In Reply.鈥擨t is true that four patients initially included in group A (TPN alone) in whom the volume of drainage decreased by less than 30% after 15 days of treatment were transferred to group B (TPN plus somatostatin). This possibility was established when the trial was designed. Three of these four patients obtained a definitive healing of the fistula when somatostatin was added.
We agree that our hypothesis would have been more strongly supported if we had included such patients in their original group. Nevertheless, this group of transferred patients was not considered to obtain closure rates or morbidity numbers in any of the groups, and, in the same way, it was not considered for statistical purposes because the patients were initially treated with TPN alone, and only after 15 days was somatostatin added. This group of transferred patients was too small to make any conclusion if we had