1386 CLINICALCHEMISTRY,Vol.24, No. 8, 1978

U Future Trends in Advanced Analytical Concepts

CLIN.CHEM.24/8, 1386-1392(1978)

Determination of Drugs in Plasma by High-Performance Thin-Layer

Chromatography

David C. Fenimore, Chester M. Davis, and Carol J. Meyer

High-performance thin-layer chromatography was usedto determine chiorpromazine, amitriptyiine, nortriptyiine,imipramine, desipramine, phenobarbital, and phenytoinin plasma, to demonstrate the utility of this technique forroutine analysis. We quantitated the separated componentsby use of ultraviolet reflectance spectrometry with de-tection limits as low as 1 zg/liter. Regressions of psy-choactive agents extracted from plasma were linear overthe range of 0 to 300 tg/liter. The anti-convulsant drugs,phenobarbital and phenytoin, were determined over arange of 0 to 50 mg/liter. Analyses were rapid, repro-ducible, and well-suited to large-scale programs. Sepa-rated components also can be identified in situ by ultra-violet spectrophotometry.

AddItIonal Keyphrases: drug assay . monitoring therapytoxicology

As evidence accumulates correlating concentrationsof psychopharmacologic agents in the blood withtreatment efficacy, there is increasing interest in ana-lytical methods which could be applied to large clinicalpopulations. Most methods currently available, how-ever, require techniques and instrumentation that areat a level of sophistication not always compatible withroutine, large-scale programs. A high order of sensitivityis usually necessary, because many of these compoundsare present therapeutically in concentrations in the lowig/liter range, and the situation may be further com-plicated by the presence of metabolites that are similarin structure to the parent compound and are a potentialsource of interference.

For these reasons, chromatographic separations usedwith sensitive detection systems have been predominantin assays of drugs in blood and other biological samples.Such techniques as electron-capture gas chromatogra-

Texas Research Institute of Mental Sciences, 1300 Moursund,Houston, Texas, 77030.

Received Mar. 29, 1978; accepted April 25, 1978.

phy (1-4), nitrogen-sensitive gas chromatography (5,6), mass fragmentography (7, 8) and, more recently,high-performance liquid chromatography (HPLC) withelectrochemical detection (9) all can provide the req-uisite sensitivity and selectivity.

The major disadvantage of these methods, especiallyfor routine clinical determinations, is the considerableinstrument time involved in sequential separationsnecessitated by column-chromatographic systems.Thin-layer chromatography (TLC), on the other hand,permits simultaneous separation of many samples.Drugs in blood have been so estimated, with densi-tometry for several years, but the problem of attainingstill-greater sensitivity has restricted this techniqueto drugs that are present in fairly high concentrations,such as the anticonvulsants (10), or to use of largesample volumes if the drug is at low concentrations, asis the case with chlorpromazine (11).

Recently introduced materials and methods for TLChave improved both resolution and sensitivity of de-tection to the point that the appellation “high-perfor-mance” may justifiably be used. Ripphahn and Halpaap(12, 13) described high-performance thin-layer chro-matography (HPTLC) and the characteristics ofHPTLC plates coated with silica gel of carefully con-trolled particle size distribution, which yielded sub-stantially better separations than other thin-layerplates. Kaiser discussed means by which samples couldbe applied to the plate in HPTLC (14), and Hezel re-ported on quantitative measurements performed withHPTLC (15). This method was used by Ritter in thedetermination of the diuretic, muzolamine, in plasmaand urine (16), and Fenimore et al. described the ap-plication of HPTLC to antipsychotic drugs (17).

This present report, a continuation of the latter study,deals with the measurement of chlorpromazine and itsmetabolites and of tricyclic antidepressants in blood.Studies of the anticonvulsant drugs phenobarbital and

CLINICALCHEMISTRY,Vol. 24, No. 8, 1978 1387

phenytoin are included, for although these drugs are notstrictly psychopharmacologic agents, they are fre-quently administered to patients receiving treatmentat mental-health facilities.

Materials and MethodsReagents

Sources of drugs and drug metabolites were: chior-promazine, chiorpromazine sulfoxide, and 7-hydroxy-chlorpromazine, Smith, Kline & French Laboratories,Philadelphia, Pa.; chlorpromazine-N-oxide dihydrate,nor-chlorpromazine, and di-nor-chlorpromazine, Psy-chopharmacology Research Branch, National Instituteof Mental Health, Rockville, Md.; loxapine, LederleLaboratories, Pearl River, N. Y.; butaperazine, A. H.Robins, Co., Richmond, Va.; perphenazine, ScheringCorp., Kenilworth, N. J.; amitriptyline, Merck, Sharp& Dohme, West Point, Pa.; nortriptyline, Eli Lilly & Co.,Indianapolis, Ind.; imipramine, Geigy Pharmaceuticals,Summit, N. J.; desipramine, USV PharmaceuticalCorp., Tuckahoe, N. Y.; sodium phenobarbital, Mal-linckrodt, Inc., St. Louis, Mo.; metharbital, AbbottLaboratories, North Chicago, Ill.; phenytoin (5,5-di-phenylhydantoin), Aldrich Chemical Co., Milwaukee,Wis.

Solvents, from Fisher Scientific Co., were all distilledbefore use in all-glass systems and stored in bottles withpoly(tetrafluoroethylene) -sleeved glass stoppers.

All glassware coming into contact with samples orextracts was silylated by use of hexamethyldisilazine atelevated temperature and reduced pressure (18).

Stock Solutions

Solutions containing internal standard (IS) andcarrier (C) compounds were prepared for the followingassays:

Chioropromazine: per liter, 40 mg of butaperazine(IS) and 200 mg of perphenazine (C) in isoamyl alco-hol/heptane (15/85 by vol).

Amitriptyline, nortriptyline, imipramine, anddesipramine: per liter, 10 mg of loxapine (IS) and 200mg of perphenazine (C) in isoamyl alcohol/heptane(15/85 by vol).

Phenobarbital and phenytoin: per liter, 100 mg ofmetharbital (IS) in ethyl acetate.

Extraction and sample preparation solutions includeNaOH (1 mol/liter), HCI (50 mmol/liter), NH4OH (1mol/liter), citrate buffer (1 mol/liter, pH 5), and isoamylalcohol/heptane (1.5/98.5 by vol).

Development solvents were hexane/acetone/dieth-ylamine, 77/20/3 by vol., and ethyl acetate/ammoniumhydroxide 97/3 by vol.

Modified Forrest reagent, for making visible thephenothiazine drugs, was prepared according to Chanand Gershon (11) by dissolving 1 g of ferric chloridehexahydrate in 100 ml of sulfuric acid/water (1/1 by vol)and 300 ml of absolute ethanol.

Thin-Layer Chromatographic Materials

HPTLC plates (HPTLC silica gel 60) were from E.

Merck, Darmstadt, G.F.R. Twin-trough developmenttanks, variable volume, and 100-nl platinum-iridiumspotting pipettes were from Camag, Inc., New Berlin,Wis. Glass spotting pipettes were made by drawingdisposable Pasteur pipettes with a flame and then si-lylating as above.

Spectrophotometry

A Model KM-3 Chromatogram Spectrophotometer(Zeiss Instruments, New York, N. Y.) was used for allquantitative measurements. Ultraviolet absorbance inthe reflectance mode was measured for all compounds,with use of a 0.5-mm slit-width and 3.5-mm slit-length.Measurement wavelengths were as follows: chlorpro-marine, 254 nm; amitriptyline and nortriptyline, 275nm; and phenobarbital and phenytoin, 230 nm. Visibleabsorbance in the reflectance mode from chlorproma-zine, after reaction with modified Forrest reagent, wasmeasured at 425 nm.

Procedures

Sample Preparation

Chiorpromazine, amitriptyline, nortriptyline, im-ipramine, or desipramine. Deliver a 100-,I portion ofstock solution (see above), diluted 10-fold and con-taining internal standard and carrier, to a 15-ml silylatedscrew-top test tube and evaporate under reducedpressure. Add 1 ml of plasma, 1 ml of NaOH (1 mol/liter), and 10 ml of isoamyl alcohol/heptane (1.5/98.5).Mix gently for 30 mm with a tube-rocker and thencentrifuge for 5 mm in a bench-top centrifuge to sepa-rate the phases. Freeze the aqueous plasma layer in anacetone-solid CO2 bath, and transfer the organic layerto another 15-ml tube. Add 1 ml of the dilute HC1, mixwith a vortex mixer, and separate by centrifugation.Aspirate the organic layer, and alkalinize the remainingaqueous layer to pH 10 by adding 0.2 ml of the diluteammonium hydroxide. Extract with 2 ml of isoamylalcohol/heptane (1.5/98.5) by mixing, then centrifuge.Freeze in the acetone-solid CO2 bath and transfer theorganic phase to a silylated Reactivial (Pierce ChemicalCo., Rockford, Ill.). Evaporate it under reduced pressureand re-dissolve the residue in 10 ,l of isoamyl alcohol/heptane (15/85) immediately before spotting theHPTLC plate.

Phenobarbital and phenytoin. Add 400 l of stocksolution containing 100 mg of metharbital per liter toa 15-ml screw-top test tube and evaporate under re-duced pressure. Add 1 ml of plasma, 1 ml of pH 5 citratebuffer, and 5 ml of diethyl ether. Mix gently on arocker-type shaker for 30 mm. Remove the ether phaseto a Reactivial and evaporate under dry nitrogen at 40#{176}C.Dissolve the residue in 50 l of dichloromethane andspot on an HPTLC plate.

Thin-Layer Chromatography.

Psychopharmacologic agents. Spot the entire re-dissolved sample on the HPTLC plate at 1.0 cm fromthe edge, using a variable-volume platinum-iridium

A B

1.5.

C

cpz

0 1 2 34

L j

P

1388 CLINICALCHEMISTRY,Vol. 24, No. 8, 1978

0 1 3 3 4 5Cm

Fig. 1. HPTLCseparation of six phenothiazine drugsMultIpledevelopmentin benzene/acetone/ammonlum hydroxIde (80/20/0.2).(1) Acetophenazlne, (2) perphenazine. (3) trifluoperazine, (4) promazine, (5)thioridazine, (6) chlorpromazine

pipette and taking care to keep the spot diameter to lessthan 2 mm. After spotting, place the plate under re-duced pressure to assure complete removal of thespotting solvent. Place the plate in a twin-troughchamber containing the hexane/acetone/diethylammnesolvent mixture for 5 mm for equilibration and then tiltthe chamber to allow the solvent to reach the plate.Allow the solvent to ascend 4 cm above the origin, andthen place the plate under reduced pressure for 30 mmto remove the solvent completely before scanning withthe chromatographic spectrophotometer.

Anticonvulsant drugs. Spot sample on the HPTLCplate with the 100-nl platinum-iridium pipette. Developas above, using the ethyl acetate/ammonium hydroxidemixture and allowing the solvent to run 4 cm above theorigin. Dry thoroughly in warm air or reduced pressurebefore scanning.

ResultsHPTLC, like any high-resolution separation tech-

nique, requires a certain amount of attention to detailfor best results. Pre-washing the plates, application ofthe sample to as small an area as possible, and devel-opment in carefully controlled, pre-saturated chambersare all requisites to satisfactory and reproducibleanalyses. Nevertheless, with a minimum of experienceseparations can be attained such as that shown in Figure1, with development times often measured in secondsrather than minutes. An appropriate TLC plate is,however, a necessary requirement, as seen in Figure 2.The resolution attained in these two separations is

Fig. 2. Chromatograms of 10 ng of phenobarbital (F), phenytoin(0), and metharbital (M) developed in ethyl acetate/ammoniumhydroxide (97/3)A, HPTLC plate; B, conventional TLC plate

0 1 2 3 4

Fig. 3. HPTLC separation of chlorpromazineA, scan of blank HPTLC plate after development; B, scan of plasma extractwithout drug: C, scan of plasma extract contaIning 10 of chlorpromazine perlIter with 400 tg of butaperazine per liter as the Internal standard (IS) and 2mg/lIter perphenazine as carrIer. Development solvent: hexane/acetone/di-ethylamine (77/20/3); wavelength. 254 nm

similar, but the background noise arising from irregu-larities in the surface of the conventional plate limits thesensitivity of measurement.

Figure 3 shows a scan of the HPTLC separation ofchlorpromazine from a 1.0-ml plasma sample containing10 zg/liter. The rising baseline is due to impurities ab-sorbed by the silica gel that were not totally removed bypre-washing the plate; the material present at 4cm fromthe origin is impurity that moved with the solvent front.This is frequently observed at high sensitivity but sel-dom causes interference. The linear regression ofpeak-height ratios of chlorpromazine to the internalstandard, butaperazine, with concentration in bloodplasma is illustrated by Figure 4. Therapeutic concen-

Q 1.5

to

05

50 100 150 200 250 300P9/L

Fig. 4. Linear regression comparing chlorpromazine to buta-perazine (internal std.) peak-height ratios at different plasmaconcentrations.y = 0.43+ .0055x, r = .992, S.E. = .061

B

A

4

0 1 2 3 4

Fig. 6. Separation of chlorpromazine and Its major metabolites,with clozapine (Sandoz Pharmaceuticals, East Hanover, N. J.)as internal standard (I.S.)Development solvent: ethyi acetate/acetic acld/water/acetone/Isopropanol(40/5/5/2.5/2.5). nsi 4 cm from origin three successive tImes. Scanned at 250nm

Fig. 5. Peak height ratios, chlorpromazine to internal standard,at various concentrations after spraying plate with modifiedForrest reagent and scanning at 425 nmA, immediately after color development; B, 30 mm after color development

2.0

CLINICALCHEMISTRY,Vol. 24, No. 8, 1978 1389

trations of the drug are well within the usable range ofconcentrations (2). With ultraviolet absorption as themeans of quantitation, the limit of detectability ofchiorpromazine in blood plasma is about 1 pg/liter,approximately the sensitivity achieved by electron-capture gas chromatography (4).

Several chromogenic reagents have been proposed formaking phenothiazines visible on thin-layer plates,including Dragendorff spray, Marquis reagent, andFPN reagent (19,20). Gershon et al. (11) used a modi-fied Forrest reagent in studies of chiorpromazine andits metabolites, and this spray appeared to yield an in-tense, stable color that could possibly serve as an al-ternative to ultraviolet absorptiometry, to enhance thesensitivity of detection of chlorpromazine. When ex-amined quantitatively, however, the sensitivity wasslightly less than that attained by ultraviolet absorption,and the intensity was observed to vary fairly quickly.Figure 5 illustrates the change in peak-height ratios ofchlorpromazine to the internal standard some 30 mmafter color development; the increase in ratio is dueprimarily to the decrease in intensity of the internalstandard with time, but both compounds faded almostcompletely in 1 to 2 h. Ultraviolet absorption, on theother hand, is relatively stable, but scans should nev-ertheless be completed as soon as possible after chro-

matographic separation, to avoid errors that might ac-company degradation of the compound by light, at-mospheric oxidation, or contamination.

The metabolic products of drugs are usually moredifficult to determine than the parent compounds be-cause of their increased polarity resulting from (e.g.)addition of hydroxyl groups or loss of methyl groups.Consequently, derivitization is usually a necessary stepin gas chromatography, to prevent sample loss throughdecomposition or adsorption to active sites in thechromatographic system. This problem is not as severewith liquid-chromatographic techniques, whether col-umn or thin-layer, and Figure 6 shows a HPTLC chro-matogram of the major metabolic products of chlor-promazine in a single chromatographic separation. Thecomponents of the mixture are resolved almost tobaseline throughout, and the solvent front is only about4 cm from the origin.

The tricyclic antidepressants are very similar to thephenothiazines with respect to sample preparation andHPTLC procedures. The principle difference is in theuse of loxapine rather than butaperazine as the internalstandard. Figure 7 shows the ultraviolet scan of thedeveloped amitriptyline and nortriptyline HPTLCseparation. Inasmuch as nortriptyline is an active me-tabolite of amitriptyline, simultaneous determinationof these drugs is an important consideration for clinicalstudies. The regression lines for amitriptyline andnortriptyline are shown in Figure 8. Linearity andstandard error appear to be acceptable throughout thereported therapeutic range of plasma concentrationsfor these drugs (21, 22).

Addition of a carrier compound that structurally issimilar to the compounds being measured significantly

IRIFTYLII#{128}y. .17+.olox

#{149}.991SL #{216}64

Table 1. Recovery and Reproducibility of HPTLCConcn. Recovery, %

In serum wuinjg/liter carrier WIthout

100 96 87

Drug

Chiorpromazine

Amitriptyline

Nortriptyline

Imipramine

Desipramine

Average values from six determinations.

10100

10100

10100

10100

10

ReproducIbilIty (CV)WIth

carrier Without

3.3 2.68.0 34.22.6 7.97.8 11.33.1

809995778391

867082

548771736286776272

8.54.04.26.0

10.7

12.418.57.9

11.313.517.7

1390 CLINICALCHEMISTRY,Vol. 24, No. 8, 1978

U,

0 1 2 3

Fig. 7. Chromatogram of amitriptyline and its demethylatedmetabolite, nortriptyline, with loxapine as the internal standard(I.S.)Development solvent: hexane/acetone/diethylamine (77/20/3). Scanned at 240nm

improves both recovery and reproducibility, particularlyat low concentration (4, 17). Perphenazine was chosenfor phenothiazine and tricyclic assays, because it con-veniently remains at the origin in the developmentsystems described in this report. The effect of this car-rier on the recovery and reproducibility of measurementis shown in Table 1. As would be expected, the influenceof carrier is more pronounced at low concentration, buteven at 100 pig/liter, the improvement is sufficient torecommend its use.

Determinations of antiepileptic drugs in blood havereceived a great deal of attention in recent years, andmethods ranging from immunoassays to nearly every

2

I12‘UI

‘U4

Fig. 8. Linear regressions of amitriptyline and nortriptyline toloxapine (IS) peak-height ratios at different plasma concentra-tions

type of chromatography have been proposed and eval-uated for clinical use (10). The therapeutic concentra-tions of most of these drugs in blood are sufficient thatsensitivity of detection is not a major problem, but onthe other hand, the capability of utilizing small samplevolumes is often a distinct advantage. Figure 9 showsa scan of phenobarbital and phenytoin extracted fromblood plasma, each being present at a concentration of10 mg/liter. The actual amount of drug on the HPTLCplate was 40 ng of each, and the amount of metharbital,the internal standard, was 120 ng. The regressions forthese compounds were somewhat curved, but predict-able. Replicate determinations of 20 mg/liter samplesof these drugs in plasma were undertaken to ascertainwhat operations in the procedure were the major con-tributors to error. The results of these determinations(Table 2) reveal that the spectrophotometry, as shownin the third column of figures, is not a very significantsource of variability. Instead, the spotting step and thesample preparation appear to be the main contributorsto error. These procedures, however, are probably themost amenable to improvement as the operator becomesmore expert.

Values for multIplespotting of single sanls

Phesobarb. Phenytain

Table 2. EvaluatIon of Sources of ErrorValues a for entire

procedurePhenoba Ph.nytoln

Values for multiplescan of single spotting

Phenobarb. Phenytoin

.811

.864

.914

1.34 .7651.581.481.481.321.091.391.331.49 .888

.636

.845

.777

1.32 .7621.33 .7561.34 .7681.33 .7681.32 .7591.32 .7561.33 .768

1.39 1.38 .8591.40 1.46 .8881.27 .8401.371.231.41

Mean 1.35 .876 1.40 .821 1.33 .765SD .0775 .025 .124 .0767 .0071 .0086CV, % 5.3 2.9 8.9 9.3 0.5 1.1

.840

.876

.852

.906

.885

.895

M

CLINICAL CHEMISTRY. Vol. 24, No. 8, 1978 1391

Peak-height ratios of phenobarbitai and phenytoln to metharbltal (internal standard) from plasma extracts.

DiscussionAlthough good resolution, sensitivity, and repro-

ducibility are all achieved by HPTLC these requisitesfor determining drugs in blood can be met by mostchromatographic techniques if sufficient attention isgiven to the components and operation of the systems.

012345

Fig. 9. Chromatogram of an extract of plasma containing 10mg/liter phenobarbital (F) and phenytoin (0)Actual amount of drugon the HPTLC plate ls4O ng for each drug, l2Ong for theInternal standard, metharbitai (i.e

Nevertheless, certain features of HPTLC merit atten-tion other than the obvious advantage of simultaneousseparations. To begin with, the selection of solventsystems for HPTLC is unrestricted by considerationsof transparency to ultraviolet radiation, a major factorin choosing solvents for use with HPLC where ultravi-olet absorption is the basis of detection. In addition, thecompound being measured need migrate only enoughto be separated from adjacent compounds. In contrast,with column chromatography the compound musttraverse the length of the column, and all other com-ponents of the sample mixture must be elu ted to avoidinterference with subsequent samples. Also, sampleclean-up does not have to be as meticulous withHPTLC, because there is no problem with columndegradation due to the accumulation of components atthe inlet.

Perhaps the most important feature of thin-layersystems is in the static rather than dynamic nature ofthe detection process. The components separated byTLC remain unmoving on the plate, to be examined byany of a number of methods at any time (stability beingthe limiting factor here). For example, an isolated spotcan be characterized by scanning with reflectance ul-traviolet spectrophotometry with incremental changesin wavelength. The ultraviolet spectra of chiorproma-zine and chlorpromazine sulfoxide obtained in thismanner are shown in Figure 10. We used such identifi-cation while studying multiple-development systems(i.e., re-running plates in the same solvent repeatedly)for chlorpromazine separation. An extraneous peak wasobserved at the position on the plate occupied bychlorpromazine before each re-development. Thiscompound, indicated by the small arrow in Figure 11,

S

400 350 250 200

1392 CLINICALCHEMISTRY,Vol. 24, No. 8, 1978

400 350 440 flO 400

Fig. 10. In situ ultraviolet spectra of spots on HPTLCplateLeft chlorpromazine. Right chlorpromazine suitoxide

0

400 350 300 250 200

F

Fig. 11. Extraneous peak (arrow) observed in chiorpromazinemultiple-development system (right) and its in situ ultravioletspectrum (left)

when subjected to in situ spectrophotometry, yieldedthe spectrum shown to the right of the chromatogram.Comparison to the spectrum of chlorpromazine sulf-oxide in the previous figure reveals identical absorptionmaxima, and there is little doubt that chlorpromazinewas oxidized on the plate under the conditions used inthe multiple-development process.

Finally, with static detection it is also possible toexamMe portions of a chromatogram with such com-puter assisted processes as signal averaging (23), digitalfiltering, edge detection, or pattern recognition-techniques not available for, or quite difficult to adaptto, detectors monitoring flowing column effluents.

References1. Curry, S. H., Determination of nanogram quantities of chlor-promazine and some of its metabolites in plasma using gas-liquidchromatography with an electron capture detector. Anal. Chem. 40,1251 (1968).

2. Curry, S. H., Chlorpromazine analysis by gas chromatography withan electron-capture detector. In The Phenothiazines and Structur-ally Related Drugs, 1.8. Forrest, C. J. Carr, and B. Usdin, Eds. RavenPress, New York, N. Y., 1974, p 335.

3. Christoph, G. W., Schmidt, D. E., Davis, J. M., and Janowsky, D.S., A method for determination of chlorpromazine in human bloodserum. Clin. Chim. Acta 38, 265 (1972).4. Davis, C. M., Meyer, C. J., and Fenimore, D. C., Improved gaschromatographic analysis of chlorpromazine in blood serum. Clin.Chirn. Acta 78, 71(1977).

5. Hucker, H. B. and Stauffer, S. C., Rapid, sensitive gas-liquidchromatographic method for determination of amitriptyline andnortriptyline in plasma using a nitrogen-sensitive detector. J. Chro-matogr. 138,437 (1977).6. Burgett, C. A., Smith, D. H., and Bente, H. B., The nitrogen-phosphorus detector and its applications in gas chromatography. J.Chromatogr. 134, 57 (1977).

7. Alfredsson, G., Wode-Helgodt, B., and Sedvall, G., A mess frag-mentographic method for the determination of chlorpromazine andtwo of its active metabolites in human plasma and C.S.F. Psycho-pharmacologia 48, 123 (1976).

8. Matin, S. B., Wan, S. H., and Knight, J. B., Quantitative deter-mination of enantiomeric compounds. I. Simultaneous measurementof the optical isomers of amphetamine in human plasma and salivausing chemical ionization mass spectrometry. Biomed. Mass Spec-trom. 4, 188 (1977).

9. Tjaden, U. R., Lankelma, J., Poppe, H., and Muusze, R. G., Anodiccoulometric detection with a glassy carbon electrode in combinationwith reversed-phase high-performance liquid chromatography. J.Chromatogr. 125, 275 (1976).

10. Kupferberg, H. J., Quantitative methods for antiepileptic druganalysis: An overview. In Antiepileptic Drugs, Quantitative Analysisand Interpretation, C. E. Pippenger, J. K. Penry, and H. Kutt, Eds.,Raven Press, New York, N. Y., 1978, p 9.

11. Chan, T., and Gershon, S., Quantitation of chlorpromazine andits metabolites in human plasma and urine by thin-layer chroma-tography. In Quantitative Thin-Layer Chromatography, J. C.Touchstone, Ed., John Wiley and Sons, New York, N. Y., 1973, p253.

12. Ripphahn, J., and Halpaap, H., Quantitation in high-performancemicro-thin-layer chromatography. J. Chromatogr. 112,81 (1975).13. Halpaap, H., and Ripphahn, J., High performance thin-layerchromatography: development, data, and results. In HPTLC HighPerformance Thin-Layer Chromatography, A. Zlatkis and R. E.Kaiser, Eds., Elsevier, Amsterdam, 1977, p 95.14. Kaiser, R. E., Dosage techniques in HPTLC. Ibid. p 85.

15. Hezel, U. B., Potential and experience in quantitative HPTLC.Ibid. p 147.16. Ritter, W., Thin-layer densitometric determination of muzol-imine, a structurally new diuretic drug, at the nanogram level in bi-ological fluids. J. Chromatogr. 142, 431 (1977).17. Fenimore, D. C., Meyer, C. J., Davis, C. M., et al., High-perfor-mance thin-layer chromatographic determination of psychophar-macologic agents in blood serum. J. Chromatogr. 142, 399 (1977).

18. Fenimore, D. C., Davis, C. M., Whitford, J. H., and Harrington,C. A., Vapor phase silylation of laboratory glassware. Anal. Chem. 48,2289 (1976).

19. Stahl, E., Thin-Layer Chromatography, Springer-Verlag, NewYork, N. Y., 1969, p 873.

20. Clark, E. G. C., Isolation and Identification of Drugs, ThePharmaceutical Press, London, 1969, p 800 and 801.21. Ziegler, V. E., Clayton, P. J., and Biggs, J. T., A comparison studyof amitripyrline and nortriptyline with plasma levels. Arch. Gen.Psychiatry 34, 607 (1977).

22. Gram, L. F., Reisby, N., Ibsen, I., et al., Plasma levels and anti-depressive effect of imipramine. Clin. Pharmacol. Ther. 19, 318(1976).

23. Ebel, S., and Hocke, J., Computer-controlled evaluation inthin-layer chromatography. J. Chromatogr. 126, 449 (1976).