EDITORIAL SYNOPSIS This paper confirms previous studies showing the increased risk of bleeding associated with blood group 0. There is an indication that this may also be true for acute perforation and indeed the relationship may be with the severity of the ulcer disease.In a previous paper (Horwich and Evans, 1966) a relationship was shown between the loss of gastric mucosal cells in 160 volunteer subjects and their ABO blood groups and ABH secretor status. Aspirin enhanced the loss from the gastric mucosa, and it was postulated that clinical overt bleeding from the gastric mucosa, whether aspirin-induced or not, may also be related to the ABO blood groups and secretor status. The present investigation concerns this problem, and it was also thought to be of value to extend the investigation to include subjects with duodenal ulcer because of the reported association with blood group 0 (Aird, Bentall, Mehigan, and Roberts, 1954) and with ABH non-secretion (Clarke, Edwards, Haddock, Howel-Evans, McConnell, and Sheppard, 1956). While the work was in progress, Langman and Doll (1965) reported an increased frequency of group 0 in gastric and duodenal ulcer subjects who had bled and an increased incidence of non-secretors in those who had undergone operation.
METHODSTwo separate series of patients were ascertained. Case records were examined for a history of salicylate ingestion during the 48 hours before the onset of bleeding and recorded at the time of admission. Details were abstracted of a barium meal examination carried out within two weeks after the bleeding episode. Where the radiologist gave a firm opinion on the presence or absence of a gastric or duodenal lesion, this diagnosis was accepted for the purpose of this investigation. Persistent duodenal deformity was accepted as being due to scarringfrom ulceration. Gastric ulcers have been divided into two categories. The first category is the classical lesser curve gastric ulcer to the left of the angulus with no other radiological lesion in the stomach or duodenum, and corresponds to type I of Johnson, Rogers, and Wyatt (1964). The second category, which corresponds to types II and III of Johnson et al. (1964), contains all gastric ulcers which areaccompanied by any abnormality of the duodenum, and all gastric ulcers to the right of the angulus. The two categories have been separated in this way because types II and III of Johnson et al. (1964) both seem to possess such features of duodenal ulcer as hypersecretion and association with blood group 0.Cases were included in this series if the barium meal was normal or showed a duodenal or gastric ulcer. Cases were excluded if there was any suspicion of bleeding from other sites, such as epistaxis, haemoptysis, oesophageal varices, hiatus hernia, carcinoma, or lesions of the large bowel. In order to be comparable with the experimental gastric-cell-shedding series, cases were excluded if there had been a previous gastric operation, except suture of a simple perforation, and for the same reason subjects of blood g...