Unusual cases and technical notes E267 Umgelter A et al. Ascending pneumonia after endotherapy of pancreatic abscess
failure caused by necrotizing pancreatitis was discharged from the intensive care unit after a 6−week stay. Two weeks later he developed a spiking fever, abdominal tenderness, and an abdominal mass. Computed tomography revealed a large pancreatic abscess that was compressing the duodenum and gastric wall (l" Fig. 1). Because of the interposition of bowel, transcutaneous drainage was risky and an endoscopic approach was chosen. After transmural duodenocystic puncture and dilation, a 7−Fr double−pigtail stent and a nasocystic tube were inserted (l" Fig. 2). Two days after the cystogas− trostomy the patient developed dyspnea and hypoxia, became drowsy, and had to be intubated and mechanically ventilat− ed. Chest radiography revealed patchy in− filtrates in the lower lobe of the right lung and bronchoalveolar lavage was per− formed. Cultures grew Klebsiella oxytoca, Enterococcus faecium, and Candida glab− rata, and these organisms were also grown from cultures of the aspirate from the pancreatic abscess. The patient died of septic multiorgan failure after a long period of intensive−care treatment. Whereas early mortality in severe acute pancreatitis is due to the systemic inflam− matory response syndrome, late mortal− ity of pancreatitis is determined by the development of secondary infection of necrotic tissue or pseudocysts . Infec− tion of pseudocysts leads to pancreatic abscess, a common complication of se− vere acute pancreatitis with a mortality of around 9 % . Several case series have proposed an en− doscopic approach for the treatment of both infected pancreatic necrosis and perigastric pancreatic abscesses. Success rates of 90 % for complete resolution of in− fected necrosis or abscess have been re− ported [3, 4], although hemorrhage and perforation have been described as com− plications of these interventions . The causal relationship linking a nosoco− mial pneumonia with an endoscopic pro− cedure performed several days before can be difficult to confirm. In this case, how− ever, the microbiological results for the cystic aspirate and the bronchoalveolar lavage effluent were identical, and the close temporal relationship between the endoscopic intervention and the pneu− monia suggests that the pneumonia must be regarded as a complication of that intervention.