Association of Dietary Patterns with Blood Heavy Metal Concentrations: Results from the Korean National Health and Nutrition Examination Survey 2012–2016

Abstract

This study investigated possible associations between dietary patterns and blood heavy metal levels in Korean adults, using data from the Korea National Health and Nutrition Examination Survey (2012–2016). To explore these associations, demographic, physical activity, anthropometric, biochemical, and dietary data, including a food frequency questionnaire, were analyzed. Foods were categorized into 19 groups, and principal component factor analysis identified three dietary patterns: Meat and processed food-enriched diet (MPD), vegetables and milk-enriched diet (VMD), and fermented and fish-enriched diet (FFD). Multivariate logistic regression was used to evaluate the association between dietary patterns and high levels of heavy metals in the blood. The results showed that a high MPD score had a positive association with high levels of blood Pb (OR = 1.470, 95% CI = 1.173–1.842) and Hg (OR = 1.559, 95% CI = 1.259–1.932); a high FFD score also showed a positive association with high levels of blood Pb (OR = 1.492, 95% CI = 1.227–1.814) and Cd (OR = 1.276, 95% CI = 1.045–1.559). In contrast, VMD score was negatively associated with high levels of blood Pb (OR = 0.760, 95% CI = 0.628–0.920) and Cd (OR = 0.948, 95% CI = 0.781–1.151). Moreover, the effect of each dietary pattern on blood heavy metal levels showed differences by sex. Some dietary patterns, such as a high intake of meat, processed foods, fermented foods, and fish, can increase blood heavy metal levels, whereas other dietary patterns, such as vegetables and milk, have a protective effect against heavy metal concentrations.

1. Introduction

Heavy metals are a group of metals and metalloids with an atomic density higher than the 4000 kg/m³. Among them, lead (Pb), mercury (Hg), and cadmium (Cd) are known as hazardous heavy metals that humans are commonly exposed to through contaminated environmental systems, food consumption, drinking water, inhalation in the air, and contact with skin [1,2]. Heavy metals that enter the body replace other useful elements and accumulate in various tissues, such as those found in blood vessels, bones, muscles, and joints, and organs, interfering with biological functions and potentially causing irreversible changes in the body [1,3]. These effects can cause serious health issues in human beings, even at a very low concentration. Previous epidemiologic studies and systematic reviews have reported an association between exposure to heavy metals and cardiovascular disease [3,4,5], metabolic syndrome [6,7,8], neurodevelopment [9,10,11], fatty liver disease [12,13], kidney function impairment [14,15], and breast cancer [16]. Notably, exposure to heavy metals during pregnancy and childbirth can result in these metals being transmitted to the fetus or newborn through the placenta and breastfeeding, resulting in a shortened chromosomal telomere length, decreased birth weight, and impaired neurodevelopment [17,18,19]. Therefore, given their non-biodegradable nature, accumulation in various parts of the body, long biological half-lives, and health effects, it is essential to investigate the sources of exposure to heavy metals.

In addition to occupational exposure, dietary exposure is a major source of heavy metal exposure [20,21]. A total dietary survey of 2976 residents of Guangzhou, China, revealed that rice, seaweed, vegetables, and raw marine products accounted for 89% of Cd exposure [22]. A study that analyzed the relationship between blood Hg (Hg-B) and Pb (Pb-B) concentrations and diet in 1026 adolescents aged 8–17 years reported that fish and fruit dietary patterns and high fast food intake patterns were significantly associated with high levels of Hg and Pb, respectively [23]. However, a protective effect against heavy metal concentrations in the blood can be obtained via dietary intake of, for example, vitamins, selenium, zinc, iron, calcium, magnesium, green tea, and jujube [24,25,26]. This suggests that dietary interventions, such as encouraging or limiting specific diets, can contribute to reducing heavy metal exposure. For example, in an intervention study targeting pregnant women and women of childbearing age, a low frequency of fish intake (12 ounces or less per week) and the consumption of specific fish species (fish from lower trophic levels in the food chain) were found to be associated with decreased levels of methylmercury [27,28].

Given the complex nature of diet and the human body, dietary interventions to control heavy metal exposure and subsequent health effects need to be based on more elaborate evidence. While studies exist on metal exposure through the consumption of individual food items [29,30,31], knowledge about the role of the overall diet is limited [32,33,34,35]. Therefore, there is a growing interest in using dietary patterns as a realistic and comprehensive approach to assessing the complex health effects of food consumption [36]. In addition, the concentrations of heavy metals such as Pb and Cd continue to be higher in Korea than in other major countries [37]. In this study, we aimed to identify dietary patterns that contribute to blood Cd (Cd-B), Pb-B, and Hg-B concentrations based on data from the National Health and Nutrition Examination Survey (KNHANES) targeting the entire Korean population.

2. Materials and Methods

2.1. Study Source and Population

This study used data from the KNHANES, conducted from 2012 to 2016 (2014–2015 data were excluded because heavy metals were not measured). The KNHANES is a cross-sectional, population-based national survey conducted by the Korea Disease Control and Prevention Agency that aims to assess the health and nutritional status of the Korean population. The KNHANES started in 1998 and is conducted annually. The key contents of this survey are as follows: a health status assessment, a nutritional status assessment, health behaviors, healthcare utilization, and an accessibility assessment. To ensure representativeness, a stratified multistage probability sampling method was adopted. A total of 16,712 adults with measured blood heavy metal levels were included in the trend analysis.

Among the 24,226 subjects who participated in the 2012, 2013, and 2016 KNHANES, we excluded subjects aged <20 years because dietary survey data were unavailable. We also excluded participants who did not have blood heavy metal levels (n = 11,750) or semi-quantitative food frequency questionnaire (SQ-FFQ) data (n = 2495). After excluding subjects who were pregnant or diagnosed with cerebrovascular disease, cancer, or diabetes mellitus, we finally included 4024 subjects in the analysis (Figure 1).

2.2. General Characteristics of Subjects

The KNHANES contains information on general characteristics, physical activity, and clinical data. General characteristics included current smoking status, region (urban or metropolitan, rural), occupation (administrators or professionals, office workers, service workers, agricultural or fishing workers, technicians or simple labor workers, and joblessness or other missing values). The physical activity of the subjects was divided into two levels: exercise (walking for at least 30 min at least 5 days per week or performing exercises such as swimming or badminton for more than 30 min at least 5 times per week) and no exercise. The following anthropometric measurement data were collected by the trained experts: body weight (kg), height (cm), body mass index (BMI) (kg/m²), waist circumference (WC) (cm), systolic blood pressure (SBP) (mmHg), and diastolic blood pressure (DBP) (mmHg). To collect clinical biochemical data, blood was sampled after 12 hours of fasting. Fasting blood glucose (mg/L), total cholesterol (mg/dL), low-density lipoprotein cholesterol (LDL-C) (mg/dL), high-density lipoprotein cholesterol (HDL-C) (mg/dL), triglycerides (mg/dL), aspartate aminotransferase (IU/L), and alanine aminotransferase (IU/L) were quantified in whole blood samples. In addition to occupation and physical activity data, we included data on sex, BMI, region, and smoking status, which could be potential covariates, for a more detailed analysis [38,39,40].

2.3. Heavy Metal Concentration in Blood

In KNHANES, to assess the concentration of heavy metals in whole blood, 3 mL of venous blood was collected from each participant using standard evacuated tubes containing sodium heparin (Vacutainer; Becton, Dickinson & Co., Franklin Lakes, NJ, USA). All analyses were performed by the Neodin Medical Institute, a laboratory accredited by the Korean Ministry of Health and Welfare and designated by the Ministry of Employment and Labor for trace metal analysis [41]. All sample collection and handling procedures were performed under standardized trace metal-free conditions to minimize external contamination. Blood Pb and Cd concentrations were determined using graphite furnace atomic absorption spectrometry (GFAAS) equipped with Zeeman background correction (AAnalyst 600; Perkin Elmer, Turku, Finland). Blood Hg levels were measured using the gold amalgamation method with thermal decomposition atomic absorption spectrophotometry (DMA-80; Milestone, Bergamo, Italy), which allows for direct analysis without chemical pre-treatment. To ensure analytical precision and accuracy, the laboratory participated in both internal and external quality-assurance programs. Internal quality control was conducted using commercial reference materials (Lyphochek Whole Blood Metals Control; Bio-Rad, Hercules, CA, USA), and the coefficients of variation for Pb, Cd, and Hg were maintained within acceptable ranges throughout the analysis period. For external validation, the laboratory regularly participated in the German External Quality Assessment Scheme (G-EQUAS) operated by Friedrich-Alexander University and the Quality Assurance Program of the Korea Occupational Safety and Health Agency. While the exact detection limits varied slightly depending on instrument sensitivity and matrix conditions, the method detection limits (MDLs) applied in the national monitoring protocol were approximately 0.12 µg/dL for Pb, 0.056 µg/L for Cd, and 0.158 µg/L for Hg. All analyzed samples in the present study had concentrations above these detection thresholds. The results were shown to have acceptable bias ranges, and inter-laboratory comparison testing indicated correlation coefficients of 0.921 for Pb and 0.932 for Cd, with equivalence confirmed at medical decision levels. For Hg, reproducibility was similarly ensured, with a correlation coefficient of 0.932 and equivalence at clinical thresholds [42].

2.4. Dietary Patterns

A total of 112 food items in the SQ-FFQ were categorized into 19 food groups based on the Korean nutrient database [43]: grains, noodles, instant noodles, fast food, potatoes, vegetables, fermented vegetables, fermented beans, fruits, red meat, white meat, processed meat, fish, fish soup, processed fish, seaweed, dairy products, alcohol, and soda (Table S1). The frequency of consumption for each food category was determined on a 9-point scale: 3 times per day, twice per day, once per day, 5–6 times per week, 2–4 times per week, once per week, 2–3 times per month, once per month, and nearly never. The amount consumed each time was estimated using the 3-point scale: 0.5 portion, 1 portion, and 1.5 portions. For the analysis, we calculated the average annual consumption by multiplying the frequency of consumption by the amount consumed each time for each item.

Factor analysis was used to identify the major dietary patterns of Korean adults. Dietary patterns were extracted using exploratory principal component factor analysis. Eigenvalues larger than 1.3 (Kaiser criterion) were applied to determine the number of factors. Scree plots were rotated using the varimax transformation to achieve a more interpretable result. After rotation, the factor scores for each dietary pattern were calculated by summing the intake of each food group weighted by the factor loading values. As shown in

Table 1. Factor loading matrix for major factors in the dietary patterns of Korean adults.

Food Groups	Component 1
	Meat and Processed Foods-Enriched Diet (MPD)	Vegetables and Milk-Enriched Diet (VMD)	Fermented Foods and Fish-Enriched Diet (FFD)
Red meat	0.661
White meat	0.622
Instant noodles	0.593
Noodles	0.575
Alcohol	0.548
Fast food	0.539
Soda	0.517
Processed meat	0.485
Fruits		0.727
Vegetables		0.662
Potatoes		0.603
Dairy products		0.522
Seaweeds		0.421
Grains			0.629
Fermented vegetables			0.613
Fermented beans			0.611
Fish soup			0.401
Processed fish			0.338
Fish			0.332
Total% of variance explained	17.468	9.882	7.745

Extraction method: principal component analysis; Rotation method: varimax with Kaiser Normalization; ¹ Factor loadings are only displayed for values <−0.30 or >0.30.

, three distinct patterns were identified based on the factor loading values. Each pattern was named according to the degree of factor loading. Pattern 1 is characterized by the consumption of meat and processed or instant food content as a meat and processed food-enriched diet (MPD); pattern 2 is characterized by the consumption of fruits, vegetables, and dairy products as a vegetables and milk-enriched diet (VMD); and pattern 3 is characterized by the consumption of typical foods in Korea, including grains, fermented foods, and fish, as a fermented and fish-enriched diet (FFD). Patterns 1, 2, and 3 displayed 17.47%, 9.88%, and 7.75% variance in the total dietary intake, respectively.

2.5. Statistical Analyses

Continuous variables were presented as means ± standard error, while categorical variables were expressed as numbers and percentages of subjects. The subjects were divided into tertiles based on their dietary pattern scores, and their nutrient intake and serum heavy metal levels were compared. Differences among the tertiles of dietary pattern groups were assessed using chi-square test, analysis of variance, and general linear model (GLM), adjusting for age and sex, followed by post hoc analysis (Bonferroni method). Using multivariate logistic regression models, odds ratios (ORs) and 95% confidence intervals (CIs) for the highest tertile group of serum heavy metal concentrations across tertiles of dietary pattern scores were compared with the lowest tertile group after adjusting for covariates, including age, sex, body mass index, region, occupation, smoking status, and physical activity. Statistical significance was set at the p-value < 0.05 level. All statistical analyses were performed using SPSS version 26.0 (SPSS Corp., Armonk, NY, USA) and Stata/SE, version 13.0 (Stata Corp, College Station, TX, USA).

3. Results

3.1. General Characteristics, Heavy Metal Concentration, and Biochemical Values According to Dietary Patterns

Table S2 presents the general characteristics, blood heavy metal concentration levels, and biochemical markers of the study participants according to the tertiles of the three dietary patterns. In MPD, the highest tertile of score had a significantly higher proportion of males and a lower mean age, both before and after adjusting for age and sex (p < 0.001). Hg-B levels were significantly higher in the highest tertile of MPD score (p < 0.001). WC and triglyceride levels tended to increase as MPD score increased. In VMD, the proportion of male subjects increased as the tertile of the VMD score increased. The Pb-B concentration and DBP tended to decrease as VMD score increased before and after adjusting for age and sex. In FFD, the proportion of female subjects and the mean age tended to increase as the tertile of FFD score increased, even after adjusting for age and sex (p < 0.001). Pb-B and Cd-B concentrations, WC, and BMI showed an increasing pattern as the FFD score increased, even after adjustment.

3.2. Association Between Blood Heavy Metal Concentrations and Dietary Patterns

Figure 2 shows a forest plot of the logistic regression results for the association between high blood heavy metal levels and tertiles of dietary patterns. The ORs for highest tertile of Pb-B and Hg-B were significantly higher in the highest tertile of the MPD score than in the lowest tertile after adjusting the models. In model 4, adjusted for covariates including age, sex, BMI, region, occupation, current smoking status, and physical activity, a high MPD score was positively associated with high levels of Pb-B (OR = 1.470, 95% CI = 1.173–1.842) and Hg-B (OR = 1.559, 95% CI = 1.259–1.932), but not with Cd-B. The increasing tertile of the VMD pattern score showed a significant association with Pb-B (OR = 0.760, 95% CI = 0.628–0.920) and Hg-B (OR = 1.327, 95% CI = 1.104–1.596), but not with Cd-B, in model 4. In the highest tertile of the FFD score, Pb-B and Cd-B concentrations showed positive associations both before and after adjustment.

Table 2. Multivariate adjusted odds ratios for high heavy metal levels according to tertiles of score for dietary patterns.

Dietary Pattern	Heavy Metal	Model	Male			Female			p for Inter Action
			Dietary Pattern Scores			Dietary Pattern Scores
			T1	T2	T3	T1	T2	T3
MPD: Meat and processed foods-enriched diet	Lead	Model 1	1	0.976 [0.744; 1.281]	0.793 [0.618; 1.016]	1	0.549 [0.439; 0.687]	0.454 [0.341; 0.605]	0.004
		Model 2	1	1.570 [1.163; 2.119]	1.984 [1.476; 2.666]	1	1.071 [0.832; 1.378]	1.422 [1.013; 1.998]	0.053
		Model 3	1	1.529 [1.119; 2.089]	1.899 [1.394; 2.586]	1	0.985 [0.759; 1.279]	1.356 [0.957; 1.923]	<0.001
		Model 4	1	1.477 [1.079; 2.023]	1.764 [1.290; 2.413]	1	0.952 [0.732; 1.238]	1.283 [0.901; 1.828]	0.082
	Mercury	Model 1	1	1.051 [0.800; 1.380]	0.979 [0.763; 1.256]	1	0.871 [0.701; 1.082]	0.838 [0.645; 1.090]	0.400
		Model 2	1	1.479 [1.104; 1.981]	1.857 [1.392; 2.477]	1	1.275 [1.004; 1.618]	1.545 [1.137; 2.099]	0.789
		Model 3	1	1.365 [1.004; 1.856]	1.588 [1.173; 2.149]	1	1.265 [0.990; 1.618]	1.514 [1.104; 2.077]	0.696
		Model 4	1	1.67 [1.004; 1.862]	1.577 [1.160; 2.143]	1	1.258 [0.983; 1.610]	1.484 [1.080; 2.041]	0.670
	Cadmium	Model 1	1	0.808 [0.899; 1.089]	0.654 [0.497; 0.862]	1	0.487 [0.402; 0.589]	0.327 [0.256; 0.418]	<0.001
		Model 2	1	1.079 [0.790; 1.474]	1.215 [0.893; 1.653]	1	1.070 [0.854; 1.340]	1.245 [0.923; 1.681]	0.009
		Model 3	1	1.070 [0.774; 1.478]	1.230 [0.894; 1.693]	1	1.018 [0.808; 1.283]	1.179 [0.866; 1.605]	0.009
		Model 4	1	0.778 [0.543; 1.114]	0.753 [0.526; 1.077]	1	0.983 [0.778; 1.241]	1.097 [0.803; 1.500]	0.356
VMD: Vegetables and milk-enriched diet	Lead	Model 1	1	0.649 [0.520; 0.811]	0.587 [0.464; 0.744]	1	1.112 [0.857; 1.442]	1.190 [0.924; 1.532]	<0.001
		Model 2	1	0.634 [0.500; 0.804]	0.583 [0.452; 0.752]	1	1.016 [0.770; 1.340]	1.020 [0.780; 1.336]	0.002
		Model 3	1	0.598 [0.466; 0.768]	0.594 [0.456; 0.775]	1	1.102 [0.827; 1.469]	1.054 [0.796; 1.395]	0.004
		Model 4	1	0.595 [0.462; 0.765]	0.579 [0.442; 0.759]	1	1.131 [0.847; 1.511]	1.066 [0.801; 1.418]	0.005
	Mercury	Model 1	1	1.178 [0.944; 1.471]	1.267 [1.002; 1.603]	1	1.485 [1.142; 1.931]	1.529 [1.183; 1.976]	0.290
		Model 2	1	1.204 [0.954; 1.518]	1.334 [1.042; 1.708]	1	1.454 [1.113; 1.898]	1.456 [1.122; 1.890]	0.629
		Model 3	1	1.104 [0.864; 1.411]	1.292 [0.994; 1.679]	1	1.530 [1.162; 2.017]	1.488 [1.135; 1.952]	0.586
		Model 4	1	1.103 [0.862; 1.411]	1.278 [0.980; 1.667]	1	1.560 [1.182; 2.058]	1.510 [1.148; 1.987]	0.574
	Cadmium	Model 1	1	0.791 [0.614; 1.018]	0.808 [0.618; 1.056]	1	0.824 [0.662; 1.102]	0.973 [0.788; 1.202]	0.284
		Model 2	1	0.804 [0.619; 1.044]	0.848 [0.642; 1.120]	1	0.672 [0.523; 0.862]	0.745 [0.584; 0.950]	0.745
		Model 3	1	0.819 [0.624; 1.075]	0.912 [0.684; 1.215]	1	0.709 [0.548; 0.917]	0.758 [0.590; 0.974]	0.559
		Model 4	1	0.930 [0.691; 1.252]	1.088 [0.793; 1.493]	1	0.722 [0.557; 0.936]	0.783 [0.607; 1.009]	0.223
FFD: Fermented and fish-enriched diet	Lead	Model 1	1	1.394 [1.058; 1.838]	2.536 [1.956; 3.288]	1	1.294 [1.026; 1.631]	1.726 [1.344; 2.215]	0.036
		Model 2	1	1.085 [0.806; 1.460]	1.813 [1.371; 2.399]	1	1.056 [0.825; 1.352]	1.303 [0.999; 1.701]	0.122
		Model 3	1	1.027 [0.755; 1.398]	1.627 [1.214; 2.180]	1	1.001 [0.776; 1.292]	1.327 [1.010; 1.743]	0.356
		Model 4	1	1.042 [0.765; 1.420]	1.613 [1.202; 2.165]	1	1.011 [0.781; 1.307]	1.361 [1.034; 1.791]	0.390
	Mercury	Model 1	1	1.515 [1.153; 1.992]	1.739 [1.345; 2.248]	1	1.174 [0.942; 1.464]	1.094 [0.850; 1.407]	0.012
		Model 2	1	1.289 [0.967; 1.728]	1.330 [1.012; 1.747]	1	1.024 [0.817; 1.285]	0.887 [0.684; 1.150]	0.009
		Model 3	1	1.272 [0.939; 1.724]	1.318 [0.984; 1.764]	1	0.993 [0.786; 1.255]	0.922 [0.707; 1.204]	0.039
		Model 4	1	1.276 [0.941; 1.730]	1.306 [0.975; 1.749]	1	0.992 [0.784; 1.254]	0.923 [0.707; 1.205]	0.043
	Cadmium	Model 1	1	1.277 [0.927; 1.759]	1.529 [1.134; 2.062]	1	1.418 [1.166; 1.723]	1.951 [1.570; 2.426]	0.196
		Model 2	1	1.042 [0.746; 1.454]	1.129 [0.825; 1.544]	1	1.102 [0.883; 1.376]	1.384 [1.083; 1.769]	0.071
		Model 3	1	0.957 [0.678; 1.351]	1.014 [0.731; 1.407]	1	1.074 [0.855; 1.349]	1.420 [1.103; 1.826]	0.022
		Model 4	1	1.041 [0.716; 1.513]	0.999 [0.700; 1.426]	1	1.093 [0.868; 1.376]	1.446 [1.122; 1.864]	0.027

ORs: odds ratios, Statistical significance is presented in red (p < 0.05). High heavy metal level was defined as the upper tertile group. Model 1: crude; model 2: adjusted for age, sex, and body mass index; model 3: same adjustments as model 2 plus region and occupation; model 4: same adjustments as model 3 plus smoking status and physical activity.

shows the odds ratios for high heavy metal levels according to tertile of dietary pattern scores by sex and p for interactions between sex and dietary pattern scores. The effect of dietary patterns on heavy metal concentrations in blood showed significant differences between the sexes. The association between MPD patterns and Pb-B was only significant for male subjects. In male, high VMD scores were significantly negatively associated with high levels of Pb-B, while in women there were positively associated, but did not show a significant association (all p for interaction <0.005).

4. Discussion

In this study, we investigated the associations between three heavy metal levels and dietary patterns in the Korean population. Our results indicate that the MPD pattern was significantly associated with increased Pb-B and Hg-B concentrations, and the FFD pattern was significantly associated with increased Pb-B and Cd-B concentrations. Interestingly, the VMD pattern was associated with lower concentrations of Pb-B and Cd-B. These results are consistent with previous studies conducted in Korea [20,23,37,44] and in other populations [45,46,47,48,49,50]. A prior study investigating the association between Korean dietary patterns and heavy metals identified the ‘balanced diet pattern’, ‘grain and kimchi pattern’, and ‘alcohol and noodle pattern’. Each of these patterns shared major food groups with our VMD, FFD, and MPD, but some food components were considered differently. The groups associated with meat and fish consumption, which were considered the main food groups in a balanced diet pattern, along with vegetables and fruits, had similar consumption patterns to those associated with the ‘alcohol and noodle’ and ‘grain and kimchi (fermented vegetables)’ groups, respectively, in our study. This difference may be caused by the rapid change in dietary patterns in Korea over the past decade due to industrial development and the westernization of lifestyles [51,52]. Additionally, the consumption of cereals, represented by rice, which has been the staple food of Koreans, and vegetables and fruits, have decreased, while meat consumption has increased dramatically, and seafood consumption has increased and then decreased [52]. The changes in the consumption of these foods, which are a major source of heavy metal exposure, may explain the declining trend in blood heavy metals in Koreans (Figure 3). The difference between our results and those of previous studies may also be explained by the fact that, unlike previous studies that used only the frequency of food intake, this study used the SQ-FFQ to reflect the average intake over a year. More importantly, the present study identified practical and current dietary risk factors for heavy metals, such as processed foods and fermented vegetables, which have become increasingly problematic in the Asian population.

To further evaluate the potential health risks associated with heavy metals detected in the blood, we reviewed the Health-based Guidance Values (HbGVs) established by the Joint FAO/WHO Expert Committee on Food Additives (JECFA). Although these are not the main findings of our study, they are presented as reference points to aid reader comprehension. The HbGVs for Pb, Cd, and Hg recommended by JECFA are as follows: 0.6 µg/kg body weight/day for Pb, associated with a loss of one IQ point in children; a provisional tolerable monthly intake (PTMI) of 25 µg/kg bw/month for Cd; a provisional tolerable weekly intake (PTWI) of 4.0 µg/kg bw/week for inorganic Hg, based on kidney weight changes [53].

Additionally, the international Codex standards for food safety set maximum levels for certain heavy metals in specific foods. For Pb, the limits include 0.2 mg/kg in polished rice, 0.1 mg/kg in potatoes, apples, and sweet potatoes, 0.2 mg/kg in bovine offal, and 0.1 mg/kg in poultry offal. For Cd, the limits are 0.4 mg/kg in polished rice, 2.0 mg/kg in squid, 0.1 mg/kg in brown rice, 0.2 mg/kg in spinach, and 2.0 mg/kg in 100% cocoa powder. Methylmercury is regulated at 1.0 mg/kg in large predatory fish and other deep-sea fish. However, international standards have not yet been established for many foods [53].

The association with dietary patterns varied according to the type of heavy metal. Pb is absorbed from the gastrointestinal tract into the blood, soft tissues, and bones; most of it is excreted through urine, but its biological half-life is approximately 27 years [54] and high blood Pb concentrations are known to be associated with an increased risk of death [55]. In the present study, increased MPD and FFD score was associated with a high Pb-B concentration, and increased VMD score was negatively associated with Pb-B concentration. Previous data in Korea also showed that high fast food consumption, low milk and tofu consumption, and alcohol and noodle consumption had the strongest associations with Pb-B [20,23]. According to the Korean Ministry of Food and Drug Safety’s heavy metal re-evaluation report, the food that makes the highest contribution to Pb levels is grains (16.8%) [53]. Their contribution to MPD is 47.1% and FFD is 31.4%. These results are consistent with those of previous studies on other populations [47,48,49]. A possible explanation for the increase in the concentration of Pb-B is that some nutrients, such as calcium and vitamin C, can interfere with the absorption of Pb. For example, Pb absorption may be reduced when calcium levels are sufficient because Pb competes with other heavy metals for the same gastrointestinal transporters [56,57]. A study targeting children reported that for every 100 mg increase in calcium, Pb-B concentration decreased by 3.8% [58]. Vitamin C (ascorbic acid) is an antioxidant that reduces oxygen-, nitrogen-, and sulfur-centered radicals and acts as a chelating agent to reduce Pb concentration in the blood [59,60]. Therefore, the U.S. Centers for Disease Control and Prevention recommend that vitamin C is used for the nutritional management of children with high Pb and B levels [60]. A recent review reported that a mixture of garlic and vitamins E and C can increase collagen formation, bone density, and excretion, thereby lowering the concentration of Pb in the body [61]. Previous studies have also shown that sex affects Pb concentration. Male respondents are known to have higher Pb-B levels than female respondents; moreover, female respondents are more vulnerable to Pb immunotoxicity, while male respondents are more vulnerable to neurotoxicity [21,62]. In the current study, the association between Pb-B concentration and dietary pattern showed a significant interaction effect with sex. Therefore, to reduce Pb-B in the body, nutrients with evidence of a chelating effect must be utilized in the daily diet considering sex differences.

Cd enters the human body via food in the general population and accumulates in all tissues and organs, particularly the liver and kidneys, with a biological half-life of >26 years [63,64,65,66]. The biomarker for Cd is the concentration of Cd in whole Cd-B and urine Cd; Cd-B reflects short-term exposure over 2–3 months and urine Cd reflects the long-term storage of Cd. In the present study, an increased FFD score was associated with increased Cd-B concentrations, and this result is similar to that of previous studies [37,44]. The FFD included grains (white rice), fermented foods (kimchi and soybean paste), and sea products (fish, mollusks, and shellfish), which are representative Korean foods. According to the Korean Ministry of Food and Drug Safety’s heavy metal re-evaluation report, foods contaminated with Cd include grains, seafood, and fermented foods, accounting for 45.2% of the total food exposure [53]. In the current study, an increased VMD score was associated with decreased Cd-B levels, which is consistent with the results of previous studies [37,45,46,50]. As dietary fiber from green vegetables passes through the intestines quickly, it may reduce the absorption of Cd due to changes in the intestinal microbial community [50]. In practice, a British cohort study conducted on pregnant women confirmed that when green leafy vegetables were consumed four or more times per week, Cd-B concentration decreased by 72% [50]. Calcium, which is present in large quantities in milk and dairy products, inhibits Cd’s low competition for absorption binding sites; as a result, a high-calcium diet lowers the concentration of Cd-B [50,67]. Studies conducted on women of childbearing age in Asian countries reported that a dairy diet and high-calcium vitamin diet lowered Cd-B levels by 14% [46]. A study of pregnant women in Canada also reported an association between increased calcium and vitamin D diets and decreased Cd-B [45]. Although the pattern of dietary fiber and calcium in VMD seems to have a Cd-chelation effect, further follow-up studies with more elaborate designs are needed. Sex differences were noted in Cd concentration. Cd retention rates are generally higher in women than in men, and this trend was similar to that of the VMD pattern in the current study [21,62].

Hg exists in elemental, inorganic, and organic forms [68]. The biological half-life of Hg is approximately 60 days. Biomarkers include Hg-B, which indicates short-term exposure, and urine Hg, which indicates long-term exposure [68,69]. In the present study, MPD and VMD scores were significantly associated with increased Hg-B. Although there is relatively little research on the association among MPD, VMD, and Hg, a relationship between alcohol consumption in MPD and Hg-B has been reported [70]. As alcohol increases the permeability of cell membranes, it can increase Pb absorption in the gastrointestinal tract and its excretion from the bone, thereby increasing Pb-B [20,71]. In previous studies, current alcohol drinkers had higher Hg-B levels than past alcohol drinkers and non-drinkers [70]. However, in the current study, FFD intake was not significantly associated with an increase in Hg-B, which is contrary to the results of previous studies. Many studies in Korea and other populations have reported an association between fish intake and Hg-B levels [20,21,23,28,70,72,73]. This difference is believed to result from the differing food groupings among studies; the FFD group in the present study included only fish, without shellfish. However, shellfish in Korea are highly contaminated with heavy metals [53]. The FFD group included fermented foods, such as kimchi and soybean paste, in addition to fish. Fermented foods are produced by the advanced conversion of food ingredients through controlled microbial growth and enzyme action, with potential probiotic effects, and can lead to the production of biogenic amines and bioactive peptides through fermentation, increasing the biological activity of phenolic compounds and their anti-inflammatory properties [74]. Considering these benefits, fermented foods may reduce Hg absorption from seafood. Although fish and seafood are the main sources of methylmercury, their health benefits should also be considered. Fish are a source of high-quality proteins and omega-3 polyunsaturated fatty acids (PUFA), such as docosahexaenoic acid and eicosapentaenoic acid, which cannot be efficiently synthesized in the body [70,75]. In particular, there is scientific evidence of an association between PUFA and cardiovascular disease, Alzheimer’s disease, and fetal neurological development [76,77,78]. Additionally, processed and fermented foods contain a large amount of sodium, which is known to be highly correlated with obesity [79,80,81], consistent with the current study, which showed that the MPD and FFD groups had positive associations with health indicators such as WC, triglycerides, and BMI. Therefore, although foods with high concentrations of Hg should be avoided, the additional health considerations of certain foods should be considered when determining the risk of heavy metal exposure.

This study had some limitations. We clustered various types of foods into three categories through factor analysis and were unable to examine the associations between individual foods and heavy metal concentrations in the blood. Recently, owing to the increase in the detection and risk of Cd, the European Commission and Codex Alimentarius have enforced maximum allowable Cd concentrations in chocolate and cocoa powders [82,83]. The Korea Ministry of Food and Drug Safety (KMFD) also concluded that standards need to be set for cocoa powder and chocolates, considering that most processed cocoa products are imported and that they are popular among sensitive population groups [53]. In the present study, cacao and chocolate were not included in the SQ-FFQ, as they were not considered major food groups; however, this distinction should be considered in future studies. In addition, because the SQ-FFQ relies on participants’ memory and fixed food categories, it may be prone to recall bias and misclassification. Considering that heavy metal exposure can vary significantly depending on food preparation methods and the specific parts of the food consumed, this approach may introduce measurement uncertainty. To enhance the accuracy of dietary data and minimize uncertainty, future research should consider the use of alternative dietary assessment tools, such as 24 h dietary recall methods. On the other hand, the use of dietary patterns derived from major food groups may still offer value, as it provides a broader perspective on the relationship between diet and blood heavy metal concentrations, accounting for the potential interactions between foods and their role in metal absorption and the metabolism.

Second, we presented only a cross-sectional study on the relationship between heavy metal blood concentrations and dietary patterns in adults because we analyzed data from the National Health and Nutrition Examination Survey conducted on the general population in Korea. As this study employs a cross-sectional observational design, it does not allow for causal inference. However, some of the conclusions imply causality without adequately addressing potential confounding factors. The authors should explicitly acknowledge the limitations of a cross-sectional approach and discuss the possibility of reverse causality or residual confounding.

Moreover, infants, young children, and pregnant women may have different dietary patterns than the general adult population, and may experience more severe and long-term health effects from heavy metals. In fact, long-term exposure to heavy metals can affect fetal health, as Pb accumulated in the bones may enter the bloodstream owing to metabolic changes during pregnancy [84]. In the future, it will be necessary to conduct long-term research targeting environmentally vulnerable groups such as pregnant women and infants. Despite these limitations, this study is significant in that it identified the dietary groups contributing to heavy metal concentrations in adults based on representative population data in Korea.

5. Conclusions

This study examined the associations between the dietary patterns and blood concentrations of heavy metals in a nationally representative sample of Korean adults. Through principal component analysis, three distinct dietary patterns were identified, each showing different relationships with internal heavy metal levels. Higher adherence to meat and processed food-based diets, as well as fermented and fish-enriched diets, was associated with increased levels of Pb, Hg, and Cd in the blood. In contrast, an increased consumption of vegetables and dairy products was linked to lower levels of Pb and Cd, suggesting a potential protective role. Notably, the associations varied by sex, indicating that dietary influences on heavy metal exposure may differ between men and women. These findings contribute novel insights into how common dietary behaviors can influence chronic exposure to environmental toxicants. By utilizing multi-year data from the Korea National Health and Nutrition Examination Survey, this study reflects temporal trends and ensures broad population representativeness. The results align with the international evidence and support the relevance of dietary patterns as modifiable risk factors for internal contamination.

From a public health perspective, our findings underscore the need for dietary recommendations that emphasize plant-based and dairy-rich eating habits while discouraging the excessive intake of certain animal-based and fermented foods, particularly in populations vulnerable to environmental exposure. These data may serve as a valuable reference for the development of national nutritional guidelines and inform policies aimed at mitigating heavy metal exposure through diet. Future longitudinal and mechanistic studies are needed to confirm causal pathways and to refine intervention strategies tailored to population-specific dietary practices and environmental conditions.