Fish consumption is relatively high and has been increasing steadily in the Asian population [^1-3]. Fish intake is a major source of mercury (Hg) exposure in humans. Hg is a heavy metal that is released into the air from coal-fired power plants and other industrial processes, and bioaccumulates in the marine food chain, binding to tissue proteins depending on body size, lifespan, or levels of environmental contamination.

The potential independent correlates of Hg have not been well-characterized in Korea. There may be differences in human Hg concentrations depending on socio-economic level [⁴], residential area [^4-6], traditional culture or ethnicity [⁴,^7-9], lifestyle characteristics [¹⁰], and the types of fish and seafood typically consumed [⁶,⁹]. Prior Korean Hg studies have generally focused on the acute toxic effects of Hg [^11-13] and have been conducted in geographically limited areas [^14-16]. Others have simply presented the average levels in subjects without considering the types of seafood or geographic variations [^17-19].

Understanding the independent dietary and non-dietary determinants of Hg status is important for elucidating the biologic mechanisms related to chronic exposure to Hg and for identifying potential interventions that might improve the health benefits of fish consumption. The present study examined both dietary and non-dietary correlates of Hg status in a cross-sectional sample of 3,397 Korean men and women aged 20–87 years who participated in the Korea National Health and Nutrition Examination Survey (KNHANES IV). Generally, Korean adults frequently consume fish in greater quantities than do Western populations [^1-3]. Thus, we identified the types of fish most frequently consumed and associated with higher levels of whole blood mercury concentration (WBHg), and investigated the independent demographic, lifestyle, and geographic determinants of WBHg levels.

Methyl mercury, mostly from fish intake, may accumulate over time in human muscle tissue, so is expected to be positively correlated with age. Region-specific dietary habits may also be related to Hg levels, such as serving whale/shark meat in dishes for a traditional ceremony or locally caught types of fish. In addition, there is now widespread concern over red meat consumption among middle aged Korean population, who tend to reject red meat in favor of other alternatives such as fish, particularly sashimi for alcohol dishes. Considering these characteristics, we hypothesized that Hg levels are closely associated with higher intakes of small marine fish as well as large predator species in the population with high and varied fish consumption, and some geographic variation can be observed, controlling diet.

Table 1 gives demographic characteristics of the study participants sorted by sex (1,994 men, 1,978 women). The median concentration of WBHg was 5.1 μg/L in men and 3.7 μg/L in women. Distributions of age, income level, and residential area did not differ by sex. The prevalence of overweight/obesity and smoking was significantly greater in men than in women. In addition, men were more likely to frequently consume alcohol and eat out and to have higher levels of physical activity (MET-h/week) and education compared to women.

Table 2 summarizes Hg levels by quartile and according to demographic characteristics. The highest levels of WBHg were recorded in individuals in the 40–59-year-old age group. Higher levels of WBHg were associated with overweight/obesity, current smoking, frequent alcohol drinking and eating out, higher levels of income and education, and residence in the mideast or southeast/island areas of South Korea.

Table 3 gives the most frequently consumed species of fish listed according to the levels of WBHg in the present study sample. Overall, higher intakes of white fish were associated with higher levels of WBHg. In particular, consumption of both yellow corvina (p < 0.01) and halibut (p < 0.001) were significantly associated with higher levels of WBHg and showed dose–response relationships. Intake of most oily fish types showed positive associations with levels of WBHg. Among oily fish, anchovy and mackerel were most frequently consumed and also showed significant positive associations with levels of WBHg (p = 0.01 and p < 0.01, respectively). Consumption of fresh tuna was significantly associated with higher levels of WBHg (p < 0.001) whereas consumption of canned tuna was not (p = 0.4). In Table 3, the values may have high within person variance due to single 24 hour recall, rather than multiple measurements.

Table 4 provides comparisons of levels of fish and shellfish consumption for weekly (FFQ) and daily (24-hour recall) intake and according to WBHg levels. Higher levels of fish and shellfish intake were associated with higher levels of WBHg, according to both consumption measures. These associations were particularly evident in the consumption of white fish, oily fish, and shellfish (p < 0.001).

To understand whether lifestyle and environmental factors were associated with WBHg after accounting for fish and shellfish consumption, we conducted multivariable-adjusted linear regression analyses for each of the identified dietary and non-dietary factors (Table 5). Both age and gender were independent correlates of WBHg levels, even after adjusting for diet as well as other lifestyle and environmental factors. WBHg levels ranged from 11.1–12.9% higher in 40–59-year-olds than in 20–29-year-olds (p < 0.001); WBHg levels were also 9.8% higher in men than in women (p < 0.001). Several lifestyle factors were also independently associated with WBHg levels. Higher BMI was associated with higher WBHg levels, with 1.0% higher levels for each additional 1 kg/m² (p < 0.001). Higher alcohol consumption was associated with higher WBHg levels with a dose–response relationship showing that, when compared to non-drinkers, levels were 9.4–9.7% higher in drinkers who had alcohol intake at least twice per week (p < 0.001). Income and education levels were also positively associated with WBHg levels, such that high-income, highly-educated participants were more likely to have higher levels of WBHg. Compared with participants in the lowest quartile of household income, those in the highest quartile had 8.1% higher WBHg levels (p < 0.001). In addition, participants with college/university education or higher had 3.7% higher WBHg levels compared to those with middle school education or less (p <0.01). Even after adjustment for each of these lifestyle factors and types of fish and shellfish consumption, we identified significant geographic variation in WBHg levels. Compared to participants living in the midwest regions of South Korea (the lowest WBHg region), those living in the southeast/island, mideast, southwest, and north, had significantly higher WBHg levels (southeast/island: 17.5% higher, p < 0.001; mideast: 16.5% higher, p < 0.001; southwest: 8.7% higher, p < 0.001; north: 5.5% higher, p < 0.001).To further examine regional differences, we calculated adjusted means of WBHg levels for each province in South Korea (Figure 1). Overall, participants in the southeast/island region appeared to have higher levels of WBHg than those in northwestern regions. The highest levels (adjusted means) of WBHg were found in Gyeongsangbuk-do (7.6 μg/L), Jeju (7.3 μg/L), Ulsan (7.1 μg/L), and Daegu (7.0 μg/L). Regions with the lowest levels (adjusted means) of WBHg were Daejeon 4.4 μg/L, Chungcheongbuk-do 4.4 μg/L, and Chungcheongnam-do 4.7 μg/L. None of these regional findings differed by age or sex.

In this cross-sectional analysis of the general Korean population, several independent correlates of WBHg levels were identified. As expected, participants with higher intakes of fish and shellfish had significantly higher levels of WBHg and the pattern of association was a dose–response relationship. In addition, independent effects of other demographic and lifestyle factors, such as gender, age, obesity, alcohol consumption, and income and education levels were identified, accounting for the levels of fish and shellfish consumption.

Park and Lee have previously examined the association between WBHg and seafood intake levels, using 9 items from the non-quantitative FFQ of KNHANES [²⁵]. We supplemented the non-quantitative and less detailed information from the FFQ with quantitative and detailed information from the single 24 h dietary recall, thus more fully using available KNHANES dietary information, to estimate the usual amounts of fish and seafood consumption and to add detail to the types of fish and seafood that are commonly consumed in this population. In addition, Lee et al. suggested that residents in coastal areas have relatively high levels of WBHg due to higher consumption of seafood [⁶]. However, our study found a significant geographic variation in WBHg levels among coastal participants even after adjusting for multiple factors.

Our study highlighted some important findings concerning Hg levels in large and small marine fish species and their relationship to Hg levels in humans. Previously, public health concern has been focused on exposure to Hg due to consumption of large predator species such as shark, swordfish, king mackerel, or tilefish [²⁶,²⁷]. However, in the present study, the most frequently consumed small marine species such as mackerel were associated with substantially higher WBHg levels in the Korean population. Thus, it is important to improve understanding of chronic exposure to Hg from frequently consumed small marine species. This is especially critical for populations such as Koreans, where there is high consumption of a variety of fish (of both small and large species) and shellfish. Given Korean meal patterns, correlation is expected between large and small fish intake. However, we observed higher WBHg in a multivariable-adjusted analysis, including simultaneous adjustment for all fish items asked in FFQ. Tuna, one of the most frequently consumed fish in Korea and worldwide, is known to be a major source of Hg [⁹,²²]. Interestingly, our study showed that higher consumption of canned tuna was not associated with elevated levels of WBHg, whereas consumption of fresh tuna, such as sashimi or sushi, was associated with significantly higher levels of WBHg. The reason is unknown, but this may be related to the types of tuna since the majority of canned tuna available in the Korean market is “light-style” tuna consisting mainly of skipjack (Katsuwonus pelamis), whereas the bluefin tuna is commonly used for sashimi or sushi [²⁸,²⁹]. Our study results are consistent with previous research indicating that bluefin tuna has higher concentrations of Hg than does canned light tuna [²²].

Several demographic and lifestyle determinants of WBHg were identified. Men had higher WBHg concentrations than women, consistent with previous findings [³⁰]. In addition, alcohol consumption, income, and education levels were significant correlates of WBHg concentrations. The association of WBHg with alcohol is grouped here with demographic variables because many Koreans have dinner/drink meetings after work. In the face of concern about meat eating, this behavior may link alcohol and fish/sashimi consumption.

In our study, WBHg levels were positively associated with BMI after controlling for other variables. While recent study findings argue that the benefits of fish consumption may outweigh the negative effects of chronic exposure to Hg [²⁶,³¹], particularly for cardiovascular health outcomes, the association between Hg exposure and overweight/obesity is not well established. Indeed, our findings contradict those of previous studies. In a study with a Brazilian sample [³²], Hg exposure from fish consumption was not associated with BMI. Further, a study of the Korean elderly population residing in coastal areas showed that the waist-hip ratio was significantly correlated with blood Hg concentration, but BMI was not [¹⁹]. Thus, the mechanism of association between Hg exposure and overweight/obesity is relatively understudied and further research is warranted.

The present study found significant geographic variation in WBHg levels. South Korea is surrounded by ocean water with a long coast line on three sides. Thus, seafood is easily accessible from almost any region. Indeed, residents in a few regions, mostly located in the center of the Korean peninsula, had relatively low levels of WBHg when compared to regions located closer to the coast line. However, even after adjustment for multiple factors including fish and shellfish intake levels, significant geographic variation remained in WBHg levels. This may be due to other uncontrolled geographic characteristics, such as the type of marine species consumed, water or soil content, local ceremonial traditions or specific dietary habits (e.g., serving dishes containing whale and/or shark meat).

Our study analyzed data from the KNHANES containing data representative of random samples of the South Korean population and therefore may be generalizable to the larger population of Korean adults. In addition, we provide multivariable-adjusted estimates of WBHg, thus minimizing residual confounding while evaluating multiple potential determinants in the Korean population. Our analysis showed how each of the recorded demographic, lifestyle, geographic, and dietary factors may correlate independently with WBHg levels.

A limitation of the present study is that Hg levels were evaluated in whole blood and thus may indicate a relatively short-term exposure to Hg compared to other biomarkers such as hair or toenails. It is also possible that, although the present study evaluated multiple variables that were estimated using standardized instruments, residual confounding could still exist due to the presence of unmeasured or imperfectly measured factors. In addition, fish and seafood intake levels were evaluated using one-day 24-hour recall and the non-quantitative FFQ questionnaire, leaving open the possibility of misclassification.

In summary, in addition to fish consumption, we identified several demographic and lifestyle determinants of WBHg in the Korean population. Furthermore, Koreans tend to consume large amounts of small marine species and consumption of these species was positively associated with levels of WBHg, showing dose–response type relationships. Geographic variation was also a significant determinant and our results suggest that local specific dietary habits, water, or soil may explain regional differences in WBHg levels. As expected, fish consumption was clearly the key determinant of Hg concentration, but together our results indicate that a dietary guideline to simply avoid certain types of fish may not reduce WBHg levels in the Korean population. Although evidence has emphasized that the benefit of fish consumption outweighs the risk posed by Hg exposure, these risk-benefit effects might vary, depending on the range of Hg exposure in the population. Thus, large-scale prospective studies with detailed dietary information, especially fish and seafood items, are warranted to make a definitive conclusion.

The authors declare that they have no competing interests.

SC developed the study design, conducted the analysis, and wrote the draft of manuscript. DRJ contributed discussion, and reviewed and edited the manuscript. KP developed the study design, conducted the analysis, contributed discussion, and reviewed and edited the manuscript. All authors read and approved the final manuscript.

The pre-publication history for this paper can be accessed here:

http://www.biomedcentral.com/1471-2458/14/527/prepub

This study was supported by a Yeungnam University research grant in 2012. The authors thank Ms. Hyo-Jin Kim for her help on the data cleaning and analysis of this manuscript.