Chapter 21. Benign Gastric Disorders

The surgical management of benign gastric disorders has evolved significantly over the past 30 years. Elective surgery for ulcer disease has largely been abandoned in favor of medical management with surgery being utilized mainly for complications after failed medical treatment. Most elective (and some emergent) gastric procedures can now be performed with laparoscopy if local expertise is available, augmented by either radiologic (mainly via intra-operative ultrasound) or endoscopic guidance for more accurate localization. These techniques can help the surgeon perform a more targeted resection because wide margins are not necessary.

When Marshall and Warren elucidated the relationship between Helicobacter pylori and peptic ulcer disease, a discovery for which they were later awarded the Nobel Prize in Medicine, they rekindled the hypothesis that this common clinical malady was an infectious disease.1H. pylori is a gram-negative spiral flagellated organism that currently infects more than half of the people in the world. The prevalence of H. pylori infection varies among populations and is strongly correlated with socioeconomic conditions. In a number of developing countries, H. pylori infection affects more than 80% of middle-aged adults. Infection rates are lower in industrialized countries. Epidemiological data indicate that the prevalence of infection in the United States has been declining since the second half of the 19th century, with the decreases corresponding to improvements in sanitation. Nonetheless, H. pylori infection is predicted to remain endemic in the United States for the next century.

Human beings are the only reservoir for H. pylori. Infection is presumed to occur by oral ingestion of the bacterium. Direct transmission from person to person occurs via saliva and feces, and infection also occurs through contact with contaminated water. In developing countries, most individuals are infected during childhood. Family members are at increased risk of infection. A number of occupations also show increased rates of H. pylori infestation, notably health care workers. Infection with H. pylori is a chronic disease and does not resolve spontaneously without specific treatment.

H. pylori has evidently adapted to the hostile gastric environment and displays a number of features that permit its entry into the surface mucus layer, attachment to gastric epithelial cells, evasion of immune responses, and persistent colonization despite luminal acidity. Up to 15% of the protein in a helicobacter organism is composed of cytoplasmic urease that converts periplasmic urea into CO2 and ammonia, the latter buffering the surrounding acid.2

Host Response to H. Pylori

H. pylori infestation is followed by continuous gastric inflammation in virtually all individuals. Because spontaneous cure is unusual for most infected individuals, this means that H. pylori gastritis is a lifelong affliction. Worldwide, H. pylori–induced gastritis accounts for 80–90% of all gastritis.

H. pylori infection is not invasive of the gastric mucosa, and the host immune response is triggered by the attachment of bacteria to surface epithelial cells. The initial inflammatory response is characterized by recruitment of neutrophils, followed sequentially by T and B lymphocytes, plasma cells, and macrophages. The resultant chronic gastric inflammation in affected individuals is characterized by enhanced expression of multiple cytokines.

The relationship between H. pylori infection and ulceration is overwhelmingly strong; multiple observations establish H. pylori as a factor in the pathogenesis of duodenal ulceration.3–7 Most of the evidence is inferential. An effective vaccine has not yet been developed.

Observations that support H. pylori as a factor in the pathogenesis of human duodenal ulceration include the following:

1. H. pylori infection is invariably followed by the development of chronic gastritis, and the organism is the primary cause of chronic active gastritis worldwide. The infectious response to H. pylori is characterized by nonerosive inflammation of the gastric mucosa. Antral gastritis is present histologically in patients with duodenal ulcer, and H. pylori can be isolated from gastric mucosa of ulcer patients.

2. H. pylori binds only to gastric-type epithelium. Gastric metaplasia of the duodenal bulb is a nonspecific response to damage, which develops after infestation of the gastric mucosa. Antral gastritis with H. pylori is preceded by active chronic duodenitis. Metaplastic gastric epithelium is colonized by H. pylori from gastric sources. Gastric metaplasia is extremely common in duodenal epithelium surrounding areas of ulceration.

3. Eradication of H. pylori with antibiotics that have no effect on acid secretion leads to ulcer healing.

4. Therapy of peptic ulceration with bismuth compounds, which eradicate H. pylori, is associated with reduced rates of ulcer relapse relative to acid suppression therapy.

5. Relapse of duodenal ulcer after eradication of H. pylori is preceded by reinfection of the gastric mucosa by the organism.

However, infection by H. pylori alone does not cause peptic ulceration in most individuals, suggesting the existence of other pathogenetic factors. Half of patients evaluated for dyspepsia have histologic evidence of bacterial infection. In developed countries, one-fifth of healthy volunteers harbor the bacteria, and the incidence of bacterial infestation increases with age in the healthy, asymptomatic population. The occurrence of peptic ulcers in only a fraction of individuals who harbor the organism suggests that other factors must also act to induce ulceration.

H. pylori infection can be diagnosed by both invasive and noninvasive means. Noninvasive methods include the urea breath test, serology, and detection of antigen in stool samples. The urea breath test is based on production of urease by H. pylori in the gastric mucosa. C14-labeled urea is ingested and C14-labeled CO2 is produced and excreted in the breath. This test has a sensitivity and specificity of greater than 90% and indicates ongoing infection. The urea breath test is useful for initial diagnosis of infection and for follow-up after eradication therapy.

The stool antigen test is another noninvasive test to detect both initial H. pylori infection as well as response to treatment. Both polyclonal and monoclonal kits have been developed. Likewise, different kits are available for both out- and in-patient settings. Overall results have been comparable to those obtained using the urea breath test method.8

Because H. pylori induces a strong immunologic response, serological testing is useful but may not be as accurate as the urea breath test or the stool antigen test. Validation with either of the two tests is recommended. It may be used for epidemiologic studies. Because H. pylori–induced serology does not return to normal after bacterial eradication, this test is not reliable in monitoring therapy.

H. pylori infection can also be diagnosed on the basis of biopsies in patients undergoing upper endoscopic examination. Individuals older than 50 years, or those with significant symptoms including gastrointestinal (GI) bleeding, anemia, and weight loss, should undergo endoscopic diagnosis. During endoscopy, antral biopsies can be obtained and the organism cultured in agar containing both urea and a pH-sensitive colorimetric agent. H. pylori hydrolysis of urea causes a diagnostic change in color. The sensitivity of this test varies from 80 to 100% and specificity exceeds 90%. The test is associated with false-negative results in patients with active or recurrent bleeding and in those taking antibiotics or antisecretory compounds. Biopsy also permits histologic examination with visualization of the organism. Culture of H. pylori is not routine and is usually reserved for recurrent infection and for antibiotic sensitivity testing when second-line therapy has failed.

Complete eradication of H. pylori infection is the goal of treatment, and recurrence of disease signifies reinfection in most circumstances. An enormous worldwide experience has developed relating to H. pylori eradication. More than 2000 articles report the results of antibiotic trials, and a large number of summary articles and meta-analyses are available. It is important to note that none of the therapeutic regimens reported to date cure H. pylori infection in 100% of patients. To be effective, antimicrobial drugs must be combined with gastric acid secretion inhibitors or bismuth salts.

In the absence of treatment, eradication of H. pylori infection is very rare. Three consensus conference meetings as well as numerous clinical guidelines in various regions have been published in the past decade to further define the approach to diagnosis and treatment of H. pylori. The Maastricht III Consensus Report brought together a multidisciplinary group in 2007 from around the world to publish an update on the initial guidelines published in 1996 (known then as the European Helicobacter Study Group) and subsequently revised in 2000 after including a review of published guidelines from North America, Europe, China, and Japan (Table 21-1).9

Current evidence indicates that eradication therapy with a proton pump inhibitor, metronidazole, and amoxicillin decreases the prevalence of metronidazole-resistant H. pylori strains. The prevalence of clarithromycin-resistant strains varies greatly from country to country, with the highest rates reported in southern Europe. In this region, clarithromycin resistance now approximates 15%. This rate is predicted to rise over the next several years with increasing use of macrolide antibiotics. In patients failing therapy, culture of H. pylori from gastric mucosa is possible for resistance testing. Also, a recent multicenter trial in Spain demonstrated that a 10-day regimen of levofloxacin (500 mg BID), amoxicillin (1 g BID), and omeprazole (20 mg BID) had 97% compliance and 77% eradication based on a negative 13C urea breath test done 4–8 weeks after completion of therapy.10

Dyspepsia is a very common symptom complex that is characterized by pain and discomfort centered in the upper abdomen. Dyspepsia is among the most common disorders encountered by primary care physicians and gastroenterologists in the United States and Western countries. It is estimated that approximately 25% of the population will experience dyspepsia and that this problem accounts for 5% of visits to primary care providers.

Dyspepsia is characterized by symptoms that are focused in the upper abdomen. Symptoms may include heartburn, but a symptom complex limited to this complaint suggests gastroesophageal reflux disease and excludes the diagnosis of dyspepsia. Nonulcerative dyspepsia is considered when no anatomic or biochemical abnormality is discovered that explains the patient's symptoms. This common disorder is not associated with increased morbidity or mortality. However, it is generally long lasting and responsible for impaired quality of life. Investigation of nonulcerative dyspepsia and its treatment represent a large economic burden. Optimal treatment is controversial.

Since the description of H. pylori as a cause of gastritis, its association with nonulcerative dyspepsia has been disputed. H. pylori infestation is always associated with histologic gastritis, but may be absent in cases of nonulcerative dyspepsia. A large number of studies have reported the efficacy of therapy for H. pylori infection on symptoms of nonulcerative dyspepsia.11–13 A recent Cochrane review of the literature showed a small but statistically significant effect of H. pylori treatment in nonulcer dyspepsia but recommended further research before making any definitive recommendations.14

For surgeons, the importance of nonulcerative dyspepsia relates to its place in the differential diagnosis of epigastric pain. There is no role for surgery in the treatment of this disorder.

Epidemiology

Peptic ulcer disease is a major public health problem in the United States and a source of substantial health care expenditure.15 Overall, peptic ulcer mortality and hospitalization rates have declined for the past two decades from over 200,000 admissions in 1993 down to a little over 150,000 in 2006. Hemorrhage continues to be the most frequent presentation at admission, followed by perforation and obstruction. A significant shift was also seen in the management of ulcer hemorrhage from surgery (21% decrease) to endoscopy (59% increase). Although overall mortality rates decreased slightly (2.7%, down from 3.8%), no change was seen in the determinants of mortality, with perforation still being associated with the highest mortality, followed by obstruction and then bleeding. The mortality from surgical intervention decreased over the time period but remains high compared to endoscopy and embolization.16

In parallel with the discovery of H. pylori and the subsequent development of improved therapies for its eradication, surgical treatment of peptic ulcer has changed dramatically, with the virtual elimination of elective operations for ulcer disease. Operative therapy is now used mostly for emergent treatment of complicated disease. Antibiotics have become primary antiulcer therapy with the realization that, in most cases, peptic ulceration is an infectious disease. A wide variety of antisecretory drugs are available for clinical practice. Endoscopic and surgical therapies are frequently integrated in the care of individual patients.

Pathophysiology

The pathogenesis of peptic ulceration is multifactorial but increasingly understood to be a consequence of H. pylori infection. Before the recognition of the role of H. pylori, ulcer disease was conceived as an imbalance between acid and pepsin secretion and mucosal defense, with the balance shifted toward peptic injury and disease. In groups of patients, increases in acid secretion are well-documented, and, although gastric acid is crucial in the development of ulcers, an acquired defect in mucosal defense exists to tip the balance away from health. Mucosal infestation with H. pylori is the factor that contributes to ulceration in most patients; nonsteroidal anti-inflammatory drug (NSAID) use is the second most important factor in ulcer pathogenesis.

Substantial evidence implicates cigarette smoking as an additive risk factor in the development of duodenal ulcers. Smokers appear to have an increased risk of developing H. pylori infection relative to nonsmokers. Cigarette smoking impairs ulcer healing and increases the risk of recurrent ulceration. Continued smoking blunts the effectiveness of active ulcer therapy. Cigarette smoking increases both the probability that surgery will be required and the risks of operative therapy. When H. pylori is eradicated in smokers, they appear to have no greater risk of peptic ulceration than nonsmokers.17,18 This observation suggests that smoking is probably not an independent risk factor for ulcer disease but acts by increasing the harmful effects of bacterial infection. Cessation of smoking is a key goal of antiulcer therapy.

Abnormalities of gastric acid secretion in patients with peptic ulceration have been recognized for more than 50 years. The formation of duodenal ulcers clearly depends on gastric secretion of acid and pepsin. This association is emphasized by the dictum “no acid-no ulcer.” H. pylori infection is now known to secondarily induce alterations in gastric acid secretion as a prerequisite for ulcer development, and a more complete and accurate statement might be “no acid and no H. pylori–no ulcer.”

Abnormalities of mucosal function have been invoked as contributing factors to peptic injury. In support of this concept, several agents that are used to treat peptic ulceration are cytoprotective. Cytoprotective agents inhibit mucosal injury at concentrations lower than threshold doses that suppress acid secretion.19 The ability of such agents to heal ulcers suggests that abnormalities in mucosal defense, in addition to abnormalities in acid secretion, cause ulceration. Most cytoprotective agents act via mucosally secreted bicarbonate or on mucosal prostaglandin production.

NSAIDs are a major risk factor for the development of acute ulceration and for hemorrhagic complications of ulceration. NSAIDs produce a variety of lesions, ranging from superficial mucosal erosions to deeper ulcerations. While the mucosal injury caused by NSAIDs is more common in the stomach than in the duodenum, ulcer complications occur with equal frequency in these two sites. H. pylori and NSAID use independently increase the risk of peptic ulcer and ulcer bleeding. These agents also act synergistically. In the duodenum, it appears likely that invasive H. pylori–associated ulcers are compounded by the direct injurious effects of NSAIDs.

The injurious actions of NSAIDs are secondary to systemic suppression of prostaglandin production. Numerous experimental models have demonstrated that NSAIDs injure the gastroduodenal mucosa. Ulcers resembling those occurring in humans can be produced by administration of NSAIDs to animals, and NSAID-associated gastric ulcers can be prevented by the coadministration of prostaglandin analogues. Ulcers associated with NSAIDs heal rapidly when the drug is withdrawn, corresponding temporally to reversal of antiprostaglandin effects.

None of the currently available NSAIDs are free of the hazard of gastroduodenal ulceration.19 Clinically significant ulceration of the stomach and duodenum is estimated to occur at a rate of 2–4% per patient-year. The risks of long-term NSAID use are increased by H. pylori infection and cigarette smoking. The incidence of NSAID-related ulcer complications is highest in older patients, as is attendant mortality rate. Peptic ulcer disease is rare in individuals who are H. pylori–negative and who do not receive NSAID medications.20

Diagnosis

Duodenal ulceration is characterized by epigastric pain. The pain is usually localized to the upper abdomen without radiation and is described as burning, stabbing, or gnawing. In the absence of complications such as perforation or penetration into the head of the pancreas, referral of pain to extra-abdominal sites is not common. Many patients report that pain is worsened by fasting. Ingestion of antacids usually provides prompt relief. In uncomplicated cases, physical examination is usually normal.

The differential diagnosis includes a variety of diseases originating in the epigastrium and upper GI tract. Common disorders to be distinguished include nonulcer dyspepsia, gastritis, gastric neoplasia, cholelithiasis and related diseases of the biliary system, neoplastic lesions of the liver, and both inflammatory and neoplastic disorders of the pancreas. In dyspeptic patients, especially those older than 50 years of age, the most important differential diagnoses are peptic ulceration and gastric cancer.

The evaluation of patients with suspected peptic ulceration usually involves endoscopic examination of the esophagus, stomach, and duodenum. In most circumstances, contrast radiography is not the preferred initial diagnostic method; endoscopy has become the standard to which other modalities are compared. Endoscopy eliminates the need for radiation, is safe, is tolerated by elderly patients, and permits both visual inspection and biopsy of the esophagus, stomach, and duodenum. In controlled trials, endoscopy was both more sensitive (92 vs 54%) and more specific (100 vs 91%) than radiographic examination.21 Endoscopy must be utilized with discretion because of the potential for perforation (approximately 1 per 5000 cases) and cost.

Endoscopically, duodenal ulceration is characterized by lesions that are erosive to the intestinal wall. When viewed endoscopically, peptic ulcers have a typical appearance, with edges that are usually sharply demarcated. The ulcer consists of the exposed underlying submucosa. With chronic ulcers, the base is usually clean and smooth. Acute ulcers and ulcers with recent hemorrhage may demonstrate clot, eschar, or adherent exudate. The surrounding duodenal mucosa may be friable, but marked inflammation is uncommon. The most frequent site for peptic ulceration is the first portion of the duodenum, with the second portion less frequently involved. Peptic ulceration of the third or fourth portions of the duodenum is distinctly unusual; occurrence of ulcers in these locations raises the possibility of gastrinoma. Peptic ulcers in the pyloric channel or the prepyloric area are similar in appearance to duodenal ulcers. Endoscopic demonstration of a duodenal ulcer does not require duodenal biopsy but should prompt mucosal biopsy of the gastric antrum to demonstrate the presence of H. pylori and guide subsequent therapy.

Contrast radiographs demonstrate retention of contrast within the ulcer. When viewed tangentially, the ulcer projects beyond the level of the duodenal mucosa. Distortion of the duodenal bulb by spasm or scarring is a secondary sign of current or previous ulceration.

The realization that peptic ulceration is an infectious disease has fundamentally altered the role of surgery in ulcer treatment.22,23 Indications for operative intervention have changed over the past 20 years as a consequence, with the virtual elimination of elective operations.24 Operative intervention is now reserved for the treatment of complicated ulcer disease. Three complications are most common and constitute contemporary indications for peptic ulcer surgery: hemorrhage, perforation, and obstruction. Evolving indications are also reflected in the forms of operative therapy and in surgical training experience.25,26

The first goal of current surgical therapy is treatment of anatomic complications, such as pyloric stenosis or perforation. The second major goal should be patient safety in the acute setting, combined with freedom from undesirable chronic side effects. The third goal in contemporary surgical treatment of complicated ulcer disease should be alteration of the ulcer diathesis so that ulcer healing is achieved and recurrence is minimized. To achieve these goals, the gastric surgeon can combine therapy through endoscopic, radiologic, or operative means, the appropriate choice depending on the clinical circumstances.

Operative Procedures

There is currently no indication for surgical treatment of uncomplicated ulcer disease. A number of operative procedures have been developed to treat peptic ulcer but have been used with decreasing frequency in the past decade. Operative treatment of gastric outlet obstruction has decreased by approximately 50%. The majority of surgical patients are currently treated emergently for the complications of bleeding or perforation.

Truncal vagotomy and drainage, truncal vagotomy and antrectomy, and proximal gastric vagotomy are the most widely utilized procedures in the operative treatment of peptic ulcer disease. However, surgical therapy of complicated peptic ulcer disease is directed increasingly at correction of the immediate problem without gastric denervation. The underlying cause of the ulcer diathesis may then be addressed after recovery from surgery by antibiotic therapy directed at H. pylori and by long-term acid suppression therapy. This approach is applicable to most patients with peptic ulcer undergoing emergent operation and is also reflected by the fact that the use of gastrectomy and vagotomy has decreased significantly from 4.4 to 2.1% (gastrectomy) and 5.7 to 1.7% (vagotomy) over the last two decades.16

Transection of both vagal trunks at the esophageal hiatus, termed truncal vagotomy, denervates the acid-producing fundus of the stomach. The procedure also denervates the remainder of the supplied viscera, including the liver and biliary tree, pancreas, small bowel, and colon to the midtransverse portion. Because denervation impedes normal pyloric coordination and impairs gastric emptying, truncal vagotomy is usually combined with a procedure to eliminate or bypass pyloric sphincter function. A pyloroplasty or gastrojejunostomy is performed for gastric drainage.

Several methods of pyloroplasty have been developed. The Heineke-Mikulicz pyloroplasty (Fig. 21-1) consists of a longitudinal incision of the pyloric sphincter extending into the antrum and the duodenum. The incision is closed transversely, eliminating sphincteric closure and increasing the lumen of the pyloric channel.

The Finney pyloroplasty (Fig. 21-2) extends the pyloric incision 5 cm onto the duodenal wall forming an inverted U-shaped incision after the placement of superior and inferior traction sutures. Once traction is applied, the two limbs of the inverted U-shaped incision are lined up and sutured to each other to complete the procedure, with the inferior suture line forming the posterior wall and the superior suture line forming the anterior wall of the pyloroplasty.

A Jaboulay gastroduodenostomy (Fig. 21-3) requires more extensive dissection beginning with a Kocher maneuver followed by corresponding incisions on the stomach and the duodenum proximal and distal to the pylorus respectively. Traction sutures are then placed between the stomach and duodenum to approximate the two incisions, and the anastomosis is then performed.

Truncal vagotomy can be combined with resection of the gastric antrum to further reduce acid secretion by removing antral sources of gastrin. The limits of antral resection are defined by external landmarks. The stomach is divided proximally along a line from a point above the incisura angularis to a point along the greater curvature midway from the pylorus to the gastroesophageal junction. Reconstruction via a gastroduodenostomy is called a Billroth I procedure. A Billroth II procedure uses a gastrojejunostomy to restore GI continuity.

Proximal gastric vagotomy, also termed highly selective vagotomy (HSV), differs from truncal vagotomy in that only the nerve fibers to the acid-secreting fundic mucosa are transected (Fig. 21-4). The hepatic and celiac divisions are not divided, and vagal nerve fibers to the antrum and pylorus remain intact. The operation has also been called parietal cell vagotomy to emphasize the intended functional consequence.

Proximal gastric vagotomy is a safe operation. The procedure has a reported operative mortality rate of less than 0.05%, lower than the reported mortality for any other gastric procedure for peptic ulcer. Truncal vagotomy and pyloroplasty has an accepted mortality rate of 0.5–0.8%, whereas mortality after truncal vagotomy and antrectomy approximates 1.5%. These statistics require an important caveat; almost all large series report the results of elective operations on patients with peptic ulceration and may not accurately reflect expected results when similar procedures are performed emergently.

Postoperative Alterations

Division of vagal nerve fibers alters gastric acid secretion by reducing cholinergic stimulation of parietal cells. Vagal denervation also decreases parietal cell responsiveness to gastrin and histamine. Basal acid secretion is diminished by approximately 80% in the immediate postoperative period and is maintained over time. The maximal acid output in response to secretagogues such as pentagastrin is reduced by approximately 70%. After 1 year, pentagastrin-stimulated maximal acid output increases to 50% of prevagotomy values but remains at this level on subsequent testing. Acid secretion due to meal stimulation is reduced by 60–70% relative to normal subjects. The inclusion of antrectomy to truncal vagotomy further reduces acid secretion. Maximal acid output is reduced by 85% relative to values recorded before antrectomy.

Both forms of vagotomy cause postoperative hypergastrinemia. Fasting gastrin values are increased to approximately twice preoperative levels. Postprandial gastrin response is exaggerated. Hypergastrinemia is due to decreased luminal acid, with loss of feedback inhibition of gastrin release. Chronic hypergastrinemia is caused by mucosal gastrin cell hyperplasia in addition to loss of inhibitory feedback. When antrectomy is performed, circulating gastrin levels are decreased. Basal gastrin values are reduced by approximately half and postprandial gastrin levels by two-thirds.

Operations that involve vagotomy affect gastric emptying. Both truncal vagotomy and proximal gastric denervation abolish vagally mediated receptive relaxation that normally allows the ingestion of a meal with no increase in intragastric pressure. After vagotomy, the intragastric pressure rise is greater for any given volume ingested, and the gastroduodenal pressure gradient higher than in normal subjects. As a result, emptying of liquids, which depends on the gastroduodenal pressure gradient, is accelerated. Because nerve fibers to the antrum and pylorus are preserved with proximal gastric vagotomy, the function of the distal stomach to mix solid food is preserved and emptying of solids is nearly normal. Truncal vagotomy affects the motor activity of the distal stomach, and solid and liquid emptying rates are usually increased when truncal vagotomy is accompanied by pyloroplasty.

Dumping is defined by a postprandial symptom complex of abdominal discomfort, weakness, and vasomotor symptoms of sweating and dizziness. Dumping occurs transiently in 10–15% of patients after truncal vagotomy and antrectomy and is persistent in 1–2%. Dumping is present initially in 10% of patients undergoing truncal vagotomy and pyloroplasty, and remains in approximately 1%. Permanent symptoms of dumping are unusual after proximal gastric vagotomy. The incidence of diarrhea, presumably caused by denervation of the pylorus and small bowel and by elimination of pyloric function, parallels the incidence of dumping when truncal vagotomy is performed. Persistent and disabling diarrhea is present in fewer than 1% of patients after proximal gastric vagotomy.

The largest surgical series examining ulcer recurrence rates were reported at a time before the pathogenic role of H. pylori was appreciated. With appropriate use of antibiotics directed against H. pylori, ulcer recurrence rates as low as 0.22% have been reported.27 Although recurrence rates (without H. pylori treatment) as low as 5% have been reported, a more generally representative figure is 10%. This rate is similar to that of reinfection with H. pylori after successful eradication. Ulcer recurrence rates after proximal gastric vagotomy can be adversely affected by the inclusion of prepyloric and pyloric channel ulcers. Proximal gastric vagotomy is significantly less effective when used to treat ulcers in this position than when used for duodenal ulceration.

Ulcer Hemorrhage

Hemorrhage continues to be a major source of morbidity in patients with peptic ulceration. Bleeding is the leading cause of death associated with peptic ulcer. The incidence of hemorrhage has not changed since the introduction of H2-receptor antagonists.23 The lifetime risk of hemorrhage for patients with duodenal ulcer who do not undergo specific therapy approximates 35%. Hemorrhage usually occurs during the initial episode of ulceration or during relapse; patients who have bled previously have a higher risk of bleeding again. Patients with recurrent bleeding and elderly patients are at greatest risk of death.25,28

The risk of mortality from bleeding peptic ulcer is surprisingly high at 10–20%. When surgery is necessary, operative risk is increased in patients who have shock at admission, recurrent bleeding, delay in operative intervention, or comorbid illnesses. Surgical delay leads to recurrent hypovolemia and subsequently multisystem organ failure.

Upper GI endoscopy is the appropriate initial diagnostic test, following resuscitation, when hemorrhage from ulceration is suspected. Endoscopy identifies the site and source of bleeding in over 90% of patients. An ulcer should be accepted as the bleeding source only if it exhibits stigmata of active or recent hemorrhage (Table 21-2). Active hemorrhage is defined by an arterial jet, active oozing, or oozing beneath an adherent clot. Signs of recent hemorrhage include adherent clot without oozing, adherent slough in the ulcer base, or visible vessel in the ulcer. Up to 30% of patients who have stigmata of recent hemorrhage experience rebleeding, and most of the patients who bleed recurrently require emergency treatment. The signs are not sufficiently accurate, nor are rebleeding rates high enough, to be indications for surgery. Endoscopic stigmata indicate that aggressive therapy is needed and close follow-up mandatory. The occurrence of hypovolemic shock, rebleeding during hospitalization, and a posteroinferior location of the ulcer are clinical features that are associated with increased risk of recurrent bleeding. Acute reduction of acid secretion by H2-receptor antagonists or proton pump inhibitors is not sufficient to prevent recurrent hemorrhage. However, the continuous infusion of proton pump inhibitors has been shown to decrease rebleeding.26

The ability to visualize bleeding duodenal ulcers endoscopically permits endoscopic treatment. Methods of endoscopic therapy include thermal coagulation by bipolar electrocoagulation or direct application of heat through a heater probe.29 Injection of epinephrine into the base of the bleeding ulcer is also an established method to control ulcer hemorrhage.

Both reduced rebleeding rates and avoidance of operation have been demonstrated for endoscopic hemostasis.29,30 Proof of efficacy for endoscopic treatment of hemorrhage is complicated by the 70% rate of spontaneous, sometimes temporary, cessation of bleeding without intervention (Table 21-3). In addition to endoscopic stigmata, hemodynamic instability, continuing transfusion requirements, red stool or hematemesis, age older than 60 years, and medical comorbidity are clinical features that mandate endoscopic therapy. Rebleeding during hospitalization and the endoscopic findings of visible vessel, oozing, or bleeding associated with an adherent clot are also indications for endoscopic hemostasis. Ulcers with clean bases and no stigmata of recent hemorrhage require no treatment.

Failure of endoscopic treatment is usually due to inaccessibility of the ulcer that is caused by pyloric scarring, rapid active bleeding, or an obscuring clot. Patients treated endoscopically should be observed closely for further hemorrhage. Those who rebleed within 72 hours of initial endoscopic control may be successfully retreated without increased risk of mortality.30

The efficacy of endoscopy diagnosis and therapy depends on timing. Early endoscopy correctly classifies patients as low risk for recurrent hemorrhage and permits safe avoidance of hospitalization. Early endoscopy also benefits high-risk patients by directing specific, active hemostatic therapy. Patients with early endoscopy have been demonstrated to have fewer episodes of rebleeding, lower rates of operation, less resource consumption, and shorter hospitalizations.31

Operative intervention is indicated for the following:

• Massive hemorrhage leading to shock or cardiovascular instability
• Prolonged blood loss requiring continuing transfusion
• Recurrent bleeding during medical therapy or after endoscopic therapy
• Recurrent hemorrhage requiring hospitalization

The need for emergency intervention significantly increases surgical risks, and not surprisingly, mortality is increased 10-fold. Emergent operative therapy should consist of duodenotomy with direct suture ligation of the bleeding vessel in the ulcer base (Table 21-4). Postoperatively, patients should receive proton pump inhibitors and antibiotics directed against H. pylori. This treatment approach is based on the observation, in medically treated patients, that peptic ulcer hemorrhage recurs in 20% of patients when H. pylori is not eradicated, while rebleeding is reduced to 3% in patients who receive H. pylori antibiotic therapy.31–34 This recommendation is an extrapolation; the studies that support this practice were not designed to evaluate postoperative hemorrhage (Table 21-5).

The lifetime risk for perforation in patients with duodenal ulceration not receiving therapy approximates 10%, while ulcer perforation is unusual if initial ulcer healing has been achieved. Duodenal ulcer perforation is followed by sudden and severe epigastric pain. The pain is caused by contact of the peritoneum with highly caustic gastric secretions. Pain is often instantaneous and remains constant. Peritoneal irritation is usually intense and causes most patients to avoid movement.

Physical examination reveals fever, diminished bowel sounds, rigidity of the abdominal musculature, and guarding. Upright abdominal radiographs demonstrate pneumoperitoneum in 80% of cases. If free air is not present, computed tomography of the abdomen is very sensitive for demonstrating perforation.

Occasional reports have described nonoperative treatment of this complication, but this approach is not appropriate for the large majority of patients with perforated peptic ulcer. Perforation is a strong indication for surgery in most circumstances. Laparotomy or laparoscopy affords the opportunity to relieve intraperitoneal contamination and to close the perforation.

The results of surgical treatment of duodenal perforation in the era preceding the recognition of H. pylori are instructive. Signs of preexisting duodenal ulceration, in terms of history of prior symptoms and anatomic evidence of duodenal scarring, should be sought, but a lack of antecedent symptoms is not protective. Patients without antecedent symptoms are at substantial risk for recurrent ulceration. By 5–6 years, symptomatic ulcer recurrence in patients with acute ulcer perforation is similar to that for patients with chronic disease. Before the role of H. pylori was appreciated, simple omental closure of duodenal perforation had not provided satisfactory long-term results; up to 80% of patients so treated had recurrent ulceration and 10% experienced reperforation. It is now known that four-fifths of all patients with perforation have H. pylori infestation and therefore are at risk of recurrent disease.

The mortality of emergency operation for ulcer perforation is most clearly correlated with the existence of preoperative shock, coexisting medical illness, and the presence of perforation beyond 48 hours.35 For stable patients who receive prompt surgical attention, the operation can be performed with safety.36 Proximal gastric vagotomy with omental patch closure of the perforation is one option in this circumstance. This procedure has been shown to be both safe and effective in preventing ulcer relapse. Incorporation of the site of perforation as part of a pyloroplasty or resection of the perforation during antrectomy can also be combined with truncal vagotomy. The performance of these operations has declined significantly, however, with the focus on H. pylori as the cause of most ulcer recurrences.

Several investigators advocate omental patch closure alone with postoperative anti–H. pylori therapy.37–41 Omental patching can also be accomplished laparoscopically in select patients.42 This approach rests upon three assumptions: (1) that most perforated duodenal ulcers are caused by H. pylori; (2) that the duodenal perforation is small enough that secure closure can be obtained; and (3) that further surgical therapy will be obviated by the effects of postoperative antibiotic therapy and acid suppression. Minimally invasive approaches are becoming frequently performed.

Gastric outlet obstruction can develop either acutely or chronically in patients with duodenal ulcer disease. Surprisingly, the incidence of H. pylori infection in this subgroup of patients may not be as high as that seen in patients presenting with hemorrhage or perforation.43 Acute obstruction, due to edema and inflammation, is associated with ulcers in the pyloric channel and the first portion of the duodenum. Pyloric obstruction causes recurrent vomiting and dehydration. Hypochloremic alkalosis, due to loss of hydrochloric acid in gastric secretions, is distinctive of gastric obstruction. Hypokalemia may develop as a secondary renal compensation for alkalosis. Acute gastric outlet obstruction is treated by nasogastric suction, rehydration, and intravenous administration of antisecretory agents. Acute obstruction due to pyloric inflammation resolves with supportive measures within a few days.

Repeated episodes of ulceration can lead to pyloric scarring and a fixed stenosis with chronic gastric outlet obstruction. In cases of recurrent duodenal ulceration, the lifetime risk of chronic pyloric stenosis approximates 10%.

Initial investigation begins with upper GI endoscopy to confirm the site of obstruction and to exclude intrinsic or extrinsic obstruction due to malignancy, the most common cause of gastric outlet obstruction in the modern era. Endoscopic balloon dilation of the area of peptic ulcer obstruction can also be attempted; with success obtained in up to 85% of patients.44 Most treated patients note immediate symptomatic improvement, but only 40% have sustained improvement by 3 months after balloon dilation. Recurrent symptoms are presumed due to residual scarring in the pyloric channel. In most cases, operative correction is required. This should include treatment of the underlying ulcer disease and relief of the anatomic abnormality. Truncal vagotomy with antrectomy and parietal cell vagotomy with gastrojejunostomy have both been used with success in this circumstance.

Diagnosis

In the United States, benign gastric ulcers are found in approximately 90,000 new patients a year, about one-fifth that of duodenal ulceration. The opposite is found in Japan where gastric ulcers are 5–10 times more common. Gastric ulcer is more common in men than women and occurs in a patient cohort approximately 10 years older than that of duodenal ulceration. In symptomatic patients, upper GI endoscopy is the preferred method for diagnosing gastric ulceration. The visual appearance of benign and malignant gastric ulcers may be identical, and differentiation may be made only by biopsy. Benign gastric ulcers appear smooth and flat and are often covered by a gray, fibrous exudate. The margin is often raised and erythematous. The ulcer margin is friable and may bleed with manipulation. All gastric ulcers should undergo multiple biopsies, obtained from the perimeter of the lesion. The addition of endoscopic brushings to multiple biopsies increases diagnostic accuracy to approximately 95%.

Although benign gastric ulcers may occur in any location in the stomach, more than half are located along the lesser curvature proximal to the incisura angularis. Fewer than 10% of benign ulcers are located on the greater curvature. Most benign gastric ulcers lay within 2 cm of the histologic transition between fundic and antral mucosa.

Similar to duodenal ulceration, H. pylori infection is the key to the pathogenesis of benign gastric ulcers. Antibiotic treatment regimens useful for duodenal ulcer have also been used for benign gastric ulceration. The response of gastric ulcers to antibiotic therapy is equivalent to that of duodenal ulcers. Recurrence of gastric ulcers after H. pylori eradication is equal to the rate of reinfection.

In addition to H. pylori infection, alterations in gastric motility have been demonstrated in some patients with benign gastric ulcers. Motility defects include delayed gastric emptying, abnormal pyloric sphincter function, prolonged high-amplitude gastric contractions, duodenogastric reflux, and alterations in the gastric migrating motor complex. These alterations have not been definitively demonstrated to be pathogenic, and their relevance to gastric ulceration is unsettled. A definite association between chronic NSAID use and benign gastric ulceration has been recognized. As with duodenal ulceration, cigarette smoking is associated with development of gastric ulceration, and continued smoking impairs medical therapy. Gastric and duodenal ulcers may occur in patients who receive hepatic artery chemotherapy if improper placement of the catheter permits perfusion of gastric and duodenal mucosa. A variety of agents, including 5-fluorouracil, cisplatin, doxorubicin, and mitomycin C, have been implicated.

Therapy

The primary therapy for benign gastric ulceration is antibiotic treatment of H. pylori infection using treatment protocols similar to those for duodenal ulceration. Antibiotic response rates are similar. Cessation of NSAID therapy is required to improve results. Operative treatment is reserved for complications of gastric ulcer, including hemorrhage and perforation. Unlike duodenal ulcer, failure of a recurrent ulcer to respond to medical therapy may be an indication for operation, usually because nonhealing raises concerns about malignant disease.

For benign gastric ulcers, the elective operation of choice is usually a distal gastrectomy with either gastroduodenal (Billroth I) or gastrojejunal (Billroth II) anastomosis. The ulcer should be excised with the gastrectomy specimen (Fig. 21-5). Performed electively, operative mortality approximates 2–3%, and ulcer recurrence rates are less than 5%. Inclusion of vagotomy does not improve recurrence rates, which is not surprising given the variability of acid secretion in patients with gastric ulcers. The occurrence of a benign ulcer near the gastroesophageal junction (type IV ulcer) represents a difficult surgical problem. The ulcer may be excised via a distal gastrectomy with an extension along the lesser curvature and reconstruction with gastrojejunostomy. Emergency treatment of hemorrhage or perforation requires ulcer excision. Distal gastrectomy, including the site of perforation or bleeding, is usually the procedure of choice. Operative mortality rates average 10–20% in the presence of hemorrhage or perforation.

Intractability or Nonhealing Ulcers

This should indeed be a rare indication for surgery performed today. Arguably, the patient referred for surgical evaluation of intractable peptic ulcer disease should raise red flags for the surgeon. Acid secretion can be totally blocked and H. pylori eradicated with modern medication; therefore, the question remains: “Why does the patient have a persistent ulcer diathesis?” The surgeon should review the differential diagnosis of nonhealing ulcer prior to any consideration of operative treatment (Table 21-6).

Surgical treatment should be considered in patients with nonhealing or intractable peptic ulcer disease who have multiple recurrences, large ulcers (>2 cm), complications (obstruction, perforation, or hemorrhage), or suspected gastric cancer. Surgery should be approached most cautiously in the thin or marginally nourished individual.

It is important that the surgeon not fall into the trap of performing a large, irreversible operation on these patients, based on the unproven theory that if all other methods have failed, a larger operation is required. Today's patients are different than those of three or four decades ago. One might argue that modern medical care has healed the minor ulcer, and that patients presenting with true intractability or nonhealing will be more difficult to treat and are likely to have chronic problems after a major ulcer operation. If surgery is necessary, less is often better. It is the practice of the authors never to perform a gastrectomy as the initial elective operation for intractable duodenal ulcer in the thin or asthenic patient. Instead, the preferred operation for this group of patients is HSV. In patients with nonhealing gastric ulcer, wedge resection with HSV should be considered in thin or frail patients. Otherwise distal gastrectomy (to include the ulcer) is recommended. It is unnecessary to add a vagotomy in patients with type I gastric ulcer.

Juxtaesophageal gastric ulcers (type IV) are pathophysiologically akin to type I gastric ulcers (ie, associated with gastric acid hyposecretion) but are often difficult to resect as part of a distal gastrectomy. A variety of techniques have been used to treat these ulcers surgically, including the Csendes operation, the Pauchet gastrectomy, and the Kelling-Madlener procedure (Fig. 21-6).

A number of syndromes have been described after gastric operations performed for peptic ulceration as well as gastric neoplasm. The occurrence of permanent disabling postoperative symptoms is uncommon, occurring only in about 1–3% of cases, and unpredictable. The two most common postgastrectomy syndromes are dumping and alkaline reflux gastritis.

Dumping

Dumping is defined as a postoperative clinical syndrome with gastrointestinal and vasomotor symptoms. The cause of dumping is uncertain but is likely related to unregulated entry of ingested food into the proximal small bowel following resection, bypass, or division of the pyloric sphincter. Early dumping symptoms occur within 1 hour of ingestion of a meal and include nausea, epigastric discomfort, tremulousness, and sometimes dizziness or syncope. Late dumping symptoms follow a meal by 1–3 hours. Late symptoms are usually due to reactive hypoglycemia.

Most patients who undergo vagotomy or gastrectomy do not experience dumping symptoms postoperatively. For patients who experience mild dumping symptoms in the early postoperative period, dietary alterations, and time bring improvement in all but approximately 1–2%. For those who remain persistently symptomatic, the long-acting somatostatin analogue, octreotide, improves dumping symptoms when administered subcutaneously before a meal.45 The effects of somatostatin on the vasomotor symptoms of dumping are summarized below (Table 21-7).

Alkaline Reflux Gastritis

Alkaline reflux gastritis is a postoperative syndrome characterized by postprandial epigastric pain associated with nausea and bilious vomiting. Endoscopic examination reveals reflux of bile into the stomach, and biopsy demonstrates histologic evidence of gastritis.

Alkaline reflux gastritis is a diagnosis of exclusion. The differential diagnosis of postoperative epigastric pain includes recurrent ulceration, calculous biliary disease, pancreatic inflammation, afferent loop obstruction, and esophagitis. Upper endoscopic examination is essential to exclude recurrent ulcer. The gastric mucosa appears inflamed, friable, and edematous. Gastric inflammation is often uneven and nonulcerative. Histologic examination shows glandular atrophy, mucosal and submucosal edema, and the presence of acute and chronic inflammatory cells in the lamina propria. Intestinal metaplasia may be present.

Postoperative alkaline reflux gastritis is resistant to medical treatment. Antacids, proton pump inhibitors, and dietary manipulations have not been definitively demonstrated to be beneficial. The most effective treatment for persistent alkaline reflux gastritis is operative diversion of intestinal contents from contact with the gastric mucosa. This solution usually requires conversion of a Billroth I or II gastrectomy to a Roux-en-Y gastrojejunostomy with an intestinal limb of 50–60 cm (Fig. 21-7). The length of the Roux limb prevents reflux of intestinal contents. This procedure is very effective in eliminating bilious vomiting. However, persistent pain is reported in up to 30% of patients, and 20% of patients develop postoperative delayed gastric emptying.

Gastritis and gastric ulceration can be induced by physiologic stress. Usually occurring in hospitalized patients with critical illness, stress gastritis can be demonstrated endoscopically in the majority of patients recovering from shock. While occult bleeding in this population is common, clinically significant hemorrhage defined by the need for blood transfusion, hypotension, or alteration in other vital signs occurs in only 0.5–5% of patients. In four recent surgical series comprising more than 28,000 patients, the incidence of clinically significant stress ulceration was 0.4%.46 In another series of 16,612 hospitalized patients, the incidence of overt stress bleeding was only 0.1%.47 In a review of patients admitted to both surgical and medical intensive care units (ICUs), the incidence of clinically significant and endoscopically proven stress ulceration was 0.17%.48

Major trauma, especially if accompanied by hypotension, sepsis, respiratory failure, hemorrhage, or multiple injuries, predisposes to acute stress gastritis (Table 21-8). Acute stress gastritis is also common after thermal injury with greater than 35% total body surface area burned. A form of gastritis similar to that following trauma may complicate central nervous system injury or intracranial hypertension. When viewed endoscopically, multiple ulcerations are observed in the proximal, acid-secreting portion of the stomach. Fewer lesions are found in the antrum, and only rare ulcerations in the duodenum.

The major complication of stress gastritis is hemorrhage. Patients with coagulopathy and those requiring mechanical ventilation are at increased risk of hemorrhage. Patients without these two risk factors have been reported to have an overall risk of hemorrhage of only 0.1%, while those with both demonstrate clinically significant bleeding in 3.7% of cases. Respiratory failure is defined as greater than 48 hours on a mechanical ventilator. Coagulopathy is defined as a platelet count less than 50,000/mm3, an international normalized ratio greater than 1.5, or a partial thromboplastin time greater than two times control.

Admission to an ICU by itself does not place patients at risk for hemorrhage, and patients undergoing major GI surgery do not have an increased risk of stress-related bleeding in the absence of complications. Increased patient age, emergency surgery, need for reoperation, and the occurrence of hypotension are risk factors for postoperative gastric bleeding. The occurrence of sepsis and respiratory failure are also risk factors. Multiple regression analysis has shown that mechanical ventilation and coagulopathy impart the greatest risk.

The diagnosis of stress ulceration requires endoscopic examination. Acute mucosal ulcerations may be observed as early as 12 hours postinsult; lesions appear as multiple shallow areas of erythema and friability, accompanied by focal hemorrhage. Histologically, the lesions consist of coagulation necrosis of the superficial endothelium with infiltration of leukocytes into the lamina propria. Signs of chronicity, such as fibrosis and scarring, are absent. With resolution of injury or sepsis, healing is accomplished by mucosal restitution and regeneration.

A survey of Society of Critical Care Medicine members showed that ranitidine, famotidine, sucralfate, and cimetidine were the drugs used most commonly for prophylaxis. The presence of bright red blood in the nasogastric tube was considered by most to define prophylaxis failure, and the addition of a second drug from a different therapeutic class was the preferred mode of treatment.49

Because hemorrhage does not occur in all patients, studies that use bloody nasogastric discharge as a sign of stress gastritis underestimate the true incidence in critically ill patients. In one endoscopically controlled study, 100% of patients with life-threatening injuries had evidence of gastric erosions by 24 hours. A high prevalence of gastric erosions is also noted in burn patients, while GI hemorrhage occurs in only 25–50% of patients with burn wound infection. Barium contrast examinations have no role in the diagnosis of stress gastritis and interfere with endoscopic examination.

It is important to distinguish stress ulceration from other causes of postoperative hemorrhage. Several recent studies have demonstrated that duodenal ulceration and gastric ulcers are common in postoperative patients. In one series, sources of clinically significant bleeding included duodenal ulcer in 26%, gastric ulcer in 13%, esophagitis in 18%, and esophageal varices in 7%. Similar results have been reported in other series, emphasizing the need for specific diagnosis.

Acute upper gastrointestinal (GI) hemorrhages are frequent medical events occurring at a rate of approximately 50 cases per 100,000 persons per year. Acute GI hemorrhage still has a significant associated mortality, approximating 10%. Although urgent endoscopy has been used for the past 20 years for the diagnosis and management of acute upper GI hemorrhage, the mortality rate has not substantially declined even with the introduction of endoscopic intervention. Patients with acute upper GI hemorrhage are increasingly of advanced age and have preexisting medical comorbidities.

Endoscopy has become the preferred method for diagnosis in patients with acute upper GI bleeding. This method is informative in most patients, correctly identifying the site and source of bleeding in 90% of cases. While the efficacy of upper endoscopy has been established for diagnosing acute upper GI tract hemorrhage, optimal timing has been controversial. The majority of existing studies support the claim that early endoscopy is both safe and effective for all risk groups.

For low-risk patients, the current evidence demonstrates that early endoscopy promotes safe patient disposition. In many instances, these patients can avoid hospitalization with a very low risk of recurrent bleeding. For high-risk patients, there is benefit of early endoscopy for outcomes, including transfusion requirements, rebleeding rate, and the need for emergency surgery. Early endoscopy directs therapy and significantly reduces length of hospitalization relative to delayed endoscopy without evidence of cost shifting to the outpatient setting. In this sense, early endoscopy provides prompt diagnosis and assists in decision making regarding clinical triage and subsequent management. Current evidence does not demonstrate, however, that early endoscopy decreases overall mortality. There is no evidence that the practice of early endoscopic intervention results in patient harm.

Based on current information, gastroduodenal ulceration accounts for approximately 40% of cases of acute upper GI hemorrhage. Other diagnoses, in decreasing frequency, include acute gastritis, esophageal variceal bleeding, esophagitis, duodenitis, Mallory-Weiss tears, and upper GI tract malignancies.

Initial treatment of upper GI tract hemorrhage begins with restoring intravascular volume. Hemodynamic monitoring is crucial. Unstable patients should be initially treated in an ICU setting. Interestingly, although numerous trials have examined the efficacy of H2-receptor antagonists in patients with bleeding peptic ulcers, none of these have demonstrated consistent therapeutic benefit either individually or when examined by meta-analysis. Sixteen prospective trials have also examined the use of proton pump inhibitors in the setting of acute ulcer bleeding. Only 7 of these 16 trials have demonstrated a statistically significant benefit in terms of rebleeding or need for urgent surgical intervention. None of the trials showed a reduction in mortality. Over half (9 of 16) of the studies did not demonstrate reduction in any of the primary outcomes that included rebleeding, surgery, or mortality.

In addition to providing diagnostic information, aggressive endoscopy also presents an opportunity for therapeutic intervention. Relative to medical therapy alone, patients with stigmata of active bleeding, visible vessels, and nonbleeding adherent clots benefit from endoscopic ulcer hemostasis. The major modalities used include bipolar electrocautery probes, heater probes, and epinephrine injection.

Mechanisms by which epinephrine injections cause ulcer hemostasis have been examined experimentally. Epinephrine causes intense vasoconstriction, platelet aggregation, and vessel sclerosis. These combined effects permit permanent control of arterial hemorrhage in most patients. Absolute alcohol has also been used for injection therapy with good results.

Potential complications of endoscopic therapy include bowel perforation and incitement of active bleeding from a nonbleeding vessel. The rate of perforation is low and has been reported to approximate 0.7%. New bleeding is induced by therapy in fewer than 1% of patients.

Under selective circumstances, repeated attempts at endoscopic therapy may also be used. In a prospective randomized trial, investigators evaluated whether emergency surgery or repeated endoscopic therapy resulted in better outcomes for patients with severe ulcer hemorrhage. Endoscopic therapy consisted of a combination of epinephrine injection and heater probe application. Definitive hemostasis was significantly higher in surgically treated patients (93 vs 73%), but the complication rate was significantly higher in the surgery group (36%) relative to the endoscopy group (15%).

Acute therapy of variceal bleeding may also be directed endoscopically. Major approaches have included variceal injection with sclerosants and band ligation. Because of efficacy and safety, endoscopic variceal ligation has largely replaced sclerotherapy as the endoscopic method of choice for acute variceal hemorrhage.50 This method has also been used for secondary prevention of esophageal variceal hemorrhage. Prospective randomized trials indicate that prophylactic variceal ligation decreases the risk of first variceal bleeding relative to no treatment or to treatment with propranolol. In addition, ligation decreases the risk of recurrent bleeding and associated mortality relative to no treatment. In this circumstance, however, relative to propranolol therapy, ligation does not improve mortality.

In patients with acute peptic ulcer as a cause for upper GI hemorrhage, H. pylori is a common etiology. After initial control of hemorrhage, eradication of infection should be a treatment imperative. Because eradication of H. pylori eliminates ulcer recurrence, it is logical to assume that it would also decrease the rate of recurrent ulcer bleeding. Randomized trials demonstrate that recurrent hemorrhage usually occurs in patients who have persistent or recurrent H. pylori infection. Without antibiotic treatment, recurrent hemorrhage occurs in as many as 20% of patients. The risk of recurrent hemorrhage can be reduced to approximately 3% in individuals treated with an effective antibiotic regimen after hemorrhage.

Polyps

Gastric epithelial polyps are the most common benign tumor of the stomach. There are essentially five types of benign epithelial polyps: adenomatous, hyperplastic (regenerative), hamartomatous, inflammatory, and heterotopic (eg, ectopic pancreas). The most common gastric polyp (~75% in most series) is the hyperplastic or regenerative polyp, which frequently occurs in the setting of gastritis and has a low but real malignant potential. Adenomatous polyps may undergo malignant transformation, similarly to adenomas in the colon. They constitute about 10–15% of gastric polyps. Hamartomatous, inflammatory, and heterotopic polyps have negligible malignant potential. Polyps that are symptomatic, larger than 2 cm, or adenomatous should be removed, usually by endoscopic snare polypectomy. Consideration should also be given to removing hyperplastic polyps, especially if large. Repeat esophagogastroduodenoscopy (EGD) for surveillance should be done following removal of adenomatous polyps.

A fourfold rise in the incidence of nonfamilial fundic gland polyps has been noted due to the increased use of proton pump inhibitors. However, no increased risk of dysplasia has been noted.51

Lipoma

Lipomas are benign submucosal fatty tumors that are usually asymptomatic, found incidentally on upper GI series or EGD. Endoscopically they have a characteristic appearance; there also is a characteristic appearance on endoscopic ultrasound. Excision is unnecessary unless the patient is symptomatic.

Bezoars

Bezoars are collections of undigestible matter that accumulate in the stomach and small bowel. They are the most common foreign body found in the stomach and may be seen in patients who have undergone prior gastric surgery, including after bariatric surgery.52–54 The most common bezoar is composed of hair (trichobezoars). It occurs most commonly in young women. Phytobezoars are composed of vegetable matter and are usually seen in association with gastroparesis or gastric outlet obstruction. Other types of bezoars include lactobezoars (concentrated milk formula), mixed medication bezoars, and food bolus bezoars.55 Bezoars may present with obstruction, ulceration or bleeding, and rarely as intussusception.56 Diagnosis is suggested by upper GI series and confirmed by endoscopy. Enzyme therapy with papain, cellulase, or acetylcysteine may be used, but most patients will need endoscopic or surgical disruption and extraction.

Dieulafoy's Lesion

Dieulafoy's lesion is a congenital arteriovenous malformation of the proximal stomach, typically on the lesser curve where it derives its supply from branches of either the left or right gastric artery. It is seen in middle-aged or elderly men and characterized by an unusually large tortuous submucosal artery. Prior to widespread endoscopy, Dieulafoy's lesions were diagnosed postoperatively but are now becoming diagnosed and treated routinely via endoscopy.57 It clinically presents as an upper GI bleed if eroded and on endoscopy appears as a stream of arterial blood emanating from what appears grossly to be a normal gastric mucosa. Patients may also present with intermittent episodes of mild upper GI bleeding, and endoscopy can miss the lesion if it is not actively bleeding. Most lesions are now treated via endoscopic therapy (injection of epinephrine or other sclerosants, electrocoagulation, hemoclipping, rubber band ligation, and photocoagulation) or via angiographic embolization. Surgery is sometimes necessary, at which time the lesion may be oversewn or resected.

Dieulafoy's lesions may occasionally be seen in the duodenum and jejunum, as well as in the colon.58–60 These lesions have also been successfully managed via endoscopy or surgery.61

Diverticula

Gastric diverticula are typically solitary and may either be congenital or acquired. Congenital diverticula are rare, true diverticula that typically occur near the gastroesophageal junction and are found on the lesser curve or in the posterior area. They will demonstrate all three layers of the gastric wall on endoscopic ultrasound.62 Acquired or pseudodiverticula usually have a negligible outer muscle layer and are due to either pulsion or traction and most are found in the antrum. Symptoms are due to inflammation and may produce pain or bleeding but perforation is rare. Symptomatic lesions should be removed and can be done laparoscopically.

Foreign Bodies

Ingested foreign bodies are usually asymptomatic, but removal of sharp or large objects should be considered to avoid bleeding, perforation or obstruction. This can usually be done endoscopically. Aspiration of the foreign body during removal may occur, as well as potential rupture of drug-containing bags in “body packers.” Both complications can be fatal. In body packers as well as in patients with large jagged objects, surgical removal is recommended.

Mallory-Weiss Tear

The Mallory-Weiss lesion is a longitudinal tear in the mucosa of the GE junction, usually due to forceful vomiting and/or retching, and is commonly seen in alcoholics. It typically presents with impressive upper GI bleeding. Endoscopy confirms the diagnosis and may be useful in controlling the bleeding, but 90% of patients stop bleeding spontaneously. In patients who continue to bleed, balloon tamponade; angiographic embolization; or selective infusion of vasopressin, systemic vasopressin, and surgery are other treatment options. At surgery, the bleeding lesion is oversewn via a long gastrotomy.

Volvulus

Gastric volvulus occurs when the stomach twists around one of its axes, usually seen with a large hiatal hernia. It can also occur in the unusually mobile stomach without a hiatal hernia. Typically, the stomach twists along its long axis (organoaxial volvulus), and the greater curvature flips up. Less frequently, it occurs around the transverse axis, called mesoaxial volvulus. It is usually a chronic condition that can be surprisingly asymptomatic and expectant nonoperative management is usually advised, especially in the elderly. The risk of strangulation and infarction has been overestimated in asymptomatic patients.

Surgery is recommended for symptomatic patients, especially if these are severe and/or progressive. These patients complain of pain and pressure related to the intermittently distending and poorly emptying twisted stomach. Dyspnea, palpitations, and dysphagia may be seen due to compressive effects of the distended stomach on the surrounding organs. Symptoms are often relieved with vomiting or, if possible, passage of a nasogastric tube. The patient who presents moribund most likely has an infarcted stomach and is a case of surgical emergency requiring resection. Elective operation may often be done laparoscopically and usually involves reduction of the stomach and repair of hiatal hernia, with or without gastropexy. Gastropexy alone may be considered for high-risk patients.

Gastroparesis is a chronic gastric motility disorder defined by delayed gastric emptying of solids without evidence of mechanical obstruction.63 Primary gastroparesis affects mostly young and middle-aged women who present with nausea, abdominal pain, early satiety, vomiting, fullness, bloating, anorexia, and weight loss, with nausea and vomiting being the most disquieting of all the symptoms. The condition is diagnosed by symptom assessment and delayed gastric emptying of a solid meal. Gastric retention of more than 10% of the standard low-fat meal at 4 hours is indicative of delayed emptying.

Severe gastroparesis might result in recurrent hospitalizations, malnutrition, and significant mortality. Patients failing medical therapy are often considered for a variety of surgical interventions, the efficacy of which is not well studied. These procedures include gastrostomy, jejunostomy, gastric pacing/stimulation, and gastrectomy or surgical drainage procedures. Completion gastrectomy seems to provide symptom relief in postsurgical gastroparesis.64

Perhaps the most common laparoscopic gastric operations performed today are for gastroesophageal reflux disease and obesity. Most of the procedures described in this chapter can be performed with minimally invasive techniques. Some (eg, partial or total gastric resection) are technically difficult or are of debatable merit (eg, laparoscopic resection for cancer).

The operations, described previously, that lend themselves most readily to minimally invasive techniques are highly selective vagotomy, vagotomy and gastrojejunostomy, and gastrostomy. Laparoscopic wedge resection, combined with either intra-operative endoscopic or radiologic localization, often is possible for most localized, benign lesions such as lipomas, or gastric diverticula, although the incision required to retrieve the specimen may be larger than the initial port incisions.65,66 Combined endoscopic and laparoscopic techniques have also been described.67 Diagnostic laparoscopy may prevent a futile laparotomy in some patients with gastric cancer. The number as well as the location of ports is determined by triangulating around the target organ, and most procedures can be performed using four to five ports. The benefits of laparoscopic surgery (less post-op pain, quicker recovery, and decreased hospital stay) are all realized without compromising surgical principles of adequate resection and tension-free anastomosis.