12 REFERENCE MATERIALS Anal. Proc. Standards for the Analyst C. A. Watson Hopkin and Williams Ltd., P.O. Box 1, Romford, Essex, RIM1 1HA Analytical chemistry is essentially a comparative science and is always dependent on the availability of suitable standards. Such standards may be fundamental, e.g., balance weights, physical, e.g., photometric, or chemical, e.g., elemental. Fortunately, some procedures are stable and frequent recourse to standards is not always necessary. This is particularly true of fundamental standards, but many procedures are dependent on the availability of chemical or physical standards, some examples of which are discussed below. Standards for Elemental Analysis The first materials that were prepared commercially, specifically as a range of standards, were for elemental analysis.This exercise was undertaken during the 1950s following recom- mendations of the Microchemical Methods Group of the Society for Analytical Chemistq7.l The objective was to make available a range of substances that were suitable for checking procedures for organic elemental analysis, which was developing rapidly at that time with increasing use of the oxygen-flask combustion method, more sophisticated combustion furnaces and, eventually, automatic analysers. Criteria for such a range were straightforward and can be summarised as follows. (a) Pwity There was no point in including materials that would not have an assay of at least 99.5% by the various methods for which the standard would be employed. As microchemical techniques have improved this figure may no longer be adequate, but no requests have been received to indicate that more stringent assay limits are needed, although the range was updated following a further report from the Microchemical Methods Group. (b) Stability It was therefore essential that, as each pack is likely to be in use for some years, its composition should remain constant.(c) Range of elements The original range of about 35 items included compounds containing 8-94y0 of carbon and 1 .5-Sy0 of hydrogen. Other elements were less comprehensively covered, but for the common elements such as nitrogen, oxygen and chlorine the coverage was still extensive. The initial range of materials commercially available was twice modified, following the recommendations of the Microchemical Methods G r o ~ p , ~ , ~ but still consisted of about 35 com- pounds.Unfortunately, the decreasing amounts of standard required for each check analysis has meant that consumption of these materials has fallen markedly in recent years, leading to a lack of commercial viability for many of the items. As a result, the range of commercially available materials had fallen to 22 items by 1976 and has since dropped to 17, i.e., less than half of the original range. Elemental standards are used in small amounts, often in a “control chart” situation. Volumetric Standards During the period from 1960 to 1975 the Standards Sub-committee of the Analytical Methods Committee studied a number of substances for their potential use as volumetricJanuary, 1982 REFERENCE MATERIALS 13 standards.This work involved not only purification of many of the substances, but also the development of suitable analytical procedures, and is described in a number of reports.*-8 Certain of the substances that were studied by the Sub-committee, together with a few other compounds that could be produced in a high state of purity, and for which assay methods of the required accuracy and precision were available, were made available on a commercial basis in 1960 by Hopkin and Williams. The range known as PVS consisted of potassium hydrogen carbonate, later replaced by sodium carbonate, benzoic acid, sulphamic acid, potassium iodate and potassium dichromate. The preparation of these substances, which were required to have an assay of 100.00 0.02% and to have no impurities detectable by normal methods of analysis that could affect the assay, presented considerable challenges, both in manufacture and analysis.Success was achieved only with careful selection of starting materials, operation under dust-free conditions and constant monitoring of each stage of the process. Many of the processes, such as rapid filtration on fine sintered-glass filters, were not easily scaled up, limiting batch sizes to only 1-2 kg. This resulted in a cost that was of the order of 100 times that of an equivalent AnalaR grade for a substance with an assay tolerance typically five times better, i.e., 100.00 & 0.02y0 rather than 100.0 &- 0.1%. This high cost proved to be commer- cially unacceptable, as analysts were not prepared to pay the premium price and the range has now, unfortunately, been withdrawn.Volumetric Solutions From the commercial point of view, this is the one area where preparation of standard sub- stances can be regarded as successful. Standard volumetric solutions and buffer solutions have long been available and, except for stability problems in a few instances, present little difficulty in manufacture or analysis. Recently these ranges have been extended by the addition of solutions of metals and anions, usually at a level of 1000 p.p.m. The main problem in producing such a range, once a suitable matrix has been selected, is ensuring that the solutions are adequately stable in the type of pack in which they are to be sold. In order to carry out stability trials and to monitor normal production, assay methods had to be developed.These ranged from simple EDTA titrations, in instances where interfering elements could be shown to be absent in significant amounts by spectroscopic means, to differential spectroscopic methods, which could achieve the required precision, i.e., about &O.lyo. It is worth noting that, although in many instances these solutions can be prepared by weighing a pure substance and exact volumetric dilution, this route leads to unacceptable costs and in any event assay methods are still required for stability testing. Other developments in this area include concentrated standard solutions and the investiga- -tion of multi-element solutions. These developments have been necessitated by the tremend- ous advances in ICP - OES, which is essentially a multi-element technique applied to solutions.There are still, however, considerable problems in this area, not least of which is the choice of element mixtures and their indiyidual concentrations, which arise owing to the very different ratios of elements which are found in different areas of analysis. Incompatibility of various ion mixtures in some matrices is also a problem, as is the development of test methods of the required precision. The apparently simple solution to these difficulties is to prepare individual solutions for each user’s needs by using pure substances and volumetric dilution; however, although this approach has been tried by at least one supplier, costs are very high and com- mercial viability is extremely doubtful.P ho top hysical Standards Many laboratories rely on the output of ultraviolet - visible spectrophotometers or spectro- fluorimeters, but do not have the facilities for the regular preparation of standards to ensure accurate and reproducible instrument operation. Two possibilities exist for overcoming this problem: the use of standard solutions and the use of solid or sealed standards. The former approach is straightforward from the supplier’s point of view, but is of limited use to the user because storage conditions are often critical, many solutions are unstable to the normal labora- tory atmosphere and many of the test substances are toxic or harmful in solution. The second14 REFERENCE MATERIALS Anal. PYOC. approach presents a number of manufacturing difficulties, but it has been possible to make available a series of solid solutions of fluorescent compounds that are suitable for standardisa- tion and checking for wavelength accuracy and resolution of spectrofluorimeters.These are available in two series, the contents of which are listed together with their general. properties in Table I. Set A is intended for the laboratory employing many different methods at a wide range of wavelengths and set B is more suitable for the laboratory engaged in quantitative analysis over a more restricted range. TABLE I STANDARDS FOR SPECTROFLUORIMETRY Substance Excitationlnm Type Anthracene and naphthalene . . Ovalene .. .. .. .. Perylene . . .. .. .. 7,8-Benzoquinoline . . .. Coronene .. .. .... Rhodamine . . .. .. Triphenylene . . .. .. Tetraphenylbutadiene . . .. p-Terphenyl . . .. .. Compound 610 . . .. .. 296 350 396 320 340 485 310 360 300 400 Mixed Sharp Mixed Sharp Sharp Broad Sharp Broad Broad Broad Two series of photometric standards in sealed 10-mm quartz cuvettes have been made avail- able for checking the performance of spectrophotometers. One of these consists of five solu- tions of potassium dichromate and a blank, which are intended primarily for checking photo- metric accuracy. This may be achieved either by standardisation using a high-accuracy spectrophotometer, or by checking the peak - trough ratios. The solutions are prepared at pH 2.8 in perchloric acid, so the data of Burke et aL9 can be used to calculate the appropriate ratios, some of which are given in Table 11.The spectra of a typical set of standards is illustrated in Fig. 1. TABLE I1 CALCULATED ABSORBANCE RATIOS FOR POTASSIUM DICHROMATE SOLUTIONS (pH 2.9 IN PERCHLORIC ACID) r----h Ratio v A 850 - A,,, & A, A313 Cuvette No. 2 2.215 2.962 2.544 3 2.215 2.970 2.551 4 2.215 2.978 2.558 5 2.215 2.986 2.565 6 2.215 3.002 2.578 The second set of standards consists of one potassium dichromate solution, which can be used as above, a lithium carbonate solution to check for stray light below 225 mm, a cuvette containing benzene vapour to check for instrument resolution (both optical and electronic in the case of a scanning instrument) and a samarium perchlorate solution, which has 28 peaks at known wavelengths between 230 and 560 mm, to be used for wavelength calibration.Photometric standards prepared in cuvettes have advantages over glass filters in that the geometry is the same as that of a normal sample and it is possible to prepare standards with sharp peaks for wavelength and resolution calibration. It should be mentioned, however, that although these standards are now available commercially, their preparation is not easy andJanuary, 1982 REFERENCE MATERIALS 15 considerable effort has gone into finding suitable filling and sealing procedures that ensure long-term stability. Conclusion Reagent suppliers have been active in the field of standards and reference materials for many years, but this activity has been of limited effectiveness for two reasons. Firstly, it has proved very difficult to ascertain exactly the users’ requirements, and secondly, the preparation of such substances is time consuming and therefore expensive.The continuing supply of standards and reference substances will depend upon establishing effective communication between the potential user and the supplier, and acceptance by users that standards and reference materials are costly to prepare. 1. 2. 3. 4. 5. 6. 7. 8. 9. 200 300 4 Wavelengthhm 0 Fig. 1. Spectra of potassium dichromate solu- Experimental ratios for instrument used to tions. record spectra : Cuvette A,,, A, No. A313 A313 A313 2 2.236 2.868 2.461 3 2.216 2.853 2.412 4 2.216 2.849 2.378 5 2.217 2.922 2.483 6 2.215 2.964 2.523 References Microchemistry Group, Analyst, 1953, 78, 258. Microchemistry Group, Analyst, 1962, 87, 304. Microchemical Methods Group, Analyst, 1972, 97, 740.Analytical Methods Committee, Analyst, 1965, 90, 251. Analytical Methods Committee, Analyst, 1967, 92, 587. Analytical Methods Committee, Analyst, 1975, 100, 675. Analytical Methods Committee, Analyst, 1977, 102, 955. Analytical Methods Committee, Analyst. 1978, 103, 93. Burke, R. W., Deardorff, E. R., and Menis, O., “Liquid Absorbance Standards,” NBS Special Publication No. 378, National Bureau of Standards, Washington, D.C., 1973.January, 1982 REFERENCE MATERIALS 15 considerable effort has gone into finding suitable filling and sealing procedures that ensure long-term stability. Conclusion Reagent suppliers have been active in the field of standards and reference materials for many years, but this activity has been of limited effectiveness for two reasons.Firstly, it has proved very difficult to ascertain exactly the users’ requirements, and secondly, the preparation of such substances is time consuming and therefore expensive. The continuing supply of standards and reference substances will depend upon establishing effective communication between the potential user and the supplier, and acceptance by users that standards and reference materials are costly to prepare. 1. 2. 3. 4. 5. 6. 7. 8. 9. 200 300 4 Wavelengthhm 0 Fig. 1. Spectra of potassium dichromate solu- Experimental ratios for instrument used to tions. record spectra : Cuvette A,,, A, No. A313 A313 A313 2 2.236 2.868 2.461 3 2.216 2.853 2.412 4 2.216 2.849 2.378 5 2.217 2.922 2.483 6 2.215 2.964 2.523 References Microchemistry Group, Analyst, 1953, 78, 258.Microchemistry Group, Analyst, 1962, 87, 304. Microchemical Methods Group, Analyst, 1972, 97, 740. Analytical Methods Committee, Analyst, 1965, 90, 251. Analytical Methods Committee, Analyst, 1967, 92, 587. Analytical Methods Committee, Analyst, 1975, 100, 675. Analytical Methods Committee, Analyst, 1977, 102, 955. Analytical Methods Committee, Analyst. 1978, 103, 93. Burke, R. W., Deardorff, E. R., and Menis, O., “Liquid Absorbance Standards,” NBS Special Publication No. 378, National Bureau of Standards, Washington, D.C., 1973.January, 1982 REFERENCE MATERIALS 15 considerable effort has gone into finding suitable filling and sealing procedures that ensure long-term stability. Conclusion Reagent suppliers have been active in the field of standards and reference materials for many years, but this activity has been of limited effectiveness for two reasons.Firstly, it has proved very difficult to ascertain exactly the users’ requirements, and secondly, the preparation of such substances is time consuming and therefore expensive. The continuing supply of standards and reference substances will depend upon establishing effective communication between the potential user and the supplier, and acceptance by users that standards and reference materials are costly to prepare. 1. 2. 3. 4. 5. 6. 7. 8. 9. 200 300 4 Wavelengthhm 0 Fig. 1. Spectra of potassium dichromate solu- Experimental ratios for instrument used to tions. record spectra : Cuvette A,,, A, No.A313 A313 A313 2 2.236 2.868 2.461 3 2.216 2.853 2.412 4 2.216 2.849 2.378 5 2.217 2.922 2.483 6 2.215 2.964 2.523 References Microchemistry Group, Analyst, 1953, 78, 258. Microchemistry Group, Analyst, 1962, 87, 304. Microchemical Methods Group, Analyst, 1972, 97, 740. Analytical Methods Committee, Analyst, 1965, 90, 251. Analytical Methods Committee, Analyst, 1967, 92, 587. Analytical Methods Committee, Analyst, 1975, 100, 675. Analytical Methods Committee, Analyst, 1977, 102, 955. Analytical Methods Committee, Analyst. 1978, 103, 93. Burke, R. W., Deardorff, E. R., and Menis, O., “Liquid Absorbance Standards,” NBS Special Publication No. 378, National Bureau of Standards, Washington, D.C., 1973.18 REFERENCE MATERIALS Anal. Proc. Reference Materials in Clinical Chemistry S.S. Brown Division of Clinical Chemistry, MRC Clinical Research Centre, Harrow, Middlesex, HA 1 3 U J and M. Hjelm Department of Chemical Pathology, Institute of Child Health and Hospital for Sick Children, Great 0rm.ond Street, London, W C l N 3JH The demands of clinical medicine for reliable, quantitative chemical assays of body fluids, tissues and excreta have increased dramatically during the past few decades, in parallel with our understanding of biochemical events and disease processes at the cellular and molecular levels. This paper sets out to trace one particular development in laboratory medicine, that of reference materials in clinical chemistry, from both the semantic and pragmatic points of view. Evolution of the Concepts of Reference, Calibration and Control Materials The recognition of the need for reference materials for the control of accuracy in clinical chemistry arose from experience of the 1950s and 1960s in the interpretation of quality assur- ance surveys that had been conducted in Europe and North America.In particular, the College of American Pathologists’ Standards Committee, in 1966, reviewed the poor relative accuracy and lack of precision in measurements of serum cholesterol. It was noted that some commercial preparations of cholesterol used for calibrating assay systems were of doubtful purity, and it was expected, reasonably, that if pure crystalline cholesterol could be made available, the performance of clinical laboratories for this assay would improve.1 As a direct result, the US National Bureau of Standards (NBS) embarked upon a programme of develop- ment of “Standard Reference Materials” (SRMs) for use in clinical laboratories; cholesterol SRM was the first such product. The College of American Pathologists and the American Association of Clinical Chemists continued to debate the whole question of “standardisation.” In a seminal paper, Radin2 considered the general status of standards in clinical chemistry and the possible adoption of proposed definitions of “standard” and “reference” materials.He described the situation as “particularly deplorable,” and noted that some manufacturers were marketing serum controls or standards with claims that could be misleading. Radin drew attention to the pure sub- stance SRMs that were then available from the NBS for the clinical laboratory: these included standard buffers for the measurement of blood pH, cyanmethaemoglobin for the measurement of total haemoglobin in blood and bilirubin, cholesterol and various biochemicals including amino acids and carbohydrates.Radin recognised a special need for standard preparations for the assay of plasma total protein and also pointed out requirements for standardised units (especially in enzyme measurements) and standardised nomenclature. In a far seeing way, he also recognised the essential link between reference materials and reference methods, and offered definitions of “definitive methods” of analysis, “reference methods” of analysis, and “acceptable (or standard) methods” of analysis, which were influential in establishing the formal definitions put forward later by the International Federation of Clinical Chemistry (IFCC).3 Radin did not actually offer a definition of a “reference material,” but Hanson,* on behalf of the College of American Pathologists’ Standards Committee, suggested that a reference material or reference control was “a control serum or similar material commonly used in control pro- grammes or which may of necessity be occasionally employed as calibration materials.” Hanson noted that the chemical composition and the physical characteristics of such materials were intended to simulate the patients’ specimens to be analysed, and that these very qualities precluded their use or definition as standards.Hanson properly drew attention to the relatively incomplete characterisation and poor stability of clinical chemistry control materials, in comparison with the reference materials issued by national standards organisations such as NBS.Nevertheless, there is no reason, both in principle and in practice, as described later, why serum preparations should not fulfilJanuary, 1982 REFERENCE MATERIALS 19 the general definition of reference materials that was put forward by COX,^ i.e., “substances or artifacts which are suitable for calibration, checking or testing measuring instruments and measuring procedures generally; reference materials are credited with certain property values and hence act to transfer a scale of measurement from one place to another.’’ In a state-of-the-art review, Mears and Youngs expounded the principles of rational measure- ment embodied in SI, and the link with a material’s purity.They emphasised that the purity of the standard that is used for calibrating a chemical system should not be the factor that limits the accuracy of the results for a given assay. In other words, standard solutions should be known with greater certainty than is required in the analysis of an actual sample, so that errors in their concentrations have a negligible effect on the total analytical error. Mears and Young listed the NBS SRMs that were then available and that were of potential use in the clinical laboratory. Most of those uses would not nowadays be regarded as of importance; thus few clinical laboratories require benzoic acid to calibrate melting-point apparatus, nor sodium oxalate for standardising the analysis of serum calcium by the Clark - Collip method.Nevertheless, Mears and Young’s paper is an important landmark in the evolution of the modern concept of a reference material. A Discussion Meeting on Reference Materials and Methods in Clinical Chemistry was held in conjunction with the First European Congress of Clinical Chemistry in 1974. The Report of that meeting3 drew attention to the urgent need for continuing and strengthening international collaboration in these areas if further advances were to be made in assuring the accuracy, as well as the precision, of clinical laboratory performance. It is gratifying to note that sessions on “Reference Technology” have been included in several other major international congresses7 and that clinical chemists have participated in the deliberations of the International Standard- ization Organization (ISO) Committee on Reference Materials (ISO-REMCO) which was set up in 1975 : ISO-REMCO recognises that a wide variety of international organisations have strong interests in reference materials.In general, however, the metrologist and technologist users of reference materials have not been aware of the ramifications of biomedical standardisation, nor of the efforts of the World Health Organization (WHO) and the international societies such as IFCC and ICSH (International Committee for Standardization in Haematology), which have been directed towards identifying and solving the manifold problems of establishing a basis of accuracy for clinical laboratory assays.Thus IFCC established an Office for Reference Materials and Methods in 1974, the objectives of which are as follows: (1) to gather and ex- change information on reference materials and reference methods in clinical chemistry, here called reference technology, at the national and international levels ; (2) to offer technical assistance and co-ordination of studies on reference technology ; (3) to promote international consensus on reference technology and written standards; and (4) to evaluate and promote international reference materials, and assist in their characterisation and distribution. A centralised information system on national and international activities concerning refer- ence methodology and the production of reference materials and written standards is currently being set up, and the establishment of IFCC Working Groups on Reference Methodology is in progress.Availability of Biomedical Reference Materials The summary of NBS Standard Reference Materials suitable for clinical measurements given by Meinkes included an extended list of organic SRMs-cholesterol, urea, uric acid, creatinine, D-glucose and bilirubin-together with the inorganic SRMs, calcium carbonate, potassium chloride and pH buffers; also neutral glass filters and orchard leaves (for the standardisation of trace metal analyses). The latest NBS catalogue of SRMs for the clinical laboratory, as out- lined by Ra~berry,~ includes cortisol, VMA, Tris and lead nitrate, together with reference material absorbance standards, a spectrophotometer cell and clinical thermometers.The availability of a human serum SRM (characterised by isotope dilution - mass spectrometrylO for several electrolytes and for cholesterol) was announced in November 1980 and a reference serum for antiepileptic drug assay is in course of preparation. All of the reference materials so far mentioned are linked to assays which are commonly performed in clinical chemistry laboratories. Certain laboratories, however, have specialised interests which range across haematology, coagulation, immunology, pharmacology, nuclear medicine and medical physics. For the purpose of standardisation in these laboratories, it is20 REFERENCE MATERIALS Anal. Proc. important to recognise the contributions which have been made by WHO and by the Inter- national Atomic Energy Agency (IAEA) .So far as biological reference materials are concerned, the network of WHO Collaborating Laboratories has for many years maintained a large catalogue,ll the chief purpose of which is to “provide a means of ensuring uniformity throughout the world in the designation of potency of preparations administered to man that are used in the prophylaxis, therapy or diagnosis of human disease.” The substances in question cover a wide range of pharmaceuticals, vaccines and diagnostic reagents, with special emphasis on products which cannot be characterised adequately by chemical or physical methods. The clinical laboratory’s requirements for these products remains strong, in that intrinsically complex substances such as proteins and hormones need to be analysed in many different centres worldwide.Radioimmunoassay is commonly used for these kinds of analyses and proper calibration of the necessary counting equipment is essential. The IAEA’s Analytical Quality Control Services issue annually a listing of Certified Reference Materials,12 which is an extension of previous activities of IAEA, the International Directories of Isotopes and of Certified Radioactive Materials. Experience in operating the Analytical Quality Control Services of IAEA has shown that the most important cause of error in the measurement of radioactivity is incorrect calibration of the measuring device. The most straightforward way to eliminate this systematic error is to calibrate the measuring apparatus with a source of known activity.The Directories facilitate the selection of the proper calibration source or solution with a certified radioactive content which is best adapted to the particular calibration problem. Terminology The term “reference material” has been promulgated in clinical chemistry for many years, with a tacit adoption of the usage of COX,^ referred to above. Cox’s definition, however, and those of ISO-REMCO* present real semantic difficulties because none of them clearly distin- guishes between the operations of accurately calibrating a measuring instrument or system and controlling the over-all accuracy of the whole analytical process, including the sample preparation steps. Hence the Reference Method for the determination of total calcium in serum13 is calibrated by means of SRM calcium carbonate but the over-all accuracy of the whole measurement process is monitored by incorporating into it aqueous and matrix reference materials.Dybkaer14 intuitively recognised the distinction when he described the term refer- ence material as “insufficient and ambigious.” As an alternative, he advocated “comparison material,” as being a “comprehensive term supplanting the misused word ‘standard.’ A given comparison material is named by at least three words describing : (1) composition type, e.g., high-purity, simple, complex ; (2) purpose, e.g., calibration, control, comparison; and (3) state, e.g., solid, liquid, gas and solution, suspension or material. Examples are pure calibration gas (with confidence interval for, e.g., the substance concentration of a specified component) and simple control liquid suspension (with confidence interval for, e.g., the substance concentration of a specified component) and simple control liquid suspension (with confidence interval for, e.g., the number concentration of erythrocytes).One of the following terms may precede these elements of the name : International, National, Local.’’ For these reasons, the IUPAC Analytical Division’s Commission on Analytical Nomenclature in Clinical Chemistry and the Clinical Chemistry Division’s Commission on Automation is considering discouraging the use of the term reference material, so as to make clear the distinc- tion between materials used for calibration on the one hand and the control of accuracy or precision on the other.A discussion paper on this topic, concerned with various aspects of nomenclature for automated and mechanised analysis, will soon be published. * Reference material (RM) : a material or substance one or more properties of which are sufficiently well established to be used for the calibration of an apparatus or for the verification of a measurement method. Generally, any reasonably small part of an RM sample should exhibit the property value(s) established for the RM as a whole, within the stated uncertainty limits. NOTE: Such material is intended to transfer the value of a measured quantity (physical, chemical, tech- nological) between one place and another. It may be in the form of a pure or mixed gas, liquid or solid, or even a simple manufactured object.One or more of the properties of a given batch of reference materials and their adequate stability will have been established before the batch is issued for use. Certified Reference Material (CRM): an RM accompanied by, or traceable to, a certificate stating the property value(s) concerned, issued by an organisation that is generally accepted as technically competent.January, 1982 REFERENCE MATERIALS Production of Calibration and Control Materials in Clinical Chemistry 21 The general way of producing such a material is little different from that required for com- parison materials for other branches of metrology and includes careful consideration of (a) the user’s particular specification of the material; (b) the pre-analytical preparation steps ; (c) the analytical method and statistical procedure used to quantify the component(s) to be analysed; and (d) the system to be established for storage and distribution.Hjelm15 has reviewed the general implications of these four aspects, but the following points deserve emphasis here. For calibration materials, the analyte should be highly purified and as well defined as the state of the art allows. Such materials are used to establish the relationship between the instru- mental response and the concentration of the analyte.16 For control materials, the analyte commonly needs to be incorporated into a biological matrix such as serum, urine or a tissue fraction. The composition and the properties of the matrix have to match as closely as possible the biological specimens to be analysed, if a matrix control material is to fulfil its functions.Such materials serve to verify the performance of the entire analytical procedure.l6 One component of the over-all model for the production of matrix control materials, the pre-analytical handling of the material, is of special importance. This is because improper handling of the material during this phase of its production may make it unsuitable as a bio- medical control material, even though procedures for subsequent characterisation, storage and distribution may be quite satisfactory. Thus, for a biomedical control material based on blood plasma, relevant factors include: (a) proper choice of human or animal blood; (b) proper conditions for specimen collection, including sex and age of the donor, fasting or non-fasting state at time of collection, type of container for the product; (c) conditions for further process- ing, e.g., sterility, preservatives, temperature ; and ( d ) handling before analysis. The reason why it is essential to take such factors into account is that biomedical matrix materials such as plasma are effectively complex interactive biological systems which at present cannot be adequately described in physical and chemical terms.Hence detailed specifications for the pre-analytical handling of the matrix must ensure that any crucial characteristics of the final material are maintained and that these particular characteristics are consistent for consecutive batches of the same type of material.A matrix, as will be proposed by the IUPAC bodies mentioned above, can be defined as “the assemblage of the components other than the analyte in a specimen or a sample.” Despite the simplicity of the definition, the complexity of blood plasma can be illustrated by the fact that there are more than 1000 compounds that can be characterised in such a matrix, including inorganic, low and high relative molecular mass organic compounds, enzymes, hormones and vitamins. Many of the compounds might react and interact if there were improper handling of the matrix due to redox processes, degradation, ligand formation, polymerisation or other structural changes, etc. Such changes might render a matrix control material inappropriate for a defined use, even if there were an accurately assayed content of the certified analyte.This would be the case, for example, if systematic error were introduced into a routine method because the matrices of the control material and the biological specimens differed. The kinds of difficulties associated with matrix materials which have been prepared in different ways can be illustrated from a recent investigation,17 in which either spiked lyophil- ised or liquid human sera were used as calibration materials for the assay of the hormone cortisol in plasma. All specimens were assayed using a radioimmunological method and with either of the two types of calibration materials in order to establish the calibration graph. In addition, all specimens and calibration materials were assayed by isotope dilution - mass spectrometry (ID - MS).In this way the difference between the value obtained by the radio- immunological method and the accurate value obtained by ID - MS could be calculated for each of the specimens assayed. It was shown that the spiked lyophilised human serum used as calibration material was unsuitable, despite the fact that the assigned value of the analyte, cortisol, was correct as determined by ID - MS. Results obtained by using liquid human sera (which had been stored deep frozen) as calibration materizl for the radioimmunological pro- cedure were acceptable. The findings of this investigation indicate that considerable syste- matic research is called for in order to develop (a) suitable procedures to handle biomedical matrix materials for calibration and control purposes, either in lyophilised or liquid form, and (b) adequate methods for characterising the critical matrix properties of such materials.22 DERIVATIVE SPECTROSCOPY Anal. Proc. References Brown, S. S., Ann. Clin. Biochem., 1973, 10, 146. Radin, N., Clin. Chem., 1967, 13, 55. Rinsler, M. G., and Mitchell, F. L., 2. Klin. Chem. Klin. Biochem., 1974, 12, 558. Hanson, D. J., Am. J . Clin. Pathol., 1970, 54, 451. Cox, J . D., Chem. Ind. (London), 1975, 420. Mears, T. W., and Young, D. S., Am. J . Clin. Pathol., 1968, 50, 411. Brown, S. S.. in Carroll, D. M., Burns, D. T., Brown, D. A., and MacDaeid, D. A., Editors, “Euro- Meinke, W. W., Anal. Chem., 1971, 43, 28A. Rasberry, S., Anal. Proc., 1982, 9, 5 . Lawson, A. M., Lim, C. K., Richmond, W., Samson, D. M., Setchell, K. D. R., and Thomas, A. C. S., in Lawson, A. M., Lim, C. K., and Richmond, W., Editors, “Current Developments in the Clinical Applications of HPLC, GC and MS,” Academic Press, London, 1980, pp. 135-153. World Health Organization, “Biological Substances : Lists of International Biological Standards, International Biological Reference Preparations, and International Biological Reference Reagents,” WHO, Geneva, 1979. International Atomic Energy Agency, “Certified Reference Materials, Reference Materials and Samples for Intercomparisons (LAB/243 Circ.) ,” IAEA, Vienna, 1980. Brown, S. S., Healy, M. J. R., and Kearns, M., J . Clin. Chem. Clin. Biochem., in the press. Dybkaer, R., in “Proceedings : International Conference on Standardization of Diagnostic Materials, June 5-8, 1973,” US Department of Health, Education, and Welfare, Atlanta, 1974, pp. 11-26. Hjelm, M., in Voller, A., Editor, “Immunoassays for the 198Os,” MTP Press, Lancaster, 1981, p. 185. Stamm, D., J . Clin. Chem. Clin. Biochem., 1979, 17, 283. Lantto, 0.. Bjorkhem, I., Blomstrand, R. F., and Kallner, A., Clin. Chem., 1980, 26, 1899. analysis I11 : Reviews on Analytical Chemistry,’’ Applied Science, London, 1979, pp. 75-92. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.