In multicentre studies of malignant and inflammatory diseases, whole blood, cell or tissue samples are often collected for analyses of gene expression to predict or monitor treatment effects. For correct analysis, sample stability during handling and transport is crucial. In developing the logistics for multicentre studies in malignant melanoma and inflammatory bowel disease, we found poor stability of a number of transcripts using the PAXgene Blood RNA System, which was advertised to maintain RNA stability for several days at room temperature. The results indicate that general statements on sample stability are not reliable and have to be verified for the specific transcripts of interest.