Occurrence of benzene as a heat-induced contaminant of carrot juice for babies in a general survey of beverages

A survey of benzene contamination of 451 beverage samples, using headspace sampling combined with gas chromatography and mass spectrometry (HS-GC/MS) with a quantification limit of 0.13 µg l^?1, was conducted. Artefactual benzene formation during headspace sampling was excluded by gentle heating at 50°C only and adjustment of sample pH to 10. The incidence of benzene contamination in soft drinks, beverages for babies, alcopops and beer-mixed drinks was relatively low, with average concentrations below the EU drinking-water limit of 1 µg l^?1. Significantly higher concentrations were only found in carrot juice, with the highest levels in carrot juice specifically intended for infants. About 94% of 33 carrot juice for infants had detectable benzene levels, with an average concentration of 1.86?±?1.05 µg l^?1. Benzene contamination of beverages was significantly correlated to iron and copper concentrations, which act as catalyst in benzene formation. The formation of benzene in carrot juice was predominantly caused by a heat-induced mechanism, which explains the higher levels in infant carrot juices that are subject to higher heat-treatment to exclude microbiological contamination.     

Benzol in Lebensmitteln – ein Überblick

Benzene is an environmental contaminant and humans are mainly exposed by ambient air, but benzene may also occur in drinking water and foods. Benzene is carcinogenic to humans causing acute non-lymphocytic leukemia. In foods, critical benzene concentrations were detected in the past mainly in alcohol-free beverages, because benzene was formed from benzoic acid. For this reason, the soft-drink industry nowadays abstains from using this preservative, and our results show that benzene is not problematic anymore in this group of beverages (contents below the EU drinking water limit of 1 µg/L). During the monitoring of beverages, however, benzene was detected in carrot juices intended for babies and young children (average 2 µg/L). In this case, a heat-induced formation mechanism is predominant. Model experiments have proven that benzene may be formed by the thermally induced degradation of different precursors contained in carrots (e.g., carotene, phenylalanine, and flavour compounds). This mechanism also leads to benzene contamination in commercial carrot-containing baby menus (complimentary foods) as well as carrots in cans, while freshly pressed carrot juices and home prepared baby foods were found to be benzene-free. Significantly lower benzene concentrations were found in commercial baby menus sold in plastic containers compared to the conventionally used glass jars. Preliminary risk assessments have shown that even for babies the exposure from foods is below thresholds, above which a health risk would be assumed. For reasons of precautionary public health protection and the minimization principle for carcinogenic contaminants, the concentrations in carrot products should be reduced, especially as these products are intended for babies and children, the most sensitive consumer group. The focus should be set on improving the sterilization conditions.     

Ochratoxin A: an improvement clean-up and HPLC method used to investigate wine and grape juice on the Polish market.

A routine method appropriate for the determination of ochratoxin A (OTA) in wine, grape juice and grape juice drinks was described, and the occurrence of the mycotoxin was investigated in the most popular red wines, grape juice and grape juice drinks available on the Polish market. After clean-up on immunoaffinity column, samples were analysed by RP-HPLC using a fluorescence detector at 330 and 460 nm. The average OTA recoveries from spiked blank wine samples varied from 60 to 82%, and RSD% ranged from 5 to 14%. The OTA recovery for spiked grape juice and grape juice drinks were 80-100%, but the RSD% was between 7 and 10%. The limit of detection and limit of quantitation for all sample types were 0.5 and 2.0 ng l(-1), respectively. Fifty-three samples of red wine and seven samples of grape juice and grape drinks were assessed by means of this analytical procedure. OTA was detected in most wine samples (92%); its concentrations ranged from 2.2 to 6710 ng l(-1). In all grape juice and drink samples, OTA levels ranged from 1.6 to 64.7 ng l(-1).     

Improved method for the determination of benzene in soft drinks at sub-ppb levels

An automated, simple, and reproducible method based on isotope dilution headspace gas chromatography/mass spectrometry developed previously for the determination of benzene in soft drinks was further improved by adding sodium sulfate to samples, lowering the gas chromatography oven starting temperature to narrow benzene peak width, and increasing sample injection volume. This improved method had a lower detection limit (0.016 µg l^-1) and good repeatability, and was used in a follow-up survey to assess benzene levels in 139 samples of soft drink products. Benzene was detected in 67% of the 139 products tested. Compared with the previous survey, the average benzene concentrations in most products from this survey were lower, and only a few products had benzene at elevated levels.     

Improved method for the determination of benzene in soft drinks at sub-ppb levels

An automated, simple, and reproducible method based on isotope dilution headspace gas chromatography/mass spectrometry developed previously for the determination of benzene in soft drinks was further improved by adding sodium sulfate to samples, lowering the gas chromatography oven starting temperature to narrow benzene peak width, and increasing sample injection volume. This improved method had a lower detection limit (0.016 µg l^?1) and good repeatability, and was used in a follow-up survey to assess benzene levels in 139 samples of soft drink products. Benzene was detected in 67% of the 139 products tested. Compared with the previous survey, the average benzene concentrations in most products from this survey were lower, and only a few products had benzene at elevated levels.     

Migration of antimony from PET bottles into beverages: determination of the activation energy of diffusion and migration modelling compared with literature data

Plastics bottles made from polyethylene terephthalate (PET) are increasingly used for soft drinks, mineral water, juices and beer. In this study a literature review is presented concerning antimony levels found both in PET materials as well as in foods and food simulants. On the other hand, 67 PET samples from the European bottle market were investigated for their residual antimony concentrations. A mean value of 224 ± 32 mg kg(-1) was found, the median was 220 mg kg(-1). Diffusion coefficients for antimony in PET bottle materials were experimentally determined at different temperature between 105 and 150°C. From these data, the activation energy of diffusion for antimony species from the PET bottle wall into beverages and food simulants was calculated. The obtained value of 189 kJ mol(-1) was found to be in good agreement with published data on PET microwave trays (184 kJ mol(-1)). Based on these results, the migration of antimony into beverages was predicted by mathematical migration modelling for different surface/volume ratios and antimony bottle wall concentrations. The results were compared with literature data as well as international legal limits and guidelines values for drinking water and the migration limit set from food packaging legislation. It was concluded that antimony levels in beverages due to migration from PET bottles manufactured according to the state of the art can never reach or exceed the European-specific migration limit of 40 microg kg(-1). Maximum migration levels caused by room-temperature storage even after 3 years will never be essentially higher than 2.5 microg kg(-1) and in any case will be below the European limit of 5 microg kg(-1) for drinking water. The results of this study confirm that the exposure of the consumer by antimony migration from PET bottles into beverages and even into edible oils reaches approximately 1% of the current tolerable daily intake (TDI) established by World Health Organisation (WHO). Having substantiated such low antimony levels in PET-bottled beverages, the often addressed question on oestrogenic effects caused by antimony from PET bottles appears to be groundless.     

Evaluation of aluminium concentrations in samples of chocolate and beverages by electrothermal atomic absorption spectrometry

Samples of chocolate, cocoa, tea infusions, soft drinks and fruit juice have been examined by, electrothermal atomic absorption spectrometry (ETA-AAS) for the presence of aluminium (Al). Fruit juices and chocolate were analysed after an adequate sample preparation; the other products were evaluated directly. Sampling was performed in duplicate for 248 independent samples. The mean Al concentration in chocolate was 9.2 +/- 7.5 mg kg(-1), and individual values were correlated with the per cent of cocoa in samples (Y = 0.63 + 0.27X, r = 0.78, p < 0.0001). Al concentration in commercial tea infusions ranged from 0.9 to 3.3 mg l(-1) (mean = 1.80 +/- 65 mg l(-1), whereas in laboratory-prepared samples it was 2.7 +/- 0.93 mg l(-1). In soft drinks, the concentrations of Al were lower, ranging from 9.1 to 179 microg l(-1); the highest values were observed in samples of orange squash (mean = 114 +/- 56 microg l(-1)). Apricot juice showed the highest Al level (mean = 602 +/- 190 microg l(-1)), being statistically, different from that of pear (mean = 259 +/- 102 microg l(-1)), but not different from that of peach juice (mean = 486 +/- 269 microg kg(-1)). Toxicologically, the amount of Al deriving from the consumption of these products is far below the acceptable daily intake of 1 mg kg(-1) body weight indicated by the FAO/WHO, and it is a verv low percentage of the normal Al dietary intake.     

Copper, zinc, calcium and magnesium content of alcoholic beverages and by-products from Spain: nutritional supply.

Levels of copper, zinc, calcium and magnesium were measured in alcoholic beverages (whiskies, gins, rums, liquors, brandies, wines and beers) and by-products (non-alcoholic liquors and vinegars) using flame atomic absorption spectrometry (FAAS). Mineral concentrations were found to be significantly different between the nine alcoholic and non-alcoholic by-products studied (p < 0.001). In distilled alcoholic beverages, concentrations measured in rums and brandies were statistically lower than those determined in gins and alcoholic liquors (p = 0.001). For Cu, measured concentrations were statistically different for each of the five groups of distilled alcoholic beverages studied (p < 0.001). In fermented beverages, Zn, Ca and Mg levels were significantly higher than those concentrations determined in distilled drinks (p < 0.005). Contrarily, Cu concentrations were statistically lower (p < 0.001). Wines designated as sherry had significantly higher Ca and Mg levels (p < 0.005). White wines had significantly higher Ca and Zn levels (p < 0.05) compared with red wines and, contrarily, Cu concentrations were significantly lower (p < 0.005). In wine samples and corresponding by-products (brandy and vinegar), statistical differences were established for all minerals analysed (p < 0.01). Remarkably, for Cu, the concentrations determined in brandies were statistically higher. On the basis of element levels and the official data on consumption of alcoholic beverages and by-products in Spain, their contribution to the daily dietary intake (DDI) was calculated to be 124.6 microg Cu day(-1) and 193.3 microg Zn day(-1), 40.3 mg Ca day(-1) and 19.9 mg Mg day(-1). From all studied elements, Cu was the one for which alcoholic beverages constitute a significant source (more than 10% of recommended daily intake). These findings are of potential use to food composition tables.     

Incidence of patulin in apple juices marketed in Turkey.

The purpose of this study was to investigate the patulin contamination of apple juices consumed by the Turkish population. Patulin was detected using high-performance liquid chromatography (HPLC) with a UV detector at 280 nm, and the identification of patulin was further confirmed by thin-layer chromatography (TLC). Using HPLC, the recoveries were 79.9 +/- 6.7% and 83.7 +/- 4.6%, and the coefficients of variation were 8.4 and 5.5% for apple juices spiked with the known amounts of patulin (60 and 120 microg/liter. respectively). The minimum patulin level detected was 5 ng in a standard solution and 5 microg/liter in apple juices. The TLC method was used only to confirm patulin levels higher than 20 microg/liter (100 ng/spot) in apple juices. The total number of samples was 45. Patulin was present in detectable levels in 60% of apple juices at concentrations ranging from 19.1 to 732.8 microg/liter. Forty-four percent of the apple juice samples had patulin contamination levels higher than 50 microg/ liter, which is the allowable upper limit in Turkey.     

Quantification of allyl hexanoate in pineapple beverages and yogurts as a case study to characterise a source of uncertainty in dietary exposure assessment to flavouring substances

One source of uncertainty in the estimation of dietary exposure to flavouring substances is the uncertainty in the occurrence and concentration levels of these substances naturally present or added to foodstuffs. The aim of this study was to assess the variability of concentration levels of allyl hexanoate, considered as a case study, in two main food categories to which it is often added: pineapple juice-based beverages and yogurts containing pineapple. Thirty-four beverages and 29 yogurts, with pineapple fruit or juice and added flavourings declared as ingredients on the package, were purchased from the local market (in Rome) and analysed. Analytical methods based on the stir bar sorptive extraction (SBSE) technique for the isolation of the target analyte, and on GC-MS analysis for final determination, were developed for the two food categories. In beverages, allyl hexanoate concentrations ranged from less than 0.01 to 16.71 mg l(-1), whereas in yogurts they ranged from 0.02 to 89.41 mg kg(-1). Average concentrations in beverages and yogurts with pineapple as the main fruit ingredient (1.91 mg l(-1) for beverages, 9.61 mg kg(-1) for yogurts) were in fair agreement with average use level data reported from industry surveys for the relevant food categories (4.5 and 6.0 mg kg(-1), respectively). Within the group of yogurts a single product was found to contain a level of allyl hexanoate more than 10-fold higher than the average reported use level. The screening techniques developed by the European Food Safety Authority (EFSA) using use level data provided by industry gave estimates of exposure that were of the same order of magnitude as the estimates obtained for regular consumers who would be loyal to the pineapple yogurt and beverage products containing the highest observed concentration of the substance of interest. In this specific case the uncertainty in the results obtained with the use of standard screening techniques for exposure assessment based on industry reported use levels is low.     

Avoiding hydrolysis of fuel ether oxygenates during static headspace analysis

The determination of fuel ether oxygenates in groundwater was found to be problematic when samples are preserved at pH < 2 and then analyzed using heated headspace sampling. Acid catalyzed the hydrolysis of tert-amyl methyl ether, ethyl tert-butyl ether, and methyl tert-butyl ether during headspace sampling when aqueous samples were heated at 80 degrees C, a typical temperature used for heated headspace sampling. Hydrochloric acid at pH 2 did not cause hydrolysis of oxygenate ethers in samples stored for 28 d at 4 degrees C. When trisodium phosphate was used to preserve the sample or to adjust the pH of samples preserved with acid before headspace sampling, the recovery of spiked ethers was excellent. The heated headspace method was also applicable for the determination of other fuel oxygenates including ethanol, tert-butyl alcohol (TBA), tert-amyl alcohol (TAA), isopropyl alcohol (IPA), acetone, and monoaromatic compounds found in gasoline including benzene, toluene, ethylbenzene, xylenes, and trimethylbenzenes. The method detection limits range from 0.1 to 0.2 microg/L for the ethers and aromatics. For alcohols and acetone, the method detection limits were 0.8 microg/L for TBA, 18 microg/L for ethanol, 1.2 microg/L for TAA, 5.5 microg/L for IPA, and 3.3 microg/L for acetone. The heated headspace method yielded accurate results for ether oxygenates in samples containing a wide range of gasoline concentrations (2500-100000 microg/L).     

Determination of polycyclic aromatic hydrocarbons in alcoholic drinks and the identification of their potential sources.

Some polycyclic aromatic hydrocarbons (PAHs), particularly those with a high molecular mass, have been classified as probably being carcinogenic to humans by the International Agency for Research on Cancer (IARC). The significance of the determination of PAHs is reflected by the special attention the European Union is paying to regulating their maximum allowed levels in various types of foodstuffs. Like tobacco and smoked meats, alcoholic drinks can also contain these carcinogenic chemicals, as the latter have been detected in the charred insides of barrels, some ingredients such as caramel or the smoke released during the drying of germinated barley in beer or whisky. This paper determined the contents of seven PAHs in alcoholic beverages of variable alcoholic strength that had been aged in charred barrels for different times (months, years). The aim was to elucidate the dependence of the formation of seven PAHs on the type of charring (traditional or convective) used and the charring intensity (light, medium or heavy). Based on the results, the way the tree raw material is toasted strongly influences PAH levels in alcoholic drinks; thus, traditional charring produces increased amounts of PAHs from the wood relative to convective toasting. The sum of the analysed PAH concentrations in the aged alcoholic beverages studied ranged from zero for a white wine to 172 ng l(-1) for a 'brandy de jerez solera'. The carcinogenic indicator benzo[a]pyrene was found at concentrations below 10 ng l(-1). These PAH concentrations in alcoholic beverages are very low relative to those in smoked and char-broiled foodstuffs. Any health hazards, however, can be minimized by using convective toasting to manufacture the barrels where the drinks are to be aged.     

Survey of caffeine levels in retail beverages in Portugal.

The caffeine content of 85 retail beverage samples purchased from local supermarkets between 1995 and 2004 was determined. The potential intake of caffeine through the consumption of these beverages (but excluding coffee) was estimated for students of the University of Coimbra, Portugal. The caffeine content of the beverages ranged from 47.5 to 282.5 mg l(-1) for teas, from 20.1 to 47.2 mg l(-1) for tea extracts samples, and from 80.7 to 168.7 mg l(-1) for cola soft drinks. Caffeine was not completely absent from caffeine-free colas, and energy drinks had a far greater caffeine content than regular drinks, ranging from 21 to 2175 mg l(-1). Soft drinks were consumed by 72% of the individuals, although 14% of the survey participants did not drink any of the different types of the beverages studied. Contrary to expectations for this age group, no consumptions of energy drinks was reported. Daily caffeine intake was estimated to range from 4.7 to 200 mg day(-1), but with only 5% reporting a daily intake around 200 mg caffeine. Cola-type beverages were an important dietary source of caffeine for the population studied. Statistical differences in the caffeine intake between the male and female populations were found, with p = 0.014, being higher for the male population. Of the beverages studied, cola-type drinks showed statistical differences for the male population, p = 0.03, and tea showed statistical differences for female population p = 0.013, respectively.     

Colour and label evaluation of commercial pasteurised red juices and related drinks

Despite growing demand by consumers for healthy beverages, artificial colours are still widely used. Levels of anthocyanins and artificial colours were determined by HPLC with UV-Vis detection in red orange juices and other red beverages (nectar, juice-based, health, carbonated and sports drinks). The contribution of pigments to the visible colour of the beverage was calculated. Red orange juice samples contained about 34?mg?l^?1 of anthocyanins, which were responsible for about 92% of the visible colour. Red juice-based drinks, containing from 0% to 30% of red orange, berry, grape or pomegranate juices, had low levels of anthocyanins (about 7?mg?l^?1) and high levels of E129 (about 32?mg?l^?1), which were responsible for about 90.7% of the colour of these beverages. Red health drinks, enriched with vitamins and polyphenols, contained from 3% to 50% of red fruit juices. Also in this case the E129 levels were higher (about 22?mg?l^?1) than anthocyanins (about 9?mg?l^?1), and were responsible for the colour of the beverages (76.1%). High levels of artificial colours were found in red orange carbonated drinks, but in comparable amounts with those found in the other beverage samples, while anthocyanins were only present in trace amounts. Although all of the beverages claimed to contain red fruits on the labels, no correlation was found between the level of anthocyanins and the declared percentage of red fruits. These labels generally conformed with the requirements of the law, but food product labels can often be misleading to consumers about the real characteristics of the product.     

Predicted intake of trace elements and minerals via household drinking water by 6-year-old children from Kraków, Poland. Part 2: Cadmium, 1997-2001.

Between 1997 and 2001, samples of drinking water (DW) were collected in about 600 households and pre-schools by a double sampling method, early in the morning (stagnant water, W1) and in the evening (flushed water, W2). The study group included about 300 5-7-year-old children living in urban (metropolitan, U) and peripheral (P) areas of Kraków and rural (R) areas in southern Poland. Cadmium concentrations were estimated by the GF-AAS method with Zeeman background correction. Mean cadmium concentrations in water did not vary substantially by year. Moreover, there was no difference between P and R group (p > 0.05). The geometric mean (GM) and 95% confidence interval (CI) of cadmium concentrations in W2 was 0.13 (0.01-1.72) microg l(-1) for U group, while for combined (P&R) non-urban group (NU) it was 0.31 (0.02-6.24) microg l(-1) (p < 0.001). Significant differences between cadmium in W1 and W2 samples were evident from the ratio of cadmium concentration in W1 to W2, i.e. R1,2; its difference (W1 - W2), i.e. D1,2; or relative difference 100 x (W1 - W2)/W1, i.e. D1,2%. For NU sites D1,2, D1,2% and R1,2 values (GM) were 0.25 microg l(-1), 34% and 2.19, respectively, and for U sites 0.19 microg l(-1), 50% and 3.49, respectively. Parameter D1,2 was not site dependent (p > 0.05) and exhibited a plateau at 0.44 microg l(-1). The WHO and Polish limit of cadmium in DW (3 microg l(-1)) was exceeded in W2 samples from 0.5, 5.8 and 10.0% of households from the U, P and R areas, respectively. The predicted cadmium intake with W2 by children was (GM) 0.68 and 2.06% of PTWI for U and NU sites, respectively. However, these contributions were higher for W1 samples by coefficient R1,2. For the cases exceeding the cadmium limit, the predicted cadmium intake with DW contributed to PTWI by 27.4-29.1%. The study proved that contamination of water with cadmium in plumbing is significant, independent of the site type and period. This contamination should be considered as a factor in public health protection programmes with special attention to children. An easy approach to limit the intake of cadmium from DW by about half would be to discard some of the first drawn portion of water from pipes after overnight or longer break of usage.     

Evaluation of Accelerated UV and Thermal Testing for Benzene Formation in Beverages Containing Benzoate and Ascorbic Acid

ABSTRACT: Under certain conditions, benzene can form in beverages containing benzoic and ascorbic acids. The American Beverage Assn. (ABA) has published guidelines to help manufacturers mitigate benzene formation in beverages. These guidelines recommend accelerated testing conditions to test product formulations, because exposure to ultraviolet (UV) light and elevated temperature over the shelf life of the beverage may result in benzene formation in products containing benzoic and ascorbic acids. In this study, the effects of UVA exposure on benzene formation were determined. Benzene formation was examined for samples contained in UV stabilized and non-UV stabilized packaging. Additionally, the usefulness of accelerated thermal testing to simulate end of shelf-life benzene formation was evaluated for samples containing either benzoic or ascorbic acid, or both. The 24 h studies showed that under intense UVA light benzene levels increased by as much as 53% in model solutions stored in non-UV stabilized bottles, whereas the use of UV stabilized polyethylene terephthalate bottles reduced benzene formation by about 13% relative to the non-UV stabilized bottles. Similar trends were observed for the 7 d study. Retail beverages and positive and negative controls were used to study the accelerated thermal testing conditions. The amount of benzene found in the positive controls and cranberry juice suggests that testing at 40 °C for 14 d may more reliably simulate end of shelf-life benzene formation in beverages. Except for cranberry juice, retail beverages were not found to contain detectable amounts of benzene (<0.05 ng/g) at the end of their shelf lives.     

Fluoride content of soft drinks,nectars, juices,juice drinks,concentrates, teas and infusions marketed in Portugal

A potentiometric method using a fluoride combination ion-selective electrode was validated and used to analyse 183 samples, including soft drinks, juices, nectars, juice drinks, concentrates, teas and infusions marketed in Portugal. The fluoride levels were higher in extract-based soft drinks, juice drinks and juice, with fluoride values of 0.86 ± 0.35, 0.40 ± 0.24 and 0.37 ± 0.11 mg l^?1, respectively. The lowest fluoride concentration was found in infusion samples (0.12 ± 0.01 mg l^?1), followed by teas and carbonated soft drinks with fluoride concentrations of 0.16 ± 0.12 and 0.18 ± 0.07 mg l^?1, respectively. Nectars, concentrates and juice-based drinks had similar fluoride concentrations of 0.33 ± 0.16, 0.29 ± 0.12 and 0.25 ± 0.14 mg l^?1, respectively. The fluoride concentrations in all these samples would only contribute intakes below the acceptable daily intake (ADI = 0.05 mg kg^?1 body weight day^?1), indicating that, individually, these beverages cannot induce fluoride toxicity in the population group of children.     

Detection of patulin in apple juices marketed in the Tohoku district, Japan

Patulin is a mycotoxin mainly produced by Penicillium and Aspergillus. We investigated the incidence of patulin contamination in 179 samples of apple juice and 9 samples of mixed juice (containing apple juice concentrate as an ingredient) commercially available in the Tohoku district of Japan. Patulin was detected in 3 of 143 samples containing domestic fruits and in 6 of 45 samples containing imported products and products produced in Japan using imported apple juice concentrate. Patulin analyses were carried out using high-pressure liquid chromatography with a detection limit of 4 microg/liter. The patulin content of contaminated domestic samples (three samples with concentrations ranging from 6 to 10 microg/liter), imported samples (one sample with a concentration of 15 microg/liter), and domestic samples produced containing imported concentrate (five samples with concentrations ranging from 6 to 9 microg/liter) was lower than the maximum limit of 50 microg/liter currently adopted by many countries, including Japan.     

Occurrence and stability of inorganic and organic arsenic species in wines, rice wines and beers from Central European market

We investigated in total 80 wine samples of different types and seven grape juice and 23 beer samples purchased from markets in Central Europe in order to understand the arsenic (As) speciation and help assess the potential As toxicity via intake of alcoholic beverages. Generally, total As concentrations in most samples investigated were below the drinking water limit 10 μg l(-1) published by the World Health Organization (WHO); ranging from 0.46 to 21.0 μg l(-1) As in red and white wines and from 0.75 to 13.4 μg l(-1) As in beers. In addition, concentrations of total As in rice wine and in rice beer were 0.63-6.07 and 3.69-8.23 μg l(-1) As, respectively. The total As concentrations in ice wine ranged from 7.94 to 18.8 μg l(-1) As, significantly higher than in white and red wine. Arsenite predominated as the As species in most of the wine samples, whereas arsenate was the dominant species in rice wine, beer and rice beer. Methyl As components were usually minor components in all wine and beer samples. Monomethylarsonic acid, dimethylarsinic acid and two additional unknown As species were frequently found in grape juice, late harvest and ice wine with higher sweetness. After air exposure, arsenite, arsenate, monomethylarsonic acid and dimethylarsinic acid were stable at 4°C for months, probably due to the acidic conditions of wine and beer samples. The presence of sulfite had little influence on As speciation in wine. Despite the predominance of more toxic arsenite and arsenate in wine and beer, the estimated weekly exposure to As (via consumption of beer, wine and rice wine) is low. The As intake per capita is 6.81 μg from beer, <1.93 μg from wine and 0.88 μg from rice wine, estimated using the median of total As concentration multiplied by the average consumption per capita of the corresponding beverage.