Abstract
In our recent project of establishing reference intervals for 13 plasma proteins following the NCCLS proposed guideline (C28‐P, 1992), confounding and interacting relationships were found among criteria for partitioning reference individuals. Therefore, we found it necessary to test for the validity of the criterion by multivariate analysis so that such relationships could be revealed. Meanwhile, to find difference appropriate for the partition, sex‐ and age‐related differences were examined in common screening tests whose reference intervals (RI) were conventionally set for each subgroup. The result led to our standard for partitioning when the difference between subgroups exceed 10% of the span of a RI derived without partitioning. Among the 13 plasma proteins, the multivariate analyses and the standard indicated gender‐based RI was for IGM, α1 acid glycoprotein, α2 macroglobulin, and transferrin, transthyretin. None of the age‐related differences were large enough. In serum IgM, ceruloplasmin, and transferrin, however, females tended to have larger differences between ages, exemplifying interaction of gender on age‐related changes of the analytes. Smoking‐related differences were > 10% in IgG and transthyretin. However, differences in the latter proved to be spurious multivariately. These results suggest a necessity for additional descriptions on the validity of the partitioning criterion in the NCCLS proposed guideline. J. Clin. Lab. Anal. 11:117–124. © 1997 Wiley‐Liss, Inc.
Keywords: partitioning criterion, multiple regression, multivariate analysis, multiple logistic regression, confounding, interaction, sex‐ and age‐related differences, smoking
References
- 1. NCCLS : How to define, determine, and utilize reference intervals in the clinical laboratory. NCCLS 12: 1–47, 1992. (NCCLS: C28‐P, 1992). [Google Scholar]
- 2. Ichihara K, Kawai T: Determination of reference intervals for 13 plasma proteins based on IFCC international reference preparation (CRM470) and NCCLS proposed guideline (C28‐P, 1992): Trial to select reference individuals by results of screening tests and application of maximal likelihood method. J Clin Lab Anal 10: 110–117, 1996. [DOI] [PubMed] [Google Scholar]
- 3. Kleinbaum DG, Kupper LL, Muller KE: Confounding and interaction in regression In Applied Regression Analysis and Other Multivariable Methods, 2nd ed. Duxbury Press, CA, 1988, p. 163–180. [Google Scholar]
- 4. Kleinbaum DG: Modeling strategy for assessing interaction and confounding In Logistic Regression. Spring‐Verlag, New York, 1994, p. 191–266. [Google Scholar]
- 5. Fisher LD, Van Belle G: In Biostatistics: A Methodology for the Health Sciences. J Wiley & Sons, New York, 1993, p. 496–595. [Google Scholar]
- 6. Baudner S: Why a new international reference preparation (IRP) for human plasma proteins. J Clin Lab Anal 7: 273–277, 1993. [DOI] [PubMed] [Google Scholar]
- 7. Sokal RR, Rohlf FJ: Biometry, 3rd ed., W.H. Freeman, New York, 1981, p. 207–217. [Google Scholar]