Log in to MyACC
Validation of a Protein Risk Score for Mortality in HF
Quick Takes
- A protein risk score using 38 proteins identified through proteomic analysis was successfully developed to predict 5-year mortality in a community-based cohort of patients with HF.
- The protein risk score performed better compared to the MAGGIC risk score and NT-proBNP levels.
Study Questions:
Can a protein risk score be developed and validated to predict risk for mortality in patients with heart failure (HF)?
Methods:
This was a community-based (Southeast Minnesota) cohort study of patients with HF enrolled between 2003 and 2012. A total of 1,351 patients were followed. In these patients, 7,289 plasma proteins were measured (high throughput proteomics). Patients enrolled between 2003 and 2007 were part of the protein risk score development cohort and patients enrolled between 2008 and 2012 were part of the validation cohort.
The protein risk score was used to assess 5-year mortality. This was compared against a clinical model that included the MAGGIC (Meta-Analysis Global Group in Chronic Heart Failure) risk score and N-terminal pro–B-type natriuretic peptide (NT-proBNP) levels across a range of predicted mortality risk groups (≤25%, 26-50%, 51-75%, >75%).
Results:
In the overall cohort, 48% were women, median age was 78 years, and 31% had a left ventricular ejection fraction (LVEF) <40%. A total of 1,013 deaths occurred during follow-up (followed through 2021), leading to a 5-year mortality rate of 52.1% (95% confidence interval [CI], 49.3-54.7%). There were 855 and 496 patients in the development and validation cohorts, respectively. The baseline characteristics of the two cohorts were in general similar.
A total of 38 proteins were identified in the development cohort as being independent predictors of mortality. After adjusting for the MAGGIC score, a 1 standard deviation increase in protein risk score was associated with an increased risk of death in both cohorts (development: hazard ratio [HR] 2.62, 95% confidence interval [CI] 2.34-2.93; validation: HR 2.01, 95% CI 1.75-2.32). In the validation cohort, the protein risk score was associated with cardiovascular mortality in the crude model (HR 1.55, 95% CI 1.34-1.79), though not significantly associated after adjusting for MAGGIC score (HR 1.12, 95% CI 0.93-1.35).
The protein risk score was well calibrated with estimated to observed mortality ratio of 1.01 (95% CI 0.92-1.10), performing better than the clinical model (E/O 1.11, 95% CI 1.02-1.19), particularly at the low- and high-risk groups. When the protein risk score was added to the clinical model, more patients were able to be re-classified to either the low or high risk for mortality groups.
Conclusions:
The development of a protein risk score using 38 proteins was successfully completed with a community-based cohort. Validation of the score demonstrated good calibration and helped stratify mortality risk in patients with HF.
Perspective:
Estimated risk of death for HF patients is important, as it may lead to an alteration in the therapeutic plan and may signal the possible need for advanced HF therapies. Current prognostic indicators (risk models, biomarkers, etc.) each have limitations but can be considered as part of the holistic assessment of the patient’s status. Novel, personalized, and potentially more accurate methods for risk assessment are needed to help advance HF care further. The authors here used a proteomics analysis approach to address this need and demonstrated a protein risk score was well calibrated when assessing mortality risk and performed better than the MAGGIC score and NT-proBNP. Specifically, this score helped to re-classify patients at high and low risk. This opens the possibility of using such a score in future practice. However, future studies will need to address key issues of generalizability in diverse cohorts and ability to inform and influence clinical therapies and outcomes. Also, how this protein risk score compares to functional testing such as cardiopulmonary exercise testing in predicting outcomes in HF is to be determined.
Clinical Topics: Heart Failure and Cardiomyopathies, Acute Heart Failure, Heart Failure and Cardiac Biomarkers
Keywords: Heart Failure, Risk Assessment
< Back to Listings
