Abstract Background Definite diagnosis of cardiac sarcoidosis (CS) requires proof of sarcoid granulomas in the heart. Endomyocardial biopsy (EMB) is considered a risky procedure with poor sensitivity (<25%) in CS (1), although comprehensive studies on its diagnostic performance are not available. Purpose We investigated the sensitivity, complications, and prognostic significance of EMB in a large cohort of patients with CS to help to choose diagnostic strategy when the disease is suspected. Methods We analysed the data of 260 consecutive patients diagnosed with CS in 1988-2022 at our institution. All met the diagnostic criteria of the Heart Rhythm Society (1). The use, findings, and complications of EMB were retrospectively noted in addition to patients’ demographics, presenting phenotype, diagnostic examinations, and future serious cardiac events. The data were retrieved from hospital records and an ongoing CS registry (2). Advanced imaging studies (cardiac magnetic resonance, positron emission tomography) were re-analysed and the follow-up information was updated until June 2023. EMB’s performance was assessed also in 30 cardiac transplant recipients having CS at the histopathologic study of the explanted heart. Results Of the 260 patients (mean age 49, 60% females), 216 (83%) underwent diagnostic EMB, 48 with repeat procedures. The sensitivity of EMB was 38%, rising to 49% after repeats. The predictors of positive EMB (Table 1) included the presenting phenotype and characteristics of the activity, extent, and location of myocardial involvement. Presentation with ventricular tachyarrhythmia, left ventricular (LV) ejection fraction ≤45%, elevation of cardiac troponins, and presence of middle or apical LV septal late gadolinium enhancement on magnetic resonance imaging were independent predictors (p<0.05) of positive biopsy. The sensitivity of EMB was directly related to the count of the predictors present (Figure 1). The rate of procedural complications was 9.7% overall and 0.7% for major events. One pericardial effusion needed drainage, but no deaths or long-term sequels followed the biopsies. Minor complications included 10 paroxysms of ventricular tachycardia and 6 small pericardial effusions. The 10-year rate (95% CI) of the composite of cardiac death, end-stage heart failure, or ventricular tachyarrhythmia was 55% (44-67%) with positive EMB vs 29% (17-44%) with negative EMB (p<0.001). When adjusted for the presenting phenotype and LV ejection fraction, EMB did not predict outcome events. In the 30 patients with CS in explanted hearts, the sensitivity of EMB, including the repeats, was 60%. Conclusion The sensitivity of EMB is better than usually presented in CS and the higher the more extensive myocardial involvement is. Risk of serious complications is <1%. In patients with suspect CS, the pre-test likelihood and value of positive EMB should be weighed against the procedural risks in shared decision-making when choosing the diagnostic pathway.