John S. Lilleyman. Depts. of Chem. Pathol. & Haematol. The Children's Hospital. Western Bank. Sheffield. SlO 2TH, UK. Comparison of Two Colorimetric.
Cyclosporin Concentrations in Whole Blood and Plasma To the Editor: Debate continues on the usefulness of measuring cyclosporin in plasma as a potential indicator of its immunosuppressive or toxic effects (1-3). In September 1982, the manufacturer (Sandoz Products Ltd., Basle, Switzerland) circulated to users details of the appar-
ent variation erythrocytes
in affinity of the drug for at various temperatures, work that has since been confirmed (4). It has been suggested that a rigorous routine of incubating the blood sample before the plasma is separated is necessary to standardize results, and an incubation temperature of 37 #{176}C has been recommended (5). We measured cyclosporin in both whole blood and plasma after a 30-mm incubation at 37 #{176}C in a single patient who was undergoing therapy with cyclosporin during bone-marrow transplantation. We used the radioimmunoassay method provided by Sandoz Products Ltd. When we compared the two sets of data, we found a poor correlation (y = 0.26x - 62.6, r = 0.58, n = 34, where y = values for plasma and x = values for whole blood), although we strictly controlled sample preparation. The whole blood/plasma concentration ratio varied from 1.6 to 3.6. These observations would appear to support suggestions that the degree of uptake of the drug by erythrocytes depends on the total drug concentration and on binding to other blood components, including lipoproteins (6). Whatever the cause, the variations are such that even strict adherence to a standardized sample-handling procedure is not sufficient to validate the use of data on concentrations in plasma from one laboratory to another-or even within the same laboratory. If figures for whole blood are used, it seems likely that the variations in sample handling, both within and outside the laboratory and among patients, will be minimized. A more widespread use of whole blood as the sample may expedite the development of a useful therapeutic range of this increasingly important drug. References 1. Bandini G, Ricci P, Visani G, et al. Measuring cyclosporin in plasma. Lancet i, 762 (1983). Letter. 2. Kennedy MS. Evaluating cyclosporin nephrotoxicity. Lancet ii, 861 (1983). Letter. 3. Van Willebrand E, Hayry P. Cyclosporin-A deposits in renal allografts. Lancet ii, 189-192 (1982). 4. Weak M, Follath F, Abisch E. Temperature dependency of apparent cyclosporin-A concentrations in plasma. Clin Chem 29, 1865 (1983). Letter.
5. Dieperink H. Temperature dependency of cyclosporin plasma
levels.
Lancet 1, 416
(1983). Letter. 6. Lemaire M, Tillement P. Role of lipoproteins and erythrocytes in the in vitro binding and distribution of cyclosporin-A in the blood. J Pharm Pharntacol 34, 715-718 (1982). Michael
J.
Bennett
Kevin H. Carpenter Eric Worthy John S. Lilleyman Depts.
of Chem.
The Children’s Western Sheffield
Pathol.
& Haematol.
Hospital
Bank
SlO 2TH, UK.
Comparison of Two Colorimetric
Assaysfor Angiotensin-Converting Enzyme Activity To the Editor: Although angiotensin-converting enzyme (ACE; peptidyldipeptide hy-
drolase, EC 3.4.15.1) plays a crucial role in blood-pressure regulation (1), its measurement in serum has also proved useful in the diagnosis and management of sarcoidosis (2, 3), the activity being increased in most patients with the active disease and declining with spontaneous remission or steroid therapy. We previously
reported
a sensitive,
precise assay for serum ACE, in which hippurate released from hippuryl-ihistidyl-L-leucine by ACE is quantified by reaction with cyanuric chloride! dioxan reagent in the presence of phosphate buffer (4). Notwithstanding a large unexplained constant error (yintercept = 10.5 UIL), results by our method have been shown (5) to correlate well (r = 0.988) with those by a radiometric assay. Recently, we had the opportunity to compare our method with the colonmetric technique of Kasahara and Ashihara (6). In this technique p-hydroxyhippuric acid, released from phydroxyhippuryl-Lhistidyl L- leucine by ACE, is hydrolyzed by hippuricase (EC 3.5.1.14) to p-hydroxybenzoic acid and glycine. Oxidative coupling of phydroxybenzoic acid with 4-aminoantipyrine produces a quinoneimine dye. ACE activity is calculated from absorbance measurements at 505 nm. This method is now available in kit form
ods. Linear regression analysis of the data for the kit method (y) vs our method (x) gave the equation y = 0.273x - 0.436 (r = 0.993, = 0.973, = 67.73, SD = 30.54,5’ = 18.08, SD = 8.40). A high degree of correlation, with minimal constant error, is evident. Although the chemical sensitivity of the kit method is about one quarter that of our method (slope = 0.273), the precision of the kit is excellent. Replicate analyses (a = 10) of two sera resulted in CVs of 2.2% at 7.8 U/L
and 1.8% at 14.6 UIL. These CVs compare favorably with data of Kasahara and Ashihara (6) and with CVs for our method. Reference intervals were not quoted by Kasahara and Ashihar, but application of the regression equation to our figures gave reference intervals for adults of 6-22 U/L (men) and 6-18 U/L (women). The ACE-color test kit should be welcomed by those clinical laboratorians who wish to assay serum ACE activity but have hitherto been hampered by the lack of a kit method.
References 1. Soffer RL. Angiotensin-converting enzyme and the regulation of vasoactive peptides. Annu Rev Biochem 45, 73-94 (1976). 2. Lieberman J. A new confirmatory test for sarcoidosis.Serum angiotensin converting enzyme: Effect of steroids and chronic lung disease.Am Rev Respir Dir 109, 743 (1974). Abstract. 3.
DeRemee RA, Rohrbach MS. Serum
angiotensin-converting enzyme activity in evaluating the clinical course of sarcoidosis. Ann Intern Med 92, 361-365 (1980). 4. Hurst PL, Lovell-Smith CJ. Optimized assay for serum angiotensin-converting enzyme activity. Clin Chem 27, 2048-2052 (1981).
5. O’Brien JF, Foreman RW, Rohrbach MS. and radiometric assays of angiotensin-converting enzyme compared. Clin Chem 29, 1990-1991 (1983). Letter. 6. Kasahara Y, Ashihara Y. Colorimetry of angiotensin-1 converting enzyme activity in Spectrophotometric
serum.
Clin Chem 27, 1922-1925 (1981).
P. L. Hurst C.
-
(“ACE-color”;
Fujizoki Pharmaceutical
Co. Ltd.,Shinjuku-ku,Tokyo, Japan; New Zealand agent,Medic DDS Ltd., Wellington,N.Z.).A gracious gift of some kits by the manufacturer enabled us to compare the measurement of serum ACE by these two methods. We assayed 40 sera with a wide range of ACE activities by both meth-
J.
Lovell-Smith
Chem. Pathol. Lab. Dunedin Hospital Dunedin New Zealand
Improved Simultaneous
Liquid-
Chromatographic Determination of Antiepileptic Drugs In Plasma To the Editor:
“High-performance” liquid chromatography (HPLC) is currently commonly used in monitoring antiepileptic
CLINICAL CHEMISTRY, Vol. 30, No. 5, 1984
817
drugs in plasma. However, most such methods are designed to involve a simple, rapid extraction. Some of these extraction procedures are poor, failing to remove all compounds that might interfere with the assay, contaminate the column and injection system (1-4), or cause low analytical recoveries (5). Here we describe a rapid, one-step
liquid-extraction procedure in which precipitation of plasma proteins is complete and interfering compounds in plasma are eliminated. The result is a rapid HPLC method for simultaneous-
ly determining ethosuximide, phenobarbital, phenytoin, and carbamazepine in plasma. We used a series 600 HPLC system (Kontron AG, Zurich, Switzerland), a visible-ultraviolet
variable-wave-
length detector (Uvikon 720LC, Kontron AG), a Kontron microprocessor, Model 200, and a Model 3390A integrator (Hewlett-Packard, PA). The reversed-phase
Avondale, analytical col5 ODS, 250 X
man was a Spherisorb 4.6 mm (Phase Separations Ltd., Queensferry, U.K.), protected with a 50 x 4.6mm guard-column and packed with CO:PELL ODS (Whatman Inc., Clifton, NJ). Chromatography was at room temperature, with a mobilephase (acetonitrile/methanol!phosphate buffer pH 4.0, 21!24/55 by vol) flow rate of 3 mL/min and a detector wavelength of 195 am. The internal-standard solution was prepared daily by diluting in water a stock standard of 5-(p-tolyl)-5-phenylhydantoin (1 g!L, in methanol) to give a concentration of 30 mg/L. Plasma was extracted as follows: In a 10-mL glass-stoppered centrifuge tube place 200 .tL of plasma and 200 L of internal-standard solution, and vortex-mix for 5 s. Add one drop of 5 mol/L HC1 and again vortex-mix for 5 s. Then add 2.5 mL of dichloromethane, stopper the tube, and vortex-mix for 1 mm.Using a calibrated spatula, add sufficient solid aminonium sulfate to saturate the aqueous layer and vortex-mix for 30s. Centrifuge (5000 rpm, about 5 mm). Aspirate and discard the
aqueous (upper) layer with a Pasteur pipette. Carefully separate the precipitate and transfer the lower layer to a 10-mL conical glass tube. Pipette 2 mL of the organic layer into another corncal glass tube and evaporate it in a
water bath at 42 #{176}C. Redissolve the residue in 50 L of a mixture of acetonitrile!methanol!water (instead of buffer) with the same volume relationships as the mobile phase and inject 20 L into the chromatograph. Quantify by using the peak-height ratios. One obtains a linear response for each of the four drugs in the following ranges: ethosuximide 10-150 mgfL, phenobarbital 5-80 mgfL, phenytoin 2.5-30 mg!L, and carbamaze818
pine 2-20 mgfL. Analytical recoveries varied from 92 to 100%, with withinday CVs from 0.9 to 4.6%, and between-day CVs from 0.9 to 4.2%. We tested the following drugs for potential interference: theophylline, acetaminophen, acetylsalicylic acid, phenylethylmalonamide, chlorpromazine, primidone, p-hydroxyphenytoin, carbamazepine-10,11-epoxide, pentobarbital, clonazepam, diazepam, meth-
aqualone,and flunitrazepam. Of these drugs, only chlorpromazine co-elutes with ethosuximide, but this is not a serious interference because these two drugs are not co-prescribed. Extraction with an organic solvent in the presence of added inorganic salt is a suitable method for plasma protein precipitation (6). Dichloromethane is a good extraction solvent because evaporation is fast at 42 #{176}C, thus minimizing losses of the highly volatile ethosuximide. The extraction with dichloromethane at low pH in combination with an excess of ammonium sulfate allows complete precipitation of plasma proteins, which otherwise would contaminate the chromatographic system, and it also eliminates some lipids and other plasma components that generally appear as early-eluting peaks and may interfere with the assay. It is advisable to add the ammoniurn sulfate before the drugs are extracted into the dichloromethane, to prevent coprecipitation of any drug trapped in the precipitate. The relatively high concentration of acetonitrile in the mobile phase diminishes back-pressure at room temperature. Moreover, the absence of interfering peaks at the beginning of the chromatogram (Figure 1) allows the use of a high flow rate, which in turn speeds the analysis and improves the shape of
the later peaks. The use of an injection solvent
with
a composition
similar
to
that of the mobile phasealso improves the shape of peaks, and the use of water
instead
of phosphate
buffer
in
this mixture avoids the problem of inorganic salt precipitating in the injection system. References 1. Freeman D, Rawal N. Serum anticonvul-
sant monitoring by liquid chromatography with a methanolic mobile phase. Clin Chem 25, 810-811 (1979). Letter. 2. Bernardo M. Improved sample preparation for measuring anticonvulsant drugs by “high-performance” liquid chromatography. Clin Chem 25, 1861 (1979). Letter. 3. Chu SY, Oliveras L, Deyasi S. Extraction procedure for measuring anticonvulsent drugs by liquid chromatography. Clin Chem 26, 521 (1980). Letter. 4. Mathies JC, AustinMA. Modified acetonitrile protein-precipitation method of sara-
ple preparation for drug assay by liquid chromatography. (1980). Letter.
Clin
Chem 26,
1760
Adams RF, Vandemark FL. Simultaneous high-pressure liquid-chromatographicdeterminationof some anticonvulsantsin 5.
serum. Clin Chem 22, 25-31 (1976). 6. Blanchard J. Evaluation of the relative efficacy of various techniques for deproteinizing plasma samples prior to high-performance liquidchromatographic analysis.J Chromatogr 226, 455-460 (1981).
Ramon Soto-Otero German Sierra-Marcu#{241}o Dept. of Biochem. School of Med. Santiago de Compostela,
Spain
Visual Detection of Abnormal A
B
‘-I
Hemoglobin with a Procedure for
Glycated Hemoglobin
#{163} C
To the Editor: The value of measuring glycated hemoglobin in the management of patients with diabetes mellitus is well documented (1), and the use of cation-
U 2 S S 0 S S
exchange chromatography to do so is well accepted (2). Abnormal hemoglo-
114$
I TIME
1141$ (mm)
Fig. 1. Liquidchromatogramsof some antiepi-
bins may give spurious values by these methods (3, 4). We find that visual inspection of the Isolab Fast Hemoglobin Test System (5) columns after the fast hemoglobin (Hb A1) fraction is eluted can lead to detection of abnormal hemoglobins.
Studying
glycated
hemoglobins
in
leptic drugs
sickle-cell
A Extract of human plasma containing, per liter, 50
gotes, we observed differences in the
heterozygotes
and homozy-
mgofethosuximide (peak 1), 20mg of phenobarbitalappearance of the columns (Figure 1) (peak 2), 10 mgof phenytoin (peak , and 8mg of after the glycated hemoglobins were carbamazepine (peak eluted (6). Hemoglobins 5, C, and D are B. Blank: extract of human plasma containingonly the internalstandard (peak 5) not eluted from the column with the
CLINICAL CHEMISTRY, Vol. 30, No. 5, 1984
