Prevention, diagnosis, and treatment of pancreatic abscess.

Since pancreatic abscesses are a highly lethal complication of acute pancreatitis, factors influencing the genesis of major peripancreatic infection (MPI) were analyzed in 330 patients with pancreatitis. MPI developed in 28 (8.5%). Management of MPI was reviewed in 35 patients, including seven admitted with MPI. Etiology and severity of pancreatitis influenced MPI frequency. MPI was common in postoperative pancreatitis (39%), compared with alcoholic (6.6%), biliary (3.6%), or other causes (15%). "Severity" of pancreatitis was estimated by 11 early prognostic signs, which were reported previously. With fewer than three signs, MPI developed in 2.7%, three to five signs, 32%; more than five signs, 50%. Treatment of pancreatitis also influenced MPI. Early laparotomy increased MPI incidence from 1.6% to 23% in mild pancreatitis and from 24% to 67% in severe pancreatitis. Early oral feeding also appeared to predispose to MPI. Prolonged nasogastric suction and avoidance of early operation reduced MPI incidence from 16% in the first 100 patients to 5% in the next 230 patients. Outcome of MPI reflected severity of underlying pancreatitis. Mortality with fewer than three signs was 14%; with three to five signs, 65%; with more than six signs, 100%. Mortality was only 26% in 19 patients treated with radical sump drainage of the entire peripancreatic retroperitoneum, compared to 75% of 16 patients treated with conventional local drainage. In summary, MPI is related to etiology and severity of pancreatitis as quantified by early signs. Early laparotomy for pancreatitis increased MPI. Treatment of MPI with wide sump drainage appears more effective than local drainage.

Clinically significant ureteral obstruction caused by inflammatory complications of severe pancreatitis.

Clinically significant ureteral obstruction caused by the inflammatory complications of severe pancreatitis is rare with only eight previously reported cases. We present two additional cases and review the world literature. Clinically significant ureteral obstruction can affect either or both ureters and present simultaneously with an episode of pancreatitis or months later. If symptomatic ureteral obstruction is present, prompt urologic drainage is recommended. Definitive correction of the obstruction is frequently required and depends on the obstructive mechanism.

The relationship of coagulation factors to clinical complications of acute pancreatitis.

Alterations in coagulation factors have been reported during acute pancreatitis. Therefore the relationship of coagulation measurements to complications of pancreatitis was evaluated prospectively in 35 patients in whom 130 serial coagulation profiles were performed, consisting of fibrinogen, platelet count (PC), fibrinogen-fibrin-related-antigen (FR-antigen), prothrombin time (PT), partial thromboplastin time, thrombin time, euglobulin clot lysis, and Factors II, V and VII-X levels. During attacks of acute pancreatitis, over-all mean initial fibrinogen and PC of 268 mg. per 100 ml. and 214,000 per cubic millimiter rose significantly (p less than 0.005) to peaks of 362 mg. per 100 ml. and 477,800 per cubic millimeter by day 6 to 10. Mean initial FR-antigen of 4.8 microgram per milliliter rose to peak 7.4 microgram per milliliter on day 5. In 21 patients with mild pancreatitis, mean highest fibrinogen, PC, FR-antigen, and PT were 329 mg. per 100 ml., 361,500 per cubic millimeter, 5.3 microng per milliliter and 14.1 seconds. These values were significantly higher (p less than 0.05 to 0.01) in patients with severe pancreatitis, being 422 mg. per 100 ml. 528,000 per cubic millimeter, 13.7 microng per milliliter, and 15.5 seconds, respectively. Evaluation of the relationship of coagulation measurements to early clinical features showd that mean initial fibrinogen levels were significantly higher (p less than 0.05 to 0.01) in patients with initial amylase greater than 1,000 Somogyi units percent, serum glutamic oxaloacetic transaminase (SGOT) greater than 250 S.F.U. percent, and initial 72 hour PAO2 less than 75 mm. Hg. Early hypoxemia also correlated (p less than 0.05) with elevated initial FR-antigen levels. Impaired early renal function correlated (p less than 0.01) with elevated initial PC only. Early hypocalcemia did not correlate with coagulation measurements. These findings demonstrate that marked changes in coagulation parameters occur during acute pancreatitis and are related to over-all morbidity. Correlation of early coagulation measurements with amylase levels and with respiratory, renal, and hepatic dysfunction suggests that enzyme-related intravascular coagulation may be implicated in the pathogenesis of these complications of pancreatitis.

Esophageal transection and paraesophagogastric devascularization for bleeding esophageal varices.

The Sugiura procedure is feasible in an unselected, high risk population of alcoholic patients with cirrhosis who have bleeding esophageal varices and poor hepatic reserve. The Sugiura procedure controlled variceal bleeding in every patient with active bleeding and prevented early rebleeding, however, the operation is tedious, time-consuming, and has a high complication rate related to the thoracic approach. The rate of anastomotic leakage of 8.6 percent (4.8 percent in elective cases) is not as high as might be anticipated, but led to death in every case. The long-term outlook for these patients is poor, and the rebleeding rate of 37 percent in our lowest risk patients is disappointingly high. Similar results can be achieved with simpler procedures.

Resection of the perforated segment. A significant advance in treatment of diverticulitis with free perforation or abscess.

As a result of improved medical management of chronic diverticular disease, perforation has become the most common indication for surgical intervention. During the past five years sixty-three patients underwent operation for colonic diverticular disease, of which forty-six were for perforation (generalized peritonitis in 8, abscess in 30, and fistula in 8). The eight patients with generalized peritonitis underwent emergency exploration for spreading peritoneal signs and were managed by resection of the perforated segment, end colostomy, and mucous fistula or Hartmann's pouch. Treatment of thirty-eight patients with abscess or fistula has also stressed primary resection of the perforated segment of colon. Resection and end colostomy without anastomosis was performed in three. Primary anastomosis with proximal diverting colostomy was performed in four. Primary anastomosis alone was done in thirty-one patients. There were no deaths. These results support primary resection of the involved colon with immediate or delayed anastomosis in the operative management of perforated diverticular disease.

Parenteral and oral antibiotics in elective colon and rectal surgery. A prospective, randomized trial.

Our evaluation consisted of a prospective, randomized clinical trial in a homogenous group of 241 patients undergoing elective colon and rectal resections. A significant decrease in wound infection was found in the patients who received intravenous cefoxitin in conjunction with standard bowel preparation. The infection rate correlated with the type of resection; rectal resections had the highest rate in each study group, but parenteral prophylaxis produced a significantly lower wound infection rate. E. coli and Staph. aureus were the most common bacterial isolates in both groups. B. fragilis was recovered in only two Group A patients, which most likely reflects the exceedingly low recovery rate of anaerobic bacteria in our laboratory. Urinary cultures were positive in a large number of patients and reflect the standard use of Foley catheterization in all patients who undergo resection of the colon or rectum. These data indicate that perioperative prophylactic administration of cefoxitin reduces the wound sepsis rate when combined with oral antibiotics and mechanical bowel preparation in patients undergoing resection of the colon or rectum.

Management of diverticulitis of the ascending colon. 10 years' experience.

Diverticulitis of the ascending colon is an uncommon disease which mimics appendicitis. The correct diagnosis is rarely made, but can be suggested by the patterns of signs and symptoms and confirmed by barium contrast study. Diverticulitis of the ascending colon should be treated by the same plan as diverticulitis of the left colon. If the diagnosis is established, nonoperative management is indicated initially. Operation is indicated when the diagnosis is in doubt, when perforation has occurred, or when the patient does not respond to nonoperative treatment. At operation, ascending colon diverticulitis can be recognized as an inflammatory mass involving the wall and mesentery of the colon. The inflammatory mass is best treated by resection with primary anastomosis of the ileum to the ascending or transverse colon in an area removed from the site of infection.

Perforated diverticula of the jejunum and lleum.

Over a ten year period, four patients with inflammation or perforation of non-Meckelian, small intestinal diverticula were treated on the surgical services of Bellevue Hospital. This entity remains uncommon but may be increasing in incidence. The patients presented with a short history of severe abdominal pain, usually accompanied by nausea and vomiting. Each patient also gave a longer preceding history of less well defined abdominal symptoms. The pathogenesis of the small intestinal diverticula is uncertain but may be related to disturbed muscular peristalsis in the small bowel analogous to the changes implicated in esophageal and colonic diverticular disease. The diverticulum may be difficult to demonstrate at operation, and careful exploration for this possibility should be carried out at the time of operation for peritonitis of obscure origin. Segmental resection and end-to-end anastomosis is the treatment of choice.

Acute pancreatitis: pathogenesis, outcome and treatment.

The aetiological associations and proposed pathogenesis of acute pancreatitis have been reviewed. Although 80 per cent of patients have underlying alcohol abuse or gallstones, the precise mechanism of induction and of progression of pancreatic injury remains uncertain. Our current approach to management is summarized in Table 8. At present, no measure designed to limit the severity of pancreatitis or to interrupt the genesis of complications has been of proven benefit. Treatment is therefore primarily supportive. Peritoneal lavage by catheters introduced under local anaesthesia appears to be a valuable adjunct to the treatment of the early cardiovascular and respiratory complications of severe pancreatitis and we continue to recommend this measure. The major unsolved problem in treatment of this disease is the prevention and treatment of infected peripancreatic abscesses.

Diagnostic standards for acute pancreatitis.

An accurate history and thorough physical examination will often raise clinical suspicion of acute pancreatitis in the differential diagnosis of a patient presenting with acute abdominal pain. An accurate diagnosis is needed to eliminate etiologies of acute abdominal pain and to appropriately direct therapy. Confirmation of the diagnosis is most often made by evaluation of serum amylase and lipase levels. Although hyperamylasemia is found in the majority of patients with acute pancreatitis, other nonpancreatic acute abdominal conditions may be present with hyperamylasemia. CT scanning provides an accurate confirmation of clinical and laboratory findings and offers excellent anatomic and morphologic representation of the pancreas and peripancreatic tissue. The following article, written by the late John H.C. Ranson, presents a discussion of the modalities available for diagnosing acute pancreatitis.

The timing of biliary surgery in acute pancreatitis.

The timing of biliary surgery remains controversial in patients with acute pancreatitis associated with cholelithiasis. Eighty hospital admissions for acute pancreatitis, occurring in 74 patients with cholelithiasis, have therefore been reviewed. Among 22 patients who underwent abdominal surgery during the first week of treatment, there were five deaths (23%) and four patients (18%) who required more than seven days of intensive care. Fifty-eight episodes of pancreatitis were managed nonoperatively during the first week of treatment, with no deaths, although six (10%) required more than seven days of intensive care. Biliary surgery was undertaken later during the same admission in 37 patients, with no deaths. Twenty-one patients were discharged without biliary operation, but seven (33%) developed further pancreatitis. Previously reported prognostic signs were used to divide pancreatitis into 57 "mild" episodes (1.8% mortality) and 23 "severe" episodes (17% mortality). Early (day 0-7) definitive biliary surgery was undertaken in 11 patients with "mild" pancreatitis, with one death (9%), and in six patients with "severe" pancreatitis, with four deaths (67%). In three recent patients with "severe" pancreatitis, early biliary surgery was limited to cholecystostomy, with no deaths. These findings suggest that although early correction of associated biliary disease may be undertaken safely in many patients with "mild" acute pancreatitis, early definitive surgery is hazardous in "severe" pancreatitis and should, if possible, be deferred until pancreatitis has subsided. In most patients biliary surgery should precede hospital discharge.

The role of surgery in the management of acute pancreatitis.

Surgical intervention in acute pancreatitis may have varied goals. Early laparotomy may be required for diagnostic purposes. There is, however, no convincing evidence that attempts to reduce the morbidity of severe pancreatitis by early operative pancreatic drainage, early formal pancreatic resection, or early biliary procedures have been effective. In fact, they may be harmful. Peritoneal lavage by catheter induced under local anesthesia may ameliorate early cardiovascular and respiratory complications in some patients. Preliminary experience suggests that early operative debridement of devitalized pancreatic tissue with postoperative lavage may be helpful in selected patients. Patients with infections of devitalized pancreatic or peripancreatic tissue require operative debridement and drainage or packing. Other complications such as colonic necrosis or pseudocysts also require operative treatment. Rarely do patients require operation to relieve protracted pancreatitis. Patients with gallstone-associated pancreatitis should usually undergo surgical correction of their cholelithiasis as soon as their pancreatitis has subsided.

Bile duct disruption by blunt trauma.

The rarity of bile duct injury secondary to blunt abdominal trauma leads to frequent delays in diagnosis and inappropriate management. An illustrative case is therefore described and 94 reported cases are reviewed. In 53% of patients, operation was delayed more than 24 hours. Early clinical findings of hypovolemia and acute abdomen are related to associated injuries. Late findings are abdominal distention and jaundice due to the biliary injury. Early diagnosis is facilitated by diagnostic paracentesis. Patients operated on during the first 24 hours after the injury had a statistically higher incidence of bile duct injury distal to the cystic duct (p less than 0.05) and of complete ductal severance (p less than 0.05). The association of location distal to the cystic duct and complete severance was highly significant (p less than 0.001). Management should include biliary exploration. Cholangiography using concentrated water-soluble contrast agents may help to find the anatomy of obscure injuries. The choice of surgical repair must be individualized according to the location and the magnitude of the injury.

Nutritional support for acute pancreatitis.

The current review has summarized current data relevant to the nutritional support of patients with acute pancreatitis. Selection of the most appropriate form of nutritional support for patients with acute pancreatitis is intimately linked to a thorough understanding of the effects of various forms of enteral and parenteral nutrition on physiologic exocrine secretory mechanisms. Two basic concepts have emerged from the multiple studies that have addressed these issues to date: 1, enteral feeds should have low fat composition and be delivered distal to the ligament of Treitz to minimize exocrine pancreatic secretion and 2, parenteral substrate infusions, alone or in combinations similar to those administered during TPN, do not stimulate exocrine pancreatic secretion. From a practical standpoint, most patients with acute pancreatitis are diagnosed by nonoperative means and will manifest some degree of paralytic ileus during the early phase of the disease. Therefore, jejunal feeds are usually not a therapeutic option early in the course of this disease. On the basis of the clinical studies reviewed herein we propose general guidelines for the nutritional support of patients with acute pancreatitis: 1, most patients with mild uncomplicated pancreatitis (one to two prognostic signs) do not benefit from nutritional support; 2, nutritional support should begin early in the course of patients with moderate to severe disease (as soon as hemodynamic and cardiorespiratory stability permit); 3, initial nutritional support should be through the parenteral route and include fat emulsion in amounts sufficient to prevent essential fatty acid deficiency (no objective data exist to recommend specific amino acid formulations); 4, patients requiring operation for diagnosis or complications of the disease should have a feeding jejunostomy placed at the time of operation for subsequent enteral nutrition using a low fat formula, such as Precision HN (Sandoz, 1.3 percent calories as fat), Criticare HN (Mead Johnson, 3 percent calories as fat) or Vivonex High Nitrogen (Norwich Eaton, 0.87 percent calories as fat), and 5, oral feedings should be low fat in composition and should be reinstituted using traditional clinical criteria, including the symptoms of the patient, physical examination and computed tomographic appearance of the pancreas (clinicians should bear in mind the well documented exocrine stimulatory effects of even low fat oral feeds and the risks of early refeeding). These general guidelines must be individualized to incorporate what is perhaps the most important clinical variable--the premorbid nutritional state of the patient.(ABSTRACT TRUNCATED AT 400 WORDS)

Long peritoneal lavage decreases pancreatic sepsis in acute pancreatitis.

Late infection of devitalized pancreatic and peripancreatic tissue has become the major cause of morbidity in severe acute pancreatitis. Previous experience found that peritoneal lavage for periods of 48 to 96 hours may reduce early systemic complications but did not decrease late pancreatic sepsis. A fortunate observation led to the present study of the influence of a longer period of lavage on late sepsis. Twenty-nine patients receiving primary nonoperative treatment for severe acute pancreatitis (three or more positive prognostic signs) were randomly assigned to short peritoneal lavage (SPL) for 2 days (15 patients) or to long peritoneal lavage (LPL) for 7 days (14 patients). Positive prognostic signs averaged 5 in both groups but the frequency of five or more signs was higher in LPL (71%) than in SPL (47%). Eleven patients in each group had early computed tomographic (CT) scans. Peripancreatic fluid collections were shown more commonly in LPL (82%) than in SPL (54%) patients. Longer lavage dramatically reduced the frequency of both pancreatic sepsis (22% LPL versus 40% SPL) and death from sepsis (0% LPL versus 20% SPL). Among patients with fluid collections on early CT scan, LPL led to a more marked reduction in both pancreatic sepsis (33% LPL versus 83% SPL) and death from sepsis (0% LPL versus 33% SPL). The differences were even more striking among 17 patients with five or more positive prognostic signs. In this group the incidence of pancreatic sepsis was 30% LPL versus 57% SPL and of death from sepsis 0% (LPL) versus 43% (SPL) (p = 0.05). In these patients, overall mortality was also reduced (20% LPL versus 43% SPL). When 20 patients treated by LPL were compared with 91 other patients with three or more positive prognostic signs who were treated without lavage or by lavage for periods of 2 to 4 days, the frequency of death from pancreatic sepsis was reduced from 13% to 5%. In those with five or more signs, the incidence of sepsis was reduced from 40% to 27% (p = 0.03) and of death for sepsis from 30% to 7% (p = 0.08). These findings indicate that lavage of the peritoneal cavity for 7 days may significantly reduce both the frequency and mortality rate of pancreatic sepsis in severe acute pancreatitis.