NT-Probnp and High-Sensitivity Troponin T Fail to Detect Cardiac Involvement in Erdheim-Chester Disease

Introduction Erdheim-Chester disease (ECD) is a rare non-Langerhans cell histiocytosis belonging to the L group of the 2016 revised histiocytosis classification¹. Although frequent and associated with clinical complications, cardiac involvement is underdiagnosed in ECD². Cardiac magnetic resonance (CMR) imaging is the most robust way to detect cardiac involvement³. However, access to this technique is limited. In other infiltrative diseases, such as amyloidosis and Fabry disease, cardiac biomarkers have been shown to be highly sensitive and specific for both diagnosis and the prediction of prognosis^4,5. We evaluated the utility of B-type natriuretic peptides and troponin for the diagnosis of cardiac involvement in ECD.

Methods We retrospectively included patients with a biopsy-proven diagnosis of ECD who were referred to the internal medicine department of a French tertiary care center or the hematology department of an Israeli tertiary care center, and that had undergone both a CMR and an assessment of B-type natriuretic peptide and troponin levels, at least once, between 2007 and 2019.

CMR was performed on a 1.5 T MR scanner (Siemens Aera). All images were re-read in a blind fashion by an experienced radiologist (M.B). Cardiac involvement was defined as abnormal epicardial infiltration or pericardial effusion, enhancement, or infiltration. Cardiac biomarkers levels were determined at the physician’s discretion, in the routine clinical care setting. Natriuretic peptide test assessment was performed using N-terminal-pro-brain natriuretic peptide (NT-proBNP) at the French center (normal < 300 pg/mL), and brain natriuretic peptide (BNP) at the Israeli center (normal < 100 pg/mL). Troponin assessment was performed using high-sensitivity cardiac troponin (hs-cTnT) at the French center (normal < 14 µg/L), and troponin I (TnI) at the Israeli center (normal < 0.04 µg/L). Only NT-proBNP and hs-cTnT were considered in quantitative analysis.

Results We included 122 patients in total (118 in France, 4 in Israel), with a mean age of 58.7 years (± 13.9 years). Biological assessment was performed a median of 12 months (1-33 months) after cardiac imaging. ECD-related cardiac involvement was present in 57 patients (46.7%).

NT-pro-BNP/BNP levels were high in 38 (32.8%) patients. High levels of NT-proBNP/BNP were not significantly associated with any of ECD-related CMR features (Table 1). Median NT-proBNP concentration was 173 [73-470] pg/mL (normal < 300 pg/mL). Median NT-pro-BNP levels were similar between patients with and without right-atrioventricular sulcus infiltration (195 vs. 154 pg/mL, P=.3, Figure 1, A). Area under the curve (AUC) was 0.68 for cardiac involvement detection using NT-proBNP (Figure 1, C). With an optimal threshold of 114 pg/mL, NT-proBNP diagnostic performances were: sensitivity: 0.79, specificity: 0.58, accuracy: 0.68, negative predictive value: 0.76 and positive predictive value: 0.62.

Troponin levels were high in 39 patients (32%). High troponin levels were not significantly associated with any of ECD-related CMR features (Table). Median hs-cTnT concentration was 9 [6-16] µg/L (normal < 14 µg/L). Median hs-cTnT levels were similar between patients with and without right-atrioventricular sulcus infiltration (10.3 vs. 9 µg/L, P=.3, Figure 1, B). AUC was 0.64 for cardiac involvement detection using hs-cTnT (Figure 1, D). With an optimal threshold of 4.8 µg/L, hs-cTnT diagnostic performances were: sensitivity: 0.96, specificity: 0.25, accuracy: 0.59, negative predictive value: 0.88 and positive predictive value: 0.54.

Conclusion NT-proBNP and hs-cTnT concentrations are not a reliable surrogate for cardiac involvement in ECD.