Comparing Screening Methods for Heart Failure in Patients with Diabetes, with Mohammad Shahzeb Khan, MD

At the 9th Annual Heart in Diabetes Conference in Philadelphia, PA, Mohammad Shahzeb Khan, MD, an assistant professor of medicine at Baylor College of Medicine, presented the results from his pooled cohort analysis comparing methods of measuring risk of heart failure (HF) in patients with type 2 diabetes without atherosclerotic cardiovascular disease (ASCVD).

A variety of screening techniques for determining HF risk in patients with diabetes have been developed, from echocardiography to clinical risk scores and cardiac biomarkers. However, no direct comparison of screening approaches, alone or in combination with each other, has been conducted. Therefore, despite the rapidly increasing prevalence of HF among patients with diabetes, the optimal screening strategy to predict the disease in a primary prevention population is unknown.1

To that end, Khan and colleagues collected deidentified data from various previous studies, including the Atherosclerosis Risk in Communities (ARIC) and Chronic Renal Insufficiency Cohort (CRIC) studies. These cohorts were collectively comprised of 6293 patients – 77.7% of these did not have prevalent ASCVD and were evaluated for HF.1

Existing screening strategies were applied among patients without a history of ASCVD, divided up between 1-step and 2-step screenings. The former group included echocardiography, troponin assessments, B-type natriuretic peptide (BNP) assessments, and so on. The latter group involved combining N-terminal pro-BNP (NT-proBNP) screening with the other named methods.1

“What we found [that] was really interesting is that if we’re using 1-step screening strategies like NT-proBNP, we are missing a lot of heart failure events,” Khan told HCPLive. “For example, if we were only using NT-proBNP as a screening tool, we would miss almost 30-40% of the heart failure events that were occurring in low risk NT-proBNP growth, meaning that even if the NT-proBNP is less than 125, a decent number of events were occurring in that patient population.”

Khan noted that combination screening, such as with NT-proBNP and the WATCH-DM score, resulted in a much lower percentage of patients missed between all 6 included cohorts.

“We also studied 2-step screening strategies, in which we used a risk score like WATCH-DM and NT-proBNP and we found that when we combined two approaches together, whether it be a biomarker and risk score, the number of patients that were left out were significantly less,” Khan said. “There were only 10-15% of patients that were missing. Indirectly, we’re predicting 85% of the heart failure events.”

Looking ahead, Khan discussed potential issues with the study’s generalizability and scope.

“We took 6 cohorts, we had 6000 patients, so there’s obviously a generalizability issue. We might not be able to generalize these findings to different continents like Asia and Africa, because patient pools are very different across the globe,” Khan said. “We also did not have a proper adjudication for heart failure incidents. It was mainly heart failure hospitalization, so we might have missed diagnoses of heart failure which were done as an outpatient.”

Khan does note, however, that the results from this trial are sufficient to indicate that 1-step screening is no longer sufficient to determine heart failure risk in patients with type 2 diabetes on a larger scale.

“The most important takeaway for this is that using 1-step screening strategies might not be enough to screen for patients with heart failure,” Khan told HCPLive. “We’ll have to use a layered strategy, at least 2-step screening in which, for earlier stratification, you can use a risk score like WATCH-DM and then use a biomarker selectively.”

Khan reports the following disclosures: Bayer

References

1. Patel KV, Segar MW, Klonoff DC, et al. Optimal Screening for Predicting and Preventing the Risk of Heart Failure Among Adults With Diabetes Without Atherosclerotic Cardiovascular Disease: A Pooled Cohort Analysis. Circulation. 2024;149(4):293-304. doi:10.1161/CIRCULATIONAHA.123.067530