Transthyretin Amyloid Cardiomyopathy Risk Evaluation in a Cohort of Patients With Heart Failure.
Angie A SuhPaul B ShawMark Y JeongKari L OlsonThomas DelatePublished in: The Permanente journal (2023)
Introduction Transthyretin amyloid cardiomyopathy (ATTR-CM) is a progressive, infiltrative form of heart failure (HF). Nevertheless, ATTR-CM is a largely underrecognized and misdiagnosed condition. This study's objective was to develop an efficient model to assess the chance of ATTR-CM in patients with HF. Methods This was an observational study of patients with HF who had a confirmed diagnosis of ATTR-CM and those with HF but without known ATTR-CM between January 1, 2019, and July 1, 2021. Patient characteristics were extracted from administrative and claims electronic databases and compared between the groups. A propensity score for having ATTR-CM was modeled. Samples of 50 control patients with the highest and lowest propensity scores were adjudicated to assess whether further workup to evaluate for ATTR-CM was warranted for each patient. The sensitivity and specificity of the model were calculated. Results Thirty-one patients with confirmed ATTR-CM and 7620 patients without known ATTR-CM were included in the study. Patients with ATTR-CM were more likely to be Black and to have atrial flutter/fibrillation, cardiomegaly, HF with preserved ejection fraction, pericardial effusion, carpal tunnel syndrome, joint disorders, and lumbar spinal stenosis and to use a diuretic (all p < 0.05). A propensity model with 16 inputs was developed (c-statistic = 0.875). The model's sensitivity and specificity were 71.9% and 95.2%, respectively. Conclusion The propensity model developed in this study provided an efficient means for identifying patients with HF who are more likely to have ATTR-CM and may warrant further workup.
Keyphrases
- ejection fraction
- heart failure
- acute heart failure
- aortic stenosis
- case report
- left ventricular
- machine learning
- multiple sclerosis
- chronic kidney disease
- end stage renal disease
- transcatheter aortic valve replacement
- health insurance
- coronary artery disease
- big data
- peritoneal dialysis
- wild type
- deep learning
- left atrial