The controlling nutritional status score predicts outcomes of cardiovascular events in patients with heart failure with preserved ejection fraction.

BACKGROUNDS: The relationship between cardiovascular outcomes and the Controlling Nutritional Status (CONUT) score in heart failure (HF) with preserved ejection fraction (HFpEF) patients is unknown. This study aimed to evaluate the relationship between the score and cardiovascular outcomes in HFpEF patients. METHODS AND
RESULTS: A total of 506 consecutive HFpEF patients were prospectively observed for up to 1500 days or until the occurrence of cardiovascular events. The mean age was 71.6 ± 9.4 years. Cardiovascular outcomes were compared between the CONUT score 0-1 group with a normal nutritional state (normal group), the CONUT score 2-4 group with a light degree of undernutrition (light group), and the CONUT score 5-8 group with a moderate degree of undernutrition (moderate group). In this study, there were no patients who scored 9-12, which was defined as a severe state of undernutrition. Overall, 238 cardiovascular events were observed during the follow-up period (median: 1159 days). Kaplan-Meier analysis showed that the moderate group was at higher risk of composite cardiovascular events than the normal group (P < 0.001) and the light group (P = 0.031). The analysis also showed that the light group was at higher risk of composite cardiovascular events than the normal group (P = 0.038). Multivariable Cox proportional hazards analysis with the significant factors from the univariate analysis showed that the CONUT score (hazard ratio: 1.12, 95% confidence interval: 1.03-1.21, P = 0.005) significantly predicted future cardiovascular events.
CONCLUSION: Nutritional screening using the CONUT score may be useful for predicting cardiovascular events in HFpEF patients.

Introduction

For the purpose of treating heart failure (HF), it is important to (1) suppress the progression of cardiac dysfunction, (2) improve symptoms, exercise skills and quality of life (QOL), and (3) prevent readmission and improve the life prognosis [1]. In the progressive stage of HF, physical function and nutritional status change in parallel, and the nutritional status of patients with HF worsens toward the end of life. Serum albumin and cholesterol levels, commonly used as indicators of nutritional status, are predictors of prognosis independent of age and severity of HF [2], [3]. It has been reported that hypoalbuminemia progression after hospitalization is associated with a worse prognosis in patients with acute HF [4].

Serum albumin has a long half-life of approximately 20 days, is susceptible to invasion and is not suitable for individuals evaluation. For this reason, the Controlling Nutritional Status (CONUT) method was developed as a tool to evaluate nutritional status using three biomarkers: protein metabolism, immunocompetence, and lipid metabolism [5]. The serum albumin level reflects protein metabolism, the total lymphocyte count reflects immunity, and the total cholesterol level reflects lipid metabolism; these items are scored, and the nutritional status is comprehensively and pleiotropically evaluated according to the three biological indices. The CONUT index has been reported as a useful index for the early screening of malnutrition in patients with HF [5]. It has been reported that undernutrition in chronic HF patients could be evaluated with the CONUT index and was correlated with subsequent cardiovascular events [6], and a correlation was found between the CONUT score and mortality in patients with HF [7], [8].

Accumulating clinical studies have demonstrated that HF with reduced left ventricular (LV) ejection fraction (EF) (HFrEF) and HF with preserved LVEF (HFpEF) are separate pathological conditions because of differences in survival rates [9], [10] and effective drug therapies; thus, we have proposed that HFrEF and HFpEF patients should be managed differently [11].

In the present study, we sought to evaluate the relationship between the CONUT score and cardiovascular outcomes in patients with HFpEF.

Methods

This study was a prospective, single-center, observational study.

Ethics statement

All procedures were conducted in accordance with the Declaration of Helsinki and its amendments. The study protocol was approved by the institutional review board of Kumamoto University (approval number, Senshin 2225). This study is registered at the University Hospital Medical Information Network (UMIN) Clinical Trials Registry (UMIN000036884). Opt-out materials are available at http://www.kumadai-junnai.com/home/wp-content/uploads/houkatsu.pdf.

Study design and patients

We prospectively investigated 948 consecutive patients with HF who were hospitalized in Kumamoto University Hospital between January 2007 and September 2013. We recorded each patient’s medical history and relevant clinical characteristics. We excluded 440 patients for the following reasons: severe valvular disease (n = 118), chronic renal failure requiring hemodialysis (n = 65), systemic inflammatory disease (n = 5), acute renal failure with dehydration (n = 1), and not meeting the diagnostic criteria for HFpEF as subsequently described (including HF with a reduced LVEF [HFrEF]; n = 251). Finally, 506 patients of the remaining 508 HFpEF patients, excluding those with insufficient data, were enrolled in this study. We subsequently calculated the CONUT score in these HFpEF patients, and the subjects were subdivided into normal- (0–1), light- (2–4), moderate- (5–8), and severe-score (9–12) groups according to the original concepts of CONUT score[5], with the occurrence of cardiovascular events followed for up to 1500 days. The study flow chart is shown in Fig. 1.

Fig. 1

Flow chart showing the enrollment protocol. Abbreviations: HF, heart failure; HFpEF, HF with preserved left ventricular ejection fraction; CONUT, Controlling Nutritional Status.

Definition of HFpEF

HFpEF was clinically defined according to the European Society of Cardiology task force as follows:

symptoms or signs of HF;

normal or mildly reduced LVEF (LVEF > 50% and LV end-diastolic volume index < 97 mL/m2);

evidence of abnormal LV relaxation, filling, diastolic distensibility, and diastolic stiffness.

We excluded HFpEF patients who had shown even a transient reduction in ejection fraction. Hence, HFpEF patients whose LVEF was <50% and was improved by optimal medical therapy were not included in the present study. In our study, we stratified patients by the E/e′ ratio, grouped by either a ≥15 ratio or >8 but <15 ratio, and by plasma B-type natriuretic peptide (BNP) levels, with a cut-off of 100 pg/mL. Physicians further confirmed that patients had HF by determining the New York Heart Association (NYHA) functional class [12], which was assessed by the standard questionnaire while the patient was in a stable condition after optimal therapy.

Calculation of the CONUT score

The CONUT score was calculated as described previously [5]. In brief, 3 parameters were used to calculate the score: (i) serum albumin level (g/dL); (ii) total cholesterol level (mg/dL); and (iii) total lymphocyte count (count/mL) (Supplemental Table 1). Thus, the CONUT score enables assessment of protein reserves, caloric depletion, and immune defenses in each patient.

Clinical parameters

The clinical parameters were described previously [11], [13], [14], [15], [16], [17]. In brief, the baseline demographic data, cardiovascular risk factors, and medications on discharge were documented. Hypertension was defined as a recorded blood pressure >140/90 mmHg or the use of any antihypertensive medications as described previously. Diabetes mellitus (DM) was defined as the presence of symptoms of diabetes and a random plasma glucose concentration ≥200 mg/dL, fasting plasma glucose concentration ≥126 mg/dL, or a 2-hr plasma glucose concentration ≥200 mg/dL according to a 75 g oral glucose tolerance test or the use of any medications for DM. Dyslipidemia was defined as low-density lipoprotein levels ≥140 mg/dL (≥3.63 mmol/L), high-density lipoprotein levels <40 mg/dL (1.04 mmol/L), triglycerides ≥150 mg/dL (≥1.7 mmol/L) or the use of medications for dyslipidemia.

Echocardiographic examinations

Echocardiography was performed while the patient was in a stable condition on admission by experienced cardiac sonographers who had no knowledge of the study data. Left ventricular ejection fraction (LVEF) was measured using a modified Simpson’s method. The LVEF and the ratio of early transmitral flow velocity to early diastolic mitral annular velocity (E/e′), which was assessed by tissue Doppler, were measured by echocardiography (Vivid 7®; GE-Vingmed Ultrasound, Horton, Norway; Aplio XG®; Toshiba, Tokyo, Japan) as previously reported [11], [14], [16].

Biomarker measurement

Blood samples were obtained in stable and fasting conditions in the early morning. The patient’s BNP levels were analyzed using a commercially available assay (Abbott Japan, Matsudo, Japan) in the hospital clinical laboratory on admission. The BNP levels were transformed into natural logarithmic levels (ln-BNP) to achieve a normal distribution. The estimated glomerular filtration rate (eGFR) was calculated using the Japanese Society of Nephrology formula [18].

Follow-up and outcomes

Patients were followed up prospectively at our outpatient clinics or by the primary care physician every month until July 2017 or until the occurrence of a cardiovascular event, including the following: cardiovascular death, hospitalization for HF decompensation, nonfatal myocardial infarction (MI), unstable angina pectoris, coronary revascularization for a new diagnosis of angina or in-stent restenosis after percutaneous coronary intervention, and nonfatal ischemic stroke. Cardiovascular death was defined as death within 30 days of documented sudden death without apparent noncardiovascular causes, MI, death from HF, or death from stroke. Hospitalization for HF decompensation was defined if the patient was admitted for at least an overnight stay in the hospital because of HF with typical symptoms and had objective signs of worsening HF requiring intravenous drug administration. MI was diagnosed by an increase or decrease in cardiac biomarkers (plasma creatine kinase-MB or cardiac troponin) above the 99th percentile of the upper limit of the normal range together with evidence of myocardial ischemia and at least 1 of the following symptoms: electrocardiographic changes (new ST-T changes, left bundle branch block, or pathological Q wave) or imaging evidence of new viable myocardial loss, or a new regional wall motion abnormality [19]. Unstable angina pectoris was diagnosed according to new or accelerating symptoms of myocardial ischemia accompanied by new ischemic ST-T-wave changes. Ischemic stroke was diagnosed according to the documented focal neurological deficit with radiological evidence of brain infarction excluding intracranial hemorrhage. Cardiovascular events were ascertained by reviewing medical records and were confirmed by direct contact with the patients, their families, and physicians or by annual telephone interview with each patient. An Events Committee comprising at least 3 independent physicians reviewed all events to avoid intraobserver biases.

Statistical analysis

Continuous variables are expressed as the mean ± standard deviation for normally distributed variables according to the Shapiro–Wilk test. Variables with a non-normal distribution are expressed as the median value with the interquartile range. Categorical variables are presented as frequencies and percentages. Differences between groups were determined using Fisher’s exact test for categorical variables. Differences in continuous variables were analyzed by the unpaired t test or the Mann–Whitney U test, as appropriate. Missing data were excluded from the analyses. A Kaplan–Meier curve was used to determine the cumulative incidence of composite cardiovascular events, and the log-rank test was used to compare the incidence of composite cardiovascular event groups. The Cox proportional hazards model was used to estimate composite cardiovascular event hazard ratios (HRs) by univariable and multivariable analyses with forced inclusion modeling. HRs and 95% confidence intervals (CIs) are presented. Furthermore, the estimates of C-statistics in the Cox proportional hazards regression models were compared after the addition of higher CONUT scores to conventional factors identified in the subanalysis of the Irbesartan in Patients with Heart Failure and Preserved Ejection Fraction (I-PRESERVE) trial (age, previous hospitalization for HF, diabetes mellitus, and ln-BNP; 4 prognostic factors [PF4]) [20]. We also assessed the incremental effects of adding a higher CONUT score to the PF4 to predict composite cardiovascular events using the net reclassification index (NRI). A P value <0.05 was considered statistically significant. The Harrell’s C-statistic, NRI and IDI were performed by R package of PredictABEL. The software Statistical Package for Social Sciences (SPSS) ver. 26.0 (IBM Japan, Tokyo, Japan) was used for other statistical analyses.

Results

Clinical characteristics of enrolled patients with HFpEF

A total of 506 patients with HFpEF were enrolled in this study. The numbers of patients (percentage) with normal, light, moderate, and severe scores were 231 (45.6%), 226 (44.6%), 49 (9.4%), and 0 (0%), respectively. Fig. 2 shows the distribution of CONUT scores among HFpEF patients. The baseline characteristics of the HFpEF patients are shown in Table 1. Overall, the patients had a mean age of 71.6 ± 9.4 years, and 54.7% were male. The prevalence of ischemic heart disease (IHD) was significantly lower in both the light group and moderate group than in the normal group. Plasm BNP levels were significantly higher, and the prevalence of DM, hypertension and dyslipidemia and the use of beta-blockers were significantly lower in the light group than in the normal group. While the tricuspid regurgitation pressure gradient (TR-PG) and pulmonary artery systolic pressure (PAP) were significantly higher in the moderate group than in the normal group, NYHA class III or IV and transthoracic echocardiographic parameters, including LVEF, E/e′, stroke volume index (SVI), and left atrial diameter (LAD), were not significantly different between these groups.

Fig. 2

Distribution of the CONUT score. Blue indicates a CONUT score of 0–1 point (normal degree of undernutrition). Green indicates a CONUT score of 2–5 points (light degree of undernutrition). Red indicates a CONUT score of 6–8 points (moderate degree of undernutrition). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1

Baseline characteristics of HFpEF patients according to group determined by CONUT (controlling nutritional status) scores.

	All HFpEF patientsn = 506	Normal groupn = 231	Light groupn = 226	Moderate groupn = 49	P value
Age, years	71.6 ± 9.4	71.7 ± 9.4	71.0 ± 9.5	74.1 ± 8.6	0.114
Male, n (%)	277 (54.7)	133 (57.5)	117 (51.7)	27 (55.1)	0.461
BMI (kg/m2)	24.1 ± 3.6	24.5 ± 3.5	23.8 ± 3.8	23.4 ± 2.9	0.061
NYHA III or IV, n (%)	86 (16.9)	35 (15.1)	39 (17.2)	12 (24.4)	0.363
Diabetes mellitus, n (%)	156 (30.8)	92 (39.8)	52 (23.0)**	12 (24.4)	<0.001
Hypertension, n (%)	396 (78.2)	198 (85.7)	163 (72.1)**	35 (71.4)	0.001
Dyslipidemia, n (%)	393 (77.6)	192 (83.1)	163 (72.1)*	38 (77.5)	0.013
IHD, n (%)	266 (52.5)	145 (62.7)	100 (44.2)**	21 (42.8)*	<0.001
Atrial fibrillation, n (%)	145 (28.6)	54 (23.3)	74 (32.7)	17 (34.6)	0.054
SBP (mmHg)	130.1 ± 21.2	129.6 ± 20.1	130.2 ± 22.4	132.7 ± 20.7	0.646
DBP (mmHg)	71.0 ± 13.1	70.2 ± 12.3	71.9 ± 13.4	70.3 ± 12.8	0.328
Hemoglobin (g/dL)	12.7 ± 1.8	12.8 ± 1.8	12.7 ± 1.9	12.5 ± 1.7	0.508
hs-CRP (mg/L)	0.44 ± 1.9	0.27 ± 0.8	0.39 ± 2.0	1.5 ± 3.9	0.073
eGFR (mL/min/1.73 m2)	62.2 ± 19.5	62.0 ± 18.6	63.6 ± 20.6	56.9 ± 17.4	0.095
BNP (pg/mL)	175.9 ± 291.2	120.3 ± 184.0	225.2 ± 374.1**	209.3 ± 291.2	<0.001
LVEF (%)	62.7 ± 5.8	63.1 ± 5.3	62.1 ± 6.0	63.1 ± 6.8	0.120
SVI	40.2 ± 9.9	40.9 ± 9.7	39.8 ± 9.5	39.3 ± 12.5	0.410
LAD (mm)	39.5 ± 7.0	39.5 ± 7.3	39.2 ± 6.9	40.8 ± 6.6	0.391
E/e′	17.5 ± 5.0	17.0 ± 4.0	17.9 ± 5.7	18.3 ± 5.7	0.054
TR-PG (mmHg)	25.3 ± 8.0	24.6 ± 8.1	25.3 ± 7.4	28.6 ± 9.7 *	0.017
PAP (mmHg)	31.6 ± 9.1	30.5 ± 9.2	31.9 ± 8.6	35.5 ± 10.2**	0.006
Diuretics, n (%)	124 (24.5)	48 (20.7)	58 (25.6)	18 (36.7)	0.082
ACE-I or ARB, n (%)	317 (62.6)	159 (68.8)	124 (54.8)**	34 (69.3)	0.006
CCB, n (%)	294 (58.1)	139 (60.1)	122 (53.9)	33 (67.3)	0.154
Beta-blocker, n (%)	225 (44.4)	121 (52.3)	86 (38.1)**	18 (36.7)	0.006
Statin, n (%)	303 (59.8)	145 (62.7)	161 (71.2)**	29 (59.1)	<0.001
CONUT score	4.3 ± 1.2	0.5 ± 0.4	2.6 ± 0.7**	5.8 ± 0.8**,‡	<0.001

Data are presented as the mean ± SD, median (interquartile range), or number (percentage).

HFpEF, heart failure with preserved ejection fraction; BMI, body mass index; NYHA, New York Heart Association; IHD, ischemic heart disease; SBP, systolic blood pressure; DBP, diastolic blood pressure; hs-CRP, high sensitivity C reactive protein; eGFR, estimated glomerular filtration rate; BNP, brain natriuretic peptide; LVEF, left ventricular ejection fraction; SVI, stroke volume index; LAD, left atrium diameter; TR-PG, tricuspid regurgitation pressure gradient; PAP, pulmonary artery systolic pressure; ACE-I, angiotensin-converting enzyme inhibitor; ARB, angiotensin Ⅱ receptor blocker; CCB, calcium channel blocker

P < 0.01 vs. light group.

P < 0.05.

P < 0.01 vs. normal group.

Cardiovascular events at follow-up

Overall, 238 cardiovascular events were recorded during the follow-up period (median: 1159 days). Table 2 shows the details of cardiovascular events during follow-up. We found significantly higher rates of composite cardiovascular events in the patients in the moderate group than in the patients in the normal (P = 0.005) and light groups (P < 0.001) and significantly higher rates of cardiovascular death in the patients in the moderate group than in the patients in the normal and light groups (P < 0.001). Moreover, the rate of hospitalization for HF decompensation was significantly higher in the moderate group than in the normal group (P = 0.003).

Table 2

Cardiovascular events according to CONUT score.

	Total(n = 506)	Normal group(n = 231)	Light group(n = 226)	Moderate group(n = 49)	P value
Total cardiovascular events, n (%)	238 (47.0)	95 (41.1)	110 (48.6)	32 (65.3)**	0.005
Cardiovascular death, n (%)	31 (6.1)	7 (3.0)	10 (4.4)	14 (28.5)**,‡	<0.001
Hospitalization for HF decompensation, n (%)	110 (21.7)	36 (15.5)	62 (27.4)**	12 (24.4)	0.003
Non-fatal myocardial infarction, n (%)	6 (1.1)	2 (0.8)	2 (0.8)	2 (4.0)	0.544
Unstable angina pectoris, n (%)	15 (2.9)	8 (3.4)	5 (2.2)	1 (2.0)	0.811
Coronary revascularization, n (%)	61 (12.0)	32 (13.8)	27 (11.9)	2 (4.0)*	0.023
Nonfatal ischemic stroke, n (%)	15 (2.9)	10 (4.3)	4 (1.7)	1 (2.0)	0.278

P < 0.01 vs. light group.

P < 0.05.

P < 0.01 vs. normal group.

Kaplan–Meier curves

We performed a Kaplan–Meier analysis and observed that the moderate group was at higher risk of composite cardiovascular events than the normal group (P < 0.001; Fig. 3) and the light group (P = 0.031; Fig. 3). It also showed that the light group was at higher risk of composite cardiovascular events than the normal group (P = 0.038; Fig. 3).

Fig. 3

Kaplan–Meier analyses after 1500 days of follow-up for cardiovascular events according to CONUT scores. The 0 time point on the x-axis indicates the discharge day of the qualifying cardiovascular events.

Cox proportional hazards analyses

Table 3 shows the results of univariate and multivariable Cox proportional hazards analyses for cardiovascular events. Univariate Cox proportional hazards analysis identified age (HR: 1.01, 95% CI: 1.00–1.03, P = 0.031), previous hospitalization for HF (HR: 1.62, 95% CI: 1.19–2.21, P = 0.002), NYHA III or IV (HR: 1.80, 95% CI: 1.33–2.43, P < 0.001), hemoglobin (HR: 0.87, 95% CI: 0.81–0.93, P < 0.001), ln-BNP (HR: 1.16, 95% CI: 1.05–1.28, P = 0.003), LAD (HR: 1.01, 95% CI: 1.00–1.03, P = 0.047), PAP (HR: 1.01, 95% CI: 1.00–1.03, P = 0.018), diuretic usage (HR: 1.48, 95% CI: 1.12–1.95, P = 0.005) and CONUT score (HR: 1.16, 95% CI: 1.08–1.24, P < 0.001) as significant factors associated with cardiovascular events. In a multivariate Cox proportional hazard analysis including PF4 by forced entry methods (model 1), previous hospitalization for HF (HR: 1.42, 95% CI: 1.02–1.99, P = 0.036) and CONUT score (HR: 1.14, 95% CI: 1.06–1.22, P < 0.001) were independently and significantly associated with cardiovascular events. Multivariable Cox proportional hazards analysis using the abovementioned 5 significant factors from the univariate analysis (model 2) identified hemoglobin (HR: 0.90, 95% CI: 0.83–0.99, P = 0.033) and CONUT score (HR: 1.12, 95% CI: 1.03–1.21, P = 0.005) as independent predictors of cardiovascular events in patients with HFpEF.

Table 3

Cox proportional hazards regression analyses for cardiovascular outcome within 1500 days follow-up.

	Univariable Regression			Multivariable Regression
				Model 1 (I-PRESERVE)			Model 2
Variable	HR	95% CI	P value	HR	95% CI	P value	HR	95% CI	P value
Age (years)	1.01	1.00–1.03	0.031	1.01	0.99–1.02	0.075	1.00	0.99–1.02	0.433
Male sex (yes)	0.90	0.73–1.17	0.453
BMI (kg/m2)	0.96	0.93–1.00	0.091
Previous hospitalization for HF (yes)	1.62	1.19–2.21	0.002	1.42	1.02–1.99	0.036	1.47	0.97–2.24	0.068
NYHA III or IV (yes)	1.80	1.33–2.43	<0.001				1.16	0.78–1.73	0.453
Diabetes mellitus (yes)	1.14	0.87–1.49	0.327	1.26	0.96–1.66	0.94
Hypertension (yes)	0.77	0.57–1.03	0.083
Dyslipidemia (yes)	0.99	0.72–1.36	0.977
IHD (yes)	1.04	0.80–1.34	0.740
Atrial fibrillation (yes)	1.18	0.89–1.59	0.237
SBP (mm Hg)	1.00	0.99–1.00	0.877
DBP (mm Hg)	0.99	0.98–1.00	0.126
Hemoglobin (g/dL)	0.87	0.81–0.93	<0.001				0.90	0.83–0.99	0.033
hs-CRP (mg/L)	1.02	0.97–1.07	0.422
eGFR (mL/min/1.73 m2)	1.00	0.97–1.01	0.351
ln-BNP	1.16	1.05–1.28	0.003	1.08	0.96–1.20	0.160	1.01	0.88–1.16	0.823
LVEF (%)	0.99	0.97–1.01	0.552
SVI (L/min)	0.99	0.98–1.00	0.555
LAD (mm)	1.01	1.00–1.03	0.047				1.00	0.98–1.02	0.690
E/e′	1.02	0.99–1.04	0.082
TR-PG (mm Hg)	1.01	0.99–1.03	0.124
PAP (mm Hg)	1.01	1.00–1.03	0.018				1.01	0.95–1.02	0.175
Diuretics (yes)	1.48	1.12–1.95	0.005				1.09	0.75–1.59	0.629
ACE-I or ARB (yes)	1.07	0.82–1.40	0.578
CCB (yes)	0.90	0.69–1.16	0.437
Beta-blocker (yes)	0.96	0.74–1.24	0.792
Statin (yes)	1.05	0.80–1.38	0.714
CONUT score	1.16	1.08–1.24	<0.001	1.14	1.06–1.22	<0.001	1.12	103–1.21	0.005

Model 1: age, previous hospitalization for HF, diabetes mellitus, ln-BNP and CONUT score.

Model 2: variables of statistical significance in the univariable analyses (P < 0.05).

Abbreviations as shown in this table. HR: hazard ratio CI: confidence interval ln-BNP: natural logarithmic transformed B-type natriuretic peptide level

Receiver operating characteristic (ROC) analysis for composite cardiovascular events and CONUT score

ROC curves were constructed to assess the ability of the CONUT score to predict composite cardiovascular events (Fig. 4). The area under the curve of the CONUT score for the detection of composite cardiovascular events was 0.599 (95% CI 0.550–0.648; P < 0.001). Using the cutoff value for the CONUT score (2.5), the sensitivity and specificity were 38.2% and 76.9%, respectively, for the detection of composite cardiovascular events.

Fig. 4

Receiver operating characteristic (ROC) curves of CONUT scores for the prediction of cardiovascular events.

C-statistic for regression models, continuous NRI and integrated discrimination improvement (IDI)

The C-statistic value for PF4 was 0.608 (95% CI: 0.559–0.658); after adding a CONUT score >2.5 as a factor, the value was 0.643 (95% CI: 0.594–0.691; P = 0.039). We reclassified the risk of cardiovascular events after adding a CONUT score >2.5 to PF4; the continuous NRI was 30.5% (P < 0.001), and the IDI was 2.2% (P < 0.001) (Table 4). The ROC curves for composite cardiovascular events are shown in Fig. 5.

Table 4

C-statistics, net reclassification improvement (NRI) and integrated discrimination improvement (IDI) for the Cox hazard model to predict cardiovascular events in patients with HFpEF by the addition of CONUT score >2.5 to the PF4.

	C-statistic			NRI			IDI
	Value	95% CI	P value	Value	95% CI	P value	Value	95% CI	P value
PF4	0.608	0.559–0.658
PF4 + CONUT score >2.5	0.643	0.594–0.691	0.039	0.305	0.146–0.456	<0.001	0.022	0.010–0.035	<0.001

PF4 (4 prognostic factors): age + DM + previous hospitalization for HF + ln-BNP.

Abbreviations as shown in this table. CI, confidence interval.

Fig. 5

ROC curves of PF4 and PF4 + CONUT score >2.5 for the prediction of cardiovascular events. The black curve indicates the PF4. The red curve indicates the PF4 + CONUT score >2.5. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Discussions

The main feature of this study is that the prognosis of HFpEF patients was classified by CONUT score, and the main findings of this study were as follows:

The Kaplan-Meier curve revealed that the higher the CONUT score, the higher the incidence of composite cardiovascular events.

Multivariate Cox proportional hazards analysis revealed that the CONUT score was an independent and significant predictor of clinical outcome in HFpEF patients.

The cutoff level of the CONUT score for composite cardiovascular events was 2.5.

The NRI and IDI were significant when a CONUT score >2.5 was added to the PF4.

Obesity has been shown to be an independent risk factor for the future development of cardiovascular disease and a risk factor for the development of HF in the general population [21]. Therefore, conventional nutritional guidance for patients with HF has mainly focused on suppressing energy intake. However, Anker et al. reported that weight loss of 7.5% or more in HF patients followed for at least 6 months is a worse prognostic factor independent of HF prognosis factors such as age, NYHA functional classification, and LVEF and that preserved body weight is a better prognostic factor [22]. The concept of better prognosis when BMI is preserved was introduced. Similar large-scale multicenter trials revealed that low BMI was associated with poor prognosis, and some guidelines warn about underweight [23], [24]. Similar phenomena were reported in Japanese patients [25], [26], [27].

In Table 1, the reason that the moderate group had higher TRPG and PAP than the normal group suggests that malnutrition was more severe for more severe HF patients. The CONUT score should be interpreted with caution when medications for treating dyslipidemia due to IHD are used. The prevalence rates of dyslipidemia and IHD were significantly higher in the normal group than in the other groups. The rate of statin use was significantly higher in the light group than in the normal group.

Our findings indicate that a higher CONUT score significantly correlates with cardiovascular events in patients with HFpEF. After adjusting for various clinical parameters, a high CONUT score is still an independent predictor. Although the underlying mechanisms of cardiovascular events in the high CONUT group in HFpEF patients remain unknown, a large registry study of heart failure in Japan, JCARE-CARD [28], showed poor prognosis for low BMI patients [25]. The percentage of patients with HF with an LVEF of 50% or more in Japan was 50.6% in the CHART-1 study [29] and 68.7% in the subsequent CHART-2 study [30]. Thus, it would be reasonable to expect HFpEF patients to have poor prognoses.

In the present study, we mentioned the importance of nutritional management for HFpEF. There is a possibility for improving the prognosis of cardiovascular disease by positively intervening against malnutrition. Therefore, our present work provides data that indicate that the CONUT score not only provides important prognostic information regarding HFpEF patients but also that targeting the optimal CONUT score might be a promising therapeutic target for HFpEF. In recent years, it has been noted that both sarcopenia [31] and frail [32] are associated with a poor prognosis when they coexist with HF. Historically, low weight and malnutrition in HF patients have been described mainly by the depletion state called cachexia [33], which was attributed to changes in humoral factors. Hence, in the treatment of HF, diet therapy, especially nutrition therapy, is important together with exercise therapy.

To the best of our knowledge, this study is the first to investigate the association of the novel CONUT score with future cardiovascular events in HFpEF patients. Each component of the CONUT score is simple, and the calculation is not only easy in clinical practice but is also well validated and has a low cost, which indicates that the score can be widely applied. If this score further predicts subsequent cardiovascular events in HFpEF patients, it can also represent a useful indicator for general clinicians, as well as cardiologists in clinical practice. Although the CONUT score is strongly expected to have clinical value, large-scale clinical studies are required to confirm its value. Therefore, additional detailed, prospective, multicenter studies are warranted to verify this precise usefulness.

Study limitations

The present study has some limitations. First, it was a single-center study with a relatively small population. Therefore, a larger multiracial and multicenter study is required. Second, there were small numbers with only 49 with a “moderate” CONUT score and none with a severe score. Third, it is unclear which factors contribute—and the extent of their contribution—to the worse HF prognosis and malnutrition. Thus, further pathophysiological and molecular physiological studies, including animal experiments, are warranted. Additional detailed, large-scale clinical studies may be required to verify our results. Finally, the study population discussed in this study had the relatively non-obese nature which is less typical than what is seen in the West where the obese phenotype of HFpEF is more prevalent. This might decrease the generalizability of the results of the Western world.

Conclusion

Despite the limitations mentioned above, the results of the present study demonstrate the following: the CONUT score may be useful for predicting cardiovascular events in HFpEF patients. The CONUT score provides important prognostic information regarding HFpEF patients, and the optimal CONUT score might be a promising therapeutic target for HFpEF.