Red Flag Screening Tool May Facilitate Earlier Detection of Transthyretin Amyloid Cardiomyopathy

An automatic screening tool that can be integrated into electronic health records (EHRs) may prompt providers to elevate their index of suspicion for transthyretin amyloid cardiomyopathy (ATTR-CM) in patients presenting with heart failure, leading to earlier initiation of diagnostic testing and therapeutic interventions in this population, according to an analysis presented at the 2026 Annual Meeting of the American Association of Heart Failure Nurses in San Diego, California. 

Transthyretin amyloid cardiomyopathy is a progressive form of cardiomyopathy that is often misdiagnosed because its symptoms overlap with common heart conditions such as ischemic heart disease or hypertrophic cardiomyopathy. Moreover, many patients living with ATTR-CM have extracardiac manifestations in addition to cardiac “red flags,” which may further complicate the diagnostic process. A lack of awareness among providers and the limited number of specialty clinics that are equipped to diagnose and treat this relatively rare condition also contribute to its underdiagnosis. 

While a diagnosis of ATTR-CM relies on specialized testing that is often conducted in specialty clinics, initiating the workup in the hospital setting, during acute heart failure episodes, could give clinicians a headstart in identifying individuals at risk for disease progression. “Heart failure nurses have an opportunity to support and drive increased identification of ATTR-CM,” said lead author Natasha Zmitrowitz, MSN-ED, RN, coordinator of the Heart Failure Program at the SUNY Upstate Medical University Hospital, in Syracuse, New York. “The goal is to initiate the diagnostic workup prior to hospital discharge, to improve outcomes for this patient population.” 

Zmitrowitz and colleagues developed an EHR-embedded screening tool that can be used by nurses caring for patients with heart failure in the hospital setting to identify those at risk for cardiac amyloidosis. “We built this in our institution in the Epic EHR,” Zmitrowitz explained.  “We were able to develop a patient column [that displays] amyloid red flags and when you hover over the number, it bolds out every single red flag that Epic is picking up. Of course, there are limitations because not everybody uses this EHR the same way, but at least we can show a screenshot to the hospitalists taking care of our patients and say, this patient has nine different red flags, I think we need to start the workup. By the time the laboratory results [are available], they go to the outpatient [provider] who can look at those results and decide whether or not they want to do genetic testing or get a nuclear scan.” 

The automatic screening tool displays green, yellow, or red flags indicating the number of amyloid-related criteria identified in a patient’s EHR, helping to identify individuals at risk for ATTR-CM. Amyloid red flags include a history of heart failure, bradycardia, orthostatic hypotension, neuropathy, age over 50 years, and preserved ejection fraction (EF > 40%), among other factors. The screening tool assigns clinical significance to isolated information that already exists in the EHR, helping clinicians performing the initial screening to identify risk through visual prompts. In addition to identifying the patient’s risk factors for ATTR-CM, the screening tool can be used to communicate the high suspicion scores to all members of the multidisciplinary clinical team. 

Nurses working in acute heart failure management can drive the early diagnosis of ATTR-CM by raising suspicion with the use of the red flag screening tool and collaborating with outpatient clinicians to continue the workup and confirm the diagnosis, Zmitrowitz said. “Back in 2024, we probably should have had 120 to 240 diagnoses [of ATTR-CM], so we knew we needed to do something [to address this gap],” she added. “Since we started using this tool, with the clinical expertise of our outpatient [collaborator], we have almost 50 patients diagnosed since last year, so we are starting to ramp up our numbers and get patients on treatment.” 

This preliminary research may represent a stepping stone for wider applications that could help hospitalists identify patients who are at risk for cardiac amyloidosis. “This has now turned into a fully validated tool,” Zmitrowitz added. “We started with a linear model and put it though machine learning, and now we use either a decision tree or random forest model, which [achieves] a mid-80s [accuracy range] for positivity for diagnosis. It has come a long way since we started.”