Meta-analysis on the immunohistological detection of inflammatory cardiomyopathy in endomyocardial biopsies

Abstract

Meta-analysis on immunohistological (IHC) concepts for the detection of inflammatory cardiomyopathy (InfCM) in endomyocardial biopsies (EMB). We included 61 publications, with 10,491 patients (mean age 47.1\xa0years; men 66%) who underwent EMB and IHC evaluation. The 460 control patients were devoid of IHC proof of InfCM. The mean IHC detection rate of InfCM was 50.8% (95% CI 47.7–53.8%; range 18.4–91.7%). A publication bias was excluded (Funnel Plot p \u2009=\u20090.4264). This IHC detection rate was significantly ( p \u2009<\u20090.0001) higher compared to the histological detection of myocarditis according to the Dallas criteria (mean 8.04%; 95% CI 5.08–12.5%; subset of 3274 patients in 30 publications). However, 13 different diagnostic IHC criteria were described in the publications, with various thresholds of diverse phenotypes of quantified infiltrates, and endothelial expression of diverse cell adhesion molecules (CAM), quantified either visually or by digital image analysis (DIA). The comparison of IHC with cardiac magnetic resonance (CMR) data available in a subset of 13 publications with 1185 patients revealed a sensitivity for CMR of 69% (95% CI 58–79%), a specificity of 73% (95% CI 59–84%), and a ROC-AUC of 0.77 (95% CI 0.73–0.81). This meta-analysis encompassing 10,491 patients confirms a mean detection rate of InfCM in 50.8% of EMB, being significantly more sensitive compared to the histological Dallas criteria. IHC cannot be fully substituted by CMR. However, standardization of the diverse IHC markers and protocols seems pertinent, especially considering the published adverse prognostic impact of IHC-confirmed InfCM and its published suitability for the selection of candidates responding favorably to immunosuppression.