Clinical profile and mortality predictors in cardiac sarcoidosis: A single centre observational study

INTRODUCTION

Cardiac sarcoidosis (CS) is a granulomatous involvement of the myocardium.¹ It is often under-recognized in clinical practice,² and most often occurs as a manifestation of systemic sarcoidosis, most commonly with lung involvement. However, isolated CS can occur in patients without evidence of sarcoidosis in other organs.³ Clinically manifested CS has been estimated to occur in approximately 5% of patients with systemic sarcoidosis. Autopsy and imaging studies of patients with systemic sarcoidosis have identified evidence of CS in 20%–29% patients.^4,5 We aimed to understand the clinical profile and survival pattern in CS, including predictors of mortality.

METHODS

Forty-three consecutive patients diagnosed with CS at our institution from January 2005 through January 2021 were included in this study. The diagnosis was made by histo-logical demonstration of non-caseating granuloma in a biopsy from cardiac or extracardiac tissue whenever possible. However,in the absence of cardiac biopsy, cardiac involvement was confirmed by MRI, 18F-fluorodeoxyglucose (FDG) positron emission tomography, and echocardiography. These criteria are in line with the Heart Rhythm Society 2014 expert consensus report,⁶ and Japanese Circulation Society 2016 clinical diagnosis group criteria,⁷ and only those patients meeting these diagnostic criteria were included. Of 45 patients, 2 with suspected viral myocarditis and 1 with severe coronary artery disease were excluded. The medical records of the eligible patients were reviewed retrospectively for demographic data, comorbid conditions, presenting clinical manifestations, laboratory tests, and imaging studies results, treatment with drugs and devices, and occurrence of refractory heart failure (HF), appropriate implantable cardioverter defibrillator (ICD) shocks, cardiac arrhythmias, and all-cause death through the end of January 2021. Identification and confirmation of arrhythmias were based on retrospective analysis of ICD interrogation reports, 12-lead electrocardiograms, 24-hour Holter recordings, and attending physicians’ documentation. The institutional ethics committee at Sree Chitra Tirunal Institute of Medical Science and Technology approved the study protocol.

RESULTS

Of the 43 patients, 31 were males, aged 49 (8.8) years. The common presenting clinical features were ventricular tachycardia (VT) and acute heart failure in 14 each (33%; Table 1). The median duration of diagnosis of CS after identification of systemic sarcoidosis was 18 (interquartile range, IQR 4.5–45) months.

All patients had histopathological and/or radiological evidence of systemic sarcoidosis (Table 2). Lymphadenopathy was seen in 28 (65%), of which 13 (32%) had detectable hilar lymphadenopathy on a chest X-ray. Pulmonary parenchymal involvement was present in 26 (60%), most commonly in isolated nodules (55%). Nearly half had left ventricular dysfunction at diagnosis. Median N-terminal-pro-brain natriuretic peptide was 1005 (371–2898) pg/ml. Median erythrocyte sedimentation rate (ESR) was 25.7 (9–35) mm/1st hour and 14% of patients had elevated angiotensin converting enzyme (ACE) levels (>twice upper limit of normal).

Electrocardiographic features (Table 3) included a complete right bundle branch block (RBBB) pattern in 9 (22%) and an incomplete RBBB in 14%, with a predominant Rsr’ pattern in lead V1. Complete left bundle branch block (LBBB) pattern was seen in 12%, fascicular block in 7%, QRS fragmentation in 32%, VT with RBBB morphology was seen in 58% and LBBB morphology in 42%; 38% had major ventricular extrasystole burden (>10/hour) on Holter monitoring. Ventricular arrhythmia was localized based on ECG to the outflow tract region in 45%, the basal septal in 25%, the free wall in 18%, and pattern break in 12%. Multiple VT morphologies were documented in 30% of patients.

On echocardiography, 22 (53%) had dilated cardiomyopathy phenotype and global left ventricular hypokinesia with basal septal thinning as the predominant feature (33%). Complete cardiac MRI data were available in 33 patients (Table 3). Late gadolinium enhancement (LGE) positivity was seen in 94%. Multiple patterns of delayed enhancement were seen in the same patient. The most commonly observed patterns were subepicardial (58%) and midmyocardial (54%). The median value of quantification of delayed enhancement of the left ventricle showed 20.8% v/v (16.3–24.7). Fragmented QRS was a 100% specific predictor of the presence of free wall scar (p=0.001; Fig. 1). Septal LGE correlated with conduction abnormalities (p=0.006; 22/26 v. 2/7). The mean LV ejection fraction by cardiac MRI was 40.1 (17.8), and a low ejection fraction (<40%) was associated with the occurrence of acute HF at presentation (p=0.041).

Close

FIG 1.

Fragmented QRS complexes in infero-lateral leads correlated with basal inferior septal (thin arrow) and lateral free wall (thick arrow) left ventricular scar in cardiac magnetic resonance imaging, oval shape indicates fragmented QRS in inferior leads

Export to PPT

The defibrillator device was implanted in 18 (42%) patients and a permanent pacemaker was implanted in 8 (19%). Appropriate ICD shock was observed in 8 (44%) patients. These patients were followed up for a mean duration of 4.3 years. On follow-up (Table 4), we observed major improvement in New York Heart Association (NYHA) functional class (p=0.011) and mean ejection fraction (p=0.155). Eleven (25%) patients died, 10 (23%) had recurrent heart failure (HF) readmissions, 5 (29%) had recurrent ICD shocks, and 6 (14%) had new onset VT.

In univariate analysis, mortality was associated with acute HF at presentation, elevated ESR, low ejection fraction at diagnosis, persistent low ejection fraction on follow-up, and recurrent HF admissions (Table 5). Multivariate analysis (Table 6) revealed that acute HF at presentation, elevated ESR at the time of diagnosis, and persistent low ejection fraction <40% during follow-up were independent predictors of mortality. Survival analysis showed that recurrent HF admission predicts early mortality, not VT (Fig. 2a and 2b).

Close

FIG 2a.

Survival curve of patients with and without recurrent admissions for heart failure

Export to PPT

Close

FIG 2b.

Survival curve of patients who did and did not receive intracardiac defibrillator shock on follow-up

Export to PPT

DISCUSSION

We observed a male predominance in cardiac involvement with systemic sarcoidosis, akin to previous studies.⁸ More than half our patients had left ventricular dysfunction. In a population-based cohort study, sarcoidosis was associated with a 2.4-fold increased hazard of HF compared with the general population.⁹ The prevalence of HF in CS patients was 25%–75%.¹⁰

Electrocardiography is essential for evaluating cardiac involvement in patients with sarcoidosis.¹¹ Conduction abnormalities occur due to involvement of the basal interventricular septum and range from complete or incomplete bundle branch block to atrioventricular block. Analysis of the QRS complex in a sarcoidosis patient is an essential clinical step. It may help identify a patient with cardiac involvement, specifically fragmented QRS, a marker of free wall scar that can be a substrate for VT.¹² Fragmented QRS is 100% sensitive and 80% specific for LGE in cardiac MRI.¹³ Multiple morphologies of ventricular ectopy or ventricular runs in Holter monitoring may suggest involvement of the myocardium in patients with CS.¹⁴

Greulich et al.¹⁵ concluded that the presence of LGE is the best independent predictor of death and other adverse events in patients with CS. We noted that the distribution of LGE predicted certain clinical features in CS patients; this correlation has not been reported so far. LGE in the septal location had a significant association with the occurrence of conduction abnormality. The septal location of the proximal conduction system explains this observation. The high incidence of delayed enhancement (94%) in our cohort may indicate delayed recognition of CS in these patients and the advanced stage of disease progression. Temporal delay in recognition of cardiac involvement is likely to have harmful consequences, leading to irreversible cardiac dysfunction, and higher morbidity (heart failure admission, VT recurrence.¹⁶

ICD therapies are frequent in patients with CS. Possible mechanisms of VT can be the inflam-matory nature of the granulomas in sarcoidosis that may lead to increased automatic ventricular activity. The subsequent scar formation is then the substrate for re-entrant ventricular arrhythmias.¹⁷ The rate of appropriate ICD therapies in our cohort of CS is higher than reported rates from primary prevention ICD trials and comparable to those for secondary prevention. In the sudden cardiac death in HF primary prevention trial, appropriate ICD shock was 21%.¹⁸ In the antiarrhythmic versus implantable defibriliators trial for secondary prevention, an appropriate ICD shock was 61%.¹⁹ Reports of ICD therapy rates from other infiltrative cardiomyopathies are also high. ICD therapy rates in arrhythmogenic right ventricular cardiomyopathy range from 24% to 74% at 5 years.²⁰

We observed an improvement in NYHA functional class and ejection fraction on follow-up, likely as a result of goal-directed medical therapy of HF with immunosuppression at the active inflammatory stage of the disease.^21,22 High mortality among patients with refractory HF indicates an advanced stage of the disease, which is refractory to pharmacotherapy. There was no difference in survival between patients who received an ICD shock with those who did not. This observation likely reflects the effectiveness of ICD therapy in preventing sudden cardiac deaths in patient cohorts with high arrhythmia risk and low ejection fraction compared to patients with no ICD shocks who likely have better ejection fraction and low arrhythmic risk.²³ In this cohort, mortality is driven by refractory HF.

The severity of congestive HF is the most powerful prognostic predictor in patients with CS.²¹ Mortality in this cohort is predicted by acute HF at presentation, elevated ESR at the time of diagnosis, and low ejection fraction during follow-up, but not VT. Survival analysis showed that recurrent HF admission predicts early mortality, which probably indicates advanced disease at diagnosis. Mortality association with elevated inflammatory markers (non-specific nature of ESR as an inflammatory marker to be considered) suggests active systemic inflammation may contribute to poor outcomes. This finding emphasises the need for optimal immuno-suppressive and guideline-directed HF therapy²⁴ initiation and maintenance in CS patients with acute HF and high risk of mortality.

Conclusion

Although arrhythmia was the most common manifestation, clinical HF was seen in nearly half of the patients with a diagnosis of CS. A high prevalence of HF, along with 25% mortality, as noted in our study, may indicate a delayed recognition of cardiac involvement in the natural history of these patients. Recurrent HF admissions predicted early mortality.