Development of TTRACK study diagnostic scores to identify elevated risk of transthyretin amyloid cardiomyopathy in older patients with unexplained hypertrophic cardiomyopathy

Background Using data from older patients with unexplained hypertrophic cardiomyopathy (HCM) who participated in the TTRACK study, we created new transthyretin amyloid cardiomyopathy (ATTR-CM) diagnostic scores to help identify patients at high disease risk who may benefit from further screening. Purpose Describe the development/validation and potential usefulness of these new diagnostic scores in predicting increased ATTR-CM risk in the TTRACK study population. Methods TTRACK study participants were age ≥50 years and had HCM (left ventricular [LV] maximum wall thickness ≥15 mm on echocardiogram [echo]) of unknown aetiology and technetium-99m-DPD/-PYP/-HMDP–labelled bone scintigraphy scans. Patients with no cardiac uptake (grade 0 [Perugini]) on scans and patients with moderate/high uptake (grades 2 and 3) and no monoclonal protein abnormalities (ATTR-CM) were assigned to derivation and validation cohorts. Candidate predictive variables were analysed by univariate logistic regression, with significant variables (P>0.05) entered in a multivariable logistic model. In the derivation cohort, variables independently associated with ATTR-CM were identified using sequential backward elimination; weighted risk prediction scores were built using variable beta estimates in the final model. In the validation cohort, area under the receiver-operating characteristic (AUC ROC) curves were created and optimal high-risk score cutoffs identified based on sensitivity/specificity. Final scores were assessed in the validation cohort. Results Two diagnostic scores were developed in the derivation cohort (grade 0, n=348; ATTR-CM, n=97) and tested in the validation cohort (grade 0, n=173; ATTR-CM, n=47): (1) a 6-variable/6.5-point overall score comprised of age >72 years, male sex, carpal tunnel syndrome (CTS), LV maximum posterior wall thickness >12 mm (echo), relative wall thickness >0.7 (echo), and LV ejection fraction <60%; and (2) a 3-variable/4-point simplified score comprised of age >72 years, male sex, and CTS (Fig. 1). Mean (SD) overall and simple scores were ≥2 times higher in patients with ATTR-CM than in those with no cardiac uptake (Fig. 1). When cutoffs of ≥3 and ≥2 were applied with the overall and simple scores, respectively, 81% and 49% of patients with ATTR-CM were identified as high risk for the disease vs 12% and 4% of patients with no cardiac uptake (Fig. 2). The TTRACK overall high-risk score performed better in identifying patients at risk for ATTR-CM than the Mayo score (AUC, 0.88 [95% CI, 0.81, 0.95] vs 0.78 [0.70, 0.87], respectively; P=0.0002); the TTRACK simple high-risk score was not significantly different than the Mayo score (0.87 [0.81, 0.92] vs (0.78 [0.70, 0.87]; P=0.6391). Conclusion The newly developed TTRACK diagnostic scores, including a simple score with only 3 clinical variables, performed well in identifying ATTR-CM risk in older patients with unexplained HCM, warranting additional investigation.