Impact of Reference Materials on Accuracy in Clinical Chemistry CARLO FRANZINI 1 and FERRUCCIO CERIOTTI 2 1 Universita ` di Milano, Istituto di Scienze Biomediche Ospedale L. Sacco, G.B.Grassi 74, 20157 Milano, Italy, and 2 Istituto Scientifico S. Raffaele, Dipartimento di Medicina di Laboratorio, Milano, Italy The analytical accuracy of the results of routine clinical chemistry measurements is contributed by a two-steps mechanism, involving transferring trueness from a higher metrological and monitoring the time-stability of trueness itself. In both operations, different materi- als are used: however, accuracy in the routine assay of genuine patient samples has to be the end product of this overall process. To such an aim, the materials must show an intermethod behavior similar to that of patient sera, i.e., they have to show commutability. Definitions of commutability and methods for assessing such a property are mentioned. The following aspects of lack of commut- ability of materials are then discussed: frequency; effects on the measured interlaboratory variability; and effects on the recalibration of analytical systems. The causes giving rise to lack of commutabil- ity are neither clear or easy to be shown. Matrix effect is one of the main causes; also, differences in the characteristics of the compo- nent being measured are often responsible for noncommutability of materials for enzyme activity measurements. Examples of these two different situations are given. It is concluded that, for an efficacious overall quality assurance process, either a set of minimally pro- cessed patient sera or commutable reference materials are to be used in the operations concerned with the control of trueness. An additional alternative approach is based on the use of materials with system-specific assigned values. Copyright © 1998 The Canadian Society of Clinical Chemists KEY WORDS: commutability; accuracy; control materi- als; reference materials. Introduction T he overall accuracy of routine clinical chemistry test results is contributed to by a two-step mech- anism, involving pursuing trueness (by means of calibration) and assuring consistency of truness through time and space (quality control). Both mechanisms involve a multisteps chain of events, assuring the traceability of the result to some higher metrological “reference” (Figure 1). The latter may be represented by the relevant SI units or by some intermediate step of the chain such as a “reference material” or a “reference method,” according to the state-of-the-art development in the specific mea- surement technology, the definition of the quantity (or the analyte) being measured, and/or other rele- vant factors (Figure 1). The transferring of trueness from the selected reference to the routine measurement result is not immediate, but includes a number of intermediate steps, building up the chain. Among the reasons for such a complex trueness-transferring mechanism are the low practicability of the measuring proce- dures belonging to the higher metrological level (generally referred to as “reference system”), and the need for the intervention of a multitude of individ- uals and technical groups such as standardization bodies, professional clinical chemists, and industrial companies. Nevertheless, the model showed practi- cable and was successfully applied on several occa- sions (1–3). At most steps of the chain, a “material” is used as the witness of truness. The materials (operatively used either as calibrators or as controls) belong to several categories (primary, secondary, reference, and so on), mainly on the basis of their composition or matrix (more or less exactly defined), and of the reliability of the quantity value(s) assigned to them. Among them, the reference material (whether certi- fied or not) plays a key role in the trueness trans- ferring mechanisms (4), because it aims at bridging the truness of the reference system with that of the routine measurements. Because the aim of the over- all process is to guarantee the truness of the routine measurements on genuine patient samples, the need for the trueness transfer to be as reliable for the patient samples as it is for the materials must be carefully considered. This may be crucial at the lower levels of the chain, where matrix-sensitive routine methods are used. Correspondence: Prof. Carlo Franzini, Universita ` di Milano, Istituto di Scienze, Biomediche Ospedale L.Sacco, Via G.B.Grassi 74, 20157 Milano, Italy. Received January 12, 1998; revised and accepted March 18, 1998. Clinical Biochemistry, Vol. 31, No. 6, 449 – 457, 1998 Copyright © 1998 The Canadian Society of Clinical Chemists Printed in the USA. All rights reserved 0009-9120/98 $19.00 + .00 PII S0009-9120(98)00054-X CLINICAL BIOCHEMISTRY, VOLUME 31, AUGUST 1998 449