A-007 Troponin I and T have similar prognostic performance for mortality in cardiac amyloidosis

Abstract Background Cardiac amyloidosis is a progressive, infiltrative cardiomyopathy that results in thickened ventricular walls and reduction in myocardial function. Most cases involve Light Chain Amyloidosis (AL) caused by deposition of free light chains (FLC). Previous studies have demonstrated that troponin T (cTnT), N-terminal proBNP (NT-proBNP) and FLC best predict survival in AL patients and have been incorporated into staging systems to adjudicate prognosis. Limited data is available comparing cTnT relative to cardiac troponin I (cTnI) for prognostication. The purpose of this study was to compare the utility of cTnI vs cTnT for prognostication of mortality in patients with AL. Methods This prospective observational study enrolled patients with a diagnosis of AL and physician ordered cTnT. Remanent specimens were collected and tested for high sensitivity cTnI on an Abbott Alinity i. Chart review was performed to assess for mortality at 24 months. Thresholds were established for cTnI using area under the receiver the operator characteristic curves (auROC), and positive predictive value (PPV) and negative predictive value (NPV) were calculated. Mayo staging scores were used for prognosis; patients were assigned a point each if cTnT was =25 ng/L, NT-proBNP =1800 pg/mL, and if FLC difference =18ng/dL. Survival curves were generated for using the Mayo score and a novel score using cTnI. Results There were 212 specimens from 96 patients with AL and two years of follow-up, of which 14 patients died. Using the first result from each patient, the median cTnT was 101 ng/L (IQR: -65182) in those that died and 31ng/L (16-50) in those that survived. Median cTnI in patients that died was 130ng/L (44-197) and 10ng/L (5-24) for those that died. The spearman r for cTnT and cTnI was 0.76 (95% CI; 0.66-0.84). The auROC for cTnT for mortality was 0.86 (0.79-0.4) and for cTnI was 0.87 (0.75-0.99). The threshold of 25 ng/L for cTnT had a sensitivity of 100% (78.5-100.0), a specificity of 44.1% (37.1-51.4), a PPV of 28.3% and a NPV of 100%. A threshold of 35 ng/L best identified patients at higher risk for mortality for cTnI with a sensitivity of 92.8% (68.5-99.6), a specificity of 84.8% (75.3-91.1), a PPV of 52.0% and NPV of 97.1%. One patient that died had a cTnI of 4 ng/L and all other 13 patients were =35ng/L. Implementation of cTnI (using 35ng/L as a threshold) into survival curves resulted in no significant difference in performance relative to cTnT based scores. For cTnI scoring, the mortality rate for stage I was: 4/114 (3.5%), Stage II: 3/38 (7.8%), Stage IIIa: 13/53 (24.5%), Stage IIIb: 2/6 (33.3%). For cTnT based scoring, the mortality rate for Stage I: 0/74 (0%), Stage II: 8/56 (14.3%), Stage IIIa: 13/71 (18.6%), Stage IIIb: 2/11 (18.2%). Conclusion cTnT and cTnI have similar performance for predicting mortality in patients with AL. The cTnT threshold used for staging maximizes the NPV but has low PPV for mortality. A higher threshold for cTnI of 35 ng/L performs similarly to cTnT for staging AL and better separates stages II from IIIa and IIIb.