Association between the Dietary Inflammatory Index and All-Cause Mortality in Adults with Obesity

Introduction: The dietary inflammatory index (DII) is associated with numerous chronic noncommunicable diseases. Previous studies have shown that the pro-inflammatory DII categories are associated with abdominal and simple obesity. However, the association between DII and mortality in patients with abdominal obesity and simple overweight or obesity remains unclear. Methods: We used data from the US National Health and Nutrition Examination Survey (NHANES) from 2007 to 2018. A DII >0 (positive DII) was defined as a pro-inflammatory diet. A restricted cubic spline curve was used to describe the trend between DII and all-cause mortality. We then examined the association between DII and all-cause mortality in different body types using a Cox regression analysis and investigated the differences between sexes. Finally, the mediating effects of systemic inflammation were explored. Results: A pro-inflammatory diet increased all-cause mortality in adults with abdominal obesity (aHR: 1.31, 95% confidence interval [CI]: 1.11–1.54; p < 0.001) and with simple overweight or obesity (aHR: 1.30, 95% CI: 1.11–1.53; p < 0.001). In addition, the most pro-inflammatory DII increased the risk of mortality by 43% (hazard ratio [HR]: Q4 vs. Q1 = 1.43, 95% CI = 1.14–1.79; p = 0.002; p for trend = 0.003) and 39% (HR: Q4 vs. Q1 = 1.39, 95% CI = 1.13–1.74; p = 0.003; p for trend = 0.009) in participants with abdominal obesity and with simple overweight or obesity, respectively. However, this association was not present in normal-sized participants. Compared with men, women resisted the effects of a pro-inflammatory diet. Mediation analysis showed that white blood cell and neutrophil were mediators of the association between DII and all-cause mortality (p < 0.001). Conclusion: A pro-inflammatory diet is associated with all-cause mortality in adults with abdominal obesity and simple overweight or obesity, and this effect differs between men and women. Systemic inflammation may mediate the association between DII and all-cause mortality.


Introduction
Obesity and overweight rates have increased since the 1980s among the adult population in the USA [1,2].Over the last few years, the prevalence of obesity has exceeded 30% in most age and sex groups, reaching 38.9% among all adults between 2013 and 2016 [3,4].Due to the increase in the prevalence of obesity, the consequent health risks have become a global concern [5].The increased rate of abdominal obesity is a sign of increased ectopic fat (on the liver, heart, skeletal muscle, and pancreas), and people with abdominal or simple obesity are significantly more likely to suffer from metabolic and cardiovascular diseases (CVD) [6].The mechanism by which obesity leads to chronic diseases may depend on factors involved in systemic inflammation, such as interleukin-6, C-reactive protein, and tumor necrosis factor-α [7].In obese adults, the number of anti-inflammatory immune cells is reduced and cytokine and leptin production is increased in adipose tissue, resulting in elevated inflammatory levels [8].Obesity and its complications are also significantly influenced by diet [9,10].
Pro-inflammatory dietary patterns are associated with many systemic diseases including CVD, obesity, type 2 diabetes mellitus, rheumatoid arthritis, and periodontal disease [11].Many dietary components are involved in the inflammatory process, and the cumulative effect of various interacting dietary components may alter inflammation levels and health outcomes [12,13].In 2014, Shivappa et al. [14] developed a dietary inflammatory index (DII) that quantifies individual diets based on the association between dietary components and biomarkers of inflammation.Life expectancy depends significantly on diet; furthermore, given the frequency of food intake, food intake plays a greater role in balancing chronic systemic inflammation than drug use [15].
Several previous studies have shown that the DII is associated with both simple and abdominal obesity [11].However, no studies have addressed the prognostic role of the DII in the people with different physical sizes; therefore, we explored this aspect.We also investigated the mediation effect of systemic inflammation (assessed through white blood cell (WBC) and neutrophil counts) on the association between all-cause mortality and DII in adults with abdominal obesity and simple overweight or obesity.We hypothesized that a higher DII is significantly associated with poorer survival in adults with abdominal obesity and simple overweight or obesity and that this burden may be mediated by systemic inflammation.

Study Population and Design
Data collected between 2007 and 2018 from the National Health and Nutrition Examination Survey (NHANES) was used in this cross-sectional study.NHANES is a multi-phase and stratified study administered by the Centers for Disease Control and Prevention (CDC) and the National Center for Health Statistics (NCHS) to residents of the District of Columbia and USA which focuses on health and nutritional status [16].Data were collected through interviews, laboratory, and physical examinations [15].The following inclusion criteria were used in this study: participants aged ≥20 years who underwent a physical examination and were interviewed about their nutritional intake.The study first included 51,472 participants with complete dietary data, and those with missing values for the following variables were excluded: death status (n = 24,079), marriage (n = 1,402), education (n = 25), chronic obstructive pulmonary disease (COPD) (n = 2), CVD (n = 2), tumor (n = 19), waist circumference (WC) (n = 979), body mass index (BMI) (n = 61), WBC (n = 896), and neutrophils (n = 42), leaving 23,965 participants in the final study cohort.Then, we divided the participants into four categories: with abdominal obesity (WC ≥102 cm for men and WC ≥88 cm for women), without abdominal obesity (WC <102 cm for men and WC <88 cm for women), with simple overweight or obesity (BMI ≥25 kg/m 2 ), and without simple overweight or obesity (BMI <25 kg/m 2 ) (Fig. 1).

Diet and DII
In the NHANES, dietary data were collected through interviews.Participants were asked to estimate the types and amounts of foods and beverages consumed during the 24 h period prior to the interview.We selected 26 dietary parameters available in the database and calculated DII scores based on previous literature [14].

Patient Characteristics
Demographic characteristics collected for the study were as follows: age, sex, ethnicity (Mexican American, Non-Hispanic Black, Non-Hispanic White, other Hispanic, and other race), education level (<9th grade, 9th-11th grade, high school graduate, some college or associate's degree, college graduate or above), and marriage status (divorced, living with partner, married, never married, separated, widowed).Laboratory data included WBC (10 9 /L) and neutrophil (10 9 /L) counts.We also included comorbidity data (CVD, hypertension, COPD, and tumors).

Diagnosis of Abdominal Obesity and Simple Overweight or Obesity
According to World Health Organization standard, participants with BMI >25 kg/m 2 were defined as overweight and obesity, and abdominal obesity was defined as WC ≥102 cm for men and ≥88 cm for women [17].

Statistical Analysis
Baseline characteristics are presented as mean (standard deviation) or median (interquartile range) for continuous variables and proportional for categorical variables.We classified DII >0 as the pro-inflammatory diet group, and DII <0 as the anti-inflammatory diet group [18].In addition, considering comparability across populations, we used the quartiles of DII in the total population for the segmentation.Cox proportional regression analysis was used to examine the association between DII and all-cause mortality in different body types and to generate hazard ratios (HRs) and 95% confidence intervals (95% CIs).We constructed four adjusted models, including the crude model, model 1 (adjusted for the demographic variables [age, sex, ethnicity, marital status, and education]), model 2 (adjusted for demographics and the inflammatory index [WBC and neutrophil counts]), and model 3 (adjusted for demographics, the inflammatory index, and comorbidities [COPD, CVD, hypertension, tumor]).Restricted spline regression was performed to evaluate the association between the DII and all-cause mortality.Kaplan-Meier curves were used to estimate survival and groups were compared using a log-rank analysis.The participants with abdominal obesity and simple overweight or obesity were divided into different subgroups according to sex for subsequent analyses.We also evaluated the role of the inflammatory index (WBC and neutrophil counts) as potential mediators.Mediation analysis was performed using the R package "mediation."WBC and neutrophil counts were used as indirect effects to assess the magnitude of the association between DII and all-cause mortality.We also conducted a trend analysis to determine the trend in mortality risk due to an increase in the DII.Additionally, we used logistic regression to additionally explore the association between DII and different comorbidities.Finally, to avoid the influence of dietary energy intake, recent death and drug usage on the results, we performed four sensitivity analyses that included dietary energy intake and obesity parameters (BMI and WC) as confounding factor, excluded deaths within 2 years and excluded people who used aspirin.Statistical analyses were performed using R version 4.2.0.All tests were two-sided, and statistical significance was set at p < 0.05.

Results
Baseline Clinical Characteristics of the Participants A total of 13,724 adults with abdominal obesity and 17,073 adults with simple overweight or obesity were included in this study, and their baseline clinical characteristics were displayed in Table 1 and online supplementary Table S1 (for all online suppl.material, see https://doi.org/10.1159/000533380).In participants with abdominal obesity, 10,308 consumed a pro-inflammatory diet and 3,416 consumed an anti-inflammatory diet, with a median DII of 1.91 and −1.04, respectively (p < 0.001).And in participants with simple overweight or obesity, 12,409 consumed a pro-inflammatory diet and 4,664 consumed an anti-inflammatory diet, with a median DII of 1.87 and −1.07, respectively (p < 0.001).BMI and WC were higher in the pro-inflammatory diet group.In addition, the values of inflammatory indicators (WBC and neutrophils) were lower in the anti-inflammatory group than those in the pro-inflammatory diet group.We observed that baseline differences between men and women and in both categories there were more women in the pro-inflammatory diet group.Although women have    Association of Inflammatory Diet with Mortality higher DII, they have a lower percentage of deaths than men, suggesting that women may better resist the adverse effects of a pro-inflammatory diet.

Association between DII and All-Cause Mortality in Adults with Different Body Types
Regardless of the adjustment method, the association between DII and all-cause mortality was not significant in participants without simple overweight or obesity or without abdominal obesity, suggesting that a proinflammatory diet did not increase the risk of mortality in a normal-sized population (online suppl.Table S2).Notably, in participants with abdominal obesity and simple overweight or obesity, elevated DII was significantly associated with an increased risk of all-cause mortality (Table 2).After full adjustment for confounders (model 3), the risk of death of the proinflammatory diet group was increased by 31% (aHR = 1.31, 95% CI = 1.11-1.54;p = 0.001) and 30% (aHR = 1.30, 95% CI = 1.11-1.53;p = 0.001) in participants with abdominal obesity and simple overweight or obesity, respectively.After dividing the DII into quartiles (Q1, <−0.43;Q2, ≥−0.43, <1.04; Q3, ≥1.04, <2.25; Q4, ≥2.25), a progressive increase in the HR was observed in participants with abdominal obesity and simple overweight or obesity.The restricted cubic spline curve also showed that the risk of death progressively increased with increasing DII in these two populations (Fig. 2a, d).Kaplan-Meier curves and log-rank test results suggested that individuals who consumed a pro-inflammatory diet had a worse prognosis (Fig. 3a, d).
Association between DII and All-Cause Mortality in Men and Women Table 3, 4 showed the results of the Cox proportional regression analysis results in men and women.The risk of death was 31% (aHR = 1.31; 95% CI = 1.02-1.56;p = 0.030) and 30% (aHR = 1.30; 95% CI = 1.04-1.76;p = 0.025) for men and women with abdominal obesity in the pro-inflammatory group, respectively.In men with abdominal obesity, the aHR was 1.37 in Q2 (95% CI = 0.99-1.90;p = 0.057), 1.53 in Q3 (95% CI = 1.12-2.10;p = 0.007), and 1.39 in Q4 (95% CI = 1.04-1.86;p = 0.025) (p for trend = 0.017).In men with simple overweight or obesity, the aHR was 1.49 in Q2 (95% CI = 1.11-2.00,p = 0.008), 1.58 in Q3 (95% CI = 1.20-2.09,p = 0.001), 1.44 (95% CI = 1.13-1.84,p = 0.003).However, in women with abdominal obesity and simple overweight or obesity, the second, third and highest quartiles of DII were not significantly associated with the risk of all-cause mortality compared to the lowest quartile.These results suggest that women show some resistance to the adverse effects of a pro-inflammatory diet.The restricted cubic curves among sexes are shown in Figure 2b, c, e, f.The Kaplan-Meier curves showed that men with abdominal obesity and simple overweight or obesity had a worse prognosis than women (Fig. 3b, c, e, f).

Mediation Analysis
After confirming the association between DII and allcause mortality, we constructed a mediation model with inflammatory indices as mediating indicators to estimate the mediating effect of systemic inflammation (Fig. 4).In participants with abdominal obesity, the mediating effect of WBC count accounted for 6.92% of the total effect (IE = −0.2295;95% CI = −0.2819 to −0.17; p < 0.001), and neutrophil count accounted for 11.2% (IE = −0.3765;95% CI = −0.4552 to −0.30; p < 0.001).And mediating effect of systemic inflammation was also present in participants with simple overweight or obesity.This suggests that dietary inflammation may affect the prognosis of adults with abdominal obesity and simple overweight or obesity by increasing the level of systemic inflammation.

Sensitivity Analysis
Previous studies have shown that dietary energy intake affected the association between DII and obesity, CVD, and metabolic disease.Therefore, we performed a sensitivity analysis in which the dietary energy intake was adjusted in each model (online suppl.Table S3).The results showed that a pro-inflammatory DII increased the risk of death by 35% (HR = 1.35; 95% CI = 1.15-1.58;p < 0.001) and 31% (HR = 1.31; 95% CI = 1.12-1.53;p < 0.001) in participants with abdominal obesity and simple overweight or obesity, respectively.After adjusted for BMI and WC, DII still had an adverse effect on all-cause mortality (online suppl.Table S4).Additionally, we excluded participants who died in the last 2 years and who took aspirin medication in the sensitivity analysis, and the results showed a consistent trend (online suppl.Table S5, S6).We additionally explored the relationship between DII and comorbidities and found an adverse relationship between pro-inflammatory diet and hypertension and CVD in the participants with different body types (online suppl.Table S7).

Discussion
In this study, we investigated the association between DII and the prognosis of adults with different body types in the USA for the first time.The results of the present study showed that the DII was negatively associated with all-cause mortality in adults with abdominal obesity and simple overweight or obesity, while this association was not present in adults with normal body size.Interestingly, a promising finding was the significant difference in the association between the DII and all-cause mortality between women and men.Our findings suggested that the pro-inflammatory diet has a greater prognostic impact on the men with abdominal obesity and simple overweight or obesity.Previous studies have shown that a positive DII was associated with an increased risk of various diseases, such as cancer [19], CVD [20], and diabetes mellitus [21], and has been shown to increase all-cause and cause-specific mortality [13,22,23].Additionally, studies have reported an association between DII and obesity.A study included 3,151 graduates and postgraduates in Universities of Minas Gerais (CUME project) cohort to explore the association between DII and overweight and obesity [24].
Another study included 7,236 participants of the PREDIMED cohort, and the results showed a positive association between DII and BMI, WC, and waist-toheight ratio [9].The Health Workers Cohort Study conducted a 13-year follow-up to evaluate the association between DII and different components of metabolic syndrome and found that participants in the highest DII quartile were associated with abdominal obesity (HR Q4 vs. Q1 = 2.68; 95% CI = 1.06-6.79;p for trend = 0.02) [25].However, to our knowledge, this study is the first to examine the association between DII and the prognosis of adults with different body types.
To date, the mechanism by which DII leads to increased all-cause mortality is not yet fully understood; however, several potential mechanisms have been proposed.First, shorter telomere length has been associated with higher all-cause mortality, and studies in Spanish and US populations have found association between DII Zheng/Ge/Ruan/Lin/Chen/Liu/Xie/Song/ Liu/Wang/Shi/Zhang/Yang/Liu/Deng/Shi  and telomere shortening [26,27].Inflammatory indicators such as C-reactive protein are elevated in those with a higher DII, resulting in excessive oxygen radical production and increased levels of oxidative stress.Telomeres are highly sensitive to oxidative stress damage; therefore, inflammation negatively affects telomere length [27,28].Second, DII is associated with increased systemic inflammation, and adipocytes and resident immune cells in the adipose tissue of obese individuals increase the level of circulating proinflammatory cytokines [29].Macrophages, the major cells in the leukocytes of adipose tissue, mediate nonspecific immune responses and produce cytokine storms [30].Systemic inflammation is the cornerstone of various physiological and pathological processes, and chronic exposure to high levels of inflammation can lead to altered homeostasis [31].
Interestingly, in the present study a progressive increase in the DII in men was associated with a worse prognosis than women.Several prior analyses have shown different associations between the DII and health outcomes between men and women.A Korean study investigated the association between DII and the risk of atherosclerotic CVD in Korean adults and found a statistical association in men but not in women [32].Based on previous literature, women tend to be less susceptible than men to environmental stresses, such as oxidative stress, and there are differences in the types and levels of sex hormones [33].In addition, estrogen has potent antioxidant and anti-inflammatory effects [34].Therefore, these differences may mediate the gender heterogeneity in the prognostic value of the DII in adults with abdominal obesity and simple overweight or obesity.To our knowledge, no studies have been performed that assessed sex differences in the association between DII and all-cause mortality in adults with abdominal obesity and simple overweight or obesity, and the exact mechanisms are unknown and need to be further investigated.
This study had several limitations.First, this was a cross-sectional study, and the causal associations between the parameters could not be clarified.Second, the dietary data were derived from self-reported 24 h dietary recall, which may not adequately reflect individual long-term dietary habits.In addition, only 26 food parameters were included to calculate the DII, which may not adequately reflect the true DII.Despite these limitations, our study had the following advantages.First, a relatively large and nationally representative database was used, and the results can be extended to the general US population owing to the random sampling design.Second, this study included a large sample size, and it is the first study on the association between the DII and the prognosis of adults with different body types.Therefore, future studies should focus on the range of DII that should be controlled under normal dietary conditions, and explore the role of sex in the association between DII and prognosis.

Conclusion
Our study suggests that DII is associated with all-cause mortality in adults with abdominal obesity and simple overweight or obesity, and this effect may be mediated by systemic inflammation.Men with abdominal obesity and simple overweight or obesity are more likely to be adversely affected by a pro-inflammatory diet.The population with high WC and BMI should consume more anti-inflammatory foods and reduce their DII to improve prognosis.

Fig. 2 . 3 (Fig. 3 .
Fig. 2. Association of DII with overall survival in participants with abdominal and simple overweight or obesity.(a) Full people with abdominal obesity; (b) men with abdominal obesity; (c) women with abdominal obesity; (d) full people with simple overweight or obesity; (e) men with simple overweight or obesity; (f) women with simple overweight or obesity.DII, dietary inflammatory index.

Fig. 4 .
Fig. 4. Mediation analysis of the effects of systemic inflammatory elements on the association of DII with overall survival.(a) WBC in participants with abdominal obesity; (b) neutrophil in participants with abdominal obesity; (c) WBC in participants with simple overweight or obesity; (d) neutrophil in participants with simple overweight or obesity.DII, dietary inflammatory index; IE, indirect effect; DE, direct effect.

Table 2 .
Association between DII and all-cause mortality in participants with simple overweight or obesity and abdominal obesity

Table 3 .
Association between DII and all-cause mortality in men and women with abdominal obesity

Table 4 .
Association between DII and all-cause mortality in men and women with simple overweight and obesity