Pancreatic abscess: An unresolved surgical problem

Pancreatic Abscess: An Unresolved Surgical Problem

Gerard V. Aranha, MD, FRCS(C), Maywood, Illinois Richard A. Prinz, MD, Maywood, Illinois Herbert B. Greenlee, MD, Maywood, Illinois

Pancreatic abscess is a rare but often lethal complication of acute pancreatitis. The story is all too familiar; in a patient who is recovering from an attack of acute pancreatitis, upper abdominal pain, fever, and leukocytosis develop, in addition to a failure to thrive. Often these symptoms and signs are thought to be due to recurrent pancreatitis, but they may be secondary to an abscess. It is critical to recognize which entity is present, because an unrecognized and undrained pancreatic abscess carries with it a high mortality [l-3]. This study was undertaken to determine if recent diagnostic and therapeutic advances have made an impact on our ability to identify the presence of a pancreatic abscess, initiate effective treatment, and thus improve the prognosis of these patients. Material

and Methods

One thousand ninety-two patients were diagnosed as having acute pancreatitis at the Veterans Administration Hines Hospital between the years 1970 and 1979. Twenty of these patients (1.8 percent) were diagnosed and treated for pancreatic abscess. In this study a pancreatic abscess was defined as an abscess arising in or contiguous with the substance of the pancreas. The patients were all men and their ages ranged from 28 to 70 years. The abscess developed as a complication of alcoholic pancreatitis in 10 patients, of abdominal surgery in 9, and of a perforating gastric carcinoma in 1 (Figure 1). Of the nine patients in whom pancreatic abscess developed after operation, six underwent procedures that involved the pancreas. These included two biliary tract procedures with pancreatic biopsy, three subtotal or near total resections of the pancreas, and one internal drainage of a pseudocyst. The three other procedures were bilateral nephrectomy with splenectomy, subtotal gastrectomy, and insertion of a peritoneal dialysis catheter. Most patients were symptomatic, but the pattern From the Surgical Service, Hines Veterans Administration Hospital, Hines, and the Department of Surgery, Loyola University Stritch School of Medicine, Maywood, Illinois. Requests for reprints should be addressed to Gerard V. Aranha. MD, Department of Surgery, Loyola University Stritch School of Medicine, 2160 South First Avenue, Maywood. Illinois 60153.

534

of presentation was quite variable (Figure 2). Abdominal pain was present in 17 patients (85 percent), back pain in 7 (35 percent), nausea and vomiting in 11 (55 percent), temperature greater than 37.8% in 18 (90 percent), weight loss in 5 (25 percent), melena in 2 (10 percent), and diarrhea in 1 (5 percent). Abdominal tenderness was present in 15 (75 percent), abdominal distention in 5 (25 percent), and abdominal mass in 4 (20 percent). Laboratory data and radiologic tests White blood cell counts ranged from 10,500 to 35,00O/mm” (mean 29,600/ mmS). The serum amylase level was measured in 18 patients but was elevated in only 7. In these seven patients the cause was alcoholic pancreatitis in three, previous surgery in three, and tumor in one. Hypoalbuminemia with a serum albumin level of less than 3.5 g/100 ml was seen in 15 (75 percent) of the patients, and liver enzymes were abnormal in 15 (75 percent) of the patients (Table I). Radiologic tests were useful in making the diagnosis in 80 percent of the patients. Plain roentgenographic examination showed an ileus in nine patients, but extraluminal gas bubbles were present in two of them, suggesting the presence of an abscess. Upper gastrointestinal series was positive in 8 of 11 patients which demonstrated anterior displacement or outlet obstruction of the stomach and effacement or extrinsic compression of the duodenum. Lower gastrointestinal series demonstrated a communication of the abscess cavity with the colon in two of the seven patients who had this study. Ultrasound examination in 14 patients demonstrated a fluid-filled cavity in 12. In two patients ultrasound was unsatisfactory for technical reasons. Computerized tomographic scanning in these instances revealed a cystic lesion which proved to be an abscess. Computerized tomographic scanning of the abdomen was positive in the three patients in whom it was done. In Figure 3 the accuracy of the various radiologic tests that were carried out is summarized. Results

All 20 patients underwent surgery. The average interval between onset of symptoms and surgery was 23 days. All patients were explored by a transperitoneal approach, and all except one were drained surgically using a sump catheter and Penrose drains or Penrose drains alone. The sump and Penrose

The American Journal of Surgery

Pancreatic

Abscess

-I

Figure 7. Etiology of pancreafitic abscess showing major causes fo be alcohol and previous surgery.

Figure 2. Presenting symptoms and physical signs in 20 patients with pancreafic abscess.

drains were brought out posteriolaterally to assure dependent drainage (Figure 4). If the anatomy did not permit, the sump was brought out anteriorly (Figure 5). Five patients needed additional procedures because of gastric (two patients) or colonic (three patients) complications. The gastric complications were bleeding and gastric outlet obstruction and were treated by vagotomy and pyloroplasty and vagotomy and gastroenterostomy, respectively. In three patients whose distal transverse colon and splenic flexures were thought to be involved in the abscess, a proximal colostomy was performed. Eight patients required reoperation because of continuing intraabdominal sepsis, and of this group, two died. Thus, reoperation to control sepsis is necessary and beneficial. Six of eight. patients survived due to persistent attempts to achieve adequate drainage. There were six postoperative deaths for a mortality rate of 30 percent. Five patients died from sequential organ failure secondary to sepsis, and one from massive upper gastrointestinal hemorrhage. Mortality was related to several factors. Only 13 percent (2 of 15 patients) of those patients drained by sump

catheter and Penrose drains died in the postoperative period compared with 75 percent (3 of 4 patients) of those drained by Penrose drains alone (Table II). Ten patients received hyperalimentation either by the enteral or the intravenous route, and in this group 2 (20 percent) died. Ten patients did not receive hyperalimentation, and 4 (40 percent) of these patients died. Postoperative death was also related to the cause of the abscess. There were two deaths in 10 patients with alcoholic pancreatitis (20 percent) and four deaths in 9 patients (44 percent) in whom the onset of pancreatitis followed other intraabdominal surgery. Bacteriologic cultures were positive in 17 patients and were polymicrobial in 14 and monomicrobial in 3. The most commonly cultured organisms were Enterococcus (la), Escherichia coli (9), Klebsiella pneumoniae (6), Pseudomonas aeruginosa (l), and Staphylococcus aureus (1). There were three instances of anaerobic infection, namely Bacteriodes fragilis in two patients and Clostridium perfringens in one. Postoperative complications: The average postoperative hospital stay for 14 surviving patients was 65 days. Postoperative morbidity was seen in 10 of the 14 (71 percent.). Only four patients (20 percent)

TABLE I

Laboratory

Data

Test

Result

Temperature (“C) Mean Range White blood cell count (mm3) Mean Range Amylase Normal Elevated Albumin