Search

Acquired fistulas between the esophagus and tracheobronchial tree are usually associated with malignancy of the esophagus, lung, or trachea. Less commonly, fistulas result from trauma or inflammation involving these structures. Untreated fistulas of any cause lead to fatal complications of aspiration. Although the prognosis in cases of malignant fistula is poor, the recognition and surgical management of nonmalignant fistulas may result in cure. An acquired esophagobronchial fistula resulting from a Barrett's ulcer of the esophagus, a previously unreported cause, is described, and the differential diagnosis and treatment of nonmalignant esophagorespiratory fistulas are discussed.

We analyzed the course of 50 consecutive patients with recurrent biliary stricture (at least one previous repair) to determine the pattern of recurrence and the outcome of additional treatment. The presenting manifestations included cholangitis in 40 percent of the patients, jaundice in 30 percent, and pain in 17 percent. Sixty-seven percent had become symptomatic within 2 years of their previous repair and 90 percent within 7 years. Ten (20 percent) patients were treated by end-to-end anastomosis of the bile duct, and 40 (80 percent) patients by bilioenteric anastomosis (hepaticojejunostomy in 32 patients and hepaticoduodenostomy in 8). Two patients (4 percent) died postoperatively. After repair of recurrent stricture, 38 (76 percent) patients had no further symptoms. In 11 (22 percent), another recurrence developed: 6 (55 percent) of the 11 did well after another operation and in 4 (36 percent), a third recurrence developed which was successfully treated by a fourth operation in 3 of the patients. Thus, two thirds of recurrent strictures were evident by 2 years and 90 percent by 7 years. The chance of another recurrence was about 25 percent after treatment of a first recurrence. There was no evidence that prolonged stenting contributed to a good result. Treatment of recurrent stricture should consist of anastomosis between the duct and the intestine, usually a hepaticojejunostomy.

We studied the effects of vagotomy on gallbladder (GB) motility in prairie dogs and humans with infusion cholescintigraphy. Twelve male prairie dogs were anesthetized and given an intravenous infusion of 120 microCi of diethyl-HIDA for 150 minutes. Images were acquired every 10 minutes. Then cholecystokinin (CCK)-8, 1.5 micrograms/kg, was given as a bolus, and images were acquired for another 30 minutes. We repeated the studies giving 300 micrograms/kg of atropine 20 minutes before administration of CCK-8. All animals underwent truncal vagotomy, and the studies were repeated 1 and 3 months later. The GB filled in a stepwise fashion; partitioning of bile varied from one 10-minute period to the next and averaged 20% +/- 2%/80% +/- 3% during the 150-minute period. Episodic partial GB emptying (ejection fraction 19% +/- 2%; intervals of 70 +/- 5 minutes) occurred during this phase. GB filling and partitioning of bile were unchanged after vagotomy. GB ejection fraction in response to CCK-8 was 69% +/- 6% in controls, 74% +/- 5% after atropine, 78% +/- 8% 4 weeks after vagotomy, and 66% +/- 6% 3 months after vagotomy. Sixteen human subjects were studied after parietal cell vagotomy (six patients) or truncal vagotomy and drainage (10 patients). GB filling average 2.5% +/- 2% per minute in patients who underwent truncal vagotomy and 3% +/- 1% per minute in patients who underwent parietal cell vagotomy. GB emptying in response to CCK-33 (0.02 U/kg/min) was 74% +/- 7% in patients who underwent truncal vagotomy and 82% +/- 4% in patients who underwent parietal cell vagotomy. Thus neither GB filling nor GB emptying in response to CCK was altered by cholinergic blockade or vagotomy.

We studied gallbladder bile flow before, during, and after cholesterol gallstone formation in the prairie dog using infusion cholescintigraphy with 99mTc-diethyl iminodiacetic acid. In 18 fasting animals partitioning of bile between gallbladder and intestine was determined every 15 min for 140 min, and gallbladder response to cholecystokinin (5 U/kg X h) was calculated from the gallbladder ejection fraction. Ten prairie dogs were then placed on a 0.4% cholesterol diet and 8 on a regular diet, and the studies were repeated 1, 2, and 6 wk later. The proportion of hepatic bile that entered the gallbladder relative to the intestine varied from one 15-min period to the next, and averaged 28.2% +/- 5.1% at 140 min. Partial spontaneous gallbladder emptying (ejection fraction 11.5% +/- 5.6%) was intermittently observed. Neither the number nor the ejection fraction of spontaneous gallbladder contractions changed during gallstone formation. By contrast, the percent of gallbladder emptying in response to cholecystokinin decreased from 72.1% +/- 5% to 25.9% +/- 9.3% (p less than 0.025) in the first week and was 14.3% +/- 5.5% at 6 wk (p less than 0.01 from prediet values, not significant from first week). Gallbladder filling decreased from 28.2% +/- 5.1% to 6.7% +/- 3% (p less than 0.01), but this change was only observed after 6 wk, when gallstones had formed. This study shows that bile flow into the gallbladder during fasting is not constant; the gallbladder contracts intermittently; gallbladder emptying in response to exogenous cholecystokinin is altered very early during gallstone formation; and gallbladder filling remains unaffected until later stages, when gallstones have formed.

We studied the effects of stasis of gallbladder bile in a dog model. Three days after cystic duct ligation, all gallbladders contained sludge, and the mucosa was covered by densely adherent mucus with solid particles 1-4 mm in diameter (gravel). Thirty percent of the animals developed stones (greater than 4 mm), which appeared grossly like human pigment stones and microscopically like condensed biliary sludge. Centrifugation of bile yielded colorless pellets (3.8 +/- 3.2 mg/ml) at day 0 and pigmented pellets (33.1 +/- 11.0 mg/ml) at day 3 (p less than 0.05). Pellets contained 73 +/- 8% mucin by weight. Dissolved mucin in supernatant bile increased from 7.46 +/- 1.19 mg/ml (day 0) to 27.36 +/- 3.05 mg/ml (day 3) (p less than 0.001), while bilirubin concentration decreased from 127 +/- 12 mg/dl (day 0) to 71 +/- 16 mg/dl (day 3) (p less than 0.001). Cholesterol concentration increased but did not reach saturation, while the concentration of bile salt and phospholipid did not change. Mucin-bilirubin complexes formed and remained suspended as sludge initially. As bile mucin content increased, sludge particles coalesced, precipitated, and eventually formed gravel and stones. We suspect that stone formation in this setting occurs because of sequestration of biliary lipids by mucin.

The technique of 24 hour esophageal pH monitoring (24 hour pH test) is described. Experience with the 24 hour pH test in 393 patients with suspected esophageal disease has shown the clinical usefulness of the test in objectively determining the presence of gastroesophageal reflux. The test was effective in evaluating atypical symptoms of gastroesophageal reflux such as respiratory symptoms and chest pain and, in children, failure to thrive and recurrent pneumonia. The 24 hour pH test was particularly useful in evaluating patients who were referred with other abdominal or thoracic disease and had, in addition, symptoms suggestive of gastroesophageal reflux on history. The test helped to unsnarl the cause of recurrent symptoms after an esophageal myotomy for achalasia or an antireflux procedure. Of 179 patients with typical symptoms of gastroesophageal reflux, 27% had normal 24 hour test results and were subsequently diagnosed as having another cause for their symptoms. Of 146 patients who had normal findings on esophagoscopy, 54% were shown to have abnormal gastroesophageal reflux on 24 hour pH monitoring, indicating lack of sensitivity of endoscopy to detect reflux. In addition, the 24 hour pH test identified patterns of abnormal reflux and indicated those patients most at risk for development of stricture. The test is well tolerated by the patients, simple to use, and dependable when performed and read as described. The clinical use of the 24 hour pH test brings objectivity to the evaluation of exophageal disease that has hitherto not been available.

To reassess the role of laparotomy and extraserosal drainage in the treatment of patients with abdominal abscess, we analyzed the course of 79 patients who underwent 97 operations to treat 120 abdominal abscesses during a five-year period. In 66 clinical episodes the abscess was drained by the most direct approach. Sepsis resolved with a single operation In 80% of these patients, five patients (8%) required a second operation for drainage for an abscess, and eight patients (12%) died. In 31 clinical episodes, the abscess was drained by a laparotomy. Sepsis resolved with a single operation in 61% of these patients, seven patients (21%) had a second abscess, six patients (19%) required a second operation to drain a metachronous abscess, and six patients (19%) died. When the location or number of abscesses was diagnosed incorrectly, the success rate of therapy fell substantially. Since most abdominal abscesses can now be accurately diagnosed preoperatively, most abscesses should be drained by a direct approach. Exploratory laparotomy is indicated when preoperative localization is unsuccessful, when sepsis has not resolved after other methods of drainage, or when the patient has a concomitant abdominal condition that must be treated surgically.

Gallstones from 82 patients were examined under a scanning electron microscope for evidence of bacteria, and the findings were compared with the clinical manifestations of the disease. Bacteria were present in 68% of pigment stones and the pigment portions of 80% of composite stones. These gallstones were referred to as infectious stones. No bacteria were found in cholesterol gallstones. Acute cholangitis was diagnosed in 52% of patients with infectious stones and in 18% of patients with noninfectious stones. Over half of the patients with noninfectious stones presented with mild symptoms. Infectious stones were more often associated with a previous common duct exploration, an urgent operation, infected bile, a common duct procedure, and complications. These data show that gallstone disease is more virulent in patients whose gallstones contain bacteria.

We studied the value of tests commonly used in diagnosing obstructive jaundice in 188 patients operated on for biliary obstruction. Ultrasonography had a diagnostic accuracy of 87 percent for gallbladder stones, 82 percent for dilated ducts, and 80 percent for pancreatic masses. Computed tomography was 93 percent accurate in diagnosing dilated ducts, and 93 percent accurate in diagnosing a pancreatic mass. Intravenous cholangiography gave useful information in only 4 of 21 patients. Among 45 patients, endoscopic retrograde cholangiopancreatography was technically successful in 91 percent and gave accurate diagnostic information in 86 percent and partially diagnostic information in 7 percent; it detected all cases of common duct stones and had a 91 percent sensitivity for detecting pancreatic cancer. Among 75 patients, transhepatic cholangiography was technically successful in 95 percent and provided accurate diagnostic information in 90 percent and partially diagnostic information in 3 percent; it detected all cases of common duct stones and 95 percent of cases of benign strictures. With persistent jaundice, ultrasonography should be the first test because it is less costly than computed tomography and provides similar information. If more information is needed, transhepatic cholangiography and endoscopic retrograde cholangiopancreatography are equally accurate; transhepatic cholangiography more often opacifies the proximal biliary tree and endoscopic retrograde cholangiopancreatography provides direct access to the ampulla of Vater and the pancreatic duct.

We examined small bowel transit of solid food in 10 patients 3 to 30 months after total gastrectomy and Roux-Y esophagojejunostomy and compared the transit pattern with that in five control subjects. All persons ate a test meal consisting of 213 g of beef stew mixed with 30 g of chicken liver which was labeled with 1 mCi of technetium 99m sulfur colloid, and they stood in front of a gamma scintillation camera and were studied for 6 to 10 hours. Solid food passed rapidly from the esophagus through the Roux-Y limb and became uniformly distributed throughout a long segment of the small intestine. Mouth-to-colon transit time was 223 +/- 18 minutes in the control subjects and 298 +/- 37 minutes in the patients with gastrectomy. Small bowel transit of the head of the meal was 187 +/- 19 minutes in the control subjects and 293 +/- 37 minutes in the patients (p less than 0.02). Transit time of the tail of the meal was 175 +/- 26 minutes in the control subjects and 396 +/- 28 minutes in the patients (p less than 0.001). After total gastrectomy and Roux-Y esophagojejunostomy, the proximal jejunum does not act as a reservoir; transit of food through the small intestine is slower than in control subjects, and because the proximal jejunum empties rapidly during eating, a meal of normal size can be consumed. These findings do not support the idea that an artificial gastric reservoir is important after total gastrectomy.

From a review of 800 patients with acute pancreatitis, eight having rare causation were selected. Definite obstruction of the common pancreatic duct was noted in four patients, and direct injury of the parenchyma, vascular lesions of pancreatic vessels and presumed autoimmune factors were responsible in the remaining four patients. Although the definitive treatment of nonobstructive pancreatitis remains controversial, surgical relief of obstruction of the pancreatic duct provides the most appropriate therapeutic approach for obstructive pancreatitis.

Using corrosion casting and scanning electron microscopy of the rat biliary tree, we investigated the site and size of the pathways that allow bacteria to reflux from bile to blood. Nonobstructed rat biliary trees were injected retrograde with methylmethacrylate resin at a constant rate of 0.04 ml/min to volumes of 40, 60, 80, 120, 160, and 200 microliters. The infusion pressure was monitored and a pressure-volume curve was constructed. After polymerization and corrosion in 30 percent potassium hydroxide, the casts were examined with scanning electron microscopy. In addition, to identify the size of the reflux pathways, ceramic particles of 150 A, 1.7 mu, or 10 mu were added to the resin, and the studies were repeated. Finally, intact livers with casted biliary trees were processed and studied by scanning electron microscopy without corrosion. Scanning electron microscopy demonstrated fine anatomic detail of the cholangiovenous reflux pathway. At 40 microliters (20 cm water pressure), normal biliary radicals were filled. Between 40 and 80 microliters (20 to 50 cm water pressure), the cast material refluxed from the bile ductules into the spaces of Mall and Disse and then into the hepatic sinusoids. Filling of sinusoids continued at volumes between 80 to 160 microliters, and filling of collecting veins was seen above 160 microliters. Particles of 1.7 mu and smaller readily refluxed, but there was no sinusoidal reflux of casting material that contained particles of 10 mu. Casting without corrosion showed that the liver parenchyma remained intact. There was no evidence of reflux across hepatocytes. This study shows that cholangiovenous reflux occurs directly from bile ductules through the spaces of Mall and Disse into the hepatic sinusoids. The pathways measure between 1.7 and 10 mu. Since this is the path of least resistance, it may be of greater importance in the reflux of bacteria and toxins than other high-resistance pathways, for example, biliary canaliculi, tight junctions, or hepatocytes.