Background: Historically, the diagnosis of cardiac sarcoidosis (CS) in the presence of extracardiac sarcoidosis has been challenging. In 2014, the HRS CS Expert Consensus Statement on diagnosing CS in the presence of proven extracardiac sarcoidosis was published. However, the diagnostic accuracy of this algorithm is unknown.
Objective: To determine the specificity, sensitivity and accuracy of the HRS CS Expert Consensus Statement using a cohort of patients referred to a specialized CS clinic.
Methods: 538 patients with suspected CS were referred to a specialized CS clinic for evaluation for the presence of CS. Patient demographics, symptoms, and investigations were collected retrospectively. The accuracy of the 2014 HRS algorithm was assessed by determining the sensitivity, specificity, accuracy, and positive and negative predictive value using our cohort. A revised algorithm was developed based on the approach used in our CS clinic.
Results: Of the 538 patients assessed, 115 had pulmonary sarcoidosis and 58 had other organ involvement. 97 patients were diagnosed with having CS using endomyocardial biopsy or advanced cardiac imaging whereas in 441 patients CS was ruled out after consult, biopsy, or advanced cardiac imaging. The sensitivity of the HRS algorithm using our cohort was 70% (95% CI 50.6-85.3) and the specificity was 37.5% (95% CI 24.0 - 52.6). To calculate positive and negative predictive values, a prevalence of 25% was assumed. The positive predictive value of the algorithm is 27.2% (95% CI 21.3 - 34.0), whereas the negative predictive value is 78.9% (95% CI 66.0 - 88.0). The overall probability (accuracy) that a patient would have been correctly diagnosed is 45.6% (95% CI 34.3 - 57.3).
Conclusion: The 2014 HRS Consensus Statement for the diagnosis of CS has modest sensitivity and NPV and poor specificity. Performing advanced cardiac imaging in asymptomatic patients without ECG or echo abnormalities can increase the detection rate of CS. These findings suggest that the current HRS algorithm should be updated.