Training surgeons in endoscopic retrograde cholangiopancreatography.

BACKGROUND: General surgeons commonly perform upper gastrointestinal endoscopy in practice, but few perform endoscopic retrograde cholangiopancreatography (ERCP), partly because of limited training opportunities. This report focuses on the value of an ERCP fellowship training program to a broad-based, mature residency in surgery and our observations on the experience required for surgeons to be trained in advanced interventional ERCP. METHODS: Since the program was initiated in 1992, 13 ERCP fellows have been trained for individual periods of 6 to 14 months. This study investigated all procedures with fellow involvement (2,008 cases) from among a total experience of 3,641 ERCPs. Data collected included type of ERCP (diagnostic/therapeutic), fellow success in cannulating the duct of interest, and faculty success in cases of fellows who failed. Of the 13 fellows, 9 had previous endoscopy experience, but none had training in ERCP. RESULTS: An 85% cannulation rate was accepted as successful, and cannulation rates for each fellow were calculated for each 3-month period. The 85% mark was reached by 4 (31%) of 13 fellows in the first period, 2 of 13 fellows (15%) in the second period, 5 of 11 fellows (45%) in the third period, 7 of 10 fellows (70%) in the fourth period, and 1 of 1 fellow (100%) in the fifth period of training. On the average, it took 7.1 months and 102 ERCPs for trainees to reach desired success levels. Success came more promptly with prior exposure to endoscopy. Fellows without prior endoscopic experience required 148 cases to reach 85% success. Resident surgical experience with major pancreatic resections increased threefold after establishment of the fellowship. CONCLUSIONS: Training in ERCP is possible within the scope of a surgical fellowship in a reasonable length of time and experience. Complication rates remain low even with fellow involvement. Establishment of an ERCP program increases the focus and experience of pancreas surgery in a surgical residency for chief residents.

Role of pancreatic duct stenting in the treatment of chronic pancreatitis.

BACKGROUND: Endoscopic retrograde cholangiopancreatography and stent placement are relatively new alternatives to surgery for the treatment of chronic pancreatitis. The objective of this study was to determine the efficacy of pancreatic duct stent placement for the treatment of chronic pancreatitis. METHODS: This study included 89 patients treated with pancreatic stents between 1993 and 2002. The patients were contacted via telephone for a personal interview with regard to pain, medication usage, weight loss or gain, and eating patterns. Additionally, medication usage before and after treatment was documented from the Kentucky Cabinet for Health Services' electronic reporting system for narcotic use. RESULTS: Of the 89 patients, 9 were deceased, 5 either refused to interview or could not be contacted, and 75 were interviewed. Significant weight gain exceeding 15 lb after treatment was experienced by 22%, whereas only 4% lost weight. A majority of the patients (68%) noted that they had less severe relapses or no relapses after treatment. The patients reported a decrease in pain level on a 10-point scale from 8.7 to 4.1 (53% decrease) after treatment. A decrease in pain medication usage was reported by 47% of the patients, and 83% considered their treatment successful. The Kentucky All Schedule Prescription Electronic Report (KASPER) was obtained before and after treatment for 55 patients. According to this statewide electronic reporting system, 63% had a documented decrease in narcotic use. CONCLUSION: The findings of this study support the use of pancreatic duct stenting as an option before surgical intervention for these difficult-to-manage patients with chronic pancreatitis.

Laparoscopic choledochojejunostomy and gastrojejunostomy in a porcine model.

BACKGROUND: Surgical extirpation remains the only known curative treatment for cancer of the pancreas. Because of locally advanced or metastatic tumor, up to 80% of patients are unresectable at the time of initial diagnosis [13]. Other investigators previously have suggested that laparoscopy before laparotomy aids in the diagnosis of unresectable pancreatic cancer in a fair number of patients even after negative computed tomography scans [3, 17]. Many surgeons are reluctant to incorporate laparoscopy into the workup of patients with cancer of the pancreas because of the frequent need for surgical bypass in the management of either biliary tract obstruction or gastric outlet obstruction [9, 13]. Previous studies have demonstrated the feasibility of laparoscopic cholecystojejunostomy combined with gastrojejunostomy in a porcine model, as well as the individual accomplishment of laparoscopic choledochojejunostomy. The purpose of this study was to document the feasibility of performing laparoscopic choledochojejunostomy with gastrojejunostomy. METHODS: Under general anesthesia, seven pigs underwent laparoscopic choledochojejunostomy and gastrojejunostomy using an intracorporeal hand-sutured technique. RESULTS: The mean operating time ranged from 150 to 450 min. All the animals recovered completely from the operation and had patent anastomoses at the time of necropsy. One pig died of gastric bleeding on postoperative day 13, and two animals had intraabdominal fluid collections discovered at the time of necropsy. CONCLUSIONS: These results suggest that synchronous laparoscopic bypass of biliary and gastric outlet obstruction is feasible, and can be performed in a manner similar to that used in open operations. We believe this lends support to the argument promoting laparoscopy in the evaluation of pancreatic cancer.

Cost comparison of endoscopic stenting vs surgical treatment for unresectable cholangiocarcinoma.

BACKGROUND: Total lifetime costs of endoscopic vs surgical treatment for obstructive jaundice due to cholangiocarcinoma are difficult to assess. The purpose of this study was to compare total costs in these two groups, including all treatment and retreatments. METHODS: This retrospective study identified patients with biopsy-proven cholangiocarcinoma treated this decade with either endoscopic biliary stenting or surgical biliary-enteric bypass with or without resection. Outcomes and hospital charges were recorded. Ten matched control patients were compared from each group. Costs included those for cost of repeat endoscopy in the endoscopic group and for the management of recurrent obstructive jaundice in the surgical group. RESULTS: The groups were similar in age and gender ratio. The surgical approach was frequently for cure; therefore, surgical patients were by and large at an earlier stage of their disease than those in the endoscopic group. Mean survival for the endoscopic group was 19 months vs 16.5 months for the surgical group. The median total lifetime cost for surgical therapy was $60,986 vs $24,251 for endoscopic therapy. CONCLUSION: Endoscopic therapy is an effective palliative therapy for unresectable cholangiocarcinoma. Endoscopic therapy for this entity costs significantly less than surgical treatment. The cost benefit of endoscopic stenting over standard surgical bypass in the management of patients with unresectable cholangiocarcinoma, when considered along with its minimally invasive approach, makes this the procedure of choice for palliative therapy.

Ventral hernia repair by the laparoscopic approach.

An analysis of these results indicates that laparoscopic hernia repair can be performed safely by experienced laparoscopic surgeons, and with lower perioperative complication rates than for open hernia repair. Although the follow-up period for the laparoscopic repair is only 2 or 3 years, the recurrence rate is likely lower than with open repair. Most patients with ventral hernias are candidates for this laparoscopic repair if safe access and trocar placement can be obtained. The choice of mesh often provokes a debate among surgeons, but little practical difference in the results seems to exist between the two types of mesh available. Although the ePTFE mesh has a good theoretic basis for promoting tissue ingrowth on the parietal side of the mesh and minimizing adhesions to the bowel side of the mesh, data indicate that no difference in outcome exists related to adhesions or fistula formation (Tables 1 and 2), so surgeon preference and cost of the prosthesis should be the deciding variables. Fistulas are of concern because of the experience with mesh in the trauma patient and in the treatment of severe abdominal wall infections, when abdominal wall reconstruction often is performed in contaminated wounds in the acute phases and leaves the mesh exposed without soft tissue coverage. These conditions do not apply for most cases of elective hernia repair. Laparoscopic ventral hernia repair offers advantages over the conventional open mesh repair and may decrease the hernia recurrence rate to 10% to 15%. When properly performed, the laparoscopic approach does not and should not compromise the principles for successful mesh repair of ventral hernias.

Bile duct carcinoma: trends in treatment in the nineties.

Surgical resection provides the only known chance of cure for cholangiocarcinoma, and even then the 5-year survival is only 10 to 20%, and only one-third of patients are resectable for cure at the time of diagnosis. In recent years we have had considerable experience with endoscopic stenting to palliate common bile duct cancers. This has prompted us to evaluate our results for both endoscopic and surgical treatment of cholangiocarcinoma. From January 1990 through June 1999, we reviewed our endoscopic retrograde cholangiopancreatography registry and the hospital records for patients we treated for cholangiocarcinoma. Fifty patients were identified: 45 with cholangiocarcinoma and five with gallbladder cancer (who were excluded). The surgical group consisted of 16 patients: in 14 patients, resection for cure was possible whereas two had palliative procedures. There was one mortality (6%) and the median survival was 16 months. There have been no long-term surgical survivors, but 2 patients are alive at 24 months. We treated 29 patients with advanced disease with endoscopic stents (the endoscopic group) mainly for relief of obstructive jaundice. Six of 29 patients in the endoscopic group were critically ill and died in less than 4 weeks, whereas 23 patients who were in better condition survived for a mean of 10 months (range 2-84 months). We conclude that for common duct bile cancer surgical resection remains the treatment of choice but is applicable in only 30 to 35 per cent of cases. Endoscopic stenting effectively relieves jaundice and can provide long-term palliation comparable with surgical bypass; 12 of 29 patients in our endoscopic group survived 12 months or longer, and one is alive at 84 months after initial stenting.

Endoscopic treatment of distal bile duct stricture from chronic pancreatitis.

BACKGROUND: Endoscopic placement of biliary stents is an effective initial treatment for jaundice and cholangitis caused by common bile duct (CBD) strictures secondary to chronic pancreatitis; however, the role of endoscopic treatment for long-term management of these strictures is less clear. In 1992, we designed a protocol of balloon dilatation and stenting for > or =12 months. This study evaluates endoscopic therapy as a definitive long-term treatment for these strictures. We have treated 25 patients with this protocol. METHODS: All patients had an endoscopic sphincterotomy, balloon dilatation of the stricture, and then placement of a polyethylene stent (7-11.5 F). Stents were exchanged at 3-4-month intervals to avoid the complications of clogging and cholangitis. We were particularly interested in how many patients would achieve resolution of the stricture and tolerate removal of the stent. RESULTS: The length of the CBD strictures ranged from 8 to 40 mm. Within days of stenting, all patients achieved relief of jaundice and cholestasis. Complications consisted of six episodes of cholangitis and nine episodes of pancreatitis. There were no deaths. Twenty of the 25 patients are now stent-free after an average stenting period of 13 months (range, 3-28). To date, there has been no recurrence of stricture, for a mean of 32 months. Three patients still have stents in place, and two patients required operation--one for persistent stricture and recurrent cholangitis after 8 months of stenting, and one for a mass in the head of the pancreas that was thought to be cancer. CONCLUSIONS: Our results indicate that these strictures will respond and dilate after a course of stenting in 80% of patients, with an acceptable morbidity. Although these are medium-term results at 32 months, we would expect most recurrences within the 1st year following stent removal. In some cases, stenting is necessary for >12 months. Thus, the data suggest that endoscopic stenting provides definitive treatment in most patients with CBD stricture due to chronic pancreatitis and may be considered a viable alternative to standard surgical bypass.

Endoscopic drainage of the pancreatic pseudocyst.

BACKGROUND: Pancreatic pseudocyst is a common complication of chronic pancreatitis occurring in 20% to 40% of cases. Pseudocysts can be treated by endoscopic cystenterostomy or transpapillary drainage, percutaneously with computed tomography guidance or operatively. METHODS: A total of 36 endoscopic pancreatic pseudocyst drainage procedures were performed in 29 patients with 34 pseudocysts. Eighty percent presented with chronic pain, 25% had recurrent pancreatitis, and approximately one half of the patients had either gastric outlet obstruction or a palpable abdominal mass. RESULTS: Thirty-six endoscopic drainage procedures were performed, 27 cystenterostomies and 9 transpapillary drainages. Endoscopic treatment achieved complete resolution of the pseudocyst in 24 of 29 patients (83%), and the other 5 (17%) eventually required surgery. Two patients required distal pancreatectomy because of their pancreatic pathology, 2 cystgastrostomies for persistence of the pseudocyst, and 1 external drainage of an infected pancreatic cyst. The mean follow-up after the initial drainage was 16 months. There were no deaths attributed to the procedures and no complication that required surgery. Only 1 nonadherent pseudocyst (cystadenoma) required immediate operation after attempted endoscopic drainage. CONCLUSIONS: The conclude that endoscopic drainage of pancreatic pseudocysts can be both safe and effective, and definitive treatment. It should be considered as an alternative option before standard surgical drainage in selected patients.

Macromolecular cytosolic delivery: cell membranes as the primary obstacle.

The "evolution" of a thing, a custom, an organ is thus by no means its progressus toward a goal, even less a logical progressus by the shortest route and with the least expenditure of force, but a succession of more or less profound, mutually independent processes of subduing, plus the resistances they encounter, the attempts at transformation for the purpose of defense and reaction, and the results of successful counteractions. The form is fluid, but the "meaning" is even more so.

Selective role for endoscopic retrograde cholangiopancreatography in abdominal trauma.

BACKGROUND: Emergency endsocopic retrograde cholangiopancreatography (ERCP) is rarely indicated in trauma patients; however, in cases of suspected pancreatic or bile duct injury or bile leak, it may be useful. The purpose of this paper is to review our ERCP experience in trauma patients. Our Level I Trauma Center admits 1800 patients annually. METHODS: Since January 1991, we have performed ERCP in 12 trauma patients, nine after blunt injury and three after penetrating injury. RESULTS: ERCP was used as a diagnostic tool to evaluate the pancreatic duct in six stable patients with equivocal CT scans and unexplained abdominal pain, fever, and an elevated amylase or a peripancreatic pseudocyst. Based on their ERCP findings-one intact pancreatic duct, one transected duct, and four pseudocysts-five of the six patients had operations. We performed ERCP in six patients for persistent bile leaks (five cases) or jaundice (one case). The findings were one case of bilemia (intrahepatic biliovenous fistula), one case of common bile duct disruption, and four cases of persistent bile leaks from the liver after liver injuries. Endobiliary stents placed in five patients successfully stopped the four bile leaks and closed the biliovenous fistula. The one case of ductal disruption required an open choledochojejunostomy. The only ERCP complication was an episode of cholangitis treated with antibiotics. The earliest ERCP was 3 days after injury, and most were performed within 2 months. CONCLUSIONS: ERCP is a helpful procedure for diagnosing biliary and pancreatic duct injury in a select group of trauma patients who do not have obvious indications for exploration. In addition, ERCP techniques are also effective for treating most bile leaks.

Intracerebral adrenoceptor agonists influence rat duodenal mucosal bicarbonate secretion.

We have studied the effects of intracerebral administration of selective alpha-adrenergic agonists on duodenal bicarbonate secretion. Duodenum free of Brunner's glands was cannulated in situ in anesthetized rats, and bicarbonate secretion into the luminal reperfusate was continuously titrated by pH stat. Infusion of the alpha 1-selective adrenoceptor agonist, phenylephrine (1,000-2,500 micrograms.kg-1.h-1), into a lateral brain ventricle increased (P < 0.01) duodenal bicarbonate secretion. Pretreatment with prazosin, an alpha 1-antagonist, significantly (P < 0.01) reduced the stimulatory effect when infused into the lateral ventricle (30 micrograms.kg-1.h-1), but not when administered intravenously (1,000 micrograms.kg-1.h-1). Hexamethonium (10 mg.kg-1.h-1 iv) abolished stimulation, whereas cervical vagotomy, epidural blockade, and naloxone were each without effect. Vasopressin, vasopressin antagonists, ts, and oxytocin did not affect basal secretion. Intracerebro-ventricular administration of the alpha 2-adrenoceptor agonist, clonidine (1,000 micrograms.kg-1.h-1), in contrast to alpha 1-receptor activation, decreased (P < 0.01) the secretion. Thus central nervous adrenoceptors influence duodenal mucosal bicarbonate te secretion, and alpha 1-adrenoceptor stimulation may provide protection against luminal acid. This potent stimulation was not mediated by the vagal nerves, spinal cord pathways, or the release of beta-endorphin but involves nicotinic, possibly enteric nervous transmission.

Management of malignant biliary stricture with self-expanding metallic stent.

BACKGROUND: Self-expanding metallic mesh stents are designed to remain patent longer than polyethylene (PE) stents, which generally clog in 3 to 4 months. Though more expensive, metal stents may therefore be a better choice for malignant strictures. METHODS: From January 1991 to October 1995, we performed ERCP in 212 patients with malignant or benign strictures, and 34 ultimately had insertion of a metallic stent. These stents were placed by the percutaneous transhepatic route in 17 patients and endoscopically in 17. RESULTS: Metallic stent insertion was successful in each case and relieved the preoperative jaundice and cholangitis. There were no procedure-related deaths; complications were pancreatitis (one) and hemorrhage (one). Overall stent patency was 6.2 months. Three of 34 stents occluded due to tumor ingrowth at 3, 4.5, and 8 months and were treated by placing a new PE stent through the blocked metal stent. The remaining 31 stents remained patent until patient death (n = 15, mean survival = 4.9 months) or are still open (n = 16, mean patency = 12.2 months). CONCLUSIONS: Self-expanding metal stents provide effective palliation of malignant biliary strictures and should be considered an alternative to open surgery. Metal stents remain patent much longer than PE stents and usually a single session of metal stenting can palliate biliary obstruction for life.

Delayed gastric emptying after gastric surgery.

BACKGROUND: The reported incidence of delayed gastric emptying (DGE) after gastric surgery is 5% to 25% and usually is based on operations for peptic ulcer disease. Ongoing improvements in perioperative care, nutritional support, and new prokinetic drugs may have had a beneficial effect on the frequency and course of postoperative DGE. METHODS: We therefore studied our recent experience with DGE in 416 patients who had gastric surgery for ulcer disease (283), cancer (92), or trauma and other indications (41) between January 1985 and December 1993. DGE was defined as inability to eat a regular diet by postoperative day 10. RESULTS: DGE occurred in 99 of 416 patients (24%). In 75 of these 99 patients, a postoperative contributing factor for DGE was identified. These factors were sepsis (32), anastomotic edema and leaks (23), obstruction (4), pancreatitis (3), multiple system organ failure (5), and miscellaneous conditions (8). In 24 patients there was no obvious cause for DGE; these patients recovered with nutritional support and time. Re-operation specifically for gastric stasis was not performed. Among the 99 patients with DGE, 67% were eating by day 21, 92% by 6 weeks, and 100% by 10 weeks. Significant risk factors for DGE were diabetes (55%), malnutrition (44%), and operations for malignancy (38%). The Whipple procedure had the highest incidence of DGE (70%), highly selective vagotomy the lowest (0%), while truncal vagotomy had no significant effect. The response to metoclopramide was 20% and unpredictable. CONCLUSION: DGE continues to affect a considerable number of our patients (24%) after gastric surgery and is particularly common in patients with diabetes, malnutrition, and gastric or pancreatic cancer. However, gastric motility does return in 3 to 6 weeks in most patients and the need for re-operation for gastric stasis is rare.

Endoscopic management of benign and malignant biliary strictures.

BACKGROUND: It is now possible to manage most extrahepatic bile duct strictures, benign or malignant, using endoscopic retrograde cholangiopancreatography (ERCP) with endoscopic dilatation and stenting. METHODS: Over a 5-year period we treated 218 patients with strictures of extrahepatic bile ducts. Eighty-six patients had benign biliary stricture. Endoscopic treatment was performed in 67 (78%) of these patients. Open surgical biliary drainage was preferred in 12 patients (14%), and 7 patients (8%) were managed conservatively without stenting or surgery. One hundred and thirty-two patients had malignant biliary stricture. One hundred and one patients (77%) underwent endoscopic stent placement. Thirty-one patients (23%) underwent surgery for potential curative resection after diagnostic ERCP. The average life span in the malignant stricture group was 5 months (range 0.1 to 25 months) after the initial endoscopic procedure. RESULTS: Altogether 313 endoscopic procedures in 218 patients were performed for benign and malignant bile duct strictures. Complications included hemorrhage in 8 (3%), pancreatitis in 10 (3%), and suspected retroperitoneal perforation in 2 (0.6%). There were no ERCP related deaths; one patient died of uncontrolled bleeding from transhepatic stenting. In benign strictures, there has been no recurrence of strictures after the last stent removal with a mean followup of 21 months (range 0.1 to 31 months). All complications were successfully treated conservatively. CONCLUSIONS: Endoscopic management of benign and malignant biliary stricture is possible with minimal morbidity and mortality and should be considered an acceptable option to surgical management.

Acute biliary pancreatitis. The roles of laparoscopic cholecystectomy and endoscopic retrograde cholangiopancreatography.

Since January 1990, we have treated 113 patients for gallstone pancreatitis; 59 with laparoscopic cholecystectomy (LC), 50 with open cholecystectomy, and 4 with ERCP/sphincterotomy only. In the LC group, 47 had LC during the index admission and 12 underwent delayed LC. Fifty patients had open cholecystectomy, 47 during the index admission. ERCPs were performed in 43 of the 113 patients; CBD stones were identified in 19/43 (44%) and removed endoscopically in 18 (95%). The ERCP complication rate was 6.5%. In total, CBD stones were identified in 29/113 patients (26%). Patients who had imaging of the CBD within the first 4 days from onset of symptoms were more likely to have stones identified than were those patients who were studied after 5 days. Recurrent pancreatitis occurred in in five of 11 patients (45%) who had a > or = 30-day delay to definitive treatment. We conclude that LC can be safely performed in most patients during the index admission for gallstone pancreatitis. This policy should reduce the 30-50% risk of recurrent pancreatitis associated with a delayed operation. ERCP is a helpful adjunct for CBD stones.

Effect of ranitidine on intragastric pH and stress-related upper gastrointestinal bleeding in patients with severe head injury.

We conducted a double-blind, placebo-controlled study to evaluate the effects of ranitidine on intragastric pH and upper gastrointestinal tract bleeding in severe head injury patients. Within 24 hr of the precipitating trauma, 34 adults with Glasgow coma scale scores < or = 10 were randomized to a 6.25 mg/hr ranitidine continuous infusion or placebo for a maximum of 72 hr. Intragastric pH was recorded via an intragastric pH electrode. Patients with hematemesis, hematochezia, bright red blood, or "coffee ground" nasogastric tube aspirates plus a 5% decrease from baseline in hematocrit were considered to have gastrointestinal bleeding. Ranitidine patients maintained a significantly greater mean pH than placebo patients (placebo 2.2, ranitidine 4.1; P < 0.01). All patients had at least two bleeding risk factors at study entry. No ranitidine patients (0/16) developed bleeding compared with five (5/18) placebo patients (P < 0.05). Ranitidine continuous infusion provided consistent intragastric pH control and significant protection from stress-related upper gastrointestinal tract bleeding in a high-risk patient population.

Laparoscopy for abdominal emergencies.

The role of laparoscopy has been reviewed for these conditions: abdominal trauma, acute abdomen, abdominal pain of uncertain etiology, appendicitis and the acute abdomen in the intensive care unit patient. Laparoscopy should only be performed in trauma patients who are hemodynamically stable and who have some evidence for abdominal injury, such as a positive peritoneal lavage or a positive CT scan. Laparoscopy is an excellent procedure for determining whether a knife or missile has penetrated the peritoneum. For penetrating wounds in the chest and upper abdomen, laparoscopy also allows excellent evaluation of the diaphragm. In blunt trauma, laparoscopy identifies the majority of injuries, but there has been a 5-15% incidence of missed injuries to the small bowel and colon. The acute abdomen is generally caused by perforation, acute inflammation or intestinal obstruction. Of the various types of perforation, diagnostic and therapeutic laparoscopy is most applicable for duodenal perforation. Acute perforation of the stomach and colon should probably be treated by standard open techniques. For acute inflammatory disorders, laparoscopy is an excellent diagnostic tool and can also provide definitive treatment in the form of drainage of an abscess or appendectomy. The role of laparoscopy for ileus and bowel obstruction is controversial; some surgeons advocate diagnostic laparoscopy and treatment, while many others still consider bowel obstruction and abdominal distention to be contra-indications. Finally, there are the intensive care unit patients in whom an acute intraabdominal process is suspected. Laparoscopy in such patients alters the clinical management in about 50% of patients.(ABSTRACT TRUNCATED AT 250 WORDS)

Impact of multiple risk factors and ranitidine prophylaxis on the development of stress-related upper gastrointestinal bleeding: a prospective, multicenter, double-blind, randomized trial. The Ranitidine Head Injury Study Group.

OBJECTIVES: To evaluate the impact of risk factors on the development of stress-related upper gastrointestinal bleeding in severe head injury patients randomized to treatment with a 6.25 mg/hr continuous ranitidine infusion or placebo. DESIGN: Prospective, multicenter, randomized, double-blind, placebo-controlled, parallel-group study. SETTING: Ten intensive care units in the United States. PATIENTS: Patients with severe head injury, defined as having a Glasgow Coma Score of < or = 10, were eligible for enrollment. INTERVENTIONS: Ranitidine 6.25 mg/hr or saline placebo was administered by continuous infusion for a maximum of 5 days. MEASUREMENTS AND MAIN RESULTS: Patients were evaluated every 8 hrs for the presence of stress-related upper gastrointestinal bleeding. Bleeding developed in 15 (19%) of 81 placebo-treated patients vs. three (3%) of 86 ranitidine-treated patients (p = .002). None of the individual risk factors had a significant effect on bleeding frequency. No bleeding occurred in the four patients with one risk factor. Placebo bleeding rates in patients with 2, 3 to 5, and > 5 risk factors were 20%, 20%, and 18%, respectively. For the ranitidine-treated patients, bleeding was reported in 0%, 5%, and 0% in the 2, 3 to 5, and > 5 risk factor subgroups, respectively. Pneumonia occurred in 19% of the placebo-treated patients vs. 14% in the ranitidine treatment group. CONCLUSIONS: The full risk to develop stress-related upper gastrointestinal bleeding was realized when two risk factors were present concomitantly. The presence of additional risk factors did not increase the occurrence of bleeding. A continuous infusion of ranitidine at 6.25 mg/hr provided significant protection from bleeding, regardless of the number of risk factors present.