Background: Crohn's disease (CD) is relatively uncommon in childhood. The clinical features of the disease in the paediatric age are well established but little is known about its long-term outcome and prognosis. Aims: The aim of the study was to document the “burden” of the disease in a cohort of children diagnosed with CD between 1990-2003 and to identify at diagnosis and at 1 year the factors predictive of a disabling course. Patients and Methods: All patients with a diagnosis of CD were identified. All clinical charts were reviewed and the data extracted were entered in a database and checked for accuracy. Inclusion criteria were a diagnosis of CD made after 1990 by a paediatric gastroenterologist, an age at diagnosis < 18 years and a disease duration > 2 years. Outcome predictors were collected at diagnosis and at 1 and 5 years of follow-up (FU), whereas outcome measures were assessed at last FU visit. Results: A total of 121 patients (M/F: 70/51) fulfilling to the inclusion criteria were identified. Median age at diagnosis was 11.4 years (0.4-16.3), median time between onset and diagnosis was 7.3 months (0.0-71.8) and median follow-up time was 5.4 years (2.1-15.7). The induction therapy was enteral nutrition (EN) in 72 pts (60%), azathioprine plus EN or steroids in 19 pts (16%) and steroids alone in 21 pts (17%). 52/121 pts (43%) and 59/121 pts (49%) were on azathioprine at 1 year of FU and at the last FU visit, respectively. Endoscopic follow-up showed a trend toward a progressive improvement of disease location over time. Considering disease behaviour, there was a 3-fold increase in stricturing/penetrating phenotype from baseline to 5 years. 14/121 pts (11.6%) and 52/121 pts (43%) underwent surgery at 1 year of FU and at the last FU visit, respectively. Male gender (P= 0.02), disease duration (P< 0.001) and stricturing disease (P= 0.001) at diagnosis and right sided (P= 0.002) and stricturing (P< 0.001) disease at 1 year were associated with increased risk for surgery. The median number of rescue therapies, inpatients days and total hospital days (in- plus out-patients admissions) per patient/year of FU were 0.6 (0.0-2.2), 2.9 (0.3-18.8) and 6.5 (2.2-23.3), respectively. The median number of rescue therapies, inpatient and total hospital days per patient/year of FU were significantly higher in the panenteric and ileocolon groups than in the other groups at 1 year (P < 0.001). The same parameters were significantly higher in the stricturing/penetrating groups than in the inflammatory group at diagnosis and at 1 year of FU (P= 0.001). Conclusions: We found that as many as 43% of our patients underwent surgery in the first 5 years of disease. Gender, disease duration, disease location and behaviour were factors predictive of a high risk of surgery. Patients who did not receive azathioprine in the 1st year were unlikely to receive this medication at 5 years. The burden of illness, measured in terms of rescue therapies, number of inpatients and total hospital days per patient/year of FU, was greater in the first 3 years and decreased thereafter. The burden of the illness was closely related to the disease location and behaviour.