A 59-Year-Old Woman With Gastroesophageal Reflux Disease and Barrett Esophagus

DR BURNS: Mrs J is a 59-year-old woman with a long history of gastroesophageal reflux disease (GERD) and documented Barrett esophagus. She lives near Boston with her husband and has 2 grown children. She sees her primary care physician, Dr W, and she has managed care insurance.

Mrs J initially developed symptoms of heartburn in the early 1990s and was prescribed ranitidine, 150 mg/d, which was later increased to twice daily without improvement. In 1993, Mrs J underwent an upper GI series, the results of which were normal; in 1994, test results for Helicobacter pylori were negative.

When Mrs J initially saw Dr W in 1996, she was taking ranitidine, 150 mg twice daily. At that time, her symptoms persisted throughout the day and night. She was drinking one glass of wine in the evening but did not find that it worsened her symptoms. Her symptoms did worsen after eating spicy foods.

Dr W decided to switch Mrs J's medication to omeprazole, 20 mg/d for 6 weeks, and to proceed with an upper endoscopy if her symptoms persisted after completing the course of therapy. Her symptoms were well controlled with omeprazole but returned when she discontinued the medication. In December 1996, she underwent an upper endoscopy with findings that were consistent with Barrett esophagus with a "single mildly dysplastic gland." She restarted omeprazole, 20 mg/d for 6 weeks, at which time the endoscopy was repeated. The biopsy showed Barrett esophagus without evidence of dysplasia. The plan was to continue taking omeprazole, 20 mg/d, and to have yearly endoscopies. Her most recent endoscopy in March 2002 again showed Barrett esophagus without evidence of dysplasia. The length of the Barrett esophagus has not been specified in the reports.

Currently, her symptoms are well controlled with omeprazole. Mrs J is 5 ft 4 in and weighs 129 lb; she has mild degenerative arthritis. She is taking hormone replacement therapy with conjugated estrogens, 0.625 mg/d, and progesterone, 100 mg/d. She is allergic to penicillin. She has a 40 pack-year smoking history; she quit smoking in the late 1980s. She exercises regularly. Her mother is alive and well and her father died of lung cancer.

Mrs J and Dr W wonder whether she needs yearly endoscopies, whether her risk of esophageal cancer is lessening given several biopsies without evidence of dysplasia, and whether her diet and medication use might influence the development of dysplasia.

For a long time, I was having different heartburn symptoms, which started relatively mild and then got worse. My physician at the time recommended Zantac (ranitidine). The symptoms were present all the time. They were worse with heavy meals and worse around dinnertime. But I thought, in my naive head, that I shouldn't be taking that much Zantac. So I would try not to take a pill until as late in the day as I possibly could. But usually around 11 AM I couldn't tolerate it anymore.

In the beginning, I was told never to have any chocolate, coffee, or wine—there was a long list—the "sin list." At first, I was very rigid about it. And then I asked for another opinion. One doctor told me that it didn't matter what I ate, I could get dysplasia anyway.

Sometimes, now, after I eat or drink, I get heartburn within an hour or two. But it's not severe like it used to be. I might be able to eat a cracker and make it go away.

I'm in the middle of evaluating what to do about hormone replacement therapy. I'm curious to know if there's any interaction, before I decide what to do. Another question I have has to do with the rigidity of how one maintains one's diet. Do you follow what bothers you more than the so-called sin list? Also, I would like to ask, if I am really vigilant, what are the chances of anything ever happening?

Mrs J has a long history of heartburn symptoms. After her upper endoscopy showed that she had Barrett esophagus, the gastroenterologist recommended some dietary changes as well as continuing omeprazole and a program of annual upper endoscopies as surveillance for the development of dysplasia.

I have several questions about her care. Does any evidence indicate that a nightly glass of wine would be harmful in terms of her risk of developing Barrett esophagus? How important is it to avoid medications that we would otherwise be using for her arthritis, such as nonsteroidal anti-inflammatory agents? Are any preventive treatments available? What treatments might be available for her were she to develop dysplasia? What statistically happens to risk over time? Does the fact that she has had 6 negative studies mean that her risk may be lower, and could we safely do these endoscopies less frequently? Is there any other way to do surveillance besides endoscopies?

What are the epidemiology and prevalence of GERD? What are the initial evaluation and treatment for patients with symptoms suggestive of GERD? What is the role for H pylori testing and treatment? What are rates of healing with histamine 2 (H₂) blockers and proton pump inhibitors (PPIs)? What is known about the long-term adverse effects of these medications? When should testing proceed to endoscopy? What are the suggested treatment and follow-up? What evaluation and treatment are recommended for patients with recurrent symptoms? Who is at risk for developing Barrett esophagus? For patients with Barrett esophagus, what medical treatment and surveillance are suggested? Is there a role for surgical treatment? What do you recommend for Mrs J?

DR SPECHLER: Mrs J has had heartburn, the cardinal symptom of GERD, for more than a decade. Results of a barium swallow were normal in 1993, but an endoscopic examination in 1996 showed Barrett esophagus, a condition that predisposes to esophageal cancer. GERD symptoms were not controlled initially by the H₂ receptor antagonist ranitidine, but a conventional dose of the PPI omeprazole has virtually eliminated her heartburn for the past 6 years. Nevertheless, Mrs J is concerned about her risk for developing cancer of the esophagus. This illustrative case raises a number of common, interesting, and unresolved issues in the management of patients with GERD and Barrett esophagus.

In GERD gastric juice refluxes into the esophagus and pharynx causing symptoms, tissue injury, or both.¹ Heartburn is the most common clinical manifestation. Surveys suggest that approximately 20% of US adults experience this symptom at least once a week,² and it has been estimated that Americans spend as much as $5 billion each year on medicines, both over-the-counter and prescription, to control heartburn.³ GERD can affect all age groups including infants, but the condition is diagnosed most frequently in individuals older than 40 years.⁴ Men and women are affected with approximately equal frequency, except during pregnancy when up to 50% of women experience heartburn.⁵

Heartburn can be especially common and troublesome at night, as it has been for Mrs J. A recent Gallup survey of 1000 adults who complained of heartburn at least once a week found that 79% experienced the symptom at night, and 45% claimed that nocturnal heartburn interfered with sleep.⁶ In another recent study, 945 patients with nocturnal heartburn who completed Short-Form 36 surveys reported considerable impairment in their health-related quality of life.⁷ Their scores for pain, vitality, social functioning, and mental health were comparable to those for patients with congestive heart failure and angina pectoris.

In addition to causing heartburn, acidic gastric juice in the esophagus can cause erosions and ulcerations in the squamous epithelial lining. Peptic esophageal ulcerations can bleed and can stimulate fibrous tissue deposition with stricture formation. In some patients, like Mrs J, the reflux-damaged squamous epithelium is replaced by a metaplastic, intestinal-type mucosa (a condition called Barrett esophagus) that is predisposed to malignancy (Figure 1).⁸ Consequently, GERD is a risk factor for esophageal adenocarcinoma,⁹ a tumor that has nearly quadrupled in frequency in the United States over the past 30 years.¹⁰ Patients with Barrett esophagus develop this adenocarcinoma at the rate of 0.5% per year.¹¹

GERD also can have a variety of "atypical" manifestations. Esophageal irritation by acid reflux can cause chest pain that mimics angina pectoris.¹² Refluxed gastric juice that reaches the mouth can erode dental enamel and cause oropharyngeal symptoms such as globus, sore throat, and burning tongue.¹³ Laryngitis and pulmonary problems such as chronic cough and asthma can result when refluxed material is aspirated into the airway, and acid in the esophagus can trigger reflex bronchoconstriction that also exacerbates asthma.^{14 - 15}

The frequency of GERD and its complications varies substantially among different ethnic groups. It has been known for decades that Barrett esophagus and esophageal adenocarcinoma are predominantly found in whites.¹⁶ Studies from the Far East have suggested that GERD is uncommon in Asians,¹⁷ and peptic esophageal ulcerations and strictures have been found more commonly in whites than in blacks.¹⁸ In a recent study my colleagues and I reviewed endoscopy reports and medical records for data on race and GERD complications in 2477 consecutive patients seen in a general endoscopy unit.¹⁹ One or more esophageal complications of GERD (peptic esophageal ulceration, stricture, or Barrett esophagus) were observed in 267 of 2174 white patients (12.3%), but in only 7 of 249 black patients (2.8%) and 1 of 54 Asian patients (1.8%). All 50 patients with peptic esophageal strictures were white, as were 61 of the 62 patients with peptic esophageal ulcerations. Although this study was not population based, it suggests that the esophageal complications of GERD (not just Barrett esophagus) are more common in whites and are uncommon in blacks and Asians. One important clinical implication of this observation is that clinicians should be especially cautious about attributing esophageal ulcerations and strictures to GERD in black and Asian patients. Other etiologies (eg, cancer, infection) must be considered strongly in these patients.

Gastric infection with H pylori causes chronic inflammation that can result in intestinal metaplasia and cancer in the stomach.²⁰ The organism does not infect the esophagus, however, and there is no positive association between H pylori infection and GERD. Indeed, a number of studies suggest that H pylori infection may protect the esophagus from GERD and its complications, perhaps by causing a chronic gastritis that interferes with acid production.²¹ H pylori strains that express cytotoxin-associated gene A (cagA) appear to be especially damaging to the stomach and especially protective for the esophagus.²² Some patients have developed GERD after antibiotic treatment for H pylori, and gastric acidity may be more difficult to control with antisecretory agents after eradication of the infection.²³ Mrs J had negative results for H pylori infection, as is typical for patients with Barrett esophagus. It has even been proposed that the declining frequency of H pylori infection in western countries underlies the rising frequency of esophageal adenocarcinoma, and that ethnic differences in the prevalence of the infection may contribute to ethnic differences in the frequency of GERD.²⁴ Nevertheless, available data on the relationship between H pylori and GERD are confusing and often contradictory. Large, well-designed prospective studies will be needed to resolve these issues. Currently, guidelines do not recommend that clinicians either seek or treat H pylori infection routinely in patients with GERD and its complications.²⁵

Patients who have typical heartburn that disappears with antacids or antisecretory therapy can be assumed to have GERD, and diagnostic tests are not necessary merely to confirm that diagnosis. Tests may be needed to evaluate patients who have atypical symptoms, or to look for complications of GERD like Barrett esophagus. The demonstration of esophageal inflammation (ie, reflux esophagitis) by endoscopy or barium swallow strongly supports the diagnosis of GERD. In patients who have endoscopic evidence of reflux esophagitis, however, barium swallow demonstrates esophagitis in only approximately 70%.²⁶ Lesions identified by barium usually require an endoscopic evaluation for confirmation and biopsy sampling, and endoscopy is needed to diagnose Barrett esophagus. Consequently, endoscopy is preferred over radiography for the initial evaluation of patients with GERD symptoms. This point is illustrated by Mrs J, for whom an endoscopy revealed Barrett esophagus, whereas an upper GI series was normal. Although an endoscopic examination is the most sensitive test available for establishing the diagnosis of reflux esophagitis and Barrett esophagus, the clinician should appreciate that acid reflux can cause heartburn without causing visible damage to the esophagus. Therefore, a normal endoscopic examination finding does not eliminate GERD as a cause of symptoms. Indeed, endoscopy shows reflux esophagitis in less than 50% of patients who complain of frequent, typical heartburn.^{27 - 28} Patients who have such heartburn with normal endoscopic examination results have nonerosive reflux disease, a disorder that may have a substantial functional component.²⁸ Symptoms for patients with nonerosive reflux disease may be more resistant to therapy than those for patients with erosive esophagitis.²⁸

The Practice Parameters Committee of the American College of Gastroenterology has recommended the following guideline on initial evaluation for patients with GERD²⁵ : "If the patient's history is typical for uncomplicated GERD, an initial trial of empirical therapy [including lifestyle modification (discussed below)] is appropriate. Patients in whom empirical therapy is unsuccessful or who have symptoms suggesting complicated disease should have further diagnostic testing" [ie, endoscopy]. Symptoms that might suggest complicated disease requiring early endoscopic evaluation include anorexia, weight loss, dysphagia, odynophagia, bleeding, and signs of systemic illness. In another publication dealing specifically with Barrett esophagus,²⁹ the Practice Parameters Committee recommends that "Patients with chronic GERD symptoms are those most likely to have Barrett esophagus and should undergo upper endoscopy." Although these guidelines can be considered the "standard of care," it is important to appreciate that they are merely committee recommendations whose efficacy has not been verified by clinical studies. Indeed, some authorities feel that there is insufficient evidence to support the practice of routine endoscopic screening of patients with chronic GERD symptoms.³⁰ In Mrs J, for example, it is not clear that the finding of Barrett esophagus has resulted in any health benefits.

Medical therapy of GERD focuses primarily on the control of acid reflux through the administration of antisecretory medications.^{28 ,31 - 32} The degree of gastric acid suppression required to effect healing varies directly with the severity of the reflux esophagitis.²⁸ For patients with GERD of mild to moderate severity (ie, those with no or only mild reflux esophagitis), it had been traditional to use a "step-up" therapeutic approach starting with antireflux lifestyle modifications and antacids, and adding progressively more potent antisecretory therapies for disease unresponsive to the lesser treatments.³³ Mrs J was treated initially with a step-up approach, but her symptoms' response to this therapy was not monitored adequately and, consequently, her symptoms were not controlled effectively for years. Modern data suggest that it may be more cost-effective to use a "step-down" approach that begins with PPIs, the most potent antisecretory medications, with later attempts to use lesser treatments after PPIs have effected the initial healing.^{34 - 35} This approach might have spared Mrs J several years of suffering from heartburn. Nevertheless, studies on this issue are not definitive, and the optimal approach remains disputed. However, for patients who are known to have severe, ulcerative, reflux esophagitis, it is appropriate to begin therapy immediately with PPIs rather than to proceed stepwise through trials of agents unlikely to effect healing.^{28 ,32}

A number of lifestyle modifications have been proposed to reduce gastroesophageal reflux in patients with GERD.^{28 ,31} The head of the bed should be elevated on 4" to 6" blocks to use gravity to reduce reflux into the esophagus. Overweight patients are advised to lose weight, with the rationale that obesity may contribute to reflux by increasing abdominal pressure. Patients are advised to avoid recumbency (that might promote reflux) for 2 hours after meals and to avoid bedtime snacks that can stimulate gastric acid production and transient lower esophageal sphincter relaxations that might result in nocturnal reflux. Tobacco and alcohol consumption are proscribed because their use may decrease lower esophageal sphincter pressure, and because cigarette smoking decreases salivation (that contributes to esophageal acid clearance). Fatty foods, chocolate, and carminatives (spearmint, peppermint) should be avoided because these foods can promote reflux by decreasing lower esophageal sphincter pressure and delaying gastric emptying. Drugs such as calcium channel blockers that can decrease lower esophageal sphincter pressure and delay gastric emptying also should be avoided if possible. Although these recommendations are based on sound physiological principles, few published data support their efficacy in controlling GERD.²⁸ For patients like Mrs J who are taking potent antisecretory medications with excellent results, rigid adherence to antireflux lifestyle modifications probably is not necessary. Few physicians would argue the general health benefits of avoiding obesity, cigarette smoking, and excessive alcohol consumption. But, if she were my patient, I certainly would allow Mrs J an evening glass of wine.

Antacids and alginic acid can effect the temporary relief of episodic heartburn, but few data are available on the utility of these agents for healing reflux esophagitis or for the long-term management of GERD symptoms.^{36 - 37} Sucralfate, a complex metal salt of sulfated sucrose, has been reported to heal reflux esophagitis within 12 weeks in approximately 40% of cases, but published data on the use of this agent in GERD are limited.^{32 ,38} Prokinetic agents like the dopamine antagonist metoclopramide might decrease gastroesophageal reflux by increasing lower esophageal sphincter pressure and by enhancing esophageal and gastric emptying.³⁹ Metoclopramide is currently the only prokinetic agent available for the treatment of GERD in the United States, and its utility is limited by frequent adverse effects such as agitation, restlessness, somnolence, and extrapyramidal symptoms.²⁸

The H₂ receptor antagonists inhibit acid secretion by blocking H₂ receptors on the gastric parietal cell. The 4 available preparations (cimetidine, ranitidine, famotidine, and nizatidine) are similar in safety and efficacy, and all can be purchased over-the-counter.⁴⁰ Mild adverse effects occur in less than 4% of patients treated with these agents, and serious adverse effects are very uncommon. Cimetidine and, to a lesser extent, ranitidine bind to the hepatic cytochrome P450 mixed-function oxidase system and can inhibit the elimination of drugs that are metabolized through this system, including theophylline, phenytoin, lidocaine, quinidine, and warfarin.⁴¹ Nevertheless, important drug interactions with these agents are uncommon.^{40 - 41}

When administered in conventional doses (Table 1), the H₂-blockers relieve GERD symptoms and heal esophagitis within 12 weeks in approximately half to two thirds of all patients.^{28 ,31 - 32} These agents are most useful for patients with mild to moderate GERD, while severe esophagitis does not respond reliably to conventional H₂-blocker therapy. Although high doses of these medications (up to 8 times the conventional dose) have been used effectively to treat severe esophagitis, tolerance to the antisecretory effects of the H₂-blockers develops frequently,⁴² and few data document their long-term efficacy. For these reasons, most authorities recommend PPIs rather than high-dose H₂-blocker therapy for patients with severe GERD.²⁸

The PPIs are substituted benzimidazoles that decrease gastric acid secretion profoundly by inhibiting H⁺,K⁺ATPase, the proton (acid) pump of the parietal cell. Five PPI preparations (omeprazole, esomeprazole [the S optical isomer of omeprazole], lansoprazole, pantoprazole, and rabeprazole) are used widely for the treatment of GERD. All are similar in efficacy and adverse effect profiles,³⁸ although very large studies (capable of detecting small differences between groups) have found that esomeprazole effects slightly higher GERD healing rates than omeprazole and lansoprazole used in conventional dosages.^{43 - 44} Headache and diarrhea are the most frequently reported PPI adverse effects, but the rate at which patients develop these symptoms does not differ significantly from that for placebo. Serious adverse effects (eg, anaphylaxis, hepatotoxicity) are rare.⁴⁰

Patients with mild to moderately severe reflux esophagitis who are treated with PPIs in conventional dosages achieve healing rates of 80% to 100% within 8 to 12 weeks.^{28 ,31 - 32} Very severe, ulcerative reflux esophagitis may persist despite conventional-dose PPI therapy in up to 40% of patients, however.⁴⁵ In such cases, the esophagitis usually can be healed by doubling or tripling the dose of the PPI.⁴⁶ Studies also have shown that GERD treatment with PPIs improves dysphagia and decreases the need for esophageal dilation in patients who have peptic esophageal strictures.⁴⁷

Proton pump inhibitors administered in conventional dosages profoundly decrease, but do not abolish, gastric acid secretion in most patients. Gastric pH monitoring studies have shown that approximately 70% of individuals who take a PPI twice daily experience nocturnal gastric acid breakthrough, during which gastric pH falls below 4 for more than 1 hour at night.⁴⁸ Gastroesophageal acid reflux has been documented during these nighttime periods of acid secretion.⁴⁹ In one study, 8 of 11 individuals with nocturnal acid breakthrough taking omeprazole, 20 mg twice daily, had the breakthrough abolished by adding ranitidine, 150 mg at bedtime.⁵⁰ Although such combination therapy has been proposed for patients whose nocturnal heartburn persists despite PPI therapy, the clinical efficacy of this approach has not been documented. Furthermore, one recent study found that tolerance to the antisecretory effect of the bedtime H₂-blocker developed within 1 month in almost 80% of patients.⁵¹

For most patients with severe reflux esophagitis, indefinite maintenance therapy with PPIs is required to maintain remission of GERD.^{46 ,52} A number of theoretical concerns have been raised regarding the long-term safety of the PPIs.⁴⁰ Chronic acid suppression can elevate the serum level of gastrin,⁵³ a hormone with trophic effects on the stomach and colon that conceivably could contribute to gastric and colonic carcinogenesis. Other reported effects of PPI-induced acid suppression that, in theory, could contribute to neoplasia in the stomach include gastric colonization with bacteria that can convert dietary nitrates to carcinogenic nitrosamines,⁵⁴ and the promotion of gastric atrophy in patients who are infected with H pylori.⁵⁵ It has been proposed that suppression of bactericidal acid by PPIs might increase susceptibility to enteric infections,⁵⁶ and PPI therapy has been shown to interfere with the absorption of vitamin B₁₂ from food.⁵⁷ Despite these theoretical concerns, to my knowledge there are no reports of tumors or important nutritional deficiencies clearly attributable to the use of PPIs after well over a decade of extensive clinical use.

Antireflux operations are designed to prevent gastroesophageal reflux by reducing the hiatal hernia that usually accompanies severe GERD; returning the distal esophagus into the abdomen with its positive-pressure environment; approximating the diaphragmatic crurae of the esophageal hiatus to restore its normal antireflux function; and wrapping a portion of the gastric fundus around the distal esophagus (fundoplication).^{58 - 59} The various antireflux procedures (eg, Nissen, Belsey, Toupet) differ primarily in the approach (eg, transthoracic vs transabdominal) and in the extent of fundoplication. Since 1991, antireflux surgery has been performed laparoscopically. The laparoscopic approach has become popular not because it is safer or because it produces a better functional result than the open procedure, but because of proposed advantages in the degree of postoperative discomfort, duration of hospital stay, and cosmetic outcome.⁵⁹

Two randomized controlled trials have compared modern medical and surgical therapies for GERD.^{60 - 63} From 1986 through 1988, investigators in the Department of Veterans Affairs randomized 247 patients with complicated GERD to receive either medical therapy (antacids, ranitidine, metoclopramide, and sucralfate) or antireflux surgery (open Nissen fundoplication).⁶⁰ For the 2-year duration of the study, surgery was found to be significantly more effective than medical therapy for healing the symptoms and signs of severe GERD. Recently, a follow-up investigation was conducted on the patients who participated in this study.⁶¹ After 10 to 13 years, there were no significant differences between the medical and surgical groups in the frequency of ulcerative esophagitis or in the frequency of esophageal strictures that required dilation. The 2 groups also did not differ significantly in overall physical and mental well-being scores and in overall satisfaction with antireflux therapy. However, 62% of the surgical patients reported that they were taking medications on a regular basis to treat GERD symptoms. This study suggests that there is no clear advantage of one therapy over the other for the prevention of GERD complications, and that antireflux surgery may not effect a permanent cure for GERD in many patients with severe disease.

A Nordic group conducted a randomized trial of PPI therapy and open antireflux surgery for 310 patients with erosive esophagitis.^{62 - 63} When the medical group was treated with omeprazole in a fixed dose of 20 mg/d, then antireflux surgery was found to be superior for maintaining GERD in remission for the 5-year duration of the study. When the physicians were permitted to titrate the dose of omeprazole as necessary for symptom control (up to 60 mg/d), as is done routinely in clinical practice, the medical and surgical groups did not differ significantly in remission maintenance for up to 5 years.

Mrs J's GERD symptoms have responded well to medical therapy, but she remains troubled by her diagnosis of Barrett esophagus. That diagnosis is established when endoscopic examination reveals columnar epithelium lining the distal esophagus (Figure 1), and esophageal biopsy specimens show specialized intestinal metaplasia.⁸ For decades, the condition was recognized almost exclusively in patients who had severe GERD with long segments of columnar epithelium extending at least 3 cm up the distal esophagus. In 1994, my colleagues and I reported that biopsy specimens taken from short segments of columnar-lined esophagus frequently showed specialized intestinal metaplasia, even in patients who had no symptoms or signs of GERD.⁶⁴ Since then, patients with 3 cm or more of specialized intestinal metaplasia in the esophagus have been categorized as having traditional or long-segment Barrett esophagus, whereas those with less than 3 cm of this mucosa have been considered to have short-segment Barrett esophagus.⁶⁵ It is not clear that these 2 types of Barrett esophagus have the same pathogenesis and the same risk for adenocarcinoma but, pending further studies, patients with short- and long-segment Barrett esophagus are managed similarly.⁶⁶

Endoscopic screening for Barrett esophagus has been proposed as a strategy to decrease deaths from esophageal adenocarcinoma.²⁹ In western countries, long-segment Barrett esophagus is found in 3% to 5% of patients who undergo endoscopy because of chronic GERD symptoms, whereas short-segment Barrett esophagus can be found in 10% to 15%.^{67 - 68} Risk factors for Barrett esophagus and adenocarcinoma include male sex, white race, advanced age, obesity, and long duration of GERD symptoms.^{67 ,69} Although it has been suggested that screening should focus on patients with these risk factors, any screening program that targets only patients with GERD symptoms will have limited impact on cancer mortality because 40% of patients with esophageal adenocarcinoma have no prior history of GERD.⁹ One recent study in which endoscopic examinations were performed at a Veterans Affairs hospital in 110 older patients who had no symptoms of GERD found that 27 (25%) had Barrett esophagus.⁷⁰ The failure of current screening practices to prevent cancer deaths was highlighted in a recent study that reviewed reports on patients who had esophageal resections for adenocarcinoma.⁷¹ Less than 5% of those patients were known to have had Barrett esophagus before they sought medical attention for symptoms of esophageal cancer.

GERD symptoms in patients with Barrett esophagus usually respond well to PPIs in the doses listed in Table 1, as did Mrs J's symptoms, although higher doses may be needed to effect healing in some cases.⁷² However, symptom relief is not a reliable index of the level of acid inhibition in Barrett esophagus. Esophageal pH monitoring studies in patients rendered asymptomatic by conventional-dose PPI therapy frequently reveal persistently pathological levels of acid reflux (ie, esophageal pH remains below 4 for more than 4.5% of the 24-hour monitoring period).⁷³ Indirect evidence has suggested that episodic acid reflux might contribute to carcinogenesis in Barrett esophagus.^{74 - 76} Consequently, a controversy now rages regarding the advisability of prescribing superaggressive antisecretory therapy (designed to eliminate rather than merely reduce acid reflux) for all patients with Barrett esophagus.⁷⁷ This approach involves PPIs given at least twice daily, often with an H₂-receptor blocker at night, with esophageal pH monitoring to document efficacy.⁵⁰ Currently, the evidence that such superaggressive therapy reduces cancer risk seems too weak to warrant its substantial expense and inconvenience in routine clinical practice.⁷⁸ For patients with Barrett esophagus, the American College of Gastroenterology does not encourage prescription of antireflux therapy beyond what is necessary to heal the symptoms and signs of GERD.²⁹

Although some have proposed that surgical antireflux therapy (fundoplication) might prevent deaths from cancer in Barrett esophagus better than medical therapy does,⁷⁹ little direct evidence supports that contention. A long-term follow-up study of 247 patients with complicated GERD who participated in a randomized trial of medical and surgical antireflux therapies found no significant difference between the treatment groups in the rate of esophageal cancer development.⁶¹ Furthermore, any potential cancer-preventive benefit of surgery was offset by an unexplained (but significant) decrease in survival for the surgical patients due to excess deaths from heart disease. In a large, population-based cohort study in which patients with GERD were followed for up to 32 years, the relative risk for developing esophageal adenocarcinoma (compared with the general population) among 35 274 men who received medical antireflux therapy was 6.3 (95% confidence interval [CI], 4.5-8.7), whereas the relative risk for 6406 men treated with fundoplication was 14.1 (95% CI, 8.0-22.8).⁸⁰ These data do not support the prescription of fundoplication solely as a means to prevent cancer deaths in Barrett esophagus.

Regular endoscopic surveillance to detect curable esophageal neoplasms has been recommended for patients with Barrett esophagus by several medical societies, including the American College of Gastroenterology.^{8 ,29} No study has verified that surveillance prevents deaths from esophageal cancer, however, and only indirect evidence supports the practice. For example, retrospective studies have documented that endoscopic surveillance can detect curable neoplasms in Barrett esophagus, and that cancers discovered during surveillance tend to be less advanced than those detected in patients who present with cancer symptoms such as dysphagia and weight loss.⁸¹

Surveillance seeks to identify patients with dysplasia, a constellation of histological abnormalities suggesting that one or more clones of cells have acquired genetic damage, rendering them neoplastic and predisposed to malignancy (Figure 2).⁸² Dysplasia may be the morphological expression of early neoplasia, but dysplasia is an imperfect predictor of malignancy in Barrett esophagus. It can be difficult to distinguish the histological changes of low-grade dysplasia from those of regeneration, and agreement among pathologists for the diagnosis of low-grade dysplasia in Barrett esophagus may be less than 50%.⁸³ Consequently, the report that Mrs J had "a single mildly dysplastic gland" provides little useful information; fortunately, her subsequent endoscopies have shown no evidence of dysplasia. Endoscopists rely on random biopsy sampling techniques to find dysplasia, and biopsy sampling error can be a major problem. Verified high-grade dysplasia in Barrett esophagus clearly is an ominous finding, however, with a substantial risk of progression to malignancy.⁸²

Dysplasia in Barrett esophagus can be ablated with thermal or photochemical energy delivered endoscopically (eg, using laser, multipolar electrocoagulation, argon plasma coagulation, or photodynamic therapy).⁸⁴ When patients are given potent antisecretory therapy, the ablated columnar mucosa heals with the new growth of squamous epithelium. For patients with localized neoplasms in Barrett esophagus, another option for removal is endoscopic mucosal resection in which an entire segment of mucosa is removed (down to the submucosa) using snare polypectomy techniques. Although preliminary data on endoscopic treatments for Barrett esophagus are promising,^{84 - 85} no study yet has established that these techniques improve survival or decrease the long-term risk for malignancy. Pending such studies, endoscopic ablation techniques must be considered experimental forms of therapy.

The management of Barrett esophagus that has been endorsed by the American College of Gastroenterology is as follows²⁹ :

Although not specifically recommended in the practice guidelines, clinicians can consider the use of experimental ablative therapies such as photodynamic therapy for their patients with high-grade dysplasia in Barrett esophagus, provided the therapy is administered as part of an established, approved research protocol. The use of ablative therapies outside of research protocols cannot be condoned at this time.

I recommend that Mrs J continue indefinitely the conventional-dose PPI therapy that has controlled her GERD symptoms so well for the past 6 years. She should be reassured that her risk of developing esophageal cancer is small, no more than 0.5% per year. Nevertheless, she should be advised to continue her program of regular endoscopic surveillance. Because there has been no evidence of dysplasia on several recent endoscopic examinations, she can be advised to have surveillance at an interval of every 3 years. With this program, her prognosis is excellent.

MRS J: I am wondering if hormone replacement therapy plays any role in Barrett's or development of cancer and whether I should consider that in making my decision to continue or to stop hormone replacement therapy.

DR SPECHLER: No specific studies have addressed the effects of hormone replacement therapy on Barrett esophagus. We do know that those hormones relax the lower esophageal sphincter, which is one reason why reflux disease is so common during pregnancy.⁵ But there are no data to suggest that hormone replacement therapy specifically is harmful for women with Barrett esophagus. Women, for reasons that are not clear, seem to be protected from esophageal adenocarcinoma, which is much more common in men.⁶⁷ So I would say that Barrett esophagus should not be a consideration in your decision regarding hormone replacement therapy.

A PHYSICIAN: Your recommendation to not test for H pylori in patients with GERD sounds a lot like "don't ask, don't tell." I understand where you're coming from, but many patients are tested nonetheless. What do you do with patients who have a positive test result?

DR SPECHLER: I agree with you, we do have a "don't ask, don't tell" policy on this. I don't think you should be routinely testing unless there are good reasons to test, such as if the patient has a history of peptic ulcer disease involving the stomach or duodenum, or if you suspect that there is a gastric lymphoma. If, for whatever reason, you do test for H pylori and the test result is positive, then I recommend that you treat the infection. We know that this bacterium is a strong risk factor for peptic ulcers and gastric malignancies, and I am uncomfortable ignoring a positive test result, for medicolegal reasons if nothing else. I don't think the data are strong enough to say that the treatment will exacerbate the reflux disease to the point that it becomes uncontrollable.

A PHYSICIAN: What is known about the complication rates of upper endoscopy?

DR SPECHLER: In large, retrospective, survey studies, major complications of upper gastrointestinal endoscopy were reported in approximately 1 in 500 procedures, and the mortality rate was approximately 1 in 10 000.⁸⁶ However, those surveys included the results of emergency endoscopic procedures performed on patients who sometimes had severe comorbidities such as advanced cardiac, pulmonary, and liver diseases. For elective endoscopy performed in patients who have no systemic illnesses, complications are very uncommon. In one study of 8270 consecutive, elective endoscopies performed in ambulatory outpatients, for example, there were no major complications or deaths.⁸⁷ Taken together, these data suggest that the risks of elective upper gastrointestinal endoscopy are very low, but clearly not zero.

A PHYSICIAN: The concern about cancer from the long-term use of PPIs frightens people. Has any evidence in humans shown a problem?

DR SPECHLER: No, there is no direct evidence for any cancer risk from the PPIs.⁴⁰ The PPIs have been available and used widely in the United States for 12 years, and we have no clear-cut evidence that any cancer incidence is rising as a result of these medications. However, I would not be cavalier about the use of the PPIs. We do not have decades of experience with them, and it could be very difficult to detect a small but significant increase in cancer risk caused by these medications. But I would say that the alternative therapies for GERD have their own problems, and for now the very real benefits of long-term therapy with PPIs for patients who have severe GERD appear to outweigh any theoretical risks.

A PHYSICIAN: Many patients go on to surgery because reflux therapy is not working; some patients just do not get control. What is the long-term success rate of fundoplication surgery? How safe and/or effective are the new endoscopic antireflux procedures?

DR SPECHLER: You raise 2 important issues. One is, what do you do for patients whose reflux disease doesn't respond to medicine? That used to be the most common reason for referrals for fundoplication.⁸⁸ With PPI therapy, however, reflux disease almost always responds. If it does not, very often the symptoms that disturb the patient are not due to acid reflux, but rather to some other problem, often functional in origin, that may not respond well to an operation. Unfortunately, those patients often end up having a fundoplication for functional symptoms blamed on GERD, and this group of patients does very poorly after surgery. Frankly, the best results for surgery are seen in patients who respond well to medical therapy, because both therapies are successful in stopping acid reflux, at least for the short-term. For patients who do not respond to medical therapy for GERD, it is very important to ascertain that the symptoms are due to gastroesophageal reflux before proceeding with surgery.

As for the new endoscopic procedures like endoscopic suturing and radiofrequency energy, I don't think they're ready for prime time application. They have been used primarily in patients who have the mildest forms of reflux disease, the kind that usually respond well to medical therapy; patients with severe GERD and large hiatal hernias generally are not candidates for these endoscopic treatments.^{89 - 90} We have no long-term data on the efficacy and safety of these endoscopic procedures. In my opinion, for a disease that responds so well to one pill in the morning for most people, I see no reason for doing an unproved and potentially hazardous invasive procedure.