Improving Diagnostic Accuracy for Acute Heart Failure

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT Acute heart failure (HF) is highly morbid, lethal, and costly. It is a difficult diagnosis to make given its symptoms and signs overlap with other cardiac and non-cardiac conditions. In the emergency department (ED), misdiagnosis of acute HF is common and associated with adverse outcomes. Biomarker testing can facilitate accurate diagnosis; however, natriuretic peptides (NP) are the only guideline recommend biomarker of HF for diagnostic testing, and are better for ruling-out, rather than ruling-in, acute HF. Even with NP testing, in contemporary clinical practice misdiagnosis of acute HF still occurs in 10 to 45% of patients presenting to the ED with dyspnea. Clinical prediction models including multiple biomarkers hold promise for improving diagnostic accuracy. The few prior studies investigating a multiple biomarker approach for diagnosing acute HF were limited by constraint to highly correlated markers from known biologic pathways, relatively small sample sizes, lack of inclusion of all a priori selected biomarkers into a single model, and absence of validation cohorts. Our study design addresses these limitations. Recent advances in “omics” enable novel biomarker discovery on a larger scale and investigations less “biased” by existing knowledge. Thus, our overarching hypothesis is a multi-marker model incorporating novel proteins discovered with plasma proteomics improves diagnostic accuracy for acute HF. In preliminary work, we performed a proof of concept study utilizing plasma proteomics to discover a multi-marker panel of 21 biomarkers which improved diagnostic accuracy for acute HF beyond current clinical practice using clinical data and NP levels. Our promising preliminary data motivate broader discovery in a larger sample size with subsequent derivation and validation of a multi-marker model for diagnosing acute HF in independent samples of adequate size. Our specific aims are to: 1) expand the discovery cohort and refine the multi-marker panel of 21 biomarkers to improve diagnostic accuracy for acute HF, 2) derive a model for diagnosing acute HF incorporating the 21-biomarker panel, 3) test performance of the multi-marker model in a prospective validation cohort, and 4) assess the incremental value of the multi- marker model for diagnosing acute HF. In aim 1, existing plasma samples from 989 patients will be used to assay 925 proteins to discover a smaller set of novel biomarkers most strongly associated with an adjudicated acute HF diagnosis. In aim 2, we will utilize an existing prospective observational cohort, EMROC-AHF, to derive the multi-marker model in 1,000 patients who presented to the ED with acute dyspnea. In aim 3, from four EDs in Detroit, MI and Nashville, TN we will prospectively recruit a new sample of 1,000 patients presenting with acute dyspnea and adjudicate the presence of acute HF by cardiologist panel review. In aim 4, we will compare our multi-marker model against the current clinical approach for diagnosing acute HF using the cohorts for Aims 2 and 3. Given the burden of HF, the frequency of inaccurate diagnosis and its adverse consequences, we will address a significant unmet need by improving diagnostic accuracy for acute HF.