**4. Discussion**

This study used the second cycle of the KoNEHS (2012–2014) to examine the obesogenic properties of blood Hg as it relates to hyperlipidemia and elevated liver enzymes. The GM of blood Hg was high, up to 3.11 μg/L, and 57.3% of the survey population had hyperlipidemia. For participants aged 40 and above, 55–66% had hyperlipidemia, whereas 38% of the participants in their 20s and 50 % of the participants in their 30s had hyperlipidemia. The mean BMI of the hyperlipidemia group was 25.0 compared to 23.3 in the non-hyperlipidemia group. Moreover, the mean BMI was 24.0 in the normal group and 25.7 in the elevated liver enzymes group.

Approximately 32.1% of males (*n* = 889) and 15.4 % of females (*n* = 566) had blood Hg levels over 5.00 μg/L, which is the acceptable level for no adverse effects (HBM-I) [22]. These results are consistent with the first cycle of the KoNEHS (2009–2011), where 33.4% of males and 16.1% of females exceeded the HBM-I [10]. Many studies have investigated the high blood Hg levels in the Korean population. Some reported significant associations between Hg, high BMI, and metabolic syndromes [23,24], while others reported no associations or even negative associations [25–27]. Metabolic syndromes are associated with many factors, including dietary habits and living patterns, and some factors may have a stronger influence than Hg exposure. For example, alcohol consumption is a major risk factor for metabolic syndromes, and raw-fish and clam soup are popular menu items that are often consumed with alcohol in Korea. Thus, blood Hg is also significantly associated with drinking alcohol, and the frequency of alcohol intake should be accounted for when evaluating the relationship between Hg exposure and metabolic syndromes in Korea. In our study, 62.3% of males and 20.9% of females drank alcohol more than once a week. Therefore, we also included the interaction of alcohol consumption frequency and sex (*p* < 0.0001 for hyperlipidemia) when examining the obesogenic effects of Hg.

It is also possible that individual treatments for obesity attenuate the effects of Hg exposure. Those diagnosed with metabolic syndromes may actively control their lipid profiles and insulin resistance. Therefore, treatments including personal medications, could affect the associated markers and the diagnoses of metabolic syndromes. In the Table 5, we showed only the ORs of blood Hg after each medication. And the same analyses provided that taking personal medications for hyperlipidemia reduced the odds of hyperlipidemia by 29% (OR: 0.710, 95% CI: 0.559, 0.902), and taking personal medications for hyperlipidemia, hypertension, and diabetes was associated with 59% lower odds of hyperlipidemia (OR: 0.410, 95% CI: 0.029, 0.733). Nonetheless, Hg significantly affected to the odds of having hyperlipidemia.

Among the different forms of Hg, alkyl Hg is more lipid soluble and passes readily through biological membranes [28]. Especially, methylmercury (MeHg) among the alkyl Hg, is the dominant form in human blood [29] because the primary exposure source for the general population is fish consumption. MeHg exposure inhibits paraoxonase-1, which prevents the atherosclerotic process by metabolizing toxic oxidized lipids associated with LDL and HDL [30]. Therefore, Hg induces oxidative stress and disrupts gluconeogenesis, resulting in systemic inflammation that affects the accumulation of abnormal adipocytes [23,31]. Our results showed that the levels of blood Hg were significantly higher (*p* <.0001) in the hyperlipidemia group (male: 4.03 μg/L, female: 2.83 μg/L) than in the non-hyperlipidemia group (male: 3.48 μg/L, female: 2.69 μg/L), and that an increase of 1 μg/<sup>L</sup> blood Hg was associated with an 11% increase in the odds of hyperlipidemia, even after adjustment for personal medications.

Though bile is the major route of excretion, Hg can be reabsorbed into the blood via the enterohepatic system [12,32]. In particular, methylated Hg makes up most of the mercury in humans and can easily bind to cysteine residues [33], such as glutathione, and penetrate the cellular membranes [34]. The MeHg-cysteine complex can then enter the bile tract and be hydrolyzed by GGT and other dipeptides [13,33]. As a result, Hg induces hydrogen peroxide, depletes glutathione, and increases GGT levels. The association between Hg exposure and GGT, a marker of oxidative stress, is supported by several animal and human studies [35–37]. Our study also showed that the levels of blood Hg were significantly higher (*p* < 0.0001) in the elevated liver enzymes group (male: 4.36 μg/L, female: 3.25 μg/L) compared to the normal group (male: 3.64 μg/L, female: 2.70 μg/L). After adjustment for personal medications, blood Hg was associated with 35% higher odds of elevated liver enzymes.

This study has several limitations and strengths. The study design is a cross-sectional survey, and each measurement was analyzed from an individual spot sample. However, the dataset is representative of the entire Korean population. According to the previous studies, the intra-class correlation for blood Hg was 0.67~0.71 [38]. Moreover, diet is the major source of Hg, so we anticipate that the individual blood Hg levels would be constant. Secondly, individual health status or medical history data were unavailable. Instead, personal medication data were adjusted for the obesogenic effects of Hg. We also performed correlation analyses between the lipid profiles and hepatic enzymes to avoid overestimating the Hg effects. There were no associations between the lipid profiles and hepatic enzymes, nor were there any differences across blood Hg groups (Figure 2). This indicates that the effects of blood Hg on the lipid profiles were irrelevant to the hepatic enzymes and that the hepatic enzymes were not affected by the lipid profiles. Thus, correlation analyses demonstrate the significant effects of blood Hg on hyperlipidemia and elevated liver enzymes.
