Right ventricular (RV) function has not been systematically assessed in sarcoidosis. The aim of this study was to assess the prevalence and associates of RV dysfunction in sarcoidosis using global longitudinal peak systolic strain (GLS). Furthermore, whether RV dysfunction was associated with clinical outcomes was investigated.
A total of 88 patients with sarcoidosis (mean age, 54 ± 13 years; 51% men) without known sarcoid-related or other structural heart disease or alternative etiologies of pulmonary hypertension were retrospectively included. RV GLS was measured using two-dimensional speckle-tracking echocardiography, and patients were stratified (using a previously defined cutoff value) as having preserved (RV GLS < −19%) or impaired (RV GLS ≥ −19%) RV function. An age- and gender-matched control group ( n = 50) was included. The main outcome was all-cause mortality or clinical heart failure (hospitalization or New York Heart Association functional class ≥ III and/or deterioration by one or more classes).
RV GLS was significantly reduced (−20.1 ± 4.6 vs −24.6 ± 1.8%, P = .001) in patients compared with control subjects. Patients with impaired RV function ( n = 41) were older and had worse pulmonary function, worse left ventricular diastolic function, and lower tricuspid annular plane systolic excursion compared with patients with preserved RV function ( n = 47). Lower tricuspid annular plane systolic excursion and diabetes were independent correlates of RV GLS. Over a median follow-up period of 37 months, 19 clinical end points occurred. Patients with impaired RV function were more likely to experience the clinical end point (log-rank P = .003).
RV contractile dysfunction, identified using RV GLS, is common in patients with sarcoidosis without manifest cardiac involvement or pulmonary hypertension and is associated with adverse outcome. RV GLS may therefore be useful to detect sarcoidosis-related RV dysfunction at an earlier and potentially modifiable stage.
Manifest RV dysfunction is known to negatively affect outcomes in patients with sarcoidosis.
RV function, assessed quantitatively by GLS, may be significantly reduced in patients with sarcoidosis even in the absence of overt clinical or otherwise apparent cardiac involvement or severe pulmonary hypertension.
A prespecified cutoff for RV GLS of −19%, shown previously to differentiate between normal and impaired RV contractile dysfunction in a general heart failure population, identified patients with sarcoidosis with impaired RV function as significantly more likely to experience death or clinical heart failure.
The ability of GLS to detect RV contractile dysfunction at an earlier and/or intermediate stage in sarcoidosis may facilitate the targeting of specific antiremodeling and/or pulmonary vasoactive therapies to these patients before the development of clinical heart failure or irreversible pulmonary hypertension.
Left ventricular (LV) systolic dysfunction is an undisputed harbinger of poor prognosis in patients with sarcoidosis. However, less is known about right ventricular (RV) function in these patients, even though RV impairment can arise through multiple mechanisms, including primary involvement of the RV myocardium or as a consequence of pulmonary hypertension. Direct granulomatous infiltration of the myocardium of the interventricular septum and/or the RV free wall not only can manifest as clinical heart failure but also can predispose to lethal bradyarrhythmias and/or tachyarrhythmias that may occur as the first mode of clinical presentation. Pulmonary hypertension, regardless of the specific etiology, is also a well-known adverse prognostic factor in patients with sarcoidosis, and the development of superimposed RV dysfunction has been shown to further negatively affect outcomes. Therefore, earlier detection of RV contractile impairment, whether primary or secondary, before the development of heart failure or significant pulmonary hypertension, is clinically relevant in this patient population.
Two-dimensional (2D) echocardiography remains the first-line tool to detect cardiac structural abnormalities in patients with sarcoidosis. However, accurate assessment of RV function by conventional echocardiography is challenged by several factors, most notably the complex geometry of the RV cavity. Meanwhile, longitudinal shortening has been shown to be the most important contributor to RV systolic function. Speckle-tracking strain echocardiography enables the direct assessment of intrinsic RV myocardial function, and RV global longitudinal peak systolic strain (GLS) has been recently validated in both normal subjects and those with RV dysfunction. It has also been demonstrated as an independent marker of prognosis in an all-comers pulmonary hypertension population.
Accordingly, the principal aim of the present study was to assess the prevalence and associates of isolated RV dysfunction in sarcoidosis in the absence of overt cardiac involvement or severe pulmonary hypertension using RV GLS as a sensitive, direct parameter of RV function. Second, we sought to investigate whether early identification of RV dysfunction using GLS is associated with clinical outcomes in these patients.
This was an observational retrospective cohort study. A total of 130 patients with sarcoidosis attending our referral center (Leiden University Medical Center, Leiden, The Netherlands) and undergoing 2D echocardiography at the time of or following their diagnoses were identified using the departmental cardiology information system (EPD-Vision). The diagnosis of sarcoidosis was made in the setting of a compatible clinical picture and the absence of an alternative disease process capable of producing a similar clinical syndrome, with or without histologic confirmation of the presence of noncaseating granulomas. Given that the aim of the present study was to identify the prevalence of RV dysfunction in patients with sarcoidosis independent of manifest systolic LV impairment or other clinical evidence of cardiac involvement, patients with known or suspected cardiac sarcoidosis on the basis of the Japanese Ministry of Health and Welfare diagnostic criteria were excluded ( n = 5 definite cardiac sarcoidosis, n = 9 one major criterion alone). Similarly, those with non-sarcoid-related etiologies of structural heart disease ( n = 16) were also excluded. Finally, patients with noncardiac conditions associated with pulmonary hypertension were also excluded to avoid possible confounding in the setting of differential causes of afterload-associated RV dysfunction ( Figure 1 ).
Control subjects were identified from the departmental echocardiographic database (EchoPAC version 112.0.0; GE Vingmed Ultrasound AS, Horten, Norway) using a dedicated search code defining the absence of cardiac structural abnormalities in subjects without histories of cardiac disease, typically referred for echocardiography in the setting of cardiovascular risk stratification or the finding of a murmur on auscultation. An age- and gender-matched comparator group ( n = 50) was selected on the basis of published comparability principles.
The current retrospective evaluation of clinically acquired data was approved by the institutional review board of the Leiden University Medical Center, which waived the need to obtain written informed consent.
Study Protocol and Clinical End Points
Clinical data, including organs involved, mode of diagnosis, serum angiotensin-converting enzyme and lysozyme levels, and results of chest radiography and pulmonary function testing, were recorded for all patients. Pulmonary function parameters were measured according to American Thoracic Society and European Respiratory Society recommendations. RV GLS was measured by 2D speckle-tracking analysis applied to baseline echocardiographic images alongside comprehensive 2D and Doppler analysis in all patients and control subjects. The main outcome measure was prespecified as a composite end point of all-cause mortality or clinical heart failure, defined as heart failure–related hospitalization, New York Heart Association (NYHA) functional class III or IV symptoms, or deterioration of NYHA functional class by one or more classes from baseline. Outcome data were assessed by retrospective review of electronic medical data and the official Dutch National Survival Registry. Follow-up was available for all included patients.
Two-dimensional grayscale and Doppler images were acquired with patients in the left lateral decubitus position using a commercially available system (Vivid 7 and e9; GE Vingmed Ultrasound AS) equipped with 3.5-MHz or M5S transducers. Image analysis was performed offline using EchoPAC version 112.0.0. LV chamber and wall thickness quantification were performed according to standard recommendations. LV ejection fraction (LVEF) was derived from LV volumes quantified according to the Simpson biplane method, as recommended. Left atrial volume was also measured using the Simpson biplane technique and subsequently indexed to body surface area. Severity of mitral regurgitation and tricuspid regurgitation was graded according to recent guidelines. For diastolic function, mitral inflow was analyzed using pulsed-wave Doppler and early (E) and late (A) diastolic velocities; E-wave deceleration time was also assessed. Mitral annular peak early velocity (average E′) was derived from color-coded tissue Doppler imaging at both the septal and lateral site of the mitral annulus, with individual site values averaged to give an overall value. The ratio of E-wave velocity to average E′ (E/E′) was subsequently calculated. Tricuspid annular plane systolic excursion (TAPSE) was measured at the RV free wall as the conventional measure of RV function. Pulmonary artery systolic pressure was estimated using RV systolic pressure calculated from the tricuspid regurgitation peak gradient added to assessment of right atrial pressure using inferior vena cava size and collapsibility, as previously described.
Two-Dimensional Speckle-Tracking Analysis
Speckle-tracking analysis was performed using commercially available software (EchoPAC version 112.0.0). For analysis of the right ventricle, the standard apical four-chamber view was used, as previously described. The endocardial border of the entire right ventricle was manually defined at end-systole, and the automatically created region of interest was then adjusted to the thickness of the myocardium. Longitudinal strain curves were generated for all six segments (basal, mid, and apical segments of both the septum and RV free wall). RV free wall longitudinal strain (FWS) was calculated as the average of the basal, mid, and apical segments of the RV free wall. RV GLS was then calculated as the average peak longitudinal systolic strain across the six regional curves ( Figure 2 ). Both measures represent the percentage of longitudinal shortening in systole relative to the original length and are expressed by convention as negative values. Therefore, a less negative value implies reduced magnitude of longitudinal strain, and a more negative value indicates greater magnitude of strain. The patient population was stratified into two groups (preserved RV function [RV GLS < −19%] and impaired RV function [RV GLS ≥ −19%]) on the basis of the previous demonstration of this cutoff for RV GLS to accurately differentiate between normal and impaired RV contractility (sensitivity, 95%; specificity, 85%).
LV GLS analysis was similarly performed by tracing the LV endocardial border at end-systole in apical four-, two-, and three-chamber views stored as digital cine loops and processed offline. LV GLS was provided by the software as the average value of peak systolic longitudinal strain across all three apical views using a 17-segment model.
Continuous variables are presented as mean ± SD and categorical variables as frequencies and percentages. Variables were compared across patient and control or RV function groups using either the unpaired Student’s t test or the χ 2 test. Univariate and multivariate linear regression analyses were then performed to assess the independent correlates of RV GLS in the total patient population. All parameters achieving significance levels of P < .10 between RV GLS groups were selected for linear regression analysis; those parameters achieving significance at the univariate level ( P < .05) were subsequently entered into the multivariate analysis. Collinearity between individual univariate parameters was checked for using the Pearson correlation coefficient. Receiver operating characteristic curve analyses were then performed to determine the area under the curve for each of the RV function parameters (TAPSE, RV GLS, and RV FWS) for association with the clinical composite end point. Sensitivity and specificity of the predefined cutoff for RV GLS of ≥−19% were determined for this outcome. Finally, event rates for the risk for death or clinical heart failure were plotted in Kaplan-Meier curves and compared between RV GLS groups using the log-rank test.
Statistical analyses were performed with SPSS Statistics (IBM, Armonk, New York). All tests were two sided, and P values < .05 were considered to indicate statistical significance.
Reproducibility of RV GLS and RV FWS were expressed as both interclass correlation coefficients and as a percentage of the absolute difference divided by the mean of the pair-repeated observations (absolute difference). For intra- and interobserver variability, 15 patients were selected at random, and measurements were repeated by the same observer on the same echocardiographic images ≥2 weeks apart, and by another independent observer.
RV GLS assessment was feasible in 91% of patients ( n = 88) eligible for inclusion, who constituted the final patient population ( Figure 1 ). The mean frame rate for the total study population was 64 ± 16 frames/sec. Baseline demographic and clinical characteristics of the patient population are shown in Table 1 . The mean age was 54 ± 13 years (control subjects, 54 ± 13 years; P = .25), and 51% of patients (control subjects, 46%; P = .56) were men. Patients and control subjects did not differ in terms of the presence of cardiovascular risk factors (hypertension, P = .22; diabetes, P = .59; hyperlipidemia, P = .08). Pulmonary involvement was present in the majority of patients (89%). Median estimated pulmonary artery systolic pressure for the overall patient population was 24 mm Hg (interquartile range, 18–32 mm Hg).
|Variable||Patient population ( n = 88)||Impaired RV function ( n = 41)||Preserved RV function ( n = 47)||P ∗|
|Age (y)||54 ± 13||59 ± 12||50 ± 13||.001|
|Male gender||45 (51%)||18 (44%)||27 (57%)||.21|
|Hypertension||22 (25%)||13 (32%)||9 (19%)||.18|
|Diabetes||8 (9%)||6 (15%)||2 (4%)||.09|
|Hyperlipidemia||26 (30%)||16 (39%)||10 (21%)||.07|
|Smoking history||21 (24%)||9 (22%)||12 (26%)||.69|
|Histologic + clinical diagnosis vs clinical diagnosis||57 (65%)||21 (51%)||36 (77%)||.07|
|Pulmonary||78 (89%)||37 (90%)||41 (87%)||.66|
|Skin||30 (34%)||10 (24%)||20 (43%)||.10|
|Eye||15 (17%)||9 (22%)||6 (13%)||.21|
|Neurologic||7 (8%)||3 (7%)||4 (9%)||.84|
|Baseline NYHA class ≥ II||11 (13%)||7 (17%)||4 (9%)||.14|
|0||10 (11%)||4 (10%)||6 (13%)|
|1||37 (42%)||15 (36%)||22 (47%)|
|2||24 (27%)||11 (27%)||13 (28%)|
|3||8 (9%)||4 (10%)||4 (8%)|
|4||9 (10%)||7 (17%)||2 (4%)|
|Serum ACE (nmol/min/mL)||64 ± 37||63 ± 39||65 ± 37||.87|
|Lysozyme (%)||240 ± 165||227 ± 175||250 ± 160||.60|
|FEV 1 (L)||2.9 ± 1.1||2.6 ± 1.1||3.1 ± 1.1||.049|
|DLCO (mmol/min/kPa/L)||7.4 ± 3.1||7.0 ± 3.0||7.8 ± 3.2||.38|
Regarding 2D echocardiographic parameters compared between patients and control subjects, as illustrated in Table 2 , patients with sarcoidosis had significantly lower LVEF and TAPSE, although for both parameters, mean values for patients remained within the normal ranges. Both quantitative global and regional RV function using RV GLS and RV FWS, respectively, were significantly reduced in patients with sarcoidosis compared with control subjects; however, 47% of the patient population ( n = 41) had impaired RV function according to the previously defined cutoff for GLS.
|Variable||Control population ( n = 50)||Patient population ( n = 88)||Impaired RV function ( n = 41)||Preserved RV function ( n = 47)||P ∗|
|LVEDD (cm)||5.1 ± 0.51||4.9 ± 0.51||4.9 ± 0.56||5.0 ± 0.45||.29|
|LVESD (cm)||3.1 ± 0.41||3.2 ± 0.46||3.2 ± 0.50||3.2 ± 0.43||.67|
|LVESV (mL)||40 ± 10||46 ± 14 †||45 ± 14||47 ± 15||.45|
|LVEDV (mL)||107 ± 21||106 ± 30||103 ± 28||108 ± 31||.42|
|LVEF (%)||63 ± 6||57 ± 5 †||57 ± 5||57 ± 5||.29|
|MR grade ≥ 2||1 (2%)||4 (5%)||1 (2%)||3 (6%)||.38|
|LAVI (mL/m 2 )||25 ± 8||24 ± 8||24 ± 8||24 ± 7||.71|
|E (cm/sec)||79 ± 19||75 ± 17||72 ± 16||79 ± 17||.06|
|A (cm/sec)||67 ± 16||70 ± 16||72 ± 17||69 ± 16||.32|
|DT (msec)||204 ± 42||199 ± 41||198 ± 37||200 ± 44||.79|
|Average E′ (m/sec)||8.3 ± 2.3||7.5 ± 2.0||6.7 ± 1.9||8.1 ± 1.9||.002|
|Average E/E′ ratio||10 ± 3||11 ± 3||12 ± 4||10 ± 2||.046|
|TAPSE (cm)||2.4 ± 0.36||2.1 ± 0.33 †||2.0 ± 0.27||2.2 ± 0.35||.02|
|TR grade ≥ 2||2 (4%)||6 (7%)||3 (8%)||3 (7%)||.83|
|PASP (mm Hg)||22 ± 7||26 ± 15||28 ± 18||25 ± 12||.36|
|RV FWS (%)||−30.6 ± 4.2||−23.9 ± 6.7 †||−19.7 ± 6.0||−27.5 ± 4.8||<.001|
|RV GLS (%)||−24.6 ± 1.8||−20.1 ± 4.6 †||−16.1 ± 2.8||−23.6 ± 2.5||<.001|