Journal of Pharmaceutical and Biomedical Analysis20 (1999) 807–814

Determination of the nutraceutical, glucosaminehydrochloride, in raw materials, dosage forms and plasma

using pre-column derivatization with ultraviolet HPLC

Zhongming Liang, James Leslie, Abimbola Adebowale, Mohammed Ashraf,Natalie D. Eddington *

Pharmacokinetics-Biopharmaceutics Laboratory, Department of Pharmaceutical Sciences, School of Pharmacy,Uni6ersity of Maryland, 100 Penn Street, AHB 540C, Baltimore, MD 21201, USA

Received 30 November 1998; received in revised form 12 April 1999; accepted 5 May 1999

Abstract

A selective and specific high performance liquid chromatography method was developed to quantitate glucosaminehydrochloride in raw materials, dosage forms and plasma. Reverse phase chromatography using pre-columnderivatization with phenylisothiocyanate, and ultraviolet detection (l=254 nm) was used to quantify the eluate. Themobile phase consisted of MeOH/H2O/CH3COOH (10:89.6:0.04) and was pumped at a flow rate of 1.2 ml/min. Thestandard curves for glucosamine hydrochloride showed linearity (r]0.99) over the selected concentration range from6.65 to 16.63 mg/ml for raw materials and dosage forms. The precision of the dosage form assay, expressed as the %relative standard deviation (R.S.D.), was B5% at all concentrations. The intra-day and inter-day accuracy, asindicated by the relative error (R.E.), ranged from −2.54 to 2.70% for glucosamine hydrochloride. For the plasmaassay, beagle dog plasma was used to prepare standard curves in the concentration range of 1.25–20 mg/ml.Precipitation of plasma proteins was accomplished with acetonitrile to separate interfering endogenous products fromthe compound of interest. The supernatant was derivatized using phenylisocyanate in phosphate buffer (pH=8.3)and subsequently evaporated to dryness under a nitrogen stream at 42°C. The residue was dissolved in 250 ml mobilephase and injected onto the chromatographic system. The assay was linear in concentration ranges of 1.25–20 mg/ml(r]0.999). Intra- and inter-day precision was 55.23 and 5.65%, respectively and the intra- and inter-day accuracy,indicated by R.E., ranged from −8.6 to 10.35%. The method was found to be specific and with excellent linearity,accuracy and precision and is well suited for the quantitation of glucosamine hydrochloride in raw materials, dosageforms, and pharmacokinetic studies. © 1999 Elsevier Science B.V. All rights reserved.

Keywords: Glucosamine hydrochloride; Nutraceutical; Phenylisothiocyanate; Arthritis

www.elsevier.com/locate/jpba

* Corresponding author. Tel.: +1-410-706-6710; fax: +1-410-706-6580.E-mail address: eddingto@pharmacy.ab.umd.edu (N.D. Eddington)

0731-7085/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved.

PII: S 0 7 3 1 - 7 0 8 5 ( 9 9 ) 0 0 0 8 5 - 0

Z. Liang et al. / J. Pharm. Biomed. Anal. 20 (1999) 807–814808

1. Introduction

Glucosamine hydrochloride is currently usedfor the treatment of degenerative joint disease insmall animals and represents a novel approach forthe treatment of degenerative conditions. Glu-cosamine, an amino monosaccharide, is a naturalcomponent of glycoproteins found in connectivetissues and gastrointestinal mucosal membranes.It is a precursor of the disaccharide unit of gly-cosaminoglycans (GAGs) which are the building-blocks of the articular cartilage, the proteoglycans[1–5]. Numerous reports have recently focused onthe utility of glucosamine for the treatment ofosteoarthritis in dogs [6,7] and horses [8,9].

Glucosamine is considered a dietary supple-ment by the Food and Drug Administration andas such the content uniformity as well as thebioavailability of glucosamine products are rarelydetermined. There is a need to accurately deter-mine the content of products containing glu-cosamine, especially since numerous products aremarketed for the treatment of osteoarthritis. Pre-vious analytical methods to determine glu-cosamine bioavailability have relied onradiolabeling the compound [10,11]. Radioactivityhas been used to quantitate the concentration ofglucosamine and metabolites in pharmacokineticstudies as well as for determining content in glu-cosamine products. Using radioactivity to quanti-tate glucosamine may potentially confound resultssince the labeled compound and any degradationproducts will be quantified. Unfortunately, thisprecludes determining the stability of glucosamineproducts under various experimental conditionsor using radioactivity to accurately determinebioavailability.

Other methods to quantify glucosamine havebeen reported for compositional analysis of aminosugars following pre-column derivatization as wellas quantifying glucosamine in certain bacterialcell walls [12–14]. The purpose of this work wasto develop a precise and accurate HPLC methodfor glucosamine that could be routinely applied tothe quantitation of this agent in raw materials andto evaluate the content uniformity of productscontaining glucosamine. In addition, to determinethe bioavailability properties of glucosamine hy-

drochloride, the second objective was to developand validate a selective, specific and reproducibleHPLC method for the quantitation of intact glu-cosamine hydrochloride in plasma.

2. Experimental

2.1. Materials and reagents

D(+ ) Glucosamine (2-amino-2-deoxy-D-glu-cose) hydrochloride (Fig. 1) was purchased fromSigma (St Louis, MO). Cosequin®/Cosamin®,Cosequin DS®/Cosamin DS® was donated by Nu-tramax Laboratories (Baltimore, MD). The cap-sule products contained 250 or 500 mg (DS) ofglucosamine hydrochloride, and chondroitin sul-fate and manganese ascorbate. Raw materials ofglucosamine were purchased or distributed byZhejiang (lotc 961013, lotc 961014, China) andMarcor Development (lotc MD-G60426, and lotI-20; China). Methanol, acetonitrile, phenylisoth-iocyanate (PITC), sodium phosphate and aceticacid were purchased from J.T. Baker (Phillips-burg, NJ). All chemicals and solvents were ACSanalytical grade or HPLC grade. Deionized waterwas prepared by an ultrapure water system Py-rosystem Plus® (Hydro, Research Triangle Park,NC).

2.2. Standard and sample preparation for rawmaterials and dosage form

Glucosamine hydrochloride standards were pre-pared in water in the concentration range from6.65 to 16.6 mg/ml. A stock solution was pre-pared at a concentration of 1.0 mg/ml and serialdilutions were made to give concentration of stan-dards of 6.65, 10.7, 12.0, 13.3, 14.6, and 16.6

Fig. 1. Chemical structure of glucosamine.

Z. Liang et al. / J. Pharm. Biomed. Anal. 20 (1999) 807–814 809

mg/ml. To each standard, 250 ml of a 0.3 Mphosphate buffer (pH 8.0) and 250 ml of 5% PITCwere added. Samples were vortexed, allowed toreact at room temperature for 10 min and wereevaporated to dryness at about 50°C under nitro-gen. After the derivatization was complete, 200 mlof mobile phase was added and 20 ml was injectedonto the equilibrated HPLC system.

2.3. Preparation of standards and plasmaextraction procedure

Standards were prepared in concentrationsranging from 1.24 to 20 mg/ml in beagle dogplasma from a stock solution of glucosamine hy-drochloride (100 mg/ml). The internal standard,galactosamine, was prepared at a concentration of200 mg/ml. Twenty microliters of the internal stan-dard (galactosamine) was added to 0.40 mlplasma. Samples were vortexed for 10 s and 500ml of acetonitrile was added. The samples wereagain vortexed and centrifuged (3000×g) for 10min. The resulting supernatant was transferred toa 3.0-ml disposable glass tube and 250 ml of 0.3 MNa2HPO4 (pH 8.3) and 250 ml of 5% PITC wereadded. Samples were vortexed, allowed to react atroom temperature for 10 min and were evapo-rated to dryness at 50°C under nitrogen. After thederivatization was complete, 200 ml of mobilephase was added and 20 ml was injected onto theequilibrated HPLC system.

2.4. Instrumentation

The samples were analyzed by an HPLC systemconsisting of two Waters 6000A solvent deliverysystems, a Waters WISP 710B automatic injector,and a Waters 486 Ultraviolet Detector (l=254nm) (Waters Associates, Milford, MA). The de-tector was coupled with an HP 3384 Integratorplotter (Hewlett-Packard, Avondale, PA) and aC18 analytical column (Bondclone (30 cm×3.9mm i.d., 5 mm, Phenomenex, MacclesfieldCheshire, UK) was used for separation. Thebuffer component of the derivatized solution wasprepared with deionized water and the pH wasadjusted to 8.0 using concentrated phosphoricacid. The mobile phase consisted of MeOH/H2O/

CH3COOH (10:89.96:0.04) and was filteredthrough a 0.45-mm nylon filter and degassed un-der ultrasound and vacuum for 15 min. The mo-bile phase was delivered at a flow rate of 1.2ml/min and the eluent was evaluated at a wave-length of 254 nm.

2.5. Assay 6alidation

2.5.1. Linearity and rangeLinearity of the dosage form assay was deter-

mined by analysis of three replicates of six con-centrations (6.65, 10.65, 11.98, 13.31, 14.64 and16.63 mg/ml) by least squares regression. The lin-ear detector response for the plasma assay wastested as follows. Three determinations (n=3)from a minimum of five concentration levels(1.25, 2.5, 5.0, 10.0 and 20 mg/ml) of the analyteconcentration were evaluated by least-squares re-gression. A weight of 1/y was used to determineslopes, intercepts and correlation coefficients. Theminimum acceptable coefficient to establish lin-earity was set at 0.95 a priori.

2.5.2. Method precision and percent reco6eryPrecision of the assay was established by analy-

sis of three replicates (n=3) of a standard solu-tion of the analyte at the followingconcentrations: 6.65, 10.65, 11.98, 13.31, 14.64,16.63 mg/ml for the dosage form assay and 1.25,2.5, 5.0, 10.0, 20 mg/ml for the plasma assay. Todetermine intra-day precision of the assay, repli-cate (n=3) blood samples of five different con-centrations were analyzed. To determine inter-dayprecision, replicate blood samples (n=3) wereanalyzed on three different days. The percentrelative standard deviations of the assay resultswere determined.

Extraction efficiency was determined by com-paring replicate (n=9) peak height ratios of ex-tracted plasma samples vs. unextracted waterstandards for the 2.5, 5, 10 and 20 mg/ml concen-trations. The percent recovery was determined bythe following equation:

% Recovery

=Peak height ratio (plasma standard)Peak height ratio (water standard)

Z. Liang et al. / J. Pharm. Biomed. Anal. 20 (1999) 807–814810

2.5.3. AccuracyAccuracy of the method was determined by

calculating the mean concentration of three repli-cate (n=3) solutions at concentrations rangingfrom 6.65 to 16.63 mg/ml for the dosage formmethod and 1.25–20 mg/ml for plasma. Methodstandards in the concentration range of 2.5–250mg/ml from three different runs performed overseveral days were used to check for accuracy. Themeans of the three runs were calculated and com-pared to the spiked value to determine the per-centage difference between the mean and thespiked value (amount added). The percentage rel-ative error was determined as:

% Relative error (R.E.)=[mean−spiked]

[spiked]×100

Between and within run accuracy were alsodetermined.

2.5.4. Specificity, stability and robustnessSpecificity was evaluated by testing samples to

ensure that no interference from components ofthe mobile phase, degradation products, excipi-ents or beagle dog plasma were obtained in thepresence of the analyte of interest. Stability of theanalyte in the dosage form standards and beagledog plasma at room temperature and at −85°Cas well as two freeze thaw cycles were determined.This was performed by analyzing the standardsolutions in triplicate at different times after stor-age. Robustness was assessed by determining theretention times of glucosamine hydrochloride us-ing two different lots of the same analyticalcolumn and two different batches of mobilephase.

2.6. Analysis of raw materials

Various lots and suppliers of glucosamine hy-drochloride raw materials were evaluated. Onehundred milligrams of the glucosamine hy-drochloride raw material samples were weighed,placed in a 100-ml volumetric flask and an appro-priate volume of water was added. The solutionwas sonicated for 20 min and an aliquot of thissolution was transferred to a test tube and dilutedto a suitable concentration within the standard

concentration range. To this aliquot 250 ml ofphosphate buffer (0.3 M, pH 8.0) and 5% PITCwere added. The analysis was continued as de-scribed above for the derivatization ofglucosamine

2.7. Analysis of capsules

Twenty capsules containing 250 mg of glu-cosamine each were weighed. The contents ofeach capsule were transferred quantitatively to a250-ml volumetric flask and about 150 ml ofwater added. The mixture was sonicated for 20min and brought to volume with water. A portionof each solution was centrifuged (3000×g for 5min) and 20 ml derivatized with PITC. Simulta-neously, a standard solution of glucosamine hy-drochloride was treated in a similar manner. Theresulting solutions were chromatographed induplicate.

2.8. Pharmacokinetic study

A study was designed to investigate the phar-macokinetics of glucosamine hydrochloride inbeagle dogs. The protocol was approved by theAnimal Research Committee at the University ofMaryland, School of Pharmacy. A quantity of2000 mg of glucosamine hydrochloride (CosequinDS/Cosamin DS, four capsules) was orally ad-ministered to a male dog (15 kg). Serial bloodsamples (2.0 ml) were collected via the jugularvein over a 24-h period into EDTA tubes. Plasmasamples were separated by centrifugation (3000×g) for 10 min and immediately frozen at −85°C.

3. Results

3.1. Resolution

Fig. 2a–c represents chromatograms of ex-tracted blank, calibration standard containing6.65 mg/ml of glucosamine hydrochloride and araw material sample containing 6.25 mg/ml ofglucosamine hydrochloride, respectively. The as-say was found to be specific for glucosaminehydrochloride and no interfering peaks from

Z. Liang et al. / J. Pharm. Biomed. Anal. 20 (1999) 807–814 811

Fig. 2. Chromatograms of (a) extracted blank sample, (b) glucosamine standard, and (c) glucosamine sample (raw material).

chondroitin, manganese ascorbate or excipientswere found. Fig. 3a and Fig. 3b represent chro-matograms of extracted blank beagle dog plasmaand a calibration standard containing 5 mg/ml ofglucosamine hydrochloride, respectively. The as-say was found to be specific for glucosaminehydrochloride and no interfering peaks fromdegradation products, EDTA, internal standardor plasma were detected. Moreover, the peaks ofglucosamine hydrochloride and the internal stan-dard were sufficiently separated with typical reten-tion times of 14.4 min for glucosaminehydrochloride and 12.2 min for the internalstandard.

3.2. Validation assay precision

The concentrations of the standards were calcu-lated using the observed peak areas and thestraight line parameters and all were within 5% ofthe nominal value. These observations and thecorrelation coefficients of greater than 0.999showed a satisfactory fit to the linear model. Thestandard curves for glucosamine hydrochlorideshowed linearity over the selected concentration

range from 6.65 to 16.63 mg/ml for raw materialand dosage form with consistent slopes and excel-lent correlation coefficients (r]0.99) throughoutthe validation runs. The dosage form intra-dayand inter-day precision data for glucosamine hy-drochloride are listed in Table 1. The intra-dayand inter-day % R.S.D. was 4.18% or less for allstandards. The precision of the assay, expressed asthe % R.S.D., was B5% at all concentrations.Table 2 summarizes the dosage form intra-dayand inter-day accuracy data for glucosamine hy-drochloride. The intra-day and inter-day accu-racy, as indicated by R.E., ranged from −2.54 to2.70% for glucosamine hydrochloride.

The standard curves for glucosamine hy-drochloride showed linearity over the selectedconcentration range from 1.25 to 20 mg/ml inplasma with consistent slopes and excellent corre-lation coefficients (r]0.99) throughout the vali-dation runs. The intra-day and inter-day precisiondata for glucosamine hydrochloride in beagle dogplasma are listed in Table 3. The intra-day andinter-day % R.S.D. ranged from 0.65 to 5.26%and 1.28 to 5.64%, respectively. Table 4 summa-rizes the intra-day and inter-day accuracy data for

Z. Liang et al. / J. Pharm. Biomed. Anal. 20 (1999) 807–814812

Fig. 3. Typical chromatograms of glucosamine hydrochloridein plasma: (A) extracted blank plasma; (B) glucosamine hy-drochloride standard, 13.3 mg/ml. Peaks 1 and 2 representglucosamine hydrochloride and galactosamine, respectively.

Table 2Intra-day and inter-day assay accuracy for the dosage formassay of glucosamine hydrochloride

Inter-dayConcentration Intra-day(mg/ml) (% R.E.) (% R.E.)

1.6516.63 2.22−0.75−2.5414.64

−1.2513.31 −1.9511.98 −2.170.90

0.851.1610.656.65 2.700.92

3.3. Stability

Stability of glucosamine hydrochloride wasevaluated by comparing the mean peak area ofthe 13.31 mg/ml dosage form standard which hadbeen in storage for 10 days at room temperaturewith the peak area of a freshly prepared standard.The mean peak area of the stored standard waswithin 99% of the freshly prepared standard. Sta-bility of glucosamine hydrochloride in plasma wasevaluated by comparing the mean peak area ofthe 2.5, 5.0 and 10 mg/ml standard and freshlyprepare standard under identical assay conditions.The glucosamine hydrochloride was stable after24 h when stored at room temperature (25°C), 3months storage at −85°C and through two freezethaw cycles from −85°C. The variation due tousing different analytical columns and mobilephases was found to be negligible under the exper-imental chromatographic conditions.

glucosamine hydrochloride in beagle dog plasma.The intra-day and inter-day accuracy, as indicatedby R.E., ranged from −1.52 to 7.44% and −8.6to 10.35%, respectively. After the comparison ofthe PARs for extracted plasma standards andwater standards the extraction recovery wasfound to be 98.68, 100.0, 99.92, 104.12 for the 2.5,5, 10 and 20 mg/ml standards, respectively.

Table 1Intra-day and inter-day assay precision for the dosage formassay of glucosamine hydrochloride

Intra-dayConcentration Inter-day(% R.S.D.)(mg/ml) (% R.S.D.)

16.63 0.992.932.0414.64 0.47

4.18 3.0113.313.1811.98 0.511.7110.65 1.27

1.45 1.386.65

Table 3Intra-day and inter-day assay precision for glucosamine hy-drochloride in beagle dog plasma

Inter-dayConcentration Intra-day(% R.E.)(mg/ml) (% R.E.)

1.25 2.19 2.902.50 5.645.26

3.255.00 3.9610.0 2.13 2.2020.0 0.65 1.28

Z. Liang et al. / J. Pharm. Biomed. Anal. 20 (1999) 807–814 813

Table 4Intra-day and inter-day assay accuracy for glucosamine hy-drochloride in beagle dog plasma

Intra-day (%Concentration Inter-day (% R.E.)R.E.)(mg/ml)

1.25 −8.64.882.50 10.351.10

−1.047.445.0010.0 0.400.08

−0.40−1.5220.0

Table 6Analysis of glucosamine hydrochloride in capsules (Cose-quin®—9602B006, Drum c2)

Capsule Weight Assay (mg) Percent of nomi-nal

243649.11 97.2649.02 255 102

3 666.6 259 103.62594 644.1 103.6

657.55 102.42566 669.1 246 98.47 233647.1 93.28 96.0240644.1

641.29 243 97.210 657.9 244 97.6

Mean 247.88.9S.D.

R.S.D. % 3.6

3.4. Raw material, dosage forms andpharmacokinetic e6aluation

The utility of the derivatized HPLC method forglucosamine was assessed by the quantitation ofglucosamine content in raw materials and dosageforms. Table 5 summarizes the content of glu-cosamine in various batches of raw materials andTable 6 presents the content uniformity of Cose-quin capsules using the validated method. Aplasma concentration vs. time profile following anoral dose of glucosamine hydrochloride (2000 mg)to a representative animal is shown in Fig. 4. Thisprofile highlights the sensitivity of the assay aswell as its utility since it is possible to follow thetime course of glucosamine after a single oraldose.

4. Discussion

Glucosamine hydrochloride does not have achromophore absorbing in the wavelength rangeuseful for liquid chromatography with ultraviolet

detection. Therefore it was decided to make aderivative of glucosamine hydrochloride thatcould be detected by ultraviolet absorption. Basedon the derivatization of amino acids and sugarswith PITC [12–15], we describe the derivatizationof glucosamine hydrochloride in dosage formsand plasma. A high performance liquid chromato-graphic method was developed and validated forthe analysis of glucosamine hydrochloride pow-der, capsules and in plasma. The method involvedderivatization of glucosamine by PITC followedby chromatography on an octadecyl (C18) column

Fig. 4. Plasma concentration vs. time profile from a represen-tative beagle dog following an oral dose of Cosequin DS®/Cosamin DS® (glucosamine hydrochloride, 2000 mg).

Table 5Content of glucosamine hydrochloride as % w/w of the labelamount in various brands of raw materialsa

Raw materialProduct

99.77 (2.8)Zhejiang, lotc 961013 (China)101.96 (1.74)Zhejiang, lotc 961014 (China)

Marcor Development Corp., lotc 99.13 (3.92)MD-60426

Lot I-20 (China) 99.95 (3.35)

a Values are mean with S.D. in parentheses.

Z. Liang et al. / J. Pharm. Biomed. Anal. 20 (1999) 807–814814

with ultraviolet detection at 254 nm. Glucosaminehydrochloride, under the optimized conditions,was eluted from the reverse phase column at aretention time of 14 min, as shown in Fig. 2.Blank solution treated with PITC showed nopeaks at the retention time of the glucosaminederivative. Linearity, precision, specificity and sta-bility of glucosamine and the PITC derivativewere tested and found to be satisfactory for theanalysis of glucosamine hydrochloride.

In initial experiments, galactosamine hy-drochloride was tested as a possible internal stan-dard. There was good resolution of the derivativesof glucosamine and galactosamine, but in subse-quent experiments it was found that there was noadvantage in using an internal standard for thedosage form analysis. Excipients that are likely tobe present in the capsule were tested and foundnot to interfere chromatographically. It is con-cluded that the method described above is suitablefor the analysis of glucosamine in dosage forms.

The assay was found to be specific for glu-cosamine hydrochloride with no interfering peaksfrom beagle dog plasma. The standard curvesdisplayed linearity over the selected concentrationrange (1.25–20 mg/ml). The precision and accu-racy data showed that the assay method is consis-tent and reliable with low values of percentrelative error. No degradation products of glu-cosamine hydrochloride were detected after 24 hat room temperature or after 3 months storage at−85°C. The method was found to be robust withrespect to deliberate variations in method parame-ters. The method can be used to quantitate glu-cosamine hydrochloride in plasma obtained frompharmacokinetic studies with beagle dogs.

5. Conclusions

In summary, a specific, precise, accurate andrugged assay for the quantitation of glucosaminehydrochloride in dosage forms and plasma hasbeen developed and validated. The accuracy andprecision data show that this method is consistent

and reliable with low values of relative error andR.S.D. for standards over the concentration rangeexamined. This chromatographic method offers adirect quantitation of intact glucosamine hy-drochloride with the use of pre-column derivatiza-tion with phenylisothiocyanate. This assay can beused for content uniformity testing of dosageforms containing glucosamine hydrochloride,quality control of raw materials and bioavailabil-ity studies.

Acknowledgements

This work was supported by a grant fromNutramax Laboratories, Incorporated.

References

[1] H.W. Greiling, W. Stuhlsatz, U. Tillmans, in: H.U.Bergmeyer (Ed.), Methods of Enzymatic Analysis, 1984,pp. 60–69.

[2] A. Conte, N. Volpi, L. Palimiera, I. Bahous, G. Ronca,Arzneim.-Forsch. Drug Res. 45 (1995) 918–925.

[3] M.L. Huber, R.L. Bill, Compend. Contin. Educ. Pract.Vet. 16 (1994) 501–504.

[4] R.J. Todhunter, G. Lust, J. Am. Vet. Med. Assoc. 204(1994) 1245–1251.

[5] R.D. Altman, D.D. Dean, O.E. Muniz, Arthritis Rheum.32 (1989) 1300–1306.

[6] P.S. McNamara, S. Barr, H.N. Erb, Am. J. Vet. Res. 57(1996) 1390–1394.

[7] P.S. McNamara, S.A. Johnston, R.J. Todhunter, Vet.Clin. North Am. Small Anim. Pract. 27 (1977) 863–867.

[8] R.R. Hanson, L.R. Smalley, G.K. Huff, S. White, T.A.Hammad, Equine Pract. 19 (1997) 9–11.

[9] R.R. Hanson, Equine Pract. 18 (1996) 18–22.[10] I. Setnikar, C. Glachetti, G. Zanoio, Pharmatherapeutica

3 (1984) 538.[11] R. Setnikar, S. Palumbo, G. Canali, G. Zanolo,

Arzneim.-Forsch. Drug. Res. 43 (1993) 1109–1113.[12] J. Diaz, J.L. Lliberia, L. Comellas, F. BrotoPuig, J.

Chromatogr. 719 (1996) 171–179.[13] S.R. Hagen, J. Chromatogr. 632 (1993) 63–68.[14] K.R. Anumula, P.B. Taylor, Anal. Biochem. 197 (1991)

113–120.[15] B.A. Bindlingmeyer, S.A. Cohen, T.L. Tarvin, J. Chro-

matogr. 336 (1984) 93–104.

.