de Dombal F T, Matharu S S, Staniland J R, Wilson D H, MacAdam W A, Gunn A A, Allan W R, Bjerregaard B
Br J Surg. 1980 Jun;67(6):413-6. doi: 10.1002/bjs.1800670610.
In a survey of 5675 patients presenting to five hospitals in England, Scotland and Denmark with acute undiagnosed abdominal pain, 106 patients later proved to have intraabdominal cancer. The risk of cancer was age-dependent: amongst patients over 50 years with 'non-specific' pain the risk of cancer was 10 per cent. The commonest primary cancer site (in 57 cases, 53.8 per cent) was the large bowel. Most cancers neither perforated nor obstructed: 73 patients merely presented with a short history of unexplained abdominal pain. Of those patients with cancer presenting with 'unexplained' pain, half (37/73, 50.7 per cent) left hospital without a diagnosis of cancer having been made. Subsequently, a computer-aided system was constructed to discriminate (in patients over the age of 50) between those with unexplained acute abdominal pain who did and did not have cancer. Overall accuracy was 84.7 per cent in 138 cases. The most helpful clinical features in making this discrimination are listed. It is suggested (a) that cancer is now a relatively common cause of acute abdominal pain, (b) that the diagnosis is frequently difficult and (c) that urgent screening of all patients over 50 with non-specific acute abdominal pain may be warranted.
