Comment

Patients with ulcerative colitis need to be able to obtain reliable advice about their likely prognosis and cancer risk. Up until now this has been difficult because the literature is imprecise. Studies determining progression over time or cancer risk have often been based on disease extent as determined by barium examinations. Moreover, some patients may have very severe distal disease while others may have very mild extensive disease. As has often been the case, a meticulous and very extensive Scandinavian study has helped to clarify the situation.

Moum et al have prospectively compared colonoscopic and histological findings at diagnosis and after 12–23 months of follow up in 384 cases of ulcerative colitis from a sample population of 496.1 Some of the findings are predictable but useful nevertheless. Extensive colitis at diagnosis may become less extensive while distal disease may become more extensive. Significant change in extent or reversion to normal endoscopic or histological appearances occurred in more than half of the patients by the median 14 month follow up. This report supports the findings of a smaller study2 which also suggested that change in colonoscopic extent of ulcerative colitis with time was the norm rather than the exception. This makes it unlikely that the striking proximal demarcation of disease that is often seen is due to the anatomical limit of the marginal artery, as suggested by Hamilton and colleagues.3

The study provides the best available data on short term progression of proctitis. Assessed on endoscopic appearances in 130 patients with proctitis, by the time of follow up 22% had progressed, 53% were unchanged, and 25% were normal. Since the follow up period was relatively short we will still have to rely on older studies based in part on radiology to determine likely long term outcome. The landmark studies came from St Mark's Hospital and showed that in patients with a diagnosis of proctitis based on a combination of rigid sigmoidoscopy and radiology, 12% extended to involve the descending colon above the iliac crest by 10 years and the cumulative risk of colectomy was 3–5% by 10 years4 in comparison with 15% overall for patients with ulcerative colitis of any extent.5 Extension of disease in 53.8% over a mean follow up of 12.7 years and a higher overall rate of surgery (37.6%) were reported in a review of 1116 patients with ulcerative colitis of any extent from the Cleveland Clinic,6 and other radiological based surveys (reviewed by Debinski and Kamm7) have shown rates of about 30% for progression of disease extent over time.

Perhaps unsurprisingly in Moum's study there was overall agreement between colonoscopy and histology in only 78% of cases at diagnosis and 60% at follow up. At follow up, 37% of those thought to have inflammation at colonoscopy had normal histology. Unfortunately, the colonic evidence of inflammation in these patients was not stated. Reddening of the mucosa has previously been shown to correlate poorly with histological inflammation in the stomach8 and this re-emphasises the importance of endoscopists reporting relatively objective features such as ulceration, contact bleeding, and loss of vascular pattern rather than a subjective change in colour intensity which may just reflect a transient change in mucosal perfusion. An impressive 24% of patients with colitis had completely normal histology at follow up. It is not stated how many biopsies had been taken per case however.

Another important conclusion from this study relates to selection of patients for cancer surveillance. Not only were 13% (34/408) of patients classified as having proctitis or left sided colitis at endoscopy reclassified as extensive colitis on histological criteria but the variation of both colonoscopic and histological criteria between diagnosis and follow up makes it impossible to determine with any accuracy which patients should be regarded as having extensive colitis with regards to long term surveillance. This reinforces the findings of an earlier UK study which reported a change from distal to more extensive colitis in 43% of patients monitored during a colonoscopic surveillance programme.9 The implication of this would seem to be that an “all or none” approach to surveillance might be more appropriate than a rather unreliable selection of patients based on disease extent. A longer term prospective study of the Scandinavian patient cohort might eventually provide clearer evidence for this.