Aspirates were obtained by introducing a 9 cm 22 gauge spinal needle connected to a 20 ml syringe into the area of the stricture, after accurate localisation, and applying suction while oscillating the needle for a few millimetres in its long axis within the lesion. At least two passes were made on each patient. Specimens were handled immediately by a cytology technician. Fine needle aspiration specimens were expressed onto glass slides and bile and cytology brushings were smeared directly on to slides. The slides were either fixed immediately in 95% alcohol solution and later stained by the Papanicolaou method or air dried and stained with the MayGrunwald-Giemsa stain. The syringe was then rinsed with Eagle's medium, centrifuged, and cytospin preparations made. Some smears were stained with the periodic acid-Schiff reagent with and without diastase digestion to identify mucin production. The slides were examined by two trained pathologists and coded as unsatisfactory (acellular), negative, highly suspicious for malignancy, or definitely positive.

Intraoperative enteroscopy was performed in 12 patients (median age 68 years) with obscure gastrointestinal bleeding probably of small bowel origin, six of whom were men. All the patients were evaluated by routine haematological, coagulation and biochemical profiles, upper and lower gastrointestinal endoscopies, visceral angiography and/or isotope scanning. All the patients were anaemic. Visceral angiography was useful on three of the 12 occasions on which it was used and isotope scanning was valuable on eight of the 11 occasions it was used. Nine patients had undergone previous laparotomy. Enteroscopy was performed successfully in all cases, with fresh blood and discrete vascular lesions being the chief findings (10 of 12 cases). Segmental resections (n = 8) and local resections (n = 2) were performed in ten patients, with two patients having more than one laparotomy for rebleeding. Five patients developed postoperative complications and there was an operative death and one late death. Three of the ten surviving patients experienced further rebleeding. Intraoperative enteroscopy is now an essential adjunct to laparotomy for gastrointestinal bleeding which has been localized to the small bowel before operation.

We reviewed 45 patients who underwent surgery for primary jejunoileal tumors over a 15-year period. There were 16 benign and 29 malignant tumors, which included 13 lymphomas, 7 adenocarcinomas, 7 carcinoid tumors, and 2 leiomyosarcomas. Eighteen patients, 13 of whom had benign tumors, presented with intestinal bleeding and 5 tumors were found incidentally at laparotomy. Benign lesions were more frequently sited in the jejunum while malignant lesions were more common in the ileum (p less than 0.001). Lesions presenting with hemorrhage were more likely to be benign than malignant (p less than 0.001) and were more commonly sited in the jejunum than in the ileum (p less than 0.05). Visceral perforation (31%), intestinal obstruction (21%), and an abdominal mass (17%) were other presenting features in patients with malignant tumors. In spite of a wide variety of investigations, the correct diagnosis was reached preoperatively in only 31% of patients. Surgical management included either limited bowel resection or segmental resection with regional lymphadenectomy. Operative mortality was 13% and morbidity was 36%. Actuarial 5-year survival for all malignant tumors was 24%, being 64% at 5 years for carcinoid tumors, 20% at 30 months for adenocarcinomas, and 10% at 42 months for lymphomas. These results reemphasize the need for a high index of suspicion and early laparotomy in patients with obscure intestinal symptoms if the prognosis of small bowel tumors is to improve.

Colonic pseudo-obstruction may have many possible causes. Some of these are well described and pose no diagnostic problems. Drug-related colonic pseudo-obstruction remains underreported, but is of importance in modern society where drugs are endemically abused. This case highlights the importance of drugs in altering colonic motility and emphasizes the nonsurgical management of this condition.

The management of 120 consecutive patients with carcinoma of the oesophagus, treated in a district general hospital over a 10-year period is reviewed. The treatment options were by endoscopic methods, radiotherapy, surgery or a combination of these. Adenocarcinoma was the most frequent histological type (51 per cent) and 41 per cent of patients had squamous carcinoma. In addition to survival the quality of swallowing and the incidence of late complications following the treatment options was assessed. Primary surgery was carried out in 21 cases (operability rate 17.5 per cent) and 81 per cent of these cases were resectable. The overall operative mortality rate was 14.3 per cent (6 per cent in resected cases) and the 5-year survival rate was 9.5 per cent. After primary radiotherapy only 5 per cent of patients survived 2 years. The majority of patients were treated by endoscopic intubation with an overall mortality of 16.6 per cent and a mean survival of 5.5 months. Patients treated surgically experienced the best symptomatic relief and had the lowest incidence of late complications, when compared with those treated differently. However the overall results of surgery were poor and there remains the need to compare the results of surgery and radiotherapy in similar groups of patients.