The aim of this uncontrolled, prospective, clinical study was to investigate the efficacy and safety of molgramostim administration in patients with severe sepsis. The subjects were 20 critically ill, mechanically ventilated patients with severe sepsis in a university intensive care unit (ICU). Molgramostim 300 microg s.c. was given every 12 h for 3 d. Treatment for severe sepsis was also administered as medically indicated. No adverse events (clinical or serum chemistry) were considered as drug related. Temperature (p = 0.334) and PaO2/FiO2 index (arterial oxygen tension/inspiratory oxygen fraction) (p = 0.178) were not significantly changed. Total leukocyte and neutrophil count increased significantly (p < 0.001) during drug administration. Simplified Acute Physiology Score II (SAPS II) was not significantly increased (p = 0.955), but there was a statistically significant decrease (p = 0.006) in Sepsis-related Organ Failure Assessment (SOFA) score. Death probability was not statistically different compared with mortality rate on day 28 and overall mortality (p = 0.238 and 0.700, respectively). There were statistically significant decreases (p < 0.01) in serum tumor necrosis factor-alpha (TNF-alpha), TNF-RII and interleukin-2 (IL-2), and an increase in TNF-RI levels between study entry and day 3. Mean ICU stay was 40.2 +/- 7.7 d. In conclusion, molgramostim administration may not affect serum chemistry and PaO2/FiO2 index, may decrease SOFA score but does not produce significant clinical benefit in terms of patients' outcome compared with death probability. It may also influence TNF-alpha, TNF-RI and TNF-RII serum complex levels. These changes may be attributed to the natural clinical course of sepsis or therapy applied.