Several studies have shown that specific mRNA sequences can be successfully detected in formalin-fixed, paraffin-embedded tissues using reverse transcriptase-polymerase chain reaction (RT-PCR). Here, we test the hypothesis that gene expression levels can be accurately quantified in formalin-fixed, paraffin-embedded tissues by determining the ratio between the copy number of the mRNA molecule of interest and the mRNA copy number of a so-called housekeeping gene. The mRNA copy numbers of the variably expressed multiple drug resistance gene (MDR)-1 and four housekeeping genes (hypoxanthine phosphoribosyl-transferase-1, glyceraldehyde-3-phosphate dehydrogenase, beta-actin, and elongation factor-1a) were quantified by real-time-quantitative RT-PCR before and after formalin-fixation and paraffin-embedding of 576 tissue samples (heart, kidney, spleen, liver) from three beagle dogs. The results indicate that fixation and embedding drastically altered the ratios between the different mRNA copy numbers and that the relative expression levels of MDR-1 per any of the housekeeping genes were artificially increased or decreased up to more than tenfold. It would thus appear questionable to normalize quantitative expression data from fixed and embedded tissues by using housekeeping genes as reference. In contrast, tissue autolysis of up to 24 h and long-term storage of embedded tissues of up to 20 years had no additional effects.