The dietary intake of chronic kidney disease according to stages: Findings from the Korean National Health and Nutritional Examination Survey

Appropriate dietary adjustment in patients with chronic kidney disease (CKD) is important, and nutritional guidelines recommend different dietary management depending on the CKD stage. However, there is no study, to our knowledge, of the characteristics of dietary intake according to CKD stages. We tried to assess the comparison of nutritional intake according to CKD stages. A cross-sectional study was conducted to reveal the characteristics of dietary intake among patients with CKD based on the Korean National Health and Nutritional Examination Survey between 2011 and 2014. Of 16,878 participants, we classified non-CKD (n = 14,952) and CKD (n = 1,926), which was stratified into five groups (I, II, IIIa, IIIb, and IV–V). We investigated the characteristics of dietary intake, such as energy, water, protein, fat, carbohydrate, sodium, potassium, calcium, and phosphorus, according to stage of CKD. We also explored nutritional intake according to CKD stage among patients with early CKD (stage I and II) and advanced CKD (stage IIIa, IIIb, and IV–V). Intake of majority of nutrients and energy tended to be decreased as CKD progressed. In early CKD stage, intake of energy, water, protein, fat, carbohydrate, potassium, calcium and phosphorus seemed to be statistically significant decreased as CKD progressed. In advanced CKD stage, intake of potassium and calcium seemed to be decreased as CKD progressed, but the intake of energy was about to be lower limit. Appropriate dietary education and CKD recognition are needed to improve nutritional intake depending on the CKD stage.


Introduction
The prevalence of chronic kidney disease (CKD) is estimated to be 8-16% worldwide [1], and over 2 million people have end-stage renal disease (ESRD) [2], which is a life-threatening outcome of CKD that requires renal replacement therapy for survival. In the Republic of Korea, the prevalence of CKD in individuals aged >20 years was 8.2% [3]. CKD is also one of the most important chronic illnesses that imposes a substantial disease in an aging society [4].
CKD treatment mainly aims to preserve renal function, as renal function continues to decrease with age and renal disease progression or complications. Other than established modifiable risk factors for CKD such as hypertension, diabetes mellitus (DM), and dyslipidemia, recent studies and guidelines suggest that nutrition is an important factor in CKD progression [2]. The significance of proper diet in CKD patients was confirmed in a large retrospective cohort study: the mortality rate was significantly lower in predialysis patients who received care from a dietitian compared to those who did not [5]. Protein-energy wasting is one of the strongest predictors of mortality in patients with CKD [6,7]. However, nutrition and dietary patterns are often neglected as a therapeutic tools for preventing and slowing CKD progression [8].
CKD is categorized into stages, with symptoms across stages. According to the Kidney Disease Improving Global Outcomes (KDIGO) guidelines, CKD is classified into stage I-V, according to the extent of albuminuria and estimated glomerular filtration rate (eGFR) [9]. Because different clinical manifestations and adverse outcomes are observed in each CKD stage, different management strategies including dietary control, are needed to prevent CKD progression. However, most studies on nutrition in CKD patients have been focused on terminal-stage CKD, or have been conducted regardless of the CKD stage [10][11][12][13].
Thus, this study examined the dietary composition of CKD patients, and investigated whether dietary intake was altered according to the CKD stage based on the Korean National Health and Nutrition Examination Survey (KNHANES) between 2011 and 2014.

Subjects
KNHANES was a population-based, cross-sectional study of health and nutritional status of the non-institutionalized Korean population. The current study obtained data from the third (2005), fourth (2007)(2008)(2009), fifth (2010-2012), sixth (2013)(2014)(2015), and seventh (2016-2017) years of KNHANES. The Korean Center for Disease Control and Prevention (KCDC) conducted the survey using a stratified, multi-stage, clustered probability design to select a representative, nationwide sample [3]. Of subjects participating in KNHANES 2011-2014 (n = 24,948), we excluded people who did not have information about kidney function (albuminuria, serum creatinine, and eGFR; n = 4,712). Moreover, we excluded subjects who did not have information about covariates, such as comorbidities (n = 1,123), and data about dietary intake (n = 2,235). In total, 16,878 subjects were included in the final analysis, and were stratified into five groups (non-CKD, stage I, stage II, stage IIIa, stage IIIb, and CKD stage IV-V) (Fig 1).
The study was performed in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of Ewha Womans University Medical Center (EUMC 2019-06-005). All subjects provided written informed consent to participate in the study.

Data collection
1. KNHANES. The nutritional survey of KNHANES consisted of dietary behavior, 24-hour recall, and food frequency questionnaires (FFQ). Data were collected by trained dietitians, one week after a face-to-face health interview. Daily energy intake was calculated using the Korean Foods and Nutrients Database of the Rural Development Administration [14].
2. Demographic and clinical characteristics. We investigated age, sex, body mass index (BMI), comorbidities such as hypertension and DM, the use of nutritional education, awareness of CKD awareness, income, occupation, residential district, and laboratory findings (serum creatinine, urinary albumin-to-creatinine ratio total cholesterol, and high-density lipoprotein).
The presence of CKD was established, based on the presence of kidney damage and level of kidney function; eGFR was calculated from standardized serum creatinine level using the Modification of Diet in Renal Disease (MDRD) equation [15]. Spot urinary albumin-to-creatinine ratio (UACR) was also calculated with mg of albumin per g of creatinine (mg/g). Based on KDIGO guidelines, we stratified CKD into five groups: stage I (eGFR �90 mL/min/1.73 m 2 , UACR �30 mg/g); stage II (eGFR 60-89 mL/min/1.73 m 2 , UACR �30 mg/g); stage IIIa (eGFR 45-59 mL/min/1.73 m 2 ); stage IIIb (eGFR 30-44 mL/min/1.73 m 2 ); and stage IV-V (eGFR <30 mL/min/1.73 m 2 ). However, there were only a few patients (n = 39) in stage IV-V, so we decided to classify these two stages as one group. Further, we stratified the CKD patients into two groups: early CKD (stage I-II) versus advanced CKD (stages III-V), and explored the characteristics of dietary intake in each group [16].
Hypertension was defined as subjects with systolic blood pressure �140 mmHg, diastolic blood pressure �90 mmHg, or who had been taking antihypertensive medication. In addition, DM was defined as patients with serum fasting glucose level �126 mg/dL, who had been taking antidiabetic medicine, or who had a previous diagnosis of DM.
The use of nutritional education was defined as nutritional counseling conducted by a dietitian in a public health center, ward office, resident center, welfare facility, school, or hospital in the previous year. CKD awareness was regarded as when subjects recognized their diagnosis of CKD or had been treated for CKD. Based on their monthly income, the populations were stratified into four groups: Q1, <$1,000; Q2, $1,000-$2,000; Q3, $2,000-$3,000; and Q4, >$3,000. Occupation status included three categories: participate worked for more than one hour per week, worked for more than eighteen hours here week at a family run business without pay, or between jobs.
When patients lived in one of the seven metropolitan cities (Seoul, Incheon, Daejeon, Daegu, Kwangju, Ulsan, and Busan), they were considered urban dwellers, whereas participants from other areas were considered rural dwellers.
3. Statistical analysis. Baseline characteristics and laboratory findings were presented as means and standard deviations or as medians and interquartile ranges. The variables were tested for normal distribution using the Kolmogorov-Smirnov test, and then differences between stages were tested using ANOVA test (normal distribution) or the Mann-Whitney test (non-normal distribution) for continuous variables. Categorical variables were compared by χ 2 -test. Based on CKD stage, the trends of nutritional intake were compared by Spearman's correlation. We used R (version 3.5.0, Statistical computing, Vienna, Austria) to reveal trends in nutrient intake according to CKD stage using graphical methods. Statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc, Cary, NC, USA), and a p value <0.05 was considered statistically significant. Table 1 shows baseline characteristics of the non-CKD and CKD stage I-V groups. Of 16,878 participants, 14,952 (88.6%) were non-CKD and 1,926 (11.4%) had CKD: 416 (2.5%) were in CKD stage I, 705 (4.2%) in stage II, 659 (3.9%) in stage IIIa, 107 (0.6%) in stage IIIb, and 39 (0.2%) in stage IV-V. In the total population of 16,878, mean age was 51.4 ± 16.2 years, and 7,109 participants (42.1%) were male. In addition, 5,364 subjects (31.8%) had hypertension, and 1,921 (11.4%) had DM. Nutritional education was given to only 708 people (4.2%). Regarding laboratory findings, median value of serum creatinine, UACR, total cholesterol, and HDL levels were 0.80 mg/dL, 3.9 mg/g, 187 mg/dL, and 49.4 mg/dL respectively. When subjects were stratified into five groups according to CKD stage and non-CKD, we found significant differences between groups in mean age, the proportion of males, BMI, the incidence of hypertension and DM, income, CKD awareness, occupation, residential district, and laboratory findings, while there was no significant difference in the incidence of nutritional education (Table 1).

Comparisons of nutritional intake between groups
We compared dietary intake between the groups using 24-hour recall and FFQ (Table 2). We found significant differences in dietary intake between groups: in particular, the intake of

<0.001
Data are expressed as mean ± standard deviations (normal distribution) or as median and interquartile range (non-normal distribution).
2) Diabetes mellitus: Serum fasting glucose level �126 mg/dL, subjects who had been taking anti-diabetic medicine, or had a previous diagnosis of diabetes mellitus.
3) Nutritional education: Nutritional counseling conducted by a dietitian in a public health center, ward office, resident center, welfare facility, school, or hospital in the previous year.

PLOS ONE
Characteristics of nutrient according to CKD stage energy, water, protein, fat, carbohydrate, sodium, potassium, calcium, and phosphorus seemed to decrease as CKD progressed. Moreover, we chose six nutrients (protein, energy, sodium, potassium, calcium, and phosphorus) recommended in the KDIGO guidelines [9]. As CKD progressed, the intakes of energy, water, protein, sodium, potassium, calcium, and phosphorus tended to decrease (Fig 2).

Nutritional intake in the early CKD and advanced CKD group
We assumed that a difference in dietary intake exists between early-stage and advanced-stage CKD, so we explored nutritional intake after stratifying the population into two groups (early CKD and advanced CKD). Moreover, non-CKD subjects were included in the early CKD group, since (except for albuminuria) early-stage CKD is characterized by normal renal function. As seen in Table 3, the intakes of energy, water, protein, fat, carbohydrate, potassium, calcium and phosphorus were significantly decreased in CKD stage II versus stage I and non-CKD. In advanced CKD, the intakes of potassium and calcium were remarkably decreased as CKD progressed which was seen in Table 4.

Discussion
Appropriate dietary intake is essential for CKD patients, to prevent or delay ESRD progression and the occurrence of CKD-related complications [13,17]. To our knowledge, no study has investigated dietary intake according to the CKD stage. In this study, we found significant differences in dietary intake according to the CKD stage and observed different patterns of dietary intake between the early CKD and advanced CKD group. The KDIGO guidelines suggest the need to stratify CKD based on kidney damage and function, which means that different clinical manifestations and adverse outcomes are expected in each stage, and physicians may need to take care of patients with different strategies, including dietary control, in accordance with the CKD status. Identification of the major nutrient sources can help establish effective dietary strategies targeted towards the nutritional care of  CKD patients. Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines recommend that CKD patients limit their intake of proteins, sodium, phosphorus, and potassium to maintain their serum nutrient levels within the normal range [18]. Dietary restriction of proteins is inevitably less than the rest of the nutrients, thereby causing nutritional imbalance. Thus, the effect of a low-protein diet remains controversial, although it proved beneficial in preventing the deterioration of renal function, it did not reduce all-cause mortality [19]. In this study, we also found that the intake of most nutrients decreases with CKD progression. Moreover, lipid profile is a potential tool to assess protein-energy wasting in CKD patients [20], and the median values of cholesterol and HDL tended to decrease with the progression of CKD stage ( Table 1). Considering that the nutrient intake and lipid profile showed a declining association with CKD progression, the risk of malnutrition has become a concerning in the context of CKD.
We explored nutritional intake after stratifying the population into two groups, early CKD and advanced CKD. Early CKD is defined by an eGFR of �60 ml/min/1.73 m 2 with substantial proteinuria, whereas advanced CKD is characterized by an eGFR of <60 ml/min/ 1.73 m 2 . Early CKD is free from nutritional restriction compared with advanced CKD, but it can be seen that the intake of most nutrients decreases with CKD progression. Moreover, in advanced CKD, the intake of potassium and calcium seemed to decrease with CKD progression, but the intake of energy was the approximate lower limit. Protein intake restriction is thought to play an important role in CKD progression. However, the published researches and meta-analyses on the effect of protein restriction cannot answer this question in clinical practice, such as kidney failure events, all-cause mortality, and nutritional status [19,21,22]. Therefore, most nephrologists recommend no restrictions or only mild restrictions on protein intake (0.8-1 g/ kg daily). We showed that protein and energy intake tended to decrease as the CKD stage progressed. Adherence to a low-protein diet can be improved by ensuring adequate energy intake [23]; therefore, fat and carbohydrate content should together account for more than 90% of the daily energy intake requirement of 30-35 kcal/kg to avoid protein-energy wasting, which can be achieved by increasing the trend of energy intake with CKD progression [24]. Although the intake of carbohydrate and fat was slightly increased in the CKD IV-V, it was insufficient to maintain the total energy intake. Appropriate glycemic control is important in advanced CKD because DM prevalence increases with the progression of CKD stage (Table 1). It is therefore necessary to tailor the diet to the patient's condition, and not just adopt a one-sizefits-all approach. Nutrient intake in CKD patients should consider the patients' overall metabolic state and comorbid conditions. Owing to such abovementioned concerns, nutritional interventions with disease-specific dietary ranges are needed. To address this, we focused on the CKD awareness rate. There were few participants with CKD awareness (41 patients, 2.1%) among all CKD patients (1,926 patients); moreover, only 33 patients (4.1%) in the advanced CKD group perceived their kidney problems in this study. In 1999-2000, the US National Health and Nutrition Examination Survey (NHANES) demonstrated that only 2% of the adult population self-reported recognition of a weak or failing kidney [25]. Moreover, between 1999 and 2004, NHANES surveys showed that CKD awareness rates were only 8% for CKD stage 3, but 41% for stage 4 [26]. The low proportion of CKD awareness in our study is consistent with previous studies. Although several trials have been proposed to improve CKD awareness, it is difficult to find evidence of such enhanced awareness [27]. In this study, we found that the rate of awareness not only in advanced CKD but also early CKD was lower than 1%. James et al. emphasized that early recognition of CKD [28], and early interventions (including dietary control), might reduce the risk of progression to kidney failure. Thus, in the future, the effectiveness of CKD awareness and changing dietary intake in CKD patients will need to be investigated in larger CKD-recognized populations. A better understanding of the renal condition is required among patients with CKD to facilitate the detection of this disease and implement therapeutic strategies for minimizing the associated complications.
Although proper nutritional education is considered a first step for adequate nutrition in CKD patients [29], only a few advanced CKD patients recognized that they had received nutritional education. Despite the higher rate of nutritional education in CKD progression, this study showed that the rate of nutritional education was generally poor. Consistent with a previous study that emphasized the importance of dietary education among predialysis patients to reduce mortality [5], more efforts are needed to provide effective education for CKD. In the Republic of Korea, insurance-covered nutritional education programs are available for CKD patients at initial diagnosis and when dialysis is started. Thus, nutritional education may have a reduced influence on altering dietary intake as CKD progresses. New educational approaches are being developed through research and quality improvement efforts to overcome the challenges of improving CKD awareness. The optimal diet for CKD patients remains controversial depending on the kidney function, type of kidney disease and the presence of other comorbidities such as DM, hypertension, or heart failure. Therefore, effective strategy to increase the effectiveness of CKD education is to provide customized education for individuals with CKD. With this approach, patients' adherence and compliance must be considered when proper nutrients are prescribed in advanced CKD. In this study, inappropriate nutrition patterns were identified across different CKD stages, and the first step toward correcting this pattern was to enhance the extent of nutritional education and CKD awareness.
There are some limitations to this study. First, the study was designed as a cross-sectional, retrospective trial, and it was challenging to validate the causal relationship between current nutritional intake and CKD deterioration. Second, although the FFQ used in this study had been validated elsewhere, there are several limitations associated with the questionnaire alone. FFQ can be underestimated because of the inadequate coverage of all available food items consumed by an individual [30]. Moreover, dietary assessments via FFQ and 24-h recall are limiting to the determination of dietary variations. Third in this study, 805 patients (4.8%) among the total participants were in the advanced CKD group. In particular, only a small number of patients (0.2%) belonged to the CKD IV-V group. The prevalence of CKD in the current study was lower than that of CKD in the Republic of Korea [3], which might have been underestimated. Fourth, we used MDRD equation for calculating eGFR. The accuracy of the CKD-EPI equation was significantly greater than the equation used in the MDRD study for patients with an eGFR of >60 ml per minute per 1.73 m 2 or a BMI of >30 kg per m 2 . Thus, we believe that it is better to use CKD-EPI equation for the South Korean population. Finally, when we investigated CKD awareness, this was regarded as when subjects recognized they had been diagnosed with CKD or had been treated for CKD. Thus, CKD recognition was somewhat dependent on memory of the participants. Moreover, nutritional education was also examined based on participant memory, which means that CKD awareness and nutritional education could be underestimated. Despite these limitations, this is the first study to reveal the characteristics of dietary intake according to CKD stage. To our knowledge, this is the first report to estimate the nutrient status across the CKD stages I-V. As dietary information was systematically collected, we were able to compare results among individuals at different stages of CKD and compare their dietary patterns with those of individuals without CKD.
In conclusion, we observed varying dietary intake patterns across several CKD stages and recognized some flaws in the nutritional status at each stage. Therefore, we believe that regular assessment of dietary protein, energy, and micronutrient intake is necessary. In the future, prospective longitudinal studies with larger populations are needed to corroborate the association of between CKD awareness, nutritional education, and CKD progression.

Conclusion
We identified differences in nutritional intake among patients with CKD who were stratified according to their CKD stage. The risk of nutritional imbalance should be recognized in light of the declining trend seen for the intake rates of a majority of nutrients depending on CKD progression. In early CKD stage, the intake of energy, water, protein, fat, carbohydrate, potassium, calcium, and phosphorus seemed to significantly decrease as CKD progressed. In advanced CKD stage, the intake potassium and calcium seemed to decrease with CKD progression, but the intake of energy was about to be lower limit. Appropriate dietary education and CKD awareness is needed to improve nutritional intake across various CKD stages.