Dyspepsia

A comprehensive search of MedLine through 1/2001 was performed using the following MeSH terms: dyspepsia, Helicobacter infections, Helicobacter pylori, Peptic ulcer.

Dyspepsia is a term used broadly to describe episodic or persistent upper abdominal symptoms believed to arise from the upper digestive tract. [1] Symptoms may or may not be related to eating and may include epigastric pain, bloating, fullness, belching, nausea or early satiety. The term dyspepsia is used to describe symptoms located in the upper abdomen (i.e. epigastrium) and is synonymous with the term indigestion.

Surveys indicate that dyspepsia is a frequent problem in the general population. A review of multiple studies of the frequency of dyspepsia in Europe estimated a prevalence of 19-41%. [2] More recently, an age and sex stratified random sample of 1120 persons from Olmsted County, Minnesota were asked to complete a previously validated self-report symptom questionnaire. Of the 82% of patients responding, 26% reported pain in the upper abdomen in the previous year. [3] Nearly one-third of these patients also had symptoms compatible with irritable bowel syndrome. Another survey in the United States using stricter criteria revealed a prevalence of dyspepsia of 13%. [4]

Frequency of patients seeking medical attention: Less than 25% of patients with dyspepsia seek medical attention for their symptoms. Nevertheless, dyspepsia has been shown to account for approximately 2-5% of all visits to general practitioners.

3.0 Differential Diagnosis of Dyspepsia:

A wide variety of underlying etiologies should be taken into consideration when evaluating the dyspeptic patient. Categories of disease known to cause dyspepsia include medication induced dyspepsia, endocrine/metabolic disturbances, luminal GI tract disease and pancreatic or biliary disease. (Table 1): Causes of Dyspepsia. A thorough medication history should also be ascertained since discontinuation of a potentially offending medication may occasionally alleviate the dyspeptic symptoms.

Notably, Helicobacter pylori gastritis in the absence of peptic ulcer disease has not been definitively shown to be a cause of dyspepsia. [9] Also, although cholelithiasis is commonly listed as a cause of dyspepsia, epidemiologic studies have shown a poor correlation between right upper quadrant pain and the presence of gallstones. [10] Attributing dyspeptic complaints to the presence of gallstones should be done with caution.

Patients without an apparent cause of dyspepsia (no medications or obvious systemic conditions) have been studied by endoscopy in large case series. A review of four series of dyspeptic patients undergoing endoscopy in the United Kingdom revealed the following diagnoses: peptic ulcer disease (20%), gastroesophageal reflux disease (24%), malignancy (2%) and gastritis or duodenitis (21%). [11]

[12] This latter group of patients with normal or near-normal endoscopies (including findings of questionable clinical significance such as gastritis or duodenitis) would typically be categorized as having non-ulcer dyspepsia. Non-ulcer dyspepsia is a term used to describe patients with no biochemical abnormality or structural findings at the time of endoscopic examination to account for dyspeptic symptoms. Gastric cancer occurs in approximately 1-2% of patients presenting with dyspepsia, and occurs almost invariably in patients greater than 45 years of age. [13,14,15,16]

Ulcers in the gastrointestinal tract are defined as focal areas of mucosal loss. When the mucosal loss is superficial and does not extend into the submucosal layer, the defect is called an erosion. When the defect extends deeper into the submucosa, it is termed an ulcer. Ulcers may occur anywhere within the GI tract. In the upper GI tract they may occur in the esophagus as a consequence of gastroesophageal reflux disease. Ulcers within the stomach or duodenum are often termed peptic ulcers. Duodenal ulcers are more common than gastric ulcers.

Approximately 90% of patients with duodenal ulcers have Helicobacter pylori infection. Most of these ulcers are probably caused by Helicobacter pylori infection because approximately 90-95% of patients with H. pylori infection and duodenal ulcers do not have recurrent ulcer disease following successful H. pylori eradication. [17] However the other 5-10% who have ulcer relapse despite successful H. pylori eradication may have other causes of the ulcer. The other major etiology for duodenal ulcer disease is non-steroidal anti-inflammatory drug use. [18] Somewhere between 3-8% of patients on chronic NSAID therapy will develop a duodenal ulcer. Other patients may have high vagal tone leading to excessive gastric acid secretion. More rare causes of duodenal ulcer include gastrinoma, Crohn's disease, and eosinophilic gastroenteritis.

Between 60 and 80% of patients with gastric ulcer have H. pylori infection. Successful eradication of H. pylori also reduces the risk of gastric ulcer relapse. NSAID use is a more common cause of gastric than duodenal ulcers. Approximately 15% of patients on chronic NSAID therapy develop gastric ulcers. [18] Whereas duodenal ulcers are almost never malignant, approximately 2% of benign-appearing gastric ulcers are actually cancers that have ulcerated. This underlines the need to biopsy all gastric ulcers, or to confirm that they have healed following appropriate treatment.

Many patients have ulcers but deny ulcer symptoms. The ulcers may be found incidentally when endoscopy is done for another reason. They also may manifest as bleeding when the unsuspecting clinician places an asymptomatic patient on aspirin or another NSAID. While it used to be thought that NSAID use was the predominant risk factor for an asymptomatic ulcer, recent studies have suggested that the absence of symptoms is not associated with NSAID use but rather with ulcer size and tobacco use. [19]

The typical symptom of an ulcer is intermittent epigastric pain. Classically, this pain awakens the patient at night or occurs a few hours after eating. It is responsive to eating or antacids, both of which neutralize gastric acid. Other associated dyspeptic symptoms of an ulcer include nausea, which may or may not accompany the pain. The presence of significant recurrent vomiting suggests the presence of a complication such as gastric outlet obstruction. Unfortunately, the large amount of overlap of individual symptoms (Table 5) makes the nature of discomfort not useful for the diagnosis of an ulcer in an individual patient.

Ulcers may bleed when the crater area erodes into a blood vessel within the wall of the stomach or duodenum. The presentation may be acute with hematemesis or melena or the bleeding may be more indolent and present with iron deficiency anemia. Signs or symptoms of GI bleeding always require investigation of the cause of bleeding.

Ulcers may cause obstruction of the distal stomach or duodenum due to acute inflammation and edema or due to chronic scarring and fibrosis. Symptoms from obstruction at either the distal stomach or duodenal bulb are similar and are often referred to as those of gastric outlet obstruction. These include post-prandial vomiting, especially following a solid meal. Other symptoms include those of heartburn and regurgitation which can be confused with gastroesophageal reflux disease.

When a posteriorly located ulcer within the duodenum or stomach erodes into the pancreas, the patient may present with pancreatitis. The symptoms are abdominal pain that often radiates to the back, associated nausea and vomiting. The presentation is not different from other causes of acute pancreatitis.

Ulcers that penetrate through the wall of the stomach or duodenum may also result in perforation of the gastrointestinal tract. Patients typically present with the sudden onset of severe pain throughout the abdomen. The resulting chemical and then bacterial peritonitis causes pain that is exacerbated by any physical movement. Fever and hypotension can also result if prompt treatment is not undertaken.

The presence of an ulcer crater in the stomach or duodenum can be detected by either upper endoscopy or upper GI X-ray. Both tests are sensitive and specific for detecting chronic ulcer craters in patients who have not had previous gastric or duodenal surgery and those who have not had a previous ulcer. If either of these situations are present, however, X-ray is less accurate than endoscopy because it cannot distinguish between a scar and an ulcer. Endoscopy is also more sensitive for detecting shallow ulcerations, particularly those that are associated with NSAID use. The relative advantages and disadvantages of these two diagnostic tests are listed in (Table 7).

Patients with complicated ulcers require special considerations. Barium should not be given to patients with a suspected bleeding ulcer because residual barium can obscure further attempts to stop the bleeding using endoscopy or angiography. The presence of a perforated ulcer should be considered when an upright abdominal radiograph reveals the presence of free air under the diaphragm. Gastric outlet obstruction is often best diagnosed with an upper GI X-ray which defines the location and extent of luminal narrowing.

There are two important considerations in the treatment of patients with peptic ulcer disease.

The first is to heal the ulcer and relieve acute symptoms by placing the patient on an acid suppressive medication (H2 receptor antagonist or proton pump inhibitor). (Table 8)

The second is to treat the underlying cause of the ulcer in order to prevent recurrence. This involves testing and treatment for Helicobacter pylori infection and discontinuation of NSAIDS.

There are a number of medications that have been used to heal ulcers. Most of these involve the suppression or neutralization of gastric acid. These agents, along with their dose and recommended duration of therapy, are listed in (Table 8).

There are specific circumstances when one ulcer-healing agent may be selected over another. Pregnant patients should be treated with sucralfate, which is not absorbed and therefore poses no risk of toxicity to the fetus. Patients taking NSAIDs who must continue these agents during ulcer therapy need more potent acid suppression. NSAIDs reduce the efficacy of H2 receptor antagonists so that a proton pump inhibitor is favored in this situation. In general, gastric ulcers are more difficult to heal than duodenal ulcers. More potent acid inhibitors such as a proton pump inhibitor and longer durations of therapy are appropriate for patients with gastric ulcer.

The side effects of the various ulcer-healing agents are also quite different and may impact therapeutic selection. Antacids are effective in healing ulcers but the frequency of dosing may lead to troublesome side effects such as diarrhea. H2 receptor antagonists are generally safe and effective although high doses of cimetidine may be associated with troublesome anti-androgen side effects such as gynecomastia. Drug interactions can also occasionally be a problem for some of the H2 antagonists and proton pump inhibitors (Table 9).

a. Helicobacter pylori: The major advance in the understanding and treatment of patients with peptic ulcer disease was the demonstration that eradication of H. pylori markedly reduced the risk of ulcer recurrence. For example, duodenal ulcers treated with antisecretory therapy have been shown to recur in 80% of patients within one year. Half doses of H2 blockers used as maintenance therapy still resulted in recurrence rates of approximately 25% per year. Recurrence can be reduced to between 5 and 10% per year by eradication of H. pylori. The effect of H. pylori eradication is even more dramatic in patients with bleeding duodenal ulcers. Recurrent bleeding can be reduced to near zero by H. pylori eradication whereas recurrent bleeding is seen within one year in about 30% of patients if H. pylori is not eradicated. [20,21] Methods of H. pylori diagnosis and treatment are discussed later in this chapter.

b. NSAID use: The risk of recurrent ulcer varies depending on the NSAID type as well as dose and duration of therapy. The major risk factors for NSAID associated ulcer complications are a history of previous ulcer or ulcer complication, and concomitant corticosteroid or anticoagulant therapy. [22,23,24] Patients in one of these high-risk groups should only be prescribed an NSAID when absolutely necessary for the treatment of an inflammatory problem. Otherwise a simple analgesic such as acetaminophen would be a much wiser choice. If an NSAID is prescribed, then prophylaxis with another agent should be considered. Misoprostol prevents gastric and duodenal ulcers caused by NSAIDs and reduces ulcer complications by 40%. [25] H-2 receptor antagonists prevent duodenal but not gastric ulcers when used in standard dosages. Double-dose H-2 antagonists and proton pump inhibitors prevent duodenal ulcers and reduce the frequency of gastric ulcers by approximately fifty percent. [26,27]

c. Patients who develop recurrent ulcers despite not having H. pylori infection or taking NSAIDs may require maintenance therapy with an antisecretory drug. H2 receptor antagonists in half doses at bedtime are most commonly used for this purpose.

Non-ulcer dyspepsia (also termed functional dyspepsia) is generally defined as dyspepsia of at least several weeks duration for which no focal or structural abnormality can be found on upper endoscopy and which cannot be explained by any structural or biochemical abnormality. [28,29] Non-ulcer dyspepsia is the most common diagnosis following investigation of dyspeptic patients by endoscopy. Fifty to sixty-five per cent of patients undergoing endoscopy for dyspepsia will have normal examinations or non-specific findings of gastritis or duodenitis and would be classified as having non-ulcer dyspepsia. [30] While non-ulcer dyspepsia has been used as a single term to describe patients with no obvious abnormality to explain dyspeptic symptoms, it likely represents a heterogeneous group of disorders including gastroesophageal reflux disease, motility disorders and pancreaticobiliary disease.

Helicobacter pylori has not been shown definitively to be a cause of chronic dyspepsia in the absence of peptic ulcer disease. Studies comparing the frequency of dyspepsia in persons with a positive H. pylori serology to a matched control population do not show a significant difference in the frequency of dyspepsia between the two groups. [30,31] Acute H. pylori infection may be a cause of transient dyspeptic symptoms. This is supported by the finding that dyspepsia has occurred in volunteers ingesting the organism, and by evidence that persons with recent seroconversion for H. pylori report more dyspeptic complaints than persons without evidence of infection or those with chronic infection. [29,31]

Studies of treating H. pylori for non-ulcer dyspepsia have failed to consistently show a therapeutic benefit. Clinical studies have shown that the vast majority of patients do not have resolution of dyspepsia following H. pylori treatment. Large randomized controlled trials have shown conflicting results regarding the effectiveness of H. pylori treatment in relieving symptoms of non-ulcer dyspepsia. [102, 103, 104, 105] In one large trial, 308 patients were randomized to H. pylori treatment or placebo and a significantly greater frequency of symptom resolution (21% vs. 7%) for patients treated for H. pylori as compared to placebo was shown at 1 year follow-up in an intention to treat analysis. [103] Three other large randomized controlled trials showed no difference in the rate of resolution of dyspeptic symptoms for patients randomized to H. pylori therapy as compared to placebo at 1 year follow-up. [102, 104, 105] In the largest study to date (860 non-ulcer dyspepsia patients randomized to H. pylori treatment or placebo) patients successfully treated for H. pylori (not an intention to treat analysis), had complete symptom relief in 44% of eradicated vs 36% of non-eradicated patients at 1 year (p=0.036). [106]

These studies show that the majority of persons with non-ulcer dyspepsia will continue to experience dyspepsia following Helicobacter pylori treatment. Two of these trials suggest that a small subset of patients with non-ulcer dyspepsia may achieve symptom resolution if successfully treated for H. pylori. [103, 106] Two meta-analyses have also reached conflicting results. A meta-analysis of 13 trials of treatment of H. pylori for non-ulcer dyspepsia showed no benefit for H. pylori treatment. [107] Another meta-analysis of 12 trials found a small but significant benefit to H. pylori treatment (relative risk reduction 9%, number needed to treat = 15 for cure of dyspepsia). [108] Although the available studies report conflicting results, two randomized trials and one meta-analysis suggest that a small subset of patients with non-ulcer dyspepsia treated for Helicobacter pylori infection may experience resolution of dyspeptic symptoms. Antibiotic treatment for patients with non-ulcer dyspepsia and evidence of H. pylori infection is a reasonable option for patients with no other explanation for their symptoms. The preferred regimen is a proton pump inhibitor bid (omeprazole 20 mg or its equivalent), clarithromycin 500 mg bid plus either amoxicillin 1 gm bid or metronidazole 500 mg bid for 10 days.

The most important initial issue in the management of patients with non-ulcer dyspepsia is to determine if pharmacologic therapy is required to alleviate symptoms. Some patients may be reassured by the knowledge that all studies are negative and not seek further medical therapy or investigation. [32]

Common sense dietary recommendations to eat a balanced diet and avoid alcohol, tobacco and coffee are also prudent, although there is no significant evidence that diet plays a significant role in non-ulcer dyspepsia. [29] Establishing a close physician-patient relationship is also an important step in the management of patients with non-ulcer dyspepsia.

Many patients may choose a trial of medical therapy in an attempt to improve their symptoms. A number of different pharmacologic agents have been studied in the treatment of non-ulcer dyspepsia.

Metoclopramide, domperidone and cisapride have been studied in randomized controlled trials for non-ulcer dyspepsia. Most trials have studied patients with dysmotility-like (bloating/belching) or reflux-like (heartburn) dyspeptic symptoms.

Cisapride has been widely studied for non-ulcer dyspepsia but has been removed from the market due to reports of serious cardiac arrhythmias and death.

Metoclopramide 10 mg tid or qid has been shown to be superior to placebo in three randomized controlled trials, with approximately 70% of patients responding as compared to 30% receiving placebo. [33] Metoclopramide is often poorly tolerated due to CNS side effects (including restlessness, drowsiness and extrapyramidal signs) limiting its long-term use.

Although not currently available in the United States, the promotility agent domperidone has undergone extensive testing for non-ulcer dyspepsia. Efficacy has been demonstrated in 10 of 11 studies (with the negative study enrolling only 11 patients) as compared to placebo. [33] A meta-analysis suggests a 56% response rate for domperidone above placebo. [34]

Antisecretory therapy with an H2 receptor antagonist or proton pump inhibitor appears to benefit a subset of patients with non-ulcer dyspepsia.

H2 receptor antagonists have been evaluated in a large number of studies, with approximately 6 studies showing no benefit and 8 studies showing a statistically significant advantage over placebo. [33] A meta-analysis of antisecretory agents for non-ulcer dyspepsia suggested a benefit of only 20% over placebo. [34] Proton pump inhibitors are much less well studied for non-ulcer dyspepsia, but preliminary results of a large randomized controlled trial of omeprazole 20 mg/day suggest a response rate of 18% over placebo. [39]

Large trials have evaluated the role of proton pump inhibitors for non-ulcer dyspepsia. Two identical randomized controlled trials (the Bond and Opera studies) have evaluated the efficacy of omeprazole for non-ulcer dyspepsia and have demonstrated a small benefit above placebo for symptom resolution (38% on omeprazole 20 mg/day as compared to 28% receiving placebo). [109] When symptom resolution was assessed according to the dominant symptom, patients with epigastric pain or heartburn had a significantly better response to omeprazole 20 mg/day as compared to patients with other complaints (including early satiety, bloating and distention.[109] Other investigators have shown that nocturnal pain is associated with a good response to omeprazole, whereas nausea is associated with a poor response. [110] In another large randomized controlled trial, the beneficial effect of 20 mg of omeprazole was shown to be limited to patients with evidence of H. pylori infection. [111] The favorable response of H. pylori positive non-ulcer dyspeptic patients omeprazole may be due to the enhanced effect of acid suppressants that has been demonstrated in H. pylori infected patients as compared to uninfected patients. Lansoprazole has been studied for non-ulcer dyspepsia in large randomized trials involving over 900 patients and a significantly greater percentage of patients achieved complete symptom relief as compared to placebo. [112] These studies suggest that PPI's benefit a subset of patients with non-ulcer dyspepsia and that symptoms are useful to select patients that are more likely to have symptom resolution (as opposed to uninvestigated dyspepsia where symptoms have not been shown to be useful in guiding therapy). The subset of patients with non-ulcer dyspepsia responding to proton pump inhibitors appears to be patients with reflux-like dyspepsia (heartburn) or ulcer-like dyspepsia (epigastric pain). A trial of a proton pump inhibitor is a reasonable option for these patients.

A subset of patients may fail to respond to the above measures and experience continued discomfort. This is often a difficult problem to manage, but a number of options may be attempted. If the patient has been prescribed a H2 blocker, changing to a proton pump inhibitor is a reasonable option. If the patient has evidence of Helicobacter pylori infection and has not yet been treated, a course of antibiotics may be beneficial. Consideration should be given to referral to a mental health professional if there is a history of depression or physical or sexual abuse. Other options that may be of benefit include stress management and relaxation training (such as yoga, meditation or biofeedback).

Perhaps the most important measure for patients with intractable symptoms is to provide regular follow-up visits and reassurance. Additional imaging or investigative measures should be avoided unless there is a change in the clinical status since this may undermine the patients confidence in the physician. The emphasis should be shifted from finding a cause of the symptoms to finding a productive coping mechanism. A trial of tricyclic antidepressants or selective serotonin re-uptake inhibitors may sometimes be useful in relieving dyspeptic symptoms, although these agents remain poortly understudied for non-ulcer dyspepsia. A small randomized controlled trial demonstrated a benefit for symptoms of non-ulcer dyspepsia treated with amitryptiline 50 mg q h.s. as compared to placebo. [40]

Behavioral therapy for non-ulcer dyspepsia remains poorly studied. A single controlled trial of relaxation therapy has shown benefit. [113] Trials of psychotherapy have also shown benefit. [114, 115] Further studies of behavioral therapies are needed in order to better assess their efficacy in non-ulcer dyspepsia. A combination of peppermint oil and caraway oil has been compared to cisapride and placebo and appears to result in symptom improvement. [116, 117]

Gastroesophageal reflux disease is also an important cause of dyspeptic symptoms. This topic is discussed more thoroughly in the chapter on heartburn and regurgitation and in the section discussing clinical features of dyspeptic symptoms.

The history should seek to define the duration, type and location of symptoms. The most important aspect of initial evaluation is to rule out the presence of "alarm symptoms" (Table 3). These alarm symptoms include weight loss, persistent vomiting, dysphagia, anemia, or bleeding and raise the suspicion of underlying malignancy or complicated ulcer disease. The presence of any of these symptoms warrants prompt investigation by endoscopy.

Clinical features of dyspeptic symptoms (such as location and nature of pain or response to eating or antacids) are unable to reliably differentiate between peptic ulcer disease, non-ulcer dyspepsia, gastroesophageal reflux disease and malignancy. [41,42,43,44,45,46,47] When attempts are made to classify patients as having ulcer-like dyspepsia (pain predominant), reflux-like dyspepsia (heartburn predominant) or dysmotility-like dyspepsia (gaseousness & bloating predominant) greater than 40% of subjects may be classified into more than 1 category. [3] The large amount of overlap between types of dyspeptic symptoms for patients with underlying gastric ulcer, duodenal ulcer, gastroesophageal reflux disease or non-ulcer dyspepsia makes the clinical history of little use in directing therapy toward a specific disease state. A study assessing the utility of specific symptoms in a mathematical predictive scoring model found that no symptom model could service as a reliable decision instrument in primary care. [48] The frequency of different symptoms in gastric ulcer, duodenal ulcer and non-ulcer dyspepsia is illustrated in (Table 5).

Symptoms are also poor predictors of underlying gastroesophageal reflux disease (GERD). A number of different dyspeptic symptoms may be present among patients with this disorder. A prospective study of 107 patients with endoscopically confirmed reflux esophagitis revealed that while a burning sensation in the epigastrium and retrosternal area are common, other symptoms such as acid regurgitation, belching, epigastric pain, postprandial fullness, retrosternal pain, pharyngeal burning, nausea and dysphagia also occur frequently. [49] The value of the clinical history has also been evaluated in large series of patients with suspected GERD undergoing 24 hour esophageal pH monitoring . A study of 304 patients with suspected GERD demonstrated that the simple presence or absence of heartburn or acid regurgitation is insensitive (73% and 66%) and non-specific (53% and 58%) for abnormal amounts of acid reflux on pH probe monitoring. [50] For patients with a clearly dominant symptom of heartburn or acid regurgitation these symptoms become more specific (89% and 95%) but are insensitive (38% and 6%) and have low positive predictive values (81% and 57%). A similar study of 220 patients of patients with suspected GERD who underwent a symptom questionnaire and pH probe monitoring revealed that a history of heartburn had a sensitivity and specificity of 49% and 69% and a positive predictive value of 59% for the presence abnormal acid reflux when compared to 24 hour pH probe monitoring. [51] A complaint of acid regurgitation had a sensitivity and specificity of 35% and 83% and a positive predictive value of 66%.

A thorough physical examination should be performed and may occasionally identify an abdominal mass lesion, evidence of gastrointestinal blood loss or organomegaly. The finding of epigastric tenderness is not specific for any of the causes of dyspepsia. [52] The cost-effectiveness of routine laboratory studies in the evaluation of dyspeptic patients remains unproven, but current recommendations include obtaining a complete blood count, electrolytes, serum calcium level and liver function tests. Thyroid function testing, serum amylase and a urine pregnancy test may also be performed when appropriate.

The optimal initial approach to the evaluation of patients with dyspepsia is unclear. The approach taken may be individualized according to a patient's desire for investigation and the cost and availability of endoscopy in an individual practitioner's setting. Possible approaches may include either an initial trial of medical therapy (H2 receptor antagonist or proton pump inhibitors), testing and treatment for Helicobacter pylori infection, or investigation by endoscopy.

Limiting endoscopic examination to groups of patients with an increased probability of abnormal findings may allow better utilization of resources. Age is commonly used as a discriminating factor in the decision to perform endoscopy since endoscopic abnormalities are more likely to be found in older persons. [53] Gastric malignancy occurs in approximately 1-2% of patients presenting with dyspepsia, but occurs almost invariably in patients greater than 45 years of age. [13,14,15,16] This has led to recommendations that initial endoscopy be limited to persons older than age 45 years and those patients presenting with alarm features, (Table 3). [54] An international working party has subsequently recommended that the age threshold for referral for endoscopy be increased to 50 years (rather than 45 years) for Western patients. [118]

The American Gastroenterological Association and World Congresses of Gastroenterology have recommended empiric testing and treatment for Helicobacter pylori for patients less than 45-50 years of age who lack alarm symptoms. [64, 118] A therapeutic trial of an antisecretory or prokinetic agent is suggested for those patients testing negative for Helicobacter pylori. For patients age > 45 investigation by endoscopy is recommended as the preferred initial strategy has been recommended by the American Gastroenterological Association. The recommendations of the AGA position paper are based largely upon decision analyses suggesting that empiric testing and treatment for Helicobacter pylori is a cost-effective initial management strategy. There is also evidence from clinical trials that empiric H. pylori therapy for uninvestigated dyspepsia is a cost-effective strategy and results in fewer endoscopies performed on long-term follow-up. [120, 121] Patients with persistent dyspeptic symptoms following an empiric treatment strategy should undergo investigation by endoscopy.

In a position paper addressing the subject of dyspepsia in 1985, the American College of Physicians recommended a 6 to 8 week trial of H2 receptor antagonists as initial therapy for dyspeptic patients without alarm symptoms. [55] Esophagogastroduodenoscopy (EGD) was recommended for those patients with persistent symptoms despite 6 to 8 weeks of therapy or for those who have no improvement of symptoms after 7 to 10 days. These recommendations were based largely upon the belief that the prescribed therapy would be the same (an H2 receptor antagonist) for all conditions likely to be found at the time of endoscopy (peptic ulcer disease, gastroesophageal reflux disease, non-ulcer dyspepsia). It is now recognized that the optimal therapy for each of these conditions is not a limited course of acid suppression. The recommended therapy for underlying peptic ulcer disease would involve testing and treatment for Helicobacter pylori if present in order to dramatically decrease the rate of ulcer recurrence. [56] H2 receptor antagonists have been studied for non-ulcer dyspepsia and have been shown to be little or no better than placebo. [57,58] Therapy for gastroesophageal reflux disease would typically involve a longer course of acid suppression as well as lifestyle and dietary modifications.

The chronic nature of peptic ulcer disease, gastroesophageal reflux disease and non-ulcer dyspepsia also raises the concern that a strategy of a course of H2 receptor antagonists of limited duration may result in high rates of repeat visits and simply defer costs to a later point in time. A randomized controlled trial has addressed this issue by assessing long-term outcomes following empirical treatment with acid suppression as compared to initial endoscopy for patients with dyspepsia. [59] Four hundred fourteen patients presenting with dyspepsia were randomized to receive either initial endoscopy or H2 receptor antagonist therapy. At one year follow-up, no differences in symptoms or quality of life were found. Interestingly, initial management with H2 receptor antagonists was associated with higher costs and lower levels of patient satisfaction as compared to initial endoscopy due to a higher number of sick leave days and continued medication use. While the results of this trial (performed in Denmark) are difficult to extrapolate to the United States, it raises the possibility that an initial trial of H2 receptor antagonists may simply defer costs to a later point in time and result in lower levels of patient satisfaction. Nonetheless, despite concerns regarding the long-term cost-effectiveness of this approach, H2 receptor antagonists remain an acceptable initial management strategy.

The recognition of Helicobacter pylori as the causative agent for most cases of peptic ulcer disease and the ability to test for the presence of Helicobacter pylori by non-invasive means has raised the question of whether testing for the presence of this organism may play a useful role in the management of patients with dyspepsia. The presence of Helicobacter pylori infection may be detected non-invasively by performing an IgG serology, H. pylori stool antigen test, or a urea breath test. These tests do not require endoscopy in order to detect the presence of Helicobacter pylori gastritis.

Studies indicate that not performing endoscopy on dyspeptic patients (age less than 45) with a negative Helicobacter pylori serology who deny NSAID use or alarm symptoms may be a satisfactory approach and is likely to result in decreased medication and endoscopy use.

Directing the management of dyspeptic patients according to Helicobacter pylori serology has been proposed as a potential means of reducing endoscopy workload. Prospective studies have shown that a negative Helicobacter pylori serology and the absence of a history of non-steroidal anti-inflammatory drug (NSAID) use virtually exclude the possibility of underlying peptic ulcer disease for patients age 45 or less. [60,61,62,63] While a positive Helicobacter pylori IgG serology does not necessarily imply the presence of peptic ulcer disease, a negative serology in a patient denying NSAID use makes peptic ulcer disease extremely unlikely. This is in agreement with the concept that nearly all ulcers result from either Helicobacter pylori infection or NSAID use. When applied to patients age 45 or less with dyspepsia referred for endoscopy, a strategy of performing endoscopy only on patients testing positive for Helicobacter pylori, using NSAIDS or with alarm symptoms would be estimated to reduce endoscopy workload by 23-48%. [60,61,62,63] This has not been studied in older patients due to concerns of missing an underlying early gastric carcinoma. A prospective study of simply reassuring patients with a negative H. pylori serology and who deny alarm symptoms or NSAID use found that very few patients were re-referred for endoscopy over 6 months follow-up and medication use was decreased as compared to a group of H. pylori negative patients undergoing endoscopy. [32]

Testing dyspeptic patients for Helicobacter pylori (serology, stool antigen or breath test) and empirically treating positive patients with antibiotics appears to be a cost-effective strategy for the management of dyspepsia. Empiric H. pylori treatment would be expected to effectively treat patients with underlying peptic ulcer disease (approximately 20% of patients) as the etiology of their dyspepsia. The role of H. pylori gastritis (without peptic ulcer disease) as a cause of non-ulcer dyspepsia remains unclear (see above) and it is likely that many patients may be "unnecessarily" treated for H. pylori. However, treatment of H. pylori may be expected to result in other long-term benefits through prevention of peptic ulcer disease and possibly gastric cancer. [119]

A position paper by the American Gastroenterological Association has endorsed a strategy of empiric testing and treatment for H. pylori infection. The University of Washington recommended strategy for persons with dyspepsia is:

The University of Washington Practice Guidelines for dyspepsia are based upon randomized controlled trials and decision analyses comparing strategies of dyspepsia management.

Evidence from clinical trials supports empiric H. pylori testing and treatment for uninvestigated dyspepsia as a strategy that results in similar or better symptoms as compared to initial endoscopy, with fewer endoscopies performed on long-term follow-up. [120, 121]

A randomized trial of 500 patients with uninvestigated dyspepsia compared prompt endoscopy to H. pylori testing plus eradication therapy. No significant difference in symptoms was found at 1 year follow-up between the two groups. The H. pylori testing group had 60% fewer endoscopies performed during the follow-up period suggesting that this approach is cost-effective. However, fewer patients randomized to the H. pylori testing strategy were satisfied with their treatment. [120] In another trial in which [104] H. pylori positive patients were randomized to eradication therapy or endoscopy, dyspepsia scores were found to be significantly better in the eradication group, and only 27% of the eradication therapy group had been referred for endoscopy at 12 months follow-up. Several other clinical trials comparing dyspepsia management strategies have been reported in abstract form and support an approach of empiric H. pylori testing and treatment for persons with uninvestigated dyspepsia. Patients with persistent dyspeptic symptoms following an empiric treatment strategy should undergo investigation by endoscopy.

Decision analyses graphically depict different clinical outcomes according to a decision tree. The probability of events for each branch of the decision tree is estimated according to the known literature, and costs are attached to each clinical outcome. As will be discussed, conflicting conclusions have been found in decision analyses regarding dyspepsia. The published clinical trials comparing clinical outcomes of dyspepsia management strategies provide stronger evidence to support the recommendation of H. pylori testing and treatment as the preferred strategy.

A comprehensive decision analysis by Silverstein et al. compared management strategies over a 1 year time-frame for patients presenting with dyspepsia. [65] Strategies compared were treatment according to EGD findings, empirical treatment with an H2 receptor antagonist and empirical treatment according to H. pylori serology result. Two different approaches were assessed for empirical treatment according to H. pylori serology result: endoscopy for patients with a positive serology or antimicrobial therapy for H. pylori. This analysis included appropriate clinical assumptions of the frequency and natural history of underlying gastroesophageal reflux disease, peptic ulcer, gastric cancer and non-ulcer dyspepsia. The results of the analysis showed that 1-year cost for the management of a patient with dyspepsia was $2122.60 for a strategy of empirical therapy with an H2 receptor antagonist, $2162.50 for initial endoscopy and $2109.26 for initial treatment of patients with a positive H. pylori serology with antimicrobials (a difference of approximately only 2%).

These findings contradict a decision analysis model by Ofman et al that compared endoscopy and empiric antibiotics for dyspeptic patients with positive H. pylori serologies. Antimicrobial therapy for H. pylori yielded an average cost savings of $456 per patient as compared with initial endoscopy. [66] The predicted cost savings was accounted for by reduced use of endoscopy which more than offset the costs associated with increased antibiotic use. It was concluded that the cost of endoscopy would have to be reduced by 96% before the cost of an initial endoscopic approach would be equal to the strategy of initial antibiotics. This model is limited to patients with a positive H. pylori serology and has been criticized because of the high endoscopy costs used in the model ($1211) and low predicted annual recurrence rates of non-ulcer dyspepsia (55%).

Another model by Sonnenberg compared the outcome of H. pylori serological testing for patients with dyspepsia. [67] This model found that a response rate of 5-10% for non-ulcer dyspepsia would make testing and treatment of H. pylori beneficial, regardless of other potential benefits. Another model assessed the costs over 10 years for a strategy of screening for and eradicating H. pylori in patients under 45 years of with dyspepsia. [69] These investigators found that cost savings accomplished by an empiric treatment strategy would take almost 8 years to be realized.

The conflicting evidence generated by these decision models makes clear the need for additional clinical trials in order to clarify whether empiric testing and treatment for H. pylori is actually cost-effective. At this time the published clinical trials indicate that endoscopy use is markedly decreased when an empiric treatment strategy is employed and that symptoms are equivalent or better than a strategy employing initial endoscopy. Testing and treatment for H. pylori on the basis of serology, urea breath test or stool antigen is the recommended approach for patients less than 50 years of age who do not have alarm features (Table 3). For additional information on H. pylori testing and treatment, refer to the sections Testing for Helicobacter pylori Infection and Treatment of Helicobacter pylori Infection.

Tests for Helicobacter pylori may be divided into those that require endoscopy (invasive tests) and those that do not require endoscopy (non-invasive tests).

Mucosal biopsies obtained at the time of endoscopy are an effective means of diagnosing H. pylori infection (sensitivity and specificity estimated to be 98% [70]). False negative biopsy results are common for patients on proton pump inhibitors since these agents has been shown to cause a decline in the density of H. pylori organisms infecting the gastric mucosa. [71] The organisms may be identified on routine H&E staining, although special stains such as the Giemsa or Genta stains increase the sensitivity. [72] The presence of histologic gastritis is a predictor of infection with H. pylori. [73] If H. pylori organisms are not seen on H&E stained biopsies and there is evidence of gastritis, special stains should be performed in order to search for H. pylori organisms. Four biopsies should be taken at the time of endoscopy (two antral and two corpus). [74]

The Helicobacter pylori organism possesses high urease activity. Mucosal biopsies may be tested for urease activity with sensitivity and specificity for H. pylori >90%. This technique involves placing a tissue specimen (obtained at the time of endoscopy) into a urea rich agar medium with a pH sensitive dye. Placing one antrum and one corpus biopsy specimen in the agar well appears to increase the sensitivity by approximately 4% as compared with one antrum biopsy alone. [75] The urease activity of H. pylori liberates ammonia from urea and raises the pH of the medium which in turn causes a change in color. Pretreatment with proton pump inhibitors may result in a false negative rapid urease test. Brand names for this test include the CLOtest, hpfast and Pylori Tek. The advantage of rapid urease testing is that these tests may turn positive before the patient leaves the endoscopy suite and treatment may be initiated at that time. The final results of the CLOtest or hpfast are read at 24 hours and the PyloriTek is read 2 hours following initial testing.

Culture is not often used as a diagnostic tool for the detection of H. pylori infection in clinical practice. This is due to the difficulty in culturing this fastidious organism. Optimal growth conditions include a moist microaerophilic atmosphere at 37° C and the use of chocolate or blood agar. [76] Experienced laboratories have achieved isolation rates estimated to be only in the range of 75-90%. [77] The relative insensitivity of this technique (and the superior results using other simpler diagnostic techniques) has resulted in the use of culture primarily as a research tool. A number of clinical factors have also been proposed as contributing to the failure to culture H. pylori. These include ingestion of topical anesthetic, use of simethicone during endoscopy, recent use of antibiotics or acid suppressive medications, contamination of biopsy forceps with glutaraldehyde or other microorganisms and the patchy distribution of H. pylori in the gastric mucosa. [77] In clinical practice, use of another diagnostic technique (as outlined in this section) is the preferred strategy for detection of H. pylori infection.

Non-invasive Testing for H. pylori (not requiring endoscopy and mucosal biopsy):

IgG serology is an indicator of ongoing H. pylori gastritis with a sensitivity and specificity of greater than 90%. Kits using either serum or whole blood are now available. [78,79,80,81] Serology is most suitable for the initial diagnosis of H. pylori infection but less useful for documenting successful treatment. A number of different enzyme-linked immunoassay (EIA) kits are commercially available in order to detect IgG antibodies to H. pylori. Different kits detect antibodies to different antigenic components of the H. pylori organism, including high-molecular-weight cell-associated proteins (HM-CAP), a partially purified H. pylori antigen (G.A.P.) and a urease-enriched antigen preparation (Pylori-Stat). The results of a serologic test are usually reported as a qualitative result (positive, negative, indeterminate) based upon a predetermined cutoff value in EIA units. The quantitative result (expressed in EIA units) falls slowly following successful treatment with antibiotics and a 20% decline in titer over 6 months has been shown to correlate with successful eradication. [82] The use of serologic testing in order to document eradication has not entered common clinical practice due to the inconvenience of waiting 6 months prior to repeat testing and the need to save acute sera for comparison to the convalescent sera in the same assay.

This non-invasive test also utilizes the urease activity of H. pylori for the purpose of diagnosing active infection. Sensitivity and specificity for both the ¹⁴C and ¹³C breath tests are greater than 90%. [80,83,84] These tests are useful both for diagnosing active infection and confirming eradication following treatment with antibiotics. For the breath test, the patient ingests urea labeled with either ¹³C or ¹⁴C. If H. pylori is present in the stomach, the urease activity of H. pylori splits urea into ammonia and carbon dioxide. The labeled carbon dioxide is then detected in the expired breath. ¹³C-labeled urea is non-radioactive whereas ¹⁴C-labeled urea contains a minute dose of radioactivity. The amount of isotope used in the ¹⁴C breath test is so small that it has been estimated to result in less radiation exposure than received from background radiation in one day. The ¹⁴C breath test has the advantage of being less expensive than the ¹³C breath test. Proton pump inhibitors, antibiotics and bismuth may cause false negative breath tests and it is recommended that patients be off of antibiotics and bismuth for 1 month prior to the breath test. Proton pump inhibitors must be discontinued 1 week prior to breath testing. [85,86] A comparison of the ¹³C and ¹⁴C breath tests is shown in (Table 6).

This non-invasive test uses an EIA to detect H. pylori antigens in stool samples. It has been shown to be a highly sensitive, specific and accurate test for H. pylori (>90%). [122] As is the case for the urea breath test, proton pump inhibitors and bismuth compounds may cause false negative stool antigen tests, whereas H2 blockers do not cause false negative stool antigen tests. [123] When the H. pylori stool antigen test is used following treatment to document successful cure, false positive tests may occur, resulting in a lower positive predictive value than in the pretreatment setting (although a negative test still has a high negative predictive value, indicating successful treatment). [124, 125] For more information on the H. pylori stool antigen test you may review the materials presented at a University of Washington Gastroenterology Section Journal Club. [View/Download Presentation]

Successful treatment of H. pylori infection requires combinations of 2 or 3 antibiotics; the use of single antibiotic agents has been shown to be ineffective.

A wide variety of treatment regimens have been studied. We endorse the following regimen for treatment: a proton pump inhibitor with two antibiotics (clarithromycin 500 mg plus either metronidazole 500 mg or amoxicillin 1 gm) all given bid for 10-14 days.

In considering the choice of therapy for H. pylori infection, it is important to recognize that results of some clinical trials may be difficult to apply to practice in the United States due to significant geographic variability in antibiotic resistance rates. Resistance to metronidazole or clarithromycin has been shown to decrease treatment success rates. Antibiotic susceptibility testing is not commonly performed in the United States. The choice of antibiotic regimen for H. pylori may be influenced by prior exposure to antibiotics and to patient compliance factors. Treatment regimens are often difficult for patients to fully comply with since they typically involve multiple different medications for one to two weeks that produce bothersome side effects. Compliance has been shown to be an important component in treatment success. [87] Eradication success rate is likely the primary factor in cost-effectiveness; failure of eradication results in increased costs by repeat physician visits and ulcer recurrence which may outweigh the initially greater cost of a particular antibiotic regimen. [88]

The following regimens have been studied as treatment for H. pylori infection (Table 4):

Bismuth subsalicylate (Pepto-Bismol) 2 tablets po QID, Metronidazole 250 mg po QID, Tetracycline 500 mg po QID.

This remains one of the most thoroughly studied and widely used regimens for H. pylori therapy. The treatment success rate with a proton pump inhibitor plus clarithromycin and either amoxicillin or metronidazole appears better than BMT and the BID dosing results in better compliance; for these reasons BMT is not considered the optimal first-line regimen. Reviews of published studies using BMT for 10-15 days have revealed overall H. pylori eradication rates of 89% [89] and 84%. [90] Meta-analyses have revealed mean eradication rates of 76-88%. [91,92] A two week course appears to be more effective than a one week course. [92,93]

Triple drug therapy is usually given with an H2 receptor antagonist (cimetidine 400 mg po BID, ranitidine 150 mg po BID, famotidine 20 mg po BID or nizatidine 150 mg po BID). The H2 receptor antagonist is included in order to promote ulcer healing and possibly lessen dyspeptic symptoms associated with therapy. Acid suppression may also improve eradication rates by enhancing the effect of the antibiotic. [94] Administration of BMT with a proton pump inhibitor (omeprazole 40 mg/day or its equivalent and bismuth subcitrate instead of bismuth subsalicylate) has been shown in one study to decrease the incidence of gastrointestinal side effects and achieve an eradication rate to 98% [95] while another study showed an eradication rate of 86% with this regimen. [96]

For patients intolerant of tetracycline, amoxicillin (not ampicillin) 500 mg po QID may be substituted, although eradication rates are lower than when tetracycline is used. [91,92] Metronidazole may produce side effects such as nausea, taste disturbances or diarrhea in approximately 30% of patients although these were severe enough to cause discontinuation of antibiotics in only 3-7% of patients. [89,90,93] These results are summarized in (Table 4).

The recommended first line therapy for H. pylori infection is a proton pump inhibitor (PPI) plus clarithromycin and either amoxicillin or metronidazole. Although significantly more expensive than the BMT regimen, these regimens are simpler to comply with since therapy is shortened to 10 days and is given BID instead of QID. Although firm data are lacking, it is felt that these regimens generally result in fewer side effects and are better tolerated than BMT therapy. Proton pump inhibitor regimens (omeprazole, lansoprazole, pantoprazole, esomeprazole, or rabeprazole) given with 2 antibiotics (a combination of metronidazole, clarithromycin or amoxicillin) have generally resulted in eradication rates in the range of 90%. [92] A large European and Canadian study of 787 patients combining a PPI with two antimicrobials for 1 week confirmed the efficacy of this approach. [97, 97A] Increasing the duration of treatment to 10 or 14 days may increase the efficacy slightly. In a randomized trial comparing 7, 10 or 14 day treatment durations for BID amoxicillin, omeprazole and clarithromycin, eradication rates were 86% at 7 days, 90% at 10 days and 92% at 14 days. [98]

Any of the available proton pump inhibitors (omeprazole 20 mg BID, lansoprazole 30 mg BID, rabeprazole 20 mg BID, pantoprazole 40 mg BID) may be used. Esomeprazole (40 mg) may be given once daily in combination with clarithromycin and amoxicillin; this appears as effective as twice daily proton pump inhibitor-based regimens. [126]

Omeprazole in combination with clarithromycin alone is an FDA approved regimen for H. pylori therapy, although this combination is not recommended due to lower eradication rates with this regimen. A combination of omeprazole, amoxicillin and metronidazole has been shown to result in lower eradication rates than PPI regimens containing clarithromycin and therefore clarithromycin containing regimens are recommended preferentially. [97,99] One should be cautioned that many of the studies of H. pylori therapy have been performed outside the United States where resistance rates may differ. The results are summarized in (Table 4).

Ranitidine bismuth citrate (RBC) is a novel compound which in combination with clarithromycin has been shown to result in acceptable eradication rates for H. pylori infection. RBC consists of covalently bound ranitidine, bismuth and citrate which readily dissociates in gastric juice and results in acid suppressive and mucosal protective effects. The theoretic advantage of this regimen is that the RBC compound results in fewer pills being taken per day and this compound may possibly reduce the emergence of resistance to clarithromycin or omeprazole. The FDA approved dosing regimen is RBC 400 mg BID for 4 weeks given together with clarithromycin 500 mg TID for the first 2 weeks only. This regimen has resulted in eradication rates of 73-84%. [100,101] This is an acceptable alternative regimen for patients intolerant to BMT, although the overall eradication rates appear slightly lower than using a PPI with two antibiotics. The results of the recommended H. pylori regimens are summarized in (Table 4).

Referral for prompt investigation by endoscopy is an acceptable initial management option for patients with dyspepsia. Patients with continued or recurrent symptoms should also be referred for investigation by endoscopy. Endoscopy should be performed rather than upper gastrointestinal X-ray studies since endoscopy is more sensitive for detection of peptic ulcer disease, reflux esophagitis and malignancy and may allow biopsies to be taken should an abnormality be found. An approach of prompt endoscopy for dyspeptic patients has the possible advantages of allowing therapy directed toward a specific etiology in approximately 30-50% of cases (peptic ulcer disease or reflux esophagitis) and has the potential to provide significant reassurance for patients with a normal endoscopy. At this time it is unclear whether long-term cost savings can be achieved by pursuing empirical medical therapy with H2 blockers or antibiotics for H. pylori infection as compared to initial endoscopy for dyspeptic patients. A randomized controlled trial comparing H2 receptor antagonists to prompt endoscopy found that after 1 year there were no significant differences in symptoms or quality of life measures. [59] However, the group treated by prompt endoscopy had lower costs, mainly due to lower number of sick-leave days and less medication use. The patients undergoing prompt endoscopy also appeared more satisfied with their medical care. While the results of this European trial suggest that an initial trial of H2 receptor antagonists may simply defer investigation and costs to a later point in time, the results are difficult to apply to patients receiving care in the United States. The three decision analysis models described in the section on medical therapy use probabilities and cost estimates based upon the population of the United States but have reached conflicting conclusions regarding the cost-effectiveness of empirical medical approaches as compared to prompt endoscopy. [65,66,68] The most comprehensive analysis of patients presenting with uninvestigated dyspepsia suggests that the long-term costs of prompt endoscopy are likely to be equivalent to empirical medical therapy. [65] While the results of clinical trials are needed in order to clarify the more cost-effective approach, prompt endoscopy remains an acceptable approach.

Indications for referral of patients with dyspepsia would include the following:

The presence of "alarm" symptoms suggests an underlying serious abnormality (e.g. malignancy, complicated peptic ulcer disease) and warrants prompt investigation by endoscopy. Alarm symptoms include dysphagia, weight loss, persistent vomiting, and evidence of GI bleeding. Patients with iron deficiency anemia and those with an abnormal abdominal examination should also be referred.

Patients not responding to an initial trial of acid suppression, prokinetic agents, or an empirical treatment approach for H. pylori warrant referral for endoscopy. Endoscopy may effectively direct therapy toward underlying peptic ulcer disease, gastroesophageal reflux disease, non-ulcer dyspepsia or rarely malignancy.

3. Patients preferring investigation as the initial step in the evaluation of dyspepsia.

It is unclear whether long-term cost savings may be achieved by pursuing medical therapy with H2 receptor antagonists or antibiotic therapy for H. pylori. Selected patients may feel more satisfied with initial investigation in order to guide therapy.

4. Patients with a diagnosis of non-ulcer dyspepsia experiencing continued symptoms despite an initial trial of reassurance and medical therapy.

Patients with a confirmed diagnosis of non-ulcer dyspepsia may benefit from re-referral to a gastroenterologist in order to further discuss additional treatment options.

Patients should have history and physical examination performed prior to referral. The presence of "alarm" symptoms of dysphagia, weight loss, persistent vomiting, and GI bleeding, or an abnormal abdominal examination should result in prompt referral to a gastroenterologist. Laboratory studies obtained as a part of the evaluation for dyspepsia should include a CBC, electrolytes, calcium and liver function tests.