Morbidity, survival, and recurrence in 203 patients treated with curative low anterior resection (LAR) were compared with those in 100 patients treated with curative abdominoperineal resection (APR). The overall 5-year survival figures for the total number of, LAR and APR patients were 75.6 +/- 5.7%, 79.8 +/- 6.4% and 67.7 +/- 9.6%, respectively. The prognosis for cancers situated low enough in the rectum to involve the anal canal was poor even when managed by APR, as evidenced by a low survival at 5 years of 59.0 +/- 9.6% and a high pelvic recurrence rate of 34%. For all except these tumors, LAR proved at least equal to, or better than APR as a curative surgical method for middle and low rectal cancers, on the basis of 5-year survival being 79.8 +/- 6.4% vs 78.7 +/- 5.2%, operative mortality being 1.5% vs 1.0%, morbidity being 39.4% vs 59.0%, and the incidence of pelvic recurrence being 8.9% vs 13.5%. When deciding upon the most appropriate surgical procedure for rectal cancer, especially for middle or low rectal lesions, the patient should not simply be condemned to a permanent colostomy. Thus, we first attempt LAR for every lesion except those which are very advanced or those with anal canal involvement, if technically feasible and suitable for the individual patient.