Association between dietary inflammatory index and mortality in adults with chronic kidney disease: a retrospective study from the National Health and Nutrition Examination Survey
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
The prevalence and mortality of chronic kidney disease (CKD) are high. Systemic inflammation often occurs in patients with CKD and limits overall survival. The dietary inflammatory index (DII), developed to assess the inflammatory potential of diet, may be an underrecognized risk factor for CKD mortality.
Methods
This study included 7,113 adult participants from the US National Health and Nutrition Examination Survey conducted between 2005 and 2018. Multivariable Cox regression analysis was used to assess the association between DII scores and mortality in patients with CKD, with hazard ratios (HRs) and 95% confidence intervals (CIs) reported. Restricted cubic spline (RCS) models were used to evaluate non-linear associations, whereas subgroup and mediation analyses explored potential effect modifications and mediation effects.
Results
The all-cause mortality rate among patients with CKD was 19.3%. After full adjustment (model 3), the highest DII quartile showed a 23% higher mortality risk (HR, 1.23; 95% CI, 1.07–1.42; Ptrend<0.001) compared with the lowest quartile. RCS confirmed a linear dose-response relationship (Poverall<0.001, Pnon-linear=0.549). No associations were observed for cardiovascular or malignant tumor mortality. The effect of DII scores on all-cause mortality was more pronounced in patients with lower body weight.
Conclusion
Higher DII scores are associated with elevated all-cause mortality in patients with CKD, supporting the potential benefit of anti-inflammatory diets.
Chronic kidney disease (CKD), defined as abnormalities of kidney structure or function that persist for more than 3 months, is a serious, progressive, and global public health issue [1]. CKD is a worldwide health problem, affecting over 10% of the general population, amounting to an estimated 843.6 million people worldwide in 2017 [2], and placing an immense burden on individuals and society [3]. Meanwhile, the mortality rate among patients with CKD is unacceptably high, with an estimated 1.2 million people dying from CKD in 2017, and the global all-age mortality rate from CKD increased by 41.5% between 1990 and 2017 [3]. Tackling the challenges posed by the increasing prevalence and mortality of CKD requires effective prevention strategies to identify risk factors.
While traditional risk factors, such as impaired fasting plasma glucose, high blood pressure, and high body mass index (BMI), account for a large proportion of CKD mortality [3], new risk factors need to be recognized. Persistent low-grade systemic inflammation often occurs in patients with CKD and limits overall survival [4]. Given that specific food ingredients have various anti- or pro-inflammatory properties, diet may contribute to immunological dysregulation and inflammatory activation in patients with CKD [5].
The dietary inflammatory index (DII) is a population-based investigation tool developed to assess the inflammatory potential of diet, with elevated DII scores indicating a notably pro-inflammatory dietary pattern. Numerous studies have linked DII to a series of diseases, such as tumors, cardiovascular disease (CVD), and type 2 diabetes mellitus [6]. Additionally, higher DII scores are associated with increased all-cause mortality in the general population [7]. Studies have also shown that a high DII score, reflecting a pro-inflammatory diet, is associated with declining kidney function and a higher prevalence of CKD [8]. However, it remains unclear whether patients with CKD and higher DII scores are at greater risk of all-cause mortality and cause-specific mortality, particularly from CVD and malignancies, conditions that are more prevalent and associated with higher mortality rates in patients with CKD than in the general population [9].
The National Health and Nutrition Examination Survey (NHANES), conducted by the National Center for Health Statistics (NCHS), is a series of research studies intended to evaluate the health and nutritional status of adults and children in the United States. With demographic, socioeconomic, dietary, and health-related questions asked and multiple biomarkers tested, NHANES has proven to be an effective tool for calculating the DII [10].
Considering the existing evidence, it is plausible that the DII could serve as an underrecognized risk factor for mortality in patients with CKD. Therefore, we aimed to assess the associations between DII scores and various mortality outcomes, including all-cause, cardiovascular, and malignancy-related mortality, in patients with CKD using data from NHANES.
Methods
Ethics approval and consent to participate
This study was a secondary analysis of publicly available, deidentified data from the National Health and Nutrition Examination Survey (NHANES). The NHANES survey protocols were approved by the NCHS Research Ethics Review Board (Protocol #2005-06 and its continuations; Protocol #2011-17 and its continuations; Protocol #2018-01). Written informed consent was obtained from all NHANES participants. No additional ethical approval was required for this secondary analysis. This study was conducted in accordance with the principles of the Helsinki Declaration.
Study population and data collection
We used data from NHANES, a series of periodic cross-sectional surveys that include home visits to administer questionnaires on demographics, diet, and other health behaviors. NHANES employs a complex multistage probability sampling method to ensure diverse geographic representation and adequate racial/ethnic coverage. Trained interviewers collect demographic, socioeconomic, dietary, and health-related data through in-home questionnaires, whereas skilled personnel conduct clinical examinations and dietary assessments at mobile examination centers.
In this study, we used existing data from NHANES collected over seven cycles from 2005 to 2018, including a total of 70,190 participants. Participants without CKD (n=57,410), those under the age of 18 years (n=4,463), those with incomplete dietary data (n=1,108), those with abnormal energy intake (n=90), and those lost to follow-up (n=6) were excluded. Individuals with extreme energy intake (<500 kcal or >5,000 kcal per day for females and <500 kcal or >8,000 kcal per day for males) were also excluded [11]. The final analytic sample included 7,113 participants (Supplement 1).
Definition of CKD
The urinary albumin to creatinine ratio (ACR) and estimated glomerular filtration rate (eGFR) are commonly used to assess kidney function in both clinical and research settings [1,9]. In this study, CKD was defined as an eGFR <60 mL/min/1.73 m2 or an ACR >30 mg/g. The eGFR was calculated using the Modification of Diet in Renal Disease Study equation [12].
Calculation of the DII
The DII is a validated tool that quantifies the inflammatory potential of a diet based on predefined inflammatory effect scores derived from global literature [13]. Although the original DII includes up to 45 food parameters, previous studies have shown that the DII calculated from 27 or 28 parameters also provides good prognostic accuracy [14]. Accordingly, we calculated DII scores using 27 nutrients available from the NHANES 24‑hour dietary recall (first day): alcohol, β-carotene, caffeine, carbohydrate, cholesterol, energy, total fat, fiber, folic acid, iron, magnesium, monounsaturated fatty acids, polyunsaturated fatty acids, n-3 fatty acids, n-6 fatty acids, protein, saturated fat, selenium, zinc, vitamin A, B1, B2, B3, B6, B12, C, and E. For each of these parameters, we applied the inflammatory effect score exactly as defined in the original DII algorithm [13]. The detailed calculation process for the DII is shown in Supplement 2.
Outcomes and follow-up
The main outcome of this study was all-cause mortality, extracted from the National Death Index records. Malignant tumor and CVD mortality were the secondary outcomes. For malignant tumor mortality, we included deaths from malignant tumors coded from C00–C97 in the International Classification of Diseases, 10th Edition, Clinical Modification System codes (ICD-10). For CVD mortality, we used the following ICD-10 codes: I00–I09, I11, I13, and I20–I51. Follow-up time was defined as the duration from the interview to either death or censoring.
Assessment of potential confounders
We considered the following characteristics as potential covariates. Demographic factors included age, sex, race, and income. Race was divided into five groups: Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, and other races. Income was evaluated using the poverty income ratio (PIR): not poor, PIR ≥3; near poor, PIR 1 to <3; and poor, PIR <1. Behavioral variables comprised smoking and drinking. Individuals who had consumed more than 0 g of alcohol in the last 12 months were defined as alcohol consumers, and those who had smoked more than 100 cigarettes in their lifetime were defined as smokers. Comorbidities included hypertension, diabetes, CVD, and malignant tumors. Hypertension was defined as a self-reported physician diagnosis of hypertension, use of antihypertensive drugs, or systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg (at least 3 times) [15]. Diabetes was defined as a self-reported physician diagnosis of diabetes, glycated hemoglobin A1c ≥6.5%, fasting glucose ≥7.0 mmol/L, or use of insulin or oral hypoglycemic medications [16]. CVD and malignant tumors were classified using the question “Have you ever been informed by a health professional or a doctor that you had CVD/malignant tumors?” BMI was calculated using the formula: body weight (kg)/height2 (m2). Laboratory indicators included eGFR, ACR, hemoglobin, white blood cell count, platelet count, albumin, calcium, phosphorus, and uric acid. ACR was divided into three groups for statistical analysis (A1, <30 mg/g; A2, 30–300 mg/g; and A3, >300 mg/g).
Statistical analysis
Characteristics of participants were descriptively analyzed for all individuals and within four DII subgroups divided by DII quartiles. The normality of the distribution was tested using the Kolmogorov-Smirnov test. Data were described as means and standard deviations (SDs) for normally distributed continuous variables, as medians and interquartile ranges (IQRs) for non-normally distributed continuous variables, and as percentages or frequencies for categorical variables. Group comparisons were conducted using the t-test, χ2 test, analysis of variance, or the Mann-Whitney U test, as appropriate. Missing data were imputed using the multiple imputation by chained equations method.
We used the Cox proportional hazards model to estimate the association between the DII scores and mortality in patients with CKD. We built three models after testing the proportional hazards assumption based on Schoenfeld residuals. Model 1 was unadjusted. Model 2 was adjusted for age, sex, race, and income. Model 3 was further adjusted for BMI, smoking, hypertension, diabetes, CVD, malignant tumors, eGFR, and ACR. Restricted cubic spline (RCS) analysis was further performed to assess the potential non-linear association between continuous DII scores and the risk of all-cause mortality, using three knots placed at the 10th, 50th, and 90th percentiles of the DII distribution. Subgroup analyses were conducted to examine the association between DII scores and mortality in patients with CKD stratified by sex, age, BMI, eGFR, ACR, hypertension, diabetes mellitus, CVD, and malignant tumors. We performed mediation analysis to explore the potential mediators (hypertension, diabetes, CVD, and malignant tumors) of the association between DII scores and all-cause mortality in patients with CKD stratified by BMI groups using the R package “mediation,” which implements quasi-Bayesian Monte Carlo simulations with 1,000 iterations.
We performed three sensitivity analyses. First, we repeated the analyses in samples without multiple imputations. Second, we applied the Fine and Gray competing risks model to account for competing mortality risks. Third, to address potential selection bias related to baseline disease history, we stratified the analysis by pre-existing CVD or malignant tumors for both all‑cause and cause‑specific mortality.
Survey sampling weights were considered throughout all analyses. All statistical analyses were conducted using IBM SPSS ver. 25.0 (IBM Corp.) and R software ver. 4.1.2 (R Foundation), and P<0.05 was considered statistically significant.
Results
Of the 7,113 NHANES participants, 2,845 (40.0%) were male, and the mean age was 57.00±18.22 years. Among these, 1,376 participants (19.3%) died during a median follow-up of 74 months (IQR, 41–119 months), of whom 387 (28.1%) died of CVD and 280 (20.3%) died of malignant tumors. The baseline characteristics and comparisons of participants are presented in Table 1. The overall DII score ranged from –5.34 to 4.82. Based on the DII score range, we divided participants into quartiles: quartile 1 (Q1: DII, –5.34 to <–1.23; n=1,778), quartile 2 (Q2: DII, –1.23 to <–0.01; n=1,778), quartile 3 (Q3: DII, –0.01 to <1.23; n=1,778), and quartile 4 (Q4: DII, 1.23 to 4.82; n=1,779). In this case, participants in Q1 consumed relatively more anti-inflammatory diets, whereas those in Q4 consumed relatively more pro-inflammatory diets. Participants with higher DII scores (pro-inflammatory diet) were significantly younger, more likely to be male, non-Hispanic Black, and had higher income, smoking, and drinking rates but a lower prevalence of comorbidities, including hypertension, diabetes, CVD, and malignant tumors. A higher DII score was also associated with a higher eGFR and lower ACR. No significant differences were observed in BMI, hemoglobin, albumin, uric acid, or mortality status across the groups (all P>0.05).
Association between DII scores and mortality
The association between DII scores and mortality is shown in Table 2. In model 1, a higher DII score was associated with decreased all-cause, CVD, and malignant tumor mortality (all Ptrend<0.05). However, after adjusting for potential confounders (in models 2 and 3), we observed an association between increased DII scores and all-cause mortality (both Ptrend<0.001). Comparing Q4 with Q1, the adjusted hazard ratios (aHRs) were 1.21 (95% confidence interval [CI], 1.05–1.39) in model 2, and 1.23 (95% CI, 1.07–1.42) in model 3. We further examined the shape of the exposure-response curve for the relationship between continuous DII scores and the risk of all-cause mortality (Figure 1). Although the exposure-response curve increased monotonically, a plateau was evident at the highest DII scores (Poverall<0.001, Pnon-linear=0.549). No statistically significant associations were observed between DII scores and either CVD or malignant tumor mortality in models 2 and 3. The major results remained stable across sensitivity analyses, including the sample analysis without multiple imputations (Supplement 3) and Fine and Gray competing risks models accounting for competing CVD and malignant tumor mortality (Supplement 4). Stratified sensitivity analyses showed that the lack of association between DII scores and CVD or malignant tumor mortality remained consistent across subgroups with or without baseline disease history after full adjustment for covariates (Supplement 5). When stratified by baseline disease history, the association between DII scores and all-cause mortality was significant in all subgroups except those with preexisting CVD (Supplement 6).
Subgroup analysis
The association between DII scores and all-cause mortality in patients with CKD stratified by different factors is shown in Figure 2. The subgroup analysis revealed that the association between DII scores and all-cause mortality was more pronounced in patients with lower BMI (<30 kg/m2) (aHR, 1.112; 95% CI, 1.064–1.163), compared with those with higher BMI (≥30 kg/m2) (aHR, 1.034; 95% CI, 0.988–1.083) (Pinteraction=0.018). In addition, a marginal interaction was observed between malignant tumor prevalence and DII scores on CVD mortality (yes vs. no: aHR, 0.915; 95% CI, 0.798–1.050 vs. aHR, 1.070; 95% CI, 0.999–1.146; Pinteraction=0.046), whereas no significant interactions were found for other covariates (all Pinteraction>0.05).
Mediation analysis stratified by different BMI groups
Considering the interactive effect of the BMI group on DII and all-cause mortality, a stratified mediation analysis was conducted between DII scores and all-cause mortality according to BMI (<30 kg/m2, ≥30 kg/m2). The result showed that the prevalence of diabetes had a partial mediating effect on the relationship between DII scores and all-cause mortality, with an effect size of –3.82%. Meanwhile, this mediating effect was more obvious in patients with CKD and diabetes and a BMI <30 kg/m2 (effect size, –4.44%). The mediating effect size may have been underestimated due to masking effects. In contrast, hypertension, CVD, and malignant tumors did not exhibit mediating effects (Table 3).
Discussion
In this study of 7,113 patients with CKD based on NHANES (2005–2018), DII score was significantly associated with all-cause mortality but not with cause-specific mortality (CVD and malignant tumor mortality). In the fully adjusted model, participants in the fourth DII quartile were 23% more likely to die from all causes than those in the first quartile. The exposure-response curve showed a positive relationship between continuous DII score and the risk of all-cause mortality. The effect of the DII score on all-cause mortality was more pronounced in patients with lower BMI. Additionally, diabetes showed a small negative mediating effect on the association between DII scores and all-cause mortality.
Participants with higher DII scores had a higher mortality risk, implying that pro-inflammatory dietary patterns may increase all-cause mortality in patients with CKD. Combined with the RCS plot, these findings provide evidence that a pro-inflammatory diet is associated with an increased risk of mortality in patients with CKD. Consistent with our findings, Huang et al. [17] previously reported a significant positive association between EDII (DII/total energy intake) and 5-year all-cause mortality in patients with CKD. Although we did not directly measure inflammatory markers in our study, the potential association may be explained by several mechanisms reported in the literature. The DII score is significantly associated with low-grade inflammation [7], which constrains the overall survival rates among patients with CKD [4]. Additionally, elevated inflammatory biomarkers, such as interleukin-6 (IL-6) and C-reactive protein (CRP), which are used to calculate the DII, are independent risk factors for mortality in CKD [18]. The results of this study are consistent with previous evidence indicating that a pro-inflammatory diet is associated with mortality [7]. Additionally, we observed a reversal of the association between DII scores and all-cause mortality after adjusting for covariates, indicating the presence of strong negative confounding, primarily by age and comorbidities. Participants with higher DII scores were younger and had fewer comorbidities, characteristics associated with lower mortality risk. These baseline differences obscured the true effect of a proinflammatory diet in the unadjusted models. After adjusting for age and comorbidities, the detrimental effect on all-cause mortality was observed.
However, associations between DII scores and CVD or malignant tumor mortality were not observed in our study. A previous NHANES-based study by Huang et al. [17] reported a significant positive association between DII scores and CVD mortality in patients with CKD, a finding not replicated here, likely due to differences in study periods, population characteristics, exposure classification, outcome windows, and covariate adjustment. However, other studies have reported mixed results. For example, a pro-inflammatory diet has been linked to an increased risk of developing and dying from cancer [19], and individuals with diabetes with higher DII scores are at a higher risk of CVD mortality [20]. However, a study of middle-aged and older Japanese adults found that DII score was associated with all-cause and CVD mortality but not with cancer mortality [21], whereas another study conducted in Italy reported no correlation between DII scores and either CVD or cancer mortality [7]. Collectively, these conflicting findings suggest that the association between DII scores and cause-specific mortality may be population-specific. For instance, the promoting effect of DII scores on CVD mortality has been observed in individuals with diabetes but not in those without diabetes [20], and Park et al. [22] found that the DII score was associated with increased all-cause and CVD mortality only in individuals with metabolically unhealthy obesity. Differences in dietary practices, baseline disease status, and study populations across studies may have contributed to these discrepancies. Future research should target specific, representative populations using standardized methodologies.
Our finding that the association between DII scores and all-cause mortality was stronger in patients with lower BMI may be attributed to several factors. First, individuals with obesity typically have chronically elevated levels of inflammatory markers, such as CRP, IL-6, and tumor necrosis factor-α [23], which may diminish the additional impact of a pro-inflammatory diet. In contrast, patients with lower BMI may have poorer nutritional status and lower baseline inflammatory activity, potentially making them more susceptible to the detrimental effects of a proinflammatory diet. Second, the so-called obesity paradox has been observed in patients with CKD [24,25], where a higher BMI is paradoxically associated with better survival, potentially masking the detrimental effects of a pro-inflammatory diet. Taken together, these mechanisms suggest that the association between DII scores and all-cause mortality may be stronger in patients with CKD and lower BMI. We also observed a marginal interaction between malignant tumor history and the association of DII scores with CVD mortality, although the stratified estimates did not reach statistical significance. Given the exploratory nature of this finding, it should be interpreted with caution and requires further validation in future studies.
Our observation showed that the relationship between DII scores and all-cause mortality was attenuated by diabetes, with a negative mediation effect of –3.82%. Previous studies have shown that pro-inflammatory diets are significantly associated with a higher risk of type 2 diabetes mellitus [26]. However, this negative and small indirect effect should not be interpreted as evidence of clinically meaningful mediation but rather as a hypothesis-generating observation. In this study, diabetes did not explain the detrimental association between DII scores and survival; instead, the observed pattern may reflect the high baseline mortality risk associated with diabetes and CKD, residual confounding, or model uncertainty. Additionally, our study revealed differences in the mediation effect across BMI groups; however, the absolute magnitude of the effect was small (–3.82% to –4.44%) in both groups. Given the small effect size, this subgroup difference should be interpreted with caution and may be attributable to chance or residual confounding rather than reflecting a true biological effect.
This population-based study has several methodological strengths, including a large sample size, detailed dietary information, a long median follow-up of 74 months, and the examination of multiple mortality outcomes. Furthermore, we adjusted for key renal confounders (eGFR and ACR) and conducted a series of sensitivity analyses to test the robustness of our findings. However, our study has some limitations. First, the use of self-reported dietary information may have led to exposure misclassification. Second, as a retrospective cohort study, our ability to establish a causal relationship between DII scores and mortality is limited, and recall bias and missing data were expected. In particular, the exploratory mediation analysis should be interpreted cautiously because the DII score and diabetes were assessed at the same time point, precluding confirmation of the temporal sequence required for causal mediation inference. In addition, diabetes was treated both as a covariate in the main Cox model and as a potential mediator, and the absolute indirect effect was clinically negligible; therefore, these findings should be considered hypothesis-generating rather than evidence of meaningful causal mediation. Third, the dietary assessment was based on a single 24-hour dietary recall at baseline, and participants’ dietary habits may have changed during follow-up. However, previous research has indicated that dietary patterns remain relatively stable over extended periods throughout adulthood [27]. Additionally, CKD was diagnosed based on a single assessment of eGFR and ACR, without confirmation of chronicity according to clinical guidelines, which may have introduced some misclassification. However, such misclassification is likely non‑differential with respect to DII score, which would tend to bias the observed associations toward the null. Finally, our study did not directly examine the relationship between DII scores and inflammatory markers, which limits our ability to confirm the hypothesized biological pathway.
In patients with CKD, a pro-inflammatory dietary pattern was associated with a significantly higher risk of all-cause mortality than an anti-inflammatory dietary pattern. These findings strongly support the adoption of anti-inflammatory dietary interventions as a potential strategy to improve clinical outcomes in patients with CKD.
Notes
Conflict of interest
No potential conflict of interest relevant to this article was reported.
Acknowledgments
The authors are grateful for the time and effort put in by participants during the NHANES data collection period. During the preparation of this work, the authors used ChatGPT to improve language. After using this tool, the authors reviewed and edited the content as needed and took full responsibility for the content of the publication.
Funding
This research work was conducted with financial support from the Kidney Disease Center, the First Affiliated Hospital, Zhejiang University School of Medicine.
Association between DII score and cardiovascular or malignant tumor mortality estimated using the Fine and Gray competing risks model (sensitivity analysis).
Association between DII score and cause-specific mortality stratified by baseline history of cardiovascular disease or malignant tumors (sensitivity analysis).
Association between DII score and all-cause mortality stratified by baseline history of cardiovascular disease or malignant tumors (sensitivity analysis).
Adjusted hazard ratio (HR) of all-cause mortality in patients with chronic kidney disease (CKD) according to dietary inflammatory index (DII) score. The graph shows HRs for patients with CKD, adjusted for age, sex, race, income, body mass index group, smoking, hypertension, diabetes, cardiovascular disease, malignant tumors, estimated glomerular filtration rate, and albumin to creatinine ratio. Data were fitted using a Cox regression model with restricted cubic splines. Red solid lines indicate HRs, and purple-filled areas indicate 95% confidence intervals.
Figure. 2.
Association between dietary inflammatory index (DII) score and mortality in patients with chronic kidney disease (CKD) stratified by different factors. The forest plots show adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs) for mortality—(A) all-cause mortality, (B) ardiovascular mortality, and (C) malignant tumor mortality—in patients with CKD stratified by sex, age, body mass index (BMI), estimated glomerular filtration rate (eGFR), albumin to creatinine ratio (ACR), hypertension, diabetes, cardiovascular disease, and malignant tumor. The models were adjusted for age, sex, race, income, BMI group, smoking, hypertension, diabetes, cardiovascular disease, malignant tumors, eGFR and ACR.
Table 1.
Characteristics of the study participants according to DII
c)Adjusted for age, sex, race, income, body mass index group, smoking, hypertension, diabetes, cardiovascular disease, malignant tumors, estimated glomerular filtration rate, and albumin to creatinine ratio.
Table 3.
Potential mediators of the association between DII score and all-cause mortality in patients with chronic kidney disease
Mediator
Subgroup
Effect size (95% CI)
Total effect of DII on all-cause mortality
Indirect effect of DII on all-cause mortality
% of effect via mediator
Hypertension
Total group
0.0090 (0.0050 to 0.0130)
0.0095 (–0.0055 to 0.0245)
NA
BMI <30 kg/m2
0.0179 (0.0110 to 0.0248)
0.0000 (–0.0002 to 0.0002)
NA
BMI ≥30 kg/m2
0.0008 (–0.0072 to 0.0088)
–0.0002 (–0.0006 to 0.0002)
NA
Diabetes
Total group
0.0131 (0.0030 to 0.0232)
–0.0005 (–0.0009 to –0.0001)
–3.82
BMI <30 kg/m2
0.0180 (0.0089 to 0.0271)
–0.0008 (–0.0015 to –0.0001)
–4.44
BMI ≥30 kg/m2
0.0002 (–0.0062 to 0.0066)
–0.0004 (–0.0009 to 0.0001)
NA
Cardiovascular disease
Total group
0.0103 (0.0050 to 0.0156)
0.0001 (0.0000 to 0.0002)
NA
BMI <30 kg/m2
0.0179 (0.0093 to 0.0265)
0.0002 (–0.0004 to 0.0008)
NA
BMI ≥30 kg/m2
0.0009 (–0.0940 to 0.0958)
–0.0001 (–0.0058 to 0.0056)
NA
Malignant tumor
Total group
0.0095 (0.0046 to 0.0144)
0.0001 (–0.0001 to 0.0003)
NA
BMI <30 kg/m2
0.0180 (0.0123 to 0.0237)
0.0001 (–0.0002 to 0.0004)
NA
BMI ≥30 kg/m2
0.0004 (–0.0085 to 0.0093)
0.0000 (–0.0003 to 0.0003)
NA
The above models are adjusted for age, sex, race, income, body mass index group, and smoking.
DII, dietary inflammatory index; CI, confidence interval; BMI, body mass index; NA, not applicable.
References
1. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2024 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int 2024;105(4S):S117-314.
3. GBD Chronic Kidney Disease Collaboration. Global, regional, and national burden of chronic kidney disease, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2020;395:709-33.
4. Mihai S, Codrici E, Popescu ID, Enciu AM, Albulescu L, Necula LG, et al. Inflammation-related mechanisms in chronic kidney disease prediction, progression, and outcome. J Immunol Res 2018;2018:2180373.
6. Phillips CM, Chen LW, Heude B, Bernard JY, Harvey NC, Duijts L, et al. Dietary inflammatory index and non-communicable disease risk: a narrative review. Nutrients 2019;11:1873.
7. Veronese N, Cisternino AM, Shivappa N, Hebert JR, Notarnicola M, Reddavide R, et al. Dietary inflammatory index and mortality: a cohort longitudinal study in a Mediterranean area. J Hum Nutr Diet 2020;33:138-46.
8. Moludi J, Fateh HL, Pasdar Y, Moradinazar M, Sheikhi L, Saber A, et al. Association of dietary inflammatory index with chronic kidney disease and kidney stones in Iranian adults: a cross-sectional study within the Ravansar non-communicable diseases cohort. Front Nutr 2022;9:955562.
10. Ahluwalia N, Dwyer J, Terry A, Moshfegh A, Johnson C. Update on NHANES dietary data: focus on collection, release, analytical considerations, and uses to inform public policy. Adv Nutr 2016;7:121-34.
11. Hao J, Zhou P, Qiu H. Association between ultra-processed food consumption and frailty in American elder people: evidence from a cross-sectional study. J Nutr Health Aging 2022;26:688-97.
12. Gera M, Slezak JM, Rule AD, Larson TS, Stegall MD, Cosio FG. Assessment of changes in kidney allograft function using creatinine-based estimates of glomerular filtration rate. Am J Transplant 2007;7:880-7.
13. 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.
14. Shivappa N, Steck SE, Hurley TG, Hussey JR, Ma Y, Ockene IS, et al. 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.
15. Unger T, Borghi C, Charchar F, Khan NA, Poulter NR, Prabhakaran D, et al. 2020 International Society of Hypertension global hypertension practice guidelines. J Hypertens 2020;38:982-1004.
17. Huang Y, Zhang L, Zeng M, Liu F, Sun L, Liu 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;9:899004.
18. Munoz Mendoza J, Isakova T, Cai X, Bayes LY, Faul C, Scialla JJ, et al. Inflammation and elevated levels of fibroblast growth factor 23 are independent risk factors for death in chronic kidney disease. Kidney Int 2017;91:711-9.
19. Nagle CM, Ibiebele T, Shivappa N, Hebert JR, DeFazio A, Webb PM, et al. The association between the inflammatory potential of diet and risk of developing, and survival following, a diagnosis of ovarian cancer. Eur J Nutr 2019;58:1747-56.
20. Yuan S, Song C, Zhang R, He J, Dou K. Dietary inflammation index and its association with long-term all-cause and cardiovascular mortality in the general US population by baseline glycemic status. Nutrients 2022;14:2556.
21. Okada E, Shirakawa T, Shivappa N, Wakai K, Suzuki K, Date C, et al. Dietary inflammatory index is associated with risk of all-cause and cardiovascular disease mortality but not with cancer mortality in middle-aged and older Japanese adults. J Nutr 2019;149:1451-9.
22. Park YM, Choi MK, Lee SS, Shivappa N, Han K, Steck SE, et al. Dietary inflammatory potential and risk of mortality in metabolically healthy and unhealthy phenotypes among overweight and obese adults. Clin Nutr 2019;38:682-8.
23. Kord Varkaneh H, Fatahi S, Tajik S, Rahmani J, Zarezadeh M, Shab-Bidar S. Dietary inflammatory index in relation to obesity and body mass index: a meta-analysis. Nutr Food Sci 2018;48:702-21.
24. Lee DH, Keum N, Hu FB, Orav EJ, Rimm EB, Willett WC, et al. Predicted lean body mass, fat mass, and all cause and cause specific mortality in men: prospective US cohort study. BMJ 2018;362:k2575.
25. Lu JL, Kalantar-Zadeh K, Ma JZ, Quarles LD, Kovesdy CP. Association of body mass index with outcomes in patients with CKD. J Am Soc Nephrol 2014;25:2088-96.
26. Namazi N, Anjom-Shoae J, Najafi F, Ayati MH, Darbandi M, Pasdar Y. Pro-inflammatory diet, cardio-metabolic risk factors and risk of type 2 diabetes: a cross-sectional analysis using data from RaNCD cohort study. BMC Cardiovasc Disord 2023;23:5.
27. Sijtsma FP, Meyer KA, Steffen LM, Shikany JM, Van Horn L, Harnack L, et al. Longitudinal trends in diet and effects of sex, race, and education on dietary quality score change: the Coronary Artery Risk Development in Young Adults study. Am J Clin Nutr 2012;95:580-6.
Association between dietary inflammatory index and mortality in adults with chronic kidney disease: a retrospective study from the National Health and Nutrition Examination Survey
Figure. 1. Adjusted hazard ratio (HR) of all-cause mortality in patients with chronic kidney disease (CKD) according to dietary inflammatory index (DII) score. The graph shows HRs for patients with CKD, adjusted for age, sex, race, income, body mass index group, smoking, hypertension, diabetes, cardiovascular disease, malignant tumors, estimated glomerular filtration rate, and albumin to creatinine ratio. Data were fitted using a Cox regression model with restricted cubic splines. Red solid lines indicate HRs, and purple-filled areas indicate 95% confidence intervals.
Figure. 2. Association between dietary inflammatory index (DII) score and mortality in patients with chronic kidney disease (CKD) stratified by different factors. The forest plots show adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs) for mortality—(A) all-cause mortality, (B) ardiovascular mortality, and (C) malignant tumor mortality—in patients with CKD stratified by sex, age, body mass index (BMI), estimated glomerular filtration rate (eGFR), albumin to creatinine ratio (ACR), hypertension, diabetes, cardiovascular disease, and malignant tumor. The models were adjusted for age, sex, race, income, BMI group, smoking, hypertension, diabetes, cardiovascular disease, malignant tumors, eGFR and ACR.
Graphical abstract
Figure. 1.
Figure. 2.
Graphical abstract
Association between dietary inflammatory index and mortality in adults with chronic kidney disease: a retrospective study from the National Health and Nutrition Examination Survey
Characteristic
Overall (n=7,113)
DII
P-value
Q1 (n=1,778)
Q2 (n=1,778)
Q3 (n=1,778)
Q4 (n=1,779)
DII (min–max)
–5.34 to 4.82
–5.34 to –1.23
–1.23 to –0.01
–0.01 to 1.23
1.23 to 4.82
-
Male sex
2,845 (40.0)
617 (34.7)
731 (41.1)
695 (39.1)
815 (45.8)
<0.001**
Age (y)
57.00±18.22
61.02±17.63
59.21±17.35
55.74±18.01
51.44±18.48
<0.001**
Race
<0.001**
Mexican American
498 (7.0)
89 (5.0)
153 (8.6)
146 (8.2)
110 (6.2)
Other Hispanic
313 (4.4)
78 (4.4)
75 (4.2)
80 (4.5)
80 (4.5)
Non-Hispanic White
4,922 (69.2)
1,304 (73.3)
1,243 (69.9)
1,211 (68.1)
1,164 (65.4)
Non-Hispanic Black
967 (13.6)
169 (9.5)
205 (11.5)
254 (14.3)
339 (19.1)
Other race
413 (5.8)
138 (7.8)
102 (5.7)
87 (4.9)
86 (4.8)
Income
<0.001**
Not poor
1,131 (15.9)
196 (11.0)
307 (17.3)
286 (16.1)
354 (19.9)
Near poor
3,037 (42.7)
713 (40.1)
730 (41.1)
771 (43.4)
831 (46.7)
Poor
2,945 (41.4)
869 (48.9)
740 (41.6)
721 (40.6)
594 (33.4)
BMI (kg/m2)
0.134
<25.0
1,981 (27.9)
524 (29.5)
509 (28.6)
448 (25.2)
500 (28.1)
25.0–27.9
1,216 (17.1)
348 (19.6)
292 (16.4)
316 (17.8)
260 (14.6)
≥28.0
3,916 (55.0)
906 (51.0)
977 (54.9)
1,014 (57.0)
1019 (57.3)
Smoking
3,400 (47.8)
708 (39.8)
855 (48.1)
851 (47.9)
986 (55.4)
<0.001**
Drinking
4,396 (61.8)
1,010 (56.8)
1,098 (61.8)
1,180 (66.4)
1,108 (62.3)
0.003*
Hypertension
3,862 (54.3)
1,015 (57.1)
1,040 (58.5)
927 (52.1)
880 (49.5)
<0.001**
Diabetes
1,707 (24.0)
422 (23.7)
531 (29.9)
386 (21.7)
368 (20.7)
<0.001**
Cardiovascular disease
590 (8.3)
159 (8.9)
191 (10.7)
123 (6.9)
117 (6.6)
0.003*
Malignant tumor
1,195 (16.8)
326 (18.3)
350 (19.7)
287 (16.1)
232 (13.0)
0.005*
eGFR (mL/min)
74.52 (55.65–98.39)
70.21 (54.76–92.81)
70.53 (54.95–96.91)
76.14 (56.09–101.18)
81.68 (57.33–102.57)
<0.001**
ACR (mg/g)
0.003*
A1
1,681 (23.6)
462 (26.0)
472 (26.5)
389 (21.9)
358 (20.1)
A2
4,739 (66.6)
1,151 (64.7)
1,110 (62.4)
1,209 (68.0)
1,269 (71.3)
A3
693 (9.7)
165 (9.3)
196 (11.0)
180 (10.1)
152 (8.5)
Hemoglobin (g/L)
13.88±1.59
13.77±1.49
13.83±1.60
13.91±1.60
14.01±1.69
0.054
White blood cell (109/L)
7.20 (6.00–8.70)
6.90 (5.70–8.30)
7.30 (6.00–8.50)
7.40 (6.10–8.81)
7.50 (6.10–9.30)
Platelet (109/L)
237.00 (197.00–284.55)
232.00 (196.00–279.00)
232.00 (191.22–280.00)
237.00 (202.00–281.00)
249.00 (205.00–298.00)
Albumin (g/L)
41.83±3.69
42.10±3.72
41.87±3.59
41.75±3.54
41.57±3.89
0.083
Calcium (mmol/L)
2.36±0.10
2.37±0.10
2.36±0.11
2.36±0.09
2.35±0.10
0.001**
Phosphorus (mmol/L)
1.22±0.20
1.23±0.18
1.21±0.19
1.21±0.19
1.22±0.22
0.05
Uric acid (μmol/L)
339.00 (273.60–404.50)
333.10 (279.60–398.50)
339.00 (279.60–404.50)
333.10 (267.70–404.50)
339.00 (273.60–410.40)
0.449
Mortality
0.071
Dead
1,376 (19.3)
363 (20.4)
380 (21.4)
318 (17.9)
315 (17.7)
Alive
5,737 (80.7)
1,415 (79.6)
1,398 (78.6)
1,460 (82.1)
1,464 (82.3)
Death cause
0.171
Cardiovascular disease
387 (28.1)
118 (32.5)
106 (27.9)
81 (25.5)
82 (26.1)
Malignant tumor
280 (20.3)
73 (20.1)
74 (19.5)
68 (21.4)
65 (20.6)
Other cause
709 (51.5)
172 (47.4)
200 (52.6)
169 (53.1)
168 (53.3)
Time to death or censoring from interview (mo)
74.00 (41.00–119.00)
71.00 (39.00–112.67)
71.00 (41.00–113.17)
73.00 (41.00–118.89)
79.00 (41.00–128.76)
0.078
Model
DII group
HR (95% CI)
All-cause mortality
Cardiovascular mortality
Malignant tumor mortality
Model 1a)
Q1
1
1
1
Q2
0.93 (0.82–1.06)
0.80 (0.63–1.02)
0.87 (0.65–1.17)
Q3
0.91 (0.80–1.03)
0.80 (0.63–1.02)
0.87 (0.65–1.17)
Q4
0.73 (0.64–0.83)**
0.63 (0.49–0.81)**
0.68 (0.50–0.93)*
P for trend
<0.001**
<0.001**
0.020*
Model 2b)
Q1
1
1
1
Q2
1.04 (0.91–1.86)
0.90 (0.71–1.15)
0.97 (0.72–1.30)
Q3
1.19 (1.04–1.36)**
1.06 (0.84–1.36)
1.13 (0.84–1.52)
Q4
1.21 (1.05–1.39)**
1.07 (0.83–1.39)
1.09 (0.80–1.50)
P for trend
<0.001**
0.270
0.359
Model 3c)
Q1
1
1
1
Q2
1.03 (0.91–1.18)
0.89 (0.70–1.13)
0.97 (0.72–1.31)
Q3
1.17 (1.03–1.34)*
1.05 (0.82–1.33)
1.11 (0.83–1.50)
Q4
1.23 (1.07–1.42)**
1.09 (0.84–1.42)
1.07 (0.78–1.47)
P for trend
<0.001**
0.274
0.538
Mediator
Subgroup
Effect size (95% CI)
Total effect of DII on all-cause mortality
Indirect effect of DII on all-cause mortality
% of effect via mediator
Hypertension
Total group
0.0090 (0.0050 to 0.0130)
0.0095 (–0.0055 to 0.0245)
NA
BMI <30 kg/m2
0.0179 (0.0110 to 0.0248)
0.0000 (–0.0002 to 0.0002)
NA
BMI ≥30 kg/m2
0.0008 (–0.0072 to 0.0088)
–0.0002 (–0.0006 to 0.0002)
NA
Diabetes
Total group
0.0131 (0.0030 to 0.0232)
–0.0005 (–0.0009 to –0.0001)
–3.82
BMI <30 kg/m2
0.0180 (0.0089 to 0.0271)
–0.0008 (–0.0015 to –0.0001)
–4.44
BMI ≥30 kg/m2
0.0002 (–0.0062 to 0.0066)
–0.0004 (–0.0009 to 0.0001)
NA
Cardiovascular disease
Total group
0.0103 (0.0050 to 0.0156)
0.0001 (0.0000 to 0.0002)
NA
BMI <30 kg/m2
0.0179 (0.0093 to 0.0265)
0.0002 (–0.0004 to 0.0008)
NA
BMI ≥30 kg/m2
0.0009 (–0.0940 to 0.0958)
–0.0001 (–0.0058 to 0.0056)
NA
Malignant tumor
Total group
0.0095 (0.0046 to 0.0144)
0.0001 (–0.0001 to 0.0003)
NA
BMI <30 kg/m2
0.0180 (0.0123 to 0.0237)
0.0001 (–0.0002 to 0.0004)
NA
BMI ≥30 kg/m2
0.0004 (–0.0085 to 0.0093)
0.0000 (–0.0003 to 0.0003)
NA
Table 1. Characteristics of the study participants according to DII
Values are presented as number (%), mean±standard deviation, or median (interquartile range) unless otherwise stated.
DII, dietary inflammatory index; BMI, body mass index; eGFR, estimated glomerular filtration rate; ACR, albumin to creatinine ratio (A1, <30 mg/g; A2, 30–300 mg/g; A3, 300 mg/g).
P<0.05.
P<0.01.
Table 2. Association between DII score and mortality
Adjusted for age, sex, race, income, body mass index group, smoking, hypertension, diabetes, cardiovascular disease, malignant tumors, estimated glomerular filtration rate, and albumin to creatinine ratio.
Table 3. Potential mediators of the association between DII score and all-cause mortality in patients with chronic kidney disease
The above models are adjusted for age, sex, race, income, body mass index group, and smoking.
DII, dietary inflammatory index; CI, confidence interval; BMI, body mass index; NA, not applicable.