Drug Information Journal - March 2009 - (Page 139) MEDICAL INFORMATION 139 Jaap H. M. Dijkman, MD, MBA, PPM Senior Director at Covance, Zeist, The Netherlands Professor Callum G. Fraser, PhD, FAACB Consultant Clinical Biochemist at the Scottish Bowel Screening Centre Laboratory, Dundee, United Kingdom F. Peter Treasure, PhD, CStat Director at Peter Treasure Statistical Services Ltd, King’s Lynn, United Kingdom Gordon F. Kapke, PhD Senior Director Biomarkers at Covance, Indianapolis, Indiana Local Laboratory Reference Intervals in Clinical Trials Assessment of whether laboratory results in clinical studies are abnormal relies heavily on the laboratory upper or lower reference limits. While central laboratories use a single reference interval for an entire study population, local laboratories apply their own reference intervals. Due to bias caused by method differences between local laboratories, such reference intervals are not interchangeable. We have compared the reference intervals of local laboratories after harmonizing their test results, using an analytical calibration step for routine clinical biochemistry analytes. Our results show that differences in local reference intervals could result in significantly variable criteria being used for assessment of normality versus abnormality—up to more than 30%. Significant differences were observed for most analytes used in routine safety monitoring. We conclude that local reference intervals are not useful for drug toxicity trend analysis. Instead, a single reference interval after data harmonization can be used with advantage in addition to assessment of changes in individuals versus their baseline results. Key Words Local laboratory; Reference interval; Calibration; Safety monitoring Correspondence Address Jaap H. M. Dijkman, Sr. Director, Covance VCL BV, Driebergseweg 26, 3708 JC, Zeist, The Netherlands (email: jaap.dijkman@covance.com) . INTRODUCTION LITERATURE STUDY Laboratory data are of major importance in drug safety assessments since they are considered objective tools to supplement clinical judgments and the transfer of results across study populations. Moreover, the spectrum of available biomarkers to assess drug safety and efficacy is continuously being developed and refined. Although central laboratories are widely used in clinical trials, local laboratories continue to play an important role for trials where the use of such central laboratories is not practical or feasible. Local laboratories are primarily performing safety testing. The collection and analysis of these data remain a challenge in view of differences in reference intervals, analytical methods, and reporting units. These challenges are widely recognized. Other factors that may account for differences between local and central laboratories are differences in population characteristics, such as ethnicity or nutrition. While such differences should be reflected by differences between local and central laboratory reference intervals, analysis of central laboratory data in global clinical studies has shown these differences to be insignificant (see next section). Al- though units are easily standardized by mathematical conversion, differences in methods and associated reference intervals continue to raise questions for statisticians. Chuang-Stein has proposed a normalization procedure based on the results of proficiency surveys by devising an algorithm to adjust results by comparing them between different laboratories and making the adjusted results comparable across laboratories (1). While such a procedure could indeed improve result comparability, its feasibility is limited by, inter alia, the difficulties of gaining regular access to proficiency testing data, interpretation of the different schemes, especially in global studies, and the degree of participation in those schemes by the involved local laboratories. Thompson et al. have used the normal result distribution of study populations as the basis for defining a commonly acceptable reference interval (2). However, this proposal is limited by the scope of the observations, which may not necessarily cover existing differences in methods in global studies, for example, the welldocumented differences for commonly tested analytes such as alkaline phosphatase or lactate dehydrogenase. Ruvuna et al. have proposed the application of generalized lab norms from a “phantom laboratory” by compiling a standard Submitted for publication: December 21, 2007 Accepted for publication: July 2, 2008 Drug Information Journal, Vol. 43, pp. 139–150, 2009 • 0092-8615/2009 Printed in the USA. All rights reserved. Copyright © 2009 Drug Information Association, Inc.
For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.