In recent years, serum levels of the prohormone procalcitonin have been investigated in a number of studies in relation to postmortem sepsis diagnostics, as macroscopic and histomorphological findings are, as a rule, nonspecific. However, due to advanced haemolysis, it is often not possible to determine serum procalcitonin (PCT) levels in cases of sepsis-related death. Moreover, the impact of postmortem interval on PCT levels is largely unclarified. In view of this, the present pilot study investigated PCT levels in the serum, aqueous humour, and cerebrospinal fluid in a study population of 25 persons who died of sepsis and a control population of 25 deaths unrelated to sepsis. Using the Mann-Whitney U test, statistically significant differences in PCT levels were determined for all the analysed samples from the study and control populations. Logistic regression analysis was used to calculate cut-off values for sepsis diagnosis for all the sample types. Furthermore, the serum elimination rates published by Tsokos et al. (Int J Legal Med 114:237-243, 2001) were used to calculate the PCT levels at the time of death for the cases with a known postmortem interval. The results of our study demonstrate that, taking account of the postmortem elimination process, it is possible to infer the value at the time of death from the procalcitonin levels measured in all three sample types and to interpret this with the aid of a defined cut-off value. The findings need to be verified based on a larger study population.