Integrating biomarkers of hypervolemia into a risk score to predict 1-year outcomes in heart failure patients with preserved ejection fraction
T Aguiar, M Silva, I Cruz, C Costa, S Carvalho, J Ribeiro, J M Bastos, A BriosaAbstract
Introduction
Heart failure (HF) with preserved ejection fraction (HFpEF) encompasses a wide range of phenotypes with different prognostic implications. There is a need for biomarkers and score systems that can identify which patients are at highest risk and require closer follow up and more intensive treatment.
Purpose
To evaluate the independent prognostic value of biomarkers indictive of hypervolemic status, and to create a novel risk score to predict future events.
Methods
We performed a cohort analysis of HFpEF patients admitted to the cardiology ward due to acute/chronic decompensated HF. We selected the following variables to assess the volemic status: 1) high estimated plasma volume status (ePVS), calculated from hematocrit and hemoglobin values, with a cut-off of ≥ 5 ml/g ; 2) hyponatremia, with a cut off of ≤134 mmol; 3) high NTpro-BNP levels, with a cut-off of ≥1000 pg/mL; 4) high blood urea to creatinine (BUN/Creat) ratio with high serum creatinine levels, with a combined cut-off of creatinine ≥1.5 mg/dL and BUN/Creat ≥ 30 mg/dL. A composite endpoint CE of cardiovascular mortality and hospital admissions at 1 year was utilized. With these variables, a novel score system was created and tested against the CE with Kaplan-Meier survival curve and multivariate Cox Regression analyis.
Results
Our cohort included 159 patients with HFpEF with at least one hospital admittance due to HF, of which 60% were male, with a mean age of 77 years old. There was a high frequency of cardiovascular risk factors (CVRF) and co-morbidities. In our analysis, ePVS had the strongest association with the CE (log-rank 4.25, P=0.39); HypoNa and BUN/Creat were also positively associated with the CE (log-rank 3.67, P=0.05 and 3.86, P=0.05, respectively). NTpro-BNP was strongly associated with cardiovascular death alone, but not with the CE. A novel score system was created, where high ePVS, hyponatremia and high BUN/Creat were each awarded 1 point. There was a strong association between this score and the CE (log-rank 10.95, P=0.01), confirmed in a multivariate adjustment for cardiovascular risk factors and comorbidities (hazard ratio 1.35, P=0.01) Interestingly, this score was also strongly associated with cardiovascular mortality alone (log-rank 10.91, P=0.01).
Conclusion
The biomarkers of hypervolemia and the novel scoring system were independent predictors of future cardiac events, and could serve as an effective tool to identify high-risk patients in this population.For image description, please refer to the figure legend and surrounding text.