Healing by fibrosis represents the end stage of a response to a variety of insults. Known antecedents to fibrosis in the colon include ischaemia, inflammation (e.g., Crohn's colitis), and noxious chemicals (e.g., irritant laxatives). Endogenous pancreatic enzymes may have a role in enhancing, even if not initiating, lesions which lead to fibrosis, as shown by animal experiments in which ligation of the pancreatic duct mitigated or delayed the response to ischaemia.1

It has been recognised fairly recently that a distinctive form of fibrosing colonopathy develops in young children with cystic fibrosis taking mega-doses of pancreatic enzymes, and it was suggested that the enzyme preparations themselves may be the cause.2 The most striking feature is a thick band of submucosal fibrosis, suggesting that some component of the enzyme preparation must gain access through the mucosal barrier, but evidence of healed mucosal ulceration has only been observed in a few cases. Vascular damage and a moderate eosinophilic infiltration were observed in some cases, and in several there was quite noticeable serosal fibrosis. Clinically, in addition to the signs of intestinal obstruction, several patients had chylous ascites. These features suggest that translocation of a component of the enzyme preparation—whether the enzymes themselves or the enteric coating used to protect them from gastric acid—occurs across the full thickness of the colon. It may be that the higher absorption of sodium, glucose, amino acids, and water observed in the small intestine of patients with cystic fibrosis3 occurs in the colon too, and physiological studies showing an increased potential difference across the rectal mucosa would support this speculation. It is not known how often subclinical colonic fibrosis occurs in cystic fibrosis, but it can be an incidental finding at autopsy.4

Against this background, Bansi et al's case report (see page 283) is of great interest because it describes typical fibrosing colonopathy in an adult who was not thought to have cystic fibrosis, but who had taken very high doses of pancreatic enzyme supplements. The lessons would seem to be that fibrosing colonopathy is not confined to cystic fibrosis, nor to children. There is no doubt that this patient did not have classic cystic fibrosis, but the authors have not ruled out one of the many atypical conditions that are also associated with malfunction of the cystic fibrosis transmembrane regulator protein (CFTR). The 12 commonest CFTR mutations found in the British population have been excluded, but without completely sequencing the gene, it is virtually impossible to rule out a CFTR related disorder, because to date more than 800 different CFTR gene mutations have been described. It should be noted that the recurrent attacks of pancreatitis which led to pancreatico-duodenectomy in this patient are often associated with cystic fibrosis mutations, even when only a single copy is present.5 The episodes of jaundice which occurred when the patient was younger could have resulted from common bile duct stenosis, which has also been attributed to cystic fibrosis.6 The surgical finding of the caecum embedded in a solid mass of fibrous tissue is reminiscent of some early descriptions of patients with cystic fibrosis and intestinal obstruction following years of abdominal pain. These reports long antedated enteric coated enzymes.7 Adhesions following surgery are a frequent problem in cystic fibrosis, and it may be that exogenous pancreatic enzymes have a damaging effect on anastomoses. In retrospect, the earlier diagnosis of Crohn's disease in this patient may have been incorrect.

Fibrosing colonopathy is usually associated with excessive doses of pancreatic enzymes. The doses of enzymes consumed by this patient were very high, and similar to those found in the UK8 and US9 case control studies, namely around 40 000 lipase units/kg/day. The American study demonstrated that the risk of developing fibrosing colonopathy increased correspondingly with increasing enzyme dosage. The British study also found an apparent association between fibrosing colonopathy and the type of enteric coating, but this was not seen in the American study and may be related to other factors such as institutional brand preference. In addition, the confounding effect of switching between different brands of enzymes was a notable feature of the UK series. High doses of the order seen in these cases were made possible by the high strength preparations, but a similar dosage can be achieved in young children with standard strength preparations, and has been associated with fibrosing colonopathy.10 None the less, some patients with cystic fibrosis seem to tolerate very high concentrations of pancreatin without developing fibrosing colonopathy, suggesting that in affected patients there may be additional factors which render their colon more vulnerable to enzyme damage.

One such factor might be diet. The patient described by Bansiet al was taking a low fat diet prior to her hemicolectomy, but no mention is made of the fibre content. It is well known that short chain fatty acids derived from fibre provide a large part of the nutrition to the colon, but in the past nutritional advice to patients with cystic fibrosis has been to recommend a high calorie, high carbohydrate, high protein, and, in recent years, normal or high fat content, and to allow this intake the component generally reduced was, by implication, fibre. In a recent study it was shown that gastrointestinal symptoms in children with cystic fibrosis were inversely correlated with the fibre content. The mean daily fibre intake in children with cystic fibrosis was less than half that of controls, and in those with severe gastrointestinal symptoms, it was only about a quarter of that of healthy controls. There was also a non-significant trend towards higher pancreatin intake in those children with more severe gastrointestinal symptoms.11 I have suggested elsewhere that depriving the colon of short chain fatty acids by limiting fibre intake, and then administering unnecessarily high doses of pancreatic enzymes, may be tantamount to adding insult to injury.4 Fibre reduces paracellular permeability in the rat colon,12 whereas increasing intestinal permeability (by oleic acid or reserpine) and then giving excessive doses of pancreatic enzymes produced intestinal eosinophilia and necrosis.13

Finally, although in their discussion Bansi et al follow convention by focusing on the dose of lipase taken by their patient, it would seem intrinsically more likely that other enzymes present in pancreatin preparations may be responsible for the damage observed. These enzymes include proteases and elastase, which are present in approximately similar proportions to lipase in all the enzyme products available. The epidemiological association between high enzyme doses and fibrosing colonopathy has been established. There is now clearly a need for experimental studies which will shed light on the pathogenesis.14