2025 Volume 7 Issue 10 Pages 904-912
Background: Malnutrition and impaired physical function are common comorbidities of heart failure (HF). We investigated the relationship between malnutrition and physical function, factors associated with these values, and their prognostic impact on clinical outcomes.
Methods and Results: We retrospectively analyzed 151 patients with HF to determine the correlation between the nutritional index, assessed using the controlling nutritional status (CONUT) score, and physical function, assessed using the short physical performance battery (SPPB). We analyzed the prognostic role of nutrition and physical function for the composite endpoints of death or HF hospitalization. The median CONUT and SPPB scores were 3 (1, 4) and 11 (8, 12), respectively. These scores showed a significant but weak correlation (r=−0.214; P=0.008). While the CONUT and SPPB scores were a significant predictor of the composite endpoint in univariable Cox analysis, only the CONUT score remained significant after adjustment for confounders. Factors associated with the CONUT score were hemoglobin and B-type natriuretic peptide levels, and those associated with the SPPB score were age, sex, and CONUT score. Using established cutoffs (i.e., CONUT ≥5, SPPB ≤9), malnutrition remained independently associated with the composite endpoint (adjusted hazard ratio 2.56; 95% confidence interval 1.46–4.48; P<0.001).
Conclusions: Malnutrition and poor physical function had a weak correlation and factors associated while these values were different. Both predicted a poor prognosis and need to be assessed in patients with HF.
The number of patients with heart failure (HF) has been increasing worldwide with an unacceptably high rate of adverse events.1,2 Frailty is commonly observed in patients with HF and is associated with impaired quality of life (QOL) and a poor prognosis.3–5 The degree of frailty has been reported to correlate with the development of new-onset HF events, HF hospitalization, and mortality.6–8 Physical function, measured using the Short Physical Performance Battery (SPPB), is a simple and reproducible indicator of physical activity and is independently associated with mortality and HF hospitalization.4,8 Malnutrition is another aspect of frailty that is also common in HF. Several metrics of malnutrition have been evaluated and reported to predict a poor prognosis.9–12 We have shown that the Controlling Nutritional Status (CONUT) score, which is calculated using serum albumin, total cholesterol, and the number of lymphocytes, was useful for the assessment and prognostication of malnutrition in HF.13 Recently, the CONUT score has been recognized not only as a nutritional indicator but also as a prognostic indicator for heart failure.14
While physical frailty and malnutrition can coexist in HF,8,13,15 their relationship is yet to be fully determined. They may share common pathophysiological processes such as decreased food intake, impaired gastroenteric function, inflammation, catabolism, and an immobile lifestyle. These can be a possible target of treatment for physical and nutritional frailty. Moreover, the prognostic value of physical frailty and malnutrition can be complementary, and assessing both may improve prognostic. Therefore, in this study, we investigated the relationship between physical function and malnutrition in patients with HF, factors associated with these values, and their prognostic role for clinical outcomes.
We retrospectively reviewed a cohort of 238 consecutive hospitalized patients who had a diagnosis of HF and underwent cardiac rehabilitation at our institution between April 2017 and March 2019. The diagnosis of HF was established according to Framingham criteria.16 The clinical course of patients was followed up for 3 years. Patients were excluded if they met the following criteria: underwent maintenance hemodialysis; B-type natriuretic peptide (BNP) <100 pg/mL or N-terminal pro BNP (NT-proBNP) <400 pg/mL; and lack of data on CONUT score and SPPB score due to the presence of any conditions that limited walking. Of 238 patients considered for inclusion in the study, 87 were excluded because of BNP <100 pg/mL or NT-proBNP <400 pg/mL, or lack of CONUT or SPPB score, leaving 151 patients in the present study (Figure 1).
Study flowchart. During the study period, 238 patients were diagnosed with heart failure (HF) and underwent cardiac rehabilitation in our institution. Nineteen patients who underwent maintenance hemodialysis (HD), 27 patients who had BNP <100 pg/mL or NT-proBNP <400 pg/mL, 32 patients without a Short Physical Performance Battery (SPPB) score, and 9 patients without a Controlling Nutritional Status (CONUT) score were excluded. A total of 151 patients were included in the final analysis. BNP, B-type natriuretic peptide; NT-proBNP, N-terminal pro B-type natriuretic peptide.
The study was approved by the Institutional Ethics Committee of Mitsui Memorial Hospital (MEC2020 No. 64; November 19, 2020) and conducted in accordance with the Declaration of Helsinki.
Data Collection and MeasurementsData on all variables were collected from electronic medical records. The recorded clinical characteristics on admission were age, sex, body weight, height, body mass index (BMI), systolic blood pressure (SBP), heart rate, presence of coronary disease, medical history, laboratory data, and echocardiographic parameters. Medication history and SPPB data at the time of discharge were recorded. The Barthel index was used to record the activities of daily living (ADL) at the time of admission and discharge.
NT-proBNP was converted to BNP using the following equation: log10(NT-proBNP) = 1.1 × log10(BNP) + 0.570.17 The CONUT score, calculated from serum albumin, total cholesterol, and lymphocyte count obtained at admission, was used to assess nutritional status. Because no laboratory data were collected at discharge, a discharge-day CONUT score could not be calculated. SPPB scores obtained at discharge were used to assess physical function. Physical function was assessed using the SPPB score, which consists of 3 components: standing balance, walking speed, and timed repeated chair rises. Each component is scored on a scale of 0–4, and the individual scores are summed up to yield a total score up to 12, with a lower score indicating greater functional impairment.18 The Barthel Index (scores 0–100) was used to evaluate disability/dependence in ADL, including the presence or absence of fecal and urinary incontinence, grooming assistance, toilet use, feeding, walking, dressing, climbing stairs, and bathing.19
The primary endpoint was a composite of death or hospitalization for HF. Each component was evaluated individually as a secondary endpoint. HF hospitalization was defined as an unexpected hospitalization with at least 1 of the following symptoms: increasing dyspnea on exertion, worsening orthopnea, paroxysmal nocturnal dyspnea, increasing fatigue or worsening exercise tolerance, or altered mental status, and at least 2 of the following symptoms: peripheral edema, elevated jugular venous pressure, radiologic signs of HF, increasing abdominal distension or ascites, pulmonary edema or crackles, rapid weight gain, hepatojugular reflex, S3 gallop, or elevated levels of BNP/NT-proBNP. These endpoints were evaluated by a retrospective review of medical records. Prognostic surveys were conducted by retrospectively reviewing the medical records of patients attending our hospital. Therefore, information on patients who visited medical institutions other than our hospital may not be included. This selection bias may limit the interpretation of the results.
Statistical AnalysisData of normally distributed continuous variables are expressed as mean±SD, and data of variables not normally distributed are presented as median (interquartile range). Data of categorical variables are expressed as numbers and percentages.
Spearman’s correlation analysis was used to assess the correlation between the CONUT and SPPB scores and their components. Receiver operating characteristic (ROC) curves of the CONUT and SPPB were generated for the clinical endpoints. The optimal cut-off point for the composite endpoint was determined using the closest point of the ROC curve for the left upper corner. Patients were divided into the malnourished group if their CONUT score was greater than or equal to the cut-off point, and the well-nourished group if their CONUT score was less than the cut-off point. The optimal cut-off value was determined using ROC curves. This differs from the cut-off values used in existing studies (e.g., ≥5 points), but was adopted because it showed higher predictive ability in the study population. Patients were also divided into the low physical functional group, defined as having a SPPB score less than or equal to the cut-off point, and the high physical functional group, defined as having a SPPB score greater than the cut-off point. The cut-off values for the CONUT and SPPB scores were determined using ROC curves to identify the values that best predicted the composite endpoint (death or heart failure hospitalization). This method has been used in previous studies and was adopted considering its clinical validity. Subanalyses used conventional thresholds (CONUT ≥5, SPPB ≤9) reported in prior heartfailure cohorts.8,18 The characteristics of the malnutrition and well-nutritional groups and the low and high physical function groups were compared with Student’s t, Mann-Whitney U, and chi-square tests as appropriate. Univariable and multivariable logistic regression analyses were performed to assess factors associated with the CONUT score greater than or equal to the cut-off point, and the SPPB less than or equal to the cut-off point. Factors with a P value <0.05 in univariable analysis were included in the multivariable model.
Kaplan-Meier, log-rank test, and Cox regression analyses were used for the clinical endpoint. As the number of clinical events was small (only 61 patients had the composite endpoint), we were not able to adjust for all possible confounding factors. Instead, the Get With the Guidelines-Heart Failure (GWTG-HF) score was used to adjust for the severity of disease.20,21 The GWTG-HF score was calculated using the 7 variables as previously reported. A risk score was established using the following 7 predictor variables: age, systolic blood pressure, heart rate, blood urea nitrogen, sodium, chronic obstructive pulmonary disease, and race. A patient’s score is obtained by summing points assigned to the value of each predictor. The values of the score are between 0 and 100. The GWTG-HF score is a score calculated from 7 clinical variables, including age, systolic blood pressure, and heart rate, and is used to predict the in-hospital mortality risk of patients with heart failure.20
All statistical analyses were performed using R, version 4.0.0, for Windows.
The mean patient age was 76±12 years, 63% were men, and 28% had coronary artery disease. The median CONUT score was 3 (1, 4), and the median SPPB score was 11 (8, 12).
Malnourished patients (CONUT score ≥3) were older, had lower systolic blood pressure, and were more frequently treated with diuretics (Table 1). The levels of hemoglobin, albumin, total cholesterol, lymphocytes, and the estimated glomerular filtration rate (eGFR) were lower, whereas the levels of C-reactive protein and BNP were higher in these patients. Malnourished patients also had lower hand grip scores and SPPB scores than well-nourished patients. Patients with lower physical function (SPPB ≤11) were older, were less frequently male, and less often had coronary artery disease. They were less frequently treated with β-blockers and were more often treated with diuretics. Patients with a lower SPPB had lower hemoglobin levels, lymphocyte count, and eGFR.
Baseline Characteristics
| CONUT <3 (n=74) |
CONUT ≥3 (n=77) |
P value | SPPB >11 (n=56) |
SPPB ≤11 (n=95) |
P value | |
|---|---|---|---|---|---|---|
| Age (years) | 74±13 | 79±11 | 0.008 | 69±14 | 81±9 | <0.001 |
| Male | 45 (61) | 50 (65) | 0.617 | 47 (84) | 48 (51) | <0.001 |
| Body weight (kg) | 60.5±31.4 | 60.1±13.3 | 0.827 | 64.8±13.4 | 57.6±12.5 | 0.001 |
| Height (cm) | 159.2±10.7 | 158.1±9.2 | 0.494 | 163.0±8.9 | 156.1±9.6 | <0.001 |
| BMI (kg/m2) | 23.7±4.1 | 23.8±3.7 | 0.903 | 24.2±3.9 | 23.5±3.9 | 0.281 |
| SBP (mmHg) | 146±37 | 132±32 | 0.020 | 146±39 | 134±32 | 0.048 |
| Heart rate (beats/min) | 88±32 | 92±29 | 0.398 | 91±31 | 89±30 | 0.673 |
| Coronary disease | 20 (27) | 22 (29) | 0.858 | 24 (43) | 18 (19) | 0.002 |
| Past medical history | ||||||
| Hypertension | 43 (58) | 47 (61) | 0.742 | 29 (52) | 61 (64) | 0.170 |
| Diabetes | 20 (27) | 18 (23) | 0.708 | 12 (21) | 26 (27) | 0.445 |
| Dyslipidemia | 22 (30) | 18 (23) | 0.461 | 12 (21) | 28 (30) | 0341 |
| Prior HF admission | 11 (15) | 17 (22) | 0.298 | 8 (14) | 20 (21) | 0.387 |
| Prior MI | 14 (19) | 21 (27) | 0.251 | 13 (23) | 22 (23) | 1.000 |
| Medication | ||||||
| ACE inhibitor | 37 (50) | 30 (39) | 0.192 | 27 (48) | 40 (42) | 0.501 |
| ARB | 15 (20) | 13 (17) | 0.677 | 13 (23) | 15 (16) | 0.283 |
| β-blocker | 39 (53) | 52 (68) | 0.069 | 40 (71) | 51 (54) | 0.039 |
| Mineralocorticoid receptor antagonist | 24 (32) | 25 (33) | 1.000 | 16 (29) | 33 (35) | 0.476 |
| Diuretic | 44 (60) | 59 (77) | 0.035 | 29 (52) | 74 (78) | 0.001 |
| Statin | 37 (50) | 39 (51) | 1.000 | 33 (59) | 43 (45) | 0.130 |
| SGLT-2 inhibitors | 5 (7) | 7 (9) | 0.818 | 3 (5) | 9 (9) | 0.554 |
| Laboratory data | ||||||
| Hemoglobin (g/dL) | 13.0±2.2 | 11.4±2.1 | <0.001 | 13.3±2.3 | 11.5±2.1 | <0.001 |
| Albumin (g/dL) | 4.1±0.3 | 3.7±0.5 | <0.001 | 3.9±0.4 | 3.8±0.4 | 0.059 |
| Total cholesterol (mg/dL) | 187.0±37.5 | 145.9±37.1 | <0.001 | 174.0±45.2 | 161.3±40.3 | 0.076 |
| Lymphocytes (count/mL) | 1,725.3±700.2 | 942.6±434.8 | <0.001 | 1,522.3±719.8 | 1,210.5±663.2 | 0.008 |
| eGFR (mL/min/1.73 m2) | 53.2±24.3 | 43.1±22.6 | 0.009 | 56.7±24.7 | 43.0±22.0 | <0.001 |
| Sodium (mEq/L) | 139.1±3.1 | 138.5±4.5 | 0.383 | 139.5±3.3 | 138.4±4.1 | 0.089 |
| CRP (mg/dL) | 1.27±2.94 | 3.00±5.32 | 0.015 | 1.45±3.86 | 2.57±4.65 | 0.130 |
| ALT (U/L) | 31.0±34.9 | 37.7±98.2 | 0.582 | 34.8±38.4 | 34.2±88.8 | 0.965 |
| BNP (pg/mL) | 512.9±464.8 | 857.9±816.4 | 0.002 | 577.6±615.3 | 754.4±722.0 | 0.127 |
| Echocardiographic parameters | ||||||
| LVEF (%) | 53±19 | 48±20 | 0.159 | 48±20 | 51±19 | 0.349 |
| Stroke volume (mL) | 56.8±18.9 | 55.3±22.8 | 0.664 | 60.2±21.7 | 53.6±20.1 | 0.066 |
| LAD (mm) | 43.5±8.6 | 46.5±9.2 | 0.049 | 43.8±8.9 | 45.7±9.0 | 0.220 |
| TRPG (mm Hg) | 30.2±10.1 | 32.1±10.2 | 0.282 | 30.0±11.1 | 31.8±9.6 | 0.336 |
| RVSP (mm Hg) | 38.3±12.1 | 41.9±12.2 | 0.094 | 40.0±12.7 | 40.3±12.0 | 0.901 |
| IVC (mm) | 16.6±4.9 | 19.8±7.0 | 0.001 | 18.4±5.4 | 18.1±6.7 | 0.802 |
| ADL (on admission) | 79±24 | 77±23 | 0.581 | 86±18 | 73±25 | <0.001 |
| ADL (at discharge) | 96±9 | 93±12 | 0.072 | 99±1 | 91±12 | <0.001 |
| Hand grip (kg) | 23.8±9.4 | 20.2±7.3 | 0.011 | 27.6±8.0 | 18.5±6.9 | <0.001 |
| Leg strength (% body weight) | 42±15 | 39±12 | 0.242 | 48±13 | 35±12 | <0.001 |
| Walking speed (m/s) | 0.92±0.31 | 0.86±0.23 | 0.167 | 1.09±0.21 | 0.76±0.23 | <0.001 |
| SPPB | 12 [8–12] | 10 [8–11] | 0.021 | 12 [12–12] | 9 [6–10] | <0.001 |
| GWTG-HF | 38.7±7.9 | 44.7±7.6 | <0.001 | 37.8±8.4 | 44.1±7.3 | <0.001 |
Data are presented as mean±SD, n (%), or median [interquartile range]. ACE, angiotensin-converting enzyme; ADL, activity of daily living; ALT, alanine aminotransferase; ARB, angiotensin II receptor blocker; BMI, body mass index; BNP, B-type natriuretic peptide; CONUT, Controlling Nutritional Status; CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; GWTG-HF, Get With The Guidelines-Heart Failure; HF, heart failure; IVC, inferior vena cava; LAD, left atrial dimension; LVEF, left ventricular ejection fraction; MI, myocardial infarction; RVSP, right ventricular systolic pressure; SBP, systolic blood pressure; SGLT-2, sodium-glucose transpoter-2; SPPB, Short Physical Performance Battery; TRPG, transtricuspid pressure gradient.
CONUT was weakly but significantly correlated with the SPPB score (r=−0.214, P=0.008). Regarding the components of these scores, lymphocyte count was significantly correlated with standing balance, walking speed, and timed repeated chair rise, and albumin level was significantly correlated with timed repeated chair rise (Table 2).
Spearman’s Correlation Coefficients for Nutrition Index and Physical Function
| Standing balance | Walking speed | Timed repeated chair rise | |
|---|---|---|---|
| Albumin | r=0.056, P=0.492 | r=0.103, P=0.208 | r=0.215, P=0.008 |
| Total cholesterol | r=−0.034, P=0.677 | r=−0.051, P=0.536 | r=0.021, P=0.798 |
| Lymphocytes | r=0.169, P=0.038 | r=0.167, P=0.040 | r=0.253, P=0.002 |
Prognostic Significance of CONUT and SPPB
During the follow-up period, composite endpoint, death, and HF hospitalization occurred in 61 (40.4%), 23 (15.2%), and 53 (35.1%) patients, respectively. The area under the ROC curve (AUC) for the composite endpoint was significantly higher for the CONUT score (0.750; 95% confidence interval [CI] 0.674–0.826) than for the SPPB score (0.576; 95% CI 0.487–0.666; P=0.002; Figure 2). The optimal cut-off points were 3 for the CONUT score and 11 for the SPPB score. A similar finding was observed with death and HF hospitalization, while only HF hospitalization showed a statistically significant difference. The AUCs for death were higher for the CONUT score (0.692; 95% CI 0.592–0.791) than for the SPPB score (0.587; 95% CI 0.477–0.697), but did not show a statistical significance (P=0.142). The AUCs for HF hospitalization were significantly higher for the CONUT score (0.748; 95% CI 0.669–0.826) than for the SPPB score (0.577; 95% CI 0.486–0.666; P=0.003).
Area under the receiver operating characteristic curve (AUC) for the composite endpoint of death or heart failure hospitalization. CI, confidence interval; CONUT, Controlling Nutritional Status; SPPB, Short Physical Performance Battery.
Kaplan-Meier curves showed that patients with higher CONUT and lower SPPB scores had the highest incidence of the composite endpoint, while those with higher SPPB and lower CONUT scores had the lowest incidence (Figure 3). In univariable Cox analysis, the higher CONUT score (≥3) was significantly associated with the composite endpoint, death, and HF hospitalization, and these relationships remained significant after adjustment for SPPB and GWTG-HF scores in multivariable analysis. While lower SPPB (≤11) predicted these outcomes in univariable analysis, this lost statistical significance after adjustment for the CONUT or GTWG-HF scores (Table 3).
Kaplan-Meier curve of the probability for the composite endpoint of death or heart failure hospitalization. The composite endpoint was more frequently observed in higher CONUT scores, and lower SPPB scores had the highest incidence for the composite endpoint, while those with higher SPPB and lower CONUT scores had the lowest incidence (log-rank P<0.001). CONUT, Controlling Nutritional Status; SPPB, Short Physical Performance Battery.
Univariable, Bivariable, and Multivariable Cox Hazard Analysis
| Variable | Univariable | Bivariable | Multivariable | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | P value | Adjusted HR (95% CI) |
P value | Adjusted HR (95% CI) |
P value | |
| Composite endpoint† | ||||||
| CONUT (≥3) | 3.74 (2.11–6.65) | <0.001 | 3.38 (1.88–6.08) | <0.001 | 2.86 (1.58–5.18) | <0.001 |
| SPPB (≤11) | 2.11 (1.18–3.78) | 0.012 | 1.59 (0.88–2.89) | 0.125 | 1.28 (0.70–2.35) | 0.420 |
| GWTG-HF | 1.08 (1.04–1.11) | <0.001 | 1.05 (1.02–1.09) | 0.006 | ||
| Death | ||||||
| CONUT (≥3) | 3.57 (1.33–9.62) | 0.011 | 3.00 (1.09–8.23) | 0.033 | 2.32 (0.84–6.46) | 0.106 |
| SPPB (≤11) | 2.99 (1.02–8.78) | 0.047 | 2.30 (0.77–6.89) | 0.138 | 1.95 (0.66–5.79) | 0.229 |
| GWTG-HF | 1.09 (1.03–1.13) | 0.002 | 1.07 (1.01–1.14) | 0.025 | ||
| HF hospitalization | ||||||
| CONUT (≥3) | 4.03 (2.15–7.58) | <0.001 | 3.38 (1.88–6.08) | <0.001 | 3.06 (1.60–5.87) | <0.001 |
| SPPB (≤11) | 2.09 (1.12–3.91) | 0.021 | 1.59 (0.88–2.89) | 0.125 | 1.25 (0.66–2.40) | 0.493 |
| GWTG-HF | 1.08 (1.04–1.12) | <0.001 | 1.06 (1.01–1.10) | 0.008 | ||
†Composite endpoint is death or HF hospitalization. CI, confidence interval; HR, hazard ratio. Other abbreviations as in Table 1.
Using established cutoffs (CONUT ≥5, SPPB ≤9), malnutrition remained independently associated with the composite endpoint (adjusted hazard ratio [HR] 2.56; 95% CI 1.46–4.48; P<0.001), whereas low SPPB showed no association (adjusted HR 1.13; 95% CI 0.68–1.88; P=0.64; Supplementary Table).
Relationship of Each Factor to CONUT and SPPBIn a multivariable logistic analysis, hemoglobin and BNP levels were significantly associated with higher CONUT scores. Significant predictors for lower SPPB were age, female sex, and CONUT score (Table 4A,B).
Predictors for Malnutrition and Impaired Physical Function
| Variable | Univariable | Multivariable | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | |
| (A) Predictor for CONUT ≥3 | ||||||
| Age (years) | 1.04 | 1.01–1.07 | 0.011 | 1.00 | 0.96–1.05 | 0.859 |
| Female | 0.84 | 0.43–1.62 | 0.600 | |||
| Body weight (kg) | 1.01 | 0.93–1.09 | 0.903 | |||
| Height (cm) | 0.99 | 0.96–1.02 | 0.491 | |||
| BMI (kg/m2) | 1.01 | 0.93–1.09 | 0.903 | |||
| SBP (mm Hg) | 0.99 | 0.98–0.99 | 0.023 | 0.99 | 0.98–1.00 | 0.174 |
| Heart rate (beats/min) | 1.00 | 0.99–1.02 | 0.396 | |||
| Coronary disease | 1.08 | 0.53–2.20 | 0.832 | |||
| Hypertension | 1.13 | 0.59–2.16 | 0.714 | |||
| Diabetes | 0.82 | 0.40–1.72 | 0.606 | |||
| Dyslipidemia | 0.72 | 0.35–1.49 | 0.377 | |||
| Prior HF admission | 1.62 | 0.70–3.75 | 0.257 | |||
| Prior MI | 1.61 | 0.75–3.46 | 0.226 | |||
| ACE inhibitor | 0.64 | 0.34–1.22 | 0.173 | |||
| ARB | 0.80 | 0.35–1.82 | 0.593 | |||
| β-blocker | 1.87 | 0.97–3.61 | 0.064 | |||
| Mineralocorticoid receptor antagonist | 1.00 | 0.51–1.98 | 0.996 | |||
| Diuretic | 2.23 | 1.11–4.51 | 0.025 | 1.17 | 0.49–2.82 | 0.725 |
| Statin | 1.03 | 0.54–1.94 | 0.936 | |||
| Hemoglobin (g/dL) | 0.70 | 0.59–0.83 | <0.001 | 0.68 | 0.53–0.87 | 0.002 |
| eGFR (mL/min/1.73 m2) | 0.98 | 0.97–0.99 | 0.011 | 1.01 | 0.99–1.03 | 0.460 |
| Sodium (mEq/L) | 0.96 | 0.89–1.05 | 0.382 | |||
| CRP (mg/dL) | 1.12 | 1.01–1.23 | 0.026 | 1.11 | 0.98–1.25 | 0.097 |
| ALT (U/L) | 1.00 | 0.99–1.01 | 0.591 | |||
| BNP† (pg/mL) | 1.51 | 1.16–1.97 | 0.002 | 1.71 | 1.18–2.48 | 0.004 |
| LVEF (%) | 0.99 | 0.97–1.00 | 0.159 | |||
| Stroke volume (mL) | 0.99 | 0.98–1.01 | 0.661 | |||
| LAD (mm) | 1.04 | 1.00–1.08 | 0.052 | |||
| TRPG (mmHg) | 1.02 | 0.99–1.05 | 0.281 | |||
| RVSP (mmHg) | 1.02 | 0.99–1.06 | 0.095 | |||
| ADL (on admission) | 0.99 | 0.98–1.01 | 0.578 | |||
| Hand grip (kg) | 0.95 | 0.91–0.99 | 0.013 | 1.00 | 0.94–1.06 | 0.965 |
| Leg strength (% body weight) | 0.99 | 0.96–1.01 | 0.242 | |||
| Walking speed (m/s) | 0.43 | 0.13–1.43 | 0.168 | |||
| SPPB | 0.93 | 0.84–1.03 | 0.166 | |||
| (B) Predictor for SPPB ≤11 | ||||||
| Age (years) | 1.11 | 1.07–1.16 | <0.001 | 1.10 | 1.04–1.17 | 0.001 |
| Female | 5.11 | 2.25–11.60 | <0.001 | 5.27 | 1.23–22.50 | 0.025 |
| Body weight (kg) | 0.96 | 0.93–0.98 | <0.001 | 1.00 | 0.96–1.04 | 0.982 |
| Height (cm) | 0.93 | 0.89–0.96 | <0.001 | 1.03 | 0.95–1.12 | 0.456 |
| BMI (kg/m2) | 0.95 | 0.88–1.04 | 0.282 | |||
| SBP (mm Hg) | 0.99 | 0.98–1.00 | 0.051 | |||
| Heart rate (beats/min) | 0.99 | 0.99–1.01 | 0.671 | |||
| Coronary disease | 0.31 | 0.15–0.65 | 0.002 | 0.46 | 0.17–1.26 | 0.132 |
| Hypertension | 1.67 | 0.85–3.27 | 0.134 | |||
| Diabetes | 1.38 | 0.63–3.02 | 0.418 | |||
| Dyslipidemia | 1.53 | 0.71–3.33 | 0.281 | |||
| Prior HF admission | 1.60 | 0.65–3.92 | 0.304 | |||
| Prior MI | 1.00 | 0.46–2.18 | 0.994 | |||
| ACE inhibitor | 0.78 | 0.40–1.52 | 0.466 | |||
| ARB | 0.62 | 0.27–1.42 | 0.259 | |||
| β-blocker | 0.46 | 0.23–0.94 | 0.032 | 0.50 | 0.20–1.27 | 0.145 |
| Mineralocorticoid receptor antagonist | 1.33 | 0.65–2.73 | 0.435 | |||
| Diuretic | 3.28 | 1.61–6.70 | 0.001 | 1.73 | 0.67–4.48 | 0.257 |
| Statin | 0.58 | 0.30–1.12 | 0.106 | |||
| Hemoglobin (g/dL) | 0.68 | 0.57–0.81 | <0.001 | 1.05 | 0.80–1.37 | 0.724 |
| Albumin (g/dL) | 0.47 | 0.21–1.04 | 0.062 | |||
| Total cholesterol (mg/dL) | 0.99 | 0.99–1.00 | 0.077 | |||
| Lymphocytes (count/mL) | 0.99 | 0.99–1.00 | 0.012 | 1.00 | 0.99–1.00 | 0.699 |
| eGFR (mL/min/1.73 m2) | 0.98 | 0.96–0.99 | 0.001 | 0.99 | 0.96–1.01 | 0.160 |
| Sodium (mEq/L) | 0.92 | 0.84–1.01 | 0.092 | |||
| CRP (mg/dL) | 1.07 | 0.98–1.18 | 0.143 | |||
| ALT (U/L) | 1.00 | 0.99–1.00 | 0.965 | |||
| BNP† (pg/mL) | 1.28 | 0.98–1.67 | 0.066 | |||
| CONUT | 1.37 | 1.12–1.68 | 0.002 | 1.45 | 1.04–2.03 | 0.029 |
| LVEF (%) | 1.01 | 0.99–1.03 | 0.347 | |||
| Stroke volume (mL) | 0.99 | 0.97–1.00 | 0.069 | |||
| LAD (mm) | 1.02 | 0.99–1.07 | 0.220 | |||
| TRPG (mmHg) | 1.02 | 0.98–1.06 | 0.334 | |||
| RVSP (mmHg) | 1.00 | 0.97–1.03 | 0.900 | |||
†BNP is a logarithmic transformation. OR, odds ratio. Other abbreviations as in Tables 1,3.
The main findings of the present study were as follows: (1) the CONUT score and SPPB showed a significant but weak correlation in patients with HF; (2) while the CONUT score and SPPB were useful for predicting prognosis in HF, the prognostic implication was better for the CONUT score; and (3) hemoglobin and BNP levels were associated with CONUT score, while age, female sex, and CONUT score were associated with SPPB.
In this study, nutritional impairment, as assessed using the CONUT score, and poor physical function, as assessed using the SPPB score, predicted poor prognosis in patients with HF. The CONUT and SPPB scores showed a significant but weak correlation, and many of their components did not correlate with each other. As shown in Table 2, lymphocyte count showed a significant correlation with standing balance, walking speed, and timed repeated chair rise. This suggests that nutritional status may affect physical function. The results of the Kaplan-Meier curve showed that patients with both impaired nutritional and physical functional status had the worst prognosis. These findings suggest that their prognostication can be complementary, and we need to assess both nutritional status and physical function in patients with HF.
In the present study, SPPB was a significant predictor of death, HF hospitalization, and the composite of these, whereas this relationship was not statistically significant after adjusting for CONUT or GWTG score. The univariate analysis was significant, but the multivariate analysis was not, and should be interpreted with caution. SPPB is a simple, easy, and versatile assessment of physical function.18,22 However, the CONUT score has the advantage that it can be assessed without placing a physical burden on the patient because it is calculated from the results of routine blood tests, such as serum albumin, total cholesterol, and lymphocyte count. In contrast, patients with severely impaired physical function could not undergo the SPPB test and thus were excluded from this study. Therefore, the relationship between nutritional status and physical function in patients with severe physical dysfunction remains unclear. Future studies are needed to investigate the relationship between nutritional status and physical function in patients for whom SPPB measurement is difficult. This difference might explain the lack of statistically significant associations of SPPB and the clinical outcomes after adjustment for the CONUT score.
We showed that the independent predictors for the CONUT score were hemoglobin and BNP levels, and predictors for the SPPB score were age, sex, and CONUT score. BNP is a well-known marker of congestion, which is central to the pathogenesis of HF and plays an important role in malnutrition.23 Intestinal congestion caused by fluid retention leads to decreased intestinal motility, impaired absorption, and reduced appetite. Increased intestinal permeability with bowel edema induces the translocation of endotoxins from Gram-negative bacteria into the bloodstream.24 This leads to systemic inflammatory activation, accelerated catabolism, and worsened nutritional status. Prolonged hypotrophy with catabolism may lead to skeletal muscle loss and impaired physical function.25 Therefore, congestion, which is fundamental to this vicious cycle, can be an important therapeutic target to improve malnutrition and physical functional status in HF. Along with diuretic therapy to reduce fluid volume, renin-angiotensin-aldosterone system inhibitors and β-blockers can ameliorate neurohumoral hyperactivity and thus may alleviate intestinal congestion and hyper-inflammatory conditions.26–28 Moreover, a combination of nutritional therapy with appropriate calorie and protein intake, and exercise training for strength, balance, and gait function, should also be considered. Exercise therapy in HF has been reported to improve skeletal muscle quality, exercise tolerance, mental status, QOL, and prognosis.29–33 Besides the general medical treatment for neurohormonal activation, assessment and therapeutic interventions for nutritional status and physical function are necessary for all patients with HF. The concordant prognostic value observed with conventional cutoffs further validates our findings across diverse clinical settings.
Study LimitationsThis was a retrospective analysis of data from a single center. Because our study excluded patients who were not able to undergo the SPPB test, the results might underestimate the impact of physical frailty on the outcome. While we conducted multivariable analysis for prognostication and factors associated with the CONUT and SPPB scores, measured and unmeasured confounding factors may have affected the results. Discharge laboratory data were not collected, precluding calculation of a discharge CONUT score; this limitation is stated in the Methods section and reiterated here.
Malnutrition and poor physical function predicted a poor prognosis in patients with HF. The CONUT and SPPB scores showed a significant but weak correlation, and many of their components did not correlate with each other. Therefore, we need to assess both of these for better prognostication and appropriate therapeutic interventions.
The authors thank the rehabilitation team members at the Mitsui Memorial Hospital for their contributions to this study.
This research received no grant from any funding agency in the public, commercial or not-for-profit sectors.
Y.H. received lecture fees from Eli Lilly and Co. J.T. received lecture fees from Abbott Medical and Boston Scientific. K.T. received lecture fees from Kaneka, Abbott Medical, and Boston Scientific. The other authors have no conflicts of interest to declare.
This study was approved by the Institutional Ethics Committee of Mitsui Memorial Hospital (MEC2020 No. 64; November 19, 2020).
The deidentified participant data will not be shared.
Please find supplementary file(s);
https://doi.org/10.1253/circrep.CR-24-0149
