Antacids Before we understood the important role of histamine in stimulating parietal cell activity, neutralization of secreted acid with antacids constituted the main form of therapy for peptic ulcers. They are now rarely, if ever, used as the primary therapeutic agent but instead are often used by patients for symptomatic relief of dyspepsia. The most commonly used agents are mixtures of aluminum hydroxide and magnesium hydroxide. Aluminum hydroxide can produce constipation and phosphate depletion; magnesium hydroxide may cause loose stools. Many of the commonly used antacids (e.g., Maalox, Mylanta) have a combination of both aluminum and magnesium hydroxide in order to avoid these side effects. The magnesium-containing preparation should not be used in chronic renal failure patients because of possible hypermagnesemia, and aluminum may cause chronic neurotoxicity in these patients.
Calcium carbonate and sodium bicarbonate are potent antacids with varying levels of potential problems. The long-term use of calcium carbonate (converts to calcium chloride in the stomach) can lead to milk-alkali syndrome (hypercalcemia, hyperphosphatemia with possible renal calcinosis and progression to renal insufficiency). Sodium bicarbonate may induce systemic alkalosis.
H2 Receptor Antagonists Four of these agents are presently available (cimetidine, ranitidine, famotidine, and nizatidine), and their structures share homology with histamine. Although each has different potency, all will significantly inhibit basal and stimulated acid secretion to comparable levels when used at therapeutic doses. Moreover, similar ulcer-healing rates are achieved with each drug when used at the correct dosage. Presently, this class of drug is often used for treatment of active ulcers (4–6 weeks) in combination with antibiotics directed at eradicating H. pylori (see below).
Cimetidine was the first H2 receptor antagonist used for the treatment of acid peptic disorders. The initial recommended dosing profile for cimetidine was 300 mg qid. Subsequent studies have documented the efficacy of using 800 mg at bedtime for treatment of active ulcer, with healing rates approaching 80% at 4 weeks. Cimetidine may have weak antiandrogenic side effects resulting in reversible gynecomastia and impotence, primarily in patients receiving high doses for prolonged periods of time (months to years, as in ZES). In view of cimetidine’s ability to inhibit cytochrome P450, careful monitoring of drugs such as warfarin, phenytoin, and theophylline is indicated with long-term usage. Other rare reversible adverse effects reported with cimetidine include confusion and elevated levels of serum aminotransferases, creatinine, and serum prolactin. Ranitidine, famotidine, and nizatidine are more potent H2 receptor antagonists than cimetidine. Each can be used once a day at bedtime for ulcer prevention, which was commonly done before the discovery of H. pylori and the development of proton pump inhibitors (PPIs). Patients may develop tolerance to H2 blockers, a rare event with PPIs (see below). Comparable nighttime dosing regimens are ranitidine 300 mg, famotidine 40 mg, and nizatidine 300 mg.
Additional rare, reversible systemic toxicities reported with H2 receptor antagonists include pancytopenia, neutropenia, anemia, and thrombocytopenia, with a prevalence rate varying from 0.01–0.2%. Cimetidine and ranitidine (to a lesser extent) can bind to hepatic cytochrome P450; famotidine and nizatidine do not.
Proton Pump (H+,K+-ATPase) Inhibitors Omeprazole, esomeprazole, lansoprazole, rabeprazole, and pantoprazole are substituted benzimidazole derivatives that covalently bind and irreversibly inhibit H+,K+-ATPase. Esomeprazole, one of the newest members of this drug class, is the S-enantiomer of omeprazole, which is a racemic mixture of both S- and R-optical isomers. The R-isomer of lansoprazole, dexlansoprazole, is the most recent PPI approved for clinical use. Its reported advantage is a dual delayed-release system, aimed at improving treatment of gastroesophageal reflux disease (GERD). These are the most potent acid inhibitory agents available. Omeprazole and lansoprazole are the PPIs that have been used for the longest time. Both are acid-labile and are administered as enteric-coated granules in a sustained-release capsule that dissolves within the small intestine at a pH of 6. Lansoprazole is available in an orally disintegrating tablet that can be taken with or without water, an advantage for individuals who have significant dysphagia. Absorption kinetics are similar to the capsule. In addition, a lansoprazole-naproxen combination preparation that has been made available is targeted at decreasing NSAID-related GI injury (see below). Omeprazole is available as nonenteric-coated granules mixed with sodium bicarbonate in a powder form that can be administered orally or via gastric tube. The sodium bicarbonate has two purposes: to protect the omeprazole from acid degradation and to promote rapid gastric alkalinization and subsequent proton pump activation, which facilitates rapid action of the PPI. Pantoprazole and rabeprazole are available as enteric-coated tablets. Pantoprazole is also available as a parenteral formulation for intravenous use. These agents are lipophilic compounds; upon entering the parietal cell, they are protonated and trapped within the acid environment of the tubulovesicular and canalicular system. These agents potently inhibit all phases of gastric acid secretion. Onset of action is rapid, with a maximum acid inhibitory effect between 2 and 6 h after administration and duration of inhibition lasting up to 72–96 h. With repeated daily dosing, progressive acid inhibitory effects are observed, with basal and secretagogue-stimulated acid production being inhibited by >95% after 1 week of therapy. The half-life of PPIs is ~18 h; thus, it can take between 2 and 5 days for gastric acid secretion to return to normal levels once these drugs have been discontinued. Because the pumps need to be activated for these agents to be effective, their efficacy is maximized if they are administered before a meal (except for the immediate-release formulation of omeprazole) (e.g., in the morning before breakfast). Mild to moderate hypergastrinemia has been observed in patients taking these drugs. Carcinoid tumors developed in some animals given the drugs preclinically; however, extensive experience has failed to demonstrate gastric carcinoid tumor development in humans. Serum gastrin levels return to normal levels within 1–2 weeks after drug cessation. Rebound gastric acid hypersecretion has been described in H. pylori–negative individuals after discontinuation of PPIs. It occurs even after relatively short-term usage (2 months) and may last for up to 2 months after the PPI has been discontinued. The mechanism involves gastrin-induced hyperplasia and hypertrophy of histamine-secreting ECL cells. The clinical relevance of this observation is that individuals may have worsening symptoms of GERD or dyspepsia upon stopping the PPI. Gradual tapering of the PPI and switching to an H2 receptor antagonist may prevent this from occurring. H. pylori–induced inflammation and concomitant decrease in acid production may explain why this does not occur in H. pylori–positive patients. IF production is also inhibited, but vitamin B12-deficiency anemia is uncommon, probably because of the large stores of the vitamin. As with any agent that leads to significant hypochlorhydria, PPIs may interfere with absorption of drugs such as ketoconazole, ampicillin, iron, and digoxin. Hepatic cytochrome P450 can be inhibited by the earlier PPIs (omeprazole, lansoprazole). Rabeprazole, pantoprazole, and esomeprazole do not appear to interact significantly with drugs metabolized by the cytochrome P450 system. The overall clinical significance of this observation is not definitely established. Caution should be taken when using theophylline, warfarin, diazepam, atazanavir, and phenytoin concomitantly with PPIs. Long-term acid suppression, especially with PPIs, has been associated with a higher incidence of community-acquired pneumonia as well as community and hospital acquired Clostridium difficile–associated disease. These observations require confirmation but should alert the practitioner to take caution when recommending these agents for long-term use, especially in elderly patients at risk for developing pneumonia or C. difficile infection. A population-based study revealed that long-term use of PPIs was associated with the development of hip fractures in older women. The absolute risk of fracture remained low despite an observed increase associated with the dose and duration of acid suppression. The mechanism for this observation is not clear, and this finding must be confirmed before making broad recommendations regarding the discontinuation of these agents in patients who benefit from them. Long-term use of PPIs has also been implicated in the development of iron and magnesium deficiency, but here again, the studies are limited and inconclusive. PPIs may exert a negative effect on the antiplatelet effect of clopidogrel. Although the evidence is mixed and inconclusive, a small increase in mortality and readmission rate for coronary events was seen in patients receiving a PPI while on clopidogrel in earlier studies. Subsequently, three meta-analyses reported an inverse correlation between clopidogrel and PPI use; therefore, the influence of this drug interaction on mortality is not clearly established. The mechanism involves the competition of the PPI and clopidogrel with the same cytochrome P450 (CYP2C19). Whether this is a class effect of PPIs is unclear; there appears to be at least a theoretical advantage of pantoprazole over the other PPIs, but this has not been confirmed. This drug interaction is particularly relevant in light of the common use of aspirin and clopidogrel for prevention of coronary events and the efficacy of PPIs in preventing GI bleeding in these patients. The FDA has made several recommendations while awaiting further evidence to clarify the impact of PPI therapy on clopidogrel use. Health care providers should continue to prescribe clopidogrel to patients who require it and should reevaluate the need for starting or continuing treatment with a PPI. From a practical standpoint, additional recommendations to consider include the following: Patients taking clopidogrel with aspirin, especially with other GI risk factors for bleeding, should receive GI protective therapy. Although high-dose H2 blockers have been considered an option, these do not appear to be as effective as PPIs. If PPIs are to be given, some have recommended that there be a 12-h separation between administration of the PPI and clopidogrel to minimize competition of the two agents with the involved cytochrome P450. One option is to give the PPI 30 min before breakfast and the clopidogrel at bedtime. Insufficient data are available to firmly recommend one PPI over another. Patients 65 years of age or older have a higher risk for some of the long-term side effects of PPIs highlighted above, in part due to the higher prevalence of concomitant chronic diseases. It is therefore important to carefully select individuals, especially among the elderly, who need long-term PPI therapy and discontinue it in those individuals who do not need it.
Two new formulations of acid inhibitory agents are being developed. Tenatoprazole is a PPI containing an imidazopyridine ring instead of a benzimidazole ring, which promotes irreversible proton pump inhibition. This agent has a longer half-life than the other PPIs and may be beneficial for inhibiting nocturnal acid secretion, which has significant relevance in GERD. A second new class of agents is the potassium-competitive acid pump antagonists (P-CABs). These compounds inhibit gastric acid secretion via potassium competitive binding of the H+,K+-ATPase.
Sucralfate Sucralfate is a complex sucrose salt in which the hydroxyl groups have been substituted by aluminum hydroxide and sulfate. This compound is insoluble in water and becomes a viscous paste within the stomach and duodenum, binding primarily to sites of active ulceration. Sucralfate may act by several mechanisms: serving as a physicochemical barrier, promoting a trophic action by binding growth factors such as EGF, enhancing prostaglandin synthesis, stimulating mucus and bicarbonate secretion, and enhancing mucosal defense and repair. Toxicity from this drug is rare, with constipation being most common (2–3%). It should be avoided in patients with chronic renal insufficiency to prevent aluminum-induced neurotoxicity. Hypophosphatemia and gastric bezoar formation have also been reported rarely. Standard dosing of sucralfate is 1 g qid.
Bismuth-Containing Preparations Sir William Osler considered bismuth-containing compounds the drug of choice for treating PUD. The resurgence in the use of these agents is due to their effect against H. pylori. Colloidal bismuth subcitrate (CBS) and bismuth subsalicylate (BSS, Pepto-Bismol) are the most widely used preparations. The mechanism by which these agents induce ulcer healing is unclear. Adverse effects with short-term use include black stools, constipation, and darkening of the tongue. Long-term use with high doses, especially with the avidly absorbed CBS, may lead to neurotoxicity. These compounds are commonly used as one of the agents in an anti-H. pylori regimen (see below).
Prostaglandin Analogues In view of their central role in maintaining mucosal integrity and repair, stable prostaglandin analogues were developed for the treatment of PUD. The mechanism by which this rapidly absorbed drug provides its therapeutic effect is through enhancement of mucosal defense and repair. The most common toxicity noted with this drug is diarrhea (10–30% incidence). Other major toxicities include uterine bleeding and contractions; misoprostol is contraindicated in women who may be pregnant, and women of childbearing age must be made clearly aware of this potential drug toxicity. The standard therapeutic dose is 200 μg qid.
Miscellaneous Drugs A number of drugs including anticholinergic agents and tricyclic antidepressants were used for treating acid peptic disorders, but in light of their toxicity and the development of potent antisecretory agents, these are rarely, if ever, used today.
THERAPY OF H. PYLORI
The physician’s goal in treating PUD is to provide relief of symptoms (pain or dyspepsia), promote ulcer healing, and ultimately prevent ulcer recurrence and complications. The greatest influence of understanding the role of H. pylori in peptic disease has been the ability to prevent recurrence. Documented eradication of H. pylori in patients with PUD is associated with a dramatic decrease in ulcer recurrence to <10–20% as compared to 59% in GU patients and 67% in DU patients when the organism is not eliminated. Eradication of the organism may lead to diminished recurrent ulcer bleeding. The effect of its eradication on ulcer perforation is unclear.
Extensive effort has been made in determining who of the many individuals with H. pylori infection should be treated. The common conclusion arrived at by multiple consensus conferences around the world is that H. pylori should be eradicated in patients with documented PUD. This holds true independent of time of presentation (first episode or not), severity of symptoms, presence of confounding factors such as ingestion of NSAIDs, or whether the ulcer is in remission. Some have advocated treating patients with a history of documented PUD who are found to be H. pylori–positive by serology or breath testing. Over one-half of patients with gastric MALT lymphoma experience complete remission of the tumor in response to H. pylori eradication. The Maastricht IV/Florence Consensus Report recommends a test-and-treat approach for patients with uninvestigated dyspepsia if the local incidence of H. pylori is greater than 20%. In addition, recommendations from this consensus report include testing and eradicating H. pylori in patients who will be using NSAIDs (including low-dose aspirin) on a long-term basis, especially if there is a prior history of PUD. These individuals will require continued PPI treatment as well as eradication treatment, because eradication of the organism alone does not eliminate the risk of gastroduodenal ulcers in patients already receiving long-term NSAIDs. Treating patients with NUD to prevent gastric cancer or patients with GERD requiring long-term acid suppression remains controversial. Guidelines from the American College of Gastroenterology suggest eradication of H. pylori in patients who have undergone resection of early gastric cancer. The Maastricht IV/Florence Consensus Report also evaluated H. pylori treatment in gastric cancer prevention and recommends that eradication should be considered in the following situations: first-degree relatives of family members with gastric cancer; patients with previous gastric neoplasm treated by endoscopic or subtotal resection; individuals with a risk of gastritis (severe pangastritis or body-predominant gastritis) or severe atrophy; patients with gastric acid inhibition for more than 1 year; individuals with strong environmental risk factors for gastric cancer (heavy smoking; high exposure to dust, coal, quartz, or cement; and/or work in quarries); and H. pylori–positive patients with a fear of gastric cancer.
Multiple drugs have been evaluated in the therapy of H. pylori. No single agent is effective in eradicating the organism. Combination therapy for 14 days provides the greatest efficacy, although regimens based on sequential administration of antibiotics also appear promising (see below). A shorter administration course (7–10 days), although attractive, has not proved as successful as the 14-day regimens. The agents used with the greatest frequency include amoxicillin, metronidazole, tetracycline, clarithromycin, and bismuth compounds.
Suggested treatment regimens for H. pylori are outlined in Table 348-4. Choice of a particular regimen will be influenced by several factors, including efficacy, patient tolerance, existing antibiotic resistance, and cost of the drugs. The aim for initial eradication rates should be 85–90%. Dual therapy (PPI plus amoxicillin, PPI plus clarithromycin, ranitidine bismuth citrate [Tritec] plus clarithromycin) is not recommended in view of studies demonstrating eradication rates of <80–85%. The combination of bismuth, metronidazole, and tetracycline was the first triple regimen found effective against H. pylori. The combination of two antibiotics plus either a PPI, H2 blocker, or bismuth compound has comparable success rates. Addition of acid suppression assists in providing early symptom relief and enhances bacterial eradication.
REGIMENS RECOMMENDED FOR ERADICATION OF H. PYLORI INFECTION
Triple therapy, although effective, has several drawbacks, including the potential for poor patient compliance and drug-induced side effects. Compliance is being addressed by simplifying the regimens so that patients can take the medications twice a day. Simpler (dual therapy) and shorter regimens (7 and 10 days) are not as effective as triple therapy for 14 days. Two anti-H. pylori regimens are available in prepackaged formulation: Prevpac (lansoprazole, clarithromycin, and amoxicillin) and Helidac (BSS, tetracycline, and metronidazole). The contents of the Prevpac are to be taken twice per day for 14 days, whereas Helidac constituents are taken four times per day with an antisecretory agent (PPI or H2 blocker), also for at least 14 days. Clarithromycin-based triple therapy should be avoided in settings where H. pylori resistance to this agent exceeds 15–20%.
Side effects have been reported in up to 20–30% of patients on triple therapy. Bismuth may cause black stools, constipation, or darkening of the tongue. The most feared complication with amoxicillin is pseudomembranous colitis, but this occurs in <1–2% of patients. Amoxicillin can also lead to antibiotic-associated diarrhea, nausea, vomiting, skin rash, and allergic reaction. Concomitant use of probiotics may ameliorate some of the antibiotic side effects (see below). Tetracycline has been reported to cause rashes and, very rarely, hepatotoxicity and anaphylaxis.
One important concern with treating patients who may not need therapy is the potential for development of antibiotic-resistant strains. The incidence and type of antibiotic-resistant H. pylori strains vary worldwide. Strains resistant to metronidazole, clarithromycin, amoxicillin, and tetracycline have been described, with the latter two being uncommon. Antibiotic-resistant strains are the most common cause for treatment failure in compliant patients. Unfortunately, in vitro resistance does not predict outcome in patients. Culture and sensitivity testing of H. pylori is not performed routinely. Although resistance to metronidazole has been found in as many as 30% of isolates in North America and 80% in developing countries, triple therapy is effective in eradicating the organism in >50% of patients infected with a resistant strain. Clarithromycin resistance is seen in 13% of individuals in the United States, with resistance to amoxicillin being <1% and resistance to both metronidazole and clarithromycin in the 5% range.
Failure of H. pylori eradication with triple therapy in a compliant patient is usually due to infection with a resistant organism. Quadruple therapy (Table 348-4), where clarithromycin is substituted for metronidazole (or vice versa), should be the next step. The combination of pantoprazole, amoxicillin, and rifabutin for 10 days has also been used successfully (86% cure rate) in patients infected with resistant strains. Additional regimens considered for second-line therapy include levofloxacin-based triple therapy (levofloxacin, amoxicillin, PPI) for 10 days and furazolidone-based triple therapy (furazolidone, amoxicillin, PPI) for 14 days. Unfortunately, there is no universally accepted treatment regimen recommended for patients who have failed two courses of antibiotics. If eradication is still not achieved in a compliant patient, then culture and sensitivity of the organism should be considered. Additional factors that may lower eradication rates include the patient’s country of origin (higher in Northeast Asia than other parts of Asia or Europe) and cigarette smoking. In addition, meta-analysis suggests that even the most effective regimens (quadruple therapy including PPI, bismuth, tetracycline, and metronidazole and triple therapy including PPI, clarithromycin, and amoxicillin) may have suboptimal eradication rates (<80%), thus demonstrating the need for the development of more efficacious treatments.
In view of the observation that 15–25% of patients treated with first-line therapy may still remain infected with the organism, new approaches to treatment have been explored. One promising approach is sequential therapy. Regimens examined consist of 5 days of amoxicillin and a PPI, followed by an additional 5 days of PPI plus tinidazole and clarithromycin or levofloxacin. One promising regimen that has the benefit of being shorter in duration, easier to take, and less expensive is 5 days of concomitant therapy (PPI twice daily, amoxicillin 1 g twice daily, levofloxacin 500 mg twice daily, and tinidazole 500 mg twice daily). Initial studies have demonstrated eradication rates of >90% with good patient tolerance. Confirmation of these findings and applicability of this approach in the United States are needed, although some experts are recommending abandoning clarithromycin-based triple therapy in the United States for the concomitant therapy or the alternative sequential therapies highlighted above.
Innovative non–antibiotic-mediated approaches have been explored in an effort to improve eradication rates of H. pylori. Pretreatment of patients with N-acetylcysteine as a mucolytic agent to destroy the H. pylori biofilm and therefore impair antibiotic resistance has been examined, but more studies are needed to confirm the applicability of this approach. In vitro studies suggest that certain probiotics like Lactobacillus or its metabolites can inhibit H. pylori. Administration of probiotics has been attempted in several clinical studies in an effort to maximize antibiotic-mediated eradication with varying results. Overall, it appears that the use of certain probiotics, such as Lactobacillus spp., Saccharomyces spp., Bifidobacterium spp., and Bacillus clausii, did not alter eradication rates but importantly decreased antibiotic-associated side effects including nausea, dysgeusia, diarrhea, and abdominal discomfort/pain, resulting in enhanced tolerability of H. pylori therapies. Additional studies are needed to confirm the potential benefits of probiotics in this setting.
Reinfection after successful eradication of H. pylori is rare in the United States (<1% per year). If recurrent infection occurs within the first 6 months after completing therapy, the most likely explanation is recrudescence as opposed to reinfection.
THERAPY OF NSAID-RELATED GASTRIC OR DUODENAL INJURY
Medical intervention for NSAID-related mucosal injury includes treatment of an active ulcer and primary prevention of future injury. Recommendations for the treatment and primary prevention of NSAID-related mucosal injury are listed in Table 348-5. Ideally, the injurious agent should be stopped as the first step in the therapy of an active NSAID-induced ulcer. If that is possible, then treatment with one of the acid inhibitory agents (H2 blockers, PPIs) is indicated. Cessation of NSAIDs is not always possible because of the patient’s severe underlying disease. Only PPIs can heal GUs or DUs, independent of whether NSAIDs are discontinued.
RECOMMENDATIONS FOR TREATMENT OF NSAID-RELATED MUCOSAL INJURY
The approach to primary prevention has included avoiding the agent, using the lowest possible dose of the agent, using NSAIDs that are theoretically less injurious, using newer topical NSAID preparations, and/or using concomitant medical therapy to prevent NSAID-induced injury. Several nonselective NSAIDs that are associated with a lower likelihood of GI toxicity include diclofenac, aceclofenac, and ibuprofen, although the beneficial effect may be eliminated if higher dosages of the agents are used. Primary prevention of NSAID-induced ulceration can be accomplished by misoprostol (200 μg qid) or a PPI. High-dose H2 blockers (famotidine, 40 mg bid) have also shown some promise in preventing endoscopically documented ulcers, although PPIs are superior. The highly selective COX-2 inhibitors, celecoxib and rofecoxib, are 100 times more selective inhibitors of COX-2 than standard NSAIDs, leading to gastric or duodenal mucosal injury that is comparable to placebo; their utilization led to an increase in cardiovascular events and withdrawal from the market. Additional caution was engendered when the CLASS study demonstrated that the advantage of celecoxib in preventing GI complications was offset when low-dose aspirin was used simultaneously. Therefore, gastric protection therapy is required in individuals taking COX-2 inhibitors and aspirin prophylaxis. Finally, much of the work demonstrating the benefit of COX-2 inhibitors and PPIs on GI injury has been performed in individuals of average risk; it is unclear if the same level of benefit will be achieved in high-risk patients. For example, concomitant use of warfarin and a COX-2 inhibitor was associated with rates of GI bleeding similar to those observed in patients taking nonselective NSAIDs. A combination of factors, including withdrawal of the majority of COX-2 inhibitors from the market, the observation that low-dose aspirin appears to diminish the beneficial effect of COX-2 selective inhibitors, and the growing use of aspirin for prophylaxis of cardiovascular events, have significantly altered the approach to gastric protective therapy during the use of NSAIDs. A set of guidelines for the approach to the use of NSAIDs was published by the American College of Gastroenterology and is shown in Table 348-6. Individuals who are not at risk for cardiovascular events, do not use aspirin, and are without risk for GI complications can receive nonselective NSAIDs without gastric protection. In those without cardiovascular risk factors but with a high potential risk (prior GI bleeding or multiple GI risk factors) for NSAID-induced GI toxicity, cautious use of a selective COX-2 inhibitor and co-therapy with misoprostol or high-dose PPI are recommended. Individuals at moderate GI risk without cardiac risk factors can be treated with a COX-2 inhibitor alone or with a nonselective NSAID with misoprostol or a PPI. Individuals with cardiovascular risk factors, who require low-dose aspirin and have low potential for NSAID-induced toxicity, should be considered for a non-NSAID agent or use of a traditional NSAID in combination with gastric protection, if warranted. Finally, individuals with cardiovascular and GI risks who require aspirin must be considered for non-NSAID therapy, but if that is not an option, then gastric protection with any type of NSAID must be considered. Any patient, regardless of risk status, who is being considered for long-term traditional NSAID therapy, should also be considered for H. pylori testing and treatment if positive. Assuring the use of GI protective agents with NSAIDs is difficult, even in high-risk patients. This is in part due to underprescribing of the appropriate protective agent; other times the difficulty is related to patient compliance. The latter may be due to patients forgetting to take multiple pills or preferring not to take the extra pill, especially if they have no GI symptoms. Several NSAID gastroprotective-containing combination pills are now commercially available, including double-dose famotidine with ibuprofen, diclofenac with misoprostol, and naproxen with esomeprazole. Although initial studies suggested improved compliance and a cost advantage when taking these combination drugs, their clinical benefit over the use of separate pills has not been established. Efforts continue toward developing safer NSAIDs, including NO–releasing NSAIDs, hydrogen sulfide–releasing NSAIDs, dual COX/5-LOX inhibitors, NSAID prodrugs, or agents that can effectively sequester unbound NSAIDs without interfering with their efficacy.
GUIDE TO NSAID THERAPY
APPROACH AND THERAPY: SUMMARY
Controversy continues regarding the best approach to the patient who presents with dyspepsia (Chap. 54). The discovery of H. pylori and its role in pathogenesis of ulcers has added a new variable to the equation. Previously, if a patient <50 years of age presented with dyspepsia and without alarming signs or symptoms suggestive of an ulcer complication or malignancy, an empirical therapeutic trial with acid suppression was commonly recommended. Although this approach is practiced by some today, an approach presently gaining approval for the treatment of patients with dyspepsia is outlined in Fig. 348-12. The referral to a gastroenterologist is for the potential need of endoscopy and subsequent evaluation and treatment if the endoscopy is negative.
FIGURE 348-12 Overview of new-onset dyspepsia. GERD, gastroesophageal reflux disease; Hp, Helicobacter pylori; IBS, irritable bowel syndrome; UBT, urea breath test. (Adapted from BS Anand and DY Graham: Endoscopy 31:215, 1999.)
Once an ulcer (GU or DU) is documented, the main issue at stake is whether H. pylori or an NSAID is involved. With H. pylori present, independent of the NSAID status, triple therapy is recommended for 14 days, followed by continued acid-suppressing drugs (H2 receptor antagonist or PPIs) for a total of 4–6 weeks. Selection of patients for documentation of H. pylori eradication (organisms gone at least 4 weeks after completing antibiotics) is an area of some debate. The test of choice for documenting eradication is the laboratory-based validated monoclonal stool antigen test or a urea breath test (UBT). The patient must be off antisecretory agents when being tested for eradication of H. pylori with UBT or stool antigen. Serologic testing is not useful for the purpose of documenting eradication because antibody titers fall slowly and often do not become undetectable. Two approaches toward documentation of eradication exist: (1) Test for eradication only in individuals with a complicated course or in individuals who are frail or with multisystem disease who would do poorly with an ulcer recurrence, and (2) test all patients for successful eradication. Some recommend that patients with complicated ulcer disease, or who are frail, should be treated with long-term acid suppression, thus making documentation of H. pylori eradication a moot point. In view of this discrepancy in practice, it would be best to discuss with the patient the different options available.
Several issues differentiate the approach to a GU versus a DU. GUs, especially of the body and fundus, have the potential of being malignant. Multiple biopsies of a GU should be taken initially; even if these are negative for neoplasm, repeat endoscopy to document healing at 8–12 weeks should be performed, with biopsy if the ulcer is still present. About 70% of GUs eventually found to be malignant undergo significant (usually incomplete) healing. Repeat endoscopy is warranted in patients with DU if symptoms persist despite medical therapy or a complication is suspected.
The majority (>90%) of GUs and DUs heal with the conventional therapy outlined above. A GU that fails to heal after 12 weeks and a DU that does not heal after 8 weeks of therapy should be considered refractory. Once poor compliance and persistent H. pylori infection have been excluded, NSAID use, either inadvertent or surreptitious, must be excluded. In addition, cigarette smoking must be eliminated. For a GU, malignancy must be meticulously excluded. Next, consideration should be given to a gastric acid hypersecretory state such as ZES (see “Zollinger-Ellison Syndrome,” below) or the idiopathic form, which can be excluded with gastric acid analysis. Although a subset of patients have gastric acid hypersecretion of unclear etiology as a contributing factor to refractory ulcers, ZES should be excluded with a fasting gastrin or secretin stimulation test (see below). More than 90% of refractory ulcers (either DUs or GUs) heal after 8 weeks of treatment with higher doses of PPI (omeprazole, 40 mg/d; lansoprazole 30–60 mg/d). This higher dose is also effective in maintaining remission. Surgical intervention may be a consideration at this point; however, other rare causes of refractory ulcers must be excluded before recommending surgery. Rare etiologies of refractory ulcers that may be diagnosed by gastric or duodenal biopsies include ischemia, Crohn’s disease, amyloidosis, sarcoidosis, lymphoma, eosinophilic gastroenteritis, or infection (cytomegalovirus [CMV], tuberculosis, or syphilis).
Surgical intervention in PUD can be viewed as being either elective, for treatment of medically refractory disease, or as urgent/emergent, for the treatment of an ulcer-related complication. The development of pharmacologic and endoscopic approaches for the treatment of peptic disease and its complications has led to a substantial decrease in the number of operations needed for this disorder with a drop of over 90% for elective ulcer surgery over the last four decades. Refractory ulcers are an exceedingly rare occurrence. Surgery is more often required for treatment of an ulcer-related complication.
Hemorrhage is the most common ulcer-related complication, occurring in ~15–25% of patients. Bleeding may occur in any age group but is most often seen in older patients (sixth decade or beyond). The majority of patients stop bleeding spontaneously, but endoscopic therapy (Chap. 345) is necessary in some. Parenterally and orally administered PPIs also decrease ulcer rebleeding in patients who have undergone endoscopic therapy. Patients unresponsive or refractory to endoscopic intervention will require surgery (~5% of transfusion-requiring patients).
Free peritoneal perforation occurs in ~2–3% of DU patients. As in the case of bleeding, up to 10% of these patients will not have antecedent ulcer symptoms. Concomitant bleeding may occur in up to 10% of patients with perforation, with mortality being increased substantially. Peptic ulcer can also penetrate into adjacent organs, especially with a posterior DU, which can penetrate into the pancreas, colon, liver, or biliary tree.
Pyloric channel ulcers or DUs can lead to gastric outlet obstruction in ~2–3% of patients. This can result from chronic scarring or from impaired motility due to inflammation and/or edema with pylorospasm. Patients may present with early satiety, nausea, vomiting of undigested food, and weight loss. Conservative management with nasogastric suction, intravenous hydration/nutrition, and antisecretory agents is indicated for 7–10 days with the hope that a functional obstruction will reverse. If a mechanical obstruction persists, endoscopic intervention with balloon dilation may be effective. Surgery should be considered if all else fails.
SPECIFIC OPERATIONS FOR DUODENAL ULCERS
Surgical treatment was originally designed to decrease gastric acid secretion. Operations most commonly performed include (1) vagotomy and drainage (by pyloroplasty, gastroduodenostomy, or gastrojejunostomy), (2) highly selective vagotomy (which does not require a drainage procedure), and (3) vagotomy with antrectomy. The specific procedure performed is dictated by the underlying circumstances: elective versus emergency, the degree and extent of duodenal ulceration, the etiology of the ulcer (H. pylori, NSAIDs, malignancy), and the expertise of the surgeon. Moreover, the trend has been toward a dramatic decrease in the need for surgery for treatment of refractory PUD, and when needed, minimally invasive and anatomy-preserving operations are preferred.
Vagotomy is a component of each of these procedures and is aimed at decreasing acid secretion through ablating cholinergic input to the stomach. Unfortunately, both truncal and selective vagotomy (preserves the celiac and hepatic branches) result in gastric atony despite successful reduction of both basal acid output (BAO; decreased by 85%) and maximal acid output (MAO; decreased by 50%). Drainage through pyloroplasty or gastroduodenostomy is required in an effort to compensate for the vagotomy-induced gastric motility disorder. This procedure has an intermediate complication rate and a 10% ulcer recurrence rate. To minimize gastric dysmotility, highly selective vagotomy (also known as parietal cell, super-selective, or proximal vagotomy) was developed. Only the vagal fibers innervating the portion of the stomach that contains parietal cells is transected, thus leaving fibers important for regulating gastric motility intact. Although this procedure leads to an immediate decrease in both BAO and stimulated acid output, acid secretion recovers over time. By the end of the first postoperative year, basal and stimulated acid output are ~30 and 50%, respectively, of preoperative levels. Ulcer recurrence rates are higher with highly selective vagotomy (≥10%), although the overall complication rates are the lowest of the three procedures.
The procedure that provides the lowest rates of ulcer recurrence (1%) but has the highest complication rate is vagotomy (truncal or selective) in combination with antrectomy. Antrectomy is aimed at eliminating an additional stimulant of gastric acid secretion, gastrin. Two principal types of reanastomoses are used after antrectomy: gastroduodenostomy (Billroth I) or gastrojejunostomy (Billroth II) (Fig. 348-13). Although Billroth I is often preferred over II, severe duodenal inflammation or scarring may preclude its performance. Prospective, randomized studies confirm that partial gastrectomy followed by Roux-en-Y reconstruction leads to a significantly better clinical, endoscopic, and histologic outcome than Billroth II reconstruction.
FIGURE 348-13 Schematic representation of Billroth I and II procedures.
Of these procedures, highly selective vagotomy may be the one of choice in the elective setting, except in situations where ulcer recurrence rates are high (prepyloric ulcers and those refractory to medical therapy). Selection of vagotomy and antrectomy may be more appropriate in these circumstances.
These procedures have been traditionally performed by standard laparotomy. The advent of laparoscopic surgery has led several surgical teams to successfully perform highly selective vagotomy, truncal vagotomy/pyloroplasty, and truncal vagotomy/antrectomy through this approach. An increase in the number of laparoscopic procedures for treatment of PUD has occurred. Laparoscopic repair of perforated peptic ulcers is safe, feasible for the experienced surgeon and is associated with decreased postoperative pain, although it does take longer than an open approach. Moreover, no difference between the two approaches is noted in postoperative complications or length of hospital stay.
Specific Operations for Gastric Ulcers The location and the presence of a concomitant DU dictate the operative procedure performed for a GU. Antrectomy (including the ulcer) with a Billroth I anastomosis is the treatment of choice for an antral ulcer. Vagotomy is performed only if a DU is present. Although ulcer excision with vagotomy and drainage procedure has been proposed, the higher incidence of ulcer recurrence makes this a less desirable approach. Ulcers located near the esophagogastric junction may require a more radical approach, a subtotal gastrectomy with a Roux-en-Y esophagogastrojejunostomy (Csendes’ procedure). A less aggressive approach, including antrectomy, intraoperative ulcer biopsy, and vagotomy (Kelling-Madlener procedure), may be indicated in fragile patients with a high GU. Ulcer recurrence approaches 30% with this procedure.
Surgery-Related Complications Complications seen after surgery for PUD are related primarily to the extent of the anatomic modification performed. Minimal alteration (highly selective vagotomy) is associated with higher rates of ulcer recurrence and less GI disturbance. More aggressive surgical procedures have a lower rate of ulcer recurrence but a greater incidence of GI dysfunction. Overall, morbidity and mortality related to these procedures are quite low. Morbidity associated with vagotomy and antrectomy or pyloroplasty is ≤5%, with mortality ~1%. Highly selective vagotomy has lower morbidity and mortality rates of 1 and 0.3%, respectively.
In addition to the potential early consequences of any intraabdominal procedure (bleeding, infection, thromboembolism), gastroparesis, duodenal stump leak, and efferent loop obstruction can be observed.
RECURRENT ULCERATION The risk of ulcer recurrence is directly related to the procedure performed. Ulcers that recur after partial gastric resection tend to develop at the anastomosis (stomal or marginal ulcer). Epigastric abdominal pain is the most frequent presenting complaint (>90%). Severity and duration of pain tend to be more progressive than observed with DUs before surgery.
Ulcers may recur for several reasons, including incomplete vagotomy, inadequate drainage, retained antrum, and, less likely, persistent or recurrent H. pylori infection. ZES should have been excluded preoperatively. Surreptitious use of NSAIDs is an important reason for recurrent ulcers after surgery, especially if the initial procedure was done for an NSAID-induced ulcer. Once H. pylori and NSAIDs have been excluded as etiologic factors, the question of incomplete vagotomy or retained gastric antrum should be explored. For the latter, fasting plasma gastrin levels should be determined. If elevated, retained antrum or ZES (see below) should be considered. Incomplete vagotomy can be ruled out by gastric acid analysis coupled with sham feeding. In this test, gastric acid output is measured while the patient sees, smells, and chews a meal (without swallowing). The cephalic phase of gastric secretion, which is mediated by the vagus, is being assessed with this study. An increase in gastric acid output in response to sham feeding is evidence that the vagus nerve is intact. A rise in serum pancreatic polypeptide >50% within 30 min of sham feeding is also suggestive of an intact vagus nerve.
Medical therapy with H2 blockers will heal postoperative ulceration in 70–90% of patients. The efficacy of PPIs has not been fully assessed in this group, but one may anticipate greater rates of ulcer healing compared to those obtained with H2 blockers. Repeat operation (complete vagotomy, partial gastrectomy) may be required in a small subgroup of patients who have not responded to aggressive medical management.
AFFERENT LOOP SYNDROMES Although rarely seen today as a result of the decrease in the performance of Billroth II anastomosis, two types of afferent loop syndrome can occur in patients who have undergone this type of partial gastric resection. The more common of the two is bacterial overgrowth in the afferent limb secondary to stasis. Patients may experience postprandial abdominal pain, bloating, and diarrhea with concomitant malabsorption of fats and vitamin B12. Cases refractory to antibiotics may require surgical revision of the loop. The less common afferent loop syndrome can present with severe abdominal pain and bloating that occur 20–60 min after meals. Pain is often followed by nausea and vomiting of bile-containing material. The pain and bloating may improve after emesis. The cause of this clinical picture is theorized to be incomplete drainage of bile and pancreatic secretions from an afferent loop that is partially obstructed. Cases refractory to dietary measures may need surgical revision or conversion of the Billroth II anastomosis to a Roux-en-Y gastrojejunostomy.
DUMPING SYNDROME Dumping syndrome consists of a series of vasomotor and GI signs and symptoms and occurs in patients who have undergone vagotomy and drainage (especially Billroth procedures). Two phases of dumping, early and late, can occur. Early dumping takes place 15–30 min after meals and consists of crampy abdominal discomfort, nausea, diarrhea, belching, tachycardia, palpitations, diaphoresis, light-headedness, and, rarely, syncope. These signs and symptoms arise from the rapid emptying of hyperosmolar gastric contents into the small intestine, resulting in a fluid shift into the gut lumen with plasma volume contraction and acute intestinal distention. Release of vasoactive GI hormones (vasoactive intestinal polypeptide, neurotensin, motilin) is also theorized to play a role in early dumping.
The late phase of dumping typically occurs 90 min to 3 h after meals. Vasomotor symptoms (light-headedness, diaphoresis, palpitations, tachycardia, and syncope) predominate during this phase. This component of dumping is thought to be secondary to hypoglycemia from excessive insulin release.
Dumping syndrome is most noticeable after meals rich in simple carbohydrates (especially sucrose) and high osmolarity. Ingestion of large amounts of fluids may also contribute. Up to 50% of postvagotomy and drainage patients will experience dumping syndrome to some degree early on. Signs and symptoms often improve with time, but a severe protracted picture can occur in up to 1% of patients.
Dietary modification is the cornerstone of therapy for patients with dumping syndrome. Small, multiple (six) meals devoid of simple carbohydrates coupled with elimination of liquids during meals is important. Antidiarrheals and anticholinergic agents are complementary to diet. Guar and pectin, which increase the viscosity of intraluminal contents, may be beneficial in more symptomatic individuals. Acarbose, an α-glucosidase inhibitor that delays digestion of ingested carbohydrates, has also been shown to be beneficial in the treatment of the late phases of dumping. The somatostatin analogue octreotide has been successful in diet-refractory cases. This drug is administered subcutaneously (50 μg tid), titrated according to clinical response. A long-acting depot formulation of octreotide can be administered once every 28 days and provides symptom relief comparable to the short-acting agent. In addition, patient weight gain and quality of life appear to be superior with the long-acting form.
POSTVAGOTOMY DIARRHEA Up to 10% of patients may seek medical attention for the treatment of postvagotomy diarrhea. This complication is most commonly observed after truncal vagotomy, which is rarely performed today. Patients may complain of intermittent diarrhea that occurs typically 1–2 h after meals. Occasionally the symptoms may be severe and relentless. This is due to a motility disorder from interruption of the vagal fibers supplying the luminal gut. Other contributing factors may include decreased absorption of nutrients (see below), increased excretion of bile acids, and release of luminal factors that promote secretion. Diphenoxylate or loperamide is often useful in symptom control. The bile salt–binding agent cholestyramine may be helpful in severe cases. Surgical reversal of a 10-cm segment of jejunum may yield a substantial improvement in bowel frequency in a subset of patients.
BILE REFLUX GASTROPATHY A subset of post–partial gastrectomy patients who present with abdominal pain, early satiety, nausea, and vomiting will have mucosal erythema of the gastric remnant as the only finding. Histologic examination of the gastric mucosa reveals minimal inflammation but the presence of epithelial cell injury. This clinical picture is categorized as bile or alkaline reflux gastropathy/gastritis. Although reflux of bile is implicated as the reason for this disorder, the mechanism is unknown. Prokinetic agents, cholestyramine, and sucralfate have been somewhat effective treatments. Severe refractory symptoms may require using either nuclear scanning with 99mTc-HIDA to document reflux or an alkaline challenge test, where 0.1 N NaOH is infused into the stomach in an effort to reproduce the patient’s symptoms. Surgical diversion of pancreaticobiliary secretions away from the gastric remnant with a Roux-en-Y gastrojejunostomy consisting of a long (50–60 cm) Roux limb has been used in severe cases. Bilious vomiting improves, but early satiety and bloating may persist in up to 50% of patients.
MALDIGESTION AND MALABSORPTION Weight loss can be observed in up to 60% of patients after partial gastric resection. Patients can experience a 10% loss of body weight, which stabilizes 3 months postoperatively. A significant component of this weight reduction is due to decreased oral intake. However, mild steatorrhea can also develop. Reasons for maldigestion/malabsorption include decreased gastric acid production, rapid gastric emptying, decreased food dispersion in the stomach, reduced luminal bile concentration, reduced pancreatic secretory response to feeding, and rapid intestinal transit.
Decreased serum vitamin B12 levels can be observed after partial gastrectomy. This is usually not due to deficiency of IF, since a minimal amount of parietal cells (source of IF) are removed during antrectomy. Reduced vitamin B12 may be due to competition for the vitamin by bacterial overgrowth or inability to split the vitamin from its protein-bound source due to hypochlorhydria.
Iron-deficiency anemia may be a consequence of impaired absorption of dietary iron in patients with a Billroth II gastrojejunostomy. Absorption of iron salts is normal in these individuals; thus, a favorable response to oral iron supplementation can be anticipated. Folate deficiency with concomitant anemia can also develop in these patients. This deficiency may be secondary to decreased absorption or diminished oral intake.
Malabsorption of vitamin D and calcium resulting in osteoporosis and osteomalacia is common after partial gastrectomy and gastrojejunostomy (Billroth II). Osteomalacia can occur as a late complication in up to 25% of post–partial gastrectomy patients. Bone fractures occur twice as commonly in men after gastric surgery as in a control population. It may take years before x-ray findings demonstrate diminished bone density. Elevated alkaline phosphatase, reduced serum calcium, bone pain, and pathologic fractures may be seen in patients with osteomalacia. The high incidence of these abnormalities in this subgroup of patients justifies treating them with vitamin D and calcium supplementation indefinitely. Therapy is especially important in females. Copper deficiency has also been reported in patients undergoing surgeries that bypass the duodenum, where copper is primarily absorbed. Patients may present with a rare syndrome that includes ataxia, myelopathy, and peripheral neuropathy.
GASTRIC ADENOCARCINOMA The incidence of adenocarcinoma in the gastric stump is increased 15 years after resection. Some have reported a four- to fivefold increase in gastric cancer 20–25 years after resection. The pathogenesis is unclear but may involve alkaline reflux, bacterial proliferation, or hypochlorhydria. The role of endoscopic screening is not clear, and most guidelines do not support its use.
ADDITIONAL COMPLICATIONS Reflux esophagitis and a higher incidence of gallstones and cholecystitis have been reported to patients undergoing subtotal gastrectomy. The latter is thought to be due to decreased gallbladder contractility associated with vagotomy and bypass of the duodenum, leading to decreased postprandial release of cholecystokinin.
Severe peptic ulcer diathesis secondary to gastric acid hypersecretion due to unregulated gastrin release from a non-β cell endocrine tumor (gastrinoma) defines the components of ZES. Initially, ZES was typified by aggressive and refractory ulceration in which total gastrectomy provided the only chance for enhancing survival. Today it can be cured by surgical resection in up to 40% of patients.
Epidemiology The incidence of ZES varies from 0.1–1% of individuals presenting with PUD. Males are more commonly affected than females, and the majority of patients are diagnosed between ages 30 and 50. Gastrinomas are classified into sporadic tumors (more common) and those associated with multiple endocrine neoplasia (MEN) type 1 (see below). The widespread availability and use of PPIs has led to a decreased patient referral for gastrinoma evaluation, delay in diagnosis, and an increase in false-positive diagnoses of ZES. In fact, diagnosis may be delayed for 6 or more years after symptoms consistent with ZES are displayed.
Pathophysiology Hypergastrinemia originating from an autonomous neoplasm is the driving force responsible for the clinical manifestations in ZES. Gastrin stimulates acid secretion through gastrin receptors on parietal cells and by inducing histamine release from ECL cells. Gastrin also has a trophic action on gastric epithelial cells. Long-standing hypergastrinemia leads to markedly increased gastric acid secretion through both parietal cell stimulation and increased parietal cell mass. The increased gastric acid output leads to peptic ulcer diathesis, erosive esophagitis, and diarrhea.
Tumor Distribution Although early studies suggested that the vast majority of gastrinomas occurred within the pancreas, a significant number of these lesions are extrapancreatic. Over 80% of these tumors are found within the hypothetical gastrinoma triangle (confluence of the cystic and common bile ducts superiorly, junction of the second and third portions of the duodenum inferiorly, and junction of the neck and body of the pancreas medially). Duodenal tumors constitute the most common nonpancreatic lesion; between 50 and 75% of gastrinomas are found here. Duodenal tumors are smaller, slower growing, and less likely to metastasize than pancreatic lesions. Less common extrapancreatic sites include stomach, bones, ovaries, heart, liver, and lymph nodes. More than 60% of tumors are considered malignant, with up to 30–50% of patients having multiple lesions or metastatic disease at presentation. Histologically, gastrin-producing cells appear well-differentiated, expressing markers typically found in endocrine neoplasms (chromogranin, neuron-specific enolase).
Clinical Manifestations Gastric acid hypersecretion is responsible for the signs and symptoms observed in patients with ZES. Peptic ulcer is the most common clinical manifestation, occurring in >90% of gastrinoma patients. Initial presentation and ulcer location (duodenal bulb) may be indistinguishable from common PUD. Clinical situations that should create suspicion of gastrinoma are ulcers in unusual locations (second part of the duodenum and beyond), ulcers refractory to standard medical therapy, ulcer recurrence after acid-reducing surgery, ulcers presenting with frank complications (bleeding, obstruction, and perforation), or ulcers in the absence of H. pylori or NSAID ingestion. Symptoms of esophageal origin are present in up to two-thirds of patients with ZES, with a spectrum ranging from mild esophagitis to frank ulceration with stricture and Barrett’s mucosa.
Diarrhea, the next most common clinical manifestation, is found in up to 50% of patients. Although diarrhea often occurs concomitantly with acid peptic disease, it may also occur independent of an ulcer. Etiology of the diarrhea is multifactorial, resulting from marked volume overload to the small bowel, pancreatic enzyme inactivation by acid, and damage of the intestinal epithelial surface by acid. The epithelial damage can lead to a mild degree of maldigestion and malabsorption of nutrients. The diarrhea may also have a secretory component due to the direct stimulatory effect of gastrin on enterocytes or the co-secretion of additional hormones from the tumor such as vasoactive intestinal peptide.
Gastrinomas can develop in the presence of MEN 1 syndrome (Chaps. 113 and 408) in ~25% of patients. This autosomal dominant disorder involves primarily three organ sites: the parathyroid glands (80–90%), pancreas (40–80%), and pituitary gland (30–60%). The syndrome is caused by inactivating mutations of the MEN1 tumor suppressor gene found on the long arm of chromosome 11q13. The gene encodes for Menin, which has an important role in DNA replication and transcriptional regulation. A genetic diagnosis is obtained by sequencing of the MEN1 gene, which can reveal mutations in 70–90% of typical MEN 1 cases. A family may have an unknown mutation, making a genetic diagnosis impossible, and therefore certain individuals will require a clinical diagnosis, which is determined by whether a patient has tumors in two of the three endocrine organs (parathyroid, pancreas/duodenum, or pituitary) or has a family history of MEN 1 and one of the endocrine organ tumors. In view of the stimulatory effect of calcium on gastric secretion, the hyperparathyroidism and hypercalcemia seen in MEN 1 patients may have a direct effect on ulcer disease. Resolution of hypercalcemia by parathyroidectomy reduces gastrin and gastric acid output in gastrinoma patients. An additional distinguishing feature in ZES patients with MEN 1 is the higher incidence of gastric carcinoid tumor development (as compared to patients with sporadic gastrinomas). ZES presents and is diagnosed earlier in MEN 1 patients, and they have a more indolent course as compared to patients with sporadic gastrinoma. Gastrinomas tend to be smaller, multiple, and located in the duodenal wall more often than is seen in patients with sporadic ZES. Establishing the diagnosis of MEN 1 is critical in order to provide genetic counseling to the patient and his or her family and also to determine the recommended surgical approach.
Diagnosis Biochemical measurements of gastrin and acid secretion in patients suspected of ZES play an important role is establishing this rare diagnosis. Often, patients suspected of having ZES will be treated with a PPI in an effort to ameliorate symptoms and decrease the likelihood of possible acid-related complications. The presence of the PPI, which will lower acid secretion and potentially elevate fasting gastrin levels in normal individuals, will make the diagnostic approach in these individuals somewhat difficult. Significant morbidity related to peptic diathesis has been described when stopping PPIs in gastrinoma patients; therefore, a systematic approach in stopping these agents is warranted (see below). The first step in the evaluation of a patient suspected of having ZES is to obtain a fasting gastrin level. A list of clinical scenarios that should arouse suspicion regarding this diagnosis is shown in Table 348-7. Fasting gastrin levels obtained using a dependable assay are usually <150 pg/mL. A normal fasting gastrin, on two separate occasions, especially if the patient is on a PPI, virtually excludes this diagnosis. Virtually all gastrinoma patients will have a gastrin level >150–200 pg/mL. Measurement of fasting gastrin should be repeated to confirm the clinical suspicion. Some of the commercial biochemical assays used for measuring serum gastrin may be inaccurate. Variable specificity of the antibodies used have led to both false-positive and false-negative fasting gastrin levels, placing in jeopardy the ability to make an accurate diagnosis of ZES.
WHEN TO OBTAIN A FASTING SERUM GASTRIN LEVEL
Multiple processes can lead to an elevated fasting gastrin level, the most frequent of which are gastric hypochlorhydria and achlorhydria, with or without pernicious anemia. Gastric acid induces feedback inhibition of gastrin release. A decrease in acid production will subsequently lead to failure of the feedback inhibitory pathway, resulting in net hypergastrinemia. Gastrin levels will thus be high in patients using antisecretory agents for the treatment of acid peptic disorders and dyspepsia. H. pylori infection can also cause hypergastrinemia. Additional causes of elevated gastrin include retained gastric antrum; G cell hyperplasia; gastric outlet obstruction; renal insufficiency; massive small-bowel obstruction; and conditions such as rheumatoid arthritis, vitiligo, diabetes mellitus, and pheochromocytoma. Although a fasting gastrin >10 times normal is highly suggestive of ZES, two-thirds of patients will have fasting gastrin levels that overlap with levels found in the more common disorders outlined above, especially if a PPI is being taken by the patient. The effect of the PPI on gastrin levels and acid secretion will linger several days after stopping the PPI; therefore, it should be stopped for a minimum of 7 days before testing. During this period, the patient should be placed on a histamine H2 antagonist, such as famotidine, twice to three times per day. Although this type of agent has a short-term effect on gastrin and acid secretion, it needs to be stopped 24 h before repeating fasting gastrin levels or performing some the tests highlighted below. The patient may take antacids for the final day, stopping them approximately 12 h before testing is performed. Heightened awareness of complications related to gastric acid hypersecretion during the period of PPI cessation is critical.
The next step in establishing a biochemical diagnosis of gastrinoma is to assess acid secretion. Nothing further needs to be done if decreased acid output in the absence of a PPI is observed. A pH can be measured on gastric fluid obtained either during endoscopy or through nasogastric aspiration; a pH <3 is suggestive of a gastrinoma, but a pH >3 is not helpful in excluding the diagnosis. In those situations where the pH is >3, formal gastric acid analysis should be performed if available. Normal BAO in nongastric surgery patients is typically <5 meq/h. A BAO >15 meq/h in the presence of hypergastrinemia is considered pathognomonic of ZES, but up to 12% of patients with common PUD may have elevated BAO to a lesser degree that can overlap with levels seen in ZES patients. In an effort to improve the sensitivity and specificity of gastric secretory studies, a BAO/MAO ratio was established using pentagastrin infusion as a way to maximally stimulate acid production, with a BAO/MAO ratio >0.6 being highly suggestive of ZES. Pentagastrin is no longer available in the United States, making measurement of MAO virtually impossible. An endoscopic method for measuring gastric acid output has been developed but requires further validation.
Gastrin provocative tests have been developed in an effort to differentiate between the causes of hypergastrinemia and are especially helpful in patients with indeterminate acid secretory studies. The tests are the secretin stimulation test and the calcium infusion study. The most sensitive and specific gastrin provocative test for the diagnosis of gastrinoma is the secretin study. An increase in gastrin of ≥120 pg within 15 min of secretin injection has a sensitivity and specificity of >90% for ZES. PPI-induced hypochlorhydria or achlorhydria may lead to a false-positive secretin test; thus, this agent must be stopped for 1 week before testing.
The calcium infusion study is less sensitive and specific than the secretin test, which, coupled with it being a more cumbersome study with greater potential for adverse effects, relegates it to rare utilization in the cases where the patient’s clinical characteristics are highly suggestive of ZES but the secretin stimulation is inconclusive.
Tumor Localization Once the biochemical diagnosis of gastrinoma has been confirmed, the tumor must be located. Multiple imaging studies have been used in an effort to enhance tumor localization (Table 348-8). The broad range of sensitivity is due to the variable success rates achieved by the different investigative groups. Endoscopic ultrasound (EUS) permits imaging of the pancreas with a high degree of resolution (<5 mm). This modality is particularly helpful in excluding small neoplasms within the pancreas and in assessing the presence of surrounding lymph nodes and vascular involvement, but it is not very sensitive for finding duodenal lesions. Several types of endocrine tumors express cell-surface receptors for somatostatin. This permits the localization of gastrinomas by measuring the uptake of the stable somatostatin analogue111 In-pentreotide (OctreoScan) with sensitivity and specificity rates of >85%.
SENSITIVITY OF IMAGING STUDIES IN ZOLLINGER-ELLISON SYNDROME
Up to 50% of patients have metastatic disease at diagnosis. Success in controlling gastric acid hypersecretion has shifted the emphasis of therapy toward providing a surgical cure. Detecting the primary tumor and excluding metastatic disease are critical in view of this paradigm shift. Once a biochemical diagnosis has been confirmed, the patient should first undergo an abdominal computed tomography (CT) scan, magnetic resonance imaging (MRI), or OctreoScan (depending on availability) to exclude metastatic disease. In addition, the positron emitter 68Ga has been used to label somatostatin analogues for positron emission tomography (PET) with some success. In addition, hybrid scanners combining CT scan with PET scan are also available in certain specialized centers. Once metastatic disease has been excluded, an experienced endocrine surgeon may opt for exploratory laparotomy with intraoperative ultrasound or transillumination. In other centers, careful examination of the peripancreatic area with EUS, accompanied by endoscopic exploration of the duodenum for primary tumors, will be performed before surgery. Selective arterial secretin injection may be a useful adjuvant for localizing tumors in a subset of patients. The extent of the diagnostic and surgical approach must be carefully balanced with the patient’s overall physiologic condition and the natural history of a slow-growing gastrinoma.
STRESS-RELATED MUCOSAL INJURY
Patients suffering from shock, sepsis, massive burns, severe trauma, or head injury can develop acute erosive gastric mucosal changes or frank ulceration with bleeding. Classified as stress-induced gastritis or ulcers, injury is most commonly observed in the acid-producing (fundus and body) portions of the stomach. The most common presentation is GI bleeding, which is usually minimal but can occasionally be life threatening. Respiratory failure requiring mechanical ventilation and underlying coagulopathy are risk factors for bleeding, which tends to occur 48–72 h after the acute injury or insult.
Histologically, stress injury does not contain inflammation or H. pylori; thus, “gastritis” is a misnomer. Although elevated gastric acid secretion may be noted in patients with stress ulceration after head trauma (Cushing’s ulcer) and severe burns (Curling’s ulcer), mucosal ischemia, breakdown of the normal protective barriers of the stomach, systemic release of cytokines, poor GI motility, and oxidative stress also play an important role in the pathogenesis. Acid must contribute to injury in view of the significant drop in bleeding noted when acid inhibitors are used as prophylaxis for stress gastritis.
Improvement in the general management of intensive care unit patients has led to a significant decrease in the incidence of GI bleeding due to stress ulceration. The estimated decrease in bleeding is from 20–30% to <5%. This improvement has led to some debate regarding the need for prophylactic therapy. The high mortality associated with stress-induced clinically important GI bleeding (>40%) and the limited benefit of medical (endoscopic, angiographic) and surgical therapy in a patient with hemodynamically compromising bleeding associated with stress ulcer/gastritis support the use of preventive measures in high-risk patients (mechanically ventilated, coagulopathy, multiorgan failure, or severe burns). Maintenance of gastric pH >3.5 with continuous infusion of H2 blockers or liquid antacids administered every 2–3 h are viable options. Tolerance to the H2 blocker is likely to develop; thus, careful monitoring of the gastric pH and dose adjustment are important if H2 blockers are used. Sucralfate slurry (1 g every 4–6 h) has also been somewhat successful but requires a gastric tube and may lead to constipation and aluminum toxicity. Sucralfate use in endotracheal intubated patients has also been associated with aspiration pneumonia. Meta-analysis comparing H2 blockers with PPIs for the prevention of stress-associated clinically important and overt GI bleeding demonstrates superiority of the latter without increasing the risk of nosocomial infections, increasing mortality, or prolonging intensive care unit length of stay. Therefore, PPIs are the treatment of choice for stress prophylaxis. Oral PPI is the best option if the patient can tolerate enteral administration. Pantoprazole is available as an intravenous formulation for individuals in whom enteral administration is not possible. If bleeding occurs despite these measures, endoscopy, intraarterial vasopressin, and embolization are options. If all else fails, then surgery should be considered. Although vagotomy and antrectomy may be used, the better approach would be a total gastrectomy, which has an exceedingly high mortality rate in this setting.
The term gastritis should be reserved for histologically documented inflammation of the gastric mucosa. Gastritis is not the mucosal erythema seen during endoscopy and is not interchangeable with “dyspepsia.” The etiologic factors leading to gastritis are broad and heterogeneous. Gastritis has been classified based on time course (acute vs chronic), histologic features, and anatomic distribution or proposed pathogenic mechanism (Table 348-9).
CLASSIFICATION OF GASTRITIS