Multi-mycotoxin exposure and risk assessments for Chinese consumption of nuts and dried fruits

WANG Yu-jiao , NIE Ji-yun , YAN Zhen , LI Zhi-xia , CHENG Yang , Saqib Farooq

1 Institute of Pomology, Chinese Academy of Agricultural Sciences, Xingcheng 125100, P.R.China

2 Laboratory of Quality and Safety Risk Assessment for Fruit (Xingcheng), Ministry of Agriculture, Xingcheng 125100, P.R.China

3 Quality Inspection and Test Center for Fruit and Nursery Stocks (Xingcheng), Ministry of Agriculture, Xingcheng 125100, P.R.China

1. Introduction

Mycotoxins are toxic secondary metabolites produced by fungi which exhibit diverse effects on human health (Marin et al. 2013; Quiles et al. 2016). They are considered to be among the most significant food contaminants of many countries, especially developing countries (FAO 2001). Studies on mycotoxin risk assessment are gaining increasing attention on the global scale, as mycotoxins are one of the most prominent risk factors of agricultural product quality and safety (Li et al. 2017). Nuts and dried fruits are increasingly popular snacks, and they are susceptible to fungal contamination because of their intrinsic characteristics of moisture and nutrient content,long storage time, and high pH and water activity, all of which favor the growth of fungi (Mello and Scussel 2007;Nie 2017; Tolosa et al. 2013). According to data from the European Commission RASFF (Rapid Alert System for Food and Feed), nuts and their products have been one of the top safety-patrolled foods since 2008 as a result of their mycotoxin contamination (Miao and Zhou 2014; Dai 2015). In China, the consumption of nuts and dried fruits,such as walnuts, pine nuts, hazelnuts, dried jujubes, etc.,have maintained a double-digit growth trend since 2005 as a result of their use as daily leisure food and as raw material input for processed foods (breakfast congee, baked goods,tea, etc.) (Chen and Liu 2015). Consequently, the health risks resulting from mycotoxins from consumption of nuts and dried fruits cannot be ignored.

There are 15 different mycotoxins commonly found in nuts and dried fruits, which belong to four categories: first,aflatoxins (AFs), including AFB1, AFB2, AFG1and AFG2(Juan et al. 2008; Asghar et al. 2017); second, trichothecene toxins (TCs), including T-2 toxin (T-2), zearalenone (ZEA),emerging enniatins (ENNs: ENA, ENA1, ENB, ENB1) and emerging beauvericin (BEA) (Varga et al. 2013); third,Alternaria toxins (ATs), including tentoxin (TEN), alternariol(AOH) and alternariol monomethyl ether (AME) (Marin et al. 2013; Patriarca 2016); and fourth, ochratoxin A (OTA)(Alghalibi and Shater 2004; Bircan 2009). According to Chinese National Food Safety Standards, only AFB1has a maximum limit in nuts which is 5 μg kg–1in cooked nuts and seeds, except for 15 μg kg–1in almonds and hazelnuts (GB2761-2017 2017). In the European Union(EU), for nuts which are to be subjected to sorting, or other physical treatment before consumption, the limit for legal regulation for AFB1is 5 μg kg–1, and for total aflatoxin(AFtotal=AFB1+AFB2+AFG1+AFG2) is 10 μg kg–1(EC 2006).The OTA levels cannot exceed more than 10 μg kg–1in dried vine fruits (raisins, etc.) according to the EC (2006). Ukraine implemented the maximum level of ZEA as 1 000 μg kg–1for all nuts (Barkai-Golan and Paster 2011) and Armenia implemented the maximum level of T-2 as 100 μg kg–1for all fruits and their products (Fung and Clark 2004).

Risk assessment is a multi-step process, in which both exposure assessment and risk characterization are included.The risk characterization is based on the results of estimated exposures to contaminants from the diet, and the exposure estimation is determined from contamination data of the foods coupled with food consumption data (Haighton et al. 2012;Marin et al. 2013). A common methodology for mycotoxin exposure assessment is the point estimation of dietary exposure based on a deterministic approach (WHO 2009;Quiles et al. 2016). The results allow comparisons between studies from different countries with a single standard value of consumer exposure: tolerable daily intake (TDI)proposed by García-Moraleja et al. (2015) and Rodríguez-Carrasco et al. (2013). Another method, the probabilistic approach, was also adopted by several previous studies which assessed the exposure to mycotoxins through several food categories (Haighton et al. 2012; Assunção et al. 2015;García-Moraleja et al. 2015). For risk characterization of individual mycotoxins, the %TDI concept and the margin of exposure (MoE) concept are usually adopted for nongenotoxic and genotoxic compounds, respectively (Benford et al. 2010; Azaiez et al. 2015; Oyedele et al. 2017).Considering the risk characterization approach to predict the combined mycotoxin hazard, cumulative risk assessments of environmental contaminants, i.e., perfluoroalkylated and polyfluoroalkylated substances (PFASs), have employed the concepts of concentration addition (CA) and independent action (IA) (Sarigiannis and Hansen 2012; Borg et al. 2013).CA assumes that the individual components act via the same mode of action, only differing in their relative potency for eliciting the toxic effects, whereas IA assumes that the individual components act independently of each other (Borg et al. 2013; Panel 2013).

Many surveys on mycotoxin contamination are developed from a varied food matrix, including nuts and dried fruits.However, reports in China regarding mycotoxins risk assessment in nuts and dried fruits are rare, and are often incomplete in that they consider only a few kinds of matrices or mycotoxins, such as AFs and OTA. As for risk assessment of the emerging mycotoxins, none has been reported for China to date. The aim of our study is to determine the dietary exposure to the aforementioned mycotoxins for Chinese consumers from nuts and dried fruits, and to assess the potential health risks resulting from the intake of individual and combined mycotoxins. Both deterministic and probabilistic approaches are used for estimating the daily intake of mycotoxins. To the best of our knowledge, this is one of only a few reports which consider the individual and combined risk assessments of multiple mycotoxins, especially the emerging mycotoxins, through the intake of nuts and dried fruits in China.

2. Materials and methods

2.1. Instruments and reagents

The Xevo TQ UPLC-MS/MS (Waters, USA), CF16RX II high speed refrigerated centrifuge (Hitachi, Japan), Milli-Q Synthesis Ultrapure Water System (Millipore, USA), and a blender (JYL-C020, JOYOUNG COMPANY LIMITED,China) were used in this study. The reversed phase BEH C18 column (2.1 mm×100 mm, 1.7 μm) for LC was produced by Waters Co. (USA).

HPLC-grade acetonitrile, methanol and formic acid were obtained from Thermo Fisher Scientific (USA,purity≥99.9%). Analytical-reagent citric acid was produced by J.T.Baker (USA, purity≥99.5%). Deionized water (＜10 MΩ cm–1, resistivity) was obtained by using the Milli-Q SP®Reagent Water System (Millipore, Beadford, MA, USA).Analytical-grade anhydrous MgSO4(purity≥99.0%) and sodium chloride (NaCl, purity≥99.5%) were obtained from Fengchuan Chemical Reagent Technology Co., Ltd. (Tianjin,China). Syringe nylon filters (0.22 μm) were produced by FINE Scientific Company (USA).

Standards of AFs, TCs, ATs and OTA were obtained from Pribolab (Singapore, purity≥98.0%). Stock standard solutions of each mycotoxin were made in acetonitrile at 200 μg mL–1concentrations, except ENNs and BEA standard solutions which were made at the same concentration in methanol. All stock standard solutions were stored in the dark at –20°C. Mixed standard stock solutions were prepared with the individual stock solutions in acetonitrile at a final concentration of 10 μg mL–1for each mycotoxin.Mixed standard working solutions were prepared daily by diluting the mixed standard stock solutions.

2.2. Sampling

A total of 233 samples were collected from local markets and supermarkets in 21 provinces, autonomous regions and municipalities of China. The 233 samples consisted of 9 different types of dried fruits and nuts: dried jujubes (n=35),dried figs (n=20), raisins (n=30), dried longans (n=15),and nuts (n=133) which included walnuts (n=35), Chinese chestnuts (n=33), hazelnuts (n=20), pine nuts (n=20) and almonds (n=25). The edible parts of all samples were milled with a blender by adding specific volumes of deionized water per 100 g edible section of samples: 50 mL for pine nut;100 mL for dried longan, walnut, raisin and chestnut; 150 mL for dried fig, almond and hazelnut; and 200 mL for dried jujube. Subsequently, processed samples were stored at–20°C until analysis.

2.3. Sample preparation and instrumental analysis

Milled samples of specific weights (7.5 g for pine nut; 10.0 g for dried longan, walnut, raisin and chestnut; 12.5 g for dried fig, almond and hazelnut; and 15.0 g for dried jujube) were placed into 50-mL PTFE centrifuge tubes. A 10-mL mixture of acetonitrile and citric acid (10 mmol citric acid per liter of acetonitrile) was added to the 50-mL tubes and shaken for 3 min. Next, 1 g of NaCl and 4 g of anhydrous MgSO4were added and the mixture was immediately shaken for 1 min. The tubes were centrifuged at 9 000 r min–1for 5 min.Subsequently, 5 mL of the acetonitrile layer was transferred to 15-mL centrifuge tubes containing 300 mg of C18 sorbent,and after shaking for 1 min, the supernatant was filtered through a nylon filter and transferred to sample vials for UPLC-MS/MS analysis.

The 15 mycotoxins were analyzed using a Waters Acquity UPLC System coupled to a Waters Xevo TQ triple quadrupole mass spectrometer. The mobile phase A was acetonitrile and the mobile phase B was 0.5% formic acid in water with 10 mmol L–1citric acid. The following linear gradient was used: 0.0–1.0 min, 5% A; 1.0–3.2 min, 5–50%A; 3.2–5.4 min, 50–95% A; 5.4–7.7 min, 95% A; 7.7–8.0 min, 95–5% A. The column temperature was maintained at 40°C. The mass spectrometer was equipped with an electrospray ionization (ESI) source and was operated in both positive and negative ion modes with the following parameters: capillary voltage 0.5 kV, source temperature 150°C, desolvation temperature 500°C, and desolvation gas and nebulizer gas (N2) were set at 800 and 50 L h–1,respectively. Quantitation was performed using multiple reaction monitoring (MRM) mode. The MS/MS conditions of the 15 mycotoxins are presented in Table 1.

2.4. Validation of the methods

The analytical method was previously optimized in our laboratory (Wang et al. 2017). Based on the guidelines in the EU Commission Decision, 2002/657/EC (EC 2002),we developed an analytic method with similar validation parameters, such as limit of detection (LOD), limit of quantification (LOQ), linearity (r2), range of matrix effects(ME), recovery and relative standard deviation (RSD), for evaluating mycotoxin contamination levels. The LOD and LOQ ranged from 0.05 to 1.00 μg kg–1and from 0.10 to 5.00 μg kg–1, respectively. All mycotoxins exhibited good linearity (r2＞0.9981). Recoveries and RSDs were used to reflect the method precision by adding mixed standard working solution to blank samples at three levels (LOQ,10LOQ and 100LOQ) according to European Commission(EC 2002). The RSD values were in the ranges of 5.88 to 17.30 for intra-day and 2.07 to 9.31 for inter-day analyses,and the experimental recovery values ranged from 80.9 to 110.7%. All of these values are within the acceptable range of EU criteria (EC 2002). According to the ME range, ion enhancement (＞100%) and ion suppression (＜100%) were both observed. Therefore, matrix-matched calibration curves were used to minimize the matrix effects for effective quantification. The detailed data of method validation and performance are shown in Table 2.

2.5. Consumption data

According to the report of Guo et al. (2014), nut consumption was calculated to be 3.7 g d–1for Chinese males and 3.2 g d–1for Chinese females. The daily consumption of dried jujubes for Chinese consumers was estimated to be 5.5 g d–1based on the data issued by Wang et al. (2015). For the consumption data of dried longans and dried figs, 0.5 g d–1was referenced from the Global Environment Monitoring System-Food Contamination Monitoring and Assessment Programme (GEMS/Food) database, which was weighted by the population size of the reporting country (Ji et al. 2017).

Table 1 MS/MS conditions for mycotoxin analysis

For raisins, the Chinese consumed 16.50 million tons in 2015 according to the report of Qi (2015), and China had 13.75 billion people in 2015 based on the population data issued by China State Statistical Bureau (NBSC 2017). Thus,the approximate annual per-capita raisin consumption was 1 200 g, calculated by dividing the raisin consumption by population size, and the daily per-capita consumption was 3.29 g d–1, calculated by the annual per-capita raisin consumption divided by the total days in a year (365 d).

2.6. Exposure assessment

Calculation of mycotoxin intakesThe estimated daily intake (EDI) for each mycotoxin analyzed was calculated for exposure assessment and expressed as ng kg–1bw d–1(Rodríguez-Carrasco et al. 2013):

EDI=(C×K)/bw

Where, C is the mean concentration of each mycotoxin expressed as μg kg–1; K is the consumption of the commodity(g d–1) and bw is the body weight. The body weight data for the male and female adult populations are assumed as 66.18 and 56.05 kg, respectively (Hou et al. 2014). The mean mycotoxin concentrations were calculated considering all contamination data, including the positive and negative samples (Quiles et al. 2016).

Risk characterization of mycotoxinsThe tolerable daily intake (TDI) is a value set by the relevant agencies (EFSA,European Food Safety Authority; JECFA, Joint FAO/WHO Expert Committee on Food Additives) and indicates the quantity of a substance that can be consumed over a lifetime without risk to health (Quiles et al. 2016). The health risk characterization of each mycotoxin (% of relevant TDI)was performed by comparing the EDI with the TDI (Azaiez et al. 2015):

%TDI=(EDI/TDI)×100

%TDI＜100 indicates a tolerable exposure level and%TDI＞100 indicates an intolerable level. The TDI values established for OTA, ZEA and T-2 are 170, 250 and 60 ng kg–1bw d–1(JECFA 2001; EFSA 2010, 2014). For the non-genotoxic mycotoxins, ENNs and BEA, no TDI or provisional tolerable daily intake (PTDI) levels are available as a result of the lack of conclusive studies about their toxicity in vitro, so their contribution to the risk assessment was approximated by the TDI values established for other Fusarium mycotoxins, such as DON (1 000 ng kg–1bw d–1)(JECFA 2001; Quiles et al. 2016). For ATs (AOH, AME and TEN), another non-genotoxic mycotoxin category for which TDI values are not yet established, a hypothetical value (100 ng kg–1bw d–1) was applied (Rodríguez-Carrasco et al. 2013).

AFB1and the mixture of AFs have been classified as human carcinogens by the International Agency for Research on Cancer (IARC 1993; 2013), and no threshold doses have been established. Several reports have recommended the margin of exposure (MoE) approach for risk characterization of genotoxic and carcinogenic mycotoxins like AFs (EFSA 2007; Benford et al. 2010; Oyedele et al. 2017), and this approach was adopted in our study. The MoE was calculated by dividing the benchmark dose lower limit (BMDL) by the EDI of AFs. The BMDL values for AFB1(170 ng kg–1bw d–1), AFB2(250 ng kg–1bw d–1),AFG1(250 ng kg–1bw d–1) and AFG2(250 ng kg–1bw d–1) were derived from the studies of Benford et al.(2010) and EFSA (2007).

2.7. Statistical analysis

The mycotoxin incidence data were analyzed using SAS (ver. 9.2) and were reported as mean±standard deviation values. According to the criteria indicated in EFSA (2010), left-censored data (results below LOD and non-detects) were assigned values of half the limit of detection (1/2LOD) for the mycotoxin dietary exposure assessment. This is considered to be a worst-case scenario that may overestimate contamination and exposure levels; however, other proposed scenarios, such as considering left-censored data as zero, could underestimate the real exposures(Quiles et al. 2016).

Point evaluation, food consumption, and other calculations were performed by using Microsoft Office Excel 2007. Probabilistic analysis was simulated by using the@RISK Software Package, ver. 5.5 (Microsoft,USA). The best fit function of@Risk Software was applied to select the best fitted probabilistic distribution of all mycotoxin concentrations and a full probabilistic model (Monte Carlo simulation) was performed.

3. Results and discussion

3.1. Mycotoxin occurrence

The mycotoxin occurrence data are summarized in Table 3. The results revealed that all 15 mycotoxins were present and 47.6% (111/233) of the samples were contaminated. There were two samples positive for AFB1(40.7 and 384.1 μg kg–1) which exceeded the maximum tolerable level standards of China(GB2761-2017 2017) and the EU (EC 2006). Higher over-standard rates have been reported in previous similar studies. For example, Luttfullah and Hussain(2011) reported that seven positive dried fruit and nut samples (9.3%) were contaminated with AFB1at levels above the suggested limit of the EU. Karaca and Nas (2006) surveyed the contamination of AFs in dried figs of Turkey and demonstrated that the positive samples contaminated with AFs ranged from 117.9 to 471.9 μg kg–1.

Table 3 Mycotoxin incidence and concentration in nut and dried fruit samples

All of the five kinds of samples had detections of one to four different AFs. Dried figs had incidences of all four AFs,while raisins only had incidences of AFG2with contamination levels of 0.6–1.2 μg kg–1. Iamanaka et al. (2007) reported that dried grapes were contaminated with AFB1and AFB2at a rate of 16%. As TCs, ZEA and T-2 were detected with the lowest frequency at levels of 1.8–49.3 and 2.8–7.0 μg kg–1, respectively. The incidence rate of ENB1(12.9%) was the highest and its contamination levels ranged from 0.1 to 134.0 μg kg–1. The maximum detection level of ENB was 247.3 μg kg–1, which was the highest among the emerging ENNs. Tolosa et al. (2013) studied ENNs in dried fruits and nuts and found that samples were contaminated with ENB at levels of 14.610 mg kg–1and that ENA was the most prominent ENNs found in nuts (45.2%). AME was the most common ATs found in 18.0% of samples and the maximum detection value of AOH (142.9 μg kg–1) was the highest among the three ATs. Wei et al. (2017) surveyed ATs in dried fruits of China and found that TEN was the most common ATs (20.5%) with the maximum detection value of 1 032.6 μg kg–1, which was much higher than that in the current study (32.9 μg kg–1).

There were 43.6% of the nut samples with mycotoxins detected, and the AME incidence in nuts was the highest(23/133) followed by AFB2(20/133) and AOH (19/133). One of the nut samples (chestnut) was contaminated with AFB1at a level of 39.3 μg kg–1, exceeding the maximum limit set by China (GB2761-2017 2017) and the EU (EC 2006). As reported previously, 58 pine nuts were all contaminated with AFs exceeding the maximum tolerable limit of 4 μg kg–1set by the EU (Sharma et al. 2015). There were two nut samples contaminated with OTA reported by Sharma et al.(2015), but OTA was not found in nuts in the present study.Considering the AT contamination in nuts, TEN, AOH and AME were detected at percentages of 8.3, 14.3 and 17.3%,respectively, with contamination levels ranging from 0.7 to 32.9 μg kg–1, from 1.4 to 142.9 μg kg–1and from 0.9 to 110.5 μg kg–1, respectively. Limited data have been reported on AT contamination of nuts, while most reports have emphasized AT contamination of fruits and their products (Asam et al.2009; Patriarca 2016; Zwickel et al. 2016).

Dried figs had the highest incidence (80.0%) of mycotoxins among the five sample species. The detection rates of AFB1,AFG1, BEA, T-2, AOH and AME were all higher in dried figs than in the other four sample species. However, no ENNs were found in dried fig samples. Azaiez et al. (2015)detected ENNs in dried figs and found high occurrences of ENA1(50.0%), ENB (50.0%) and ENB1(50.0%). As for dried jujubes, 42.9% (15/35) samples had detectible ENB1in the range of 0.7–32.9 μg kg–1. The maximum detection level of ENB (247.3 μg kg–1) was the highest among the 10 mycotoxins found in dried jujubes. Dried longans had the highest incidence of AFG2among the five sample types and had no incidences of ATs or OTA. Previous data on mycotoxin contamination in dried jujubes and dried longans from other countries are negligible because these two kinds of foods are almost exclusive to China. Domestic studies regarding mycotoxin contamination on dried jujubes and dried longans have seldom been reported. By contrast,the mycotoxin occurrence in raisins, a kind of dried fruit produced worldwide, has been reported extensively. Wei et al. (2017) found 19.3% of raisins containing OTA in the range of 0.17–8.8 μg kg–1, and none of the samples exceeded the maximal level set by the EU (10 μg kg–1) (EC 2006). Han et al. (2016) reported higher OTA contamination levels (56.5%) with a range of 0.4–65.7 μg kg–1in raisins of China, and found15.6% of positive samples exceeding the maximum limit of the EU (EC 2006). In this study, none of the raisin samples were contaminated with OTA levels exceeding the maximum limit of the EU (EC 2006).

3.2. Deterministic estimation

Exposure assessment of aflatoxins (AFs)As discussed previously, the MoE approach is the most common method for assessing the risk characterization of AFs. The MoE concept suggests that an MoE value of ＞10 000 should be considered as ‘safe’ while an MoE value ≤10 000 could pose a potential risk to public health and the lower the value, the higher the risk (Cartus and Schrenk 2017; Heshmati et al.2017).

As shown in Table 4, the EDI of AFs through consumption of nuts, dried jujubes, dried figs, raisins and dried longans were calculated to range from 0.0004 to 0.0663 ng kg–1bw d–1. The EDI value of AFB1through consumption of dried figs was the highest among the five sample types,and the AFB1mean contamination level in dried figs was the highest among the five sample types. The MoE values of AFB2, AFG1and AFG2for all of the five sample types and the MoE values of AFB1for raisins and dried longans were ＞10 000, meaning that the intake of these mycotoxins posed no health risk. However, the MoE values for AFB1for nuts, dried jujubes and dried figs were in the range of 2 549–9 502, which indicates that the consumption of nuts and dried jujubes and figs puts Chinese adult consumers at an exposure risk to AFB1.

Only limited information regarding the AF MoE of nuts,dried jujubes, raisins and dried figs has been reported previously, though there have been several cases. For example, the MoE value for AFB1for dried fig consumption in Iran was reported to be 4 250 (Heshmati et al. 2017).Ding et al. (2012) published that the MoE value for AFB1for peanut consumption in China was 1 273. In addition,the estimated MoE value for AFB1from the intake of bakery products and pasta was reported to be 24.6 (Bol et al. 2016).

Exposure assessment of trichothecene toxins (TCs)To the best of our knowledge, the exposure assessment of emerging mycotoxins, the ENNs (ENA, ENA1, ENB and ENB1) and BEA, in nuts and dried fruits has never been reported in China. Data on the risk of TC exposure for Chinese adults by consuming nuts and dried fruits are presented in Table 5. The results showed that none of the calculated EDI exceeded the TDI(%TID＜100) indicating that there is no health risk from individual TCs to Chinese consumers of nuts and dried fruits. Nuts and dried jujubes had higher incidences of ENNs than raisins, dried figs and dried longans. As a result, the EDI values of ENNs were higher for nuts and dried jujubes than for the other three sample species. These results are in agreement with results of García-Moraleja et al. (2015) and Rodríguez-Carrasco et al. (2013) who investigated ENNs in coffee, wheat, rice and maize. Furthermore,García-Moraleja et al. (2015) reported that the EDI values of ENNs were higher than other mycotoxins (AFs, OTA, etc.). In our study,however, the rank of EDI values among different mycotoxins was complicated because of the varied types of matrices analyzed. For example, the EDI values of ENNs are higher than those of AFs for nuts and dried jujubes, while the EDI values of AFs are higher than those of ENNs for dried figs, raisins and dried longans.

Exposure assessment of Alternaria toxins (ATs)The estimated daily intake data for ATs are summarized in Table 6. Nuts had the highest incidence of ATs among the five sample types (Table 3).However, the dietary exposure values of ATs through nuts were lower than the values in the remaining samples. None of these values exceeded the TDI. Therefore, no risk to consumer health existed through the dietary consumption of nuts and dried fruits and subsequent exposure to ATs. Zhao et al. (2015) reported dietary exposure to ATs through consuming wheat flour and wheat-based foods in China. In that report, the threshold of toxicological concern(TTC) approach was employed to assess the health risk resulting from exposure to the ATs. According to the results of that study,the daily intakes of TEN ranged from 41.3 to 469 ng kg–1bw d–1,which was much lower than the TTC value (1 500 ng kg–1bw d–1).The daily intakes of AOH and AME, however, were higher than the corresponding TTC value (2.5 ng kg–1bw d–1). Hence, that study concluded that the exposure to TEN posed no health risk to Chinese consumers, while the exposure to AOH and AME posed potential health risks. Comparing the results of Zhao et al. (2015) and the current study, the health concerns caused by the intakes of ATs were much lower through the consumption of nuts and dried fruits than through the consumption of wheat flour and wheat-based foods.

Exposure assessment of ochratoxin A (OTA)The dietary exposure to OTA through nuts and dried fruits ranged from 0.0038 to 0.2510 ng kg–1bw d–1, all below the TDI, indicating that they posed no health risk to Chinese adult consumers (Table 6). The exposure intake of OTA was the highest through consuming raisins followed by dried jujubes. Azaiez et al. (2015) evaluated OTA in dried fruits and found OTA only in raisins with EDI values of 0.14 ng kg–1bw d–1. OTA has been classified as group 2B, “possible carcinogens in humans”, according to the report of IARC (2013). Consequently,Heshmati et al. (2017) used MoEs to calculate dietary exposure to OTA in figs and reported that the MoE values were ＞10 000.

Considering the individual mycotoxins, the dietary exposure did not pose health risks to Chinese adults except for AFB1by consuming nuts, dried jujubes and dried figs.The RASFF studied the distribution of contamination by AFs in food sources originating throughout the world and concluded that the proportions of contaminated samples among dried figs and nuts (hazelnuts and pistachios) were the highest (55%) (Kabak 2016). Therefore, the health risks posed by the daily intake of AFB1through nuts and dried fruits should not be ignored, although the EDI values of AFB1were calculated in the deterministic approach,which may overestimate the exposure level. Furthermore,efficient regulations need to be updated to protect Chinese consumers against the health effects caused by AFB1.Besides, every EDI value for females was higher than for males, though there was no significant difference between sexes (P＞0.05) based on the analyzed results of the t-test by SAS. This conclusion was in accordance with the previous results presented by Cano-Sancho et al. (2012).

3.3. Probabilistic analysis

The point approach commonly used to assess mycotoxin exposure usually does not consider the variability and uncertainty of the food consumption and contamination level parameters, resulting in unrealistic exposure calculations in some cases (Han et al. 2014). In contrast, the probabilistic approach takes into account every possible value that each variable can assume, and weighs each possible scenario for the probability of its occurrence, allowing for a more accurate characterization of mycotoxin intake distribution(Assunção et al. 2015).

As stated earlier, the mycotoxin exposure was higher for females than for males based on the risk assessment results of the point approach, and there was no significant difference between the two population groups. Thus, the EDI values of exposure to the 15 mycotoxins through consumption of nuts and dried fruits were calculated by the probabilistic approach only for females and the results are shown in Table 7. The %TDI values for TCs, ATs and OTA acquired by both the deterministic approach (Tables 5 and 6) and the probabilistic approach (Table 7) are all＜100,indicating that the daily intake of these specific mycotoxins through consumption of nuts and dried fruits do not pose a health risk to Chinese consumers. Comparing the mycotoxin MoE values of AFs acquired by the deterministic approach(Table 4) and the probabilistic approach (Table 7), there are differences in some cases, indicating different risk characterizations. For example, the risk characterization results based on the deterministic approach indicates that only the daily exposure to AFB1through consumption of nuts, dried jujubes and dried figs posed health risks to Chinese females and males. Based on the probabilistic approach, however, the MoE to AFB1through nuts at all three percentiles were ＞10 000 indicating that there were no health risks from exposure to AFB1by consuming nuts.The MoE values of AFB2in nuts at the 90th (5 204＜10 000)and 95th percentiles (4 514＜10 000), and the MoE value of AFG1in dried jujubes at the 95th percentile (8 185＜10 000),also demonstrated different results compared to the results based on the point approach. Han et al. (2014) assessed the health risk of exposure to deoxynivalenol mycotoxin,and its acetyl derivative, via dietary intake of wheat and maize from Shanghai, China, employing both point and probabilistic approaches. The results obtained using these two approaches were found to be very similar.

Most risk assessments of mycotoxins in nuts and dried fruits have been made based on the point approach (Azaiez et al. 2015; Kabak 2016; Heshmati et al. 2017), while the probabilistic approach has been adapted more often to estimate the health risk of mycotoxins from cereals and other foods. For example, a report on risk assessment of mycotoxins present in breakfast cereals consumed by children in Portugal adopted the probabilistic method for performing mycotoxin exposure assessment (Assunção et al. 2015). García-Moraleja et al. (2015) revealed the daily intake of 21 mycotoxins in coffee by calculating the EDI in different scenarios corresponding to population consumption habits, including the mean population and highly exposed consumers (the 95th, 97.5th and 99th percentiles). García-Moraleja et al. (2015) concluded that the EDI values in highly exposed segments were considerably higher than in the mean population.

3.4. Cumulative risk assessment of mycotoxins analyzed

The mycotoxin occurrence data in nuts and dried fruits confirmed that the population consuming these foods are exposed simultaneously to multiple mycotoxins. Various methods for the hazard and risk assessment of mixtures have been developed to predict the combined toxicity of mixtures and their risks (Han et al. 2014; Assunção et al.2015). According to the report of Assunção et al. (2015),two methods based on the CA concept, the Hazard Index(HI) and the combined Margin of Exposure index (MoET),were used for cumulative risk assessment in this study. The HI, an estimate of the total risk based on the individual risks of each component, was used for non-genotoxic mycotoxin analyses. The MoET was used for the cumulative risk assessment of genotoxic and carcinogenic mycotoxins.The mycotoxins analyzed were grouped into their chemical classes, and the cumulative risk assessments were performed for AFs (AFB1, AFB2, AFG1and AFG2), TCs (ENA,ENA1, ENB, ENB1, BEA, T-2 and ZEA) and ATs (TEN, AOH and AME).

Table 8 Results of the mycotoxins risk assessment in combination for female population performed by deterministic approach(DA) and probabilistic approach (PA)

Table 8 presents the results of the cumulative risk characterization by using MoET for AFs, and HI for TCs and ATs, derived from the estimated exposure to these mycotoxins performed by both point and probabilistic approaches. The values of MoET＞10 000 or HI＜100 indicate that there is no cause for concern for the combination exposure to mycotoxins through the consumption of nuts and dried fruits. According to the data in Table 8, all of the HI values for TCs and ATs were＜100, demonstrating that there is no health risk from combined TCs or combined ATs through consuming nuts and dried fruits. The cumulative exposure to AFs through consumption of nuts, dried jujubes and raisins poses a potential health risk to Chinese adults indicated by the MoET values lower than 10 000. Few studies have reported the combined exposure to AFs through the intake of nuts or dried fruits. Assunção et al. (2015) studied the cumulative risk assessment of AFs (AFM1, AFB1, AFB2and AFG1) in breakfast cereals based on a probabilistic approach and found MoET＜10 000 at the 90th, 95th and 99th percentiles. To the best of our knowledge, this paper is the first report to study the cumulative risk assessment of multiple mycotoxins, especially the emerging mycotoxins, in nuts and dried fruits consumed by Chinese adults.

3.5. Uncertainty

There were several uncertainties in these assessments. The major uncertainty was the crudely estimated consumption data of nuts and dried fruits, which could increase the uncertainty considering that consumption is surely variable with time and the geographical distribution of the population. For example, it was reported that the consumption of dried jujubes was higher in the production area than in non-production areas and higher during the traditional festival than other times of the year (Yin and Meng 2017). Another uncertainty was the step of setting concentration data reported below the LOD (left-censored data) as 1/2LOD. This scenario is known to represent the worst-case for mycotoxin dietary exposure assessment,and might overestimate both contamination and exposure levels (Jacxsens et al. 2016; Kabak 2016; Quiles et al.2016). In addition, the present study only concerns the risks associated with adults’ exposure to mycotoxins through consuming nuts and dried fruits. It is expected that the exposure resulting from the consumption of all foods present in the Chinese diet would increase the intake of mycotoxins(Assunção et al. 2015).

4. Conclusion

In this study, 15 individual mycotoxins were analyzed in five different types of nuts and dried fruits of China. AFs were detected in all five types of samples, and TCs and ATs were each detected in four sample types with the exceptions of raisins and dried longans, respectively. OTA was found in only two raisin samples with contamination levels of 4.6 and 7.4 μg kg–1. AME had the highest detection incidence of 42/233 samples, followed by ENB1(30/233) and AOH(29/233). The daily exposure and risk characterization of the 15 mycotoxins through consumption of nuts and dried fruits were assessed for Chinese adult males and females. According to our survey of the literature, the risk assessment of multiple mycotoxins including AFs, ATs, OTA,ZEA, T-2, and especially ENNs and BEA, in nuts and dried fruits from China has not been reported before. The risk characterization results based on both deterministic and probabilistic approaches for TCs, ATs and OTA indicated no health concerns for individuals based on expected exposures to multiple mycotoxins (%TDI and HI＜100). As for individual and combined AFs, the MoE below 10 000 and the MoET below 10 000 suggested potential health concerns for either the worst-case scenario or the highest percentiles of intake (the 90th and 95th percentiles). As a whole, the health risks posed by the 15 mycotoxins to Chinese adults consuming nuts and dried fruits appeared to be relatively low. However, mycotoxin intake through the whole diet should be considered to acquire a complete view of the risk assessment, particularly because nuts and dried fruits are non-staple foods according to common consumption customs. The results of this study also corroborate the need for further studies in the domain of the mycotoxin cumulative risk assessment for establishing legal protective values to achieve an amelioration of the people’s health.

Acknowledgements

This work was supported by the National Program for Quality and Safety Risk Assessment of Agricultural Products of China (GJFP2016003 and GJFP2017003) and the Scientific and Technological Innovation Project of the Chinese Academy of Agricultural Sciences (CAAS-ASTIP).

Alghalibi S M, Shater A R M. 2004. Mycoflora and mycotoxin contamination of some dried fruits in Yemen Republic.Assiut University Bulletin for Environmental Researches,7, 19–27.

Asam S, Konitzer K, Schieberle P, Rychlik M. 2009. Stable isotope dilution assays of alternariol and alternariol monomethyl ether in beverages. Journal of Agricultural and Food Chemistry, 57, 5152–5160.

Asghar M A, Ahmed A, Zahir E, Asghar M A, Iqbal J, Walker G. 2017. Incidence of aflatoxins contamination in dry fruits and edible nuts collected from Pakistan. Food Control, 78,169–175.

Assunção R, Vasco E, Nunes B, Loureiro S, Martins C, Alvito P. 2015. Single-compound and cumulative risk assessment of mycotoxins present in breakfast cereals consumed by children from Lisbon region, Portugal. Food and Chemical Toxicology, 86, 274–281.

Azaiez I, Font G, Mañes J, Fernández-Franzón M. 2015. Survey of mycotoxins in dates and dried fruits from Tunisian and Spanish markets. Food Control, 51, 340–346.

Barkai-Golan R, Paster N. 2011. Mycotoxins in Fruits and Vegetables. Academic Press, USA.

Benford D, Bolger P M, Carthew P, Coulet M, DiNovi M, Leblanc J C, Renwick A G, Setzer W, Schlatter J, Smith B, Slob W,Williams G, Wildemann T. 2010. Application of the margin of exposure (MoE) approach to substances in food that are genotoxic and carcinogenic. Food and Chemical Toxicology,48, S2–S24.

Bircan C. 2009. Incidence of ochratoxin A in dried fruits and cooccurrence with aflatoxins in dried figs. Food and Chemical Toxicology, 47, 1996–2001.

Bol E K, Araujo L, Veras F F, Welke J E. 2016. Estimated exposure to zearalenone, ochratoxin A and aflatoxin B1 through the consume of bakery products and pasta considering effects of food processing. Food and Chemical Toxicology, 89, 85–91.

Borg D, Lun B O, Lindquist N G, Håkansson H. 2013.Cumulative health risk assessment of 17 perfluoroalkylated and polyfluoroalkylated substances (PFASs) in the Swedish population. Environment International, 59, 112–123.

Cano-Sancho G, Marin S, Ramos A, Sanchis V. 2012.Occurrence of zearalenone, an oestrogenic mycotoxin, in Catalonia (Spain) and exposure assessment. Food and Chemical Toxicology, 50, 835–839.

Cartus A, Schrenk D. 2017. Current methods in risk assessment of genotoxic chemicals. Food and Chemical Toxicology,106, 574–582.

Chen F, Liu X X. 2015. Roasted seeds and nuts industry market share will reach 100 billion yuan in 2020. Chinese Business Daily. p. 6. (in Chinese)

Dai Y, Li Q, Liu W, Liu A L, Liu P. 2015. Analysis on China food safety by notifications of EU rapid alert system for food and feed (RASFF) in 2014. Food Research and Development,36, 177–180.

Ding X X, Li P W, Bai Y Z, Zhou H Y. 2012. Aflatoxin B1 in postharvest peanuts and dietary risk in China. Food Control,23, 143–148.

EC (European Commission). 2002. Commission regulation(EC) No. 1058/2012 of 12 november 2012 amending regulation (EC) No. 1881/2006 as regards maximum levels for aflatoxins in dried figs. Official Journal of the European Union, L313, 14–15.

EC (European Commission). 2006. Commission regulation(EC) No. 1881/2006 of 19 december 2006 setting maximum levels for certain contaminants in foodstuffs. Official Journal of the European Union, L364, 5–24.

EFSA (European Food Safety Authority). 2007. Opinion of the scientific panel on contaminants in the food chain on a request from the commission related to the potential increase of consumer health risk by a possible increase of the existing maximum levels for aflatoxins in almonds,hazelnuts and pistachios and derived products. EFSA Journal, 446, 1–127.

EFSA (European Food Safety Authority). 2010. Management of left-censored data in dietary exposure assessment of chemical substances. EFSA Journal, 8, 1557.

EFSA (European Food Safety Authority). 2014. Scientific opinion on the risks for human and animal health related to the presence of modified forms of certain mycotoxins in food and feed. EFSA Journal, 12, 1–107.

FAO. 2001. Manual on the application of the HACCP system in mycotoxin prevention and control. FAO Food and Nutrition Paper, 73, 1–114.

Fung F, Clark R F. 2004. Health effects of mycotoxins: A toxicological overview. Journal of Toxicology (Clinical Toxicology), 42, 217–234.

García-Moraleja A, Font G, Mañes J, Ferrer E. 2015. Analysis of mycotoxins in coffee and risk assessment in Spanish adolescents and adults. Food and Chemical Toxicology,86, 225–233.

GB2761-2017. 2017. National Standards of the People’s Republic of China National Food Safety Standard: Limits for Mycotoxins in Food. National Health and Family Planning Commission of the People’s Republic of China, China Food and Drug Administration. (in Chinese)

Guo Y D, Liu Y Y, Yuan Y H, Yue T L. 2014. Fumonisins contamination in major food products and its dietary risk assessment in China. Journal of Northwest A & F University(Nature Science), 42, 78–82.

Haighton L A, Lynch B S, Magnuson B A, Nestmann E R. 2012.A reassessment of risk associated with dietary intake of ochratoxin A based on a lifetime exposure model. Critical Reviews in Toxicology, 42, 147–168.

Han Z, Dong M, Han W, Shen Y, Nie D, Shi W, Zhao Z.2016. Occurrence and exposure assessment of multiple mycotoxins in dried fruits based on liquid chromatographytandem mass spectrometry. World Mycotoxin Journal, 9,465–474.

Han Z, Nie D, Ediage E N, Yang X, Wang J, Chen B, Li S, On S L W, Saeger S D, Wu A. 2014. Cumulative health risk assessment of co-occurring mycotoxins of deoxynivalenol and its acetyl derivatives in wheat and maize: Case study,Shanghai, China. Food and Chemical Toxicology, 74,334–342.

Heshmati A, Zohrevand T, Khaneghah A M, Nejad A S M,Sant’Ana A S. 2017. Co-occurrence of aflatoxins and ochratoxin A in dried fruits in Iran: Dietary exposure risk assessment. Food and Chemical Toxicology, 106, 202–208.

Hou J, Qu Y H, Ning D L, Wang H. 2014. Characteristic of human exposure factors in China and their uncertainty analysis in health risk assessment. Environmental Science& Technology, 37, 179–187. (in Chinese)

Iamanaka B T, de Menezes H C, Vicente E, Leite R S, Taniwaki M H. 2007. Aflatoxigenic fungi and aflatoxins occurrence in sultanas and dried figs commercialized in Brazil. Food Control, 18, 454–457.

IARC (International Agency for Research on Cancer).1993. Some naturally occurring substances, food items and constituents, heterocyclic aromatic amines and mycotoxins. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, 56, 489–521.

IARC (International Agency for Research on Cancer). 2013.Agents classified by the IARC monographs, volumes 1–120.[2013-09-07]. http://monographs.iarc.fr/ENG/Classification/List_of_Classifications.pdf

Jacxsens L, Uyttendaele M, de Meulenaer B. 2016. Challenges in risk assessment: Quantitative risk assessment. Procedia Food Science, 6, 23–30.

JECFA (Joint FAO/WHO Expert Committee on Food Additives).2001. Safety evaluation of certain mycotoxins in food prepared by the fifty-sixth meeting of the Joint FAO/WHO Expert Committee on Food Additives. Food and Agriculture Organization, 47.

Ji X, Li R, Yang H, Qi P, Xiao Y, Qian M. 2017. Occurrence of patulin in various fruit products and dietary exposure assessment for consumers in China. Food Control, 78,100–107.

Juan C, Zinedine A, Moltó J C, Idrissi L, Mañes J. 2008.Aflatoxins levels in dried fruits and nuts from Rabat-Salé area, Morocco. Food Control, 19, 849–853.

Kabak B. 2016. Aflatoxins in hazelnuts and dried figs:Occurrence and exposure assessment. Food Chemistry,211, 8–16.

Karaca H, Nas S. 2006. Aflatoxins, patulin and ergosterol contents of dried figs in Turkey. Food Additives and Contaminants, 23, 502–508.

Li Z X, Nie J Y, Yan Z, Zhang X N, Guan D K, Shen Y M. 2017.Progress in research of detection, risk assessment and control of the mycotoxins in fruits and fruit products. Scientia Agricultura Sinica, 50, 332–347. (in Chinese)

Luttfullah G, Hussain A. 2011. Studies on contamination level of aflatoxins in some dried fruits and nuts of Pakistan. Food Control, 22, 426–429.

Marin S, Ramos A, Cano-Sancho G, Sanchis V. 2013.Mycotoxins: Occurrence, toxicology, and exposure assessment. Food and Chemical Toxicology, 60, 218–237.

Mello F R D, Scussel V M. 2007. Characteristics of in-shell brazil nuts and their relationship to aflatoxin contamination:Criteria for sorting. Journal of Agricultural & Food Chemistry,55, 9305.

Miao T S, Zhou Q J. 2014. Empirical study on food safety of EU RASFF database. Journal of Food Science and Technology,32, 76–82.

NBSC (National Bureau of Statistics of China). 2017. Statistical Yearbook of China. China Statistics Press, Beijing. (in Chinese)

Nie J Y. 2017. Occurrence, control and determination of patulin contamination in fruits and fruit products. Scientia Agricultura Sinica, 50, 3591–3607. (in Chinese)

Oyedele O A, Ezekiel C N, Sulyok M, Adetunji M C, Warth B,Atanda O O, Krska R. 2017. Mycotoxin risk assessment for consumers of groundnut in domestic markets in Nigeria.International Journal of Food Microbiology, 251, 24–32.

Panel E C. 2013. Scientific Opinion on risks for animal and public health related to the presence of nivalenol in food and feed. EFSA Journal, 11, 3262.

Patriarca A. 2016. Alternaria in food products. Current Opinion in Food Science, 11, 1–9.

Qi G M. 2015. Overview of the world fresh and dried grape production and consumption. Sino-Overseas Grapevine &Wine, 1, 62–64. (in Chinese)

Quiles J M, Saladino F, Mañes J, Fernández-Franzón M, Meca G. 2016. Occurrence of mycotoxins in refrigerated pizza dough and risk assessment of exposure for the Spanish population. Food and Chemical Toxicology, 94, 19–24.

Rodríguez-Carrasco Y, Ruiz M J, Font G, Berrada H. 2013.Exposure estimates to Fusarium mycotoxins through cereals intake. Chemosphere, 93, 2297–2303.

Sarigiannis D A, Hansen U. 2012. Considering the cumulative risk of mixtures of chemicals - A challenge for policy makers. Sarigiannis and Hansen Environmental Health,11, S18.

Sharma S, Gupta D, Sharma Y. 2015. Natural incidence of aflatoxins, ochratoxin A, patulin and their co-occurrence in chilgoza pine nuts marketed in Jammu, India. Proceedings of the National Academy of Sciences, India (Section B:Biological Sciences), 85, 45–50.

Tolosa J, Font G, Mañes J, Ferrer E. 2013. Nuts and dried fruits:Natural occurrence of emerging Fusarium mycotoxins. Food Control, 33, 215–220.

Varga E, Glauner T, Berthiller F, Krska R, Schuhmacher R,Sulyok M. 2013. Development and validation of a (semi-)quantitative UHPLC-MS/MS method for the determination of 191 mycotoxins and other fungal metabolites in almonds, hazelnuts, peanuts and pistachios. Analytical &Bioanalytical Chemistry, 405, 5087–5104.

Wang H X, Guo X J, Yan J S, Jin W, Li C X, Sun J, Liu L G, Xin Q Q. 2015. Domestic red jujube research present situation and developing trend of drying technology. Journal of Agricultural Mechanization Research, 8, 249–253.

Wang Y J, Nie J Y, Li Z X, Yan Z, Cheng Y, Zhang X N. 2017.Simultaneous determination of 16 kinds of mycotoxins in common dried fruits by UPLC-MS/MS. Chinese Journal of Analytical Chemistry, 45, 1556–1563. (in Chinese)

Wei D, Wang Y, Jiang D, Feng X, Li J, Wang M. 2017. Survey of alternaria toxins and other mycotoxins in dried fruits in China. Toxins, 9, 200.

WHO (World Health Organization). 2009. Principles and methods for the risk assessment of chemicals in food.Environmental Health Criteria 240. [2017-01-02]. http://www.who.int/foodsafety/chem/principles/en/index1.html

Yang J L. 2008. Investigation and risk assessment of ochratoxin a in mian foods in China. Ph D thesis, Northwest Agriculture and Forestry University, China. (in Chinese)

Yin H S, Meng Z X. 2017. Study on consumer demand and production supplies of jujube. Journal of Shanxi Agricultural Sciences, 43, 465–468. (in Chinese)

Zhang C. 2008. Investigation and risk assessment of aflatoxin B1 in mian foods of China. Ph D thesis, Northwest Agriculture and Forestry University, China. (in Chinese)

Zhao K, Shao B, Yang D, Li F, Zhu J. 2015. Natural occurrence of Alternaria toxins in wheat-based products and their dietary exposure in China. PLoS ONE, 10, e0132019.

Zwickel T, Klaffke H, Richards K, Rychlik M. 2016. Development of a high performance liquid chromatography tandem mass spectrometry based analysis for the simultaneous quantification of various Alternaria toxins in wine, vegetable juices and fruit juices. Journal of Chromatography (A),1455, 74–85.