Article Text

Download PDFPDF

Presence of gas in the refluxate enhances reflux perception in non-erosive patients with physiological acid exposure of the oesophagus
  1. S Emerenziani1,
  2. D Sifrim2,
  3. F I Habib3,
  4. M Ribolsi1,
  5. M P L Guarino1,
  6. M Rizzi1,
  7. R Caviglia1,
  8. T Petitti1,
  9. M Cicala1
  1. 1
    Department of Digestive Diseases, University Campus Bio Medico, Rome, Italy
  2. 2
    Gasthuisberg University Hospital, Centre for Gastroenterological Research, Leuven, Belgium
  3. 3
    Department of Clinical Science, University La Sapienza, Rome, Italy
  1. Professor M Cicala, Via Longoni 83, 00155 Roma, Italy; m.cicala{at}


  • Competing interests: None.

View Full Text

Statistics from

The pathogenesis of heartburn and acid regurgitation still remains to be fully elucidated, particularly in the large group of gastro-oesophageal reflux disease patients not exhibiting oesophageal mucosal injury at endoscopy, namely the non-erosive reflux disease (NERD) patients, whose 24 h pH test often remains within the normal range.13 Indeed, several studies have demonstrated that NERD patients are less sensitive to proton pump inhibitor (PPI) treatment than patients with erosive reflux disease (ERD).46 One of the most common clinical findings in these patients is hypersensitivity to visceral stimulation, which is characterised by a reduced pain threshold to experimental stimulation.7 8 It has been suggested that dysfunction of visceral neural pathways and/or alterations in central pain processing of visceral stimuli may contribute to produce and mediate oesophageal visceral hypersensitivity in patients with symptoms but no evidence of endoscopic or pH monitoring abnormalities.9 Recent pH–impedance studies have shown that weakly acidic reflux may be associated with symptoms of oesophageal and extraoesophageal gastro-oesophageal reflux disease, particularly during PPI treatment.10 11 However, so far, no studies have focused on a comparison of the relationship between weakly acidic reflux and typical gastro-oesophageal reflux disease symptoms in NERD and in ERD patients.

Heartburn does not appear to be stimulus specific. Prolonged pH monitoring studies reveal a poor correlation between acid reflux episodes and heartburn sensation;12 it has been reported that acute acid exposure of the oesophagus enhances the oesophageal perception to intraluminal balloon distension and that oesophageal distension induces heartburn and chest pain.1315 Moreover, it has recently been demonstrated that heartburn, but not chest pain, in normal subjects is the most common response to oesophageal distension, and it seems that the distension-induced sensation is mediated through stretch mechanoreceptors.16 The composition and volume of refluxate can affect its proximal extent and the degree of oesophageal distension.17 18

Oesophageal pH–impedance monitoring can accurately identify less acidic or even non-acid reflux, with or without a gas component.19 This technique detects at least 95% of all reflux episodes and it has been demonstrated to be reproducible.20 Therefore, the aim of this study was to assess the symptom–reflux association according to acidity and composition (liquid–gas) of refluxate in NERD patients, with and without excessive oesophageal acid exposure, when compared with ERD patients.


From January 2005 to January 2006, 60 consecutive patients complaining of typical gastro-oesophageal reflux disease symptoms—heartburn and/or acid regurgitation—lasting >12 months, with a history of a favourable response to PPI treatment (>50% symptom improvement at standard or double doses) and of an early relapse after PPI discontinuation, were invited to take part in the study. All patients underwent upper endoscopy, stationary oesophageal manometry and 24 h pH–impedance monitoring, the latter performed 3–14 days (median 5 days) after endoscopy. Following upper endoscopy, five patients were excluded from the study due to duodenogastric peptic disease and/or Barrett’s oesophagus, and three patients presenting with NERD were excluded because of evidence of erosive oesophagitis at previous (3–5 years) endoscopy. Of the remaining 52 enrolled patients, 32 were NERD patients (16 male, 16 female, median age 41 years, range 31–57) and 20 patients (11 male, 9 female, median age 54 years, range 34–59) showed erosive oesophagitis (grade I, n = 12; grade II, n = 8, according to a modified Savary–Miller classification). Impedance–pH data were compared with those of 10 asymptomatic, hospital staff volunteers (6 male, 4 female, median age 34 years, range 31–47) (healthy control group), all non-smokers.

Healthy controls did not differ from NERD patients in term of mean age, but they were significantly younger than ERD patients.

None of the patients had undergone previous gastrointestinal surgery or was taking medication known to influence oesophageal motor function. Patients on antisecretory drugs stopped the treatment at least 2 weeks prior to the study (mean 5 weeks, range 2.8–8).

All NERD patients were invited to undergo a repeated pH–impedance study while on PPI treatment (esomeprazole 40 mg once a day); 12 of them agreed (5 male, 7 female, median age 42 years, range 32–50), and the study was performed following at least 4 weeks (range 4–5) of treatment.

The study protocol was approved by the Ethics Committee of University Campus Bio Medico of Rome, and written informed consent was obtained from all individuals.

Intraluminal electrical impedance and pH

Intraluminal electrical impedance was recorded with a 2.3 mm diameter polyvinyl assembly containing a series of cylindrical electrodes, each 4 mm in axial length, spaced at 2 cm intervals. Each pair of electrodes formed a measuring segment, 2 cm in length, corresponding to one recording channel (Sandhill Scientific Inc., Highlands Ranch, CO). The signals from the impedance and pH channels were digitised at 50 Hz and stored in a separate data logger (Sandhill Scientific Inc.). Oesophageal and gastric pH were measured with an antimony pH electrode. The pH electrodes were calibrated using pH 4.0 and pH 7.0 buffer solutions before beginning recording and at the end of the recording.

Patients and controls were studied after an overnight fast of at least 10 h. Prior to the ambulatory study, all subjects underwent stationary oesophageal manometry to locate the lower oesophageal sphincter (LOS). Following stationary manometry, the combined pH–impedance assembly was passed through the nose under topical anaesthesia and positioned with the pH electrodes at 5 cm above the LOS and 10 cm below the LOS. In this position, impedance was measured at 3, 5, 7, 9, 15 and 17 cm proximal to the LOS. Patients and controls were asked not to lie down during the day, but to lie down only at their usual bedtime. Furthermore, patients were instructed to consume three meals and two beverages at fixed times during the 24 h measurement period. Event markers, on the monitor, recorded meal times and posture changes.

Data analysis

In the analysis of impedance tracings, liquid reflux was defined as a retrograde 50% drop in impedance starting distally (above the LOS) and propagating at least to the next two more proximal impedance measuring segments. Gas reflux was defined as a rapid (3 kΩ/s) increase in impedance >5000 Ω, occurring simultaneously in at least 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. Pure gas reflux was defined as a rapid (3 kΩ/s) and pronounced rise in impedance that moved in a retrograde direction over at least two consecutive impedance sites. Changes in oesophageal pH during reflux detected by impedance allowed classification of reflux into: (1) acid reflux—refluxed gastric juice with a pH <4, which can either reduce the pH of the oesophagus to <4 or occur when 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 7.21

Acid exposure time (AET) was defined as pathological if the time at pH <4 exceeded 5% of total recording time. Heartburn and acid regurgitation were considered in the analysis of symptoms. Reflux episodes were classified as symptom-related if they occurred ⩽2 min before the onset of the symptom. The symptom association probability (SAP) index was calculated according to the formula described elsewhere.22

Statistical analysis

Data were expressed as mean, range or SEs, or 95% CIs when required. Analysis of variance (ANOVA) was used to compare the characteristics of reflux events between the groups and in NERD patients off and on PPI treatment. The relationships between the presence of gas in the refluxate and symptoms was analysed using logistic regression models. As reflux events were intrapatient dependent, patients were treated as clusters and the reflux episodes were considered as random samples within the clusters. Logistic regression models for surveys were used. The model was controlled for acidity of refluxate. Moreover, a sampling weight was added as (1)/(number of reflux events within a patient), so that the sum of the weights for a patient was 1.23 Odds ratio (OR) results of logistic regression were used as relative risk estimators. All statistical analyses were performed using STATA® Statistical Software (Stata Corporation, College Station, TX).


Acidity and composition of reflux in gastro-oesophageal reflux disease patients and healthy controls

Of the 32 NERD patients, 17 (8 male, 9 female, mean age 42 years) showed a pathological AET at the distal oesophagus (NERD pH-positive, mean AET 15%, range 5.1–27.2%) and 15 patients (4 male, 11 female, mean age 37 years) showed a normal pH-metric profile (NERD pH-negative). All ERD patients showed a pathological AET (mean 21%, range 6.7–35.5%). A total of 2984 reflux events were detected at the distal oesophageal level: mean 28 (range 18–43) in the healthy controls, 56 (14–104) in NERD, 41 (14–83) in NERD pH-negative, 56 (26–109) in NERD pH-positive and 81 (32–176) in ERD. The frequency of pure gas and weakly alkaline reflux was very low and, therefore, not further analysed. The frequencies of acidic and weakly acidic reflux in gastro-oesophageal reflux disease groups and healthy controls are shown in fig 1. Gastro-oesophageal reflux disease patients presented significantly higher numbers of acidic reflux episodes compared with healthy controls; weakly acidic reflux was significantly less frequent in patients than in healthy controls. Nearly two-thirds of weakly acidic reflux occurred during the post-prandial periods.

Figure 1 Frequency (mean (SE)) of acidic and weakly acidic reflux in gastro-oesophageal reflux disease subgroups and controls. *p<0.01 vs controls.

The composition of reflux in gastro-oesophageal reflux disease groups and controls is shown in fig 2. Mixed reflux accounted for 45–55% of total refluxes in all groups.

Figure 2 Frequency of mixed and liquid reflux (mean (SE)) in controls and gastro-oesophageal reflux disease patients.

Symptom–reflux analysis

Of the 32 NERD patients, 25 reported 247 symptoms during the study (mean 9, range 1–25), and of the 20 ERD patients, 13 reported 72 symptoms (mean 4, range 1–16). Heartburn accounted for 70% and regurgitation for 30% of symptoms. Heartburn accounted for 68% of pure liquid and 76% of mixed symptom-related refluxes. Belching was rarely reported (2 out of 25 patients) and was not included in the symptom analysis. SAP was positive in 16 of the 25 NERD patients (7 out of 11 NERD pH-negative patients, 9 out of 14 NERD pH-positive patients) and in 6 of the 13 ERD patients.

The frequency of symptom-related and non-symptom-related reflux, according to composition of refluxate, is shown in table 1.

Table 1 Frequency (mean (SE)) of symptom-related reflux and non-symptom-related reflux, according to composition of refluxate

The frequency of symptom-related reflux according to acidity and composition of refluxate is shown in figs 3 and 4. In NERD pH-negative patients, weakly acidic reflux accounted for 32% (10%) (vs 22% (6%) in NERD pH-positive and 12% (8%) in ERD patients, p<0.05 vs ERD) and mixed reflux accounted for more than two-thirds of all symptom-related refluxes. Multivariate logistic analysis showed that, controlling for acidity of refluxate, only in NERD pH-negative patients was the risk of reflux perception significantly higher when gas was present in the refluxate (p<0.01) (table 2).

Figure 3 Frequency of reflux-related symptoms (mean (SE)) according to the acidity of the refluxate in non-erosive reflux disease (NERD) and erosive reflux disease (ERD) patients. *p<0.05 vs ERD.
Figure 4 Frequency of reflux-related symptoms (mean (SE)) according to the composition of the refluxate in non-erosive reflux disease (NERD) and erosive reflux disease (ERD) patients. *p<0.05 vs ERD.
Table 2 Odds ratio (OR) of reflux perception in mixed refluxes with respect to liquid refluxes

Reflux pattern off and on PPI treatment

In all 12 patients who repeated the study on a PPI, the analysis of gastric pH confirmed the compliance and the effectiveness of treatment, and the AET at the distal oesophagus was within the normal range. The number of total reflux episodes decreased, non-significantly, from 492 (mean 41, range 15–108) off therapy to 428 (mean 36, range 6–95) on PPIs; the frequency of weakly acidic reflux significantly increased from 22% (95% CI: 19% to 26%) to 69% (95% CI: 64% to 73%) (p<0.001). Mixed reflux accounted for 52% of all reflux episodes (95% CI: 41% to 62%) off therapy and decreased to 40% (95% CI: 29% to 51%) on therapy (p<0.05). The decreased frequency of mixed reflux was observed in 9 out of the 12 patients and it was unchanged in the remaining 3 patients.

All patients who underwent repeat pH–impedance investigation reported typical symptoms during the first study (112 symptoms, mean 10, range 3–33). Of these, 4 patients experienced symptoms during the study while on PPI treatment (8 symptoms: 6 regurgitation and 2 heartburn episodes; mean 2, range 1–3).


Although gastric acid plays a pivotal role in the pathogenesis of gastro-oesophageal reflux disease, other stimuli are believed to be involved in the pathogenesis of typical symptoms. This is more likely in the large group of NERD patients, approximately 50% of them showing a normal oesophageal acid exposure, and/or not exhibiting a significant association between acid reflux and symptoms at ambulatory pH test, but the majority of them responding, even if to a lesser degree than ERD patients, to a PPI given at a single or double dose.

In agreement with a recent multicentre study carried out on patients reporting typical and atypical gastro-oesophageal reflux disease symptoms, with no information available concerning endoscopic findings,11 weakly acidic reflux in our gastro-oesophageal reflux disease population (NERD and ERD patients) was less frequent than in healthy controls. This same finding was first reported by Sifrim et al in a well selected gastro-oesophageal reflux disease population even if a liquid meal was chosen during the test.24 Of interest, in our series, although the large majority of symptoms were related to acid reflux, NERD patients, particularly those with a physiological acid exposure, were significantly more sensitive to weakly acidic reflux than ERD patients: weakly acidic reflux was responsible for 24% and 32% of all symptomatic episodes, respectively, in all NERD patients and in the pH-negative subgroup. Previous studies carried out on patients off therapy, complaining of symptoms suggestive of gastro-oesophageal reflux disease,11 or selected on the basis of a positive symptom–reflux correlation,25 26 showed that weakly acidic reflux accounts for a minority of typical symptoms. Similar findings would have been observed in our entire gastro-oesophageal reflux disease population if patients had not been classified according to endoscopic and pH results.

In the present study, in keeping with previous findings,23 no difference in terms of reflux composition was observed between healthy controls and patients or between disease subgroups. Moreover, we analysed, for the first time in NERD, the reflux composition in the same patients off and on PPI treatment. The results derived from a small group of patients indicated a decreased frequency of mixed reflux in most patients following PPI treatment, although the total number of reflux episodes was fairly similar. We have no explanations for this finding, and its possible role in symptom resolution remains unclear in our population of PPI responder patients.

Interestingly, in NERD patients with physiological acid exposure of the oesophagus, the presence of gas in the refluxate significantly enhanced the probability of reflux perception. At present, only two studies have focused on the relationship between reflux composition and symptoms in gastro-oesophageal reflux disease patients. Bredenoord et al, in patients complaining of typical symptoms, observed a gaseous component in the refluxate with comparable frequency in both asymptomatic and symptomatic reflux episodes.23 In the second study, from the same group, patients were enrolled if they exhibited a significant association between acid reflux and symptoms (SAP).24 Compared with controls, the authors found a higher incidence of liquid and mixed reflux only in patients with excessive acid exposure. In these two studies, patients were not classified, as here, according to endoscopic findings. In the present study, patients were consecutively enrolled on the basis of their typical symptoms and their positive, although not always complete, response to acid-suppressive treatment, reflecting the majority of cases with NERD and still classified as having gastro-oesophageal reflux disease, irrespective of their symptom–reflux association, in the new Rome III consensus report for functional oesophageal disorders.27 Indeed, a recent report, aimed at assessing the accuracy of symptom–reflux association tests in predicting PPI response, revealed their poor negative predictive values.28 In order to assess the relationship between reflux characteristics and its perception, tracings were carefully analysed for each reflux event and symptom occurrence; moreover, the probability of reflux perception (OR) was calculated by means of multivariate logistic regression.

At present, direct estimation of reflux volume is not feasible in physiological conditions. It is likely that the presence of gas in addition to liquid increases the volume of refluxate and, thus, the degree of oesophageal distension. Previous studies evaluating the effect of oesophageal acid perfusion on perception of oesophageal distension in healthy controls and patients showed that acid perfusion of the oesophagus enhances sensitivity to oesophageal balloon distension.13 29 In contrast, DeVault et al reported that a 15 min acid perfusion had no significant effect on pain perception during oesophageal distension;30 Trimble et al reported an enhanced sensation and discomfort following oesophageal distension in gastro-oesophageal reflux disease patients with normal acid exposure time compared with patients with pathological pH-metry or Barrett’s oesophagus.8 In a more recent study, the effect of acid infusion on sensitivity to oesophageal distension was related to gastro-oesophageal reflux disease severity: although short-term acid exposure of the healthy oesophagus resulted in sensitisation of both mechanoreceptors and chemoreceptors, only the chemical hypersensitivity persisted following long-term acid exposure associated with moderate tissue injury, whereas mechanosensitivity returned to normal.14 In keeping with these latter two studies, our results showed that the presence of gas in the refluxate significantly enhances the probability of reflux perception only in NERD patients with physiological acid exposure of the oesophagus, in whom weakly acidic reflux is also highly perceived.

One of the most common clinical findings in NERD patients is hypersensitivity to visceral stimulation, and results from the present study—that is, hypersensitivity to weakly acidic reflux and to mixed reflux—further support this finding.8 9 Accordingly, increasing evidence indicates an overlap between NERD and functional gastrointestinal disorders, such as functional dyspepsia and irritable bowel syndrome.31 32 Further studies, evaluating the relationship between composition and perception of reflux in patients not responding to PPIs, would better demonstrate the clinical impact of our findings.

At present, impedance–pH monitoring is the only method to detect gas and weakly acidic reflux and, therefore, seems to be useful to confirm the diagnosis in NERD patients, particularly in the pH-negative patients, who are also known to respond to a lesser degree to antireflux treatment.5

In conclusion, the large majority of symptoms, in all patients, are related to acid reflux. In NERD patients with physiological acid exposure of the oesophagus, the presence of gas in the refluxate significantly enhances the probability of reflux perception. NERD pH-negative patients are more sensitive to weakly acidic reflux than ERD patients. It can be hypothesised that these findings offer an explanation for the reduced response to PPI treatment frequently observed in these patients.


The authors are grateful to Mrs Marian Shields for help with the preparation of the manuscript.


View Abstract


  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles