|
Vol: 56(70) No: 4 / December 2011 Effects of Obesity: A Multivariate Analysis of Laboratory Parameters Tamás Ferenci Department of Control Engineering and Information Technology, Budapest University of Technology and Economics, Magyar tudósok krt. 2., 1117, Budapest, phone: (361) 463-4027, e-mail: ferenci@iit.bme.hu Zsuzsanna Almássy Department of Toxicology, Heim Pal Children’s Hospital, Üllői út 86., 1089, Budapest, Hungary, e-mail: almassy.zsuzsa@t-online.hu Adalbert Kovács Dept. of Mathematics, “Politehnica” University of Timisoara, Piaţa Victoriei nr. 2, 300006 Timisoara, Romania, e-mail: adalbert.kovacs@mat.upt.ro Levente Kovács Department of Control Engineering and Information Technology, Budapest University of Technology and Economics, Magyar tudósok krt. 2., 1117, Budapest, e-mail: lkovacs@iit.bme.hu Keywords: Obesity, Biostatistics, Multivariate statistics, Principal Components Analysis, Cluster Analysis Abstract It is well-known that obesity has a marked effect on many of the routinely measured laboratory parameters. An obvious example is the serum level of various blood lipids: hyperlipidemia, hypercholesteremia are often observed in obese people. Our research aims to provide a more thorough understanding of the effects of obesity on laboratory parameters, concentrating on every laboratory parameter (not just those that are already established as being related to obesity) and their correlational structure. We focus on adolescent population, as they are the most important from the public health point of view. Material and methods: A cross-sectional clinical study was performed that included the observation of n=163 male children (aged 14-18), consisting of healthy volunteers from Hungarian secondary schools and obese patients treated with E66.9 “Obesity, unspecified” diagnosis (ICD-10). The observation included the recording of 33 laboratory parameters from blood sample. To explore this database, we performed Principal Components Analysis (PCA) and Factor Analysis (FA) to ease the understanding on correlations of the laboratory parameters by identifying those groups of variables that have strong stochastic connection. Such connections between laboratory parameters were further analyzed by Cluster Analysis (CA). Results: The applied methods all reveal similar patterns of association between different laboratory parameters. Variables that are found to be stochastically connected, also share physiologic similarities. The effects of obesity can also be exposed. Conclusion: Stochastically connected laboratory parameters – with different physiological interpretation – can be in fact statistically identified and used to draw conclusions about the multivariate structure of laboratory results. References [1] R.E. Andersen, Obesity: etiology, assessment, treatment, and prevention, Champaign: Human Kinetics Publishers, 2003. [2] SRI for Health, “Hungary’s healthcare and social system”, Budapest: Strategic Research Institute for Health, 2004. [3] M.M. Avram, A.S. Avram, and W.D. James, “Subcutaneous fat in normal and diseased states: 1. Introduction”. Journal of the American Academy of Dermatology, vol. 53, pp. 663–670, 2005. [4] Hungarian Central Statistical Institute, “Mortality by common death causes (1990- )” [http://portal.ksh.hu/pls/ksh/docs/hun/xstadat/ xstadateves/iwnh001.html]. [5] M. Mamtani, and H. Kulkarni, “Predictive Performance of Anthropometric Indexes of Central Obesity for the Risk of Type 2 Diabetes”, Archives of Medical Research, vol. 36, pp. 581–589, 2005. [6] S.A. Jebb, and M. Elia, “Techniques for the measurement of body composition: A practical guide”, Int. J. Obesity Related Metabolic Disorders, vol. 17, pp. 611–621, 1993. [7] T. Ferenci, “Biostatistical analysis of obesity related parameters in Hungarian children”, (MSc Thesis, in Hungarian), Budapest University of Technology and Economics, Department of Informatics and Control Engineering, 2009. [8] T. Ferenci, L. Kovács, Zs. Almássy, L. Szilágyi, B. Benyó and Z. Benyó, Differences in the Laboratory Parameters of Obese and Healthy Hungarian Children and Their Use in Automatic Classification. In Proc. of 32th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Buenos Aires, Argentína, pp. 3883-3886. , August 2010. [9] R Development Core Team (2009). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, URL http://www.R-project.org/. [10] P.N. Tan, M. Steinbach and V. Kimar, Introduction to Data Mining, Addison-Wesley, New York, 2006. [11] B. Everitt, S. Landau, M. Leese and D. Stahl, Cluster Analysis (Wiley Series in Probability and Statistics), Wiley, New York, 2011. [12] I.T. Jolliffe, Principal Component Analysis. Springer-Verlag, 1986. [13] B. Flury, A first course in multivariate statistics. New York: Springer, 1997. [14] I.K. Fodor., A survey of dimension reduction techniques, LLNL, URL: https://computation.llnl.gov/casc/sapphire/pubs/148494.pdf, Accessed: 14. 03. 2011. [15] R.L. Gorusch, Factor Analysis. Lawrence Erlbaum Associates, Hillsdale, 1983. [16] T. Ferenci, Zs. Almássy, A. Kovács and L. Kovács: Effects of obesity: a multivariate analysis of laboratory parameters. In: 6th International Symposium on Applied Computational Intelligence and Informatics (SACI 2011). Timisoara, Románia, 2011.05.19-2011.05.21. Budapest: pp. 629-634. (ISBN: 978-1-4244-9107-0). [17] C. Dziuban and E. Shirkey, “When is a correlation matrix appropriate for factor analysis? Some decision rules”. Physiological Bulletin, vol. 81, pp. 358-361, 1974. [18] H.F. Kaiser, “The application of electronic computers to factor analysis”. Educational and Physiological Measurement, 1960, vol. 20, pp. 141-151. [19] H.F. Kaiser, “The varimax criterion for analytic rotation in factor analysis”. Psychometrika, vol. 23 (3), 1958. |