Refractory Gastroesophageal Reflux Disease: What Next?

ABSTRACT: As many as 70% of patients with symptoms of gastroesophageal reflux disease have nonerosive disease (including functional heartburn). These patients are far less likely than those with erosive esophagitis to respond to a standard course of proton pump inhibitor (PPI) therapy. Noncompliance is the most likely cause of persistent symptoms after a course of PPI therapy. Other causes include duodenogastroesophageal reflux, delayed gastric emptying, and a lower pain threshold. Available diagnostic modalities are unlikely to provide significant clues to the exact cause. The usual approach to patients with refractory symptoms is to double the dose of the PPI. In patients with symptoms of regurgitation or a sour or bitter taste, the addition of a transient lower esophageal sphincter relaxation reducer may be useful. Pain modulators, such as tricyclic antidepressants and selective serotonin reuptake inhibitors, may also be considered. Key words: gastroesophageal reflux disease, proton pump inhibitor, nonerosive reflux disease _______________________________________________________________________________________________________________ Although proton pump inhibitors (PPIs) are highly effective, clinical failure in patients with gastroesophageal reflux disease (GERD) is seen regularly—not only in GI clinics but also in primary care offices. In fact, the prevalence of failure with PPIs has increased in proportion to the expanding indications for their use. Patients are considered to have refractory GERD if PPI therapy is unsuccessful. About 30% of GERD patients remain symptomatic while taking standard (once-daily) doses of a PPI.1,2  Most of these patients will continue to experience GERD symptoms even with higher doses of PPIs. We define PPI failure as the inability to achieve complete esophageal healing and/or satisfactory symptomatic response after a full course (ie, 2 months) of once-daily PPI therapy. This definition allows for the inclusion of patients who perceive symptoms that persist despite PPI therapy as bothersome, independent of frequency or severity. In this article, we address the possible causes of PPI failure and review possible diagnostic and therapeutic approaches. EFFECTIVENESS OF PPI THERAPY The 3 usual presentations of GERD are nonerosive reflux disease, erosive esophagitis, and Barrett esophagus. As many as 70% of patients with typical symptoms of GERD in a primary care setting have nonerosive reflux disease, the most common presentation of GERD.3 The proportion of patients with nonerosive disease who respond to a standard dose of a PPI is approximately 20% to 30% lower than that of patients with erosive esophagitis. Nonerosive reflux disease. In a systematic review of the literature, the PPI symptomatic response pooled rate was 36.7% (95% confidence interval [CI], 34.1% to 39.3%) in patients with nonerosive reflux disease and 55.5% (95% CI, 51.5% to 59.5%) in those with erosive esophagitis.4  The therapeutic gain was 27.5% in nonerosive disease and 48.9% in erosive esophagitis. In patients with nonerosive disease, the time to sustained symptom response was 2 to 3 times longer than it was in patients with erosive esophagitis. About 30% to 50% of patients with nonerosive disease demonstrate esophageal acid exposure within the physiologic range. Patients with nonerosive disease and normal pH test results are considered to have functional heartburn, defined as episodic retrosternal burning in the absence of pathologic gastroesophageal reflux, pathology- based motility disorders, or structural abnormalities.5  Patients with functional heartburn have the lowest symptom response rate to once-daily PPI therapy of all those with nonerosive disease.6 Erosive esophagitis. Studies of patients with erosive esophagitis who were treated with once-daily PPI therapy showed 88% to 96% healing rates after 8 weeks of therapy, regardless of which PPI was used or the severity of erosive esophagitis.7-10  Thus, it appears that only 4% to 12% of patients with erosive esophagitis do not respond to once-daily PPI therapy. Barrett esophagus. The prevalence of Barrett esophagus is 6% to 12% among patients who undergo endoscopy for GERD-related symptoms, and 0.25% to 3.9% among unselected patients who undergo upper endoscopy.11  Patients with Barrett esophagus have higher levels of acid exposure in the distal esophagus than do those with nonerosive disease or erosive esophagitis,12  which suggests the need for more aggressive antireflux treatment. PPI failure per se in patients with Barrett esophagus has rarely been studied. Many studies used high doses of PPIs in these patients as the initial therapeutic strategy. In one study, all patients with Barrett esophagus who received omeprazole, 20 mg/d, were maintained symptomfree.13  However, several studies using high doses of PPIs (omeprazole, 80 mg; lansoprazole, 60 mg; and omeprazole, 60 mg once daily) demonstrated complete heartburn resolution in 80% to 85% of patients.14-16 CAUSES OF PPI FAILURE Several factors may contribute to the persistence of GERD symptoms despite PPI therapy. Some are important clinically; others are relatively uncommon and have little clinical significance. Poor compliance. This is probably the most common cause of reported PPI failure. Before further evaluation, all patients in whom PPI failure is suspected should be assessed for compliance. Several factors may contribute to patients’ willingness to adhere to longterm treatment. These include adequate knowledge about the disorder and the prescribed drug, perceived severity of symptoms, the incidence and severity of adverse events, and the convenience or complexity of the regimen (Table 1).17  Older patients and those with certain personality traits may be less compliant than others.  GERD is primarily a symptomdriven disease, and many patients continue to take medications as long as they experience symptoms. When symptoms resolve, many discontinue treatment. According to a large population- based survey, only 55% of GERD patients took their PPI once daily for 4 weeks as prescribed, while 37% took their PPI on 12 days of the month or fewer.18 Even patients who take their medication once a day may not take it correctly. For maximum efficacy, PPIs should be taken about 30 minutes before a meal. This timing is supported by a study that demonstrated significantly better gastric acid control when omeprazole or lansoprazole was taken 15 minutes before breakfast than if it was taken without breakfast.19  Many patients are not aware that a PPI must be taken before a meal because they have not received proper instructions. (If the dose is doubled, one tablet is taken in the morning and the second in the evening, 30 minutes before a meal.) An important indication of poor compliance is the recurrence of GERD-related symptoms after a period of complete resolution. Unlike H2 blockers, PPIs have not been associated with the development of tolerance.  Other proposed causes of refractory GERD, such as Helicobacter pylori infection, bioavailability of the PPI, PPI resistance, rapid metabolism of a PPI, and nocturnal acid breakthrough, appear to play a very limited role in PPI failure. Duodenogastroesophageal reflux (DGER) and nonacidic reflux. These conditions, which involve the reflux of duodenal and gastric contents through the stomach and into the esophagus,20  have been implicated as possible causes of heartburn symptoms in patients who are unresponsive to PPI therapy. The introduction of the multichannel intraluminal impedance with pH sensor (to assess nonacidic reflux) and spectrophotometric probe (to assess DGER) allows the detection of reflux of gastric contents into the esophagus without a concomitant drop in pH below 4. The role of nonacidic reflux and DGER, as well as their composition, in PPI failure remains to be elucidated. There is evidence for association, but the extent of causality is unknown. The degree of overlap between nonacidic reflux and DGER is also unclear. Other pain mechanisms. The mechanisms of pain in patients with functional heartburn appear to be diverse; acid exposure within the physiologic range is the underlying cause in only a subset of these patients. A number of studies that used esophageal balloon distention or electrical stimulation have shown that patients with functional heartburn consistently have a lower perception threshold for pain than do those with other presentations of GERD.21,22  This suggests that alteration in pain perception may play an important role in generation of symptoms. Delayed gastric emptying. This is often viewed as a significant factor in the pathophysiology of GERD.23 Several studies have demonstrated that the prevalence of delayed gastric emptying in patients with GERD is approximately 40%.24,25  Some investigators have postulated that delayed gastric emptying may contribute to PPI failure in these patients.26 DIAGNOSTIC TESTING The value of upper endoscopy in patients with refractory GERD has not been widely investigated. Although this procedure is often performed regardless of whether alarm symptoms are present, the yield in patients with refractory GERD is probably very low. The presence of erosions in these patients may suggest a high grade of esophageal inflammation before treatment, poor compliance, use of medication that can damage the esophageal mucosa, or alcohol abuse. Upper endoscopy is unequivocally indicated in patients with alarm symptoms, such as dysphagia, weight loss, or anorexia (Table 2 ). Ambulatory 24-hour esophageal pH monitoring has been widely used in evaluating patients with GERD for abnormal distal esophageal acid exposure. However, this technique appears to have limited value in patients who experience failure of once-daily PPI therapy and to be noncontributory in those in whom twice-daily therapy fails.27,28 TREATMENT The Algorithm suggests a possible therapeutic approach to patients in whom once-daily PPI therapy is unsuccessful. Double dose. The most common strategy, which has become the standard of care in clinical practice, is to double the PPI dose.29  Studies suggest that most patients who experience failure with once-daily PPI will continue to be symptomatic with twice-daily PPI therapy.30  A double dose of a PPI appears to benefit patients with functional heartburn.31 However, the maximum dose of a PPI that will relieve symptoms or increase the number of responders is not yet known. Furthermore, because most GERD patients who continue to be symptomatic with twicedaily PPI have normal esophageal acid exposure,32  it is highly unlikely that increasing the PPI dose to 3 or more times daily will provide significant additional benefit. Alternative agents. Little information exists about potential therapeutic approaches to GERD patients who remain symptomatic with twicedaily PPI therapy. Other diagnoses may be considered at this point. In GERD patients with symptoms such as regurgitation or a sour or bitter taste, the addition of a transient lower esophageal sphincter relaxation (TLESR) reducer may be useful. Baclofen, a -aminobutyric acid (GABA)- B agonist, has been used as a TLESR reducer in several trials,33  but anecdotal clinical experience with the drug in patients with PPI failure has been relatively disappointing. In addition, baclofen has been associated with adverse effects, such as confusion, dizziness, light-headedness, drowsiness, weakness, and trembling. Tegaserod, a partial serotonin (5HT4 ) agonist, has some effect on TLESR.34  However, no placebo-controlled trial has shown whether the addition of tegaserod to the regimen of patients who have experienced once-daily PPI failure is effective. The role of pro-motility agents in patients who have experienced oncedaily PPI failure is unknown. However, in patients with PPI failure who demonstrate delayed gastric emptying, the addition of a pro-motility drug—such as metoclopramide or tegaserod—is an attractive option. It is unclear whether adding a pro-motility agent to the regimen of patients who have no evidence of delayed gastric emptying might be beneficial. Pain modulators. Some experts suggest adding a pain modulator to a PPI or prescribing a pain modulator alone if a patient has not obtained adequate relief of symptoms with a PPI. This approach is based on the supposition that many patients in whom PPIs are ineffective are likely to have functional heartburn. Pain modulators, such as tricyclic antidepressants and selective serotonin reuptake inhibitors, are effective in patients with noncardiac chest pain of presumed esophageal origin.35-37  These visceral analgesics are used in non–moodaltering low doses to relieve esophageal pain. No studies have yet demonstrated their value in patients with PPI failure, but they may provide a therapeutic alternative until more novel and GI-specific pain modulators are available. Bile acid binders. The role of adjunctive bile acid binders, such as cholestyramine, in PPI failure has not been elucidated. Controversy still exists as to whether this modality should be considered in GERD. Surgery. Antireflux surgery in patients with PPI failure has been discouraged because of evidence that positive response to medical therapy predicts surgical success.38,39  For patients who experience PPI failure because of symptoms that suggest volume reflux, such as regurgitation or sour or bitter taste in the mouth, surgery may be effective. However, evidence to support such intervention is still lacking. Several studies have reported that endoscopic techniques may reduce or eliminate the need for a PPI in patients who demonstrate only partial response to PPI therapy.40-42  However, the role of endoscopic therapy in patients with GERD and refractory GERD has been under close scrutiny recently because of adverse effects and evidence of improvement only on subjective parameters when compared with sham intervention. Dr Shapiro is a research fellow with the Neuroenteric Clinical Research Group, Section of Gastroenterology, at the Southern Arizona VA Health Care System in Tucson. Dr Moore is a clinical fellow in the gastroenterology section at the University of Arizona School of Medicine in Tucson. Dr Fass is associate professor of medicine at the University of Arizona School of Medicine and director of the gastrointestinal motility laboratories at the Southern Arizona VA Health Care System and the University of Arizona Health Sciences Center. REFERENCES: 1. Carlsson R, Dent J, Watts R, et al. Gastro-oesophageal reflux disease in primary care: an international study of different treatment strategies with omeprazole. International GORD Study Group. Eur J Gastroenterol Hepatol. 1998;10:119-124. 2. Inadomi JM, Jamal R, Murata GH, et al. Stepdown management of gastroesophageal reflux disease. Gastroenterology. 2001;121:1095-1100. 3. Galmiche JP, Barthelemy P, Hamelin B. Treating the symptoms of gastro-oesophageal reflux disease: a double-blind comparison of omeprazole and cisapride. Aliment Pharmacol Ther. 1997;11:765-773. 4. Dean BB, Gano AD Jr, Knight K, et al. Effectiveness of proton pump inhibitors in nonerosive reflux disease. Clin Gastroenterol Hepatol. 2004;2:656-664. 5. Clouse RE, Richter JE, Heading RC, et al. Functional esophageal disorders. In: Drossman DA, Corazziari E, Talley NJ, et al; The Rome II Multinational Working Teams, eds. Rome II, The Functional Gastrointestinal Disorders. 2nd ed. Lawrence, Kan: Allen Press, Inc; 2000:275. 6. Lind T, Havelund T, Carlsson R, et al. Heartburn without oesophagitis: efficacy of omeprazole therapy and features determining therapeutic response. Scand J Gastroenterol. 1997;32:974-979. 7. Bardhan KD, Hawkey CJ, Long RG, et al. Lansoprazole versus ranitidine for the treatment of reflux oesophagitis. UK Lansoprazole Clinical Research Group. Aliment Pharmacol Ther. 1995;9:145-151. 8. Castell DO, Richter JE, Robinson M, et al. Efficacy and safety of lansoprazole in the treatment of erosive reflux esophagitis. The Lansoprazole Group. Am J Gastroenterol. 1996;91:1749-1757. 9. Richter JE, Bochenek W. Oral pantoprazole for erosive esophagitis: a placebo-controlled randomized clinical trial. Pantoprazole US GERD Study Group. Am J Gastroenterol. 2000;95:3071-3080. 10. Sharma VK, Leontiadis GI, Howden CW. Metaanalysis of randomized controlled trials comparing standard clinical doses of omeprazole and lansoprazole in erosive oesophagitis. Aliment Pharmacol Ther. 2001;15:227-231. 11. Fass R, Sampliner RE. Barrett’s oesophagus: optimal strategies for prevention and treatment. Drugs. 2003;63:555-564. 12. Martinez SD, Malagon IB, Garewal HS, et al. Nonerosive reflux disease (NERD)—acid reflux and symptom patterns. Aliment Pharmacol Ther. 2003;17:537-545. 13. Cooper BT, Neumann CS, Cox MA, Iqbal TH. Continuous treatment with omeprazole 20 mg daily for up to 6 years in Barrett’s oesophagus. Aliment Pharmacol Ther. 1998;12:893-897. 14. Fass R, Sampliner RE, Malagon IB, et al. Failure of oesophageal acid control in candidates for Barrett’s oesophagus reversal on a very high dose of proton pump inhibitor. Aliment Pharmacol Ther. 2000;14:597-602. 15. Sampliner RE. Effect of up to 3 years of highdose lansoprazole on Barrett’s esophagus. Am J Gastroenterol. 1994;89:1844-1848. 16. Malesci A, Savarino V, Zentilin P, et al. Partial regression of Barrett’s esophagus by long-term therapy with high-dose omeprazole. Gastrointest Endosc. 1996;44:700-705. 17. Hungin AP, Rubin G, O’Flanagan H. Factors influencing compliance in long-term proton pump inhibitor therapy in general practice. Br J Gen Pract. 1999;49:463-464. 18. The Gallup Organization. Gallup Study of Consumers’ Use of Stomach Relief Products. 2000. Available at: www.gallup.com. 19. Kuo B, Castell DO. Optimal dosing of omeprazole 40 mg daily: effects on gastric and esophageal pH and serum gastrin in healthy controls. Am J Gastroenterol. 1996;91:1532-1538. 20. Vaezi MF. Duodenogastroesopahgeal reflux. In: Castell DO, Richter JE, eds. The Esophagus. 4th ed. Philadelphia: Lippincott, Williams & Wilkins; 2004:434-450. 21. Fass R, Ofman JJ. Gastroesophageal reflux disease— should we adopt a new conceptual framework? Am J Gastroenterol. 2002;97:1901-1909. 22. Fass R, Tougas G. Functional heartburn: the stimulus, the pain, and the brain. Gut. 2002;51:885-892. 23. Castell DO, Murray JA, Tutuian R, et al. The pathophysiology of gastro-oesophageal reflux disease— oesophageal manifestations. Aliment Pharmacol Ther. 2004;20:14-25. 24. Buckles DC, Sarosiek I, McMillin C, McCallum RW. Delayed gastric emptying in gastroesophageal reflux disease: reassessment with new methods and symptomatic correlations. Am J Med Sci. 2004;327:1-4. 25. McCallum RW, Berkowitz DM, Lerner E. Gastric emptying in patients with gastroesophageal reflux. Gastroenterology. 1981;80:285-291. 26. Kudara N, Chiba T, Orii S, Suzuki K. Gastric emptying of patients with persistent reflux symptoms and erosive esophagitis under PPI therapy. Neurogastroenterol Motil. 2004;16:654. Abstract 23. 27. Charbel S, Khandwala F, Vaezi MF. The role of esophageal pH monitoring in symptomatic patients on PPI therapy. Am J Gastroenterol. 2005;100:283-289. 28. Bautista JM, Wong WM, Pulliam G, et al. The value of ambulatory 24-hour esophageal pH monitoring in clinical practice in patients with persistent GERD-related symptoms on standard dose of antireflux medications. Dig Dis Sci. In press. 29. Hetzel DJ, Dent J, Reed WD, et al. Healing and relapse of severe peptic esophagitis after treatment with omeprazole. Gastroenterology. 1988;95:903-912. 30. Fass R, Murthy U, Hayden CW, et al. Omeprazole 40 mg once a day is equally effective as lansoprazole 30 mg twice a day in symptom control of patients with gastro-oesophageal reflux disease (GERD) who are resistant to conventional-dose lansoprazole therapy: a prospective, randomized, multi-centre study. Aliment Pharmacol Ther. 2000;14:1595-1603. 31. Watson RG, Tham TC, Johnston BT, Mc- Dougall NI. Double blind cross-over placebo controlled study of omeprazole in the treatment of patients with reflux symptoms and physiological levels of acid reflux—the “sensitive oesophagus.” Gut. 1997;40:587-590. 32. Vaezi MF, Richter JE, Stasney ER, et al. A randomized, double-blind, placebo-controlled study of acid suppression for the treatment of suspected laryngopharyngeal reflux. Gastroenterology. 2004; 126:A-22. Abstract 160. 33. Koek GH, Sifrim D, Lerut T, et al. Effect of the GABA(B) agonist baclofen in patients with symptomsand duodeno-gastro-oesophageal reflux refractory to proton pump inhibitors. Gut. 2003;52:1397-1402. 34. Kahrilas PJ, Quigley EM, Castell DO, Spechler SJ. The effects of tegaserod (HTF 919) on oesophageal acid exposure in gastro-oesophageal reflux disease. Aliment Pharmacol Ther. 2000;14:1503-1509.  35. Clouse RE. Psychotropic medications for the treatment of functional gastrointestinal disorders. Clin Perspect Gastroenterol. 1999;2:348-356. 36. Varia I, Logue E, O’Connor C, et al. Randomized trial of sertraline in patients with unexplained chest pain of noncardiac origin. Am Heart J. 2000; 140:367-372. 37. Prakash C, Clouse RE. Long-term outcome from tricyclic antidepressant treatment of functional chest pain. Dig Dis Sci. 1999;44:2373-2379. 38. So JB, Zeitels SM, Rattner DW. Outcomes of atypical symptoms attributed to gastroesophageal reflux treated by laparoscopic fundoplication. Surgery. 1998;124:28-32. 39. Spechler SJ. The management of patients who have “failed” antireflux surgery. Am J Gastroenterol. 2004;99:552-561. 40. Chuttani R, Sud R, Sachdev G, et al. A novel endoscopic full-thickness plicator for the treatment of GERD: a pilot study. Gastrointest Endosc. 2003;58:770-776. 41. Triadafilopoulos G, Dibaise JK, Nostrant TT, et al. Radiofrequency energy delivery to the gastroesophageal junction for the treatment of GERD. Gastrointest Endosc. 2001;53:407-415. 42. Feretis C, Benakis P, Dimopoulos C, et al. Endoscopic implantation of Plexiglas (PMMA) microspheres for the treatment of GERD. Gastrointest Endosc. 2001;53:423-426.