CLIN. CHEM. 38/12, 2365-2371 (1992)

CLINICAL CHEMISTRY, Vol. 38, No. 12, 1992 2365

lnterlaboratory Study of the IFCC Method for Alanine Aminotransferase Perlormedwith Use of a Partly Purified Reference Material

Fran#{231}oise Schiele,’ Jocelyne Muller,’ Elisabeth Colinet,2 Gerard Siest,’ P. Arzoglou,3 H. Brettschneider,4D. H. Calam,5 Ferruccio Ceriotti,6, Georges F#{233}rard,7 Jorg Frei,8 A. Hadjivassiliou,9 J. C. M. Hafkenscheid,lO�tMogens H#{248}rder,” F. Javier GeHa,’2 Robert Rej,’3 Ellen Schmidt,’4 Anne Vassault,’5 and V. Viette’6

We present the results of a study on performance of areference material for alanine aminotransferase (ALT, EC2.6.1 .2) and the corresponding IFCC-approved method inan interlaboratory trial involving 13 laboratories. The ALTmaterial was partly purified from pig heart (specific activ-ity, 1 50 kU/g) and was essentially free of six potentiallycontaminating enzyme activities, including aspartate ami-notransferase (EC 2.6.1 .1). The partly purified ALT waslyophilized in a triethanolamine-buffered matrix, pH 6.4,

containing bovine serum albumin and saccharose. Underthese conditions, the predicted yearly loss of activity was0.02% at 4 #{176}Cand <0.01% at -20 #{176}C.The final blank-corrected results of the accepted set of data gave a mean(SD) of 128.5 (5.1) U/L. The among-laboratory SD was4.6 U/L and the within-laboratory SD was 2.0 U/L. Thecertified ALT catalytic concentration in the reconstitutedmaterial was 1 29 U/L with a 0.95 confidence interval of ±

4 U/L.

1 Laboratoire du Centre de M#{233}decinePreventive et Centre du

M#{233}dicament, UA CNRS N#{176}597, 30 rue Lionnois, 54000 Nancy,

France.

2 Commission des Communaut#{233}s Europeennes, BCR, rue de laLoi 200, 1049 Bruasels, Belgium.

3 University of Salonika, Salonika, Greece.4 Boehringer Mannheim GmbH, Penzberg, Germany.5 National Institute for Biological Standards and Control, Pot-

tore Bar, UK.6 Istituto Scientifico, HS Raffaele, Milan, Italy.

7 Universit#{233} Louis Pasteur, Strasbourg, France.8 CHU Vaudois, Lausanne, Switzerland.9 Metaxas Memorial Cancer Hospital, Piraeus, Greece.10 Sint Radboudzienkenhuis, Nijmegen, The Netherlands.11 Odense University Hospital, Odense, Denmark.12 Umversitat Autonoma, Barcelona, Spain.

13 Wadsworth Center for Laboratories and Research, Albany,NY.

14 Medizinische Hochschule, Hannover, Germany.15 H#{244}pital Necker-Enfants Malades, Paris, France.

16 HOpital de la Chau.x-de-Fonds, La-Chaux-de-Fonds, Switzer-

land.t deceased.‘7 Nonstandard abbreviations: ALT, alanine aminotranaferase

(L-alanine:2-oxoglutarate aminotransferase, EC 2.6.1.2); AST, as-partate aminotransferase (L-aspartate:2-oxoglutarate aminotrans-ferase, EC 2.6.1.1); IFCC, International Federation of ClinicalChemistry; ERM, enzyme reference material; NIBSC, NationalInstitute for Biological Standards and Control; EC BCR, Commu-nity Bureau of Reference of the European Communities; DEAE,thethylaininoethyl; BSA, bovine serum albumin; CM, carboxyme-thyl; PBE, polybuffer exchanger; PP, pyridoxal 5’-phosphate; andSFBC, Soci#{233}t#{233}Francaise de Biologie Clinique.

Received December 10, 1990; accepted July 24, 1992.

Serum alanine aminotransferase (ALT, EC 2.6.1.2) is,along with aspartate aminotransferase (AST, EC

2.6.1.1), one of the most frequently measured enzyme

activities in clinical laboratories.17 The importance of

ALT has also increased because it is commonly deter-

mined in blood banks to identify blood samples poten-

tially contaminated with non A-non B hepatitis virus

(hepatitis C).

In 1980, the Expert Panel on Enzymes ofthe Interna-

tional Federation ofClinical Chemistry (IFCC) proposed

a method for measuring ALT activity in human serum;this method was officially approved in 1985 (1 ). Inter-

national cooperation has led to enzyme reference mate-

rials (ERMs) being available for most of the currently

used enzymes in clinical chemistry. The procedure for

value assignment for catalytic concentration of theERMs uses the most accurate and reliable measurement

technique available through strictly defined and con-

trolled interlaboratory studies. However, there is rela-

tively little information concerning the transferability

of the IFCC ALT method and, to our knowledge, no

interlaboratory study using purified ALT has been per-

formed, although this was done for AST (2, 3) and

alkaline phosphatase (4, 5).

Therefore, we report here the results of a study of anERM for ALT where the IFCC-recommended method

was used in an interlaboratory trial run under the

auspices of the Commission of the European Communi-

ties. The objectives of this study were to test the possi-

bility of a partly purified ALT being an ERM and to

assign a value to the material by using the IFCCmethod for ALT measurement through an intenlabora-tory study.

MaterIals and Methods

Instrumentation

The spectrophotometers used in the different pantici-

pating laboratories, grouped by temperature-monitor-

ing system, were (a) those with a thermistor in the

reaction solution: Beckman Model DU 8 (bandpass, 0.5nm; Beckman Instruments, Fullerton, CA); Cary Model

219 (bandpass, 1 nm; Vanian Instrument Div., Palo

Alto, CA); Eppendorf 1101 M (bandpass <1 nm; Eppen-dorf GerAtebau Netherlert Hinz GmbH, Hamburg, Ger-

many); Kontron Uvikon Models 860, 930, and 940

(bandpass, 2 nm; Kontron Instruments, Everett, MA);

2366 CLINICAL CHEMISTRY, Vol. 38, No. 12, 1992

and Shimadzu UV 190 (bandpass, 2 nm; ShimadzuEuropa GmbH, Duisberg-Grossenbaum, Germany); (b)those with a thermistor mounted in the wall of thecuvette holder: Cary Model 219 (bandpass, 1 nm) and

Gilford Stasar Ill (bandpass, 8 nm; Gilford InstrumentLabs. Inc., Oberlin, OH); and (c) one with a thermistorin water bath: Beckman Model 25 (bandpass 1 nm).

Chemicals

Potassium phosphate, disodium EDTA, ammonium

sulfate, potassium chloride, imidazole, glycine, histi-dine, Ti-is, saccharose, and triethanolamine were ob-

tamed from Merck (Darmstadt, Germany). Dithiothrei-

tol and bovine serum albumin (BSA; purity 98-99%, cat.no. A-7030) were purchased from Sigma Chemical Co.

(St. Louis, MO). Diethylaminoethyl (DEAE)-tris-acryland carboxymethyl (CM)-tris-acryl were from Industrie

Biologique Francaise (Paris, France). Ultrogel ACA 34

was purchased from LKB Produkter AB (Bromma, Swe-den). Polybuffer exchanger (PBE) and cyanogen bro-mide-activated Sepharose 4-B were obtained from Phar-

macia (Uppsala, Sweden). Sodium chloride was fromProlabo (Paris, France), and pyridoxal 5’-phosphate

(PP) and 2-oxoglutaric acid were from BoehringerMannheim (Meylan, France).

Chemicals used for measuring catalytic concentration

of ALT (provided by Boehringer Mannheim GmbH,Mannheim, Germany) were supplied to all participating

laboratories. One batch of each reagent [Tris buffer,L-alanine, 2-oxoglutaric acid, disodium NADH Grade I

(purity, 100%), lactate dehydrogenase from pig skeletal

muscle in a glycerol matrix, and PP] was used for thestudy. In addition, each laboratory was provided with a

single batch of D-alanine (purity >99%, D:L >99.5:0.5;

Fluka AG, Buchs, Switzerland) for sample blank-reac-tion measurements. The purity of the reagents was

controlled and was shown to meet the requirements ofthe IFCC-approved method (1 ). Reagent mixtures were

prepared in each laboratory according to the IFCC

method.To assess the linearity of the spectrophotometers in

the absorbance range 0.39-1.64 at 339 run and to revealany gross errors in their accuracy, each laboratory used

a commercially stabilized solution of NADH (UV-Trol;

BioM#{233}rieux, Charbonni#{232}res-les-Bains, France).

Methods

We measured ALT activity during the preparationand characterization of the ALT material with a Cary

Model 219 spectrophotometer or with a Cobas Bio cen-trifugal analyzer (Hoffmann-La Roche, Base!, Switzer-

land) with kits from Boehringer Mannheim, at 30 #{176}C(1 ). Protein concentrations were determined by the

method of Lowry et al. (6) with BSA as standard.We examined the partly purified ALT for various

contaminating enzyme activities by using commerciallyavailable kits. We used kits from BioM#{233}rieux for mea-surements of alkaline phosphatase (EC 3.1.3.1), cre-atine kinase (EC 2.7.3.2), lactate dehydrogenase (EC1.1.1.27), and AST activities according to the methods

recommended by the Societe Fran#{231}aise de Biologie Cli-nique (SFBC) (7) and for acid phosphatase (EC 32.3.2)

activity according to Hillmann’s method (8). ‘y-Glu-

tamyltransferase (EC 2.3.2.2) activity was determinedwith a kit from Boehringer Mannheim (9).

Preparation of the ALT Material

ALT was partly purified from pig heart according tothe procedures ofKojima (10) and Noguchi et al. (11 ), as

modified in our laboratory. One part tissue (200 g) was

homogenized with three parts phosphate buffer (50mmol/L, pH 6.0) containing, per liter, 0.2 mmol ofEDTA, 0.1 mmol ofdithiothreitol, 0.05 mmol ofPP, and

5 mrnol of 2-oxoglutaric acid in a Waring Blendor atmaximal velocity for 3 mAn (3 x 1 mm) at 4 #{176}C.Themixture was heated to 65 #{176}Cand kept at that tempera-

tare for 1 mm, immediately cooled at 4 #{176}Cin an ice bath,

and centrifuged at 12 000 x g for 20 mm at 4 #{176}C.The

supernate was precipitated with 20-55% ammoniumsulfate saturation followed by centrifugation at 12 000x g for 20 mm at 4 #{176}C.The pellet was solubilized in a

solution containing imidazole (25 mmolIL, pH 7.6) and

EDTA (5 mmolJL) and dialyzed overnight at 4 #{176}C

against the same buffer.The material was chromatographed at room temper-

ature on a 5 x 40 cm column ofDEAE-tris-acryl, elutedwith a linear gradient of KC1 (0-0.3 molfL) in the

imidazole buffer, and dialyzed at 4 #{176}Cagainst phosphatebuffer (50 mmol/L, pH 6.8).

The active fractions were concentrated at room tern-perature by ultrafiltration in a stirred cell (ArniconCorp., Lexington, MA) with a PM 10 membrane, dia-

lyzed at 4 #{176}Cagainst a histidine buffer (25 mmol/L, pH6.2), and filtered at 4 #{176}Con a 2.5 x 100 cm column of

Ultrogel ACA 34 equilibrated with Tris HC1 buffer (25mmol/L, pH 7.5). Next the material was chrornato-

graphed at room temperature on a 1 x 6 cm immunoaf-finity column prepared from anti-cytoplasmic AST an-

tibodies that we produced in our laboratory and bound

to a bromide cyanogen-activated Sepharose-4B columnin NaHCO3 buffer (100 mmolfL, pH 8.3) containingNaC1, 50 mmol/L. After equilibrating the column withphosphate buffer (100 mrnolIL, pH 7.0) containing NaClat 150 mmolIL, we collected directly the fractions con-taming ALT activity by elation with the same buffer.The remaining AST in the preparation was retained onthe column and eluted with a glycine HC1 buffer (20mmol/L, pH 2.5).

We analyzed for potentially contaminating enzymes

at different stages during the purification process. Thepartly purified ALT was added to the following matrixto yield a catalytic concentration of -�130 UIL: per liter,

30 g of BSA, 10 g of saccharose, and 50 mmol oftriethanolamine hydrochloride, pH 6.4. Before dilution,we measured the activities of possible contaminatingenzymes in the BSA. The liquid material was frozen and

shipped on solid carbon dioxide to the National Institutefor Biological Standards and Control (NIBSC, Potters

Bar, UK) for lyophilization. The enzyme solution wasthawed at 4 #{176}C,then filtered at the same temperature

Table I . Partial PurifIcatIon of ALT from PIg Heart

1400 141 9.93 32 66

740 46 16.1 52 35

555 15 36.8 119 26

470 3.2 147 474 22

CLINICAL CHEMISTRY, Vol. 38, No. 12, 1992 2367

through two 0.22-pin pore-size membrane filters (Santo-rius Ltd., Epsom, UK). It was dispensed at 4 #{176}Cinto

ampules at a nominal volume ofl.0 mL per ampule. The

ampules were processed as a single batch essentially as

described by Campbell (12) by using a Minifast freeze-

drier (Edwards High Vacuum, Crawley, UK). Lyophili-

zation commenced at a shelf temperature of -50 #{176}C.When it was complete, the batch was submitted to

secondary desiccation over phosphorus pentoxide for 5days. The ampules were then filled with pure, dry

nitrogen and sealed by fusion of the glass. The total

number of ampules (1 mL) so prepared was 1768.

Characterization of the ALT Material

The precision and accuracy of dispensing the ALT

liquid material was monitored by weighing 34 arnpules

taken at regular intervals throughout the ifiling proce-dure and before and immediately after dispensing the

1-mL volumes. The residual moisture content of the

lyophilized material was determined in three ampules

by an automated Fischer method (13). The uniformity(ampule-to-ampule variability) of the lyophilized prod-

uct was assessed by measuring ALT catalytic concen-

tration in 20 different ampules picked at random.

The long-term stability studies of the lyophilized

preparation were coordinated by NIBSC and were con-ducted in the same manner as for -y-glutamyltransferase

Certified Reference Material 319 by using an acceler-ated thermal degradation procedure (14). ALT catalyticconcentration was measured after 65 and 189 days of

storage at different temperatures. Each ampule was

assayed in duplicate with and without PP, and thepredicted loss of activity was calculated by NIBSC (14).

The stability of the reconstituted material at room

temperature during a working day was examined bymeasuring ALT catalytic concentration in duplicate bythe SFBC method (7) with a Cobas Bio centrifugal

analyzer at 1-h intervals for 8 h.

We examined some of the major kinetic properties ofthe partly purified ALT. We determined the apparentMichaelis constants (Km) for L-alarnne and 2-oxogluta-rate at 30 #{176}Cby using the IFCC method except forvarying the substrate concentration.’8 Simultaneous

measurements were performed in duplicate with theALT material and with a pooled specimen of human

serum. L-Alanine concentrations were 25-700 mmol/L,

and 2-oxoglutarate concentrations were 1-24 mrnol/L.We also tested the effect on ALT activity of preincuba-tion with PP, 0.1 mmol/L, for 15 mm.

Interlaboratory Study

Thirteen laboratories were involved in the interlab-

oratory study: three in France, two in Germany, two inGreece, and one each in Denmark, Italy, Spain, Swit-

zerland, The Netherlands, and the United States. All

laboratories were familiar with the IFCC method for

18 � the published IFCC method (p. 483, ref. 1 ), the catalytic

concentration oflactate dehydrogenase (solution V) should be 2.52x 106 nkat x L’.

Homogenateb

Heat treatment

Pptn with(NH4)2S04(20-55%saturation)

Chromatography onDEAE-tns-acryl

Chromatography onCM-tns-acryl

Gel filtration onUltrogel ACA 34

Immunoaffinftychromatography

TotalcatalytIcacty., U

Totalprotsinmass,

mg

SpecIficcatalytic

acty.,kU/g

Purtfl-cation,

n-foldV1.Id

%�

3000 - - - -

2120 6775 0.31 (1) 100

1 700 1 180 1 .44 4.7 80

a Percent of total catalytic activity.b For 200 g of fresh material.

measuring catalytic concentration of ALT because theyhad been involved in a preliminary analytical runduring a feasibility study with a pilot batch of ALT

material.

Each participating laboratory was provided with five

ampules ofALT material, a set ofreagents, and a set of

documents, including a copy of the IFCC method formeasuring ALT catalytic concentration, a protocol forreconstituting the lyophilized material, a protocol forchecking instrumentation, a form requesting detailsconcerning instruments and ALT-material reconstitu-lion, and a sheet for reporting results. Each laboratorywas required to perform duplicate analyses on 3 differ-

ent days, a different ampule being freshly reconstituted

and analyzed on each day.The Community Bureau ofReference ofthe European

Communities (EC BCR) collected all data for statistical

evaluation.

Results

ALT Material

Our procedure for purification of ALT from pig heart

yielded a 470-fold-purified enzyme with a specific activ-

ity of �-150 kU/g protein (Table 1). y-Glutamyltrans-

ferase and lactate dehydrogenase activities were not

detectable after the DEAE-tris-acryl chromatography.After the gel filtration on Ultrogel ACA 34, activities of

alkaline and acid phosphatases were not measurable;creatine kinase activity was 1 U per 100 U ofALT. Thefinal product after the immunoaffinity chromatography

with anti-cytoplasmic AST antibodies contained only0.1 U of AST activity per 100 U of ALT.

The apparent Km values obtained for L-alanine were49.1 and 27.9 mmol/L for the pig-heart ALT and thehuman serum pool, respectively. For 2-.oxoglutarate, theapparent Km values were 041 and 1.26 mmol/L, respec-tively.

We observed an increase of -35% ofthe ALT activity

after preincubating the partly purified enzyme with PP,

lized specimens, the participants weighed the ampules

before and immediately after adding 1 mL of distilled

water with a class A volumetric pipette (15). The

ActIvIty %� variation in the reconstitution of the lyophiuized sam-Catalytic cone pies showed an overall mean value of 1.0035 g with an

20 ‘C 37 #{176}C 45 #{176}C 56 �C at -20 C, U/L interlaboratory SD of 0.0087 g (extreme values: 0.9881

and 1.0258 g).- 96 85 31 129 All participants used the recommended wavelength of99 89 58 - 138 339 nm except that one laboratory used 334 nm (Hg

line). Appropriate corrections were made for this labo-- 96 85 30 125 ratory’s final results. The spectral bandpass used by the99 90 57 - 129 participants was generally �2 nm. All participants

monitored the reaction temperature by using a ther-_________________________________________ mister directly in the reaction solution and (or) mounted

in the wall of the cuvette holder, except for one labora-

tory. The results ofthis laboratory agreed well with theothers, however.

Six of the laboratories calculated automatically theabsorbance change per minute. A recorder was used byone participant, and digital readout and stopwatch wereused by four laboratories. The IFCC method recorn-mends a sample volume of 200 p.L for a final volume of2 mL. The within-laboratory CV for the sampling p1-pette derived from the mass of water delivered during

eight pipettings, was between 0.11% and 0.52%. Partic-ipants in this study reported the values for change inabsorbance per minute (&4/�t) measured for the overallreaction with complete reagent (A) and for the three

blank reactions given in the IFCC method (1 ): a reagentblank for the overall reaction (B), in which sodiumchloride (154 mrnol/L) is substituted for the specimen;an overall sample blank (C), in which D-alanine issubstituted for L-alanine; and a reagent blank for the

sample blank (D), in which D-alanine is substituted forL-alanine and sodium chloride (154 mmol/L) is substi-bited for the specimen. The blank-corrected change inabsorbance per minute was then calculated from these

four measurements as follows (1):

2368 CLINICAL CHEMISTRY, Vol. 38, No. 12, 1992

Table 2. Effect of Storage Temperature on theLyophlllzed ALT Material as Measured with and without

Prelncubatlon with Pyridoxal 5-Phosphate

Storage,

days

With PP

65189

Without PP

65189

a Expressed as a percentage of the activity measured in ampules stored at

-20 #{176}Cfor the same time and determined in the same assay.

0.1 mmolfL, for 15 mm. This increase reached an equi-librium within 5-7 mm. This could indicate that the

partly purified enzyme is predominantly holoenzyrne.

We added the partly purified enzyme to the matrixcontaining BSA (40 gIL). No activities of aspartate and

alanine aminotransferases, lactate dehydrogenase, cre-

atine kinase, alkaline phosphatase, and y-glutarnyl-transferase were measurable in the BSA solution, andacid phosphatase activity was <0.8 UIL.

The 1-mL samples of the liquid ALT material dis-pensed during the lyophilization procedure had a mean

mass of 1.0028 g (range 1.0023-1.0033 g, n = 34). Theresidual moisture content of the ampules was 0.025%.The total variability (CV) ofthe ALT activity measuredin 20 reconstituted specimens was 1.3%. By a one-wayanalysis of variance, the between-ampule component(SD = 0.44 U/L) was substantially smaller than the

within-ampule (SD = 1.75 U/L) component of variance.

The effect of storage temperature on the lyophilizedspecimens is given in Table 2. The results after 65 and189 days of storage at various temperatures were rea-sonably consistent with the Arrhenius model. Relativeto results for samples stored for the same times at-20 #{176}C,the predicted annual loss ofcatalytic concentra-tion was -0.02% at 4 #{176}Cand <0.01% at -20 #{176}C,regard-less of whether the ALT catalytic concentration was

measured with or without PP. This lack of difference

means that the molecule does not lose any of its cata-lytic properties under such conditions.

The reconstituted material was stable throughout a

working day (i.e., for 8 h at room temperature), theobserved CV of the measured ALT catalytic concentra-tion being within the limits of the analytical variabilityof the method used.

lnterlaboratory Study

Thirteen laboratories originally submitted results.

One participant withdrew his contribution because of alinearity problem with the reaction. A second laboratoryencountered problems with the temperature-control sys-tern and the results were obtained at an assay temper-ature that was too low. The results of these two labora-tories were not included in the evaluation of theinterlaboratory results.

To detect errors during reconstitution of the lyophi-

(�A)/i�t)corrected = [(�A)/�t)A - (�A)/�t)B]

- [(M/�t)� - (M/�t)DJ

Final ALT catalytic concentration was then calculatedby using the value of the molar linear absorbance ofNADH given in the IFCC method (630 m2/mol) (1),

except for the laboratory making its measurement at334 nrn (which used 618 m2/mol).

Table 3 gives the corrected results by laboratory,ampule, and duplicate. The laboratories are listed in theorder of their numerical mean results. The final cor-rected results showed a mean (SD) value of 128.5 (5.06)

UIL. The within-laboratory SD, including the day-to-day variability, was 2.36 U/L. The between-laboratoriesSD and between-days SD were 4.58 and 1.97 U/L,respectively, as given by a two-way analysis of variance.

Table 4 compares the results from this interlaboratory

study with those obtained in the feasibility study in-volving a pilot batch of ALT material prepared in the

Table 3. Final Corrected Results for ALT CatalyticConcentration from Participating Laboratories

DaIly values, U/LMsan (SD),

Un-

120.0 (2.28)cv, %

I .9

122.7 (0.68) 0.6

Mean

Table 4. Comparison of the Sources of Varlablifty Inthe Preilminary and Final interlaboratory Studies

Between-day

Within-day 1.59

Final study

SDofactlvlty,WL CV,%

4.58 3.61.97 1.5

1.2

PrelImInary studyaSource ofvariatIon SD of actlvfty, u/L CV, S

Between-lab 8.10 7.0

3.31 2.8

1.4 1.57

Variance components were calculated from duplicate measurements of oneampule on 3 days in 13 part�padng laboratoiles for the preliminary study and11 laboratories for the final study. Degrees of freedom for mean squares forthe preliminaiy study: between lab, 12; between day, 26; and overall, 77.Degrees of freedom for mean squares for the final study: between lab, 10;between day, 22; and overall, 65.

a fl� preliminary Interlaboratory study used a partly purified materialprepared in the same way as the material used for the final study and havinga similar catalytic concentration (1 16.5 U/L).

same way as the reference material. Clearly, results

obtained in this second interlaboratory trial are greatlyimproved from those of the first trial, both for between-laboratory and between-day variability.

CLINICAL CHEMISTRY, Vol. 38, No. 12, 1992 2369

Lab Dayl Day2 Day3A 118.9

123.4

120.4

118.9

116.8

121.3

B 123.7

123.3

122.1

122.9

122.3

122.0

C 124.1 122.6 124.1

124.1 124.1 125.1

0 122.5 128.8 126.5

122.7 130.1 122.2

E 128.0 130.3 131.5

128.4 127.6 132.6

F 130.8 129.0 129.5

131.2 125.4 132.6

G 129.6 127.7 131.5

131.5 125.8 133.5

H 133.1 130.8 132.9133.3 131.6 133.5

135.8 131.8 130.6

134.6 130.6 132.3

J 135.5 136.3 129.6

137.1 131.0 127.0

K 134.2 136.2 135.1132.6 134.1 134.1

Statistical analysis (15) yielded no outlying valueamong the set of results accepted on technical grounds.

Because the analysis of variance showed that the be-tween-laboratory variability was the major source ofvariability, each laboratory’s mean value (and not the

individual results) was used to calculate the certified

value and uncertainty (16). The certified ALT catalytic

concentration ofthe reconstituted (with 1 mL of distilled

water) material and the corresponding uncertainty at

the 0.95 confidence level was 219 ± 4 UIL. This material

124.0 (0.80) 0.7 15 now available through the EC BCR as CRM 426.

Discussion

125.5 (3.49) 2.8 In this interlaboratory study we used partly purified

ALT from pig heart. Pig heart ALT was said to have

129 7 (2 05) 1 6 catalytic properties very similar to those of the humanserum enzyme (1 7), as was also shown by our results.Particular attention was paid to the absence of poten-

129.8 (2.49) 1.9 tially contaminating enzyme activities. Although thespecific activity that we obtained (150 kU/g protein) wasfar from that reported for preparations from the same

129.9 (2.82) 2.2 organ by Saier and Jenkins (18) and Kojima (10) (i.e.,498 and 340 kU/g protein, respectively), our preparationwas free ofthe main potentially contaminating enzyme

132.5 (1.08) 0.8 activities. The major difficulty was, in fact, to obtain thepartly purified ALT free of AST, because these two

enzymes have very similar chemicophysical properties132.6 (2.14) 1 .6 (i.e., isoelectric point and relative molecular mass). This

separation was achieved by using immunochromatogra-phy with antibodies directed against the cytoplasmic

132.8 (4.13) 3.1 AST isoenzyme.

Another difficulty was to obtain a stable material.

Some interlaboratory studies deathbed difficulties

134.4 (1 .20) 0.9 linked to the instability ofthe enzyme preparations (19,20). We chose a triethanolamine-buffered matrix at pH

1285 506 39 6.4,containingBSAandsaccharose.ThispHvaluewas. ( . ) . selected because purified rat-liver ALT includes sulihy-

dryl groups, some of them directly involved in theenzyme activity (21 ). Similarly, a relatively acidic pHvalue was shown to ensure good storage stability of

creatine kinase isoenzymes (22), which also containthiol groups. The partly purified ALT was predomi-nantly saturated with its cofactor, as shown by the factthat incubation with PP increased its activity by <5%;

this increase was completed within the 10-mAn preincu-bation time required by the IFCC method (1 ). In addi-tion, previously described interaction between albumin

present in the AST RM 8430 samples and PP was said tobe responsible for a 339-nm absorbance change both inthe overall reaction and in the sample blank reaction for15 mm (3, 23). Therefore, we preferred not to add PP inthe matrix to avoid possible sources of imprecision inmeasuring ALT catalytic concentration, specifically forblank reactions.

The ALT material appeared suitable in terms ofstability, uniformity, and contaminating enzyme activ-

ities for use in an interlaboratory trial. The ampule-to-

ampule variability for ALT catalytic concentration in

the reconstituted material was 1.3%. The predicted

2370 CLINICAL CHEMISTRY, Vol. 38, No. 12, 1992

change in ALT catalytic concentration in the lyophilized

material at -20 #{176}Cwas <0.01% per year.

The ALT catalytic concentration of the material wasdetermined by using the IFCC-approved method (1 ). All

the participating laboratories were involved in a prelim-mary trial. This preliminary trial pointed out some

problems encountered with the implementation of theIFCC-approved method, in particular with the purity ofreagents, especially the D-alanine, and with participat-

ing laboratories not being sufficiently familiar with themethod. Possible spurious values of sample blankcaused by contamination of D-alanine with L-alanine

were previously noted by Okorodudu et al. (24). We

measured the overall sample blank reaction with D-ala-nine that was �99% pure from six different manufac-

turers. The values obtained ranged from 2.1 to 11.4 UIL.

We decided to use one lot of reagents for this interlab-oratory trial and to test all reagents’ purity. In addition,

to ensure that the laboratories were themselves suffi-ciently familiar with the IFCC method, the participants

in this second trial were the same as those involved inthe feasibility study. However, most of the published

interlaboratory data from IFCC methods has needed

several trials before acceptable between-laboratoriesvariations were obtained.

The importance of preinstrumental and instrumental

errors in the measurement of catalytic concentration of

enzymes has been stressed in many other interlabora-

tory studies (2-5, 15, 25). We used a well-defined proto-

col to minimize and possibly to identify all the sources ofvariation in the measurement procedure. The volumes

delivered by the sampling pipettes, derived from the

weight ofdistilled water delivered during pipetting (mL

= mass/density corrected for temperature), ranged from

198.5 to 202.0 p.L (99.25-101.00% of the expected val-ue). Each laboratory used its own sampling volume

when calculating the ALT catalytic concentration. Thevariation in the reconstitution of the lyophilized sam-

ples was estimated to be ± 0.86%. The total variabilityofthe results accepted for certification was 3.9%. Bowers

et al. (26) obtained a CV of 7.4% but for a much lowermean value (24.2 U/L) and without sample blank cor-

rection with lyophilized human serum material NBS

SRM 909 (National Institute of Standards and Technol-

ogy, Gaithersburg, MD). They identified between- and

within-laboratory CVs, including the day-to-day van-

ability, of6.5% and 3.5%, respectively. In this study the#{149}respective CVs were 3.6% and 1.8% and are comparablewith those obtained in other interlaboratory studiesinvolving IFCC methods. The between-laboratory CVswere 2.2% for AST (3), 2.6-6.3% (4) and 4.3% (5) for

alkaline phosphatase, and 3.1% for y-glutamyltrans-ferase (15). The within-laboratory CVs, including theday-to-day variability, were 1.2% for AST, 1.2-3.6% and

1.6% for alkaline phosphatase, and 1.6% for -y-glutamyl-

transferase.The small within-laboratory CV (1.8%) including the

day-to-day variation (CV = 1.5%) is evidence of the

excellent reproducibility achievable by the IFCC

method. The variation between laboratories (CV =

3.6%) reflects the current state of the art. The consist-ency of the results between laboratories in severaldifferent countries, including the United States, clearlydemonstrates the transferability of the IFCC-approvedmethod for measuring ALT catalytic concentration.

We are very indebted to R. Games Das (National Institute forBiological Standards and Control, Potters Bar, UK) for calculatingthe predicted degradation rates ofthe material. We thank C. Cossyand M. P. Recouvreur for their excellent technical assistance.

References

1. Bergmeyer HU, H#{248}rder M, Rej R. Approved recommendation(1985) on IFCC methods for the measurement ofcatalytic concen-tration of enzymes. Part 3. IFCC method for alanine aminotrans-ferase (L-alanine:2-oxoglutarate aminotranaferase, EC 2.6.1.2). JClin Chem Clin Biochem 1986;24:481-95.

2. Sampson EJ, Whitner VS, Burtis CA, McKneally SS, Fast DM,Bayse DD. An interlaborathry evaluation ofthe IFCC method foraspartate aminotransferase with use ofpurified enzyme materials.Clin Chem 1980;26:1156-64.3. Bowers GN Jr, McComb RB, Syed D, et al. Primary humanaspartate aminotransferase material. The catalytic activity con-centration ofAST in RM 8430 as measured by the IFCC Reference

Method for AST in 10 enzyme standardization laboratories. ClinChem 1988;34:450-9.4. Tiet.z NW, Rinker AD, Burtis C, et al. Transferability studiesfor the AACC reference method and the JFCC method for mea-surement of alkaline phosphatase activity. Cliii Chem 1984;30:704-6.5. Schiele F, Muller J, Colinet E, Siest G. Certification of anenzyme reference material for alkaline phosphatase (CRM 371).Clin Biochem 1991;24:159-.68.6. Lowry OH, Rosebrough NJ, Farr AL, Randall 1W. PrOteinmeasurement with the Folin phenol reagent. J Biol Chem 1951;193:265-75.7. Soci#{233}t#{233}Fran#{231}aisede Biologie Clinique. Travaux et recomman-dations dela Commission Enzymologie. Ann Biol Clin 1982;40:87-164.8. Hilhnann G. Fortlaufende photometrische Messung der aaOrenPhosphatase Aktivitat. J Clin Chem Clin Biochem 1971;9:273-4.

9. Persijn PJ, Van der Silk WA. A new method for the determi-nation ofgamma-glutamyltransferase in serum. J Clin Chem ClinBiochem 1976;14:421-7.

10. Kojima H. Large-scale preparation ofcrystalline alanine ami-notransferase from pig heart and some of its properties. Kuma-moth Med J 1979;32:23-34.11. Noguchi T, Takada Y, Kido R. Glutamate-glyoxalate ami-notransferase in rat liver cytosol. Purification, preperties andidentity with alanine-2-oxoglutarate aminotransferase. Hoppe-

Seyler’s Z Physiol Chem 1977;358:1533-42.12. Campbell PJ. International biological standards and referencepreparations. II. Procedures used for the production of biologicalstandards and reference preparations. J Biol Stand 1974-,2:259-67.13. Fischer K. Neues Verfahren zur massanalytiachen Bestim-mung des Wassergehaltes von Flussigkeiten and festen Korpen. ZAngew Chem 1935;8:394-6.14. Schiele F, Muller J, Colinet E, Siest G. Production andcertification ofan enzyme reference material for gamma-glutamyl-transferase (CRM 319). Part 1. Preparation and characterization.Clin Chem 1987;33:1971-7.15. Schiele F, Muller J, Colinet E, Siest G. Production andcertification ofan enzyme reference material for ga.mma-glutamyl-transferase (CRM 319). Part 2. Certification campaign. Clin Chem1987;33:1978-82.16. Marchandise H, Colinet E. Assessment of methods of assign-ing certified values to reference materials. Fresenius Z Anal Chem1983;316:669-72.17. Gruber W, MOllering H, Perras L. Isolation, pH optima andapparent Michaelis constants of highly purified enzymes fromhuman and animal sources. J Clin Chem Clin Biochem 1977;15:565-73.18. Baler MH Jr, Jenkins WT. Alanine aminotransferase. 1.Purification and properties. J Biol Chem 1967;242:91-100.

CLINICAL CHEMISTRY, Vol. 38, No. 12, 1992 2371

19. Jansen RIP, Jansen AP. Standards versus standardized meth-ods in enzyme assays. Ann Clin Biochem 1983;20:52-9.20. Bullock DG, Moss DW, Whitehead 1?. EXternal quality as-sessment of serum enzyme activity assays and the effect of cali-bration on interlaboratory concordance. Ann Clin Biochem 1986;23:577-84.21. Matsuzawa T, Segal HL. Rat liver alanine aminotransferase.Crystallization, composition, and role of sulfhydryl groups. J BiolChem 1968;25:5929-34.22. Nealon DA, Pettit SM, Henderson AR. Effect of serum pH onstorage stability and reaction lag phase of human creatine kinaseisoenzymes. Clin Chem 1980;26:1165-9.23. Syed D, McComb RB, Bowers GN Jr. A study of absorbancechanges in the aspartate aminotransf’erase reaction prior to addi-tion of 2-oxoglutarate [Abstract]. Clin Chem 1986;32:1138.

24. Okorodudu AO, Pelletier PR, Valcour AA, Bowers GN Jr,McComb RB. Evaluation of the IFCC Reference Method for ala-nine aminotransferase: spurious blank ALT activity due to con-tamination ofD-alarnne with L-alanine, and recommendations fora correction. Clin Chem 1989;35:153-6.25. Koedam JC, Van Dreumel HJ, van der Ham W, Steentjes GM,Terlingen JBA. Production and certification of secondary enzymereference materials (ERMs). Part 2: the need for meaningfulprOtOCOlS that assure photometric accuracy and a well-definedprocess for value assignment. Clin Chem 1986;32:1906-13.26. Bowers GN Jr, Alvarez R, Cali JP, et al. The measurement ofthe catalytic (activity) concentration of seven enzymes in NBShuman serum SRM 909. Gaithersburg, MD: National Bureau ofStandards, 1983. (NBS special publication 260-83.)