Background: Predictive factors for a mild course of Crohn's disease (CD) may have therapeutic consequences, but as yet have not been identified. Aims: To identify baseline factors that predict mild CD and design a predictive scoring system. Methods: A retrospective, multicenter study of newly diagnosed CD patients allocated to mild CD (no therapy, mesalazine only, or mesalazine with a single initial short course of low-dose prednisone) or moderate CD (all other patients including resected patients). Results: 162 patients (median follow-up 43. months) were analyzed: 47 mild CD and 115 moderate CD. For mild CD versus moderate CD, mean age at first diagnosis was higher (41.1 versus 33.9. years, p=0.02), mean C-reactive protein (CRP) concentration was lower (1.6 versus 3.6. mg/L, p < 0.01), and perianal lesions were less frequent (0% versus 10.4%, p=0.02). The combined incidence of complications (stenosis, any type of fistula, extraintestinal complications or fever) was 21.3% in mild CD versus 35.7% in moderate CD (p=0.07). A scoring system based on age, CRP, endoscopic severity (adapted Rutgeert's score), perianal lesions and combined incidence of complications was developed which can predict a mild prognosis at the initial diagnosis, giving patients the chance of simplified therapy and accelerated step-up in the event of treatment failure. Conclusions: Approximately a third of CD patients experience a mild disease course and require only basic therapy. A possible scoring system to predict mild CD which may avoid overtreatment and unnecessary risks for the patient and costs is suggested.