Evauation of Short and Efficient Reversed-Phase Columns for the Fast Analysis of Theophylline by HPLC

Chromato gra ph y , the Stat e of the Ar t H. Kala sz and L.S. Ettre (Ed s)

EVALUATION OF SHORT AND EFFICIENT REVERSED-PHASE COLUMNS FOR THE FAST ANALYSIS OF THEOPHYLLINE BY HPLC ROY EKSTEEN a nd JERRY J. THOMA*

Supelco Inc., Supelco Park, Bellefonte, PA 16823, USA *Sout h Be nd Medical Foundat i on, 530 North Lafayette Boulevard, South Bend, I N 46601, USA

ABSTRACT

Separation of theophylline from theobromine, 1, 7-dimethyl­xa nth ine , S -hydroxyethyltheophylline /internal standard/ and caffein~was accomplished within three minutes using a 5cm su­PELCOS I L LC-18 column and a ternary mob i l e phase consisting of 1% acetonitrile, l % tetrahydrofuran and 98% 1.75 mH H P04 . Moderate fl ow rate and low co lumn back-pr essure contribu~ed to long co lumn lifet i me , reduc e d solvent consumpt ion a nd decreased cost per ana l ys i s. A sil ica column extr action technique was used f or cleanup of serum sampl es. Fifty-four drugs were inves­tigated with respect to chromatographic retention and/or e xtrac­tabi lity. Chlorothiazide and metronidazole i nterfered with the internal s t a ndard a~d theophylline, respectively . Levels o f 1,7-dimethylxanthine were genera lly below 1 pg /ml, and i nde­pendent o f the caffeine concentration.

INTRODUCTION

The deve l opment a nd commercial availability of second-gene­

rat ion HPLC columns have resulted in the widespread application

of high-performance liquid chromatography /HPLC / in therapeut ic

drug monitoring /TD!1 /. Efficiences of 18 1 000 the oretical plates

fo r a 25 em co l umn, and 10,000 for a 15 em column packed with

spherical 5 pm /bonded phase/ si l ica particles can be obt ained

under the manufacturer 's test condit ions. Although t hese effi­

c i encies are generally not real ized in "real " samples, it is

apparent that, in many TDM analyses, only 1,500 to 3,000 theo­

retical plates are requ i red for separation. Und e r optimum con­

ditions, 5 em columns can genera t e abo ut 3 1 500 theor etical

51

Proceedings of the Budapest Chromatography Symposium, June 1-3, 1983. Published in Volume 1, Chromatography the State of the Art, Budapest, 1985. Edited by H. Kalasz and L. Ettre

plates. Consequently, instead of 25 or 15 em columns, a 5 em

column is sufficient for most TDt1 analyses. Examples of drugs

that are readily analyzed with these short columns include pro­

cainamide, quinidine, lidocaine, disopyramide, theophylline ,

acetaminophen, propranolol and chloramphenicol. For the analysis

of these drugs, 5 em columns can provide faster analysis time

and thus reduced cost per analysis. Additional advantages in­

clude a reduction in solvent consumption and operation of the

column at lower back pressure. Finally, short columns are less

expensive than conventional columns and are generally more

stable .

In this paper, we will discuss our experience with the use

of short columns for routine therapeutic drug monitoring of

theophylline and related xanth ines in a cl inical setting . Eva­

luation of 5 em SUPELCOSIL LC-18 columns for therapeutic moni­

toring of theophy lline was carried out at the Toxicology D~rt­

ment of the South Bend Medical Foundation over the past two

years.

~illTERIALS AND METHODS

Chromatography

HPLC analyses were performed with a Model SP-8000 liquid

chromatograph /Spectra Physics, San Jose, CA 95134/ equipped

with a Model 7010 injector /Valco, Houston, TX 77055/, a Mo­

d el SF-740 var iable wavelength UV detector /Kratos, Ramsey,

NJ 0744 6/, and with a system consisting of a Constametric III

pump /LDC, Riviera Beach, FL 33404/, a Model 7125 injector

/Rheodyne , Cotati, CA 94928/, a Model UV-10 variable wavelength

UV detector /Varian, Palo Alto, CA 94303/, and an Omniscribe

B-5000 recorder /Houston Instruments, Austin, TX 78753/. The

mobile phase was made up from HPLC-grade solvents by adding

10 ml acetonitrile or methanol and 10 ml tetrahydrofuran to

980 ml 1.75 ~M H3Po4 /pH= 2.8/. All columns /50 x 4.6 mm and

150 X 4.6 mm ID/ contained SUPELCOSIL~LC-18 /Supelco, Inc. ,

Bellefonte, PA 16823 / , spherical 5 pm silica bonded with octa-

52

decyl-dimethylchlorosilane and "end capped" wi th trimethylchlo­

rosilane.

Extraction

Serum /0.1 ml/ was extracted with the aid o f Clin-Elut

1000 M columns / Analytichem, Harbor City, CA 90710/ according

to the method described by Frethold /1 /. The evaporated sample

e xtra ct was reconstituted in 150 pl mobile phase of which

40 pl was injected. Quantitation of serum theophyl lin e le­

vel s was done by peak height using the internal s tandard ratio

method. The absolute recove ry of theophylline was 78 % over

the entire therapeutic range.

RESULTS AND DISCUSSION

One of the main advantages o f HPLC in TDM is its ability

t o separate the drug of interest from its metabolites and

related chemicals . Since t heophylline b e longs to the same che­

mical class as caffe i ne and theobromine /common in ma ny bever­

ages / , the analysis must allow for adequate separat ion o f c af ­

feine, theophylline, theobr omine, and an appropriate internal

standard. Metaboli sm of caffeine can also lead to other poten­

tial interference such as those found with the caffeine meta­

bolite 1,7-dimethy lxanthine /paraxanthine/ . Finally, potential

inte rferences from medications often prescribed concurrently

with theophyll ine, such as antibi o tics, diuretics , ana l ges ics,

and phenobarbital, must be el i minated.

As other a uthors /2, 3 / have reported, we found i t neces­

sary to mix a small percentage of tetrahydrofuran with a ceto­

nitrile or methanol to af fect separation of paraxanthine from

theophilline. A quaternary mobile phase consisting of aceto­

nitrile, methanol, tetrahydrofuran and buffer did not lead to

fur ther improvements . A standard mixture of theobromine ,acet­

aminophen, paraxan~hine, theophylline , diphylline, S - hydroxy­

ethyltheophylline and caffeine was injected on a 5 em and 15

em SUPELCOSIL LC-18 column {see Fig. 1/. At a flow rate of

53

~

Figure 1.

A 3 B 3 4 1 4

2 56 7 5

~ 2

I I I I I I I I 0 1 2 3 0 1 2 3 ~in. ~in .

Theophylline analysis on 5 and 15 em LC-18

columns.

Mobile Phase: 1 % tetrahydrofuran, 1 % acetonit­

rile, 98 % 1.75 mM H3Po4 ; Temperature: 3o0 c;

Detection: UV at 274 nm; Injection: 10 pl of

each compound at 10 pg/ml in mobile phase;

Sample: theobromine /1/, acetaminophen /2/,

paraxanthine /3/, theophylline /4/, diphylline

/5/, 6 -OH-ethyltheophylline /6/, caffeine /7/.

~ ~ SUPELCOSIL LC-18, 50 x 4.6 mm; Flow Rate:

2 ml/minute; Sensitivity: 0.2 AUFS; Pressure:

740 psi.

A ~ SUPELCOSIL LC-18, 150 x 4.6 mm; Flow Rate:

6 ml/minute; Sensitivit y: 0.1 AUFS, Pressure:

4800 psi.

2 ml minute, caffeine elutes from the 5 em column within three

minutes. Although the same mixture was better resolved on a

15 em column, excessive pressure at a flow rate of 6 ml/minute

was necessary ·to obtain the same analysis time as on the 5 em

column. Thus, for fast analysis of a rel?tively small number of

compounds, one benefits from the use of short columns in terms

54

n

\\A 1\,.s B

I 2

3

I

Figure 2. Retention behaviour before and after wetting.

SUPELCOSIL LC-18, 50 x 4.6,mm; Mobile Phase: 1 %

tetrahydrofuran, 1% methanol, 98 % 1.76 roM H3

Po4

;

Flow Rate: 2 ml/minute; Pressure: 815 psi; Tem­

perature: 30°C; Detection: UV at 254 nm; Sensitivi­

ty: 0 . 4 AUFS; Injection: 10 pl of each compound

at 0.2 mg/ml in mobile phase; Sample: theobromine

/1/, paraxanthine /2/, theophylline /3/, S-OH­

ethyltheophylline /4/, caffeine /5/.

A. Dry column.

B. After wetting with 100 pl methanol /see text/ .

of lower pressure drop, solvent consumption, and price. The ad­

ditional advantage of increased sensitivity for a fixed sample

volume is of minor importance for this particular analysis, al­

though Figure 1 illustrates a two-fold gain for the 5 em column

when using a constant injection volume.

Some minor precautions must be taken to assure satisfactory

results on a 5 em column. Since the 5 em column is 3 to 5 times

55

shorter than the 15 and 25 ern columns, i t dries out quicker

during shipping and storage. Therefore, the column should be

was hed with l rnl methanol b e fore equilibrating with the theo­

phylline mobile phase /l % acetonitrile, l % t e trahydrofuran ,

98 % 1.75 mM H3P0 4 /. This procedure assure s that the surface

of the Cl8 reversed-phase s i lica is properly wetted. No reten­

t ion of the anal y t e s was observed when a dry column was equi ­

librated directl y with the ternary mobile phase, as shown i n

Figure 2A.A 100 pl injection of methanol can be used as an

alternative to t h e l rnl methanol wash. Once properly wetted,

a reversed phase column can be operated at any desired p ercent­

age of organic modifier . Note that the described wetting pro­

cedure is similar to that used with reverse d phase type sample

cleanup columns and hydrophobic membrane filters.

With solid phase extraction for sample preparation , at

least 1 7000 serum extracts could be injected without the use

of a guard column before reduction of column efficiency warrants

a co lumn c hange. Since the column p acking does not show a de­

pression during its us e ful lifetime, we believe that contamina­

tions from the sample and the sampling valve are the cause of

gradual degradation of resolution. Althoug h not restricted to

the use of short columns, we noticed ~ occasional loss of ef ­

ficiency due to part ial blockage of the column inlet frit . This

blockage does not necessarily give rise to increasing column

pressure. In such cases, we were able to restore the original

sepa ration after reversing the flow through the column.

During this operation, the column is disconnected from the

detector to prevent particulates from entering the detector

c e ll. It is s u fficient to operate the column in this rever s ed

flow direction for only a short time at a high flow rate. The

anal y siscan then be continued with the flow going in either

direction. Protection of the analytical 5 ern column by a 2 ern

disposable guard column packed with either pellicular /40 pm/

or porous /5 prn/ Cl8 bonded silica r esulted in less instrument

downtime and extended column life. Using a 2 ern disposable pel­

licular guard column /a total of three were actually used/ more

than 3,ooo serum extracts were injected without serious loss of

resolution.

56

TABLE I. Extractable compounds with HPLC retention times less

than 5 minutes

Compound

Acetaminophen B-OH-ethyltheophylline

Caffeine ~ Chlorothiazide /Diuril/~ Diphylline Hydrochloroth~azide r1etronidazole3 Paraxanthine Theobromine Theophylline

Retention Time in ~1inutes

1.2 2.0 2.8 2.0 1.8 2.4 1.2 1.5 1.0 1.7

Serum 1 Concentration in p.g/ml

100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

1 Using a sample volume of 100 p.l of serum in the extraction procedure, 10 pg/ml of compound in mobile phase would be equivalent to 100 pg/ml of compound in the extracted serum, assuming 100 % recovery.

2 Coelutes with S-OH-ethyl theophylline,

3 Tailing edge coelutes with theophylline.

Clinically useful and cost effective monitoring of theo~yl­

line requires rapid sample preparation, adequate separation of

analytes and interferences, and short chromatographic analysis

time. Theophylline sample preparation for HPLC ranges from di­

rect injection of serum after precipitation of protein by ace­

tonitrile to complex extraction procedures. A multitude of pub­

lished methods have appeared over the past three years /4-25/.

Precipitation and extraction methods each have inherent advant­

ages. The authors feel that the elimination of nonextracted in­

terferences justifies the additional time required to prepare

samples using a solid phase or solvent extraction procedure.

Therefore, in the study presented here, a solid phase adsorp­

tion extraction system was used to prepare all samples /1/.

A variety of compounds were tested for potential interfe­

rences and are listed in Tables I and II. The retention times

57

TABLE II. compounds with retention times greater than 10 mi n ­

utes1

Amount Arrtount Injected Injected

Compound I ,ug I in Compound I pg I in Mobile Mobile Phase Phase

a-Phenylglutarimide 12.5 Mesantoin 12.5 Amobarbital 12 .5 Methaqualone 12.5 Barbital 12.5 Metharbital 12.5 Butabarbital 12.5 Methsuximide 12.5 Butalbital 12.5 4-Methylprimidone 12.5 Carbamazepine 12.5 Methylprylon 12.5 Cefamandole 20.0 Motrin 20.0 Cephalothin 20.0 Nirvanol 25.0 Cephapirin2 20.0 Nordiazepam 12.5 Chloramphenicol 20.0 Oxazepam 6 . 25 Chlordiazepoxide 12.5 Papaverine 20.0 Cimetidin 20.0 Pentobarbita l 12 . 5 N-Desmethsuximide 25.0 Phenacetin 12.5 Diallybarbiturid Acid 12 .5 Phenobarbital 20 . 0 Diazepam 12.5 Phensuximide 12.5 Dilantin 17.5 Phenylbutazone 12.5 Diphenylhydantoin 20.0 PEMA 15.0 Ethinamate 12.5 Primidone 12.5 Ethosuximide 50.0 Secobarbital 12.5 Ethotoin 12.5 Sulfamethoxazole 20.0 Furosimide 20.0 Sulfisoxazole 20.0 Glutethimide 12. 5 Talbutal 12.5 Lidocaine 12.5 Tobramycin 20.0 Mephobarbital 12.5 Trimethoprim 20 . 0 Meprobamate 18.75

1 Since these drugs did not interfere with the chromatography,

we did not test them i n our extraction procedure.

2 Cephapirin has a retention time of 3.5 minutes but did

not extract at a serum concentration of 100 pg/ml in the

procedure outlined.

of those drugs that are extracted and have chromatographic re­

tention times of less than 5 minutes /under the conditions of

Figure 1/ are given in Table I. Other drugs that showed reten­

tion times longer than 10 minutes are listed in Table II to­

<Jether with the concentration /in mobile phase/ t hat was jn-

58

3

A 4

1

5

I I I I 0 1 2 3

Min.

Figure 3.

4 4

B c 3

3

1 2 1 2 5

I l I I I I 0 2 3 0 1 2

Min. Min .

Theophylline patient serum samples.

Column and conditions as in Figure lA except

for 40 pl sample volume. Peaks are labeled

as in Figure 1. Theophylline concentrations:

I 3

A = 11.0 pg/ml, B = 4.3 pg/ml, C = 8.3 pg/ml.

jected. These drugs were not taken through the extraction pro­

cedure. Cephapirin eluted in 3 . 5 minutes but did not extract.

The chromatograms from three patient serum samples containing

varying amount of theophylline, theobromine and paraxanthine

are shown in Figure 3. In our experience, the concentra tion of

paraxanthine rarely exceeds 1 pg/ml regardless of the concen­

tration of caffeine.

The analysis of theophylline by HPLC using short 5 em col­

umns, combined with a solid phase serum extraction procedur~,

has proven to be reliable and cost effective with a minimum of

interferences.

59

ACKNOWLEDGEHENT

The authors thank Ann Marie Litka, M.T. /ASCP/, South Bend

Medical Foundation, for collecting the data reported in the in­

terference study.

REFERENCES

14.

15.

P . J.

363.

D.A.

H.H.

Naish, M. /1979/

Asdeppa,

Farrish, S.D. Lipton, Clin . Chern. 26 788. /1980/

W.A. Wargin, Clin. Chern. 2 524. /1980/

16. N. Weidner, D.N. Dietzler , J.H. Ladenson , G. Kessler ,

L. Larson, C.H. Smith, T. Janes, J.M. McDonald , Am. J. Clin. Pathol. 7 79. /1980/

17. P.J. Naish, R.E. Chambers, M. Cooke . Ann. Clin. Biochem.

16 254. / 1979/

60

18. J.R. Miksic, B. Hodes, J. Pharrn. Sci. 68 1200. /1979/

19. G.P. Butrirnovitz, V.A. Raisys, C1in. Chern. ~ 1461 /1979/

20. L.T. Mann, Jr., Clin. Chern.~ 1336./1979/

21. L.W. Bond, D.L. Thornton, C1in . Chern. 25 1186. /1979/

22. C.A. Robinson, Jr., J. Dobbs, Clin. Chern. 24 2208. /19 78/ 23. K. Nakatsu, J . A. Owen, K.Scu11y, C1in. Biochern. 1 148

/1978/

24. C.A. Robinson, Jr., B. Mitchell, J. Vasi1iades, S.L. Sie­

sel, Clin. Chern. 24 1847. /1978/

25. F.L.S. Tse, D.W. Szeto , J. Chrornatogr. 226 231. /1981/

61