Levels and potential health risk of heavy metals in marketed vegetables in Zhejiang, China

The present study analyzed 5785 vegetables for concentrations of As, Cd, Cr, Pb, Ni and Hg, and estimated the health risk to local consumers by deterministic (point estimates) approaches. Levels of elements varied in different vegetables. Average levels of As, Cd, Cr, Ni, Hg and Pb were 0.013, 0.017, 0.057, 0.002, 0.094 and 0.034 mg/kg (fresh weight), respectively. The samples with 0.25% for Cd and 1.56% for Pb were exceeding the maximum allowable concentrations (MACs) set by the Chinese Health Ministry. No obvious regular geographical distribution for these metals in vegetables was found in areas of Zhejiang, China. The mean and 97.5 percentile levels of heavy metal and metalloid were used to present the mean and high exposure assessment. The health indices (HIs) were less than the threshold of 1 both in mean and high exposure assessment. It indicates that for the general people there is very low health risk to As, Cd, Cr, Pb, Ni and Hg by vegetable intake.

Detailed levels of metals and metalloid in vegetables. Different vegetables may accumulate different heavy metals, and the absorption ability varies in different biological species due to their diverse physiological character. The detailed levels of these metals in vegetables of 28 species can be found in Table S1. As shown in Fig. 1, Coriandrum sativum L. (coriander) contained the highest total mean levels of selected 6 heavy metals (0.550 mg/kg). In contrast, Cucumis sativus L. (cucumber) had the lowest total mean levels (0.120 mg/kg). The highest average levels of As and Pb were both found in Coriander with 0.037 and 0.114 mg/kg. The highest average levels of Cd, Cr, Ni, and Hg were in Chicorium endiva L. (romaine lettuce) with 0.041 mg/kg, Spinacia oleracea L. (spinach) with 0.142 mg/kg, Phaseolus vulgaris L. (kidney bean) with 0.331 mg/kg, and Ipomoea aquatica Forssk (swamp cabbage) with 0.0034 m/kg, respectively.
Brassica campestris L. (pakchoi) showed average levels of As and Cd with 0.012 and 0.017 mg/kg which were lower than previous study 12 . But mean level of Pb with 0.079 mg/kg in spinach was higher than that previous value of 0.056 mg/kg. It is hard to explain the differences between two studies. Some non-laboratory factors, such as reason, farming type and sampling areas may lead to the data diversity. Notwithstanding, data from large sampling number is more reliable than small one considering the variety in different vegetables. Similarly, other studies also revealed the high Pb in Spinach 19 .
The differences of the metal contents in these vegetables depend on the physical and chemical nature of the soil or water and absorption capacity of each metal by the plant 19 , which is altered by various factors like environmental and human interference, and the nature of the plant. The magnitude of heavy metal deposition on vegetable surfaces varied with morpho-physiological nature of the vegetables 20 . AI Jassir et al. 21 have shown that unwashed leafy vegetables sold on roadside of Riyadh city, Saudi Arabia had higher levels of heavy metals as compared to washed leafy vegetables. Demirezen and Aksoy 22 have reported higher concentrations of Pb, Cd and Cu in Abelmoschus esculentus collected from urban areas of Kayseri, Turkey as compared to those from rural areas. The partitioning of heavy metals is well known with accumulation of greater concentrations in the edible portions of leafy or root crops than the storage organs or fruits 23,24 .  Spatial distribution of heavy metals and metalloid. The spatial distribution of these metals and metalloid in vegetables from Zhejiang province was shown in Fig. 2. It was drawn by the software of MapGIS K9 SP2 free trial edition, and different metal and metalloid levels were marked with different color in selected areas. The different concentration ranges of metals and metalloid were classified for comparing among 11 areas. In a whole, the highest average levels of As, Cd, Cr, Pb, Ni and Hg were all found in the middle area. Furthermore, the second highest levels of As, Cr, Pb and Ni were observed in the south zones of Zhejiang. However, no obvious regular distribution for these elements in vegetables was revealed in areas of Zhejiang, China. The sources of heavy metal pollution in vegetables are diverse. It can occur due to factors including irrigation with contaminated water, the addition of fertilizers and metal-based pesticides, industrial emissions, transportation, harvesting process, storage and/or sale. The middle of Zhejiang is the main manufacturing location for metal-origin products. So, it may have higher potential for heavy metals contamination than other region in Zhejiang.
Exposure assessment. Based on the data of food consumption survey 25 , the estimated vegetable intake is 273.3 g/day per person. The average and P97.5 levels of heavy metal are respectively used to present the common and high exposure. Table 2 shows the estimated exposure to general population in vegetables from Zhejiang province and the health hazard index. The mean daily intakes of As, Cd, Cr, Hg, Pb and Ni by vegetables were 0.063, 0.083, 0.278, 0.010, 0.459 and 0.166 μ g/kg bw/day, respectively. The high exposure of As, Cd, Cr, Hg, Pb and Ni were 0.293, 0.396, 1.221, 0.040, 2.002 and 0.830 μ g/kg bw/day. Comparing with previous study of Huang et al. 12 , the mean exposure of As, Cd and Pb are higher in this study, but the high exposures of above metals were lower. It may be caused by the increased sample numbers. Larger number of samples probably leads to the lower dispersed degree of analyzed values.
Our mean exposure data of As and Pb were lower than those report in Turkey 22,26 , but little higher than in Greece and Egypt 27,28 . To estimate the health risk, the health hazard indices (HIs) were calculated by dividing daily intake of heavy metals by their reference doses. HI is usually adopted to assess the health risk of hazard materials in foods 29,30 . An HI more than 1 is considered as not safe for human health 31 . As shown in Table 2, HIs at the mean level were all less than 1. The mean exposure of intake just accounted for no more than 11% of the reference doses. It indicates that there is low health risk to common heavy metal exposure by intake of vegetables. Meanwhile, for the high exposure (P97.5 level), HIs of all tested metals were also less than 1. The HIs of Cr near to zero are the lowest, which may be ascribed to its higher tolerance dose.

Uncertainty analysis.
Deterministic approaches used here are the norm in chemical risk assessment, for instance to determine whether any risk may arise from consumption of a single food containing heavy metal. Although, using bioaccessible metals concentrations 34 to conduct human health risk evaluations is considered to be the most reliable and accurate method 35 , it is hard to perform actually. Animal experiments to quantify the bioaccessible metal levels are costly, whereas in vitro digestion models are not suitable for estimating the complicated dermal contact.
In this study, the total ingested metal levels were considered as the real absorption values and taken to estimate the health risks. Thus, the health risk assessments in this study could be somewhat overestimated. Furthermore, detailed consumption data of every tested vegetable were not obtained here. We just adopted the sum of vegetable consumption data to calculate exposure values, which may lead to the uncertainty of the estimates.
In conclusion. This study revealed the different levels of As, Cd, Cr, Ni, Hg and Pb in 5785 vegetable samples of 28 species. Comparing with the MAC of China, only 0.25% samples for Cd and 1.56% for Pb were out of range. No regular geographical distribution for these metals in vegetables was found. After the dietary exposure assessment, we conclude that that there is low health risk to As, Cd, Hg and Pb for general people in Zhejiang, China.
However, in terms of uncertainty in our assessment, further specific and detailed exposure estimates need to be performed in targeted area. In addition, the regular monitoring of heavy metals in vegetables is still necessary in this area to ensure the dietary safety.

Materials and Methods
Vegetable Consumption data. The vegetable consumption data used in this report was extracted from Chemical analysis. The concentrations of As, Cd, Cr, Pb, Ni and Hg were tested as described by Huang et al. 12 Briefly, the samples were digested as follows: a 5-10 g sample was added to a 100 mL round-bottom flask. Then, 10 mL concentrated nitric acid was mixed to the sample and heated for 220 °C. Hydrogen peroxide was periodically added with 1 ml until the digestion step was complete, i.e., a clear solution was reached. Finally, we diluted the solution up to a 50 mL in the volumetric flask with distilled water. The solution was analysis by Thermo SOLAAR model iCE3000 atomic absorption spectrometry (AAS) with a graphite furnace for Cd Cr, Pb and Ni, and hydride generation-atomic fluorescence (HG-AFS 9230, Jitian Co., Beijing, China) for As and Hg. Maximum allowable concentrations of contaminants in foods in China are based on the established limits by Chinese health ministry 36,37 .
Validation of analytical method. The accuracy of the analytical procedures was verified by analysis of appropriate certificated reference materials (CRMs) using the same digestion and analytical methods. Two CRMs (    38 . Thus, a value of 1 / 2 LOD was assigned to all results below the LOD, where the proportion of < LOD results is not > 60%. The provisional tolerable weekly intake (PTWI) of As is 21 μ g/kg bw (equivalent to 3 μ g/kg bw/day) according to JECFA 39 . PTMI of Cd is 25 μ g/kg bw on a monthly basis (0.8 μ g/kg bw/day) according to JECFA 40 . Rfd of Cr (1500 μ g/kg bw/day) and Ni (20 μ g/kg bw/day) is based on US EPA and WHO 41,42 . PTWI of Hg is 1 μ g/kg bw per week according to JECFA 39 . Reference value of 1.5μ g/kg bw/day for Pb is based on cardio-vascular effects according to EFSA 43 Deterministic (point estimates) approaches were adopted here for quantification of the health risk. Exposure from vegetable was obtained by combining its consumption data and the heavy metal concentrations of the specific item and then dividing by body weight. The average body weight in this study was considered as 55.9 g 14 . The mean and 95th percentile of the daily exposure levels were used to represent the dietary exposure for average and high consumers, respectively 44 . The health risk index was calculated by dividing daily intake of heavy metals by their safe limits 32 .  Table 3. Determination of certified materials of vegetables (n = 6, fresh weight).