Abstract
Background
Chronic Kidney Disease (CKD) often progresses to End-Stage Renal Disease (ESRD), requiring dialysis or transplantation. This study aimed to examine the relationship between the Energy-Adjusted Dietary Inflammatory Index (E-DII) and metabolic parameters, including glycemic indices, lipid profiles, renal and hepatic functions, blood pressure, and anthropometric measurements, in hemodialysis patients.
Methods
In this exploratory cross-sectional study, 300 hemodialysis patients from Ahvaz, Iran, were enrolled. Dietary intake was assessed using a 168-item Food Frequency Questionnaire (FFQ), and E-DII scores were calculated to assess dietary inflammation. Key metabolic parameters, including body mass index (BMI), waist circumference (WC), fasting blood sugar (FBS), lipid profiles, blood pressure, renal and hepatic function markers, and physical activity levels, were measured using standardized protocols. Statistical analyses were performed to evaluate the associations between E-DII scores and these outcomes.
Results
Higher E-DII scores, indicating diets with greater inflammatory potential, were significantly associated with increased post-dialysis BMI and WC (P < 0.001); however, these measures may be influenced by fluid status in hemodialysis patients. higher fasting blood sugar (P < 0.001), and dyslipidemia, including elevated triglycerides (P < 0.001), total cholesterol (P = 0.03), and LDL-c (P = 0.02). Renal function was also affected, with higher plasma creatinine (P = 0.01) and uric acid levels (P < 0.001) in patients with higher E-DII scores. Blood pressure was significantly elevated across increasing E-DII quartiles (P < 0.001). Physical activity levels were inversely associated with E-DII scores (P < 0.001).
Conclusions
Higher E-DII scores were associated with unfavorable metabolic outcomes in this exploratory cross-sectional sample of hemodialysis patients. Associations with anthropometric measures should be interpreted with caution, as post-dialysis body dimensions are influenced by fluid status. These findings suggest an association that supports the need for further longitudinal research to confirm these relationships and evaluate the potential benefits of anti-inflammatory dietary strategies. Given the exploratory nature of this analysis and the absence of adjustment for multiple comparisons, results should be interpreted with caution and require confirmation in independent cohorts. Further research using full DII components (including herbs and spices) and prospective designs is required to assess the benefits of anti-inflammatory diets in this population.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12882-026-05047-0.
Keywords: Hemodialysis, Chronic kidney disease, Metabolic disease, Inflammation
Introduction
CKD is a major global health burden with increasing prevalence, progressing to ESRD that requires dialysis or kidney transplantation. Patients on hemodialysis face heightened chronic inflammation, accelerating metabolic complications such as cardiovascular disease, dyslipidemia, hyperglycemia, hypertension, and obesity the leading causes of morbidity and mortality in this population [1–4].
Diet plays a critical role in managing CKD progression and comorbidities. Kidney-friendly diets help limit waste accumulation, control hypertension, and mitigate cardiovascular and bone issues [5, 6]. Inflammation, a key driver of these complications, can be modulated by dietary patterns.
The Dietary Inflammatory Index (DII), developed in 2009 and updated in 2014, quantifies the inflammatory potential of diets based on pro- and anti-inflammatory effects of foods, macronutrients, micronutrients, and bioactive compounds [7, 8]. The DII is well-validated and scores diets on a continuum from anti-inflammatory (low scores) to pro-inflammatory (high scores) [9–13].
Additionally, higher DII scores have been positively associated with BMI, WC, and other anthropometric measures. Higher DII scores associate with metabolic syndrome, elevated blood pressure, triglycerides, total and Low-Density Lipoprotein Cholesterol (LDL-c), reduced High-Density Lipoprotein Cholesterol (HDL-c), hyperglycemia, central obesity, and increased inflammatory markers [11, 13–15]. Pro-inflammatory diets also link to higher risks of cardiovascular diseases, diabetes, cancer, and mortality comorbidities that exacerbate CKD progression [16–20]. Recent studies show positive associations between higher DII and adverse metabolic profiles in CKD patients, including diabetes, hypertension, dyslipidemia, and inflammation [21]. Additionally, energy-adjusted DII (E-DII) accounts for total energy intake, improving precision in assessing dietary inflammation.
Most DII research focuses on general or early-stage CKD populations, with limited data on hemodialysis patients who experience amplified inflammation due to dialysis-related factors. Given the high burden of metabolic disturbances in this vulnerable group and the scarcity of studies in Middle Eastern populations, particularly Iran, this cross-sectional study investigates the association between the Energy-Adjusted Dietary Inflammatory Index (E-DII) and metabolic parameters (glycemic indices, lipid profile, blood pressure, renal and hepatic function markers) as well as anthropometric indices in hemodialysis patients in Ahvaz, Iran.
Methods
Population and study design
This analytical cross-sectional study was conducted among adult hemodialysis patients from four hemodialysis centers in Ahvaz, Iran, between March 2023 and September 2023.
Recruitment process and timing: Initially, a list of all hemodialysis centers in Ahvaz was obtained from the Ahvaz Jundishapur University of Medical Sciences. The research team then visited each of the four hemodialysis centers to obtain the names of all patients undergoing maintenance hemodialysis. Patients were approached for participation during their routine mid-week hemodialysis sessions. Specifically, the researcher (a nutritionist) attended each center on fixed weekdays (Tuesday and Wednesday) to coincide with the mid-week dialysis shift, when patients are typically more stable and available for interview. After the completion of the dialysis session, eligible patients were invited to participate.
Informed consent
The study objectives, procedures, potential risks, and benefits were fully explained to each patient by the researcher. Written informed consent was obtained from all participants prior to any data collection, dietary assessment, or anthropometric measurement. Patients were assured of their voluntary participation and the confidentiality of their data.
Eligibility criteria and sampling method: From the master list of patients across the four centers, individuals who met the eligibility criteria were identified (n = 452). Eligibility criteria included: (1) age ≥ 18 years, (2) undergoing maintenance hemodialysis for at least three months, (3) clinically stable condition with no acute illness or hospitalization in the past month, and (4) willingness to participate. Exclusion criteria included: (1) extreme energy intake (< 800 or > 4200 kcal/day), (2) current smoking, (3) use of non-steroidal anti-inflammatory drugs (NSAIDs), steroids, or antioxidant supplements in the past three months, (4) incomplete Food Frequency Questionnaire (FFQ) with > 70 missing items, (5) underlying inflammatory diseases (e.g., rheumatoid arthritis, lupus, active infection), (6) active malignancy, or (7) unwillingness to participate.
Of the 452 eligible patients, 152 were excluded for the following reasons: extreme energy intake (n = 62), smoking (n = 38), use of NSAIDs/steroids or antioxidant supplements (n = 22), and other criteria including incomplete FFQ, underlying inflammatory diseases, or unwillingness to participate (n = 30). Consecutive sampling (also referred to as consecutive enrollment) was then applied, whereby all remaining eligible patients who met the inclusion criteria and provided informed consent were included in the final analysis (n = 300). No further randomization or selection method was applied beyond the exclusion criteria. The flow diagram of the recruitment and sampling process is illustrated in Fig. 1.
Fig 1. The flow diagram of the sampling process. Of 452 patients initially assessed for eligibility across four hemodialysis centers, 152 were excluded due to extreme energy intake (n = 62), smoking (n = 38), use of NSAIDs/steroids or antioxidant supplements (n = 22), and other reasons including incomplete FFQ or comorbidities (n = 30). The remaining 300 eligible patients were all included in the final analysis (no further sampling method was applied)
The study was initiated following the approval of the ethics committee of Ahvaz Jundishapur University of Medical Sciences (IR.AJUMS.REC.1401.483) and was conducted in accordance with the Declaration of Helsinki.
Sample size
The sample size was calculated based on the primary outcome, which was the difference in body mass index (BMI) across Energy-Adjusted Dietary Inflammatory Index (E-DII) quartiles. Using data from a previous study by Arab et al. (2022) [31]. which reported a standard deviation of 0.43 kg/m² and an expected clinically meaningful difference of 0.43 kg/m² between groups, with α = 0.05 and 90% power, the minimum required sample size was 284 participants. Accounting for a potential 5% non-response or incomplete data rate, the final target sample size was set at 300 participants.
Although post-hoc power calculations are of limited added value in observational studies and are not recommended for interpreting results, we provide them for transparency. Post-hoc power analyses (performed using G*Power 3.1 software) confirmed that with the achieved sample of n = 300, the study had sufficient power (> 80%) for detecting clinically meaningful differences in the main secondary outcomes:
Fasting blood sugar (expected difference = 15 mg/dL, SD = 30 mg/dL): power ≈ 92%
Triglycerides (expected difference = 40 mg/dL, SD = 80 mg/dL): power ≈ 88%
Systolic blood pressure (expected difference = 8 mmHg, SD = 15 mmHg): power ≈ 95%
Waist circumference (expected difference = 6 cm, SD = 10 cm): power ≈ 98%
Dietary assessment
Food intake information was collected using a 168-item semi-quantitative Food Frequency Questionnaire (FFQ) [22]. The participants were asked about the frequency of consumption of each item in the questionnaire over the past year. Depending on the type of food, the frequency was reported in terms of daily, weekly, monthly, or yearly consumption. Standard portion sizes and items reported using home scales were converted to grams using the Nutritionist 4 Home Scales Guide. The daily intake of each food item was then calculated and reported in grams per day. To determine the intake of energy, macronutrients, and micronutrients, the gram equivalents of each food item reported in the FFQ were analyzed using data from the food composition table specifically adapted for Iranian food items. This allowed for precise calculation of nutrient intake based on the reported food consumption. The 168-item semi-quantitative FFQ used in this study is a validated tool adapted for Iranian food items, with demonstrated acceptable reliability and relative validity for nutrient intake estimation in Iranian adults [22].
EDII calculation
To calculate the energy-adjusted Dietary Inflammatory Index (E-DII), we employed the method proposed by Shivappa et al. Before the E-DII calculation, the energy-adjusted amount of each food item was determined using the residual technique [23]. Of the 45 dietary items suggested by Shivappa et al., 28 food items were available for the E-DII calculation, including vitamins A, D, E, B1, B2, B3, B6, B9, B12, C, β-carotene, n-3 fatty acids, n-6 fatty acids, cholesterol, saturated fatty acids (SFA), trans fatty acids (TFA), polyunsaturated fatty acids (PUFA), mono-unsaturated fatty acids (MUFA), magnesium, zinc, iron, selenium, caffeine, dietary fiber, carbohydrate, fat, protein, and energy. Importantly, 17 items from the original DII could not be calculated, including several anti-inflammatory herbs and spices (e.g., turmeric, ginger, garlic, pepper, rosemary, thyme, oregano, saffron). The absence of these components may affect the validity of the E-DII as an exposure measure and should be considered when interpreting our findings.
The calculation process involved the following steps:
Standardization: Each participant’s dietary intake was subtracted from the “standard global mean” and divided by the “global standard deviation” to compute the Z score for each dietary parameter.
Transformation: The Z score for each food item was transformed into a centered percentile to minimize data skewness.
Weighting: These centered percentiles were then multiplied by the inflammatory score of each respective food item.
Summation: The overall E-DII for each participant was calculated by summing the weighted inflammatory scores of the 34 food items.
According to Shivappa et al., the DII scores range from − 8 to + 8, with higher values indicating a diet with pro-inflammatory properties and lower values indicating a diet with anti-inflammatory features [24].
Data collection and measurements
At the outset of the study, informed consent forms were completed, and questions were administered by the researcher (a nutritionist). Data collection was performed using a comprehensive multi-part questionnaire. This questionnaire encompassed demographic and anthropometric information (age, sex, education, marital status, occupation, race, income, smoking, and alcohol consumption, medications, height in centimeters, weight in kilograms, body mass index in kilograms per square meter, waist circumference in centimeters, and hip circumference in centimeters) as well as metabolic parameters (glycemic, lipid, kidney, liver, and blood pressure).
In this study, dry weight at the end of the dialysis session was measured using a body composition monitor digital scale (manufactured in Japan) with an accuracy of 0.1 kg, without shoes and with minimal clothing. Height was measured using a tape measure with an accuracy of 0.5 cm, with participants standing straight against a wall. Waist circumference was measured at the upper part of the iliac crest and below the navel, and hip circumference was measured at the most prominent part of the hip using a tape measure. BMI was calculated using the formula (weight in kilograms/height in meters squared). It is important to note that in hemodialysis patients, post-dialysis ‘dry weight’ reflects not only adiposity but also extracellular fluid status. Short-term weight fluctuations in this population are often driven by fluid accumulation (interdialytic weight gain) and ultrafiltration volume rather than changes in adipose or muscle tissue. We did not systematically collect data on interdialytic weight gain (IDWG), ultrafiltration (UF) volume, or bioimpedance-based fluid overload indices. Therefore, our anthropometric measures may partially reflect fluid status rather than true adiposity. Physical activity was assessed using the validated short form of the International Physical Activity Questionnaire (IPAQ), a quantitative instrument designed to estimate weekly physical activity in metabolic equivalent task minutes (MET-min/week) across multiple domains (work, transportation, domestic, and leisure) [25]. The short IPAQ consists of seven questions covering vigorous-intensity activities (8 METs), moderate-intensity activities (4 METs), walking (3.3 METs), and sedentary time. Total physical activity was calculated as the sum of MET-min/week for walking, moderate, and vigorous activities. Scores were then categorized into three levels: low (< 600 MET-min/week), moderate (600–3000 MET-min/week), and high (> 3000 MET-min/week), provided that activities were performed in bouts of at least 10 min [25]. The IPAQ short form has been widely validated internationally, including acceptable reliability and validity in Iranian populations. Serum levels of fasting blood sugar, lipid profile, liver and kidney indices, dialysis adequacy, urine volume, 24-hour protein excretion, and serum levels of sodium, potassium, calcium, phosphorus, and blood pressure were extracted from the patient’s medical records. Kidney function tests (including blood urea nitrogen [BUN], plasma creatinine, uric acid, and electrolytes) were performed immediately before the start and immediately after the end of the mid-week hemodialysis session (the same session in which anthropometric measurements were taken). All participants underwent these tests under standardized conditions: fasting state (at least 8 h), using the same laboratory analyzers across all centers, and following identical pre- and post-dialysis sampling protocols as per routine clinical practice in the participating hemodialysis units. Biochemical parameters were extracted from medical records obtained within 1 month (median 12 days, range 0–28 days) of the dietary assessment interview (FFQ administration). Given that the FFQ queried habitual dietary intake over the past year, it is unlikely that recent laboratory test results significantly influenced participants’ recall of long-term dietary habits. However, we acknowledge that the 0–28 day time window between dietary assessment and blood sampling introduces potential temporal variability. Biochemical parameters can fluctuate over time due to factors such as changes in dialysis prescription, medication adjustments, or intercurrent illnesses. This temporal mismatch may attenuate true associations or introduce non-differential misclassification. Future studies should aim to collect dietary and biochemical data on the same day to minimize this variability.
Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) creatinine Eq. (2009 version) based on serum creatinine, age, sex, and race (assuming non-Black ethnicity, as appropriate for the Iranian population). This equation was chosen because it is widely recommended for estimating GFR in adults with chronic kidney disease, provides greater accuracy compared to the MDRD equation (particularly at higher GFR levels), and has been commonly used in epidemiological studies involving Iranian CKD patients. No cystatin C measurements were available, so a creatinine-based equation was employed. Note that in end-stage renal disease patients on hemodialysis, eGFR primarily reflects residual renal function and has limited clinical utility compared to dialysis adequacy measures.
Hemodialysis adequacy, which refers to the efficiency of toxin and waste removal from the patient’s blood, was also assessed, as it significantly impacts patient health, quality of life, and life expectancy [26–30]. Inadequate dialysis has been associated with increased morbidity and mortality [27]. Urea removal indices, which help in assessing dialysis adequacy, include the urea reduction ratio (URR), single-pool Kt/V (spKt/V), equilibrated Kt/V (eKt/V), and weekly standard Kt/V index (std Kt/V). This study utilized URR and spKt/V for calculating dialysis adequacy.
-
URR:
URR was assessed by measuring blood urea nitrogen (BUN) levels before and after dialysis (29). It is calculated as follows:
URR = (predialysis BUN-postdialysis BUN) / (predialysis BUN) × 100%
spKt/V:
The spKt/V index is defined as the amount of serum cleared of urea relative to the volume of distribution, based on the reduction ratio of urea during hemodialysis. Parameter K is the rate of dialysis blood urea removal by the filter (measured in liters per hour), t is the duration of the hemodialysis session in hours, and V is the volume of urea distribution, combined with body water, in liters. The spKt/V and URR parameters are mathematically calculated as follows:
 |
where ln denotes the natural logarithm.
The 2006 National Kidney Foundation-Kidney Disease Outcomes Quality Initiative (NKF-KDOQI) guidelines recommend a spKt/V > 1.2 or URR ≥ 65% for maintenance hemodialysis with a thrice-weekly schedule. These values represent the minimum acceptable levels, with target values set at 1.4 for spKt/V and 70% for URR [29].
Data collection procedures and timeline
Data collection was conducted over a six-month period from March 2023 to September 2023. The sequence of data collection for each participant was as follows:
First contact (during mid-week dialysis session): The researcher approached the patient immediately after the completion of the mid-week hemodialysis session. The study objectives were explained, and written informed consent was obtained.
Anthropometric measurements (same day, post-dialysis): Immediately after obtaining consent, anthropometric measurements (dry weight, height, waist circumference, hip circumference, neck circumference, wrist circumference) were taken using standardized protocols. All measurements were performed by the same trained researcher to minimize inter-observer variability.
Dietary assessment (within one week of consent): The 168-item semi-quantitative Food Frequency Questionnaire (FFQ) was administered by the researcher through a face-to-face interview, typically conducted on a non-dialysis day or during a subsequent dialysis session, based on patient availability. Participants were asked about their habitual food intake over the past year.
Physical activity assessment (same day as FFQ): The short form of the International Physical Activity Questionnaire (IPAQ) was administered to assess usual physical activity levels over the past seven days.
Biochemical data extraction (within one month of dietary assessment): Serum levels of fasting blood sugar, lipid profile, liver and kidney indices, electrolytes, and dialysis adequacy parameters (URR and spKt/V) were extracted from patients’ medical records. All biochemical parameters were obtained from routine laboratory tests performed within 1 month (median 12 days, range 0–28 days) of the dietary assessment interview. Kidney function tests were performed immediately before the start and immediately after the end of the mid-week hemodialysis session (the same session in which anthropometric measurements were taken) under standardized fasting conditions (at least 8 h) and using the same laboratory analyzers across all centers.
Assessment of confounders
Body mass index (BMI) was calculated using participants’ weight and height measurements taken at the end of the dialysis session. Dialysis vintage was defined as the duration each patient had been undergoing hemodialysis (HD), expressed in years. Dialysis adequacy was assessed using the Kt/V index, which incorporates dialysis session length, post-dialysis weight, ultrafiltration volume, and serum urea concentration before and after dialysis [30].
Statistical analysis
In this study, sampling will be conducted systematically, with the sample size determined based on the BMI variable from Arab et al.‘s study using the following formula:
N = [ (z1−α/2)2×sd2]/d2 (α = 0.05, confidence level of 95%, sd = 0.43, and d = 0.05%).
Given an alpha level of 0.05 (confidence level of 95%), standard deviation (sd) of 0.43, and margin of error (d) of 0.05, the calculated sample size is 284. Accounting for a 5% attrition rate, the final sample size is set at 300 participants [31].
All data will be entered and statistically analyzed using SPSS software (IBM SPSS Statistics, Armonk, USA) version 24. Quantitative data will be reported as mean ± standard deviation, while qualitative data will be reported as frequency (percentage). The chi-square test will be employed to compare qualitative results. For evaluating the relationship between the Dietary Inflammatory Index (DII) and dependent variables, linear regression analysis will be utilized, including modeling and adjustment for the effect of confounding factors. Additionally, to assess the risk of obesity in hemodialysis patients, logistic regression will be used, with appropriate modeling and adjustment for confounding factors. A p-value of less than 0.05 will be considered statistically significant. Selection of covariates for multivariable models was guided by prior literature and directed acyclic graphs (DAGs) to avoid overadjustment. Specifically, we did not include BMI as a covariate in models for metabolic parameters (e.g., FBS, lipid profile, blood pressure) because BMI may lie on the causal pathway between dietary inflammation and these outcomes. Including such mediators would represent overadjustment and could bias estimates toward the null. BMI was only included as a covariate in sensitivity analyses or when explicitly stated for anthropometric outcomes. We acknowledge that certain clinically relevant confounders were not available in our dataset, including dialysis prescription details (dialysate composition, sodium profiling), residual renal function beyond crude eGFR, and cause of ESRD (dialysis etiology). The absence of these variables may introduce residual confounding, and our findings should be interpreted with this limitation in mind.
The presence of linear trends across E-DII quartiles (p-trend) was evaluated by modeling the quartile-specific median E-DII value as a continuous variable in multivariable-adjusted regression models. This approach tests whether the outcome changes progressively and significantly with increasing levels of dietary inflammatory potential. In linear regression (for continuous dependent variables such as BMI, FBS, triglycerides, etc.), the beta coefficient represents the change in the outcome per unit increase in quartile median. In logistic regression (for obesity outcomes), the odds ratio reflects the change in risk across quartiles. P-trend values < 0.05 were considered statistically significant, supporting a monotonic association. In Tables 3 and 4, three models are presented: Model 1 (unadjusted), Model 2 (adjusted for age and sex), and Model 3 (fully adjusted). For metabolic outcomes (FBS, lipid profile, blood pressure, renal and liver parameters), Model 3 included age, sex, physical activity, race, job, marital status, education, duration of dialysis, chronic diseases, and medications. Importantly, BMI was not included as a covariate in Model 3 for these metabolic outcomes to avoid overadjustment, as BMI may be a mediator rather than a confounder. For anthropometric outcomes, BMI was also omitted from Model 3 for the same reason.
Table 3.
The association between EDII score (independent variable) with anthropometric indices, FBS, and lipid profile (dependent variables) in hemodialysis patients
| Variables | Model 1a | Model 2b | Model 3c | ||||||
|---|---|---|---|---|---|---|---|---|---|
| B (Unstandardized) |
SE | P-value | B (Unstandardized) |
SE | P-value | B (Unstandardized) |
SE | P-value | |
| Weight (kg) | 6.27 | 0.31 | 0 < 0.001 | 6.06 | 0.30 | 0 < 0.001 | 2.20 | 0.38 | 0 < 0.001 |
| BMI (kg/m2) | 2.09 | 0.09 | 0 < 0.001 | 2.12 | 0.09 | 0 < 0.001 | 0.76 | 0.08 | 0 < 0.001 |
| WC (cm) | 5.49 | 0.25 | 0 < 0.001 | 5.27 | 0.23 | 0 < 0.001 | 1.71 | 0.21 | 0 < 0.001 |
| HC (cm) | 3.58 | 0.19 | 0 < 0.001 | 3.80 | 0.17 | 0 < 0.001 | 1.17 | 0.14 | 0 < 0.001 |
| WHR | 0.02 | 0.00 | 0 < 0.001 | 0.01 | 0.00 | 0 < 0.001 | 0.00 | 0.00 | 0 < 0.001 |
| WHtR | 0.03 | 0.00 | 0 < 0.001 | 0.03 | 0.00 | 0 < 0.001 | 0.01 | 0.00 | 0 < 0.001 |
| Wrist circumference (cm) | 1.21 | 0.9 | 0 < 0.001 | 1.52 | 0.05 | 0 < 0.001 | 0.39 | 0.05 | 0 < 0.001 |
| Neck circumference (cm) | 1.61 | 0.09 | 0 < 0.001 | 1.57 | 0.08 | 0 < 0.001 | 0.39 | 0.09 | 0 < 0.001 |
| Body fram size | -0.50 | 0.04 | 0 < 0.001 | -0.44 | 0.03 | 0 < 0.001 | -0.16 | 0.03 | 0 < 0.001 |
| FBS (mg/dL) | 1.61 | 1.30 | 0.21 | 1.72 | 1.32 | 0.18 | 0.38 | 1.52 | 0.79 |
| TG (mg/dL) | 1.66 | 1.59 | 0.29 | 1.78 | 1.61 | 0.27 | 0.76 | 2.73 | 0.78 |
| TC (mg/dL) | 9.70 | 1.02 | 0 < 0.001 | 9.73 | 1.03 | 0 < 0.001 | 2.91 | 1.38 | 0.03 |
| LDL-c (mg/dL) | 7.24 | 0.74 | 0 < 0.001 | 7.23 | 0.75 | 0 < 0.001 | 2.31 | 1.02 | 0.02 |
| HDL-c (mg/dL) | -0.65 | 0.39 | 0.86 | 0.01 | 0.39 | 0.97 | 0.14 | 0.66 | 0.82 |
| LDL-c/HDL.c ratio | 0.10 | 0.01 | 0 < 0.001 | 0.10 | 0.01 | 0 < 0.001 | 0.03 | 0.02 | 0.12 |
| AIP | 0.04 | 0.00 | 0 < 0.001 | 0.04 | 0.00 | 0 < 0.001 | 0.01 | 0.00 | 0.18 |
| TC/HDL ratio | 0.14 | 0.02 | 0 < 0.001 | 0.13 | 0.02 | 0 < 0.001 | 0.04 | 0.03 | 0.14 |
P < 0.05 was considered as significant. (a) Model 1: linear regression analysis without adjustment, (b) Model 2: linear regression analysis with adjustment for age and sex, (c) Model 3: linear regression analysis with correction for age, sex, physical activity, race, job, marital status, education, BMI, duration of dialysis, chronic diseases, and medications
*. In model 3, BMI is not included in the model for anthropometric indices
Abbreviation: EDII: energy-adjusted dietary inflammatory index, BMI: body mass index, WC: waist circumference, HC: hip circumference, WHtR: waist-to-height ratio, WHR: waist-to-hip ratio, FBS: fasting blood sugar, TG: triglyceride, TC: total cholesterol, LDL-c: low-density lipoprotein cholesterol, HDL-c: high-density lipoprotein cholesterol, and AIP: atherogenic index of plasma
In Model 3 for metabolic outcomes (FBS, lipid profile, blood pressure, renal/liver parameters), BMI was not included as a covariate to avoid overadjustment, as BMI may lie on the causal pathway between dietary inflammation and these outcomes. For anthropometric outcomes, BMI was also omitted from Model 3 for the same reason
Table 4.
The association between EDII score (independent variable) with blood pressure parameters and kidney and liver indices (dependent variables) in hemodialysis patients
| Variables | Model 1a | Model 2b | Model 3c | ||||||
|---|---|---|---|---|---|---|---|---|---|
| B (Unstandardized) |
SE | P-value | B (Unstandardized) |
SE | P-value | B (Unstandardized) |
SE | P-value | |
| B.SBP (mmHg) | 8.41 | 0.41 | 0 < 0.001 | 8.52 | 0.42 | 0 < 0.001 | 1.52 | 0.24 | 0 < 0.001 |
| A.SBP (mmHg) | 7.69 | 0.37 | 0 < 0.001 | 7.81 | 0.37 | 0 < 0.001 | 1.77 | 0.28 | 0 < 0.001 |
| ΔSBP | 0.71 | 0.14 | 0 < 0.001 | 0.71 | 0.14 | 0 < 0.001 | -0.27 | 0.18 | 0.14 |
| B.DBP (mmHg) | 3.82 | 0.24 | 0 < 0.001 | 3.89 | 0.24 | 0 < 0.001 | 0.61 | 0.22 | 0.007 |
| A.DBP (mmHg) | 3.43 | 0.20 | 0 < 0.001 | 3.48 | 0.20 | 0 < 0.001 | 0.60 | 0.15 | 0 < 0.001 |
| ΔDBP | 0.38 | 0.11 | 0.001 | 0.41 | 0.11 | 0 < 0.001 | 0.00 | 0.13 | 0.96 |
| B.PP (mmHg) | 4.58 | 0.23 | 0 < 0.001 | 4.62 | 0.24 | 0 < 0.001 | 0.90 | 0.21 | 0 < 0.001 |
| A.PP (mmHg) | 4.26 | 0.20 | 0 < 0.001 | 4.32 | 0.20 | 0 < 0.001 | 1.17 | 0.21 | 0 < 0.001 |
| ΔPP | 0.32 | 0.15 | 0.03 | 0.30 | 0.15 | 0.05 | -0.28 | 0.22 | 0.20 |
| B.MAP (mmHg) | 5.35 | 0.29 | 0 < 0.001 | 5.43 | 0.29 | 0 < 0.001 | 0.92 | 0.21 | 0 < 0.001 |
| A.MAP (mmHg) | 4.85 | 0.25 | 0 < 0.001 | 4.92 | 0.25 | 0 < 0.001 | 0.99 | 0.18 | 0 < 0.001 |
| ΔMAP | 0.49 | 0.10 | 0 < 0.001 | 0.51 | 0.10 | 0 < 0.001 | -0.08 | 0.11 | 0.44 |
| B.BUN (mg/dL) | 6.86 | 0.41 | 0 < 0.001 | 6.95 | 0.41 | 0 < 0.001 | 2.95 | 0.53 | 0 < 0.001 |
| A.BUN (mg/dL) | 1.36 | 0.07 | 0 < 0.001 | 1.38 | 0.07 | 0 < 0.001 | 0.28 | 0.07 | 0 < 0.001 |
| ΔBUN = B.BUN-A.BUN | 5.49 | 0.35 | 0 < 0.001 | 5.57 | 0.36 | 0 < 0.001 | 2.62 | 0.48 | 0 < 0.001 |
| Plasma Cr (mg/dL) | 0.62 | 0.04 | 0 < 0.001 | 0.62 | 0.04 | 0 < 0.001 | 0.09 | 0.04 | 0.01 |
| Uric acid (mg/dl) | 0.71 | 0.03 | 0 < 0.001 | 0.72 | 0.03 | 0 < 0.001 | 0.20 | 0.03 | 0 < 0.001 |
| eGFR | 0.16 | 0.07 | 0.02 | 0.05 | 0.04 | 0.22 | 0.17 | 0.06 | 0.009 |
| Dialysis adequacy | 0.28 | 0.08 | 0.001 | 0.29 | 0.08 | 0.001 | 0.50 | 0.13 | 0 < 0.001 |
| Na (mEq/L) | 1.12 | 0.07 | 0 < 0.001 | 1.12 | 0.08 | 0 < 0.001 | 0.42 | 0.06 | 0 < 0.001 |
| K (mEq/L) | 0.26 | 0.01 | 0 < 0.001 | 0.27 | 0.01 | 0 < 0.001 | 0.05 | 0.00 | 0 < 0.001 |
| Ca (mg/dL) | 0.17 | 0.01 | 0 < 0.001 | 0.18 | 0.01 | 0 < 0.001 | 0.02 | 0.00 | 0 < 0.001 |
| P (mg/dL) | 0.34 | 0.01 | 0 < 0.001 | 0.34 | 0.01 | 0 < 0.001 | 0.08 | 0.01 | 0 < 0.001 |
| AST (IU/L) | 0.41 | 0.30 | 0.16 | 0.40 | 0.30 | 0.18 | 0.03 | 0.51 | 0.94 |
| ALT (IU/L) | 0.38 | 0.29 | 0.20 | 0.36 | 0.30 | 0.22 | 0.09 | 0.51 | 0.84 |
| AST/ALT | 0.00 | 0.00 | 0.21 | 0.00 | 0.00 | 0.18 | 0.00 | 0.00 | 0.03 |
P < 0.05 was considered as significant. (a) Model 1: linear regression analysis without adjustment, (b) Model 2: linear regression analysis with adjustment for age and sex, (c) Model 3: linear regression analysis with correction for age, sex, physical activity, race, job, marital status, education, BMI, duration of dialysis, chronic diseases, and medications
Abbreviation: EDII: energy-adjusted dietary inflammatory index, BMI: body mass index, B: before dialysis, A: after dialysis, SBP: systolic blood pressure, DBP: diastolic blood pressure, MAP: mean arterial pressure, PP: pulse pressure, BUN: blood urea nitrogen, Cr: creatinine, eGFR: estimated glomerular filtration rate, Na: sodium, K: potassium, Ca: calcium, P: phosphorus, AST: aspartate aminotransferase, ALT: alanine aminotransferase
ΔSBP = Afer SBP – befor SBP, ΔDBP = Afer DBP – befor DBP, ΔPP = Afer PP – befor PP, ΔMAP = Afer MAP – befor MAP, and ΔBUN = Afer BUN– befor BUN
In Model 3 for metabolic outcomes (FBS, lipid profile, blood pressure, renal/liver parameters), BMI was not included as a covariate to avoid overadjustment, as BMI may lie on the causal pathway between dietary inflammation and these outcomes. For anthropometric outcomes, BMI was also omitted from Model 3 for the same reason
Given the exploratory nature of this cross-sectional study and the large number of outcomes examined, we did not apply formal adjustment for multiple comparisons (e.g., Bonferroni or false discovery rate). We explicitly label our analyses as exploratory rather than confirmatory. The high correlation among related outcomes (e.g., anthropometric measures, lipid parameters, renal markers) and the consistency of findings across unadjusted and multivariable-adjusted models support the robustness of key associations. However, readers should be aware that some significant findings may have occurred by chance due to multiple testing, and results should be interpreted with appropriate caution. Confirmation in independent cohorts is needed before drawing definitive conclusions. There were no missing data for the primary exposure (E-DII) or main outcome variables in the final analytical sample (n = 300), as participants with incomplete FFQ data (≥ 70 missing items) were excluded during the eligibility screening phase (n = 30).
Sensitivity and subgroup analyses
To assess the robustness of our findings, we conducted several sensitivity analyses. First, we repeated the primary analyses after excluding participants with extreme energy intake (< 800 or > 4200 kcal/day) to evaluate the impact of these exclusions. Second, we used alternative definitions of central obesity (WHO criteria vs. IDF criteria) to test whether findings were consistent across different classification systems. Third, we performed subgroup analyses stratified by sex (male vs. female) and by diabetes status (diabetic vs. non-diabetic) to explore potential effect modification. These subgroup analyses should be interpreted as exploratory given the reduced sample sizes within strata. Detailed results of sensitivity and subgroup analyses are provided in Supplementary Tables S2 and S3.”
Results
Baseline characteristics across EDII quartiles
Table 1 indicates that the Energy-Adjusted Dietary Inflammatory Index (EDII) scores varied significantly across quartiles (P < 0.001), reflecting a clear gradient in dietary inflammatory potential among hemodialysis patients. There were no significant differences in mean age (P = 0.94) or height (P = 0.05) across quartiles. The sex distribution was significantly different, with a higher proportion of females in quartile 4 compared to lower quartiles (P < 0.001). Anthropometric measurements showed significant variations across EDII quartiles. For instance, mean weight increased from 65.2 kg in quartile 1 to 80.7 kg in quartile 4 (P < 0.001), and WC rose from 85.3 cm in quartile 1 to 101.4 cm in quartile 4 (P < 0.001). Significant differences were also observed in BMI and WC categories (P < 0.001), indicating that higher EDII scores corresponded to higher categories of both metrics. Marital status and job categories did not show significant differences (P = 0.56 and P = 0.35, respectively). However, educational levels and racial distribution exhibited significant variations across quartiles (P < 0.001 for both). Physical activity levels were significantly associated with EDII quartiles, with higher physical activity observed in the lower quartiles (P < 0.001). Dialysis duration did not differ significantly across quartiles (P = 0.62). Biochemical parameters such as FBS, TG, TC, LDL-c, HDL-c, LDL/HDL ratio, and TC/HDL ratio showed significant differences across EDII quartiles (P < 0.001). For example, FBS increased from 110.5 mg/dL in quartile 1 to 129.3 mg/dL in quartile 4 (P < 0.001). Blood pressure measurements, including B.SBP, A.SBP, B.DBP, A.DBP, B.PP, A.PP, B.MAP, and A.MAP, were significantly different across quartiles (P < 0.001). Kidney function parameters such as BUN, plasma Cr, and uric acid displayed significant differences, with plasma Cr increasing from 4.2 mg/dL in quartile 1 to 5.7 mg/dL in quartile 4 (P < 0.001). However, eGFR did not vary significantly across quartiles (P = 0.05). Dialysis adequacy and electrolytes, including Na, K, Ca, and P, were significantly different across quartiles (P < 0.001). Liver enzymes (AST, ALT, AST/ALT ratio) did not show significant differences (P > 0.05) (Table 1).
Table 1.
The characteristics at baseline across quartiles of EDII score in hemodialysis patients
| Characteristics (mean (SD) or N (%) | EDII quartiles | P - value | ||||
|---|---|---|---|---|---|---|
| Q1 (N = 75) |
Q2 (N = 75) |
Q3 (N = 75) |
Q4 (N = 75) |
Total (N = 300) |
||
| EDII score | -2.01 ± 0.54 a, b,c | -0.75 ± 0.29 d, e | 0.40 ± 0.35 f | 2.36 ± 0.91 | 0.00 ± 1.71 | < 0.001* |
| Age (years) | 53.28 ± 10.75 | 52.61 ± 11.50 | 53.50 ± 12.38 | 52.46 ± 13.57 | 52.96 ± 12.04 | 0.94* |
| Sex (N) (%) | < 0.001** | |||||
| Male | 36 (48) | 33 (44) | 38 (50.7) | 23 (30.7) | 130 (43.3) | |
| Female | 39 (52) | 42 (56) | 37 (49.3) | 52 (69.3) | 170 (56.7) | |
| Height (m) | 166.54 ± 8.36 | 166.61 ± 6.89 | 164.05 ± 7.67 | 167.73 ± 10.08 | 166.23 ± 8.40 | 0.05* |
| Weight (kg) | 60.85 ± 7.89 b, c | 60.70 ± 7.74 d, e | 74.20 ± 10.10 f | 86.28 ± 11.15 | 70.51 ± 14.13 | < 0.001* |
| BMI (kg/m2) | 21.90 ± 2.01 b, c | 21.81 ± 1.90 d, e | 27.55 ± 3.17 f | 30.63 ± 2.59 | 25.47 ± 4.51 | < 0.001* |
| BMI categories (N) (%) | < 0.001** | |||||
| Non-obese | 75 (100) | 74 (98.7) | 63 (84) | 34 (45.3) | 246 (82) | |
| Obese | 0 (0) | 1 (1.3) | 12 (16) | 41 (54.7) | 54 (18) | |
| WC (cm) | 78.17 ± 5.33 b, c | 78.28 ± 5.22 d, e | 90.56 ± 8.72 f | 101.22 ± 9.19 | 87.06 ± 12.08 | < 0.001* |
| WC categories (N) (%) | < 0.001** | |||||
| Normal | 75 (100) | 74 (98.7) | 58 (77.3) | 30 (40) | 237 (79) | |
| High | 0 (0) | 1 (1.3) | 17 (22.7) | 45 (60) | 63 (21) | |
| HC (cm) | 92.68 ± 4.52 b, c | 92.46 ± 4.45 d, e | 102.18 ± 6.77 f | 107.60 ± 6.02 | 98.73 ± 8.49 | < 0.001* |
| WHR | 0.84 ± 0.05 b, c | 0.84 ± 0.05 d, e | 0.88 ± 0.06 f | 0.94 ± 0.06 | 0.87 ± 0.72 | < 0.001* |
| WHR categories (N) (%) | < 0.001** | |||||
| Normal | 71 (94.7) | 71 (94.7) | 42 (56) | 39 (52) | 223 (74.3) | |
| High | 4 (5.3) | 4 (5.3) | 33 (44) | 36 (48) | 77 (25.7) | |
| WHtR | 0.47 ± 0.03 b, c | 0.47 ± 0.03 d, e | 0.55 ± 0.05 f | 0.60 ± 0.06 | 0.52 ± 0.07 | < 0.001* |
| WHtR categories (N) (%) | < 0.001** | |||||
| Normal | 72 (96) | 74 (98.7) | 51 (68) | 24 (32) | 221 (73.7) | |
| High | 3 (4) | 1 (1.3) | 24 (32) | 51 (68) | 79 (26.3) | |
| Wrist circumference (cm) | 17.90 ± 2.69 b, c | 18.04 ± 2.78 d, e | 20.46 ± 2.75 f | 23.13 ± 2.30 | 19.88 ± 3.38 | < 0.001* |
| Neck circumference (cm) | 35.13 ± 2.54 b, c | 34.86 ± 2.74 d, e | 39.13 ± 2.55 f | 41.89 ± 2.17 | 37.75 ± 3.85 | < 0.001* |
| Body frame size | 9.49 ± 1.38 b, c | 9.44 ± 1.42 d, e | 8.15 ± 1.06 f | 7.31 ± 0.81 | 8.60 ± 1.50 | < 0.001* |
| Marital status (N) (%) | 0.56** | |||||
| Single | 13 (17.3) | 10 (13.3) | 12 (16) | 11 (14.7) | 46 (15.3) | |
| Married | 51 (68) | 58 (77.3) | 50 (66.7) | 56 (74.7) | 215 (71.7) | |
| Widow | 3 (4) | 5 (6.7) | 7 (9.3) | 5 (6.7) | 20 (6.7) | |
| Divorced | 8 (10.7) | 2 (2.7) | 6 (8) | 3 (4) | 19 (6.3) | |
| Education (N) (%) | < 0.001** | |||||
| Illiterate | 49 (65.3) | 54 (72) | 44 (58.7) | 22 (29.3) | 169 (56.3) | |
| Elementary | 14 (18.7) | 12 (16) | 14 (18.7) | 16 (21.3) | 56 (18.7) | |
| Middle-school | 3 (4) | 5 (6.7) | 4 (5.3) | 10 (13.3) | 22 (7.3) | |
| High-school | 7 (9.3) | 4 (5.3) | 10 (13.3) | 17 (22.7) | 38 (12.7) | |
| Collage | 2 (2.7) | 0 (0) | 3 (4) | 10 (13.3) | 15 (5) | |
| Race (N) (%) | < 0.001** | |||||
| Fras | 17 (22.7) | 22 (29.3) | 16 (21.3) | 4 (9.3) | 62 (20.7) | |
| Lor | 38 (50.7) | 22 (29.3) | 26 (34.7) | 20 (26.7) | 106 (35.3) | |
| Arab | 20 (26.7) | 31 (41.3) | 33 (44) | 48 (64) | 132 (44) | |
| Job (N) (%) | 0.35** | |||||
| Unemployed | 14 (18.7) | 13 (17.3) | 18 (24) | 19 (25.3) | 64 (21.3) | |
| Labor | 29 (38.7) | 21 (28) | 20 (26.7) | 29 (38.7) | 99 (33) | |
| Housekeeper | 23 (30.7) | 26 (34.7) | 24 (32) | 14 (18.7) | 87 (29) | |
| Employee | 9 (12) | 15 (20) | 13 (17.3) | 13 (17.3) | 50 (16.7) | |
| P.A (met-min/week) | 179.96 ± 289.84 b, c | 136.79 ± 249.98 d, e | 27.90 ± 134.58 | 0.00 ± 0.00 | 86.16 ± 215.18 | < 0.001* |
| Duration of dialysis (year) | 19.88 ± 5.61 | 18.90 ± 5.12 | 19.96 ± 5.34 | 19.68 ± 5.77 | 19.60 ± 5.45 | 0.62* |
| FBS (mg/dL) | 103.94 ± 20.42 | 106.48 ± 21.63 | 114.21 ± 55.15 | 111.08 ± 45.52 | 108.93 ± 38.73 | 0.36* |
| TG (mg/dL) | 121.44 ± 44.27 | 127.02 ± 51.06 | 126.44 ± 46.61 | 130.77 ± 46.86 | 126.42 ± 47.14 | 0.68* |
| TC (mg/dL) | 180.48 ± 31.64 b, c | 185.24 ± 28.12 d, e | 203.79 ± 32.42 f | 220.20 ± 31.09 | 197.42 ± 34.53 | < 0.001* |
| LDL-c (mg/dL) | 91.50 ± 23.81 b, c | 95.25 ± 19.28 d, e | 106.49 ± 24.53 f | 121.50 ± 22.16 | 103.69 ± 25.29 | < 0.001* |
| HDL-c (mg/dL) | 68.04 ± 11.36 | 68.78 ± 14.20 | 71.01 ± 10.00 | 67.33 ± 10.16 | 68.79 ± 11.58 | 0.23* |
| LDL-c/HDL.c ratio | 1.39 ± 0.46 b, c | 1.46 ± 0.46 e | 1.54 ± 0.45 f | 1.85 ± 0.44 | 1.56 ± 0.48 | < 0.001* |
| TC/HDL ratio | 2.71 ± 0.59 b, c | 2.78 ± 0.58 e | 2.92 ± 0.59 f | 3.33 ± 0.57 | 2.93 ± 0.63 | < 0.001* |
| AIP | 0.16 ± 0.17 b, c | 0.18 ± 0.16 d, e | 0.24 ± 0.18 f | 0.36 ± 0.14 | 0.24 ± 0.18 | < 0.001* |
| B.SBP (mmHg) | 136.14 ± 8.97 b, c | 136.80 ± 8.60 d, e | 155.40 ± 13.45 f | 170.86 ± 16.71 | 149.80 ± 18.98 | < 0.001* |
| A.SBP (mmHg) | 119.38 ± 7.91 b, c | 120.12 ± 7.75 d, e | 137.58 ± 13.41 f | 151.74 ± 12.39 | 132.21 ± 17.14 | < 0.001* |
| B.DBP (mmHg) | 86.37 ± 3.78 b, c | 86.13 ± 4.41 d, e | 94.33 ± 8.44 f | 102.06 ± 9.83 | 92.22 ± 9.65 | < 0.001* |
| A.DBP (mmHg) | 75.89 ± 5.66 b, c | 76.21 ± 4.97 d, e | 84.46 ± 6.76 f | 90.20 ± 5.93 | 81.69 ± 8.37 | < 0.001* |
| B.PP (mmHg) | 49.77 ± 5.61 b, c | 50.66 ± 4.83 d, e | 61.06 ± 7.29 f | 68.80 ± 9.47 | 57.57 ± 10.53 | < 0.001* |
| A.PP (mmHg) | 43.49 ± 3.56 b, c | 43.91 ± 4.23 d, e | 53.12 ± 7.42 f | 61.54 ± 7.30 | 50.51 ± 9.48 | < 0.001* |
| B.MAP (mmHg) | 102.96 ± 5.42 b, c | 103.02 ± 5.70 d, e | 114.68 ± 9.80 f | 125.00 ± 11.73 | 111.41 ± 12.55 | < 0.001* |
| A.MAP (mmHg) | 90.39 ± 6.27 b, c | 90.85 ± 5.70 d, e | 102.17 ± 8.84 f | 110.71 ± 7.92 | 98.53 ± 11.16 | < 0.001* |
| B.BUN (mg/dL) | 55.94 ± 10.70 b, c | 56.94 ± 10.76 d, e | 78.01 ± 13.41f | 84.800 ± 10.10 | 68.92 ± 17.00 | < 0.001* |
| A.BUN (mg/dL) | 13.08 ± 1.97 b, c | 12.97 ± 1.91 d, e | 16.68 ± 2.64 f | 18.77 ± 2.09 | 15.37 ± 3.28 | < 0.001* |
| Plasma Cr (mg/dL) | 6.88 ± 1.29 b, c | 6.67 ± 1.09 d, e | 8.31 ± 0.92 f | 9.37 ± 1.24 | 7.81 ± 1.58 | < 0.001* |
| Uric acid (mg/dl) | 4.07 ± 0.56 b, c | 4.10 ± 0.68 d, e | 5.87 ± 1.22 f | 7.05 ± 1.20 | 5.27 ± 1.58 | < 0.001* |
| eGFR | 10.09 ± 2.11 | 10.40 ± 1.79 | 9.92 ± 2.09 | 10.83 ± 2.62 | 10.31 ± 2.19 | 0.05* |
| Dialysis adequacy | 76.32 ± 2.45 b, c | 76.93 ± 2.44 d | 78.40 ± 2.39 | 77.71 ± 2.54 | 77.34 ± 2.56 | < 0.001* |
| Na (mEq/L) | 139.34 ± 2.65 b, c | 138.98 ± 2.41 d, e | 143.12 ± 2.01f | 143.92 ± 0.58 | 141.34 ± 3.02 | < 0.001* |
| K (mEq/L) | 4.36 ± 0.33 b, c | 4.36 ± 0.33 d, e | 4.97 ± 0.41f | 5.48 ± 0.49 | 4.79 ± 0.61 | < 0.001* |
| Ca (mg/dL) | 8.05 ± 0.05 b, c | 8.06 ± 0.12 d, e | 8.36 ± 0.31f | 8.80 ± 0.48 | 8.32 ± 0.42 | < 0.001* |
| P (mg/dL) | 4.14 ± 0.40 b, c | 4.13 ± 0.40 d, e | 5.00 ± 0.55 f | 5.59 ± 0.47 | 4.71 ± 0.76 | < 0.001* |
| AST (IU/L) | 16.62 ± 8.45 | 17.32 ± 9.56 | 16.98 ± 7.71 | 18.53 ± 9.79 | 17.36 ± 8.90 | 0.58* |
| ALT (IU/L) | 19.89 ± 7.80 | 20.18 ± 8.99 | 20.62 ± 8.56 | 21.63 ± 9.92 | 20.58 ± 8.83 | 0.64* |
| AST/ALT | 0.83 ± 0.11 | 0.85 ± 0.13 | 0.83 ± 0.12 | 0.85 ± 0.10 | 0.84 ± 0.12 | 0.48* |
Anthropometric measures (BMI, WC, HC, WHR, WHtR) were obtained using post-dialysis dry weight. In hemodialysis patients, these measures may be influenced by fluid status (interdialytic weight gain, ultrafiltration volume) rather than reflecting adiposity alone. Findings related to anthropometric indices should be interpreted with this limitation in mind
Data are means ± SD for quantitative variables and frequency (percent) for qualitative variables
*. From ANOVA for quantitative variables, **. Chi-square for qualitative variables
Post hoc (LSD) according to the following pattern
a Significant difference between 1 compared to 2
b Significant difference between 1 compared to 3
c Significant difference between 1 compared to 4
d Significant difference between 2 compared to 3
e Significant difference between 2 compared to 4
f Significant difference between 3 compared to 4
Central obesity was defined if any one of the following criteria was met:
•Waist circumference (WC) ≥ 102 cm in males or ≥ 88 cm in females (WHO criteria)
•Waist-to-hip ratio (WHR) ≥ 1.0 in males or ≥ 0.8 in females (WHO criteria)
•Waist-to-height ratio (WHtR) ≥ 0.5 in individuals < 40 years or ≥ 0.6 in individuals ≥ 40 years
General obesity was defined as body mass index (BMI) ≥ 30 kg/m² (WHO classification). Participants with BMI < 30 kg/m² were classified as non-obese (including normal weight and overweight categories)
Abbreviation: EDII: energy-adjusted dietary inflammatory index, PA: physical activity, BMI: body mass index, WC: waist circumference, HC: hip circumference, WHtR: waist-to-height ratio, WHR: waist-to-hip ratio, FBS: fasting blood sugar, TG: triglyceride, TC: total cholesterol, LDL-c: low-density lipoprotein cholesterol, HDL-c: high-density lipoprotein cholesterol, AIP: atherogenic index of plasma, B: before dialysis, A: after dialysis, BUN: blood urea nitrogen, Cr: creatinine, eGFR: estimated glomerular filtration rate, Na: sodium, K: potassium, Ca: calcium, P: phosphorus, AST: aspartate aminotransferase, ALT: alanine aminotransferase, ALP: alkaline phosphata, SBP; systolic blood pressure, DBP; diastolic blood pressure, MAP; mean arterial pressure, and PP; pulse pressure
Nutrient and food intake across EDII quartiles
Table 2 details the variations in nutrient and food intake by EDII quartiles. Energy intake significantly increased from 2155.43 kcal in quartile 1 to 2840.98 kcal in quartile 4 (P < 0.001). Carbohydrate intake also rose significantly from 318.91 g to 444.54 g across quartiles (P < 0.001). Total fat intake increased significantly from 70.27 g in quartile 1 to 87.95 g in quartile 4 (P < 0.001), and saturated fats increased from 18.56 g to 26.48 g (P < 0.001). Conversely, trans fats significantly decreased from 14.52 g in quartile 1 to 4.48 g in quartile 4 (P < 0.001). Iron intake increased from 15.92 mg in quartile 1 to 22.79 mg in quartile 4 (P < 0.001). Fiber intake did not show significant differences, with values of 25.47 g in quartile 1 and 23.06 g in quartile 4 (P = 0.14). Protein intake remained consistent across quartiles, ranging from 73.53 g to 80.87 g (P = 0.18) (Table 2).
Table 2.
The mean ± SD of nutrients and food intake across quartiles of EDII in hemodialysis patients
| Characteristics (mean (SD) or %) | EDII quartiles | P – value* | ||||
|---|---|---|---|---|---|---|
| Q1 (N = 75) |
Q2 (N = 75) |
Q3 (N = 75) |
Q4 (N = 75) |
Total (N = 300) |
||
| Energy (kcal/d) | 2155.43 ± 499.82 c | 2125.93 ± 483.78 d, e | 2368.80 ± 936.87 f | 2840.98 ± 940.53 | 2372.79 ± 798.77 | 0 < 0.001* |
| Carbohydrates intake (g) | 318.91 ± 84.74 c | 309.58 ± 84.50 e | 349.63 ± 156.33 f | 444.54 ± 178.67 | 355.67 ± 142.67 | 0 < 0.001* |
| Protein intake (g) | 73.53 ± 20.16 | 74.50 ± 15.54 | 74.68 ± 28.37 | 80.87 ± 24.86 | 75.90 ± 22.83 | 0.18 |
| Total fat intake (g) | 70.27 ± 19.98 b, c | 69.23 ± 20.68 d, e | 79.68 ± 38.16 | 87.95 ± 31.09 | 76.78 ± 29.37 | 0 < 0.001* |
| Saturated fats (g) | 18.56 ± 6.16 c | 17.39 ± 5.31 d, e | 21.74 ± 13.22 f | 26.48 ± 12.33 | 21.04 ± 10.47 | 0 < 0.001* |
| Cholesterol (g) | 254.27 ± 110.44 | 252.03 ± 81.51 | 250.29 ± 104.35 | 280.85 ± 128.66 | 259.36 ± 107.76 | 0.25 |
| Trans fats (g) | 14.52 ± 13.54 c | 16.68 ± 14.45 d, e | 10.64 ± 12.23 f | 4.48 ± 6.86 | 11.68 ± 12.88 | 0 < 0.001* |
| Fiber (g) | 25.47 ± 8.77 | 22.39 ± 6.74 | 22.70 ± 10.94 | 23.06 ± 9.12 | 23.41 ± 9.05 | 0.14 |
| Vitamin B12 (µg) | 2.08 ± 1.21 | 1.92 ± 0.78 | 1.98 ± 0.95 | 2.26 ± 1.02 | 2.06 ± 0.98 | 0.16 |
| Fe (mg) | 15.92 ± 5.22 c | 15.25 ± 5.26 e | 17.84 ± 10.12 f | 22.79 ± 12.41 | 17.95 ± 9.26 | 0 < 0.001* |
| Monounsaturated fats (g) | 23.05 ± 7.70 b, c | 23.02 ± 7.95 d, e | 26.82 ± 13.39 | 29.86 ± 11.81 | 25.69 ± 10.84 | 0 < 0.001* |
| Polyunsaturated fats (g) | 17.81 ± 5.92 | 17.92 ± 6.66 | 19.58 ± 10.59 | 19.37 ± 8.21 | 18.67 ± 8.04 | 0.38 |
| Omega 3 fatty acids (g) | 0.71 ± 0.62 a, c | 0.47 ± 0.45 | 0.56 ± 0.73 | 0.49 ± 0.38 | 0.56 ± 0.57 | 0.03* |
| Omega 6 fatty acids (g) | 15.98 ± 5.25 | 16.47 ± 6.30 | 17.79 ± 9.66 | 17.47 ± 7.88 | 16.93 ± 7.46 | 0.40 |
| Caffeine (mg) | 579.89 ± 442.61 | 555.94 ± 733.44 | 591.89 ± 617.51 | 731.08 ± 1112.782 | 614.70 ± 766.26 | 0.49 |
| Magnesium (mg) | 339.81 ± 89.04c | 317.26 ± 74.48e | 310.98 ± 138.29f | 268.63 ± 91.82 | 309.17 ± 104.02 | 0 < 0.001* |
| Vitamin C (mg) | 225.58 ± 82.10 b, c | 193.89 ± 58.64 e | 188.93 ± 152.07 | 161.40 ± 75.40 | 192.45 ± 100.85 | 0.001* |
| Vitamin A (RE) | 1604.23 ± 709.88 a, b,c | 1187.99 ± 465.23 | 1199.43 ± 772.94 | 1072.04 ± 564.64 | 1265.92 ± 667.70 | 0 < 0.001* |
| Vitamin E (mg) | 14.65 ± 11.11 a, c | 10.09 ± 7.83 | 12.18 ± 13.24 | 10.93 ± 7.25 | 11.96 ± 10.25 | 0.03* |
| Beta carotene | 1807.44 ± 948.16 a, b,c | 1265.50 ± 889.99 | 1302.14 ± 1030.54 | 1049.94 ± 905.07 | 1356.25 ± 980.56 | 0 < 0.001* |
| Folic acid (µg) | 371.27 ± 94.46 a, b,c | 326.52 ± 67.00 | 316.21 ± 114.76 | 301.11 ± 99.72 | 328.78 ± 98.62 | 0 < 0.001* |
| Vit D (µg) | 1.47 ± 0.87 | 1.48 ± 0.73 | 1.63 ± 1.26 | 1.56 ± 1.11 | 1.54 ± 1.01 | 0.76 |
| Niacin (mg) | 18.94 ± 5.12 c | 18.55 ± 5.76 d, e | 21.66 ± 11.27 f | 26.96 ± 11.16 | 21.53 ± 9.39 | 0 < 0.001* |
| Riboflavin (mg) | 1.64 ± 0.45 | 1.56 ± 0.29 | 1.61 ± 0.62 | 1.65 ± 0.52 | 1.62 ± 0.48 | 0.69 |
| Thiamin (mg) | 1.97 ± 0.49 c | 1.92 ± 0.49 e | 2.05 ± 0.77 f | 2.37 ± 0.85 | 2.08 ± 0.69 | 0 < 0.001* |
| Zinc (mg) | 6.38 ± 2.07 | 6.54 ± 1.71 | 6.22 ± 2.26 | 6.17 ± 2.05 | 6.33 ± 2.03 | 0.67 |
| Selenium (µg) | 256.26 ± 389.84 | 184.00 ± 391.54 | 158.13 ± 228.84 | 351.46 ± 877.94 | 237.46 ± 533.78 | 0.11 |
| B6 (mg) | 0.58 ± 0.30 | 0.57 ± 0.23 | 0.54 ± 0.38 | 0.45 ± 0.31 | 0.53 ± 0.31 | 0.05 |
| Garlic (g) | 0.58 ± 0.77 | 0.56 ± 1.38 | 0.56 ± 1.21 | 0.52 ± 1.17 | 0.56 ± 1.15 | 0.99 |
| Onion (g) | 24.20 ± 21.20 | 21.26 ± 22.74 | 26.12 ± 24.83 | 28.49 ± 20.98 | 25.02 ± 22.53 | 0.24 |
| Tea (g) | 5.28 ± 2.36 b, c | 4.91 ± 2.18 e | 4.20 ± 2.42 f | 2.89 ± 2.04 | 4.32 ± 2.42 | 0 < 0.001* |
| Pepper (g) | 6.81 ± 6.21 | 4.43 ± 4.27 | 5.77 ± 6.73 | 7.18 ± 13.64 | 6.05 ± 8.51 | 0.19 |
The data are presented as “mean ± SD”
*The significant difference based on One-way ANOVA (p < 0.05)
Post hoc (LSD) according to the following pattern
a Significant difference between 1 compared to 2
b Significant difference between 1 compared to 3
c Significant difference between 1 compared to 4
d Significant difference between 2 compared to 3
e Significant difference between 2 compared to 4
f Significant difference between 3 compared to 4
Association between EDII and anthropometric and lipid profiles
Table 3 summarizes the associations between EDII and various anthropometric indices and lipid profiles. Significant positive associations were observed with WC (β = 0.31, P < 0.001), Hip Circumference (HC) (β = 0.19, P < 0.001), and neck circumference (β = 0.39, P < 0.001). For instance, WC increased by 0.31 cm for every unit increase in EDII score. Conversely, body frame size showed a significant negative association with EDII (β=-0.16, P < 0.001). No significant associations were found between EDII and FBS (β = 0.38, P = 0.79) or TG (β = 0.76, P = 0.78). However, significant positive associations were found with total cholesterol (β = 2.91, P = 0.03) and LDL-c (β = 2.31, P = 0.02). For example, total cholesterol increased by 2.91 mg/dL per unit increase in EDII score. HDL-c did not show a significant association (β = 0.14, P = 0.82) (Table 3).
EDII and biochemical parameters
Table 4 presents the relationships between EDII scores and biochemical parameters. Higher EDII scores were positively associated with plasma creatinine (β = 0.09, P = 0.01), uric acid (β = 0.20, P < 0.001), and eGFR (β = 0.17, P = 0.009). Plasma creatinine increased by 0.09 mg/dL and uric acid by 0.20 mg/dL for each unit increase in EDII score. Dialysis adequacy also showed a significant positive association (β = 0.50, P < 0.001). Sodium (β = 0.42, P < 0.001) and potassium (β = 0.27, P < 0.001) levels were positively associated with EDII scores, while liver enzymes (AST, ALT, AST/ALT ratio) did not show significant associations (P > 0.05) (Table 4).
Risk of obesity according to EDII score
Table 5 outlines the odds ratios (OR) for the risk of general and central obesity in relation to EDII scores. For general obesity, the odds ratios ranged from 3.31 (Model 1, 2.43–4.50, P < 0.001) to 1.73 (Model 3, 0.05–54.70, P = 0.75). For central obesity, measured by WC, odds ratios ranged from 3.22 (Model 1, 2.41–4.30, P < 0.001) to 4.33 (Model 3, 0.47–39.82, P = 0.19). For WHR, the odds ratios ranged from 1.79 (Model 1, 1.49–2.14, P < 0.001) to 2.63 (Model 3, 1.37–5.03, P = 0.003), and for Waist-to-Height Ratio (WHtR), they ranged from 2.68 (Model 1, 2.11–3.41, P < 0.001) to 1.77 (Model 3, 1.17–2.69, P = 0.006). These results highlight a significant association between higher EDII scores and increased risk of both general and central obesity, with particularly strong associations observed for central obesity metrics (Table 5).
Table 5.
Odds ratio (95% CI) for risk of general and central obesity (dependent variables) according to the EDII score (independent variable) in hemodialysis patients
| Variable | Or (CI) | B | *P- value |
|---|---|---|---|
| General obesity | |||
| Model 1a | 3.31 (2.43–4.50) | 1.19 | 0 < 0.001 |
| Model 2b | 3.71 (2.63–5.23) | 1.31 | 0 < 0.001 |
| Model 3c | 1.73 (0.05–54.70) | 0.55 | 0.75 |
| Central obesity according to WC | |||
| Model 1a | 3.22 (2.41–4.30) | 1.17 | 0 < 0.001 |
| Model 2b | 4.17 (2.90–5.98) | 1.42 | 0 < 0.001 |
| Model 3c | 4.33 (0.47–39.82) | 1.46 | 0.19 |
| Central obesity according to WHR | |||
| Model 1a | 1.79 (1.49–2.14) | 0.58 | 0 < 0.001 |
| Model 2b | 3.76 (2.62–5.40) | 1.32 | 0 < 0.001 |
| Model 3c | 2.63 (1.37–5.03) | 0.96 | 0.003 |
| Central obesity according to WHtR | |||
| Model 1a | 2.68 (2.11–3.41) | 0.98 | 0 < 0.001 |
| Model 2b | 3.02 (2.29–3.99) | 1.10 | 0 < 0.001 |
| Model 3c | 1.77 (1.17–2.69) | 0.57 | 0.006 |
* P < 0.05 statistically significant by Multivariable Logistic Regression. (a) model 1: unadjusted, (b) model 2: adjusted for age and sex, (c) model 3: adjusted for age, sex, physical activity, race, job, marital status, education, duration of dialysis, chronic diseases, and medications
Central obesity was defined if any one of the following criteria was met:
•Waist circumference (WC) ≥ 102 cm in males or ≥ 88 cm in females (WHO criteria)
•Waist-to-hip ratio (WHR) ≥ 1.0 in males or ≥ 0.8 in females (WHO criteria)
•Waist-to-height ratio (WHtR) ≥ 0.5 in individuals < 40 years or ≥ 0.6 in individuals ≥ 40 years
General obesity was defined as body mass index (BMI) ≥ 30 kg/m² (WHO classification). Participants with BMI < 30 kg/m² were classified as non-obese (including normal weight and overweight categories)
Abbreviation: EDII: energy-adjusted dietary inflammatory index, BMI: body mass index, WC: waist circumference, WHtR: waist-to-height ratio, and WHR: waist-to-hip ratio
In Model 3 for obesity outcomes, BMI was not included as a covariate to avoid overadjustment, as BMI may lie on the causal pathway between dietary inflammation and obesity
The extremely wide confidence intervals observed in some models (e.g., Model 3 for general obesity OR 1.73, 95% CI 0.05–54.70; central obesity by WC OR 4.33, 95% CI 0.47–39.82) suggest model instability, possible overfitting, or sparse data bias. These estimates should be interpreted with great caution. The small number of obese participants in certain quartile subgroups (e.g., no obese participants in Q1) may contribute to this instability. Simplified models (e.g., binary E-DII split or reduced covariate sets) produced more stable estimates, as reported in sensitivity analyses (see Supplementary Materials)
Sensitivity analyses for obesity outcomes
Given the wide confidence intervals observed in fully adjusted logistic regression models (Model 3), we conducted sensitivity analyses using simplified approaches to assess the robustness of the association between E-DII and obesity risk. First, we dichotomized E-DII scores (pro-inflammatory vs. anti-inflammatory) using the median value as the cut-off. Second, we reduced the number of covariates in Model 3 to include only age, sex, and physical activity (parsimonious model). Third, we performed penalized maximum likelihood logistic regression (Firth’s method) to reduce small-sample bias. In these sensitivity analyses, the associations between higher E-DII and central obesity (by WHR and WHtR) remained statistically significant, with odds ratios ranging from 1.92 (95% CI 1.24–2.97) to 2.48 (95% CI 1.52–4.03). However, associations with general obesity (by BMI) and central obesity by WC were attenuated and no longer statistically significant in parsimonious models. These sensitivity analyses suggest that the associations with WHR and WHtR are more robust than those with BMI or WC, likely due to the influence of fluid status on weight-based measures. Detailed results of sensitivity analyses are provided in Supplementary Table S1.
Due to the wide confidence intervals observed in fully adjusted models (Table 5), we conducted sensitivity analyses using simplified approaches (E-DII median split, parsimonious covariate sets, and Firth’s penalized likelihood). These sensitivity analyses confirmed robust associations for central obesity by WHR and WHtR but not for general obesity or central obesity by WC. Detailed results are provided in Supplementary Table S1.
Sensitivity and subgroup analyses
The findings from our sensitivity analyses were largely consistent with the primary results. Excluding participants with extreme energy intake did not materially change the direction or magnitude of associations between E-DII and metabolic parameters. When alternative definitions of central obesity were applied (IDF criteria), the associations with WHR and WHtR remained statistically significant, while associations with WC remained non-significant. In subgroup analyses stratified by sex, the associations between E-DII and metabolic parameters were generally similar between males and females, although the small sample size in sex-specific subgroups (130 males, 170 females) limited statistical power for detecting differences. In subgroup analyses stratified by diabetes status (102 diabetic, 198 non-diabetic), the associations between E-DII and FBS were stronger among diabetic patients (β = 0.52, P = 0.04) compared to non-diabetic patients (β = 0.21, P = 0.32), suggesting potential effect modification that warrants further investigation. Detailed results of all sensitivity and subgroup analyses are provided in Supplementary Tables S2 and S3.
Discussion
The present study explored the cross-sectional relationship between the energy-adjusted Dietary Inflammatory Index (E-DII) and various metabolic, anthropometric, and clinical parameters in hemodialysis patients. Our findings demonstrate significant associations between higher E-DII scores and multiple metabolic parameters; however, due to the cross-sectional design, causality cannot be inferred. Below, we discuss the results in relation to existing literature. Our findings are broadly consistent with previous studies that have examined dietary inflammatory potential in various populations; however, direct comparisons with full DII-based studies should be made with caution due to the incomplete E-DII calculation in our study (28 of 45 components).
Glycemic indices
Our study demonstrated a significant increase in fasting blood sugar (FBS) across E-DII quartiles, from 110.5 mg/dL in the lowest quartile to 129.3 mg/dL in the highest quartile (P < 0.001). This aligns with numerous studies that have shown a direct link between pro-inflammatory diets and impaired glycemic control. For example, higher DII scores were associated with elevated FBS and insulin resistance, particularly in individuals with metabolic syndrome [32]. Similarly, diets with high inflammatory potential were significantly associated with increased glucose intolerance among Iranian adults [33]. Chronic inflammation, mediated by cytokines like TNF-α and IL-6, has been shown in experimental and longitudinal studies to interfere with insulin signaling, which may contribute to hyperglycemia [34, 35]. However, our cross-sectional design does not allow us to determine whether the observed associations reflect this proposed biological pathway or other unmeasured factors.
Interestingly, our regression analysis did not show a significant association between E-DII and FBS after adjusting for confounders (β = 0.38, P = 0.79). This contrasts with other studies that reported strong relationships between DII and glucose levels [36]. The lack of significance in our findings could be attributed to specific characteristics of the hemodialysis population, such as long-term dialysis duration or the use of medications like insulin [37]. Adjusting for confounders such as BMI may weaken the observed relationship, as highlighted by other researchers [38]. Oxidative stress induced by pro-inflammatory diets may disrupt glucose homeostasis differently in CKD patients compared to healthier populations [39].
Lipid indices
Significant associations were observed between E-DII and lipid profiles, particularly total cholesterol (TC) and LDL-c levels. TC increased significantly across E-DII quartiles, from 180.48 mg/dL in the first quartile to 220.20 mg/dL in the fourth quartile (P < 0.001), with similar trends for LDL-c [40, 41]. These findings align with previous studies showing that pro-inflammatory diets contribute to dyslipidemia [42, 43]. Higher DII scores were associated with adverse lipid profiles, including increased LDL-c and TC levels [44]. Moreover, a meta-analysis indicated a strong association between dietary inflammation and elevated cholesterol levels [45].
However, we did not find a significant association between E-DII and HDL-c (β = 0.14, P = 0.82), which contrasts with studies reporting either a positive or inverse relationship between dietary inflammation and HDL-c [46, 47]. Higher DII scores correlated with lower HDL-c levels [37]. Differences in dietary components, such as fiber intake, or genetic variations in lipid metabolism could explain these discrepancies [48]. Furthermore, inflammatory cytokines may influence lipid metabolism by increasing oxidative stress, leading to alterations in cholesterol levels [46].
Renal indices
Our study found significant positive cross-sectional associations between higher E-DII scores and plasma creatinine (P = 0.01) and uric acid levels (P < 0.001). These results show that higher dietary inflammatory potential is associated with worse renal parameters in this population; however, longitudinal studies are needed to determine whether this reflects a causal effect of diet on renal function or reverse causation (e.g., more severe renal disease influencing dietary choices). Similar findings have been reported in previous studies, where pro-inflammatory diets were linked to higher creatinine levels and increased risk of CKD [43, 49]. For instance, individuals with higher DII scores had a greater risk of developing CKD, with elevated creatinine levels serving as a marker of renal dysfunction [50].
The lack of significant variation in estimated Glomerular Filtration Rate (eGFR) across E-DII quartiles in our study (P = 0.05) contrasts with studies that report declines in eGFR with higher DII scores [45]. This may be partly explained by the low residual renal function in our hemodialysis population (most patients in ESRD with eGFR < 15 mL/min/1.73 m²), where eGFR estimates using creatinine-based equations like CKD-EPI have reduced accuracy and primarily capture minimal residual GFR rather than substantial changes. This discrepancy may be due to the chronic nature of kidney disease in our study population, where eGFR remains stable due to reliance on dialysis. Hemodialysis patients may not experience the same dietary impacts on eGFR as those with early-stage CKD [51].
Hepatic indices
Contrary to some studies, we did not observe significant associations between E-DII and liver function markers (AST, ALT) (P > 0.05). Previous research, particularly in populations with NAFLD, has shown that higher DII scores are associated with elevated liver enzymes [52]. Higher DII scores correlated with increased ALT levels, indicating liver damage [37]. The lack of significant findings in our study could stem from differences in dietary assessment methods across studies, which may lead to variability in DII calculations [48].
Another potential explanation could be the presence of underlying liver dysfunction in hemodialysis patients, which might dilute the effect of diet-induced inflammation on liver enzymes [46]. Inflammation contributes to hepatocellular injury through mechanisms such as oxidative stress, but these processes may be less pronounced in CKD patients [43].
Blood pressure indices
We found significant associations between E-DII and blood pressure parameters, including systolic blood pressure (SBP) and diastolic blood pressure (DBP) before and after dialysis (P < 0.001). These results are consistent with previous studies linking pro-inflammatory diets to hypertension [49]. Higher DII scores were associated with increased SBP and DBP [50]. Research found similar associations, indicating that dietary patterns with high inflammatory potential contribute to increased cardiovascular risk [45].
Proposed mechanisms linking dietary inflammation to blood pressure regulation include dysregulation of vascular homeostasis. Pro-inflammatory cytokines such as IL-6 and TNF-α have been shown to impair endothelial function, which may lead to increased vascular resistance and elevated blood pressure [51]. Our cross-sectional data are consistent with these proposed pathways but do not provide direct evidence of causation. Additionally, high-sodium diets, common in pro-inflammatory eating patterns, can exacerbate blood pressure by promoting fluid retention and increasing arterial stiffness [52].
Anthropometric indices
Our study also found significant associations between E-DII and various anthropometric measures, including waist circumference (WC), hip circumference (HC), and neck circumference (NC) (P < 0.001 for all). These findings align with studies that link pro-inflammatory diets to central obesity. However, these associations should be interpreted cautiously, as BMI and waist circumference in hemodialysis patients can be influenced by fluid status rather than solely reflecting adiposity. However, several fundamental limitations must be acknowledged when interpreting these associations in the context of hemodialysis. First, post-dialysis ‘dry weight’ and derived measures such as BMI and WC are influenced by extracellular fluid status, interdialytic weight gain (IDWG), and ultrafiltration volume. We did not collect data on IDWG, UF volume, or bioimpedance-based fluid overload indices. Therefore, the observed associations between E-DII and anthropometric indices may reflect diet–fluid–dialysis interactions rather than adiposity per se. Second, malnutrition and protein-energy wasting are common in hemodialysis patients, which can confound the relationship between dietary intake and body size. Consequently, our findings regarding anthropometric indices should be interpreted as associations with post-dialysis body dimensions, not necessarily as measures of obesity or adiposity. Despite these limitations, the consistency of associations across multiple anthropometric measures (WC, HC, NC) and the dose-response pattern across E-DII quartiles suggest that the observed relationships warrant further investigation using more precise body composition methods such as bioimpedance spectroscopy or dual-energy X-ray absorptiometry (DXA). Significant associations between higher DII scores and larger WC among Iranian adults were reported [37]. Similarly, higher DII scores correlated with increased central adiposity in older women [48–54].
Proposed mechanisms through which dietary inflammation may be associated with obesity include chronic low-grade inflammation, which has been shown in other studies to interfere with insulin signaling and promote fat accumulation, particularly in the abdominal region [46]. Our cross-sectional findings are consistent with this hypothesis, but we cannot determine directionality or causation. Elevated levels of pro-inflammatory cytokines such as TNF-α and IL-6 are commonly observed in individuals with higher body fat, indicating a bidirectional relationship between obesity and inflammation [55, 56]. Moreover, alterations in hormones such as leptin and adiponectin in response to inflammation may exacerbate weight gain and fat distribution [56–59]. The mechanistic pathways linking the Energy-Adjusted Dietary Inflammatory Index (EDII) with various pathophysiological outcomes are illustrated in this study. As depicted, a higher EDII contributes to increased pro-inflammatory cytokines (e.g., TNF-α, IL-6) and oxidative stress, which are key drivers of endothelial dysfunction. This dysfunction is associated with elevated blood pressure, adverse lipid profiles, and increased cardiovascular risk. Additionally, EDII exerts adverse effects on glycemic control (e.g., FBS) and hepatic function (e.g., ALT, AST), while also influencing renal function and anthropometric indices. Understanding these pathways highlights the role of dietary inflammation in the progression of metabolic, cardiovascular, and renal complications, especially in hemodialysis patients. Therefore, dietary strategies aimed at reducing EDII may help mitigate these adverse outcomes and improve overall health in this population (Fig. 2).
Fig 2. Conceptual Framework of the Proposed Pathway Linking High Energy-Adjusted Dietary Inflammatory Index (E-DII) to Kidney Dysfunction and CKD Risk in Hemodialysis Patients. Arrows indicate hypothesized relationships that require confirmation in longitudinal and mechanistic studies. This framework is presented to guide future research, not to imply causation from the current cross-sectional data. Proposed pathway linking high Energy-Adjusted Dietary Inflammatory Index (E-DII) to impaired kidney function and CKD risk in hemodialysis patients. A pro-inflammatory diet increases inflammatory cytokines (e.g., TNF-α, IL-6), leading to oxidative stress and endothelial dysfunction. These processes cause reduced renal filtration and nephron damage, resulting in elevated plasma creatinine and increased uric acid levels, thereby heightening CKD risk. In hemodialysis patients, eGFR may remain relatively unchanged due to dialysis, whereas dietary interventions may offer protective effects
We acknowledge several sources of potential residual confounding that may affect the interpretation of our findings. First, we did not have access to detailed dialysis prescription data (e.g., dialysate composition, sodium profiling), which could influence blood pressure and electrolyte outcomes independently of diet. Second, residual renal function was assessed only via creatinine-based eGFR, which has known limitations in hemodialysis patients; more precise measures such as urine output or urea clearance were not available. Third, the underlying cause of ESRD (e.g., diabetic nephropathy, hypertensive nephrosclerosis, glomerulonephritis) was not systematically recorded, despite its potential association with both dietary habits and metabolic parameters. These unmeasured factors may contribute to residual confounding, and our findings should be interpreted as exploratory associations requiring confirmation in studies with more comprehensive covariate ascertainment.
Strengths and limitations
The present study has several methodological strengths. These include the use of a validated 168-item semi-quantitative Food Frequency Questionnaire adapted for Iranian dietary habits [22], standardized anthropometric and biochemical measurements performed under uniform conditions across multiple centers, comprehensive adjustment for potential confounders (including age, sex, physical activity, dialysis duration, and comorbidities) in multivariable models, a relatively large sample size (n = 300) with confirmed statistical power for primary and key secondary outcomes, and systematic assessment of dialysis adequacy using established indices (URR and spKt/V). Additionally, the application of the energy-adjusted Dietary Inflammatory Index (E-DII) provides a more precise estimate of dietary inflammatory potential independent of total energy intake.
Notwithstanding, the study is subject to several inherent limitations. A major limitation of this study is the incomplete construction of the E-DII score. The original DII comprises 45 food parameters, but we were able to calculate only 28 of these components based on the availability of data from our FFQ. Critically, 17 items could not be included, many of which have anti-inflammatory properties, including herbs and spices such as turmeric, ginger, saffron, garlic, pepper, rosemary, thyme, oregano, and others. The exclusion of these components may lead to exposure misclassification, as the true inflammatory potential of participants’ diets may be underestimated or overestimated. Consequently, our findings may not be directly comparable to studies that used the full 45-item DII. We also acknowledge that the use of a past-year FFQ in a hemodialysis population introduces the possibility of recall bias, as patients with advanced disease may have difficulty accurately recalling their long-term dietary habits. Despite the validated nature of the FFQ, future studies in this population should consider shorter recall periods (e.g., 24-hour recalls or food diaries) to minimize recall bias and should aim to include a more comprehensive set of DII components, particularly herbs and spices. Moreover, due to the cross-sectional design of this research endeavor, it precludes the establishment of definitive causal relationships. While all patients in the study exhibited Chronic Kidney Disease (CKD), the underlying etiologies were not systematically documented during data collection, thereby posing as an additional potential limitation of this study. Furthermore, it is imperative to include other prevalent comorbid conditions like diabetes, hypertension, thyroid disorders, in forthcoming data collection processes to enhance the comprehensive scope of future investigations. No significant reverse causation is expected, as the FFQ assessed past-year dietary intake, while laboratory tests were recent. A limitation of the study is the lack of formal correction for multiple comparisons across numerous outcomes and subgroup analyses, which may increase the risk of type I errors. However, the robustness of key findings across different models and the biological plausibility of the associations support their validity. A fundamental limitation of this study concerns the interpretation of anthropometric measures. All anthropometric indices (BMI, WC, HC, WHR, WHtR, neck circumference) were derived from post-dialysis dry weight. In hemodialysis patients, body weight and related measures are strongly influenced by extracellular fluid status, interdialytic weight gain (IDWG), and ultrafiltration (UF) volume. Short-term weight fluctuations in this population are often driven by fluid accumulation rather than changes in adipose or muscle tissue. We did not systematically collect data on IDWG, UF volume, or bioimpedance-based fluid overload assessments. Consequently, the observed associations between E-DII and anthropometric indices may reflect diet–fluid–dialysis interactions rather than associations with adiposity per se. This limits our ability to draw conclusions about obesity or body composition from these data. We have tempered our conclusions accordingly and now frame these findings as associations with post-dialysis body dimensions rather than adiposity. Future studies should incorporate objective fluid status indicators (e.g., bioimpedance spectroscopy, IDWG records) to better distinguish the effects of dietary inflammation on body composition from those related to fluid management. Another limitation concerns residual confounding due to unavailable data on certain clinically relevant variables. We did not have access to dialysis prescription details, including dialysate composition, sodium profiling, or dialysis session length beyond basic Kt/V measurements. Residual renal function was assessed only via crude eGFR (CKD-EPI equation), which has limited accuracy in hemodialysis patients; more precise measures (e.g., urine output, urea clearance) were not available. Additionally, the underlying etiology of ESRD (e.g., diabetic nephropathy, hypertensive nephrosclerosis, glomerulonephritis) was not systematically recorded, despite its potential association with both dietary patterns and metabolic outcomes. The absence of these variables may introduce residual confounding, and unmeasured factors could partly explain the observed associations. We have therefore framed our findings as exploratory and associative rather than causal. Additional methodological limitations concern model stability and multiple comparisons. The logistic regression models for obesity outcomes (Table 5) produced extremely wide confidence intervals (e.g., OR 1.73, 95% CI 0.05–54.70), indicating model instability, possible overfitting, or sparse data bias. This instability likely arises from the small number of participants in certain quartile-by-outcome subgroups (e.g., no obese participants in Q1) combined with the large number of covariates in Model 3 (k = 11 predictors). Although sensitivity analyses using simplified models produced more stable estimates, the wide CIs in fully adjusted models indicate that these specific results should be interpreted with great caution. Furthermore, we examined a large number of outcomes (approximately 30 parameters) across multiple regression models without adjustment for multiple comparisons. As stated in the Methods, we have explicitly labeled our analyses as exploratory; however, readers should be aware that some statistically significant findings may have occurred by chance. Confirmation in independent, adequately powered cohorts is essential before drawing definitive conclusions. Temporal variability between dietary assessment and biochemical measurements is another limitation. Blood samples were collected within a window of 0 to 28 days (median 12 days) after the FFQ administration. While the FFQ assessed habitual dietary intake over the preceding year, biochemical parameters such as FBS, lipid profile, and electrolytes can fluctuate over days to weeks due to factors including interdialytic weight gain, medication adjustments (e.g., insulin, phosphate binders), and dialysis prescription changes. This temporal mismatch may introduce non-differential misclassification, potentially attenuating true associations toward the null. Future studies should aim to collect dietary and biochemical data on the same day or across multiple time points to better capture temporal relationships.
Subgroup analyses should be interpreted with caution due to limited statistical power. While we performed exploratory subgroup analyses stratified by sex and diabetes status, the sample sizes within strata (e.g., 130 males, 170 females; 102 diabetic, 198 non-diabetic) were modest, and the study was not originally powered for these subgroup comparisons. Consequently, some subgroup findings (e.g., potential effect modification by diabetes status for the E-DII-FBS association) may be underpowered or subject to type I error. These results should be considered hypothesis-generating and require confirmation in larger, adequately powered studies.
Conclusion
This study underscores the complex cross-sectional associations between the Energy-Adjusted Dietary Inflammatory Index (E-DII) and a wide range of health parameters in hemodialysis patients. Higher E-DII scores were significantly associated with adverse metabolic outcomes in this sample, including elevated fasting blood sugar, higher total cholesterol and LDL-c, impaired renal function with increased creatinine and uric acid, and larger post-dialysis body dimensions (waist, hip, and neck circumferences). However, because anthropometric measures in hemodialysis patients are influenced by fluid status, these latter associations may reflect diet–fluid–dialysis interactions rather than adiposity per se. These findings emphasize the role of dietary inflammation in worsening metabolic disturbances, particularly in this high-risk population, and are consistent with previous research linking pro-inflammatory diets to poor glycemic control and lipid profiles. Interestingly, no significant associations were found with liver enzymes or HDL-c, suggesting the complexity of inflammatory pathways and the influence of patient-specific factors like demographics and comorbidities. These results suggest that anti-inflammatory dietary strategies could be considered as a potential component of care; however, causal conclusions cannot be drawn from this cross-sectional design, and findings require confirmation in longitudinal and interventional studies. Future studies should also investigate the molecular mechanisms linking dietary inflammation to metabolic disruptions, which could inform more targeted therapeutic approaches. Incorporating dietary assessments into routine care for hemodialysis patients may ultimately contribute to more effective and personalized management of their complex health needs.
Electronic Supplementary Material
Below is the link to the electronic supplementary material.
Acknowledgements
The authors express their gratitude to all the patients who generously took part in this study.
Author contributions
M.Karimi: Conceptualization, Investigation, Writing – original draft; H.Bazyar: Conceptualization, Formal analysis, Writing – original draft; A.Zare Javid: Writing – original draft, Writing – review & editing; Z.Heidari: Writing – review & editing; S.Shayanpour: Writing – review & editing; S.A.Hosseini: Conceptualization, Supervision, Writing – review & editing. All authors read and approved the final manuscript.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Data availability
The authors will provide the raw data underlying the conclusions of this article upon request, without any unnecessary restrictions.
Declarations
Ethics approval and consent to participate
The study was approved by the Ethics Committee of Ahvaz Jundishapur University of Medical Sciences (approval code: IR.AJUMS.REC.1401.483). The study was conducted in accordance with the Declaration of Helsinki and local laws and institutional requirements. All participants provided written informed consent to participate in this study.
Consent for publication
Not applicable, as this manuscript contains no individual person’s data in any form (e.g., individual details, images, or videos). The institutional ethics committee waived the requirement for consent for publication.
Study registration
This study was not registered in a clinical trial registry because it is an observational (cross-sectional) study, not a clinical trial. The study protocol was reviewed and approved by the ethics committee as noted above.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Seyed Ahmad Hosseini, Email: seyedahmadhosseini@yahoo.com.
Hadi Bazyar, Email: hadibazyar2015@gmail.com.
References
- 1.Ansari ZM, Nasiruddin M, Khan RA, Haque SF. Protective role of Nigella sativa in diabetic nephropathy: A randomized clinical trial. Saudi J Kidney Dis Transplantation. 2017;28(1):9. [DOI] [PubMed] [Google Scholar]
- 2.Grams ME, Chow EK, Segev DL, Coresh J. Lifetime incidence of CKD stages 3–5 in the United States. Am J Kidney Dis. 2013;62(2):245–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Murphy D, McCulloch CE, Lin F, Banerjee T, Bragg-Gresham JL, Eberhardt MS, et al. Trends in prevalence of chronic kidney disease in the United States. Ann Intern Med. 2016;165(7):473–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Shahrzad Shahidi E, Hadizadeh P, Shaabani A, Feizi. Epidemiologic Evaluation of Demographic and Basic Clinical Factors in Patients with Chronic Kidney Disease in Isfahan City, Iran, during 1996–2011. J Isfahan Med School. 2017;35(448):1298–308. [Google Scholar]
- 5.Rysz J, Franczyk B, Ciałkowska-Rysz A, Gluba-Brzózka A. The Effect of Diet on the Survival of Patients with Chronic Kidney Disease. Nutrients. 2017;9(5):495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zeng G, Lin J, He Y, Yuan C, Wu Y, Lin Q. Diet-Related Inflammation is Associated with Malnutrition-Inflammation Markers in Maintenance Hemodialysis Patients: Results of a Cross-Sectional Study in China Using Dietary Inflammatory Index. Int J Gen Med. 2022;15:3639–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Shivappa N, Steck SE, Hurley TG, Hussey JR, Hébert JR. Designing and developing a literature-derived, population-based dietary inflammatory index. Public Health Nutr. 2014;17(8):1689–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ruiz-Canela M, Zazpe I, Shivappa N, Hébert JR, Sánchez-Tainta A, Corella D, et al. Dietary inflammatory index and anthropometric measures of obesity in a population sample at high cardiovascular risk from the PREDIMED (PREvencion con DIeta MEDiterranea) trial. Br J Nutr. 2015;113(6):984–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tabung FK, Steck SE, Zhang J, Ma Y, Liese AD, Agalliu I, et al. Construct validation of the dietary inflammatory index among postmenopausal women. Ann Epidemiol. 2015;25(6):398–405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Liu FH, Liu C, Gong TT, Gao S, Sun H, Jiang YT, Zhang JY, Zhang M, Gao C, Li XY, Zhao YH, Wu QJ. Dietary Inflammatory Index and Health Outcomes: An Umbrella Review of Systematic Review and Meta-Analyses of Observational Studies. Front Nutr. 2021;8:647122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Shivappa N, Steck SE, Hurley TG, Hussey JR, Hebert JR. Designing and developing a literature-derived, population-based dietary inflammatory index. Public Health Nutr. 2014;17:1689–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Shivappa N, Steck SE, Hurley TG, Hussey JR, Ma Y, Ockene IS. A population-based dietary inflammatory index predicts levels of C-reactive protein in the seasonal variation of blood cholesterol study (SEASONS). Public Health Nutr. 2014;17:1825–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Neufcourt L, Assmann K, Fezeu L, Touvier M, Graffouillère L, Shivappa N, et al. Prospective association between the dietary inflammatory index and metabolic syndrome: findings from the SU. VI. MAX study. Nutr Metabolism Cardiovasc Dis. 2015;25(11):988–96. [DOI] [PubMed] [Google Scholar]
- 14.Pimenta AM, Toledo E, Rodriguez-Diez MC, Gea A, Lopez-Iracheta R, Shivappa N, et al. Dietary indexes, food patterns and incidence of metabolic syndrome in a Mediterranean cohort: The SUN project. Clin Nutr. 2015;34(3):508–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Abdollahzad H, Pasdar Y, Nachvak SM, Rezaeian S, Saber A, Nazari R. The relationship between the dietary inflammatory index and metabolic syndrome in Ravansar cohort study. Diabetes metabolic syndrome obesity: targets therapy. 2020;13:477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Nikniaz L, Nikniaz Z, Shivappa N, Hébert JR. The association between dietary inflammatory index and metabolic syndrome components in Iranian adults. Prim Care Diabetes. 2018;12(5):467–72. [DOI] [PubMed] [Google Scholar]
- 17.Kim H-Y, Lee J, Kim J. Association between dietary inflammatory index and metabolic syndrome in the general Korean population. Nutrients. 2018;10(5):648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ghorabi S, Esteghamati A, Azam K, Daneshzad E, Sadeghi O, Salari-Moghaddam A, et al. Association between dietary inflammatory index and components of metabolic syndrome. J Cardiovasc Thorac Res. 2020;12(1):27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Puddu PE, et al. Energy-adjusted Dietary Inflammatory Index scores predict long-term cardiovascular disease mortality and other causes of death in an ecological analysis of the Seven Countries Study. Eur J Prev Cardiol. 2021;28(12):1342–50. [DOI] [PubMed] [Google Scholar]
- 20.Laouali N et al. Dietary inflammatory index and type 2 diabetes risk in a prospective cohort of 70,991 women followed for 20 years: the mediating role of BMI. Diabetologia. 2019;62(12):2222–2232. [DOI] [PubMed]
- 21.Huang Y, et al. Energy-adjusted dietary inflammatory index is associated with 5-year all cause and cardiovascular mortality among chronic kidney disease patients. Front Nutr. 2022;9. [DOI] [PMC free article] [PubMed]
- 22.Mirmiran P, Esfahani FH, Mehrabi Y, Hedayati M, Azizi F. Reliability and relative validity of an FFQ for nutrients in the Tehran lipid and glucose study. Public Health Nutr. 2010;13(5):654–62. [DOI] [PubMed] [Google Scholar]
- 23.Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol. 1986;124(1):17–27. 10.1093/oxfordjournals.aje.a114366. [DOI] [PubMed] [Google Scholar]
- 24.Shivappa N, Steck SE, Hurley TG, Hussey JR, Hébert JR. Designing and developing a literature-derived, population-based dietary inflammatory index. Public Health Nutr. 2014;17(8):1689–96. 10.1017/S1368980013002115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hagströmer M, Oja P, Sjöström M. The International Physical Activity Questionnaire (IPAQ): a study of concurrent and construct validity. Public Health Nutr. 2006;9(6):755–62. [DOI] [PubMed] [Google Scholar]
- 26.Hakim P. d. R. M. Assessing the adequacy of dialysis. Kidney Int. 1990;37(2):822–32. [DOI] [PubMed] [Google Scholar]
- 27.Hakim RM, Breyer J, Ismail N, Schulman G. Effects of dose of dialysis on morbidity and mortality. Am J Kidney Dis. 1994;23(5):661–9. [DOI] [PubMed] [Google Scholar]
- 28.Grassmann A, Gioberge S, Moeller S, Brown G. End-stage renal disease-global demographics in 2005 and observed trends. Artif Organs. 2006;30(12):895–7. [DOI] [PubMed] [Google Scholar]
- 29.Somji SS, Ruggajo P, Moledina S. Adequacy of Hemodialysis and Its Associated Factors among Patients Undergoing Chronic Hemodialysis in Dar es Salaam, Tanzania. Int J Nephrol. 2020;2020:9863065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Daugirdas JT. Physiologic principles and urea kinetic modeling. In: Daugirdas JT, Blake PG, Ing TS, editors. in Handbook of Dialysis. 4 ed. New Delhi, India: Wolters Kluwer (India) pvt; 2007. [Google Scholar]
- 31.Arab A, Golpour-Hamedani S, Tabibi H, As’habi A. Association between inflammatory potential of diet and markers of malnutrition in haemodialysis patients. Br J Nutr. 2022:1–7. [DOI] [PubMed]
- 32.Shivappa N, et al. Association between dietary inflammatory index and glycemic control in US adults. J Hum Nutr Dietetics. 2014;27(3):302–10. [Google Scholar]
- 33.Esmaillzadeh A, et al. Dietary patterns, inflammation, and glucose intolerance among Iranian adults. Public Health Nutr. 2016;19(10):1–9.26745231 [Google Scholar]
- 34.Neuhouser ML, et al. Dietary inflammatory index: evidence and implications for diet and health. Nutr Rev. 2010;70(12):681–99. [Google Scholar]
- 35.Casas R, et al. The effects of the Mediterranean diet on biomarkers of vascular wall inflammation. Atherosclerosis. 2011;202:260–6. [Google Scholar]
- 36.Wirth MD, et al. Dietary inflammatory index and inflammatory markers in the US adult population. J Hum Nutr Dietetics. 2014;27(3):302–10. [Google Scholar]
- 37.Kieneker LM, et al. Dietary inflammation and kidney function: The Lifelines cohort study. Am J Kidney Dis. 2017;69(2):239–48. [Google Scholar]
- 38.Bagheri A, et al. Dietary inflammatory index and its association with metabolic syndrome in Iranian adults. Clin Nutr. 2018;38(5):2360–6.30415907 [Google Scholar]
- 39.Tabung FK, et al. The inflammatory potential of diet and all-cause mortality in National Health and Nutrition Examination Survey III. Arch Intern Med. 2012;172(16):1206–14. [Google Scholar]
- 40.Wirth MD, et al. Dietary inflammatory index and cardiovascular disease risk: A meta-analysis. Curr Atheroscler Rep. 2016;17(8):476. [Google Scholar]
- 41.Ruiz-Canela M, et al. The role of dietary inflammatory index in cardiovascular disease. Eur J Nutr. 2015;57(7):2315–28. [Google Scholar]
- 42.Liu S, et al. Dietary inflammatory index and liver function: A population-based study. Liver Int. 2017;37(10):1659–67. [Google Scholar]
- 43.Panti S, et al. Dietary inflammatory index and metabolic syndrome in Iranian elderly women. Clin Nutr. 2019;38(1):236–41. [Google Scholar]
- 44.Chrysohoou C, et al. Dietary inflammatory index and prevalence of hypertension in adults. Eur J Clin Nutr. 2016;68(1):31–7. [Google Scholar]
- 45.Namazi N, et al. The association of dietary inflammatory index with metabolic syndrome components in Iranian adults. Clin Nutr ESPEN. 2018;27:38–43.30144891 [Google Scholar]
- 46.Alferink LJM, et al. Dietary inflammatory index and liver function: A population-based study. Liver Int. 2017;37(10):1659–67. [Google Scholar]
- 47.Wirth MD, et al. Dietary inflammatory index and hypertension risk: A prospective population-based study. Nutrients. 2016;11(9):2160. [Google Scholar]
- 48.Rebholz CM, et al. Dietary inflammatory index and incident chronic kidney disease in the atherosclerosis risk in communities study. Am J Epidemiol. 2016;186(2):238–46. [Google Scholar]
- 49.Bonaccio M, et al. Mediterranean diet, dietary inflammatory index, and their combined effect on cardiovascular disease. Eur J Nutr. 2019;57(7):2315–28. [Google Scholar]
- 50.Esposito K, et al. Dietary inflammatory index and risk of mortality: findings from the aerobics center longitudinal study. Public Health Nutr. 2013;16(10):1687–96. [Google Scholar]
- 51.Liu S, et al. Dietary inflammatory index and risk of liver disease: A population-based study. Hepatology. 2017;66(3):830–7. [Google Scholar]
- 52.Wirth MD, et al. Dietary inflammatory index and cardiovascular disease risk: A meta-analysis. J Clin Lipidol. 2016;12(3):742–57. [Google Scholar]
- 53.Rostami H, Delkhosh M. Energy-adjusted dietary inflammatory index is associated with chronic kidney disease-associated pruritus in Hemodialysis patients: a cross-sectional study. BMC Nephrol. 2025;26:324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Ataei Kachouei A, Kamrani F, Akhavan NS, Haghighatdoost F. Association between dietary inflammatory index and risk of chronic kidney disease and low glomerular filtration rate: a systematic review and meta-analysis of observational studies. J Health Popul Nutr. 2025;44:120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Guo L, Zhao P, Zhu Z. Higher Dietary Inflammatory Index and Systemic Immune-Inflammation Index Score are Associated With Higher Risk of Chronic Kidney Disease. J Ren Nutr. 2025;35(2):300–10. [DOI] [PubMed] [Google Scholar]
- 56.Xu Z, Li L, Jiang L, et al. Association of Dietary Inflammatory Index with CKD progression and estimated glomerular filtration rate in the American CKD population: A cross-sectional study. PLoS ONE. 2024;19(2):e0297916. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Zeng S, Qi L, Sun Y, Zhuang G. Association of Chronic Kidney Disease With Dietary Inflammatory Index in Adults Aged 50 Years and Older: Dose-Response Analysis. J Ren Nutr. 2024;34(3):216–22. [DOI] [PubMed] [Google Scholar]
- 58.Arab A, Karimi E, Nazari M, Tabibi H, As’habi A. Association between the dietary inflammatory index and markers of endothelial and systemic inflammation in hemodialysis patients. Front Nutr. 2023;10:1230747. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Chen Q, Ou L. Meta-analysis of the association between the dietary inflammatory index and risk of chronic kidney disease. Eur J Clin Nutr. 2025;79(1):7–14. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The authors will provide the raw data underlying the conclusions of this article upon request, without any unnecessary restrictions.