Association between dietary inflammation index with anemia in Americans: a cross-sectional study with U.S. National health and nutrition examination survey

ABSTRACT

Objectives

Dietary inflammatory index (DII) is utilized to determine the inflammatory effects of nutrients and foods on various diseases. Inflammation is a potential risk factor for anemia. We hypothesize that pro-inflammatory diets boost the incidence of anemia, as indicated by high DII.

Methods

41, 360 Americans were included in this study from the U.S. National Health and Nutrition Survey (NHANES) from 2003–2018. Multivariable logistic regression models were employed to examine the association between DII and anemia.

Results

After adjustment for all the covariates, the odds ratios (ORs) (95% CI) between the risk of anemia and DII across tertile 3 were 1.2556 (95% CI 1.0621, 1.4843; P = 0.0077), and the trend test was statistically significant (P for trend = 0.009). Furthermore, in the subgroup analysis stratified by gender. The ORs (95% CI) between the risk of anemia and DII across tertile 2 and 3 were 1.8071 (95% CI 1.1754, 2.7783; P = 0.0070) and 2.1591 (95% CI 1.4009, 3.3278; P = 0.0005) in men after multivariable adjustment. However, in women, this association was only significantly different (P < 0.05) across tertile 3 in the crude model. In the subgroup analysis stratified by race, this association was significant (P < 0.05) between the risk of anemia and DII for Non-Hispanic Whites/Blacks after adjustment.

Discussion

Together, anemia was significantly associated with DII using logistic regression. In stratified analyses, higher DII scores were linked to an increased incidence of anemia in men, while no association was found in women after adjustment. Additionally, anemia may be associated with greater pro-inflammatory diets in Non-Hispanic Whites/Blacks.

Conclusion

In the present study, we evaluate the potential relationship between DII and anemia using data from NHANES. This cross-sectional study confirmed the hypothesis that the higher DII was significantly associated with a higher risk of anemia in the U.S. population.

KEYWORDS:

• Dietary inflammatory index
• NHANES
• Anemia
• Hemoglobin
• Diet
• ‌Cross-sectional study
• ‌Inflammation
• Pro-inflammatory diets

1. Introduction

Anemia is a worldwide general medical condition that can affect any individual, and it can be brought about by various pathophysiologic mechanisms [1–4]. Following the World Organization estimates, anemia affects 30% of women aged 15–49, 37% of pregnant women, and 40% of children aged 6–59 months globally. All varieties of anemia have a reduced ability to carry oxygen, resulting in serious health implications including fatigue, weakness, feeling dizzy, being drowsy, shortness of breath, and even affecting both morbidity and mortality [5]. In addition, anemia also harms social and economic progress. In 2010, 68.4 million years, or 9% of all disability worldwide due to all illnesses, were attributed to anemia [6]. According to the NHANES database, the prevalence of anemia rises from 4.03% during the 1999–2000 survey cycle to 6.49% during the 2017–2020 survey cycle [7]. This indicates that the incidence of anemia has not decreased with socioeconomic development.

A vital modifiable factor linked to the emergence of chronic diseases is dietary consumption. Relative to traditional public health research that focuses on individual nutrition, examining dietary patterns may be a more effective way to explore how nutrition interacts with the prevalence of chronic disease [8]. In the previous study, a particular dietary pattern was discovered to be connected to low-grade inflammation [9]. It is reported that children’s anemia and iron deficiency are correlated with nutrient density [10]. By accounting for dietary variations around the world, the Dietary Inflammatory Index (DII) was developed to thoroughly investigate the inflammatory potential of diet in the general population. Pro-inflammatory effects are indicated by higher DII scores, and anti-inflammatory effects are shown by lower DII scores [11,12]. Numerous studies have demonstrated an association between DII and various diseases, including cancer, diabetes, hypertension, and cardiovascular disease [13–15]. A link between maternal anemia and a pro-inflammatory diet and gestational diabetes mellitus was indicated by a previous study [16]. However, no studies have looked at its connection to anemia based on big data analysis, like the U.S. National Health and Nutrition Survey (NHANES) database.

Therefore, we hypothesize that inflammatory diets may promote the incidence of anemia, pro-inflammatory diets are the risk factor for anemia, while anti-inflammatory diets prevent the development of anemia. To update current knowledge on the correlation of anemia with inflammatory diets, this study examined the general U.S. population from 2003 to 2018 of the NHANES database and investigated the relationship of anemia with DII. This analysis further investigated periodic trends and prevalence within particular subgroups: sex, and race/ethnicity.

2. Methods and Materials

2.1. Study population

The data of the present study were obtained from the NHANES (https://www.cdc.gov/nchs/nhanes/index.htm). For determining disease prevalence and risk factors in the U.S. population, NHANES, a cross-sectional survey was conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC). NHANES has been conducted continuously in two-year cycles since 1999. The data that were provided by eligible participants included two 24-hour dietary recalls, health behaviors, laboratory tests, and physical examinations. These health measurement data are collected at mobile exam centers (MECs). Data from eight 2-year NHANES cycles (2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, and 2017–2018) were combined in this study. Figure 1 depicts the thorough inclusion and exclusion procedure. NHANES was approved by the National Center for Health Statistics Ethics Review Board, and all participants provided written informed consent.

Figure 1. Flow diagram of inclusion criteria and exclusion criteria. 41, 360 participants were included after excluding incomplete data of interest.

2.2. Calculation of DII

In-depth details regarding the development and validation of the DII can be accessed in previous literature [12]. Briefly, The Z-score is produced by individual daily intake subtracting the global average daily intake and dividing by the standard deviation. Then, it is converted to a percentile score, which is subsequently doubled and subtracted by ‘1’ to achieve a symmetrical distribution. Further, the percentile value is multiplied by the corresponding overall inflammation effect score. Finally, by adding up each DII score, we can obtain an individual ‘overall DII score’. To gather dietary data for the current investigation, we used two 24HRs from the NHANES database. The DII score was calculated using 26 of the 45 food parameters, including carbohydrates; protein; fat; alcohol; fiber; cholesterol; saturated, monounsaturated, and polyunsaturated fatty acids; omega3 and omega6 polyunsaturated fatty acids; niacin; vitamins A, B1, B2, B6, B12, C, E; iron; magnesium; zinc; selenium; folic acid; beta carotene; and caffeine. Importantly, even if the nutrients applied for the calculation of DII are less than 30, the DII scores are still available [12]. A low DII is indicative of an endocrine anti-inflammatory diet, and a high DII of a pro-inflammatory diet [17].

2.3. Anemia

The complete Blood Count with a 5-part Differential in laboratory data from the NHANES database was used to collect the Hemoglobin data. The definition of anemia was based on the serum hemoglobin (Hb) threshold (g/dL) as recommended by the WHO.

2.4. Covariates

Demographic characteristics presented on the NHANES website as potential confounders were considered as covariates in our study analyses, including age, gender (male, female), race/ethnicity (Mexican Americans, other Hispanic, non-Hispanic White, non-Hispanic Black, other race), educational level (less than 9th grade, 9–11th grade, high school, some college, college graduate, NA(refused or do not know)), marital status (married, widowed, divorced, separated, never married, living with partner, NA(refused)) and income status classified into three levels of the poverty income ratio (PIR) (≤1, 1–3, >3). PIR was calculated by dividing family (or individual) income by the poverty guidelines specific to the survey year. Other co-variate data, such as Body Mass Index (BMI) (<25, 25–30, >30), hypertension status (yes, no or NA(do not know or refused)), diabetes status (yes, no or NA(do not know or refused)), and smoking status (yes, no or NA(do not know or refused)) were obtained from the corresponding questionnaire based on a review of the literature and the clinical experience.

2.5. Statistical analyses

All analyses employed the recommended sampling weights to explain the intricate sample survey design of NHANES unless otherwise indicated in the Tables. First, categorical variables are presented as frequencies (percentages), and the chi-square test was used to compare the differences between groups. In the logistic regression analysis, the reference tertile for DII was the lowest. All data analyses were conducted using the software SAS Version 9.4 (SAS Institute, Cary, NC; 2017), STATA Version 15 (Stata Corporation, College Station, TX, USA), and Excel Version 2010 (Microsoft Corporation, Redmond, WA; 2010). All reported probabilities (P values) were two-sided, with P < 0.05 being considered statistically significant.

3. Results

3.1. Characteristics of participants

The detailed process of inclusion and exclusion is shown in Figure 1. Initially, 80313 potential participants were identified from eight cycles of NHANES (NHANES 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, and 2017–2018). After excluding 38953 participants with missing DII data, hemoglobin data, or covariate data who did not meet the inclusion criteria, we recruited the remaining 41360 eligible participants. Table 1 describes the basic characteristics of the study participants with or without anemia. Of the total 41360 participants, 1015 individuals met the criteria for the diagnosis of anemia in this study and the weighted prevalence of anemia was 2.45% overall. There were significant differences between the two groups in the distribution of DII, age, gender, race, education level, marital status, PIR, BMI, diabetes status, hypertension status, and smoking status (P < 0.05). Importantly, compared with the participants without anemia, those with anemia had higher DII.

Table 1. Descriptive characteristics of participants with and without anemia in the enrolled population of NHANES.

Variables	Anemia	Without Anemia	P-value
N (%)	1015 (2.45%)	40345 (97.55%)
Age, mean ± SD	45.5 ± 19.0	43.0 ± 18.9	0.0004
DII			<0.0001
Overall, mean ± SD	0.3 ± 1.5	−0.2 ± 1.5
Tertile 1 (%)	25.5	37.8
Tertile 2	36.5	33
Tertile 3	38	29.2
Gender (%)			<0.0001
Male	13.7	49.2
Female	86.3	50.8
Race (%)			<0.0001
Mexican American	11.7	8.9
Other Hispanic	6.0	5.1
Non-Hispanic White	38.9	68.8
Non-Hispanic Black	36.1	10.8
Other Race	7.3	6.5
Education (%)			<0.0001
Less than 9th grade	8.6	4.9
9–11th grade	14.9	9.6
High school graduate/GED or equivalent	21.1	20.3
Some college or AA degree	28.4	27.5
College graduate or above	17.5	24.8
NA	9.5	12.9
Marital Status (%)			<0.0001
Married	42.3	49.6
Widowed	9.3	4.9
Divorced	8.3	8.9
Separated	4.3	2.0
Never married	20.2	17.8
Living with Partner	7.9	6.8
NA	7.7	9.9
PIR (%)			<0.0001
≤1	23.9	14.2
1–3	45.8	39.8
>3	30.3	46
BMI (%)			0.019
<25	34.1	35.1
25–30	27.5	32.2
≥30	38.4	32.6
Diabetes (%)			<0.0001
Yes	18.8	9.5
No	81.2	90.5
NA	0	0.1
Hypertension (%)			<0.0001
Yes	36.3	27.4
No	60.0	65.8
NA	3.7	6.7
Smoking (%)			<0.0001
Yes	31.6	40.4
No	59.9	47.4
NA	8.5	12.1

Note: NHANES National Health and Nutrition Examination Survey, PIR family poverty income ratio, BMI body mass index, DII dietary inflammatory index.

3.2. Associations between DII and anemia

Associations between DII and anemia using logistic regression are shown in Table 2. In the crude model without adjustment, odds ratios (ORs) (95% CI) between the risk of anemia and DII across tertile 2 and 3 compared with tertile 1 were 1.4374 (95% CI 1.2183, 1.6959; P < 0.0001) and 1.7568 (95% CI 1.4980, 2.0603; P < 0.0001), respectively. ORs (95% CI) of model 1 between the risk of anemia and DII after adjustment for age, gender, and race across tertile 2 and 3 compared with tertile 1 were 1.2277 (95% CI 1.0379, 1.4523; P = 0.0167) and 1.3083 (95% CI 1.1109, 1.5409; P = 0.0013), respectively. While, after adjustment for all the covariates of interest in model 2, the ORs (95% CI) between the risk of anemia and DII across tertile 2 and 3 compared with tertile 1 were 1.1821 (95% CI 0.9978, 1.4006; P = 0.0531) and 1.2556 (95% CI 1.0621, 1.4843; P = 0.0077), respectively. Additionally, the trend test also showed that the risk of anemia and DII were all statistically significant in these three models (P for trend <0.001 in the crude model; = 0.002 in model 1; = 0.009 in model 2). Moreover, in spline analyses of hemoglobin content, the smooth curve fitting addressed the non-linear relationship between hemoglobin content and DII (Figure 2).

Figure 2. Spline analysis shows the positive association between DII and hemoglobin. a Each black point represents a sample. b Red line represents the smooth curve fit between variables. Blue lines represent the 95% of confidence interval from the fit. Race, age, gender, education, marital status, PIR, BMI, diabetes, hypertension, and smoke in life were adjusted. DII, Dietary Inflammatory Index; PIR, Poverty Income Ratio; BMI, Body Mass Index.

Table 2. Weighted odds ratios (95% confidence intervals) of anemia across tertiles of DII in the enrolled participants of NHANES 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, 2017–2018 cycles.

Variable	Crude Model		Model 1		Model 2
	OR (95%CI)	P-value	OR (95%CI)	P-value	OR (95%CI)	P-value
DII
Tertile 1	Reference		Reference		Reference
Tertile 2	1.4374 (1.2183, 1.6959)	<0.0001	1.2277 (1.0379, 1.4523)	0.0167	1.1821 (0.9978, 1.4006)	0.0531
Tertile 3	1.7568 (1.4980, 2.0603)	<0.0001	1.3083 (1.1109, 1.5409)	0.0013	1.2556 (1.0621, 1.4843)	0.0077
P for trend	<0.001		0.002		0.009

Note: In sensitivity analysis, dietary inflammatory index was converted from a continuous variable to a categorical variable (tertiles). Crude Model adjust for: None. Model 1 adjust for: Gender; Age; Race. Model 2 adjust for: Gender; Age; Race; Education; Marital Status; PIR; BMI; Diabetes; Hypertension; Smoking.

3.3. Stratified analyses

Furthermore, we conducted stratified analyses by subgroups defined by gender or race (Table 3). In the subgroup analysis stratified by gender, this association was significant (P < 0.05) between the risk of anemia and DII in males across tertile 2 and 3 compared with tertile 1 in the crude model (without adjustment), model 1^a (adjustment for age and race) and model 2^a (adjustment for all the covariates of interest). However, in women, this association was only significantly different (P < 0.05) across tertile 3 in the crude model. No association was observed between the risk of anemia and DII in women across tertile 2 in the crude model and across tertile 2 and tertile 3 in model 1^a and model 2^a (P > 0.05). These data indicate that the association between the risk of anemia and DII is more obvious in men than in women.

Table 3. Association between anemia and tertiles of DII by gender or race in the enrolled participants of NHANES 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, 2017–2018 cycles.

Variable	DII
	Tertile 1		Tertile 2		Tertile 3
	OR (95%CI)	P-value	OR (95%CI)	P-value	OR (95%CI)	P-value
Crude Model
Gender
Overall	Reference		1.2621 (1.0685, 1.4907)	0.0061	1.4009 (1.1926, 1.6455)	<0.0001
Male	Reference		1.9824 (1.3015, 3.0195)	0.0014	2.6413 (1.7488, 3.9893)	<0.0001
Female	Reference		1.1439 (0.9542, 1.3713)	0.1462	1.2346 (1.0374, 1.4693)	0.0176
Model 1^a
Gender
Overall	Reference		1.2277 (1.0378, 1.4523)	0.0167	1.3083 (1.1109, 1.5409)	0.0013
Male	Reference		1.9300 (1.2631, 2.9490)	0.0024	2.5205 (1.6570, 3.8341)	<0.0001
Female	Reference		1.0913 (0.9085, 1.3110)	0.3501	1.1112 (0.9303, 1.3272)	0.2448
Model 2^a
Gender
Overall	Reference		1.1821 (0.9978, 1.4006)	0.0531	1.2556 (1.0621, 1.4843)	0.0077
Male	Reference		1.8071 (1.1754, 2.7783)	0.0070	2.1591 (1.4009, 3.3278)	0.0005
Female	Reference		1.0537 (0.9051, 1.3015)	0.5795	1.0853 (0.9051, 1.3015)	0.3769
Crude Model
Race
Overall	Reference		1.3486 (1.1420, 1.5926)	0.0004	1.5284 (1.3010, 1.7957)	<0.0001
Mexican American	Reference		1.1069 (0.7698, 1.5917)	0.5835	0.9749 (0.6658, 1.4274)	0.8960
Other Hispanic	Reference		0.8090 (0.4479, 1.4611)	0.4821	1.3215 (0.7676, 2.2751)	0.3145
Non-Hispanic White	Reference		1.9684 (1.3749, 2.8181)	0.0002	2.0905 (1.4612, 2.9908)	<0.0001
Non-Hispanic Black	Reference		1.3632 (1.0466, 1.7757)	0.0216	1.6149 (1.2610, 2.0681)	0.0001
Other Race	Reference		1.2322 (0.7077, 2.1457)	0.4605	1.3486 (1.1420, 1.5926)	0.1282
Model 1^b
Race
Overall	Reference		1.2277 (1.0379, 1.4523)	0.0167	1.3083 (1.1109, 1.5409)	0.0013
Mexican American	Reference		1.0024 (0.6944, 1.4471)	0.9898	0.7804 (0.5305, 1.1481)	0.2081
Other Hispanic	Reference		0.6574 (0.3618, 1.1945)	0.1686	1.0693 (0.6162, 1.8557)	0.8116
Non-Hispanic White	Reference		1.8596 (1.2966, 2.6669)	0.0007	1.9832 (1.3801, 2.8498)	0.0002
Non-Hispanic Black	Reference		1.2483 (0.9543, 1.6328)	0.1055	1.3843 (1.0754, 1.7820)	0.0116
Other Race	Reference		1.0596 (0.6048, 1.8565)	0.8396	1.1759 (0.6657, 2.0771)	0.5766
Model 2^b
Race
Overall	Reference		1.1821 (0.9978, 1.4006)	0.0531	1.2556 (1.0621, 1.4843)	0.0077
Mexican American	Reference		1.0065 (0.6946, 1.4585)	0.9728	0.8259 (0.5560, 1.2268)	0.3434
Other Hispanic	Reference		0.6598 (0.3603, 1.2083)	0.1780	1.0980 (0.6197, 1.9457)	0.7486
Non-Hispanic White	Reference		1.7861 (1.2381, 2.5768)	0.0019	1.7827 (1.2199, 2.6051)	0.0028
Non-Hispanic Black	Reference		1.2070 (0.9203, 1.5830)	0.1739	1.3789 (1.0662, 1.7832)	0.0143
Other Race	Reference		1.0265 (0.5795, 1.8181)	0.9287	.0342 (0.5694, 1.8784)	0.9121

Note: In sensitivity analysis, dietary inflammatory index was converted from a continuous variable to a categorical variable (tertiles). Crude Model adjust for: None. Model 1a adjust for: Age; Race. Model 2a adjust for: Age; Race; Education; Marital Status; PIR; BMI; Diabetes; Hypertension; Smoking. Model 1b adjust for: Age; Gender. Model 2b adjust for: Age; Gender; Education; Marital Status; PIR; BMI; Diabetes; Hypertension; Smoking.

In the subgroup analysis stratified by race, the association between the risk of anemia and DII was significant (P < 0.01) for Non-Hispanic White across tertile 2 and 3 compared with tertile 1 in the crude model (without adjustment), model 1^b (adjustment for age and gender) and model 2^b (adjustment for all the covariates of interest). While, for Non-Hispanic Black, this association was significant (P < 0.05) across tertile 2 and 3 in the crude model, and across tertile 3 in model 1^b and model 2^b. However, no association was observed between the risk of anemia and DII for Mexican Americans, Other Hispanics, and other races in these three models (P > 0.05). Together, these data suggest that the association between the risk of anemia and DII varies in different races. Among them, this association is significant for Non-Hispanic White and Non-Hispanic Black.

4. Discussion

In the present study, we evaluate the potential relationship between DII and anemia using data from NHANES. This cross-sectional study confirmed the hypothesis that the higher DII was significantly associated with a higher risk of anemia in the U.S. population. The spline analysis of the association between DII and hemoglobin further demonstrated that lower DII was correlated to higher hemoglobin content, while higher DII was correlated to lower hemoglobin content. Considering that a higher DII indicates a pro-inflammatory diet and a lower DII represents an anti-inflammatory diet, our data suggests that the pro-inflammatory diet may be the potential risk factor for anemia. The pro-inflammatory diet has been reported to be associated with the development of some chronic diseases, such as diabetes [18], cognitive function impairment [18], and osteoporosis [19,20]. Because of their impact on oxidative stress and specific physiological pathways, phytochemicals may exert anti-inflammatory effects, whereas saturated fats, refined carbohydrates, and red meat could display anti-inflammatory effects [21]. Chronic inflammation is regarded as the potential reason for anemia, which could cause the dysfunction of red blood cells. Herein, They cannot absorb and utilize iron effectively. Additionally, under inflammatory status, the body also cannot respond normally to erythropoietin, a hormone produced by the kidneys that encourages bone marrow to generate red blood cells [22,23]. Given that IL-6 raises hepcidin to inhibit iron absorption [24], studies have been carried out to evaluate the effectiveness and security of IL-6 monoclonal antibodies, such as sultuximab, and anti-IL-6 antibodies, such as tocilizumab in the treatment of anemia. Although the former has been found to raise hemoglobin levels, it also increases the risk of infections [25].

Furthermore, we conducted stratified analyses by subgroups defined by gender. In men, we found that the DII was positively and significantly associated with the incidence of anemia in the unadjusted, micro-adjusted, and fully-adjusted models. Whereas, in women, this association was only significantly different across tertile 3 of DII in the unadjusted model. These data indicate that the relationship between DII and anemia is more obvious in men than in women. Therefore, in only considering the risk of anemia associated with DII, men should be more disciplined in their dietary patterns to mitigate a pro-inflammatory diet, which is beneficial for the prevention of anemia. Furthermore, the subgroup analysis stratified by race was also conducted. The positive association between the risk of anemia and DII is significant for Non-Hispanic White in unadjusted, micro-adjusted, and fully-adjusted models. While, for Non-Hispanic Black, this association was significant in the unadjusted model, and across tertile 3 of DII in micro-adjusted, and fully-adjusted models. However, no association was observed between the risk of anemia and DII for other races. These data suggest that the correlation of the risk of anemia to DII varies in different races. Among them, this association is most obvious for Non-Hispanic White. It is reported that Non-Hispanic Blacks had a significantly higher prevalence of anemia compared to Non-Hispanic Whites [26,27]. Similarly, the black race in the U.S. was significantly associated with a higher risk of anemia in pregnancy [28]. The reason why Non-Hispanic Blacks are more likely to suffer from anemia than Non-Hispanic Whites is currently unclear. One study showed that Naphthalene, an important contaminant in indoor and outdoor air, is associated with hemoglobin and hematocrit, the exposure to which was highest in Non-Hispanic Blacks [29]. Another study demonstrated that American Blacks with anemia had a higher prevalence of inflammation and microcytosis than Whites [30]. However, the risk of anemia in Non-Hispanic Whites is more sensitive to DII, as indicated by this study. The study we conducted has several advantages. On one hand, to the best of our knowledge, this is the first analysis to examine the relationship between anemia and DII in the U.S. population based on the NHANES database. On the other hand, to increase the sample size and improve the accuracy of DII assessments, we combined data from eight two-year survey cycles from the large, nationally representative database. There are a few potential limitations concerning our analysis. First, our study is a cross-sectional analysis based on the NHANES database, so we cannot establish a causal association. Second, 38,953 participants were excluded due to missing data, and dietary data was obtained using two 24-hour reviews from a cross-sectional survey design, which could not accurately reflect regular nutrient intake and may have influenced the results. Third, a common risk of anemia is nutritional-deficiency that can happen if the body does not ingest enough of certain supplements. In this study, the DII score was calculated using 26 of the 45 food parameters, some of the nutrients included in working out the DII score are themselves likely to have an impact on risk of anemia. For instance the intake of iron and B12 would reduce the risk of anemia, while the intake of caffeine would raise it. Additionally, due to the data being collected from the database, we are unable to include all potentially significant variables.

5. Conclusions

Although the conclusions of the present study are partially limited by some nonsignificant comparisons after multicovariates adjustment, The findings of this exploratory cross-sectional study suggest that the risk of anemia is significantly associated with DII using logistic regression in the U.S. population. Moreover, In stratified analyses, higher DII scores were linked to an increased incidence of anemia in men, while no association was found in women after adjustment, indicating that dietary patterns may be more important in preventing anemia in the male population in the U.S.. Additionally, anemia may be associated with greater pro-inflammatory diets in both Non-Hispanic White and Non-Hispanic Black people.

Author contributions

Xue Liu contributed to the conception of the study; analysis and manuscript preparation.

Consent to participate

Not applicable. NHANES was approved by the National Center for Health Statistics Ethics Review Board, and all participants provided written informed consent.

Disclosure statement

No potential conflict of interest was reported by the author(s).