Cardiomyocyte stretch mediates the relation between left ventricular amyloid burden and adverse outcomes in light chain amyloidosis: a 18F-florbetapir positron emission tomography study

Abstract

Abstract Background Patients with light chain (AL) amyloidosis and cardiac involvement have poor prognosis. Mayo stage accounts for severity of plasma cell dyscrasia and cardiac biomarker release, and provides powerful risk stratification. Myocardial amyloid burden can be quantified by 18F-florbetapir positron emission tomography (PET), but its prognostic value is not known. Purpose To test our hypothesis that (1) myocardial amyloid burden predicts adverse outcomes and (2) the relationship between amyloid burden and adverse outcomes is mediated by cardiomyocyte stretch and injury. Amyloid burden was estimated by left ventricular 18F-florbetapir retention index (RI) and cardiomyocyte stretch and injury by NT proBNP and troponin T respectively. Methods We performed 18F-florbetapir PET (median dose 9.05 mCi) in prospectively enrolled subjects with newly diagnosed AL amyloidosis with abnormal cardiac biomarkers or with normal cardiac biomarkers and normal left ventricular wall thickness (NCT02641145). Left ventricular RI was calculated as the activity concentration between 10 and 30 min. after injection divided by the integral of the left atrial blood time-activity curve from 0 to 20 min. RI was categorized as normal (<0.06/min, based on controls), increased (0.06–0.12/min), or high risk (>0.12/min, based on log-rank statistic maximization). Mayo stages I–IV were based on elevated serum cardiac biomarkers: NT-proBNP ≥1800 pg/ml, troponin T ≥0.025 ng/ml, and difference in free light chains ≥180 mg/l. Adverse outcomes of all-cause death or heart failure hospitalization were evaluated. Survival analysis was performed using Kaplan-Meier and Cox regression including Mayo stage and RI. Mediation analysis was used to elucidate the role of cardiomyocyte stretch (as NT-proBNP) and injury (as troponin T) in the association between amyloid burden estimated by RI and adverse outcomes. Results We studied 80 subjects with median age 62 years (IQR 57–67), 46 men (57%), 60 with abnormal cardiac biomarkers (75%), and median RI of 0.10/min (IQR 0.06–0.16). At follow-up (median 15 months), adverse outcomes occurred in 34 subjects (42%), with 17 deaths (21%) and 23 heart failure hospitalizations (29%). The incidence of adverse outcomes increased across Mayo stages from 9% to 44% (log-rank p<0.001), and across RI levels from 29% to 57% (log-rank p=0.037, Figure 1). In multivariable Cox regression, only Mayo stage independently predicted adverse outcomes (HR 2.0 [95% CI 1.4–3.0], p<0.001). Multivariable mediation analysis showed that 83% of the association between RI and adverse outcomes was mediated by NT-proBNP (p<0.001, Figure 2), without contribution from troponin T. Conclusion Myocardial amyloid burden estimated by F-18 florbetapir RI predicts adverse outcomes in AL amyloidosis, but not independently of Mayo stage. Cardiomyocyte stretch mediates the relationship between myocardial amyloid burden and adverse outcomes in AL amyloidosis. Funding Acknowledgement Type of funding sources: Private company.