The goal of the study was to analyse plasma procalcitonin (PCT) concentrations during infectious events of burns in ICU. Clinical and laboratory data were collected at admission and twice a week in burned patients admitted with a total body surface area (TBSA) >20%. Procalcitonin was determined using both a semi-quantitative detection (PCT-Q) and a quantitative immunoluminometric method (PCT-Lumi). A total of 359 time points in 25 consecutive patients with 40+/-17% (20-86%) TBSA burned, defined as a procalcitonin concentration associated with an inflammatory status according to society critical care medicine definition, were made. The principal site of infection was the respiratory tract (84% of patients required mechanical ventilation). PCT-Lumi values corresponded to the four semi-quantitative ranges of PCT-Q and statistically reflected the simultaneously observed inflammatory status (Kruskall-Wallis test). The area under the receiver operating characteristic curve for C-reactive protein (CRP) was higher than those for PCT and white blood cell (WBC) count, but this difference was not significant. The optimum PCT cut-off value was 0.534 ng/ml with sensitivity and specificity of 42.4% and 88.8%, respectively. However, PCT does not appear to be superior to C-reactive protein (CRP) and white blood count (WBC) as diagnosis marker of sepsis in burns. PCT is not sufficient to diagnose and to follow infection in burns admitted in ICU.