Sepsis is one of the leading causes of mortality in critically ill patients, and providing a timely diagnosis and early intervention is necessary for successful treatment. Procalcitonin (PCT) may be a better marker of sepsis than conventional inflammatory markers. The aim of this study was to evaluate the clinical utility of the PCT level as a marker of sepsis. Forty-five patients with sepsis, 24 patients with pneumonia who did not meet the SIRS criteria (PN) and 56 controls were enrolled in this study. The levels of PCT and other serum markers were measured, and their utility as markers of sepsis was assessed. The serum PCT levels exhibited statistically significant differences between the three groups (p<0.0001). The PCT levels in the sepsis group (29.3 ± 85.3 ng/mL) were significantly higher (p<0.001) than those observed in the PN group (0.34 ± 8.6 ng/mL) and the control group (0.74 ± 2.1 ng/mL), according to a post hoc analysis. There were no differences in the white blood cell (WBC) counts or C-reactive protein (CRP) levels between the three groups. Fourteen of the 45 patients with sepsis had positive microbiological blood cultures (Gram-positive cocci [GPC] in seven patients, Gram-negative rods [GNR] in six patients, other types of bacteria in one patient). The 13 patients with GNR or GPC were categorized into the GNR group or GPC group according to the identified pathogens. The serum PCT levels were significantly higher in the GNR group (149.8 ± 199.7 ng/mL) than in the GPC group (19.1 ± 41.8 ng/mL) (p<0.05), although there were no differences in the WBC counts or CRP levels between these groups. When the cut-off value for the PCT level was set at 16.9 ng/mL, the sensitivity and specificity for the detection of GNR infection were 85.7% and 83.3%, respectively. The PCT level is a potentially useful marker of the type of causative pathogen in patients with sepsis whose measurement may facilitate the selection of appropriate empiric antibiotic treatment.