- Email: [email protected]
Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome
G Model
JJCC-1316; No. of Pages 7 Journal of Cardiology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Cardiology journal homepage: www.elsevier.com/locate/jjcc
Original article
Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome Masashi Fujino (MD)a,b, Hiroyuki Takahama (MD, PhD)a,*, Toshimitsu Hamasaki (PhD)c, Kenichi Sekiguchi (MD, PhD)a, Kengo Kusano (MD, PhD, FJCC)a,b, Toshihisa Anzai (MD, PhD, FJCC)a,b, Teruo Noguchi (MD, PhD)a,b, Yoichi Goto (MD, PhD, FJCC)a, Masafumi Kitakaze (MD, PhD, FJCC)a, Hiroyuki Yokoyama (MD, PhD)a, Hisao Ogawa (MD, PhD, FJCC)a,d, Satoshi Yasuda (MD, PhD, FJCC)a,b a
Department of Cardiovascular Medicine, National Cerebral and Cardiovascular Center, Suita, Japan Division of Advanced Cardiovascular Medicine, Graduate School of Medical Sciences, Kumamoto University, Kumamoto, Japan c Office of Biostatistics and Data Management, Department of Advanced Medical Technology Development, National Cerebral and Cardiovascular Center, Suita, Japan d Division of Cardiovascular Medicine, Graduate School of Medical Sciences, Kumamoto University, Kumamoto, Japan b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 4 January 2016 Received in revised form 17 April 2016 Accepted 7 May 2016 Available online xxx
Background: Several blood tests are commonly used to assess nutritional status, including serum albumin levels (SAL) and lymphocyte counts (LC). The aim of this study is to investigate whether nutritional screening on admission can be used to determine risk levels for adverse clinical events in acute heart failure syndrome (AHFS) patients. Methods: In 432 consecutive AHFS patients, we measured SAL and LC and prospectively followed the patients for their combined clinical events (all-cause death and re-hospitalization for heart failure) for three years from admission. The classification and regression tree (CART) tool identified the cut-off criteria for SAL and LC to differentiate among patients with different risks of clinical events as 3.5 g/dl and 963/mm3, respectively. Results: The CART tool classified 15.5% patients as high risk, 15.7% as intermediate risk, and 68.8% as low risk. The CART for nutritional status (CART-NS) values were strongly correlated with combined clinical events [hazard ratio of 2.13 (low vs high risk), 95% confidence interval of 1.42–3.16, p < 0.001], even after adjusting for plasma brain natriuretic peptide levels. The CART-NS analysis improved the specificity (89.5%) of predictions of clinical outcomes with the comparable sensitivity (36.3%) compared with the use of a single criterion (SAL <3.5 g/dl: 70.2, 42.4% or LC <963/mm3: 73.4, 41.7%, respectively). Conclusion: A substantial proportion of AHFS patients are at risk of malnutrition, and this risk is associated with poor clinical outcomes. We demonstrate that this algorithm for nutritional screening, even in emergency clinical settings, can determine risk levels for further adverse events in AHFS patients. ß 2016 Japanese College of Cardiology. Published by Elsevier Ltd. All rights reserved.
Keywords: Nutrition Acute heart failure Albumin Lymphocyte counts Prognosis
Introduction Malnutrition is often observed in patients with heart failure (HF) and associated with adverse clinical outcomes [1–3]. Several
* Corresponding author at: Department of Cardiovascular Medicine, National Cerebral and Cardiovascular Center, Suita 565-8565, Osaka, Japan. Tel.: +81 6 6833 5012; fax: +81 6 6872 7486. E-mail address: [email protected] (H. Takahama).
studies have demonstrated the significance of malnutrition in chronic HF because of its association with long-term clinical outcomes [4]. However, the clinical significance of identification of malnutrition on admission in patients with acute heart failure syndrome (AHFS) has not been fully recognized in the clinical setting. Several examinations have been proposed to assess nutritional status in a clinical setting, including serum albumin levels, which are widely used for the assessment of nutritional status. Because
http://dx.doi.org/10.1016/j.jjcc.2016.05.004 0914-5087/ß 2016 Japanese College of Cardiology. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx
2
the half-life of albumin is 14–18 days, serum albumin levels can provide information on nutrition status related to prolonged malnutrition [5]. Horwich et al. reported that the presence of hypo-albuminemia (3.4 g/dl) was significantly associated with one-year survival in HF patients with systolic dysfunction [6]. Recently, a relationship between long-term mortality and hypo-albuminemia in HF patients has been reported [7]. However, serum albumin levels are influenced by multiple factors, such as hepatic synthesis [8]. Rapid and simple, but accurate, screening for nutritional status is necessary for the management of HF in emergency clinical settings, such as in patients with AHFS. Lymphocyte counts are also used as a standard measure of nutritional status [3,9–12] that is independent of hepatic synthesis. Therefore, we combined two variables for nutritional screening in AHFS patients. To obtain maximal accuracy of risk stratification within a short time, we used the classification and regression tree (CART) for nutritional status (CART-NS) tool and analyzed serum albumin levels and lymphocyte counts. Several studies for risk stratification of AHFS patients are reported [13– 15]; however, little has been known about the risk stratification of AHFS using assessment for nutritional status. The primary aim of the present study was to investigate the extent to which AHFS patients are at risk of malnutrition on admission and to determine whether the CART-NS algorithm can differentiate among highrisk patients with AHFS with regard to their long-term mortality and likelihood of re-hospitalization for HF.
Clinical follow-up and data analysis In the present study, we prospectively followed clinical events in AHFS patients for three years. Combined clinical events were defined as either all-cause death or re-hospitalization for HF, and such information for HF patients was obtained via medical records or communication by telephone and written correspondence. Risk stratification Risk stratification by the CART tool has been described previously [15]. Briefly, risk was allocated through a series of two binary decisions based on lymphocyte counts and serum albumin levels on patient admission (Fig. 1). The first allocation was for lymphocyte counts 963/mm3 vs lymphocyte counts <963/mm3, which was followed by serum albumin levels 3.5 g/dl vs serum albumin levels <3.5 g/dl, and these threshold values were identified through a recursive partitioning analysis (CART method) [15,20,21]. Ethics This study was approved by the National Cerebral and Cardiovascular Centre Institutional Ethics Committee and conducted in accordance with the ethical principles of the Declaration of Helsinki. Statistical analysis
Methods Study population From July 2006 to June 2009, 662 consecutive AHFS patients who met the Framingham criteria [16] were prospectively enrolled in our database. In this single-center registry [17], AHFS patients with acute coronary syndromes and patients who were admitted to our institute during this follow-up period more than twice (n = 117) were excluded. Additionally, patients whom we could not follow-up (n = 10) and for whom laboratory data were unavailable (n = 17) were not included in the present study. Because we used lymphocyte counts to build CART-NS, it was necessary to exclude the patients with abnormal values of white blood cell (WBC) counts, indicating the presence of infectious disease. Based on the previous study [18], we excluded the patients with either high WBC counts (12,000/mm3) or low WBC counts (<4000/mm3).
Data are expressed as the median and interquartile range (IQR), or percentages as appropriate. The Wilcoxon’s rank sum test was used for the comparison of continuous variables between the two groups; categorical variables were compared with x2 statistics or Fisher’s exact test as appropriate. For analysis of survival and hospitalization for HF, Kaplan–Meier method was used to estimate
Data collection and definition of nutritional impairment For each patient, baseline clinical data were collected, including (1) demographic, (2) etiological, and (3) comorbidity data, (4) laboratory tests and (5) echocardiographic findings. In addition to the echocardiographic examination, biochemical tests following blood sampling were performed on admission, including tests to determine the WBC count, lymphocyte count, and hemoglobin, albumin, C-reactive protein, creatinine, and blood urea nitrogen levels, which were obtained in the first 24 h after admission. Plasma brain natriuretic peptide (BNP) levels were obtained in the first three days after admission. All of the biochemical analyses were performed using routine hospital analytical facilities. An echocardiograph was also performed on admission, and the body mass index (BMI) was calculated as the body weight in kilograms divided by the square of the height in meters. The estimated glomerular filtration rate was calculated according to the following published equation: 194 serum creatinine 1.094 age 0.287 0.739 (for females) [19].
Fig. 1. The relationship of serum albumin levels and lymphocyte counts with clinical outcomes. The figure shows the values of serum albumin levels and lymphocyte counts in acute heart failure syndrome patients and indicates the presence of combined clinical outcomes, which are indicated by red closed circles for combined clinical events and blue open circles for no occurrence of events. A classification and regression tree for nutritional status analysis determined the cutoff values for lymphocyte counts and serum albumin levels, which were 963/mm3 and 3.5 g/dl, respectively.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx
survival curves and log-rank test was used to compare the survival curves. The multivariable, stepwise Cox proportional regression was used to determine which factors were significantly related to cardiac events after adjustment for other variables. The hazard ratio (HR) with 95% confidence interval (CI) and probability (p) value by likelihood ratio test are presented, where HRs for continuous variables apply per unit of the analyzed variable. Values of p < 0.05 were considered statistically significant. All statistical analyses were performed using JMP1 9 or 10 statistical analysis software (SAS Institute Japan Inc, Tokyo, Japan) or the CART was in-house validated Fortran programs. Results Classification and regression trees for nutritional status to differentiate risk According to these cut-off values of WBC counts, we excluded 86 patients. Ultimately, the remaining 432 patients constituted the study population with long-term clinical follow-up. Over the three years of follow-up, 104 patients died from any cause and 35 patients were re-hospitalized because of worsening HF. Fig. 1 shows the serum albumin levels and lymphocyte counts in AHFS patients as well as the presence of combined clinical outcomes, which are indicated by red closed circles (occurrence of combined clinical events) or blue open circles (no occurrence of combined clinical events). The CART-NS analysis provided the best lymphocyte count and serum albumin level cut-off values (963/mm3 and 3.5 g/dl, respectively) for differentiating among patients with clinical events. Fig. 2 shows the CART-NS risk categories assigned to the 432 patients, which indicates that 67 (15.5%) were stratified as high risk, 68 (15.7%) were stratified as intermediate risk, and 297 (68.8%) were stratified as low risk. Baseline characteristics in the stratified AHFS patients Table 1 shows the baseline characteristics in the stratified AHFS patients. The BMI tended to be lower in the intermediate-risk and
AHFS patients enrolled in the study (n=518) Exclusion (WBC counts≥12,000, <4000)
Lym < 963/mm3
Low risk N=297 Serum albumin ≥ 3.5 g/dl Intermediate risk n=68
high-risk groups than in the low-risk group, and the age was higher in the intermediate- and high-risk groups than the low-risk group (both, p < 0.05). The New York Heart Association (NYHA) class was comparable in the three groups, and no differences in the etiology of HF were observed among the groups. In the laboratory data, there is a tendency of lower hemoglobin and higher blood urea nitrogen in the intermediate-risk and high-risk groups compared with the low-risk group. Plasma levels of BNP were higher in the high-risk and intermediate-risk groups compared with the lowrisk group (p < 0.05). Echocardiography showed that systolic function and left ventricular chamber size did not differ among the three groups. The frequency of beta-blocker intake tended to be higher in the low-risk group compared with the other two groups. Additionally, there was a weak correlation between serum albumin levels and lymphocyte counts (r = 0.103, p = 0.032), and serum albumin levels and lymphocyte counts were weakly associated with plasma BNP levels (r = 0.099, p = 0.041 and r = 0.122, p = 0.011, respectively). Relationship between nutritional status and clinical outcomes Kaplan–Meier survival analysis revealed differences in the frequency of combined clinical events among the groups (p < 0.001, Fig. 3). Table 2 shows the predictive values of the CART-NS for combined clinical events. Model 2, which is adjusted for age, gender, and log BNP levels, shows that the HR was 2.13 (95% CI: 1.42–3.16, p < 0.001) between the low-risk and high-risk groups and 1.99 (95% CI: 1.17–3.46, p = 0.010) between the intermediate-risk and high-risk groups. Table 3 shows the results of the univariate and multivariate analyses. The multivariate analysis shows that the CART-NS algorithm has strong predictive power, even after adjusting for other factors. Improved predictability of clinical events by the combined use of serum albumin levels and lymphocyte counts Table 4 shows the sensitivity and specificity of low serum albumin levels (<3.5 g/dl), low lymphocyte counts (<963/mm3) and high-risk classification in CART-NS for combined clinical events. Use of a single criterion (low lymphocyte count or low serum albumin level) predicted combined clinical events with a sensitivity of 41.7% and 42.4% and specificity of 73.4% and 70.2%, respectively. The use of CART-NS (high-risk in CART-NS) increased specificity to 89.5% with comparable sensitivity of 36.3%. Risk stratification by nutritional screening in patients with de novo hospitalization vs re-hospitalization because of acute heart failure
After exclusion(n=432) Lym ≥ 963/mm3
3
Serum albumin < 3.5 g/dl High risk n=67
Fig. 2. Flow chart of enrolled patients. The flow chart shows the enrolment of acute heart failure syndrome (AHFS) patients and cut-off criteria for lymphocyte counts (Lym) and serum albumin levels assigned to the stratified groups: low-risk group (lymphocyte counts 963/mm3), intermediate-risk group (lymphocyte counts <963/mm3 and serum albumin levels 3.5 g/dl) and high-risk group (lymphocyte counts <963/mm3 and serum albumin levels <3.5 g/dl). Exclusion criteria of the patients with abnormal values of white blood cell (WBC): 12,000/mm3 or <4000/ mm3.
Of the enrolled AHFS patients, the number of the patients who were de novo hospitalized or had already past history of hospitalization due to HF was 232 and 200, respectively, and different distributions of stratified groups were observed between the de novo hospitalization and re-hospitalization groups because of AHFS, as shown in Fig. 4 (p = 0.042). Additionally, we could identify the cause of death in 75 patients (72% of overall deceased patients). Of the patients (n = 75), cardiovascular death occurred in 53 patients (71%), and non-cardiovascular death occurred in the remaining patients (22 patients, 29%). The sub-group analysis using CART-NS demonstrated that frequency of non-cardiovascular death tended to be higher in the high-risk group (n = 10, 42%) compared with the low-risk group (n = 8, 21%), although there was no statistical difference among the groups (p = 0.190), as shown in Supplemental Figure. Nonetheless, in the high-risk group, cardiovascular death accounted for more than a half of the causes of death (n = 14, 58%).
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx
4
Table 1 Baseline characteristics in acute heart failure syndrome patients. Low risk Pts number Age (year) Male (%) Body mass index (kg/m2) NYHA III–IV (%) History Prior MI (%) HF hospitalization (within six months) (%) Hypertension (%) DM/IGT (%) Etiology Ischemic cardiomyopathy (%) Non-ischemic cardiomyopathy (%) Valvular (%) Hypertensive (%) Others (%) Vital signs on admission Systolic blood pressure (mmHg) Heart rate (beat per minute) Laboratory data on admission WBC count (/mm3) Lymphocyte count (/mm3) Hemoglobin (g/dl) Serum albumin (g/dl) BUN (mg/dl) eGFR (ml/min/1.73 m2) C-reactive protein (mg/dl) BNP (pg/ml) Echocardiography LVEDd (mm) %FS (%) Medications ACEi or ARB (%) Beta-blockers (%) Loop diuretics (%) Spironolactone (%)
297 73 (64–82) 66 22.6 (20.5–25.0) 89
Intermediate risk
High risk
68 77 (64–82)* 68 21.9 (19.1–24.6) 92
67 77 (68–85)* 63 21.7 (18.9–25.5) 91
26 13 69 34
30 15 72 40
35 21 72 50
32 22 30 13 3
36 16 41 6 2
39 20 20 13 8
140 (117–163) 94 (77–120) 7400 (5800–9150) 1482 (1190–1950) 12.5 (10.9–14.1) 3.7 (3.4–3.9) 20 (16–28) 54 (39–69) 0.41 (0.17–1.43) 660 (369–1162)
137 (115–153) 86 (70–104)* 6450 (5150–7375)* 755 (504–856) 11.1 (9.7–12.5)* 3.8 (3.6–4.0) 27 (18–38)* 41 (30–55)* 0.59 (0.19–1.33) 835 (375–1412)*
132 (117–163) 87 (74–102) 6600 (5400–10,280) 720 (536–845) 10.9 (9.6–12.1)* 3.1 (2.9–3.3) 27 (20–38)* 42 (28–56)* 2.18 (0.46–7.11)*,# 1044 (431–1783)*
56 (47–62) 18 (13–26)
55 (46–62) 20 (15–28)
57 (51–62) 18 (13–27)
67 63 82 37
63 52 70 45
60 52 81 41
Pts, patients; NYHA, New York Heart Association; MI, myocardial infarction; HF, heart failure; DM, diabetes mellitus; IGT, impaired glucose tolerance; WBC, white blood cell count; BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate; BNP, brain natriuretic peptide; LVEDd, left ventricular end-diastolic diameter; FS, fractional shortening; ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor antagonists. Values are expressed as the median (interquartile range, IQR) or percentage. Lymphocyte counts and serum albumin levels were not selected for the analysis owing to the category criteria. Analysis is unadjusted. * p < 0.05 vs low-risk group. # p < 0.05 vs intermediate-risk group.
Discussion In a cross-sectional and prospective analysis of 432 AHFS patients, we demonstrated that initial nutritional screening using CART analysis can stratify risk levels for further adverse clinical events over three years. In addition, the CART analysis determined the cut-off values of serum albumin levels and lymphocyte counts. According to this algorithm, we found that a substantial proportion of AHFS patients were at risk of malnutrition on admission (15.5%), and this state was strongly associated with poor clinical outcomes. Although the severity of HF indicated by NYHA class was similar among all three groups, the highest frequency of combined clinical events was observed in the high-risk group. The multivariate analysis revealed that classification as high-risk was strongly associated with the frequency of clinical events over the three years following admission, even after adjusting for age, gender, and plasma BNP levels. Therefore, we demonstrated that an initial nutritional assessment in the acute phase of AHFS can provide accurate predictions of future clinical outcomes. Interestingly, differences in body size were relatively small among the groups in the present study. The finding suggests that our algorithm can detect early warning signs of malnutrition in patients with HF before body weight loss is apparent, such as in cardiac cachexia. We also can raise the significance of non-cardiac factors, such as
nutritional status, for determining systemic pathophysiology in HF patients. Impact of nutritional screening in patients with AHFS in emergency room There is accumulating evidence that malnutrition is common and affects disease progression in patients with chronic HF [1– 3]. Importantly, cardiac cachexia, an advanced manifestation of nutritional impairment, has been recognized as an important factor in deteriorating HF [22]. However, a definition of malnutrition in patients with HF has not been established. Although it is widely known that body weight loss is an important indicator of advanced HF, body weight alone may be insufficient to predict the risk of progression to cardiac cachexia. Indeed, CART-NS was a better predictor than BMI in the present study (Table 2). Based on previous studies, we selected serum albumin levels and lymphocyte count, and both parameters were tested in emergency room settings. Subsequently, we performed CART analysis (Fig. 1) to determine the patient’s nutritional status in a clinical setting. The classification showed that on admission, a substantial proportion of the AHFS patients were at risk of malnutrition (15.5%), which is associated with poor clinical outcomes. The present algorithm has several strengths for assessing nutritional status. First, the required
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx Table 4 Sensitivity and specificity of the variables for combined clinical events.
100
Event-Free Rate (%)
5
80 3
Low lymphocyte count (<963/mm ) Low serum albumin (<3.5 g/dl) Nutritional risk tree (high risk)
60
Sensitivity (%)
Specificity (%)
41.7 42.4 36.3
73.4 70.2 89.5
40 High risk group 20
Intermediate risk group Low risk group
0 0
100
200
300
400
500
600
700
800
900 1000
Time to Event (days) Low risk group Int risk group High risk group
297 68 67
264 54 48
187 33 36
119 23 17
Fig. 3. Kaplan–Meier analysis of patients with acute heart failure syndrome stratified according to classification and regression tree for nutritional status (CART-NS). Rate of combined clinical events as determined by the Kaplan–Meier analysis of patients classified according to the CART-NS algorithm. Green line, red line, and blue line indicate the low-risk group (lymphocyte counts 963/mm3), intermediate-risk group (lymphocyte counts <963/mm3, serum albumin levels 3.5 g/dl) and high-risk group (lymphocyte counts <963/mm3, serum albumin levels <3.5 g/dl), respectively. Kaplan–Meier survival analysis revealed differences in the frequency of combined clinical events among the groups (p < 0.001).
Table 2 Predictive values of the classification and regression tree for nutritional status for combined clinical events.
Low risk vs high risk Model 1 Model 2 Intermediate risk vs high risk Model 1 Model 2
HR (95% CI)
p-Value
2.33 (1.56–3.44) 2.13 (1.42–3.16)
<0.001 <0.001
2.11 (1.25–3.67) 1.99 (1.17–3.46)
0.005 0.010
Model 1 was adjusted for age and gender, and Model 2 was adjusted for age, gender and log B-type natriuretic peptide levels. HR, hazard ratio.
blood tests can be performed easily, even in the emergency room, and the results can be obtained within a short time. Second, our algorithm is strongly associated with future clinical outcomes. Moreover, serum albumin levels have been shown to be influenced by hepatic synthesis, whereas lymphocyte counts are not influenced by this factor. Indeed, the correlation between serum albumin levels and lymphocyte counts was weak in the present study. Thus, combining these criteria can be recommended to
Fig. 4. Risk stratification by nutritional screening in patients with de novo hospitalization vs re-hospitalization because of acute heart failure syndrome (AHFS). Of the enrolled patients, de novo hospitalizations because of AHFS (de novo HOSP) occurred in 232 patients and re-hospitalization because of AHFS (Re-HOSP) occurred in 200 patients. In the de novo HOSP group, the patients assigned to lowrisk (green square), intermediate-risk (red square) and high-risk groups (blue square) accounted for 73%, 12%, and 14%, whereas in the Re-HOSP group, patients assigned to the low-risk, intermediate-risk and high-risk groups accounted for 63%, 20%, and 17%, respectively. Different distributions of stratified groups were observed between the de novo HOSP and Re-HOSP groups because of AHFS (p = 0.042).
improve the accuracy of clinical risk predictions, and in this study, the combined use of both criteria increased specificity to 89.5% with comparable sensitivity of 36.3% compared with their use as single criteria (serum albumin level, 42.4% and 70.2%, respectively; lymphocyte count, 41.7% and 73.4%, respectively). Third, the associations of these factors with plasma BNP levels are weak; thus, we can evaluate non-cardiac factors that represent different aspects of HF pathophysiology from circulating BNP levels. Therefore, we demonstrated the potential of initial assessment of nutritional status for risk stratification in AHFS patients. Mechanism of malnutrition in HF patients Although it is widely known that hypo-albuminemia can serve as a biomarker for prolonged malnutrition, serum albumin levels are influenced not only by food intake but also by hepatic synthesis, protein degradation, and increased vascular permeability [8]. Moreover, studies have suggested that hypo-albuminemia in HF patients could be caused by increased inflammatory activity in addition to decreased intake of food [6,23]. In clinical studies, increased circulating inflammatory cytokines and chemokines have been significantly correlated with increasing HF severity or deteriorating left ventricular systolic function, suggesting that systemic inflammatory activity plays an important role in HF
Table 3 Univariate/multivariate analysis for combined clinical events. Individual measures Age (per year increase) BMI (per kg/m2 increase) Log BNP eGFR (per ml/min/1.73 m2 increase) BUN (mg/dl increase) Hb (mg/dl increase) CART-NS (intermediate vs high) CART-NS (low vs high)
Univariate HR (95% CI) 1.01 0.92 2.71 0.99 1.02 0.90 2.14 2.58
(1.00–1.02) (0.88–0.96) (1.73–4.28) (0.98–0.99) (1.01–1.02) (0.84–0.97) (1.27–3.73) (1.73–3.78)
p-Value 0.005 <0.001 <0.001 0.001 <0.001 0.006 0.004 <0.001
Multivariate HR (95% CI) 1.02 0.96 1.89 1.00 1.01 1.02 2.27 2.14
(1.01–1.04) (0.91–1.00) (1.17–3.10) (0.99–1.01) (1.00–1.02) (0.93–1.11) (1.32–4.02) (1.38–3.27)
p-Value 0.006 0.062 0.009 0.677 0.016 0.697 0.003 0.001
HR, hazard ratio; CI, confidence interval; BMI, body mass index; BNP, B-type natriuretic peptide; eGFR, estimated glomerular filtration rate; BUN, blood urea nitrogen; Hb, hemoglobin; CART-NS, classification and regression trees tool for nutritional status.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 6
M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx
patients [24–28]. Interestingly, it has also been reported that tumor necrosis factor alpha (TNF-a) is significantly increased in patients with cardiac cachexia [29]. Experimental studies have demonstrated that inflammatory cytokines (e.g. TNF-a, interleukin-1) and chemokines (e.g. monocyte chemotactic protein 1) directly depress cardiac function and were related to impaired nutritional status [30–33]. The evidence also suggests a strong association between a systemic inflammatory state and malnutrition, both of which might play a role in aggravating HF. Indeed, the patients in the high-risk group in the present study showed the lowest levels of hemoglobin and highest levels of C-reactive protein among the groups. Because this was an observational study, further investigation is necessary to address this issue. Another possible mechanism of malnutrition in HF patients is related to the aggravation of nutritional status by protracted HF itself. Indeed, the high-risk group included a greater number of rehospitalized patients than de novo HF hospitalized patients. Several interplaying factors may be involved in the relationship between nutritional state and HF aggravation. Decreased food intake is often observed in hospitalized patients with HF, which may also cause malnutrition in patients with HF. Patients with chronic HF and cachexia exhibit gastrointestinal fat malabsorption [30]. It should also be noted that statistical relationships of plasma levels of BNP with nutritional parameters, such as plasma albumin levels and lymphocyte counts, were weak. This finding also implies that malnutrition and plasma BNP are indicative of different aspects of worsening HF. Indeed, a frequency of non-cardiovascular death tended to be higher in the high-risk group, as shown in the Supplemental Figure. Furthermore, it is widely believed that HF may activate inflammation, potentially causing vicious cycles in HF patients. The interplay among malnutrition, inflammation, and deterioration of HF per se may increase not only cardiovascular events but also non-cardiovascular death. Because the present study was observational, further investigation is necessary to clarify the temporal relationship between HF and worsening nutritional status. This study had several limitations. First, the ‘gold-standard’ evaluation for nutritional status in patients with HF has not been established. Nonetheless, BMI at hospital discharge tended to be lower in the high-risk group than the low-risk group (19.0 vs 20.3, respectively p = 0.057). These findings suggest that our algorithm is associated with future body weight changes. Second, this study was based on a single-center registry with a moderate number of enrolled patients. Third, although differences in the frequency of beta-blocker intake were found between the low-risk group and the other two groups, the reason for this finding remains unclear. In the high-risk group, the patients were older than in the other two groups. We could not exclude the possibility that these patient characteristics might influence beta-blocker use in the population. An association of low relative lymphocyte counts with mortality in HF has also been reported, suggesting a relationship with elevated serum cortisol [10]. In the present study, we did not obtain cortisol measurements in all patients; therefore, we did not address this relationship. Additionally, we did not include cholesterol values in the algorithm because it was unable to evaluate the details of food intake or effects of medications, such as statins in all AHFS patients in emergency clinical settings. Additionally, we excluded the patients with abnormal values of WBC counts indicating severe infectious disease. In the patients with mild degree of infectious diseases, unless they influence the WBC count values, theoretically our algorithm (CART-NS) can be applicable to these patients. However, since it is difficult to distinguish infection and AHFS only by WBC counts [34] in acute phases, careful interpretation and follow-up would be necessary to apply this algorithm for patients with symptoms of serious infection representing abnormal values of WBC counts in the clinical setting. Lastly, we could not identify
the precise cause of death in 22 patients who did not die in our hospital. Conclusions On admission, a substantial proportion of AHFS patients are at risk of malnutrition, which is associated with poor clinical outcomes. We demonstrated that nutritional screening on admission can be used to stratify risk levels for further clinical events in AHFS patients despite limited differences in body size among the patient groups. These findings suggest that our algorithm can detect early warning signs of malnutrition in patients with HF before body weight loss is apparent. More importantly, we determined that this novel algorithm for nutritional screening can be performed within a short time and can even be completed in an emergency room. Where is Figure 4? I cannot find the Figure 4. Conflicts of interest The authors declare that there is no conflict of interest. Funding sources The present study was supported by the Intramural Research Fund, grant number 24-4-1, for Cardiovascular Diseases of the National Cerebral and Cardiovascular Center. Disclosure No financial supports exist in this study. Acknowledgment The authors thank Yoko Sumita, our secretary, for assistance in managing data.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jjcc.2016.05.004. References [1] Schwengel RH, Gottlieb SS, Fisher ML. Protein-energy malnutrition in patients with ischemic and nonischemic dilated cardiomyopathy and congestive heart failure. Am J Cardiol 1994;73:908–10. [2] Berger MM, Mustafa I. Metabolic and nutritional support in acute cardiac failure. Curr Opin Clin Nutr Metab Care 2003;6:195–201. [3] Carr JG, Stevenson LW, Walden JA, Heber D. Prevalence and hemodynamic correlates of malnutrition in severe congestive heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol 1989;63:709–13. [4] Sarma S, Gheorghiade M. Nutritional assessment and support of the patient with acute heart failure. Curr Opin Crit Care 2010;16:413–8. [5] Omran ML, Morley JE. Assessment of protein energy malnutrition in older persons, Part II: Laboratory evaluation. Nutrition 2000;16:131–40. [6] Horwich TB, Kalantar-Zadeh K, MacLellan RW, Fonarow GC. Albumin levels predict survival in patients with systolic heart failure. Am Heart J 2008;155:883–9. [7] Bonilla-Palomas JL, Gamez-Lopez AL, Moreno-Conde M, Lopez-Ibanez MC, Anguita-Sanchez M, Gallego de la Sacristana A, Garcia-Catalan F, Villar-Raez A. Hypoalbuminemia in acute heart failure patients: causes and its impact on hospital and long-term mortality. J Card Fail 2014;20:350–8. [8] Ballmer PE. Causes and mechanisms of hypoalbuminaemia. Clin Nutr 2001;20:271–3. [9] Dzieniszewski J, Jarosz M, Szczygiel B, Dlugosz J, Marlicz K, Linke K, Lachowicz A, Ryzko-Skiba M, Orzeszko M. Nutritional status of patients hospitalised in Poland. Eur J Clin Nutr 2005;59:552–60. [10] Ommen SR, Hodge DO, Rodeheffer RJ, McGregor CG, Thomson SP, Gibbons RJ. Predictive power of the relative lymphocyte concentration in patients with advanced heart failure. Circulation 1998;97:19–22.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx [11] Huehnergarth KV, Mozaffarian D, Sullivan MD, Crane BA, Wilkinson CW, Lawler RL, McDonald GB, Fishbein DP, Levy WC. Usefulness of relative lymphocyte count as an independent predictor of death/urgent transplant in heart failure. Am J Cardiol 2005;95:1492–5. [12] Acanfora D, Gheorghiade M, Trojano L, Furgi G, Pasini E, Picone C, Papa A, Iannuzzi GL, Bonow RO, Rengo F. Relative lymphocyte count: a prognostic indicator of mortality in elderly patients with congestive heart failure. Am Heart J 2001;142:167–73. [13] Raposeiras-Roubin S, Abu-Assi E, Lopez-Lopez A, Bouzas-Cruz N, CastineiraBusto M, Cambeiro-Gonzalez C, Alvarez-Alvarez B, Virgos-Lamela A, Varela-Roman A, Garcia-Acuna JM, Gonzalez-Juanatey JR. Risk stratification for the development of heart failure after acute coronary syndrome at the time of hospital discharge: predictive ability of grace risk score. J Cardiol 2015;66:224–31. [14] Okayama D, Minami Y, Kataoka S, Shiga T, Hagiwara N. Thyroid function on admission and outcome in patients hospitalized for acute decompensated heart failure. J Cardiol 2015;66:205–11. [15] Fonarow GC, Adams Jr KF, Abraham WT, Yancy CW, Boscardin WJ. Risk stratification for in-hospital mortality in acutely decompensated heart failure: classification and regression tree analysis. JAMA 2005;293:572–80. [16] Ho KK, Anderson KM, Kannel WB, Grossman W, Levy D. Survival after the onset of congestive heart failure in Framingham heart study subjects. Circulation 1993;88:107–15. [17] Takahama H, Yokoyama H, Kada A, Sekiguchi K, Fujino M, Funada A, Amaki M, Hasegawa T, Asakura M, Kanzaki H, Anzai T, Kitakaze M. Extent of heart rate reduction during hospitalization using beta-blockers, not the achieved heart rate itself at discharge, predicts the clinical outcome in patients with acute heart failure syndromes. J Cardiol 2013;61:58–64. [18] Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, Sevransky JE, Sprung CL, Douglas IS, Jaeschke R, Osborn TM, Nunnally ME, Townsend SR, Reinhart K, Kleinpell RM, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med 2013;41:580–637. [19] Matsuo S, Imai E, Horio M, Yasuda Y, Tomita K, Nitta K, Yamagata K, Tomino Y, Yokoyama H, Hishida A. Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis 2009;53:982–92. [20] Friedman JH, Meulman JJ. Multiple additive regression trees with application in epidemiology. Stat Med 2003;22:1365–81. [21] Rovlias A, Kotsou S. Classification and regression tree for prediction of outcome after severe head injury using simple clinical and laboratory variables. J Neurotrauma 2004;21:886–93.
7
[22] Anker SD, von Haehling S. Inflammatory mediators in chronic heart failure: an overview. Heart 2004;90:464–70. [23] Arques S, Roux E, Stolidi P, Gelisse R, Ambrosi P. Usefulness of serum albumin and serum total cholesterol in the prediction of hospital death in older patients with severe, acute heart failure. Arch Cardiovasc Dis 2011;104:502–8. [24] Yndestad A, Damas JK, Oie E, Ueland T, Gullestad L, Aukrust P. Systemic inflammation in heart failure – the whys and wherefores. Heart Fail Rev 2006;11:83–92. [25] Aukrust P, Ueland T, Lien E, Bendtzen K, Muller F, Andreassen AK, Nordoy I, Aass H, Espevik T, Simonsen S, Froland SS, Gullestad L. Cytokine network in congestive heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol 1999;83:376–82. [26] Aukrust P, Ueland T, Muller F, Andreassen AK, Nordoy I, Aas H, Kjekshus J, Simonsen S, Froland SS, Gullestad L. Elevated circulating levels of C–C chemokines in patients with congestive heart failure. Circulation 1998;97: 1136–43. [27] Damas JK, Gullestad L, Ueland T, Solum NO, Simonsen S, Froland SS, Aukrust P. Cxc-chemokines, a new group of cytokines in congestive heart failure – possible role of platelets and monocytes. Cardiovasc Res 2000;45:428–36. [28] Testa M, Yeh M, Lee P, Fanelli R, Loperfido F, Berman JW, LeJemtel TH. Circulating levels of cytokines and their endogenous modulators in patients with mild to severe congestive heart failure due to coronary artery disease or hypertension. J Am Coll Cardiol 1996;28:964–71. [29] Torre-Amione G, Kapadia S, Lee J, Durand JB, Bies RD, Young JB, Mann DL. Tumor necrosis factor-alpha and tumor necrosis factor receptors in the failing human heart. Circulation 1996;93:704–11. [30] Mann DL. Inflammatory mediators and the failing heart: past, present, and the foreseeable future. Circ Res 2002;91:988–98. [31] Aukrust P, Gullestad L, Ueland T, Damas JK, Yndestad A. Inflammatory and anti-inflammatory cytokines in chronic heart failure: potential therapeutic implications. Ann Med 2005;37:74–85. [32] Mann DL. Stress-activated cytokines and the heart: from adaptation to maladaptation. Annu Rev Physiol 2003;65:81–101. [33] Okonko DO, Anker SD. Anemia in chronic heart failure: pathogenetic mechanisms. J Card Fail 2004;10:S5–9. [34] Maisel A, Neath SX, Landsberg J, Mueller C, Nowak RM, Peacock WF, Ponikowski P, Mockel M, Hogan C, Wu AH, Richards M, Clopton P, Filippatos GS, Di Somma S, Anand I, et al. Use of procalcitonin for the diagnosis of pneumonia in patients presenting with a chief complaint of dyspnoea: results from the BACH (Biomarkers in Acute Heart Failure) trial. Eur J Heart Fail 2012;14:278–86.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
JJCC-1316; No. of Pages 7 Journal of Cardiology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Cardiology journal homepage: www.elsevier.com/locate/jjcc
Original article
Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome Masashi Fujino (MD)a,b, Hiroyuki Takahama (MD, PhD)a,*, Toshimitsu Hamasaki (PhD)c, Kenichi Sekiguchi (MD, PhD)a, Kengo Kusano (MD, PhD, FJCC)a,b, Toshihisa Anzai (MD, PhD, FJCC)a,b, Teruo Noguchi (MD, PhD)a,b, Yoichi Goto (MD, PhD, FJCC)a, Masafumi Kitakaze (MD, PhD, FJCC)a, Hiroyuki Yokoyama (MD, PhD)a, Hisao Ogawa (MD, PhD, FJCC)a,d, Satoshi Yasuda (MD, PhD, FJCC)a,b a
Department of Cardiovascular Medicine, National Cerebral and Cardiovascular Center, Suita, Japan Division of Advanced Cardiovascular Medicine, Graduate School of Medical Sciences, Kumamoto University, Kumamoto, Japan c Office of Biostatistics and Data Management, Department of Advanced Medical Technology Development, National Cerebral and Cardiovascular Center, Suita, Japan d Division of Cardiovascular Medicine, Graduate School of Medical Sciences, Kumamoto University, Kumamoto, Japan b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 4 January 2016 Received in revised form 17 April 2016 Accepted 7 May 2016 Available online xxx
Background: Several blood tests are commonly used to assess nutritional status, including serum albumin levels (SAL) and lymphocyte counts (LC). The aim of this study is to investigate whether nutritional screening on admission can be used to determine risk levels for adverse clinical events in acute heart failure syndrome (AHFS) patients. Methods: In 432 consecutive AHFS patients, we measured SAL and LC and prospectively followed the patients for their combined clinical events (all-cause death and re-hospitalization for heart failure) for three years from admission. The classification and regression tree (CART) tool identified the cut-off criteria for SAL and LC to differentiate among patients with different risks of clinical events as 3.5 g/dl and 963/mm3, respectively. Results: The CART tool classified 15.5% patients as high risk, 15.7% as intermediate risk, and 68.8% as low risk. The CART for nutritional status (CART-NS) values were strongly correlated with combined clinical events [hazard ratio of 2.13 (low vs high risk), 95% confidence interval of 1.42–3.16, p < 0.001], even after adjusting for plasma brain natriuretic peptide levels. The CART-NS analysis improved the specificity (89.5%) of predictions of clinical outcomes with the comparable sensitivity (36.3%) compared with the use of a single criterion (SAL <3.5 g/dl: 70.2, 42.4% or LC <963/mm3: 73.4, 41.7%, respectively). Conclusion: A substantial proportion of AHFS patients are at risk of malnutrition, and this risk is associated with poor clinical outcomes. We demonstrate that this algorithm for nutritional screening, even in emergency clinical settings, can determine risk levels for further adverse events in AHFS patients. ß 2016 Japanese College of Cardiology. Published by Elsevier Ltd. All rights reserved.
Keywords: Nutrition Acute heart failure Albumin Lymphocyte counts Prognosis
Introduction Malnutrition is often observed in patients with heart failure (HF) and associated with adverse clinical outcomes [1–3]. Several
* Corresponding author at: Department of Cardiovascular Medicine, National Cerebral and Cardiovascular Center, Suita 565-8565, Osaka, Japan. Tel.: +81 6 6833 5012; fax: +81 6 6872 7486. E-mail address: [email protected] (H. Takahama).
studies have demonstrated the significance of malnutrition in chronic HF because of its association with long-term clinical outcomes [4]. However, the clinical significance of identification of malnutrition on admission in patients with acute heart failure syndrome (AHFS) has not been fully recognized in the clinical setting. Several examinations have been proposed to assess nutritional status in a clinical setting, including serum albumin levels, which are widely used for the assessment of nutritional status. Because
http://dx.doi.org/10.1016/j.jjcc.2016.05.004 0914-5087/ß 2016 Japanese College of Cardiology. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx
2
the half-life of albumin is 14–18 days, serum albumin levels can provide information on nutrition status related to prolonged malnutrition [5]. Horwich et al. reported that the presence of hypo-albuminemia (3.4 g/dl) was significantly associated with one-year survival in HF patients with systolic dysfunction [6]. Recently, a relationship between long-term mortality and hypo-albuminemia in HF patients has been reported [7]. However, serum albumin levels are influenced by multiple factors, such as hepatic synthesis [8]. Rapid and simple, but accurate, screening for nutritional status is necessary for the management of HF in emergency clinical settings, such as in patients with AHFS. Lymphocyte counts are also used as a standard measure of nutritional status [3,9–12] that is independent of hepatic synthesis. Therefore, we combined two variables for nutritional screening in AHFS patients. To obtain maximal accuracy of risk stratification within a short time, we used the classification and regression tree (CART) for nutritional status (CART-NS) tool and analyzed serum albumin levels and lymphocyte counts. Several studies for risk stratification of AHFS patients are reported [13– 15]; however, little has been known about the risk stratification of AHFS using assessment for nutritional status. The primary aim of the present study was to investigate the extent to which AHFS patients are at risk of malnutrition on admission and to determine whether the CART-NS algorithm can differentiate among highrisk patients with AHFS with regard to their long-term mortality and likelihood of re-hospitalization for HF.
Clinical follow-up and data analysis In the present study, we prospectively followed clinical events in AHFS patients for three years. Combined clinical events were defined as either all-cause death or re-hospitalization for HF, and such information for HF patients was obtained via medical records or communication by telephone and written correspondence. Risk stratification Risk stratification by the CART tool has been described previously [15]. Briefly, risk was allocated through a series of two binary decisions based on lymphocyte counts and serum albumin levels on patient admission (Fig. 1). The first allocation was for lymphocyte counts 963/mm3 vs lymphocyte counts <963/mm3, which was followed by serum albumin levels 3.5 g/dl vs serum albumin levels <3.5 g/dl, and these threshold values were identified through a recursive partitioning analysis (CART method) [15,20,21]. Ethics This study was approved by the National Cerebral and Cardiovascular Centre Institutional Ethics Committee and conducted in accordance with the ethical principles of the Declaration of Helsinki. Statistical analysis
Methods Study population From July 2006 to June 2009, 662 consecutive AHFS patients who met the Framingham criteria [16] were prospectively enrolled in our database. In this single-center registry [17], AHFS patients with acute coronary syndromes and patients who were admitted to our institute during this follow-up period more than twice (n = 117) were excluded. Additionally, patients whom we could not follow-up (n = 10) and for whom laboratory data were unavailable (n = 17) were not included in the present study. Because we used lymphocyte counts to build CART-NS, it was necessary to exclude the patients with abnormal values of white blood cell (WBC) counts, indicating the presence of infectious disease. Based on the previous study [18], we excluded the patients with either high WBC counts (12,000/mm3) or low WBC counts (<4000/mm3).
Data are expressed as the median and interquartile range (IQR), or percentages as appropriate. The Wilcoxon’s rank sum test was used for the comparison of continuous variables between the two groups; categorical variables were compared with x2 statistics or Fisher’s exact test as appropriate. For analysis of survival and hospitalization for HF, Kaplan–Meier method was used to estimate
Data collection and definition of nutritional impairment For each patient, baseline clinical data were collected, including (1) demographic, (2) etiological, and (3) comorbidity data, (4) laboratory tests and (5) echocardiographic findings. In addition to the echocardiographic examination, biochemical tests following blood sampling were performed on admission, including tests to determine the WBC count, lymphocyte count, and hemoglobin, albumin, C-reactive protein, creatinine, and blood urea nitrogen levels, which were obtained in the first 24 h after admission. Plasma brain natriuretic peptide (BNP) levels were obtained in the first three days after admission. All of the biochemical analyses were performed using routine hospital analytical facilities. An echocardiograph was also performed on admission, and the body mass index (BMI) was calculated as the body weight in kilograms divided by the square of the height in meters. The estimated glomerular filtration rate was calculated according to the following published equation: 194 serum creatinine 1.094 age 0.287 0.739 (for females) [19].
Fig. 1. The relationship of serum albumin levels and lymphocyte counts with clinical outcomes. The figure shows the values of serum albumin levels and lymphocyte counts in acute heart failure syndrome patients and indicates the presence of combined clinical outcomes, which are indicated by red closed circles for combined clinical events and blue open circles for no occurrence of events. A classification and regression tree for nutritional status analysis determined the cutoff values for lymphocyte counts and serum albumin levels, which were 963/mm3 and 3.5 g/dl, respectively.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx
survival curves and log-rank test was used to compare the survival curves. The multivariable, stepwise Cox proportional regression was used to determine which factors were significantly related to cardiac events after adjustment for other variables. The hazard ratio (HR) with 95% confidence interval (CI) and probability (p) value by likelihood ratio test are presented, where HRs for continuous variables apply per unit of the analyzed variable. Values of p < 0.05 were considered statistically significant. All statistical analyses were performed using JMP1 9 or 10 statistical analysis software (SAS Institute Japan Inc, Tokyo, Japan) or the CART was in-house validated Fortran programs. Results Classification and regression trees for nutritional status to differentiate risk According to these cut-off values of WBC counts, we excluded 86 patients. Ultimately, the remaining 432 patients constituted the study population with long-term clinical follow-up. Over the three years of follow-up, 104 patients died from any cause and 35 patients were re-hospitalized because of worsening HF. Fig. 1 shows the serum albumin levels and lymphocyte counts in AHFS patients as well as the presence of combined clinical outcomes, which are indicated by red closed circles (occurrence of combined clinical events) or blue open circles (no occurrence of combined clinical events). The CART-NS analysis provided the best lymphocyte count and serum albumin level cut-off values (963/mm3 and 3.5 g/dl, respectively) for differentiating among patients with clinical events. Fig. 2 shows the CART-NS risk categories assigned to the 432 patients, which indicates that 67 (15.5%) were stratified as high risk, 68 (15.7%) were stratified as intermediate risk, and 297 (68.8%) were stratified as low risk. Baseline characteristics in the stratified AHFS patients Table 1 shows the baseline characteristics in the stratified AHFS patients. The BMI tended to be lower in the intermediate-risk and
AHFS patients enrolled in the study (n=518) Exclusion (WBC counts≥12,000, <4000)
Lym < 963/mm3
Low risk N=297 Serum albumin ≥ 3.5 g/dl Intermediate risk n=68
high-risk groups than in the low-risk group, and the age was higher in the intermediate- and high-risk groups than the low-risk group (both, p < 0.05). The New York Heart Association (NYHA) class was comparable in the three groups, and no differences in the etiology of HF were observed among the groups. In the laboratory data, there is a tendency of lower hemoglobin and higher blood urea nitrogen in the intermediate-risk and high-risk groups compared with the low-risk group. Plasma levels of BNP were higher in the high-risk and intermediate-risk groups compared with the lowrisk group (p < 0.05). Echocardiography showed that systolic function and left ventricular chamber size did not differ among the three groups. The frequency of beta-blocker intake tended to be higher in the low-risk group compared with the other two groups. Additionally, there was a weak correlation between serum albumin levels and lymphocyte counts (r = 0.103, p = 0.032), and serum albumin levels and lymphocyte counts were weakly associated with plasma BNP levels (r = 0.099, p = 0.041 and r = 0.122, p = 0.011, respectively). Relationship between nutritional status and clinical outcomes Kaplan–Meier survival analysis revealed differences in the frequency of combined clinical events among the groups (p < 0.001, Fig. 3). Table 2 shows the predictive values of the CART-NS for combined clinical events. Model 2, which is adjusted for age, gender, and log BNP levels, shows that the HR was 2.13 (95% CI: 1.42–3.16, p < 0.001) between the low-risk and high-risk groups and 1.99 (95% CI: 1.17–3.46, p = 0.010) between the intermediate-risk and high-risk groups. Table 3 shows the results of the univariate and multivariate analyses. The multivariate analysis shows that the CART-NS algorithm has strong predictive power, even after adjusting for other factors. Improved predictability of clinical events by the combined use of serum albumin levels and lymphocyte counts Table 4 shows the sensitivity and specificity of low serum albumin levels (<3.5 g/dl), low lymphocyte counts (<963/mm3) and high-risk classification in CART-NS for combined clinical events. Use of a single criterion (low lymphocyte count or low serum albumin level) predicted combined clinical events with a sensitivity of 41.7% and 42.4% and specificity of 73.4% and 70.2%, respectively. The use of CART-NS (high-risk in CART-NS) increased specificity to 89.5% with comparable sensitivity of 36.3%. Risk stratification by nutritional screening in patients with de novo hospitalization vs re-hospitalization because of acute heart failure
After exclusion(n=432) Lym ≥ 963/mm3
3
Serum albumin < 3.5 g/dl High risk n=67
Fig. 2. Flow chart of enrolled patients. The flow chart shows the enrolment of acute heart failure syndrome (AHFS) patients and cut-off criteria for lymphocyte counts (Lym) and serum albumin levels assigned to the stratified groups: low-risk group (lymphocyte counts 963/mm3), intermediate-risk group (lymphocyte counts <963/mm3 and serum albumin levels 3.5 g/dl) and high-risk group (lymphocyte counts <963/mm3 and serum albumin levels <3.5 g/dl). Exclusion criteria of the patients with abnormal values of white blood cell (WBC): 12,000/mm3 or <4000/ mm3.
Of the enrolled AHFS patients, the number of the patients who were de novo hospitalized or had already past history of hospitalization due to HF was 232 and 200, respectively, and different distributions of stratified groups were observed between the de novo hospitalization and re-hospitalization groups because of AHFS, as shown in Fig. 4 (p = 0.042). Additionally, we could identify the cause of death in 75 patients (72% of overall deceased patients). Of the patients (n = 75), cardiovascular death occurred in 53 patients (71%), and non-cardiovascular death occurred in the remaining patients (22 patients, 29%). The sub-group analysis using CART-NS demonstrated that frequency of non-cardiovascular death tended to be higher in the high-risk group (n = 10, 42%) compared with the low-risk group (n = 8, 21%), although there was no statistical difference among the groups (p = 0.190), as shown in Supplemental Figure. Nonetheless, in the high-risk group, cardiovascular death accounted for more than a half of the causes of death (n = 14, 58%).
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx
4
Table 1 Baseline characteristics in acute heart failure syndrome patients. Low risk Pts number Age (year) Male (%) Body mass index (kg/m2) NYHA III–IV (%) History Prior MI (%) HF hospitalization (within six months) (%) Hypertension (%) DM/IGT (%) Etiology Ischemic cardiomyopathy (%) Non-ischemic cardiomyopathy (%) Valvular (%) Hypertensive (%) Others (%) Vital signs on admission Systolic blood pressure (mmHg) Heart rate (beat per minute) Laboratory data on admission WBC count (/mm3) Lymphocyte count (/mm3) Hemoglobin (g/dl) Serum albumin (g/dl) BUN (mg/dl) eGFR (ml/min/1.73 m2) C-reactive protein (mg/dl) BNP (pg/ml) Echocardiography LVEDd (mm) %FS (%) Medications ACEi or ARB (%) Beta-blockers (%) Loop diuretics (%) Spironolactone (%)
297 73 (64–82) 66 22.6 (20.5–25.0) 89
Intermediate risk
High risk
68 77 (64–82)* 68 21.9 (19.1–24.6) 92
67 77 (68–85)* 63 21.7 (18.9–25.5) 91
26 13 69 34
30 15 72 40
35 21 72 50
32 22 30 13 3
36 16 41 6 2
39 20 20 13 8
140 (117–163) 94 (77–120) 7400 (5800–9150) 1482 (1190–1950) 12.5 (10.9–14.1) 3.7 (3.4–3.9) 20 (16–28) 54 (39–69) 0.41 (0.17–1.43) 660 (369–1162)
137 (115–153) 86 (70–104)* 6450 (5150–7375)* 755 (504–856) 11.1 (9.7–12.5)* 3.8 (3.6–4.0) 27 (18–38)* 41 (30–55)* 0.59 (0.19–1.33) 835 (375–1412)*
132 (117–163) 87 (74–102) 6600 (5400–10,280) 720 (536–845) 10.9 (9.6–12.1)* 3.1 (2.9–3.3) 27 (20–38)* 42 (28–56)* 2.18 (0.46–7.11)*,# 1044 (431–1783)*
56 (47–62) 18 (13–26)
55 (46–62) 20 (15–28)
57 (51–62) 18 (13–27)
67 63 82 37
63 52 70 45
60 52 81 41
Pts, patients; NYHA, New York Heart Association; MI, myocardial infarction; HF, heart failure; DM, diabetes mellitus; IGT, impaired glucose tolerance; WBC, white blood cell count; BUN, blood urea nitrogen; eGFR, estimated glomerular filtration rate; BNP, brain natriuretic peptide; LVEDd, left ventricular end-diastolic diameter; FS, fractional shortening; ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor antagonists. Values are expressed as the median (interquartile range, IQR) or percentage. Lymphocyte counts and serum albumin levels were not selected for the analysis owing to the category criteria. Analysis is unadjusted. * p < 0.05 vs low-risk group. # p < 0.05 vs intermediate-risk group.
Discussion In a cross-sectional and prospective analysis of 432 AHFS patients, we demonstrated that initial nutritional screening using CART analysis can stratify risk levels for further adverse clinical events over three years. In addition, the CART analysis determined the cut-off values of serum albumin levels and lymphocyte counts. According to this algorithm, we found that a substantial proportion of AHFS patients were at risk of malnutrition on admission (15.5%), and this state was strongly associated with poor clinical outcomes. Although the severity of HF indicated by NYHA class was similar among all three groups, the highest frequency of combined clinical events was observed in the high-risk group. The multivariate analysis revealed that classification as high-risk was strongly associated with the frequency of clinical events over the three years following admission, even after adjusting for age, gender, and plasma BNP levels. Therefore, we demonstrated that an initial nutritional assessment in the acute phase of AHFS can provide accurate predictions of future clinical outcomes. Interestingly, differences in body size were relatively small among the groups in the present study. The finding suggests that our algorithm can detect early warning signs of malnutrition in patients with HF before body weight loss is apparent, such as in cardiac cachexia. We also can raise the significance of non-cardiac factors, such as
nutritional status, for determining systemic pathophysiology in HF patients. Impact of nutritional screening in patients with AHFS in emergency room There is accumulating evidence that malnutrition is common and affects disease progression in patients with chronic HF [1– 3]. Importantly, cardiac cachexia, an advanced manifestation of nutritional impairment, has been recognized as an important factor in deteriorating HF [22]. However, a definition of malnutrition in patients with HF has not been established. Although it is widely known that body weight loss is an important indicator of advanced HF, body weight alone may be insufficient to predict the risk of progression to cardiac cachexia. Indeed, CART-NS was a better predictor than BMI in the present study (Table 2). Based on previous studies, we selected serum albumin levels and lymphocyte count, and both parameters were tested in emergency room settings. Subsequently, we performed CART analysis (Fig. 1) to determine the patient’s nutritional status in a clinical setting. The classification showed that on admission, a substantial proportion of the AHFS patients were at risk of malnutrition (15.5%), which is associated with poor clinical outcomes. The present algorithm has several strengths for assessing nutritional status. First, the required
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx Table 4 Sensitivity and specificity of the variables for combined clinical events.
100
Event-Free Rate (%)
5
80 3
Low lymphocyte count (<963/mm ) Low serum albumin (<3.5 g/dl) Nutritional risk tree (high risk)
60
Sensitivity (%)
Specificity (%)
41.7 42.4 36.3
73.4 70.2 89.5
40 High risk group 20
Intermediate risk group Low risk group
0 0
100
200
300
400
500
600
700
800
900 1000
Time to Event (days) Low risk group Int risk group High risk group
297 68 67
264 54 48
187 33 36
119 23 17
Fig. 3. Kaplan–Meier analysis of patients with acute heart failure syndrome stratified according to classification and regression tree for nutritional status (CART-NS). Rate of combined clinical events as determined by the Kaplan–Meier analysis of patients classified according to the CART-NS algorithm. Green line, red line, and blue line indicate the low-risk group (lymphocyte counts 963/mm3), intermediate-risk group (lymphocyte counts <963/mm3, serum albumin levels 3.5 g/dl) and high-risk group (lymphocyte counts <963/mm3, serum albumin levels <3.5 g/dl), respectively. Kaplan–Meier survival analysis revealed differences in the frequency of combined clinical events among the groups (p < 0.001).
Table 2 Predictive values of the classification and regression tree for nutritional status for combined clinical events.
Low risk vs high risk Model 1 Model 2 Intermediate risk vs high risk Model 1 Model 2
HR (95% CI)
p-Value
2.33 (1.56–3.44) 2.13 (1.42–3.16)
<0.001 <0.001
2.11 (1.25–3.67) 1.99 (1.17–3.46)
0.005 0.010
Model 1 was adjusted for age and gender, and Model 2 was adjusted for age, gender and log B-type natriuretic peptide levels. HR, hazard ratio.
blood tests can be performed easily, even in the emergency room, and the results can be obtained within a short time. Second, our algorithm is strongly associated with future clinical outcomes. Moreover, serum albumin levels have been shown to be influenced by hepatic synthesis, whereas lymphocyte counts are not influenced by this factor. Indeed, the correlation between serum albumin levels and lymphocyte counts was weak in the present study. Thus, combining these criteria can be recommended to
Fig. 4. Risk stratification by nutritional screening in patients with de novo hospitalization vs re-hospitalization because of acute heart failure syndrome (AHFS). Of the enrolled patients, de novo hospitalizations because of AHFS (de novo HOSP) occurred in 232 patients and re-hospitalization because of AHFS (Re-HOSP) occurred in 200 patients. In the de novo HOSP group, the patients assigned to lowrisk (green square), intermediate-risk (red square) and high-risk groups (blue square) accounted for 73%, 12%, and 14%, whereas in the Re-HOSP group, patients assigned to the low-risk, intermediate-risk and high-risk groups accounted for 63%, 20%, and 17%, respectively. Different distributions of stratified groups were observed between the de novo HOSP and Re-HOSP groups because of AHFS (p = 0.042).
improve the accuracy of clinical risk predictions, and in this study, the combined use of both criteria increased specificity to 89.5% with comparable sensitivity of 36.3% compared with their use as single criteria (serum albumin level, 42.4% and 70.2%, respectively; lymphocyte count, 41.7% and 73.4%, respectively). Third, the associations of these factors with plasma BNP levels are weak; thus, we can evaluate non-cardiac factors that represent different aspects of HF pathophysiology from circulating BNP levels. Therefore, we demonstrated the potential of initial assessment of nutritional status for risk stratification in AHFS patients. Mechanism of malnutrition in HF patients Although it is widely known that hypo-albuminemia can serve as a biomarker for prolonged malnutrition, serum albumin levels are influenced not only by food intake but also by hepatic synthesis, protein degradation, and increased vascular permeability [8]. Moreover, studies have suggested that hypo-albuminemia in HF patients could be caused by increased inflammatory activity in addition to decreased intake of food [6,23]. In clinical studies, increased circulating inflammatory cytokines and chemokines have been significantly correlated with increasing HF severity or deteriorating left ventricular systolic function, suggesting that systemic inflammatory activity plays an important role in HF
Table 3 Univariate/multivariate analysis for combined clinical events. Individual measures Age (per year increase) BMI (per kg/m2 increase) Log BNP eGFR (per ml/min/1.73 m2 increase) BUN (mg/dl increase) Hb (mg/dl increase) CART-NS (intermediate vs high) CART-NS (low vs high)
Univariate HR (95% CI) 1.01 0.92 2.71 0.99 1.02 0.90 2.14 2.58
(1.00–1.02) (0.88–0.96) (1.73–4.28) (0.98–0.99) (1.01–1.02) (0.84–0.97) (1.27–3.73) (1.73–3.78)
p-Value 0.005 <0.001 <0.001 0.001 <0.001 0.006 0.004 <0.001
Multivariate HR (95% CI) 1.02 0.96 1.89 1.00 1.01 1.02 2.27 2.14
(1.01–1.04) (0.91–1.00) (1.17–3.10) (0.99–1.01) (1.00–1.02) (0.93–1.11) (1.32–4.02) (1.38–3.27)
p-Value 0.006 0.062 0.009 0.677 0.016 0.697 0.003 0.001
HR, hazard ratio; CI, confidence interval; BMI, body mass index; BNP, B-type natriuretic peptide; eGFR, estimated glomerular filtration rate; BUN, blood urea nitrogen; Hb, hemoglobin; CART-NS, classification and regression trees tool for nutritional status.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 6
M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx
patients [24–28]. Interestingly, it has also been reported that tumor necrosis factor alpha (TNF-a) is significantly increased in patients with cardiac cachexia [29]. Experimental studies have demonstrated that inflammatory cytokines (e.g. TNF-a, interleukin-1) and chemokines (e.g. monocyte chemotactic protein 1) directly depress cardiac function and were related to impaired nutritional status [30–33]. The evidence also suggests a strong association between a systemic inflammatory state and malnutrition, both of which might play a role in aggravating HF. Indeed, the patients in the high-risk group in the present study showed the lowest levels of hemoglobin and highest levels of C-reactive protein among the groups. Because this was an observational study, further investigation is necessary to address this issue. Another possible mechanism of malnutrition in HF patients is related to the aggravation of nutritional status by protracted HF itself. Indeed, the high-risk group included a greater number of rehospitalized patients than de novo HF hospitalized patients. Several interplaying factors may be involved in the relationship between nutritional state and HF aggravation. Decreased food intake is often observed in hospitalized patients with HF, which may also cause malnutrition in patients with HF. Patients with chronic HF and cachexia exhibit gastrointestinal fat malabsorption [30]. It should also be noted that statistical relationships of plasma levels of BNP with nutritional parameters, such as plasma albumin levels and lymphocyte counts, were weak. This finding also implies that malnutrition and plasma BNP are indicative of different aspects of worsening HF. Indeed, a frequency of non-cardiovascular death tended to be higher in the high-risk group, as shown in the Supplemental Figure. Furthermore, it is widely believed that HF may activate inflammation, potentially causing vicious cycles in HF patients. The interplay among malnutrition, inflammation, and deterioration of HF per se may increase not only cardiovascular events but also non-cardiovascular death. Because the present study was observational, further investigation is necessary to clarify the temporal relationship between HF and worsening nutritional status. This study had several limitations. First, the ‘gold-standard’ evaluation for nutritional status in patients with HF has not been established. Nonetheless, BMI at hospital discharge tended to be lower in the high-risk group than the low-risk group (19.0 vs 20.3, respectively p = 0.057). These findings suggest that our algorithm is associated with future body weight changes. Second, this study was based on a single-center registry with a moderate number of enrolled patients. Third, although differences in the frequency of beta-blocker intake were found between the low-risk group and the other two groups, the reason for this finding remains unclear. In the high-risk group, the patients were older than in the other two groups. We could not exclude the possibility that these patient characteristics might influence beta-blocker use in the population. An association of low relative lymphocyte counts with mortality in HF has also been reported, suggesting a relationship with elevated serum cortisol [10]. In the present study, we did not obtain cortisol measurements in all patients; therefore, we did not address this relationship. Additionally, we did not include cholesterol values in the algorithm because it was unable to evaluate the details of food intake or effects of medications, such as statins in all AHFS patients in emergency clinical settings. Additionally, we excluded the patients with abnormal values of WBC counts indicating severe infectious disease. In the patients with mild degree of infectious diseases, unless they influence the WBC count values, theoretically our algorithm (CART-NS) can be applicable to these patients. However, since it is difficult to distinguish infection and AHFS only by WBC counts [34] in acute phases, careful interpretation and follow-up would be necessary to apply this algorithm for patients with symptoms of serious infection representing abnormal values of WBC counts in the clinical setting. Lastly, we could not identify
the precise cause of death in 22 patients who did not die in our hospital. Conclusions On admission, a substantial proportion of AHFS patients are at risk of malnutrition, which is associated with poor clinical outcomes. We demonstrated that nutritional screening on admission can be used to stratify risk levels for further clinical events in AHFS patients despite limited differences in body size among the patient groups. These findings suggest that our algorithm can detect early warning signs of malnutrition in patients with HF before body weight loss is apparent. More importantly, we determined that this novel algorithm for nutritional screening can be performed within a short time and can even be completed in an emergency room. Where is Figure 4? I cannot find the Figure 4. Conflicts of interest The authors declare that there is no conflict of interest. Funding sources The present study was supported by the Intramural Research Fund, grant number 24-4-1, for Cardiovascular Diseases of the National Cerebral and Cardiovascular Center. Disclosure No financial supports exist in this study. Acknowledgment The authors thank Yoko Sumita, our secretary, for assistance in managing data.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jjcc.2016.05.004. References [1] Schwengel RH, Gottlieb SS, Fisher ML. Protein-energy malnutrition in patients with ischemic and nonischemic dilated cardiomyopathy and congestive heart failure. Am J Cardiol 1994;73:908–10. [2] Berger MM, Mustafa I. Metabolic and nutritional support in acute cardiac failure. Curr Opin Clin Nutr Metab Care 2003;6:195–201. [3] Carr JG, Stevenson LW, Walden JA, Heber D. Prevalence and hemodynamic correlates of malnutrition in severe congestive heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol 1989;63:709–13. [4] Sarma S, Gheorghiade M. Nutritional assessment and support of the patient with acute heart failure. Curr Opin Crit Care 2010;16:413–8. [5] Omran ML, Morley JE. Assessment of protein energy malnutrition in older persons, Part II: Laboratory evaluation. Nutrition 2000;16:131–40. [6] Horwich TB, Kalantar-Zadeh K, MacLellan RW, Fonarow GC. Albumin levels predict survival in patients with systolic heart failure. Am Heart J 2008;155:883–9. [7] Bonilla-Palomas JL, Gamez-Lopez AL, Moreno-Conde M, Lopez-Ibanez MC, Anguita-Sanchez M, Gallego de la Sacristana A, Garcia-Catalan F, Villar-Raez A. Hypoalbuminemia in acute heart failure patients: causes and its impact on hospital and long-term mortality. J Card Fail 2014;20:350–8. [8] Ballmer PE. Causes and mechanisms of hypoalbuminaemia. Clin Nutr 2001;20:271–3. [9] Dzieniszewski J, Jarosz M, Szczygiel B, Dlugosz J, Marlicz K, Linke K, Lachowicz A, Ryzko-Skiba M, Orzeszko M. Nutritional status of patients hospitalised in Poland. Eur J Clin Nutr 2005;59:552–60. [10] Ommen SR, Hodge DO, Rodeheffer RJ, McGregor CG, Thomson SP, Gibbons RJ. Predictive power of the relative lymphocyte concentration in patients with advanced heart failure. Circulation 1998;97:19–22.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004
G Model
JJCC-1316; No. of Pages 7 M. Fujino et al. / Journal of Cardiology xxx (2016) xxx–xxx [11] Huehnergarth KV, Mozaffarian D, Sullivan MD, Crane BA, Wilkinson CW, Lawler RL, McDonald GB, Fishbein DP, Levy WC. Usefulness of relative lymphocyte count as an independent predictor of death/urgent transplant in heart failure. Am J Cardiol 2005;95:1492–5. [12] Acanfora D, Gheorghiade M, Trojano L, Furgi G, Pasini E, Picone C, Papa A, Iannuzzi GL, Bonow RO, Rengo F. Relative lymphocyte count: a prognostic indicator of mortality in elderly patients with congestive heart failure. Am Heart J 2001;142:167–73. [13] Raposeiras-Roubin S, Abu-Assi E, Lopez-Lopez A, Bouzas-Cruz N, CastineiraBusto M, Cambeiro-Gonzalez C, Alvarez-Alvarez B, Virgos-Lamela A, Varela-Roman A, Garcia-Acuna JM, Gonzalez-Juanatey JR. Risk stratification for the development of heart failure after acute coronary syndrome at the time of hospital discharge: predictive ability of grace risk score. J Cardiol 2015;66:224–31. [14] Okayama D, Minami Y, Kataoka S, Shiga T, Hagiwara N. Thyroid function on admission and outcome in patients hospitalized for acute decompensated heart failure. J Cardiol 2015;66:205–11. [15] Fonarow GC, Adams Jr KF, Abraham WT, Yancy CW, Boscardin WJ. Risk stratification for in-hospital mortality in acutely decompensated heart failure: classification and regression tree analysis. JAMA 2005;293:572–80. [16] Ho KK, Anderson KM, Kannel WB, Grossman W, Levy D. Survival after the onset of congestive heart failure in Framingham heart study subjects. Circulation 1993;88:107–15. [17] Takahama H, Yokoyama H, Kada A, Sekiguchi K, Fujino M, Funada A, Amaki M, Hasegawa T, Asakura M, Kanzaki H, Anzai T, Kitakaze M. Extent of heart rate reduction during hospitalization using beta-blockers, not the achieved heart rate itself at discharge, predicts the clinical outcome in patients with acute heart failure syndromes. J Cardiol 2013;61:58–64. [18] Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, Sevransky JE, Sprung CL, Douglas IS, Jaeschke R, Osborn TM, Nunnally ME, Townsend SR, Reinhart K, Kleinpell RM, et al. Surviving sepsis campaign: international guidelines for management of severe sepsis and septic shock: 2012. Crit Care Med 2013;41:580–637. [19] Matsuo S, Imai E, Horio M, Yasuda Y, Tomita K, Nitta K, Yamagata K, Tomino Y, Yokoyama H, Hishida A. Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis 2009;53:982–92. [20] Friedman JH, Meulman JJ. Multiple additive regression trees with application in epidemiology. Stat Med 2003;22:1365–81. [21] Rovlias A, Kotsou S. Classification and regression tree for prediction of outcome after severe head injury using simple clinical and laboratory variables. J Neurotrauma 2004;21:886–93.
7
[22] Anker SD, von Haehling S. Inflammatory mediators in chronic heart failure: an overview. Heart 2004;90:464–70. [23] Arques S, Roux E, Stolidi P, Gelisse R, Ambrosi P. Usefulness of serum albumin and serum total cholesterol in the prediction of hospital death in older patients with severe, acute heart failure. Arch Cardiovasc Dis 2011;104:502–8. [24] Yndestad A, Damas JK, Oie E, Ueland T, Gullestad L, Aukrust P. Systemic inflammation in heart failure – the whys and wherefores. Heart Fail Rev 2006;11:83–92. [25] Aukrust P, Ueland T, Lien E, Bendtzen K, Muller F, Andreassen AK, Nordoy I, Aass H, Espevik T, Simonsen S, Froland SS, Gullestad L. Cytokine network in congestive heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol 1999;83:376–82. [26] Aukrust P, Ueland T, Muller F, Andreassen AK, Nordoy I, Aas H, Kjekshus J, Simonsen S, Froland SS, Gullestad L. Elevated circulating levels of C–C chemokines in patients with congestive heart failure. Circulation 1998;97: 1136–43. [27] Damas JK, Gullestad L, Ueland T, Solum NO, Simonsen S, Froland SS, Aukrust P. Cxc-chemokines, a new group of cytokines in congestive heart failure – possible role of platelets and monocytes. Cardiovasc Res 2000;45:428–36. [28] Testa M, Yeh M, Lee P, Fanelli R, Loperfido F, Berman JW, LeJemtel TH. Circulating levels of cytokines and their endogenous modulators in patients with mild to severe congestive heart failure due to coronary artery disease or hypertension. J Am Coll Cardiol 1996;28:964–71. [29] Torre-Amione G, Kapadia S, Lee J, Durand JB, Bies RD, Young JB, Mann DL. Tumor necrosis factor-alpha and tumor necrosis factor receptors in the failing human heart. Circulation 1996;93:704–11. [30] Mann DL. Inflammatory mediators and the failing heart: past, present, and the foreseeable future. Circ Res 2002;91:988–98. [31] Aukrust P, Gullestad L, Ueland T, Damas JK, Yndestad A. Inflammatory and anti-inflammatory cytokines in chronic heart failure: potential therapeutic implications. Ann Med 2005;37:74–85. [32] Mann DL. Stress-activated cytokines and the heart: from adaptation to maladaptation. Annu Rev Physiol 2003;65:81–101. [33] Okonko DO, Anker SD. Anemia in chronic heart failure: pathogenetic mechanisms. J Card Fail 2004;10:S5–9. [34] Maisel A, Neath SX, Landsberg J, Mueller C, Nowak RM, Peacock WF, Ponikowski P, Mockel M, Hogan C, Wu AH, Richards M, Clopton P, Filippatos GS, Di Somma S, Anand I, et al. Use of procalcitonin for the diagnosis of pneumonia in patients presenting with a chief complaint of dyspnoea: results from the BACH (Biomarkers in Acute Heart Failure) trial. Eur J Heart Fail 2012;14:278–86.
Please cite this article in press as: Fujino M, et al. Risk stratification based on nutritional screening on admission: Three-year clinical outcomes in hospitalized patients with acute heart failure syndrome. J Cardiol (2016), http://dx.doi.org/10.1016/j.jjcc.2016.05.004