BACKGROUND: Necrotizing soft tissue infections (NSTIs) are associated with significant mortality if not promptly diagnosed and surgically treated. AIM: This study aims to compare patients with severe skin and soft tissue infection treated with or without a surgical intervention and to identify risk factors that can predict the need for early surgery. METHODS: Demographics, clinical, laboratory, Risk Indicator for Necrotizing Fasciitis (LRINEC) and imaging results were retrospectively collected. RESULTS: There were 91 non-NSTI (group 1), 26 NSTI who were operated (group 2) and eight suspected NSTI who were not operated (group 3). In the multivariate analysis, skin necrosis, tachycardia, CRP value and hyperglycemia were predictive for surgery. A performance analysis revealed AUC of 0.65 (95%CI: 0.52-0.78) as to the LRINEC score for the use of surgery. The AUC for a predictive model associating four variables (heart rate, skin necrosis, CRP and glycemia at admission) was 0.71 (95%CI: 0.59-0.84). In terms of outcome, the median length of stay (LOS) was statistically higher in group 2 vs. group 1 (seven days (5-15) vs. 34 days (20-42), p < .001) and in group 2 vs. group 3 (34 days (20-42) vs. 14 days (11-19), p = .005). The overall in-hospital mortality at 30 days was 3.2% and did not statistically differ between the three groups. CONCLUSIONS: Although the LRINEC score performed well in predicting surgery, the AUC of a model combining four predictive variables (glycemia, skin necrosis, CRP and heart rate) was superior. Further research is needed to validate this model.