In the area of proteomic, results of analysis were for a long time dependant on analytical variations. Nowadays, due to the emergence of new technologies and controls of data, these variations are less important than those due to preanalytical conditions, which are difficult to overcome. The reasons are due to the number of parameters and to the fact that the biologist is not always fully informed. In this document, we present the main preanalytical factors of variation, and their effects on the results of analysis. New technologies involving mass spectrometry are very promising, but they are very sensitive to parameters like the stability of samples and the choice of the clotting agent. Thus, it is more and more necessary to take these data into account.