features (eg, myositis). Participants in the ASCS do not undergo a protocolized HRCT; thus, we may have underestimated the prevalence of ILD in our cohort and selected patients with more severe ILD. We do not collect data within the ASCS about duration of traditional CV risk factors, including HTN.
To our knowledge, this is the first study to describe the prevalence and clinical correlates of LVDD specifically in patients with SSc-ILD, and to explore extrapulmonary causes of breathlessness in this group. We demonstrated that 26% of individuals with SSc-ILD had concomitant LVDD; older age, HTN, IHD, and multimorbidity were all associated with LVDD, as were specific SSc disease features including skeletal muscle disease, GI involvement, and raised inflammatory markers. Those with ILD-LVDD had a higher mortality than those with ILD alone. Patients with ILD-LVDD were more breathless, highlighting the multifactorial nature of symptomatic and functional decline in individuals with SSc-ILD. Further studies are required to elucidate the mechanisms of LVDD in SSc-ILD and the effect of treatment on symptom burden and prognosis.
Footnotes
The Australian Scleroderma Cohort Study is supported by Janssen, Boehringer Ingelheim, Scleroderma Australia, Scleroderma Victoria, Arthritis Australia, Musculoskeletal Australia (muscle, bone, and joint health), Australian Rheumatology Association (ARA), and St. Vincent’s Hospital Melbourne IT Department and Research Endowment Fund. JLF holds a National Health and Medical Research Council (NHMRC) Postgraduate Scholarship (2013842) and an Australian Government Research Training Program Scholarship. LJR holds an Arthritis Australia/ARA Victoria Fellowship. MN holds an NHMRC Investigator Grant (GNT1176538).
The Australian Scleroderma Interest Group received research support grants from Janssen and Boehringer Ingelheim. WS has received consulting fees from Boehringer Ingelheim. MN has received consulting fees from Janssen, Boehringer Ingelheim, AstraZeneca, and GSK, and honoraria from Janssen, Boehringer Ingelheim, and GSK. JLF has received honoraria from Boehringer Ingelheim and conference sponsorship from Pfizer. DP has received speaker fees from Novartis and Janssen. The remaining authors declare no conflicts of interest relevant to this article.
- Accepted for publication December 9, 2023.
- Copyright © 2024 by the Journal of Rheumatology
