Mineral analyses in which samples are prepared via

Transcription

1 536 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, 2018 INFANT FORMULA AND ADULT NUTRITIONALS Minerals and Trace Elements in Milk, Milk Products, Infant Formula, and Adult/Pediatric Nutritional Formula, Method: Collaborative Study, AOAC Final Action , ISO/DIS 21424, IDF 243 LAWRENCE H. PACQUETTE and JOSEPH J. THOMPSON 1 Abbott Nutrition, 3300 Stelzer Rd, Columbus, OH Collaborators: I. Malaviole; R. Zywicki; F. Woltjes; Y. Ding; A. Mittal; Y. Ikeuchi; B. Sadipiralla; S. Kimura; H. Veltman; A. Miura AOAC Final Action Official Method SM Minerals and Trace Elements in Milk, Milk Products, Infant Formula and Adult/Pediatric Nutritional Formula, Method was collaboratively studied. Note that milk, milk products has now been added to the title of the Final Action method because whole milk and several dairy ingredients were successfully incorporated into the collaborative study for the purpose of developing an International Organization for Standardization/International Dairy Federation standard (ISO/DIS 21424; in progress). The method determines sodium, magnesium, phosphorus, potassium, calcium, iron, manganese, zinc, copper, chromium, molybdenum, and selenium by inductively coupled plasma (ICP)-MS after microwave digestion. Ten laboratories participated in the study, and data from five different model units were represented. Thirteen products, five placebo products, and six dairy samples were tested as blind duplicates in this study, along with a standard reference material, for a total 50 samples. The overall repeatability and reproducibility for all samples met Standard Method Performance Requirements put forth by the AOAC Stakeholder Panel on Infant Formula and Adult Nutritionals, with a few exceptions. Comparisons are made to ICP-atomic emission data from a collaborative study of AOAC Official Method carried out concurrently on these same samples. Received August 15, Accepted by SG September 13, The method was approved by the AOAC Official Methods Board as Final Action. See Standards News, (2016) Inside Laboratory Management, March/April 2017 issue. The AOAC Stakeholder Panel on Infant Formula and Adult Nutritionals (SPIFAN) invites method users to provide feedback on the Final Action methods. Feedback from method users will help verify that the methods are fit for purpose and are critical to gaining global recognition and acceptance of the methods. Comments can be sent directly to the corresponding author. 1 Corresponding author s joseph.thompson@abbott.com Supplemental Information is available on the J. AOAC Int. Web site, DOI: Mineral analyses in which samples are prepared via closed-vessel microwave digestion and analyzed by either inductively coupled plasma (ICP)-atomic emission spectrometry (AES) or have become commonplace in the analytical laboratory, often replacing older methods that use flame or graphite furnace atomic absorption spectrophotometry because of the ICP s multielement capability and excellent sensitivity. As ICP technology and the sample preparation technique have evolved, a need to update official methods was noted. For example, AOAC Official Method SM (, after closed-vessel microwave digestion of the sample in nitric acid hydrogen peroxide) was introduced as a more modern version of AOAC Official Method (open-vessel digestion with dangerous perchloric acid), which became a Final Action method in These methods covered analytes Na, Mg, P, K, Ca, Fe, Mn, Zn, and Cu. The AOAC Stakeholder Panel on Infant Formula and Adult Nutritionals (SPIFAN) has recommended that the newest ICP methods be specifically examined for their performance on infant formula and child/adult matrixes so that there can be no doubt about the reliability of these methods in the case of regulatory or trade disputes. The goal is to develop and collaboratively study test methods that have the reproducibility/robustness needed by infant formula manufacturers as they market their products in many different countries. Such test methods would be recommended as new or replacement Codex Type II methods and be recognized as dispute-resolution methods. SPIFAN developed a set of Standard Method Performance Requirements (SMPRs SM ) for minerals. These contain the analytical range, recovery, LOQ, repeatability, and reproducibility of the test methods that are needed to be fit-forpurpose. Table 1 shows the two SMPRs (1, 2) developed for minerals combined into a single table. A collaborative study (also referred to as multilaboratory testing, or MLT; publication pending) has recently been concluded for Method , the method mentioned above, to test method performance relative to the SMPR parameters shown in Table 1. On the side, one MLT study of AOAC Official Method has already been completed and published for ultratrace elements Cr, Mo, and Se (3). AOAC Final Action Official Method extended Method to include the nine major/trace elements mentioned above in addition to the three ultratrace elements. We report herein the results of an MLT study of this 12-element method. The samples used in this study were identical to those used in the MLT study of Method

2 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, Table 1. SPIFAN SMPR parameters for Methods and Element Analytical range LOQ RSD r, % Recovery, % RSD R,% Na mg/100 g 10 mg/100 g Mg mg/100 g 3 mg/100 g P mg/100 g 15 mg/100 g K mg/100 g 10 mg/100 g Ca mg/100 g 20 mg/100 g Fe mg/100 g 0.01 mg/100 g Mn mg/100 g mg/100 g Mn mg/100 g mg/100 g Cu mg/100 g mg/100 g Cu mg/100 g mg/100 g Zn mg/100 g 0.1 mg/100 g Cr µg/kg 20 µg/kg Se µg/kg 10 µg/kg Mo µg/kg 20 µg/kg Thus, the accuracy of the and data can be ascertained by comparison of the mean results of the two independent studies. Other accuracy and validation data for Method can be found in the publication of the singlelaboratory validation (SLV) data (4) and in a separate publication specifically focused on the performance of this method at very low levels (5). Note that one other comparison is also available: the reproducibility for Cr, Mo, and Se reported for the present MLT compared to that reported for the same elements by the same method in the initial MLT (3). Different samples were used in the two independent MLTs, but it is interesting to compare the reproducibility of the original MLT that required known duplicate determinations to be averaged with the present MLT that required single determinations of each sample. Multilaboratory Collaborative Study Table 2 lists the laboratories and equipment involved in the study. Although 17 laboratories originally signed up for the study, several dropped out for various reasons and did not submit any data; the 10 laboratories shown in the table all completed the study. There is representation from seven countries, five different model units, and six different kinds of microwave units. All instruments were equipped with modern collision/reaction cell (CRCs) that are thought to be necessary for this method to avoid low-mass molecular interferences. In the first phase of the study, participating laboratories set up the method and checked the linearity, LOQ, and accuracy of their respective instruments. These same qualification tests were used in the first MLT (3), as these techniques have consistently been shown to lead to robust method transfers within Abbott Nutrition s network of laboratories. Specifically, laboratories were asked to transfer the Abbott method provided (which is more detailed than the published First Action Method ) and translate the given operating conditions for Agilent units into those suitable for their system. It was mandatory to keep the prescribed analyte and internal standard (ISTD) masses, as well as the nominal standard concentrations (made from a reputable vendor s master multielement standard solution), but other settings, such as ion lens voltages, collision cell gas flows, argon gas flow rates in plasma and the nebulizer, type of nebulizer, microwave settings, etc., could be varied to suit the laboratory s need and to obtain the necessary qualification data. The goal was to develop robust pulse/analog detector responses that would yield highly linear calibration curves with correlation coefficients all greater than and all calibration residuals below 4%. Next, the laboratories were asked to make up independent standard solutions and to run them as samples over 3 days. These samples were spaced at 10%, 40%, 800%, and 2400% of the Table 2. Participating laboratories and equipment for Method MLT Lab No. Lab name model Microwave model Country 4 Laboratory Aquanal Agilent 7800 CEM MARS 6 France 5 Syngene International, Ltd Agilent 7700x CEM MARS 5 India 6 Covance Laboratories Thermo icap Q CEM MARS 6 United States 9 Abbott Nutrition Agilent 7700x CEM MARS 5 Singapore 12 Qlip, Netherland Thermo icap Q Milestone Ultrawave The Netherlands 13 Eurofins Agilent 7500ce CEM MARS 6 The Netherlands 14 Fonterra PE NexION 300D NovaWAVE FA New Zealand 15 Meiji Co., Ltd Agilent 7700x PE Multiwave 3000 Japan 16 Morinaga Milk Industry Co., Ltd Agilent 7800 TOP Analytikjena Japan 17 Megmilk Snow Brand Co., Ltd PE NexION 300 TOP Analytikjena Japan

3 538 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, 2018 low standard calibration level. The mean 3 day result was required to be within % of the standard s nominal concentration for levels within the calibration range. Note that there were two check standards that were deliberately set at a level below the low standard. These served to identify the practical LOQ (PLOQ) for each laboratory or the concentration level at which linearity breaks down and there is >5% calibration bias. The third step in the laboratories preparation was the measurement of the actual LOQ. Reagent blanks (nitric acid, hydrogen peroxide, and ISTD) were analyzed on 5 different days, and the analyte concentrations were averaged. The mean result of the blank plus 10 times the SD of the mean was taken to be the LOQ. These solution concentrations were converted to sample concentrations using the default dilution factor of 1 g to 50 ml for the method. The criterion was that all LOQs had to be at or below the SMPR criteria (Table 1). In practice, this was difficult to achieve for Mn, Fe, and Cu, as explained in a publication (5). Ref. 5 also explains why undigested blanks were used for this practice exercise, as well as in Method After the laboratories had completed the above steps, a standard reference material (SRM 1849a) was analyzed in duplicate per the method, on 1 or more days. From the accuracy relative to the certified values and from the repeatability of the duplicate samples, the Study Director was able to judge whether the laboratories could begin analysis of the study samples. Study samples included the SPIFAN array of liquid and powder products made by the infant formula manufacturers for this purpose. These were shipped to the laboratories by Covance Laboratories (Madison, WI) who distributed them on behalf of AOAC INTERNATIONAL and SPIFAN. Thirteen products were included that are, by SPIFAN consensus, representative of most infant, pediatric, and adult formulas on the market. Five placebo products were included in which the manufacturer omitted any individual or premix additions of vitamins and minerals. Six dairy ingredients/products were added to the program at the request of the International Organization for Standardization (ISO)/International Dairy Federation (IDF) communities. These matrixes were reference materials supplied and shipped to the participating laboratories by Muva Kempten (Germany). Both Covance and Muva Kempten shipped blind duplicate samples with random codes on the labels for identification throughout the study. Only the Study Directors knew the assignments of which codes corresponded to which SPIFAN/IDF matrix, although in some cases (butter and cheese, for example), it was obvious to the laboratories which samples were paired. Lastly, the Covance shipments included samples of SRM 1849a coded as samples. The Study Directors made the decision to use this SRM as a type of control sample and not as another study sampleruninblindduplicateformat.thecodewasmade known to the laboratories so that they could us the SRM in their preliminary work as described above, and later as the first sample on each day of running the test samples. All told, there were 13 products, 5 placebos, and 6 dairy products/ingredients tested with their blind duplicates for a total of 48 samples. Because of the large number of samples that would have to be microwave digested, the decision was made to force the laboratories to split these over 2 days. A mandatory sample sequence was provided that started with the SRM 1849a as a control sample and put the placebos near the start of the sequence on day 1 and then the potentially messy (high-fat or high-concomitant elemental concentrations) ingredients toward the end of day 2. Blind duplicates were all sequenced on the same day so that true repeatability could be determined. This identical sequence was shared with and enacted by the Study Directors of the MLT (Method ). Reminders were given in the protocol sample sequence about how to prepare each sample per the method. Liquid products were weighed at 1.0 g directly into the microwave vessel. Powder products were 25.0 g powder reconstituted with 200 g water (11.1%, w/w), and a 1.8 g sample aliquot was taken from this slurry and weighed into a microwave vessel. For the dairy samples, the whole milk was weighed directly at 1.0 g like an infant formula. The butter, cheese, whey protein concentrate, and whey powder were weighed directly at 0.3 g into the microwave vessel. The whole-milk powder was reconstituted at 11.1% by weight like a powder product. The sample sizes given above were lower than for the MLT (pending publication), but the scheme was the same as far as what samples were weighed directly, and the reconstitution rates of powders were identical. Laboratories were given a results template spreadsheet in which to submit their data for both the preliminary (linearity and LOQ) and study sample data. Participants were also asked to report any deviations from the method and any relevant system suitability parameters and to add any comments about the method to the data form. All 10 laboratories completed all 48 samples and two SRMs, although one sample at one laboratory was accidently destroyed, and the data could not be recovered. The study protocol did not require the analysis of Cr, Mo, or Se because that was handled by the earlier MLT for Method , and SPIFAN was only asking for a nine-element MLT. However, many laboratories were interested in performing a 12-element method, and so 8 of the 10 laboratories completed this optional part of the protocol by adding these 3 ultratrace elements to their method. All data were transferred and statistically analyzed in a spreadsheet (6) using AOAC guidelines to determine overall mean, s r,rsd r,s R,RSD R, and the Horwitz ratio (HorRat). Cochran (P = 0.025; one-tail) and Grubbs (P = 0.025; single and double, two-tail) tests were used to determine statistical outliers. METHOD The initial First Action/Final Action method (2015/2017) has been revised to add references to the dairy products in the title, the scope, and other places. Some other minor edits were made to clarify that sample duplication is not required and to reference this MLT. Importantly, Table D has been added to summarize and clarify the recommended analytical range of the method. The range is derived from data presented in ref. 5. Although the PLOQ of the does not meet SPIFAN SMPR requirements for Cu, Mn, and Fe, note that the LOQ (about 10 times lower) does meet all SMPR and Codex criteria. AOAC Official Method Minerals and Trace Elements in Milk, Milk Products, Infant Formula, and Adult/Pediatric Nutritional Formula Method First Action 2015 Final Action 2017 ISO/DIS IDF 243 (Applicable for the determination of Na, Mg, P, K, Ca, Cr, Mn, Fe, Cu, Zn, Se, and Mo in infant formula and adult/pediatric

4 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, nutritional formula. Also applicable to the determination of Na, Mg, P, K, Ca, Mn, Fe, Cu, Zn, Se and Mo in milk and milk products. See Table D for analytical ranges.) Caution: The chemicals used are common-use solvents and reagents that are harmful if inhaled, swallowed, or absorbed through the skin. Refer to the adequate manuals or Material Safety Data Sheets to ensure that the safety guidelines are applied before using chemicals. Microwave operation involves a hot pressurized acid solution. Use appropriate personal protective equipment, such as a laboratory coat, safety glasses, rubber gloves, and a fume hood. Dispose of all materials according to federal, state, and local regulations. A. Principle This method is an extension of AOAC Official Method to determine nine additional elements. The method is also extended to support analysis of dairy products (milk, milk powder, whey powder, whey protein concentrate, butter and cheese) for eleven of the twelve elements (Cr is excluded because none of the dairy products contained Cr above the LOQ, and so reproducibility could not be determined.). Nitric acid, internal standard (ISTD), and hydrogen peroxide are added to the sample in microwave vessels, and the samples are digested using preprogrammed temperature control. The addition of hydrogen peroxide helps reduce carbon and nitrous oxide levels in the digestate. The presence of carbon in the samples causes signal enhancement of Se. Therefore, to matrix-match the samples, carbon in the form of methanol is added to both standard solutions and the digestate before analysis. Ge (for 11 elements) and Te (only for Se) are used as the ISTDs. Analysis is performed by inductively coupled plasma (ICP)-MS. Polyatomic interferences with low mass elements are reduced or eliminated by analysis in He collision mode using kinetic energy discrimination (KED). For Se measurements, the H 2 gas mode is preferred for increased sensitivity. Quantitation of the 12 elements is achieved essentially simultaneously by comparing the analyte ISTD response ratios in the unknown samples with a standard curve constructed from the response ratios of calibration standards. B. Apparatus (a) ICP MS. With quartz spray chamber, quartz torch, Ni/Pt sample cone, Ni/Pt skimmer cone, autosampler, and printer. The must have collision/reaction cells. In two independent multilaboratory testing (MLT) studies four different instrument models from three major vendors delivered equivalent performance. (b) Microwave oven. A commercial microwave designed for laboratory use at C, with a closed-vessel system and controlled temperature ramping capability. Use manufacturerrecommended vessels. In the MLT studies, five different microwave designs delivered equivalent performance. (Caution: Microwave operation involves a hot pressurized acid solution. Use appropriate face protection and laboratory clothing.) (c) Hydrogen generator (hydrogen is recommended for better Se sensitivity). On-demand supply of >99.999% pure hydrogen at >150 ml/min. Alternatively, a high-pressure cylinder (99.999% purity) may be used. (d) Magnetic stir plate. (e) polytetrafluoroethylene (PTFE)-coated magnetic stir bars. (f) Analytical balance. Capable of weighing to g. (g) Fume hood. (h) Common laboratory glassware/plasticware. (i) Repipetter. 50 ml. (j) Bottle-top dispenser. PTFE; Adjustable volume ml. (k) Volumetric pipets. Class A, assorted sizes. (l) Digital pipets. 1 ml adjustable, to deliver 500 µl with accuracy tolerance of better than 0.8% and precision of better than 0.2% RSD. C. Reagents (a) Multielement standard stock solution. National Institute of Standards and Technology (NIST) or NIST-traceable, containing Se at 20 µg/l, Cr and Mo at 40 µg/l, Mn and Cu at 0.25 mg/l, Zn at 1 mg/l, Fe at 2.5 mg/l, Mg at 10 mg/l, P at 25 mg/l, Ca and K at 50 mg/l, and Na at 25 mg/l in 2% HNO 3 + trace hydrofluoric (HF) acid. This stock standard solution expires on the date given by the manufacturer. (b) Multielement ISTD stock solution. NIST or NISTtraceable, containing Ge and Te at 5 mg/l in 2% HNO 3 + trace HF acid. This stock standard solution expires on the date given by the manufacturer. (c) Tuning and pulse/analog (P/A) factor tuning stock solutions. NIST or NIST-traceable containing various elements at concentration levels recommended by the manufacturer. Because this method determines the major elements at relatively high concentrations, it is important to understand the solutions needed and the procedure to obtain high-quality calibration curves in which the detector is used in both pulse-counting and analog modes. A properly calibrated instrument will deliver the linearity requirements of the method, for example, so that calibration residuals are <4% (see section F). (d) QC sample (QCS). Standard Reference Material (SRM) 1849a (NIST) milk-based hybrid infant/adult nutritional powder with certified values for Ca, Cu, Cr, Fe, Mg, Mn, Mo, P, K, Se, Na, and Zn. Supplied as a unit of 10 packets each containing approximately 10 g material. This is the recommended control material for this analysis, but other suitable SRMs could be substituted. (e) Methanol %, analytical reagent grade. (f) Nitric acid. Concentrated, ultrapure reagent grade. (g) Nitric acid. Concentrated trace metal grade. (h) Hydrogen peroxide, 30%. ACS reagent grade. (i) Laboratory water. metal-free, organic-free, pyrogen-free, filtered 18 MV cm quality. (j) Tergitol. Type 15-S-9, Sigma, or equivalent surfactant (optional). (k) Argon gas % purity. (l) Helium gas % purity. D. Preparation of Standards and Solutions (a) Tergitol solution (optional, approximately 5%, v/v). Add about 700 ml laboratory water to a 1 L plastic bottle containing a PTFE-coated stirring bar. Place the bottle on a magnetic stirrer and begin stirring at a moderate speed. Slowly add 50 ml Tergitol from a graduated cylinder. When the Tergitol is

5 540 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, 2018 dissolved, fill the bottle to approximately 1000 ml with laboratory water. Transfer to a 1 L plastic bottle fitted with a PTFE-constructed dispenser with adjustable volume from 0.5 to 5 ml. This solution is added to the autosampler rinse solution to minimize residue buildup in the spray chamber. It does not otherwise affect the analysis. Expiration: 6 months; store at room temperature. (b) Nitric acid rinse solution (2%, v/v) for autosampler rinse port with Tergitol added. Mix 20 ml concentrated nitric acid (ultrapure reagent grade) with 20 ml Tergitol solution (a) and laboratory water to prepare a total volume of 1000 ml. Expiration: 3 months; store at room temperature. (c) P/A factor tuning working solution. Dilute and/or combine P/A factor tuning stock solutions (or equivalent) to the manufacturer s recommended dilution level with laboratory water for use with the instrument. Expiration: 6 months; store at room temperature. (d) Calibration blank (Cal Blk) and preparation blank (PB) solution. Add approximately 15 ml laboratory water to a 50 ml volumetric flask. Dispense (using a bottle dispenser or pipet) 5 ml nitric acid (ultrapure reagent grade) into the same volumetric flask. Pipette (using a digital pipet) ml ISTD stock and ml methanol into the flask. Dilute to volume with laboratory water. This solution serves as both the Cal Blk and PB. The Cal Blk is used as the initial calibration point, whereas the PB is used as the QCS (see below). Use the same lots of reagent for samples. Expiration: 2 days; store at room temperature. (e) Calibration standard solution set. Prepare Cal Blk, Cal Std 1, Cal Std 2, Cal Std 3, and Cal Std 4 standard solutions by pipetting (with a Class A glass pipet) 0.00, 1.00, 5.00, 20.00, and ml, respectively, multielement standard stock solution into separate 50 ml volumetric flasks or sample tubes. Add ml ISTD stock (using a Class A pipet or digital pipet calibrated at point-of-use to 0.8% accuracy), 5 ml (using repipetter or PTFE bottle dispenser) nitric acid (ultrapure reagent grade), and 0.5 ml methanol to each flask. Dilute the flasks to volume with laboratory water. Expiration: 2 days; store at room temperature. The analyte and ISTD concentrations in the calibration standard solutions are shown in Table A. E. Preparation (a) Make a single sample preparation. In sample vessels, weigh test portions to the nearest g. For liquid products (includes milk), the test portion size is 1.0 g. Liquid samples shall be thoroughly shaken (5 min in a mechanical shaker is appropriate), the container opened and the contents dumped into a plastic container into which a magnetic stir bar is placed. While stirring, remove the 1 g sample with a disposable pipet for weighing directly into the tared microwave vessel. For ingredients such as whole milk powder, whey powder, or whey protein concentrate, use a direct weight of 0.3 g. For powder products, the test portion size is net 0.20 g powder sample, which should be taken from a 25 g powder g warm (60 C) laboratory water reconstitution (i.e., 1.8 g of the 11.1% reconstitution). For butter or processed cheese (take a moldfree portion) use a direct weight of 0.3 g. After weighing the sample, add ml ISTD stock using a calibrated digital pipet, 5 ml nitric acid (ultrapure reagent grade), and 2 ml of 30% hydrogen peroxide. (Note: the PB/Cal Blk solution prepared with the standards is the correct sample blank for this method. Specifically, do not microwave-digest the sample blank, which can subject the blank to contamination. Also note that the digital pipet used for the addition of ISTD solution must be calibrated at the point of use to ensure that it delivers a nominal volume of ml within a tolerance of ±0.8% and precision better than an RSD of 0.2%). (b) Seal the vessels and place into the microwave oven. Execute a heating program equivalent to that shown in Table B, suitable for total digestion of the sample. (c) After digestion, place the vessels in a fume hood, unscrew the cap/venting nut slowly to gradually release pressure, and then completely remove the cap. (d) Slowly add approximately 20 ml laboratory water to the contents of the vessel, swirl to mix, and transfer contents to a 50 ml sample vial. Add 0.5 ml methanol to the sample vial and dilute to approximately 50 ml with laboratory water. Shake briefly. The transfer or final volume does not need to be quantitative because ISTDs were added prior to digestion; therefore, the analyte ISTD ratios will be constant. F. Determination (a) Using the appropriate tuning solutions, tune the instrument for optimal sensitivity in KED mode and/or reaction mode according to the instrument design. Also, tune the instrument to find the P/A calibration factors needed for those calibration curves that will extend above roughly 100 µg/l (depending on instrument type). Table C summarizes typical instrument parameters for analysis. (b) Analyze test solutions using an instrument standardized with the indicated standard solutions (Table A). Ge is used as the ISTD for the 11 elements not including Se. Those 11 elements are Table A. Concentrations of standards and ISTDs in calibration standard solutions Standards and calibration standard solutions Na, mg/l Mg, mg/l P, mg/l K, mg/l Ca, mg/l Cr, µg/l Mn, mg/l Fe, mg/l Cu, mg/l Zn, mg/l Se, µg/l Mo, µg/l Cal Blk Cal Std Cal Std Cal Std Cal Std ISTD (at 50 µg/l) Ge Ge Ge Ge Ge Ge Ge Ge Ge Ge Te Ge

6 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, Table B. Microwave operating parameters: Stages 1 and 2 are operated sequentially, without removing vessels from the oven determined in He collision mode, using KED. Te must be used as the ISTD for Se determinations, and we recommend that low levels of Se be determined in H 2 mode, i.e., reaction mode. Depending on the instrument model, it may not be possible to easily switch between helium and hydrogen mode. In such a case, follow the instructions of the instrument manufacturer for changing from helium to hydrogen mode and analyze Se separately from the other elements. Alternatively, verify in separate experiments that the practical limit of quantification (PLOQ) for Se is at or below 10 ng/g in the sample when using an alternate collision/reaction gas. (c) Typical calibration correlation coefficients are or better for all analytes, but suitability is determined by calibration residuals as follows. Analyze calibration standard (Cal Std) 3 or another suitable QC solution every 10 test portions to monitor for instrument drift and linearity (the result must be within 4% of the standard s nominal concentration). The inclusion of a method blank (run as a sample; its measured concentration must be < half of the lowest calibration standard), and known reference materials serving as control samples (recovery check within control or certified limits) are mandatory for good method performance. Duplicate samples are optional. If used, the mean result is reported and appropriate criteria based upon the data would be a relative percent difference within 10% for Cr, 7% for Se, and 5% for all other elements. If any of these QC checks fail, results should be considered invalid. (d) The order of analysis should be calibration standards, followed by rinse, blank check, check standard, control sample, sample, sample duplicate (if used), and, finally, a repeated check standard. G. Calculations Stage 1 sample digest 1 Power, W 1600 (100%) 2 Ramp to temperature, min 20 3 Hold time, min 20 4 Temperature, C Cool down, min 20 Stage 2 sample digest 1 Power, W 1600 (100%) 2 Ramp to temperature, min 20 3 Hold time, min 20 4 Temperature, C Cool down, min 20 Total, h 2 concentrations in nanograms per gram are automatically calculated by the software using a nonweighted least-squares linear regression calibration analysis to produce a best-fit line: y = ax +blank Note that, the sample blank is identical to the Cal Blk in this method and is essentially zero because high-purity reagents are used. Table C. 7700x a The analyte concentration in the sample is calculated: x = y blank DF a where x = analyte concentration (nanograms per gram); y = analyte-to-istd intensity ratio, which is the measured count of each analyte s standard solution data point in the calibration curve divided by the counts of the ISTD at the same level; similarly, blank = analyte-to-istd intensity ratio, which is the measured count of the blank standard solution data point in the calibration curve divided by the counts of the ISTD at the same level as the blank standard solution; a = slope of the calibration curve (ml/ng); and DF = the dilution factor, or the volume of the sample solution (milliliters) divided by sample weight (grams). H. Method Validation Typical parameters for the Agilent RF power, W b 1600 RF matching, V b 1.8 Sampling depth, mm 9 Extract 1 lens, V 0 Carrier gas, L/min 0.9 Make-up gas, L/min 0.2 Nebulizer (glass concentric) MicroMist Spray chamber temperature, C 2 Interface cones Ni He cell gas flow rate, ml/min 4.5 H 2 cell gas flow rate, ml/min 4.2 Nebulizer pump rate, rps 0.1 (0.5 ml/min) Peristaltic pump tubing White/white, 1.02 mm id Drain tubing Blue/yellow, 1.52 mm id a b The isotopes used for analysis are 23 Na, 24 Mg, 31 P, 39 K, 44 Ca, 52 Cr, 55 Mn, 56 Fe, 63 Cu, 66 Zn, 78 Se, and 95 Mo, with 72 Ge and 130 Te as internal standards. RF = Radiofrequency. This method is very well characterized. It has undergone a thorough Single Laboratory Validation (SLV) using AOAC guidelines to probe its linearity, LOQ, specificity, precision, accuracy, and ruggedness/robustness. Two independent MLT protocols were carried out to measure the reproducibility of the method, and a special study focused on the performance of the method at very low levels between the lowest calibration standard and the actual LOQ as measured from digested and undigested blanks. Accuracy of the results was confirmed by comparing the mean MLT results from Method to those from Method , which employed testing on the same sample set (data to be published). Based upon all these data, Table D summarizes the latest figures of merit for this method. The last row gives the recommended analytical ranges of this method to fully meet the SMPR requirements. The minimum limits for Mn, Cu, and Fe are slightly above the SMPR criteria, which were very aggressive for these analytes well into the inherent level for the kinds of products that apply to this method. Note that the method can be used for analyte concentrations in the sample below this range down to the

7 542 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, 2018 Table D. Performance of Method in the SLV and MLT and recommended analytical ranges Parameter Na Mg P K Ca Mn Fe Cu Zn Cr Se Mo Low standard, µg/l PLOQ, µg/l a PLOQ, mg/100 g b PLOQ meets SMPR? c Yes Yes Yes Yes Yes No No No Yes Yes Yes Yes PLOQ meets Codex? d Yes Yes Yes Yes Yes No Yes Yes Yes Yes No Yes LOQ, mg/100 g e LOQ meets SMPR? c Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes LOQ meets Codex? d Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Avg. repeatability, % f Avg. reproducibility, % f Recommended analytical range, mg/100 g g h i h a b c d e f g h i The PLOQ is the lowest level at which the measured 3 day mean of a standard concentration, run as an unknown against the calibration curve, is within 5% of its nominal concentration. This is the preferred lower limit of the method, as it limits the amount of calibration bias in the result. However, the sensitivity of typically affords a lower actual LOQ, as measured in the traditional way of measuring blank concentrations over several days. PLOQ converted to a product concentration level using the typical dilution factor of 1 g to 50 ml for an RTF. The lower limit of the analytical range as set forth in the SMPRs (Table 1). Codex STAN minimum levels for infant formulas and Food for Special Medical Purposes products. LOQs derived from measuring several digested and undigested blanks, each run on 5 separate days. Both sets of blanks yielded similar results. For 18 samples tested at 10 laboratories in the MLT not counting the Adult High-Fat RTF sample or the samples in which the analyte was at or below the quantitation limit (as opposed to <PLOQ). To meet SMPR requirements. These are extrapolated from the MLT study, relying on reproducibility data for usually one to two samples that are no more than a factor of 3 in concentration from the indicated lower limit. Concentrations apply to (1) milk and RTF liquid as-is, using a typical sample size of 1 g per final analytical solution volume of 50 ml; (2) reconstituted milk powder, reconstituted infant formula powder, and reconstituted adult nutritional powder (25 g in 200 g water) using a typical sample size of 1.8 g reconstituted slurry per final analytical solution volume of 50 ml; and (3) ranges for nonreconstituted dairy ingredients (butter, cheese, whey powder, and whey protein concentrate) will be adjusted proportionally upward from these values based on the sample size used for the ingredient. For example, if 0.3 g cheese is digested, the ranges will be 1/0.3 g, i.e., 3.3 times higher. Mn and Se meet repeatability and reproducibility criteria at this level, but, unlike Cu, there is some additional calibration bias that might affect the accuracy of the result below the PLOQ. Although Cr has an adequate PLOQ and LOQ when measured under SLV conditions, the SMPR criteria for reproducibility (<15% RSD) are not consistently met below the SMPR lower limit of mg/100 g (20 ng/g). Slight contamination of the samples is a likely possibility. If only a single analysis is performed, testing Cr below 20 ng/g is not recommended. LOQ limits given in the table. This would allow measurement at the CODEX minimum limits for Mn and Cr, for example, but the expected repeatability would be above 5% RSD, and the expected reproducibility would be above 15% RSD, with nonlinearity of the calibration curves expected to produce a bias of >10% in the result. Collaborative Study Results and Discussion Method MLT statistical parameters for each of the 12 elements are given in Table 3 for the SPIFAN matrixes and Table 4 for the ISO/IDF matrixes. Parameters include the number of outliers, overall mean, s r and s R, RSD r and RSD R, repeatability and reproducibility limits, and the HorRat value. The HorRat value (7) is included for legacy purposes as a point of comparison with past collaborative studies, but the authors opinion is that it is a meaningless figure of merit in comparison with reproducibility, per se, and does not need to be included in collaborative studies. The premise that reproducibility is expected to get worse as the concentration of the analyte diminishes is clearly not supported by the data of the present study nor most of the SPIFAN MLTs published to date: For example, studies on ultratrace minerals (3), pantothenic acid (8), iodine (9), and inositol (10) reported an RSD R of most samples in the same 3 8% range over a wide range of concentrations. Rather, it is the relative level of the analyte above the LOQ that matters, and any analyte level above about 10 times the LOQ can be determined with roughly the same reproducibility that is due to the sources of variance of the method. For example, in this study, Mo had an average overall mean concentration of mg/100 g in the 18 SPIFAN matrixes and had an average reproducibility of 3.6% (both numbers can be calculated from the data in Table 3), whereas Na had an average level of 42.6 mg/100 g in these products times higher than Mo and was determined with about the same reproducibility of 3.7% (Table 3). Indeed, if one does not include results lower than the PLOQ and results from a problematic high-fat sample, the average RSD R (see Table 6) for all 12 elements in all 18 SPIFAN matrixes and 6 IDF matrixes ranged from a low of 3.7% for Mg to a high of 7.9% for Fe all in a narrow range because the sensitivity of the allows for determinations at analyte levels well above the LOQ. In Tables 3 and 4 repeatability and reproducibility that failed the SMPR criteria are indicated with a footnote. If one excludes adult high-fat ready-to-feed (RTF) sample 2, there are very few cases in which the method did not meet the criteria for products. 2 is a difficult one to prepare because it has a high fat level and a lot of insoluble tricalcium phosphate in it, and thus it separates easily and is difficult to homogenize. A different batch

8 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, Table 3. MLT statistics for the 12 elements in the SPIFAN matrixes Parameter 1 a 2 b 3 c 4 d 5 e 6 f 7 g 8 h 9 i 10 j 11 k 12 l 13 m 14 n 15 o 16 p 17 q 18 r Calcium Year of interlab test No. of labs submitting results No. of outliers (individual replicates) No. of accepted results Overall mean of all data (grand mean) t S r S R RSDr s s RSDR s Repeatability limit r (r = 2.8 sr) Reproducibility limit R (R = 2.8 sr) HorRat value Repeatability limit SMPR Reproducibility limit SMPR Copper Year of interlab test No. of labs submitting results No. of outliers (individual replicates) No. of accepted results Overall mean of all data (grand mean) t S r S R RSDr s s s 0.9 RSDR s s s 4.4 Repeatability limit r (r = 2.8 sr) Reproducibility limit R (R = 2.8 sr) HorRat value Repeatability limit SMPR Reproducibility limit SMPR Iron Year of interlab test No. of labs submitting results No. of outliers (individual replicates) No. of accepted results

9 544 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, 2018 Table 3. (continued) Parameter 1 a 2 b 3 c 4 d 5 e 6 f 7 g 8 h 9 i 10 j 11 k 12 l 13 m 14 n 15 o 16 p 17 q 18 r Overall mean of all data (grand mean) t Sr SR RSD r s s s 0.8 RSD R s s s 5.4 Repeatability limit r (r = 2.8 s r ) Reproducibility limit R (R = 2.8 s R ) HorRat value Repeatability limit SMPR Reproducibility limit SMPR Potassium u Year of interlab test No. of labs submitting results No. of outliers (individual replicates) No. of accepted results Overall mean of all data (grand mean) t Sr SR RSD r RSD R Repeatability limit r (r = 2.8 s r ) Reproducibility limit R (R = 2.8 s R ) HorRat value Repeatability limit SMPR Reproducibility limit SMPR Magnesium u Year of interlab test No. of labs submitting results No. of outliers (individual replicates) No. of accepted results Overall mean of all data (grand mean) t Sr SR RSD r

10 PACQUETTE & THOMPSON: JOURNAL OF AOAC INTERNATIONAL VOL. 101, NO. 2, Table 3. (continued) Parameter 1 a 2 b 3 c 4 d 5 e 6 f 7 g 8 h 9 i 10 j 11 k 12 l 13 m 14 n 15 o 16 p 17 q 18 r RSDR Repeatability limit r (r = 2.8 sr) Reproducibility limit R (R = 2.8 sr) HorRat value Repeatability limit SMPR Reproducibility limit SMPR Manganese Year of interlab test No. of labs submitting results No. of outliers (individual replicates) No. of accepted results Overall mean of all data (grand mean) t S r S R RSD r s 5.3 s s 0.7 RSDR s s 1.8 Repeatability limit r (r = 2.8 sr) Reproducibility limit R (R = 2.8 sr) HorRat value Repeatability limit SMPR Reproducibility limit SMPR Sodium u Year of interlab test No. of labs submitting results No. of outliers (individual replicates) No. of accepted results Overall mean of all data (grand mean) t S r S R RSD r RSDR Repeatability limit r (r = 2.8 sr) Reproducibility limit R (R = 2.8 sr) HorRat value Repeatability limit SMPR Reproducibility limit SMPR