AOAC SPIFAN Nutrients ERP Reviewer Forms (July 28, 2022)

AOAC Stakeholder Program on Infant Formula and Adult Nutritionals

EXPERT REVIEW PANEL (ERP) (NUTRIENTS) REVIEWER FORMS Thursday, July 28, 2022

AOAC INTERNATIONAL 2275 Research Blvd., Suite 300 Rockville, MD, 20850 USA

dboyd@aoac.org 301.924.7077 x126

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition Evaluation of Method_ Amino 05

Title: An Ion Exchange Liquid Chromatography with Post Column Derivatization Method for the Determination of Taurine in Infant Formula and Adult/Pediatric Nutritional Formula

Author: Mérieux Reviewer Name:

Reviewer 1 Summary of Method: Powdered, semi liquid or liquid samples are extracted with a citrate buffer in the presence of Norleucine as internal standard. The extracted reacted samples are filtered and analyzed using Amino Acid Analyzer ionic chromatography with post column derivatization with Ninhydrin.

Method Scope/Applicability: Determination of Taurine amino acid in infant formula and adult/pediatric nutritional formulas

General comments about the method: This method uses well-established methodology for analyzing amino acids using cation-exchange column and post-column derivatization with Ninhydrin reagent. The authors also submitted a method that can be used to analyze free and total amino acids but the validation study was completed for Taurine only. For Taurine a simple extraction sample preparation procedure was used but hydrolysis step is required to determine the total amino acids. The chromatographic conditions described in the method are unique to a particular amino acids analyzer so buffers and Ninhydrin reagent have to be purchased from the manufacturer of the analyzer. Different brands of amino acids analyzers are available on the market and their performance should be comparable to the one described in the method. The method should emphasize general performance requirements that can be used to assess the performance of different amino acids analyzers.

Method Clarity:

Method is clearly written

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Pros/Strengths: Well-known analytical approach using Ninhydrin reagent for analysis of amino acid. The same chromatographic conditions can be used to determine all the amino acids listed in SMPR but sample hydrolysis step is needed if analysis of total amino acids is required

Cons/Weaknesses The exact composition of buffers and reagents is not available so all the solutions have to be purchased from the manufacturer of the amino acids analyzer.

Long run time. If no other amino acids besides Taurine and Norleucine are present on the chromatogram it would be beneficial to optimize the method for shorter run time.

Supporting Data • General Comment:

Method Optimization:

• Performance Characteristics:

Analytical Range : 0.5 – 2500 mg/100 g for liquids or reconstituted powders – meets SMPR

LOQ: 0.5 mg/100g for reconstituted powder samples – meets SMPR

Accuracy/Recovery: meets SMPR – 2 samples with 3 spike levels were used for accuracy study

Precision (RSD r ): meets SMPR – 5 samples used for precision study

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Reproducibility (RSD R ): NA

• System suitability: Acceptance criteria for calibration curve and chromatographic separation are provided

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

1. Is the Validation Study Report in a format acceptable to AOAC? Yes

2. Is the method described in sufficient detail so that it is relatively easy to understand, including equations and procedures for calculation of results (are all terms explained)? Yes

3. Are the figures and tables sufficiently explanatory without the need to refer to the text? Yes

4. Are all the figures and tables pertinent? Yes

5. Could some be omitted and covered by a simple statement?

No

6. Are the references complete and correctly annotated?

Yes

7. Does the method contain adequate safety precaution reference and/or statements? Yes

Recommendation: The method describing analysis of Taurine has met SMPR requirements and I would recommend it as AOAC first action official method. Notes should be made in the method that equivalent equipment, columns, buffers and Ninhydrin reagent from other manufacturers can be used, providing system suitability requirements are met. Since the validation study was only competed for Taurine the method describing analysis of free and/or total amino acids should not be considered at this time. It would be beneficial if authors could complete validation for free and total amino acids so the same method could be used for all the amino acids listed in SMPR.

4

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition Evaluation of Method Amino-05

Title: An Ion Exchange Liquid Chromatography with Post Column Derivatization Method for the Determination of Free and/or Total Amino Acids and Taurine in Infant Formula and Adult/Pediatric Nutritional Formula Author: Giampaolo Perinello, Paolo Tonini and Lorenzo Papa. Mérieux Nutrisciences Italia. Reviewer Name: Reviewer 2 Summary of Method: The method allows the determination of free, total and sulfur amino acids and Taurine in infant formula and adult/pediatric nutritional formula. For free amino acids and Taurine samples are extracted with a citrate buffer in the presence of Norleucine as an internal standard. For the determination of total non-sulfur amino acids the samples is hydrolyzed in oven with hydrochloric acid and dilute for chromatographic analysis. For the determination of sulfur amino acids such as Cystine + Cysteine and methionine, the sample is reacted with an oxidizing mixture of formic acid and phenol to convert them into Cysteic acid and Methionine sulfone. This is followed by hydrolysis in oven with hydrochloric acid and diluted for chromatographic analysis. The extracted samples are filtered and analyzed using Amino Acid Analyzer Ionic Chromatography with post column derivation with Ninhydrin and detected with UV-VIS detector. General comments about the method: The determination is performed by extracting a portion of the sample using citrate buffer and Norleucine as internal standard. The extracted sample is analyzed using Amino Acid Analyser Biochrom 30+ with post column derivatization with Ninhydrin and UV-VIS detector at 570 nm wavelength. This procedure is applicable to powdered samples, semi-liquid or liquid samples. Although the method have a very good resolution. I consider it is inconvenient to use equipment from an exclusive supplier. Method Clarity: The method is clear, simple and easy to follow. Method Scope/Applicability: The method is performed to determine only Taurine in SPIFAN samples kit.

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Pros/Strengths: This Step Gradient system, allows to obtain the maximum reproducibility of the retention times and resolution of the analysis and allows one hour saving in analysis times compared to other standard methods for amino acids. The ion exchange analytical column is completely regenerable, so it does not have to be periodically replaced and allows a workflow exceptionally low (25ml / hour) for a reduced consumption of eluent and reaction solvents that means a significant reduction of the analysis costs.

Cons/Weaknesses

Equipment from exclusive supplier.

Supporting Data • General Comment:

Method Optimization:

Three calibration curves have to be tested with independently prepared standards. For testing the calibration models 6 calibration samples are used ranging from 0,1 – 11 mg/L. Working samples and calibration samples are both analized in the Amino Acid Analyser Biocrom 30+ with post column derivation with Ninhydrin.

• Performance Characteristics:

Analytical Range: SMPR: 0.5–2500 mg/100 g reconstituted final product it is necessary to describe the calculation procedure.

LOQ: SMPR: ≤0.5 mg/100 g reconstituted final product LOQ : 0.43 mg/100g for reconstituted powder samples.

The laboratory has proceeded by preparing solutions at least at four concentration levels between LOD and two times LOQ. Three different runs using different instruments and from three different analyst were conductes in order to evaluate LOQ.

Accuracy/Recovery : SMPR: 0.5–5.0

±12% 5.0–150 ±10% 150–2500 ±7%

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

The NIST SRM 1869 was analized through spikes at two ranges in the middle/high field of SMPR. Spiked and unspiked samples were analyzed in triplicate on each three days. Results: 99.6 % ±12% for low Taurine content (0.5-5.0)

96.8 % ±3.2% for medium Taurine content (5.0-150) 98.4 % ±3.1% for high Taurine content (150-2500)

Precision (RSD r ): SMPR: 0.5–5.0

≤7%

5.0–150 ≤5% 150–2500 ≤3%

The precision was assessed by repeated measurements carried out on 6 different days, in duplicate on the each sample of SPIFAN kit. The precision was evaluated also on spiked samples NIST 1869 and NIST 8261 to cover all the analytical range. Results:

0.5–5.0 1.4% 5.0-150 2.0 % 150-2500 0.9% Reproducibility (RSD R ): No data

• System suitability:

Blank acceptance criteria No peaks detected except that of Norleucine internal standard.

Acceptance of the Calibration curve The calibration curve is considered acceptable if the Linear Regression coefficient is ≥ 0.995 .

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

1. Is the Validation Study Report in a format acceptable to AOAC? Yes. 2. Is the method described in sufficient detail so that it is relatively easy to understand, including equations and procedures for calculation of results (are all terms explained)? Yes. 3. Are the figures and tables sufficiently explanatory without the need to refer to the text? Yes. But it is necessary 4. Are all the figures and tables pertinent? Yes. 5. Could some be omitted and covered by a simple statement? Yes. It is necessary to describe the SPIFAN samples analyzed and the procedure used to determine the analytical range 6. Are the references complete and correctly annotated? Yes. 7. Does the method contain adequate safety precaution reference and/or statements? Yes, in the method there is a warning about the security measures to be adopted.

Recommendation:

In my opinion, is necessary to make a review of SMPR parameters. It is necessary to describe the SPIFAN samples be analyzed and the procedure used to determine the analytical range. After that method could be recommended to move to First Action .

4

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition Evaluation of Method FLUOR-05 (Resubmission)

Title: Development and Validation of a Microdiffusion Potentiometric Method for the Determination of Fluorides in Infant Formula and Adult/Pediatric Nutritional Formula Author: Giampaolo Perinello and Lorenzo Papa; Mérieux NutriSciences Italia Reviewer Name: Reviewer 1 Please Provide a Brief Summary of the Resubmitted Method: Powdered samples are extracted at 60°C in a micro-diffusion cell in the presence of an acidic solution containing Hexamethydisilazane (HMDS). The released fluoride reacts with sodium hydroxide within the cell and is subsequently measured by potentiometry using a selective electrode. Liquid samples are freeze-dried before being treated in an analogous manner as the powdered samples. Was Method Scope/Applicability Updated? Yes (see highlighted text below). Previous ranges read 30 to 200 µ g/100g for liquids, 30 to 400 µ g/100g for powders. The method is applicable for the determination of free and slightly bound Fluoride in a variety of Infant Formula and Adult/Pediatric Nutritional Formula at concentrations from 3.33 to 200 μg/100g for liquids (30 to 18 00 μg/100g for powders) .

Has the resubmitted method met the requirements outlined as per the ERP recommendation/comment(s)? Summary of previous reviews and ERP recommendations: ERP Review ERP Recommendation Author Response Apr 2021 Include raw data & information including sample size on infant formula tested.

Addressed – see similar Dec 2021 recommendation Addressed – see similar Dec 2021 recommendation Somewhat addressed – see similar Mar 2022 recommendation Addressed – see similar Mar 2022 recommendation

Is method measuring total or free/weakly bound fluoride (please comment). Need information about microdiffusion cells and their specific requirements Need information about 1000 ppm stock standard used to prepare the calibrants

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

ERP Review ERP Recommendation

Author Response

Dec 2021

Supporting raw data for LOQ, analytical range, spike condition Need data to demonstrate the selectivity of the method Please provide the form of the fluoride reported Compare data to common samples Submit method in AOAC format

Addressed – see similar Mar 2022 recommendation Completed with Mar 2022 submission Completed with Mar 2022 submission Completed with Mar 2022 submission Addressed – see similar Mar 2022 recommendation Addressed – see similar Mar 2022 recommendation Product information included as a footnote [This item is commercially available from ORION Scientific S.r.l., (p/n OS-CCT90023], as well as a statement of “Equivalent products may be used if they can be shown to lead to the same results.” While this is an improvement on previous method versions, I am concerned about how widely available this is (I was not able to find it on the Orion website). Also, without adequate specifications listed in the method, it’s hard to know what would be considered ‘equivalent’. Text added as a footnote: This standard is commercially available from Romil Ltd (Cat. SI194) and contains a water solution of NaF with 1000 ppm as F. Romil Ltd is an example of suitable product available commercially. Equivalent products may be used if they can be shown to lead to the same results. All tables updated to reflect calculations in both powder and reconstituted forms. Analytical range extended to 30-1800 ug/100 g powder (3.33-200 ug/100 g reconstituted product). 2 NIST samples spiked at medium/high levels (72-190 ug/100 g F). RSDrs were 3.3% or less. Completed using SRM 1869 with and without addition of the TISAB III solution (Table XV). I do not understand the results presented in this table, unless the intention is that the method doesn’t work on reconstituted formula.

Clarify the levels used in spiking

Mar 2022

Questions on microdiffusion cells; specifications on how to acquire

Need requirements for stock standard

Calculations need to be in SPIFAN units (reconstituted products) Analytical range needs to be extended to meet the SMPR Need to see spiked recovery and precision on higher level samples

Requesting data in reconstituted formula

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Requesting information on timing analysis Timing analysis described in Annex E. Review minor & editorial comments from the reviewers I believe these were addressed.

Please provide documentation: See table above.

Given the past concerns about accuracy of F methods and lack of reference materials with assigned values, I compared the results provided by authors of this method (FLUOR-05) with those provided by the other method being reviewed (FLUOR-06) on the same samples. Unclear if FLUOR-06 data is provided as powder or reconstituted, so I’ve assumed here it’s powder.

Provide General comments: The supporting data has been presented well and nearly all ERP concerns have been answered. My only remaining concern is around the proprietary nature and/or availability of the microdiffusion cells. I have yet to see evidence that this critical piece of equipment is widely available or have enough information to indicate that a suitable replacement could be found. Additionally, while the request for data in reconstituted formula was met, the implication of the results is unclear to me. I have no issue with a method that requires analysis directly from powder or for liquid powders to be lyophilized prior to analysis, but others found this to be important. Final Recommendation: I am not yet prepared to recommend this method as First Action based on the proprietary nature of the microdiffusion cell and the concerns of other ERP members. I would like to discuss these concerns as an ERP prior to making a motion.

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition Evaluation of Method__Fluor-05-SLV (July 2022) (Resubmission)

Title: A Microdiffusion Potentiometric Method for the Determination of Fluorides in Infant Formula and Adult/Pediatric Nutritional Formula Author: Giampaolo Perinello, Lorenzo Papa and Roberta Mazzonetto Mérieux Nutrisciences Italia, Via Fratta 25, 31023, Resana (Treviso) Reviewer Name: Reviewer 2 Please Provide a Brief Summary of the Resubmitted Method: There are some minor changes of the method comparing to the last version (March 2022). These changes provide more details on reagents sources, calibration concentrations for medium and high fluoride samples, and quality control. Additional supporting data and detail info are also provided in the SLV report. Was Method Scope/Applicability Updated? There is no change in method scope/applicability. Has the resubmitted method met the requirements outlined as per the ERP recommendation/comment(s)? I believe the resubmitted method and the SLV report meet the requirements outlined in the last meeting (March 2022) Specifically, the following additional info or data have been added: 1. The method has been revised to include additional info on the sources of reagents, calibration solutions for medium and high content samples, and residual of x (in Quality control section). 2. In the SLV report, additional data on the medium and high content samples, such as the calibration, (Table V), repeatability (Table VIII), and accuracy (Table X and XII), are provided. 3. Additional info that was missing in the previous report on the LOQ evaluation is now provided. 4. Additional results are also included in the specificity and the robustness sections in the SLV report. Please provide documentation: Provide General comments: This method has been revised with more relevant info and the procedure is easier to understand. The SLV report contains more supporting data to address the medium and high content samples. More accuracy study results for the low level of fluoride content are also included. One concern about the accuracy is that all the evaluation was by spiking experiments. There is no certified value for those NIST samples for a direct comparison.

Minor comment:

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

In the Section D, Preparation of calibration curve, please clarify what is the 0.05 mg/L (Cf) solution.

Final Recommendation: I would like to recommend it for the First Action method.

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition Evaluation of Method___Fluor-06______

Title: HS-GC analysis of fluoride in infant formula and adult nutritionals Author: TRISKELION Reviewer Name: Reviewer 1

Summary of Method:

A portion of sample is suspended in 3M sulfuric acid to liberate bound fluoride. Derivatization with TMCS is then completed. The TMFS derivative is then analyzed by headspace GC-FID. Method Scope/Applicability: Method notes applicability to infant formula and adult nutritional products. Validation data include precision data on several infant formula and adult nutritional NIST SRMs (NIST 1869 (IF/AN), 8260 (hydrolyzed milk IF), and 8261 (AN), 1869a (IF/AN)). SLV does not indicate the matrix used for spiking. It appears only powders were tested during SLV. General comments about the method: Method states sulfuric acid is employed to release fluoride bound to matrix. The SMPR defines fluoride as the ion or anionic salt, which indicates free fluoride. Acid treatment may liberate additional bound fluorine. Method Clarity: The method is missing several critical elements, including calibration standard preparation, system suitability, and analytical quality control. The other portions of the method are clearly written. Pros/Strengths: • Straight forward method steps. Cons/Weaknesses • Setpoints for described headspace sampler (PE Turbomatrix 40) will likely require optimization if utilizing a different headspace system vendor (Tekmar, Gerstel, etc.). Transferability between headspace samplers might be challenging. • Scope limited to powders only. Reconstitution (25g + 200 g water) is an option in the method, but it has not been validated. • Incomplete details on calibration standard prep. • System suitability and QC requirements not defined.

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• No safety information on TMCS and sulfuric acid.

Supporting Data • General Comment:

Table 1 is confusing for LOQ and Analytical Range.

The calibration curves for the Day 1 -3 used a narrower range of 0.05 – 1 µ g F - /vial. Only one calibration curve demonstrated the wider range of 0.05 – 10 µ g F - /vial. The lowest calibration levels (0.05 and 0.1 µ g F) for the wider calibration range showed significant residual deviations (49.9% and 19.8%, respectively). Additional explanation of the deviation and analysis on more days for the wider calibration range is recommended. Specificity of this method/technique is not presented. A statement or testing demonstrating the specificity, including the impact of the acid treatment on bound forms of fluoride, is recommended.

Method Optimization:

Matrix matched standard calibration is needed to address the impact of background components on the efficiency of headspace sampling. The method may already use this calibration approach, but it is not described in the procedure.

• Performance Characteristics:

Analytical Range:

According to Table 1 the analytical range is 15-3000 ug/100 g powder. Using the target sample weight (0.3g) and the lowest (0.05 µ g F - ) and highest (10 µ g F - ) concentrations in the calibration tables, the measured fluoride range correlates to 17 – 3,333 µ g F - /100g powder. Using the 25 g powder + 200 g water reconstitution, the range corresponds to 2 – 370 µ g F - /100g recon. The wider calibration range appears to exceed the SMPR range. Recommendation is to provide additional explanation of how the analytical range in Table 1 is derived.

LOQ:

The LOQ from Table 1 is 17 µ g F - /100g powder, 22 µ g F - /100g reconstituted powder, and 2.2 µ g F - /100g liquid. Explanation of how these LOQ were determined is recommended.

Accuracy/Recovery:

Typo in calculated recovery table. QC-LOW corresponds to 50 µ g F - /100 powder. 150 µ g F - /100 powder is listed for this column. All recoveries meet SMPR. Additional description of how the average innate concentrations were determined is recommended. The QC-MED concentration is outside the narrower calibration range. An explanation of the calibration curves used for each spiking level is recommended.

Precision (RSD r ):

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Intermediate precision is presented. All precision data meet SMPR for RSD r , except for lowest measured amount. All measured concentrations are below the SMPR LOQ (30 µ g F - /100g reconstituted powder). Precision data support method provides acceptable precision.

Reproducibility (RSD R ):

RSD R data not submitted.

• System suitability:

No system suitability data submitted.

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

1. Is the Validation Study Report in a format acceptable to AOAC? No. 2. Is the method described in sufficient detail so that it is relatively easy to understand, including equations and procedures for calculation of results (are all terms explained)? No. Calibration standard preparation, system suitability, and analytical quality control are omitted. Recommend updating method to include these elements. 3. Are the figures and tables sufficiently explanatory without the need to refer to the text? Clarification of Table 1 is recommended. 4. Are all the figures and tables pertinent? Recommend adding chromatograms of mid-level calibration, matrix without F - spike, and matrix with F - spike. 5. Could some be omitted and covered by a simple statement? No. 6. Are the references complete and correctly annotated? Yes. 7. Does the method contain adequate safety precaution reference and/or statements? No safety precautions included in the method. While this method presents promising data for the requested measurement, additional clarification and details are recommended before First Action approval. Recommended updates include: • Revise method with calibration standard preparation, system suitability, quality control procedures, and safety statement. • Validation of reconstitution procedure. • Clarify Table 1 for Analytical range and LOQ. • Add information on method specificity and the impact of acid treatment on bound fluorine. • Describe how innate fluoride concentrations were obtained for the spike experiments. • Explain the calibration curves used for each spiking level. • Add chromatograms of mid-level calibration, matrix without F - spike, and matrix with F - spike. Recommendation:

4

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition Evaluation of Method: Fluor-06

Title: Quantification of fluoride by HS-GC-FID

Author: (Triskelion)

Reviewer Name:

Reviewer 2

Summary of Method: Infant formula samples are dissolved in 3M sulfuric acid followed by derivatization of free fluoride trimethylfluorosilane using trimethylchlorosilane in an GC headspace vial. The sample is then analyzed by HS-GC-FID

Method Scope/Applicability: The method study covers the fluorine in infant formula

General comments about the method: To an aliquot of infant formula 3M sulfuric acid is added to release bound fluoride. Fluoride is derivatized to trimethylfluorosilane by the addition of trimethylchlorosilane. The complete procedure is performed in an amber colored headspace vial and directly subjected to HS-GC-FID analysis. Concerns on how derivatization done in aqueous solution. Generally, these are done in dry conditions and TMCS is unstable in water. Method Clarity: Clearer instruction on sample preparation provided however, details on construction of calibration standards over appropriate range are lacking. The equation given to calculate final results has error.

Pros/Strengths: Simple method within vial derivatization applying a commonly used analytical technique.

Cons/Weaknesses

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

There are insufficient details presented to justify the appropriateness of using TMCS in an aqueous environment. TMCS reacts with nucleophiles and water is present in large quantities likely will react to TMCS. Additionally, insufficient data to show that fluoride is stronger nucleophile than chloride. In polar protic solvents (ie water), chloride is a stronger nucleophile than fluoride. Fluoride is only stronger nucleophile in polar aprotic solvents.

Supporting Data

General Comment: Would like to see chromatogram for sample, standards, and blank. Performance Characteristics: While performance characteristics look good, more details needed.

Analytical Range: Difficult to assess the analytical range as this is given for standards only and without information on makeup of standards and how these related to sample can only go by amount of analyte in vial and not to amount in sample. Plot given is linear showing dose response.

LOQ: No data on LOQ available to assess nor how calculated.

Accuracy/Recovery: Calculated recovery ranges from 92.0–106.1% Meets requirements of SMPR

Precision (RSD r ): Repeatability range 1–2% Meets requirements of SMPR

Reproducibility (RSD R ): No reproducibility data.

System suitability: No system suitability

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

1. Is the Validation Study Report in a format acceptable to AOAC? Key details needed for validation parameters other than precision and recovery.

2. Is the method described in sufficient detail so that it is relatively easy to understand, including equations and procedures for calculation of results (are all terms explained)? Detail is sufficient, however underlying science behind method needs more evidence.

3. Are the figures and tables sufficiently explanatory without the need to refer to the text? No.

4. Are all the figures and tables pertinent? Yes.

5. Could some be omitted and covered by a simple statement? No.

6. Are the references complete and correctly annotated? Yes.

7. Does the method contain adequate safety precaution reference and/or statements? No safety section or comments on safety written within method.

Recommendation:

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition Evaluation of Method: Phospho-02

Title: Quantification of phospholipids by UPLC-MS/MS Author: Francesca Giuffrida and Isabelle Tavazzi (Nestlé)

Reviewer Name:

Reviewer 1

Summary of Method: Phospholipids are extracted from infant formula samples into chloroform/methanol solution. After centrifugation the liquid phase is transferred to another tube with separation of organic phase by addition of potassium chloride solution. This organic phase is blown down under nitrogen, after which the residue is resolubilized in chloroform: methanol solution and filtered by centrifugal filtration. The residue is once again blown down under nitrogen and resolubilized in chloroform methanol. Analysis is performed by UPLC-MS/MS with a range of MRM transitions monitored for each phospholipid class according to fatty acid profile. Calibration of each phospholipids class is performed against a variety of different sources of phospholipids.: PC is calibrated against Egg PC PE is calibrated against 5 particular PE standards or varying fatty acid composition SM is calibrated against a SM milk standard

PI is calibrated against 4 particular PI standards of varying fatty acid composition PS is calibrated against 3 particular PS standards of varying fatty acid composition

Method Scope/Applicability: The method study covers the measurement of PC, PE, PI, PS and SM in infant formula

General comments about the method: The method uses a standard technique for extraction of phospholipids from samples using Folch extraction. The specific molecular species chosen for standard may or may not represent the typical distribution in different products which may contains different source of phospholipids, thus may end up with more bias than a natural standard mixture or bovine source. The proportions of different molecular species have been optimized for the particular samples used in this study. The appropriateness of these proportions may be of concern particularly when using real-world samples rather than reference materials. Question around method robustness:

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Long washing up to 4 hours between batches Short column age (700 injections, that is roughly 35 injections per day, including approximately 10 standards and 20 samples plus few blanks, for 20 days, so equivalent to 1 column very month???) Peak tailing caused by range of molecular species? And this may get worse as column getting older thus affect peak integration? Method Clarity: Clearer instruction on making up of standard solutions needed. For example, Table 2 references to volume of “mix of standards” to be blown down and reconstituted in 250uL. Not clear what this mix of standards refers to. The MS method include 3 periods of MRM (section 5.2.3), but lack of detail of time range, also what is the MRM in the last period? Pros/Strengths: Standard technique for extraction of phospholipids from samples using Folch extraction. Use of tandem MS provide high degree of selectivity to method (see Cons below) Cons/Weaknesses The specific molecular species chosen for standard may or may not represent the typical distribution in different products which may contains different source of phospholipids, thus may end up with more bias than a natural standard mixture or bovine source. Are users fixed to using this proportions? Multiple steps used including sample preparation, chromatographic separation, and MS analysis, could bring in experimental variations of different species of each class in each step. These variations impact on accurate quantification of phospholipid class. These errors are unlikely to be corrected with the standard curves that are produced under different conditions to sample analysis. An internal standard would in part correct for some of these issues in an MS method. Some LC systems may not be suited for analysis as a larger column compartment to fit 250mm column is required as well as chloroform resistance seals. Limited range of sample materials used in SLV analysis. General Comment: Not quite sure how to interpret the “recovery” part of the raw data. What is Ref value? I did assume that this was NMR value, but this does not tally with NMR values given in table Confused if raw data given for recovery is only 1–10%-ish? How do they link to the last few tables in Section 7.6? When I calculated the repeatability using same formula given in appendix using raw data, I can’t get the same names as given in report. Difference not huge but may be significant near SMPR limit, but confused over difference. Supporting Data

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Performance Characteristics: Precision as repeatability and intermediate precision was assessed by replicate analyses of 4 different infant formula materials.

Analytical Range: Analytical range specified for each class is:

PC 0.02–0.73 mg/100g PE 0.02–0.72 mg/100g PI 0.02 –0.48 mg/100g PS 0.01–0.24 mg/100g SM 0.02–0.72 mg/100g

All ranges meet SMPR except PS. Calibration appears to be linear but method states polynomials used. No regression equations given so cannot assess curves (either linear or polynominal). Confidence interval of intercept excludes zero, hence interpolation at lower concentrations may have a higher error.

LOQ: Acceptable LOQ values obtain for all 5 phospholipid classes.

Accuracy/Recovery: Bias against 31P NMR showed no statistically significant difference between methods for the samples measured Recovery range:

PC 96–105% PE 96–106% PI 87–102% PS 85–100% SM 92–103% Meets requirements of SMPR

Precision (RSD r ): Repeatability range:

PC 2.5–5.5% PE 1.8–6.0% PI 5.0–7.4% PS 3.3–8.4% SM 0.8–4.8% Repeatability exceeds SMPR for PI and PS

Reproducibility (RSD R ): No reproducibility data.

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

System suitability: Precautions given on setting up system, but no system suitability criteria are given. Check of results against QC sample verified by NMR expected, but no criteria.

4

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

1. Is the Validation Study Report in a format acceptable to AOAC? Yes

2. Is the method described in sufficient detail so that it is relatively easy to understand, including equations and procedures for calculation of results (are all terms explained)? No equations given explaining how the measured results are calculated. Phospholipid content is extrapolated from calibration curve with polynominal calibration curve used. However, no details provided on order of polynominal or limits on characteristic of acceptable non-linear curve.

3. Are the figures and tables sufficiently explanatory without the need to refer to the text? For most part yes.

4. Are all the figures and tables pertinent? Yes.

5. Could some be omitted and covered by a simple statement? No.

6. Are the references complete and correctly annotated? No published references to method are given.

7. Does the method contain adequate safety precaution reference and/or statements? No safety section or comments on safety written within method.

Recommendation:

5

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition Evaluation of Method Phospho-02 July 2022

Title: Quantification of phospholipids by UPLC-MS/MS Author: Francesca Giuffrida and Isabelle Tavazzi, Societe des Produits Nestle S.A. Reviewer Name: Reviewer 2

Summary of Method: Phospholipids are extracted using organic solvents and separated according the PL head groups on a silica column using a gradient mobile phase. Chromatography and detection by LC-MS/MS using electrospray ionization source. Quantification is conducted using external calibration method. All molecular species from each phospholipid class are grouped and quantified together. Method Scope/Applicability: Applicable to the determination of phospholipids in Infant formula and Adult nutrition (phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol (PI), phosphatidylserine (PS), and sphingomyelin (SM)) General comments about the method: The method takes advantage of MS technology to selectivity detect and quantify classes of PL’s by their head groups. The method was developed and published in 2018 and has been used since then. Compared with 2018 separation of Phosphatidylserine was improved. The quantification using external calibration method uses selected calibrants expected for infant formulas and adult nutritional’s and is straightforward.

Method Clarity: The method procedure is easy to follow. A system suitability section is included. We have not seen a specific safety section.

Pros/Strengths: • Selective detection of PL classes using LC-MS/MS • High detection sensitivity • Mix of standards designed and applicable for infant formulas and adult nutrition (and ingredients is characterized).

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• System suitability section • SPIFAN kit including NIST 1849b, NIST 1869, RM 8260 is used to obtain recovery data and precision data. •

Cons/Weaknesses • No (isotope labeled) internal standard in this LC-MS quantification method • No safety section

• Sample intake reflecting amount for 0.1 g fat = 0.3-0.5 g powder Only described for homogenous powders. Please describe additional procedure for re-constituted powders (25 g in 200 g of water). In pa 7.6 describes that re-constitution is used. • Rather long run time including washing steps etc.

Supporting Data • General Comment: Supporting data are well reported. Mix of standards for infant formulas (and ingredients is characterized. Sphingomyelin (SM) milk standard was selected because species proportion similar to infant formula is considered. Egg PC standard was selected, because species proportion similar to infant formula is considered, phosphatidylethanolamine PE mix standards (34:1; 34:2; 36:2; 36:4; 38:4) and phosphatidylinositol PI mix (34:1; 36:2) because species proportion similar to infant formula is considered. The SLV report include data to demonstrate LOQ, (analytical range) , recovery , accuracy against P-NMR data and precision. Typical chromatograms for mix of PL standards and IF sample are included. Original data are provided in appendices.

Method Optimization: Selectivity for PS is according to requirements. Polynomial calibration is used, not forced through zero.

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• Method Performance requirements

• Performance Characteristics:

Analytical Range: SMPR 0.4-50 mg PC /100g RTF or 0.4-44 mg other PL’s /100 g RTF

Calculation/expression towards mg PL/100g RTF is not clear in the report. Assuming that range is covered.

LOQ: The limit of quantification was determined as 0.1, 0.1, 0.06 ,0.03 and 0.1 mg/100 g RTF for Pc, PE, PI, PS and SM respectively, meets SMPR 0.4 mg/100 g RTF

Accuracy/Recovery: Spikes recovery for 3 matrices on one level and for one IF matrix on two spike recovery target levels at 6 days analyzed in duplicate did meet SMPR (SLV guideline: Each selected matrix will be spiked at three levels. Use spike levels covering the analytical range specified in the SMPR. Spiked and unspiked samples will be analyzed in triplicates on each of three (3) days).

PC 96-105% (SMPR 90-110%) PE 96-106% (SMPR 90-110%) PI 87-104% (SMPR 85-115%) PS 85-100% (SMPR 85-115%) SM 92-103% (SMPR 90-110%)

Accuracy was compared with P-NMR form external laboratory for PC, PE, PI and SM. P-NMR is seen as reference in this field. Data for PS are missing. Data show Accuracy (after deletion of outliers) on average for PC =90%; SM = 102%; PE = 119 % and PI = 101%, both methods with distinct technology gave similar result. Accuracy compared with other technique/laboratory as such is not a requirement in SMPR

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Precision (samples were analyzed in duplicate on each of six (6) days) Repeatability (RSD r ):

PC 2.5-4.9% (SMPR<7%) PE 1.8-6.0% (SMPR <7%) PI 5.0-7.4% (SMPR <7%) PS 3.3-8.4%* (SMPR <7%) SM 0.8-4.8% (SMPR <7%) *very low level

Repeatability figures meet SMPR requirements

Reproducibility (RSD R ): intermediate PC 9.4-13.9% (SMPR<14%) PE 6.6-11.4% (SMPR <14%) PI 13.5-14.6% (SMPR <14%) PS 10.9-13.9 (SMPR <14%) SM 6.5-13% (SMPR <14%)

• System suitability:

In section 5.2.4 system suitability aspects are mentioned. They seemed based on years of experience . It is not clear what the definition is of a batch sequence. No criteria are mentioned for a suitable QC sample

Additional comments and questions. 1. The method author mentioned sample intake related to 0.1 fat. In accuracy section reconstitution (25 g in 200 g water) is applied. What sample intake was applied or reconstitution for precision and recovery experiments? 2. In section 4.1.5 Acetonitrile/Acetone/isopropanol (4/3/3) is stated, but seem not used in the method. Please clarify? 3. It is not clear at which point in the process spikes are added in the recovery experiments. Please clarify. 4. How does a batch sequence look like, in terms of number of samples, blancs? 5. What would be criteria for a suitable QC sample in terms of minimal or max PL concentrations? 6. Data seemed obtained by duplicate analysis. Do method authors advice to run samples in duplicate in routine? Please clarify? 7. In accuracy data related to P-NMR no data for Phosphatidylserine are shown, Please clarify?

4

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

1. Is the Validation Study Report in a format acceptable to AOAC? YES 2. Is the method described in sufficient detail so that it is relatively easy to understand, including equations and procedures for calculation of results (are all terms explained)? Yes, 3. Are the figures and tables sufficiently explanatory without the need to refer to the text? Not completely, eg missing sample intakes for different sections and abbreviations in table 5. 4. Are all the figures and tables pertinent? YES 5. Could some be omitted and covered by a simple statement? No, figures and tables show and explain the SLV data 6. Are the references complete and correctly annotated? No, references on SMPR and publication (2018) could be included 7. Does the method contain adequate safety precaution reference and/or statements?

No, safety section should be included.

Recommendation: The method shows excellent performance characteristics according to the SMPR. The source of PL standards is provided including its specification and how this is related to forms measured in samples The accuracy is assessed by recovery against P-NMR as no infant formula certified reference materials are available. Method meets requirements of SMPR. We recommend to move this method forward to first action, after questions have been answered satisfactorily.

5