AOAC SPIFAN ERP Reviewer Forms (March 16, 2017)

AOAC INTERNATIONAL Stakeholder Panel on Infant Formula and Adult Nutritionals (SPIFAN) REVIEWER FORMS

Thursday, March 16, 2017

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

UNITED STATES 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-1

Title: Single Laboratory Validation Report for Total Amino Acids by UHPLC-UV in Infant Formulas and Adult Nutritionals

Author: Greg Jaudzems, Sabine Lahrichi and Christophe Fuerer (Nestle)

Summary of Method: Proteins are hydrolyzed in 6 M HCl for 24 h at 110 °C in presence of phenol, 3-3’- Dithiodipropionic acid (DDP) and norvaline. Phenol (0.1%) is added to prevent halogenation of tyrosine. Norvaline is added as an internal standard. DDP is added to convert cystine and cysteine to S-2-carboxyethylthiocysteine (XCys) and the resulting derivative can be separated from other amino acids for quantification. After neutralization, amino acids and converted cysteine (XCys) are derivatized with 6- aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC). Derivatized amino acids are separated using reversed phase UHPLC with UV detection at 260 nm. Method Scope/Applicability: This method allows the determination of the following amino acids: alanine, arginine, aspartic acid (combined with asparagine), cystine (dimer of cysteine, combined with cysteine), glutamic acid (combined with glutamine), glycine, histidine, isoleucine, leucine, lysine, methionine, phenylalanine, proline, serine, taurine, threonine, tyrosine, and valine. This method is not suitable for the determination of tryptophan. This method is applicable to infant and adult/pediatric nutritional formulas and other matrices such as infant cereals and pet foods.

General comments about the method:

Method is clearly written and should be easy to follow

Method uses established methodology of acidic hydrolysis and pre-column derivatization to determine amino acids All SPIFAN II kit matrices are analyzed and results for repeatability, intermediate reproducibility and recoveries are presented.

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Data are well organized and clearly presented

Method Clarity:

Method is clearly written.

Method lists all the necessary supplies and analysis steps

Method Safety Concerns:

No major safety concerns are noted

Pros/Strengths:

• Uses established methodology and commonly available equipment • Fast chromatographic analysis • Method describes alternative eluants and reagents preparations procedures – from commercially available kits as well as “from scratch” using common chemicals • Data for all SPIFAN matrices are presented • Linearity and other parameters are evaluated and compared with SMPR

Cons/Weaknesses

Accuracy data and intermediate reproducibility data for many amino acids don’t meet SMPR

Method is not suitable for analysis if Tryptophan

Supporting Data

- Method Optimization: Some method optimization may be needed if baseline separation for amino acids is not achieved using chromatographic conditions described in the method

• Performance Characteristics:

Analytical Range and Linearity:

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

The linear calibration range from 0.25 to 50 pmol/μL was established for all amino acids except Cystine (corresponds to average concentrations between 2 and 820 mg/100 g, depending an amino acid).

For Cystine linear calibration range was between 0.25 – 10 pmole/uL.

It was determined that about 2 % of measured values in SPIFAN II kit were above the calibration range.

The linear range doesn’t meet SMPR requirements

LOQ:

Was determined as 2.5 mg/100 g. This doesn’t meet SPMR requirements of 0.4 mg/100 g but based on actual amino acids concentrations found in the matrices this level seems to be sufficient for the scope of analyzing amino acids in baby formula and adult nutritionals.

Accuracy/Recovery: Accuracy was evaluated by doing spike-recoveries studies on SPIFAN matrices as well as analyzing NIST standard reference material 1849a and comparing the results with CofA. For SRM 1849a results for 3 amino acids (Aspartic Acid, Lysine and Tyrosine) fell outside of certified range. Spike recoveries showed that recoveries for 23 % of analyzed amino acids didn’t meet SMPR. The authors don’t described what spike level was used

Precision (RSD r ):

Repeatability data for majority of amino acids in all the matrices met SMPR requirements. Only a few were outside the range but still were very close to specifications and none exceeded 4%.

Reproducibility (RSD R ):

Intermediate reproducibility was determined for all SPIFAN II matrices. Intermediate reproducibility for majority (67%) of amino acids didn’t meet SMPR requirements with considerable number of values being well above the highest acceptance criteria of 5%. Intermediate reproducibility for many amino acids was higher than 10 %

• System suitability: Notes about performance of the system are included in the method

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Recommendation:

The method is well written and all the required validation data are collected and clearly presented for review. Though the method is very promising and showed good repeatability there are following method performance problems that need to be addressed before this method could be recommended for first action official method: 1. Intermediate reproducibility data for majority of amino acids don’t meet SMPR requirements with some RSDiR exceeding 10 or even 20%. Providing that repeatability data were within SMPR specifications and the chromatograms look very clean with well separated peaks explanation for such variations in results is needed or this will be a problem if the method moves to multi-lab study. 2. Results of analysis for SRM 1849a don’t meet SMPR with 3 amino acids falling outside the certified range. 3. The method has narrow analytical range that doesn’t meet SMPR. It was noted by the authors that several results from the SPIFAN kit matrices were falling outside the calibration range. This limitation should be clearly noted in the method and recommendations provided of how to deal with the samples outside the linear calibration range.

This method is not recommend as the first action at this time

4

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method ___Amino-01____________

Title: SLV Report for Total Amino Acids by UHPLC-UV in Infant Formula and Adult Nutritionals

Author: Nestle – G. Jaudzems

Summary of Method:

Samples are hydrolyzed in 6M HCL for 24 Hours at 110C, neutralized and then derivatized with 6- aminoquinolyl-N-hydroxysuccinimidyl. The derivatized sample is then analyzed by UHPLC-UV at 260 nm.

Method Scope/Applicability:

The method is capable of determining all SMPR required amino acids except for tryptophan.

General comments about the method: The method has become a commonly used one in the industry and is utilized in multiple laboratories around the world, however, method requires specialized equipment and supplies from a single vendor. We have rejected good methods in the past that require specialized equipment.

Method Clarity: Method was easy to follow

Method Safety Concerns: No major Safety concerns

Pros/Strengths:

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• Single method to determine majority of SMPR required analytes

Cons/Weaknesses • Specialized equipment and reagents

Supporting Data • General Comment:

- Method Optimization:

• Performance Characteristics:

Analytical Range: 2.5 to 500 mg/100g (fails SMPR of 0.4 to 2500)

LOQ: Fails LOQ of 0.4

Accuracy/Recovery: Low recovery on Methionine across the board. 25% of recovery experiments fail to meet SMPR requirements.

Precision (RSD r ): Precision is mostly acceptable across the board (+90% of all analytes across all matrices), however, there were a few that failed but significantly. I would be ok with the precision obtained.

Reproducibility (RSD R ): N/A

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• System suitability:

N/A

Recommendation: Due to past practice, I have difficulty recommending this for 1 st action unless it can be less one Vendor specific.

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method AMINO-02 Title: HPLC Determination of Total Tryptophan in Infant Formula and Adult/Pediatric Nutritional Formula Following Enzymatic Hydrolysis

Author: Abbott Nutrition, Division of Abbott Laboratories

Summary of Method: Sample preparation consists of adding a weighed sample, the enzyme solution, internal standard (5- methyl-DL-tryptophan) and Trizma buffer into a tube. The preparation is mixed and incubated at 50°C for sixteen hours (overnight), to assure complete hydrolysis of all sample types. After hydrolysis, the sample-enzyme mixture is diluted to 50mL with methanol/water and filtered. The sample is injected onto a C-8 column with reference standards and enzyme blank preparations and the Tryptophan and 5-Methyl-DL-tryptophan are detected by native fluorescence at an excitation wavelength of 295 nm and emission wavelength of 345 nm. Tryptophan is quantified by internal standard calibration mode using reference standard to internal standard amount and response ratios.

Method Scope/Applicability: This HPLC method allows for the determination of total tryptophan in infant, pediatric, and adult nutritional products.

General comments about the method: The method is clear, simple and easy to follow.

Method Safety Concerns: No specific concerns on safety.

Pros/Strengths Method ruggedness were established run the samples on two HPLC systems, use of two different stationary phase batches, precisions runs prepared by two analysts. The results obtained show the ruggedness of the method.

Supporting Data

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• General Comment:

• Performance Characteristics:

Analytical Range: SMPR: 0.4 – 2500mg/100g

0.18 – 300mg/100g.

LOQ: SMPR: ≤0.4mg/100g

0.18mg/100g

Accuracy/Recovery: Spike recovery requirements were met for all SPIFAN matrices.

Recovery data ranged from 93.8-104.9%.

Precision (RSD r ): SMPR : 0.4 - 0.5 : ± 4% 5.0 - 50: ± 3% 50 - 2500: ± 2%

Results ranged from 0.2 –2.0 %

All fourteen fortified SPIFAN matrices were prepared and analyzed in duplicate on six days.

Repeatability requirements were met for all SPIFAN matrices with the exception of the partially hydrolyzed soy-based powder.

Reproducibility (RSD R ): -----

• System suitability:

- The SRM 1849a mean result of 181.3 mg/100g is 98.5% of the 184 mg/100g reference value.

- The use of several blank enzyme solutions per run.

-

Linearity:

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Three independent experiments were run with eight levels of tryptophan ranging from approximately 0.2 to 30 mg/L. The method demonstrated good linearity over a standard range of 0.2-30 mg/L Tryptophan with eight dilution levels having average percent deviation from theoretical value of 0.3%.

Recommendation:

The method meets the SMPR requirements.

I recommend move the method to First Action for determination of Total Tryptophan.

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method Amino-02

Title: HPLC Determination of Total Tryptophan in Infant Formula and Adult/Paediatric Nutritional Formula Following Enzymatic Hydrolysis

Author: Abbott Nutrition

Summary of Method: Internal standard, 5-methyl-DL-tryptophan, and protease enzyme solution, (containing at least 10 proteolytic enzymes), are added to weighed samples diluted in trisma buffer. Duplicate blanks are included with each analysis set. The samples set is incubated overnight at 50 O C, cooled and made up to 50 mLs with methanol/water. Tryptophan is quantitated on a C-8 RP column on an isocratic HPLC system using a 4 point ISTD calibration with FLD detection, (Excitation 295nm, Emission 345nm). Method Scope/Applicability: This method is applicable to the determination of total tryptophan in infant formula and adult/pediatric nutritional formulas. The LOQ of a typical ready to feed formula is 0.18 mg/100g and the analytical range of the method is 0.18 – 300 mg/100g. Method is clearly written but needs more detail; part numbers or at minimum, CAS numbers, safety section and more detail in the sample preparation section re sample weight Method Clarity: A 25 g in 200 g water should have been performed as per SLV protocol to ensure homogeneity. Possible two weighing options, one for homogeneous and one for non-homogenous products. General comments about the method:

The following calculation could be used so that analyst is clear on the sample weight to use:

w = 1000 * 200 P m

W = amount of sample to be weighed 1000 = factor to convert from g to mg P

= amount of protein in the sample (g/100g)

M

= Sample wt (~ 25g) = Wt of dilution water

200

Method is for total tryptophan but free tryptophan is mentioned in the method also!

Recommend to add chromatograms to method – can use ones from SLV report.

Method Safety Concerns:

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

General safety section needs to be added to the method

Pros/Strengths: • Well laid out method with good detail. • ISTD added at beginning of sample preparation • Instrument and reagents (other than enzyme) are commonly used in the lab • No hush sample treatment is involved. Cons • Enzyme effectiveness will be critical for the assay. Need to mention how to ensure the effectiveness of the enzyme and how it is controlled. • More specific information required for sample weight to be used. • System suitability: As per SMPR, System Suitability Tests and/or Analytical Quality Control should be included in the method. An enzyme blank is included but others required, for example, standard repeatability, slope accuracy, linearity > 0.995

Supporting Data • General Comment:

- Method Optimization: Optimized with enzyme blank subtraction

Method performance requirements:

SMPR: Applicability: Determination of free and/or total proteinogenic L-α-amino acids and taurine (as shown in Table 1) in all forms of infant, adult, and/or pediatric formulas (powders, ready-to-feed liquids, and liquid concentrates).

Analytical range

0.4–2500 mcg/100g RTF

Limit of quantitation (LOQ)

≤ 0.4 mcg/100g RTF

Recovery

90–110% 0.4–5.0 5.0 – 50 50 – 2500 0.4–5.0 5.0 – 50 50 – 2500

Repeatability (RSDr)

≤ 4% ≤ 3% ≤ 2 % ≤ 6% ≤ 4% ≤ 3 %

Reproducibility (RSDR)

Linearity: Minimum of six levels (levels to be agreed upon by SDs) that span the desired working range. (b) Relative error of back-calculated concentrations determined within the desired working range. (No specifi c criterion in standard method performance requirement. Recommend calibration errors to be <5%.) (c) Minimum of three independent experiments. (Independently prepared standards, if feasible.)

Analytical Range: 0.18-300mg/100g of RTF sample (including reconstitute sample)

Linearity: 0.2 – 30 mg/L tryptophan. 8 levels in triplicate with max 1.35 % calibration error

LOQ: 0.18 mg/100g. 0.007 mg/L * 10 = 0.07 mg/L

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

0.07 m/L * 0.05 L * 100 * 1000 = 0.0175 mg/100g powder/Liquid 2 (g/mL) * 1000

Accuracy/Recovery: 93.8 – 104.9 %

SRM 181.3 mg/100g vs Ref value of184 mg/100g – 98.5 % recovery

Precision (RSD r ):

5.0 – 50 mg/100g 50 – 2500 mg/100g

≤ 0.2 – 2.0% except HA with soy at 5 %

≤ 0.3 – 1.9 %

Intermediated Precision:

5.0 – 50 mg/100g 50 – 2500 mg/100g

≤ 1.0 – 2.4% except HA with soy at 6.9 %

≤ 1.9 – 3.7 %

System suitability:

Method didn’t mention system suitability

Only enzyme blank included – recommend more are added

Recommendation:

1. 5-M-trp as the internal standard added before hydrolysis is not necessary, since it didn’t monitor the enzymatic hydrolysis loss. If weighing and dilutions are performed accurately (analytical balance and volumetric flask), the internal standard can be added late or not needed at all. 2. The performance of this method is very good (precision and accuracy) in most of matrices except a little out of SMPR for soy hydrolyzed infant formula product (precision). 3. The design of spike-recovery is very reasonable, both using BSA to the samples matrices and comparing to SRM 1849a and SRM927e (BSA). The recovery for BSA demonstrates the accuracy better than SRM1849a, since the content of tryptophan in BSA is decided by the amino acid profile of the protein. 4. Recommend for First Action and MLT Study .

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method ___CAROT-3 _____

Title: Determination of beta-Carotene, Lutein and Lycopene in Infant and Adult/Pediatric Nutritional Formula Using High Performance Liquid Chromatography With UV Detection. Author: Abbot Nutrition

Summary of Method: After protected (sodium ascorbate and BHT) saponification using 3 % KOH in 10:4:1 MeOH:H2O:THF, lutein, beta-carotene, and lycopene are extracted from infant, pediatric, and adult nutritionals using 75:25 hexane:MTBE. After evaporation to dryness, the carotenoids are dissolved 75:25 MeOH with 10% BHT:MTBE, then separated and quantitated by HPLC using a methanol:MTBE gradient mobile phase through a C30 column terminated with UV detection at 445 nm.

Method Scope/Applicability: I nfant, pediatric, and adult nutritionals in powder, ready to feed liquid, and liquid concentrate forms.

General comments about the method: Author(s) should consider rewriting the method summary to reflect actual conditions of saponification (see above). Is the methanol used in Sample Prep step d the methanol solution with BHT added? If so, that should be so stated. If not, does the author think the BHT present in the THF is sufficient to protect the carotenoids through the procedure up to step o? Determination of Standard Concentrations step b may not be adequate for purpose. What if a standard of any one of the analytes is purchased that has an impurity that absorbs @ 445 and is not in the respective time range? Worse yet, what if that impurity falls in another analytes time range, thus overstating the purity of the analyte in consideration, and understating the purity of the other analyte? I believe each of the analytes should be injected independently, and data collected at its respective absorbance wavelength shown in 6a. I am assuming the data in columns 3 and 4 of the Method Performance vs SMPR requirements are reversed, otherwise this method is clearly out to lunch and deserves no further consideration.

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Assuming the columns are reversed, I believe this is an excellent method with results for carotenoids better than the current state of analytical practice.

Method Clarity: Needs some minor clarifications-see above.

Method Safety Concerns: None

Pros/Strengths: • Excellent method, better than the current state of analytical practice.

Cons/Weaknesses • Write up needs some attention to detail (see above)

Supporting Data • General Comment: This is all in the tables in the report, so I won’t spend time retyping it.

- Method Optimization:

• Performance Characteristics:

Analytical Range:

LOQ:

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Accuracy/Recovery:

Precision (RSD r ):

Reproducibility (RSD R ):

• System suitability:

Recommendation:

I recommend this method be adopted as AOACI Official First Action.

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method __CAR-03___________

Title: Carotenoids Lutein, beta-Carotene, Lycopene Single Laboratory Validation

Author: Abbott Nutrition

Summary of Method: This method extracts the carotenoids from the product using hexane/MtBE following saponification of the sample. The sample is then dried and redissolved in a BHT/MtBE mixture prior to analysis using HPLC-UV with C30 chromatography utilizing a gradient.

Method Scope/Applicability: Method appears to be ideal for compounds of interest in SPIFAN initiative.

General comments about the method: Simple method for the determination of the carotenoids using common laboratory supplies and equipment. Could be easily set up anywhere in the world.

Method Clarity: The Method was clear and easily understood.

Method Safety Concerns: No major safety concerns about this method.

Pros/Strengths: • Method has the ability to determine lutein, lycopene and beta carotene in a single injection that is not >45 minutes/injection • Common Lab instrumentation and supplies used for analysis

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• No major concerns over the chemicals involved •

Cons/Weaknesses • Spike recoveries on the lutein are all below 100%. Is the method under recovering lutein or is the saponification too harsh on the sample destroying the lutein?

Supporting Data • General Comment: Positive that multiple analysts used in SLV

- Method Optimization:  Were multiple column manufacturers used for C30  Multiple Instruments?

• Performance Characteristics:

Analytical Range: Method meets SMPR requirement

LOQ: Method Meets SMPR requirement

Accuracy/Recovery: Recoveries were acceptable except for Lutein. Though they meet the SMPR of 90-110% requirement, the recoveries are all below 100%.

Precision (RSD r ):

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Concerned with high RSD with Child Elemental Powder being 14.88% RSD on a sample with levels around 34 mcg/100g. Other high RSDs are with samples sub 1 mcg/100 levels, so not as concerned. RSDs on remaining products meet the SMPR requirements of <8% for the levels found in the product.

): N/A

Reproducibility (RSD R

• System suitability: System suitability was not stated in the method.

Recommendation: Move to First action after investigation of low lutein spike recoveries.

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method Caro-03

Title: Determination of β-Carotene, Lutein and Lycopene in Infant and Adult/Paediatric Nutritionals Formula using HPLC with UV Detection

Author: Abbott Nutrition

Summary of Method: Carotenoids are extracted from samples with 75:25 hexane: MTBE following saponification with a mix of KOH, THF and methanol. Following extraction a portion of the organic layer is evaporated to dryness and the residue dissolved in 75:25 10% BHT:MTBE. Analysis is by HPLC with separation of the cis and trans isomers, beta carotene and lycopene on a C30 carotenoid column with a methanol: MTBE gradient and detection by UV spec at 445nm.

Method Scope/Applicability: Applicable to the determination of cis and trans -lutein, cis and trans -β- carotene, and cis and trans lycopene in infant formula and adult nutritionals.

General comments about the method:

Need to include scope/applicability – can use sentence from introduction in SLV report

Requires more detail on sample weigh – sample weight for liquid products included but no detail on powder or reconstituted products. Assume 1 g is used for all but need to be stated.

To comply with SPIFAN guidelines, add sample preparation for homogeneous vs non-homogeneous samples – 25g plus 200 mL water etc.

Add note indicating that liquid and reconstituted samples should be prepared fresh prior to analysis.

Add chromatograms to the method - (in the SLV report).

Suggest laying out standard and sample concentration calculations in a clearer format with glossary.

Consider two curves – one for low concentration and one for high carotenoid concentrations or allow for optimizing curve to suit concentration of samples being analyzed.

Method Safety Concerns:

General safety section needs to be added to the method.

Pros/Strengths:  Good detail on materials, chemicals, reference standard and sample preparation.  Both UV Spectrophotometric and Chromatographic purity performed.

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Cons 

No ISTD

 4 point curve used – minimum of 6 levels is recommended for SLV

Method performance requirements:

Lutein Analytical Range: Trans Lutein 10-250 µg/L reference standard

6.25 – 156.25 mcg/100g product

Linearity: 0.99991 and <1 % error

LOQ: 0.40 µg/100g RTF

Accuracy/Recovery: 90.3 -95.3 %

Precision (RSD r ):

1-100 mcg/100g RTF - 1.89 – 7.32 %

14.88 % for child formula?

Intermediated Precision: 1-100 mcg/100g RTF - 1.93 – 8.06 %

13.99 % for Child formula?

Beta Carotene Analytical Range: Trans Lutein 25 - 500 µg/L reference standard

15.6 – 312.5 mcg/100g product

Linearity: 0.99993 and <1 % error

LOQ: 0.10 µg/100g RTF

Accuracy/Recovery: 89.3 -108.1 %

Precision (RSD r ):

1-100 mcg/100g RTF - 1.81 – 6.77 %

Intermediated Precision: 1-100 mcg/100g RTF - 3.07 – 16.20 %

Lycopene

Analytical Range: Trans Lutein 5-100 µg/L reference standard

3.125 – 62.5 mcg/100g product

Linearity: 0.9998 and <1.5 % error

LOQ: 0.30 µg/100g RTF

Accuracy/Recovery: 97.3 – 108.8 %

Precision (RSD r ):

1-100 mcg/100g RTF - 3.01 – 6.37 %

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

≤ 1.0 mcg/100g RTF 4.67 %

Intermediated Precision: 1-100 mcg/100g RTF - 4.29 – 10.29 % ≤ 1.0 mcg/100g RTF 7.73 %

System suitability:

Correlation of standard curves must be greater than 0.999 and standard response (drift) of not more than 10%. This is a bit high, suggest it should be no more than 3 %. Slope accuracy could be used instead of standard area to assess drift.

Suggest adding a peak resolution solution as fairly critical for this assay – 5 peaks of interest and depending on sample matrix, some very close to interfering peaks.

Recommendation:

 Recommend to be considered for First Action and MLT Study

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method MET -01/03 AOAC 2011.14

Title: Determination of Ca, Cu, Fe, Mg, Mn, P, K, Na and Zn by ICP-AES in Infant Formula and Adult Nutritionals

Author: Hans Cruijsen, FrieslandCampina Eric Poitevin, Nestle

Summary of Method: Microwave digestion, ICP-AES determination

Method Scope/Applicability: • SPIFAN Matrices • Does not work reliably at the low end of the concentration range specified in the SMPR. • The performance with the SPIFAN Placebo samples and the IDF dairy samples did not meet the RSDR specification in the SMPR for Cu, Mn, Fe and Zn.

General comments about the method: • Good method. Great multi-lab trial. The low end of the concentration range specified in the SMPR is too optimistic for the ICP-AES technique for Cu, Mn, Fe and Zn. The Reproducibility RSD shows this clearly with values as high as 56%.

Method Clarity:

Method Safety Concerns:

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Pros/Strengths: • ICP-AES is a proven, mature technology • MLT run with a great variety of instruments • Variety of microwave digestion systems were used o 6 CEM MARS o 3 Milestone Ethos

o 1 Analytikjana TOPWave o 1 Milestone UltraClave

o 1 NOVAWAVE o 2 unidentified • Variety of ICP instruments – o Old and new models

o Both Eschelle based optics and Paschen Runge/Roland Circle optics o 8 Perkin Elmer instruments – 4 models of varying age o 1 Thermo

o 1 Varian o 1 Agilent o 2 Spectro o 1 Unknown • Both Axial and radial view of plasma used

o 10 labs used axial view for Fe, Cu, Mn and Zn • Great performance on NIST 1849a. No bias for any element. • Not many outliers considering the size of the data set

Cons/Weaknesses

Supporting Data • General Comment:

 14 Laboratories from 11 countries  Included SPIFAN matrices plus IDF SAMPLES: milk, milk powder, whey powder, whey protein concentrate, butter and processed cheese

- Method Optimization:

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• Performance Characteristics:

Analytical Range:

LOQ:

Accuracy/Recovery : Great recovery with NIST 1849a. No bias

Reproducibility (RSD R

): and Precision (RSD r ):

• The overall RSDR for Mn was 12.2%. The SMPR limit is 10%. • The following IF and Adult Nutritional samples did not meet SMPR: Ca: Adult RTF Hi Fat RSDr=9.3 RSDR=27.8 Infant Elemental Powder RSDr=8.46 RSDR=12.9

Cu: Infant PH Powder Milk RSDr=7.52 Infant Powder Milk RSDSR=12.3

Fe: Adult RTF Hi Fat RSDr=7.81 RSDR=12.8 Adult RTF Hi Protein RSDr=6.14 Infant PH Powder RSDR=14.3 Infant Powder fos-gos RSDR=15.0

Mg: Infant Elemental Powder RSDR=12.1

Mn: Adult RTF Hi Fat RSDr=6.5 RSDR=17.5 Mean = 0.357 Infant PH Powder Milk RSDR20.7 Mean=0.012 Infant Powder fos-gos RSDR=13.0 Mean =0.012 Infant Powder Milk RSDr=6.30 RSDR=23.7 Mean=0.006 Toddler Powder RSDR=11.44 Mean=0.88

 IDF Dairy Samples: Cu: 4 failed both RSDr and RSDR, 1 failed RSDR Fe: 3 failed botgh RSDr and RSDR, 2 failed RSDR K: 2 failed RSDR Mg: 1 failed both RSDr and RSDR Mn: 2 failed both RSDr and RSDR, 1 Failed RSDR Zn: 1 failed both RSDr and RSDR, 1 failed RSDR

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• System suitability:

Recommendation: Approve. However, I think the SMPR low limits for Cu, Fe, Mn and Zn need to be raised. The current values are not realistic with current technology.

4

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method _MTE-01/MTE-03

Title: AOAC 2011.14 ISO/CD 15151 | IDF 229 Infant formula, Adult nutrition Milk and milk products — Determination of calcium, copper, iron, magnesium, manganese, phosphorus, potassium, sodium and zinc contents – Inductive coupled plasma atomic emission spectrometric method (ICP-AES) (January 2017)

Author: Hans Cruijsen, FrieslandCampina, The Netherlands Eric Poitevin, Nestle, Switzerland

Summary of Method: The organic matter is decomposed by wet digestion using nitric acid in a pressurized microwave assisted wet digestion system or any appropriate instrumentation in wet digestion and diluted. The test and calibration solutions are atomized into plasma of an inductively coupled plasma spectrometer and the emission is measured at appropriate wavelengths using external calibration. Method Scope/Applicability: This method is applicable to calcium, copper, iron, magnesium, manganese, phosphorus, potassium, sodium, and zinc contents in milk and milk products. The method has been tested with the SPIFAN sample set as well as formula milk, milk powder, whey powder, whey protein concentrate, butter and processed cheese. General comments about the method: 1) Method is well written. 2) Use microwave to assist sample digestion, sample preparation time is about 70 min. 3) Various ICP-AES instrument systems were used in the study. These include ICP-AES from Perkin Elmer, Thermo, Agilent, Varian, Spectro. Different models from Perkin Elmer were used also. 4) Comparing to the existing AOAC Final method that using ICP-MS, both methods has similar sample prep (microwave, nitric acid), and both tested with butter, cheese, whey powder, etc. in addition to the SPIFAN sample set. The ICP-MS method has been validated on additional three elements (chromium, selenium, and molybdenum).

Method Clarity: Method is well written, clear to follow

Method Safety Concerns: Safety hazards are addressed.

Pros/Strengths: • ICP-AES instrument from different vendors and different models from the same vendor were used in the MLT.

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

• ICP-AES is more common in labs (Availability / instrument affordability).

Cons/Weaknesses • Performance at low concentration levels for some elements (Mn, Fe, Cu)

Supporting Data • General Comment:

Well written and comprehensive report on the MLT. The report provides details in the study design and results analysis. It contains collaborators, study designs, equipment used, data results, comments from the participating labs, and final conclusion. The conclusion is fairly reasonable. - Method Optimization: • Performance Characteristics:

Analytical Range:

LOQ: Meet the SMPR requirement on Ca, Mg, P, K, Na, and Zn.

Accuracy/Recovery: Good performance on the NIST Reference material (1849a)

Precision (RSD r ) and Reproducibility (RSD R ): Good repeatability and reproducibility for all elements, except the three elements, Mn, Fe, and Cu at low concentrations. For manganese, iron, and copper, the RSDs at or near the LOQ (up to 50 times of the LOQ) are larger than the SMPR requirement. 1) Copper, at about 0.05 mg/100g RTF, the RSDs are about 10% (which is the requirement limit). The SMPR LOQ is 0.001 mg/100g RTF. 2) Manganese, at 0.008 the RSD R still above 16% (the limit is 10%), the SMPR LOQ is 0.001. The report has pointed out that if look at these RSDs in terms of HorRat value, they are not too bad, all are less than 2.0. • System suitability: Recommendation: The MLT results of this method showed excellent accuracy, precision and reproducibility results except that for Cu, Mn, and Fe at low concentration levels the RSD R values do not meet the SMPR requirements. I would reluctant to recommend this method as the final method against the existing SMPR. On the other hand, based on the MLT results, I am confident that this method will allow accurate and precise determination of these nine elements at the majority of the SMPR analytical ranges. Only for the two or three elements (Copper, Manganese, and Iron) at near their SMPR required LOQ levers, they show large determination uncertainties, or RSD. I think this method is worth to be recommended for slightly modified analytical ranges.

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method _2011.14__________________

Title: MLT for 9 Minerals by ICP

Author: Nestle/Freisland Campina

Summary of Method:

Samples are digested by microwave under acidic conditions. Then analyzed by ICP-AES

Method Scope/Applicability:

Method is applicable to SPFIAN

General comments about the method:

Unnecessary to include Dairy products.

Method Clarity:

Method Safety Concerns:

Pros/Strengths: • Use of several labs with multiple vendors used for both ICP and microwave

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Cons/Weaknesses • High RSDs.

Supporting Data • General Comment:

- Method Optimization:

• Performance Characteristics:

Analytical Range:

LOQ:

Accuracy/Recovery:

Precision (RSD r ):

Reproducibility (RSD R ): Many of the elements fail to meet SMPR requirements in several products. Calcium had >10% RSD on several products that would be high level

• System suitability:

Recommendation: Method does not go to final action based on data generated.

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Expert Review Panel for Infant Formula and Adult Nutrition

Evaluation of Method 2016.02 MLT

Text in RED : SKC review for First Action, 3/2016. Text in BLUE : SKC review for Final Action, 3/2017.

Title: DETERMINATION OF BIOTIN BY LIQUID CHROMATOGRAPHY COUPLED WITH IMMUNOAFFINITY COLUMN CLEAN-UP EXTRACTION Author: George Joseph, Ranjani Devi and Raj Naganaboyina AsureQuality Ltd, PO Box 41, Shortland Street, Auckland 1140, New Zealand

Summary of Method: A portion of sample is dispersed in phosphate buffered saline(pH 7.4) and autoclaved at 121±2⁰C for 25 minutes, cooled to room temperature and diluted to 100mL. The extract is clarified (centrifuged, filtered) and a portion of the resulting solution loaded to a Biotin immunoaffinity column mounted in an SPE manifold. The SPE column is washed w PBS then biotin is eluted with methanol, which is evaporated, and the residue re-constituted in water. A portion of the resulting sample preparation is then injected onto a reversed-phase HPLC system using an isocratic separation (0.1% phosphoric acid / acetonitrile) with a column wash performed after elution of the peak of interest. Detection is by UV at 200nm. Quantification is by external standard against a multi- point standard curve. The method does not specify whether peak height or peak area is used. Method Scope/Applicability: Updated 3/2017: The SMPR refers to the analysis of “total biotin”. The fortificant used in formulated nutritional products is free biotin. Biotin in foods may be covalently bound to L-lysine, as biocytin, or to proteins. The method used for the SLV addressed only free biotin. The method used for the MLT was updated to include quantification of biocytin: a separate standard is prepared and its response is monitored. The conditions of the separation are unchanged. Biocytin elutes early in the run and is positioned atop a baseline disturbance. Method Clarity: Sample Preparation, Table 1, specifies the mass of product to process based on the expected biotin content in “ug/100g”; the values appear to be “as fed” or “as reconstituted” (rather than “as powder e.g.”); the basis of the expression should be stated. Updated 3/2017: The 3/2016 review of the Bio-02 method listed several points requiring clarification. These have been largely addressed in the method which accompanied the MLT: General comments about the method:

1

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

The method does not include: • Example calculations

o Height or area? Calculations are based on sample area (MLT Step 3.3.12, page 8). o How the standard curve data is treated to create a calibration? Injection sequence is specified (MLT Step 3.3.11, page 8). o Sample calculation. Sample calculation added (MLT Step 3.3.12, page 8) • System suitability o What sequence of injections should be used for calibrants and unknowns; o Performance criteria for standard curve (r) and read-backs. o These points are addressed in MLT Step 3.3.11, page 8. The Study Director noted that a clarification was made to the original method, to one step of the sample preparation, concerning elution from the immunoaffinity column: • Bio-02 (3/2016), page 3 Step 13: Introduce a reacti-vial and elute the analyte under gravity with 2mL methanol. Elute further with additional 1mL of methanol. Back flush at least 3 times when eluting. • 2016.02-MLT (3/2017), page 7 step 12, italics added: Introduce a reacti-vial and elute the analyte under gravity with 2mL methanol. Elute further with additional 1mL of methanol. Back flush at least 3 times when eluting and this can be achieved by gentle up and down motion of the syringe plunger to maximize the elution . Pros/Strengths: • The sample preparation is straightforward and uses conventional, widely practiced techniques • The quantification uses conventional HPLC w UV detection. • The validation report documented similar analytical results for the majority of samples from a second brand of immunoaffinity column; most samples produced results within ± 5%. The milk based IF powder result was 10% lower - 3 mcg/100g powder, or c 0.4 mcg/100mL (4 mcg/L) as fed; although the absolute difference is small, further investigation is warranted because of the commonness of this matrix. (Table 5) Cons/Weaknesses • Text needs to be added to describe system suitability and calculations. Done, see above. • The values in the SLV are not consistently reported in the SMPR units (mcg/100g reconstituted) The MLT did not use the units of the SMPR either. Method Safety Concerns: None

Supporting Data • General Comment:

The validation report stated that while a reagent blank and a matrix blank prepared from infant elemental powder were devoid of interferences, the placebo (non-fortified) products child formula powder, adult nutritional RTF high fat, and infant formula RTF milk based showed a response near the retention time of

2

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

biotin. The report does not show these chromatograms nor clarify if the “response” was attributed to (endogenous) biotin or was an interference. - Method Optimization: • Performance Characteristics: Analytical Range: • The method describes the processing of samples with biotin content over the entire concentration range stated in the SMPR (0.1 – 150 mcg/100g as fed) • In practice, the concentration encountered in the SLV using the SPIFAN sample set was ~1.5 – 65 reconstituted. LOQ: • The SMPR LoQ (0.1 mcg/100g reconstituted) is achieved at the lowest standard curve concentration using the typical sample weight specified in the method for this concentration of analyte.

Accuracy/Recovery: SMPR:

0.1 – 1 mcg/100g reconstituted: 80 - 120% > 1 mcg/100g reconstituted: 90 – 110%

• Spike and recovery: Four placebo samples (Child Formula, Infant Elemental, Adult RTF High Fat, IF Milk RTF) were spiked at +50% and +150% of typical target, 3 days in duplicate, with recoveries meeting the SMPR requirements ranging from 95.2% – 100%. One fortified sample (Adult RTF High Protein) was spiked similarly and showed recovery of 96.7% - 104%, also meeting the SMPR requirement. (Table 3) • SRM 1849a: the sample was analyzed on 6 different days; the results (range 196.5 – 200.0 mcg/100g as is) were all well within the Certified Mass Fraction range (199 ± 13 mcg/100g as is). (Table 2)

Precision (RSD r ): SMPR:

0.2 – 1 mcg/100g reconstituted: ≤ 8% > 1 mcg/100g reconstituted: ≤ 6%

Reproducibility (RSD R ): SMPR:

0.1 – 1 mcg/100g reconstituted: ≤ 16% > 1 mcg/100g reconstituted: ≤ 12%

RSD IR was characterized by duplicate analysis of the 12 fortified SPIFAN samples across 6 days (2 analysts, 2 instruments). All values were below 3.2%. Although they were not calculated individually, none of the within day precision values appeared to be greater than this value. The method could likely meet the SMPR requirements if extended to MLT. (Table 4) 3/2017: The MLT consisted of 12 laboratories from 10 different countries. Three laboratories only reported results for the practice samples. One of the remaining laboratories produced results

3

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

which appeared to have a low bias, likely related to incomplete elution of biotin from the immunoaffinity column, however their results were included in the data set. 14 different product types were analyzed in duplicate over two separate days. Two practice samples were analyzed by all 12 laboratories. The 12 remaining samples were analyzed by 9 laboratories. For biotin, relative to the SMPR, all 14 samples appeared to fall into the higher mass fraction range of > 1mcg/100g reconstituted product. • The values for RSDr ranged from 2.3% to 7.0%. Without excluding any data, 12 of the 14 samples met the SMPR requirement of RSDr ≤ 6%. The Study Director noted that for each of the 2 samples which exceeded this limit, the removal of 1 laboratory’s data set would reduce the RSDr to < 6%. No formal outlier testing was performed, and the data was simply presented as a whole. With respect to accuracy, SRM-1849a was included in the MLT sample set and was analyzed by 9 laboratories on 2 different days: 17 of the 18 values produced fell within the certified mass fraction range for the material (199 ± 13 mcg/100g), and the grand mean was within 1.5% of the certified mass fraction value. With respect to biocytin, the Study Director noted that the compound was not detected by any laboratory in any samples. Since biocytin had not been included in the original SLV, it was unclear whether the compound was present in the sample matrices but not recovered, or simply absent in the samples. Upon inquiry, the Study Director provided SLV data in which 2 SPIFAN matrices were spiked on 1 day, in duplicate, at each of 2 levels, in which average recovery ranged from 90% - 99%, appearing to satisfy the SMPR criteria. • System suitability: - Not stated. Recommendation: Add system suitability and calculations. Although the entire SPIFAN set was not spiked, acceptable for First Action status. 3/2017: To remain faithful to the SMPR requirement for “determination of total biotin” the quantification of biocytin should be added to the existing 2016.02. The updated language concerning sample elution of the immunoaffinity column, calculations, and system suitability should all be added to the method: Acceptable for Final Action status. • The values for RSD R ranged from 3.0% to 9.5%. Without excluding any data, all 14 samples met the SMPR requirement of RSD R ≤ 12%.

From: George Joseph [mailto:George.Joseph@asurequality.com] Sent: Tuesday, March 07, 2017 1:50 PM To: Scott Christiansen Subject: RE: Biotin / Biocytin

4

FOR EXPERT REVIEW PANEL USE ONLY NOT FOR DISTRIBUTION

Thanks Scott, We have extended the SLV to carry out recovery studies with biocytin. The recoveries confirm the affinity of IAC for biocytin. The data from SLV is given below.

Recovery (%)

Placebo

Spike Levels

Day 1

Day 2

Day 3

Average

Biotin 50% Biotin 150%

99.18 99.40 90.20 99.42 94.91 97.10 97.98 95.41 97.15 94.87 90.1 92.4

99.66 98.89 98.40

101.00

100.95

99.39 99.97

99.72 96.65

99.98 98.71 94.30 90.30 99.63 94.81 98.65 91.50 98.23 95.40 97.32 95.20

99.28

98.05

Child Formula Powder

Biocytin 50% Biocytin 150%

90.5

Biotin 50% Biotin 150%

99.94 94.38

100.44

98.90 95.33

99.91 94.47

99.14 94.47

95.28

Infant Elemental Powder

Biocytin 50% Biocytin 150%

100.20

90.6

Biotin 50% Biotin 150% Biotin 50% Biotin 150%

98.56 95.25 97.73 95.13

98.17 95.11 96.51 95.51

97.58 94.61 97.50 95.86

99.21 97.35 97.18 95.06

97.85 94.66 97.85 94.77

Adult Nutritional RTF, High Fat

Infant Formula RTF, Milk Based *Infant Formula RTF, Milk Based

Biotin 50%

97.49

98.07

93.52

94.50

97.51

98.19

96.55

Sample

Spike Levels

Day 1

Day 2

Day 3

Average

Biotin 50%

103.00

103.50

103.98

105.21

103.42

103.01

103.69

**Adult Nutritional RTF High Protein

Biotin 100%

96.25

96.95

97.65

97.04

97.24

95.24

96.73

Kind Regards George

From: Scott Christiansen [ mailto:Scott.Christiansen@perrigo.com ] Sent: Wednesday, 8 March 2017 5:13 a.m. To: George Joseph < George.Joseph@asurequality.com > Subject: Biotin / Biocytin Hello George,

I read with interest the summary of the biotin MLT. The results for biotin were outstanding . I was interested to see that no biocytin was detected in any of the samples. Did you perform any experiments which would confirm biocytin would be quantitatively trapped on the immunoaffinity column / eluted from it under the method conditions? The open question would be whether the lysine side chain would interfere with the process, i.e. are the negative findings in the samples from the inability to adhere to the column, or be eluted from it? I trust we will see you next week. Tx, /Scott

Scott Christiansen | Senior Manager Perrigo Nutritionals | Quality Control Direct: 802.528.8604 (x358604) | Perrigo Cell: 802.309.3047 scott.christiansen@perrigo.com www.perrigo.com | Quality Affordable Healthcare Products

5