CLIN. CHEM.

36/1,

113-115 (1990)

Non-Extraction HPLC Method for Simultaneous Measurement of Dyphylline and Doxofylline in Serum
Franco

Tagliaro,RomoloDorlzzl,2 AlbertoFrIgerlo,3 and MarioMango1

and HPLC grade water were purchased from Carlo Erba, Milano, Italy. Serum and plasma were used either fresh or after storage at -20 #{176}C. For sample clean-up we used 0.45-pPm (pore size) disposable ifiters purchased from Millipore, Bedford, MA. Apparatus. The isocratic HPLC system we used was composed of a Jasco (Tokyo, Japan) Model 880-PU HPLC pump, a 7125 Rheodyne (Cotati, CA) irjector equipped with a 10-giL loop, a Jasco Model 875 ultraviolet detector, and a Model 3390 A integrator (Hewlett-Packard, Palo Alto, CA). We used an internal surface reversed phase Pinkerton column (150 x 4.6 mm) from Regis Chemical Co., Morton Grove, IL, packed with 5-.an silica particles derivatized with glycine-phenylalanine-phenylalanine. A Type 73XX column inlet filter (0.5-pm pore size, 3 mm diameter; Rheodyne) was connected between the injector and the
column. Chromatographic conditions and sample preparation. We used a mobile phase of phosphate buffer (0.1 mol/L, pH 6.8)

A direct injection HPLC method for the simultaneous measurement of dyphylline and doxofylline in serum is reported.

Chromatography is carried out with a Pinkerton internal surface reversed-phase column and phosphate buffer (0.1 moi/L, pH 6.8) as mobile phase. Absorbance at 275 nm is
monitored. Only 10 L of serum is required to detect less than 1 mg of each drug per liter in biological samples. The standard curve for the method is linear over the range 6-100 mg/L, and precision is acceptable for both drugs, with CVs of
2% and 1.5% for respective concentrations of 6.2 and 25

mg/L. None of the 76 drugs tested for interference affected the measurement of either drug. Four samples can be injected per hour. This method provides a fast, accurate way to monitor two interesting therapeutic agents used in chronic
obstructive airway diseases.

AdditIonal Keyphrasea: dimethyixanthines . chromatography, reversed-phase . drug assay Theophylline (1,3-dimethylxanthine), especially in sustained-release preparations, is widely prescribed in the therapy of asthma (1). Among the theophylline analogs recently introduced for such therapy are dihydroxypropyl theophylline (dyphylline) and 2-(7-theophyllinemethyl)1,3-dioxolane (doxoflline), which are claimed to retain the therapeutic properties of theophylline but have a lower incidence of side effects (2-4). However, they have encountered limited success, mainly because of the difficulty of adequately monitoring their concentrations in patients sera. Indeed, no immunoassays are available for determination of these drugs, and chromatographic methods are too complex for routine use in clinical laboratories (5-7). In this paper we describe a very simple HPLC method, not requiring sample treatment, for simultaneously determining dyphylline and doxorlline. The method is comparable with immunoassays in terms of operating simplicity, while retaining the precision, accuracy, and ruggedness of chromatographic methods.

at a flow rate of 0.3 mL/min, at room temperature. The ultraviolet detector was set at 275 nm with a full-scale deflection of0.08A. To reduce the risk of column clogging, we periodically reversed the column according to the manufacturers recommendations. Fresh samples (either serum or plasma) were filtered through 0.45-pm (pore size) disposable filters and injected. Samples that had been frozen were centrifuged before ifitration.

Results
Under the described conditions, dyphylline and doxofylline were eluted in quite symmetrical peaks and were well resolved from each other and from the peaks for theophylline and caffeine (Figure 1). Nevertheless, the efficiency of the column apparently depended on flow rate in the range 0.1-1 mL/min, suggesting problems with the mass transfer as reported by Cook and Pinkerton (8). Thus, we used a relatively low flow rate, 0.3 mU/mm, which still assures acceptable retention times. In practice, samples could be injected every 15 min. Retention times for dyphylline and doxofylline were 8.43 and 13.94 mm, respectively. Coefficients of variation (CVs) were respectively 2.6% and 4.0% during the same day (n = 5) and 9.6% and 9.2% for different days (n = 5). The linearity of response was calculated over the range 6-100 mg/L. The calibration curve data, obtained with serum with added dyphylline and doxofylline, were analyzed by a least-squares regression method, giving the following equations: y = 20752x 6565 andy = 17615x 17707 for dyphylline and doxofylline, respectively, where x is the concentration in mgfL and y is the peak area in arbitrary units. For both drugs a close parallelism was observed between calibration curves prepared from samples in serum and in buffer (Figure 2). Because no extraction procedures requiring the monitoring of the recovery were involved, we used external standardization, interpolating the concentrations of unknown
-

Materials and Methods

Reagents. Theophylline, caffeine, and dyphylline were purchased from Sigma Chemical Co., St. Louis, MO. Doxofylline was a kind gift of ABC S.p.A., Torino, Italy. Standards of therapeutic and abused drugs, supplied desiccated on glass microfiber disks impregnated with silicic acid (Toxi-Disc Library), were purchased from Analytical Systems, Laguna Hills, CA, and used to investigate possible interferents. Analytical grade chemicals and solvents Institute 2Cth Italy. of Forensic Medicine, University of Verona, Italy. Chemistry Laboratory, Hospital of Legnago (VR),

31t.alian Group for Scientific Research and Study, Milano, Italy. Presented in part at the American Association for Clinical

Chemistry, 41st national meeting, Atlanta, GA, July 1989. Received April 24, 1989; accepted September 1, 1989.

CLINICALCHEMISTRY, Vol. 36, No. 1, 1990 113

samples
fl,

from

a plot of results

for standards

vs their

concentrations.

The reproducibility (CV) of results, evaluated at two concentrations (6.2 and 25 mg/L, five replicates each) was 2.05% and 1.5% for dyphylline and 2.10% and 1.5% for
doxofylline.

Fig. 1. Chromatograms of (nghl) a human serum sample supplemented with dyphylline(whichwas elutedat 8.41mm),theophylline (10.88 mm), caffeine (13.03 mm), and doxofylline(13.93 mm) (6.2 mg/L of each drug), and (left) blank serum (same conditions, except flow rate = 1 mL/mln)

um
300-

The classical definition of minimum detectable quantity-i.e., twicethe instrumentalnoise-does not fit with our chromatographic system when real samples are injected. In this case the main factor limiting sensitivity is the negative drift of the baseline, owing to the tail of the front peak caused by endogenous components. This hampers the use of sensitivity ranges lower than 0.08 A. Under these conditions instrumental noise is not measurable. The limit of detection (defined as the least concentration of drug giving a peak surely distinguishable from other peaks of the matrix) was less than 1 mgfL for each drug, whether in serum or in plasma. Table 1 lists more than 70 drugs that, in a concentration of 20 mg/L, did not interfere in the HPLC determination of dyphylline and doxofylline (i.e., no spurious peaks were detectable at the retention times of dyphylline and doxofylline).

Dyphylline

$u,um
7060 5040 30

Doxofyllina

DiscussIon
Manufacturers of drug assays primarily have focused on developing immunoassays, especially for drugs with a large market. New or not widely prescribed drugs are often ignored. On the other hand, clinicians are reluctant to use drugs that cannot be monitored. Therefore, clinical chemists are often left to develop in-house methods to measure specific therapeutic agents. We prefer the use of directiijection HPLC methods [as recently reviewed (9, 10)] for such purposes. The method we propose shows precision, accuracy, and

200-

100 20-

y.1.158.0.455

0-

1O

200 0 10 20 30 40 50 60

Fig. 2. Comparison between peak heights (mm) of dyphylline and doxofyllmne in serum and in buffer

Table 1. Drugs InvestIgated and Found Not to interfere wIth the IIPLC at 20 mg/L
Opiatesand antagonists Codeine
Mepenidine Terpin hydrate

Dextromethorphan Methadone

Diphenoxilate Papaverine

Hydromorphone Propoxyphene

Central nervous system active drugs Amphetamine Amitnptyline Caffeine Chlorprothixene Phenytoin Diazepam Loxapine Imipramine Methyiphenidate Methaqualone
Phenmetrazine
Phentermine

Benztropine
Chlopromazine Doxepin

Carbamazepine
Cocaine Flurazepam

Meprobamate
Nordlazepam

Prazepam
Thiothixene

Protriptyline Trifluperazine

Phencyclidine Strychnine Triftupromazine Benzoylecgonine

Diphenhydramine Erythromycin Udocaine Orphenadrine

Methamphetamine Nortrlptyline Thiondazine

Miscellaneous
Acetaminophen Chlorpheniramine

Atropine
Cimetidine Emetine Hydroxyzine

Carisoprodol
Disopyramide

Doxylamine
Hydrocortisone

Glutethimide Methapyrilene
Pentazocine Phenylpropanolamine Pseudoephedrine

Methocarbamol Phenacetin Propranolol

Quinine

Nicotine Pyrilamine Procaine

Salicylamide Tnimeprazine

Phenolphthalein
Procainamide Spironolactone Tnmetobenzamide

Triexyphenidyl
Tnpelennamine

Tniamterene Trimethoprim

114 CLINICAL CHEMISTRY, Vol.36, No. 1, 1990

sensitivity comparable with other HPLC methods for methyixanthines (11, 12) and is suitable for both therapeutic drug monitoring and pharmacokinetics studies. Because the specificity of the assay, in the absence of any extraction procedure, relies on the chromatographic separation, we have given special care to examining the possibility of co-eluting peaks. Theophylline and caffeine are well resolved from dyphylline and doxofylline, and none of 76 drugs we examined (Table 1) has shown significant interferences. However, paracetamol (acetaminophen) partly overlaps the theophylline peak. According to Dawson et al. (13), use of a 250-mm-long column should resolve the problem. It has been suggested that strong protein binding theeretically could hamper the use of direct-injection HPLC methods because of disturbances of the absorption equilibria of the analytes (9). The parallelism between the calibration curve for both the drugs, in buffer and in serum, suggests that they are not so strongly linked to proteins as to interfere with separation on this column. We saw no significant decrement in column performance or pressure increase after more than 500 injections of samples. We consider this direct-injection HPLC method a simple, quick, accurate, and inexpensive way to measure methylxanthines of therapeutic interest. References 1. Uden DL, Wyatt RA, Zaske DE. Theophylline therapy. Postgrad Med 1984;75:247-56. 2 Simons FER, Simons KJ, Bierman CW. The pharmacokinetics of dihydroxypropyltheophylline: a basis for rational therapy. J

Allergy Clin Immunol 1975;56:347-55. 3. Simons KJ, Simons FER, Bierman CW. Bioavailability of a sustained-release dyphilline formulation. J Clin Pharmacol 1977;17:237-42. 4. Grossi E, Biffignandi P, Franzone JS. La doxoffihina: profflo del farmaco e rassegna degli studi clinici. Eur Rev Med Pharmacol Sci 1988;10:1-16. 5. Paterson N. High-performance liquid chromatographic method for the determination of diprophylline in human serum. J Chromatogr 1982;232:450-5. 6. Badini C, Masera F, Franzone JS. Dosaggio del 2-(7-teo fillinmetil)-1,3-diossolano mediante HPLC. Farmaco 1982;37:3204.

7. Kester MB, Saccar CL, Mansmann Jr HC. Microassay for the

simultaneous determination of theophyllme and dyphylline in serum by high-performance liquid chromatography. J Chromatogr

1987;416:91-7. 8. Cook SE, Pinkerton TC. Characterization of internal surface reversed-phase silica supports for liquid chromatography. J Chromatogr 1986;368:233-48. 9. Shihabi ZK. Review of drug analysis with direct serum injection on the HPLC column [Review]. J Liq Chromatogr 1988;11:1579-93. 10. Westerlund D. Direct injection of plasma into column liquid chromatographic systems [Review]. Chromatographia 1987;24:155-64. 11. Broussard LA. Theophylline determination by high-pressure liquid chromatography. Clin Chem 1981;27:1931-3. 12. Butrimovitz GP, Raisys VA. An improved micromethod for theophylline determination by reversed-phase liquid chromatography. Clin Chem 1979;25:1461-4. 13. Dawson CM, Wang TWM, Rainbow SJ, Tickner TR. A nonextraction HPLC method for the simultaneous determination of serum paracetamol and salicylate. Ann Clin Biochem 1988;25:661-7.

CLIN. CHEM.36/1, 115-118(1990)

Three Commercial Polyclonal Immunoassays for Cyclosporine in Whole Blood Compared: 1. Results with Patients Specimens
MIchaelJ. Strassman, Gary L Lensmeyer, DonaldA. Wlebe,and Ian H. Carison We assessed the performance of three commercially available polyclonal immunoassays for apparent cyclosponne in 120 whole-blood specimens collected from transplant recipients just before their next dose of cyclospormne (C5A). The

tation of the analytical results, we assayed serial specimens from six post-transplant patients. These showed significantly dissimilar, but parallel, results among the methods for any
single sample. Occasionally, however, a result would not fit

assays were (a) Abbotts TDx fluorescent polarization immunoassay for GsA and its metabolites in whole blood; (b) the Sandoz radioimmunoassay (RIA); and (C) lncstars CycloTrac RIA. Mean respective CVs were 3.8%, 9.3%, and 24.3%. Analyticalrecovery was nearly 100% for concentrations up to 1000 ug/L for lncstar and up to 1500 pg/L for Abbott and Sandoz; linearity was compromised at greater concentrations. We also quantified the parent C5A concentrations by HPLC. Moreover, to follow day-to-day fluctuations in patients cyclosporine concentrations with each method and to assess the impact these differences have on interpre-

the established trend. Biases observed among the assays can be explained in part by the nonspecific antisera crossreacting with GsA metabolites. Most important, we demonstrate that patients results are not reliably interchangeable among the methods. Polyclonal radioimmunoassays (RIAs) and an assortment of in-house-developed HPLC procedures have formerly been the laboratorians only tools for therapeutic monitoring of cyclosporine (CsA). Unfortunately, these assays can be labor-intensive, time-consuming, and (in the case of RIA) present a health risk from radioactive tracers. Sandoz Pharmaceuticals and Incstar Corp. remain the commercial sources for RIAs, and both companies now offer a more-specific monoclonal RIA. The monoclonal assays purportedly measure only the parent drug, with negligible CLINICAL CHEMISTRY, Vol. 36, No. 1, 1990 115

Department of Pathology and Laboratory Medicine, and the Clinical Laboratories, University of Wisconsin Hospital & Clinics, 600 Highland Ave., Madison, WI 53792. Received June 27, 1989; accepted September 11, 1989.