Inverse association between dietary antioxidant indices and dietary inflammatory potential in jordanian adults: a cross-sectional study Authors: Lana Agraib, Buthaina Alkhatib, Asem Shawabkeh, Ahmad Suleiman Aslaih Published: 2026 ABSTRACT AbstractBackgroundThe Dietary Antioxidant Index (DAI) and Dietary Inflammatory Index (DII) measure the antioxidant capacity and inflammatory potential of diets, respectively, and are inversely associated with the risk of chronic disease.AimsThis study examines DAI characteristics and their association with DII in Jordanian adults, adjusting for sex, body mass index (BMI), and waist circumference (WC).MethodsA cross-sectional study was conducted, and 391 adults were randomly included. Sociodemographic, lifestyle, and anthropometric data (BMI and WC) were collected using standardized protocols. Dietary intake was assessed using a validated 157-item FFQ, from which DAI, Dietary Antioxidant Quality Score (DAQS), and DII were calculated using established international methods.ResultsFemales, individuals with overweight/obesity, and those with central obesity had significantly higher DAI and DAQS scores (p < 0.001). Unexpectedly, participants with overweight/obesity and central obesity were more frequently represented in higher antioxidant score categories, whereas normal-weight and normal-WC individuals clustered in lower antioxidant score categories (p < 0.001). DAI and DAQS decreased progressively across increasing DII quartiles (p < 0.001), showing a moderate, significant inverse correlation with DII (r ≈ − 0.32 to − 0.37). Regression analysis confirmed that higher antioxidant indices were associated with lower DII scores, accounting for 14–16% of the variance after adjustment.ConclusionDietary antioxidant indices were inversely and gradually associated with DII, independent of sex and adiposity measures. These findings indicate substantial overlap between antioxidant-rich dietary patterns and lower inflammatory dietary profiles. Keywords: Diet, Anti-Inflammatory; Antioxidants; Inflammation; Body Mass Index; Waist Circumference Introduction The Dietary Antioxidant Index (DAI) and the Dietary Inflammatory Index (DII) are two critical measures that assess the health implications of dietary patterns, focusing on antioxidant capacity and inflammatory potential, respectively [1, 2]. The DAI evaluates the cumulative ability of ingested antioxidants from various foods—primarily fruits, vegetables, and whole grains—to neutralize free radicals, which are linked to chronic diseases such as cardiovascular conditions, cancer, mental health, and type 2 diabetes [1, 3, 4]. Generally, DAI measures the overall antioxidant capacity of the diet, typically based on intake of vitamins (A, C, E), minerals (zinc, selenium), and carotenoids [5]. Research shows that a higher DAI is linked to better cardiovascular health. Specifically, studies have found that consuming antioxidants can increase serum antioxidant capacity, reduce oxidative stress, and lower total cholesterol levels, particularly LDL cholesterol [6]. Higher DAI scores indicate diets richer in antioxidants [4]. Another index used to assess antioxidant-nutrient intake is the Dietary Antioxidant Quality Score (DAQS) [7]. The methodologies for calculating these indices vary; the DAI employs scores based on food antioxidant capacities [1]. Conversely, the DII scores assess the inflammatory potential of consumed foods, guiding individuals toward anti-inflammatory diets that may reduce the risk of diseases such as obesity, cardiovascular disease, and certain cancers [8–10]. Accumulating evidence links dietary intake to inflammatory processes involved in the development of non-communicable diseases (NCDs). Higher DII scores are associated with elevated inflammatory markers and increased disease risk [11, 12]. Moreover, pro-inflammatory diets (higher DII score) are associated with both metabolically healthy and unhealthy obesity, but the risk is higher for those with metabolic complications [13, 14]. Multiple large studies and reviews confirm that higher DII scores are directly associated with higher body mass index (BMI) and waist circumference (WC), and with an increased risk of both general and abdominal obesity in adults, even after adjusting for confounders such as physical activity and adherence to healthy dietary patterns [12, 15]. Additionally, Researchers have found that lower DII scores (indicating anti-inflammatory effects) are associated with a 12–40% higher likelihood of multiple long-term conditions [16]. Multiple epidemiological studies have demonstrated that the DII can measure the inflammatory potential of dietary habits. As a result, the DII may serve as a useful tool for assessing an individual’s nutritional inflammatory capacity and, in turn, their risk of developing NCDs later in life [12, 17]. The DII uses a comprehensive assessment of 45 dietary parameters to assign inflammation-related scores to food items [18]. Several studies concluded the inverse relationship between DII and DAI; they found that high DII (pro-inflammatory) is linked to increased risk and mortality in COPD, cancer survivors, cardiovascular disease, kidney disease, and gallstones. On the other hand, high CDAI/DAI (antioxidant-rich) is associated with reduced risk and better outcomes in these conditions [2, 7, 19–21]. An indirect effect of DII on the DAI is that higher (pro-inflammatory) DII scores are consistently linked to poorer rankings on other dietary indices, which are considered to have high antioxidant contents, such as the Healthy Eating Index (HEI), Alternative HEI (AHEI), and DASH [22, 23]. Furthermore, anti-inflammatory diets (lower DII score) are associated with a lower glycemic index and healthier eating habits [22]. However, it remains an unexplored area of research regarding the association between DII and DAI among individuals with overweight and obesity (general and abdominal obesity). Additionally, the effects of several factors remain unexamined. Therefore, this study aims to investigate the characteristics of DAI and DAQS in the Jordanian population, their correlation with DII, and the influence of sociodemographic and anthropometric factors, such as sex, BMI, and WC, on this correlation. Materials and methodology Study design A cross-sectional study was conducted in the Hashemite Kingdom of Jordan between May 2023 and December 2024. A convenience sample of individuals regularly visiting outpatient clinics for routine evaluation at two governmental hospitals in the central region of Jordan (Al-Basheer Hospital and Zarqaa Governmental Hospital) was invited to participate, along with their caregivers. Any individual who was Jordanian, aged 18 years or above, able to communicate, and willing to participate was included in this study. Exclusion criteria included individuals who were non-Jordanian and under 18 years old, or who lacked a medical file in the Hakeem system (Jordan’s national electronic health record (EHR) system implemented across the Ministry of Health (MOH) facilities). It is a centralized, government-operated digital platform designed to standardize, document, and integrate patient medical records) and were pregnant women. Participants reporting physician-diagnosed chronic diseases (e.g., diabetes, cardiovascular disease, renal disease, inflammatory disorders) were excluded from the analysis to minimize potential bias related to medically prescribed dietary modifications. Ethical approvals and consent to participate The study protocol was reviewed and approved by the Institutional Review Board of The Hashemite University (No. 1/7/2022/2023) and the Jordanian Ministry of Health (MOH/REC/2023/375), in accordance with the Declaration of Helsinki. Before participating in the study, each willing and eligible subject provided written informed consent for enrolment. Sample size The Raosoft calculator (available online) was used to determine the required sample size, based on 2020 data from the Department of Statistics. Adults constituted 52.2% (5,621,970) of Jordan’s overall population (10,806,000). With a margin of error set at 5%, a confidence interval (CI) of 95%, and a response distribution estimated at 30%, the minimum necessary sample size was calculated to be 323 participants. Ultimately, a sample size of 472 individuals was successfully recruited. Only 391 participants have full data for DII analysis. (Fig. 1). Flow diagram of sample size calculation, recruitment, and the final analyzed sample Data collection The data were collected from participants through an interview conducted by a qualified nutritionist. The necessary data and anthropometric data were gathered and verified for accuracy under the supervision of the principal investigators. Collected data included demographic characteristics (sex, age, marital status, educational level, employment status, number of family members), lifestyle factors (smoking status, hours of sleep, physical activity), and public health and medication information. Anthropometric and body composition measurements Seca stadiometers (Seca, USA) were used to measure body height (measurement was recorded to the nearest 0.1 cm). In contrast, weight (kg) was measured using bioelectrical impedance analysis (BIA) with an InBody H20N body fat scale (FSA, US), with subjects wearing light clothing and barefoot (measurements were recorded to the nearest 0.1 kg). Furthermore, BMI was calculated in kilograms per square meter (kg/m²). RIEDER Body Measure tape, Lock Pin, and Push-Button Retract (Inct. Bonus Kit, REIDHK, China) was used to measure WC to the nearest 0.1 cm. The subjects were classified based on the WHO classification of excess weight (obese as a BMI ≥ 30 kg/m2, and overweight 29.9 ≥ BMI ≥ 25), and normal as a BMI ≥ 24.9 [24], and particularly abdominal obesity (abnormal WC) as a WC ≥ 102 cm for men and ≥ 88 cm for women [25]. Dietary Antioxidant Quality Score (DAQS) and Dietary Antioxidant Index (DAI) To calculate DAQS, a validated Arabic version of a semi-quantitative food frequency questionnaire (FFQ) comprising 157 items was used [26], and the FFQ was adapted by adding items needed to estimate DII components. For the current study, minor adaptations were made to the questionnaire/content processing to enable calculation of DII-related components. These modifications included questions about tea, ginger, thyme, pepper, some herbs, garlic, onion, saffron, curcumin, etc., and other DII components. Although the original instrument had been previously validated, the adapted version used in the present study was not formally revalidated, which is acknowledged as a limitation. A photographic Jordanian food atlas was used to determine portion size [27], then the collected dietary data analyzed by an automated nutrition analysis program (ESHA’s Food Processor, Nutrition Analysis Software, version 11.0; ESHA’s Food Processor® Nutrition Analysis Software (version 11.15.94, database structure 11.15.1) to estimate the daily nutrient intake of six antioxidant vitamins and minerals, that have been recognized to act as dietary antioxidants: vitamin A, vitamin C, vitamin E, selenium, manganese, and zinc. The score reflects intake of these specific vitamins and minerals. The daily intake of these vitamins and minerals was compared with the Recommended Dietary Allowance (RDA). Thus, the intake of each of the five evaluated antioxidant nutrients at the RDI was assessed separately, with a value of 0 or 1 assigned to each nutrient as follows: when intake was below 2/3 of the RDI, a value of 0 was assigned. Similarly, when intake exceeded 2/3 of the RDI, a value of 1 was assigned. Consequently, the DAQS ranged from 0 (pro-oxidative diet) to 5 (anti-oxidative diet) [7]. The DAI was calculated using the method developed by Wright et al. [28]. To estimate the DAI, each of the same six dietary vitamins and minerals was standardized in two ways. The first method involves subtracting the global mean and then dividing the result by the global standard deviation (Table 1). The second method consists of subtracting the Study population mean intake and then dividing the result by the Study population standard deviation. Then the DAI was calculated by summing up the standardized intakes of these vitamins and minerals and equal weight, as follows [5, 28, 29]: Table 1 The mean and standard deviation of nutrients included in the Dietary Antioxidant Index calculation [5, 28]NutrientUnitsEstimated Global Adults’ Mean IntakeEstimated Global adults Standard Deviation (SD)Study population mean intakeStudy population Standard Deviation (SD Vitamin C mg/day8025102.7470.65 Vitamin E mg/day1046.685.38 Zinc mg/day936.222.54 Selenium µg/day551577.0624.28 Vitamin A (RAE) µgRAE/day600200415.59294.99 Manganese mg/day2.20.74.112.46 The mean and standard deviation of nutrients included in the Dietary Antioxidant Index calculation [5, 28] Dietary Inflammatory Index (DII) The method developed by Shivappa and colleagues (2013) was used to calculate DII, which uses dietary intake data linked to a global database and provides means and standard deviations for various parameters [18]. Forty-five parameters were included in the DII and extracted from the FFQ. These parameters cover specific food items, fiber, tea flavonoids, spices, and macro- and micronutrients. The latter include vitamins like thiamin (B1), riboflavin (B2), niacin (B3), pyridoxine (B6), folic acid (B9), and cobalamin (B12), as well as vitamins A and C, and the minerals magnesium, selenium, iron, and zinc [30]. Each person’s intake is compared to the global average using a Z-score, which is calculated by subtracting the mean and dividing by the standard deviation. To account for skewness, this Z-score is converted to a percentile score ranging from − 1 (anti-inflammatory) to + 1 (pro-inflammatory). Each adjusted score is multiplied by its corresponding specific inflammatory effect score to obtain a specific DII score, and all these scores are summed to create an overall DII score [18]. Only 38 out of 45 items were calculated because not all items were provided by the ESHA food processor (7 items were missing: Eugenol, Flavan-3-ol, Flavones, Flavanols, Flavanones, Isoflavones, and Anthocyanidins). Statistical analysis SPSS software (IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Corp) was used to analyze the data. To calculate frequencies and percentages for general characteristics, descriptive statistics were used. Normality was assessed using the Shapiro–Wilk test and Q–Q plots. Homogeneity of variance was evaluated using Levene’s test. To assess differences in proportions, the chi-square test was used. One-way ANOVA or an independent t-test was used to compare the means. To determine the direction and strength of the association between the variables under investigation, Pearson correlation coefficients were computed in bivariate and partial analyses (adjusted for sex, age, BMI, and WC). Linear regression analysis was used to calculate the Beta and p-value for the trend. The Model 2 was adjusted for age, sex, body mass index, waist circumference, and smoking status. Outliers were excluded, and sensitivity analyses excluding extreme inflammatory values yielded similar results. The p-value of < 0.05 was considered statistically significant. Results Table 2 describes the study population classified into quartiles of the DII. The mean age decreased progressively from the lowest to the highest DII quartile: 46.46 ± 1.17 years in Q1, 44.56 ± 1.37 years in Q2, 42.60 ± 1.41 years in Q3, and 40.99 ± 1.23 years in Q4 (p = 0.020). When grouped by age category, middle-aged adults (36–55 years) predominated in Q1 (59.2%), whereas young adults (19–35 years) were more frequent in Q4 (36.5%). The proportion of older adults (56–65 years) declined from 21.4% in Q1 to 8.3% in Q4 (p = 0.010). The mean body weight ranged from 79.3 ± 2.01 kg in Q3 to 82.2 ± 2.04 kg in Q1, with no significant difference (p = 0.743). Height was similar across quartiles (≈ 160–162 cm; p = 0.179). Similarly, BMI did not differ significantly across quartiles, with an average of approximately 31 kg/m² (p = 0.906). Likewise, mean WC was comparable (95–98 cm; p = 0.479). BMI categories also showed no significant differences: normal weight ranged from 19.8% (Q2) to 24.5% (Q3–Q4), overweight from 21.4% (Q3) to 30.2% (Q2), and obesity from 48.0% to 54.1% across quartiles (p = 0.831). Females constituted the majority in all quartiles (≈ 80–88%), with no significant difference among groups (p = 0.455). Educational level showed similar distributions across DII quartiles (p = 0.745): approximately 55% had completed high school, approximately 20–28% had university education, and approximately 6% were illiterate. Moreover, marital status was also comparable (p = 0.788), with married individuals comprising about two-thirds of each quartile. Smoking status did not differ significantly (p = 0.752). Roughly one-third of participants in each quartile were current smokers, another third were non-smokers, and about one-third were passive smokers. Physical activity levels were similar as well, with approximately 26–32% reporting regular activity and 68–74% reporting none (p = 0.817). Table 2Descriptive characteristics of study population based on DII quartiles (n = 391)VariablesMean ± SEMp-value*Q1Q2Q3Q4Age (years)46.46 ± 1.1744.56 ± 1.3742.6 ± 1.4140.99 ± 1.23 0.020 Weight (kg)82.23 ± 2.0479.71 ± 2.1979.33 ± 2.0181 ± 1.990.743Height (cm)161.68 ± 0.79159.76 ± 0.68160.65 ± 0.68161.65 ± 0.70.179Body mass index (kg/m2)31.41 ± 0.7531.42 ± 0.9730.68 ± 0.7431.08 ± 0.770.906Waist Circumferences (cm)98.42 ± 1.7195.43 ± 1.6995.97 ± 1.6195.08 ± 1.650.479 Sex n (%) Male20 (20.2)12 (12.4)14 (14.1)14 (14.6)0.455Female79 (79.8)85 (87.6)85 (85.9)82 (85.4) Age groups (years) Young Adults (19–35)19 (19.4)27 (27.8)36 (36.4)35 (36.5) 0.010 Middle-aged adults (36–55)58 (59.2)46 (47.4)45 (45.5)53 (55.2)Older adults (56–65)21 (21.4)24 (24.7)18 (18.2)8 (8.3) Body mass index categories Normal weight20 (20.6)19 (19.8)24 (24.5)23 (24.5)0.831Overweight25 (25.8)29 (30.2)21 (21.4)21 (22.3)Obese52 (53.6)48 (50.0)53 (54.1)50 (53.2) Educational level Illiteracy6 (6.1)6 (6.2)5 (5.1)5 (5.2)0.745Primary11 (11.2)16 (16.5)15 (15.3)13 (13.5)High school53 (54.1)56 (57.7)55 (56.1)53 (55.2)University28 (28.5)19 (19.6)23 (24.4)25 (26.1) Marital status Single16 (16.2)18 (18.6)19 (19.2)24 (25.0)0.788Married65 (65.7)68 (70.1)65 (65.7)62 (64.6)Divorced9 (9.1)6 (6.2)7 (7.1)6 (6.3)Widow9 (9.1)5 (5.2)8 (8.1)4 (4.2) Smoking status Smoker29 (29.3)27 (27.9)38 (38.4)31 (32.3)0.752None-smoker34 (34.3)33 (34.0)27 (27.3)38 (39.6)Previous smoker1 (1.0)1 (1.0)0 (0.0)1 (1.0)Passive smoker35 (35.4)36 (37.1)34 (34.3)26 (27.1) Physical activity Yes26 (26.3)31 (32.0)31 (31.3)28 (29.2)0.817No73 (73.7)66 (68.0)68 (68.7)68 (70.8)* p-value <0.05 is considered statistically significant. DII: Dietary Inflammatory Index. Q: quartile; SEM: standard error of mean Descriptive characteristics of study population based on DII quartiles (n = 391) * p-value <0.05 is considered statistically significant. DII: Dietary Inflammatory Index. Q: quartile; SEM: standard error of mean Table 3 shows that DAI (global and sample-based) and DAQS were significantly higher in females compared with males (p = 0.004 and p < 0.001, respectively). Participants with overweight and obese had significantly higher DAI and DAQS scores compared with normal-weight individuals (all p < 0.001). Similarly, individuals with abnormal WC exhibited significantly higher antioxidant indices than those with normal WC (all p < 0.001). Table 3DAI description for study population based on Sex, BMI, and WC (n = 391)IndicesSex (Mean ± SEM)BMI (Mean ± SEM)WC (Mean ± SEM)MaleFemalep-valueNormalOverweightObesep-valueNormalAbnormalp-valueDAI Based on Global-1.19 ± 0.701.42 ± 0.47 0.004* -2.92 ± 0.631.71 ± 0.762.84 ± 0.59 < 0.001** -1.84 ± 0.512.85 ± 0.54 < 0.001** DAI Based on Sample-1.87 ± 0.33-0.56 ± 0.24 0.004* -2.76 ± 0.32-0.29 ± 0.390.11 ± 0.29 < 0.001** -2.18 ± 0.270.14 ± 0.27 < 0.001** DAQS1.55 ± 0.092.89 ± 0.08 < 0.001** 2.03 ± 0.112.48 ± 0.162.94 ± 0.10 < 0.001** 2.10 ± 0.102.89 ± 0.10 < 0.001** DAI Global (n(%)) Q128 (32.9)52 (21.9) 0.004* 48 (49.0)16 (19.8)15 (10.9) < 0.001** 62 (42.2)18 (10.3) < 0.001** Q227 (31.8)54 (22.8)26 (26.5)17 (21.0)36 (26.1)32 (21.8)49 (28.2)Q320 (23.5)61 (25.7)14 (14.3)27 (33.3)38 (27.5)31 (21.1)50 (28.7)Q410 (11.8)70 (29.5)10 (10.2)21 (25.9)49 (35.5)22 (15.0)57 (32.8) DAI Sample (n(%)) Q128 (32.9)52 (21.9) 0.004* 48 (49.0)15 (18.5)15 (10.9) < 0.001** 61 (41.5)19 (10.9) < 0.001** Q227 (31.8)54 (22.8)24 (24.5)18 (22.2)38 (27.5)31 (21.1)50 (28.7)Q320 (23.5)61 (25.7)14 (14.3)25 (30.9)40 (29.0)31 (21.1)50 (28.7)Q410 (11.8)70 (29.5)12 (12.2)23 (28.4)45 (32.6)24 (16.3)55 (31.6)**Association is significant at the 0.01 level (2-tailed). *. Association is significant at the 0.05 level (2-tailed). Percentages are column percentages; minor discrepancies may occur due to rounding. BMI: body mass index; DAI: Dietary Antioxidant Index; DAQS: Dietary Antioxidant Quality Score; DII: Dietary Inflammatory Index; WC: waist circumference.; Q: quartile; SEM: standard error of mean DAI description for study population based on Sex, BMI, and WC (n = 391) **Association is significant at the 0.01 level (2-tailed). *. Association is significant at the 0.05 level (2-tailed). Percentages are column percentages; minor discrepancies may occur due to rounding. BMI: body mass index; DAI: Dietary Antioxidant Index; DAQS: Dietary Antioxidant Quality Score; DII: Dietary Inflammatory Index; WC: waist circumference.; Q: quartile; SEM: standard error of mean When stratified by BMI, participants with normal BMI had a significantly lower global DAI (− 2.92 ± 0.63) than those who were overweight (1.71 ± 0.76) or obese (2.84 ± 0.59) (p < 0.001). A similar pattern was observed for the sample-based DAI (normal: −2.76 ± 0.32; overweight: −0.29 ± 0.39; obese: 0.11 ± 0.29; p < 0.001). DAQS also increased progressively across BMI groups (normal: 2.03 ± 0.11; overweight: 2.48 ± 0.16; obese: 2.94 ± 0.10; p < 0.001). Likewise, participants with abnormal WC showed a significantly higher global DAI (2.85 ± 0.54) than those with normal WC (− 1.84 ± 0.51; p < 0.001). The same trend was evident for sample-based DAI (0.14 ± 0.27 vs. −2.18 ± 0.27; p < 0.001) and DAQS (2.89 ± 0.10 vs. 2.10 ± 0.10; p < 0.001). Analysis of DAI quartiles further supported these findings. For the global DAI, the lowest quartile (Q1) had a higher percentage of males (32.9%) than females (21.9%), whereas the highest quartile (Q4) had a higher percentage of females (29.5%) (p = 0.004). By BMI, 49.0% of normal-weight participants were in Q1, whereas 35.5% of obese individuals were in Q4 (p < 0.001). A similar pattern appeared when grouped by WC: 42.2% of participants with normal WC were in Q1, and 32.8% of those with abnormal WC were in Q4 (p < 0.001). Similar trends were observed for sample-based DAI quartiles. Figure 2 illustrates the distribution of study participants across five DAQS levels (1 = lowest to 5 = highest) according to sex (A), WC (B), and BMI categories (C). Females predominated in the higher antioxidant score categories (26.4% in score 3, 22.1% in score 4, and 10.6% in score 5), with a progressive rise in proportion from score 1 to score 5, whereas males were more concentrated in the lower scores (61.2% in score one vs. 11.1% for females) (p < 0.001). Participants with abnormal WC (central obesity) were more frequently represented in the higher DAI scores (26,6%, 23.1%, and 11.0%, for scores from 3 to 5), compared with those with normal WC who were more frequent in the lowest scores (34.9% in scores 1 and 2) (p < 0.001). A similar gradient appeared across BMI categories; participants with overweight or obesity were disproportionately represented in the upper DAI score levels (25.6%, 21.5%, and 10.0% for scores 3 to 5), while normal-weight individuals clustered at the lower scores (36.1% and 41.2% for scores 1 and 2 (p < 0.001). Panel (A) shows the distribution of participants across five DAQS levels (score 1 = lowest antioxidant quality; score 5 = highest), stratified by sex (male, female). Panel (B) presents the DAQS distribution by waist circumference categories (normal vs. abnormal WC). Panel (C) displays DAQS distribution according to BMI categories (normal weight, overweight, obesity). Values are expressed as column percentages. The y-axis represents the percentage of participants within each category, and the x-axis represents DAQS score levels (1–5). Asterisks indicate statistically significant differences across categories (χ² test, p < 0.001) As presented in Table 4a, DII quartiles showed strong and consistent inverse associations with all antioxidant indicators. The global DAI decreased steadily from Q1 to Q4 (5.30 ± 1.66, 4.50 ± 0.91, 1.28 ± 0.67, and − 1.23 ± 0.57, respectively; p < 0.001). The sample-based DAI followed a similar downward gradient (1.47 ± 0.81 to − 2.04 ± 0.27; p < 0.001). Also, DAQS declined significantly across DII quartiles (3.18 ± 0.25 in Q1 to 2.17 ± 0.10 in Q4; p < 0.001). Chi-square analyses confirmed these patterns. For global DAI quartiles, 47.5% of participants in the lowest DII quartile (Q1) were in the highest DAI quartile (Q4), while only 11.4% of those in DII Q4 were in Q4 (p < 0.001). Similar trends appeared for sample-based DAI: 50% of Q1 participants fell into DAI Q4, compared to 10.2% in DII Q4 (p < 0.001). DAQS distributions also shifted toward lower scores as DII increased; the highest DAQS category (score 5) included 32.5% of individuals in DII Q1 but none in DII Q4 (p < 0.001). Table 4b. The association between DAI as dependent and DII as independent (n = 391)IndexesDII(mean ± SEM)p-valueDII (n (%)Q1Q2Q3Q4p-value DAI Global Q11.50 ± 0.26 < 0.001** 9 (20.0)8 (20.0)12 (20.0)22 (40.0) < 0.001** Q21.81 ± 0.226 (10.0)5 (10.0)16 (30.0)32 (50.0)Q31.20 ± 0.176 (10.0)16 (20.0)27 (40.0)24 (30.0)Q40.32 ± 0.1519 (20.0)31 (40.0)19 (20.0)10 (10.0) DAI Sample Q11.68 ± 0.24 < 0.001** 7 (10.0)8 (20.0)11 (20.0)24 (50.0) < 0.001** Q21.63 ± 0.238 (10.0)4 (10.0)17 (30.0)31 (50.0)Q31.23 ± 0.175 (10.0)18 (20.0)26 (40.0)24 (30.0)Q40.31 ± 0.1520 (30.0)30 (40.0)20 (30.0)9 (10.0) DAQS 11.36 ± 0.30 < 0.001** 7 (20.0)6 (20.0)9 (20.0)17 (40.0) < 0.001** 21.45 ± 0.2212 (20.0)9 (10.0)16 (20.0)36 (50.0)31.56 ± 0.151 (0.0)16 (20.0)24 (40.0)24 (40.0)40.72 ± 0.157 (10.0)18 (30.0)21 (40.0)7 (10.0)5-0.57 ± 0.2313 (50.0)10 (40.0)4 (10.0)0 (0.0)**. Association is significant at the 0.01 level (2-tailed). *. Association is significant at the 0.05 level (2-tailed). DAI: Dietary Antioxidant Index; DAQS: Dietary Antioxidant Quality Score; DII: Dietary Inflammatory Index; Q: quartile; SEM: standard error of mean b. The association between DAI as dependent and DII as independent (n = 391) **. Association is significant at the 0.01 level (2-tailed). *. Association is significant at the 0.05 level (2-tailed). DAI: Dietary Antioxidant Index; DAQS: Dietary Antioxidant Quality Score; DII: Dietary Inflammatory Index; Q: quartile; SEM: standard error of mean Table 4a. The association between DII as dependent and DAI as independent (n = 391)IndicesDII (Mean ± SEM)p-valueQ1Q2Q3Q4DAI Based on Global5.30 ± 1.664.50 ± 0.911.28 ± 0.67-1.23 ± 0.57 < 0.001** DAI Based on Sample1.47 ± 0.811.08 ± 0.47-0.58 ± 0.34-2.04 ± 0.27 < 0.001** DAQS3.18 ± 0.253.29 ± 0.162.93 ± 0.132.17 ± 0.10 < 0.001** DAI Global (n(%)) Q19 (22.5)6 (15.0)6 (15.0)19 (47.5) < 0.001** Q28 (13.3)5 (8.3)16 (26.7)31 (51.7)Q312 (16.2)16 (21.6)27 (36.5)19 (25.7)Q422 (25.0)32 (36.4)24 (27.3)10 (11.4) DAI Sample (n(%)) Q17 (17.5)8 (20.0)5 (12.5)20 (50.0) < 0.001** Q28 (13.3)4 (6.7)18 (30.0)30 (50.0)Q311 (14.9)17 (23.0)26 (35.1)20 (27.0)Q424 (27.3)31 (35.2)24 (27.3)9 (10.2) DAQS (n(%)) 17 (17.5)6 (10.2)9 (12.2)17 (19.5) < 0.001** 212 (30.0)9 (15.3)16 (21.6)36 (41.4)31 (2.5)16 (27.1)24 (32.4)24 (27.6)47 (17.5)18 (30.5)21 (28.4)7 (8.0)513 (32.5)10 (16.9)4 (5.4)0 (0.0)**. Association is significant at the 0.01 level (2-tailed). *. Association is significant at the 0.05 level (2-tailed). DAI: Dietary Antioxidant Index; DAQS: Dietary Antioxidant Quality Score; DII: Dietary Inflammatory Index; Q: quartile; SEM: standard error of mean a. The association between DII as dependent and DAI as independent (n = 391) **. Association is significant at the 0.01 level (2-tailed). *. Association is significant at the 0.05 level (2-tailed). DAI: Dietary Antioxidant Index; DAQS: Dietary Antioxidant Quality Score; DII: Dietary Inflammatory Index; Q: quartile; SEM: standard error of mean When the analyses were reversed (Table 4b), examining mean DII values across DAI categories yielded a similarly strong inverse relationship. Mean DII values decreased significantly from the lowest to the highest global DAI quartile (1.5 ± 0.26 in DAI Q1 to 0.32 ± 0.15 in DAI Q4; p < 0.001). The sample-based DAI showed a similar pattern (1.68 ± 0.24 to 0.31 ± 0.15; p < 0.001). Distributional data supported this relationship; in the highest global DAI quartile (Q4), 40% of participants were in the lowest DII quartile (Q1), compared to only 10% in DII Q4 (p < 0.001). Similar findings were seen for sample-based DAI and for DAQS categories: the highest DAQS group (score 5) included 50% of participants in DII Q1 but none in DII Q4 (p < 0.001). In Table 5, correlation and linear regression analyses examined the bidirectional relationship between DII and DAI (DAI based on global and sample references, as well as DAQS). The results show that there is a significant moderate inverse association between DII, DAI (based on global: r=-0.342, p < 0.001; based on sample: r = -0.369, p < 0.001), and DAQS (r = -0.323, p < 0.001), which means that DII reduces with increasing antioxidant content. A similar result was revealed after adjustment for confounding factors (sex, age, BMI, WC, and smoking). Linear regression analysis shows that a one-unit increase in DAI and DAQS is associated with a decrease in DII, with β ranging from − 0.08 (95% CI: -0.10 to -0.05) to -0.43 (95% CI: -0.59 to -0.28). This relationship accounts for 10.5%- 13.6% of the variation in DII. After adjustment for age, sex, BMI, WC, and smoking status, the β value showed minimal change. In contrast, the percentage of DII variance explained by DAI and DAQS increased to 14.1–16% (adjusted R² = 0.14–0.16), indicating that age, sex, and body composition may mask the effect of DAI and DAQS on DII. On the other hand, the DII increasing by one unit significantly leads to β=-1.5 (95%CI: -(-2.00; -1.00) reduction in DAI based on global cut points, β=-0.81 (95%CI: -1.06; -0.56) in DAI based on sample cut points, and β=-0.24 (95%CI: -0.33; -0.16) in DAQS. Furthermore, DII explains 11.7% (adjusted R² = 0.117) of the change in DAI based on Global, 13.6% of the change in DAI based on Sample, and 10.5% of the change in DAQS. The effect and alterations in DAI and DAQS, as explained by DII, increased after adjusting for age, sex, BMI, WC, and smoking status, ranging from 25.8% to 30.9% (adjusted R² =0.258–0.309). This indicates that the impact of DII on DAI and DAQS exceeds the inverse effect, and that the change in DII with respect to DAI and DAQS is related to the antioxidant concentration. Table 5The Correlation and linear regression analysis for DII with DAI (n = 391)IndexesPearson rp-valueβ (95%CIs)p-valueR2 change%of Changep-value DII as dependent and DAI as independent: Model 1 DAI Based on Global -0.342 < 0.001** -0.08 (-0.1; -0.05) < 0.001** 0.11711.7 < 0.001** DAI Based on Sample -0.369 < 0.001** -0.17 (-0.22; -0.12) < 0.001** 0.13613.6 < 0.001** DAQS -0.323 < 0.001** -0.43 (-0.59; -0.28) < 0.001** 0.10510.5 < 0.001** DII as dependent and DAI as independent: Model 2 DAI Based on Global -0.328 < 0.001** -0.08 (-0.11; -0.06) < 0.001** 0.14114.1 < 0.001** DAI Based on Sample -0.354 < 0.001** -0.18 (-0.24; -0.12) < 0.001** 0.16016.0 < 0.001** DAQS -0.333 < 0.001** -0.52 (-0.69; -0.35) < 0.001** 0.15115.1 < 0.001** DAI as dependent and DII as independent: Model 1 DAI Based on Global - - -1.50 (-2.00; -1.00) < 0.001** 0.11711.7 < 0.001** DAI Based on Sample - - -0.81 (-1.06; -0.56) < 0.001** 0.13613.6 < 0.001** DAQS - - -0.24 (-0.33; -0.16) < 0.001** 0.10510.5 < 0.001** DAI as dependent and DII as independent: Model 2 DAI Based on Global - - -1.37 (-1.84; -0.90) < 0.001** 0.25825.8 < 0.001** DAI Based on Sample - - -0.74 (-0.98; -0.51) < 0.001** 0.26826.8 < 0.001** DAQS - - -0.24 (-0.32; -0.16) < 0.001** 0.30930.9 < 0.001** **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). “- “: used instead of repetition of the value. DAI: Dietary Antioxidant Index; DAQS: Dietary Antioxidant Quality Score; DII: Dietary Inflammatory Index. Model 1: unadjusted value. Model 2: the values are adjusted for age, sex, body mass index, waist circumference, and smoking status The Correlation and linear regression analysis for DII with DAI (n = 391) **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). “- “: used instead of repetition of the value. DAI: Dietary Antioxidant Index; DAQS: Dietary Antioxidant Quality Score; DII: Dietary Inflammatory Index. Model 1: unadjusted value. Model 2: the values are adjusted for age, sex, body mass index, waist circumference, and smoking status Discussion The current study examined DII, DAI, and DAQS among Jordanian adults by sex, BMI, and WC. Females consistently showed higher antioxidant scores than males. Surprisingly, higher BMI and abnormal WC were also associated with markedly greater DAI and DAQS values, and lower DII scores. Moreover, higher antioxidant consumption (higher DAI and DAQS) was consistently linked with lower DII scores, demonstrating a strong, graded inverse relationship between DII quartiles and DAI measures. As DII increased from Q1 to Q4, the mean values of global and sample-based DAI and DAQS decreased significantly. Higher inflammatory scores are associated with lower antioxidant intake, and this relationship becomes stronger when demographic factors are included. Diets rich in antioxidants are linked to a reduced risk of high inflammation. This indicates that increasing the antioxidant content of foods may help reduce inflammation, thereby lowering the risk of chronic diseases. Adjusting for demographic and anthropometric factors strengthened these relationships, with explained variance rising to roughly 14–31% (adjusted R² =0.14–0.31). Regarding associations between the DII and body composition, Correa-Rodriguez et al. reported findings consistent with those of the current study: DII was significantly associated with fat-free mass and weight, but not with BMI or fat mass, in adults [31]. In contrast to the present findings Oliveira et al. have shown that participants both men and women in the fourth DII quartile had the highest prevalence of overweight and obesity (prevalence ratio [PR], 1.35; 95% confidence interval [CI], 1.14–1.59 and PR, 1.97; 95% CI, 1.20–3.22, respectively, in men; PR, 1.38; 95% CI, 1.17 to 1.65 and PR, 1.95; 95% CI, 1.31–2.90, respectively, in women). Moreover, participants in the highest DII quartile (most proinflammatory diet) were more likely to consume less antioxidant-rich foods such as red and ultra-processed meats, fats and oils (excluding olive oil), bottled fruit juices and soft drinks, sugars, sweets, and higher overall caloric intake, compared with the first quartile of DII [32]. Furthermore, in a cross-sectional study, 151 abdominally obese subjects were recruited from northwestern Iran. Three distinct dietary patterns emerged from principal component analysis. BMI and fat mass were found to be directly and significantly associated with the Western dietary pattern (characterized by lower antioxidant content and higher inflammatory content). On the other hand, serum lipopolysaccharide-binding protein (LBP) was negatively associated with the Healthy dietary pattern, even after adjustment for potential confounders. The Traditional pattern (characterized by higher intake of anti-inflammatory rich antioxidant foods such as poultry, vegetables, whole, juice or puree fruits) was linked to a decrease in DII (b = -0.3, p < 0.001) [33]. Another study has demonstrated that higher scores of DII are associated with a higher risk of obesity and overweight in childhood; The hazard ratio for subjects in the highest quartile (most pro-inflammatory diet (higher DII score)) was 1.32 (95% confidence interval 1.08–1.60) compared with participants in the lowest quartile (most anti-inflammatory diet (lower DII score)), with a significant linear dose-response relationship (p = 0.004). Consistently, increases in average yearly weight gains were significantly associated with proinflammatory diets. A proinflammatory diet was significantly associated with a higher annual weight gain and a higher risk of developing new-onset overweight or obesity [9]. Moreover, Aslani et al. reported that pro-inflammatory diets (higher DII scores) were associated with higher obesity indices, including hip circumference, WC, neck circumference, and abdominal obesity, in young children and adolescents [34]. Additionally, a study by Kendal et al. involving 81 participants found that anti-inflammatory diets (lower DII score) can be effective in managing obesity. The study also showed significant weight loss and reductions in visceral adipose tissue in the group that received anti-inflammatory diets (lower DII score). Following an anti-inflammatory diet (lower DII score) led to significant reductions in body weight and visceral fat, as well as improvements in participants’ cardiovascular health and inflammatory markers [35]. Ruiz-Canela et al. have shown direct associations between the DII and obesity indices, including BMI and WC, and have reported that diet may play a role in the development of obesity through mechanisms of inflammatory modulation [10]. These conflicting results may be due to participants’ age and DII, which could play significant roles in obesity and body fat percentage. Additionally, the studies mentioned earlier discussed the causality of obesity with higher DII and lower dietary antioxidant indicators, but did not describe the usual intake of overweight and obese participants. Consequently, the situation of normal-weight Jordanian adults is concerning, as they consume more proinflammatory components (higher DII) and lower DAI and DAQS compared to their overweight counterparts. Adults with excess weight tend to modify their intake to improve health, indicating greater awareness of healthy choices among these groups. Consistent with the present findings, some studies reported lower BMI in the highest versus the lowest dietary total antioxidant capacity (TAC) categories [28, 29, 36]. In contrast, some studies reported no significant differences in obesity prevalence or BMI across TAC categories [37, 38]. Central obesity, as measured by WC, has a varying prevalence. Some studies have reported a lower prevalence of central obesity or larger WC in the highest TAC categories than in the lowest [39, 40]. However, other studies found no statistically significant differences [41, 42]. Farhangi and colleagues found that higher categories of TAC were significantly linked to WC (WC; WMD: −1.17; 95% CI: −1.47, − 0.87; p < 0.001), while there was no significant change in BMI (BMI; WMD: −0.17; 95% CI: −0.35, 0.01; p = 0.120) [43]. By contrast, in a cross-sectional study using data from the National Health and Nutrition Examination Survey (NHANES) 2009–2018, CDAI was inversely associated with obesity (OR: 0.97, 95% CI: 0.95–0.99). Participants in the highest CDAI quartile had a 23% lower risk of obesity compared with the lowest quartile (OR: 0.77, 95% CI: 0.62–0.95) [44]. Our observation that participants with overweight/obesity and central obesity exhibited higher DAI/DAQS scores seems counterintuitive and requires careful interpretation. Since DAI/DAQS are based on intake and do not account for absorption, bioavailability, or utilization of antioxidants, higher antioxidant intake may not correspond to increased antioxidant activity in the body. In individuals with obesity, heightened oxidative stress and inflammation can accelerate antioxidant turnover, meaning that higher reported intake doesn’t necessarily reflect improved biological antioxidant status. Consequently, diet-based antioxidant indices may not align with actual physiological antioxidant capacity in states of adiposity-related metabolic stress [45]. Additionally, reverse causality is a potential concern in a cross-sectional design. Individuals with obesity or central obesity may have received dietary counselling or self-initiated dietary modifications (e.g., increased fruit/vegetable intake or supplement use) after weight gain or medical advice. This can elevate antioxidant index scores even when adiposity remains high, particularly if the changes are recent or partial. Third, differential misreporting may contribute: underreporting of energy-dense foods is more common in higher-BMI groups, which can artifactually improve diet-quality scores, including antioxidant-related indices, when based on FFQ self-report. Relatedly, residual confounding from unmeasured factors (dieting history, supplement use, socioeconomic gradients, comorbidity-driven restrictions) may inflate antioxidant indices in heavier participants even after adjustment for age/sex/BMI/WC/smoking. These explanations may also help reconcile our results with reports showing inverse associations between antioxidant indices and obesity-related outcomes. For example, population-based analyses of composite dietary antioxidant indices have reported lower odds of obesity (defined as BMI- and WC-based criteria) among individuals with higher antioxidant scores [46]. Differences in index construction (nutrient-based vs. food-based components), energy adjustment, and outcome definition (BMI-only vs. dual criteria including WC) may partly explain the divergent findings. Additionally, DAQS-focused work in obese cohorts often evaluates metabolic correlates within obesity rather than cross-weight comparisons, which can yield distinct patterns [45]. The results on the correlation between DII and dietary antioxidant indicators are consistent with published findings. It has been found that diets high in antioxidants and aligned with the Mediterranean diet were strongly associated with lower DII scores (i.e., more anti-inflammatory). In contrast, energy-dense, low-quality diets were associated with higher DII scores [47]. A scoping review of 65 studies found that diets emphasizing fruits, vegetables, whole grains, and legumes consistently reduced inflammation, whereas red/processed meats and added sugars consistently increased it [48]. Moreover, vegan and Mediterranean diets scored lowest (most anti-inflammatory) on the DII and highest on diet quality scores. In contrast, the ketogenic and Dukan diets scored the highest for pro-inflammatory activity and the lowest for antioxidant capacity [49]. Likewise, in a French RCT follow-up, people with higher DII scores had higher mortality risk from cancer and cardiovascular disease. Still, this risk was reduced in those receiving antioxidant supplementation [50]. Additionally, NHANES analyses showed that adherence to anti-inflammatory diets (lower DII score) was consistently associated with higher Healthy Eating Index (HEI) scores, supporting the association between diet quality and inflammation [51]. In a study among COPD patients, the multivariate Cox proportional hazards model showed that increased DII was associated with higher all-cause mortality (HR (95% CI): 1.11 (1.04, 1.18), p = 0.002). In contrast, composite dietary antioxidant index (CDAI) was negatively correlated with all-cause mortality (HR (95% CI): 0.95(0.91, 0.99), p = 0.010). Moreover, a nonlinear correlation was observed between DII or CDAI and all-cause mortality [19]. However, discrepancies across studies may reflect differences in population characteristics and methodology. Clinical cohorts (e.g., COPD or cardiometabolic patients) often exhibit higher baseline inflammation, potentially strengthening observable diet–inflammation gradients. In contrast, our sample was community-based, and inflammatory burden was not measured directly. Additionally, variations in DII parameter inclusion (38 vs. 45 components), energy adjustment procedures, food composition databases, and categorization of antioxidant indices may influence effect sizes. Prospective designs incorporating objective inflammatory biomarkers are necessary to determine whether the observed dietary overlap translates into measurable biological effects. Among the strengths of the study are the use of different antioxidant indicators from various references. Additionally, this is the first study to describe the dietary intake of apparently healthy Jordanian adults, assessed using DII, DAI, and DAQS, by sex and anthropometric indices. Also, we used a locally validated questionnaire to assess dietary intake. An experienced dietitian administered the FFQ via an interview to reduce measurement error. To account for any missed components of DII, we added those to the FFQ, and an updated ESHA cloud version was used. Several limitations should be considered; As a cross-sectional study, we cannot establish cause-and-effect relationships or determine the sequence of events. It may also be affected by recall bias and confounding factors, which can complicate the interpretation of the results. Although several potential confounders were adjusted for in the analysis, and no significant differences were observed between groups with respect to available socioeconomic variables, residual confounding by unmeasured factors—such as education level—cannot be entirely excluded. Moreover, the DII was calculated using 38 of the original 45 food parameters, as flavonoid subclasses were not available in the food composition database. Although previous studies have shown that partial DII calculations remain valid, omission of certain antioxidant-related components may have led to underestimation or attenuation of the inflammatory potential, potentially biasing associations toward the null. Dietary intake was assessed using a food FFQ, which is subject to recall bias and measurement error. Misclassification of dietary intake may have affected the precision of estimates of DAI, DAQS, and DII. Although dietary intake was assessed using a previously validated Arabic FFQ, the questionnaire was modified by adding items required for DII calculation. Because the adapted version was not formally re-validated, some degree of measurement error cannot be ruled out. Additionally, participants were recruited from government health facilities, which may limit generalizability. This hospital-based sampling approach may not fully represent the broader Jordanian adult population, particularly individuals who do not routinely access public healthcare services. Finally, the observation of higher dietary antioxidant scores among individuals with overweight or obesity may appear counterintuitive. This finding could reflect reporting bias, dietary modification following weight-related counseling, or residual confounding. Longitudinal studies incorporating objective inflammatory and antioxidant biomarkers are required to clarify these relationships. In conclusion, overall, dietary antioxidant intake and quality were significantly influenced by sex and anthropometric status. Women and individuals with higher BMI or larger WC showed diets richer in antioxidants, as indicated by both global and sample-based DAI and DAQS. These findings suggest that nutritional guidance should account for sex and body composition to promote antioxidant-rich dietary patterns and support metabolic health. Together, these analyses demonstrate a robust, graded, inverse relationship between antioxidant intake/quality and dietary inflammatory potential. Higher inflammatory scores correlate with lower antioxidant intake, and this association is stronger after controlling for demographic factors. Diets high in antioxidants are associated with a reduced risk of elevated inflammation. This suggests that improving the antioxidant capacity of foods may help reduce elevated inflammation, thereby lowering the risk of chronic diseases. Acknowledgements We extend our sincere gratitude to the study participants. Author contributions Buthaina Alkhatib conceived the research idea, managed overall scientific tasks, and prepared the manuscript. Lana M. Agraib conceived the research idea, performed data analysis, and prepared the manuscript. Ahmad Suleiman Aslaih participated in manuscript preparation and data analysis. Asem Taher Shawabkeh participated in manuscript preparation, data collection, and analysis. All authors critically revised the manuscript, approved the final version to be published, and agreed to be accountable for all aspects of the work. Funding The project has been granted by the Dean of Scientific Research at the Hashemite University (15/1/2022/2023). Data availability The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval The study protocol was reviewed and approved by the institutional review board of The Hashemite University (No.1/7/2022/2023) and the Jordanian Ministry of Health (MOH/REC/2023/375) in accordance with the Helsinki Declaration. Consent to participate All participants were informed of the study objectives at the beginning of the questionnaire and provided a written informed consent to participate. Their data was kept anonymous. 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. References Associated Data Data Availability Statement The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.