Dietary intake was assessed in 24 children with active Crohn's disease. Seventeen of the children were sex and height matched with healthy siblings of other patients. The mean energy intake was 1,764 KJ/day (420 kcal/day) lower in patients during relapse than controls (p < 0.05). All 24 children entered a randomised controlled trial of high-dose steroids compared to an elemental diet. The elemental diet was well tolerated orally in most cases; only one patient required nasogastric administration. There was a similar improvement in disease activity and duration of remission in both groups regardless of the site of disease. Subsequent growth velocity was significantly better in the group treated with the elemental diet despite a greater and more sustained increase in energy intake in the group treated with steroids. Reintroduction of specific foods in the elemental diet group initially caused symptoms in three children but were subsequently tolerated. During remission, there was no significant difference in energy intake, but vegetables were consumed significantly less frequently (p < 0.01) and confectionery significantly more frequently (p < 0.05) when patients were compared with their own siblings.