Heart failure (HF) is a global public health concern, especially in countries with aging populations where diagnosis, treatment, and prevention of re-hospitalization are challenging [1]. Unfortunately, repeated hospitalization exerts a huge cost on national healthcare budgets and diminishes quality of life [2]. It is important to predict a patient’s clinical course as early as possible by selecting evidence-based management strategies that improve the care of patients with acute HF. Risk stratification in these patients is challenging [3].The use of natriuretic peptides in the diagnosis of acute HF is well established [4]. Moreover, a peptide that correlates with intra-cardiac pressure is expected to reflect prognosis and predict both short- and long-term outcomes of HF [5]. According to the guidelines of the European Society of Cardiology (ESC) and American College of Cardiology Foundation/American Heart Association (ACCF/AHA), the biomarker B-type natriuretic peptide (BNP) is the most reliable tool for diagnosing HF, establishing prognosis or disease severity, and guiding treatment planning by indicating the need for treatment intensification [6, 7]. Natriuretic peptides such as BNP are counter-regulatory hormones involved in volume homeostasis and cardiovascular remodeling; they have also become promising cardiac markers in various HF settings, especially since the advent of rapid assays [8, 9]. Serial changes in BNP can act as surrogate markers in patients with progressive HF, and these values give incremental prognostic information and help track therapeutic response [10, 11]. A change in BNP during hospitalization appears to predict outcomes in patients with acute decompensated HF (ADHF), providing incremental prognostic value beyond baseline measurement [12]. Also, understanding BNP levels through the early period after discharge may help expedite treatment initiation and allow lower-risk patients to transit more rapidly through the system. However, most studies have evaluated patients only during hospitalization [13, 14]. Variation in BNP level during short-term follow-up is associated with both the degree of congestive status and clinical outcomes [15]. Therefore, this study aimed to evaluate the prognostic value of short-term follow-up BNP and the change in BNP after discharge for predicting mortality of hospitalized patients with HF.

We found that both short-term follow-up BNP after discharge and percent change in BNP between admission and follow-up were powerful prognostic markers of mortality in hospitalized patients with HF. In contrast, BNP at admission was not associated with mortality risk. Furthermore, we observed that the high median of follow-up BNP after discharge and percent change in BNP was associated with higher mortality than the below median. This association persisted after adjustment for potential confounders. BNP after discharge and percent change in BNP were independent predictors of mortality.

Many studies have found that BNP level at admission provides useful information for the diagnosis of HF, the assessment of HF severity, and the prognosis of high-risk patients during the early follow-up period [10, 18]. One study that used data from the KorAHF registry found a difference in risk stratification pattern, especially in patients with low NT-proBNP levels [19]. In general, patients with low BNP usually have better prognoses than those with high BNP [19].

Patients hospitalized for HF who are identified as high risk could be scheduled for earlier outpatient visits with a physician and may require follow-up appointments for better outcomes [20]. As expected, BNP concentration 1 month after hospital discharge was the strongest prognostic factor. Another study suggested that HF patients with the highest percentage decrease in BNP 4 months after discharge have the lowest mortality and morbidity [5]. However, no randomized controlled trials have tested whether such data can be used to improve outcomes or have determined the best time for BNP follow-up [18]. In the current study, we found the best cut-off value for monitoring the prognosis of HF to be 294.01 pg/ml for short-term follow-up BNP (within 3 months of discharge; median, 22 days). Median BNP level decreased from 816.5 pg/ml to 369.7 pg/ml (median percent change in BNP, −52.2%) over the 3-month follow-up period. Moreover, short-term follow-up BNP after discharge and percent change in BNP from admission to discharge were associated with mortality. These data may motivate clinicians to measure BNP after discharge to monitor patient outcomes [21]. In Asian countries, including Korea, China, and Japan, the prevalence of HF in the population has increased due to aging of the population and adoption of a Western lifestyle [22]. A few studies have been performed in Asian populations and have revealed that BNP at admission is the most powerful prognostic factor of mortality in patients with HF, and that high levels of BNP are significantly associated with poor outcomes [19, 23]. Most of the studies were designed to investigate the significance of admission BNP for HF patients. Very few studies have explored short-term follow-up BNP. In this study, we mainly focused on Korean HF patients and the effects of short term (90-day) follow-up BNP on prognosis. We also compared the prognostic value of admission BNP, follow-up BNP and percent change in BNP. We measured BNP at admission and at the first follow-up visit to the outpatient clinic after discharge. During the 90-day short term follow-up period after discharge, many patients were lost to follow-up. Thus, unfortunately, our study population is small. However, we analyzed BNP at admission and after discharge and the percent change in BNP and found that the high median for both follow-up BNP and percent change in BNP after short-term follow-up were associated with greater mortality than the below median. We did not find any association between BNP at admission and mortality. After adjustment for covariates of age, sex, clinical characteristics, Hb, Na, and echocardiographic measurement, short-term BNP after discharge and percent change in BNP still predicted mortality. In previous studies, various factors have limited the standard time for evaluating plasma BNP in the prediction of long-term outcomes after discharge [24]. A number of studies have been conducted at single centers or with small sample sizes, limiting the evaluation of the relationship between BNP and other important clinical characteristics; other studies have had variable follow-up times and heterogeneous endpoints [24]. Among those studies that explored the prognostic value of BNP at admission and after discharge, several have also concluded that measurement of BNP at discharge was a strong prognostic marker in HF [25, 26]. One well-designed prospective cohort study of patients with acute HF found that the pre-discharge BNP and percent change in BNP provided conclusive prognostic information and predicted mortality and readmission [27]. Another large, single-center study revealed that high pre-discharge BNP is the most significant predictor of 6-month outcomes and is a more relevant predictor than change in BNP during acute care [2]. However, those studies had small sample sizes, which limited their ability to adjust for covariates [2]. Other studies have found that the percent reduction in natriuretic peptide during HF admission is a powerful predictor of outcome, but the sample sizes were relatively small (less than 200 patients), limiting their ability to control for confounding variables [13, 28]. In the Valsartan Heart Failure Trial (Val-HeFT), BNP was measured at randomization and at the fourth month in patients with chronic HF. The results suggested that percent change in neurohormones such as BNP is associated with mortality and morbidity, thus supporting their role as important surrogate markers in HF [29, 30]. Although there is currently no standard follow-up duration after discharge, our results suggest that short-term follow-up BNP and change in BNP are important metrics for predicting mortality and morbidity in patients with HF. Baseline BNP reflects hemodynamic stress regardless of cause or severity, while follow-up BNP and change in BNP reflect treatment response and hemodynamic status after treatment. Therefore, follow-up BNP and change in BNP are more powerful than baseline BNP for predicting morality or morbidity [18, 31].

HFpEF, defined as symptomatic HF with a normal or almost normal ejection fraction and diastolic dysfunction, accounts for about one-half of patients hospitalized for acute HF [5, 32, 33]. HFpEF patients have comorbidities that can drive myocardial dysfunction and concentric remodeling due to progressive loss of cardiomyocytes with increased relative wall thickness and a relatively preserved LV diameter, resulting in a high ratio of mass to volume. Patients with HFrEF have an enlarged LV cavity but fairly normal LV wall thickness, and exchange of dead cardiomyocytes by collagen creates patchy areas of fibrosis [34, 35]. BNP is a cardiac neurohormone that is primarily exudated from the ventricles in response to increases in wall tension and volume overload [36]. Patients with HFpEF who have lower wall tension thus have lower natriuretic peptide levels than those with HFrEF [37]. The KorAHF registry revealed that the plasma level of natriuretic peptide at baseline or at admission is the most powerful prognostic factor in both HFpEF and HFrEF, and it is also useful for risk-stratifying patients with HF [19]. We used similar inclusion criteria for patients from the same registry; however, we selected whole acute HF patient without any differentiation; interestingly, we found no relationship between BNP at admission and mortality. These differences might be due to the role of baseline BNP in reflecting baseline hemodynamic change, regardless of the cause of HF and severity of underlying disease. Moreover, multiple factors affect the BNP level [18]. Furthermore, when we used new statistical analysis methods (NRI and IDI), we found that BNP after discharge (IDI of 0.072, P < .0001 and NRI of 0.707, P < .0001) and percent change in BNP (IDI of 0.113, P < .0001 and NRI of 0.782, P < .0001) demonstrated the greatest increase in discrimination and net reclassification for mortality. This study was limited by a small population. Though the sample size was small but the study was chosen as a cohort study and most of the analytic data set had able to give significance in both FU-BNP and percent changes BNP with little variations. Moreover, a greater decrease in the percent change of BNP from admission to discharge increased the survival rate.

Short-term follow-up BNP after discharge and percent change in BNP are significant prognostic markers of mortality for hospitalized patients with HF. These values are clinically useful for the evaluation of patient prognosis.