Diagnostic materials for clinical analysis*

STANDARDIZATION OF DIAGNOSTIC MATERIALS 1. Diagnostic materials for clinical analysis* ROBERT SCHAFFER 1 Standardization of diagnostic materials has lagged behind their development and application. The need to improve the accuracy of the results obtained with them has stimulated collaborative efforts to upgrade the quality ofthe reagents, control materials, and standards used, with the result that specifications for reagents and standards and authorita- tive standard materials are now available for a number of clinical tests. Achievements thus far and the efforts now in progress are reviewed. The clinical laboratory has too often been inade- quately equipped or has borne too heavy a workload to undertake to verify the quality of most of the chemicals and prepackaged test reagents it uses, and has relied on its suppliers of these materials to provide this necessary evaluation. Notwithstanding, it is the clinical laboratory that has the full responsi- bility for the test results provided to the physician, and hence it must be expected to judge critically the quality of the diagnostic materials it uses. The basis of these judgments should be criteria that are widely accepted by clinical laboratories and manufacturers alike. Where such standards of evaluation do not exist, judgments remain tenuous. Reliance on recog- nized standards of acceptability is, of course, mani- festly superior to naive trust. At times the quality of diagnostic test materials has been called in question or shown to be unsatisfactory. This paper, dealing in general with the chemical materials used in clinical analyses as reagents, controls, and standards, has as its underlying concern the designation of appropriate materials whose use would contribute to the reliable performance of diagnostic tests. REAGENT CHEMICALS Reagent solutions employed for diagnostic tests combine preset proportions of chemical compounds, each compound having an identifiable purpose, i.e., to serve as a buffer, catalyst, cofactor, reactive reagent, stabilizing agent, solvent, etc. Each sub- * Presented at the International Conference on Standard- ization of Diagnostic Materials, Atlanta, USA, 5-8 June 1973. 1 Chief, Bio-organic Standards Section, Analytical Chem- istry Division, National Bureau of Standards, Washington, D.C., USA. stance added to the reagent introduces accompany- ing impurities some of which may produce unex- pected, undesired effects that may go unrecognized or become evident only on use in the test. Provided that the chemicals used in the preparation of the reagent solution conform in specifications to those described, for example, in Reagent chemicals (Ameri- can Chemical Society, 1968) and in Analar stan- dards,2 knowledge is available of the maximum limits to which impurities may be introduced with each added chemical. Supplementary sources of specifica- tions, although primarily for drugs, are international and national pharmacopoeias. Such specifications designate only the maximum permissible content of impurities, and a reagent preparation containing less than that amount of one or more impurities is not recognized as better than others on that account. However, some commercial suppliers provide re- agent chemicals that are better characterized and/or purer, according to the measurements displayed in analytical detail on the manufacturer's label, than is required by existing specifications. Such high-quality chemicals from reputable manufacturers should be considered for use in clinical work whenever additional quality can be beneficial to the test mea- surement. Many substances are not so well characterized that specifications-accurate descriptions of high-quality commercial preparations-have as yet been elabo- rated. Notable examples are the enzymes, coen- zymes, and radioactive-isotope-labelled biochemicals (including labelled biochemicals for the nonlabelled forms of which rigorous specifications are available). In the USA, the National Academy of Sciences ' Analar Standards, Ltd, London, 1967. 3071 715 6 STANDARDIZATION (1972), which has provided specifications and criteria for about 500 biochemicals, offers only criteria for enzymes and coenzymes to help users to judge for themselves the quality of purchased commercial preparations of these substances. As a minimum, where adequate criteria or specifications for a sub- stance employed in a diagnostic test are not provided in a recognized reference, the identity and biological source of that substance; a detailed description of the method used in its preparation; and, if appropriate, its activity (and how this was estimated) must be made known to the clinical laboratory user. The purity of the water used in clinical laborato- ries must be of continuous concern. The detrimental influence that contaminants in water exert on clinical test results was clearly documented by Winstead (1967). Even purified water needs to be tested to ensure its suitability. For reagent water, specifica- tions are available (American Chemical Society, 1968; American Society for Testing and Materials, 1970). Stier & Miller (1972) have considered these specifications with reference to requirements for a variety of applications in the clinical laboratory, and offer methods for quality control. To establish speci- fications that clearly apply to the water used in the clinical laboratory, the Area Committee for Clinical Chemistry of the (US) National Committee for Clinical Laboratory Standards (NCCLS) has under- taken to formulate such specifications as a Proposed Standard. (The Committee's proposal will neces- sarily undergo review by NCCLS member organiza- tions for acceptance through consensus procedures, next as a tentative standard and then finally as an approved standard.) STANDARDS Each standard used in clinical analysis may be categorized by the degree to which it provides assurance as a basis for accurate measurement. Standard materials, as one frame of reference, are here divided into chemical preparations that are: (a) issued by recognized national and international standards organizations that designate them for use as clinical analytical standards; (b) selected by the clinical laboratory or the manufacturer from " self- judged" preparations for use as standards; and (c) mixtures composed of complex substances, such as control sera, which must meet recognized analyti- cal specifications to qualify as standards. Specifications for standards adopted by recognized standards organizations provide the second frame of reference. Standard materials and specifications for materials used as standards are, of course, intimately associated. Standards from recognized standards organizations The International Hemiglobincyanide Reference Preparation, prepared by the National Institute of Public Health (Bilthoven, Netherlands) on behalf of the World Health Organization, exemplifies a stan- dard material conforming to internationally accepted specifications (International Committee for Stan- dardization in Haematology, 1967). Samples of the preparation are provided by the above-mentioned institute. The US National Bureau of Standards (NBS) is the source of a number of other standard materials for the clinical laboratory. Impetus for its work in this field was provided by the cooperation of the National Institute of General Medical Sciences of the National Institutes of Health. In 1967, NBS began to issue a series of chemicals as Standard Reference Materials (SRMs) for use as standards in clinical analysis. The purity of these chemicals is certified after extensive testing to determine their homogene- ity as well as their composition, and NBS monitors them for stability and maintains them in continuing supply. Table 1 lists by name and number the clinical SRMs certified so far. The table includes also neu- tral-density glass filters and solutions that NBS issues as SRMs for spectrophotometric absorbance-lin- earity in the visible wavelength range. These SRMs are in use as standards for the comparison of methods, for evaluating working standards, and for calibrating instruments. A most significant use was that of SRM 915 (calcium carbonate) as the basis of standardization in the development of the reference method for serum calcium (Cali et al., 1972). Meinke (1971) has reviewed the NBS programme of SRMs for clinical chemistry. The standard materials described so far are the basic standards for the clinical test measurements to which they pertain. Preparations corresponding to these basic standards, but originating from sources other than the standards laboratories, should be checked against the recognized standard. Materials of this kind are currently available from some manufacturers of high-quality analytical products. So far, specifications have been developed for only 3 clinical standard materials: (1) The acceptable range for the spectrophotometric absorbance of crys- talline bilirubin preparations to be used for prepar- ing a serum standard as provided in the recom- 716 DIAGNOSTIC MATERIALS Table 1. Clinical standard reference materials (SRMs) issued by the National Bureau of Standards Material SRM No. cholesterol 911 urea 912 uric acid 913 creatinine 914 calcium carbonate 915 bilirubin 916 D-glucose 917 potassium chloride 918 sodium chloride 919 mannitol (for glycerol from triglycerides) 920 cortisol 921 " Tris a 922 " Tris- HCl" 923 lithium carbonate 924 "VMA b 925 glass filters for spectrophotometry 930 & 930a liquids for spectrophotometry 931 a 2-Amino-2-(hydroxymethyl)-1,3-propanediol. b a,4-dihydroxy-3-methoxybenzeneacetic acid. mendations of a committee (Clin. Chem., 1962) composed of representatives from the American Academy of Pediatrics, the College of American Pathologists, the American Association of Clinical Chemists, and the National Institutes of Health-has been recognized widely. The NBS standard material counterpart to this specification for bilirubin, SRM 916, shows a molar absorptivity that corre- sponds to the upper limit of the range recommended by that committee. The Expert Panel on Bilirubin of the International Federation of Clinical Chemists is using SRM 916 in the development of a recom- mended procedure for preparing bilirubin-protein solutions as standard solutions. (2) Criteria for certification of a pure cholesterol preparation for standardizing serum cholesterol measurements were prepared by the Standards Committee of the College of American Pathologists (1967). Such a standard cholesterol preparation was issued by NBS as SRM 911. (3) Specifications for a standard for the performance of total protein determinations were adopted as Approved Standard ACC-1 by the Na- tional Committee for Clinical Laboratory Standards (1972a). In this standard, developed by the Area Committee for Clinical Chemistry, the protein con- centration of the Standardized Protein Solution (Bovine Serum Albumin) is defined as the weight of lyophilized powder less the sum of its analysed content of ash, carbohydrate, lipid, nonprotein amino compounds, and water (Peters, 1968). Al- though a number of specified properties of the lyophilized albumin and of the solution are to be analytically evaluated for the user in meeting the required specifications, the Standardized Protein Solution is approved for use as the primary standard to which the total protein content of clinical samples is to be related through the use of a designated biuret method. The Expert Panel on Proteins of the Inter- national Federation of Clinical Chemists has adopted this standard. NBS is at present developing the SRM meeting these specifications. Specifications for other clinical standards that are now being developed deserve particular attention. The Scandinavian Committee on Enzymes is cur- rently engaged in the development of specifications for reduced ,B-nicotinamide adenine dinucleotide (NADH) to be used in NADH-coupled enzyme assays. Measurements of the physicochemical prop- erties of the NADH and of its relative reaction rates with specified substrates are being incorporated to provide specifications for acceptable purity and in- hibitor content of NADH. A similar activity is in progress at NBS, where experimental work is under way having as its initial objective the delineation of criteria for high-purity, inhibitor-free NADH. These criteria are intended for subsequent use in the selec- tion of a preparation of NADH for certification as the SRM. As a complementary effort, SRM sodium pyruvate will also be established. Another programme in its experimental phase is a joint effort between the Area Committee for Clinical Chemistry of the NCCLS and the Standards Com- mittee of the American Association of Clinical Chemists to develop specifications for two enzymes isolated from human source material. These are red- cell lactate dehydrogenase (1.1.1.27) and placental alkaline phosphatase (3.1.3.1). Studies of highly puri- fied preparations of these enzymes are providing the experimental data needed as the basis of a proposed standard. Experimental studies are being directed towards the future certification of other standard materials at 717 STANDARDIZATION NBS. Among these are SRMs for determining spec- trophotometric linearity in the ultraviolet range and for the standardization of spectrofluorometric mea- surements, and SRM-cuvettes having specified path- lengths and parallel sides accurate to 1 part in 104. Selected chemicals as standards The selection of a commercial or a laboratory- purified chemical preparation for use as a standard is facilitated if the counterpart standard material is available so that an experimental comparison of physical properties and analytical behaviour can be performed to establish its applicability. Without the standard material for comparison, but where analyti- cal specifications for that chemical as a reagent have been established, the purity of the selected chemical should conform to or surpass those specifications. However, where neither the standard material nor analytical specifications exist, the chemical prepara- tion must be selected only after evidence has been obtained to show its purity or its correspondence with the highest quality preparations reported in the literature. Where literature values serve as the points of reference, measurements on the selected prepara- tion must be obtained by use of appropriate and properly calibrated instruments. Complex mixtures as standards The use of highly complex mixtures that simulate clinical samples is often necessary for calibrating clinical analyses. Their use is essential for quality control (i.e., for monitoring precision and detecting systematic deviations). However, calibration and control are distinctive functions and the same mix- ture must not be used for both purposes. In either function, the mixture must be homogeneous and the period of time for which each of its analytes is stable must be known. The necessity for using a complex mixture for calibration is justified only if direct use of a standard material is not feasible or is otherwise inappropriate. On the other hand, to be justified for use in calibra- tion, the concentrations of its constituents should be assigned by determinations against a standard refer- ence material. (The latter condition should be ful- filled if at all possible.) The National Committee for Clinical Laboratory Standards (1972c) has adopted a " Standard for calibration reference and control materials in clinical chemistry " as a tentative standard for the specifications that complex mixtures are to fulfil to be appropriate for calibration or control purposes. According to this tentative standard, developed by the Area Committee for Clinical Chemistry and scheduled for adoption in 1973 as an approved standard, a preparation designated as a " calibration reference material " would have concentration values assigned for given constituents by the use of officially designated analytical methods (when they exist) or by the most accurate methods available. The preci- sion of these values for normal or below-normal assigned values is restricted to an overall uncertainty interval (i.e., ±2 standard errors) of 8% of the 95% normal range. For above-normal assigned values, the allowable uncertainty interval is increased by the ratio of the assigned value to the mid-point normal- range value for that constituent. Preparations whose constituent concentrations are known with some- what less precision are to be labelled as " control materials with assigned values " if the overall uncer- tainty interval does not: (a) exceed 20% of the 95% normal value or (b) exceed the standard deviation between single measurements made on the same day in the same laboratory using optimum methods. A third classification designated in the tentative stan- dard is that for control materials without assigned values. To meet the specifications for any of the classifications, the variability in the concentration (or amount) of substance in sample vials must be de- monstrated not to exceed ± 1 % of the mean in more than 5% of the vials. One final point should be made: few, if any, of the "control" materials currently used as calibration reference materials qualify as such according to the requirements speci- fied in the tentative standard. LABELLING For the proper and effective use of the chemical materials that are supplied for clinical tests, it is essential not only that they conform to specifications ensuring reliable performance but also that they be accompanied by appropriate information. Full de- tails of the composition of reagents; the species, tissue, or body-fluid source of material used as the matrix (e.g., human erythrocytes, bovine serum); the identity of anticoagulants, chemical preservatives, antibiotics, etc.; and the species, tissue source, and specific activity of enzymes, are essential to the user. For the information that must accompany re- agents and kits, the NCCLS Area Committee on Labeling has developed a proposed standard entitled " Labeling of laboratory reagents " (National Com- mittee for Clinical Laboratory Standards, 1972b). 718 DIAGNOSTIC MATERIALS 719 The identity of the product by name and lot numbes, its manufacturer, instructions for recommended user, known contraindications to its applicability, and any hazards associated with its use are to appear in the labelling. Directions as detailed as necessary for proper use; necessary storage conditions and expira- tion date; and instructions for reconstituting the product, as well as information on its stability in the reconstituted form are to be supplied with the product. There is unquestioned efficiency in a system in which the necessary reagents, controls, and standards for designated tests are manufactured centrally as kits for distribution to many laboratories. However, the efficacy of the system can be destroyed unless these kits and kit components arrive in user labora- tories without having suffered deterioration. The manufacturer needs to use appropriate means for ensuring the preservation of sensitive materials while they are in transit. RECOMMENDATIONS (1) Chemicals requiring specifications for use in clinical analysis need to be identified. (2) Standards required as specifications or as stan- dard materials (issued by recognized standards orga- nizations) need to be identified. (3) Official methods of assigning values to sub- stances in complex mixtures used as calibration reference materials and as control materials with assigned values need to be developed. (4) Standards organizations to participate in the development of specifications or standard materials for clinical analysis should be identified. Any work now in progress should be made widely known. (5) There should be an international clearing- house for information on identified needs, potentially helpful standards organizations, and current efforts of standards organizations working on these sub- jects. RESUMEt MATtRIELS DE DIAGNOSTIC POUR L'ANALYSE CLINIQUE Le technicien travaillant dans un laboratoire clinique et le fabricant qui foumit le mat6riel servant aux analyses ont en commun le souci de connaitre la qualit6 des subs- tances chimiques employ6es pour les epreuves de diagnos- tic afin d'assurer la fiabilit6 des r6sultats. L'utilisation de produits chimiques conformes a des sp6cifications bien d6finies garantitl'absence de tout contaminant ind6sirable. Si ces sp6cifications n'existent pas, il importe de choisir avec soin les r6actifs. L'utilisateur doit connaitre la nature et l'origine biologique d'un reactif, son mode de pr6para- tion ainsi que son activit6 et les m6thodes employ6es pour la mesurer. Dans beaucoup d'applications cliniques, les normes de purete de l'eau doivent etre rigoureusement re3p.ct6es si l'on veut 6viter les erreurs dues a la presence d'impuret6s. Divers organismes ont pr6par6 des substances etalons pour des 6preuves de laboratoire clinique: dosages de l'h- moglobine, du cholest6rol, de 1'ur6e, de I'acide urique, de la cr6atinine, de la bilirubine, du calcium, du potassium, etc. I1 existe des 6talons pour les filtres et les solutions destin6s aux examens spectrophotom.triques. D'autres sont en pr6paration comme ceux de la lactate-de'shydrog6nase 6rythocytaire et de la phosphatase alcaline placentaire. Tous les r6actifs utilises dans les te3ts cliniques doivent etre accompagn6s d'une notice d6taillee afin de per- mettre leur identification et leur utilisation judicieuWe. REFERENCES American Chemical Society (1968) Reagent chemicals, 4th ed., Washington, D.C. American Society for Testing and Materials (1970) Stan- dards, Part 23: Industrial water; atmospheric analysis, Philadelphia Cali, J. P. et al. (1972) US National Bureau of Standards Special Publication 260-36 Clin. Chem., 1962, 8, 405 College of American Pathologists, Standards Committee (1967) Amer. J. clin. Path., 47, 654 720 STANDARDIZATION International Committee for Standardization in Haema- tology (1967) Brit. J. Haemat., 13 (Suppl.), 71 Meinke, W. W. (1971) Analyt. Chem., 43, 28A National Committee for Clinical Laboratory Standards (1972a) Approved standard ACC-J, standardizedprotein solution (bovine serum albumin), Los Angeles National Committee for Clinical Laboratory Standards (1972b) Proposed standard, labeling of laboratory re- agents, Los Angeles National Committee for Clinical Laboratory Standards (1972c) Tentative standard, standard for calibration of reference and control materials in clinical chemistry, Los Angeles Peters, T. Jr (1968) Clin. Chem., 14, 1147 Stier, A. R. & Miller, L. K. (1972) Coll. Amer. Path., 41, March US National Academy of Sciences (1972) Specifications and criteria for biochemical compounds, 3rd ed., Wash- ington, D.C. Winstead, M. (1967) Reagent grade water, Austin, Tex., American Society of Medical Technologists