Measurement of Glycosylated Hemoglobin on Cellulose Acetate by jennyyingdi


									CLIN. CHEM. 29/2, 340-343 (1983)

Measurement of Glycosylated Hemoglobin on Cellulose Acetate Membranes
by Mobile Affinity Electrophoresis
Jeffrey    Ambler,1 Borek Janik,2and GeoffreyWalker’

In this method for separating glycosylated from nonglycosy-                     tnchloroacetic    acid solution, 75 g/L), “Super Sepraclear”
lated  hemoglobin in blood by electrophoresis on cellulose                      clearing solution, and “Super Sepraphore”          membranes.   In
acetate membranes, we exploit the affinity of low-molecular-                    addition we also used the following solutions: sodium chlo-
mass dextran sulfate for the nonglycosylated fraction, which                    ride solution, 8.5 g/L; hemolyzing solution (per liter, 1 g each
increases the mobility of the latter relative to that of glycosy-               of saponin and EDTA); affinity     electrophoresis     buffer, pH
lated hemoglobin. After the membrane strips are cleared and                     6.4, 33 mmol/L citrate containing 2 mol of dextran sulfate
stained, the two fractions are quantified densitometrically. As                 and    8      zmol   of disodium         EDTA   per   liter   (obtainable   as
evaluated by use with blood from diabetics, results compare                     “Glyco-Phore   Buffer” from Gelman Sciences, prod. no.
well with those by chromatography on short columns and by                       51261).
                                                                                   Sample preparation. Blood was collected into tubes con-
electrophoresis in commercial agar gel films.
                                                                                taining EDTA as anticoagulant.    We added 100 L of whole
                                                                                blood to 5 mL of sodium chloride    solution in a centrifuge
   Glycosylation      of hemoglobin,         a nonenzyme-mediated,
                                                                                tube, mixed gently, and centrifuged. The supernatant      fluid
post-translational       modification, occurs slowly          in erythro-
                                                                                was aspirated        as completely      as possible    and 350 L of
cytes (1). Glucose     is bound by a stable ketoamine           linkage to
                                                                                hemolyzing      solution was added.
the N-terminal valine residue of the hemoglobin beta-chain
                                                                                   Electrophoresis.     Membranes      were equilibrated   with affin-
(2). The extent of glycosylation depends on the blood glucose
                                                                                ity electrophoresis       buffer for at least       10 mm and the
concentration   and measurements    of it are made to monitor
                                                                                chambers     were filled with buffer. The applicator wells were
long-term control of blood glucose in patients with diabetes
                                                                                loaded with hemolyzed          samples    (7 L per sample well) by
                                                                                use of a Gilson “Pipetman”            or similar device. The mem-
   There have been several approaches to the measurement
                                                                                branes     were blotted    and placed in the chamber          and the
of glycosylated    hemoglobin    concentration.   Long ion-ex-
                                                                                samples     were applied.     To ensure an even application,       the
change columns (1) and isoelectric focusing (4) are too
                                                                                applicator was not removed from the membrane until the
complex and tedious for routine use. “High-performance”
                                                                                sample was totally absorbed. After electrophoresis at 150 V
liquid chromatography    (5) can be automated   to carry out 50
                                                                                for 40 mm the strips were promptly placed in stain for 10
analyses per day with good precision, provided the equip-
                                                                                mm. Excess  stain was washed out by constantly    agitating in
ment is available in the laboratory. Estimations      involving
                                                                                changes of 50 mL/L acetic acid for a maximum    of 5 mm. The
short ion-exchange columns (6) and an automated thiobarbi-
                                                                                membranes    were cleared in “Super Sepraclear”     for 2 mm,
tune acid (7) method have been used in clinical laboratories,
                                                                                transferred  to a glass plate, and placed in an oven at 80-
but both have disadvantages.              The short ion-exchange         col-
                                                                                      for about 10 mm until transparent.
                                                                                90 #{176}C                                  The separations
umn method is capricious,           being affected by small variations
in temperature         (8), pH (9), and buffer (10). The thiobarbi-
                                                                                were scanned from the anode with an ACD-18 Densitometer
turic acid method, although it has been optimized                 (11), has
                                                                                at a wavelength of 520 nm.
not performed       well in our hands. More recently,          an agar gel      Agar Electrophoresis
electrophoretic        method     with direct     densitometry       of un-
stained    gels has been introduced          (12). This method        is not      Prepared      agar films             were obtained as a kit (Glytrac;
affected by temperature,          buffer, or pH variations      (13) and is     Corning    Ltd., Halstead,             Essex, U.K.) and the method per-
extremely       simple     to use. However,       films are tedious        to   formed according to the manufacturer’s       instructions, with
prepare    and costly when bought ready made.                                   the Corning cassette    electrophoresis  cell, variable power
   We present       a method       based on a new concept,          mobile      supply, applicator syringe, and drying oven. Whole-blood
affinity electrophoresis.       It requires no special apparatus         and    hemolysates (1 L) were applied directly to the preformed
exploits the convenience and economy of cellulose acetate                       wells in the agar. After electrophoresis  at 60 V for 40 mm,
membranes.     Large batch analyses may be carried out and                      the gel was dried and scanned, unstained,      with a Beckman
results correlate   well with those obtained   by the short-                    Densitometer    (R-112) fitted with a special 420-nm filter
column method and by agar gel electrophoresis.                                  donated       by Corning        Ltd.

Materials and Methods                                                           Short Ion-Exchange                Columns
                                                                                  These were obtained from BioRad Ltd., Watford, Herts,
Mobile Affinity      Electrophoresis                                            U.K., and analyses were performed according to the manu-
   Reagents   and equipment.        The following     equipment  and            facturer’s      instructions.
methods were kindly supplied by Gelman Sciences Inc., Ann
Arbor, MI 48106: Sepratek Electrophoresis              Chamber, Sep-            Results
ratek Applicator    (eight-sample       size), Power Supply, ACD-18               A typical separation     of specimens containing   various
Computing    Densitometer,       protein stain (Ponceau S, 5 g/L in             amounts of glycosylated hemoglobin is shown in Figure 1;
                                                                                densitometric scans of tracks 1 and 2 are shown in Figure 2.
   ‘Department  of Clinical Chemistry, University Hospital, Not-
tingham NG7 2UH, U.K.                                                           Sample Preparation
   2Gelman Sciences Inc., 600 South Wagner Rd., Ann Arbor, MI
48106.                                                                             The results were not affected by the anticoagulant      used
   Received May 25, 1982; accepted Oct. 26, 1982.                               for blood collection. Blood from five unselected  individuals

340 CLINICAL CHEMISTRY, Vol. 29, No. 2, 1983
                                                                                     achieve       good separations              of components,           one      should     use

                                                                                     hemolysates           with about 30-40           g of hemoglobin              per liter.
                      HbA       -                                        -

                                                                                     Effect of Variations              in Temperature              and Buffers
                      HbA1 -                                                               The   performance           of the     method      was       little     affected     by
                                                                                     changes       in temperature          or buffer pH and ionic strength, as
      Sample application
                                                                                     shown       by the following          tests. Eight samples from diabetic
                         Sample               12345678
                                                                                     patients   were chosen    without                 conscious     bias,       applied  to
                                                                                     membranes,     and then subjected                to electrophoresis          under the
                                                                                     conditions outlined in Table 1. It being                               impractical   to
Fig. 1. Electrophoretic separation of glycosylated hemoglobin (HbA1)                 examine all possible variations, an alteration in one param-
from nonglycosylated hemoglobin (HbA) in eight samples on cellulose                  eter was examined while retaining the others at optimum
acetate                                                                              conditions. The assay was very little affected by these
                                                                                     changes, indicating a wide margin for error in dissolving the
                                                                                     buffer      powder.     Change        in the buffer         concentration          altered
                                                                                     the distance       of the separation          rather     than the distribution               of
                                                                                     protein      between      the two bands,         as shown          in Figure       3. High
                                                                                      ionic strength, i.e., >75 mmol/L, produced no separation,
                                                                                     just a single, very tight band. Progressive      dilution of the
                                                                                     buffer produced greater      separation  of the bands and an
                                                                                      increasingly   diffuse band for nonglycosylated     hemoglobin.
                                                                                     We found the optimum concentration         of buffer was 30-40
                                                                                     mmol/L on the Super Sepraphore          membrane.      Other mem-
                                                                                     branes may require     slightly different  conditions;    for exam-
                                                                                     ple, a buffer molarity  of 25 mmol/L gave a good separation        of
                                                                                     the two bands             with     little    diffusion      on a Sepraphore                III

                                                                                     Hemoglobin             Variants
      HbA1         NbA                              HbA1        HbA
                                                                                           Only hemoglobin            F (HbF),     if present,        interfered      with      the
                                                                                     quantification by giving a spuriously    high glycosylated
Sample track 1 9.3% HbA1                       Sampletrack 2 23%HbA1
                                                                                     fraction. When high concentrations   of HbF were present,
                                                                                     however, the glycosylated hemoglobin (HbA1) values could
Fig. 2. Densitometnc scans at 520 nm of sample tracks 1 and 2 of the                 be corrected simply by subtracting      the HbF percent from the
separation shown in Fig. 2
                                                                                     apparent     HbA1 value. When hemoglobin          C was present,
                                                                                     double bands were obtained,          and the variant     was easily
was   collected  into tubes containing  EDTA, oxalate, or                            recognizable    by its slightly   anodic mobility    in comparison
heparin as anticoagulants.  Hemolysates prepared from each                           with nonglycosylated        hemoglobin.    By comparing        results
sample were run in duplicate. The mean ± SD percent of                               from the column        and agar-electrophoretic        methods,      we
glycosylated       hemoglobin     in the EDTA-, oxalate-,  and hepa-
rim-treated      blood was 7.9     ± 0.2, 7.8 ± 0.2, and 7.9 ± 0.4%,
respectively.  These        values fall within acceptable   precision
limits for the procedure.
   A solution of 1 g each of saponin                and EDTA per liter was
routinely      used to hemolyze         blood. However, other concentra-
tions of saponin (1-10              g/L) and EDTA (0.5-5 g/L) worked
equally well, as did the Gelman “Hemolyzate        Reagent”                            Buffer              0.025M        0.033M       O.05M           0.066M           0.075M
(saponin, 10 g/L, with no EDTA).                                                       Strength
   The concentration  of total hemoglobin in the hemolysate
had little     if any effect on results.         To assure   completeness      of    Fig. 3. Effect of varying the buffer strength (mol/L) on the appearance of
the hemolysis,         reduce       the     viscosity   of the sample,       and     the separation

       Table 1. Effect of Varying the Electrophoresis Conditions on Glycohemoglobin Results for Eight
                                                                         % glycohemogiobinat
                                                Temp, #{176}C                                      pH                                             Bufferconcn, mmoi/t.
conditions                            4                       22                     6.2                         6.6                             25                             50
      11.1                           11.4                    11.1                   11.0                        10.7                          11.4                            11.4
      24.0                           23.6                    23.0                   23.4                        24.4                          21.0                            21.2
       8.5                            8.1                     9.0                    8.5                         9.4                           7.7                             7.1
      13.3                           13.3                    13.7                   12.5                        13.4                          13.0                            13.4
      14.5                           14.0                    14.9                   14.3                        14.6                          12.4                            13.4
       8.5                            8.5                     8.4                    7.8                         8.8                           6.6                             7.2
      14.0                           13.7                    14.3                   13.4                        14.5                          11.9                            13.3
      12.0                           12.8                    12.2                   11.5                        13.2                          10.8                            11.6
      13.2                           13.2                    13.3                   12.8                        13.6                          11.9                            12.3
  Optimum       condition s are buffer. 33 mmol/L, pH 6.4 and temperature 18 C.

                                                                                                             CLINICAL CHEMISTRY, Vol. 29, No. 2, 1983                           341
found that glycosylated  hemoglobin S (HbS1) co-migrated                                                                            or entrapped    in the matrix of the gel. In the method we
with HbA1 and that the two nonglycosylated     fractions, HbA                                                                       describe,  the active ligand is in the mobile phase and no
and HbS, also co-migrated electrophoretically.                                                                                      special modification     of the support medium is necessary.
                                                                                                                                    This allows use of a simple procedure that is ideally suitable
Precision                                                                                                                           for routine use.
   Coefficients of variation of 4.9 and 3.8% were obtained for                                                                        The membrane we used, “Super Sepraphore,” is a Mylar-
16 replicate analyses of samples with 7.5% and 15% glycosy-                                                                        supported cellulose strip. We obtained better separation of
lated hemoglobin content, respectively. The precision of the                                                                       the bands with this particular          membrane      than with the
method was therefore satisfactory for the routine estimation                                                                       more usual type of nonsupported          membrane. Although the
of glycosylated hemoglobin in the clinical chemistry labora-                                                                       electrophoresis    could be carried     out in any electrophoresis
tory.                                                                                                                              chamber,    we used the Gelman Sepratek          System because it
                                                                                                                                   allowed the simple automatic application of eight samples at
Comparison with the Agar Gel Electrophoresis                                                                                       one time. Running four chambers from one power pack, we
Method                                                                                                                             could analyze 32 specimens per batch. Alternatively,               one
   The hemoglobin bands obtained by the new method were                                                                            might use the Gelman Semi Micro II Chamber, which takes
more compact and better defined than the agar gel electro-                                                                         three membranes.
phoresis pattern (Figure 4). Excellent correlation was ob-                                                                            We did not stain for protein in the initial development
mined between paired results of the two methods (r = 0.94),                                                                        work. Instead, a whole-blood hemolysate             was taken and
as shown in Figure 5, and results by the two methods can be                                                                        applied to the membrane. After electrophoresis           the proteins
used interchangeably.                                                                                                              were fixed in trichloroacetic       acid and the membrane         was
                                                                                                                                   dried in an oven. The separation           was scanned at 420 nm
Comparison              with the Short Column                                                  Method                              without protein staining, but the precision of densitometry
                                                                                                                                   was not good for a noncleared           membrane       and a normal
   The correlation    between 76 paired results by chromatog-
                                                                                                                                   glycosylated    hemoglobin      value. Because the membrane
raphy on short columns and by mobile affinity electrophore-
                                                                                                                                   could not be cleared        without protein staining,        we used
sis was also excellent (r = 0.94) when the columns were
temperature     controlled at 22 #{176}C.
                                      Operating the columns at                                                                     washed erythrocyte       hemolysates      to avoid interference     by
                                                                                                                                   serum protein staining.
room temperature        and correcting for temperature  did not
                                                                                                                                      It is becoming      increasingly   recognized   that   most   simple
produce such good results (r = 0.89).
                                                                                                                                   methods for measuring      total glycosylated     hemoglobin are
Discussion                                                                                                                         subject to interference  by an unstable     intermediate,    which
                                                                                                                                   may cause a spuriously high glycosylated percentage to be
  Affinity electrophoresis    is a relatively new concept and
                                                                                                                                   recorded, especially where grossly above-normal concentra-
hitherto  has usually    been carried out in agarose (14) or
                                                                                                                                   tions of blood glucose are present just before the blood is
polyacrylamide    gels (15) with the active ligand either bound
                                                                                                                                   withdrawn. The intermediate      may be removed by incubat-
                                                                                                                                   ing the blood in a glucose-free medium before estimation
           Application           -                                                                  electrophoresis                (16), a step that has been applied to the agar gel electropho-

                                                                                                                     t             resis method     (17) in which whole-blood
                                                                                                                                   normally used. Our preliminary
                                                                                                                                                                         results confirm that the

                                                                                                    -‘-Appl          ication       unstable intermediate     may be removed by incubating          the
                                                              A                  B                                                                   for
                                                                                                                                   cells at 37 #{176}C5 h in the saline wash before they are
                                                                                                                                   centrifuged      and hemolyzed.
                                                                                                                                      The new method performed well in comparison          with agar
Fig. 4. Electrophoresis on agar (A) and cellulose acetate (B) compared
                                                                                                                                   electrophoresis,      as expected, because the principles of the
      20                                                                                                             o         0   two methods are similar. In the agar electrophoretic         meth-
      19                                                                                                                           od, the nonglycosylated         hemoglobin  binds to the sulfate
                                                                                                                     o             groups in the medium and its mobility is retarded,       while the
                                                                                                                                   glycosylated      fraction is carried towards the cathode by
      17                                                                                                                           electroendosmosis.        In the cellulose acetate method, the
                                                                                          OO                o0
      16                                                                                                                           nonglycosylated       fraction combines with the sulfate groups
                                                                                                                                   on the carrier molecule and is separated from the nonglycos-
      15                                                               00         000
                                                                                                                                   ylated hemoglobin by electrophoresis.       A real advantage of
      14                                                                        Oc$
                                                                                                                                   our method is the economy of cellulose acetate membranes.
.13                                                                                   0                                            The total cost per test is about 20-30 p, compared with
=                                                        00             #{176}        0                                            about 80 p per test for the agar electrophoretic    method. The
      12                                             o
                                                                                                                                   new method, however, has the disadvantage            of involving
      11                                         0000
                                                                                                                                   more technical       steps, e.g., preparing washed cell hemoly-
      10                                   o8                     00

                                             0                                                                                     sates and protein staining. In comparison with the short ion-
       9                             %00                                                                                           exchange columns, the cellulose acetate method is easier to
                             0                                                                                                     use for large batches, has better precision, and shows no
       8                                                                                            R       0.94
                                                                                                                                   interference     from changes in temperature,      pH, or buffer
           -0                                                                                                                      concentration.
       6                                                                                                                              The principle may be applied to any electrophoretic
                I   I        I        I          I            I    I        I         I         I       I        I                 medium such as agarose or polyacrylamide       gel that does not
                6   7        8        9         10 11 12 13                           14 15 16 17 18 19 20                         bind the carrier molecule. Furthermore,     other active groups
                                                                                                                                   such as antibodies, enzyme substrates,    or lectins might also
                                                     Cellulose acetate % HbA1                                                      be bound by existing methods to dextran             sulfate. The
Fig.5. Comparison                of results obtained by the two electrophoretic                                                    charged nature of the carrier molecule could then succeed in
methods                                                                                                                            producing the desired specific separation.

342 CLINICALCHEMISTRY, Vol. 29, No. 2, 1983
   We thank Margaret Villiers, Department of Medicine, University         9. Gruenholz A, Wood R, Rogers P, Mangiardi V. Rapid quantita-
Hospital, Nottingham,    for carrying out all the column assays of        tive determination of hemoglobin A1 using ion exchange chroma-
glycosylated hemoglobin.                                                  tography. Clin Chem 25, 1075 (1979). Abstract.
                                                                          10. Jongeneel     J, Van Wissen M. Optimizing measurement              of
References                                                                glycosylated hemoglobins. Clin Chem 25, 642 (1979). Letter.
1. Trivelli LH, Ranney HM, Lai HT. Hemoglobin components in               11. Menard L, Dempsey ME, Blankstein LA, et al. Quantitative
patients with diabetes mellitus. NEngIJ Med 284,353-357(1971).            determination    of glycosylated hemoglobin A, by agar gel electro-
2. Bunn   HF, Shapiro R, McManus M, et al. Structural heterogene-         phoresis. Gun Chem 26, 1598-1602 (1980).
ity of human hemoglobin A due to nonenzymatic glycosylation. J            12. Aleyassine H, Gardiner RJ, Blankstein           LA, Dempsey ME.
Biol C/tern 254, 3892-3898 (1979).                                        Agar gel electrophoretic determination     of glycosylated hemoglobin:
3. Gabbay KH, Hasty K, Breslow JL, et al. Glycosylated hemoglo-           Effect of variant hemoglobins, hyperlipidemia, and temperature.
bins and long term blood glucose control in diabetes mellitus. J Clin     Clin C/tern 27, 472-475 (1981).
Endocrinol Metab 44, 859-864 (1977).                                      13. Parker KM. England JD, DaCosta J, et al. Improved colorimet-
4. Spicer KM. Allen RC, Buss MG. A simplified assay of hemoglo-           nc assay for glycosylated       hemoglobin. C/in Chem 27, 669-672
bin A1 in diabetic patients by use of isoelectric focusing and            (1981).
quantitative   microdensitometry.     Diabetes 27, 384-388 (1978).                            TC,
                                                                          14. B#{216}g-Hansen Bjerrum OJ, Ramlau J. Identification             and
5. Cole RA, Soeldner JS, Dunn PJ, Bunn HF. A rapid method for             quantitation   of glycoproteins by affinity electrophoresis. Anal Bio-
the determination    of glycosylated hemoglobins using high pressure      chem 81, 78-87 (1977).
liquid chromatography.      Metabolism 27, 289-301 (1978).                15. Horejsi V, Ticha M, Kocourek J. Determination of dissociation
6. Abraham EC, Hutt TA, Cape JD, et al. Determination            of the   constants of lectin sugar complexes by means of affinity electropho-
glycosylated (HbA1) hemoglobins with a new micro-column proce-            resis. Biochim Biophys Acta 499, 290-300 (1977).
dure. Diabetes 27, 931-937 (19781.                                        16. Goldstein DE, Peth S, England JD, et al. Effect of acute changes
7. Fluckeger R, Winterhatter        KH. In vitro synthesis of HbA1c.      in blood glucose on HbA,. Diabetes 29, 623-628 (1980).
FEBS Lett 71, 356-360 (1976).                                             17. Nathan DM. Labile glycosylated hemoglobin contributes to
8. Dix D, Cohen P, Kingsley S, et al. Evaluation of a glycosylated        hemoglobin A, as measured by liquid chromatography           or electro-
hemoglobin kit. Clin C/tern 24, 2073 (1978). Letter.                      phoresis. Gun C/rem 27, 1261-1263 (1981).

CLIN. CHEM. 29/2, 343-345      (1983)

Assessment of the Benzethonium Chloride Method for Routine
Determination of Protein in Cerebrospinal Fluid and Urine
Ellsabeth Flachaire,1Odlie Damour,2 Jacques Bienvenu,2Tarak Aoulti,2and Roger Later’

We have tested the characteristics of the method of Iwata                 that the   turbidity  produced  is the same for albumin and
and Nishikaze    (C/in Chem 25: 1317, 1979). The linearity,               gamma-globulins,     and is more stable than in Meulemans’s
sensitivity, and precision are satisfactory and the reactivity of         method (3) or in the method of Bossak et al. (4).
benzethonium chloride with various proteins (albumin, immu-                  We have evaluated their method, and we have compared
noglobulins) is the same. The method has been compared                    the results with those by the method we routinely use in our
with Meulemans’s technique (Gun Chim Acta 5: 757, 1960),                  laboratories,   the sulfosalicylic    acid method     (3).
routinely used in our laboratories, by analysis of 82 samples
                                                                          Materials and Methods
of cerebrospinal fluid (CSF) and 119 samples of urine. Our
results for cerebrospinal fluid agree well with those of Iwata            Preparation and Storage of Specimens
and Nishikaze (r = 0.976; y = O.992x - 0.013), but we find
                                                                             CSF and urine specimens were promptly centrifuged after
their method unsuitable for urinary protein determination,                collection. Protein was determined concurrently     by the two
probably because of interfering compounds in urine.                       methods, either immediately                                (the
                                                                                                         or after storage at -20 #{176}C
                                                                          specimen    being centrifuged just before analysis). Spectro-
AdditionalKeyphrase:turbidimetry                                          photometers     used were the Model 25 Beckman (at Pierre
                                                                          Benite2) and Model 1800 Philips (at Lyon’).
   We have been very interested         in the new turbidimetric
method described by Iwata and Nishikaze          (1) for determina-       Reagents
tion of proteins in cerebrospinal         fluid (CSF) and urine.             IwatalNishikaze       (I/N) method:    Benzethonium          chloride
Benzethoniuxn      chloride in alkali is used as a flocculation
                                                                          (Sigma Chemical Co., St. Louis, MO 63178) 2 g/L, sodium
reagent.   Iwata and Nishikaze       say that this method has a           hydroxide     0.5 mollL in 33 mmol/L tetrasodium           ethylenedi-
higher sensitivity    than the classic method of Lowry et al. (2)         aminetetraacetate       (EDTA) (Prolabo, Paris, France).
and   shows   satisfactory   reproducibility    and   recovery,    and
                                                                             Meulemans’s      method: (a) Dissolve 6 g of sulfosalicylic acid
                                                                          in 100 mL of distilled water. (b) Dissolve 14 g of anhydrous
   1 Service de Biologie (Professeur  Cl. Quincy), H#{244}pital
                                                           Neurologi-     sodium sulfate in 100 mL of water. Mix equal volumes of a
que, 59, Bd. Pinel, 69003 Lyon, France.                                   and b for use in the protein determination.
   2 Service de Biologie Secteur Jules Courmont    (A. Roullet.), Cen-       Standard      and control. Human    albumin     (crystallized      and
tre Hospitalier Lyon Sud, 69310 Pierre Benite, France.                    lyophilized;    Sigma A9511 and human globulin (Gamma 16;
   Received January 6, 1982; accepted November 15, 1982.                  Institut   Merieux, Marcy l’Etoile, France).

                                                                                            CLINICAL CHEMISTRY, Vol. 29, No. 2, 1983 343

To top