Abstract Background Intestinal permeability (IP) is increased in Crohn’s disease (CD) patients and their first degree relatives. The causes of barrier disruption remain unclear but likely relate to inflammation with possible effects of nutrients and microbes. Infliximab has been shown to improve IP in CD but the impact of nutritional therapy on IP is unknown. We prospectively assessed the effects of nutritional therapy on IP in a randomised controlled trial, comparing the Crohn’s disease exclusion diet (CDED) to the gold standard exclusive enteral nutrition (EEN) in children with CD. Mannitol is an easily absorbed small sugar that reflects the small bowel (SB) surface area, whereas the disaccharide lactulose is only absorbed through larger pores and reflects permeability; therefore, the ratio of lactulose/mannitol (L/M) represents SB relative permeability. Methods The CDED study was a 12-week prospective, international, multi-centre, randomised controlled trial in children with mild-to-moderate active luminal CD, comparing CDED to EEN. During the first 6 weeks of the study patients in the CDED group received CDED Stage 1 diet + 50% calories from liquid formula (Modulen, Nestle) whereas the EEN group were fed exclusively with Modulen. A L/M test for intestinal permeability was performed at weeks 0 and 3 by administering a sugar solution containing lactulose (5 g) and mannitol (1 g) and then collecting urine for LC-MS/MS analysis. A cut-off L/M ratio of 0.015 was chosen, based on published literature (McOmber et al. JPGN 2010). Results L/M ratios were available at both time points for 39 patients (23 CDED and 16 EEN). At baseline, 9/23 (39%) CDED and 8/16 (50%) EEN patients had a normal L/M ratio, whereas at 3 weeks of treatment this increased to 15/23 (65%) and 9/16 (56%), respectively. Using generalised estimating equation analysis there was no difference in change of L/M between groups (p = 0.193). In both groups, 50% of those with abnormal L/M ratio at baseline (ratio > 0.015) normalised at Week 3 (7/14 for CDED; 4/8 with EEN). Interestingly, 1/14 (7%) CDED cases with normal L/M ratio became abnormal at 3 weeks, vs. 3/16 (19%) in the EEN group (NS). These findings indicate an improvement in IP with 3 weeks of nutritional therapy; however, there was no correlation between change in IP status and failure to respond to therapy or poor compliance to the diet at 3 weeks. Conclusions Although both EEN and CDED are associated with improved IP, this was not observed in all patients, despite clinical improvement. This suggests that small bowel IP alone may not be a primary mechanism for early clinical response; the effect of IP on sustaining remission following dietary therapy will require further study.