Patients with extensive or complicated Crohn's disease (CD) at diagnosis should be treated straightaway with immunosuppressive therapy according to the most recent guidelines. In patients with localized and uncomplicated CD at diagnosis, early use of immunosuppressive therapy is debated for preventing disease progression and limiting the disabling clinical impact. In this context, there is a need for predictors of benign or unfavourable subsequent clinical course, in order to avoid over-treating with risky drugs those patients who would have experienced spontaneous mid-term asymptomatic disease without progression towards irreversible intestinal lesions. At diagnosis, an age below 40 years, the presence of perianal lesions and the need for treating the first flare with steroids have been consistently associated with an unfavourable subsequent 5-year or 10-year clinical course. The positive predictive value of unfavourable course in patients with 2 or 3 predictors ranges between 0.75 and 0.95 in population-based and referral centre cohorts. Consequently, the use of these predictors can be integrated into the elements that influence individual decisions. In the CD postoperative context, keeping smoking and history of prior resection are the strongest predictors of disease symptomatic recurrence. However, these clinical predictors alone are not as reliable as severity of early postoperative endoscopic recurrence in clinical practice. In ulcerative colitis (UC), extensive colitis at diagnosis is associated with unfavourable clinical course in the first 5 to 10 years of the disease, and also with long-term colectomy and colorectal inflammation-associated colorectal cancer. In patients with extensive UC at diagnosis, a rapid step-up strategy aiming to achieve sustained deep remission should therefore be considered. At the moment, no reliable serological or genetic predictor of inflammatory bowel disease clinical course has been identified.