Physiological Reference Values: A Shared Business?
by Anders Kallner
The theme of the Second European Symposium of the Clinical Laboratory and in vitro Diagnostic Industry —held 6-7 February 2003 in Barcelona, Spain— was Physiological Reference Values: A Shared Business? The symposium, chaired by X. Fuentes-Arderiu, addressed several of the problems encountered in establishing, teaching, and using physiological reference values.
The format of the symposium was unique: questions were posed to a panel of experts who were then expected to address every question. The audience was also invited to participate in the discussion. The audience was comprised of scientists from Spain and Europe and representatives of the major manufacturers of reagents and instruments. The symposium was arranged by the Catalan Association for Clinical Laboratory Sciences and sponsored by IUPAC, IFCC, and major industries. Detailed proceedings of the symposium will be published by Hyltoft-Petersen.1
Better understanding of reference intervals is vitally important because diagnosis of disease is frequently based on measurement of biochemical quantities (components) of blood, serum, urine, etc. Conclusions from obtained result are usually drawn by comparison with results obtained in healthy individuals or, less common, with individuals with specified diseases or conditions. The acceptance of a reference point or reference interval is as important for the physicians as a fix point was for Archimedes: ‘Dòs moí poü stö kaì kinö tën gen’ (‘give me a fixed point and I will move the earth’). Originally the profession used the term ‘normal value’ but about 25 years ago it was agreed that normal was an improper term since it was very hard to find any ‘normal’ individuals. The term was changed to ‘reference interval’ or ‘reference value’, depending on its use.2 Also the concept was changed and broadened by introducing the concept of a ‘reference population’ and even ‘reference individuals’. We can now refer a reference interval to a defined group of individuals who need not be healthy but may suffer from a specific condition. The concept also allows for relating the reference interval to age, gender and ethnicity, as appropriate. The age and gender variation is known for many commonly used quantities. 3
All approaches to establishing reference intervals require large groups of individuals (e.g., a minimum of 120 individuals in the IFCC recommendation). It is an insurmountable task for any individual laboratory to establish its own reference values. Many published reference values are taken from textbooks and lack an acceptable "audit trail" to the original source. Thus, a reference interval is important for diagnosis and screening of populations, but as soon as the individual becomes a patient and undergoes treatment, the changes in the values of certain quantities might be more important than the absolute values. In these cases the patient becomes his or her own reference.
Considering the nature of the problem and the impact on health, diagnosis, and economics, it is important that reference intervals are scientifically based and appropriate for the actual situation of the patient and the physician. Up until quite recently establishment of reference intervals has been very loosely regulated. However, the "EU IVD Directive 98/794 on in vitro diagnostic devices" can be interpreted as placing the responsibility for establishing reference intervals on the manufacturer of reagent kits and/or instruments:
"Where appropriate, the instructions for use must contain [...] the reference limits for the quantities being determined, including a description of the appropriate reference population: […]."
Nonetheless, the newly published international standard that is used for the accreditation of medical laboratories EN/ISO 151895 states,
"Biological reference intervals shall be periodically reviewed. If the laboratory has reason to believe that a particular interval is no longer appropriate for the reference population, then an investigation shall be undertaken, followed, if necessary, by corrective actions. A review of biological reference intervals shall also take place when the laboratory changes an examination procedure or pre-examination procedure, if appropriate."
The USA has a similar set of rules, whereby manufactures are obliged to provide reference intervals and the laboratories are obliged to prove that they are applicable in their environment.
The symposium made clear that although the EU IVD Directive 98/794 puts the responsibility on industry to disseminate information about physiological reference intervals for their measurement systems, the industry is not well equipped or in a position to collect the necessary raw data without collaboration from the profession and medical laboratories. Panellists also argued that reference intervals need to be locally modified by the clinical laboratories. There are biological reasons for this, such as ethnicity or dietary habits, and also methodological reasons. Thus, there are many methods or method modifications available that do not give the same results although they set out to measure the same quantity (i.e., a deficient transferability of results between laboratories).
From a metrological point of view, transferability of results would be achieved if the same calibrator were used. Globally that requires that each calibrator is traceable to the same primary standard, preferentially the realization of the SI unit. In biological systems this hypothesis does not work very well. A major reason is the complex nature of the sample and the influence of the "matrix" defined as "all components of a material system, except the analyte.6" The matrix may not be consistent and may therefore have different influences on the results from patient to patient. For years the profession has tried to cope with the situation by external quality assessment schemes (EQA, Europe) and proficiency testing (PT, US). The effect has largely been to identify that there are different measurement procedures on the market and that their results differ. Most of these schemes use artificial samples based on bovine or equine serum that may introduce bias due to the changed matrix. Furthermore the EQA or PT schemes are primarily designed to monitor the trueness of results and measure the bias between laboratories and it serves this purpose from a general point of view (e.g., CLIA 88 in the USA). However, its usefulness for the individual laboratory is limited.
An alternate solution to approach the trueness, reduce the bias, and meet the transferability demand would be to establish a network of laboratory comparisons using patient material. A NCCLS recommendation7 describes how such comparisons can be made and a simplified and practical, yet powerful procedure has been developed.8 A very extensive experiment to align results of laboratories and to create identical reference values in a region has been carried out in the Nordic countries.9 Other multicenter projects have been carried out.10-13 During the symposium, it was emphasized that establishment or appropriate reviewing of reference intervals are very cumbersome and expensive tasks and much would be gained if the reference intervals could be established in a sustainable way. The costs for establishing reference intervals can be exceedingly high and it was argued that it could be an obligation of the heath authorities to finance this procedure.
The information technology is rapidly developing and most health care systems either have or are developing laboratory information systems (LIS) and hospital information systems (HIS). This allows the laboratory to give more detailed information about the patient and already many LIS give differentiated reference intervals depending on the age and gender of the patient. Most LIS provide cumulative reports and mark values outside the reference intervals to the benefit of the physician and patient.
It might be useful to develop reference intervals for other than healthy populations, e.g specific disease groups that can be characterised by one or more biological properties. Typical diseases are liver and kidney diseases, metabolic disturbances like diabetes or screening procedures e.g. for Downs syndrome (trisomy 21). Such reference intervals are usually based on multivariate procedures, resulting in algorithms that require special software, usually available in the LIS. The interpretation of the results for a diseased individual is, however, much more complex because most diseases develop – in a positive or negative direction – and the values of the quantity varies accordingly in a fashion that we presently do not have sufficient structured information about. A solution might be to report the deviation from a value or distribution of values.
Conclusions from the Symposium
The rulings by the EU Directive 98/79/EC (Annex I 8.7) and the EN/ISO 15189 (5.5.5) on the responsibilities of industry and laboratories to create reference intervals are not clear and—as written—difficult to interpret and follow. Symposium attendees proposed that a request is sent to CEN TC 140 and ISO TC 212 to clarify the demands of the directive and the standard.
It was concluded that a document resolving this question should address at least the following items:
In addition, it was concluded that collaboration should be established between professional organizations, industry, and health authorities to achieve common reference intervals for homogeneous groups, including multicenter reference intervals. Also, the concept and interpretation of reference intervals should be taught to pre- and post-graduate students of health sciences and physicians.
1. Hyltoft-Petersen P., Report from the Second European Symposium Clinical Laboratory and in vitro Diagnostic Industry "Physiological Reference Values: A Shared Business?"
Clin Chem Lab Med 2003; 41(6):825-8
2. Solberg HE, PetitClerc C. International Federation of Clinical Chemistry (IFCC), Approved recommendation (1988) on the theory of reference values. Part 3. Preparation of individuals and collection of specimens for the production of reference values. J Clin Chem Clin Biochem. 1988;26(9):593-8.
3. Kallner A, Gustavsson E, Hendig E. Can Age and Sex Related Reference Intervals be Derived for Non-Healthy and Non-Diseased Individuals from Results of Measurements in Primary Health Care? Clin Chem Lab Med 2000;38(7):633-654.
4. EU IVD Directive 98/79 on in vitro diagnostic medical devices. Official J European Communities L331/1 98-12-07.
5. EN/ISO 15189 Medical laboratories – Particular requirements for quality and competence. ISO, Geneva 2003.
6. Dybkaer R.Vocabulary for use in measurement procedures and description of reference materials in laboratory medicine. Eur J Clin Chem Clin Biochem 1977;35:141-73.
7. NCCLS EP9-A. Method comparison and bias estimation Using Patient Samples; Approved Guideline. 1995 ISBN 1-56238.283-7.
8. Kallner A, Khorovskaya LA, Groth T. Can the method comparison procedure in NCCLS recommendation EP9-A be simplified? Clin Chem 2002;48:A174.
9. Nordic Reference Interval Project.
10. Andrew CE, Hanning I, McBain AM, Moody D, Price A. A model for a multicentre approach to the derivation of reference intervals for thyroid hormones and testosterone for laboratories using indentical analysers. Clin Chem Lab Med 2000;38:1013-9.
11. Ferre-Masferrer M, Fuentes-Arderiu X, Goma-Llongueras M, Aluma-Trullas A, Aramendi-Ramos M, Castano-Vidriales JL, Esteban-Salan M, Graells-Ferrer M, Jimenez-Gonzalez A, Jimenez-Lobo C, Lopez-Esparza M, Pares-Pollan L, Regulez-Uranga M, Riesco-Prieto M, Rodriguez-Camba A, Vidal-Martinez J. Regional reference values for some quantities measured with the ADVIA Centaur analyser. A model of co-operation between the in vitro diagnostic industry and clinical laboratories. Clin Chem Lab Med. 2001;39:166-9.
12. Fuentes-Arderiu X, Ferre-Masferrer, M, Gonzalez-Alba JM, Escola-Aliberas J, Balsells-Rosello D, Blanco-Cristobal C, Gonzalez-Cruz E, Ibarz-Escuer M, Latorre-Marcellan P, Lugo-Arocena J, Mar-Medina C, Muros-de-Fuentes M, Vicens-Manero M. Related Articles, Links Multicentric reference values for some quantities measured with Tina-Quant reagents systems and RD/Hitachi analysers. Scand J Clin Lab Invest. 2001;61:273-6.
13. Fuentes-Arderiu X, Ferre-Masferrer, González-Alba JM, Villarino-González MI, Arrimadas-Esteban E, Cabrero-Olivé D, Cándenas-Arroyo M, García-García D, García-Lario JV, Idoate-Cervantes I, León-López C, López-Lazareno N, Mar-Medina C, Martí-Marcet I, Mauri-Dot M, Pérez-Valero V, Reta-Manterola A, Sánchez-Eixeres MR. Multicentric reference values for some quantities measured with the Elecsys 2010 analyser. Clin Chim Acta 2001;304:143-6.
Anders Kallner <firstname.lastname@example.org> is from the Department of Clinical Chemistry of the Karolinska Hospital in Stockholm, Sweden. He is the current president of the IUPAC Division of Human and Human Health.
Page last modified 30 June 2003.
Copyright © 2002-2003 International Union of Pure and Applied Chemistry.
Questions regarding the website, please contact email@example.com