Single Laboratory Validation of a Method for Determination of Glucosamine in Raw Materials and Dietary Supplements Containing Glucosamine Sulfate and/or Glucosamine Hydrochloride by High-Performance Liquid Chromatography with FMOC-Su Derivatization

Abstract

Single laboratory validation of a method for determination of glucosamine in raw materials and dietary supplements containing glucosamine sulfate and/or glucosamine hydrochloride by with high-performance liquid Chromatography FMOC-Su derivatization. Tests with 2 blank matrixes containing SAMe, vitamin C, citric acid, chondroitin sulfates, methylsulfonylmethane, lemon juice concentrate, and other potential interferents showed the method to be selective and specific. Eight calibration curves prepared over 7 working days indicated excellent reproducibility with the linear range at least over 2.0–150 μg/mL, and determination coefficients >0.9999. Average spike recovery from the blank matrix (n = 8 over 2 days) was 93.5, 99.4, and 100.4% at respective spike levels of 15,100, and 150%, and from the sample matrix containing glucosamine (n = 3) was 99.9 and 102.8% at respective levels of 10 and 40%, with relative standard deviations <0.9%. The method was also applied to 12 various glucosamine finished products and raw materials. The stability tests confirmed that glucosamine–FMOC-Su derivative once formed is stable at room temperature for at least 5 days. Limit of quantitation was 1 μg/mL and limit of detection was 0.3 μg/mL. The method is ready to proceed for the collaborative study.

As a part of the AOAC effort to develop Official Methods^SM for dietary supplements, funded by the U.S. Food and Drug Administration (FDA) and the National Institutes of Health (NIH), an AOAC expert review panel (ERP) on glucosamine reviewed many analytical methods from different sources for the rapid and accurate determination of glucosamine in various dietary supplement products. The panel concluded that the glucosamine high-performance liquid Chromatography (HPLC) method developed at NOW Foods, Inc. was the most appropriate method to recommend for further laboratory validation (Waszkuc et al., in preparation). Glucosamine and chondroitin sulfates have been widely promoted as a treatment for osteoarthritis, and the products as dietary supplements have become increasingly popular in recent years in the United States.

Glucosamine is an amino sugar. Because the molecule structure lacks a UV-absorbing chromaphore, derivatization is necessary for HPLC UV detection. In this method, glucosamine finished products or glucosamine sulfate/hydrochloride raw materials are dissolved in aqueous solution. Triethylamine (TEA) is added to neutralize the H₂SO₄/HCl salts. The glucosamine free base is then derivatized with N-(9-fluorenylmethoxycarbonyloyxy-succinimide (FMOC-Su) at 50°C for 30 min, and analyzed by HPLC with UV detection [Furst, P., Pollack, L., Graser, T.A., Godel, H., & Stehle, P. (1990) J. Chromatogr. 499, 557–569; You, J., Fan, X., Wang, H., Wang, G, Su, J.X., & Zhou, C.L. (1998) J. Liq. Chromatogr. and Rel. Technol. 21, 2103–2115]. Because glucosamine has 2 natural stereoisomers (α and β), and the interconversion of these 2 in aqueous solution is not preventable, 2 peaks are shown in the chromatogram. The sum of the areas of these 2 peaks is used for the quantification of the glucosamine free base.

The method was applied to 12 representative glucosamine brands from different manufacturers in many forms, including tablets, capsules, combination products with chondroitin, SAMe (S-adenosyl-L-methionine) and MSM (methylsulfonylmethane), liquid drink or drink mix, chews, and raw materials from shellfish and non-shellfish sources.

This report only covers the method validation work. The detail method development will be published separately.

METHOD

Apparatus

1. LC system.—Agilent HPLC 1100 series with pump, degasser, autosampler, thermostatted column compartment, photodiode array detector, and Chemstation software for system control and data acquisition (Agilent Technologies, Inc., Palo Alto, CA). The LC system was operated under the following conditions: mobile phase flow rate, 0.8 mL/min: detection wavelength, 265 nm; column compartment temperature, 30°C; and injection volume, 10 μL.

2. LC column.—Phenomenex Prodigy (MidBore™) ODS-3 100 Å, 5 μm, 150 × 3.2 mm, Phenomenex order No. 00F-4097-R0 (Phenomenex, Torrance, CA).

3. LC guard column.—Phenomenex Prodigy SecurityGuard™ Cartridges ODS-3 100 Å, 4 × 3.0 mm, Phenomenex order No. AJO-4287.

4. Analytical balance.—Ohaus AS60, accurate to ±0.0001 g (Ohaus, Florham Park, NJ).

5. Sonicator.—Branson 8210 ultrasonic cleaner (Branson Ultrasonic Corp., Danbury, CT).

6. Vortex.—Type 16700 mixer (Barnstead International, Dubuque, IA).

7. pH Meter.—Beckman Φ40 (Beckman Instruments, Inc., Irvine, CA).

8. Grinder.—One-Touch Coffee Grinder (General Electric Co., Fairfield, CT).

9. Volumetric flasks.—5 and 100 mL, class A (Fisher Scientific, Pittsburgh, PA).

10. LC solvent filters.—0.45 μm nylon membrane (Gelman Sciences, Ann Arbor, MI).

11. Syringe filters.—PTFE, 0.45 μm × 13 mm (Restek Corp., Bellefonte, PA) and 0.45 μm × 25 mm (Fisher Scientific).

12. Syringe.—Luer-Lok™, 3 mL, from Becton Dickinson Co. through VWR International (South Plainfield, NJ).

13. Eppendorf pipets and tips.—Repeater^® Plus Pipettor and tips (various sizes), available from Fisher Scientific.

14. HPLC injection vials.—Screw-cap vials with Teflon-coated caps (Agilent Technologies).

Reagents

1. Reference standard.—D(+)-glucosamine HCl, minimum 99% pure, available from Sigma (St. Louis, MO; Product No. G4875).

2. Derivatization reagent.—N-(9-fluorenylmethoxy-carbonyloxy)succinimide (FMOC-Su), 97% pure, available from Lancaster (Windham, NH; Cat. No. 6908).

3. Solvents.—Acetonitrile, HPLC grade; trifluoroacetic acid (TFA), minimum 99.0% pure; TEA, minimum 99% pure: water, HPLC grade; all from Fisher Scientific.

4. FMOC-Su derivatization solution (15mM).—Dissolve 50 mg FMOC-Su in 10 mL acetonitrile.

5. Mobile phases.—Mobile phase A.—Water containing 0.05% TFA, pH 2.4. Mobile phase B.—Pure acetonitrile.

Standard and Unknown Sample Preparation for Derivatization

(a) Sampling

Accurately weigh (solid samples) or measure (liquid samples) the amount, as indicated below, into different 100 mL volumetric flasks: glucosamine HCl standard (for standard stock solution), 240 mg; glucosamine HCl contained in raw materials or products: the amount of ground powder or liquid containing about 240 mg glucosamine HCl (refer to product label claim); glucosamine sulfate contained in raw materials or products: the amount of ground powder or liquid containing about 340 mg glucosamine sulfate (refer to product label claim). Note: For tablets and/or capsules, find and record the mean (fill) weight of 20 tablets/capsules; then grind/mix and weigh. For liquid products, shake well before taking the sample.

(b) Dissolving

Add 90 mL water, mix on a Vortex mixer for 1 min, and sonicate for 5 min or until all samples dissolve. (Note: Some of the products’ excipients, e.g., silicon dioxide, may never dissolve.) Pipet 750 μL TEA to each of the flasks to neutralize HCl or H₂SO₄, and dilute to volume with water. Filter about 1.5 mL of each solution through a 0.45 μm × 25 mm PTFE filter into an HPLC injection vial.

Derivatization Procedures

Note: Both standard and unknown samples must be derivatized simultaneously in 5 mL volumetric flasks.

(a) Sample transfer

Transfer exact amount, as specified below, of the above-filtered solutions from HPLC injection vials into different 5 mL volumetric flasks: glucosamine standard working solutions (3-point calibration): 50,100, and 150 μL, respectively; then add 150, 300, and 450 μL 15mM FMOC-Su solution in the same order into the flasks, respectively; for all other (unknown) samples: 100 μL, then add 300 μL 15mM FMOC-Su solution to each flask.

(b) Derivatization

Cap the flasks tightly with Teflon stoppers, mix well with Vortex mixer, and sonicate all flasks in sonicator water bath at 50°C for 30 min.

(c) Derivatization quench

Remove flasks from bath, let cool to room temperature, and dilute to volume with mixture of mobile phases A/B (1/1, v/v). Mix well with Vortex mixer. Filter each sample through 0.45 μm × 13 mm PTFE into HPLC vial for injection.

Determination

(a) System suitability tests

Equilibrate HPLC system with mobile phase for at least 30 min. Make 5 replicate injections of second (mid-concentration) glucosamine HCl working standard. The relative standard deviation (RSD) of the sums of peak area of glucosamine peaks 1 and 2 from the 5 injections should be <2.0%.

(b) Mobile phase gradient program

Elute samples with the following gradient mode of mobile phases A and B: 0.0–6.0 min, held isocratic at 70A:30B; 6.0–11.0 min, change toOA:100B; 11.0–13.0 min, to 70A:30B; and 13.0–15.0 min, held isocratic at 70A:30B.

(c) Run time

15 min.

(d) Sample injection

Make single injection of each standard working solution and unknown solution.

Calculations

(a) Expected retention times

Glucosamine–FMOC derivative peak 1: ca 4.9 min, and glucosamine–FMOC derivative peak 2: ca 6.1 min.

(b) Concentrations of glucosamine working standard solutions

Calculate concentrations of glucosamine free base in working standard solutions after FMOC derivatization

$$\text{STD}{\text{n}}_{\text{G}·\text{Freebase},}\text{mg}/\text{mL}=0.83091\times \text{d}\times \text{W}\times \text{F}$$

where n = 1, 2, 3 for 3 different standard working solutions: 0.83091 is the conversion factor from glucosamine HCl to glucosamine free base: 179.17/215.63; d = the dilution factor: v/(100 mL × 5 mL), with v = 0.050, 0.100, and 0.150 mL for STD1, STD2, STD3, respectively; W = the amount of glucosamine HCl standard weighed, mg; and F = the purity factor of glucosamine HCl standard used.

(c) Percentage of glucosamine HCl in finished products or raw materials

Calculate % glucosamine HCl in glucosamine HCl finished products or raw materials

$$\%\text{mg}/\text{mg}\text{G}·\text{HCl}=(\text{P}-\text{b})\times 100/(0.83091\times \text{a}\times \text{D}\times \text{W})$$

where P = the sum of peak area of glucosamine peaks 1 and 2 of the unknown sample; a = slope of the calibration curve; b = intercept of the calibration curve; D is the dilution factor: 0.100 mL/(100 mL × 5 mL) = 2 × 10⁻⁴/mL; W = the amount of unknown sample weighed, mg.

For liquid samples:

$$\text{mg}/\text{mL}\text{G}·\text{HCl}=(\text{P}-\text{b})/(0.83091\times \text{a}\times \text{D}\times \text{V})$$

where V = the amount of liquid unknown sample used, mL.

(d) Percentage of glucosamine sulfate (2GlucosamineFreeBase·H₂SO₄) in finished products or raw materials

Calculate % glucosamine sulfate in glucosamine sulfate finished products or raw materials:

$$\%\text{mg}/\text{mg}\text{G}·\text{sulfate}=(\text{P}-\text{b})\times 100/(0.78511\times \text{a}\times \text{D}\times \text{W})$$

where 0.78511 is the conversion factor from glucosamine sulfate (2GlucosamineFreeBase·H₂SO₄) to glucosamine free base:(2 × 179.17)/456.418.

For liquid samples:

$$\text{mg}/\text{mL}\text{G}·\text{sulfate}=(\text{P}-\text{b})/(0.78511\times \text{a}\times \text{D}\times \text{V})$$

(e) Amount of glucosamine per unit

Calculate glucosamine sulfate (or HCl) in mg per tablet (or capsule), or glucosamine sulfate (or HCl) in mg per serving volume (mL):

$$\begin{array}{c}\text{mg}\text{G}·\text{Sulfate}(\text{HCl})\text{per}\text{Tablet}(\text{Capsule})=\\ \%\text{mg}/\text{mg}\hspace{0.17em}\text{G}·\text{Sulfate}(\text{HCl})\times \text{mgOneTab}(\text{Cap})/100\end{array}$$

$$\begin{array}{c}\text{mg}\text{G}·\text{Sulfate}(\text{HCl})\text{per}\text{Serving}\text{Volume}=\text{mg}/\text{mL}\\ \text{G}·\text{Sulfate}(\text{HCl})\times \text{mLOneServing}\end{array}$$

where mgOneTab (Cap) = the average weight of one tablet or the average fill weight of one capsule, in mg; and mLOneServing = the serving volume for liquid products, in mL.

Validation Parameters

Selectivity and Specificity Studies

(a) Blank checking

The blank materials shown in Table 1 are representative matrixes used for various glucosamine products in the current market by different manufacturers, and should be effective to test the selectivity and specificity of the glucosamine method. Apply exactly the same procedures as for the glucosamine finished products or raw material to the blank materials, and inject the blank samples along with some glucosamine finished product or raw material samples to ensure no interfering or carryover on glucosamine peaks.

Table 1.

Samples for glucosamine method single laboratory validation

Sample	Form	Serving size	Glucosamine per serving size	Other ingredients and excipients
Finished products
Glucosamine sulfate, 1000 mg	Tablet	1 Tablet	Glucosamine sulfate, 1 g	Microcrystalline cellulose, vegetable cellulose, titanium dioxide, silica, vegetable magnesium stearate, vegetable glycerin
Glucosamine complex	Capsule	2 Capsules	Glucosamine sulfate 200 mg, glucosamine HCl, 600 mg glucosamine, N-acetyl 200 mg	Rice powder, silica, magnesium stearate, gelatin
Move free glucosamine and chondrotin	Drink mix powder	1 ½Tbsp (12.125 g)	Glucosamine HCl, 1.5 g	Chondroitin sulfate 500 mg, hydrolyzed gelatin, 10 g
Glucosamine and chondroitin	Tablets	2 Tablets	Glucosamine sulfate, 1.5 g	Chondroitin sulfates, 1.2 g, cellulose, stearic acid, silica, and magnesium stearate
Glucosamine plus MSM	Tablet	3 Tablets	Glucosamine complex (D-glucosamine hydrochloride, D-glucosamine sulfate, N-acetyl D-glucosamine), 1.5 g	MSM (methylsulfonyl methane) 1.5 g, dextrose, cellulose, copolyvidone, coating (hydroxypropyl methylcellulose, titanium dioxide, polydextrose, triacetin, and polyethylene glycol), stearic acid, magnesium stearate, and potassium chloride
Joint Action SAM-e, 200 mg glucosamine 500 mg	Tablet	2 Tablets	Glucosamine hydrochloride, 1000 mg	SAM-e (S-adenosylmethionine) 400 mg, may contain: cellulose gel, glycerol behenate, methacrylic acid copolymer, sodium starch glycolate, povidone, talc, silica, magnesium stearate, triethyl citrate, polyethylene glycol, glycerol palmitostearate, medium chain triglyceride, artificial colors (iron oxide, titanium dioxide), polysorbate 80, sodium hydroxide, simethicone
Move Free Plus SAM-e	Tablet	3 Tablets	Glucosamine hydrochloride, 1.5 g	S-adenosyl-L-methionine (SAM-e) 400 mg, cellulose, coating (methacrylic acid copolymer, magnesium trisilicate, titanium dioxide, polysorbate 80, propylene glycol), copolyvidone, croscarmellose sodium, and magnesium stearate
Glucosamine Lemonade liquid drink mix	Liquid	1 Pouch (59 mL}	Glucosamine HCl, 1500 mg	Vitamin C 60 mg, high fructose corn syrup, water, lemon juice concentrate, natural flavors, citric acid, sodium benzoate, potassium sorbate
Flexi-Licious	Chew	1 Chew	Glucosamine HCl, 500 mg	Corn syrup, sugar, dry whole milk, water, chocolate liquor, partially hydrogenated soybean oil, whey, mono- and diglycerides, soya lecithin, salt, artificial flavor
Raw materials
Glucosamine HCl	Powder from shellfish	N/A	Min 98.5% glucosamine HCl
Glucosamine sulfate	Powder from shellfish	N/A	Min 71% glucosamine sulfate	KCl 23–25%. Glucosamine sulfate potassium chloride complex empirical formula: 2GlucosamineFreeBase·H2SO4·2KCl
Glucosamine HCl	Powder from non-shellfisha	N/A	Minimum 98.5% glucosamine HCl
Blank matrixes
Combination blank matrix for tablets and capsules	Powder	N/A	No glucosamine contained	Chondroitin sulfates (28.8)b, MSM (6.1), vitamin C (1.3), SAMe (4.8), manganese amino acid chelate (hydrolyzed vegetable proteins) with citric acid in rice flour (1.1), microcrystalline cellulose (1 6.4), vegetable grade stearic acid (1.9), vegetable grade magnesium stearate (0.1), silica (0.2), and calcium carbonate (39.3)
Lemonade juice all natural	Liquid	N/A	No glucosamine contained	Water, high fructose com syrup, lemon juice concentrate, citric acid, potassium citrate, natural flavor

^aAt present, glucosamine sulfate raw material from non-shellfish source is still not available anywhere.

^bNumber shown in parentheses are weight percentages in the mixture.

(b) Peak purity and ID confirmation

Use photo diode array (PDA) detector to collect the UV spectra of 2 glucosamine peaks. Compare spectra from different spots of the same peak to check peak purity. Also compare spectra and retention times of peaks between the standard and unknowns to ensure ID match.

Calibration Curve

Use glucosamine reference standard to prepare a new calibration curve every day for 5 days. Use another glucosamine reference standard (may be from the same reference standard vendor, but at least in a different package) to prepare a control sample to check accuracy and precision of the calibration. The control sample will be injected and analyzed daily with the new calibration curve. The data from the control sample can also provide information about the stability of glucosamine–FMOC-Su derivative. Our previous studies showed that the glucosamine–FMOC-Su derivative once formed is stable without change at room temperature for at least 5 days.

(a) Linearity

The determination coefficient of each calibration curve for glucosamine analysis must be ≥0.999. This data also indicates something about the precision of the calibration curve.

(b) Calibration curve precision and accuracy

RSD of the 5 experimental values of the control sample obtained in 5 days with the independent calibration curves should be ≤3%: and the difference between the average experimental value and true value of the control sample should be within 3%.

Test Material Precision (Repeatability Standard Deviation)

Twelve glucosamine finished product and raw material samples shown in Table 1 were chosen for this test (n = 12). Precision will be determined by the following 2 parts of the test: General test.—Each of 12 samples will be tested once a day for 3 days (m = 3, r = 1, d = 3). Specific test.—Glucosamine sulfate tablets and glucosamine·HCl raw material from shellfish (2 of the most popular sample types) will be tested 12 times for each with 3 times a day for 4 days (m = 12, r = 1, d = 4). See Table 2 for expected precision.

Table 2.

AOAC recommended recovery and precision limits for single laboratory validation

Concentration	Repeatability, %	Recovery, %
100%	1	98–101
10%	1.5	95–102
1%	2	92–105
0.1%	3	90–108
0.01%	4	85–110
10 μg/g, 10ppm	6	80–115
1 μg/g, 1 ppm	8	75–120
10 ng/g, 10 ppb	15	70–125

Test Accuracy

(a) Analysis of known amount of glucosamine reference standard

This is performed in section (b) of Calibration Curve, where the purpose is to check the accuracy of the calibration curve.

(b) Recovery of glucosamine spiked to the blank matrix

In the combination blank matrix for tablets and capsules (containing all ingredients except glucosamine), spike with the amount of glucosamine HCl reference standard as follows: 15% level: 734 mg blank powder with 36 mg glucosamine HCl standard; 100% level: 530 mg blank powder with 240 mg glucosamine HCl standard; and 150% level: 410 mg blank powder with 360 mg glucosamine HCl standard.

Prepare replicate of 4 samples at each level, and repeat experiment on a separate day (total of 8 replicates at each level). The acceptable recovery levels are found in Table 2. Note: The base for the above spike percentage level is 240 mg glucosamine HCl. This amount dissolved in the solvent in a 100 mL flask (see Standard and Unknown Sample Preparation for Derivatization) is the expected concentration designed for this method to analyze various glucosamine samples.

(c) Recovery of glucosamine spiked to the sample matrix containing glucosamine (method of standard addition)

To further confirm the spike recovery, fortify the glucosamine sulfate tablet powder used in the section on Test Material Precision, above, with glucosamine HCl reference standard, at 10 and 40% of the expected concentrations. Prepare samples in triplicate at each level. See Table 2 for expected recovery.

Derivative Stability

The control sample used (see Calibration Curve above) and tested daily for 5 days will indicate the stability of the derivative. The glucosamine–FMOC-Su derivative once formed should be stable without change at room temperature for at least 5 days.

Analytical Range

The testing range of the spike recovery combined with the method linear range will be a good indicator of the method analytical range. To prove the method linear range covers at least 2.0–150 μg/mL, accurately weigh 60 mg glucosamine HCl standard into 25 mL volumetric flask. Add 20 mL water, mix on a Vortex mixer for 1 min, and sonicate 5 min. Pipet 150 μL TEA to the flask. Dilute to volume with water, and mix. Filter about 1.5 mL of the solution through 0.45 pm PTFE filter into HPLC injection vial; then transfer, respectively, the exact 5.0, 50, 100, 150, 250, and 375 μL filtered solution into 6 different 5 mL volumetric flasks (6-point calibration). Add 15, 150, 300, 450, 750, and 1125 μL 15mM FMOC-Su solution in the same order into the flasks. Follow the procedures in the method to finish the derivatization and inject the samples for HPLC analysis.

Repeat the test 2 more times. The linearity of all 3 calibration curves should be ≥0.999.

Limits of Quantitation (LOQ) and Detection (LOD)

(a) LOQ

Using the least concentrated glucosamine sample prepared (see Analytical Range section), sequentially dilute the sample until the signal-to-noise (S/N) ratio of the glucosamine peaks is ca 10. Make 6 injections of the solution at the LOQ, and calculate precision.

(b) LOD

Continue dilutions of the solution used for the determination of LOQ until the S/N is approximately 3.

Validation Samples

Fourteen glucosamine samples were chosen and approved by glucosamine study monitors for this validation work (Table 1). These samples cover all representative forms of finished products, raw materials, and blank samples, as well as different brands and manufacturers.

Statistical Evaluation

Mean (A), standard deviation (SD), and RSD used for the statistical analysis are defined as follows:

$$\text{Mean}=\text{A}=(\sum {\text{A}}_{\text{i}})/\text{N}$$

$$\text{Standard}\text{deviation}=\text{SD}=[\text{(}\sum {{\text{(A}}_{\text{i}}-\text{A})}^2)/(\text{N}-1){]}^{1/2}$$

$$\text{Relative}\text{standard}\text{deviation}=\text{RSD}\%=100\text{SD}/\text{A}$$

where A_i is the individual measurement of the sample, and N is the number of data points needed to achieve ΣA_i.

Results and Discussion

System Suitability

All system suitability tests showed that the HPLC systems used for the validation work, before each analysis, met the system suitability acceptance criteria, i.e., RSD of glucosamine peak area from 5 replicate injections of the working standard solution should be <2.0%.

Selectivity and Specificity Studies

Specificity can be defined as the ability to measure accurately the concentration of an analyte in the presence of all other sample materials [Snyder, L.R., Kirkland, J.J., & Glajch, J.L. (1997) Practical HPLC Method Development, 2nd Ed., John Wiley & Sons, Inc., New York, NY]. If specificity is not assured, method accuracy, precision, and linearity all are seriously compromised. Figure 1 shows chromatograms of glucosamine standard and 2 blank matrixes containing SAMe, vitamin C, citric acid, chondroitin sulfates, MSM, lemon juice concentrate, and other potential interferents, after all the samples were analyzed with the exact same procedures and injected in sequence. It is evident that the baseline of 2 blank matrixes is clean and flat in the region of glucosamine peaks, indicating no interference from other ingredients and excipients, or carryover from the previous injections to glucosamine quantitation under the method conditions.

Figure 1.

Combination chromatograms of glucosamine standard and 2 blank matrixes. Blank 1 is a combination blank matrix for tablets and capsules containing chondroitin sulfates, MSM, vitamin C, SAMe, manganese amino acid chelate (hydrolyzed vegetable proteins) with citric acid in rice flour, microcrystalline cellulose, vegetable grade stearic acid, vegetable grade magnesium stearate, silica, and calcium carbonate. Blank 2 is lemonade juice all natural containing lemon juice concentrate, water, high fructose corn syrup, citric acid, potassium citrate, and natural flavor.

To check the peak purity further, the UV spectra of the apex, up- and down-slope, of each glucosamine peak from a finished product were compared by using the diode array data processing software (total 5 points for each peak: 2 on the bottom of 2 sides, 2 on the middle point of up- and down-slope, and 1 on the apex). The 5 spectra had an excellent match from 200 to 400 nm, again indicating no interference or coelution on both glucosamine peaks.

Stability tests revealed that the glucosamine–FMOC-Su derivative once formed is very stable without change at room temperature for at least 5 days (see detail results in Derivative Stability), ruling out the possibility of glucosamine degradation products that might interfere with the analysis if samples are analyzed within 5 days.

Because modifying the conditions of the original HPLC method can change selectivity so as to resolve previously overlapping peaks, the mobile phase ratios (% of A and B or gradient steepness) were adjusted. The tests did not show coelution and provide another evidence of peak purity.

Calibration Curve

Linearity

Five 3-point calibration curves were prepared over 6 working days. The determination coefficient of each curve is >0.9999, indicating an excellent reproducibility of the calibration curve.

Precision and accuracy

The experimental values of the control sample obtained in 5 days with the independent calibration curves were 242.1 mg glucosamine HCl (first day); 241.0 mg (third day); 242.6 mg (fourth day); 242.1 mg (fifth day). Note: There was no test on the second day because of the instrument break-down. The true value of the control sample was 241.2 mg. Therefore, the percentages of recovery of the above 4 tests were 100.4, 99.92, 100.6, and 100.4, respectively. The average was 241.9 ± 0.7 (SD) with an RSD of 0.3%. The difference between the average experimental value and true value of the control sample was 0.3%. The data showed high precision and accuracy of the calibration curves.

Test Material Precision (Repeatability Standard Deviation)

Twelve glucosamine finished products and raw material samples were tested (n = 12). Ten were tested once a day for 3 days (m = 3, r = 1, d = 3). Glucosamine sulfate tablets and glucosamine HCl raw material from shellfish were specially tested 3 times a day for 4 days (m = 12, r = 1, d = 4). Table 3 shows that all tests met AOAC recommended precision limits for single laboratory validation (Table 2) except for 2 samples. For the drink mix powder, the test results were 99.9 (first day), 94.6 (second day), and 98.3% of the labeled amount (third day), respectively, with the second day’s result slightly off possibly due to uniformity or analyst mistake. However the precision (RSD) of the 3 results is still within 3%, an industry standard. For the product of chew, 6 test results were respectively obtained from different extraction methods: two 77.3% by this method, 77.6% by multiple extraction (4 times), 74.9% by this method combining with chloroform extraction, and 72.2 and 74.0% by this method combining with hexane single and multiple (3 times) extractions. Chloroform and hexane are used to improve recovery because the product contains soy lecithin, although the chew has a water-soluble formula. Based on the above results, we are confident that the method reflects the true amount of glucosamine in the chew. Therefore, it is concluded that the glucosamine method presented here works for all 12 samples. No noticeable interference to the glucosamine peaks was found in each product analysis. All tests were performed without any extraordinary difficulty, and the chromatograms in the region of glucosamine peaks were clean. Figure 2 shows a sample chromatogram of the glucosamine Joint Action tablets containing SAMe.

Table 3.

Test material precision (repeatability standard deviation)

	Labeled amount found, %
Sample	Day 1	Day 2	Day 3	Day 4	Precision RSD, %a
Finished products, with GFB concentrationb
Glucosamine tablet, ca 46%	144.85	144.72	143.2		0.6
Glucosamine capsule, ca 56%	113.2	113.0	111.6		0.8
Glucosamine and chondroitin drink mix powder, ca 10%	99.93	94.6	98.34		2.8
Glucosamine and chondroitin tablet, 26%	106.13	107.2	103.39	108.3	1.4
	106.56	106.2	103.7	103.6
	105.7	106.8	104.76	107.0
Glucosamine Plus MSM tablet, 31%	100.3	100.3	99.23		0.6
Joint Action with SAMe tablet, 32%	99.54	98.6	98.97		0.5
Move Free Plus SAMe, 36%	104.0	101.7	104.22		1.3
Joint Juice Glucosamine Drink Mix, 2.1%	119.6	121.3	120.82		0.7
Flexi-Licious Chew, 5.1%	77.3	77.6	74.9	72.7	2.7
				77.3
				74.0
Raw materials, (with GFB concentrationb (Note: Not glucosamine HCl or sulfate concentration)
Glucosamine HCl from shellfish, 81.8%	99.46	100.2	99.45	98.85	0.5
	100.23	99.8	99.74	98.76
	98.7	99.8	99.58	99.39
Glucosamine sulfate from shellfish, 55%	99.2	98.3	99.3		0.6
Cargill glucosamine HCl from non-shellfish; GFB concentration, 81.8%	98.62	99.1	97.96		0.6

^aRSD is relative standard deviation or repeatability standard deviation.

^bTo evaluate the repeatability of tests against AOAC recommended recovery and precision limits for single laboratory validation shown in Table 2, the concentration of glucosamine free base (GFB) in products was calculated based on the label claim.

Figure 2.

Sample chromatogram of the glucosamine Joint Action tablets containing SAMe.

Test Accuracy

Analysis of known amount of glucosamine reference standard

The 4 tests over 5 days of the control sample (see Calibration Curve, Precision and Accuracy section) showed an average recovery of 100.3 ± 0.3% (SD), indicating in part the high accuracy of the method.

Recovery of glucosamine spiked to the blank matrix

The results of this study are shown in Table 4. The spike recovery at 15% level was slightly lower because of the low concentrations of glucosamine free base in the spiked samples, but still met AOAC recommended recovery and precision limits for single laboratory validation. All other tests at 100 and 150% levels revealed excellent spike recoveries and precisions. The blank matrix used was the combination blank matrix for tablets and capsules listed in Table 1, which contain a complex mixture of ingredients. This study proved not only the high accuracy, but also the robustness of the method.

Table 4.

Recovery of glucosamine spiked to the blank matrix

		Spike recovery, %
Spike level, %	Concentration	Day 1	Day 2	Average spike recovery, %
15	734 mg blank powder with 36 mg glucosamine HCl standard; GFB concentration, 3.9%	92.6	93.3	93.5 (±0.8)a
		95.1	93.3
		94.1	92.8
		93.6	93.5
	HORRAT value	0.65	0.18
100	530 mg blank powder with 240 mg glucosamine HCl standard; GFB concentration, 26%	98.6	99.3	99.4 (±0.5)
		99.0	100.1
		99.8	99.5
		99.2	99.6
	HORRAT value	0.41	0.28
150	410 mg blank powder with 360 mg glucosamine HCl standard; GFB concentration, 39%	101.0	100.9	100.4 (±0.6)
		99.3	100.3
		100.6	100.6
		99.7	100.4
	HORRAT value	0.68	0.23

^aStandard deviation.

Recovery of glucosamine spiked to the sample matrix containing glucosamine (method of standard addition)

Table 5 shows recovery of glucosamine spiked to the tablet powder of glucosamine and chondroitin, one of the products in Table 1. In general, it is difficult to do spike recovery by the method of standard addition with great accuracy and precision because it relies on the accuracy of standard addition, which is usually very low, e.g., 10%. It also relies on the accuracy of all data on the sample matrix to be spiked, such as sample uniformity and predetermination of the potency before spike. In this study, the sample matrix was the tablet powder containing glucosamine sulfate. To improve stability, the current glucosamine sulfate raw materials in the market contain potassium chloride in a ratio as in 2GlucosamineFreeBase·H₂SO₄·2KCl. Challenges in production of the exact salt may make many lots not uniform. However despite these difficulties, the spike recovery results presented here are very good in both reproducibility and accuracy. It is obvious that the method is reliable.

Table 5.

Recovery of glucosamine spiked to the sample matrix (tablet powder) containing glucosamine

Spike level, %	Tablet powder used, mga	GFB from tablet powder used, mg (A)b	GFB spiked, mg (B)	Total GFB expected (A+B)	Total GFB found, mg(C)	Spike recovery, % [C/(A+B)]×100	Average spike recovery, %
10	747.4	207.7	19.9	227.5	227.9	100.2	99.9 (±0.9}c
	751.0	208.7	19.2	227.9	225.2	98.8
	748.8	208.1	20.3	228.3	229.8	100.7
40	750.6	208.6	80.9	289.5	297.1	102.6	102.8 (±0.8)
	752.8	209.2	81.1	290.3	296.4	102.1
	750.2	208.5	80.3	288.8	299.3	103.6

^aGlucosamine and chondroitin tablets listed in Table 1 were used in this study.

^bNumber determined from previous tests (see section on Test Material Precision).

^cStandard deviation.

Derivative Stability

The test results shown in the section Calibration Curve, Precision and Accuracy on the control sample indicated that the glucosamine–FMOC-Su derivative once formed was stable without change at room temperature for at least 5 days.

Analytical Range

The spike recovery test also showed in part the analytical range of the method. As described above, the spike recovery tests were performed at spike levels of 15, 100, and 150% (glucosamine spiked to the blank matrix), and the recoveries were all very good (Table 4). For the analytical range purpose, the 15, 100, and 150% levels correspond to 36, 240, and 360 mg glucosamine HCl in 770 mg tablet/capsule powder, respectively. The 100% level is the expected target measurement level of the method. The additional supporting information is good spike recovery of the control sample made with 240 mg glucosamine HCl standard reference. The prepared three 6-point calibration curves proved that the linear range of the method at least covers 2.0–150 μg/mL. The determination coefficients of all these 3 are >0.9999.

Limits of Quantitation and Detection

The LOQ (S/N = 10:1) and LOD (S/N = 3:1) of the method were 1 and 0.3 μg/mL, respectively. They were determined from the smaller glucosamine peak at a retention time of ca 4.9 min.

Method Validation

The following are points about the method and its validation:

1. The derivatization reaction takes place under alkaline conditions. We found that 750 μL organic base TEA under the method conditions was enough to make the sample solutions alkaline for derivatization. Avoid using too much TEA (e.g., >1000 mL). Always use the same amount of TEA for both standard and unknown samples.

2. The method is easy to use and robust. However, because it is a balanced analytical system, all conditions in the method are connected. Follow the method exactly to obtain accurate results. For example, for standard and unknown sample preparation, it is necessary to use the amounts specified in the method. Too much deviation from the amounts indicated may increase error %.

3. This method determines the amount of glucosamine free base in raw materials and dietary supplements, but the results in glucosamine free base can be converted to the amount of glucosamine HCl or glucosamine sulfate based on the stoichiometry of chemistry, and then to confirm the label claims of raw materials or finished products. In the current market, almost all glucosamine products claim either glucosamine HCl or glucosamine sulfate, and most contain only one of the 2. For a mixture of glucosamine HCl/sulfate, if both forms are individually claimed on the label, the total amount of glucosamine can be confirmed by converting the claims to glucosamine free base and comparing it with the experimental value by this method.

4. We used both Agilent 1100 and Waters 2695 HPLC systems in the method validation. The method works on both systems. However, LOQ and LOD vary with instrument. The LOQ and LOD presented above were from Waters 2695. With Agilent 1100, LOQ and LOD were 0.1 and 0.02 μg/mL, respectively. Although the Agilent in NOW laboratories showed more sensitivity, our Waters gave more reproducible results.

Replies to Some of Reviewers’ Comments

(1) Comment: In the section Derivatization Procedures, different amounts of the derivatization reagent (DR.) was used for 3 levels of the calibration samples (i.e., 150, 300, and 450 μL 15mM FMOC-Su), but 300 μL for the unknown sample preparation. The same amount of DR should be used for both the calibration and unknown samples.

Response: The method will be changed to use the same amount of DR for all the sample preparation in the method for the collaborative study. Note: In this SLV study, the amount of the DR used for all unknown sample preparation (300 μL) was the same amount for the middle point of the calibration solutions. Since the amount of DR used is excessive, it is essentially unnecessary within the analytical range of the method to use different amounts of DR for the other 2 calibration levels.

(2) Comment: An Eppendorf Repeater Plus Pipettor was used in the sample preparation. Can regular Eppendorf pipettors be used?

Response: Although we used an Eppendorf Repeater Plus Pipettor for the study, regular pipettors (Eppendorf or equivalent) should also work. However, be sure that the pipettors are properly calibrated.

Conclusions

The single laboratory validation of this method was successfully performed. The obtained data proves the high accuracy and precision of the method. The method is applicable to the analysis of various glucosamine raw materials and finished products. The method is ready to proceed for the collaborative study.