455 29 Turanitz, K., and A. Wottawa, Institute of Biology, Forschungszentrum Seibersdorf, A-2444 Seibersdorf (Austria)
Sister-chromatid exchanges in lymphocytes of smokers SCE exchanges proved to be a useful method in investigating genetic damage in humans exposed to environmental mutagens and carcinogens. We analysed SCE frequencies in lymphocytes of 12 healthy persons with differing smoking habits. Over a period of 1 year, 8 blood samples were taken per person, cultivated and their SCE frequencies determined. The smokers, males and females, aged 23-41, were divided into 2 groups: heavy smokers ( 9 10 cigarettes/day) and moderate smokers (< 10 cigarettes/day). Per sample, 35-50 metaphases were analysed. SCE frequencies were tested with the Student test. Non-smokers had average SCE frequencies of 5.73 --- 2.44, moderate smokers 6.16 -+-2.68 and heavy smokers 7.06 --- 3.13. The comparison of smokers-non-smokers, heavy and moderate smokers yielded highly significant results with clearly increased SCE frequencies in heavy smokers. Temporary overlappings of group means advise analyses over an extended period of time with either a homogeneous group of persons or a relatively large group.
30 Vollmar, J., Boehringer Mannheim GmbH, D-6800 Mannheim (F.R.G.)
On the statistical analysis of the SCE test The statistical bases for analysis of SCE tests were investigated by using extensive results obtained from 9 different laboratories (altogether 15 600 cells from control investigations in vivo and in vitro). The SCEs did not fit any of the known distribution types (i.e. binomial, Poisson, negative binomial or normal distribution). For this reason, non-parametric methods of analysis are proposed (e.g. Jonckheere test or one-tailed Wilcoxon rank sum test). The arithmetic mean of the SCE counts per cell should be used as the appropriate parameter of location for each individual (-- sample unit). The median of the individual means should be calculated for the control and treatment groups. The necessary sample sizes were determined on the basis of the statistical method of analysis, the probabilities of a and fl errors, and the biological relevant difference in the data available. The following approximate numbers are given for a = fl = 0.05. For bone marrow of the Chinese hamster: