Management of Gastroesophageal Reflux Disease that does not Respond Well to Proton Pump Inhibitors

Tiberiu Hershcovici and Ronnie Fass


Curr Opin Gastroenterol. 2010;26(4):367-378. 

In This Article

Diagnostic Tests

Management of a patient with 'refractory GERD' requires a high level of certainty about the initial diagnosis of GERD that prompted the PPI therapy. It should be determined if the diagnosis of GERD was based solely on symptoms (subjective findings) or if objective tests, such as upper endoscopy or pH testing, were utilized. A management algorithm for GERD patients who failed PPI once daily is presented in Fig. 3.

Figure 3.

Management algorithm of GERD patient who failed PPI once daily (* complete or partial)
GERD, gastroesophageal reflux disease; PPI, proton pump inhibitor. Reproduced from [8].

Various mechanisms have been shown to contribute to the failure of PPI treatment. At present, much of the research conducted in this area is focused on weakly acidic reflux and esophageal hypersensitivity.[9]

Compliance with treatment and proper PPI dosing also are important considerations. Several surveys have demonstrated that poor compliance with PPI treatment is not uncommon among patients with GERD. In a large, population-based study, it was demonstrated that the main factors influencing compliance were the presence or absence of symptoms, the severity of symptoms, and a personal preference about when to take treatment.[10] In addition to compliance, timing and frequency of dosing are critical for maximum efficacy of the medication. In a study of 100 patients with persistent GERD symptoms, only 46% dosed optimally the PPIs (i.e. took PPIs within 30 min of a meal).[11] Of those who dosed suboptimally, 39% consumed their PPI more than 60 min before meals, 30% after meals, 28% at bedtime and 4% as needed. Thus far, there is still no clear evidence that optimizing PPI dosing results in better symptom control for patients with refractory GERD.

Recently, weakly acidic and alkaline reflux has been implicated as a cause for refractory GERD-related symptoms. The mechanism by which weakly acidic reflux causes symptoms remains poorly understood. Some authorities have proposed that mechanical distension of the esophagus and/or esophageal hypersensitivity are the culprits. Thus far, there is no evidence that weakly acidic reflux is more commonly associated with increased volume of refluxate than acidic reflux. Although esophageal impedance can document an association between weakly acidic reflux and symptoms, esophageal impedance is unable to measure volume.

Several studies have demonstrated that the proximal extent of weakly acidic reflux (a possible surrogate of volume reflux) was the most important determinant of symptomatic reflux events in patients who failed PPI treatment.[12,13] However, these studies also demonstrated a considerable overlap in the proximal extent between symptomatic and asymptomatic weakly acidic reflux episodes. In addition, symptomatic reflux episodes in patients who failed PPI twice daily were often composed of both gas and liquid.[12] There are several potential explanations for the association between proximal migration of a reflux event and symptoms. These include increased sensitivity of the proximal esophagus as compared with the distal esophagus and/or 'summation effect' due to greater recruitment of sensitized pain receptors along the esophagus. Recently it was demonstrated that the transition zone between the striated and smooth muscle of the esophagus is more sensitive to mechanical stimulation than the distal esophagus, which has only smooth muscle.[14]

Diagnostic tests in refractory GERD are primarily utilized to identify residual reflux (acidic, nonacidic or bile), anatomical and histological abnormalities of the upper gastrointestinal tract and functional heartburn. The currently available diagnostic techniques for refractory heartburn patients are summarized below:

  1. Upper gastrointestinal endoscopy

  2. Esophageal biopsies for dilated intercellular spaces (DIS)

  3. Ambulatory 24-h esophageal pH monitoring/wireless pH capsule ('Bravo')

  4. Ambulatory 24-h esophageal impedance and pH monitoring

  5. Esophageal Bilitec

Upper Gastrointestinal Endoscopy

Upper endoscopy is commonly used in clinical practice to evaluate patients with GERD who have failed PPI treatment. This clinical strategy has been endorsed by the American Society of Gastrointestinal Endoscopy (ASGE).[15] The hope is to identify anatomical and histological abnormalities that can explain the treatment failure. However, the benefit of this practice has been challenged by a recent study that evaluated GERD-related endoscopic and histologic findings in patients with refractory GERD (failure to once-daily PPI therapy) vs. patients with heartburn who had not received antireflux treatment.[16] A total of 105 patients (mean age 54.7 years; 71 men, 34 women) were enrolled into the refractory GERD group and 91 (mean age 53.4 years; 68 men, 23 women) into the no-treatment group. Anatomical findings during upper endoscopy were significantly more common in the no-treatment group compared with the refractory GERD group (55.2 vs. 40.7%, P < 0.05). GERD-related findings were significantly more common in the no-treatment group compared with the refractory GERD group (erosive esophagitis: 30.8 vs. 6.7%, respectively, P < 0.05). Refractory GERD was associated with a significantly decreased odds ratio (OR) of erosive esophagitis as compared with no treatment when adjusted for age, sex, and body mass index (BMI) [adjusted OR 0.11, 95% confidence interval (CI) 0.04–0.30]. Eosinophilic esophagitis was found in only 0.9% of refractory GERD patients.

In general, the value of endoscopy in discovering GERD-related findings in patients with refractory GERD is very low. This is primarily due to the predominance of NERD and functional heartburn patients among this group of patients and the high efficacy of PPIs in healing erosive esophagitis. Studies have demonstrated that the healing rates of erosive esophagitis in patients receiving standard-dose PPI are high (ranging from 75 to 95%), and endoscopic healing of erosive esophagitis is commonly accompanied by symptom improvement.[17,18] Persistence of symptoms despite PPI treatment may reflect failure in healing esophageal mucosal injury, almost always in those with more severe grades of erosive esophagitis (C and D). In rare cases, endoscopy in heartburn patients who have failed PPI treatment may reveal a non-GERD-related cause of the symptoms (e.g. eosinophilic esophagitis, ulcers due to Zollinger-Ellison syndrome, pill-induced esophagitis, achalasia, gastroparesis and skin diseases with esophageal involvement).[19]

Esophageal Biopsies and Dilated Intercellular Spaces

Dilated intercellular spaces (DIS) in the esophageal mucosa can be identified in virtually all GERD patients.[20] The presence of DIS is associated with a reduced potential difference, diminished transepithelial resistance, and increased esophageal mucosal permeability. DIS are the result of repeated mucosal exposure to refluxed acid and digestive enzymes, although the exact mechanism of damage to the intercellular junctions remains unclear.[21] It has been proposed that DIS are pivotal for symptom generation in GERD.[22,23] DIS enable the diffusion of refluxed gastric acid into the intercellular spaces. The acid then can reach and subsequently activate chemosensitive nociceptors located in the mucosa. Information from esophageal sensory afferents is then transmitted via the dorsal horn of the spinal cord to the brain.[24] Nociceptors sensitive to changes in pH, such as transient receptor potential vanilloid subtype 1 (TRPV1) and acid-sensing ion channels (ASIC), are located on the sensory afferents of the esophageal mucosa.[25]

Other reflux components like bile acids, trypsin, and lipase have also been shown to result in DIS. In addition, experimental exposure of the esophageal mucosa to weakly acidic solutions with or without bile acids (similar to what could happen in refractory GERD patients) can lead to DIS. Interestingly, such perfusions of the distal esophagus can lead to DIS, not only in the 'exposed' mucosa, but also in the more proximal, 'nonexposed' esophageal mucosa. However, induction of DIS may not coincide with reports of heartburn.[26••] Moreover, DIS are not a specific feature of GERD and can be found in up to 30% of asymptomatic patients, as well as in patients with esophageal candidasis, food allergy, eosinophilic esophagitis, and esophageal cancer.[27–30]

A recent preliminary study suggests that NERD patients refractory to PPI demonstrate persistence of DIS.[31] In this study, 10 refractory-NERD patients on PPI twice daily underwent 24-h esophageal impedance and pH monitoring as well as upper endoscopy with biopsies. Measurements of DIS diameter in the failure group on PPIs were compared with those from 33 NERD responders off therapy and 12 asymptomatic volunteers. The mean DIS diameter in both NERD groups was significantly greater, both in the distal and proximal esophagus, than that of the asymptomatic volunteers. Importantly, the DIS diameter in the distal esophagus of the NERD responders off therapy was greater than that of the PPI failure group. No differences in DIS diameter were observed in the proximal esophagus between the two NERD groups. This study suggests that refractory NERD patients demonstrate abnormal DIS diameter throughout the esophagus, although less than that of NERD responders off any antireflux treatment.

Ambulatory 24-h Esophageal pH Monitoring

Both catheter and wireless esophageal pH monitoring allow the quantification of esophageal acid exposure and the assessment of the temporal relationship between symptoms and acid reflux events.[32,33] Esophageal pH monitoring is commonly used in the evaluation of patients with refractory GERD. In the assessment of such patients, pH monitoring can be performed off PPI to test if the initial diagnosis was correct (i.e. heartburn was due to acid reflux) or on PPI to test whether the symptoms are due to residual acid reflux. Inclusion of a symptom-reflux correlation measure such as symptom index and/or symptom association probability (SAP) helps to determine the relationship between heartburn episodes and acid reflux events, regardless if the pH test is normal or abnormal. Some authorities have suggested that the threshold for abnormal pH test should be lowered in patients undergoing pH testing on PPI therapy. Kuo and Castell[34] proposed a cut-off of 1.6% based on a study in healthy individuals treated with omeprazole 40 mg once daily. However, most gastrointestinal motility laboratories use the classic DeMeester and Johnson criteria (i.e. normal is esophageal pH <4 for <4.2% of the monitoring period) irrespective of PPI treatment.

A positive pH test on PPI suggests that patients' persistent heartburn might be related to ongoing acid reflux. If the pH test is normal on PPI treatment, but the symptom index is abnormal, then heartburn induced by physiological levels of acid exposure could be the explanation. A normal pH test on PPI with a negative symptom index suggests that the patient's heartburn is unlikely to be related to ongoing acid reflux. However, a negative pH test may result from the patient's poor tolerance of the pH probe that causes them to limit reflux-provoking activities.[35]

Residual acid reflux has been documented in GERD patients with persistent heartburn despite treatment with PPI once or twice daily. In one study, 38.6% of the GERD patients undergoing pH testing for persistent symptoms while on PPI once a day were found to have an abnormal pH test.[32] There was no correlation between a negative pH test and age, sex, or brand of PPI. In another study, 31 and 4% of the GERD patients with refractory symptoms on PPI once daily and PPI twice daily, respectively, had an abnormal pH test.[36] Recently, Karamanolis et al.[37] demonstrated that 16 and 32% of the symptomatic patients on double dose and standard dose PPI, respectively, had an abnormal pH test. Overall, positive symptom index was documented in 40% and 7–11% of the patients who remained symptomatic on PPI once or twice daily, respectively.[38–40]

Extending the usage of the wireless pH monitoring to 4 days has been suggested to provide a better comparison of esophageal acid exposure and symptom reflux association between off and on PPI treatment periods.[41] However, studies evaluating this diagnostic strategy in refractory GERD patients are still missing.

Nocturnal acid breakthrough (NAB), defined as gastric pH below 4 for at least 1 h during the night, has been observed in 75% of all individuals (GERD patients as well as healthy individuals) taking a PPI twice daily.[42] It has been proposed that NAB might be a cause of PPI-refractory GERD. However, studies have shown that NAB does not exhibit a temporal relationship with reflux-related symptoms. In one study, 71% of patients with GERD who did not respond to twice-daily PPI experienced NAB, but only 36% showed a correlation between symptoms and NAB.[43] Furthermore, no relationship between NAB and nocturnal heartburn has ever been established.

Esophageal Multichannel Intraluminal Impedance-pH Monitoring

Multichannel intraluminal impedance-pH (MII-pH) monitoring allows the detection of virtually all reflux events and the distinction between acidic, weakly acidic, and weakly alkaline reflux.[44] Studies using this technique have suggested that persistent typical as well as atypical GERD symptoms in patients taking PPIs might be due to nonacidic reflux (weakly acid or alkaline).

The first stationary, postprandial impedance-pH study in patients who failed PPI twice daily documented a shift from primarily acidic reflux at baseline before PPIs to mostly weakly acidic reflux during PPI therapy. Furthermore, regurgitation became the predominant symptom in patients who failed PPI twice daily.[45] Further studies using ambulatory MII-pH monitoring in patients who failed PPI twice daily demonstrated that acidic reflux was associated with 7–28% of the persistent GERD symptoms, whereas weakly acidic reflux preceded 30–40% of the symptoms. Between 30 and 60% of symptoms were not preceded by any reflux.[39,46,47]

A recent MII-pH study in refractory GERD patients on PPI therapy showed that up to 68% of heartburn episodes were associated with weakly acidic reflux.[13] High esophageal proximal extent was the only important factor associated with reflux perception. Furthermore, as compared with regurgitation, weakly acidic reflux episodes resulting in heartburn were more frequently pure liquid, slightly more acidic with a lower nadir pH (4.8 vs. 5.5), and were more commonly associated with preceding acid reflux episodes.

An important controversy in patients with refractory GERD is whether to conduct testing on or off PPI therapy. Zerbib et al.[47] reported the results of MII-pH monitoring in a group of patients off vs. those on twice a day PPI treatment. In patients off PPI, combined MII-pH monitoring added little to pH monitoring alone (5–10%). In patients on PPI therapy, however, adding impedance to pH monitoring improved the diagnostic yield by 15–20%, with better symptom correlation analysis than that during pH testing alone.[47]

Another recent study compared the results of MII-pH monitoring in 39 patients with refractory GERD on twice-a-day PPI therapy with those of wireless pH monitoring in the same patients off PPI therapy.[48•] On PPI therapy, all patients had normal esophageal acid exposure, and 64% had normal impedance parameters for weakly acidic and alkaline reflux, suggesting that causes other than GERD were responsible for symptoms. Of the patients with abnormal MII-pH results on therapy, 93% also had abnormal esophageal acid exposure when studied by wireless pH monitoring off therapy. Furthermore, patients with abnormal weakly acidic or alkaline reflux by MII-pH were significantly more likely to have a greater degree of abnormal esophageal acid exposure at baseline than patients with normal impedance results. The study suggests that abnormal MII-pH on twice-daily PPI predicts abnormal esophageal acid exposure at baseline, and functional heartburn is unlikely in those patients.

Another recent study compared the yield of MII-pH monitoring on twice-daily PPI vs. off-treatment PPI in refractory heartburn patients. In contrast to the aforementioned studies, the investigators found that off-treatment PPI MII-pH testing provided a higher diagnostic yield than on-treatment PPI MII-pH testing for refractory GERD.[49] The main limitations of this study are the focus on SAP and the cessation of PPI treatment for only 7 days prior to MII-pH testing.

Esophageal Bilitec

Esophageal Bilitec monitoring detects refluxed bilirubin, which is a surrogate marker for bile reflux. One should note that nonacidic reflux and bile reflux are distinct phenomena. A recent study evaluated 20 patients with refractory GERD with both Bilitec and MII-pH monitoring.[50] The authors demonstrated abnormal duodeno-gastroesophageal reflux (DGER) in nine cases, and six of those had normal values for nonacidic reflux. Also, there was no correlation between the two types of reflux.

There is evidence that esophageal exposure to bile acids can cause heartburn. For example, esophageal perfusion with bile salt solutions at nonacidic pH has been shown to provoke heartburn.[51] In addition, exposure of rabbit esophageal mucosa to weakly acidic solutions containing bile acids resulted in increased mucosal permeability and induced DIS.[52]

Tack et al.[53] suggested a possible role for DGER in a subset of patients with difficult to manage, symptomatic reflux. In a study that included 65 patients with persistent heartburn and regurgitation while on single or double-dose PPI therapy, the authors demonstrated that 64% of the patients experienced symptoms that were associated with bile reflux alone or in combination with acid. The role of adding Bilitec to conventional pH monitoring was also assessed in 347 patients (157 men, mean age 49.4 years) who underwent pH studies on PPI therapy (28% half-dose, 67% single-dose, and 5% double-dose PPI) for persistent typical (53%) or atypical (75%) GERD symptoms.[37] The addition of Bilitec increased the number of abnormal results over pH monitoring alone, from 38 to 69% on half-dose, from 27 to 69% on single-dose, and from 0 to 38% on double-dose PPI.

Most studies evaluating patients with refractory GERD tend to compare reflux patterns at baseline and during PPI treatment. However, these studies do not provide the answer to the 'burning' query if residual reflux (acidic, nonacidic or bile) is a unique PPI failure phenomenon or a general phenomenon that occurs in all patients on PPIs, irrespective of their symptomatic response. To answer this question, the reflux pattern of responders and nonresponders to the same dose of PPI should be compared. Recently, a study evaluated 24 patients who failed and 23 patients who fully responded to a PPI once a day using both 24-h esophageal Bilitec and pH monitoring during treatment.[54•] The authors demonstrated that abnormal DGER occurred in 82% of the responders vs. 67% of the nonresponders (P = n.s.). Indeed, all pH testing and Bilitec parameters were similar between the two groups. However, GERD symptoms in the PPI-failure group were more commonly associated with acid reflux than with DGER. This study demonstrated for the first time that the level of bile and acid exposure is similar among patients who fail and those who respond to treatment with a PPI once daily. In addition, in patients who failed to respond to PPI once daily, acidic reflux still appears to play an important role in symptom generation. Finally, although PPI therapy reduces both acid and bile reflux,[55,56] complete acid suppression does not guarantee elimination of DGER.[57,58]