The acid component of the refluxate plays a major role in the pathophysiology of gastro-oesophageal reflux disease (GORD), especially in eliciting symptoms.1 Ambulatory pH monitoring allows assessment of the temporal relationship between symptoms and reflux episodes. Although oesophageal sensitivity may vary widely, it has been shown that, in most patients, only a minority of acid reflux episodes will evoke symptoms, especially when high proximal extent occurs, nadir pH is low or the magnitude of the pH drop is high.2^–5

Although proton pump inhibitors (PPIs) have a remarkable efficacy for mucosal healing and symptom relief,6 some patients are refractory to adequate acid-suppressive therapy.7 Many factors may be responsible for unresponsiveness to PPI in patients with symptoms suggestive of GORD. When combined with pH monitoring, oesophageal multichannel intraluminal impedance allows characterisation of reflux episodes as acid, weakly acidic or weakly alkaline, as well as assessment of their proximal extent within the oesophagus.8 Recent studies performed with oesophageal pH–impedance monitoring have shown that many patients on PPIs (up to 63%) report symptoms which cannot be associated with any type of gastro-oesophageal reflux,9 10 suggesting that the hypothesis of GORD as a cause of their symptoms should reasonably be abandoned. In contrast, the same studies have shown that 30–50% of patients on PPIs do have symptoms which can—at least according to symptom association indices—be attributed to persistent gastro-oesophageal reflux, mainly weakly acidic reflux. These studies highlight the diagnostic yield of pH–impedance monitoring which can demonstrate an association between symptom and weakly acidic reflux in approximately one-third of patients on PPIs, while pH studies alone can demonstrate a pathological acid reflux in only 3–11% of patients.9^–11

To our knowledge, the determinants of reflux perception in patients on PPI therapy have never been reported. Considering the major role of weakly acidic reflux in eliciting symptoms in patients on PPIs, our aim was to assess the factors associated with the perception of gastro-oesophageal reflux episodes detected by means of ambulatory 24 h oesophageal pH–impedance monitoring.

PATIENTS AND METHODS

Patients

A total of 59 patients (41 women, mean age 47.3 years, range 18–77) were referred to two different centres (Bordeaux and Lyon University Hospitals) for pH–impedance study because of persistent typical GORD symptoms—that is, regurgitation and/or heartburn, despite adequate PPI therapy. Regurgitation was defined as the sensation of effortless return of gastric or oesophageal fluid into the mouth or throat; heartburn was defined as a burning and ascending retrosternal sensation. Only patients with either a positive symptom index (⩾50%) or a positive symptom association probability (>95%) for any type of reflux (acid and/or weakly acidic) were selected for further analysis. Thirty-nine patients were excluded because of negative symptom–reflux association (n = 35) or no symptom occurrence during the 24 h recording (n = 4). In total, 20 patients (14 women, mean age 44.6 years, range 18–77) with typical GORD symptoms were finally included in the study. Exclusion criteria were: history of thoracic, oesophageal or gastric surgery, and primary or secondary severe oesophageal motility disorders (eg, achalasia, scleroderma). Patients could have either a previously documented GORD by endoscopy and/or pH study (ie, symptoms refractory to PPI) or persistent symptoms on PPI prescribed empirically for at least 2 weeks before the pH–impedance study. All patients were prescribed a double dose of PPI—that is, twice-daily omeprazole 20 mg, lansoprazole 30 mg, rabeprazole 20 mg, pantoprazole 40 mg or esomeprazole 40 mg.

Impedance–pH equipment

Oesophageal impedance–pH monitoring was performed using a Sleuth^® Multi-channel Intraluminal Impedance ambulatory system (Sandhill Scientific, Inc., Highland Ranch, CO). The system includes a portable data logger with impedance–pH amplifiers and a catheter containing one antimony pH electrode and eight impedance electrodes at 2, 4, 6, 8, 10, 14, 16 and 18 cm from the tip of the catheter. Each pair of adjacent electrodes represents an impedance-measuring segment, 2 cm in length, corresponding to one recording channel. The impedance amplifier delivers AC voltage in a range of 1–2 kHz with resulting current flow variations in response to intraluminal impedance changes. The six impedance and pH signals were recorded at 50 Hz on a 128 MB CompactFlash card for further analysis.

Study protocol

The studies were performed on an outpatient basis after an overnight fast. Before the start of the recordings, the pH recorder was calibrated using pH 4.0 and pH 7.0 buffer solutions.

After lower oesophageal sphincter (LOS) location by oesophageal manometry, the impedance–pH catheter was passed transnasally under topical anaesthesia and positioned in the oesophageal body to record pH at 5 cm and impedance at 3, 5, 7, 9, 15 and 17 cm proximal to the LOS.

Subjects were discharged and were encouraged to maintain normal activities and sleep schedule, and eat their usual meals at their normal times. They were asked to remain upright during the day, and lie down only during their usual bedtime. Event markers on the data logger recorded symptoms, meal times and posture changes.

Data analysis

The methods of analysis have been reported in detail previously.12 Prior to the study, investigators from the two centres participated in a series of workshops aimed at training on visual characterisation of gastro-oesophageal reflux by oesophageal impedance–pH. These training sessions were organised in order to reduce interobserver variability and improve the interpretation of impedance–pH tracings.

The data stored on the CompactFlash card were downloaded onto a personal computer and visually analysed using the assistance of dedicated software (Bioview Analysis^®, version 5.0.9, Sandhill Scientific, Inc.). Analysis included identification, enumeration and characterisation of individual reflux events, and measure of clearance times (bolus and pH clearance). Meals were excluded for the analysis.

Definitions of reflux episodes

Liquid reflux was defined as a retrograde 50% drop in impedance starting distally (at the level of the LOS) and propagating to at least the next two proximal impedance measuring segments. Only liquid reflux episodes lasting at least 3 s were taken into account. Gas reflux was defined as a rapid (3 kΩ/s) increase in impedance >5000 Ω, occurring simultaneously at least in two oesophageal measuring segments, in the absence of swallowing. Mixed liquid–gas reflux was defined as gas reflux occurring immediately before or during a liquid reflux.

Gas reflux events without liquid (belches) were not considered for the purpose of this study.

Reflux episodes were characterised by pHmetry as acid, weakly acidic or weakly alkaline according to a recently published consensus report on detection and definitions of gastro-oesophageal reflux13: (1) acid reflux: refluxed gastric juice with a pH <4 which can either reduce the pH of the oesophagus to <4 or occur when the oesophageal pH is already <4; (2) weakly acidic reflux: reflux events that result in an oesophageal pH between 4 and 7; and (3) weakly alkaline reflux: reflux episodes during which nadir oesophageal pH does not drop below 7.

Gastro-oesophageal reflux variables

All reflux events were analysed in both the upright and supine position. Each individual reflux event was analysed separately and classified as symptomatic or not: reflux events were considered to be symptomatic if their onset occurred within a 2 min time window before the activation of an event marker by the patient.

For each reflux event, the following variables were determined: body position, composition (liquid or mixed liquid–gas), chemical composition (acid, weakly acidic or weakly alkaline), proximal reflux (reflux reaching the 15 cm impedance site), median reflux bolus clearance time (determined 5 cm above the LOS), bolus burden (cumulative oesophageal exposure to the bolus 1 h before the onset of the reflux episode), number of preceding (1 h) acid and weakly acidic reflux episodes, acid clearance time (defined for acid reflux events as the time during which the pH was <4), acid burden (cumulative oesophageal acid exposure 1 h before the onset of reflux episode) and nadir pH. The time window of 1 h was chosen because it was recently shown that it was adequate as far as acid reflux was concerned.3

Statistical analysis

All data were collected prospectively. Quantitative and qualitative data are given, respectively, as median (interquartile range) and as numbers (percentage).

Comparisons between symptomatic and asymptomatic reflux episodes were performed on a per subject basis using the Wilcoxon signed rank tests. Symptoms were considered globally and separately for heartburn and regurgitation. Comparison of the characteristics of reflux episodes preceding heartburn and regurgitation was performed by Mann–Whitney tests and by χ2 tests or Fisher exact tests when appropriate. To determine the proportion of patients with an abnormal number of reflux episodes, data were compared with those obtained in 20 healthy subjects studied on therapy (esomeprazole 40 mg twice daily).14 A p value <0.05 was considered statistically significant.

RESULTS

Patients

Among the 20 patients included in this study, six reported regurgitation, two reported heartburn and 12 reported both symptoms. Endoscopy was normal in 14 patients, showed hiatal hernia without oesophagitis in two, low grade oesophagitis in one and short segment Barrett’s oesophagus in three (including one with hiatal hernia). No patient had a previous history of severe oesophagitis.

The median total number of reflux episodes per patient was 45.5 (30.0–80.3). The median numbers of acid, weakly acidic and weakly alkaline reflux episodes were 2 (1.0–12.5), 33.5 (21.3–57.3) and 0.0 (0.0–1.0), respectively. Overall, 45% (9/20) of patients had a total number of reflux episodes above normal values,14 including 30% (6/20) and 15% (3/20) with abnormally high numbers of weakly acidic/weakly alkaline and acid reflux episodes, respectively.

Characteristics of reflux episodes

A total of 1273 reflux episodes were detected, including 243 (19.1%) acidic, 1018 (80.0%) weakly acidic and 12 (0.9%) weakly alkaline reflux episodes. The majority of reflux episodes occurred while patients were in the upright position (1039 vs 234). A total of 705 (59.3%) were pure liquid reflux and 703 (55.2%) reached the proximal oesophagus. A total of 312 (24.5%) refluxes were symptomatic according to our definition. The majority of symptomatic reflux episodes were weakly acidic, but the rates of symptomatic acid and weakly acidic reflux events were not significantly different (27.6% and 23.6%, respectively).

Symptomatic versus asymptomatic reflux episodes

The characteristics of the 312 symptomatic and 961 asymptomatic reflux episodes are indicated in table 1. Compared with asymptomatic reflux episodes, symptomatic reflux episodes had similar characteristics except for high proximal extent (reflux reaching the 15 cm impedance site above the LOS) which occurred more frequently for symptomatic episodes (53.3% vs 33.8%, p = 0.037) (fig 1). The positive predictive value of high proximal extent was 32.4%.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Rates of symptomatic and asymptomatic reflux events with high proximal extent in 20 patients with persisting symptoms on twice-daily PPIs. Each white circle represents the percentage of reflux events with high proximal extent in an individual patient. The filled circles and associated lines are the median values for each group.

Supine versus upright position

The characteristics of the 312 symptomatic reflux episodes according to the patient’s position are indicated in table 2. As compared with the supine position, symptomatic reflux episodes occurring in the upright position more frequently reached the proximal oesophagus (p = 0.009), and were more frequently preceded by a weakly acidic reflux episode (p = 0.003) and an acid reflux episode, although the latter did not reach the level of statistical significance (p = 0.06).

Heartburn versus regurgitation

The 312 symptomatic reflux episodes consisted of 246 episodes of regurgitation and 76 episodes of heartburn (table 3). Compared with regurgitation, the reflux episodes associated with heartburn occurred more frequently in the supine position (p<0.001), were more frequently pure liquid (p = 0.009) and acidic (p = 0.027), and had a lower nadir pH (p<0.001). In addition, reflux episodes associated with heartburn were more frequently preceded by acid reflux episodes as evidenced by a significantly higher acid burden (p<0.001), and occurrence (p<0.001) and number of acid reflux episodes (p<0.001) during the preceding hour. Finally, reflux episodes associated with heartburn had a longer reflux bolus clearance time (p<0.001) and had a significantly higher “bolus burden” (p = 0.034)—that is, a cumulative oesophageal exposure to bolus 1 h before the onset of reflux episode.

DISCUSSION

Weakly acidic reflux episodes play a major role in eliciting symptoms in 30–40% of patients on PPI therapy, as demonstrated by two recent studies with ambulatory 24 h oesophageal pH–impedance monitoring.9 10 Therefore, this technique was used to assess the determinants of reflux perception in patients with symptoms despite adequate antisecretory therapy—that is, double-dose PPIs taken twice daily. We observed that high proximal extent of the refluxate was the only factor associated with symptom perception. However, there were significant differences in the characteristics of reflux episodes when heartburn and regurgitation were considered separately.

For the purpose of this study, the patients’ selection is a crucial issue. We therefore included a relatively homogenous group of patients with persisting typical GORD symptoms and no severe mucosal damage at endoscopy since only one had mild oesophagitis and three had short segment Barrett’s oesophagus. Moreover, to be included, patients had to have a positive correlation between symptoms and reflux episodes. These are similar to the inclusion criteria recently used by Bredenoord et al who performed the same study in patients off therapy,3 thus allowing comparisons between the two studies. Similarly, in order to minimise the possible interindividual variations in oesophageal perception and/or occurrence of reflux episodes, we performed a per-patient statistical analysis—that is, for each patient, symptomatic and asymptomatic reflux episodes were compared. Finally, our patients can be reasonably considered as representative of the group of patients with persisting typical symptoms on PPIs since we observed a higher rate of reflux episodes associated with regurgitation (n = 246) than with heartburn (n = 76), although most of them had both types of symptoms. From a pathophysiological standpoint, these rates allow meaningful comparisons to be made to assess the characteristics of reflux episodes eliciting both types of symptoms. Although in clinical practice they represent the most difficult group of patients to deal with, including patients presenting with exclusively or predominantly reflux episodes associated with heartburn would have probably biased the results of this physiological study.

Overall, 24.5% of all reflux episodes were symptomatic according to our definition. Although the majority of symptomatic reflux episodes were weakly acidic, rates of both symptomatic acid (27.6%) and weakly acidic (23.3%) reflux events were relatively high. Previous studies with oesophageal pH monitoring alone showed much lower proportions of perceived acid reflux events in patients without treatment, between 5% and 14%.2 4 Most available pH studies in patients on PPIs address the issue of oesophageal exposure but not reflux–symptom association.11 15 16 Data on perception of weakly acidic reflux events are scarce. Bredenoord et al have recently reported that only 14.8% of all symptomatic reflux episodes were weakly acidic in patients off therapy, with therefore an overall proportion of symptomatic weakly acidic reflux events of 5.5%.3 To our knowledge, such data obtained in patients on PPIs have never been reported to date. These results may reflect a high sensitivity to reflux in patients on therapy since most symptomatic patients, despite PPI therapy, do not have excessive gastro-oesophageal reflux. Indeed, we recently reported that, in a cohort of 71 symptomatic patients on PPIs, only 33% had an abnormally high number of weakly acidic reflux, and 21% an increased oesophageal bolus exposure.10 In the present study, only 9/20 patients had an abnormally high number of reflux episodes. However, to demonstrate definitely increased oesophageal sensitivity in patients with persistent symptoms, patients should first have been studied off therapy, and comparisons performed according to symptomatic response to PPIs.

We observed that high proximal extent of the refluxate was the only important factor associated with reflux perception in patients on therapy, as was previously shown for patients off therapy.2^–5 The role of proximal extent has been recently pointed out by an impedance study in patients off therapy with hypersensitive oesophagus—that is, with normal acid exposure but positive symptom association probability.17 Our results further favour the hypothesis of an increased sensitivity of proximal oesophagus in the subgroup of patients with persisting symptoms on double-dose PPI therapy. There is, however, a considerable overlap in proximal extent of symptomatic and asymptomatic reflux episodes. Therefore, it is not possible to establish individual thresholds for the extent above which reflux episodes will consistently provoke symptoms.

In patients off therapy, Bredenoord et al reported that not only higher proximal extent but also a larger pH drop, lower nadir pH and longer volume and acid clearance were associated with reflux perception.3 In contrast, in patients on PPI twice daily, apart from a higher proximal extent, we observed no other significant reflux characteristic associated with symptom perception when symptoms were considered globally, thus confirming that the role of chemical composition of the refluxate may not be so relevant in patients with adequate acid-suppressive therapy. However, we found significant differences in factors associated with perception of heartburn and regurgitation analysed separately. Compared with regurgitation, we observed that reflux episodes eliciting heartburn were more frequently pure liquid and acidic, had a lower nadir pH and were more frequently associated with preceding acid reflux episodes. These results suggest that the perception of heartburn is more sensitive to the chemical composition of the refluxate than regurgitation. Together with the significantly higher “bolus burden” of the reflux episodes eliciting heartburn, these results suggest that sensitisation of oesophageal mucosa may occur by preceding acid exposure. In addition, we also observed that reflux episodes associated with heartburn had a longer reflux bolus clearance time, which reflects an increased contact time between the refluxate and the mucosa. Similar results regarding the gaseous component of the refluxate associated with regurgitation have been observed in Bredenoord’s study in patients off therapy.3 In contrast, in this study, proximal reflux events were more frequent in regurgitation than in heartburn, while there was no difference regarding chemical composition probably because in patients off therapy a large majority of reflux episodes eliciting symptoms are of acidic composition.

In conclusion, our study demonstrates the major role of proximal spread of reflux episodes as a determinant of perception in patients on PPI therapy, although, compared with regurgitation, the chemical composition of the refluxate appears to be involved in the perception of heartburn. It is therefore tempting to speculate that, in patients with “refractory GORD”, second intention antireflux therapies (eg, surgery, endoscopic procedures or drugs) should aim at not only reducing the number of reflux episodes but also limiting their proximal extent within the oesophagus.