British Journal of Surgery 1998, 85, 1118–1120 Outcome after ‘curative’ surgery for carcinoma of the lower third of the rectum B. TOPAL, F. PENNINCKX, L. KAUFMAN*, L. FILEZ, R. AERTS, N. ECTORS† and R. KERREMANS Departments of Abdominal Surgery and †Pathology, University Clinic Gasthuisberg, Leuven and *Department of Statistics, University Clinic Vrije Universiteit Brussels, Brussels, Belgium Correspondence to: Professor F. Penninckx, Department of Abdominal Surgery, University Clinic Gasthuisberg, Herestraat 49, 3000 Leuven, Belgium Background Controversy exists about the optimal surgical resection for lower third rectal carcinoma. The aim of this retrospective study was to analyse whether the type of surgery is a significant predictor of outcome after curative surgery alone. Methods Eighty-two consecutive patients underwent abdominoperineal rectum excision (APRE, 41 patients) or sphincter-saving operation (SSO, 41 patients) for adenocarcinoma at 3·5–7·5 cm from the anal margin. Cox proportional hazards technique with univariate and corrected (multivariate) analyses and the Kaplan–Meier life-table method were used to evaluate the data. Results Tumour wall penetration and lymph node involvement, but not the tumour level or the type of surgery, were found to be significant predictors of outcome. The local recurrence rate at 1, 2 and 5 years was 10, 22 and 26 per cent respectively after APRE, and 5, 13 and 21 per cent after SSO. The disease-free survival rate at 1, 2 and 5 years was 85, 67 and 58 per cent respectively after APRE, and 88, 78 and 62 per cent after SSO. Conclusion Tumour-related factors are significant predictors of outcome. The type of surgery (APRE or SSO) did not seem to be a significant variable in this non-randomized study. Controversy surrounds the optimal curative resection for carcinoma of the lower third of the rectum. Published series on the results of abdominoperineal rectum excision (APRE) and sphincter-saving operations (SSOs) in rectal carcinoma are difficult to assess. Most authors report on cancer of the whole rectum, while others have reported only on middle and lower third rectal cancers together. For an appropriate comparison of techniques, however, only a limited part of the rectum should be studied be- cause outcome deteriorates with decreasing distance be- tween tumour and anal verge 1,2 . It is agreed that complete mesorectal excision and a distal margin of 2 cm are adequate to contain local cancer spread 3,4 . Strict appli- cation of the 2-cm distal margin rule, however, means that every carcinoma in the lower third of the rectum has to be treated by APRE. Indeed, if one takes into account smooth muscle shortening in the fresh, unpinned specimen of about 50 per cent of its in situ length and a 2–4-cm length of the anal canal in the conscious patient, the lower edge of a rectal cancer would have to be located at 6–8cm above the anal verge for an SSO to be indicated. In clinical practice, however, this theoretical rule is not always respected. The outcome in patients with cancer in the lower third of the rectum was therefore reviewed to identify pre- dictors of oncological outcome and, more specifically, to evaluate whether the type of operation (APRE or SSO) is related to failure, i.e. local or distant disease recurrence. Patients and methods Eighty-two consecutive patients with carcinoma of the lower third of the rectum were studied. The lower third of the rectum was defined as the zone extending from the anorectal ring to 7·5 cm from the anal margin 5 and all patients underwent primary ‘curative’ surgery (R 0 ) between January 1984 and December 1993. Exclusion criteria were rectal tumours extending into the anal canal (21 patients), intraoperative tumour breach or soiling (eight patients) and perioperative adjuvant (chemo)radiation (12 patients). A sphincter-saving procedure was performed in 41 patients and an APRE in another 41. Five patients had a syn- chronous gastrointestinal carcinoma, two in the APRE group and three in the SSO group, treated at the time of rectal surgery. Comparative analysis showed no significant difference with regard to sex ratio, tumour diameter, depth of tumour wall pene- tration, lymph node involvement, tumour node metastasis (TNM) stage, differentiation grade, synchronous carcinoma and operating time between the APRE and SSO groups. Mean(s.d.) age of patients in the APRE and SSO groups was 68(10) and 60(10) years respectively P  0·001, t test). The mean(s.d.) dis- tance from the anal margin, measured with a rigid proctoscope, was 4·5(1·2) cm in the APRE group compared with 5·3(1·1) cm in the SSO group ( P = 0·001, t test). Of four surgeons, one performed 24 of the 41 SSOs ( P  0·001,  2 test). A sphincter-saving procedure consisted of low anterior resec- tion (11 patients), restorative rectal excision with straight coloanal anastomosis (12) or colon pouch–anal anastomosis (18). The anastomosis was hand sewn in 22 patients and stapled in the remainder. A protective stoma was constructed in 23 patients. The mean(s.d.) duration of hospital stay was 19·7(9·5) days after APRE and 17·7(9·2) days after SSO ( P not significant). Postoperative complications occurred in 40 patients: in 18 of 41 after APRE and in 22 of 41 after SSO ( P not significant). Post- operative complications after APRE were wound infection (five patients), bladder dysfunction (four), ileus (three), urinary tract infection (UTI) (two), bleeding (two), pulmonary complications (two), deep venous thrombosis (one), urethral fistula (one) and colostomy retraction (one). In the SSO group symptomatic anastomotic leakage occurred in nine patients, UTI in four, ileus in four, pulmonary complications in three, wound dehiscence in one, bladder dysfunction in one and delirium in one. Follow-up was complete up to 31 December 1995. All patients were re-examined at 6-month intervals, including serum carcino- Paper accepted 20 January 1998 1118 © 1998 Blackwell Science Ltd