161
Clinica Chimica Acta, 70 (1976) 161-169 0 Elsevier Scientific Publishing Company,
Amsterdam
- Printed in The Netherlands
CCA 7922
QUALITY CONTROL IN HAEMOGLOBINOMETRY WITH SPECIAL REFERENCE TO THE STABILITY OF HAEMIGLOBINCYANIDE REFERENCE SOLUTIONS
O.W. VAN ASSENDELFT and W.G. ZIJLSTRA 4*
a, A. BUURSMA
a, A.H. HOLTZ b, E.J. VAN KAMPEN =
a Laboratory of CheFical Physiology, University of Groningen, bNational Institute of Public Health, Bilthoven and CClinical Chemical Laboratory, Diakonessenhuis, Groningen (The Netherlands) (Received
February 26,1976)
Summary
In haemoglobinometry grave errors are still being made even though an internationally accepted standardized method is available for the determination of the haemoglobin content of blood. Up to the present only haemiglobincyanide reference solutions have been available on a wide scale to check the measuring stage of the standardized haemiglobincyanide method. These reference solutions are shown to remain stable, under proper storage conditions, for more than 10 years. Concentrated haemoglobin solutions have become available recently, offering the possibility to control the dilution and conversion steps of the haemiglobincyanide method. Such a solution is shown to remain stable, under proper storage conditions, for at least a year. Using both haemiglobincyanide reference solutions and concentrated haemoglobin solutions, as well as having the possibility of checking the cyanide content of the reagent used, an acceptable intra-laboratory control program may now be set up.
Introduction
Recommendations for haemoglobinometry [l] , with the haemiglobincyanide (HiCN) method as the method of choice, were adopted at the world level during the 11th International Congress of Haematology in 1966. Although a number of papers detailing errors encountered when employing the HiCN method have been published, it is only recently that quality control in the haema* To whom correspondence should be addressed.
162
tological laboratory has gained more widespread attention [2,3] . As yet, however, quality control reference materials are available only for the measurement of haemoglobin. In 1962 Zijlstra and Van Kampen [4] described the preparation of a HiCN reference solution. The haemoglobin content (cub) of this reference solution is calculated from spectrophotometric measurements and is based on a value of 11.0 for the millimolar extinction coefficient of HiCN at h = 540 nm (E?& = 11.0) [ 5,6] . Preliminary studies showed properly ampoulated HiCN solutions, when stored at 4”C, to remain stable for at least three years [7,8]. In 1964 the Dutch National Institute of Public Health (Rijksinstituut voor de Volksgezondheid; R.I.V.) started preparing HiCN reference solutions [9] on behalf of the International Committee for Standardization in Haematology (I.C.S.H.), to be issued as international HiCN reference solutions [1]. Such a solution was accepted in 1968 by the World Health Organization (W.H.O.) as International HiCN Reference Preparation [lo] . The R.I.V. has been designated holder of this preparation by W.H.O. and issues international HiCN reference solutions which are available on request [ 111. A new batch of such reference solutions is made yearly; content and purity [l] are determined and checked by I.C.S.H.-appointed reference laboratories in Germany, Great Britain, Italy, Japan, The Netherlands and the U.S.A. The solutions are issued some nine
TABLE1 STABILITYMEASUREMENTSOFINTERNATIONALHiCNREFERENCESOLUTIONS Time
years
0 0 1 1 2 2 3 3 4 4 5 5 6 7 6 9 10 11 12
40400 month
0
3 6 9 0 3 6 9 0 3 6 9 0 6 0 6 0 6 0 0 0 0 0 O 0
A 540
0.391
60400 p Aso
A-J
1.61
0.386 0.387 0.387 0.381 0.388
0.389
1.61
0.391
1.61
0.391
1.61
0.390
1.61
0.391
1.62
1.61
1.61 1.61 1.61
1.61
0.393
1.61
0.386
0.389
1.61 1.61 1.61 1.61 1.61
0.381
0.389
0.394 0.390 0.389 0.390
fi
A 540
A504
0.388 0.381
0.390
70400
00400 e As04
0.386
1.61
0.381
1.61
e Aso
$383 0.381 0.386 0.383
1.60 1.60 1.61s 1.61
0.386 0.385
1.61s 1.61
0.386
1.61
1.61
0.386
1.61
0.405 0.406
1.61 1.60s
0.388
1.62
0.384
1.61
0.404
1.62
0.402 0.405 0.404
1.61 1.60 1.61
0.382 0.383 0.383
1.60 1.60 1.61
0.406
1.62
0.404
1.62
0.405 0.403 0.401 0.405 0.405 0.406 0.404
1.595 1.61 1.62 1.62 1.62 1.62 1.62
0.404
1.62 1.61 1.60 1.60 1.60 1.59
A54
163
months after preparation and, at present, being available simultaneously. Stability
of HiCN reference
are valid for 2 years, thus two batches
solutions
All batches of international HiCN reference solutions issued have been stored at 4°C in one of our laboratories and have been checked regularly. Measurements have been performed with an Optica CF, grating spectrophotometer at X = 750, 540 and 504 nm (total bandwidth 0.5, 0.13 and 0.13 nm, respectively) in a layer thickness of 1.000 cm against water as blank. The instrument is calibrated regularly as to wavelength using the mercury emission spectrum and as to optical density using a certified (National Bureau of Standards) Carbon Yellow filter [ 121. The results obtained for HiCN solutions followed for more than 3 years after the date of preparation are shown in Table I. Because no values greater than 0.002 have been encountered for A&$. [l] , these measurements have not been included. The results (Table I) show that, when stored in the dark at 4”C, properly prepared and ampoulated HiCN reference solutions seem to be stable for at least 10 years. It would thus be reasonable to extend the period of validity of international HiCN reference solutions over the present period of two years.
00400
90400 A540
0.389
0.389 0.388
10400 A540
AS40 Ii=
A540
1.61 1.61 1.62
0.395 0.395
0.396
1.61 1.61 1.61
1.60
0.400
1.62
A5401
A504
20400 A540
A540
1.61 1.60 1.615
0.414 0.414 0.415 0.413
1.61 1.605 1.61 1.61
A540
A540 A504
A540
0.408 0.408 0.408
1.61 1.61 1.61
0.403 0.402 0.403
1.61 1.61 1.62
0.408
1.62
0.402 0.401 0.403
1.61 1.61 1.60
0.406
1.61 0.401
1.61
A504
0.413 0.413 0.415
30400
P
0.388 1 0.389 0.389
1.62 1.61'
0.398 0.397
1.62 1.61
0.390
1.61
0.396 0.393
1.62 1.60
0.412
1.61
0.404
1.61
0.400
1.61
0.390
l.605 1.60 1.61 1.61
0.411 0.412 0.412
1.61 1.61 1.60
1.60
0.403
1.60
1.61 1.61
0.393 0.393 0.393
0.407
0.381 0.387
0.405
1.60
0.367
1.60
0.393
1.60
0.412
1.60
0.386 0.386
1.60 1.60
0.393
1.60
164
The HiCN method and an interlaboratory trial
In the concept of the standardized HiCN method [1,13] for the determinaof cnb in human blood, it was thought that the availability of a HiCN reference solution to provide a calibration point would suffice to optimize measuring results. This, in fact, did not prove to be the case. A restricted international interlaboratory trial was organized by the R.I.V. in 1973 at the request of the I.C.S.H. Expert Panel on Haemoglobinometry. The invitation to participate was sent to some 95 laboratories in Africa and Europe, primarily recipients of the international HiCN reference solution, but also some other, presumably interested, persons most of whom had some function within the structure of I.C.S.H. In the end, 54 HiCN recipients participated (46 in Europe and 8 in Africa), as well as 13 I.C.S.H. functionaries. The materials to be tested consisted of two fresh blood samples, two glycerolcontaining haemolysates and a HiCN solution. *The blood contained 1.4 g/l EDTA (disodium salt) as anticoagulant. The haemolysates had been prepared according to a method used at the Center for Disease Control, Atlanta, Ga., U.S.A. The HiCN solution was identical to the material used as International HiCN Reference Preparation [lo]. Of the blood samples and the haemolysates, cnb had to be determined, of the HiCN solution AZ&. In general the samples reached their destination within 3-4 days after despatch. This interval was short enough to prevent any detrimental changes in the whole blood samples. The “true” values were derived from the results obtained by the members of the I.C.S.H. Expert Panel on Haemoglobinometry. Their method was, of course, according to the I.C.S.H. recommendations [l] . The values are given in Table II. The results obtained by the participating laboratories are shown as Youden plots [14] in Figs. 1 and 2 for samples 1, 2 and 3, 4, respectively. In both figures the results have first been subdivide into two categories: me~~ments in which the dilution has been made manually using pipettes (panel A in Figs. 1 and 2) and where dilutors have been used (panel B in Figs. 1 and 2). It is evident from the elongation of the cluster of points along the diagonal that many systematic errors have been made. Even after excluding some evident outliers, errors range up to k1.5 g/dl. The results of measuring the HiCN solution are shown in Fig. 3, after convertion
TABLE cm
II
AND A@N
VALUES
OF SAMPLES
Sample
USED IN INTERLABORATORY cub &/dl)
No.
Material
1 2 3 4 5
Whole blood Whole blood Haemolysate Haemoiysate HiCN solution * * Measured in a 1.00 cm square cuvette.
A540 HiCN
17.2 13.8 16.6 12.5 0.406
TRIAL
165
8
. l
. . l
;:;
_:i
.
,
.
.
,
,
,
f
>’ lb ’ l+st.lotL~o a g/l%’ somplrl
7l-dl&~o
17.
19.0
g/dKl somphl
Fig. 1. Youden plots of samples 1 and 2. whole blood. Reference laboratories’ value indicated by arrow. Solid lines denote mean value of all results. A. Dilution made with pipettes. B. Dilution made with a dilutor.
sion into the corresponding cnb of a hypothetical blood sample. For comparison the histograms of the results obtained on samples l-4 have been included. It is immediately evident that the values thus obtained have a much smaller spread than those of the blood samples and the haemolysates. When one looks into the different steps of the HiCN method prone to systematic errors, the following points should be considered. 1, Sampling; 2, diluting; 3, conversion to HiCN; 4, photometry. 1. Errors due to insufficient mixing prior to sampling (inhomogeneity of the sample) may be ruled out because the range of results is practically equal for the whole blood and the haemolysate samples.
Fig. 2. Youden plots of samples 3 and 4. haemolysate. Reference laboratories’ value indicated by arrow. Solid linesdenote meanvalue of all results. A. Dilution made with pipettek. B. Dilution made with dilutor.
sample 1 5 /,
n
Arn$-Ln nn
1 g/l00 ml Fig. 3. Histogram of results of the determination of cub by 67 laboratories in an international trial. Samples 1 and 2, whole blood; samples 3 and 4. haemolyzed and partlypurified blood; sample 6. a HiCN reference Solution. Arrows indicate value determined by reference laboratories.
2. The use of non- or falsely calibrated pipettes as well as the use of incorrectly adjusted dilutors may cause grave systematic errors. In this trial, errors in diluting can account for about one half of the total error made (Fig. 3). 3. Loss of CN- from a (faultily prepared or aged) reagent has been found to yield consistently low results because part of the haemoglibin is not converted into HiCN. Because of the spread of results around the mean together with the absence of many obviously too low values in this trial, it was considered improbable that appreciable errors were caused by lack of CN-. 4. Errors in photometry may be due to reading errors and/or to c~ibratiou errors. Clearly this has not been the major source of error in this trial as is evidenced by the smaller spread of results obtained on measuring the HiCN solution included in the trial. For this sample a coefficient of variation of 1.7% has been calculated. Because coefficients of variation from 2.9% up to 3.3% have been calculated for the whole blood samples and the haemolysates, the photometry error has been estimated to account for about one half of the total error, diluting errors for the other half.
167
It is thus demonstrated by these results that, in addition to a HiCN reference solution to calibrate the instrument, a whole blood-like preparation is necessary to control the HiCN method for the determination of haemoglobin in daily practice. Concentrated
haemolyzed samples for ~t~abo~to~
control
As early as 1968 Vanzetti and Franzini [ 151 stressed the need to control both the measuring and the dilution steps in routine haemoglobinometry. To this end they prepared concentrated solutions of HiCN and haemiglobinazide (HiN3), which they found to exhibit reasonably good stability. However, especially when kept at room temperature, these solutions often proved contaminated. The preparation of sterile solutions by means of c~bontet~c~o~de haemolysis and membrane ~ltmtion resulted in ~provement of keeping qualities both at 4°C and at room temperature. A commercially available haemolysate (Merz und Dade A.G., Bern, Switzerland), prepared by means of toluene haemolysis of human blood, sterility being ensured by membrane filtration, has been checked. The samples are available in aliquots of 1 ml to prevent too frequent sampling from one and the same bottle. One lot, kept at 4”C, has been checked over a one year period, a new bottle being opened for each measu~ment series. D~utio~ 1 : 200 (0.5 ml to 100 ml) and 1 : 251 (0.02 ml in 5.0 ml) have both been used; measuring has been performed on an Optica CF4 grating spectrophotometer at h = 540 nm. Purity of the resulting HiCN solutions has been checked by also measuring at h = 750 and 504 nm, as well as by recording an absorption spectrum (450 < X < 750 nm). All the resulting HiCN solutions were found to fall within the I.C.S.H. purity requirements for HiCN reference solutions [l].The results are given in Table III. The haemoly~~ checked has the advantage over Vanzetti’s material that not only the dilution step can be controlled, but that errors in the conversion of haemoglobin to HiCN will show up. When such haemolysates are used for the day to day internal quality control of routine cub determinations, the samples should be stored at 4°C and one bottle should not be used for periods exceeding one week.
TABLE III STABILITY OF PARTLY PURIFIED HAEMOGLOEIN MEASUREMENTS AFTER CONVERSION TO HiCN
Time
A750
A540
0.001 0.0015 0.001 0.0015 0.001 0.001 0.0015
0.3015 0.302 0.300 0.300 0.375 0.375 0.297
(months)
0 1 2 3 4 6 12
1.59 1.58 1.68 1.59 1.59 1.59 1.58
SOLUTION
Dilution factor
QIb WJl)
261 251 261 200 200 200 251
11.1 11.1 10.95 11.1 11.0 11.0 10.9
168 TABLE IV RESULTS METHOD
OF CN CONTENT
Reagent NO.
1 2 3 4 5 6 7 a 9 10
DETERMINATION
OF REAGENT
SOLUTIONS
USED IN THE HiCN
E (mV) Reference *
Reagent
-230 -217 -242 -224 -213 -213 -227 -230 -230 -225
-217 -202 -243 -216 -211 -212 -212 -222 -214 -215
47.2 46.6 50.1 48.2 49.6 49.8 46.7 43.3 46.5 47.6
* Freshly prepared reagent containing 50 mg KCN per litre.
Intralaboratory
quality
control
of haemoglobinometry
With the aid of HiCN reference solutions to calibrate the measuring instrument and concentrated haemoglobin solutions to check the HiCN method as a whole, it should be possible for every haematological laboratory to insure that valid results are obtained in haemoglobinometry. It should be stressed, however, that the use only of concentrated haemoglobin solutions does not suffice for an intralaboratory control program because such solutions do not offer the possibility of differentiating between errors in the actual measurement and errors in the dilution and conversion steps of the HiCN method; HiCN reference solutions remain necessary to exclude errors in the measuring stage. For those laboratories which might also wish to check separately on the reagent used, a method is available to screen the CN- content. In this method [ 161 use is made of a solid-state, ion-selective electrode with a sensing membrane of AgCN. The results of checking commercially available (Merz und Dade A.G., Bern, Switzerland) concentrated stock solutions of reagent, after appropriate dilution, are given in Table IV. In the application of the ion-selective electrode, the reagent to be tested was always compared to a freshly made up reagent prepared in the laboratory by weighing-in analytic grade chemicals. Full intralaboratory control of haemoglobinometry should thus incorporate daily use of concentrated haemoglobin solutions, frequent use of HiCN reference solutions and the possibility of checking reagent solutions as to CNcontent. References 1 International Committee for Standardization in Haematology (1967) Rr. J. Haematol. 13 SUPPI., 71 2 Lewis. S.M. and Caster. J.F. (eds.) Quality Control in Haematology, (1975) Academic Press, London, New York 3 Van Ksmpen. E.J. and Van Assendelft. O.W. (1975) in Quality Control in Clinical Chemistry (Anido, G., Van Kampen, E.J.. Rosalki. S.B. and Rubin. M.. eds.), P. 325 De Gruyter. Berlin and New York; Van Amend&t, O.W. and Van Kampen. E.J. ibid. p. 335
169 4 Ziilstra, W.G. and Van Kampen. E.J. (1962) Ciin. Chim. Acta 7.96 6 Zijlstra. W.G. and Van Kampen. E.J. (1960) Ciin. Chim. Acta 6,719 6 ZijIstra, W.G., Van Kampen. E.J. and Van Assendelft. O.W. (1969) Proc. KoninkI. Ned. Akad. Wetenschap. Ser. C 72,231 7 Van Assendelft, O.W.. ZijIstra, W.G.. Van Kampen, E.J. and Holtz. AH. (1966) CIin. Cbim. Acta 13. 521 8 Van Assendelft. O.W.. Holtr. AH.. Van Kampen. E.J. and Ziilstra. W.G. (1967) Chn. Chim. Acta 18, 78 9 Holtz. A.H. (1965) Bibliotheca Haematol. 21,75 10 Wodd Health Organization (1968) W.H.O. Technical Report Series 384.85 11 International Committee for Standardization in Haematology (1972) Br. J. Haematol. 23.123 12 National Bureau of Standards, Letter circular LC 929 from the U.S. Dept. of Commerce, National Bureau of Standards (1948) 13 Van Kampen, E.J. and ZiiIstra. W.G. (1961) Clin. Chim. Acta 6, 538 14 Youden, W.J. (1969) Statistical Techniques for Collaborative Tests, The Association of Official Analytical Chemists, Washington, D.C. 1B Vanzetti, G. and Franzini. C. (1969) Ciin. Chim. Acta 24,417 16 Zijlstra, W.G.. Van Assendelft, O.W.. Buursma. A. and Van Kampen, E.J. (1972) in Modern Concepts in Hematology (Izak, G. and Lewis, S.M., eds.) P. 54. Academic Press. New York and London
Clinica Chimica Acta, 70 (1976) 161-169 0 Elsevier Scientific Publishing Company,
Amsterdam
- Printed in The Netherlands
CCA 7922
QUALITY CONTROL IN HAEMOGLOBINOMETRY WITH SPECIAL REFERENCE TO THE STABILITY OF HAEMIGLOBINCYANIDE REFERENCE SOLUTIONS
O.W. VAN ASSENDELFT and W.G. ZIJLSTRA 4*
a, A. BUURSMA
a, A.H. HOLTZ b, E.J. VAN KAMPEN =
a Laboratory of CheFical Physiology, University of Groningen, bNational Institute of Public Health, Bilthoven and CClinical Chemical Laboratory, Diakonessenhuis, Groningen (The Netherlands) (Received
February 26,1976)
Summary
In haemoglobinometry grave errors are still being made even though an internationally accepted standardized method is available for the determination of the haemoglobin content of blood. Up to the present only haemiglobincyanide reference solutions have been available on a wide scale to check the measuring stage of the standardized haemiglobincyanide method. These reference solutions are shown to remain stable, under proper storage conditions, for more than 10 years. Concentrated haemoglobin solutions have become available recently, offering the possibility to control the dilution and conversion steps of the haemiglobincyanide method. Such a solution is shown to remain stable, under proper storage conditions, for at least a year. Using both haemiglobincyanide reference solutions and concentrated haemoglobin solutions, as well as having the possibility of checking the cyanide content of the reagent used, an acceptable intra-laboratory control program may now be set up.
Introduction
Recommendations for haemoglobinometry [l] , with the haemiglobincyanide (HiCN) method as the method of choice, were adopted at the world level during the 11th International Congress of Haematology in 1966. Although a number of papers detailing errors encountered when employing the HiCN method have been published, it is only recently that quality control in the haema* To whom correspondence should be addressed.
162
tological laboratory has gained more widespread attention [2,3] . As yet, however, quality control reference materials are available only for the measurement of haemoglobin. In 1962 Zijlstra and Van Kampen [4] described the preparation of a HiCN reference solution. The haemoglobin content (cub) of this reference solution is calculated from spectrophotometric measurements and is based on a value of 11.0 for the millimolar extinction coefficient of HiCN at h = 540 nm (E?& = 11.0) [ 5,6] . Preliminary studies showed properly ampoulated HiCN solutions, when stored at 4”C, to remain stable for at least three years [7,8]. In 1964 the Dutch National Institute of Public Health (Rijksinstituut voor de Volksgezondheid; R.I.V.) started preparing HiCN reference solutions [9] on behalf of the International Committee for Standardization in Haematology (I.C.S.H.), to be issued as international HiCN reference solutions [1]. Such a solution was accepted in 1968 by the World Health Organization (W.H.O.) as International HiCN Reference Preparation [lo] . The R.I.V. has been designated holder of this preparation by W.H.O. and issues international HiCN reference solutions which are available on request [ 111. A new batch of such reference solutions is made yearly; content and purity [l] are determined and checked by I.C.S.H.-appointed reference laboratories in Germany, Great Britain, Italy, Japan, The Netherlands and the U.S.A. The solutions are issued some nine
TABLE1 STABILITYMEASUREMENTSOFINTERNATIONALHiCNREFERENCESOLUTIONS Time
years
0 0 1 1 2 2 3 3 4 4 5 5 6 7 6 9 10 11 12
40400 month
0
3 6 9 0 3 6 9 0 3 6 9 0 6 0 6 0 6 0 0 0 0 0 O 0
A 540
0.391
60400 p Aso
A-J
1.61
0.386 0.387 0.387 0.381 0.388
0.389
1.61
0.391
1.61
0.391
1.61
0.390
1.61
0.391
1.62
1.61
1.61 1.61 1.61
1.61
0.393
1.61
0.386
0.389
1.61 1.61 1.61 1.61 1.61
0.381
0.389
0.394 0.390 0.389 0.390
fi
A 540
A504
0.388 0.381
0.390
70400
00400 e As04
0.386
1.61
0.381
1.61
e Aso
$383 0.381 0.386 0.383
1.60 1.60 1.61s 1.61
0.386 0.385
1.61s 1.61
0.386
1.61
1.61
0.386
1.61
0.405 0.406
1.61 1.60s
0.388
1.62
0.384
1.61
0.404
1.62
0.402 0.405 0.404
1.61 1.60 1.61
0.382 0.383 0.383
1.60 1.60 1.61
0.406
1.62
0.404
1.62
0.405 0.403 0.401 0.405 0.405 0.406 0.404
1.595 1.61 1.62 1.62 1.62 1.62 1.62
0.404
1.62 1.61 1.60 1.60 1.60 1.59
A54
163
months after preparation and, at present, being available simultaneously. Stability
of HiCN reference
are valid for 2 years, thus two batches
solutions
All batches of international HiCN reference solutions issued have been stored at 4°C in one of our laboratories and have been checked regularly. Measurements have been performed with an Optica CF, grating spectrophotometer at X = 750, 540 and 504 nm (total bandwidth 0.5, 0.13 and 0.13 nm, respectively) in a layer thickness of 1.000 cm against water as blank. The instrument is calibrated regularly as to wavelength using the mercury emission spectrum and as to optical density using a certified (National Bureau of Standards) Carbon Yellow filter [ 121. The results obtained for HiCN solutions followed for more than 3 years after the date of preparation are shown in Table I. Because no values greater than 0.002 have been encountered for A&$. [l] , these measurements have not been included. The results (Table I) show that, when stored in the dark at 4”C, properly prepared and ampoulated HiCN reference solutions seem to be stable for at least 10 years. It would thus be reasonable to extend the period of validity of international HiCN reference solutions over the present period of two years.
00400
90400 A540
0.389
0.389 0.388
10400 A540
AS40 Ii=
A540
1.61 1.61 1.62
0.395 0.395
0.396
1.61 1.61 1.61
1.60
0.400
1.62
A5401
A504
20400 A540
A540
1.61 1.60 1.615
0.414 0.414 0.415 0.413
1.61 1.605 1.61 1.61
A540
A540 A504
A540
0.408 0.408 0.408
1.61 1.61 1.61
0.403 0.402 0.403
1.61 1.61 1.62
0.408
1.62
0.402 0.401 0.403
1.61 1.61 1.60
0.406
1.61 0.401
1.61
A504
0.413 0.413 0.415
30400
P
0.388 1 0.389 0.389
1.62 1.61'
0.398 0.397
1.62 1.61
0.390
1.61
0.396 0.393
1.62 1.60
0.412
1.61
0.404
1.61
0.400
1.61
0.390
l.605 1.60 1.61 1.61
0.411 0.412 0.412
1.61 1.61 1.60
1.60
0.403
1.60
1.61 1.61
0.393 0.393 0.393
0.407
0.381 0.387
0.405
1.60
0.367
1.60
0.393
1.60
0.412
1.60
0.386 0.386
1.60 1.60
0.393
1.60
164
The HiCN method and an interlaboratory trial
In the concept of the standardized HiCN method [1,13] for the determinaof cnb in human blood, it was thought that the availability of a HiCN reference solution to provide a calibration point would suffice to optimize measuring results. This, in fact, did not prove to be the case. A restricted international interlaboratory trial was organized by the R.I.V. in 1973 at the request of the I.C.S.H. Expert Panel on Haemoglobinometry. The invitation to participate was sent to some 95 laboratories in Africa and Europe, primarily recipients of the international HiCN reference solution, but also some other, presumably interested, persons most of whom had some function within the structure of I.C.S.H. In the end, 54 HiCN recipients participated (46 in Europe and 8 in Africa), as well as 13 I.C.S.H. functionaries. The materials to be tested consisted of two fresh blood samples, two glycerolcontaining haemolysates and a HiCN solution. *The blood contained 1.4 g/l EDTA (disodium salt) as anticoagulant. The haemolysates had been prepared according to a method used at the Center for Disease Control, Atlanta, Ga., U.S.A. The HiCN solution was identical to the material used as International HiCN Reference Preparation [lo]. Of the blood samples and the haemolysates, cnb had to be determined, of the HiCN solution AZ&. In general the samples reached their destination within 3-4 days after despatch. This interval was short enough to prevent any detrimental changes in the whole blood samples. The “true” values were derived from the results obtained by the members of the I.C.S.H. Expert Panel on Haemoglobinometry. Their method was, of course, according to the I.C.S.H. recommendations [l] . The values are given in Table II. The results obtained by the participating laboratories are shown as Youden plots [14] in Figs. 1 and 2 for samples 1, 2 and 3, 4, respectively. In both figures the results have first been subdivide into two categories: me~~ments in which the dilution has been made manually using pipettes (panel A in Figs. 1 and 2) and where dilutors have been used (panel B in Figs. 1 and 2). It is evident from the elongation of the cluster of points along the diagonal that many systematic errors have been made. Even after excluding some evident outliers, errors range up to k1.5 g/dl. The results of measuring the HiCN solution are shown in Fig. 3, after convertion
TABLE cm
II
AND A@N
VALUES
OF SAMPLES
Sample
USED IN INTERLABORATORY cub &/dl)
No.
Material
1 2 3 4 5
Whole blood Whole blood Haemolysate Haemoiysate HiCN solution * * Measured in a 1.00 cm square cuvette.
A540 HiCN
17.2 13.8 16.6 12.5 0.406
TRIAL
165
8
. l
. . l
;:;
_:i
.
,
.
.
,
,
,
f
>’ lb ’ l+st.lotL~o a g/l%’ somplrl
7l-dl&~o
17.
19.0
g/dKl somphl
Fig. 1. Youden plots of samples 1 and 2. whole blood. Reference laboratories’ value indicated by arrow. Solid lines denote mean value of all results. A. Dilution made with pipettes. B. Dilution made with a dilutor.
sion into the corresponding cnb of a hypothetical blood sample. For comparison the histograms of the results obtained on samples l-4 have been included. It is immediately evident that the values thus obtained have a much smaller spread than those of the blood samples and the haemolysates. When one looks into the different steps of the HiCN method prone to systematic errors, the following points should be considered. 1, Sampling; 2, diluting; 3, conversion to HiCN; 4, photometry. 1. Errors due to insufficient mixing prior to sampling (inhomogeneity of the sample) may be ruled out because the range of results is practically equal for the whole blood and the haemolysate samples.
Fig. 2. Youden plots of samples 3 and 4. haemolysate. Reference laboratories’ value indicated by arrow. Solid linesdenote meanvalue of all results. A. Dilution made with pipettek. B. Dilution made with dilutor.
sample 1 5 /,
n
Arn$-Ln nn
1 g/l00 ml Fig. 3. Histogram of results of the determination of cub by 67 laboratories in an international trial. Samples 1 and 2, whole blood; samples 3 and 4. haemolyzed and partlypurified blood; sample 6. a HiCN reference Solution. Arrows indicate value determined by reference laboratories.
2. The use of non- or falsely calibrated pipettes as well as the use of incorrectly adjusted dilutors may cause grave systematic errors. In this trial, errors in diluting can account for about one half of the total error made (Fig. 3). 3. Loss of CN- from a (faultily prepared or aged) reagent has been found to yield consistently low results because part of the haemoglibin is not converted into HiCN. Because of the spread of results around the mean together with the absence of many obviously too low values in this trial, it was considered improbable that appreciable errors were caused by lack of CN-. 4. Errors in photometry may be due to reading errors and/or to c~ibratiou errors. Clearly this has not been the major source of error in this trial as is evidenced by the smaller spread of results obtained on measuring the HiCN solution included in the trial. For this sample a coefficient of variation of 1.7% has been calculated. Because coefficients of variation from 2.9% up to 3.3% have been calculated for the whole blood samples and the haemolysates, the photometry error has been estimated to account for about one half of the total error, diluting errors for the other half.
167
It is thus demonstrated by these results that, in addition to a HiCN reference solution to calibrate the instrument, a whole blood-like preparation is necessary to control the HiCN method for the determination of haemoglobin in daily practice. Concentrated
haemolyzed samples for ~t~abo~to~
control
As early as 1968 Vanzetti and Franzini [ 151 stressed the need to control both the measuring and the dilution steps in routine haemoglobinometry. To this end they prepared concentrated solutions of HiCN and haemiglobinazide (HiN3), which they found to exhibit reasonably good stability. However, especially when kept at room temperature, these solutions often proved contaminated. The preparation of sterile solutions by means of c~bontet~c~o~de haemolysis and membrane ~ltmtion resulted in ~provement of keeping qualities both at 4°C and at room temperature. A commercially available haemolysate (Merz und Dade A.G., Bern, Switzerland), prepared by means of toluene haemolysis of human blood, sterility being ensured by membrane filtration, has been checked. The samples are available in aliquots of 1 ml to prevent too frequent sampling from one and the same bottle. One lot, kept at 4”C, has been checked over a one year period, a new bottle being opened for each measu~ment series. D~utio~ 1 : 200 (0.5 ml to 100 ml) and 1 : 251 (0.02 ml in 5.0 ml) have both been used; measuring has been performed on an Optica CF4 grating spectrophotometer at h = 540 nm. Purity of the resulting HiCN solutions has been checked by also measuring at h = 750 and 504 nm, as well as by recording an absorption spectrum (450 < X < 750 nm). All the resulting HiCN solutions were found to fall within the I.C.S.H. purity requirements for HiCN reference solutions [l].The results are given in Table III. The haemoly~~ checked has the advantage over Vanzetti’s material that not only the dilution step can be controlled, but that errors in the conversion of haemoglobin to HiCN will show up. When such haemolysates are used for the day to day internal quality control of routine cub determinations, the samples should be stored at 4°C and one bottle should not be used for periods exceeding one week.
TABLE III STABILITY OF PARTLY PURIFIED HAEMOGLOEIN MEASUREMENTS AFTER CONVERSION TO HiCN
Time
A750
A540
0.001 0.0015 0.001 0.0015 0.001 0.001 0.0015
0.3015 0.302 0.300 0.300 0.375 0.375 0.297
(months)
0 1 2 3 4 6 12
1.59 1.58 1.68 1.59 1.59 1.59 1.58
SOLUTION
Dilution factor
QIb WJl)
261 251 261 200 200 200 251
11.1 11.1 10.95 11.1 11.0 11.0 10.9
168 TABLE IV RESULTS METHOD
OF CN CONTENT
Reagent NO.
1 2 3 4 5 6 7 a 9 10
DETERMINATION
OF REAGENT
SOLUTIONS
USED IN THE HiCN
E (mV) Reference *
Reagent
-230 -217 -242 -224 -213 -213 -227 -230 -230 -225
-217 -202 -243 -216 -211 -212 -212 -222 -214 -215
47.2 46.6 50.1 48.2 49.6 49.8 46.7 43.3 46.5 47.6
* Freshly prepared reagent containing 50 mg KCN per litre.
Intralaboratory
quality
control
of haemoglobinometry
With the aid of HiCN reference solutions to calibrate the measuring instrument and concentrated haemoglobin solutions to check the HiCN method as a whole, it should be possible for every haematological laboratory to insure that valid results are obtained in haemoglobinometry. It should be stressed, however, that the use only of concentrated haemoglobin solutions does not suffice for an intralaboratory control program because such solutions do not offer the possibility of differentiating between errors in the actual measurement and errors in the dilution and conversion steps of the HiCN method; HiCN reference solutions remain necessary to exclude errors in the measuring stage. For those laboratories which might also wish to check separately on the reagent used, a method is available to screen the CN- content. In this method [ 161 use is made of a solid-state, ion-selective electrode with a sensing membrane of AgCN. The results of checking commercially available (Merz und Dade A.G., Bern, Switzerland) concentrated stock solutions of reagent, after appropriate dilution, are given in Table IV. In the application of the ion-selective electrode, the reagent to be tested was always compared to a freshly made up reagent prepared in the laboratory by weighing-in analytic grade chemicals. Full intralaboratory control of haemoglobinometry should thus incorporate daily use of concentrated haemoglobin solutions, frequent use of HiCN reference solutions and the possibility of checking reagent solutions as to CNcontent. References 1 International Committee for Standardization in Haematology (1967) Rr. J. Haematol. 13 SUPPI., 71 2 Lewis. S.M. and Caster. J.F. (eds.) Quality Control in Haematology, (1975) Academic Press, London, New York 3 Van Ksmpen. E.J. and Van Assendelft. O.W. (1975) in Quality Control in Clinical Chemistry (Anido, G., Van Kampen, E.J.. Rosalki. S.B. and Rubin. M.. eds.), P. 325 De Gruyter. Berlin and New York; Van Amend&t, O.W. and Van Kampen. E.J. ibid. p. 335
169 4 Ziilstra, W.G. and Van Kampen. E.J. (1962) Ciin. Chim. Acta 7.96 6 Zijlstra. W.G. and Van Kampen. E.J. (1960) Ciin. Chim. Acta 6,719 6 ZijIstra, W.G., Van Kampen. E.J. and Van Assendelft. O.W. (1969) Proc. KoninkI. Ned. Akad. Wetenschap. Ser. C 72,231 7 Van Assendelft, O.W.. ZijIstra, W.G.. Van Kampen, E.J. and Holtz. AH. (1966) CIin. Cbim. Acta 13. 521 8 Van Assendelft. O.W.. Holtr. AH.. Van Kampen. E.J. and Ziilstra. W.G. (1967) Chn. Chim. Acta 18, 78 9 Holtz. A.H. (1965) Bibliotheca Haematol. 21,75 10 Wodd Health Organization (1968) W.H.O. Technical Report Series 384.85 11 International Committee for Standardization in Haematology (1972) Br. J. Haematol. 23.123 12 National Bureau of Standards, Letter circular LC 929 from the U.S. Dept. of Commerce, National Bureau of Standards (1948) 13 Van Kampen, E.J. and ZiiIstra. W.G. (1961) Clin. Chim. Acta 6, 538 14 Youden, W.J. (1969) Statistical Techniques for Collaborative Tests, The Association of Official Analytical Chemists, Washington, D.C. 1B Vanzetti, G. and Franzini. C. (1969) Ciin. Chim. Acta 24,417 16 Zijlstra, W.G.. Van Assendelft, O.W.. Buursma. A. and Van Kampen, E.J. (1972) in Modern Concepts in Hematology (Izak, G. and Lewis, S.M., eds.) P. 54. Academic Press. New York and London
