Construction of a loop ileostomy is usually advised in patients having an ileal pouch-anal anastomosis to minimize the complication of chronic pelvic sepsis. Formation and closure of a loop ileostomy was associated with a 41 percent and 30 percent complication rate, respectively, in a prospective series of 34 patients. This morbidity must now be assessed in relation to the benefits of avoiding temporary fecal diversion in restorative proctocolectomy.

The crypt cell production rate was measured in 14 patients with adenomatous colorectal polyps, 17 patients with colorectal cancer and 12 control subjects. The median (interquartile range) rate (cells per crypt per hour) was found to be significantly higher (P less than 0.001) in the polyp (2.45 (1.94-3.20)) and cancer (3.01 (2.35-3.68)) groups compared with controls (1.25 (0.70-1.85)). A double-blind cross-over study was performed in patients with adenomatous polyps consisting of 2 months' treatment, 2 weeks' washout, followed by 2 months' treatment with dietary calcium supplementation (1.25 g day-1) versus placebo. A significant reduction in the crypt cell production rate occurred with calcium treatment compared with the placebo (1.25 (0.6-2.25) versus 2.15 (1.58-3.08) cells per crypt per hour, P = 0.035). This study demonstrates a significant reduction in mucosal cell proliferation by dietary calcium supplementation in patients with adenomatous polyps. Such treatment may be worthy of further investigation in patients at high risk of developing colorectal polyps.

No abstract

3500 Background: Intensive long-term follow-up after surgery for colorectal cancer is common practice but neither the actual benefit nor the optimal methodology is known. Methods: Pragmatic factorial randomised controlled trial in 39 UK hospitals, comparing minimum follow-up (which included a single CT scan at 12-18 months) with 3-6 monthly blood carcinoembryonic antigen (CEA) testing and 6-12 monthly computerised tomography (CT) imaging of the chest, abdomen and pelvis following 1202 participants for 3-5 (mean 3.7) years. Results: The proportion of participants with recurrence treated surgically with curative intent was lower than predicted (6.0% overall) but was about 3x higher in the more intensive than minimum follow-up arms (p=0.019). The adjusted odds were 2.7 for CEA only (p=0.035) and 3.4 for CT only (p=0.007); the absolute differences in detection rate in the more intensive arms compared to minimum follow-up were 4.3-5.7% (5.8-8.0% per protocol). Combining CEA and CT provided no additional benefit (adjusted odds for CT+CEA arm = 2.9). The absolute difference in the proportion of participants with recurrence treated surgically with curative intent in the factorial comparison was 1.4% for CEA (p=0.28) and 2.8% for CT (p=0.04). There was no statistical difference in colorectal cancer deaths nor overall deaths in the minimum compared to the intensive follow-up arms. Conclusions: Both regular CEA measurement and CT scanning result in significantly higher rates of diagnosis of operable recurrent colorectal cancer compared to minimal follow up. There is no benefit in monitoring with both CEA and CT. To date no difference in the overall mortality has been demonstrated. CEA monitoring combined with a single CT scan at 12-18 months seems likely to be cost effective. Clinical trial information: 41458548.

We agree with Mr Widdison's comment that perineal drains can be brought out more anteriorly with a consequent reduction in postoperative discomfort of group 2 patients (undergoing primary closure of the perineum and perineal drainage).With respect to the comments of Lewis et al., our results show a longer postoperative hospital stay for group 2 patients than in their series. We do not know their rate of infection but our patients had a high rate of perineal wound infection (47 per cent) with method 2, which made partial or complete opening of the perineal wound necessary. This led to a longer hospital stay.Our patients were not randomized, although the study was done prospectively, requiring that the three series be homogeneous with respect to age, sex and histological degree of malignancy. A non-parametric test could be used for the differences in hospital stay but the statistical test used is robust and the conclusions identical with either method.