A multicenter, international, North American study published in the Journal of the American College of Cardiology has indicated that N-terminal pro–B-type natriuretic peptide (NT-proBNP) testing may be valuable in the identification or exclusion of heart failure in emergency department patients with shortness of breath.
Although NT-proBNP testing is widely supported at this point (Class I in the guidelines), all of the data focusing on NT-proBNP testing in acute dyspnea is based on patient cohorts who were enrolled in the early 2000s. Therefore, subsequent changes in demographics and medical backgrounds in patients with heart failure necessitated a more contemporary look. In an interview with Dr. Gibson, the principal investigator, Dr. James Januzzi, a Cardiologist at Massachusetts General hospital stressed on the need for reassessment, saying that “With the rise in heart failure with preserved ejection fraction, the cut-offs that we use for NT-proBNP might have to be decreased (to increase sensitivity).” He also added, “In the aging population, cut-offs may need to be raised.” To determine whether the cutoffs that were developed in the early 2000s are still valid, Januzzi et al. enrolled 1471 patients in 19 sites with a vast array of medical conditions, backgrounds, balanced demographics, including 50 percent women and a substantial percentage of people of color (very characteristic of the United States of America or North America). Of those patients, about 20 percent had acute heart failure, adjudicated by the committee that reviewed all the cases. The gold standard was an adjudicated diagnosis based on the clinical presentation but blinded to the NT-proBNP.
[perfectpullquote align=”full” bordertop=”false” cite=”” link=”” color=”” class=”” size=””]“BNP is certainly part of the story, just as troponin is part of the story.”- Dr. James Januzzi, M.D.[/perfectpullquote]
In the study, the investigators found that the marker was highly discriminatory in the identification or exclusion of heart failure. The area under the receiver operator characteristic (ROC) curve was 0.91, which was remarkably high for an all commerce emergency department population. Commenting on these findings, Dr. Januzzi remarked, “We were very careful of not restricting our inclusion and exclusion criteria to have an overly optimistic experience and yet, the results were really smashing.” There was a very high negative predictive value of almost 98 percent for the rule-out cutoff of 300 pg/ml. When patients were categorized by age <50, 50 to 75, and >75 years, NT-proBNP cutoff levels of 450, 900, and 1,800 pg/ml had high sensitivity and specificity. The focus was also on likelihood ratios because “Only one-fifth of the patients in the study actually had acute heart failure, so the prevalence was relatively low,” Dr. Januzzi said. This surprising finding could have been suggestive of the fact that patients with heart failure were more likely to seek sources of acute care for their management such as office-based intravenous diuretics, instead of going to the emergency department. However, the positive and the negative likelihood ratios of 5.99 and 0.09, used to identify or exclude heart failure based on NT-proBNP, were superior to any other study that has been performed in this space.
Dr. Januzzi reinforced the importance of the study, saying that the focus was not on a single number but on validating prior experience using age-stratified cut offs to verify and exclude heart failure. Not only was the study a confirmation of the cut off strategy, but an affirmation of the fact that NT pro-BNP could remain highly useful in the evaluation of a contemporary population of patients. Therefore, it was reasonable to expect that these cutoffs would receive regulatory approval for global use. However, he emphasized that this diagnostic tool was merely a support to clinical judgment, which needed to be integrated with a balanced history and physical examination. He was hopeful that in the future, discordance of values with clinical behavior could be minimized with the advent of new ways to integrate diagnostics with treatment strategies like artificial intelligence and machine learning. “BNP is certainly part of the story, just as troponin is part of the story,” he concluded.