Journal of Automatic Chemistry, Volume 4, Number 2 (April-June 1982), pages 90-106 International Congress on Automation in the Clinical Laboratory: Editorial The last number of Journal of Automatic Chemistry marked the first edition produced by our new publisher, likewise this second issue also marks a new departure. Part of this issue is devoted to the publication of abstracts for the International Congress on Automation in the Clinical Laboratory. This meeting is being held in Barcelona from 19 to 22 April 1982, and the abstracts are devoted to various aspects of automation as applied to the clinical laboratory. However, the problems of applying computers to instrumentation, or to the general work-flow in any laboratory, present similar problems to the user and to the systems designer and there is considerable merit in sharing experiences. Further details on the papers abstracted can be obtained directly frgm their autho, although hope that several of them will submit the full text of their papers for publication in the Journal. Every participant at the congress will be presented with a copy of the Journal with their registration documents. If it is the first time that they have seen the Journal, we welcome them as readers and hope that within these pages they will find much of value to their work. also hope that they have a valuable and enjoyable congress and gain much of value from their participation. On a personal note, have just returned from a brief visit to the 33rd Annual Pittsburgh Conference and the large exhibition will be presented in the next issue. There, during a discussion with two members of my main editorial board--Professor Howard Malmstadt and Dr Rolf Arndt--about inviting Professor G. Horlick to the editorial team, a very interesting suggestion was raised. It was proposed that the formation of a society of laboratory automation would be an attractive concept and that it should be directly linked to .the Journal. This should encourage a broader subscription base for the Journal and promote conferences on laboratory automation where clinical chemists and industrial chemists, along with those in academic life, can discuss the many problems involved in introducing computerization and automation into various work environments. The proposal certainly has considerable appeal and would welcome reactions from our readers. A world-wide conference on laboratory automation, along the lines of the congress in Barcelona, would be a valuable forum for discussing the problems and advantages of automation. An interesting development at Pittsburgh was the introduction of an automated system based on robotics. The Zymark Corporation was formed in March 81 to develop, manufacture and market instrumentation for automated laboratory-scale processes used in chemistry and biochemistry. The Zymate Laboratory Automation system combines robotics and laboratory stations to automate the procedures used in sample preparation. For example liquid handling, sample conditioning, separation and chemical modification. In addition to the instrumentation, the Zymark Corporation has developed an educational program to provide a foundation in modern techniques of sample preparation--this program is available on a subscription basis. Whilst the approach is not totally new--the Robot Chemist and similar systems were launched more than 10 years ago--the technology now available makes the chances of success of this new venture more likely. Future developments along similar lines are eagerly awaited. Peter B. Stockwell Congress Abstracts Session 20.1" Clinical Chemistry General 20.1.1: A colorimetric alpha-amylase method on automated analysers By J. L. Derocque, Boehringer Mannheim GmbH, Mannheim, FR Germany A survey of the various methods for the assay of alpha-amylase activity in serum and urine is presented. Their suitability for application to automated instruments is discussed, and a new kinetic alpha- amylase method using para-nitrophenyl- maltoheptaoside as a chromogenic substrate is described. This method is free from interference from endogenous glucose and can easily be adapted to most of the common automated systems. An international inter-laboratory survey including more than 1000 labora- tories in 14 European countries was car- ried out. The participants in this survey were requested to report simultaneously the results they obtained with the new method and the results of the method routinely used in their laboratory. Evaluation of the results showed that the inter-laboratory coefficient of variation (CV=12), obtained with the new method was good and comparable with those obtained for other enzymes (for example alkaline phosphatase, gamma- GT) in surveys carried out in Germany. The inter-laboratory precision was also far better than that obtained with most of other alpha-amylase methods. Results are also discussed according to the various instruments used in this survey. 20.1.2: Determination of serum urea with use of o-phthaldehyde reagent in the Coulter chemistry CA-3 and ABA-100 By J. M. Paz, A. Lopez-Urrutia, J. C. Tutor and Del Rio, Laboratorio Central, Hospital General de Galicia, Universidad de Santiago de Compostela, Spain The o-phthaldehyde procedure, which was developed by Jung for the measure- merit of urea, was adapted to the Coulter 90 chemistry CA-3 and ABA-100 and evaluated for accuracy, precision, linearity and sample-to-sample inter- action. The results obtained were com- pared with those for the urease/GLDH and diacetyl methods. The precision obtained was good (CP > 98"5) and the total analytical error was not medically significant. The linearity throughout was 1"33-44.96mmol/1. A sample-to-sample interaction was ob- served only for Coulter chemistry. For 121 random samples diacetyl (Coulter Chemistry) versus o-phthaldehyde (Coulter chemistry) gave a Pearson coef- ficient (r) of 0"998 (y =0"98 +0"10). For 63 samples o-phthaldehyde (Coulter chem- istry) versus urease/GLDH (ABA-100), r=0’998 (y=0.97x+0.18). For 75 samples diacetyl (Coulter chemistry) versus o-phthaldehyde (ABA-100), r =0"999 (y 0"99x +0.13). The Jung procedure can be simply adapted to the Coulter chemistry and ABA-100. It provides an accurate, precise and economical alternative to the diacetyl and urease/GLDH methods.
Transcript
Journal of Automatic Chemistry, Volume 4, Number 2 (April-June 1982), pages 90-106
International Congress on Automation inthe Clinical Laboratory: EditorialThe last number of Journal ofAutomatic Chemistry marked the first edition produced by our new publisher, likewise this second issuealso marks a new departure. Part ofthis issue is devoted to the publication ofabstracts for the International Congress on Automation inthe Clinical Laboratory. This meeting is being held in Barcelona from 19 to 22 April 1982, and the abstracts are devoted to variousaspects ofautomation as applied to the clinical laboratory. However, the problems ofapplying computers to instrumentation, or to thegeneral work-flow in any laboratory, present similar problems to the user and to the systems designer and there is considerable merit insharing experiences. Further details on the papers abstracted can be obtained directly frgm their autho, although hope that several ofthem will submit the full text of their papers for publication in the Journal. Every participant at the congress will be presented with acopy of the Journal with their registration documents. If it is the first time that they have seen the Journal, we welcome them as readersand hope that within these pages they will find much ofvalue to their work. also hope that they have a valuable and enjoyable congressand gain much of value from their participation.
On a personal note, have just returned from a brief visit to the 33rd Annual Pittsburgh Conference and the large exhibition will bepresented in the next issue. There, during a discussion with two members ofmy main editorial board--Professor Howard Malmstadtand Dr RolfArndt--about inviting Professor G. Horlick to the editorial team, a very interesting suggestion was raised. It was proposedthat the formation of a society of laboratory automation would be an attractive concept and that it should be directly linked to .theJournal. This should encourage a broader subscription base for the Journal and promote conferences on laboratory automation whereclinical chemists and industrial chemists, along with those in academic life, can discuss the many problems involved in introducingcomputerization and automation into various work environments. The proposal certainly has considerable appeal and wouldwelcome reactions from our readers. A world-wide conference on laboratory automation, along the lines of the congress in Barcelona,would be a valuable forum for discussing the problems and advantages of automation.
An interesting development at Pittsburgh was the introduction of an automated system based on robotics. The ZymarkCorporation was formed in March 81 to develop, manufacture and market instrumentation for automated laboratory-scaleprocesses used in chemistry and biochemistry. The Zymate Laboratory Automation system combines robotics and laboratory stationsto automate the procedures used in sample preparation. For example liquid handling, sample conditioning, separation and chemicalmodification. In addition to the instrumentation, the Zymark Corporation has developed an educational program to provide afoundation in modern techniques of sample preparation--this program is available on a subscription basis. Whilst the approach is nottotally new--the Robot Chemist and similar systems were launched more than 10 years ago--the technology now available makes thechances of success of this new venture more likely. Future developments along similar lines are eagerly awaited.
Peter B. Stockwell
Congress AbstractsSession 20.1" ClinicalChemistry General
20.1.1: A colorimetric alpha-amylasemethod on automated analysersBy J. L. Derocque, Boehringer MannheimGmbH, Mannheim, FR GermanyA survey of the various methods for theassay of alpha-amylase activity in serumand urine is presented. Their suitabilityfor application to automated instrumentsis discussed, and a new kinetic alpha-amylase method using para-nitrophenyl-maltoheptaoside as a chromogenicsubstrate is described. This method is freefrom interference from endogenousglucose and can easily be adapted to mostof the common automated systems.
An international inter-laboratorysurvey including more than 1000 labora-tories in 14 European countries was car-ried out. The participants in this surveywere requested to report simultaneouslythe results they obtained with the newmethod and the results of the method
routinely used in their laboratory.Evaluation of the results showed that theinter-laboratory coefficient of variation(CV=12), obtained with the newmethod was good and comparable withthose obtained for other enzymes (forexample alkaline phosphatase, gamma-GT) in surveys carried out in Germany.The inter-laboratory precision was alsofar better than that obtained with most ofother alpha-amylase methods.
Results are also discussed accordingto the various instruments used in thissurvey.
20.1.2: Determination of serum urea withuse of o-phthaldehyde reagent in theCoulter chemistry CA-3 and ABA-100By J. M. Paz, A. Lopez-Urrutia, J. C.Tutor and Del Rio, Laboratorio Central,Hospital General de Galicia, Universidadde Santiago de Compostela, SpainThe o-phthaldehyde procedure, whichwas developed by Jung for the measure-merit of urea, was adapted to the Coulter
90
chemistry CA-3 and ABA-100 andevaluated for accuracy, precision,linearity and sample-to-sample inter-action. The results obtained were com-pared with those for the urease/GLDHand diacetyl methods.
The precision obtained was good (CP> 98"5) and the total analytical error wasnot medically significant. The linearitythroughout was 1"33-44.96mmol/1. Asample-to-sample interaction was ob-served only for Coulter chemistry. For121 random samples diacetyl (CoulterChemistry) versus o-phthaldehyde(Coulter chemistry) gave a Pearson coef-ficient (r) of 0"998 (y =0"98 +0"10). For 63samples o-phthaldehyde (Coulter chem-istry) versus urease/GLDH (ABA-100),r=0’998 (y=0.97x+0.18). For 75samples diacetyl (Coulter chemistry)versus o-phthaldehyde (ABA-100),r =0"999 (y 0"99x +0.13).
The Jung procedure can be simplyadapted to the Coulter chemistry andABA-100. It provides an accurate, preciseand economical alternative to the diacetyland urease/GLDH methods.
20.1.3: An automated procedure of serumiron on the Technicon AutoAnalyzer IIBy Jau-Wen Chu and Clara V. SumeThy,Henry Ford Hospital, Detroit, MichiTan48202, USAComparison of Technicon AutoAnalyzerII method of serum iron (method No. 25,March 1972) to that of AutoAnalyzer I(method N-62P) yields an average of5higher in serum iron. The spuriously highresult of the former is due to the positiveDonnan effect resulting from the imbal-ance of ionic species between sample andrecipient streams. This can be correctedby reformulating the reagents as follows:iron acid reagent, 5 NaC1 and 0.01Tween 20 in 0.11 N HC1; acid diluent, 1ascorbic acid in the iron acid reagent; ironcolour reagent, 0.035 ferrozine and0.01 Tween 20 in 0.11 N HC1; andsodium acetate, 1"0 N. The flow diagramsremained unchanged. Lag phase and half-wash time, studied by the decay ofsteady-state curve, were found to be 4.1 s and 5 s,respectively.
The modified method compares wellwith AutoAnalyzer I, with a correlationcoefficient of0"995, a slope of 1"05, and anintercept of -3.7/zg/dl. The recoveryrate, studied by the addition of aqueousiron solution to patient sera, was found toaverage 98.2. The day-to-day coefficientof variation during the two months’study was 3.6. The method is notaffected by the normal concentrations ofcopper present in the sera. Thioglycollicacid, which chelates to serum copper, canreplace ascorbic acid in the acid diluentand provides a longer reagent shelf-life.This automated procedure for serum ironhas been used in the authors’ laboratoryfor over three years, and has proved to bea simple and reliable method.
20.1.4: An improved creatinine procedureof high specificity for the Cobas-BioAnalyserBy M. Sheehan and J. Salmon, GoodSamaritan Hospital, Portland, Oregon97210, USAThe authors have adapted the end-pointprocedure of Slot (see ScandinavianJournal of Clinical LaboratoryInvestilation, 17 [1965]) to the Cobas-Bio centrifugal analyser. The two-reagentaddition capability of the instrumentallows complete automation of thismanual procedure. Sample and alkalinepicrate reagent are added to the reactioncuvette and allowed to react for 5 min.After an absorbance reading is taken, anacidic reagent is added and anotherabsorbance reading is taken 60s later.The second absorbance reading (colourdue to interfering substances) is sub-tracted from the first (colour due topicrate complexes plus interferences),and the analyser calculates creatinineconcentration of samples based on thisdifference relative to standards. Within-day (N= 20) and day-to-day (N 10) pre-
International Congress on Automation in the Clinical Laboratory: Abstracts
cision studies exhibited percentage CVsof less than and 1.7, respectively, at the6" 3 mg/dl level. The method is linear to atleast 15mg/dl and exhibits good corre-lation with the ASTRA-8 kinetic alkalinepicrate method. No interference wasseen for samples from diabetics in ketoacidosis.
20.1.5: Pyridoxai-5-phosphate amino-transferase method for Autoanalyser II YSMA 12/60By J. Perez Garcia-Buela, ServicioAnalisis Clinicos, Ciudad Sanitaria JuanCanalejo, La Corufia, SpainA pyridoxal-5-phosphate (P5P) activatedaminotransferase (AST) method for useon Technicon’s continuous-flow equip-ment is described (the AA II and SMA12/60). The method is in line with therecommendations of the InternationalFederation of Clinical Chemistry (IFCC).A substrate, according to the IFCC’smethod, with 30mM pyridoxal-5-phosphate was used. No pre-incubationof the sample with P5P or manifoldchanges are required. With this procedurethe linearity, at 37C, was to 500 U/1. Thefollowing results were obtained:
Level Level 2
Number of assays 51 54Media (U/I 37C) 63 215SD (U/l) 2.24 1.93CV (%) 3.55 (89
Samples from a normal donor group andfrom a group of patients with chronicliver disease were studied with the ASTmethod. The normal donor group (N 42)showed an increase of AST to 7-10 U/1with added P5P. Patients in the chronicliver disease group who had near-normalAST values showed the same increase asthe normal group, while the patients withelevated AST values showed increases ashigh as 100.
20.1.6: Automated calculation of gammaglutamyl transpeptidase in alcoholicsBy P. Bataller Martinez and V. Sanchis-Bayarri, Clinical Analysis Service,Provincial Hospital, Valencia, SpainIt is sometimes difficult to evaluate theeffects of alcohol on the hepatic cell.Patients have a tendency to minimizetheir consumption of alcohol in theirclinical histories. One of the objectivemethods is to calculate quantitively thegamma glytamyl transpeptidase (GGTP).The results of analyses of GGTP inpatients from various hospital centreswere studied over a three-month period.An auto-analyser (ABA-100) with a 450filter was used, samples were incubated at37C, sample volume was 5/d and theanalysis time was 5min. The results
showed an increase in the GGTP inpatients with a clinical history of pre-sumed alcoholism, even in the absence ofa rise in other enzymes, such as serumglutamate-oxalacetate (SGOT), serumglutamate-pyruvate (SGPT) and alcalinephosphatase (AP). In some patients,where the aim is to reduce the intake ofalcohol rather than total abstinence, theresults of a series of measurements ofGGPT provides an objective control ofalcohol consumption.
20.1.7: Automated determination of serumfl-N-acetyI-D-glucosaminidaseBy G. Johannsen, D. Maruhn and D. Paar,Division ofClinical Chemistry, Departmentof Medicine, University of Essen, FRGermanyPottet al.’s discontinuous manual assayfor serum fl-N-acetyl-D-glucosaminidase(EC 3.2.1.30, AGS) was adapted to theACP 5040 Eppendorfdiscrete automatedanalyser. The instrument was run in end-point mode at a throughput rate of 120samples/h. The analyser was equippedwith a thermojacketed flow-throughdevice to provide the dispensers with pre-heated (at 40C) reagents. In the assay,25#1 of sample was added to 250#1 ofbuffer-substrate solution (citrate buffer,0.05mol/1, pH 4"2; 4-nitrophenyl-fl-N-acetyl-D-glucosaminide, final concen-tration 9mmol/l). After 6.5min. incu-bation at 37C the reaction was termin-ated by the addition of 250/d of 2-amino-2-methyl-propan-l-ol/HC1 (AMP) bufferat pH 10.25. Absorbance of liberated4-nitrophenol was read against a blank towhich AMP buffer was added beforeincubation.
A linear relationship was found bet-ween the amount of enzyme in the assaymixture and enzyme acitivity up to100U/1. Within-run imprecision (CV)ranged from 0"89 to 4"78%. Between-dayimprecision, as determined with Hyland N11 accuracy control serum, amounted to4.09% (N =9, CV). Comparison with themanual method yielded a correlationcoefficient of0.9885. The automated ACP5040 procedure is fast, precise and reliablefor the determination of AGS in humanserum.
20.1.8: Determination of non-pregnancytotal urinary oestrogens by fluorescencephotometry: comparison between aclassical extraction and an automatedKober-Ittrich methodBy R. Wolfrum, M. Trapp, H. G. Bohnet,and F. Leidenberyer, Chemistry Depart-ment, Allgemeines Krankenhaus Heidber9and Institute for Hormone and FertilityDisorders, Hamburg, FR GermanyHalket and Wolfrum published a paper in1980 which showed that a significantcorrelation exists between a manual andan automated continuous flow techniquefor 24 h urinary total oestrogen determi-nation in pregnancy (rag range). The
91
International Congress on Automation in the Clinical Laboratory: Abstracts
present study is a further contribution tothe development of a reference methodand describes the analytical results in 93urine samples, in the concentration range6-200/g/24 h, using the same automatedsystem as the previous project (BredaScientific) with a preceding Sephadex G-10 step (van Kessel) and Brown’s semi-automatic method. Regression analysis ofthe paired values shows a linear relation-ship, y= 1.14x 3.36 (y= flow technique),and a correlation coefficient (r) of 0"97(p < 0.001). At present the automated flowtechnique is the only system for thesimultaneous determination of dU-mgand -/g concentrations. Its use shouldallow external quality-control proceduresto be implemented.
20.1.9: Chamber analytical technique--itsapplication to enzymatic and immuno-logical assays in the ultra micro rangeBy E. Hoffmann-Blume and A. Horn,Institute of Physiological Chemistry,Friedrich-Schiller University, Jena,German Democratic RepublicThe chamber analytical technique allowsthe performance of colorimetric andfluorimetric tests in the form of fixed-point or kinetic measurements in sets of50 samples with test volumes between 8and 80/A at light paths between 0.2 and2 mm.
A new device is described for tem-perature-controlled automated fluori-metric and colorimetric measurementswith increased sensitivity by directing thelight beam twice through the solution forcolorimetry. Precisions are in the range of5-8" 10- ’ units of absorbance and 0"5for fluorimetric measurements. Thesystem is accomplished by automatedvolume dispensing and accessories forpre-analytical steps.
Application examples are reported:determinations of alanine aminotrans-ferase activities as well as of KM and Vm,xof alkaline phosphatase in 80#1 testsresult in day-to-day CVs of 2-5and 6-7, respectively. An ultra microELISA for e-fetoptotein in serum withtest volumes of 101 and a CV of < 10demonstrates the applicability forfluorimetry. About 150 assays can becarried out with ml of antiserum forturbidimetric determinations of albuminand apoprotein B (inter-assay CV 4-8).With the measuring and volume dispens-ing automation 100-200 miniaturizedtests/h can be performed.
authors’ clinical laboratory computersystem (PRIMULAB) with its prime ob-jective towards the service requirementsof the laboratory. Therefore it includespossibilities of simultaneous optical read-ing of request forms and on-line captur-ing, processing and printing oflaboratorytest data. Priority request forms, whichallow the clinician to indicate the intervalafter which emergency test results shouldbe available, are registered by an opticalreader and arranged according tourgency by the computer. Work-sheetsare replaced by displaying all the infor-mation required for accurate specimenanalyses on a large, colour TV screen. Theindividual processing status of all tests,from as many as 30 request forms, isdisplayed in a colour code. For processcontrol the updated delay time for testperformance is faded in.
The computer capability for manag-ing the sample throughout and dataprocessing minimizes the stress on tech-nicians. This results in a reduction of theturnaround time of tests. 95 of allrequested tests are performed and re-ported within the interval indicated. Incritical situations, test results are avail-able within 3-10 min.
20.1.11: Experiences in a clinical lab-oratory gained by a Hungarian-madeprogrammable automatic chemicalanalyserBy L. Bartalits and B. Bak, Hospital inPterfy S, str. 8, Budapest 1076, Hungaryand Central Research Institutefor Physics,Budapest, HungaryThe necessary conditions for the auto-matic instrumentation ofenzyme activitydeterminations are examined, the oper-ational details of the CRIP clinicalchemical analyser are introduced and 18methods for measuring enzyme activityand metabolite-concentration of bodyfluids are presented. The analyser is in adiscrete system operated by compressedair and has a throughput of 200 samples/h. The measurement of metabolites isperformed at equilibrium point, themeasurement of enzyme activity is per-formed by two-point kinetics, accordingto the Trayser-Saligson principle in anoptimized reaction environment. Theiron-free components of the analysermake measurement of iron by a photo-metric method possible.
20.1.10:. An emergency laboratory inte-grated within a laboratory data processingsystemBy D. Neumeier, H. Sator, G. Rindfleischand M. Knedel, Institut flit KlinischeChemic, Klinikum Grol3hadern, Ludwig-Maximilians- Universitgtt Miinchen, D-8000 Miinchen 70, FR GermanyThe data-processing system for the emer-gency laboratory was integrated in the
92
Session 20.2: ClinicalChemistry--Evaluation
20.2.1: The evaluation of a high-speed(9000 tests/h) multichannel (30 channels)automated analyser for clinical chemistryBy Yasuyuki Hayashi and MasalazuHineno, Department ofClinical Pathology,Juntendo University School of Medicine,3-1-1 Hongo, Bunlyo-lu, Tokyo and R/DEngineering Department, AnalyticalInstruments Division, ShimadzuCorporation, Kyoto, JapanThe Shimadzu Multichannel AnalyserCL-30 was developed by the ShimadzuCorporation (Kyoto, Japan); Japaneseprofessionals co-operated in the develop-ment--this co-operation is a guide-line ofthe Ministry of International Trade andIndustry of Japan. The new analyser isintended to realize high capacity withminimized volumes of specimen and re-agents, as well as to be supplied at a lowprice. It has a characteristic disposablesheet made fro.m polyester resin, which isused for reaction vessels and cuvettes, andis termed a ’micro-cuvette-sheet’. Therepertoire of this analyser includes serumenzymes, lipids, sugars, proteins, nitrogencompounds and other contents by kineticassay, visible colorimetry or flame pho-tometry. Evaluation data are discussed;the machine has been studied since May1980. Within-day precision (as CV) was0"5-3 and day-to-day precision was2-8, these precisions are acceptablewhen compared with the usual method inthe authors’ laboratory. Approximately5 carry-over was found, which is pro-bably due to the Teflon transportationtube and so may be improved on.
Analytical programmes are change-able by altering the position of a photo-meter or by adding optional accessories.The Shimadzu CL-30 is capable of per-forming fast analyses with small specimenvolumes and reagent consumption.
20.2.2: Evaluation of a new centrifugalanalyser: Flexigem, developed by Electro-Nucleonics Inc.By Y. Gourmelin, J. Herfent, G. Glikmanafand A. Truchaud, Department of Bio-chemistry, Meaux Hospital, FranceThe instrument was evaluated for twomonths in the authors’ laboratory.(1) Precision at three different levels ofurea was determined over a period of20 days, using a kinetic urea reagent.Within-run CV was better than 2"5 andday-to-day precision was better than 4.(2) Contamination was tested by means ofa kinetic Jaff6 method for creatinine.(3) Temperature accuracy and stabilitywere tested at 25, 30 and 37C, using acresol-red solution. (4) The linearity andaccuracy of the optics were tested at 340,405 and 500 nm. (5) Special attention wasgiven to the micro-computer allowing theuser to choose between four curve-fit
calculation modes: 4 parameter log-logit,5 parameter logit, 5 parameter ex-ponential, 5 parameter polynomial. Aquantitative turbidimetric IGG methodwas used to test the polynomial mode.The 5 parameter logit mode was testedby means of EMIT gentamicin. In bothcases the program proved to operatesatisfactorily.
20.2.3: Application of the Flexigem ana-lyser for routine laboratory analysisBy E. Knoll, K. Dettmer and H. Wisser,Robert-Bosch-Krankenhaus, Departmentof Clinical Chemistry, Auerbachstrasse110, 7000 StuttTart 50, FR GermanyThe authors have eight years’ experiencewith a centrifugal analyser of the firstgeneration--Gemsaec. They investigatedthe performance of the newly developedFlexigem, a centrifugal analyser of thethird generation, and compared it withGemsaec. It was tested over a period ofthree months, its practicability, accuracyand precision were examined. Threeenzyme determinations (GOT, GPT andGamma-GT), four substrate determi-nations (glucose, creatinine, triglyceridesand urea) as well as the turbidimetricdeterminations of IGG, IGA and IGMwere tested. Precision was establishedusing control sera. Accuracy was de-termined by comparing analyses ofpatient samples, covering as wide a con-centration range as possible. The authorsreport a good correlation between theFlexigem methods and those ofGemsaec.
20.2.4: Investigations of the activity ofsome enzymes by microcentrifugal ana-lyser: IL Multistat IIIBy J. Jordanova, Higher Medical Institute,Central Clinical Laboratory, Sofa,BulgariaThe activity of aspartat aminotransferase(AsAT), alanin aminotransferase (A1AT),lactate dehydrogenase (LDH), -hydroxibutira dehydrogenase (HBDH)and creatine phosphokinase (CPK) inserum was studied with the micro-centrifugal analyser IL Multistat IIIprogrammed for Boerhinger’s tests. Thepossibilites of the Multistat III increasingthe number ofthe samples and decreasingassay time was investigated. Multistat IIIis shown to reduce the volume of thebiological material necessary, as well asthe volume of the reagents. The within-day and day-to-day reproducibility andaccuracy (N 20) were examined usingfive commercial control sera.
20.2.5: Evaluation of the Hitachi 705automatic analyserBy C. Ferre, M. J. Castifieiras, M. J.Alsina, J. Riera and R. Galimany,Oepartmento de Analisis Clinicos (C.S.),’Principes de Espafia’, Hospitalet deLlobregat, Barcelona, SpainThe Hitachi 705 automatic analyser wastested against the standard procedure laid
International Congress on Automation in the Clinical Laboratory: Abstracts
down by the Instrument Commission ofthe ’Sociedad Espafiola de QuimicaClinica’. The parameters investigated forthe evaluation were: (1) linear ranges;(2) within-run and inter-run imprecision,using three different concentration levels;(3) contamination effects.
The serum constituents used in thetrials were: glucose (GOD-PAP), urea(urease UV), creatinine (Jaff6 withoutdeproteinization), calcium (cres-olphtalein), AST (optimized standardmethod) and CK (optimized standardmethod, NAC activated). In addition theHitachi 705 was assessed for inaccuracyagainst the Technicon SMAC.
2, 3, and 4. Precision data obtained atthree levels for each analyte and methodcomparison data are presented. Precisiondata for key assays is:
20.2.6: Serum urate determinations on theHitachi ABCA 706 DBy L. Lepoutre and S. Pauwels, U.1.A.,Academic Hospital, Wilrijkstraat, 10B2520 Edegem, BelgiumSerum urate determinations using achromogen-based method were perfor-med on the Hitachi 706 D and comparedwith an end-point analysis on theCentrifiChem 400. Analyses were per-formed at 25C on the Hitachi and at 30Con the centrifugal analyser. Precision andlinearity were studied on appropriatedilutions prepared from pooled serum,saturated with uric acid, while calibrationwas based on external quality-controldata. The following results were ob-tained: within-run precision: 0.13mmol,N= 30, CV 1"47%; 0.25 mmol,N=31, CV=2"08%; 0.47mmol,N= 30, CV =0"69%; 0"88 mmol, N 24,CV =0"99%. Between-run precision:(N =consecutive days) 0"26 mmol, N 30,CV 2.1%; 0.62 mmol, N 30, CV 2"0%.Linearity was found to exceed 1.98 mmol.Correlation between the two methods canbe represented by y= 1"090 x-0"656 withr=0"978 on 163 serum samples (wherey=chromogen method, x=centrifugaltechnique). Carry-over was observed fromconcentrations higher than 0"82 mmol.
20.2.7: Evaluation of chemistry and Emitassays on Flexigem, a new centrifugalanalyserBy F. Van Lente and V. M. Leitz, Electro-Nucleonics Inc., 368 Passaic Ave.,Fairfield, New Jersey 07006, USAFlexigem is a microprocessor-controlledcentrifugal analyser, which in addition toperforming routine chemistry procedures,can be programmed to perform immuno-assay and special chemistry procedures.Data for non-linear immuno-assays canbe fitted to generate standard curvesusing one of four mathematical models:log-logit, four parameter polynomial,exponential, and five parameter logit.Thirty such curves can be stored inmemory at any one time. The perform-ance of 30 routine chemistry procedures,four Emit procedures and three immun-oglobulin procedures has been evaluatedusing protocols based on NCCLS PSEPs
Comparison studies have been performedat three sites for key assays. Data onTheophylline was essentially identical atthe two sites: site 1, r=0’97, x--18"2,= 18"1; site 2, r=0"97, x 12"8, y= 13"2,and precision was statistically equivalent.
In conclusion, the Flexigem analyserhas been shown to give precise resultsand values obtained have been shown tobe in good agreement with those obtainedusing established procedures.
20.2.8: Experience with multichannel ana-lysers in 26 Swedish hospitalsBy Kas Levin and Torsenten Aronsson,Clinical Chemistry Central Laboratory,Central Hospital, S 721 89 Vasterts andClinical Chemical Central Laboratory,Academic Hospital, S 750 14 Uppsala,SwedenIn October 1980 a workshop on multi-channel analysers was arranged under thesponsorship of the Swedish Society forClinical Chemistry. A mulitchannel ana-lyser was defined as an instrument cap-able ofcarrying out more than 10 differentanalytical procedures on one sample,with a capacity of analysing at least 60samples-standards-controls/h. The usersof eight different systems were present atthe meeting. The participants had beeninvited to analyse one common batch ofacontrol serum in a standardized fashionbefore the meeting. The data wereanalysed using a statistical model, wherethe imprecision could be separated intowithin-day imprecision, between-dayimprecision and total imprecision.
At the meeting the results werecollected and discussed within the users’group, and the mean values for each typeofmultichannel analyser were calculated.The differences found between the per-formance of different analysers are small.These differences and the differencesbetween different types of analysers arediscussed by the authors.
93
International Congress on Automation in the Clinical Laboratory: Abstracts
20.2.9: Aurora, a new clinical analyserBy Lars-Ake Carlsson, Clinicon AB, Box148, S 161 26 Bromma, SwedenA novel instrument for kinetic and end-point measurements, the Aurora, isdescribed. The instrument is a completelynew development, having unique featuresin liquid handling and photometry. TheAurora analyser is able to selectively per-form up to 16 analyses on a patient sample.It is equipped with a computer witha multi-task operating system making itpossible to perform several tasks simul-taneously, for example process control,input of requests, and printing of results.Sample and reagent handling is made bystepper-motor driven pumps. Thesepumps are highly precise (better than0"5%) and accurate (better than 1%) andhave negligible dead volume and very lowcarry-over (sample-sample 0"05-0.2%,method-method less than ppm). Sampleand reagent are incubated in tubes in arotor and after 10min. are sucked into aphotometer for measurement. Threedifferent reagent pumps and a specialstepper mode of the reaction rotor makesit possible to add reagents twice at 12different moments during an lmin.period. The temperature of the two in-dependent temperature control systemsfor the photometer and the incubationrotor are held within +0.05C from thestated temperature. The four-channelphotometer is optimized for enzymemeasurements (short path-length 2"5 ram)and shows low sensitivity to turbidity inpatient samples and controls due to aunique optical system with a large accept-ance angle. Calculation of the slope ofzero-order reactions is made by linearregression, while first-order reactions areevaluated with an integration method.Rate curves are displayed on a VDU andare sorted on a discette if a quality figureis exceeded. Patient results are acceptedonly after passing an acceptance pro-cedure, where for example, a VDUpresentation of controls is made. Fromthis picture drift any excessive variationduring an analysis is easily seen. Resultsfrom an external evaluation of theinstrument are presented.
20.2.10: Analytical experiences withAstra-8By A. SzabO, Postgraduate MedicalSchool, Department of Clinical Chemistry,H1389 Budapest, P.O. Box 112, HungaryThe Astra-8 analyser (Automated Stat-Routine Analyser, Beckman Company)with modules was used for assayingserum and urine sodium, potassium,chloride, urea, carbon dioxide, glucoseand creatinine, respectively. Duringinstallation the following parameterswere investigated. (1) The maximumdeviation between measured values andthose declared by the manufacturersranged between +3.4 and -6"2%. Tendifferent control sera were used to deter-mine accuracy (300 assays). (2) To
investigate the precision, 586 assays wereperformed. The variation coefficientvaried from 0"3 to 3-0. (3) In 42 assaysthe carry-over effect was found to belower than 0"1. (4) As compared todifferent routine methods the correlationcoefficient was between 0.988 and 0’999(800 assays).
Astra-8 uses the most recent results ofanalytical chemistry and data processing.The analyser, with all its built-in pro-grams, is completely satisfactory andreliable for the requirements of modernclinical chemistry.
20.2.11: Performance of three kinetic a-amylase methods on the Kone CD clinicalanalyserBy G. A. Harff and F. Lamie, AcademicHospital, Free University, Amsterdam,The Netherlands, and Hoechst HollandN.V., Behring Diagnostics, Amsterdam,The NetherlandsThe maltotetraose method of GeneralDiagnostics, the carboxymethylatedstarch method of Pharmacia Diagnosticsand the p-nitrophenyloligosacharidemethod of Behring Werke were evaluatedfor use on the Kone CD. The deter-minations were performed at 30C,with 5 rain. lag time and 3 rain. kineticmeasurement. Kone D Analysis No.27.2 was used. The photometric responsefor Ortho Normal control serum (lot No.11T222) was limA/rain, at a volumefraction of the sample of 0"0476 with the’maltotetraose’ method; 14 mA/min, at afraction of 0"00662 with the ’starch’method; and 13 mA/min, at a fraction of0"0385 with the ’p-nitrophenyl’ method.Patients’ sera showed almost the sameratio of mA/min. The volume fractionsare according to the manufacturer’s in-structions. A disadvantage of the ’starch’method is the need to predilute thesample because of the small volumefraction. Linearity and day-to-day pre-cision are presented. The method ofcomparison was the Phadebas Amylase(tablet) method, in which results are ex-pressed at 37C. Least squares regressionequations were: Phadebas=x(U/1) andkinetic method y(U/1), N 60
’maltotetraose’ method: y=0.095x-0"04 r=0"973
’starch’ method: y=0-434 x-0.09r =0"993
’p-nitrophenyl’ method: y=0"198x- 1.0 r=0"992
Interference was found for pyruvate(from retool/l), -ketobutaric acid (from0.3retool/I), -ketoglutaric acid (from0"5 retool/l) and H + (from 20retool/l)with the ’maltotetraose’ method. Glucosegave an elevated initial absorbance withthe ’starch’ method. Bilirubin (from350 #tool/l), H + (from 10mmol/1) andEDTA (from 5retool/l) interfered withthe ’p-nitrophenyl’ method. No inter-ference was found for heparin, citrate,
20.2.12: Evaluation of the BeckmanElectrolyte 2, an indirect potentiometricinstrument for analysis of sodium andpotassium in serum and urineBy G. A. Harff and C. van Leeuwen,Department of Clinical Chemistry,Academic Hospital, Free University, POBox 7057, 1007 MB Amsterdam, TheNetherlandsThe authors evaluated and compared theBeckman Electrolyte 2 to the IL 243 flamephotometer. For serum, the Beckmanliquid control sera, Decision Levels 1, 2and 3 were used to assess the within-runand between-day precision.
Within-run precision
x SD N
Serum Na
Serum K +
125’8 0"5 38137.4 0"5 38149.4 1"2 38
2.73 0.01 384"22 0"02 385-76 0.05 38
Between-day precision
x SD N
Serum Na
Serum K
125.7 1.1 21137.6 1.2 21151.4 1.3 21
2-68 0.02 214.19 0.03 215.87 0.05 21
For urine, within-run and between-dayprecision were assessed with the HylandQ-Pak Chemistry Urine Control.
Within-run precision
Urine Na 87-3 0"3 20Urine K 30.8 0"2 20
Between-day precision
Urine Na 87.3 0.9 20Urine K 30.7 0" 3 20
Method comparison was performed withthe IL 243 flame photometer. Deming’scalculation of regression equations wereused. Serum Na4: y(ISE) 1.024 x(flame)+0"0. Syx=0"39 and serum K/: y= 1.024x-0.093 Syx=0.12. For urine modeleast squares regression equations wereNa4:y=0.9995 x-0"95, r=0.9997 andK/: y=0"9838 x-0.15, r=0.9987. TheNa/ electrode was influenced by the K/
concentration in urine samples. At aconcentration for Na/ of 50mmol/1 thiseffect was y Na / and x K /; y 0"0205x + 50"4, Sr,=0"22. The higher the Na /
concentration the larger the absolutecontribution of K+ to the Na+ result.
94
20.2.13: Evaluation of sodium and pot-assium determinations in serum and urineon KONE OY’s MicrolyteBy G. A. Harff and C. van Leeuwen,Department of Clinical Chemistry,Academic Hospital, Free University, POBox 7057, 1007 MB Amsterdam, TheNetherlandsDirect potentiometric determinations ofthe electrolytes sodium, potassium andionized calcium in whole blood, plasma,serum and urine can be performed on theMicrolyte. Dimensions of the compactapparatus are (height, depth, width) 255x 210 x 315mm. The authors assessedsodium and potassium in serum andurine. The instrument tested containedPROM version 2.1 NCCLS protocolsPSLP-2 and 3 were followed to evaluatethe precision. The method ofcomparisonwas flame photometry on the IL 143. Theregression equations for the serum modewere calculated according to the methodof Deming.
For the urine mode the samples weremanually pre-diluted five times accordingto the manufacturer’s recommendations.Least squares regression equations arepresented. Some relevant substances weretested for interference.
20.2.14: The evaluation of biochemicalparameters among the Novi Sad districtpopulation, Yugoslavia, by using theTechnicon System: SMA IIBy A. Gavrilovff:, Lj. Risti, Z. Radujkov,M. Vuurevi, Dj. Krsti?. and D. Peni?.,Institute of Pathological Physiology andLaboratory Diagnostic, Medical Faculty,Hajduk Veljkova 1, Novi Sad, Yugoslavia
During two years’ experience ofoperatingthe Technicon System SMA II, theauthors examined about 25 000 patientsfrom the Novi Sad district in Yugoslavia.The following parameters weremeasured: iron, creatinine, uric acid,triglycerides, phosphorus, cholesterol,calcium, glucose, total bilirubin, urea,alkaline phosphatase and total protein.At the same time the authors determinedthe reference values of the above-mentioned parameters and area. In thecourse of routine operations, the authorshave performed permanent, daily qualitycontrols, by using different biochemicalcontrol sera.
20.2.15: Application of innovative tech-nology in a new random access analyserBy J. Harem, 39 Boulevard de la Muette,95140 Garges-les-Gonesse, FranceNew technology for discrete aspirationand delivery of micro volumes of liquidswith effective elimination of contami-
International Congress on Automation in the Clinical Laboratory: Abstracts
nation allows rapid analysis in a randomaccess, while maintaining a high qualityof test results. This technology and itsapplication with the new Technicon RA-1000 analyser is described.
Performance data for clinicalchemistry methods are also given.
20.2.16: The Technicon SMAC II C 9000seriesBy Jean Gaumeton, 39 Boulevard dela Muette, 95140 Garges-les-Gonesse,FranceThe Technicon SMAC II is new conceptdesigned to support the analytical anddata handling needs of the laboratory. Itconsists of two major components: theanalytical console and data-managementpackage. All analyses are performedautomatically from 12 up to 20 chem-istries using improved techniques, forexample ion-selective electrodes forsodium and potassium, bound enzymetechnology for glucose and uric acid,multi-point measurement of enzymaticactivity. The samples, identified by a newbar-code identification label, are treatedin batches or in stat (for emergencyspecimens). A new statistical programusing Westgard’s algorithm checks theaccuracy of each result on-line. With theintegrated Syslab comput6r of the SMACII, the entire laboratory organizationincluding sample or patient reception,analytical process, secretarial tasks andlaboratory management is achieved.SMAC II is a major advance in lab-oratory efficiency.
Editorial note: The abstract of poster No.20.2.17 is not in English and so is notpublished here.
20.2.18: Evaluation of a third-generationcentrifugal analyser: the Flexigem systemBy Augusto Corominas Vilardell,Residencia del lnsalud and AnalysisLaboratory of Badalona, SpainAfter having realized the quality-controlof the Gemeni system, integrated in thesecond generation of centrifugal ana-lysers, the author initiated a preliminaryevaluation of the Flexigem system. Thefollowing aspects were considered. (1)Precision, at three different levels, of themost common tests made in any lab-oratory for biochemical analysis (glucose,cholesterol and urea), during a period ofthree weeks. (2) A study ofcontaminationin tests like creatinine (Jaff6method) calcium/cresolphthalein com-plexone. (3) Correlation of kinetic resultsworking with three possible analysertemperatures: 25C, 30C and 37C. (4)The use of the calibration curve memorywhich allows work on parameter log-logit, parameter logit, parameter ex-ponential and parameter polynomial.
The author reports that the resultsobtained in the evaluation have beensatisfactory.
20.2.19: Dry chemistry: preliminaryevaluation of the SeralyzerBy Francesco Aguzzi (1), AntonioMontinaro (2), Ferruccio Ceriotti (2),Rossana Colomi (1) and MuroneMichelangelo (2), (1) Laboratorio AnalisiOspedale di Broni e Stradella, ltaly, (2)Laboratorio di Ricerche Cliniche Ospedalesan Raffaele (Istituto di Ricovero e Cura aCarattere Scientifico), 20132 Milano, ItalyThe Seralyzer is a small instrument fordry-chemistry semi-automatic determin-ation of the most relevant biochemicalparameters. The paper illustrates theresults of a two-centre evaluation trial ofthe Seralyzer. Glucose, urea, bilirubin,uric acid, cholesterol, creatinine andcreatinine-phosphokinase have beenstudied. After a familiarization period,the precision, accuracy, influence of anti-coagulants have been assessed. Theresults are discussed and the problems ofstaff training are reported.
Session 20.3: Systems forKinetic Measurements
20.3.1: Determination of free haemoglobinin serum and plasma by an automatedkinetic assayBy Kurt Bauer, Institute for ClinicalChemistry and Laboratory Medicine,University of Vienna, A 1090 Wien,Lazarettg. 14, Austria
A kinetic assay for the determination offree haemoglobin in serum as well as inplasma is described. The method issuitable for use in automated systems.Commercially available reagent solu-tions, containing 4-aminophenazone aschromogen, and haemoglobin cyanidestandard solutions were used. Pro-cedures, reference intervals, statistics, andlimitations are shown for the determina-tion in serum and heparin plasma.Because of its speed and accuracy, thistest is valuable for human and animalapplications.
20.3.2: An automated method for themeasurement of serum phosphodiesteraseI activityBy C. H. Konings, Department of ClinicalChemistry, Academic Hospital, FreeUniversity, Amsterdam, The NetherlandsA kinetic assay for determining phos-phodiesterase (E.C. 3.1.4.1.) in humanserum was developed using a Gemsaeccentrifugal analyser. 40/1 of serum and80 gl of water were mixed with 320 #1of reagent consisting of thymidine-5’-monophosphate-p-nitrophenol (T5MN)in a Tris-HC1 buffer at pH 9.65. Therelease of paranitrophenol from theT5’MN by the catalytic action of theenzyme is monitored for 90 at 405 nm.The assay is carried out at 30C and is
95
International Congress on Automation in the Clinical Laboratory: Abstracts
linear for phosphodiesterase I activity to260 U/1. Within-run precision for apooled serum sample and a commercialcontrol serum gave CVs of 4.0 and 2.7(mean values of 35 and 69 U/1 respectively).A comparison study between the presentedmethod and a manual assay using thesame substrate is discussed. Results ofscientific studies on patients’ sera usingthis rapid and sensitive automatedtechnique are given.
20.3.3: Automated assay of serum acidphosphatase and its use in the diagnosis ofprostatic carcinomaBy H. Caillens, T. D. Nguyen, R. Chiche,O. G. Ekindjian, L. Boccon-Gibod andJ. Yonger, Laboratory of ClinicalChemistry and Urology Department,Cochin Hospital, Paris, FranceA kinetic method of serum acid phos-phatase determination has been adaptedto the Cobas Bio centrifugal analyser,using -naphtylphosphate as a substrateand concurrent diazotation of thereleased -naphtol with fast red TR. Theprostatic acid phosphatase activity is in-hibited by L( + tartrate. The time necess-ary for the test was 6 min. 30 s. Under theconditions described the response waslinear up to 340 UI/1, repeatability andreproducibility were good. The resultswere compared with those obtained by amanual method using p-nitrophenyl-phosphate as substrate. There was a closecorrelation between the two methods forboth total acid phosphatase (r =0"99) andprostatic acid phosphatase (r =0"97). Thediagnostic value of the proposed methodfor the detection of prostatic carcinomahas been tested on selected patients in theurology department at Cochin Hospital.A preliminary study on 86 subjectsshowed no false positive results. Therelation between acid phosphataseactivities and the clinical stages of thedisease are discussed for 21 cases ofconfirmed prostatic carcinoma.
20.3.4: Amniotic fluid cholinesterasemeasurement with a centrifugal analyserBy F. Moreau, C. Guermeur, F. Papa,O. G. Ekindjian, R. Henrion and J. Yonger,Cochin Hospital, Paris, FranceAmniotic fluid total cholinesterase(TChE) and acetylcholinesterase (ACHE)activities are usually measured for theexclusion of neural-tube defects. Ellmanbased determinations of TChE on thehydrolysis ofacetylthiocholine and detec-tion of the free thiol released by DTNB(5-5’ dithiobis [2-nitrobenzoic acid]);for a direct assay of ACHE, and Daleproposed ethopropazine hydrochloride(’Lysivane’), as an inhibitor of non-specificcholinesterase. A method has beenadapted for a Cobas Bio analyser. Thesample volume for each determination is20/ul. TChE is measured for 6 min. 30 s at30C in the presence of the substrate aftera min. pre-incubation with DTNB in a
phosphate buffer at pH 8"0. For ACHE,min. pre-incubation time with 14/m
inhibitor has been considered as optimal,following preliminary studies with 1.4 #mto 28 m Lysivane for 30 to 300 s. Thereactions are linear up to 200 UI/1.Within-run (CV <3) and between-run(CV < 5) precisions are excellent. Non-NTD amniotic fluids were used to estab-lish usual values. The technique proposedis easy, simple, and rapid enough toprovide results during the course of anamniocentesis clinic.
20.3.5: The bichromatic determination oftriglycerides on the Multistat III micro-centrifugal analyserBy Claudio Musitelli, Antonio Montinaro,Michelangelo Murone and PierangeloBonini, Laboratorio di Ricerche Cliniche,Ospedale San Raffaele, 20132 Milano,ItalyThe Trinder reaction has been widelyapplied in clinical chemistry in the last fewyears. This procedure was introduced forthe automatic determination of triglycer-ides on the Multistat III microcentrifugalanalyser. The presence of surfactants inthe reaction mixture, and the peculiardesign of the Multistat rotors, can resultin a premature mixing of sample andreagents, thus preventing the use of thefixed-time procedure. The Multistat III’ssoftware requires at least two readings, attwo different times or at two differentwavelengths (bichromatic procedure).The authors adopted the end-pointbichromatic procedure and the resultscompare well with those obtained by themanual determination.
20.3.6: The use of a bacteric cholesterol-esterase for the automatic determinationof cholesterolBy Pierangelo Bonini, Claudio Musitelli,Ferruccio Ceriotti and MichelangeloMurone, Laboratorio di Ricerche Cliniche,Ospedale San Rqffaele, 20132 Milano,ItalyThe enzymatic determination of choles-terol using commercial kits requires from10 to 12 min. incubation at 37C, for fulldevelopment of the colour. The incub-ation time can be reduced by increasingthe amount of the pancreatic cholesterol-esterase in the reaction mixture (whichincreases the cost of the determination),or by replacing it with a more active one.By introducing a bacteric cholesterol-esterase, the 37C incubation time wasreduced to 4 min. This method has beensuccessfully applied to the Multistat IIImicrocentrifugal analyser with a bichro-matic procedure. The results correlatewell with those obtained by the manualprocedure.
96
Session 21.1: Microbiology21.1.1: Laboratory evaluation ofthe Kirby-Bauer Disc diffusion test and the Autobac-I systemBy E. Boquet, C. Romeu, D. XairO andJ. L. Artigalgts, Centro Hospitalario deManresa, C Bases 6y8, Manresa, Spain
The results obtained with the standard-ized disc diffusion test were compared tothose obtained by an automated suscept-ibility testing system, the Autobac-I(Pfizzer). The comparative investigationinvolved over 517 micro-organisms iso-lated from clinical specimens. The anti-microbial agents tested were: nalidixicacid, amikacin, ampicillin, carbenicillin,cephalosporin, chloramphenicol, cloxa-cillin, colistin, sulfamethoxazole + tri-methoprim, erytromycin, gentamicin,nitrofurantoin, penicillin, tetracyline, andtobramycin. The similarities and differ-ences of the methods were investigated.Discrepancies were classified as (a) verymajor- susceptible by Autobac I,resistant by disc diffusion; (b) major--resistant by Autobac I, susceptible by discdiffusion; (c) minor--intermediate in oneof the two methods.
The overall average for the 15 anti-microbial agents tested was 91-96--nalidixic acid was 95"14, amikacin82.74, ampicillin 82.21, carbenicillin95.38, cephalosporin 91.65, chloram-phenicol 100, cloxacillin 92"86,colistin 97.26, sulfamethoxazole + tri-methoprim 92.80, erytromycin 97.27,gentamicin 97.01, nitrofurantoin88.82, penicillin 77" 17, tetracycline96"86 and tobramycin 92.46.
The values of medium discrepancieswere: very major 1.27, major 2.16 andminor 4.61%.
21.1.2: An evaluation ofthe Replireader forthe identification and sensitivity determi-nation of Gram negative bacilliBy G. Mascart, G. Gazon, and Ph.Mascart, Department of ClinicalPathology, Edith Cavell Clinic, rue EdithCavell 32, 1180 Brussels, BelgiumThe authors present an evaluation ofthe Replireader, a new semi-automaticsystem for identification and antibioticsusceptibility determination of Gramnegative bacilli. The basic principle ofthissystem is replication on agar plates. 17biochemical plates and 22 antibioticplates were used, and 641 Gram negativebacilli, most of which were Entero-bacticereae were tested. 94.5 of identifi-cations were correct. 2"34 of the strainswere not identified and 3.12 weremisidentified. In this last category half ofthe strains were misidentified only as faras the species, and not the genus, wereconcerned. The antibiotic susceptibilityresults were in close agreement with thoseobtained with the Kirby-Bauer method.Indeed, there was a 97.1 agreement
between the two methods. Minor dis-crepancies amounted to 1"1 and majordiscrepancies 1"7. The Replireaderappears to be an excellent alternative toconventional methods. It is very easy touse. Quality of work can be higher thanwith conventional methods as a result ofthe systematic quality control of theplates and more complete biochemicaldata is used. One single inoculum is usedfor identification and susceptibility deter-mination. It is possible to save time,media and reagents.
21.1.3: Rapid automated susceptibilitytesting with the MS-2 systemBy G. Szilagyi and V. Aning, Hospital ofthe Albert Einstein College of Medicine,Bronx, New York, USAEarly reporting of antimicrobial suscept-ibility test result of bacterial isolates fromclinical specimens is an important factorin the better care of hospitalized patients.Several semi-automated and automatedinstruments have been developed forrapid susceptibility testing. The MS-2microbiology system (Abbott Labora-tories) is a computerized automatedsystem which performs this analysis byphotometrically monitoring turbidi-metric changes that result from bacterialgrowth. The MS-2 system was evaluatedfor the accuracy of the antimicrobialsusceptibility test results. A total of 482clinical isolates (396 gram negative and 86gram positive) were tested and the resultscompared with those obtained from theconventional disc diffusion method. Dis-crepancies were categorized as verymajor, major and minor. Agreementwas determined by organism and byantimicrobial agent. Full accord (all dis-crepancies considered) of 96.6 andessential accord (minor discrepanciesexcluded) of 98.1 was found with gramnegative organisms. With gram positiveorganisms 84.0 full accord and 92.9essential accord was observed. Mean testtime for all isolates was 4.2 h. Suscept-ibility testing with the MS-2 system wasfound to be an accurate and rapidmethod, readily adaptable to the workflow in the authors’ clinical microbiologylaboratory.
21.1.4: Comparative study of antibioticsensitivities of bacteria with manual andautomated methodsBy Lszl6 Sa196, Ervin Sziilliksi andErzskbet Nagy, Department ofPaediatrics,Department of Clinical Microbiology,Central Laboratory, University MedicalSchool, Szeged, HungaryOnly during the past few years has auto-mation made a significant impact in theclinical bacteriology laboratory. TheABAC (Sclavo, Siena, Italy) automaticsystem was tested. A 4 h broth-culturewas prepared from the bacteria isolatedfrom 400 clinical patients; this served asthe standard starting material for thevarious methods. Depending on the
International Congress on Automation in the Clinical Laboratory: Abstracts
bacteria isolated, the sensitivity wasexamined for four different antibiotic orchemotherapeutic groups. The disc-diffusion test method generally employedin Hungary was compared with theresults given by the ABAC system, andwith the micro-organism drug-sensitivi-ties determined under quantitative con-ditions; the latter were also obtained bymicro and macro procedures. The resultswere evaluated, and the applications ofthe methods and the advantages of theautomatic technique were assessed.
21.1.5: Antimocrobial susceptibility testingof Streptococcus pneumoniae by twomethodsBy J. L. Prez, J. Lifiares, M. C. Marneand S. Romero, Servicio de Microbiologia,C. S. ’Principes de Espafia’, Hospitalet deLlobregat, Barcelona, Spain
Thirty clinical isolates of S. pneumoniaewere tested for susceptibility to penicillin,ampicillin, tetracycline, chloramphenicol,erythromycin and clindamycin by thestandard agar dilution method. Resultswere compared to those obtained usinga commercial microdilution product,Sensititre, in which Todd-Hewitt brothwas supplemented with 5 defibrinatedwhole horse blood. Among the 30 strains,two were resistant (MICs 2 mcg/ml), twowere relatively resistant (0" 12-0.5 mcg/ml)and 25 were susceptible to penicillin byboth methods. Only one strain showed athr.ee dilutional step difference by micro-dilution broth and agar dilution, resultingin categorization as susceptible by theformer method but relatively resistant bythe latter.
The Sensititre and agar dilutionminimum inhibitor, concentrations(MICs) were equivalent within +_ dilu-tion in 96?/o of the comparable test resultsfor penicillin, ampicillin and chloramphe-nicol. The correlation was excellent forerythromycin, clindamycin and tetra-cycline.
The micro-broth dilution technique isa reliable, simple and convenient semi-automated method of determining MICsofS. pneumoniae, with the advantage ofanextended shelf-life when stored at roomtemperature.
21.1.6: Microdilution aminoglycosidesusceptibility testing of Pseudomonasaeruginosa with divalent cation sup-plemented inoculumBy M. C. Marne, J. Lifiares, A. Gas6s,N. Parrero and J. L. Pbrez, Servicio deMicrobiologia, C. S. ’Prncipes de Espaa’,Hospitalet de Llobregat, Barcelona, SpainThe concentration of magnesium andcalcium in media has been cited as afactor affecting susceptibility results ofaminoglycoside testing of P. aeruginosa.The purpose of this study was to comparethe minimum inhibitory concentrations(MICs) obtained for three aminoglyco-sides (gentamicin, tobramycin and ami-kacin) by the sensititre technique using
unsupplemented and supplemented in-oculum, with the agar dilution MICs.
Thirty one clinical strains of P. aeru-ginosa, previously checked as aminogly-coside susceptible by the disc diffusionmethod, were used. Tests were run inparallel with Mueller-Hinton broth un-supplemented and supplemented with 25mg ofmagnesium and 50 mg ofcalcium/1.Agar dilution MICs were determined byconventional methods. The susceptibilityin supplemented broth had 100 agree-merit _+ one doubling dilution) with agardilutions for the three aminoglycosides.With the unsupplemented broth P. aeru-ginosa was found to be more susceptiblethan with the agar dilution method.Differences ranged from two to fourdoubling dilutions for gentamicin, two tofour for tobramycin and one to three foramikacin.
Divalent cation content of the testingmedia must be controlled for validcomparison between antipseudomonalantibiotics. The use of the sensititremicrodilution procedure for MICs deter-minations with supplemented brothcompares well with standard agardilution methods, but is simpler toperform.
21.1.7: Evaluation of automated proce-dures for monitoring antibiotic treatmentof acute bacterial infectionsBy A. Visconti and U. Marini, Laboratoryof Clinical Microbiology and DepartmentofInternal Medicine VI, Ospedale S. CarloBorromeo, Milan, ltalyThe outcome of treatment in acute bac-terial infections is very dependent on therational choice of proper antimicrobialagents. For that purpose previoussusceptibility testing of clinical isolatesshould be mandatory whenever possible.The most useful and important inform-ation from such testing is expected whenseveral parameters, such as minimalinhibitory (MIC) and bactericidal (MBC)concentrations, MBC:MIC ratio (ortolerance), killing rate, growth curvesinhibition, drug level in body fluids,bacterial activity of the host serum andopsonic requirements, are simultaneouslyevaluated at the right moments duringtherapy.
A number of patients with acutebacterial infections (pneumonia, menin-gitis, sepsis, pyelonephritis etc.) werestudied by testing all these parametersusing automated procedures, such asturbidity (MS-2 Research System)and bioluminescence (Luminometer)methods. The effects of inoculum, media,time of incubation, and other experi-mental variables were also evaluated.
The results are presented and dis-cussed by the authors.
97
International Congress on Automation in the Clinical Laboratory: Abstracts
Session 21.2: ClinicalChemistry Immunoassay21.2.1: Comparision of automated andmanual techniques for the determinationof creatine kinase-MBBy J. P. Chapelle, Department of ClinicalChemistry, University of Liige, rue desBonnes-Villes, B 4020 Liege, BelgiumThe immuno-inhibition method for thedetermination of creatine kinase-MB(CK-MB) activities has been automatedby the author and the results have beencompared with those of two manualtechniques: ion-exchange chromat-ography and polyacrylamide gel electro-phoresis (PAGE). The precision of theautomated technique was evaluatedthrough the measuring range of CK-MB,providing values of 0"8 to 4-3% CV. Thelinearity limit was 250 IU/L.
The automated immunologicalmethod correlated well with the chromat-ographic technique (r=0"98). For someserum specimens, however, the twotechniques led to conflicting results, theimmunological method indicating highCK-MB values (> 15% of total CK)whereas activities determined afterchromatographic separation were withinthe normal range. PAGE indicated inthose cases the absence of CK-MB andthe presence of abnormally migratingCK-MM sub-bands. These forms werenot inhibited by the anti-CK-M anti-bodies and were therefore responsible foran apparent CK-MB activity when theimmunological technique was used. Theabnormal CK-MM sub-bands were notassociated with myocardial damage.
The automated immuno-inhibitiontechnique offers advantages of timesaving, increased precision and accuracy,and is easily adaptable to the laboratoryroutine. However, when elevatedCK-MB activities (15-30% of total CK)are found, the presence of this isoenzymemust be confirmed by electrophoresis.
11.6% for T3-binding capacity. Thenormal range was identical to that of themanual test-kit.
The results obtained for serum thy-roxine concentration and T3-bindingcapacity from more than 200 patientswere compared with all others in vitro (bythe TRH stimulation test); in addition,clinical in vivo investigations were made.Significant differences were found amongthe groups with euthyroid goiter withoutand under oestrogen medication, preg-nancy, treatment with thyroxine, normals,hyper- and hypothyroid diseases. Toevalute the diagnostic validity of theresults the authors compared the per-centage deviation from the normal rangewithin each group.
Although the determination of thy-roxine, T-binding capacity and FT4-index make it easy to recognize diseases,further in vitro and in vivo parametersshould be investigated in some cases.
21.2.3: An automated enzyme immuno-assay for serum ferritinBy T. Carter, R. Bunce and T. P.Whitehead, Wolfson ResearchLaboratories, QEMC, Birmingham, UK’Sandwich’ enzyme immunoassay (EIA)is becoming an increasingly popularalternative to radioimmunoassay. Whenperformed with polystyrene tubes ormicro-titration plates as the solid phase,precision is often unacceptably poor and,moreover, these are unsuitable for auto-mation. An EIA system using antibody-coated polystyrene balls and a specially-designed incubation module for usetogether with components of the KoneCD Parallel Analyser has been devel-oped. The system is versatile and allowsthe simultaneous analysis of96 specimensin approximately 3 h. In an assay forserum ferritin, the detection limit wasabout ng/ml with a working range upto 300 ng/ml and a precision of 6.5 at100 ng/ml.
21.2.2: Automated RIA determination(125j) of serum thyroxine concentrationand T3-binding capacityBy S. Knapp and N. Panitz, MedizinischDiagnostisches Institut Dr Kapp, 6500Mainz, FR Germany and Deutsche Klinikf. Diagnostik, FB Nuklearmedizin, 6500Wiesbaden, FR GermanyTo reduce the difficulties with incubationand separation ofmechanized RIA tests anew kit was developed. After incubationin a continuous-flow system the boundfraction was separated by magnetificationofantibody-carrying particles (TechniconSTAR). The bound fraction of the 125j.labelled antigen was then measured.
Precision, accuracy and correlationwere the same as in a manual test-kit(PEG separation). The inter-assay coef-ficients of variation ranged from 3.1% to5.3% for thyroxine and from 4.6% to
21.2.4: Automated labelled antibodyimmunoassaysBy Heather A. Kemp, Rhys John andJ. Stuart Woodhead, Department ofMedical Biochemistry, Welsh NationalSchool of Medicine, Card!if, UKA range of labelled antibody assays foruse on a fully automated microprocessor-controlled immunoassay system (Kemtek3000) has been developed. These auto-mated assays are capable of rapidlyprocessing large numbers of samples.Examples ofassays where this technologyhas proved particularly valuable arethose of el-fetoprotein (AFP) and thyro-tropin (TSH), which are both used ex-tensively in population screening. Bothare measured by the two-site method inwhich serum samples are reacted eithersimultaneously or sequentially withexcess labelled antibody and solid phaseantibody. Radioactivity bound to the
98
solid phase is a function of the concen-tration of antigen present. Using acellulose-linked antibody and 125I-labelled antibodies, a sensitive AFP assay(detection limit 6 KU/1) can be obtainedwith a h reaction, followed by separ-ation of the solid phase by means offiltration. Up to 100 patient samples canbe processed in a total of 3 h. A screeningtest for congenital hypothyroidism hasbeen developed which uses 6 mm discs ofdried blood on filter paper from neonates.Samples are reacted overnight with 25I-labelled antibody and then for a further2 h with cellulose-linked antibody, andfinally subjected to automated filtration.The assay has a detection limit of 3-75 #/1and a CV of only 6% at a TSH concen-tration of 30 //1. Of 14500 samplesscreened so far, only six repeats have beennecessary; of these two were confirmedcases ofhypothyroidism. The flexibility ofthe Kemtek 3000 and the advantages oflabelled antibody technology make theprocedures suitable for high-throughputrapid-turnround immunoassays.
Session 21.3: Immunology21.3.1: Circulating immunocomplexes.Contribution to a new automated tech-nique for their quantificationBy F. L. Elorza, F. Malagon, M. E.Dorado and A. Moreno, Dpto. BioquimicaClinica, Hospital Universitario de Sevilla,SpainNephelometry is a turbidimetric tech-nique, which is based on the dispersion ofa beam of light when it falls on aprecipitate dispersed in a liquid. Quantifi-cation ofcirculating immunocomplexes isvery useful for the diagnosis and follow-up of a great number of pathologies(hepatic disease, collagen disease etc.).Adopting a low-cost automated tech-nique for this type of quantificationwould be of great utility to patient care inlarge hospitals. The authors present theirexperience with 1300 patients, classifiedaccording to their pathologies, andcorrelate the results of the technique withBach’s manual technique.
21.3.2: Determination of alpha-fetoprotein(AFP) in the serum of patients withpsoriasis vuigaris by the particle countingimmunoassay (PACIA)By W. Krotz, A. Leek, Angelika Welterand V. Hochstein, Heinrich-Pette-Institutffir Experimentelle Virologie undImmunologie an der Universitiit Hamburg,Martinistrasse 52, D 2000 Hamburg 20;and Technicon Chemicals Company,Animalerie Centrale 56.20, AvenueHipocrate 56, B 1200 Brussels, BelgiumPUVA treatment with psoralen (P) andlong-wave ultra-violet light (UVA) of
psoriasis vulgaris is associated with anincalculable long-term risk. For example,a certain structural similarity exists bet-ween photosensitiziers of this type andthe aflatoxins. This gave rise to an in-vestigation of the AFP concentration inthe serum ofpsoriasis patients, employingthe PACIA system developed by A. Leekand using modified DAKO antisera. Theprinciple behind this new technique is theagglutination of coated latex particleswith the corresponding immunochemicalreactant. The latex particles were coatedwith antibodies of AFP. PACIA is acontinuous-flow analysis with an auto-matic particle count before and after theimmunochemical reaction. Using thisassay, AFP concentration in the serum ofmore than 100 psoriasis patients wasexamined. Patients who had received truePUVA long-term treatment (at least 10years) were not included. The PACIA-AFP values of health subjects were up to20 ng/ml; on average these were approxi-mately 10 ng/ml. In contrast, one-third ofthe psoriasis patients examined hadvalues of 20 to 350 ng/ml. The study willbe continued with long-term controls ofpsoriasis patients treated with PUYA.
Session 21.4: ClinicalChemistry Chroma-tography
21.4.1: Liquid chromatographic assay ofsome anthraquinone glycosidesBy O. O. Komolafe, Department ofPharmaceutical Chemistry, University ofIre, Ile-Ife, NigeriaThe current methods of assay of someanthracene purgative drugs, which deter-mine both the free and combined anthr-acene derivatives, lack sensitivity andspecificity because the main pharmacolo-gical activity of the anthracene derivat-ives is shown by the glycoside form. Thiswork reports paired-ion chromatogra-phic analysis of the major glycosides ofsenna pod and some pharmaceutical pre-parations of senna using a Corasil C18column with methanol (30 in water)containing 0.005 M tetrabutylmmoniumphosphate (at pH 7"5) as solvent. Liquidchromatographic analysis of the samecrude drug and pharmaceutical prepar-ation therefrom have been carried outusing a weak anion-exchange resin withammonium nitrate solution (0"1 M, pH7’5) as the mobile phase. These systemsallowed for the determination of senno-sides A and B in nanogram quantities,free from other adjuncts and excipientspresent and without derivatization and itsinherent problems. The chromatographicmethods used in this work cut the assaytime for the drugs to about half that
International Congress on Automation in the Clinical Laboratory: Abstracts
required by current assay procedures,and they are recommended for use inautomatic systems.
21.4.2: Significance of CK values in anadult and paediatric population seen in auniversity medical centreBy Horacio A. Perez, Ronald Busch,Gary E. Blank, and Ajit Sanghvi, ClinicalChemistry Laboratory, University HealthCenter of Pittsburgh, 203 De Soto Street,Pittsburgh, PA 15261, USAThe central laboratory of the authors’health centre has been, for several yearsnow, measuring levels of total CK andCK isoenzymes in serum with theABA-100 bichromatic analyser (AbbottLaboratories) at 30C. Isoenzyme separ-ation is obtained using ion-exchangechromatography columns based onMercer’s method. A retrospective studywas conducted comprising 342 patientsconsecutively admitted to the hospitalsduring a six-month period, in whomabnormally elevated serum levels of totalCK activity were present on admissionand/or on subsequent days. A total of 107patients (31) comprised the paediatricgroup admitted to the Children’sHospital. Elevated CK-MM isoenzymelevels were present in accidents involvingtrauma to skeletal muscle or brain.Elevated CK-MB levels in this groupwere seen in patients with Duchenne’smuscular dystrophy, Reye’s syndrome,and patients who had undergone com-plete liver transplantation. One patientwith biliary atresia showed increasedCK-BB activity in the serum. The adultpatient group (69)consisted largely ofpatients admitted to the Coronary CareUnit with suspected and/or diagnosedacute myocardial and/or subendocardialinfarction, or infarct extension. Othercases in which CK-MB activity was alsodemonstrated include patients withdermatomyositis, and patients who hadundergone complete liver transplant-ation. Finally, patients in whom elevatedtotal CK activity was present in the serumwithout concommitant elevation of CKisoenzymes were found to have pul-monary embolism, true angina pectoris,severe pneumonia, pericarditis, cerebro-vascular accidents, general surgery andcardiac catheterization. Also, cases ofcardiac defibrillation and intramuscularinjection were identified.
Session 21.5:MathematicsClinical Chemistry21.5.1: An automated system of qualitycontrolBy Wieslaw Piechota, NorbertSymonowicz and Wojciech Minta, Centreof Postgraduate Medical Education,00-909 Warsaw 60, Szaserow 128, Poland
No single method of quality control iscapable of detecting all possible errorswhich may affect laboratory diagnosticinformation. Therefore several methodsshould be used routinely. Such a task,however, poses a great burden on alaboratory. In order to ease this task,several basic methods of quality controlwere combined into one coherent systemusing a minicomputer. These methods arebased on: (1) results of measurements ofcontrol specimens; (2) very unlikely andabsurd values; (3) averages of patients’normal results (for a given number orbatch) and for daily values; (4) com-parison of a result with a patient’sprevious result; and (5) physiological andpathological correlation of results of thesame patient. The algorithms used in thesystem, with the criteria of correctness ofresults, are presented. The results to bechecked are listed with marks indicatingwhich of the criteria is (are) not met. Therelation between the methods employedare discussed (statistical and arbitrarynature of some accepted criteria, forexample precision and comparison of thesame patient’s results etc.) as well as theadvantages for technicians and chemicalpathologists.
21.5.2: Indirect establishing of referenceintervalsqis its value always questionable?By Wieslaw Piechota, NorbertSymonowicz and Zofia Pastusiak, Centreof Postgraduate Medical Education,00-909 Warsaw 60, Szaserow 128, Poland
The introduction of an operational term’reference intervals’ instead ofan idealisticand abstract ’normal ranges’ has revivedthe idea of seeking indirect methods tocompute reference intervals. However,despite a few attempts, the simplest andthe oldest method utilizing probabilitypaper has not yet been evaluated com-pletely. This method was evaluated tak-ing into account the following conditions:(1) application of the method to popu-lations which were supposed to be’healthier’ than hospital patients, namelysubjects who underwent preventive check-ups (N 1606) and ambulatory patients(N 1768); (2) application of the methodto Technicon SMA 12/60 profiles toavoid bias due to selection of a test by aphysician; (3) taking into considerationdemographic factors by establishing sexand age dependent ranges; (4) analysing(prior to establishing reference ranges)frequency distributions of 11 constituents
99
International Congress on Automation in the Clinical Laboratory: Abstracts
in healthy subjects by means of the chi-square test and coefficients of skewnessand kurtosis and using transformationsif necessary. According to Cherian’scriterion, which is based on precision ofanalytic methods, the reference rangesobtained with the indirect method wereno different from the intervals establishedin the conventional way for 875 healthyindividuals.
21.5.3: Quality control of densitometermeasurements in protein assayBy A. Lopez, F. Lema, R. Perez, S. Bandinand R. Alvareg, Laboratorio de Bioquimica,R.S. ’Almirante Vierna’, Vigo, Spain
For the measurement of serum proteinprofiles, commercial control material wasused to investigate the within-day and thebetween-day reproducibility of the com-plete procedure and of the densitometrycomponent. Linearity and material stab-ility were also studied.
(1) Densitometry accounted for approxi-mately half of the total CV.
(2) The between-day CVs for the wholeprocedure and for the densitometrycomponent are both increased pro-portionately compared with corres-ponding within-day CVs.
(3) Linearity for the different proteinfractions is not the same.
(4) When the reconstituted controlmaterial is kept at room temperatureand at 4C, the CVs remained withinexpected limits over the time of thestudy.
study has been performed on each of thedistributions, followed by longitudinaladjustment of the same to polynomicfunctions. Results and conclusions arepresented.
21.5.5: Measurement of creatinine bycontinuous flow. Contribution of statisticsto obtaining accuracyBy O. Houot, L. Lepage and R. Gueguen,Centre de Mb.decine Prb.ventive, 2 Avenuedu Doyen Jacques Parisot, 54500Vandoeuvre-les-Nancy, France
The measurement of creatinine in plasmaor serum by Jaff6’s method is known to beaffected by numerous interferences. Manymodifications of the method have beenproposed to partially eliminate the reac-tion of other endogenous compoundswith the picric acid. The authors discuss astatistical method which enables thecalculation ofknown interferences due toprotein, glucose and urea. The referencemethod used for this calculation is basedon the specific adsorption ofcreatinine byLloyd’s reagent (Haeckel). The formula isbased on the correlation coefficientsmeasured between the different con-stituents, and the results obtained bycalculation are shown to be comparablewith those of the reference method. Theformula is applied to results determinedfor creatininemia on the SMA II(Technicon) on supposedly healthy sub-jects. The authors evaluate the relation-ship between the calculated creatinineand physiological parameters (height andweight for example).
21.5.4: Longitudinal study of 20biochemical parameters (SMAC) inchildren:By P. Gomez, C. Coca, C. Vargas andA. Martinez, Servicio de Bioquimica:Hospital Materno-Infantil, C.S.S.S., deOctubre, Ctra. de Andalucia Km 5.3Madrid, Spain
Serum samples were taken from 1800children attending the out-patient clinicsat the Primero de Octubre Maternal-Child Hospital from July 1980 to October1981. The children’s ages ranged from afew days old to 14 years and they weregrouped by year. The following weredetermined simultaneously for eachserum: glucose creatinine, uric acid, in-organic phosphorus, calcium, sodium,potassium chloride, iron, cholesterol, tri-glycerides, total protein, albumin, totalbilirubin, creatinphosphokinase, lacticdehydrogenase, alkaline phosphatase,total carbon dioxide, alanine aminotrans-ferase, and aspartic aminotransferase in acontinuous flow autoanalyser (SMAC).
The statistical method includes codifi-cation of data and processing on an IBM370, carrying out a transverse study ofeach of the biochemical parameters, ana-lysis of respective distributions, and elim-ination of outlying observations by theGrubbs method. Likewise, a percentile
Session 21.6:Mathematics--Haematology21.6.1: Erythropoietic model in evaluatinghypokinetic effectsBy Y. G. Zorbas, Yu. M. Portnoy andM. D. Dilek, Life Science Division,European Institute of EnvironmentalCybernetics, Athens 514/1, Greece
Modelling the system of erythropoiesis isa basic component in presenting or de-scribing processes connected with anorganism’s reliability (adaptive capacity)to hypokinesia (HK) (limited motoractivity) or weightlessness. The aim ofthisreport is to describe a mathematicalmodel of erythropoiesis and its regu-lation, which takes the age structure ofthered blood cell population into account. Aflowchart of the model containing theprincipal elements and factors was pre-pared. Biological and medical conceptswere applied in order to construct theflowchart. It was concluded that thismathematical model can be considered in
100
evaluating the influence of prolongedhypokinetic conditions, both in examin-ing the state of the erythropoietic systemproper and describing a certain set ofmodels that simulate the integral effect ofhypokinesia on the organism.
Session 22.1:Haematology Evaluation22.1.1: Comparative study between thehematimetric values obtained with auto-matic methods (Haemalog 8/90 andCoulter S systems) and manual methodsBy A. Viloria, M. Fernitndez de Castro, C.Larrocha, M. C. Jimbnez Herrdtez, A.Ferndndez Zamorano, A. Sicilia and J. L.Ferndtndez-Chacdn, Departmento deLaboratoria, C.S.S.S. ’La Paz’, Madrid,Spain
Values obtained for leukocytes, red cells,haemoglobin and packed cell volumewith a Hemalog 8/90 system are com-pared with the results obtained with aCoulter S system. 204 samples were pro-cessed for the leukocyte study, the corre-lation index was 0.97 and the linearregression line was y 0.35 +0"98 x (wherey=values obtained with the Coulter Ssystem and x values obtained with theHemalog 8/90 system). 203 samples wereused for the red cell study, the correlationindex was 0"90 and the linear regressionline was y=0.03 +0.97 x. 207 sampleswere used for the haemoglobin study, thecorrelation index was 0"99 and the linearregression line was y=0"16+ 1"01 x. 200samples were used for the packed cellvolume study, the correlation index was0.92 and the linear regression line was y=1"03+0"96 x. Also, the values forhaemoglobin and packed cell volumeobtained with a Hemalog 8/90 systemwere compared with the values obtainedby manual methods (cianmetahaemo-globin and microhematocrit). 95 sampleswere used in the haemoglobin study, thecorrelation index was 0"99 and the linearregression line was y=0.12+0.97 x. 93samples were used in the packed cellvolume study, the correlation index was0"99 and the linear regression line wasy 1"46 + 1"01 x.
There is a good correlation betweenall the parameters studied, especially inthe case of haemoglobin (Hemalog 8/90versus Coulter S and Hemalog 8/90versus manual methods), in the case ofleukocyte (Hemalog 8/90 versus CoulterS) and that of packed cell volume(Hemalog 8/90 versus manual methods).
22.1.2: Comparison between four auto-mated systems for leukocyte differentialcountBy J. M. Jou, J. L1. Vives-Corrons, J. L1.Aguilar, M. J. Insa, A. Ester, J. Del Olmoand E. Colomer, Laboratori Central
Four automated systems for leukocytedifferential count were evaluated" (a)Hematrak-360, (b) Diff-3, (c) ADC-500and (d) Hemalog D/90. In systems a, b andc an automated leukocyte differentialcount was based on morphological inter-pretation from Wright stained smears,and in system d it was based on flowcytochemistry and size analysis ofEDTA-treated whole blood. Blood samples wereobtained from 890 ambulatory adultsand 1080 hospitalized non-haematologicpatients. The manual counts were inter-preted by four technologists over a periodof 20 months. For each instrument theparameters studied were’(1) sample pre-paration for analysis; (2) reproducibility(c); (3) correlation coefficient (r) betweenthe automated differential counters andthe manual method; (4) comparison ofthevariations in percentage for all cell classesdetected by the manual method versusthe automated differential counters. Ininstruments a and d red blood cellmorphology and semi-quantitative plate-let appreciation were also studied.
The results obtained showed good cand r for neutrophils and lymphocytes inall systems analysed but deficient for’bands’ (given by instruments a, b and c)and monocytes. Sample preparation wascumbersome in system b and differencesin all cell class percentage between countand automated systems varied from10.5 to 25.
22.1.3: Evaluation of the Technicon H600:preliminary data on utilization and signi-ficance of the red cell distribution width(ROW)By C. Izzo, L. Miano and A. M. Ferrari,Centro Diagnostico ltaliano, Via SaintBon 20, 20147 Milan, Italy
Around 80000 whole blood counts andleucocyte differential counts are performedat the Centro Diagnostico Italiano eachyear. A Technicon H6000 was acquired in1981. The instrument determines the sevenconventional haematological para-meters (WBC, RBC, HGB, HCT, MCV,MCH, MCHC), leucocyte differentialcount, platelet count and indices (meanplatelet volume and plateletcrit), sizehistograms and distributional width forred cells and platelets simultaneously oneach blood sample. Analyser imprecisionand correlation between H6000 andCoulter S results were evaluated. Clinicalutilization and significance of the redcell distribution width (RDW) in somehaematological disorders was also studied.
Imprecision was determined for low,medium and high values of all para-meters, testing four replicates of 60 bloodspecimens each day for 20 days. Resultsobtained in all three groups of valuesshowed a CV lower than 2.8 for
International Congress on Automation in the Clinical Laboratory: Abstracts
conventional haematological parameters,3"6 for leucocyte differential count,6"6 for platelet count and indices. Dataon correlation between the H6000 andCoulter S, analysed by the Deming’smethod, showed a correlation coefficientranging from 0.907 to 0.960. Clinicalsignificance ofRDW was evaluated on 65patients with B-Thalassemia (TH), 62with iron deficiency (ID), 61 with mis-cellaneous anaemia (MA) in comparisonwith 1500 normal controls (NC). Resultswere respectively: 23.2+1.4 for TH,19.3+1.6 for ID, 15.9_+2.1 for MA,16.5+_ 1"0 for NC. To investigate thesignificance of these findings a study oncorrelation between RDW and otherparameters (MCV, osmotic fragility,serum iron and o Hb A2) was carried out.A good negative correlation was observedonly with MCV and osmotic fragility (r
0"907 and 0" 894 respectively).The Technicon H6000 is shown to be
a reliable instrument, which, despite tech-nical differences, shows good correlationto Coulter S. The RDW index on theH6000 seems to be related to red celldistribution width as well as to red cellmean volume; this feature makes itparticularly suitable for some clinicalpurposes.
22.1.4: Evaluation of Technicon H6000haematology system to handle all routinehaematology tests in a rapid, accurate andcost-effective mannerBy P. W. Stiffler and C. Joson, DamonClinical Laboratories, Inc., 4720 W.Montrose Ave., Chica9o, Illinois 60641,USA
The Technicon H6000 system is a onework-station, fully automated haema-tology analyser that provides in a singlereport the RBC, WBC, Hgb, HCT, MCV,MCH, MCHC, platelet count, plateletsize histogram, red cell size histogram anda 10000 cell WBC differential (bothabsolute numbers and percentages ofleukocyte population) plus a permanentmonolayer slide all from a 0"55 ml wholeblood sample. Samples are processed at arate of up to 90/h with lag time of 10 min.by flow-through cytometry with electro-optical counting and sizing technologyplus cytochemical staining.
For the 400 samples analysed, theoverall correlation for the H6000 CBCand platelet counts with the Coulter S +
PLUS was 0"95. The correlation for theH6000 10000 cell differential with thestandard manual 100 cell differential was0.85 for polymorphonuclear leukocytesand lymphocytes for the 160 normalsamples analysed. This was mainly due tothe imprecision of the manual counts.Precision studies on the H6000 on dup-licate samples was far superior to themanual readings. The H6000 system isreliable and accurate for screening out thenormal human blood samples in a patientpopulation (80o) so that only the speci-mens having blood cell abnormalities
(20o) need to be screened by haematol-ogists and/or pathologists.
Therefore in the high-volume refer-ence laboratory, when the majority ofblood samples are normal human blood,the H6000 makes it possible to increaseprecision and accuracy significantly andshorten throughput time while reducingthe number of haematologists needed todo the work.
22.1.5: H-6000--a unique haematologyanalyserBy Jean Pierre Ganon, 39 Boulevard dela Muette, 95140 Garges-les-Gonesse,France
The H6000 provides a complete haema-tology profile from a blood sample. The12 CBC parameters with platelet count,RBC size histogram (Price Jones curve),platelet size histogram along with the XYWBC distribution pattern are reportedon one form. With the autoslide option aWright stained smear is also providedwith the results. The design and technicalfeatures of the instrument are describedand its application to medicine isdiscussed.
22.1.6: The evaluation of two years’ ex-perience with Technicon’s Hemalog 8 andHemalog DBy A. Lu?:i?, R. Gruji6, K. Beri(, Z.Radujkov, D. Kerac, L. Mani6, D. Divjakand K. Anti:, Institute of PathologicalPhysioloty and Laboratory Diaonostics,Medical Faculty, Novi Sad, I-IajdukVeljkova 1, YugoslaviaBasic haematological profiles, includingdifferential leukocyte counting, have beencompiled for a group of 25 000 personswith the automated systems--Hemalog 8and Hemalog D. In the pathologicalreliability the haematological parametersobtained, have been evaluated and com-pared with results obtained by manualmethods. The authors have demonstrateda positive correlation between resultsobtained by automated systems and thosefrom the manual technique. Useful dataon patients with blood malignant diseaseswere obtained with the Hemalog D.
Session 22.2:Heamatology General
22.2.1: Correlation of the HPX and thetotal white count by optic microscopyBy B. Cuesta, E. Rocha, J. A. Ptramo,A. L@ez Fernandez, R. Sefrbs Corderoand J. Fernandez, Clinica Universitaria,Facultad de Medicina de la Univerida deNavarra, Pamplona, Spain
1000 blood samples analysed with aTechnicon Hemalog-D and by conven-tional microscopy were compared. Corre-lations between the HPX values with data
101
International Congress on Automation in the Clinical Laboratory: Abstracts
given by the Hemalog-D were establishedfor (a) Neutrophils; (b) Lymphocytes;(c) LUC (large unstained cells); (d)remainder; (e) total white cell count. Andcorrelations between the HPX valuesobtained by conventional microscopy for(1) number of granulocytes (neutrophils,bands, immature granulocytes); (2) num-ber of neutrophils; (3) number of bands;(4) immature granulocytes (HPX) werefound. The average and standard devi-ations of HPX were obtained for thoseindividuals who are out of the twostandard deviations, with their neutrophilcount (given by the Hemalog-D) and thenumber of granulocytes obtained bymicroscopy.
22.2.2: Automated cytochemistry in thediagnosis of several haematologicaldisordersBy M. Fern{mdez de Castro, A. Viloria,M. C. Jimknez Herrdez, C. Larrocha,C. Alonso, J. L. Ferndndez-Chacdn andJ. M. Gdmez Mantilla, Departmento deLaboratorio, C.S.S.S. ’La Paz’, Madrid,Spain
Corresponding features from diversehaematological disorders, obtained froma study on automated cytochemistry, aredescribed. The Hemalog D system wasused for the analyses; this provides adifferential count of the diverse groups ofblood leucocytes, based on their cyto-chemical behaviour. Different types ofleukaemia, both acute and chronic, havebeen assayed, establishing the quanti-tative measure of leukaemic and non-leukaemic cells with greater accuracy andprecision than is possible with the manualdifferential. Different types of lympho-proliferative and myeloproliferativediseases were studied, as well as infectiousmononucleosis, Hodgkin’s disease,myeloma and other disorders likemyeloperoxidase deficiency, erythroblas-tosis and the eosinophilia due to intoxi-cation by toxic colza oil. Characteristicfeatures in the Hemalog D results, eachone corresponding to a pathologicalentity were obtained in all cases. Theautomated cytochemical method is agood system for detection of severalhaematological disorders.
22.2.3: Hereditary myeloperoxidase de-ficiency: study of 12 casesBy M. C. Jimnez Herrez, A. Viloria,M. Fernandez de Castro, C. Larrocha,J. L. Ferndndez-Chaon, M. Salinas andJ. M. Gbmez Mantilla, Departamento deLaboratorio, C.S.S.S. ’La Paz’, Madrid,Spain
This report concerns 12 cases ofhereditarymyeloperoxidase (MPO) deficiency. MPOdeficiency in the azurophil granulationsof the neutrophils and monocytes wasdetected by an automated cytochemistrysystem (the Hemalog D system). Cyto-chemical stainings included peroxidase,
Sudan Black B, chloroacetate esteraseand alkaline phosphatase. Quantitativeenzymatic assay of peroxidase, beta-glucuronidase, acid phosphatase, lyso-zyme and alkaline phosphatase, wereperformed, as well as the granulocyticfunction and electron microscopicalstudy. Five different families were studied,including two generations in each family.Automated cytochemical methods areconsidered the best way to detectmyeloperoxidase deficiency.
22.2.4: Pseudothrombocytopenia (PTC)produced by platelet agglutinins in EDTA-treated bloodBy N. Gdmez Gdmez, M. Lozano Alvarez,F. Olmeda, F. G6mez Reino, M. J.Ferhandez-Villalta and J. M. Ferhandez-Rafiada, Gran Hospital del Estado, Diegode Le6n, 62, Madrid, Spain
The authors observed a PTC in EDTA-treated blood from three patients whoseplatelet counts were obtained withCoulter S Plus counters. The smearsshowed a normal platelet number and thepresence of large-size platelet aggregates.The platelet volume distribution curve inblood treated with EDTA at standard orlow concentrations showed the typicalflattening observed in platelet anysoci-tosis. This result was confirmed by theplatelet DW. The mean platelet volumeobtained with Coulter S Plus Counterwas higher than anticoagulants. The ad-dition of platelet-free plasma from thesepatients to a nor-PRP produced a PTCfollowing incubation at 4C and 22C, butnot when incubated at 37C. Tests withneutralizing antibodies and denaturaliz-ation with 2-Dithiotreitol suggest that theplasma factor responsible for this pheno-menon is an IgM type antibody. Giventhe high incidence of this phenomenon itis necessary to confirm all thrombocyto-penias by a second procedure (chamberor Fonio’s) to avoid erroneous diagnosisand subsequent treatment of possibleWerlhof’s disease.
22.2.5: A comparison of automatic andmanual blood-group antibody screening inblood donors for routine testingBy R. L. McShine, P. C. Das, C. Th.Smit Sibinga, Regional Red Cross BloodBank Gronigen-Drenthe, Groningen, TheNetherlands
The Regional Red Cross Blood Bank inGroningen serves a population of 1.2 Mand 13 hospitals, drawing about 56 000donations annually. Therefore, auto-mation of routine screening tests likeABO and Rhesus grouping, hepatitis andsyphilis serology, and screening forirregular blood-group antibodies is ofimportance for running the laboratory.Preliminary studies have shown thatwhen two types ofchannels are combined(low ionic strength polybrene andbromelin methyl cellulose), most of theclinically important blood-group anti-
bodies can be detected, but most of thesestudies were performed on samples frompatients. This paper describes how thechannels (Marsh and Ldezari) were builtfor antibody screening in donor blood.The authors compare auto-analyserresults with manual techniques for salineRT, albumin 37C, indirect antiglobulintest, enzyme 37C and the low ionicstrength method. In January 1981 routineantibody screening was changed frommanual testing to automatic screeningwith an autoanalyser.
22.2.6: Hemalog D and Hemalog 8 appli-cation in peritonitis (P) diagnosis andtreatment on continuous ambulatory peri-toneal dialysis (CAPD)By R. Selgas, M. Fernandez de Castro,A. Rodriguez Carmona, P. Gomez, J. M.Beberide, O. Ortega and L. Sanchez Sicilia,C.S. La Paz, Madrid, Spain
Recurrent P has been found to be themost frequent complication with CAPD;under these circumstances the number ofleucocytes in the peritoneum increases. Acount of leucocytes is very useful forCAPD diagnosis and follow-up. 24patients on CAPD have passed 69episodes of P; white cells count out of Pin 154 samples (Hemalog-8) was: 276 +_ 36(: +/- SEM)/mm
Results: Leucocyte count in P/mmwas 3461 +__ 460 (r 400- 18800) and dif-ferential count (Hemalog-D): neutrophil:88 + 6% (+ S.D.); Linphoc" 6.5 + 6;Monoc: 3-4___ 3; Eosin: 0" 8. No signifi-cant differences between types of P(according to culture) have been found;eosinophilia is not higher in negativeculture P. There is no correlation betweencell counts and delay in sampling orsymptoms (r=0" 12). Previous P, sex andage do not correlate with count. Countsin the first six months are slightly higher(p=0"05) but neutrophil is no different.The results of cultures have only a smallinfluence on the count. Intra-peritonealantibiotics provoke a progressive countdecrease: (+SEM) two-three days846+ 189/mm3; six to eight days 200+ 14;10-12 days 155-9"4. Evolution is differentin those P which have a non-effectivetreatment" two to three days 1685 708;six-eight days 1171+463 (p=0"05) and10-12 days 933+213 (p=0.01). Finally,some cases with a good initial responsehave been followed by an increased count.The use of the automatic counter is asignificant aid for diagnosis and follow-up of peritonitis on CAPD.
22.2.7: Usefulness of EDTA-blood assamples or controls for haematology testsBy I. M. Penttilfi, E. Puhakainen andT. Rantanen, Department of ClinicalChemistry, Kuopio University CentralHospital, SF 70210 Kuopio 21, Finland
A Coulter Counter S-Plus analyser(Coulter Electronics Ltd, Harpenden,
102
UK) was used for one and a half years forthe measurement of parameters in wholeblood. During the first months the sur-vival times of blood cells in EDTA-bloodwere measured--all cell values werefound to be constant for up to 6 h at roomtemperature. Some difficulties with thistime have been experienced becausehealth centres are often remote. Thesecond reason for this study was a needfor quality control information. Samplesanalysed daily were stored at +4C untilthe following morning, when they werereanalysed for the most importantparameters with the Coulter CounterS-Plus analyser. Table presents theresults.
Table 1. Effect.of the storage onparameters ofEDTA-blood samples (N
950). Mean I isfor initial values andmean 2for results after 18 h storage at+4C. P-values are calculated by thepaired t-test for analyses.
WBC RBC Hgb09/L 02/
Mean 8.44 4"46 133.2Mean 2 8"66 4"47 133.2p < 0"001
Although most differences are signficant,it is evident that in routine haematologypractice the daily samples can be used forthe control ofthe Coulter Counter S-Plusanalyser, in addition to commercialpreparations. EDTA-samples can beanalysed the morning after sampletaking.
22.2.8: Haemoglobin levels of patientsattending an under-fives clinic in BeninCity, NigeriaBy O. N. Ogbeide and O. Ogbeide,Department of Chemistry, University ofBenin, Benin City, Nigeria
The importance of anaemia is wellknown, but has undoubtedly been under-estimated. The high incidence of thisdisease in infancy, especially in develop-ing countries, warrants routine lab-oratory diagnosis. This is essential forappropriate clinical management, im-proved surveillance and effective healthcontrol. In order to assess the effect ofsome childhood diseases on haemoglobinvalues, all patients attending an under-fives clinic in Benin, Nigeria were ex-amined and their Hb levels measured.
International Congress on Automation in the Clinical Laboratory: Abstracts
Such laboratory information using asingle haemoglobinometer is presentedand the usefulness of this type of routinetest is discussed.
deviate’ calculated over reference seraresults.
An efficient and easy monitoring ofthe chemical procedures and good resultreliability is achieved. Results fromdifferent automatic analysers can alsobe collected, processed, controlled andreported.
Session 22.3:Electronic Data Processing
22.3.1: Real time control procedures anddata sample random management usinga minicomputer on-line to automaticanalysersBy G. Barbaresi, M. L. Gozzo, G. E.Martorana and C. Zuppi, ClinicalChemistry Laboratory, Policlinieo A.Gemelli, Institute ofBiological Chemistry,Universitit Cattolica del Sacro Cuore,Rome, Italy
The automatic analysers in a clinicalchemistry laboratory will normally takecare of the analytical work-load. At thesame time they usually require a dupli-cation of management procedures.Modern computerized instruments arenot able to communicate directly with thecentral data-bank and do not fit intomanual or automatic clerical procedures,which depend upon traditional work-sheets. Moreover, security controls mustmatch the speed of the analytical outputand follow up instrument performancewithout delay. For this reason a localinformation system was developed usingan Olivetti P6060 minicomputer on-lineto the analysers and exchanging inform-ation with the Hospital Data ProcessingCenter (HPDC) by means of floppy discs.An automatic pattern recognition system(optical reader, OPR) performs positiveidentification of samples which arerandomly analysed. Patient numbersequence is automatically linked to ana-lytical output. Identified or ’personalized’result records thus obtained are thenmatched with request records, loaded ona floppy disc from the HDPC whichprocessed the test requests from clinicalwards and the patient personal datapresent in the bank. Thus, a completeclinical report is automatically assembledfor each patient, it can be immediatelyprinted in the laboratory and transferredto the data-bank for storing. Moreover, inorder to relieve the technical staff fromcontinuous surveillance of analyticaloutput quality, the programs are builtto provide alarms of troubles detected bythe instrument’s own software and toperform additional data validation tests.These latter control result consistencywith mathematical comparisons betweenchemical parameters (discrepancy tests)and with confidence limits suggestedby clinical experience (alert tests). Con-currently, a real-time quality-controlprocedure supplies a statistical indexsimilar to the ’standardized normal
22.3.2: Specimen identification in theEDP-supported chemical centrallaboratoryBy H. J. Gibitz and D. Hauch,Zentrallaboratorium der Landeskran-kenanstalten, A 5020 Salzburg, Austria
Mark-sense forms are used in theZentrallaboratorium requests, the formscontain the name of the patient, amachine-readable patient number withseven digits is applied by the nurse withthe help of an ADREMA card. Eachmark-sense form is preprinted with oneletter and a four-digit specimen numberwhich is machine readable. This specimenidentification is also preprinted on sixself-adhesive, differently coloured labelson the right margin ofthe form but is onlynormally readable. These labels are stuckon to the specimen tubes by the nurse. Inthe specimen receiving office the tubes aredistributed amongst laboratory depart-ments following the lead colours of thelabels. The mark-sense forms are read byan IBM 370. The printed working filescontain the specimen identifications in anascendant order (letters and numbers) aswell as the corresponding names andnumbers of the patients.
The advantages ofthis system are thatthe specimen, sample and patient identifi-cation is given, without change, fromspecimen collection to the printing of thepatient report. No additional number, aday number for example, is necessary.And an expensive and time-consumingcentral sample distribution, which is sus-ceptible to the risk of confusion, is notrequired. Every functional section of thelaboratory gets the exactly identifiedspecimen directly from the patientwithout a time delay.
Details about the production and thepreprinting of the mark-sense forms andexperiences ofthis system over three yearsare presented.
22.3.3: Are patient and biological sampleidentifications compatible?By P. Valdiguib, Laboratoire de Biochimie,C. H. U. de Rangueil, 31054 ToulouseCedex, France
Correct identification of the manysamples handled in different containerswithin a clinical laboratory is obviouslynecessary, but it is difficult. This identifi-cation must be adapted to the generalprocedures already in use in the lab-oratory and must be easily readable byautomated machines which are popular.The ’positive’ identification systemswhich are available are not very common
103
International Congress on Automation in the Clinical Laboratory: Abstracts
yet because they are installed mainly onlarge and expensive analysers and they donot fit in with the normal organization ofthe laboratory. Moreover, they use smallnumbers. Patient identification, on theother hand, usually needs multidigitnumbers which are difficult to handledirectly during the biological process inthe laboratory. The connection betweenthese two incompatible identifications isdiscussed. A scheme which will be im-plemented in a large university hospital inToulouse is reported.
22.3.4: Bar-code label reading of bloodsamples from donors simultaneously withblood-grouping in an autoanalyser, for theprevention of clerical errors in a routineblood-bank laboratoryBy R. L. McShine, A. G. J. de Stair, Z. Smitand C. Th. Smit Sibinga, Regional RedCross Blood Bank Groningen-Drenthe,Groningen, The Netherlands
The Regional Red Cross Blood Bank is alarge and comprehensive blood centre inthe northern Netherlands, which serves13 hospitals and receives 56 000 donationsa year. Automation has been organized toprevent clerical errors in identificationand reading of samples, and to increasethe standard and accuracy of routinequality control. ABO/Rhesus grouping ofdonor blood has been automated using aTechnicon BG-15 and a connected self-designed autoanalyser with a Marsh andLalezari channel for automated irregularantibody detection. The sampler has beenadapted with a moving white-light sourcereading device for bar-code identificationlabels (Codabar). Results are read anddecoded in a data-processor, stored ondiscette and printed for immediate con-trol. The daily results, on discette, arepresented to the computer department forupdating the donor file and machine-readable quarantine release using a bar-code reader-pen. This procedure has hada significant impact on clerical safety ofroutine quality control in the laboratoryand the quarantine release division ofcomponent department.
22.3.5: Desk-top computer in the clinicallaboratory interfaced to automatic multi-channel biochemistry analysersBy Yngve Berqqvist and Lars Funding,Department of Clinical Chemistry, FalunCentral Hospital, S-791 82 Falun, Sweden,Brendan Henderson, Infax Data H.B., Box1169, S-163 12 Spanga, Sweden, and PerEriksson, Scandia Metric AB, Box 1307,S-171 25 Solna, Sweden
New analytical instruments of increasingsophistication are continuously beingintroduced into the clinical laboratory.Some drawbacks with these new instru-ments are inflexible result print-outs,meagre analytical method control reportsand the absence of statistics for billingcosts. With the help of Scandia MetricAB of Solna, Sweden, two Metric-85 desk
top computers were successfully linked toa Greiner G-300 and a Technicon SMAIIin order to collect results automaticallyand to improve performance of theabove-mentioned functions. The systemhardware consists of two Metric-85computers each containing a Zilog Z80CPU, 60 k bytes RAM and two Micro-polis mini-floppy disc drives with astorage capacity of 315 k bytes per drive.The hardware system costs less than70.000 Sw. kr. (12 000 US $). The pro-grams, which are written in BASIC, weredeveloped by Brendan Henderson. Thecomputerized system has the followingfeatures: automatic data acquisition; out-put of patient results on self-adhesivelabels; quality control reports for fivecontrols, daily and between days; stat-istics on the number of analyses orderedby each ward; registration and retrieval ofdata from disc using patient I.D.; possi-bility of manual data entry and dataediting; data security--passwords inhibitunauthorized access to data. Softwareand hardware can be used by all lab-oratory staff after only a few hours’training. This approach demonstrates theuse of low-cost desk-top computers forautomation in the clinical laboratory.
22.3.6: Desk-top processor-based interfacefor chemical analysersBy J. I. Dydula, O. S. Lauritsen and T. H.Nielsen, Department ofClinical Chemistry,Frederiksberg Hospital, Copenhagen,Denmark
Data in a routine laboratory appear inmany ways, according to the type ofanalyser. In the interface area betweenanalyser and computer, small desk-topprocessors (DTPs) facilitate data hand-ling. Price and flexibility of DTPs allowthe decentralized data handling to betailored to the needs of various work-stations. Benefits from this solution are:flexible data capture (on-line/off-line; I.D.numbers/results), quality control anderror corrections under the control of theanalyst and powerful back-up in case ofcomputer breakdown. This presentationwill focus on criteria which could be usedwhen selecting a proper DTP configur-ation: input/output facilities, program-ming language, memory size, programlibrary, vendor reliability. The criteriawill be illustrated by some experienceswith the use of an Apple II.
22.3.7: On-line data acquisition from aflame photometerBy R. A. Lutz, Medical ChemistryLaboratory, Kantonsspital, CH 8401Winterthur, Switzerland
The interfacing ofa Motorola 6800 micro-processor to a IL-Flame Photometer 343is described. This microprocessor canstore results on an RAM memory modulewith battery back-up. If the mainlaboratory minicomputer (NOVA 3) isrunning and ready, the microprocessor
104
transmits the data, together with theserum number and additional inform-ation entered by the laboratory techni-cian. This communication uses thestandard RS-232C protocol. If the mini-computer is turned off" or is not ready, themicroprocessor keeps the results in itsown memory until data transfer ispossible. This system can be easilymodified for other digital laboratoryinstruments and, therefore, is uniformlyapplicable. The system has been runningfor over a year and has proved reliable.Other microprocessors interfaced toother instruments can also be branched toit, simply by adding additional RS-232Cinterface boards to the microprocessorbus. Only one line to the minicomputer isoccupied for up to eight instrumentsrequiring minor software modifications.
22.3.8: A low-cost data processing systemfor clinical laboratoriesBy Rafael Clemente, IZASA, S. L., AragOn,90, Barcelona, Spain
A low-cost data processing system forclinical laboratories is described, stressingthe following functions: information flowin the laboratory; automation of clericaltasks: filing, invoicing, etc.; time distri-bution study and improvements achievedwith the inclusion of the system in anumber of laboratories. Expansions ofthe basic equipment, up to reaching multi-user capability and ’on-line’ connection ofseveral automatic analysers, is discussed.
22.3.9: Rentability and qualityadvantages--computer and sample distri-bution in laboratoriesBy Dr Knipps, GFC--Gesellschaft fiirComputersysteme in der Medizin mbH,Holzhauser Str. 159-165, 1000 Berlin 27,FR Germany
The GFC-MELAS system has succeededin gaining a top position in the marketduring an extremely short period. Thesystem is used in the data processingdepartments of hospital laboratories.
Only those computers which corres-pond to most modern technology andflexibility as required for one or morechannel measuring instruments are used.
The software is based on modulartechnique. The high-level programmablelanguage can easily be understood evenby personnel who are unfamiliar withdata processing. Changes which arenecessary in the laboratory, i.e. methodsand measuring equipment can easily beperformed by the personnel. The PVSSample Distribution Station 911 is anautomatic system which can be connectedon-line to laboratory computer systems.The system allows nearly 700 pipettingsevery hour if one module is used, usingthree, 2000 pipettings are possible.
Segments can be easily installedduring the distribution process. Thesystem saves approximately one to twohours a day by automating the sample-handling stage.
22.3.10: On-line computers; application inquality control of resultsBy J. M. Navarro and M. J. Alsina,Ambulatorio ’Jaime I’, Vilanova la Geltru,Barcelona, Spain
Errors that arise in clinical laboratoriesare produced in three areas: adminis-tration, pre-analysis, and the analysisitself. Conventional quality control pro-grammes normally show only the ana-lytical error. Now, by using computersconnected on-line to autoanalysers, it ispossible to identify a high percentage ofall three types of error. The program isdivided into the following modules,applied in sequence: analytical qualitycontrol; outliers values set by thelaboratory according to the type ofpatient; parameter interrelations; com-patibility with diagnosis (abnormalranges); delta check.
The purpose of this program is tocarry out in real time the same operationsthat would normally be done by thebiochemist, leaving him free to deal withthose results which do not fit the require-ments listed above. A considerable savingin time and greater laboratory efficiencyis achieved.
22.3.11: Integrating a fast computerizedanalyser (PRISMA) in the organization ofa computer-assisted centralized clinicalchemical laboratoryBy C. Pietrzyk, W. Schmaderer, H.-J.Engelhardt, W. Vogt and M. Knedel,Institut fiir Klinische Chemie, KlinikumGroflhadern, Ludwig-Maximilians-Univer-sitfit Miinchen, D-8000 Miinchen 70, FRGermany
A fast multichannel analyser in a clinicallaboratory should supply precise, accuratedata, quickly and at low cost. The conceptof the analytical system determines pre-cision and accuracy of the measurement,while cost and turnaround time for testsare optimal only when the analyticalsystem is fitted to the organization of thelaboratory. Bidirectional communicationbetween analyser and laboratory com-puter system has been realized in thePRIMULAB system developed by theauthors. The system offers the followingfeatures: requesting analyses and reruns ispossible without expending time; patientdata do not have to be entered by labor-atory personnel; results can be controlledat two levels--analysis-oriented directlyon the display unit of the analyser andpatient-oriented using the laboratorycomputer system, that stores more ex-tensive information; long-term storage ofquality control data is possible in thelaboratory computer.
The laboratory organization musttake into account partial or total failure ofthe multichannel analyser, so that otheranalysers must replace it. Samples have tobe divided and distributed according tothe current need, this can be controlleddirectly by the computer.
International Congress on Automation in the Clinical Laboratory: Abstracts
22.3.12: Evaluation of a computer systemin a private laboratoryBy F. Echevarne and J. I. Hornos,Laboratorio de Anfilisis del Dr F.Echevarne, Barcelona, Spain
This study investigated the main featuresthat a computer should offer to fit into aprivate laboratory, the authors discussfeatures which are absolutely necessaryand those which are of marginal im-portance. The economic advantages ofhaving a computer in a private laboratoryare also presented.
22.3.13: Demands and limitations of theon-line connections between laboratoryanalysers and computersBy V. Seco Torrecillas and J. Aznar Lucca,Laboratorio de Anlisis Clinicos; C.S.S.S.’La Fe’, Valencia, Spain
The authors describe the problems theyhave found when connecting a computeron-line to their automatic equipment.The features a computer should have arediscussed together with the type of equip-ment that should be connected. They alsoreport on the way data is processed--from the analyser sending it to thecomputer to the definitive print-out ofresults. These points are related to thelaboratory’s work structure and costs andbenefits of computerization are given.
22.3.14: Experiences of ,a turn-key lab-oratory computer systemBy Stanley Wilson, North ManchesterGeneral Hospital, Manchester, UK
The author analyses his experience with aturn-key system in a laboratory of bio-chemistry in a university hospital over aperiod of one and a half years, since itsinstallation. The advantages and dis-advantages of an on-line connection of aTechnicon SMAC, as well as the expan-sion of the system to the haematologylaboratory, are emphasized.
22.3.15: Six years’ experience with an on-line data-base in haemostaseologyBy Th. Gergely, Institute for ClinicalChemistry and Laboratory Diagnosis oftheUniversity of Vienna, Ch. Korninger, 1stMedical Department of the UniversityHospital of Vienna, and W. Dorda,Institute for Computer Medicine at theUniversity of Vienna, Austria
The department of haemostaseology ofthe 1st Medical University Hospitalis unique in Vienna, advising otherhospitals and supervising the sensitivecontrol of inborn and acquired haemo-staseological disorders. To improveefficiency and precision ofpatient manage-ment and control, the traditional data-base was computerized and linked withthe university hospital informationsystem: WAMIS. DOKU is a subsystemof WAMIS and handles the documen-tation of medically relevant information,such as anamnesia, physical examination
and results ofdiagnostic tests. During thelast six years, extensive data on thecoagulation system of more than 8000people have been collected. These dataare accessible on-line for routine patientmanagement, such as working schedulesfor nurses, drug prescriptions, medicalreports for referring physicians as well asreminders to physicians and/or patientsabout necessary control check-ups andrequired changes in therapy. In addition,a variety of standard and sophisticatedmathematical programs are available foron-line scientific work. On-line dataprocessing has proven to be a useful toolto educate personnel--from physicians tonurses, to decrease the paper-work loadand to facilitate the use ofpatient data forscientific and teaching purposes.
Session 22.4: Cost Analysis
22.4.1: Economic and quality evaluationobtained with the a1tomation of five im-munologic parameters. Experience of sixyears and 85 000 deterninationsBy F. Malagon, F. L. Elorza, M. E.Dorado and J. M. Corralejo,Departamento Bioquimica Clinica,Hospital Universitario de Sevilla, Spain
The authors present their experiences,accumulated over six years, on the pro-gressive automation of a laboratory ofimmunology in a hospital with 1000 bedsand 1000 out-patient visits per day. Theresults obtained with the first automatedparameters (IgA, IgG, IgM, C3 and C4)are evaluated; and the technological appli-cations, automated or not, which they setup during the same period of time andwith the same personnel are reported. Theresults are calculated on the basis of85000 determinations during the sixyears ofthe study, taking into account theevolution undergone by the various para-meters (cost, time, etc.) from year to year,with the object of forseeing future needs.The results of automation have beenextraordinarily positive, the savingsduring this time being 16 M pesetas.
22.4.2: Doubled efficiency by modernanalysersBy R. Asper, J. W. S. Fliickiger and D. J.Vonderschmitt, Central Laboratory ofClinical Chemistry, University Hospital,CH 8091 Ziirich, Switzerland
In university hospital with 1300 beds,automation was brought up to date byreplacing 10 batch, or discrete, analyserswith five selective or continuous flowanalysers. With half of the staff, eightlaboratory technicians instead of 16,efficiency was doubled; the average lagbetween sample arrival and result de-livery decreased from three to one and ahalf hours. Also the emergency results,
105
International Congress on Automation in the Clinical Laboratory: Abstracts
25 of the total, are accomplished faster:90o of these results are delivered in lessthan one hour. The types of instrumentand new organization are discussed.
22.4.3: Contribution to the setting up, costanalysis and management of a clinicallaboratoryBy F. Lema and R. Alvarez, Laboratorio deBioquimica, R. S. ’Almirante Vierna’, Vi9o,Spain
The authors discuss the steps which needconsideration in the setting up of aclinical laboratory. A plan is presentedwhich covers the following points: (1)short-, medium- and long-term objectives.(2) Necessary structures, with referenceto: internal structure, personnel, instru-mentation, storage, etc. (3) Organizationand planning: plans, ways of communi-cation, work, making use ofresources, etc.(4) Study ofexpenses: permanent expenses,variable expenses, expected invoices,starting point, income. (5) Once thelaboratory is functioning, an inventorybalance ofits goods should be made. (6) Astudy should be made on the efficiency ofthe personnel. (7) Determine the economiccourse of the laboratory, by consecutiveanalysis by ratios of the balance results,permanent active, debts and financialautonomy, in comparison with theexpected economic objectives.
20.2.20: Evaluation of an autoanalyser fordetermination of glucose by immobilizedenzymesBy J. Mateo and S. Valle, Department ofClinical Chemistry, University Hospital,Sevilla, Spain
The aim of this presentation is to publishthe results obtained in an evaluation ofanautoanalyser (Model C-202) for the de-termination of glucose by immobilizedenzymes (glucose-oxidase). The protocoldescribed by Broughton et al. (Annals ofClinical Biochemistry, 11 [ 1974], 207-218)was followed for the following para-meters: precision at three concentrationsover 20 days; accuracy by means ofprimary standards; linearity at variousconcentrations from 0 to 400mg; carry-over and drift. Similarly, cost per sample,response time and correlation with otherenzymatic methods was studied.Although the results obtained show thatthe values are satisfactory, the smallsample volume (2#1) may introduceerrors necessitating some precautions.
Editor’s note: this abstract is not insequence because it arrived late.
Notes for Authors
Journal of Automatic Chemistry covers all aspects ofautomation and mechanization in analytical, clinicaland industrial environments. The Journal publishesoriginal research papers; short communications oninnovations, techniques and instrumentation, orcurrent research in progress; reports on recentcommercial developments; and meeting reports, bookreviews and information on forthcoming events. Allresearch papers are refereed.
ManuscriptsTwo copies of articles should be submitted to theEditor. All articles should be typed in double spacingwith ample margins, on one side ofthe paper only. Thefollowing items should be sent: (1) a title-pageincluding a brief and informative title, avoiding theword ’new’ and its synonyms; a full list ofauthors withtheir affiliations and full addresses; (2) an abstract ofabout 250 words--this should succinctly describe thescope of the contribution and highlight significantfindings or innovations; it should be written in a stylewhich can easily be translated into French andGerman; (3) the main text with sections andsubsections numbered; (4) appendices (if any);(5) references; (6) tables, each table on a separate sheetand accompanied by a caption; (7) illustrations(diagrams, drawings find photographs) numbered in asingle sequence from upwards and with the author’sname on the back of every illustration; captions toillustrations should be typed on a separate sheet.
ReferencesReferences should be indicated in the text by numbersfoll9wing the author’s name, i.e. Skeggs [6]. In thereference section they should be arranged thus:
to a journal
MANKA, D. P., Journal of Automatic Chemistry, 3(1981), 119.
to a book
MALMSarADar, H. V., in Topics in Automatic Chemistry,Ed. Stockwell, P. B. and Foreman, J. K. (Horwood,Chichester, 1978), p. 68.
IllustrationsLine diagrams are preferred to photographs. Originalcopies of diagrams and drawings should be supplied,and should be drawn to be suitable for reduction to thepage or column width of the Journal, i.e. to 85 mm or179mm, with special attention to lettering size.Photographs may be sent as glossy prints or asnegatives.
Proofs and offprintsThe principal or corresponding author will be sentgalley proofs for checking and will receive 50 offprintsfree ofcharge. Additional offprints may be ordered ona form which accompanies the proofs.
Manuscripts should be sent to the Editor: Dr Peter B.Stockwell, Plasma-Therm Ltd, 6 Station Road, Londons:o ,f2.
