PURPOSE:The long‐term prognosis after curative surgery for colorectal cancer was evaluated in relation to age and life expectancy as a possible basis for assessing the risk to benefit ratios in the elderly.METHODS:Data relating to 1,256 patients operated on from 1976 to 1994 were stored in a computer database prospectively from 1987. Patients were subdivided into four age groups (A=<60 years; B=60‐69; C=70‐79; D=≥80). Distribution of general contraindications to curative surgery was examined. In the 869 patients who underwent curative treatment (A=206; B=256; C=289; D=118), distribution of tumor stage and elective/emergency surgery and the operative mortality rate were evaluated. Crude and age‐corrected survival curves were calculated in 794 patients. The median crude survival of each group was related by gender and tumor stage to demographic life expectancy, assuming as “relative median survival index” the ratio between the two values.RESULTS:General contraindications to curative surgery increased significantly with age. The operative mortality rate was higher in Group D than in Groups A, B, plus C over the total series (P<0.001) and in both elective (P<0.001) and emergency surgery (P<0.05). Intergroup analysis of long‐term survival rates showed significant differences between “crude” (P=0.0057) but not age‐corrected (P=0.66) curves. The relative median survival index increased with age, up to approximately 1 in the local stages of Groups C and D.CONCLUSIONS:To evaluate long‐term results, elderly patients should be compared with unaffected, same‐age subjects. Because the risks may be very high, the surgical policy in the elderly should be carefully weighed and related to life expectancy and actual results.