Background: Patients with primary rectal cancer undergoing low anterior resection are often reconstructed using a pouch procedure. The aim of this trial was to compare colon J pouch (CJP) with transverse coloplasty pouch (TCP) reconstruction with regard to functional results, perioperative mortality and morbidity. As there is considerable uncertainty over the true anastomotic leak rate in patients with a TCP, the study analysed short-term outcome data. Methods: Elective patients suitable for either procedure after sphincter-saving low anterior resection were eligible. Randomization took place during surgery. The primary endpoint was the rate of late evacuation problems after 2 years; secondary endpoints were anastomotic leak rate, perioperative morbidity and mortality. Results: Between 21 October 2002 and 5 December 2005, 149 patients were randomized. All 76 patients randomized to TCP had the procedure compared with 68 of the 73 patients (93 per cent) randomized to CJP. Both groups were comparable with regard to demographic and clinical characteristics. Surgical complications (CJP: 19 per cent; TCP: 18 per cent) and the overall anastomotic leak rate (8 per cent) were equally distributed in both groups. Conclusion: This trial demonstrated a comparable early outcome for TCP and CJP. This contradicts previous reports suggesting a higher leak rate after TCP. Registration number: ISRCTN78983587 (http://www.controlled-trials.com).