Stereospecific analysis of flurbiprofen and its major metabolites in plasma and urine by chiral-phase liquid chromatography

Stereospecific Analysis of Flurbiprofen and its Major Metabolites in Plasma and Urine by ChiraI-Phase Liquid Chromatography

2003, 57, 7-18

B. K. Patel 1. / J.Valentova 2 /A .J . Hutt 1

1 Department of Pharmacy, King's College London, Franklin-Wilkins Building, 150 Stamford Street, London SE 19NN, United Kingdom; E-Maih bhavesh.patel@ kcl.ac.u k

2 Faculty of Pharmacy, Comenius University, Odbojdrov 10, 832 32 Bratislava, Slovak Republic

Key Words Column liquid chromatography Enantiospecific chromatography Enantioselective disposition Flurbiprofen enantiomers and metabolites

Summary Direct chiral-phase HPLC methods have been developed for the determination of flurbiprofen and its major metabolites, namely 4'-hydroxyflurbiprofen and 3'-hydroxy-4'-methoxyflurbiprofen, in biological fluids using a derivatized amylose chiral stationary phase (CSP; Chiral- pakAD). Quantification of all three analytes, both free and conjugated, in urine was carried out following liquid-liquid extraction using tandem ultraviolet (UV) and fluorescence detection. Determination of flurbiprofen and the 4'-hydroxy- metabolite in plasma utilized the same CSP but required modification in the mobile phase composition and sole use of fluorescence detection. The urine assay was linear (r> 0.998) bel',,veen 0.05- 10 I~g mL 1, 0.1 -20 I~g mL 1 and 0.01 - 2 I~g mL 1 for the enantiomers of flurbiprofen, 4'-hydroxyflurbiprofen and 3'-hydroxy-4'-methoxyflurbiprofen respectively. The plasma assay was linear (r > 0.997) bel',,veen 0.1 - 6 I~g mL 1 and 0.01 -0.6 I~g mL 1 for the enantiomers of flurbiprofen and 4'-hydroxyflurbiprofen respectively. Both assays, typically yielded within- and bel',,veen-day imprecision and accuracy values less than 10% for the enantiomers of the different analytes. Initial volunteer studies suggest that the disposition of flurbiprofen displays modest enantioselectivity in humans.

Introduction

Flurbiprofen (rac-2-[2-fluoro-4-biphenyl]- propionic acid; Figure 1) is a non-steroi- dal anti-inflammatory drug (NSAID) that is widely used for the treatment of pain and inflammation in rheumatic diseases and other musculoskeletal disorders. [1, 2]. Similarly to the other members of the 2-arylpropionic acid (2-APA) group of NSAIDs the main pharmacological activ-

ity of flurbiprofen, inhibition of arachido- nic acid binding to cyclooxygenase and thus prevention of the formation of proin- flammatory prostaglandins, resides prin- cipally in the enantiomer of the S-configuration [3]. The R-enantiomer of flurbiprofen, however, is nearly as effective as (S)-flurbiprofen as an antinociceptive agent [4] and all preparations to date are marketed as the racemate.

Flurbiprofen, similarly to other 2- APAs, undergoes metabolic chiral inversion from the R- to the S-enantiomer which shows species variability [5] and does not appear to take place in humans [6]. In humans, flurbiprofen is eliminated predominantly through metabolic oxidation and conjugation reactions [7 9]. Flurbiprofen undergoes oxidation to yield 4'-hydroxy- and 3',4'-dihydroxyflurbiprofen, the latter undergoing methylation to form 3'-hydroxy-4'-methoxyflurbiprofen, all of which retain the chiral centre (Figure 1). These metabolites, in addition to the parent compound are subjected to conjugation by either glucuronidation or in the case of the phenols, sulphation [10]. Urinary recovery studies by Szpunar et al. [8] indicate that approximately 22% of a 100 mg oral dose of flurbiprofen is excreted in urine as the parent drug, 48% as 4'-hydroxyflurbiprofen and only 7% as 3 '-hydroxy-4'-methoxyflurbiprofen, with each primarily present as conjugates rather than in the free form.

Achiral chromatographic analysis of the metabolites of flurbiprofen in biological fluids has received relatively little at- tention in the literature [8, 11, 12]. Furthermore, the development of enantiospecific methods has been restricted due to the difficulties associated with the chromatographic separation and resolution of flurbiprofen and its metabolites, coupled with the lack of authentic standards of the individual enantiomers of the metabolites. Adams et al. [13] presented, in abstract form, a methodology for the stereospecific analysis of flurbiprofen and the 4'-hydroxy- metabolite in urine and

Original

0009-5893/00/02 7-12 $ 03.00/0

Chromatographia 2003, 57, January (No. 1/2)

�9 2003 Friedr. Vieweg & Sohn Verlagsgesellschaft mbH

7

F CH3 ~ H- - -COOH

flurbiprofen

F

4'-hydroxyflurbiprofen

F F-- F --7 ', ~ ells ,

7 "3 i

HO HO 1

Y-hydroxy-4'-metboxyflurbiprofen L__ 3 ' , 4 ' - d i h y d r o x y f l u r b i p r o fen __1

Figure 1. Metabolism of flurbiprofen (* denotes chiral centre).

plasma, based on precolumn derivatiza- tion with ( )-(S)-l-phenylethylamine and separation of the diastereoisomeric deri- vatives by HPLC using an octyl column. A similar indirect approach was subse- quently adopted by Knadler and Hall [14] for the bioanalysis of flurbiprofen and both major metabolites, which circum- vented chromatographic separation difficulties by employing different methods for the analysis of (R)- and (S)-flurbiprofen and for the enantiospecific analysis of 4'-hydroxyflurbiprofen and 3'-hydroxy- 4'-methoxyflurbiprofen. Furthermore, the determination of flurbiprofen in plasma or urine required different conditions, the analysis of urine extracts required an adjustment in the mobile phase composition to allow for the simultaneous mea- surement of 3',4'-dihydroxyflurbiprofen and required the use of fluorescence detection, rather than UV, to minimise interfer- ence from endogenous compounds. How- ever, it is worth noting that UV detection was still essential for the monitoring of the internal standard in this assay.

We have recently reported the simultaneous chromatographic separation and enantiomeric resolution of flurbiprofen and its major metabolites, 4'-hydroxyflurbiprofen and 3'-hydroxy-4'-methoxyflurbiprofen on a amylose tris(3,5-dimethyl- phenylcarbamate) CSP (Chiralpak AD) and by a combination of semi-preparative isolation and chiroptical characterisation established the enantiomeric elution order for the metabolites [15]. The ability to re- solve all three analytes within a reason- able run time would indicate that this method is ideally suited for application in enantiospecific bioanalysis. This report therefore describes the development and validation of enantiospecific assays, based on the Chiralpak AD CSP, for the quanti-

fication of flurbiprofen and its metabolites in urine and plasma. The utility of these assay methods for dispositional investigations was assessed by performing some preliminary urinary recovery studies following the oral administration of the drug to three healthy male volunteers and the establishment of a plasma concentration-time profile for one of these volunteers. Part of this work has been presented previously in abstract form [16].

Experimental Materials

Acetonitrile, dichloromethane, ethanol, ethyl acetate, hexane and isopropanol (HPLC grade) were purchased from Rathburn (Walkerburn, UK). Trifluoroa- cetic acid (TFA) and sodium hydrogen phosphate (GPR grade) were purchased from BDH (Poole, Dorset, UK). Sodium acetate pHix Buffer (4.0 M, pH 5.5 • 0.03 at 25~ was purchased from Pierce (Rockford, Illinois, USA). J3-Glucuroni- dase (EC 3.2.1.31) type H-5, with 530 units of J3-glucuronidase and 30 units of sulphatase activity per mg solid, was purchased from Sigma Chemicals (Poole, Dorset, UK). (S)-Naproxen was obtained from Aldrich Chemicals (Gillingham, Dorset, UK). rac-Benoxaprofen was kindly supplied by Lilly Research Centre Ltd. (Windlesham, Surrey, UK) and rac-, (S)-, (R)-flurbiprofen, rac-4'-hydroxyflurbiprofen and rac-3 ' -hydroxy-4'-methoxyflurbiprofen were the generous gifts of Boots Company PLC (Nottingham, UK).

Chromatographic Column

The chiral stationary phase was a Chiral- pak AD (amylose tris(3,5-dimethylphe- nylcarbamate)) column (250 • 4.6 mm, 10 ixm), used with a matching guard column (50 • 4.6 mm, 10 ixm), and was supplied by HPLC Technology Ltd. (Mac- clesfield, UK).

Instrumentation

Chiral-phase HPLC was performed using an LDC Constametric 3000 pump (Stone, UK) linked to a Kontron SFM-25 fluorescence detector (Watford, UK) and a LDC Spectromonitor 3100 UV detector (Stone, Staffs., UK). Data from the two detectors was acquired using LDC CI- 4000 and LDC CI-4100 computing inte- grators (Stone, Staffs., UK). Samples were injected on column using a Perkin Elmer ISS-100 autosampler (Beacons- field, Bucks., UK).

Urine Assay

Urinary calibration standards (1 mL) containing racemic flurbiprofen, 4'-hydroxyflurbiprofen and 3'-hydroxy-4'-methoxyflurbiprofen were prepared in drug-free control urine to give single enantiomer flurbiprofen concentrations of 0.05, 0.1, 0.25, 0.5, 1.0, 2.5, 5.0, 7.5 and 10 ixgmL 1 and single enantiomer 4'-hydroxy- and 3 '- hydroxy-4'-methoxy- metabolite concentrations 2-fold more concentrated and 5- fold more dilute than those of flurbiprofen respectively.

Determination of Free Levels

(S)-Naproxen (2.5mg; 501xL of a 50 ixg mL 1 solution in acetonitrile) as internal standard, was added to 1 mL samples of urine standard, quality-control urine or volunteer urine. The samples were acidified by the addition of hydrochloric acid (1.0 M; 100 ixL) and buffered to pH 3.8 with 1.0 mL of phosphate buffer (pH 3.8; 1.0 M). Liquid-liquid extraction was performed by adding 6 mL of hexane- isopropanol (95:5 v/v), the solution was vortex mixed and then gently shaken on a test-tube rocker for 20 minutes. Phase separation was achieved by centrifugation for 5 minutes at 1000 g and the organic layer was then transferred into a clean glass tube and evaporated under a gentle

8 Chromatographia 2003, 57, January (No. 1/2) Original

stream of nitrogen at 40 ~ on a dry heat- (a) ing block. The residue was then reconsti- ~ tuted in 150 ixL mobile phase and 100 ixL 1 2 injected into the chiral phase HPLC system.

Separation and enantiomeric resolution of analytes was performed using a Chiralpak AD CSP column protected by a guard column containing similar material. The mobile phase consisted of hexane- ethanol (90:10 v/v) containing TFA (0.05% v/v) as modifier, at a flow rate of /

1.0 mL min 1. UV detection at 254 nm was used for monitoring flurbiprofen, 4'-

I hydroxyflurbiprofen and the higher con- 0 centrations of 3'-hydroxy-4'-methoxyflurbiprofen (calibration range: 0.1 2.0 ixg mL 1 for each enantiomer). How- ever, fluorescence detection was used for lower concentrations of 3'-hydroxy-4'- (b) methoxyflurbiprofen (calibration range: J~ 0.01 1.0 ixg mL 1 for each enantiomer), .~ the excitation wavelength was 288 nm and the emission cut-off filter was 340 nm. .~

On each day of analysis one set of urin- ,= ary standards was analysed together with the urine samples. Calibration curves were constructed by plotting the ratio of the peak area of each analyte to (S)-naproxen ~'~ (peak area ratio) against the concentration of each analyte enantiomer and subjecting the data to linear regression analysis.

Determination of Free Plus AcyI-Glucuronide Levels

Aliquots (0.1 mL) of urine blank or volunteer urine samples were diluted to 1.0 mL with distilled water and treated with sodium hydroxide (1.0 M; 200 ixL) and the hydrolysis reaction left to proceed for 2 hours at room temperature. Subse- quently, the base was neutralised and the samples acidified by the addition of hydrochloric acid (1.0 M; 300 ixL). Then (S)- naproxen (2.5 ixg), 1.0 mL of phosphate buffer (pH 3.8; 1.0 M) and 6.0 mL ofhex- ane-isopropanol (95:5 v/v) added before the samples were extracted and analysed as described above for the free nonconju- gated analytes.

Determination of Total (i.e. Free, AcyI-Glucur- onide Plus Phenolic Conjugates) Levels

To aliquots (0.1 mL) of urine blank or volunteer urine samples, for optimal hydrolysis of phenolic conjugates, was added 13- glucuronidase (EC 3.2.1.31) type H-5 (1000 units of 13-glucuronidase plus 57 units of sulphatase) in 0.9 mL acetate buffer (pH 5.0; 0.05 M) and incubated at

4

5

t I I I I I I I I I I

20 40 0 20 40 0

Retention t ime (min)

5

I I I I

20 40

I I I I I I I I I I I

0 20 40 0 20 40 0

6 7

6

I I I I 20 40

Retention time (min)

i ii iii

Figure 2. Chiral-phase chromatograms of (i) reference standards, extracts of (ii) blank drug free urine and (iii) urine standard; using (a) UV detection and (b) fluorescence detection. Peaks: 1 = (R)- flurbiprofen; 2 = (S)-flurbiprofen; 3 = (S)-naproxen (I.S.); 4 = (R)-4'-hydroxyflurbiprofen; 5 = (S)- 4'-hydroxyflurbiprofen; 6 = (R)-3'-hydroxy-4'-methoxyflurbiprofen and 7 = (S)-3'-hydroxy-4'- methoxyflurbiprofen [Mobile phase hexane-ethanol (90:10 v/v) containing TFA (0.05% v/v); Flow rate, 1.0mLmin 1].

37~ for 16 hours. Following this the samples were treated with sodium hydroxide (1.0 M; 200 ixL) and the procedure followed as described above for base hydrolysis.

Plasma Assay

Plasma calibration standards (1 mL) containing racemic flurbiprofen and 4'-hydroxyflurbiprofen were prepared in drug- free control plasma to give single enantiomer flurbiprofen concentrations of 0.1, 0.2, 0.4, 0.6, 0.8, 1.0, 2.0, 4.0 and

6.0 ixg mL 1 and single enantiomer 4'-hydroxyflurbiprofen concentrations 10-fold more dilute than those of flurbiprofen.

rac-Benoxaprofen (0.25 Ixg; 50 IxL of a 5.0 Ixg mL 1 solution in acetonitrile) as internal standard, was added to 1 mL samples of plasma standard, plasma blank or volunteer plasma. Subsequent treatment of the samples was carried out as described above for the determination of the free analyte concentrations in urine, with the exception that hexane-ethyl acetate (90:10 v/v; 6 mL) was used as the extraction solvent. Chromatographic analysis was performed using the Chiralpak AD

Original Chromatographia 2003, 57, January (No. 1/2) 9

Table I. Within- and between-day imprecision and accuracy of the urine assay for the enantiomers of flurbiprofen and its major metabolites after analysis of standard samples (n = 6).

Concentration R - enmltiomer S-enantiomer

Analyte Enantiomer Determined Determined concentration, pg n-lL -I pg mL -1 CV % MD" % ~tg mL-' CV % MD %

Within-day Flurbiprofen

4'- Hydr oxyflurbipr ofen

3'-Hydroxy-4'-methoxyflurbiprofen b

Between-day Flurbiprofen

4'-Hydr oxyflurbipr o fen

�9 b Y -Hydr oxy-4'-m eth oxyflurblprofen

0.1 0.108 7.47 8.0 0.103 6.39 3.0

0.5 0.47 4.05 -6.0 0.50 3.49 0.0

5 5.04 2.97 0.8 5.03 2.88 0.6

0.2 0.211 4.25 5.5 0.203 3.98 1.5

1 1.05 2.86 5.0 1.04 2.68 4.0

10 10.10 2.84 1.0 10.10 2.90 1.0

0.02 0.020 6.32 0.0 0.019 6.10 -5.0

0.1 0.099 2.63 - 1.0 0.098 2.54 -2.0

1 1.05 3.86 5.0 1.04 3.80 4.0

0.1 0.103 10.51 3.0 0.098 11.61 -2.0

0.5 0.49 2.10 -2.0 0.49 3.20 -2.0

5 5.01 2.64 0.2 4,98 2.02 -0.4

0.2 0.207 3.89 3.5 0.201 4.35 0.5

1 1.02 4.40 2.0 0.99 5.86 - 1.0

10 9.88 4.15 - 1.2 9.87 4.60 -1.3

0.02 0.021 8.03 5.0 0.020 11.91 0.0

0.1 0.098 7.71 -2.0 0.096 7.62 -4.0

1 1.03 4.20 3.0 1.02 4.52 2.0

a b MD, mean difference; data based on fluorescence (0.02 and 0.1 btg mL 1) and UV (1 btg mL 1) detection.

Table II. Liquid-liquid extraction efficiencies for the enantiomers of flurbiprofen and its metabolites from urine and plasma (mean • SD, n = 6).

Extraction Efficiency %

Media" Analyte Enantiolncr concentration ,ug mI, -1 R enantiomer S-cnant iomcr

Urine

Plasma

Flurbiprofen

4'-Hydroxyfl urbiprofen

3'-Hydr oxy-4'-m ethoxyfl urbiprofen ~

Flurbiprofen

4'-Hydroxyfl ur biprofen

0.1 8 9 . 5 i 6 . 8 88.9•

0.5 89.2• 89.0•

5 86.8• 86.5•

0.2 74.5• 74.2~3,8

1 75.2• 74.9~2.8

10 75.4• 75.1•

0.02 78.3• 77.8•

0.1 79.1~4.8 78.8•

1 77.4• 77.1•

0.2 9 0 . 4 t 1 0 . 8 89.3•

1 92.5• 91 .9•

5 9 t . 4 • 91.0•

0.02 85.4• 86.0•

0.1 86.4• 85.9•

0.5 87.4• 87.1•

a extraction was performed using hexane-isopropanol (95: 5 v/v) and hexane-ethyl acetate (90:10 v/v) b 1 for urine and plasma samples respectively; data based on fluorescence (0.02 and 0.1 btg mL ) and

1 UV (1 btg mL ) detection.

CSP with a mobile phase of hexane-ethanol (87:13 v/v) containing T F A (0.05% v/v) at a flow rate of 1.0 mL min 1. Column

eluate was monitored using a fluorescence detector set at excitation and emission wavelengths of 288 and 340 nm respectively.

On each day of analysis one set of plasma standards was analysed together with the plasma samples. Calibration curves were constructed by plotting the ratio of the peak heights of each analyte to first eluting enantiomer of benoxaprofen (peak height ratio) against the concentration of

each analyte enantiomer and subjecting the data to linear regression analysis. The analyte concentrations in plasma samples were determined by comparing their respective peak height ratios to the appropriate calibration curves.

Validation of Analytical Methodologies

Extraction Efficiency

The recoveries of the extraction proce- dures were determined for the appropriate

analytes in urine and plasma at three concentrations (n = 6), within the respective calibration ranges, using racemic flurbiprofen, 4 '-hydroxyflurbiprofen and 3'- hydroxy-4'-methoxyflurbiprofen.

Within- and Belween-DayAssayAccuracy and Precision

The urine and plasma assays were assessed by six replicate analyses on the same day at three different concentrations of standard to yield within-day precision, expressed as the coefficient of variation, and within-day accuracy, expressed as percentage of expected values, at these concentrations. The between-day precision and accuracy of the assays were de-


termined by analysing the above standard

samples on six consecutive days.

Limit of Quantification

The limit of quantification ( k O q ) for the analytes in the different methodologies was expressed as the lowest concentration that could be quantified with an imprecision < 16% and was established as the lowest point of the respective calibration curve.

Enantiomeric Excess and Quantification

To determine whether the analytical pro- cedures could produce accurate data with respect to the enantiomeric composit ion of flurbiprofen over a wide range of concentrations, a series of urine and plasma quality-control samples of varying concentration and enantiomeric composit ion were prepared using the individual flurbi-

profen enantiomers. "Total" flurbiprofen concentrations of 0.5, 2.0 and 10 ixg mL 1 in urine and 0.5, 2.0 and 6 ixg mL 1 in

plasma were prepared with the following enantiomeric compositions: R.'S; 20:80; 40:60; 60:40 and 80:20 for each concentration. The samples were extracted and analysed in triplicate and the precision and accuracy of the method was determined for each enantiomeric composit ion at the chosen concentrations as described above.

Precision of Urinary De-Conjugation Approaches

The applicability of the de-conjugation methodologies for quantitative analysis was assessed by determining the precision following the replicate analysis (n = 6) of a volunteer 's 4 6 h post-dose urine collection, using the enzyme plus base approach and the base approach.

Preliminary In Vivo Investigations

Three healthy male volunteers (aged 48, 35 and 28 years; weight 68, 81 and 70 kg respectively) were given a single oral dose of racemic flurbiprofen (1 • 100 mg Fro-

ben | tablet) with 150 mL water following an overnight fast. One volunteer was also administered 50 mg (R)-flurbiprofen and 50 mg (S)-flurbiprofen with a washout in-

terval between each dose of at least two weeks.

A blank, drug-free, urine sample was obtained prior to drug administration and sequential samples were collected at two

Table III. Imprecision and accuracy data for the determination of a series of different enantiomeric compositions of flurbiprofen in urine at different total concentrations (n = 3).

Enantiomer Nomina l De te rmined

conc. a /ag mL "1 conc. gg mL 4 CV % M D b %

Concentration, 0.5 tag mL ~ R 0.1 0.111 7.64 11.3

S 0.4 0.431 1.16 7.8

R 0.2 0.209 2.91 4.5

S 0.3 0.318 2.10 5.9

R 0.3 0.305 4.10 1.7

S 0.2 0.217 2.08 8.3

R

S

Concentration, 2 lag mL -I R

S

R

S

Concentration, 10 gg mL q R

S

0.4 0.410 3.33 2.6

0.1 0.109 1.41 8.7

0.4 0.402 2.07 0.4

1.6 1.595 1.05 -0.3

0.8 0.800 1.52 0.0

1.2 1.188 1.61 -1.0

1.2 1.190 0.46 -0.9

0.8 0.788 1.88 -1.5

1.6 1.610 0.40 0.6

0.4 0.390 1.68 -2.5

2 2.17 2.44 8.6

8 7.89 2.44 -1.3

4 3.94 2.54 -1.4

6 5.99 2.52 -0.1

6 6.02 1.20 0.3

4 4.01 0.77 0.3

8 7.94 0.56 -0.7

2 2.03 0.93 1.3

b conc., concentration. MD, mean difference.

hourly intervals up to 12 hours post dosing, together with a 12 24 h sample. The individual urine volumes were recorded and 50 mL aliquots retained for analysis. Blood samples were also taken from one volunteer following the administration of the 100 mg racemic dose. A venous blood sample (10 mL) was collected prior to flur-

biprofen dosing and at 0.25, 0.50, 0.75, 1, 1.5, 2, 2.5, 3, 3.5, 4, 6, 8, 10 and 24 h post administration. The procedure was ap- proved by the Research Ethics Committee of the School of Medicine and Dentistry, King's College London and each subject gave his written consent before participa-

tion. All urine and plasma samples were stored at 20 ~ until required for chromatographic analysis.

Determinat ion of pharmacokinetic

parameters was performed by conven- tional non-compartmental analysis of the enantiomeric serum concentration-time profiles. The maximum observed drug and 4 '-hydroxy- metabolite enantiomer plasma concentrations (Cm~x) and the time to attain them (tm~x) were obtained from an examination of the individual data points. The elimination half-lives (tin) were calculated using the method of least squares from the terminal linear phase of the semi-


Table IV. Imprecision data for the determination of the enantiomers of flurbiprofen and its metabolites following enzymatic plus base and base hydrolytic treatments of a urine sample (n = 6).

Analytc Enantiomcr Determined

concentration p.g m1-1 CV %

Enzyme & base hydrolysis

Flurbiprofen R 69.9 2.66 S 56.0 2.77

4'-Hydroxyflurbiprofen R 126.5 3.40 S 109.2 3.31

3 '-Hydroxy-4'-methoxyflurbipro fen R 7.33 1.55 S 8.72 1.72

Base hydrolysis Flurbiprofen R 68.3 2.22

S 54.8 2.28

4'-Hydroxyflurbiprofen R 124.0 5.55 S 108.8 5.93

3'-Hydroxy-4'-methoxytlurbiprofen R 0.83 1.37

S 1.33 2.15

(a) (b)

J F I I

0

1

I I I

10 20

2 1

V- u u u g I

0 10 20

Retention time (rain)

Figure 3. Chiral-phase chromatograms of extracts from a plasma sample (8 h) obtained from a volunteer following the oral administration of 100 mg racemic flurbiprofen using a mobile phase composition of (a) hexane-ethanol (90:10 v/v) with TFA (0.05% v/v) and (b) hexane-ethanol (87:13 v/v) with TFA (0.05% v/v) and fluorescence detection. Peaks: 1 = (R)-flurbiprofen; 2 = (S)-flurbiprofen; 3 = (S)-naproxen (I.S.); 4 = (R)-4'-hydroxyflurbiprofen; 5 = (S)-4'-hydroxyflurbiprofen; *, additional unknown peaks.

logarithmic plasma concentration vs time curves. Areas under the plasma enantiomer concentration-time curves (AUC0 24) were estimated using the trapezoidal method up to 24 h post drug administration. The values of the AUC were extrapolated to time infinity (AUC0 o~) using C24/k, where C24 is the plasma concentration at 24 h post drug administration and k is the terminal dispositional rate constant. The apparent oral clearance of each flurbiprofen enantiomer (CL/F = D/AUC0 o~) and volume of distribution (VD/F= CL/k) were calculated by considering the dose (D) to

be equal to one-half the administered dose of the racemate, and F to be the systemic availability.

Results and Discussion

UrineAssay

The simultaneous baseline separation and enantiomeric resolution of flurbiprofen and its metabolites within a run time of only 45 minutes using a Chiralpak AD CSP, has been described recently with en-

antiomeric resolution values of 1.67, 3.67 and 3.44 for flurbiprofen, 4'-hydroxyflurbiprofen and 3'-hydroxy-4'-methoxyflurbiprofen respectively, the elution order being the R- before the S-enantiomer for each analyte [15], and thus this CSP has been evaluated for use in bioanalytical studies.

The presence of the drug and its metabolites at widely different concentration ranges in urine [8], make it almost impos- sible to use a single detector with accept- able sensitivity for the analysis of all three analytes. However, the fluorescent nature of flurbiprofen and its metabolites allow for the use of UV and fluorescence detectors in tandem to overcome this pro- blem. The fluorescence detector was set at excitation and emission wavelengths of 288 nm and 340 nm respectively to pro- vide an optimum response for the detection of 3'-hydroxy-4'-methoxyflurbiprofen, the enantiomers of this analyte are present at the lowest concentrations in urine samples and being the later eluting- peaks on the Chiralpak AD CSP exhibit the poorer peak shapes in terms of band broadening. The UV detector was used at 254 nm and was of sufficient sensitivity to allow for the simultaneous analysis of the enantiomers of flurbiprofen, 4'-hydroxyflurbiprofen and also higher concentrations of 3'-hydroxy-4'-methoxyflurbiprofen (> 0.1 ixg mL 1 for each enantiomer in urine).

(S)-Naproxen was selected as internal standard due to its fluorescence and UV- absorption characteristics, together with its favourable chromatographic proper- ties, being baseline resolved between the peaks of (S)-flurbiprofen and (R)-4'-hydroxyflurbiprofen with a retention time of 10.1 minutes (Figure 2). Flurbiprofen and its metabolites were isolated from urine by liquid-liquid extraction using a hexane- isopropanol (95:5 v/v)mixture following acidification. Typical chromatograms obtained using the Chiralpak AD CSP for extracts from a blank urine sample and a urine standard are shown in Figure 2.

Previous assay methods have quantified the conjugates of flurbiprofen, and its metabolites, by using either base [11] or acid hydrolysis [12, 14]. However, in addition to acyl glucuronidation, the metabolites of flurbiprofen undergo phenolic conjugation to form ether glucuronides and sulphate conjugates and the dis- advantage of the above methodologies is that they do not differentiate between conjugation at the different sites. There-


Table V. Within- and between-day imprecision and accuracy of the plasma assay for the enantiomers of flurbiprofen and 4'-hydroxyflurbiprofen after analysis of standard samples (n = 6).

Concentration

R -enantiomer S -enantiomcr

Analyte Enantiomer Dctcnnincd Determined

concentration gg mL "l gg lnL -1 CV % M D " % g g mL a CV % MD %

Within-day Flurbiprofen

4'-Hydroxyflurbiprofen

Between-day

Flurbiprofen

4'-Hydr oxyflurbipr o fen

0.2 0.199 3.96 -0,5 0.197 3.85 -1.5

1 1.01 3.04 1,0 1.00 3.41 0.0

5 5.04 1.00 0.8 5.04 0.96 1.0

0.02 0.021 2.67 5.0 0.021 1.96 5.0

0.1 0.099 3.55 -1.0 0.101 4.82 1.0

0,5 0.51 1.78 2.0 0.50 2.11 0.0

0.2 0.203 5.36 1.5 0.204 3,77 2.0 1 1.02 2.54 2.0 1.01 2.29 1.0 5 5.11 1.46 2.2 5.12 1.63 2.4

0.02 0,021 4.72 5.0 0.021 4.26 5.0 0.1 0.099 3.90 - 1.0 0.100 5.26 0.0 0.5 0.51 2.06 2.0 0.51 2.18 2.0

a MD, mean difference.

fore, we have developed an enzymatic- base hydrolysis combinat ion method, in addition to a base hydrolysis procedure, that allows for the quantification of both acyl- and phenolic-conjugation of the flurbiprofen metabolites. Phenolic conjugates tend to be relatively stable and so mild basic conditions are suitable only for the liberation of the free acids from their ester conjugates, including any isomers that may have arisen from intramolecular rearrangement (acyl migration). A par- tially purified form of J3-glucuronidase, namely J3-glucuronidase (EC 3.2.1.31) type H-5, was selected for enzymatic hydrolysis as it also contains sulphatase activity and so will hydrolyse both phenolic- and 1-O-acyl- glucuronides in addition to sulphate conjugates. But it is worth noting that the glucuronides derived via acyl migration are not susceptible to enzymatic hydrolysis [17]. Conditions for the J3-glucuronidase hydrolysis of the conjugates of flurbiprofen and its metabolites were optimised by performing enzyme ki- netic experiments using volunteer urine samples to determine the influence of enzyme content and incubat ion time (data not shown). On the basis of the results obtained from these experiments, treatment of the urine samples with a 1000 units of J3-glucuronidase (also containing 57 units of sulphatase activity) for a durat ion of 16 hours, which is a convenient time period for over-night incubation, was deemed most appropriate for application in the urine assay method. Furthermore, the time-course studies suggested that the

(a) (b) (c)

I !

0

5

6

3

I! II

[ I I I I I I I I I l I

10 20 0 l0 20 0 10 20

Re ten t ion t ime (min)

Figure 4. Chiral-phase chromatograms of (a) reference standards, extracts of (b) blank plasma and (e) plasma standard; using fluorescence detection. Peaks: 1 = (R)-flurbiprofen; 2 = (S)-flurbiprofen; 3 = (R)-4'-hydroxyflurbiprofen; 4 = (S)-4'-hydroxyflurbiprofen and 5 and 6, enantiomers of benoxaprofen (I.S.) [Mobile phase hexane-ethanol (87:13 v/v) containing TFA (0.05% v/v); Flow rate, 1.0 mL min 1].

enantiomeric composition of all three analytes remained essentially constant throughout, indicating that there was no stereoselective hydrolysis or racemization under the experimental conditions em- ployed. In addition, the methodology could be extended by the use of inhibitors of J3-glucuronidase and/or sulphatase in order to determine the exact nature of the phenolic conjugates should the need arise.

Calibration curves for the analysis of the enantiomers of flurbiprofen, 4 '-hydro-

xyflurbiprofen and 3'-hydroxy-4'-methoxyflurbiprofen isolated from urine, were linear between 0.05 10 pg mL 1, 0.1 2 0 p g m L l a n d 0 . 1 2 . 0 p g m L l respectively using UV detection. The optimisa- tion of the fluorescence detector for 3 '-hydroxy-4'-methoxyflurbiprofen allowed for the sensitive analysis of its enantiomers in urine between 0.01 1.0 pg mL 1. Linear regression analysis of the calibration curves for all the analytes routinely gave correlation coefficients better than 0.998 with limits of quantification of


(a)

I I 0 20

3 4

I I I J t

0 20 40

Retent ion t ime (rain)

6 7

I I I I I I I

40 0 20 40

(b)

12

6 7

I I I I I I I I

20 40 0 20 40

4 5

3

12

J~. L..,.. 6 7

I I I I I

0 20 40

Retent ion t ime (min)

i ii iii

41 3

2

Figure 5. Chiral-phase chromatograms of (i) non-treated, (ii) base treated and (iii) enzymatic and base treated extracts of a urine sample from a volunteer obtained 4 6 h following the oral administration of racemic flurbiprofen (100 mg); using (a) UV detection and (b) fluorescence detection. Peaks: 1 = (R)-flurbiprofen; 2 = (S)-flurbiprofen; 3 = (S)-naproxen (I.S.); 4 = (R)-4'-hydroxyflurbiprofen; 5 = (S)-4'-hydroxyflurbiprofen; 6 = (R)-3 '-hydroxy-4'-methoxyflurbiprofen and 7 = (S)-3 '- hydroxy-4'-methoxyflurbiprofen [Mobile phase hexane-ethanol (90:10 v/v) containing TFA (0.05% v/v); Flow rate, 1.0 mL min 1].

(a) (b) 5

3 at

I I I I [

0 10 20

tsv._ r -

i t I I i

20 0 10

Retention time (min)

0.051xg m L 1, 0.1 Ixg mL 1 and 0.01 Ixg m L 1 for the enantiomers offlurbiprofen, 4 '-hydroxyflurbiprofen and 3'-hydroxy- 4 '-methoxyflurbiprofen respectively. Ad-

ditional validation was also carried out for the overlap region between the UV and fluorescence calibration curves for the 3 ' -hydroxy-4 '-methoxy- metabolite, comparison of enantiomer concentrations determined using both detection methods yielded lines with gradients of 1.02 and 1.05, and correlation coefficients of 0.996 and 0.995, for (R)- and (S)-3 ' -hydroxy-4'- methoxyflurbiprofen respectively.

The analytical procedure showed ac- ceptable within and between day precision and accuracy at all concentration levels examined for the different analytes with coefficients of variation less than 12% and mean differences no greater than 8% (Ta- ble I). Extraction efficiencies were determined at three different concentrations for each analyte and essentially by employing the selected extraction system, optimal recovery of the polar metabolites was sacrificed for the maintenance of "contaminant-free" extracts (Table II). Extraction recoveries were generally greater than 86%, 74% and 77% for the enantiomers of flurbiprofen, 4 '-hydroxyflurbiprofen and 3 '-hydroxy-4'-methoxyflurbiprofen respectively.

The methodology was further validated by analysis of a series of drug-free urine-samples to which mixtures of the individual enantiomers of flurbiprofen at

three different concentration levels were added. This validation approach is necessary as biological samples from pharmacokinetic and metabolic studies will con- tain non-racemic mixtures of analytes due to stereoselectivity in drug metabolism and disposition [18]. The precision and accuracy values calculated for these analyses are presented in Table III. The measured enantiomeric compositions were in good agreement with the expected values and the variation involved was within accepta- ble limits at all three " total" concentrations examined. These data indicate that

Figure 6. Chiral-phase chromatograms of plasma samples (a) 3 h and (b) 8 h following oral administration of racemic flurbiprofen (100 mg) to a healthy volunteer; using fluorescence detection. Peaks: 1 = (R)-flurbiprofen; 2 = (S)-flurbiprofen; 3 = (R)-4'-hydroxyflurbiprofen; 4 = (S)-4'-hydroxyflurbiprofen and 5 and 6, enantiomers of benoxaprofen (I.S.); *, additional unknown peaks [Mobile phase hexane-ethanol (87:13 v/v) containing TFA

0 1 (0.05'/o v/v); Flow rate, 1.0 mL min ].


racemization is not occurring during sample manipulation and that the concentration of flurbiprofen does not appear to af- fect the measured enantiomeric composition. Due to the unavailability of sufficient quantities of the individual enantiomers of the metabolites, this validation approach was not performed for 4'-hydroxyflurbiprofen and 3'-hydroxy-4'- methoxyflurbiprofen.

The methodologies established for the determination of the conjugates of flurbiprofen and its metabolites were assessed for their suitability by determining the precision following the replicate analysis (n = 6 in each case) of a volunteer urine sample after enzymatic plus base treatment and base treatment alone. As expressed in Ta- ble IV, the coefficient of variation was generally less than 6% for all the analytes following either de-conjugation procedure and thus indicated the utility of the approaches for quantitative analysis. The close agreement of the determined concentrations for the enantiomers of flurbiprofen between the two hydrolytic treatments can also be seen from these validation experiments further indicating the applicability of the approach (Table IV).

Plasma Assay A preliminary investigation of volunteer plasma samples post flurbiprofen administration using the methodology adopted for the analysis of urine samples indicated that the approach would require modification, in terms of the detection method, mobile phase composition and choice of internal standard, for application in plasma analysis.

Flurbiprofen and the 4'-hydroxy- me- tabolite have previously been quantified in human serum or plasma, concentrations of 4'-hydroxyflurbiprofen being 10 to 20 times less than those of the parent drug, whereas 3'-hydroxy-4'-methoxyflurbiprofen has not been detected [11,19]. Similarly, we were unable to detect the 3 '- hydroxy-4'-methoxy- metabolite in plasma samples of a volunteer following administration of racemic flurbiprofen. The development of the plasma assay was therefore restricted to the quantitative analysis of the enantiomers of flurbiprofen and the 4'-hydroxy- metabolite using fluorescence detection set at an optimum sensitivity for 4'-hydroxyflurbiprofen.

Examination of the chromatograms obtained following analysis of extracts of volunteer plasma post drug administra-

Table VI. Imprecision and accuracy data for the determination of a series of different enantiomeric compositions of flurbiprofen in plasma at different total concentrations (n = 3).

E n a n t i o m e r N o m i n a l D e t e r m i n e d

c o n c . a p g m L "1 conc , g g m L "1 C V M D b %

Concentration, 0.5 lag mL "l R 0.1 0.097 1.79 -3.0 S 0.4 0.393 1.28 -1.7

R 0.2 0.200 3.26 0.0 S 0.3 0.295 1.60 -1.6

R 0.3 0.316 2.34 5.2 S 0.2 0.203 2.33 1.3

R

S

Concentration, 2 lag mL ~ R

S

R

S

Concentration, 6 p,g mL -1 R

S

0.4 0.409 1.59 2.2 0.1 0.105 6.78 4.7

0.4 0.402 1.37 0.6 1.6 1.615 1.15 1.0

0.8 0.811 1.16 1.4

1.2 1.202 2.37 0.2

1.2 1.193 0.67 -0.6 0.8 0.813 1.72 1.6

1.6 1.615 0.56 1.0 0.4 0.408 1.47 2.0

1.2 1.19 0.60 -0.8 4.8 4.82 0.50 0.3

2.4 2.40 0.42 0.1

3.6 3.63 0.84 0.8

3.6 3.61 0.48 0.2 2.4 2.39 0.40 -0.4

4.8 4.80 0.19 0.0 1.2 1.19 0.42 -0.7

a b conc., concentration. MD, mean difference.

tion revealed the presence of a slight shoulder on the peak of (S)-4'-hydroxyflurbiprofen (Figure 3a). As the "shoulder" was not observed in chromatograms of plasma standards, and extracts of blank plasma did not show additional peaks in this region of the chromatogram, it was thought that the "shoulder" could be due to co-eluting drug-related material. Also if the peak is due to drug-related material, then as the drug is chiral it is highly likely that the unknown is also chiral and that two, rather than one, additional peaks may be present. It was therefore necessary to adjust the mobile phase composition in

an attempt to separate the interfering component from the peak due to (S)-4'- hydroxyflurbiprofen. Increasing the ethanol content of the mobile phase from 10% to 13% resulted in the separation of the two components with the additional peak eluting just after the tail of the (S)-4'-hydroxyflurbiprofen peak (Figure 3b). Ex- amination of hydrolysed and non-hydrolysed volunteer urine samples under these chromatographic conditions did not un- mask a similar peak after the S-4'-hydroxy- metabolite peak, suggesting that this particular product was not detectable in urine (data not shown).


Table VII. Urinary excretion (0 24 h) of flurbiprofen and its metabolites following the administration of the racemic drug (100 mg) to three healthy volunteers (data expressed as a mean percentage of the administered dose • SD, n = 3).

Analyte R-enantiomer S -enantiomer S/R ratio

Flurbiprofen

free: 0.56 • 0.41 0.50 • 0.41 0.89

acyl-conjugate: 9.95 • 3.24 8.33 • 2.85 0.84

4'-Hydroxyflurbiprofen

free: 1.13 • 1.25 1.25 • 1.15 1.10

acyl-conjugate: 13.17 • 0.83 11.32 • 1.37 0.86

phenol-conjugate: 2.10 • 0.83 1.44 • 0.67 0.69

3'-Hydr oxy-4'-methoxyflurbiprofen

free: 0.004 • 0.005 0.005 • 0.005 1.25

acyl-conjugate: 0.034 4- 0.026 0.0414- 0.043 1.21

phenol-conjugate: 1.30 • 0.17 1.63 • 0.43 1.25

Total Recovery 28.24 4- 1.03 24.68 • 0.30 0.87

Table VIII. Urinary excretion (0 24 h) of flurbiprofen and its metabolites following the administration of (R)-flurbiprofen (50 mg), rac-flurbiprofen (100 mg) and (~S)-flurbiprofen (50 mg) to a healthy volunteer (data expressed as a percentage of the enantiomeric dose).

Dose 50 mg 100 mg Raccmatc 50 mg

Analyte R-cnantiomcr R-cnantiomcr S-enantiomer S-cnantiomcr

Flurbiprofen

free: 0.52 0.36 0.26 2.61

acyl-conjugate: 19.94 19.34 15.36 15.89

4'-Hydroxyfl urbipro fen

free: 0.94 0.68 1.50 4.30

acyl-conjugate: 28.05 28.14 25.70 17.31

phenol-conjugate: 0.73 3.64 2.64 1.3

3'-Hydroxy-4'-methoxyflurbiprofen

flee: ND ~ 0.01 0.0t ND

acyl-conjugate: ND 0.01 0.09 ND

phenol-conjugate: 2.47 2.96 3.56 4.82

Total Recovery 52.65 55.14 49.12 46.23

a ND, not detected.

Table IX. Pharmacokinetic parameters for the enantiomers of flurbiprofen and 4'-hydroxyflurbiprofen following the oral administration of rac-flurbiprofen (100 mg) to a healthy volunteer.

Parameter Flurbiprofcn 4'-Hydroxy flurbiprofcn

(units) R-enantiomer S-cnantiomcr R-cnantiomcr S-cnantiomcr

C r ~ (p-g mL -l) 7.8 7.6

t ~ (h) 3.0 3.0

t In (h) 5.5 5.6

AUC (p.g mL -~ h) 45.0 46.3

CL/F (mL min -1) t8.5 18.0

VWF (L) 8.8 8.8

0.36 0.52

3.5 3.5

5.8 5.9

3.1 4.4

A consequence of the alteration in the mobile phase composit ion was partial coalescence between the peaks due to (R)- 4 '-hydroxyflurbiprofen and (S)-naprox-

en, the internal standard (Figure 3b); it was therefore necessary to find an alterna- tive fluorescent internal standard for application in the plasma assay. Racemic benoxaprofen displayed suitable chromatographic characteristics with baseline resolution of its enantiomers and elution after (S)-4'-hydroxyflurbiprofen (Figure 4).

Either enantiomer of benoxaprofen could have been applied for quantitative analysis thus the choice of the first eluting enan- t iomer as the internal standard was an ar- bitrary decision. Due to the non-availability of the individual enantiomers the elution order is unknown. However, previously Booth and Wainer [20] have shown that the elution order for a series of 2-APAs, including flurbiprofen and be-

noxaprofen, was always the R- before the S-enantiomer on the amylose tris(3,5-di- methylphenylcarbamate) CSP and so it is likely that (R)-benoxaprofen elutes prior to (S)-benoxaprofen.

Employment of the same extraction solvent as used for urine samples was found to be unsuitable as it resulted in the co-extraction of plasma contaminants, the presence of which may cause problems in analyte retention, resolution and column stability [21, 22]. A hexane-ethyl acetate (90:10 v/v) mixture displayed more suitable characteristics as it gave clean extracts and recovery values greater than 85% for all analytes at the various concentrations examined (Table II). Typical chromatograms obtained using the Chir- alpak A D CSP for extracts from a blank plasma sample and a plasma standard are shown in Figure 4 and illustrate the ab- sence of interfering peaks at the retention times of the analytes and internal standard.

Quantitative analysis of flurbiprofen and 4 '-hydroxyflurbiprofen in plasma was performed by determining peak-

height ratios instead of peak-area ratios as peak height measurements are less likely to be influenced by the presence of slight overlap between neighbouring peaks, which may be an issue in the chromatographic analysis of volunteer plasma extracts due to the exposure of the additional peak at the tail of the (S)-4'-hydroxyflurbiprofen peak. Single enantiomer calibration curves for flurbiprofen (range: 0.1 6.0 ixg mL 1) and 4'-hydroxyflurbiprofen (range: 0.01 0.6 ixg mL 1) typi-


cally yielded correlation coefficients greater than 0.997 with limits of quantification of 0.1 ixg mL 1 and 0.01 ixg mL 1 for the enantiomers of flurbiprofen and 4'-hydroxyflurbiprofen respectively. The within-day and between-day precision and accuracy values for the enantiomers of flurbiprofen and 4'-hydroxyflurbiprofen, as presented in Table V, are generally good with coefficients of variation and mean differences less than 6% at the concentrations examined. Additional validation using non-racemic mixtures of flurbiprofen at various concentrations exhibited good a correlation between the measured and theoretical enantiomeric compositions with variability typically limited to less than 7% (Table VI).

Volunteer Studies

The utility of the urine assay was demon- strated by analysing urine samples collected at 2 hour intervals from 0 to 12 h followed by a pooled urine sample from 12 to 24 h for three healthy volunteers following the oral administration of 100 mg of racemic flurbiprofen. Typical chromatograms of extracts obtained from non- treated and the various hydrolytic treatments of urine samples from this study are shown in Figure 5. The presence of 4'-hydroxyflurbiprofen at much higher concentrations than 3'-hydroxy-4'-methoxyflurbiprofen in urine and the necessity for a dual detection approach is clearly illu- strated by the large differences observed in the size of the peaks between the enantiomers of the two metabolites.

The amounts of flurbiprofen and metabolites excreted in urine were determined in terms of free and both acyl- and phenol- conjugated levels for all the volunteers and the mean urinary excretion data is presented in Table VII. Of the administered dose some 19.3% was excreted as flurbiprofen, 30.4% as 4'-hydroxyflurbiprofen and 3.0% as 3'-hydroxy-4'-methoxyflurbiprofen over the 24 hour collection period, primarily as conjugates. These results are in good agreement with values previously reported [8, 14]. It is noteworthy that the two metabolites have different preferential sites for conjugation; the 4'-hydroxy- metabolite tends to form acyl-conjugates similarly to the parent drug whereas the 3 '-hydroxy-4'-methoxy- metabolite is excreted mainly as phenolic conjugates (see also Table IV). A lar- ger percentage of the dose was recovered

lO - - - 0 - - (R)-flurbiprofen

(S)-flurbiprofen ( R ) - 4 ' - h y d r o x y f l u r b i p r o f e n

_ ,_

1

0.1

0.01

0 4 8 12 16 20 24

Time (h)

Figure 7. Plasma concentration-time profiles for the enantiomers of flurbiprofen and 4'-hydroxyflurbiprofen following the oral administration of racemic flurbiprofen (100 mg) to a healthy volunteer.

with the R-configuration of the propionic acid moiety over the 24 hour collection period with total flurbiprofen and total 4'-hydroxyflurbiprofen having SIR ratios of 0.84 and 0.85 respectively, these findings are consistent with the observations of Knadler and Hall [14]. However, the minor metabolite, 3'-hydroxy-4'-methoxyflurbiprofen, seemed to display preferential excretion of the S-enantiomer with a SIR ratio of 1.25.

In an additional exploratory study, one of the volunteers was also dosed oral- ly with 50 mg (R)-flurbiprofen and 50 mg (S)-flurbiprofen with a washout interval between each dose of at least two weeks. Comparison of the urinary excretion data following administration of the individual enantiomers with that following dosing with the racemate provides an insight into whether, or how, enantiomeric interac- tions influences drug disposition. The products detected in the urine samples retained the same stereochemical configuration as the administered enantiomer. This highlights the lack of chiral inversion of flurbiprofen in humans and is consistent with the findings of Jamali et al. [9] and Geisslinger et al. [6]. The total recoveries of flurbiprofen and its metabolites following administration of the single enantio-

mers and the racemate to a healthy volunteer are presented in Table VIII. It is worth noting that the recovery values following dosing with rac-flurbiprofen are expressed as a percentage of the enantiomeric dose, rather than administered dose (as in Table VII), for ease of comparison. A greater proportion of the dose was recovered in 24 hours following administration of (R)-flurbiprofen than with (S)- flurbiprofen, this is in agreement with preferential elimination of products with the R-configuration following dosing with the racemate. The total recoveries of material in the R-isomeric form after administration of 50 mg (R)-flurbiprofen and 100 mg rac-flurbiprofen were consistent and there was good agreement between the amounts of the individual metabolites. Similar observations were also made for (S)-flurbiprofen with the only disparages being between the quantities of free (S)-flurbiprofen and acyl-conjugated (S)-4'-hydroxyflurbiprofen recovered. In general, this would suggest that there is little dispositional interaction between the enantiomers of flurbiprofen.

Following the administration of 100mg racemic flurbiprofen, blood as well as urine samples were collected from one of the volunteers and typical chroma-


tograms for the analysis of sample extracts from the 3 and 8 h collections are shown in Figure 6. The additional peaks at 10.0 and 13.2 minutes were not observed in pre-dose plasma or plasma standard extracts and appear to become more prominent with time. These peaks could possibly correspond to the enantiomers of the 3',4'-dihydroxy- intermediate involved in the formation of 3'-hydroxy-4'- methoxyflurbiprofen, it is noteworthy that the separation factor (~) between these two peaks (~ = 1.5) is not too dissim- ilar to the enantioseparation of flurbiprofen (~ = 1.7) and 4'-hydroxyflurbiprofen (~ = 1.8) and that the resolution value (Rs) of 3.08 is close to the resolution of the enantiomers of the 4'-hydroxy- metabolite (Rs = 3.48). However, an authentic sample of 3',4'-dihydroxyflurbiprofen was not available for confirmation and any further investigations of the suspected metabolite were considered to be outside the scope of the current investigations.

Semilogarithmic plots of plasma concentration-time profiles of flurbiprofen and 4'-hydroxyflurbiprofen for the volunteer are shown in Figure 7. Visual examination of the profiles for (R)- and (S)-flurbiprofen suggested that the drug exhibited non-stereoselective pharmacokinetics and that disappearance of the drug from systemic circulation was biphasic. The enantiomers of the 4'-hydroxy- metabolite were present in plasma at concentrations between ten to twenty times less than flurbiprofen enantiomer concentrations at the corresponding time points and their rates of disappearance from plasma ap- peared to be comparable to that of the parent enantiomers.

Pharmacokinetic data analysis of the plasma concentrations revealed only min- or differences in the values of the parameters between (R)- and (S)-flurbiprofen and between (R)- and (S)-4'-hydroxyflurbiprofen, which reiterates that neither the drug or metabolite appear to undergo ex- tensive stereoselective disposition in this individual (Table IX). The observation that (R)- and (S)-flurbiprofen have equivalent half-lives is consistent with the findings of Jamali et al. [9]. The AUC for the (S)-4'-hydroxy- metabolite was 9.5% of the AUC of (S)-flurbiprofen and the corresponding value for the R-enantiomer was 6.9%, emphasising that the plasma concentrations of the enantiomers of this metabolite are relatively low. The half- lives of (R)- and (S)-4'-hydroxyflurbiprofen were essentially equivalent to those of

the parent drug enantiomers suggesting that metabolite elimination is formation rate-limited [23].

Conclusion In summary, reliable, sensitive chiral- phase HPLC assays for flurbiprofen and its major metabolites in urine and plasma have been developed and validated. The compounds being isolated from urine or plasma using liquid-liquid extraction techniques prior to separation and resolution on an amylose tris(3,5-dimethylphe- nylcarbamate) CSP (Chiralpak AD). The use of fluorescence and UV detectors in series allows for the concurrent determination of the enantiomers of flurbiprofen, 4'-hydroxyflurbiprofen and 3'-hydroxy- 4'-methoxyflurbiprofen in urine extracts. The employment of a fluorescence detector is sufficient for the quantitative analysis of the enantiomers of flurbiprofen and the 4'-hydroxy- metabolite in plasma. The methods were applied in preliminary urinary excretion and plasma disposition studies following the administration of flurbiprofen to healthy volunteers. The urinary recovery studies indicate that the acyl- conjugates of flurbiprofen and the 4'-hydroxy- metabolite are the major components in urine and the excretion profile of the drug shows modest enantioselectivity for products with the R-configuration. The analysis of plasma samples from a volunteer suggests that enantioselectivity in the disposition of flurbiprofen in humans is limited. To our knowledge this is the first report of methodology allowing simultaneous determination of the enantiomers of flurbiprofen and its major metabolites in biological fluids using chiral- phase chromatography.

Acknowledgements

During the course of this work BKP was the recipient of a University of London Triangle Trust Postgraduate Studentship and JV was funded by a joint grant from the British Council and Comenius Univer- sity. The authors thank Professor F. De- vinsky of Comenius University for his continued encouragement.

References [1] Adams, S.S.; McCullough, K.F.; Nichol-

son, J.S. Arzneim. Forsch. 1975, 25, 1786 1791.

[2] Brogden, R.N.; Heel, R.C.; Speight, T.M.; Avery, G.S. Drugs 1979,18, 417 438.

[3] Carabaza, A.; Cabr6, F.; Rotllan, E.; Gd- mez, M.; Gutierr6z, M.; Garcia, L.; Mau- ledn, D. J. Clin. Pharmacol. 1996, 36, 505 512.

[4] Brune, K.; Beck, W.S.; Geisslinger, G.; Menzel-Soglowek, S.; Peskar, B.M.; Pes- kar, B.A. Experientia 1991, 47, 257 261.

[5] Menzel-Soglowek, S.; Giesslinger, G.; Beck, W.S.; Brune, K. J. Pharm. Sci. 1992, 81,888 891.

[6] Geisslinger, G.; L6tsch, J.; Menzel, S.; Kobal, G.; Brune, K. Br. J. Clin. Pharma- col. 1994, 37, 392 394.

[7] Risdall, P.C.; Adams, S.S.; Crampton, E.L.; Marchant, B. Xenobiotica 1978, 8, 691 704.

[8] Szpunar, G.J.; Albert, K.S.; Bole, G.G.; Dreyfus, J.N.; Lockwood, G.F.; Wagner, J.F. Biopharm. DrugDi~pos. 1987, 8, 273 283.

[9] Jamali, F.; Berry, B.W.; Tehrani, M.R.; Russell, A.S.J. Pharm. Sci. 1988, 77, 666 669.

[10] Davies, N.M. Clin. Pharmacokin. 1995, 28, 100 114.

[11] Adams, W.J.; Bothwell, B.E.; Bothwell, W.M.; Van Giessen, G.J.; Kaiser, D.G. Anal. Chem. 1987, 59, 1504 1509.

[12] Hutzler, J.M.; Frye, R.F.; Tracy, T.S.J. Chromatogr. B2000, 749, 119 125.

[13] Adams, W.J.; Van Giessen, G.J.; Kaiser, D.G. Pharm. Res. 1988, 5, S-229.

[14] Knadler, M.P.; Hall, S.D.J. Chromatogr. 1989,494, 173 182.

[15] Patel, B.K.; Valentova, J.; Hutt, A.J. Chromatographia 2002, 55, 135 142.

[16] Patel, B.K.; Hutt, A.J.J. Pharm. Pharma- col. 1999, 51 (Suppl), 75.

[17] Caldwell, J.; Hutt, A.J.; Marsh, M.V.; Sin- clair, K.A. In Drug Metabolite isolation and Determination: Reid, E.; Leppard, J.P., Eds, Plenum Publishing Corpora- tion, London, 1983.

[18] Hutt, A.J. Anal. Proc. 1991,28, 185 186. [19] Knadler, M.P.; Brater, D.C.; Hall, S.D.

Br. J. Clin. Pharmacol. 1992,33, 369 375. [20] Booth, T.D.; Wainer, I.W.J. Chromatogr.

A 1996, 737, 157 169. [21] Ducharme, J.; Fernandez, C.; Gimenez,

F.; Farinotti, R. J. Chromatogr. B 1996, 686, 65 75.

[22] Hutt, A.J.; Patel, B.K. In Drug Metabo- lism: Towards the Next Millennium: Gooderham, N.J., Ed, IOS Press, Amster- dam, 1998.

[23] Rowland, M.; Tozer, T.N. In Clinical Pharmaeokineties, 3rd ed., Lea and FeN- ger Books, Philadelphia, 1995.

Received: Apr 26, 2002 Revised manuscript received: Oct 14, 2002 Accepted: Oct 15, 2002


Documents

Stereospecific analysis of flurbiprofen and its major metabolites in plasma and urine by chiral-phase liquid chromatography