PubMed1993
Ninety patients with esophageal perforations were operated on at our institutions between 1970 and 1992. Thirty-four of them were seen after delayed diagnosis (> 24 hours) with mediastinal sepsis caused by perforation of the thoracic esophagus. There were 18 patients with spontaneous ruptures, 11 with instrumental perforations (including one caused during laparotomy), and 3 perforations caused by foreign bodies. One patient had perforation of an esophageal ulcer into the pericardium and another had perforation of an esophageal diverticulum into the mediastinum. Nineteen patients underwent primary repair of the perforation with cleansing and drainage of the mediastinum and the pleural cavity. The remaining 15 had primary extirpation of the thoracic esophagus, irrigation of the mediastinum with antibiotics, cervical esophagostomy, gastrostomy, and drainage of the mediastinum and pleural cavity. Nineteen of the 34 patients survived (hospital mortality 44%). Of patients with primary repair, only six survived (in-hospital mortality 68%), whereas only two patients treated with esophagectomy died (in-hospital mortality 13%). The difference was highly significant (p = 0.001). The most common cause of death was multiorgan failure resulting from sepsis. Postoperative complications developed in four patients treated with primary repair (two sepsis, one empyema, and one anuria) and in seven patients treated with esophagectomy (two empyema, two sepsis, one pneumonia, one mediastinal abscess, and one brain abscess). After healing of the mediastinitis, the esophagogastric continuity was reconstructed with colon in 11 patients and stomach in two patients. In the management of delayed esophageal perforation with mediastinal sepsis, esophagectomy is superior to primary repair alone, which often leads to mediastinal leakage, continued sepsis, and death.