Upper GI foreign body: an adult urban emergency hospital experience.

BACKGROUND: The efficacy of flexible endoscopy by a single endoscopist in the therapy of foreign body ingestion was assessed at an adult urban emergency hospital. METHODS: Fifty-one adult patients with upper GI foreign body ingestion treated at Detroit Receiving Hospital from 1988 to 2004 were identified. Endoscopic and hospital medical records were reviewed to evaluate etiology, treatment, and outcomes for these patients. RESULTS: The etiology was related to eating in 38(75%) patients, most of whom were eating meat; phytobezoars were seen in four, often after previous upper GI surgery. True foreign bodies were found in 13 patients (25%) and included a screwdriver, a ballpoint pen, spoons, coat hanger pieces, batteries, and latex gloves. Dysphagia was the most common symptom (75%); pain was common in patients with true foreign bodies, and 62% of this group had psychiatric difficulties or problems with drug abuse. Nearly 80% of the food-related group had post-surgical or other upper GI pathology. One patient had an esophageal stricture secondary to previous Sengstaken-Blakemore tube insertion. Flexible endoscopy was successful in extracting the foreign body in almost all (49) patients, with snare extraction the most common therapeutic modality. Both failures were of true foreign bodies that could not be safely removed. In one of these cases, it became necessary to employ the gallstone lithotripter, and the overtube was required in patients with metallic or sharp foreign bodies to protect the upper aerodigestive structures. CONCLUSIONS: Most upper GI foreign bodies are related to food impaction, with meat most often found. Underlying pathology is the rule and should be dealt with immediately. Flexible endoscopy is the treatment of choice for upper GI foreign body removal with near perfect success.

Endoscopic repair of gastrostomy after inadvertent removal of percutaneous endoscopic gastrostomy tube.

Percutaneous endoscopic gastrostomy (PEG) tube placement has become the preferred method of enteral feeding for many patients. Neurologic disease and cancer are the most frequent indications for PEG tube placement. PEG tubes are also becoming more frequent in trauma patients for early initiation of enteral feeding. Inadvertent PEG tube removal is a well-known complication of PEG tubes. Patients undergoing PEG tube placement are frequently malnourished and in poor general medical condition, making them relatively high risk for surgical intervention. In the past, after early inadvertent PEG removal, patients underwent laparotomy for surgical repair of the gastrostomy site. Recently, laparoscopic replacement of the PEG tube has been described. We present a new technique of endoscopic repair of the gastrostomy site with hemoclip placement followed by later PEG tube placement.

Current therapy for nonvariceal upper gastrointestinal bleeding.

Upper gastrointestinal bleeding continues to plague physicians despite the discovery of Helicobacter pylori and advances in medical therapy for peptic ulcer disease. Medical therapy with new nonsteroidal anti-inflammatory medications and somatostatin/octreotide and intravenous proton pump inhibitors provides hope for reducing the incidence of and treating bleeding peptic ulcer disease. Endoscopic therapy remains the mainstay for diagnosis and treatment of upper gastrointestinal bleeding. Many methods of endoscopic hemostasis have proven useful in upper gastrointestinal hemorrhage. Currently, combination therapy with epinephrine injection and bicap or heater probe therapy is most commonly employed in the United States. Angiography and embolization play a role primarily when endoscopic therapy is unsuccessful.

Endoscopic sphincterotomy for stenosis of the sphincter of Oddi.

BACKGROUND: Sphincter of Oddi dysfunction (SOD) is one of the causes of postcholecytectomy syndrome and biliary pain. Endoscopic sphincterotomy (EST) is recommended in some cases for patients refractory to conservative treatment. By the Milwaukee classification, patients with biliary pain can be divided into three groups. Group I patients show all the objective signs suggestive of a disturbed bile outflow-i.e., elevated liver function tests, dilated common bile duct (CBD), and delayed contrast drainage during endoscopic retrograde cholangio pancreatography (ERCP). Group II patients have biliary-type pain along with one or two of the criteria from group I. Group III patients have only biliary pain, with no other abnormalities. This study confirms the effectiveness of EST for the relief of symptoms in group I patients (papillary stenosis). METHODS: Between 1989 and 1999, we treated eight patients clinically diagnosed as having group I papillary stenosis by EST. Their ages ranged from 52 to 73 years. In addition to biliary pain, all patients were found to have dilated CBD, elevated enzyme levels, and delayed contrast drainage at ERCP. None of the patients had CBD stones or other causes of obstruction. Sphincter of Oddi manometry was not performed. RESULTS: EST was successfully performed in eight patients. Each patient had a very large papilla. A false orifice was found in one patient. In five patients, endoscopic cannulation of the bile duct was very difficult. The use of a long, tapered catheter and guidewire papillotomy was necessary in four patients. A precut papillotomy was performed in one patient. All patients achieved resolution of their symptoms after EST. There were no complications. The average length of the follow-up period was 26 months. CONCLUSIONS: SOD is a real entity that continues to pose a diagnostic dilemma. EST is an effective and safe modality for the treatment of papillary stenosis (group I patients). SOD manometry is not necessary before EST in group I patients.

Experimental studies of thermal therapy for severe nonvariceal bleeding in dogs.

BACKGROUND: There are several methods of achieving endoscopic hemostasis for nonvariceal hemorrhage, including use of a heater probe, bipolar electrocoagulation, use of a Gold probe, and injection therapy with epinephrine or ethyl alcohol. However, due to clinical variations, clinical studies comparing thermal with injection therapy have yielded conflicting results. Therefore, we used a canine model of acute bleeding from gastric serosal vessels to examine the efficacy of the heater probe and the Gold probe in achieving hemostasis and to compare the injurious effects of these methods with injection therapy. METHODS: Seven mongrel dogs were used in the study. Four were assigned to acute experiments in which transected blood vessels were allowed to bleed profusely. Two dogs of this group were treated with either a large or small Gold probe, while the other two were treated with either a large or small heater probe. In the other three dogs, we tested the chronic effects of the heater probe, the Gold probe, and injection therapy with dilute epinephrine. RESULTS: Complete hemostasis was achieved for all four dogs in the acute experiments. Dogs treated with either a large or small heater probe had coagulation necrosis that extended to the serosa and muscularis but not to the mucosa. The large Gold probe had similar results, but the small Gold probe caused tissue damage to the serosa, muscularis, submucosa, and mucosa at several of the application sites. Both probes caused scarring of the gastric wall. In the chronic experiments, we found that the Gold probe caused larger mucosal ulcers than the heater probe. All ulcers healed in 3 weeks. The epinephrine injection caused localized swelling and discoloration, but after 1 week the tissue returned to normal. CONCLUSIONS: Both the heater probe and the Gold probe are effective in achieving hemostasis in a canine model of nonvariceal hemorrhage, and both methods are superior to injection therapy. For active bleeding ulcers, we currently recommend a combination therapy, using first injection therapy and then a heater or Gold probe. However, clinicians should be aware of the potential for tissue damage.

Bleeding esophageal ulcers caused by NSAIDs.

This report describes four patients with NSAID-induced esophageal ulcers documented by endoscopy. The cause of injury was ibuprofen alone in two patients, aspirin in one patient, and a combination of aspirin and ibuprofen in one patient. The most common findings were anemia, retrosternal pain, and dysphagia. Three patients had bleeding esophageal ulcers requiring blood transfusions. One patient had massive bleeding which was controlled by endoscopic hemostasis. Three patients were followed up by endoscopy, which showed healing in 3-4 weeks. These NSAID-induced ulcers had characteristic endoscopic features, namely, a large, shallow, discrete ulcer in the midesophagus near the aortic arch with normal surrounding mucosa. These findings suggest that the injury resulted from mucosal contact with NSAIDs. A precise history and immediate endoscopic examination were most important in establishing the diagnosis of esophageal ulcer. Healing occurs if drug-induced injury is recognized early and treatment is appropriately started with antacids and H2 blockade. Offending medication should be discontinued and patients should be counseled to take pills in an upright posture with liberal amounts of fluids well before retiring for the night.

Endoscopic management of nonvariceal gastrointestinal bleeding.

Endoscopic management of upper gastrointestinal bleeding has been expanded from a purely diagnostic role to a therapeutic role in many patients. In addition to controlling active bleeding, it is an option in a patient who is clinically at a high risk of rebleeding, or in patients who have peptic ulcers with visible vessels or stigmata indicating high risk. Several methods have been studied, and currently the most useful include thermal cautery with the heater probe or bipolar electrocoagulation, and injection using epinephrine and/or sclerosants. Endoscopic hemostasis can effect permanent control of bleeding in many patients, but should be considered complementary to conventional surgical control in other patients, where temporary control to stabilize the patient is a desired end.

Endoscopic interventional management of bleeding duodenal and gastric ulcers.

Bleeding duodenal and gastric ulcers continue to be a common and serious problem. Definition of the precise appearance and location of the ulcer by endoscopy gives important information about the source of bleeding and additional information about the risk of rebleeding and the indications for surgery. Several endoscopic hemostatic methods are available. The nonerosive contact probes (heater and BICAP) are preferred. Injection therapy with vasoconstrictors or sclerosing agents can also be recommended as a safe, efficacious, and economical means of treatment. Several hemostatic modalities should be available for use depending on the anatomic location and type of bleeding ulcers. The collaboration of skilled interventional endoscopists with their traditional surgical colleagues offers the patient with bleeding peptic ulcer disease the optimum probability of a successful outcome, with minimum treatment-associated morbidity.

Effect of hemostatic agents in canine gastric serosal blood vessels.

The efficacy of various sclerotherapeutic agents in the control of acute bleeding, via subserosal injection, was assessed in 10 dogs. Blood flow rate (BFR) from severed gastric serosal vessels (diameter, 1.6 to 2.2 mm) was measured for 3 min (ml/min) for a control group and the agent used. The agents tested were 5 ml of normal saline (NS), 5 ml of 3% hypertonic saline (HS), 5 ml of 1:10,000 epinephrine in NS, 5 ml of 1:10,000 epinephrine/HS, 5 ml of 1:20,000 epinephrine/HS, 2 ml of old thrombin "cocktail" (thrombin, cephapirin + 1% tetradecyl), and 2 ml of fresh thrombin cocktail (total seven). One agent was tested per dog; there were one to two dogs in each subgroup. All of the agents showed significant reduction in BFR (except old thrombin) when compared with BFR of control vessels. The reduction ranged from 30% to more than 75% after 1:10 epinephrine/HS. Complete hemostasis was achieved in up to 47% of vessels using 1:20 epinephrine/HS. Overall, the epinephrine solutions achieved the best results. No systemic effects were observed with the use of any of the agents. Histological studies showed that epinephrine caused mild tissue damage, whereas the cocktail caused significant tissue necrosis. This serosal vessel model permits comparison of the effectiveness of each agent; however, clinical extrapolation to mucosal vessels in a patient and the long-term histological changes are not known.

Sclerotherapy of bleeding esophageal varices using a thrombogenic cocktail.

The short- and long-term efficacy of a thrombogenic sclerosant (1% tetradecyl sulfate, thrombin, and cefazolin) was studied in 101 patients. The majority of patients had alcoholic cirrhosis with Child's C classification (84/101). Bleeding was controlled in 94 per cent of patients with the first sclerotherapy. In-hospital and early (within 6 weeks) mortality were 14 per cent and 19 per cent, respectively. There was a strong correlation with hospital mortality and the severity of hepatic disease. Long-term follow-up in 70 patients (mean of 16 months) showed that survival correlated with compliance to follow-up sclerotherapy and abstention from further alcohol intake. Mortality in patients compliant with follow-up was 5 per cent (1/19), as compared with 24 per cent (12/51) in patients who were not compliant with follow-up sclerotherapy. The mortality in alcoholic cirrhotic patients who abstained from further alcoholic intake was 6 per cent (1/17), as compared with 23 per cent (10/44) in those who continued to abuse alcohol. No systemic thrombotic or allergic events related to the use of bovine thrombin were noted during a total of 349 sclerotherapy sessions.

Hemostasis activation during esophageal variceal sclerotherapy with thrombin in cirrhotics.

Sclerotherapy of bleeding esophageal varices in liver cirrhotics is a common procedure, but little is known about the possible entry of sclerosants into the systemic circulation. We injected a mixture of thrombin, sodium tetradecyl, and cefazolin and studied the effect of this sclerosant on selected hemostasis parameters. Twenty-four patients with liver cirrhosis (Child's Classification C) were studied 29 times. Blood samples were drawn before and immediately after the injection of the sclerosant. In seven patients we collected a sample 30 minutes and 24 hours after treatment. Before injection, almost all patients had elevated D-dimer, t-PA and PAI-1 levels. Fibrinogen, antithrombin, alpha-2 antiplasmin, and protein C were decreased. Only thrombin/antithrombin III complex (TAT) levels were within normal ranges. Immediately after the injection, TAT, D-dimer, and t-PA levels rose significantly (P less than 0.001, P less than 0.01, P less than 0.001), PAI-1 and PC levels decreased (P less than 0.01), while antithrombin, alpha-2 antiplasmin, and fibrinogen concentrations were unchanged. TAT and D-dimer levels were still elevated after 24 hours (P less than 0.05). These data indicate that thrombin entered the systemic circulation (elevated TAT) and that the hemostasis system was briefly systemically activated (elevated D-dimer). In spite of these changes in the hemostasis system, clinically there were no detectable thrombotic or hemorrhagic complications.

The role of endoscopic retrograde pancreatography (ERP) in blunt abdominal trauma.

The diagnosis of pancreatic injury is often difficult because it lies retroperitoneally in a protected area. Delayed diagnosis and treatment of blunt pancreatic trauma can result in significant morbidity and mortality. Endoscopic Retrograde Pancreatography (ERP) is infrequently used in the diagnosis of pancreatic injury. We reviewed our experience with the use of ERP in patients with blunt pancreatic injury. Two stable patients with traumatic pancreatitis underwent ERP shortly after injury. CT scans revealed a transverse fracture of the distal pancrease in one and fluid accumulation in the other around the pancreas extending to the right kidney and left hepatic lobe. The absence of ductal disruption on ERP allowed nonoperative management of the pancreatitis. Resolution was documented by the absence of symptoms on regular oral intake, normal serum amylase levels, and normal follow-up CT scans. A third patient with persistent fistulae three months postinjury underwent preoperative ERP revealing ductal obstruction. This facilitated the planning of a distal pancreatectomy and subsequently the fistulae healed. A fourth patient underwent an exploratory laparotomy on the basis of clinical and CT scan findings that could have been circumvented with preoperative ERP. ERP in selected patients allows nonoperative treatment in the absence of ductal injury or earlier operative treatment of ductal injury. It also aids the treatment of late complications by delineating ductal anatomy.

Upper GI bleeding in an urban hospital. Etiology, recurrence, and prognosis.

Acute upper gastrointestinal bleeding (UGIB) continues to be a common cause of hospital admission and morbidity and mortality. This study reviews 469 patients admitted to a surgical service of an urban hospital. There were 562 total admissions because 53 patients were readmitted 93 times (recurrence rate, 20%). The most common causes of bleeding, all endoscopically diagnosed, included acute gastric mucosal lesion (AGML) (135 patients, 24%), esophageal varices (EV) (121 patients, 22%), gastric ulcer (108 patients, 19%), duodenal ulcer (78 patients, 14%), Mallory-Weiss tear (61 patients, 11%), and esophagitis (15 patients, 3%). Nonoperative therapy was sufficient in 504 cases (89.5%). Endoscopic treatment was used in 144 cases. Operations were performed in 58 cases (10.5%), including 29% of ulcers. Emergency operations to control hemorrhage were required in only 2.5% of all cases. The rate of major surgical complications was 11% and the mortality rate was 5.2%. There were 58 deaths (12.6%), with 36 deaths directly attributable to UGIB. Factors correlating with death include shock at admission (systolic blood pressure less than 80), transfusion requirement of more than five units, and presence of EV (all p less than 0.001). Most cases of UGIB can be treated without operation, including endoscopic treatment, when diagnostic endoscopy establishes the source. Subsequent operation in selected patients can be done with low morbidity and mortality rates.

Oxygen saturation monitoring during endoscopy.

The role of monitoring during endoscopy is not clearly defined. We have prospectively investigated continuous arterial oxygen saturation (SaO2) monitoring in 326 patients undergoing upper endoscopy (EGD) and 90 undergoing colonoscopy. Automated blood pressure recording was evaluated in 278 of these patients. SaO2 desaturation (less than 90%) occurred in 17.8% of patients undergoing EGD and 12.9% undergoing colonoscopy. Systolic blood pressure abnormalities (greater than 200 or less than 90 mmHg) occurred in 19.8% of EGD patients and 19.6% of colonoscopy patients. Treatment based on these abnormalities was required in 4.3% of patients during EGD and 8.8% during colonoscopy. A history of pulmonary or cardiac disease predicted increased risk during colonoscopy, while cardiac disease and age 60 years or above predicted desaturation during EGD. Pulse oximetry and automated blood pressure monitoring was especially valuable during endoscopy in the elderly and patients with cardiac or pulmonary disease. It may be used as a guide to therapeutic intervention and to avert major cardiopulmonary complications.