| dc.description.abstracteng | Animal experiments are an important part of medical research. On the other hand, since their application there has been a discussion about the benefits of animal testing in relation to animal rights, animal welfare and animal ethics. Out of this discussion, the principles of the three R's of Russel and Burch have arisen in 1959. These form the basis for animal welfare laws in the field of research worldwide. They say that care must be observed in animal experiments:
• to keep the required number of animals as low as possible,
• to avoid as much as possible the suffering of the animals from side effects and pain,
• if possible, to use non-vertebral animals or other forms of research, such as cell lines or model calculations.
In the animal experiments on osteoporosis research of the research laboratory of the Department of Trauma Surgery, Orthopedics and Plastic Surgery of the University Medical Center Göttingen, on whose data this work is based, the experimental setup and the examinations have remained the same for several years. As part of an approval procedure for this series of experiments, it, therefore was suggested to waive the current control groups in future trails. This could minimize the number of animals used. In future experiments the evaluation could then be done on the basis of the data collected from control groups of previous years.
Thus, the aim of this study was to investigate whether the use of the already existing control group data from the earlier experiments allows a future renunciation of current control groups.
The data of this dissertation comes from animal experiments on rats, which were carried out between 2009 and 2014 by doctoral students and laboratory research staff of the Department of Trauma Surgery, Orthopedics and Plastic Surgery of the University Medical Center Göttingen. In osteoporosis research, effects of various anti-osteoporotic therapies (medically and physically by means of vibration) on the vertebral bodies, femora, tibiae as well as muscles were investigated. There always were two control groups, divided into ovariectomized and intact animals and 3-5 treated groups per experiment. The group size was between 15 and 12 animals. There were 108 measurement parameters, which were collected for each group of the 8 experiments and were examined with respect to the question concerning the need for current control groups.
Using the STATISTICA-program it could be shown in this dissertation that in the direct, retrospective comparison of the control group data, many of the 108 study parameters analysed in the 8 experiments show distinct differences in their distribution. Using the Kruskal-Wallis test, the distributions of the individual parameters of the individual experiments were compared. It was found that in 14 parameters there was no significant difference in the distribution of the results in both control groups. By means of a statistical meta-analysis, reference values were created for these 14 parameters, which could replace the control groups in the future (using R-Studio program).
As possible causes for the respective fluctuations in results, various reasons are discussed. As possible causes are discussed amongst others changes in equipment, different sources of animal supply and alternating investigators.
Furthermore, a retrospective analysis using as example the cortical thickness of the tibia as a single study parameter showed that the comparison of the treated animals in some cases led to different results compared to all control groups. This would have led to a misinterpretation of the result of the studied therapy.
After weighing all the factors to be considered in recommendations using historical control data, the present investigations show that not a single parameter fulfills the criteria for using historical control data. A reduction in the number of laboratory animals by omitting the current control group seems therefore not possible.
In addition, to be able to find treatment effects, the concurrent control group is always the most appropriate and, for scientific reasons, should be favored over the historical control group. | de |