Liquid Chromatographic Determination of Penicillins by Postcolumn

7/18/2019 Liquid Chromatographic Determination of Penicillins by Postcolumn

http://slidepdf.com/reader/full/liquid-chromatographic-determination-of-penicillins-by-postcolumn 1/9

ANALYTICAL BIOCHEMISTRY 168, 132- 140 ( 1988)

Liquid Chromatographic Determination of Penicillins by Postcolumn

Alkaline Degradation Using a Hollow-Fiber Membrane Reactor

JUN HAGINAKA’ AND JUNKO WAKAI

Faculty of Pharm aceutical Scie nces , Mukogawa Women S University,

4-16 Edagawa-cho, Nishinomiya , Hyogo 663, Japan

Received May 18, 1987

A high-performance liquid chromatographic method using a hollow-fiber membrane reactor

is described for the determination of penicillins. This method involves separation of penicillins

on a Cl8 column, postcolumn reaction with sodium hydroxide and mercury(I1) chloride intro-

duced into the main flow stream using sulfonated hollow-fiber membrane reactors immersed in

each solution (4 M sodium hydroxide and 3 X lo-’ M mercury(I1) chloride plus lo-* M nitric

acid), and detection at 290 nm based on the uv absorbance of the degradation products. At

penicillin concentrations of 5 pg/ml, within- and between-run precisions (relative standard

deviation) were 0.24-2.39 and 1.19-4.13%, respectively. The detection limits of the proposed

method were l-5 ng at a signal-to-noise ratio of 3. The method was applied to assays of

ampicillin and its metabolites in human serum and urine.

o 1988 Academ ic press, Inc .

KEY WORDS: chromatography; HPLC; penicillins; penicillin metabolites; hollow-fiber

membrane reactor; postcolumn reaction.

A number of high-performance liquid

chromatographic methods combined with

precolumn ( l-6) and postcolumn (4,7- 18)

derivatization to enhance sensitivity and se-

lectivity have been developed for the assay of

penicillins in pharmaceutical preparations

and in body fluids. However, most of the

derivatization methods developed so far are

sensitive for unchanged penicillins, but are

less sensitive or insensitive for their degrada-

tion products and metabolites. In previous

papers, we reported an HPLC method for the

determination of penicillins (12) and of am-

picillin and its metabolites in human urine

( 13) using postcolumn alkaline degradation

with sodium hydroxide, mercury(I1) chlo-

ride, and ethylenediaminetetraacetic acid.

The drawback of the method is that it needs

an additional pump for delivering the re-

agent solution and a mixing unit and a reac-

tion coil as the reactor. They lead to an in-

’ To whom correspondence should be addressed.

crease in detector noise, band broadening,

and dilution, resulting in a reduction in the

sensitivity gained through derivatization and

deteriorating the ultimately obtainable limit

of detection.

In ion chromatography, a hollow-fiber

membrane (i.e., membrane suppressor) has

been used for removing most of the back-

ground conductance of the eluant (19-2 1).

Recently, hollow-fiber membrane reactors

were employed for HPLC (22-26) and flow

injection analysis (27,28). The reactors can

eliminate or control the problems as de-

scribed above.

This paper deals with an HPLC method

for the determination of penicillins using two

sulfonated hollow-fiber membrane reactors

for introducing sodium and mercury(I1) ions

into the main flow stream for the post-

column alkaline degradation reaction. The

method was successfully applied to the deter-

mination of ampicillin and its metabolites in

human serum and urine.

0003-2697/88 3.00

Copyright 0 1988 by Academic Press, Inc.

All righ ts of reproduction in any form reserved.

132



CHROMATOGRAPHIC DETERMINATION OF PENICILLINS

133

MATERIALS AND METHODS

Reagents and materials. Ampicillin

(ABPC),* phenethicillin (PEPC), phenoxy-

methylpenicillin (PCV), and ciclacillin

(ACPC) were kindly donated by Meiji Seika

Kaisha (Tokyo, Japan) and Takeda Chemi-

cal Industries Co. (Osaka, Japan). Benzyl-

penicillin (PCG) was purchased from Sigma

Chemical Co., Ltd. (St. Louis, MO). ABPC

metabolites ((5R,6R)-ampicilloic acid (l),

the (S,S,6R)-epimer (2) and (2R)-piperazine-

2’,5’-dione (3)) were prepared according to

the methods reported previously (29,30).

The structure and abbreviations of penicil-

lins and ABPC metabolites used in this study

are listed in Figs. 1 and 2, respectively. So-

dium heptanesulfonate and other chemicals

of analytical reagent grade were obtained

from Nakarai Chemicals, Ltd. (Kyoto,

Japan). Control serum (Control Serum I

Wako) was purchased from Wako Pure

Chemical Industries, Ltd. (Osaka, Japan). A

sulfonated hollow-fiber membrane (AFS-2

fiber) (ca. 0.3 mm i.d.) was purchased from

Dionex Co. (Sunnyvale, CA).

Deionized, glass-distilled water and dis-

tilled methanol were used for the prepara-

tions of sample solutions and HPLC eluants.

Chromatography. Figure 3 illustrates the

instrumentation and arrangement of HPLC

and hollow-fiber membrane reactors: an

LC-SA pump (Shimadzu Co., Kyoto, Japan)

for delivering an eluant; a Model 7 125 loop

injector (Rheodyne, Cotati, CA) equipped

with a loo-p1 loop for the loading of the sam-

ples; 150 X 4.6-mm-i.d. columns packed

with Develosil ODS (5-pm particle size, No-

mura Chemicals, Seto, Aichi, Japan) for the

separation of penicillins; and Nucleosil Cl8

(5-pm particle size, Macherey-Nagel, Dtiren,

West Germany) for the separation of ABPC

and its metabolites (these columns were pro-

’ Abbreviations used: ABPC, ampicillin; PEPC, phe-

nethicillin; PCV, phenoxymethylpenicillin; ACPC, ci-

clacillin; PCG, benzylpenicillin; cr,, band broadening;

RSD, relative standard deviation. PTFE, polytetrafluo-

roethylene.

ampicillin f ABPC )

ciclacillin (ACPC )

NH,

NH,

A

benzyipenicillin ( PCG ) uCH,-

phenoxymethylpenicillin (PCV )

OCH,-

phenethicillin ( PEPC )

FIG. 1. Structure and abbreviations of penicillins.

tected by the guard columns (30 X 4.6 mm

i.d.) packed with the same materials); coiled,

sulfonated hollow-fiber membranes at

lengths of 80 and 30 cm immersed in sodium

hydroxide and mercury(I1) chloride solu-

tions as the reactor; an SPDdAV spectro-

photometer (Shimadzu Co.) equipped with a

8-PL flow-through cell for detection; a

C-R3A recorder-integrator (Shimadzu Co.)

for recording and integrating chromato-

graphic peaks. The eluants used were as fol-

lows: eluant A, 2 mM sodium dihydrogen

phosphate:2

InM

disodium hydrogen phos-

phate:methanol(O.8:0.8: 1, v/v); eluant B, 10

mM sodium dihydrogen phosphate: 10

mM

disodium hydrogen phosphate:methanol

(l:l:l, v/v); eluant C, 15 mM sodium hep-

tanesulfonate:24 mM phosphoric acid:6

mM

sodium dihydrogen phosphate:methanol

(2:2:2:3.75). The flow rates were maintained

at 0.8 ml/min. Eluant A was used for the



HAGINAKA AND WAKAI

(5R.6R)- ampicillolc acid

(1)

(BS.GR)-ampicilloic acid

2)

(2R)-piperazine-215’-dione (3)

FIG. 2. Structure of ampicillin metabolites.

separation of penicillins; eluant B, ABPC in

serum samples; and eluant C, ABPC and its

metabolites in urine samples. The hollow-

fiber membrane reactors, inserted between

the column and the detector, were immersed

in 50-ml beakers containing 4

M

sodium hy-

droxide and 3

X

lo-’

M

mercury(B) chloride

plus IO-* M nitric acid solutions, respec-

tively. Detection was performed at 290 nm

for unchanged penicillins and 265 nm for

ABPC metabolites (1, 2, and 3).

Pretreatment procedures of serum and

urine samples. Two hundred microliters of

serum samples was ultrafiltered using a Mol-

cut II (Nihon Millipore, Tokyo). A 2Oq.d

portion of the ultrafiltrate was loaded onto a

column.

Urine samples, diluted IO-fold with water,

were filtered with 0.45-pm acrylate copoly-

mer membrane (Gelman Science Japan,

Tokyo). A 20-~1 portion of the filtrate was

loaded onto a column.

RESULTS AND DISCUSSION

Reaction Conditions for Hollow-Fiber

Membrane Reactor

Comparison of detection method, The In previous papers (12,13), we reported an

band broadening (at) due to the postcolumn

HPLC method for the determination of pen-

reactor was estimated by the following

icillins and their metabolites (penicilloates)

PLIHP INJECTOR

methods: method A, detection at 230 nm

without a postcolumn reactor; method B,

detection at 290 nm with an open-tubular

postcolumn reactor; method C, detection at

290 nm with hollow-fiber membrane reac-

tors. For method B, the additional reaction

devices used were as follows: a double-

plunger pump (NP-DX-2, Nihon Seimitu

Kagaku, Tokyo) for delivering the post-

column reagent (0.75

M

sodium hydroxide, 2

X

1Op3

M

mercury(I1) chloride, and lo-*

M

EDTA solution) at a flow rate of 0.2 ml/min,

a mixing tee made of Diflon (each angle,

120”), and a reaction coil of 0.5 mm i.d. X 2

m PTFE tube for the postcolumn reaction.

For method C, the postcolumn reaction con-

ditions were the same as described above.

The band broadening due to the column

(column plus injector, connector, and detec-

tor) and the reactor was calculated.

COLUHN

ELURNT NaOH

bC12

FIG. 3. Experimental setup used in this study.

YASTE



CHROMATOGRAPHIC DETERMINATION OF PENICILLINS 135

using the postcolumn degradation reaction

with sodium hydroxide, mercury(I1) chlo-

ride, and EDTA in the presence of methanol:

penicillins were initially degraded with hy-

droxide ion in the presence of methanol to

the corresponding methyl penicilloates, the

methyl penicilloates and penicilloates were

further converted to the corresponding

methyl penamaldates and penamaldates

(which have uv absorption maxima at ca.

290 and 280 nm, respectively) in the pres-

ence of mercury(I1) ion, and EDTA was

added to prevent the precipitation of mer-

cury(I1) oxide. In this study, we used two sul-

fonated hollow-fiber membrane reactors im-

mersed in sodium hydroxide and mer-

cury(I1) chloride solutions for the above

reaction system.

The postcolumn reaction conditions for

penicillins and ABPC and its metabolites

were examined with respect to the length of

the hollow-fiber membrane reactor and the

concentration of sodium hydroxide and mer-

cury(I1) chloride. Two hollow-fiber mem-

brane reactors were connected in series: the

first one was immersed in sodium hydroxide

solution and the second, in mercury(I1) chlo-

ride plus nitric acid solution. Nitric acid was

added to prevent the precipitation of mer-

cury(I1) oxide (which is caused by sodium

ion leak from the second hollow-fiber mem-

brane to the mercury(I1) chloride solution).

A 20-11 portion of the solution of penicillins

or ABPC and its metabolites was loaded onto

the column and the peak heights were mea-

sured. When the length of the first hollow-

fiber membrane was 80 cm, much the same

peak height was obtained at the second hol-

low-fiber membrane lengths of 30 and 50

cm. At the second hollow-fiber length of 30

cm, the length of the first hollow-fiber mem-

brane was changed from 50 to 120 cm. Much

the same peak height was obtained at the

lengths of 80, 100, and 120 cm. Thus, the

lengths of the first and second hollow-fiber

membranes were fixed at 80 and 30 cm, re-

spectively.

Figures 4A and B show the effects of the

concentrations of sodium hydroxide and

mercury(I1) chloride on the uv absorbance of

the degradation products of penicillins, re-

spectively. The concentration of sodium hy-

droxide was varied from 1 to 5

M

at a mer-

cury(I1) chloride concentration of 3 X 1Op2M

(Fig. 4A). The maximum uv response was

obtained at a sodium hydroxide concentra-

tion of 4 M. When the concentration of so-

dium hydroxide was fixed at 4 M, the maxi-

mal uv absorbance was obtained at a mer-

cury(I1) chloride concentration of 3 X 1Oe2

M

and above (Fig. 4B). Figures 5A and B show

the effects of concentrations of sodium hy-

droxide and mercury(I1) chloride on the uv

absorbance of the degradation products of

ABPC and its metabolites, respectively. The

results obtained were similar to those of pen-

icillins except that there were almost no ef-

fects on the uv response of 3. Thus, the post-

column reaction conditions were selected as

described under Materials and Methods. The

optimum detection wavelength, which is ex-

amined by using the HPLC detector, was

285-305 nm for unchanged penicillins, 265

nm for 1 and 2, and 355 nm for 3. In a

previous paper (13) we suggested that 3

might be detected as the corresponding pen-

amaldate. Taking into account the absorp-

tion maximum of the degradation product(s)

of 3,3 should not be the corresponding pen-

amaldate but the other degradation prod-

uct(s) as suggested by Bundgaard and Lar-

sen (3 1).

The sodium hydroxide and mercury(I1)

chloride solutions were used for about 20 h

without loss of their activity. It is advisable to

immerse the second hollow-fiber membrane

in 0.1

M

nitric acid after use to avoid depres-

sion of membrane permeability. Neverthe-

less, when the peak height obtained was

lower, the inward and outward surfaces of

the membrane were thoroughly cleaned with

0.5 M nitric acid.

Comparison of Detection Method

Table 1 shows the data for band broaden-

ing (q) due to the column and the reactor.

This result reveals that the band broadening



I

I

0 1 2 3 4 5

Concentration of NaOH (MI

/

,/’

,’

,’

d’

0 i 2 3 4

Concentration of HgCt* (x16 )

FIG.

4. Ef fect of the concentrations of sodium hydroxide (A) and mercury(I1) chloride (B) on the uv

response of the degradation products of penicillins. The concentrations of mercury(B) chloride (A) and

sodium hydroxide (B) were kept constant at 3 X lo-’ and 4 M, respectively. A 20-~1 portion of a mixture of

ACPC (10 &ml), PCG (10 &ml), PCV (10 &ml), and PEPC (20 pg/ml) was loaded onto the column.

Detection was performed at 290 nm and at sensitiv ity o f 0.064 AUFS. Key is as follows: a, ACPC; b, PCG;

c, PCV; d, PEPC. Other conditions are given in the text .

A

B

i i i i

& 0

Concentration of NaOH (MI

I

1

2 3

4

Concentration of HgCll (x 16 )

FIG. 5. Ef fect of the concentrations of sodium hydroxide (A) and mercury(B) chloride (B) on the uv

response of the degradation products of ABPC and its metabolites. The concentrations of mercury(B)

chloride (A) and sodium hydroxide (B) were kept constant at 3 X 10m2and 4 M, respectively. A 20-~1

portion of a mixture of ABPC (20 &ml), 1, (20 pg/ml), 2 (20 &ml), and 3 (10 &ml) solutions was

loaded onto the column. Detection was performed at 290 nm and at sensitivity of 0.064 AUP?% Key is as

follows: a, ABPC; b, 1; c, 2; d, 3. Other conditions are given in the text.

136



CHROMATGGRAPHIC DETERMINATION OF PENICILLINS

137

TABLE 1

BAND BROADENING IN THE POSTCOLUMN F&ACTOR~

4)

Reactor Total

Peni- Method Method Method Method

cillin Column” B C B C

ABPC 4.2 8.9 2.4 9.8 4.8

ACPC 4.8 9.2 2.2 10 .4 5.3

PCG 5.4 9.0 2.1 10.5 6.0

PCV 8.4 9.6 1.6 12.7 8.5

PEPC 9.9 8.7 0.8 13.1 9.9

‘Open-tubular and hollow-fiber postcolum n reactors were

used in methods B and C, respectively.

‘Band broadening due to column plus injector, connector,

and detector.

‘The peak at short retention time was used for estimating the

band broadening.

of the hollow-fiber membrane reactor

(method C) is much less than that of the

conventional postcolumn reactor (method

A 0

B). Figures 6A, B, and C shows the chro-

matograms of penicillins followed by three

different detection methods (methods A, B,

and C). Penicillins were more sensitively de-

tected at 290 nm followed by the postcolumn

reaction (methods B and C), compared with

a native uv detection mode (method A);

method C gave a 3.3 to 7.1 times higher re-

sponse than method A. Peak heights of peni-

cillins obtained in method C were 1.6 to 1.9

times higher than in method B, and the reso-

lution between ACPC and PCG (peaks 2 and

3 in Fig. 6) was 0.45 and 0.75 in methods B

and C, respectively. These results are due to

the fact that the band broadening of method

C is much less than that of method B as de-

scribed above (in method C there is almost

no dilution due to the mixing of eluant and

reagent) and that the postcolumn reaction

conditions are independently examined with

the concentrations of sodium hydroxide and

mercury(I1) chloride in method C. In addi-

C

I I

I

1

I

1

0 10

0

10 0

10

Time (mln) Time (mid Time (mln)

FIG. 6. Comparison of the three detection methods for penicillins. (A) Detection at 230 nm without a

postcolumn reactor (method A); (B) detection at 290 nm with an open-tubular postcolumn reactor

(method B); (C) detection at 290 nm with sulfonated hollow-fiber membrane reactors (method C). A 20-~1

portion of a mixture of ABPC ( 10 &ml), ACPC ( 10 &ml), PCG ( 10 &ml), PCV ( 10 &ml), and PEPC

(20 &ml) was loaded onto the column. Sensitivity: 0.0 16 AUFS. Peak assignments: 1, ABPC; 2, ACPC; 3,

PCG, 4, PCV, 5, (lOR)- and (lOS)-epimers of PEPC. Other conditions are given in the text .



138 HAGINAKA AND WAKAI

tion, in method C the baseline noise and drift

were less than in method B.

Reproducibility, Linearity, and Detection

Limits

Table 2 lists within- and between-run pre-

cisions (relative standard deviation (RSD))

for measured peak heights of penicillins. The

results in Table 2 reveal good reproducibility

for all penicillins. Peak heights were found to

be scattered at random around a mean value;

that is, no trends (constant decrease in peak

height with time) were observed. This reveals

that the optimal concentrations of hydroxide

and mercury(I1) ions are maintained in spite

of continuous depletion of sodium and mer-

cury(I1) ions. The calibration graphs con-

structed by peak height versus concentration

for each penicillin were linear in the concen-

tration ranges 0.05-10 and lo-500 pg/ml

with a correlation coefficient of 0.999 or

above and passed through the origin. The

detection limits of the proposed method were

l-5 ng at a signal-to-noise ratio of 3.

Application to the Determination of ABPC

and Its Metabolites in Serum and Urine

On the basis of above findings, we at-

tempted to apply the present method to the

determination of ABPC and its metabolites

in serum and urine. Figures 7 and 8 show the

TABLE 2

PRECISIONOF THE ASSAY OF PENICILLINS~

Penicillin

Within-runb Between-runC

(%) @)

ABPC 0.79 1.90

ACPC 1.62 3.45

PCG 1.21 2.46

PCV 0.24 1.19

PEPC 2.39 4.13

a The concentration of each penicillin was 5.0 &ml.

b Relative standard deviation of five analyses.

’ Relative standard deviation of three analyses.

A

-

0

-

1

I

\

1

0 5 10

Time (mid

I

I

0

5

10

Time (mid

FIG. 7. Separation of ABPC from the background

components of serum. A 20-~1 portion of the ultrafiltrate

of serum samples was loaded onto the column. (A) De-

tection at 230 nm without a postcolumn reactor; (B)

detection at 290 nm with sulfonated hollow-fiber mem-

brane reactors. Peak 1 is ABPC. Concentration: 5.0

pg/ml. Sensit ivity: 0.016 AUFS. Other conditions are

given in the text.

separation of ABPC and of ABPC and its

metabolites from the background compo-

nents of serum and urine, respectively: Figs.

7A and 8A, detection at 230 nm without a

postcolumn reactor; Figs. 7B and 8B, detec-

tion at 290 nm (or 265 nm) with hollow-fiber

membrane reactors. ABPC and its metabo-

lites were detected at 290 nm (or 265 nm)

following the postcolumn reaction 2.5 to 4

times more sensitively compared with detec-

tion at 230 nm. At an ABPC concentration

of 2 Kg/ml in serum samples, RSD was

2.57%

(n

= 15), and at ABPC, 1, 2, and 3

concentrations of 5 pg/ml in urine samples,

RDSs were 2.01, 3.12, 3.87, and 0.83%

(n

= 15), respectively. The calibration graphs

constructed by peak height versus concen-

tration for ABPC and its metabolites in

serum and urine were linear in the concen-



CHROMATOGRAPHIC DETERMINATION OF PENICILLINS

139

A

i

B

i

r

I 1

I

I I

0

10 20

30

0

10 20 30

Time (mid

Time (mid

FIG. 8. Separation of ABPC and its metabolites from the background components of urine. A 20-~1

portion of the filtrate of urine samples was loaded onto the column. (A) Detection at 230 nm without a

postcolumn reactor; (B) detection at 265 (O-22 min) and 290 nm (after 22 min) with sulfonated hollow-

fiber membrane reactors. Peak assignments: 1, ABPC 2,1; 3, 2; 4,3. Concentrations: ABPC, 1, and 2,20

pg/ml; 3, 10 &ml. Sensitivity: 0.016 AUFS. Other conditions are given in the text .

tration ranges 0.5-100 pg/ml with a correla-

tion coefficient of 0.999 or above and passed

through the origin. The limits of accurate de-

termination were 0.1 pg/ml for ABPC in

serum samples with a 204 injection, 0.25

Fg/ml for ABPC, 1 O pg/ml for 1 and 2, and

0.5 pg/ml for 3 in neat urine samples.

The proposed method will be applicable

for the pharmacokinetic studies of ABPC

and its metabolites after therapeutic dose

and be used for their determination in bile,

cerebrospinal fluid, and tissues with a slight

modification to the chromatographic condi-

tions.

ACKNOWLEDGMENTS

The authors are grateful to T. Uno and H. Yasuda,

Mukogawa Women’s University, for their interest and

support. Thanks are also due to R. Sakurai for her tech-

nical assistance.

REFERENCES

1. Lam, S., and Grushka, E. (1978) J. Liq. Chroma-

togr. 1, 33-41.

2. Rogers, M. E., Adlard, M. W., Saunders, G., and

Holt, G. (1983) J. Liq. Chromatogr. 6,

2019-2031.

3. Miyazaki, K., Ohtani, K., Sunada, K., and Arita, T.

(1983) J. Chromatogr. 276,478-482.

4. Rogers, M. E., Adlard. M. W., Saunders, G., and

Holt, G. (1984) J. Chromatogr. 297, 385-39 1.

5. Haginaka, J., and Wakai, J. (1985) Analyst (Lon-

don) 110, 1185-l 188.

6. Haginaka, J., and Wakai, J. (1985) Analyst (Lon-

don) 110, 1277-1281.

7. Lee, T. L.. D’arconte, L., and Brooks, M. A. (1979)

J. Pharm. Sci. 68,454-458.

8. Westerlund, D., Carlqvist, J ., and Theodorsen, A.

(1979) Acta Pharm. Suet.

16,

187-214.

9. Carlqvist, J ., and Westerlund, D. (1979) J. Chroma-

togr.

164, 373-381.

10. Rogers, M. E., Adlard, M. W., Saunders, G., and

Holt, G. (1983) J. Chromatogr. 257,91-100.



140

HAGINAKA AND WAKAI

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

Buchbcrger, W., Winsauer, K., and Nachtmann, F.

suura, T., and Nanba, A. (1982) J. Chromatogr.

(1983) Fresenius’Z. Anal. Chern. 315,525-527.

239,537-548.

Haginaka, J., and Wakai, J. (1985) Anal. Chem. 57,

22. Davis, J . C., and Peterson, D. P. (1985) Anal. Chem.

1568-1571.

57,768-77 1.

Haginaka, J., and Wakai, J. (1987) J. Pharm. Phar-

macol. 39,5-g.

Kok, W. Th., Halvax, J. J., Voogt, W. H., Brink-

man, U. A. Th., and Frei, R. W. (1985) Anal.

Chem. 57,2580-2583.

Carlqvist, J., and Westerlund, D. (1985) J. Chroma-

togr. 344,285-296.

23. Haginaka, J., Wakai, J., and Yasuda, H. (1986)

Chem. Pharm. Bull. 34, 1850-1852.

24. Haginaka, J., Wakai, J., and Yasuda, H. (1987)

Anal. Chem. 59,324-327.

25. Haginaka, J., and Wakai, J. (1987) J. Chromatogr.

39Ik421-428.

Selavka, C. M., Krull, I. S., and Bratin, K. (1986) J.

Pharm. Biomed. Anal. 4,83-93.

Haginaka, J., and Wakai, J. (1986) Anal. Chem. 58,

1896-1898.

26. Haginaka, J., and Wakai, J. (1987) J. Chromatogr.

3%,297-305.

27. Hwang, H., and Dasgupta, P. K. (1986) Anal. Chem.

58, 1521-1524.

Haginaka, J., and Wakai, J. (1987) J. Chromatogr.

413,219-226.

Stevens, T. S., Davis, J. C., and Small, H. (1981)

Anal. Chem. 53, 1488-1492.

Stevens, T. S., Jewett, G. L ., and Bredeweg, R. A.

(1982) Anal. Chem. 54,1206-1208.

Hanaoka, Y. , Murayama, T., Muramoto, S., Mat-

28. Dasgupta, P. K., and Yang, H. C. (1986) Anal.

Chem. S&2839-2844.

29. Bird, A. E., Cutmore, E. A., Jennings, K. R., and

Marshall, A. C. (1983) J. Pharm. Pharmacol. 35,

138-143.

30. Bundgaard, H., and Larsen, C. (1979) Int. J. Pharm.

3, l-l 1.

31. Bundgaard, H., and Larsen, C. (1983) J. Pharm.

Biomed. Anal.& 29-37.

Documents

Liquid Chromatographic Determination of Penicillins by Postcolumn