Distribution of metals and metalloids in dried seaweeds and health risk to population in southeastern China

Concern about metals and metalloids, especially heavy metals in seaweeds has risen due to potential health risk. This study investigated the distribution of 10 metals and metalloids in 295 dried seaweeds (brown and red) and estimated the possible health risk via hazard index (HI). Elements in seaweeds can be sequenced in descending order by mean values: Al > Mn > As > Cu > Cr > Ni > Cd > Se > Pb > Hg. The levels of Cd, Cu, Mn and Ni in red seaweeds were significantly higher than those in brown seaweeds (P < 0.01). Correlation analysis showed contents of Ni-Cr (r = 0.59, P < 0.01) in seaweeds had moderate positive correlations. Seaweeds from different geographical origins had diverse element distribution. Risk assessment showed that HI at mean level was less than the threshold of 1. It indicates that for the general people there is low health risk to these elements by the intake of seaweeds. Furthermore, in terms of the confirmative toxicity of some metals, such as Cd, Pb and Hg, surveillance of metals in seaweeds should be performed continuously.


Results and Discussion
Validation of analysis method. The results for element levels of certified reference materials (CRMs) (GBW10023 and GBW08517) (n = 6) are summarized in Table 1. Quantitative results (within 10% of the certified value) were obtained for all elements in each CRM. Mean recoveries were ranged between 82.6-99.4%. For the quality control experiment, reference materials were analyzed along with samples. The relative standard deviations of reference materials between tested value and labeled were less than 10% (except for Hg, it is 15%), so we considered these data of samples be acceptable. Table 2, the levels of Cd, Cu, Mn and Ni in red seaweeds were significantly higher than those in brown seaweeds (P < 0.01). Levels of Hg and Se in red seaweeds were significantly lower than in brown seaweeds (P < 0.05). Similarly, Rubio et al. 13 pointed out red seaweeds showed a tendency to have higher concentrations of Cd than brown seaweed. For all seaweed samples, elements with mean values can be sequenced in descending order, Al > Mn > As > Cu > Cr > Ni > Cd > Se > Pb > Hg. Our results were close to report of Desideri et al. 14 with mean levels of Al (736 mg/kg), As (19.3 mg/kg) and Cd (1.8 mg/kg) in Italy. But, Al (553.8 mg/kg) and Pb (0.595 mg/kg) in this study were higher than studies in Spain with Al (<50 mg/kg) and Pb (<0.2 mg/kg) 13,15 . Cu (6.135 mg/kg) was similar with in Chile with Cu (7.46 mg/kg) 16 . Pb concentration (0.595 mg/kg) was lower than in India and Algeria 17,18 . The difference of these elements content in countries may be caused by the diverse growing environment, as well as the seaweed species. Previous study revealed that eight genera of seaweed had significant differences (P < 0.05) in levels of most of trace elements 19 . Furthermore, the data of each metal and metalloid level have large standard deviations. It may be caused by the samples collected from different origins.

Levels of metals and metalloids in seaweeds. As shown in
The present study only tested the levels of total arsenic. Usually organic arsenics, for example arsenosugars, have less toxicity than inorganic arsenics. Seaweeds contain arsenic primarily in the form of arsenosugars, which can be metabolized to a wide range of arsenic compounds. The metabolite, such as thio-dimethyl arsenic    (Table 3). It was observed that no strong correlation (r > 0.6) between the contents of studied elements. Weak correlation was found for Ni-Cr (r = 0.59, P < 0.01), Mn-Cu (r = 0.29, P < 0.01) and Mn-Pb (r = 0.22, P < 0.01). The content of Al-Cd, Cd-Cr, Cr-Cu, Hg-Mn, Ni-Pb and Al-Se showed very weak correlation (0.1 < r < 0.2, P < 0.05).
Seaweeds from different geographical origins. The levels of elements in brown and red seaweeds from different origin were showed in Table 4. The results for combination of above brown and red seaweeds were also calculated. For brown seaweeds, levels of Al, Cd, Cu, Hg, Mn and Se existed significant differences in different origin areas (P < 0.05). For red seaweeds, elements with obvious difference in areas were Al, As, Cu, Mn and Pb. When the data of red and brown seaweeds were combined, the Al, Hg and Pb showed significant difference in terms of origin areas.
Investigators such as Larrea-Marı'n et al. 20 and Akcali et al. 21 have emphasized the importance of the culture region to changes of elemental levels in seaweeds. To identify the geographical origin of seaweeds, we performed linear discriminant analysis (LDA) with stepwise procedure. LDA can maximize the variance between groups and minimize the variance within the group by creating new variables Canonical plot of linear discriminant analysis were showed in Fig. 2. Samples from different origin focus in different labeled circles. The size of the circle corresponds to a 95% confidence limit for the mean. Groups that are significantly different tend to have non-intersecting circles. " + " sign is a marker for the center of the circle. On the whole, 67.4% seaweed samples were correctly classified. The origin area 1 had the best performance with only 6.3% samples misclassified. The discrimination functions were listed as bellow. These functions can be used for predicting the origins of unknown seaweeds. So when we want to predict the place of a sample, we just calculate the value by the function equation and observe the distribution in circles. Table 5, the exposure doses of elements were calculated at the mean and P97.5 level. For As, we adopt 5% of value of total As to calculated the exposure, because in seaweed about 90% total As were organic arsenic, especial arsenosugars which shown very little toxicity 5 . The targeted hazard quotient (THQ) for single element was less than 1. Hazard index (HI) was 0.49 for mean exposure and 1.68 for high exposure. It indicates that there is low health risk for toxic elements by intake of seaweeds. However, in terms of the high exposure level (P97.5), they may be the potential adverse effect on human health. Several studies also claimed that total elemental intake does not appear to pose any threat to the consumers in Italy, South Korea and Spain 13,14,22 , though different safety reference values, such recommended reference dose (RfD) from US EPA and provisional tolerable weekly intake (PTWI) from the Joint FAO/WHO Expert Committee were used for assessment.

Exposure assessment. As shown in
For calculating exposure dose, we adopted 5.2 g/capita/day as the consumption of seaweeds in China, which is higher than in Japan for 4 g/adult/day 23 and lower than in South Korea for 8.5 g/adult/day 22 . In addition, it should be noted that the risk assessment on arsenic was not performed because the arsenic species were not analyzed here. The toxicity of organic arsenic is lower than inorganic species 24 . The study about organic arsenic exposure including arsenolipids and arsenosugars was recently reported by Taylor et al. 6 .
Uncertainty of the risk assessment. It should be noted that HI used for risk assessment was based on concentration/dose addition of different metals and metalloid. The addition of each value is a non-interactive process,  Table 2. The levels of analyzed elements in red and brown seaweeds. "*" and "**" mean that theelemental content between red seaweed and brown seaweed is significantly different at the level of P < 0.05 and P < 0.01, respectively.
SCientifiC REPORtS | (2018) 8:3578 | DOI:10.1038/s41598-018-21732-z which means that metals and metalloid in the mixture do not affect their toxicity. Actually, multiple metals in human body may have the interactive effect. Furthermore, loss of elements caused by the cooking and absorption ratio of human intestine was not considered here. So, the above factors may lead to the uncertainty of the estimates. Usually, the average level of targeted chemicals was adopted for dietary exposure assess. For avoiding the underestimation of the intake, we used both mean and P97.5 value of element concentration for the exposure assessment.

Conclusion
The present study revealed the different levels of Al, As, Cd, Cr, Cu, Hg, Mn, Ni, Pb and Se in dried seaweeds from southeastern China (Zhejiang province). It indicates that element concentration changes with different species of seaweeds and origin areas. For example, the levels of Cd, Cu, Mn and Ni in red seaweeds (Phorphyra) were significantly higher than those in brown seaweeds (Laminaria and Undaria) (P < 0.01). The estimate of health risk   Table 4. The levels of elements in seaweed from different origins. a,b,c,A,B,C,x,y and z mean data with upper letters where do not share one same alphabet shows that there are significant differences among areas by Oneway ANOVA and Tukey HSD multiple comparisons (P<0.05).   showed that there was low health risk for potential toxic elements by intake of these seaweeds. However, in terms of the hazard of some metals to human body, continuous surveillance in edible seaweeds is necessary for health protection to consumers 25 . Furthermore, for Asian countries, regulations on maximal concentration of heavy metals in seaweed products should be set up.

Materials and Methods
Sampling. Dried seaweeds were collected in Zhejiang, China, where the latitudes range from 27°09′ to 31°11′N, and the longitudes from 118°02′ to 122°57′E from April to November, 2016. Four places of seaweed origin in Zhejiang coast were marked as 1, 2, 3 and 4 in Fig. 1. The simple map (Fig. 1.)  Health risk assessment. The targeted hazard quotient (THQ) and hazard index (HI) were used toestimate health risk according to US EPA's IRIS database 29 . We used the mean and 97.5th percentile of obtained elements concentration to represent the consumers with average and high exposure, respectively 30 . The sum of all THQs for each element is referred to as the HI. The formulas are as follows: Ci is the average or P97.5 concentration of the element in the seaweeds (mg/kg); Di is the daily intake of seaweeds (5.2 g/capita/day); Ed is the average exposure duration (e.g., 70 years); Bw is the average weight (e.g., 70 kg); At is the average lifetime (e.g., 70 years). RfD is the recommended reference dose (RfD); According to US EPA guidelines for assessing conservative risk, HI were calculated by sum of the THQ. When HI < 1, no health risk is expected to occur; If HI ≥ 1, there is moderate or high risk for adverse human effects.
Statistics. One-way ANOVA and Tukey HSD for comparing means was used for observe differences between different seaweeds or different origins. Pearson correlation was used to analyze linear relationships among different elemental concentrations. Linear discriminant analysis (LDA) was used for identify seaweeds from different growing regions by quantified elements. All these statistical methods were performed on software of JMP 10.00 free-try edition (SAS Institute Inc.). In addition, data used for exposure estimates were according to the recommendation of the report Reliable Evaluation of Low-Level Contaminations of Food issued by WHO 31 . Thus, a value of 1 / 2 LOD was assigned to all results below the LOD, where the proportion of <LOD results is not > 60%.