Clin Chem Lab Med 2012;50(5):841–844 © 2012 by Walter de Gruyter • Berlin • Boston. DOI 10.1515/cclm-2011-0868 Age dependence of within-subject biological variation of nine common clinical chemistry analytes Anna Carobene 1, *, Maria Stella Graziani 2 , Claudia Lo Cascio 2 , Livia Tretti 2 , Eveline Cremonese 2 , Teowoldemedhn Yabarek 3 , Giovanni Gambaro 4 and Ferruccio Ceriotti 1 1 Diagnostica e Ricerca San Raffaele, Istituto Scientifico Universitario San Raffaele, Milan, Italy 2 Clinical Chemistry Laboratory, Azienda Ospedaliera Universitaria Integrata di Verona, Verona, Italy 3 Department of Medicine, Division of Nephrology, Azienda Ospedaliera Universitaria Integrata di Verona, Verona, Italy 4 Nephrology and Dialysis, Columbus-Gemelli University Hospital, Università Cattolica, Rome, Italy Abstract Background: The knowledge of biological variation (BV) data is important for clinical decisions and as a basis for defining analytical quality specifications. However, in gene- rating reliable data of biological variation there are still some unsolved problems, such as age dependence. The aim of our work is to verify this aspect. Methods: Twenty-six subjects divided into three groups by age were studied. Blood samples were collected in lithium heparin tubes for four weeks at one week intervals, on the same day of the week (Tuesday) and at the same time of day (8–9 a.m.) by the same phlebotomist. They were analysed in duplicate for creatinine, urate, calcium, albumin, total cho- lesterol, high density lipoprotein (HDL) and low density lipoprotein (LDL)-cholesterol, triglycerides and iron. After outlier exclusion by Cochran’s test, components of biological variation were calculated by ANOVA. The significance of the differences between results of the classes was also calculated with the Student’s test (t-test) and the Fisher’s test (F-test). Results: Excluding albumin, the group 3 results (age range from 78 to 98 years) showed significantly lower CV within subjects (CV W ) than the other two groups. Conclusions: Our data seem to highlight the relevance of the age when choosing the reference subjects for biological varia- tion studies. The level of within-subject biological variation of the elderly group may have been further reduced by the homogeneity of the group constituted by individuals living together in the same nursing home. Keywords: analytical quality; biological variation; quality specification. Introduction The knowledge of biological variation (BV) data is essential for clinical decision-making, to calculate the reference change value (1). Moreover, since 1970 many authors (2) have rec- ommended the use of BV as a basis for defining analytical quality specifications (3). In the 1999 Stockholm consensus conference (4), where a hierarchy of models to set quality specifications was presented, BV was set at the second level. The first level being practically unobtainable for most of the analytes, the data of biological variation represent the most effective and objective manner to define analytical quality specification, whereas the “state of the art” is highly influenced by the context and could be different according to the groups of data used (5). In 1999 Ricos et al. realized a popular database (6) collect- ing all the experimental data known in the literature. Since then, the database has been expanded and updated (7), but there are limitations: Data are not available for all the analytes (there are about • 1000 laboratory tests and data of BV are present only for about 300 analytes); For many analytes the data of BV come from just one or a • few publications; In some cases, the quality specifications derived from the • data are very restrictive if compared with the state-of-the- art; There are only few papers about the possible differences • between healthy and diseased population data (8); There is no mention of the possible data differences based • on gender and age of the subjects (9). The problems of limitations of the BV database were also discussed in occasion of a recent meeting of over 40 medi- cal laboratory opinion leaders. The participants recognized no limitation of the BV concept but several limitations in the available data (10). BV being increasingly relevant, it is fundamental to under- stand all the sources of variability that could influence the BV data. Even though Fraser (11) proposed a protocol for calcula- ting BV, the literature data are frequently divergent and the reasons for the differences are not exactly known. One of the causes might be the dependence of within-subjects BV on the age of the subjects. The aim of our work is to verify this aspect. *Corresponding author: Anna Carobene, Laboratorio di Standardizzazione, Diagnostica e Ricerca San Raffaele S.p.A., via Olgettina 60, 20132 Milan, Italy Phone: +39-02-26432850, Fax: +39-02-26434178, E-mail: carobene.anna@hsr.it Received November 21, 2011; accepted December 25, 2011; previously published online January 20, 2012 Brought to you by | University of Georgia Libraries Authenticated Download Date | 6/1/15 3:49 AM