Determination of sulfite in foods and beverages by ion exclusion chromatography with electrochemical detection: collaborative study

A liquid chromatographic (LC) method for determination of total sulfite in foods and beverages by alkali extraction followed by ion exclusion chromatographic separation and electrochemical detection (IEC-EC) was collaboratively studied by 9 laboratories. Blind duplicate samples of starch, diluted lemon juice, wine cooler, dehydrated seafood, and instant mashed potatoes were analyzed without spiking and with added sulfite at 2 levels. The initial sulfite levels varied from 0 to 384 ppm SO2, and the levels added varied from 10 to 400 ppm. The initial sulfite levels determined by the IEC-EC method and the Monier-Williams method were in good agreement. Recovery of added sulfite by the IEC-EC method was generally higher than that by the Monier-Williams method. Within-laboratory repeatability (RSDr) for the IEC-EC method varied from 4.4 to 26.0%, and overall reproducibility (RSDR) varied from 8.5 to 39.3%. The collaborators found the method to be fast, sensitive, and easy to use, which makes it a useful alternative to the Monier-Williams method. The method has been adopted official first action.     

Ion chromatographic determination of sulfites in foods

Ion chromatography (IC) is shown to be a promising technique for the determination of sulfites (SO2, SO2/3-) in foods. Results of a 10 min flash distillation and 10 min IC determination compare favorably with the results from the conventional Monier-Williams method for total sulfite in a variety of food matrices. The IC technique also provides a wealth of additional information, such as (1) sulfite and sulfate (oxidized sulfite) content of the spiking or treatment solution, (2) residual sulfite applied to the food after oxidation losses in the treatment process, (3) free sulfite in foods, and (4) total sulfite in foods. As a further check on the Monier-Williams method, the sulfate content of the trapping solution can be determined by IC. Because the IC technique traps the liberated SO2 in a non-oxidizing rather than an oxidizing medium, it is considered free from interfering sulfides and organic sulfur-containing groups which can give false positives in the Monier-Williams method. IC thus offers a high speed, more sensitive, and cost-effective alternative to conventional techniques for the determination of sulfite in foods.     

Differential pulse polarographic determination of sulfites in foods: collaborative study

A differential pulse polarographic (DPP) method for the determination of "free" and "total" sulfite in foods was collaboratively studied by 8 laboratories. The collaborators analyzed blind duplicate samples of shrimp, orange juice, peas, dried apricots, and dehydrated potatoes. Collaborators also analyzed the same samples spiked with sulfites at 2 levels, which ranged from 10 to 1100 micrograms added SO2/g. The variability of free SO2 results was excessive for quantitative analysis, but the method can be used for qualitative detection of free SO2. The method for total SO2 determination was suitable for as low as approximately 10 micrograms/g. Recoveries are comparable to those for the official Monier-Williams method at high levels and are superior at low levels. The method has been adopted official first action for determination of total SO2 in the foods studied.     

Determination of total sulfite in shrimp, potatoes, dried pineapple, and white wine by flow injection analysis: collaborative study

A method for the determination of total sulfite in shrimp, potatoes, dried pineapple, and white wine by flow injection analysis (FIA) was collaboratively studied by 8 laboratories. In the method, the sample solution is reacted with sodium hydroxide to liberate aldehyde-bound sulfite. The sample stream is acidified to produce SO2 gas, which diffuses across a Teflon membrane in the gas diffusion cell into a flowing stream of malachite green. The degree of discoloration of the malachite green is proportional to the amount of sulfite in the sample solution. Red wine was included in the study but interlaboratory precision for these samples was not satisfactory and correlation with Monier-Williams results was poor. The present method is not recommended for use with these samples. For shrimp, potatoes, dried pineapple, and white wine, average reproducibility (RSDR) of results was 25% for samples at 10 ppm SO2 and 10% for samples at greater than 50 ppm. Overall average reproducibility was 14%. Recoveries of sulfite added to samples averaged 80%. Comparison of FIA with the Monier-Williams method indicated comparable results by the 2 methods. The FIA method has been adopted official first action for determination of greater than or equal to 5 ppm total sulfite in shrimp, potatoes, dried pineapple, and white wine.     

Reevaluation of Monier-Williams method for determining sulfite in food

The Monier-Williams distillation procedure has a long history of successful use for determining sulfite in fruit products and wine; however, a systematic evaluation of its accuracy and precision with other food matrices has not been undertaken. We found that the Monier-Williams distillation yielded greater than 90% recovery of sulfite added to foods such as table grapes, hominy, dried mangoes, and lemon juice. Less than 85% recovery was obtained with broccoli, soda crackers, cheese-peanut butter crackers, mushrooms, and potato chips. These results may, in fact, accurately reflect the residual levels of sulfite if a portion of the sulfite undergoes irreversible reaction with some food components. Analysis of commercial food products gave sulfite levels ranging from 1400 ppm in dried apple slices to 25 ppm in cream sherry.     

Analysis for total sulfite in foods by using rapid distillation followed by redox titration

A rapid and accurate analysis for total sulfite as sulfur dioxide has been developed for foods and food products. The method, which combines a selective distillation cleanup procedure with the selective redox titration of sulfite ion by iodine, has been applied to a variety of foods and food products over a period of time with no significant interference encountered in any matrixes other than garlic and leeks. For the foods analyzed, the method typically shows a detection limit of 10 ppm, a relative standard deviation of 7.5% (compared with 10.4% for similar matrixes by the Monier-Williams method), and recoveries of 97.9 +/- 6.4%. Comparison of results for this method with those obtained using the Monier-Williams method showed a mean value for the distillation/titration method of 241 ppm compared with 242 ppm for the Monier-Williams method. A correlation of 0.991 and odds of a difference between methods of 10.7% (Student's paired t-test (1-alpha) X 100) were obtained for those matrixes where no interferences were encountered with either method.     

Evaluation of capillary gas chromatography for pesticide and industrial chemical residue analysis. II. Comparison of quantitative results obtained on capillary and packed columns

Results of pesticide and industrial chemical residue determinations, using both capillary and packed column gas chromatography (GC), in 3 Food and Drug Administration (FDA) laboratories have been compiled and compared. Samples consisted of food products collected for routine residue screening by the respective laboratories. Extracts were prepared by conventional multiresidue methodology. Capillary column systems and operating conditions were selected at the discretion of each laboratory and were therefore variable, although split/splitless injectors in the split mode were used with prescribed precautions in all cases. Packed column systems were operated as specified in the FDA Pesticide Analytical Manual (PAM). Overall correlation between the 2 systems, expressed as the average ratio of packed column result to capillary column result, was 0.99 for 120 determinations in 41 samples. The higher resolving power of the capillary systems allowed quantitation of several residues that were incompletely separated and therefore unquantifiable using the packed columns. Capillary column GC with the split injection technique, used with appropriate precautions, was found to be both reliable and advantageous for regulatory determination of pesticide and industrial chemical residues in foods and feeds.     

Comparison of three liquid chromatographic methods with FDA optimized Monier-Williams method for determination of total sulfite in foods

Three liquid chromatographic (LC) methods employing amperometric detection were compared with the collaboratively studied FDA optimized Monier-Williams distillation method for the determination of total sulfite in 5 food types. The foods included lemon juice, white wine, instant mashed potatoes, golden raisins, and onion flakes. Two of the LC methods (one employing headspace sampling and the other direct injection) used ion-exchange chromatography with a basic mobile phase (pH about 10.8) and a glassy carbon electrode; the third (employing direct injection) used ion-exclusion chromatography with an acidic mobile phase (pH about 2) and a platinum electrode. All 4 methods produced similar results for the wine, lemon juice, and raisins. Results were different for instant mashed potatoes and onion flakes. The headspace-LC method and direct ion-exclusion LC method, both of which employed an alkaline sample extraction, yielded significantly higher values for sulfite in instant potatoes than did the other 2 methods. A large interfering peak with both direct LC methods prevented quantitation of sulfite in the onion flakes. All methods can detect sulfite as low as about 1 microgram/g in 4 of 5 food types examined.     

Comparison of the ion exclusion chromatographic method with the Monier-Williams method for determination of total sulfite in foods

Experimental data comparing the alkali extraction/ion exclusion chromatographic method with the Monier-Williams method for determination of total sulfite are presented in (a) enzymatic and nonenzymatic browning systems, (b) vegetables containing naturally occurring sulfite, and (c) a carbohydrate-type food additive, erythorbic acid. Excellent agreement, with a linear correlation coefficient of 0.99, was observed in fresh potato samples homogenized with sulfite and allowed to react for different time intervals (enzymatic browning system). A good overall correlation was observed in dehydrated, sulfited apple samples heated for different times (nonenzymatic browning system); however, as heating time increased, higher results were obtained by the Monier-Williams method than by the alkali extraction/ion exclusion chromatographic method. The results of determining sulfite in the alkali trapping solution following acid distillation or acid treatment without heat suggested that this deviation was due to a fraction of sulfite bound to the browning reaction products in such a way that it was released by acid distillation but not by alkali extraction or acid treatment without heat. Similar behavior was demonstrated in cabbage with naturally occurring sulfite, which was released by acid distillation but not by alkali extraction or acid treatment without heat. The ion exclusion chromatographic method could overcome interference by the volatile caramelization reaction products in the Monier-Williams determination of erythorbic acid.     

Determination of sulfites in foods by simultaneous nitrogen purging and differential pulse polarography

An improved technique has been developed for determination of sulfites in food by differential pulse polarography. A Teflon sleeve is fitted to the dropping mercury electrode capillary so that SO2 is purged from the sample and simultaneously detected at peak potential. Bound sulfite in the sample is released at room temperature by addition of base in the absence of oxygen. For some foods, the prepared sample was passed through a Sep-Pak C-18 cartridge to remove naturally occurring sulfur compounds so that only added sulfite is measured. The level of detection was approximately 1 microgram SO2/g. Results agreed with those obtained by the optimized Monier-Williams method for a variety of foods.     

Determination of sulfur dioxide in grapes: comparison of the Monier-Williams method and two ion exclusion chromatographic methods

Results for determination of sulfur dioxide in grapes were compared by 3 methods: the modified Monier-Williams method, acid distillation/ion exclusion chromatography with electrochemical detection (AD/IEC-EC), and alkali extraction/ion exclusion chromatography with electrochemical detection (AE/IEC-EC). An unusual positive response was observed during the later stage of the Monier-Williams distillation of both control grapes and sulfited grapes. Development of volatile acidic compounds in parallel with this Monier-Williams response and darkening of sample was also observed by collection in an alkali trap and analysis using anion exclusion chromatography and photodiode array detection. No parallel increase in sulfite was observed by the more selective AD/IEC-EC method, which clearly demonstrated that the response observed during the later stage of the Monier-Williams method is a false positive, probably due to caramelization reaction products. Monier-Williams results for grapes containing ca 10 ppm sulfite were in reasonably good agreement with those by either the AD/IEC-EC or AE/IEC-EC methods, presumably because the false positive response in the Monier-Williams analysis compensated for the somewhat incomplete recovery of sulfite. The AE/IEC-EC method is recommended because it is rapid, sensitive, straightforward, and free from interference. Accurate results by Monier-Williams analysis could be obtained by limiting distillation to 60 min and correcting for recovery.     

Gas chromatographic-thermal energy analysis method for determination of volatile N-nitrosamines in baby bottle rubber nipples: collaborative study

A collaborative study was conducted on the U.S. Food and Drug Administration (FDA) dichloromethane extraction method for determining volatile N-nitrosamines in baby bottle rubber nipples. Following dichloromethane extraction, N-nitrosamines were determined by gas chromatography-thermal energy analysis. Six pairs of blind duplicate rubber nipple samples representing 6 lots were analyzed by 11 collaborating laboratories. All samples were portions taken from equilibrated composites of cut-up rubber nipples obtained from manufacturers in the United States. Recoveries of the internal standard (N-nitrosodipropylamine) at approximately 20 ppb ranged from 10 to 120%. Reproducibility relative standard deviations (RSDx) were between 35 and 45% for N-nitrosamine levels from 10 to 20 ppb. However, when data from laboratories with recoveries less than 75% were excluded (this is now specified in the method), RSDx values were between 11 and 32% for N-nitrosamine levels from 6 to 26 ppb. Values were consistent with or better than those reported for other analytical techniques designed to quantitate trace contaminants at the low ppb level, e.g., aflatoxin in foods. The method has been adopted official first action for the quantitation of volatile N-nitrosamines in baby bottle rubber nipples.     

State findings on pesticide residues in food--1988 and 1989

Findings of pesticide and related chemical residues are presented for 27,065 samples of foods collected and analyzed in 10 state food laboratories over 1988 and 1989 (fiscal years (FY) 88 and 89). These laboratories conduct food regulatory programs compatible with national programs of the U.S. Food and Drug Administration. Of the findings, 6325 samples contained detectable levels of 1 or more pesticide analytes and 418 (or 1.5%) of the total number of samples were deemed to be of regulatory significance.     

Immunodiffusion screening method for detection of motile Salmonella in foods: collaborative study

A collaborative study was performed to validate the performance of the 1-2 TEST for detection of motile salmonellae in foods. Detection is based on observation of an immobilized band of cells. Twenty-three laboratories participated in the study. The 1-2 TEST (immunodiffusion test) was compared with the standard culture procedure (BAM/AOAC; FDA Bacteriological Analytical Manual) for detection of Salmonella in 6 food types: ground black pepper, soy flour, dried whole egg, milk chocolate, nonfat dry milk, and raw deboned turkey. Uninoculated and inoculated samples were included in each food group analyzed. After the tests on the 6 foods were completed, analysis of the data for turkey and soy flour showed that certain collaborators obtained data inconsistent with the data from the majority of collaborators. No specific method deviations to account for the inconsistencies were reported by those collaborators, so the collaborative testing of these 2 foods was repeated. Analysis of data for pepper, chocolate, nonfat dry milk, dried whole egg, and the second set of soy flour and turkey indicated 96.1% agreement between the BAM/AOAC and immunodiffusion test methods. The false negative rates for the immunodiffusion test and BAM/AOAC methods were 3.6 and 1.7%, respectively. There was no significant difference in the productivity of the immunodiffusion test and BAM/AOAC method at the 5% level for any of the 6 foods. The immunodiffusion screening method has been approved official first action for detection of motile Salmonella in foods.     

Food and Drug Administration pesticide residue monitoring of foods: 1983-1986

Pesticide residues in foods are reported for the 4-year period 1982-1986 [fiscal years (FY) 83-86]. Results were summarized from the 2 complementary approaches that make up the Food and Drug Administration's (FDA) pesticide residue monitoring program. Under regulatory monitoring, which focuses on residues in raw agricultural commodities, a total of 49,055 samples (27,700 domestic and 21,355 import) that included fresh fruits and vegetables, grains and grain products, milk and dairy products, seafoods, and a variety of processed foods were analyzed. No residues were found in 60 and 48% of the domestic and import samples, respectively, compared with 55 and 44% in FY78-82. About 3% of the domestic and 5% of the import samples were violative. In FY78-82, about 3 and 7% were violative, respectively. The other FDA monitoring approach, the Total Diet Study, was revised in April 1982 to expand coverage of age/sex groups, use updated diets, and provide for analysis of individual foods. Results from monitoring under this modified approach and from regulatory monitoring continued to demonstrate that pesticide residues in the U.S. food supply were well below regulatory limits, and dietary intakes were many times lower than the Acceptable Daily Intakes established by international agencies.     

Effect of cooking on levels of ethylene dibromide residues in rice

Two studies were conducted to determine the effect that cooking has on the level of residues of ethylene dibromide (EDB) in rice. In the first study, 4 samples of long and medium grain polished white rice containing 113, 295, 956, and 1568 ppb EDB were cooked according to typical label directions. Three batches of cooked rice were prepared from each sample of polished rice and frozen until analysis; each batch was analyzed in duplicate. EDB levels in all cooked rice samples were less than 10 ppb. In the second study, conducted jointly by the Food and Drug Administration (FDA) and the Environmental Protection Agency (EPA), a sample of medium grain polished white rice containing about 1600 ppb EDB was cooked by each laboratory. Overall average EDB levels in rice analyzed immediately after cooking were 16 and 37 ppb for FDA and EPA, respectively. The corresponding frozen samples contained 8 and 39 ppb EDB. The 2 laboratories exchanged these frozen samples and reanalyzed them to check variability in the analytical procedure. FDA found 49 ppb EDB in the sample cooked by EPA and EPA found 8 ppb EDB in the sample cooked by FDA, thus indicating that analytical methodology was not a major source of variability. The range of EDB levels was therefore attributed to minor differences in the way the rice was cooked or handled immediately after cooking.     

Collaborative study of the glass wool filtration method for the recovery of virus inoculated into ground beef.

A method for estimating viral population levels in ground beef was studied collaboratively in 7 laboratories. The collaborators recovered virus from 6 inoculated samples. Three samples were replicates of the high virus concentration 050 plaque-forming units (pfu)/g) and 3 replicates represented the low concentration (10 pfu/g). Six of the 7 collaborators recovered acceptable levels of virus from the samples. The per cent of variation was 30.6 for the high concentration and 18.5 for the low concentration. Collaborators did not differ from one another significantly in the results obtained for the 10 pfu/g samples, but results from one collaborator were significantly low for the recovery of virus from the 50 pfu/g samples. The results indicate that the glass wool filtration method is adequate for the detection of a number of viruses that may be found in foods. The method has been adopted as official first action.     

Food and Drug Administration pesticide residue monitoring of foods: 1978-1982

Pesticide residues in foods are reported for the 5-year period 1978-1982 [fiscal years (FY) 78-82]. Results were compiled from the 2 complementary elements that comprise the Food and Drug Administration's (FDA) program for monitoring pesticide residues in foods. Under regulatory monitoring, which focuses on residues in raw agricultural commodities, a total of 49,877 samples (30,361 domestic and 19,516 import) that included fresh fruits and vegetables, grains, milk and dairy products, seafoods, and a variety of processed foods were analyzed. No residues were found in about 55 and 44% of the domestic and import samples, respectively. About 3% of the domestic and 7% of the import samples were classed as violative. Data from the Total Diet Study, which is conducted to determine dietary intakes of a variety of chemicals, showed that residues of 42 pesticides were found in 1044 composites of table-ready foods. Results of FDA's monitoring for FY78-82 demonstrate that pesticide residue levels in the U.S. food supply were generally well below regulatory limits, and dietary intakes were manyfold lower than the Acceptable Daily Intakes established by international agencies.     

Collaborative studies on cytogenetic and dominant lethal assays.

Two in vivo cytogenetic studies and one dominant lethal study were jointly conducted by members of government, academic, industrial, and independent commercial laboratories between 1970 and 1975 to determine the validity and reproducibility of the techniques. The cytogenetic studies consisted of preparations and analyses of bone marrow cells from male rats for chromosome abnormalities. In the first study, participants from 4 laboratories jointly prepared slides at a Food and Drug Administration (FDA) laboratory according to a prescribed procedure and independently analyzed the slides in their own laboratories. In the second study involving 6 laboratories, a workshop was held at Dow Chemical Co. to reduce scoring differences and to develop a joint protocol. All participants then independently performed the exact procedure in their own laboratories, using animals from a common source. In the dominant lethal study involving 6 laboratories, a workshop was held at the FDA to discuss protocol and to practice the technique. All participants then independently conducted the study in their own laboratories using rats and test chemicals from common sources. In all 3 studies, significant interlaboratory variability existed, and the differences varied with the parameters analyzed.     

Determination of sulfite in food by flow injection analysis

A method is described for the determination of sulfite levels in food products by flow injection analysis (FIA). The method is based on the decolorization of malachite green by SO2, which is isolated from the flowing sample stream by means of a gas diffusion cell. The FIA method has a detection limit in food sample extracts of 0.1 ppm SO2 (3 times peak height of blank), which corresponds to 1-10 ppm SO2 in a food product, depending on the extraction procedure used. At the 5 ppm SO2 level in a food extract, the precision of replicate injections is +/- 1-2%. The method was tested on a variety of both sulfite-treated and untreated food products and the results compared favorably with those obtained by the Monier-Williams, colorimetric (pararosaniline), and enzymatic (sulfite oxidase) methods. The average differences from the FIA results were 19, 11, and 12%, respectively, for those samples (n = 12) above 50 ppm SO2. At lower levels the results were somewhat more erratic due to inaccuracies of the various methods at low concentrations.