INTRODUCTION: There is considerable controversy about timing of surgery in acute pancreatitis. Scoring system is widely used in Intensive Care Units (ICU) but above all in order to assess the severity of disease. AIM OF THE STUDY: In this prospective clinical study, modification of clinical evolution and scores (APACHE II, SAPS II and SOFA) are analyzed. Scores were computed, daily or every other day. Two particular phases were observed: the day of ICU admission and the day of surgical treatment. MATERIAL AND METHODS: Twenty-one patients are studied; they were all surgically treated only after identification of positive pancreatic coltures. Open-packing was performed 2-6 weeks after the beginning of acute pancreatitis. In six patients who died, a progressive deterioration was noticed between admission and identification of infected necrosis. All scores were higher in patients who died, in particular a significant difference was found between SOFA score at admission and at treatment (mean ± SD: 5.0 ± 3.2 vs 8.5 ± 3.0, p < 0.05). CONCLUSIONS: In conclusion when clinical conditions are critical a high SOFA score could contribute to indicate surgical treatment even without identification of sepsis