Associations between in‐hospital daily protein intake and adverse clinical outcomes in older patients with heart failure

Abstract

Aims

The adverse effects of low daily protein intake (DPI) on clinical outcomes in patients with heart failure (HF) are known; however, an optimal DPI to predict event adverse outcomes remains undetermined. Moreover, whether protein restriction therapy for chronic kidney disease is applicable in patients with HF and renal dysfunction remains unclear.

Methods and results

In this single‐centre, ambispective cohort study, we included 405 patients with HF aged ≥65 years (mean age, 78.6 ± 7.5 years; 50% women). DPI was estimated from consumption over three consecutive days before discharge and normalized relative to the ideal body weight [IBW, 22 kg/m² × height (m)²]. The primary outcome was a composite of all‐cause mortality and HF‐related readmission within the 2 year post‐discharge period.

Results

During an average follow‐up period of 1.49 ± 0.74 years, 100 patients experienced composite events. Kaplan–Meier survival curves revealed a significantly lower composite event‐free rate in patients within the lowest quartile of DPI than in the upper quartiles (log‐rank test, P = 0.02). A multivariate Cox proportional hazards analysis after adjusting for established prognostic markers and non‐proteogenic energy intake revealed that patients in the lowest DPI quartile faced a two‐fold higher risk of composite events than those in the highest quartile [hazard ratio (HR), 2.03; 95% confidence interval (CI), 1.08–3.82; P = 0.03]. The composite event risk linearly increased as DPI decreased (P for nonlinearity = 0.90), with each standard deviation (0.26 g/kg IBW/day) decrease in DPI associated with a 32% increase in composite event risk (HR, 1.32; 95% CI, 1.10–1.71; P = 0.04). There was significant heterogeneity in the effect of DPI, with the possible disadvantage of lower DPI in patients with HF with cystatin C‐based estimated glomerular filtration rate <30 mL/min/1.73 m². The cutoff value of DPI for predicting the occurrence of composite events calculated from the Youden index was 1.12 g/kg IBW/day. Incorporating a DPI < 1.12 g/kg IBW/day into the baseline model significantly improved the prediction of post‐discharge composite events (continuous net reclassification improvement, 0.294; 95% CI, 0.072–0.516; P = 0.01).

Conclusions

Lower DPI during hospitalization is associated with an increased risk of mortality and HF readmission independent of non‐proteogenic energy intake, and the possible optimal DPI for predicting adverse clinical outcomes is >1.12 g/kg IBW/day in older patients with HF. Caution is warranted when protein restriction therapy is administered to older patients with HF and renal dysfunction.

Introduction

Heart failure (HF) is a global health concern that affects over 64 million individuals worldwide, with a considerable number being older individuals.¹ In this demographic, muscle wasting and physical frailty are common, contributing to increased morbidity and mortality.^2–4

The interplay among HF-induced protein imbalance, age-associated decline in muscle protein synthesis, and nutritional inadequacy places older patients with HF at a high risk of developing conditions, such as sarcopenia and frailty.⁵ In addition to exercise, protein intake also plays a pivotal role in maintaining muscle health and stimulating protein synthesis.^6,7 To optimize these processes, the European Society for Clinical Nutrition and Metabolism (ESPEN) guidelines recommend that older adults with chronic or acute diseases, including HF, should consume 1.2–1.5 g/kg/day of daily protein intake (DPI).⁸ Moreover, individuals diagnosed with sarcopenia are advised to maintain a DPI of 1.0–1.5 g/kg/day, alongside regular physical exercise.⁷ However, there are limited data showing the optimal DPI level for improving clinical outcomes in patients with HF. A recent multicentre cohort study by Streng et al. demonstrated that an estimated lower protein intake was independently associated with a higher risk of mortality in 2282 patients with HF who were included in the BIOSTAT-CHF multicentre prospective observational study.⁹ Although this study showed a close association between DPI and clinical outcomes in patients with HF, the study used a formula to estimate protein intake based on urinary nitrogen concentrations. However, it remains unclear whether the formula can be applied to older patients with HF who have a high prevalence of sarcopenia and renal dysfunction, as blood and urine urea nitrogen concentrations are affected by the status of protein catabolism and tubular absorption of urea in addition to protein intake. In addition, an optimal DPI for preventing adverse clinical outcomes was not proposed in that study.⁹ Finally, the association between DPI and adverse clinical outcomes in patients with HF and renal dysfunction, a common comorbidity in older patients with HF, was not addressed. This is a critical issue because guidelines recommend protein intake restriction for patients with chronic kidney disease (CKD) to prevent the development of end-stage kidney disease and death.

To resolve these issues, the present study aimed to examine the association between DPI estimated using the visual approximation technique over three consecutive days before discharge and adverse clinical outcomes in older patients with HF. To determine the optimal DPI for predicting adverse events, the dose–response association between DPI and the probability of adverse events was analysed. This investigation aimed to elucidate whether the effects of DPI on prognosis remain consistent across different clinical conditions, including muscle wasting, frailty, and renal dysfunction. Our findings contribute to the development of more precise and effective nutritional strategies tailored for this vulnerable patient population.

Methods

Study participants

This was a single-centre, ambispective (combined retrospective and prospective) cohort study. Consecutive patients diagnosed with HF and admitted to our institute for treatment between 1 April 2016 and 31 December 2020 were enrolled in this study (Figure 1). We selected this time frame for participant enrolment because of the initiation of routine assessment of nutritional status, body composition, and physical frailty on 1 April 2016. HF diagnosis adhered to the guidelines stipulated by the Japanese Circulation Society/Japanese Heart Failure Society.¹⁰ The study used the following inclusion criteria: included individuals (i) age ≥65 years and (ii) enrolled in a multidisciplinary cardiac rehabilitation programme for HF management during their hospital stay. This programme has been in implementation since August 2010 and includes exercise training, patient education on self-monitoring and medication, and nutritional counselling by a team of cardiologists, nurses, physical therapists, pharmacists, dietitians, and social workers, as previously described.¹¹ The exclusion criteria included patients undergoing maintenance haemodialysis (n = 3), those with Stage 5 CKD (n = 18), those with pulmonary hypertension (n = 3), and those who experienced in-hospital death (n = 6).

[IMAGE OMITTED. SEE PDF]

This study adhered to the Declaration of Helsinki and was approved by the Clinical Investigation Ethics Committee of Sapporo Medical University Hospital (Number 302-243). Before 10 April 2019, the participants' consent was obtained through an opt-out strategy. After 11 April 2019, the participants were required to provide written informed consent.

Evaluation of dietary intake and nutritional status

Dietary intake was estimated by quantifying the intake over a 3 day period prior to discharge, whereas the Mini Nutritional Assessment—Short Form (MNA-SF) was used to evaluate the nutritional status just before discharge.

Daily energy and DPI were calculated according to the visual estimation method, validated previously through comparison with the weighed food estimation method.^11–14 The patients were catered for meals offering between 1400 and 2200 kcal, depending on their ideal body weight (IBW) during their hospital stay. Visual approximation techniques were used by nursing staff and physical therapists to gauge the proportion of each consumed meal item over a span of three consecutive days prior to discharge. The observations were subsequently used to calculate daily energy intake (kcal/day) and DPI (g/day). A previous study validated the efficacy of visual approximation by nursing staff, establishing a strong correlation with estimates obtained through dietary intake weighing for various types and textures.^13,14 To adjust for potential spurious estimations of daily energy and DPI in patients with varying body weights, daily energy intake (kcal/kg/day) and DPI (g/kg/day) were normalized to the IBW.¹⁵ The IBW was determined to be 22 kg/m² × [height (m)].² Subsequently, DPI was classified into four discrete groups based on sex-specific quartile distributions.

Nutritional status was assessed using the MNA-SF, as detailed in previous studies.^11,12,16 This assessment tool comprises six questions about reduction in food intake over the past 3 months, weight loss during the preceding 3 months, mobility, psychological stress or acute disease occurrence within the last 3 months, neuropsychological problems, and body mass index (BMI), with a total score ranging from 0 to 14.

Body composition analysis

Body composition was analysed using dual-energy X-ray absorptiometry (DEXA) Horizon-A DXA System (Hologic, Waltham, MA, USA).¹⁷ The procedure enabled a detailed examination of whole and regional fat and lean masses. Appendicular skeletal muscle mass (ASM) was calculated by adding the bone-free lean masses of the limbs. Furthermore, the ASM index (ASMI) was computed by dividing the ASM (kg) by the square of the height (m). Muscle wasting was identified when the ASMIs were <7.00 kg/m² in men and <5.40 kg/m² in women.

Assessment of physical frailty

Physical frailty was assessed by trained physical therapists using the revised Japanese version of the Cardiovascular Health Study.¹⁸ Patients manifesting at least three of five indicators—shrinking, weakness (indicated by diminished handgrip strength), exhaustion, slowness (represented by decreased usual gait speed), and low physical activity—were classified as physically frail.

Handgrip strength was measured using a Smedley spring-type digital dynamometer (TKK 5401 GRIP-D; Takei, Niigata, Japan).^19,20 The maximum force exerted over two trials, each trial with a different hand, was recorded as an absolute value (kg). Diminished handgrip strength was defined as readings <28 kg for men and <18 kg for women.

The usual gait speed was evaluated by timing the patients as they traversed a marked intermediate distance of 10 m within a total path of 14 m at their habitual pace. Walking aids were allowed, if required. Gait speed <1.0 m/s was deemed indicative of a decreased usual gait speed.^19,20

Collection of data for other clinical parameters

A comprehensive dataset comprising demographic information, comorbidities, medication regimens, laboratory data, and echocardiographic data was collected from the patients' medical records.

Within the week preceding discharge, laboratory data, including N-terminal pro-brain natriuretic peptide (NT-proBNP), serum albumin, haemoglobin, cystatin C, cystatin C-based estimated glomerular filtration rate (eGFRcys), blood urea nitrogen, and uric acid levels, were obtained. The eGFRcys was calculated using a formula specifically tailored for Japanese individuals.²¹

Transthoracic echocardiography was performed following standard protocols, and the left ventricular ejection fraction (LVEF) was measured using the modified Simpson method.

Comorbidities were ascertained based on comprehensive medical information, including patient history, collected data, and prescribed medications. CKD was diagnosed when the eGFRcys was <60 mL/min/1.73 m².

Clinical endpoints

The primary clinical outcome measure was the first adverse event within a 2 year follow-up period after enrolment. This was defined as a composite of all-cause death and unplanned readmission due to the exacerbation of HF symptoms. An episode of worsening HF was characterized by either an unplanned hospital admission specifically for HF management or an emergency consultation triggered by escalating HF symptoms.

Sample size calculations

Based on our estimation, the composite event rate over 2 years and the hazard ratio (HR) of the first quantile for DPI were projected to be 25% and 2.0, respectively. The correlation coefficient squared between the lower DPI and confounding variables, as inferred from the dataset of our previous study,¹² was estimated to be 0.12. With these estimations, the sample size required to detect statistically significant differences between the groups with at least 80% power at the 5% significance level was 397. The sample size was calculated using the R software Version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria; ), specifically employing the ssizeEpi function Version 0.1.3 from the powerSurvEpi package.

Statistical analyses

Data are presented as means ± standard deviations, medians with inter-quartile ranges (25th–75th percentile), or numerical quantities supplemented by percentages. Baseline characteristics were compared using Welch's t-test, Mann–Whitney U test, or χ² test, as appropriate.

Missing data points were imputed via multiple imputation analyses, with the imputation model incorporating both the outcome and all exposure and adjustment variables. Presuming that the missing data were random, 100 imputations were conducted using chained equations, and the estimates were pooled according to Rubin's rule.²²

Survival curves were derived using the Kaplan–Meier method, and the statistical significance of differences between the curves was assessed using log-rank statistics. Both univariate and multivariate Cox proportional hazards analyses were performed to gauge prognostic predictive ability. The multivariate Cox proportional hazards model was adjusted for a range of variables, including age, sex, BMI, New York Heart Association (NYHA) Functional Class III, prior admission due to HF, LVEF < 40%, logarithmic NT-proBNP (log NT-proBNP) level, eGFRcys < 45 mL/min/1.73 m², hypertension, diabetes mellitus (DM), use of beta-blocker, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blocker or angiotensin receptor–neprilysin inhibitor, and use of mineralocorticoid receptor antagonist. Additionally, we incorporated non-proteogenic energy intake into the adjustment to mitigate the confounding effects of energy intake.²³ To explore the presence of effect modification, subgroup analyses were performed after stratification for age (<75 or ≥75 years), sex, BMI (<25.0 or ≥25.0 kg/m²), NYHA functional class (I–II or III), DM, eGFRcys level (<45 or ≥45 mL/min/1.73 m², <30 or ≥30 mL/min/1.73 m²), muscle wasting, physical frailty, and nutritional status as indexed using the MNA-SF (≤7 or >7 points). The dose-dependent association between DPI and the probability of composite events was examined using a Cox proportional hazards model integrated with a restricted cubic spline function featuring four knots.

To ascertain whether the DPI information contributes to a significant incremental prognostic value beyond the risk factors for composite events, several models were constructed. A baseline model that included the variables used for adjustment in the Cox proportional hazards models was constructed. Expanded models that included the variables in the baseline model along with the DPI were also constructed. The optimal DPI cutoff value for predicting the incidence of composite events was determined based on the Youden index. Harrell's C-index was calculated and compared between the baseline and expanded models, following the methods of DeLong et al.²⁴ Additionally, the significance of the incremental discriminative value added by the DPI was assessed by calculating the continuous net reclassification improvement (cNRI). For sensitivity analyses, univariate and multivariate Cox proportional hazards analyses were performed for the main analyses using non-normalized DPI (per g/day).

A P value <0.05, based on a two-tailed test, was considered statistically significant. The statistical analyses were performed using JMP Pro Version 17.0.0 (SAS Institute Inc., Cary, NC, USA) and R software Version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria; ).

Results

Among the 435 patients with HF who were initially screened, 30 were excluded based on the stipulated exclusion criteria, resulting in 405 patients whose data were analysed (Figure 1).

Baseline characteristics

The mean age of the patients was 78.6 ± 7.5 years, with an equal distribution of men and women (Table 1). The average actual and IBW of the patients were 54.0 ± 11.0 and 54.2 ± 6.5 kg, respectively, and the average BMI of the patients was 21.9 ± 3.6 kg/m². Upon discharge, 36% of the patients were classified as NYHA Functional Class III or IV. The mean LVEF was 50.0 ± 15.8%, with 30% of patients exhibiting LVEF < 40%. Hypertension, dyslipidaemia, DM, atrial fibrillation, and CKD were prevalent in 72%, 59%, 41%, 43%, and 68% of the patients, respectively. Furthermore, prior hospitalization for HF was part of the medical history of 41% of the patients. Valvular heart disease was the most common aetiology of HF (42%), followed by cardiomyopathy (28%), and ischaemic heart disease (17%). The average MNA-SF score was 8.1 ± 2.8 points. The average daily energy intake was 1430 ± 291 kcal/day, and the average daily energy intake per IBW was 26.5 ± 5.4 kcal/kg IBW/day. In total, 63% of the 405 patients were diagnosed with physical frailty, and 74% were found to have muscle wasting, which is consistent with the results of previous studies.^3,25

Table 1 Baseline characteristics according to sex-specific quantile of daily protein intake

			Daily protein intake (g/kg IBW/day)
		Q1 (lowest)	Q2	Q3	Q4 (highest)
Missing		M: <1.00 W: <1.02	M: 1.00–<1.18 W: 1.02–<1.24	M: 1.18–<1.27 W: 1.24–<1.39	M: ≥1.27 W: ≥1.39
n (%)	n = 405	n = 101	n = 102	n = 101	n = 101	P value
Age (years)		78.6 ± 7.5	78.8 ± 7.1	77.4 ± 7.8	78.1 ± 7.4	80.1 ± 7.4	0.054
Women, n (%)		203 (50)	51 (51)	51 (50)	50 (50)	51 (51)	1.00
Height (cm)		156.7 ± 9.4	156.8 ± 8.5	160.6 ± 11.0	157.5 ± 8.5	152.0 ± 7.3	<0.01
Body weight
Actual (kg)		54.0 ± 11.0	53.2 ± 11.1	56.1 ± 13.1	54.6 ± 10.5	52.2 ± 8.6	0.06
Ideal (kg)		54.2 ± 6.5	54.2 ± 5.9	57.0 ± 7.8	54.7 ± 5.8	51.0 ± 4.9	<0.01
BMI (kg/m2)		21.9 ± 3.6	21.5 ± 3.8	21.6 ± 4.0	21.9 ± 3.4	22.5 ± 3.0	0.19
Systolic blood pressure (mmHg)		119 ± 19	119. ± 20	113 ± 17	119 ± 20	125 ± 19	<0.01
NYHA functional class
I, n (%)		15 (4)	2 (2)	4 (4)	5 (5)	4 (4)	0.64
II, n (%)		246 (61)	58 (57)	67 (66)	57 (56)	64 (63)
III, n (%)		144 (36)	41 (41)	31 (30)	39 (39)	33 (33)
LVEF (%)		50.0 ± 15.8	50.9 ± 15.8	48.5 ± 16.2	48.6 ± 15.9	51.9 ± 15.3	0.33
<40%, n (%)		120 (30)	29 (29)	30 (29)	33 (33)	28 (28)	0.88
Past or current smoker, n (%)		187 (46)	51 (51)	53 (52)	43 (43)	40 (40)	0.22
Prior HF hospitalization, n (%)		168 (41)	49 (49)	35 (34)	43 (43)	41 (41)	0.23
Aetiology, n (%)							0.15
Cardiomyopathy		114 (28)	22 (22)	31 (30)	28 (28)	33 (33)
Valvular heart disease		170 (42)	41 (41)	35 (34)	45 (45)	49 (49)
Ischaemic heart disease		68 (17)	21 (21)	19 (19)	18 (18)	10 (10)
Comorbidity, n (%)
Hypertension		291 (72)	80 (79)	69 (68)	65 (64)	77 (76)	0.06
Dyslipidaemia		239 (59)	69 (68)	57 (56)	53 (52)	60 (59)	0.12
Diabetes mellitus		168 (41)	51 (51)	39 (38)	40 (40)	38 (38)	0.20
Atrial fibrillation		174 (43)	48 (48)	40 (39)	40 (39)	37 (37)	0.22
Chronic kidney disease	11 (3)	266 (68)	84 (86)	60 (61)	58 (60)	64 (64)	<0.01
Physical frailty	95 (23)	196 (63)	56 (82)	43 (54)	44 (59)	53 (60)	<0.01
Laboratory data
NT-proBNP (pg/mL)		1202 (576–2645)	1429 (698–3681)	1086 (495–2365)	1195 (532–2645)	1027 (597–2087)	0.08
Albumin (g/dL)		3.5 ± 0.4	3.3 ± 0.5	3.5 ± 0.4	3.5 ± 0.4	3.6 ± 0.4	<0.01
Haemoglobin (g/dL)		11.6 ± 1.7	11.1 ± 1.5	11.9 ± 1.8	11.8 ± 1.8	11.7 ± 1.6	<0.01
eGFRcys (mL/min/1.73 m2)	11 (3)	50.4 ± 19.0	40.2 ± 17.1	54.0 ± 17.6	54.7 ± 18.4	52.7 ± 19.2	<0.01
Blood urea nitrogen		24.5 ± 12.2	28.0 ± 15.8	22.3 ± 9.6	24.1 ± 11.5	23.7 ± 10.5	<0.01
Uric acid		5.8 ± 1.6	6.1 ± 1.6	6.0 ± 1.7	5.6 ± 1.6	5.5 ± 1.4	0.02
Medication, n (%)
Beta-blocker		276 (68)	72 (71)	76 (75)	69 (68)	59 (58)	0.08
ACE-I or ARB or ARNI		218 (54)	49 (49)	60 (59)	56 (55)	53 (52)	0.50
MRA		179 (44)	47 (47)	45 (44)	48 (48)	39 (39)	0.58
Loop diuretics		248 (61)	70 (69)	63 (62)	62 (61)	53 (52)	0.11
Tolvaptan		77 (19)	28 (28)	14 (14)	19 (19)	16 (16)	0.06
SGLT2 inhibitor		27 (1)	4 (4)	9 (9)	3 (3)	11 (11)	0.07
Nutritional status
MNA-SF score		8.1 ± 2.8	6.8 ± 2.4	8.0 ± 2.8	8.5 ± 2.9	9.0 ± 2.7	<0.01
Energy intake
kcal/day		1430 ± 291	1157 ± 369	1439 ± 218	1528 ± 156	1596 ± 143	<0.01
kcal/kg IBW/day		26.5 ± 5.4	21.2 ± 6.2	25.3 ± 2.9	28.0 ± 2.4	31.4 ± 2.7	<0.01
Non-proteogenic energy intake
kcal/day		1176 ± 251	974 ± 342	1180 ± 195	1249 ± 141	1302 ± 130	<0.01
% energy intake		82.5 ± 2.6	84.5 ± 3.8	82.0 ± 1.9	81.7 ± 1.4	81.7 ± 1.2	<0.01
Protein intake
g/day		62.1 ± 14.0	42.4 ± 10.4	64.1 ± 8.5	69.4 ± 5.6	72.6 ± 4.1	<0.01
g/kg IBW/day		1.15 ± 0.26	0.78 ± 0.17	1.13 ± 0.06	1.27 ± 0.06	1.43 ± 0.11	<0.01
% energy intake		17.5 ± 2.6	15.5 ± 3.8	18.0 ± 1.9	18.3 ± 1.4	18.3 ± 1.6	<0.01
Carbohydrate intake (g/day)		198.5 ± 45.2	166.4 ± 61.2	198.9 ± 36.3	210.1 ± 29.7	218.4 ± 26.4	<0.01
Fat intake (g/day)		42.5 ± 9.7	34.3 ± 12.9	42.7 ± 8.1	45.4 ± 4.7	47.6 ± 4.4	<0.01
Salt intake (g/day)		5.5 ± 1.1	4.3 ± 1.4	5.6 ± 0.7	5.9 ± 0.4	6.1 ± 0.5	<0.01
Physical function
Handgrip strength (kg)	91 (22)	21.6 ± 7.9	19.1 ± 7.5	24.1 ± 9.0	20.8 ± 7.0	21.8 ± 7.2	<0.01
Low handgrip strength, n (%)	91 (22)	203 (65)	56 (80)	46 (57)	50 (67)	51 (58)	0.01
Gait speed (m/s)	22 (5)	0.825 ± 0.260	0.744 ± 0.243	0.855 ± 0.279	0.846 ± 0.249	0.848 ± 0.255	<0.01
Slow gait speed, n (%)	22 (5)	292 (76)	78 (86)	69 (71)	68 (71)	77 (78)	0.06
ASMI (kg/m2)	95 (23)	5.71 ± 0.95	5.66 ± 1.05	5.82 ± 0.97	5.88 ± 1.14	6.08 ± 0.97	0.10
Muscle wasting, n (%)	95 (23)	281 (74)	50 (78)	58 (67)	49 (64)	40 (48)	<0.01
Clinical endpoints, n (%)		100 (25)	37 (37)	21 (21)	21 (21)	21 (21)	0.02
All-cause death		41 (10)	18 (18)	9 (9)	8 (8)	6 (6)	0.045
HF-related readmission		59 (15)	19 (19)	12 (12)	13 (13)	15 (15)

Association between daily protein intake and clinical parameters

The mean DPI was 62.1 ± 14.0 g/day and averaged at 1.15 ± 0.26 g/kg IBW/day. Upon stratification of patients into sex-specific quartiles based on DPI per IBW, no significant differences were observed in BMI, incidence of NYHA Functional Class III or IV, LVEF, prevalence of prior HF hospitalization, hypertension, dyslipidaemia, DM, atrial fibrillation, aetiology of HF, log NT-proBNP level, and medication usage across the quartiles (Table 1). However, patients in the Q1 group (i.e. the group with the lowest DPI) exhibited an elevated prevalence of physical frailty and muscle wasting, diminished levels of albumin and haemoglobin, and reduced eGFRcys. In contrast, blood urea nitrogen and uric acid levels were higher in the Q1 group compared with the upper quartile groups. Concurrently, a decline in the DPI was associated with a reduction in the MNA-SF score, energy intake, non-proteogenic energy intake, and intake of carbohydrates, fat, and salt. Additionally, the Q1 group exhibited weaker handgrip strength and slower gait speed than the other groups, underscoring the correlation between lower DPI and impaired physical performance.

Impact of daily protein intake on composite endpoints in older patients with HF

During an average follow-up period of 1.49 ± 0.74 years, 100 (25%) patients experienced composite events. Kaplan–Meier survival curves revealed that the composite event-free rate was significantly lower for patients in the Q1 group than for those in the upper quartile groups (composite event rate, 37%, 21%, 21%, and 21% for the first, second, third, and fourth quartiles, respectively; log-rank test, P = 0.02; Figure 2).

[IMAGE OMITTED. SEE PDF]

To compare the effects of DPI on the incidence of composite events between quartiles, we used Q4 (the highest quartile) as the reference category. The univariate Cox proportional hazards analysis showed that patients in the Q1 group had a two-fold higher risk of composite events compared with those in the Q4 group [HR, 2.06; 95% confidence interval (CI), 1.21–3.53; P < 0.01; Table 2]. A significant increase in risk of composite events in the Q1 group was also found after adjustment with potential confounders, including markers of HF severity, such as NT-proBNP level, NYHA functional class, and non-proteogenic energy intake (HR, 2.03; 95% CI, 1.08–3.82; P = 0.03; Table 2).

Table 2 Cox proportional hazards analyses for composite events

	DPI, sex-specific quartile	DPI, continuous
Q1 (lowest)	Q2	Q3	Q4 (highest)	per 1.0 SD decrease
HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value	HR	HR (95% CI)	P value
Unadjusted model	2.06 (1.21–3.53)	<0.01	0.96 (0.52–1.75)	0.88	0.98 (0.54–1.80)	0.96	1.00 (ref.)	1.37 (1.10–1.71)	<0.01
Adjusted modela	2.03 (1.08–3.82)	0.03	1.23 (0.65–2.34)	0.52	1.03 (0.56–1.92)	0.92	1.00 (ref.)	1.32 (1.03–1.71)	0.03

To further examine the effects of DPI on composite events, we assessed the fully adjusted dose-dependent association between DPI per IBW and the composite event rate using a Cox proportional hazards model with a restricted cubic spline function comprising four knots. As expected, the risk of composite events linearly increased with decreasing DPI per IBW (P for nonlinearity = 0.90, Figure 3). In the univariate Cox proportional hazards model, each standard deviation (0.26 g/kg IBW/day) decrease in DPI per IBW was associated with a 37% increase in composite event risk (HR, 1.37; 95% CI, 1.10–1.71; P < 0.01; Table 2). This association persisted even after adjusting for covariates, with a 33% increase in composite event risk (HR, 1.33; 95% CI, 1.02–1.70; P = 0.04; Table 2). Sensitivity analyses confirmed the similar results: each standard deviation (14.0 g/day) decrease of DPI per g/day was associated with higher cumulative composite event rates in both univariate (HR, 1.37; 95% CI, 1.10–1.71; P < 0.01) and multivariate (HR, 1.29; 95% CI, 1.03–1.60; P = 0.03) models.

[IMAGE OMITTED. SEE PDF]

We examined the association between DPI and the incidence of composite events across the subgroups of interest (Figure 4). In a multivariate model in which potential confounders were included, there was significant heterogeneity in the effect of DPI with the possible disadvantage of lower DPI in patients with HF with eGFRcys < 30 mL/min/1.73 m² (P for interaction = 0.03, Figure 4). The possible detrimental effect of inadequate DPI in patients with HF and severe renal impairment was confirmed by the results of the fully adjusted dose-dependent association between DPI per IBW and the composite event rates according to subgroups of patients with different levels of renal dysfunction (eGFRcys: <45 vs. ≥45 mL/min/1.73 m² or <30 vs. ≥30 mL/min/1.73 m², Supporting Information, Figure S1).

[IMAGE OMITTED. SEE PDF]

Effects of daily protein intake on the prediction of composite events in older patients with HF: determination of optimal daily protein intake

Considering the finding that DPI independently predicts composite events in older patients with HF, we investigated whether evaluating DPI according to IBW improves the accuracy of predicting the incidence of composite events in this group. Incorporation of DPI into the baseline model did not enhance the C-index but yielded a significant improvement in cNRI (Table 3). A DPI of 1.12 g/kg IBW/day was the optimal cutoff point, as determined by the Youden index (Figure 5A). Incorporating a DPI < 1.12 g/kg IBW/day into the baseline model significantly improved the cNRI (Table 3). Furthermore, the dose–response analyses of the C-index and cNRI across various cutoff points of DPI revealed the highest improvements in a DPI of ~1.1 g/kg IBW/day (Figure 5B–D).

Table 3 Effects of daily protein intake on the prediction of composite outcomes in older patients with HF

Model	C-index (95% CI)	P value	cNRI (95% CI)	P value
Baseline model (ref.)	0.744 (0.687–0.795)	Ref.		Ref.
+ DPI, continuous	0.750 (0.693–0.800)	0.43	0.290 (0.067–0.514)	0.01
+ DPI, sex-specific quartiles	0.750 (0.693–0.800)	0.47	0.271 (0.047–0.495)	0.02
+ DPI, <1.12 g/kg IBW/day (vs. ≥1.12 g/kg IBW/day)	0.757 (0.701–0.806)	0.17	0.294 (0.072–0.516)	0.01

[IMAGE OMITTED. SEE PDF]

Discussion

Several studies have analysed the association between DPI and clinical outcomes in patients with HF. A prospective study by Streng et al. found an association between a lower estimated protein intake and a higher risk of mortality; however, the analyses were not adjusted for energy intake, which is a powerful predictor of all-cause mortality in older patients with HF.⁹ In addition to the methodological limitation by protein intake estimation based on urinary nitrogen concentrations, the applicability of this study's finding to older patients with HF may be limited, given the significant differences in participant demographics from ours: the average age was 68 years (vs. our average of 79 years), only 27% of participants were women (vs. 50% in our study), and median BMI between the first and fourth quartiles ranged from 24.2 to 30.0 kg/m² (vs. our mean BMI that ranged from 21.5 to 22.5 kg/m²). A single-centre, retrospective cohort study by Saijo et al., in which 165 older patients with HF (median age, 82.0 years; 49% women) were at risk of malnutrition, found that a low DPI (<1.2 g/kg actual body weight/day) during hospitalization was an independent risk factor for readmission within 1 year after discharge.²⁶ However, the study results were not adjusted for key established prognostic markers, such as NT-proBNP level, prior HF admission, NYHA functional class, and medication usage. To the best of our knowledge, this is the first study to demonstrate that the association between low DPI and increased adverse events in older patients with HF is independent of insufficient non-proteogenic energy intake and HF severity. Therefore, integrating the DPI into pre-established prognostic predictors is a useful strategy for accurate risk stratification in older patients with HF.

Optimal daily protein intake in older patients with HF

There are no recommendations and statements for DPI to improve prognosis in patients with HF, although the Academy of Nutrition and Dietetics guideline recommends a protein intake of at least 1.1 g/kg actual body weight to prevent catabolism in patients with HF.²⁷ In the present study, a protein intake of 1.12 g/kg/day was the optimal cutoff value to predict favourable clinical outcomes, and a protein intake of ≥1.12 g/kg/day improved the prognostic ability of established prognostic factors. The DPI associated with better clinical outcomes in the present study was consistent with the ESPEN guidelines for older adults with chronic diseases.^8,28 Thus, the ESPEN guideline recommendations seem reasonable for older patients with HF; however, prospective studies that include prognosis, physical function, and activities of daily living as endpoints are clearly required. In addition, whether higher protein intake is better for older patients should be analysed in the future.

The underlying mechanism of the association between low daily protein intake and poor clinical outcomes

Low DPI is not only a surrogate marker of malnutrition and low energy intake as low DPI is an independent prognostic factor even after adjusting for non-proteogenic energy intake. The results of previous randomized controlled studies have revealed that supplementation with proteins or amino acids improves physical function and muscle strength and prevents hospitalization, but conflicting findings have also been reported. An increase in DPI may affect the improvement of physical function and the maintenance of skeletal muscle mass. The possibility that the resulting increase in myokines, cytokines, and other peptides produced and released by muscle fibres may lead to an improvement in HF and its comorbidities is an interesting hypothesis. However, data on the direct cardiac effects of protein intake are limited. For example, the consumption of fish and shellfish may contribute to increased taurine intake, and taurine improves various diseases, including HF, by modulating calcium metabolism and its antioxidant/anti-inflammatory properties.²⁹ Thus, the bystander effects of macronutrients/micronutrients obtained from protein intake should be considered.

Optimal daily protein intake in older patients with HF and renal dysfunction

Of particular note are our findings on the interaction between DPI and CKD severity (eGFRcys < 30 or ≥30 mL/min/1.73 m²). Some of the catabolic products of proteins, such as p-cresyl sulfate and indoxyl sulfate, serve as uraemic toxins. Furthermore, protein load has a detrimental effect on nephron function, primarily through glomerular hyperfiltration. Certainly, the results of several clinical studies have suggested a favourable effect of protein restriction on the decline in renal function,^30–32 although conflicting findings regarding its effect on mortality have been reported.³³ For this reason, recent Kidney Disease Outcomes Quality Initiative clinical practice guideline recommends a protein intake of 0.55–0.6 g/kg/day for patients with CKD Stages 3–5 and a protein intake of 0.6–0.8 g/kg/day for those with DM in the absence of any active diseases, such as inflammatory/infectious diseases and cancer, a recent history of hospitalization, and short-term loss of body weight.^34,35 In contrast, a higher DPI was not associated with a faster decline in estimated glomerular filtration rate (eGFR) among Japanese community-dwelling older adults.³⁶ Taken together with our findings in older patients with HF, older patients with CKD taking a DPI of >0.8 g/kg/day have a significantly lower risk of all-cause mortality compared with those taking a protein intake of 0.6–0.8 g/kg/day, although relatively high DPI was related to a faster decline in the eGFR. These discrepant results are not easily reconciled; however, the maintenance of body mass, including muscle mass, through sufficient protein intake may overcome the suppressive effect of protein restriction on eGFR decline in older patients. Nevertheless, caution is warranted when protein restriction therapy is administered to older patients with HF and CKD, particularly those with sarcopenia or frailty.

Study strengths and limitations

This study has several notable strengths. First, the sample size was determined using rigorous calculations to ensure sufficient statistical power for the primary outcome. Second, our study incorporated a subgroup analysis to delineate the severity of renal dysfunction. This stratified approach allowed us to examine the effect of DPI across a spectrum of renal impairment levels in older patients with HF, thus providing a nuanced understanding of the role of DPI in diverse patient cohorts. Fourth, we performed multiple imputations using a chained equation to account for potential selection bias and bolster the efficiency of the analysis, thereby offsetting any inaccuracies resulting from missing data. Finally, we calculated eGFR using cystatin C, which is currently the most accurate marker of renal function and is not influenced by glomerular hyperfiltration.^37,38 Moreover, unlike creatinine, cystatin C levels were not influenced by dietary meat intake and muscle mass.^39–42

Notably, this study has some limitations. First, because this was a retrospective observational study and patients had missing data in a single centre, there could have been a selection bias in the study participants even after multiple imputation techniques. Additionally, the present cohort included patients aged ≥65 years with a higher prevalence of physical frailty and muscle wasting. Therefore, the results may not be generalizable to younger patients or those without such comorbidities. Second, due to the small number of older patients with HF, the present study may not have enough power to detect differences between specific groups. Therefore, we performed post hoc analyses to separately analyse patients with LVEF < 40%, that is, HF with reduced ejection fraction (HFrEF), and those with LVEF ≥ 40%, that is, non-HFrEF (Supporting Information, Tables S1–S4). Patients with non-HFrEF were older than those with HFrEF, had a higher proportion of females, and had a lower prevalence of muscle wasting. There were significant differences in nutritional status and dietary intake: patients with non-HFrEF had higher MNA-SF scores than those with HFrEF, whereas patients with non-HFrEF had lower DPI than those with HFrEF (Supporting Information, Table S1). Multivariate Cox proportional hazards analysis showed that non-HFrEF patients in the Q1 group had a three-fold higher risk of composite events than those in the Q4 group (HR, 2.99; 95% CI, 1.37–6.55; P < 0.01); however, DPI had no impact on the incidence of composite events in patients with HFrEF (HR, 0.96; 95% CI, 0.39–2.36; P = 0.93; Supporting Information, Table S4). However, given the small number of older patients with HFrEF (n = 120), drawing conclusions from these results appears challenging. Therefore, the effects of DPI on clinical outcomes in specific groups (men vs. women, HFrEF vs. non-HFrEF, and patients with and without muscle wasting) should be analysed in a future study with a large sample size. Third, patients with HF who died in the hospital could not be included in the analyses because DPI was assessed for three consecutive days before discharge. Fourth, race-dependent variations in anthropometric parameters, including BMI, enable the application of the results of this study to other ethnicities. Fifth, protein intake was calculated using amounts visually estimated and not directly measured. Sixth, DPI during the three consecutive days prior to discharge may not necessarily reflect post-discharge DPI. Future longitudinal studies are essential to examine the long-term clinical impacts of post-discharge DPI changes. Seventh, IBW is generally defined as body weight divided by height, showing the lowest risk of mortality, but the lowest BMI-associated mortality is higher in older individuals than in younger individuals.⁴³ Although our analyses were adjusted for BMI, the optimal DPI should be analysed according to different BMI categories in future studies. Finally, we could not collect data on the dietary sources of protein (i.e. animal or plant protein), which were differently associated with mortality in the geriatric population and patients with CKD.⁴⁴

Conclusions

A lower DPI during hospitalization is associated with an increased mortality risk and HF readmission independent of non-proteogenic energy intake. A DPI > 1.12 g/kg IBW/day was observed alongside lower rates of adverse clinical outcomes in older patients with HF, suggesting a potential area for further investigation. These insights may inform nutritional management considerations for older patients with HF including those with severely impaired renal function, where the question of enhancing or restricting DPI frequently arises. Further interventional studies are required to verify these findings and explore the potential benefits and safety of protein supplementation in this high-risk population.

Acknowledgements

We are grateful to the participants of the study and the staff at Sapporo Medical University Hospital.

Conflict of interest

None declared.

Funding

This study was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI [Grant Numbers JP20K19313 (R. Nagaoka) and JP22K11288 (S.K.)], a grant from the Yuasa Memorial Foundation, the Hokkaido Heart Association Grant for Research, and the Kondou Kinen Medical Foundation. The funding body played no role in the design of the study; in the collection, analysis, and interpretation of data; and in writing the manuscript.

References

Savarese G, Becher PM, Lund LH, Seferovic P, Rosano GMC, Coats AJS. Global burden of heart failure: A comprehensive and updated review of epidemiology. Cardiovasc Res 2022;118:3272‐3287. doi:10.1093/cvr/cvac013

Chen R, Xu J, Wang Y, Jiang B, Xu X, Lan Y, et al. Prevalence of sarcopenia and its association with clinical outcomes in heart failure: An updated meta‐analysis and systematic review. Clin Cardiol 2023;46:260‐268. doi:10.1002/clc.23970

Matsue Y, Kamiya K, Saito H, Saito K, Ogasahara Y, Maekawa E, et al. Prevalence and prognostic impact of the coexistence of multiple frailty domains in elderly patients with heart failure: The FRAGILE‐HF cohort study. Eur J Heart Fail 2020;22:2112‐2119. doi:10.1002/ejhf.1926

Katano S, Honma S, Nagaoka R, Numazawa R, Yamano K, Fujisawa Y, et al. Anthropometric parameters‐derived estimation of muscle mass predicts all‐cause mortality in heart failure patients. ESC Heart Fail 2022;9:4358‐4365. doi:10.1002/ehf2.14121

Josiak K, Jankowska EA, Piepoli MF, Banasiak W, Ponikowski P. Skeletal myopathy in patients with chronic heart failure: Significance of anabolic‐androgenic hormones. J Cachexia Sarcopenia Muscle 2014;5:287‐296. doi:10.1007/s13539‐014‐0152‐z

Damluji AA, Alfaraidhy M, AlHajri N, Rohant NN, Kumar M, Malouf CA, et al. Sarcopenia and cardiovascular diseases. Circulation 2023;147:1534‐1553. doi:10.1161/circulationaha.123.064071

Bauer J, Morley JE, Schols AMWJ, Ferrucci L, Cruz‐Jentoft AJ, Dent E, et al. Sarcopenia: A time for action. An SCWD position paper. J Cachexia Sarcopenia Muscle 2019;10:956‐961. doi:10.1002/jcsm.12483

Deutz NEP, Bauer JM, Barazzoni R, Biolo G, Boirie Y, Bosy‐Westphal A, et al. Protein intake and exercise for optimal muscle function with aging: Recommendations from the ESPEN Expert Group. Clin Nutr 2014;33:929‐936. doi:10.1016/j.clnu.2014.04.007

Streng KW, Hillege HL, Maaten JM, Veldhuisen DJ, Dickstein K, Ng LL, et al. Clinical implications of low estimated protein intake in patients with heart failure. J Cachexia Sarcopenia Muscle 2022;13:1762‐1770. doi:10.1002/jcsm.12973

Tsutsui H, Isobe M, Ito H, Ito H, Okumura K, Ono M, et al. JCS 2017/JHFS 2017 guideline on diagnosis and treatment of acute and chronic heart failure—Digest version. Circ J 2019;83:2084‐2184. doi:10.1253/circj.cj‐19‐0342

Katano S, Hashimoto A, Ohori K, Watanabe A, Honma R, Yanase R, et al. Nutritional status and energy intake as predictors of functional status after cardiac rehabilitation in elderly inpatients with heart failure—A retrospective cohort study. Circ J 2018;82:1584‐1591. doi:10.1253/circj.cj‐17‐1202

Katano S, Yano T, Kouzu H, Ohori K, Shimomura K, Honma S, et al. Energy intake during hospital stay predicts all‐cause mortality after discharge independently of nutritional status in elderly heart failure patients. Clin Res Cardiol 2021;110:1202‐1220. doi:10.1007/s00392‐020‐01774‐y

Kawasaki Y, Sakai M, Nishimura K, Fujiwara K, Fujisaki K, Shimpo M, et al. Criterion validity of the visual estimation method for determining patients' meal intake in a community hospital. Clin Nutr 2016;35:1543‐1549. doi:10.1016/j.clnu.2016.04.006

Kawasaki Y, Akamatsu R, Tamaura Y, Sakai M, Fujiwara K, Tsutsuura S. Differences in the validity of a visual estimation method for determining patients' meal intake between various meal types and supplied food items. Clin Nutr 2019;38:213‐219. doi:10.1016/j.clnu.2018.01.031

Ikizler TA, Cano NJ, Franch H, Fouque D, Himmelfarb J, Kalantar‐Zadeh K, et al. Prevention and treatment of protein energy wasting in chronic kidney disease patients: A consensus statement by the International Society of Renal Nutrition and Metabolism. Kidney Int 2013;84:1096‐1107. doi:10.1038/ki.2013.147

Kaiser MJ, Bauer JM, Ramsch C, Uter W, Guigoz Y, et al. Validation of the Mini Nutritional Assessment Short‐Form (MNA®‐SF): A practical tool for identification of nutritional status. J Nutr Health Aging 2009;13:782‐788. doi:10.1007/s12603‐009‐0214‐7

Katano S, Yano T, Tsukada T, Kouzu H, Honma S, Inoue T, et al. Clinical risk factors and prognostic impact of osteoporosis in patients with chronic heart failure. Circ J 2020;84:2224‐2234. doi:10.1253/circj.cj‐20‐0593

Satake S, Arai H. The revised Japanese version of the Cardiovascular Health Study criteria (revised J‐CHS criteria). Geriatr Gerontol Int 2020;20:992‐993. doi:10.1111/ggi.14005

Ohori K, Yano T, Katano S, Nagaoka R, Numazawa R, Yamano K, et al. Coexistence of sarcopenia and self‐reported weight loss is a powerful predictor of mortality in older patients with heart failure. Geriatr Gerontol Int 2023;24:95‐101. doi:10.1111/ggi.14778

Nagaoka R, Katano S, Yano T, Numazawa R, Yamano K, Fujisawa Y, et al. Optimal serum 25‐hydroxyvitamin D level to prevent sarcopenia in patients with heart failure: Insights from a dose‐response relationship. Nutr Metab Cardiovasc Dis 2023;34:606‐617. doi:10.1016/j.numecd.2023.10.003

Horio M, Imai E, Yasuda Y, Watanabe T, Matsuo S, Collaborators Developing the Japanese Equation for Estimated GFR. GFR estimation using standardized serum cystatin C in Japan. Am J Kidney Dis 2013;61:197‐203. doi:10.1053/j.ajkd.2012.07.007

Austin PC, White IR, Lee DS, van Buuren S. Missing data in clinical research: A tutorial on multiple imputation. Can J Cardiol 2020;37:1322‐1331. doi:10.1016/j.cjca.2020.11.010

Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epidemiologic studies. Am J Clin Nutr 1997;65:1220S‐1228S. doi:10.1093/ajcn/65.4.1220s

DeLong ER, DeLong DM, Clarke‐Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics 1988;44:837. doi:10.2307/2531595

Konishi M, Akiyama E, Matsuzawa Y, Sato R, Kikuchi S, Nakahashi H, et al. Prognostic impact of muscle and fat mass in patients with heart failure. J Cachexia Sarcopenia Muscle 2021;12:568‐576. doi:10.1002/jcsm.12702

Saijo T, Yasumoto K, Ryomoto K, Momoki C, Habu D. Association of protein intake during hospitalization with readmission in older adult patients with heart failure at risk of malnutrition. Nutr Clin Pract 2023;38:686‐697. doi:10.1002/ncp.10957

Kuehneman T, Gregory M, de Waal D, Davidson P, Frickel R, King C, et al. Academy of Nutrition and Dietetics evidence‐based practice guideline for the management of heart failure in adults. J Acad Nutr Diet 2018;118:2331‐2345. doi:10.1016/j.jand.2018.03.004

Volkert D, Beck AM, Cederholm T, Cruz‐Jentoft A, Hooper L, Kiesswetter E, et al. ESPEN practical guideline: Clinical nutrition and hydration in geriatrics. Clin Nutr 2022;41:958‐989. doi:10.1016/j.clnu.2022.01.024

Qaradakhi T, Gadanec LK, McSweeney KR, Abraham JR, Apostolopoulos V, Zulli A. The anti‐inflammatory effect of taurine on cardiovascular disease. Nutrients 2020;12: [eLocator: 2847]. doi:10.3390/nu12092847

Yan B, Su X, Xu B, Qiao X, Wang L. Effect of diet protein restriction on progression of chronic kidney disease: A systematic review and meta‐analysis. PLoS ONE 2018;13: [eLocator: e0206134]. doi:10.1371/journal.pone.0206134

Jiang Z, Zhang X, Yang L, Li Z, Qin W. Effect of restricted protein diet supplemented with keto analogues in chronic kidney disease: A systematic review and meta‐analysis. Int Urol Nephrol 2016;48:409‐418. doi:10.1007/s11255‐015‐1170‐2

Esmeijer K, Geleijnse JM, de Fijter JW, Kromhout D, Hoogeveen EK. Dietary protein intake and kidney function decline after myocardial infarction: The Alpha Omega Cohort. Nephrol Dial Transpl 2019;35:106‐115. doi:10.1093/ndt/gfz015

Naghshi S, Sadeghi O, Willett WC, Esmaillzadeh A. Dietary intake of total, animal, and plant proteins and risk of all cause, cardiovascular, and cancer mortality: Systematic review and dose‐response meta‐analysis of prospective cohort studies. BMJ 2020;370: [eLocator: m2412]. doi:10.1136/bmj.m2412

Ikizler TA, Burrowes JD, Byham‐Gray LD, Campbell KL, Carrero J‐J, Chan W, et al. KDOQI clinical practice guideline for nutrition in CKD: 2020 update. Am J Kidney Dis 2020;76:S1‐S107. doi:10.1053/j.ajkd.2020.05.006

Kistler BM, Moore LW, Benner D, Biruete A, Boaz M, Brunori G, et al. The International Society of Renal Nutrition and Metabolism commentary on the National Kidney Foundation and Academy of Nutrition and Dietetics KDOQI clinical practice guideline for nutrition in chronic kidney disease. J Ren Nutr 2021;31:116‐120.e1. doi:10.1053/j.jrn.2020.05.002

Sekiguchi T, Kabayama M, Ryuno H, Tanaka K, Kiyoshige E, Akagi Y, et al. Association between protein intake and changes in renal function among Japanese community‐dwelling older people: The SONIC study. Geriatr Gerontol Int 2022;22:286‐291. doi:10.1111/ggi.14355

Dharnidharka VR, Kwon C, Stevens G. Serum cystatin C is superior to serum creatinine as a marker of kidney function: A meta‐analysis. Am J Kidney Dis 2002;40:221‐226. doi:10.1053/ajkd.2002.34487

Huang S‐HS, Sharma AP, Yasin A, Lindsay RM, Clark WF, Filler G. Hyperfiltration affects accuracy of creatinine eGFR measurement. Clin J Am Soc Nephrol 2011;6:274‐280. doi:10.2215/cjn.02760310

Ishigo T, Katano S, Yano T, Kouzu H, Ohori K, Nakata H, et al. Overestimation of glomerular filtration rate by creatinine‐based equation in heart failure patients is predicted by a novel scoring system. Geriatr Gerontol Int 2020;20:752‐758. doi:10.1111/ggi.13959

Butani L, Polinsky MS, Kaiser BA, Baluarte HJ. Dietary protein intake significantly affects the serum creatinine concentration. Kidney Int 2002;61:1907. doi:10.1046/j.1523‐1755.2002.00342.x

Tangri N, Stevens LA, Schmid CH, Zhang Y, Beck GJ, Greene T, et al. Changes in dietary protein intake has no effect on serum cystatin C levels independent of the glomerular filtration rate. Kidney Int 2011;79:471‐477. doi:10.1038/ki.2010.431

Nair S, O'Brien SV, Hayden K, Pandya B, Lisboa PJG, Hardy KJ, et al. Effect of a cooked meat meal on serum creatinine and estimated glomerular filtration rate in diabetes‐related kidney disease. Diabetes Care 2014;37:483‐487. doi:10.2337/dc13‐1770

Bhaskaran K, dos‐Santos‐Silva I, Leon DA, Douglas IJ, Smeeth L. Association of BMI with overall and cause‐specific mortality: A population‐based cohort study of 3·6 million adults in the UK. Lancet Diabetes Endocrinol 2018;6:944‐953. doi:10.1016/s2213‐8587(18)30288‐2

Bernier‐Jean A, Prince RL, Lewis JR, Craig JC, Hodgson JM, Lim WH, et al. Dietary plant and animal protein intake and decline in estimated glomerular filtration rate among elderly women: A 10‐year longitudinal cohort study. Nephrol Dial Transplant 2020;36:1640‐1647. doi:10.1093/ndt/gfaa081