Investigation of food products containing garlic or onion for a false positive sulphite response by LC-MS/MS

ABSTRACT

In the US, sulphites must be declared on the label if they are present in concentrations greater than 10 mg/kg (determined as) SO₂ because an allergic-like response has been reported in a small subset of the population upon consumption of sulphite-containing products. The most widely used method for sulphite determination, the optimised Monier-Williams (OMW), produces false positive results with vegetables from the Allium (garlic) and Brassica (cabbage) genera due to extraction conditions that are thought to cause endogenous sulphur compounds to release SO₂. Recently, an LC-MS/MS method was developed for sulphites but has only been tested with samples that are 100% Allium or Brassica. Since regulatory samples may contain these vegetables as ingredients, additional investigations were necessary to determine the potential extent of false positives. Four blank matrices, chips, phyllo shells, hummus, and quinoa were spiked with various concentrations of onion and garlic powders. The sulphite concentrations were determined using an LC-MS/MS method. The matrix is extracted with a buffered formaldehyde solution, converting free and reversibly bound sulphite to the stable formaldehyde adduct, hydroxymethylsulfonate (HMS). It was determined that even at concentrations up to 8% garlic powder or 2% onion powder, the measured sulphite concentration was below the 10 mg/kg SO₂ labelling threshold. Commercial dried garlic powders were evaluated to determine the variation in responses that might be encountered in future regulatory samples. Recovery studies were conducted to determine if these methods would detect added sulphite. The ability to eliminate false positives due to these ingredients will result in a greater reliability in the accurate determination of added sulphite to ensure compliance with labelling requirements.     

Determination of Sulphiting Agents in Raw and Processed Meat: Comparison Between a Modified Monier-Williams Method and the Direct Analysis by Ion Chromatography with Conductometric Detection

Sulphiting agents are a class of compounds that are used as preservatives since they release SO₂ in food. Regarding meat products, the legislation restricts the use of these food additives, due to some toxic effects that they may have in humans. The most employed analytical procedure for the determination of sulphiting agents in foodstuffs is the Monier-Williams (M-W) method, but the reliability of this method was called into question by several authors. In this work, the M-W method was modified by replacing both the distillation unit to shorten the extraction time (from hours to 5 min) and the final titration with a chromatographic separation followed by conductometric detection of sulphate ion (m-M-W/IC-CD). This method was then validated, and the performance parameters were compared with those of the method based on the direct analysis of sulphite ion by ion chromatography with conductometric detection (DIC-CD). Linearity, accuracy at 40 and 80 mg kg^?1 of SO₂ and measurement uncertainty resulted comparable. Accuracy at 10 mg kg^?1 of SO₂ resulted higher for the m-M-W/IC-CD method, but this parameter could be influenced by traces of other sulphur-containing compounds that may be present in meat. The limit of determination of the m-M-W/IC-CD method was slightly higher than that obtained by the DIC-CD method. Finally, through spiking tests, it was proved that sulphide, 2-methyl-3-furanthiol and l-methionine cause “false-positive” responses, by using M-W-based methods.     

Carrageenan analysis. Part 2: Quantification in swine-adapted infant formula

An LC-MS/MS method was validated in accordance with 21CFR Part 58 Good Laboratory Practice for Non-clinical Laboratory Studies to measure the concentration of carrageenan in dose formulations used in a 28-day piglet dietary feeding study of swine-adapted infant formulations stabilised with carrageenan. Carrageenan concentrations in the test formulations were 0, 300, 1000 and 2250 mg kg^–1 formula. The method for the measurement of carrageenan was LC-MS/MS coupled with ESI in negative-ion mode for detection. Linearity was established over the range 1.00–7.50 µg ml^–1. Carrageenan dose formulation samples ranging from 0 to 2250 mg kg^–1 of carrageenan were diluted to within the linearity range for measurement.     

Improvement of wine aromatic quality using mixtures of lysozyme and dimethyl dicarbonate,with low SO2 concentration

The use of sulphur dioxide (SO₂) in the treatment of foodstuffs presents some problems as it could lead to pseudo-allergies in some people. The aim of this research work was to study the addition of different preservative mixtures and their influence on the concentration of volatile compounds and sensorial quality in wine. To do so, vinifications were carried out using Garnacha must to which lysozyme, dimethyl dicarbonate (DMDC) and mixtures of these with SO₂ were added at different doses (25 and 50 mg l^?1). The results were compared with a control sample to which only SO₂ had been added (50 mg l^?1). In general, mixtures of SO₂ with lysozyme and DMDC favoured the formation of volatile compounds in the wines. Wines obtained from the mixtures of lysozyme and DMDC with 25 mg l^?1 of SO₂ had better sensorial quality than the wines obtained with 50 mg l^?1 as the only preservative used.     

Measurement of ampicillin residue levels in chicken eggs during and after medicated feed administration by LC-MS/MS

A simple and sensitive LC-MS/MS method was developed and validated for the determination of ampicillin (ABPC) in chicken eggs. Residues were extracted by reverse-phase solid-phase extraction. Chromatographic separation was performed using a reverse-phase column with an elution gradient. The limits of detection and quantification were 0.01 and 0.1 ng g^?1, respectively. For the 0.1–50 ng g^?1 concentration range, mean recovery and accuracy values were 93.9–98.5% and 100.2–118.0%, respectively. ABPC residue concentrations in eggs before, during and after 7 days of medicated feeding of maximum dosage (40 mg kg^?1 body weight day^?1) of ABPC were determined with the LC-MS/MS method. The maximum concentration of ABPC in eggs was 3.6 ± 1.7 ng g^?1 (mean ± SD) on the last day of the administration period. Residue concentrations of ABPC in eggs during and after ABPC administration were not over the Japanese maximum residue limit of 0.01 mg kg^?1.     

Pharmacokinetics of tulathromycin in edible tissues of healthy and experimentally infected pigs with Actinobacillus pleuropneumoniae

The aim of this study was the comparison of the tissue pharmacokinetics of tulathromycin in healthy pigs and pigs experimentally infected with Actinobacillus pleuropneumoniae (App). Tulathromycin was given to 24 healthy and 24 infected pigs by intramuscular injection at a single dosage of 2.5 mg kg^?1 body weight (b.w.). Pigs were euthanised at each group and then samples of liver, kidney, muscle, injection site and skin with fat were taken at scheduled time points. Drug concentrations were determined by LC-MS/MS. In this study, higher values of the area under the concentration–time curves (AUC) were calculated in all tissue samples taken from infected than healthy pigs. In pigs with App the AUCs of liver, kidney, muscle, skin with fat and injection site were 1111, 1973, 235, 181 and 2931 mg kg^?1 h, while in pigs without inflammation they were 509, 1295, 151, 111 and 1587 mg kg^?1 h, respectively. Maximum drug tissue concentrations (C_max) in infected animals were 2370, 6650, 2016, 666 and 83 870 µg kg^?1, while in healthy pigs they were 1483, 6677, 1733, 509 and 55 006 µg kg^?1, respectively. The eliminations half-times (T_1/2) were respectively longer in all tissue samples taken from infected animals (from 157.3 to 187.3 h) than in healthy ones (from 138.6 to 161.2 h). The tulathromycin tissue concentrations were significantly higher (p < 0.05) in all tissue samples of the infected pigs compared with the healthy animals at 360 h (from 0.0014 to 0.0280) and at 792 h (from 0.0007 to 0.0242) after drug administration. The results suggest that the tissue pharmacokinetic properties and residue depletion of tulathromycin can be influenced by the disease state of animals.     

Determination of halosulfuron-methyl residues and dissipation in wheat by QuEChERS and LC-MS/MS

A sensitive and rapid analytical method based on QuEChERS (quick, easy, cheap, effective, rugged and safe) sample preparation and LC-MS/MS detection was developed for the analysis of halosulfuron-methyl residues in wheat. The recoveries of halosulfuron-methyl in both the wheat plant and grain ranged from 87% to 119% and from 75% to 97%, respectively, with relative standard deviations (RSDs) of 3–9%. The limit of quantification (LOQ) was 0.005 mg kg^?1 for wheat plant and 0.001 mg kg^?1 for wheat grain. The half-life of halosulfuron-methyl in the wheat plant was 0.9–9.5 days. The terminal residue levels of halosulfuron-methyl in wheat grain were below 0.01 mg kg^?1 at harvest.     

A Robust Transferable Method for the Determination of Glyphosate Residue in Liver After Derivatization by Ultra-high Pressure Liquid Chromatography–Tandem Mass Spectrometry

Glyphosate determination in liver is challenging due to this particular molecule/matrix combination. Glyphosate is a very polar molecule and liver composition is highly variable between individuals and species. Since 2014, the Multiannual Control Program (MACP) of the European Union (EU) demands to analyse glyphosate in food of animal origin on a voluntary basis. Moreover, this analysis will be mandatory in 2017. This paper describes a robust and easily transferable method for glyphosate quantification in liver of animal origin by means of liquid chromatography–tandem mass spectrometry (LC-MS/MS). An intensive clean-up was used to eliminate matrix interferences and was combined with a derivatization step which ensures good retention of glyphosate on a conventional reverse-phase LC column. This method allows to meet the MACP requirements without a time-consuming change in the set-up of the routinely used LC-MS/MS system. Furthermore, it allows the use of an LC column and mobile phases often used in multi-residue analysis. The analytical method was validated according to the SANCO/12571/2013 criteria. Isotopic dilution was used to quantify glyphosate, leading to mean apparent recoveries of 115 and 101 % for the low (0.025 mg kg^?1) and the high (0.250 mg kg^?1) fortification levels, respectively. At both levels, the relative standard deviation was below 10 %. The limit of quantification of 0.025 mg kg^?1 was found to be satisfactory as it was below the maximum residue level (MRL) value set at 0.050 mg kg^?1 for glyphosate in liver. It is also the lowest MRL for all commodity types.     

Developing a microbiological growth inhibition screening assay for the detection of 27 veterinary drugs from 13 different classes in animal feedingstuffs

Many regulations prohibit using veterinary drugs in feedingstuffs to protect consumers and animals alike. Within this investigation we developed a simple, cost-efficient primary screening method for detecting antibiotics and coccidiostats in animal feeds. Thirty-two veterinary drugs were originally considered. Following matrix-free testing to optimise detection, an assay based on matrix extraction with methanol/acetonitrile/phosphate buffer followed by inoculation and diffusion in agar plates was developed. Final validation was performed with 14 representative drugs (one per drug class) and four bacteria (Escherichia coli ATCC11303 and ATCC27166, Staphylococcus aureus ATCC6538P, Micrococcus luteus ATCC9341) in bovine, lamb and swine fodder, measuring growth inhibition zones. Of the original drugs tested, 27 remained detectable in feed matrices at or below 20 mg kg^?1. Of the 14 validated representatives, two had estimated minimum detectable concentrations of 10–11 mg kg^?1, others of 5 mg kg^?1 or lower, an earlier minimum European Union inclusion rate for many veterinary drugs. No significant matrix effect on inhibition zones was detected. Per cent wrong negative deviations ranged from 0% (nine of 14 compounds) to 20–27% (two of 14), while inter-day precision based on inhibition zones had relative standard deviations (RSDs) of 6–109% (mean of 40%). When setting a 1 mm inhibition zone, the maximum observed for negative controls, as a cut-off level, no false-positives were found. While not all targeted antibiotics were detectable in complex matrices, the majority of veterinary drugs were detected with reasonable sensitivity, indicating that this method could be suitable for screening feedingstuffs prior to further confirmatory investigation of positive findings such as by LC-MS/MS.     

Validation and quantification of neonicotinoid insecticide residues in rice whole grain and rice straw using LC-MS/MS

An efficient analytical method was developed and validated using a modified QuEChERS method and LC-MS/MS for the detection and quantification of neonicotinoid insecticide residues in rice whole grain and rice straw. The samples were extracted using acetonitrile and cleaned up by dispersive solid-phase extraction. Validation based on five fortification levels showed good recoveries of neonicotinoids ranging from 75% to 116 % and 60% to 105 % for rice whole grain and straw, respectively. The precision ranged between 3% and 17 %, and 2% and 10 % for grain and straw, respectively. The limit of detection was from 0.007 to 0.0084 mg kg^?1 and 0.005 to 0.15 mg kg^?1 and the limit of quantification was in the range of 0.024–0.028 mg kg^?1 and 0.016–0.051 mg kg^?1 for rice whole grain and rice straw, respectively. Monitoring of farm gate samples indicated that, out of 24 samples, 1 rice whole grain sample was contaminated with thiamethoxam residues (0.07 mg kg^?1).     

High-performance liquid chromatography with fluorescence detection for the determination of nitrofuran metabolites in pork muscle

A simple and sensitive HPLC method with fluorescence detection (HPLC-FLD) is reported for the simultaneous determination of metabolites of four nitrofuran drugs (furazolidone, furaltadone, nitrofurantoin and nitrofurazone) in pork muscle. The method involves acid hydrolysis of the protein-bound drug metabolites and the conjugation of the released side-chains with a novel fluorescence agent 2-hydroxy-1-naphthaldehyde. After liquid–liquid extraction and effective separation of the derivatives on a YMC-Pack Polymer C18 column at 40°C under alkaline conditions, the high fluorescence intensity of these derivatives at emission wavelength λ_em = 463 nm enables their simultaneous determination in pork muscle at concentrations as low as 1 µg kg^?1. The method was validated using blank pork muscle fortified with all four metabolites at 0.5, 1.0 and 2.0 µg kg^?1. Recoveries were > 92.3% with RSDs < 8.5% for all four metabolites. The results obtained with HPLC-FLD and LC-MS/MS methods showed very good agreement for pork muscle samples.     

Comparison of four different methods for the determination of sulfites in foods marketed in South Korea

Sulfites in foods were analysed using four methods: optimised Monier–Williams (official method), modified Rankine, HPLC and ion-exchange chromatography (IEC). The modified Rankine and HPLC methods were performed according to the previously reported methods but with some modifications. The IEC method was carried out through a combination of a modified Rankine apparatus and an anion-exchange column for the first time. In false-positive response tests, false-positive results with acetic acid and propionic acid were not observed in the modified Rankine, HPLC or IEC methods, unlike the optimised Monier–Williams method. All methods were evaluated for accuracy, precision and simple correlations. Modified Rankine, HPLC and IEC methods were determined to be suitable for foods with less than 10 mg kg^–1 of sulfur dioxide (SO₂). The modified Rankine and HPLC methods were suggested to be the most appropriate for the determination of sulfites in foods due to their high correlation coefficient with the optimised Monier–Williams method (R² = 0.9138 and 0.9011, respectively).     

Deposition and depletion of decoquinate in eggs after administration of cross-contaminated feed

Decoquinate, a chemical coccidiostat used as a feed additive, can occur in eggs due to cross-contamination of feedstuffs for laying hens. An experiment was designed to assess the transfer of decoquinate to hen eggs and its distribution between egg yolk and egg white. Hens were given the feed containing decoquinate at a cross-contamination level (0.34 mg kg^–1) and collected eggs were analysed using an LC-MS/MS method. The plateau level was reached on the eighth day of the experiment and averaged 8.91 µg kg^–1, which is far below the maximum level established at 20 µg kg^–1 for whole eggs. Decoquinate was deposited mostly in egg yolks (26.2 µg kg^–1) and did not deplete completely during 14 days of administration of decoquinate-free feed. The results confirmed that administration of cross-contaminated feed is associated with very low risk of non-compliant residue levels of decoquinate in eggs.     

Occurrence of chloramphenicol in cereal straw in north-western Europe

Two surveys are presented of straw analysed for naturally occurring chloramphenicol (CAP), a drug banned for use in food-producing animals. In the first study, CAP was analysed by LC-MS/MS and detected in 37 out of 105 straw samples originating from the Netherlands, France, the UK, Germany and Denmark. The highest level found was 6.3 µg kg^?1, the average 0.6 µg kg^?1 and the median 0.2 µg kg^?1. The second study included a method comparison between ELISA and LC-MS/MS and a survey of CAP in cereal straw sampled at farms in all areas of Sweden. A total of 215 samples were screened by ELISA and a subset of 26 samples was also analysed by LC-MS/MS. Fifty-four of the samples contained more than 1 µg kg^?1 CAP and the highest level found was 32 µg kg^?1 (confirmed by LC-MS/MS). The highest contents of CAP in this study were allocated to the Baltic sea coast in the south-eastern part of Sweden (the county of Skåne and the Baltic Sea isle of Gotland). These results indicate a high incidence of CAP in straw in north-west Europe and have a severe impact on the enforcement of European Union legislation.     

Selective determination of thiram residues in fruit and vegetables by hydrophilic interaction LC-MS

Thiram belongs to the most important class of dithiocarbamate (DTC) fungicides including dimethyldithiocarbamates (DMDs), ethylenebis(dithiocarbamtes) (EBDs) and propylenebis(dithiocarbamates) (PBDs). During the surface extraction of fruit and vegetables for the LC-MS determination of residues of DMDs, EBDs and PBDs, thiram is reduced by the penicillamine buffer to the DMD anion, thus resulting in false-positive findings of DMD fungicides like ziram. Therefore, an alkaline sulfite buffer was applied for surface extraction, quantitatively transforming thiram into the DMD anion and a stable DMD–sulfite adduct that was used as a selective marker for thiram. Separation was performed isocratically on a ZIC-pHILIC column with acetonitrile–10 mM ammonium hydroxide solution (85/15). Mass selective detection was carried out on a single-quadrupole mass spectrometer coupled to an electrospray ionisation interface operating in negative mode. Using d₁₂-thiram as the internal standard, recoveries of 80–108% were obtained from apples, tomatoes, grapes and sweet peppers, spiked in the range of 0.02–1 mg kg^?1. Limits of detection and quantification were 0.6 and 2 µg kg^?1, respectively.     

Tissue deposition and residue depletion of melamine in fattening pigs following oral administration

The adulteration of animal feed as well as milk products with melamine has led to concerns about the ability to establish appropriate withdrawal intervals to ensure food safety. Two experiments were conducted in this study. The first was to investigate the deposition and depletion of melamine in blood and tissues of pigs exposed to adulterated feed with high doses of melamine. A total of 500 or 1000 mg kg^–1 melamine was added to the diet for fattening pigs (initial BW = ±60.24 kg). Melamine residues were detected in tissues (brain, duodenum, liver, heart, muscle and kidney) by LC-MS/MS. Dose-dependent effects were found between melamine residual concentration and its dose in feed. Five days after the withdrawal of melamine from the diets, the residue concentration in tissues fell below 2.5 mg kg^–1. In the second experiment, blood samples were taken at different time points from fattening pigs (BW = 100 kg) fed with adulterated feed with 1000 mg kg^–1 of melamine for 42 days. Results from the pharmacokinetics analysis showed that it would take 83 h for the melamine level in plasma depleting to the safe level of 50 ng ml^–1 after an expose of 1000 mg kg^–1 melamine contaminated feed for 42 days.     

Occurrence of 3-MCPD and glycidyl esters in edible oils in the United States

Fatty acid esters of 3-monochloropropanediol (3-MCPD) and glycidol are processing contaminants found in a wide range of edible oils. While both 3 MCPD and glycidol have toxicological properties that at present has concerns for food safety, the published occurrence data are limited. Occurrence information is presented for the concentrations of 3-MCPD and glycidyl esters in 116 retail and/or industrial edible oils and fats using LC-MS/MS analysis of intact esters. The concentrations for bound 3-MCPD ranged from below the limit of quantitation (<LOQ) to 0.09 mg kg^?1 (ppm) in 22 unrefined oils and from 0.005 to 7.2 mg kg^?1 (ppm) in 94 refined oils. The concentrations for bound glycidol ranged from <LOQ to 0.03 mg kg^?1 (ppm) in unrefined oil samples and from <LOQ to 10.5 mg kg^?1 (ppm) in processed oil samples. The highest concentrations for both 3-MCPD and glycidol were seen in refined palm oil and palm olein samples. Palm olein samples also contained a higher percentage of 3-MCPD in mono-ester form than any other type of oil.     

Analysis of sterigmatocystin in cereals,animal feed,seeds, beer and cheese by immunoaffinity column clean-up and HPLC and LC-MS/MS quantification

A method is reported for the analysis of sterigmatocystin in various food and feed matrices using a commercial sterigmatocystin immunoaffinity column (IAC) for sample clean-up prior to HPLC analysis by UV with mass spectrometric detection (LC-MS/MS). Cereals (wheat, oats, rye, maize and rice), sunflower seeds and animal feed were spiked with sterigmatocystin at levels from 0.75 to 50 µg kg^?1 to establish method performance. Using acetonitrile/water extraction followed by IAC clean-up, and analysis by HPLC with detection at 325 nm, recoveries ranged from 68% to 106%, with repeatability from 4.2% to 17.5%. The limit of quantification with UV detection in these matrices was 1.5 µg kg^?1. For the analysis of beer and cheese the sample preparation prior to IAC clean-up was changed to accommodate the different properties of the matrix, prior to analysis by LC-MS/MS. For beer and cheese spiked at 5.0 µg kg^?1 the recoveries were 94% and 104%, and precision (RSDs) were 1.9% and 2.9% respectively. The limits of quantification by LC-MS/MS in beer and cheese were 0.02 and 0.6 µg kg^?1 respectively. The sterigmatocystin IAC was demonstrated to provide an efficient clean-up of various matrices to enable this mycotoxin to be determined by either HPLC with UV detection or LC-MS/MS.     

Presence of mycotoxins in sorghum and intake estimation in Tunisia

Sorghum samples (n = 60) from Tunisian markets were analysed for the occurrence of 22 of both traditional and emerging mycotoxins. Samples were extracted with a QuEChERS-like method and mycotoxins were detected by LC-MS/MS. This method was validated and adequate analytical parameters were obtained. All samples had contamination with mycotoxins and several samples had higher contamination levels than European Union legislative limits (MLs). The most frequently found mycotoxins were ENB (100%), OTA (98%), ENA₁ (63%), ENB₁ (56%), BEA (48%), AFB1 (38%) and STG (33%). Mean contaminations were 30.7, 1.93, 33.2, 51.0, 15.4, 1.49 and 20.5 µg kg^–1, respectively. While two samples were contaminated with FB2 and FB3 at mean values of 16.2 and 45.9 µg kg^–1, respectively, one sample was contaminated with AFB2 and ZEA at levels of 0.82 and 45.0 µg kg^–1, respectively. The results were used to estimate the daily intake of mycotoxins through sorghum consumption with regard to normal consumers (low-risk population) and high consumers such as babies (high-risk consumers) who are facing an alarming situation.     

Estimate of dietary exposure to sulphites in child and adult populations in the Basque Country

Sulphites are widely used as a preservative and antioxidant additive in food. The aim of this study was to assess dietary sulphite intake in adults aged 35–65 years and in children aged 4–18 years living in the Basque Country, northern Spain. We determined sulphite concentrations in 909 samples covering 16 food types. The maximum permitted levels were exceeded in 17% of samples. Making recommended assumptions for non-quantifiable results, estimates of mean lower and upper bounds were calculated for sulphite concentrations in each food type. These sulphite data were combined with consumption data derived from 8417 adults from the European Prospective Investigation in Cancer and Nutrition cohort in Gipuzkoa, recruited in 1992–1995 using a diet history method, and 1055 children from the Basque Country Nutrition Children Survey, conducted in 2004–2005 using two 24-h recall questionnaires to assess diet. The results were compared with the acceptable daily intake (ADI) proposed by the Joint Expert Committee on Food Additives (JECFA). The mean dietary exposure to sulphites was 0.08 mg kg^?1 bw day^?1, only 11% of the ADI in the overall group of children (4–18 years old), but the acceptable intake was exceeded by 4% of 4–6 year olds. For the adults (35–65 years old), the mean dietary exposure was 0.31 mg kg^?1 bw day^?1, 45% of the ADI, but the acceptable intake was exceeded in 14.6% of cases. The major contributing foods were minced meat and other meat products for children and wine for adults.