Oesophageal perforation at fibreoptic gastroscopy

In a survey of 173 endoscopy units conducted through the British Society for Digestive Endoscopy (which merged with the British Society of Gastroenterology in 1980) 31 of the 101 units who replied reported serious complications (R Cockel, unpublished data). A supplementary questionnaire was sent to these centres requesting further details, and replies were received from 24. The table records the information concerning oesophageal perforation at oesophagogastroduodenoscopy and related procedures. Perforation at diagnostic oesophagogastroduodenoscopy was rare (one in 5474 examinations; seven cases altogether). These seven patients were relatively elderly (mean age 72 years) compared with others undergoing the procedure (mean age 53 years). Barium swallow had been performed in four of the seven patients before endoscopy. In five cases the perforations were high and associated with difficult intubation; in four of these cases intubation was performed by inexperienced operators. The remaining two patients had radiologically identified strictures (one peptic, one carcinoma). No particular type of fibrescope was associated with perforation and no obvious medical factors contributed. Perforation was recognised almost immediately in six patients, but one outpatient died two days later at home. Thoracotomy was performed in two patients, one of whom died. Perforation during oesophageal dilatation was more common (one in 109 procedures; 1 1 patients) and occurred despite previous radiological visualisation of the stricture in nine patients. Eight patients had peptic strictures, and perforation occurred at first dilatation in seven and at second dilatation in one. On three occasions perforation occurred during the first stage of a planned two-stage procedure, and defective dilators contributed in one case. Two oesophageal cancers, both misinterpreted as benign strictures, and one invading bronchial carcinoma were perforated. X-ray screening was used in only three of the 11 patients. Four patients died, two after thoracotomy, which was performed in only four patients. Heavy sedation was used in many of the patients with perforation. Perforation occurred at oesophageal dilatation in two additional patients not recorded in the table. One patient died after perforation of the stomach during dilatation of a peptic oesophageal stricture when the metal flexible finger end of the Eder-Puestow dilator became unwound. Another patient died from jejunal perforation after dilatation of a benign oesophagojejunal stricture. Perforation during palliative oesophageal intubation for carcinoma was common (one in 13 procedures; 18 cases). Barium swallow had been performed in 15 of the 18 patients before the procedure. Except for one bronchial carcinoma, all were lesions of the mid or lower third of the oesophagus. Eight of the strictures were so tight that the guidewire passed only with difficulty. Despite perforation intubation was successful in 16 of the patients. The remaining two died soon after the procedure. Screening was used in 13 of the patients. All were managed conservatively, cardiorespiratory problems contributing to death in five.