Simultaneous determination of bisphenol A-diglycidyl ether, bisphenol F-diglycidyl ether and their hydrolysis and chlorohydroxy derivatives in canned foods

A new method for the simultaneous determination of bisphenol A-diglycidyl ether (BADGE), bisphenol F-diglycidyl ether (BFDGE) and their hydrolysis and chlorohydroxy derivatives in canned foods is presented. Oily and aqueous food samples were extracted with tert-butyl methyl ether and acetonitrile, respectively. The compounds in both extracts were determined by using reverse-phase gradient high-performance liquid chromatography with fluorescence detection. Optimization of extraction and chromatographic determination is outlined in detail. After validation the method was used to analyze various canned food samples, such as tuna and sardine in oil, vegetables, fruit cocktails, etc. In none of the samples were significant amounts ( >100 μg/kg) of BADGE or BFDGE found, whereas in most samples BADGE/BFDGE chlorohydroxy compounds were detected. These originate most probably from the use of organosol varnishes instead of epoxy resins. Risk assessment and regulations of these compounds by the European Union are urgently needed. Additionally, the syntheses and characterization of the not available standard compounds bisphenol A-p-glycidyl-p′-(3-chloro-2-hydroxypropyl) ether (BADGE.HCl) and bisphenol A-p-(2,3-dihydroxypropyl)-p′-(3-chloro-2-hydroxypropyl) ether (BADGE.HCl.H₂O) are presented. Received: 28 July 1999 / Revised version: 29 October 1999     

Stability of bisphenol A diglycidyl ether and bisphenol F diglycidyl ether in water-based food simulants

Varnishes used for the inner coatings of food cans are often based on epoxy resins or vinylic organosols. The epoxy resins can be produced from bisphenol A (BPA) and bisphenol F (BPF), and these also contain bisphenol A diglycidyl ether (BADGE) or bisphenol F diglycidyl ether (BFDGE) as stabilising components. These compounds may break down during storage and also by the influence of food simulants. The stability of BADGE and BFDGE were studied using reverse-phase gradient high-performance liquid chromatography (RP-HPLC) with ultraviolet detection (UV). Three experimental conditions for spiked simulants were compared: (1) the storage at 25 °C (C1), (2) the storage at 40 °C (C2) and (3) the storage at 25 °C after 15 min heating at 120 °C (C3). Distilled water, 3% acetic acid and 10% ethanol were used as food simulants. It was observed that BADGE is more stabile than BFDGE. The loss of BADGE and BFDGE were minimal in 10% ethanol (39 and 46% at 25 °C, 60 and 69% at 40 °C, respectively) and highest in 3% acetic acid (60 and 63% at 25 °C, 76 and 82% at 40 °C, respectively). At experiment (C3), the hardest conditions, significant degradation was not shown in comparison with conditions (C1) and (C2), contrariwise BADGE and BFDGE in 10% ethanol were minimal degradated at conditions (C3) from all these experiments (loss of 5 and 8%, respectively).     

A LC-MS/MS method for the determination of BADGE-related and BFDGE-related compounds in canned fish food samples based on the formation of [M+NH(4)](+) aducts

A new and simple liquid chromatography tandem mass-spectrometry method for the determination of different bisphenol A (BPA) derivatives such as bisphenol A diglycidyl ether (BADGE), bisphenol F diglycidyl ether (BFDGE) and their reaction products with water and hydrochloric acid in different fish food products was developed. The extraction procedure and the chromatographic conditions were optimised for complex food matrices such as fish products. Food samples were homogenised and extracted with a 1:1 solution of acetonitrile-hexane, the solvent was eliminated in a N(2) stream and the extract was reconstituted with 0.5mL of a 0.01M solution of ammonium formate. The sample solution obtained was directly measured by LC-MS/MS without any further purification under the developed conditions. The use of a mobile phase composed by ammonium formate-methanol in a binary gradient mode produced [M+NH(4)](+) aducts for the different BADGEs and BFDGEs. These aduct's fragmentations were employed for the LC-MS/MS quantification of BPA derivatives in canned fish samples. The results of the validation were appropriate: the method was linear for BADGE and its hydrolysed derivatives up to 1000μgkg(-1), for the remaining compounds linearity achieved up to 100μgkg(-1). Quantification limits were in the range 2-10μgkg(-1). RSD (intra and inter-day) was 6-12% and the recovery was comprised between 89% and 109%. Under the optimised conditions, the chromatographic separation was performed in 8min per sample. The method was applied to the determination of BADGE, BFDGE and their reaction products in different samples of canned fish from Spanish origin. Migration results obtained were in compliance with the EU regulations.     

金属食品罐内涂层中双酚类物质的迁移及检测研究进展

鉴于双酚A及其他双酚类物质作为食品罐内涂层材料有潜在的毒性与危害,中国、欧盟、美国等国家均已严格限制它们在金属食品罐内涂料中的使用。本文介绍双酚A、双酚A二缩水甘油醚、双酚F、双酚F二缩水甘油醚的结构、应用、危害、检测方法及迁移结果,并简述固相萃取在样品前处理中的应用。     

肉类罐头食品中双酚A二缩水甘油醚及其衍生物的迁移

本实验应用高效液相色谱-电喷雾串联质谱法分析检测肉类罐头食品中双酚A二缩水甘油醚(BADGE)及其衍生物的残留,重点研究了肉类罐头中的内容物、存储时间(6个月、9个月、12个月)以及存储温度(4、20、100℃)对双酚A二缩水甘油醚(BADGE)及其衍生物的迁移量的影响,单因素方差分析(One-way ANOVA)检验实验结果表明,不同内容物罐头中从内壁涂层迁移至样品的BADGE及其衍生物存在显著性差异(p<0.05),Student-Newman-Keuls法检验表明存储12个月后目标化合物的迁移量与6个月、9个月存在显著性差异,存储温度为4、20℃时化合物迁移量没有显著差异,但是罐头加热到100℃后目标化合物迁移量是最大的。     

食品金属罐中14种双酚类物质迁移量的同时测定及杀菌过程对其迁移的影响

为了调查国内食品金属罐中双酚类物质的迁移水平,本文建立了同时测定食品金属罐中14种双酚类物质迁移量的高效液相色谱法,并研究了杀菌过程对双酚类物质迁移的影响及食品金属罐中双酚类物质的迁移规律。结果表明,14种双酚类物质在4种食品模拟物（4%体积乙酸、10%体积乙醇、50%体积乙醇和异辛烷）中线性关系良好,加标回收率为83.67%～107.05%,精密度范围为2.32%～7.67%,该方法的精密度和准确度较好,可用于食品金属罐中14种双酚类物质迁移量的检测分析。对市场上9种罐头和饮料用金属罐中的14种双酚类物质的迁移量进行了同时测定,发现有7种双酚类物质检出,无双酚A及其类似物的检出。基于欧盟限量标准要求,发现午餐肉罐头中的双酚A-（2-3-二羟基丙基）缩水甘油醚（BADGE·H₂O）、牡蛎葛根饮料中的双酚A-二（2-3-二羟基丙基）醚（BADGE·2H₂O）和黑莓罐中的双酚F-（2-3-二羟基丙基醚）（BFDGE·2H₂O）存在一定的迁移风险。此外,杀菌过程影响食品金属罐中双酚类物质的迁移种类和迁移量。迁移试验结束后,经过杀菌的金属罐中仅检测出BADGE·2H₂O,而未经杀菌的金属罐中同时检测到BADGE·2H₂O和双酚A-（3-氯-2羟丙基）（2-3-二羟基丙基）缩水甘油醚（BADGE·H₂O·HCl）。另外,食品模拟物的类型影响着双酚类物质的迁移量和迁移种类,且随着迁移温度的升高和迁移时间的增加,双酚类物质的迁移量逐渐升高直至平衡。     

Food and beverage can coatings: A review on chemical analysis,migration, and risk assessment

The internal surface of food and beverage cans is generally covered with polymeric coatings to preserve food and protect metal substrate from corrosion. Coating materials are complex formulations that contain different starting substances (e.g., monomers, prepolymers, additives, etc.) and in addition during the manufacture of the material several compounds can be formed (e.g., reaction products, degradation products, etc.). These substances have the potential to migrate into the food. Many of them have not been identified and only some have been toxicologically evaluated. This article aims to provide a comprehensive review on the analytical methods used for the identification of potential migrants in can coatings. The migration and exposure to chemicals migrating from can coatings are also reviewed and discussed so far, which is essential for risk assessment. Moreover, a brief section on the current status of the legislation on varnishes and coatings for food contact in Europe is also presented. Liquid chromatography coupled to diode array and fluorescence detectors and particularly to mass spectrometry and gas chromatography–tandem mass spectrometry seem to be the techniques of choice for the identification of potential migrants in can coatings. Some studies have reported migration levels of BPA (bisphenol A) and BADGE (bisphenol A diglycidyl ether) and derivatives exceeding the specific migration limits set in the European legislation. On the whole, low dietary exposure to migrants from can coatings has been reported. However, it is interesting to highlight that in these studies the combined exposure to multiple chemicals has not been considered.     

Migration from can coatings: Part 3. Synthesis, identification and quantification of migrating epoxy-based substances below 1000 Da

Bisphenol A-derived glycidyl ethers as well as its reaction products with other lacquer components can migrate into the packed food from epoxy-based can coatings. A sensitive and selective method is presented using high-performance liquid chromatography coupled with ultraviolet light, fluorescence and electrospray ionization-mass selective detection for the identification and quantification of all migrants with a bisphenol A backbone and a molecular weight below 1000 Da, an estimated boundary for the absorption in the gastrointestinal tract. The identification of migrants was confirmed by microreactions of technical bisphenol A diglycidyl ether with solvents and phenols, which provided the fragmentation pattern of the mass selective detection and relative retentions of 42 different bisphenol A-related substances. It was shown by calibration of different isolated and synthesized bisphenol A derivatives that the fluorescence response relies on the amount of bisphenol A moiety in the respective molecule. Therefore, all migrating bisphenol A-related substances below 1000 Da were determined as bisphenol A diglycidyl ether equivalents using a calibration (fluorescence detection) of the commercially available bisphenol A diglycidyl ether monomer. The limit of quantification was set at 5 μg bisphenol A diglycidyl ether equivalents kg^-1 (or 0.8 μg dm^-2). This method was validated for epoxy coatings (0.1 μg dm^-2 limit of detection and 24 μg bisphenol A-related substances below 1000 Da dm^-2 standard deviation, corresponding to 4.4% relative standard deviation). The quantification could be extended by combining the fluorescence response and structural information gained from the mass spectra, which provides more accurate results for each migrant. The calculation is based on the calibration of the bisphenol A chromophore content of the molecule. According to this method, the amount of migrating bisphenol A-related substances below 1000 Da in the acetonitrile extract (assuming a worst case) varied from about 0.4 to 0.7 mg dm^-2 in the examined coatings. The determined amounts comply with about 50% of the total migrate below 1000 Da.     

Development and Validation of a LC-FD Method for the Simultaneous Determination of Eight Bisphenols in Soft Drinks

Bisphenol A (BPA) is a chemical widely used as a monomer in the production of polymers of plastics. It acts as an endocrine-disrupting agent and thus its contaminations of food and beverage should be carefully monitored in order to assess consumers’ risk. In this study, we propose a liquid chromatography-fluorescence detection (LC-FD)-validated method for the simultaneous determination of BPA and seven analogues, i.e., bisphenol AF, bisphenol B, bisphenol E, bisphenol F, BPA diglycidyl ether, bisphenol F diglycidyl ether, and Bisphenol M in soft drinks. A one-step solid-phase extraction (SPE) was effective at reducing the interferences, obtaining good purification of the samples, and consequently good recoveries of all analytes. The separation was obtained on a C18 column by using acetonitrile/water 55:45 (v/v) under isocratic conditions. Method validation was performed according to the European Commission Decision 2002/657/EC criteria, providing good results regarding the analytical parameters of linearity, selectivity, sensitivity, precision, recovery, decision limit (CCα), detection capability (CCβ), limit of detection (LOD), limit of quantification (LOQ), stability, and robustness. The method allows the detection of BPA and BADGE at levels much lower than their legal limits in the food, which are 0.06 and 9.00 mg kg^?1, respectively.     

食品包装材料中有害物质双酚A二缩水甘油醚 体外代谢研究

目的通过模拟体内代谢,对双酚A二缩水甘油醚(BADGE)的体外基本代谢情况进行研究。方法采用肝微粒体、肝S9 2种体外代谢试剂,通过模拟体内肝脏代谢,对BADGE的代谢行为及其代谢产物进行研究。通过代谢试剂浓度及代谢时间条件的优化,采用高效液相色谱-串联质谱(HPLC-MS/MS)作为检测手段对BADGE的体外代谢产物进行分析确证。结果体外代谢最佳孵化时间为60 min,最佳体外代谢试剂浓度为0.5mg/m L,在肝S9及肝微粒体2种体外代谢试剂的作用下,BADGE发生显著的代谢反应。结论本研究与传统的动物试验相比,节约了时间、精力,对食品包装材料的毒理学研究和安全性评价有重要的推动作用。     

超高效液相色谱-串联质谱法同时测定鱼、肉类罐头食品中的BADGE、NOGE及其衍生物

建立了一种罐头食品中的BADGE(bisphenol A diglycidyl ether,双酚A-二环氧甘油醚)、NOGE(novolacglycidyl ether,酚醛清漆甘油醚)及其衍生物含量的超高效液相色谱-串联质谱检测法。前处理包括正己烷/丙酮微波辅助萃取,Varain-PS-DVB固相萃取柱净化等。样品经BEH C18色谱柱(100 mm×2.1 mm,1.7μm)分离,以乙腈和0.2%甲酸水为流动相,进行梯度洗脱,在API 4000-QTRAP质谱仪电喷雾正离子、多反应监测(MRM)模式下进行检测。分析物检出限达到0.010 2 ng/g到0.197 2 ng/g,三水平加标回收率在65.7%～99.1%。     

Improved sample extraction and clean-up for the GC-MS determination of BADGE and BFDGE in vegetable oil.

A straightforward method was established for the determination of migration contaminants in olive oil with a special focus on the two can-coating migration compounds bisphenol A diglycidyl ether (BADGE) and bisphenol F diglycidyl ether (BFDGE). The preferred sample preparation was a single liquid-liquid extraction of compounds from the oil into 20% (v/v) methanol in acetonitrile, followed by clean-up with solid-phase extraction on aminopropyl bonded to silica. This purification procedure selectively removed all free fatty acids from the extracts without removing phenolic compounds of interest. The solid-phase extraction columns were used many times by implementing a procedure of washing out the strongly retained fatty acids with 2% acetic acid in methanol. Gas chromatography coupled with full scan (m/z 33-700) electron ionization mass spectrometry was used for the determination of several model compounds in olive oil samples. BADGE and BFDGE could be determined in the 0.05-2 mg kg(-1) range in oil samples with a relative SD of <6% (six replicates). The method was used in an enforcement campaign for the Norwegian Food Control Authority to analyse vegetable oil samples from canned fish-in-oil.     

Migration of bisphenol-A diglycidyl ether (BADGE) and its reaction products in canned foods

Bisphenol-A diglycidyl ether (BADGE) is used as an additive or starting agent in coatings for cans. The presence of hydrochloric acid in the organosol (PVC-based) lacquers results in formation of chlorohydroxy compounds of BADGE. These compounds, as well as BADGE itself, are potential migrants into the preserved food and are of toxicological concern. In the present investigation the presence of BADGE and the chlorohydroxy compounds (BADGE.HCl and BADGE.2HCl) in various kinds of canned foods from 30 brands have been determined by HPLC with fluorescence detection. BADGE was found in levels up to 5.1mg/kg in the food and only in food from cans containing BADGE.HCl and BADGE.2HCl in the lacquers. BADGE was found both in fish in oil and in fish in tomato sauce, however, the highest amounts were found in the fatty foodstuffs. BADGE.HCl and BADGE.2HCl were found in concentrations up to 2.4mg/kg and 8.3mg/kg, respectively. Unlike BADGE, BADGE.2HCl was found in similar concentrations in fish in oil and in fish in tomato sauce. In aqueous and acidic foodstuffs BADGE readily hydrolyses into mono- and dihydrolysed products (BADGE.H₂O and BADGE.2H₂O). In this study BADGE.H₂O was not found in any food sample, whereas BADGE.2H₂O was found in levels up to 2.6mg/kg. The Scientific Committee for Food (SCF) of the European Commission has proposed that a limit of restriction of 1mg/kg food shall include BADGE itself and BADGE.H₂O, BADGE.HCl, BADGE.2HCl and BADGE.HCL.H₂O. The present results indicate that the migration of BADGE.HCl and BADGE.2HCl, compounds with almost no data on toxicity, implies a greater problem than BADGE.H₂O and BADGE.2H₂O.     

Are Canned Beverages Industries Progressively Switching to Bisphenol AF?

Seven bisphenols, endocrine‐disruptor chemicals, were analytically determined for risk assessment in 52 large‐consumption beverages collected from the Italian market. The analytes under examination were bisphenol A, bisphenol F, bisphenol E, bisphenol B, bisphenol AF, bisphenol A diglycidyl ether, and bisphenol M. The concentration levels of all bisphenols detected ranged from <LOQ to 1,358 ng/mL in beers and from <LOQ to 76 ng/mL in energy drinks. The results of this monitoring study demonstrate the high presence of some congeners in beers, such as bisphenol AF, for which a European regulation is not yet available. Although the concentrations of the investigated bisphenols and that are under European regulations for migration into the food (Bisphenol A and BADGE) resulted below the legal limits in all screened beverages, the importance of their presence in foods should not be underestimated. Indeed, the safety of these analogues has not entirely been demonstrated and they could contribute to the total daily intake of endocrine disruptors, with special regards to specific demographics.     

Bisphenol A (BPA) and its source in foods in Japanese markets

The determination of bisphenol A (BPA) and/or bisphenol A diglycidyl ether (BADGE) in foods sold in Japanese markets and in water leached from six epoxy resin cans with similar diameters was carried out using high-performance liquid chromatography (HPLC) with electrochemical detection (LC/ECD), LC-mass spectrometric detection (LC/MS) and LC-tandem mass spectrometric detection (LC/MS/MS). BPA concentrations were 0-842 ng g^-1 for 48 canned foods, 0-14 ng g^-1 for 23 foods in plastic containers, and 0-1 ng g^-1 for 16 foods in paper containers. No BADGE was detected in three canned foods. There was no difference in leaching concentrations of BPA into glycine buffers at pHs 8 and 11, and water. The amounts of BPA leached into water from six epoxy resin cans held at 121°C for 20 min were almost the same as the cans' contents and were much higher than the amounts leached from cans held at or below 80°C for 60 min. The amount leached depended on the type of can, but not on the amount of BADGE leached from the cans. Considerably more BPA than BADGE leached to water from six cans. Two cans whose contents had high concentrations of BPA showed no BADGE leaching even at 121°C, suggesting the different kinds of epoxy resin can linings from others. The results imply that the main source of human exposure to BPA is food from cans with linings that contain high percentages of BPA as an additive or an unforeseen contaminant.     

Compliance work for food contact materials: feasibility of the legally required safety assessment of an epoxy/amine-based coating for domestic water pipe restoration

Options were explored for fulfilling the legally required safety assessment for a widely applied epoxy/amine coating used for restoring corroded domestic drinking water supply systems. The coating was made up of two components mixed shortly before application, the first mainly consisting of bisphenol A diglycidyl ether (BADGE), the second of various amines. The analytically identified starting substances were all authorised, but only constituted a small proportion of the low molecular mass material left after curing and potentially migrating into water. Reaction products synthesised from constituents of the starting components (expected oligomers) could not be eluted from GC even after derivatisation, indicating that standard GC-MS screening would miss most potential migrants. They were detectable by size exclusion chromatography (SEC) after acetylation. HPLC with MS or fluorescence detection was possible for constituents including a BADGE moiety, but phenalkamines could not be detected with adequate sensitivity. Possibilities for determining long-term migration relevant for chronic toxicity are discussed. Analysis in water shortly after application of the coating overestimates migration if migration decreases over time and requires detection limits far out of reach. Analysis of a solvent extract of the coating is easier and provides an upper estimate of what could migrate into the drinking water over the years. However, to satisfy the regulatory requirements, components of the complex mixture need to be identified at lower proportions than those accessible. In vitro testing of the whole mixture for genotoxicity is expected to fail because of the required sensitivity and the glycidyl functions probably wrongly resulting in positive tests. The difficulties in dealing with this situation are discussed.     

Bisphenol A diglycidyl ether (BADGE) migrating from packaging material 'disappears' in food: reaction with food components.

Bisphenol A diglycidyl ether (BADGE) is widely used as a monomer for coatings and adhesives for food-contact applications. Previous publications indicate that, after migration from packaging into foodstuffs, BADGE undergoes various reactions with unidentified food components. In order to elucidate the fate of BADGE, losses were determined after incubation with different foodstuffs and food components. Food proteins were identified as the main reaction partner with BADGE. Adduct formation was found with nucleophilic side-chains of amino acids. In vitro, cysteine exhibited significant activity. The previously reported occurrence of methylthio-derivatives of BADGE in foodstuffs was shown to originate from the reaction of BADGE with methionine. BADGE-methylthio derivatives can, therefore, be used as marker substances in foodstuffs for protein reactions with BADGE. The reported results offer a new viewpoint on the evaluation of BADGE migration. The hydrolysis and hydrochlorination derivatives subject to European legislation make up only a fraction of the totally migrated BADGE, and a further concern is that the toxic or allergenic potential of the protein adducts are unknown.     

Bisphenol A diglycidyl ether (BADGE) migrating from packaging material ‘disappears’ in food: reaction with food components

Bisphenol A diglycidyl ether (BADGE) is widely used as a monomer for coatings and adhesives for food-contact applications. Previous publications indicate that, after migration from packaging into foodstuffs, BADGE undergoes various reactions with unidentified food components. In order to elucidate the fate of BADGE, losses were determined after incubation with different foodstuffs and food components. Food proteins were identified as the main reaction partner with BADGE. Adduct formation was found with nucleophilic side-chains of amino acids. In vitro, cysteine exhibited significant activity. The previously reported occurrence of methylthio-derivatives of BADGE in foodstuffs was shown to originate from the reaction of BADGE with methionine. BADGE-methylthio derivatives can, therefore, be used as marker substances in foodstuffs for protein reactions with BADGE. The reported results offer a new viewpoint on the evaluation of BADGE migration. The hydrolysis and hydrochlorination derivatives subject to European legislation make up only a fraction of the totally migrated BADGE, and a further concern is that the toxic or allergenic potential of the protein adducts are unknown.     

Bisphenol A diglycidyl ether (BADGE) migrating from packaging material 'disappears' in food: reaction with food components

Bisphenol A diglycidyl ether (BADGE) is widely used as a monomer for coatings and adhesives for food-contact applications. Previous publications indicate that, after migration from packaging into foodstuffs, BADGE undergoes various reactions with unidentified food components. In order to elucidate the fate of BADGE, losses were determined after incubation with different foodstuffs and food components. Food proteins were identified as the main reaction partner with BADGE. Adduct formation was found with nucleophilic side-chains of amino acids. In vitro, cysteine exhibited significant activity. The previously reported occurrence of methylthio-derivatives of BADGE in foodstuffs was shown to originate from the reaction of BADGE with methionine. BADGE-methylthio derivatives can, therefore, be used as marker substances in foodstuffs for protein reactions with BADGE. The reported results offer a new viewpoint on the evaluation of BADGE migration. The hydrolysis and hydrochlorination derivatives subject to European legislation make up only a fraction of the totally migrated BADGE, and a further concern is that the toxic or allergenic potential of the protein adducts are unknown.     

Determination of Bisphenol A diglycidyl ether (BADGE) and its hydrolysis products in canned oily foods from the Austrian market

 Bisphenol A diglycidyl ether (BADGE) is determined in canned oily foods from Austria using a new simplified HPLC method. Samples are extracted with pentane, back extracted with methanol, and finally dissolved in the mobile phase (cyclohexane/tert–butyl methyl ether). Separation is performed on a normal-phase HPLC column using fluorescence detection. Verification of the BADGE-containing peak is carried out by using off-line GC-MS. Additionally, the synthesis and determination of BADGE hydrolysis products, Bisphenol A bis(2,3-dihydroxypropyl) ether (BADGE.2H₂O) and Bisphenol A glycidyl (2,3-dihydroxypropyl) ether (BADGE.H₂O) are presented. From 67 analyzed cans, containing various fatty meat or fish products, 16% were above the maximum quantity of 1 mg/kg tolerated by the European Community, 45% were in the range between 0.1–1 mg/kg, 24% between 0.02 and 0.1 mg/kg, and in 15% the BADGE concentrations were below the detection limit of 0.02 mg/kg. The hydrolysis product BADGE.H₂O was not detected in any sample, whereas BADGE.2H₂O was found in some samples up to a concentration of 0.5 mg/kg. Received: 11 May 1998 / Revised version: 1 July 1998