Among the 83 patients who met the histologic and clinical criteria of CS according to the guidelines of the HRS consensus statement,(12) 25 patients had a definite CS diagnosis (granuloma detected at endomyocardial biopsy) and 58 a probable CS diagnosis (24 with perfusion defects and FDG uptake on FDG-PETand 34 patients with positive CMR findings, in which 16 had an inflammatory pattern with increased signal intensity T2 sequences and 18 had patchy delayed enhancement with non-coronary distribution). The main symptoms associated with the diagnosis were dyspnea (52%), palpitations (21%), and chest pain (11%). The mean (range) age of the cohort at diagnosis of CS was 53.6±10.8 (25-75) years, 34.9% were women, 14.5% were African American, 67.5% had lung involvement, and 68.7% had lymph node involvement. A few patients had eye (9.6%), skin (12.0%), or central nervous system (5.0%) involvement. The mean age of the patients at the initial diagnosis of extracardiac sarcoidosis was 50.6±11.3 years, and myocardial involvement was detected after a median (range) of 26.3 (0-1.291) months. On ECG, 67 patients had the following arrhythmias: type 2 second-degree atrioventricular (AV) block (3 patients, 3.6%), third-degree AV block (22 patients, 26.5%), right bundle branch block (14 patients, 16.9%), left bundle branch block (2 patients, 2.4%), and ST-segment and T-wave alterations (26 patients, 31.3%). Acute coronary syndrome or other cardiomyopathies not related to sarcoidosis were ruled out in this cohort of patients. Other general characteristics of the CS patients and comorbidities are shown in Table 1.
Echocardiographic Parameters
All 83 patients underwent transthoracic echocardiographic evaluation. Mean LV mass indexed for body surface area was 111.9±32.3 g/m2 in men and 104.2±29.8 g/m2 in women; mean RWT was 0.38±0.08 in men and 0.37±0.08 in women. Thirty-one patients (37.3%) had eccentric hypertrophy, and 8 (9.6%) had concentric hypertrophy. Of the patients, 27% had normal LVEF, and 13.3% had severe LV systolic dysfunction (LVEF<30%). Wall motion abnormalities were detected in two-thirds of patients, in particular in the basal (up to 22%) and mid-segments of the LV (up to 33%). Parameters of LV diastolic function were abnormal in most patients (grade 1: 39.3%, grade 2: 23.8%, grade 3: 7.1%). Of the patients, 22.9% had abnormal RV systolic function, and 36.1% had RV dilatation. Thirty-six (43%) patients had pulmonary hypertension: Of those, 23 had concurrent sarcoid pulmonary involvement; and 3 had severe tricuspid regurgitation. Other echocardiographic parameters are shown in Table 2.
Typical findings for advanced CS were found in 15 patients: 11 (13.3%) had interventricular septal thinning; and 4 (4.9%), wall aneurysms; all of these patients had severe LV and RV systolic dysfunction.
Strain Analysis
Mean LV GLS, GCS and GRS and RV GLS of CS patients are shown in Table 3. The percentage of reliability of 2D-STE analysis was 91% for longitudinal strain, 89% for circumferential and radial strain. Figure 1 shows representative bulls eye display of LV segmental longitudinal strain; the lowest mean values were detected in the LV basal and mid interventricular inferoseptum as well as in the inferior wall. CS patients with a positive endomyocardial biopsy had more compromised LV GLS (-10.5%±3.7), GCS (-13.3%±6.7), global longitudinal systolic strain rate (GLSRs) (-0.6±0.2 s-1), and global circumferential systolic strain rate (GCSRs) (-0.8±0.5 s-1) as well as lower global longitudinal early diastolic strain rate (GLSRe) (0.6±0.2 s-1) and global circumferential early diastolic strain rate (GCSRe) (0.9±1.3 s-1) values; furthermore, RV and free wall RV GLS were also reduced more in this group of patients (-11.5%±4.5 and -13.0%±4.4, respectively). In addition, no differences in strain parameters were detected for patients who had never been treated compared with those treated with past or concurrent immunosuppressant therapy (including steroids) before the diagnosis of CS (data not shown), or those with high blood pressure.
Comparison With Patients Without Sarcoidosis
Twenty-three patients with early stage of CS and LV systolic function within normal limits (LVEF>52% for men; >54% for women, mean value: 57.3%±3.8%) and no wall motion abnormalities or RV dysfunction at the time of CS diagnosis (mean age, 52.9±10.0 years; 21.7% women; mean body mass index, 29.9±4.7 kg/m2) were selected and their parameters compared with those of 97 controls (4:1 ratio) with normal echocardiographic findings and no history of coronary artery disease or other cardiovascular and metabolic comorbidities. At the time of cardiac evaluation, arterial blood pressure, linear dimensions of LV walls and LV mass were within normal limits. Only 16% of these CS patients had diastolic dysfunction (grade 1), and only 4 patients had mildly elevated LV filling pressures. No wall aneurysms or septal thinning were detected. Parameters of RV systolic function were normal in all patients (mean tricuspid S′, 0.13±0.02 m/s and fractional area change 39.6%±9.4%). Mean pulmonary artery pressure was 52±17 mmHg; no severe tricuspid or mitral valve impairment was detected, and only two CS patients had moderate tricuspid regurgitation. All 2D–STE and conventional systolic and diastolic echocardiographic parameters were significantly reduced when they were compared to those of controls (Table 4).
Diagnostic Value of Speckle Tracking Analysis
ROC analysis was conducted evaluating the diagnostic value of LV GLS and RV GLS for the identification of cardiac involvement of sarcoidosis. As shown in Figure 2A&B, a LV GLS value of -16.3% provided 82.2% sensitivity and 81.2% specificity for the diagnosis of CS (area under the curve, [AUC] 0.91); An inferoseptal GLS value of -16.3% provided 84.9% sensitivity and 67% specificity for the diagnosis of CS (AUC 0.85); when the inferior wall was added to the model, the sensitivity increased to 86.3% and specificity to 70.1% (AUC: 0.90). A RV GLS value of -19.9% provided 88.1% sensitivity and 86.7% specificity (AUC 0.93), while a free wall RV GLS of -21.4% provided 86.4% sensitivity and 80.6% specificity (AUC 0.91) Figure 3. LVEF, LV GCS, and LV GRS did not offer additional diagnostic value for CS.
Intra-class correlation coefficients were the highest for global longitudinal strain measurements, followed closely by circumferential strain measurements, and the lowest for global radial strain measurements. GLS 0.998 (0.994-0.999), 0.96 (0.989-0.998), GCS 0.997 (0.994-0.999), 0.993 (0.983-0.997) and GRS 0.857 (0.751-0.933) and 0.845 (0.727-0.927) for intra-observer and inter-observer respectively.
Outcomes
Median follow-up was 26.3 months (IQ25% 9.59, Q75% 68.2 months, range 1-226.2). An LV GLS value more positive than −14% was found to be related to a higher rate of hospital admission for cardiac complications (OR, 4.17 [95% CI, 1.34-12.98]; P=.01) and heart failure (OR, 5.03 [95% CI, 1.04-21.32]; P=.04) in the univariate logistic regression analysis. Reduced LVEF or abnormal ECG findings did not correlate with cardiac events (data not shown).