Twenty-five patients with traumatic diaphragmatic hernia discovered at least five months after injury are described, of whom 18 were male and seven female. All but one hernia occurred on the left side. Stab wounds were the etiological factor in 22 patients and blunt trauma in three. The diagnosis was most often made by a chest or abdominal radiograph, but barium ingestion confirmed the diagnosis in ten patients. Intercostal drainage of gastric contents provided the diagnosis in two patients. In all nine patients initially approached by a thoracotomy or a thoracoabdominal incision, the hernia was easily reduced and the defect repaired. Although reduction and repair were easily accomplished by the abdominal route in seven patients, this approach was unsatisfactory or inadequate in six others. The colon and stomach were usually in the chest, and strangulation occurred in five patients. The mortality was 20% but rose to 80% when gangrene was present.

Tuberculous infection of the oesophagus is rare. This is confirmed by our present review of cases managed in our teaching hospitals over a period of 18 years which uncovered only 11 patients. The main presentation is that of dysphagia whose algorithm of investigation should seek to differentiate tuberculosis from carcinoma, the more common cause of this symptom. Of the 11 patients, 9 presented with dysphagia while 2 had haemorrhage; 7 had an abnormal plain chest radiograph, of whom 4 had a mediastinal mass lesion (3 were lymphadenopathy and one an abscess). All but one had an abnormal radio-contrast oesophagogram, including a mediastinal sinus in two and a traction diverticulum in another two. The mainstay of investigation was oesophagoscopy through which diagnostic biopsy material was obtained in half of the patients. In the other half diagnosis was by either biopsy of associated mediastinal (3) or cervical (1) lymph node masses or by acid fast bacilli positive sputum (1). The diagnosis was established post-mortem in one patient. Treatment was primarily non-operative with standard anti-tuberculosis drug therapy. Two patients underwent a diagnostic thoracotomy and one a drainage of mediastinal abscess together with resection and repair of oesophago-mediastinal sinus during the early part of the series. Outcome of management was very rewarding in 9 patients and death occurred in 2 patients, one of whom had his anti-tuberculosis drug therapy interrupted by severe hepatitis B virus infection. The other death occurred in a patient whose haemorrhage from an aorta-oesophageal fistula was not established ante-mortem. It is recommended that when biopsy material of the oesophagus is unobtainable or non-diagnostic in patients with dysphagia, especially with an abnormal chest radiograph or human immunodeficiency virus infection, effort should be made to obtain biopsy material from associated lymph nodes, even by thoracotomy if necessary, or culture of biopsy from the radiologically abnormal part oesophagus and sputum for mycobacteria, in order to establish the diagnosis of this rare but eminently treatable cause of dysphagia. Clinicians should be aware of tuberculosis of the oesophagus as a possible cause of haematemesis in patients with otherwise unexplained upper gastrointestinal haemorrhage.

Massive haemorrhage from a pancreatic pseudocyst is an uncommon cause of upper gastrointestinal bleeding. The condition carries a high mortality and presents a major clinical problem both in terms of timely diagnosis and appropriate surgical therapy. Four patients are presented with pseudocyst-related bleeding arising from the gastroduodenal artery. In all instances the bleeding was successfully controlled by transcatheter embolization with gelfoam. No untoward sequelae were noted and the occlusion remained effective as shown by follow-up angiography 2 months later. The importance of early diagnostic angiography is stressed and it is concluded that interventional angiography may obviate the need for high risk emergency surgery.

One hundred and eighty-one patients with carcinoma of the upper thoracic oesophagus were intubated perorally using a Procter-Livingstone tube. The mortality was 16-6 per cent but, in the patients who survived, the palliation achieved, as judged by improved swallowing, was considered satisfactory. Factors influencing the success of intubation are also considered.

Malignant esophago-respiratory fistula is an incurable condition calsing severe distress to those patients suffering from it. Sixty two successive patients with esophago-respiratory fistulae, secondary to squamous cell carcinomas of the esophagus, were intubated for palliation. The first 14 were intubated by a traction technique using the Celestin tube, and the remaining 48 were intubated by the Procter-Livingstone tube inserted by a pulsion method. Celestin intubation had a mortality of 64.3% but 75% of those intubated by the Procter-Livingstone tube were discharged from hospital, swallowing satisfactorily and relieved of their respiratory distress. In our experience the insertion of the Proctor-Livingstone tube by a pulsion technique, can be performed in a few minutes, with little morbidity and allows the patient to be discharged from hospital within 3-4 days of the procedure. It is a successful and acceptable method of treating malignant esophago-respiratory fistulae.