Intra-oesophageal distribution and perception of acid reflux in patients with non-erosive gastro-oesophageal reflux disease


Dr M. Cicala, Department of Digestive Disease, University Campus Bio Medico, Rome, Italy.


Background: The majority of patients with gastro-oesophageal reflux disease do not present with erosive oesophagitis and make up a heterogeneous group. Patients with non-erosive gastro-oesophageal reflux disease are less responsive than patients with oesophagitis to acid-suppressive therapy.

Aim: To assess the role of acid reflux in gastro-oesophageal reflux disease symptoms.

Methods: The spatio-temporal characteristics of reflux events were analysed and related to reflux perception in 45 patients with non-erosive gastro-oesophageal reflux disease and 20 patients with erosive oesophagitis.

Results: Compared with healthy controls, all patients showed a higher intra-oesophageal proximal spread of acid, which was prominent in patients with non-erosive gastro-oesophageal reflux disease (> 50% of events lasting for 1–2 min). Irrespective of mucosal injury, the risk of reflux perception was very high when acid reached proximal sensors (odds ratio, 7.6; 95% confidence interval, 4.6–12.5), being maximal in patients with non-erosive gastro-oesophageal reflux disease with normal acid exposure time (odds ratio, 11; 95% confidence interval, 5.2–22.3).

Conclusions: Patients with non-erosive gastro-oesophageal reflux disease are characterized by a significantly higher proportion of proximal acid refluxes and a higher sensitivity to short-lasting refluxes when compared with patients with oesophagitis. The highest proximal acid exposure and highest perception occurred in patients with non-erosive gastro-oesophageal reflux disease presenting with a normal pH-metric profile. The assessment of acid distribution and its perception in the oesophageal body can better identify reflux patients who should benefit from acid-suppressive treatment.


Pathophysiological studies have provided substantial data on the mechanisms responsible for the oesophageal reflux of gastric contents in both physiological and pathological conditions.1–5 The presence of gastro-oesophageal reflux disease (GERD) is currently defined on the basis of typical symptoms and/or their impact on the health-related quality of life.6 However, the main determinants of GERD symptoms still remain to be fully elucidated. Studies using intraluminal impedance monitoring in combination with pH and bile monitoring have confirmed that, of the gastric contents, acid reflux plays a major role in eliciting symptoms and, at the same time, is related to disease complications.7–9 Patients with GERD not exhibiting oesophageal mucosal injury at endoscopy, namely patients with non-erosive gastro-oesophageal reflux disease (NERD), account for up to 70% of the total and represent a heterogeneous group, according to the 24-h pH profile and symptom–reflux association indices; indeed, it has recently been reported that, in 30–50% of symptomatic patients, results of currently available diagnostic techniques remain within the normal range.10–12 In patients with NERD, the quality of life impairment is similar to that in patients with oesophagitis,6, 13 but the symptomatic response to proton pump inhibitors is reported to be less favourable; indeed, most studies have shown a significantly reduced efficacy of standard-dose proton pump inhibitors when compared with that in patients with oesophagitis.14–17 So far, psychosocial factors, which are known to contribute to reflux perception, have not, to our knowledge, been well defined in long-term outcome studies. Despite the clinical relevance of NERD, most studies evaluating the spatio-temporal characteristics and perception of acid reflux have been limited to patients with mucosal injury and/or extra-oesophageal manifestations.18, 19

In the present study, in order to investigate whether patients with NERD with typical symptoms show any specific pathophysiological findings responsible for the reduced efficacy of proton pump inhibitors, the intra-oesophageal distribution and perception of acid reflux episodes were assessed in patients with and without abnormal acid exposure of the distal oesophagus and compared with those of patients with erosive oesophagitis. The dynamic characteristics of gastro-oesophageal reflux in patients were also compared with those in asymptomatic controls.

Materials and methods

Consecutive patients attending our out-patient unit (between February 2001 and May 2003) for recurrent typical GERD symptoms (heartburn and/or acid regurgitation) lasting for more than 6 months were invited to take part in the study. At upper gastrointestinal endoscopy, 54 patients did not present any oesophageal mucosal injury; of these, nine were excluded from the study as previous endoscopic investigations (range, 3–11 years) showed evidence of oesophageal erosive disease; the remaining 45 patients (33 females, 12 males; mean age, 41 years), 39 with negative findings at previous endoscopy and six with no history of antisecretory treatment for more than 2 weeks, defined as NERD patients, were enrolled. Twenty consecutive patients (10 females, 10 males; mean age, 45 years) presenting with reflux oesophagitis (grade A in 10, grade B in eight and grade C in two, according to the Los Angeles classification) were enrolled for comparative purposes (erosive oesophagitis patients). Oesophageal manometry and ambulatory 24-h gastro-oesophageal pH monitoring were performed consecutively in all patients, 3–14 days (mean, 5 days) after endoscopy. Manometric and pH-metric data were compared with those of 12 asymptomatic, hospital staff volunteers (eight females, four males; mean age, 34 years) (healthy control group), all of whom were non-smokers.

Oesophageal manometry was performed with a perfused Dentsleeve device assembly that incorporated a 6-cm sleeve sensor (Dentsleeve, Medtronic, Adelaide, Australia) and side-hole recording sites at 5, 10 and 15 cm above and 5 cm below the sleeve. The sleeve catheter was passed transnasally and placed within the lower oesophageal sphincter high-pressure zone. The lower oesophageal sphincter resting pressure was measured at the end of the expiratory phase.

Ambulatory 24-h oesophageal pH monitoring was performed using a double probe, each with two antimony sensors, with a separate skin reference (Zinetics Medical Inc., Salt Lake City, UT, USA). Data were stored on a single portable digital recorder (Digitrapper pH200, Medtronic, Minneapolis, MN, USA).

Before each study, the pH probe was calibrated in buffer solutions of pH 7 and pH 1. The four pH sensors were placed, according to the manometric findings, at gastric level, 5 cm above the lower oesophageal sphincter and 10 and 3 cm below the upper oesophageal sphincter.

Patients and controls completed a diary card recording the times of meals and position, time and type of symptoms. Clinical data were obtained using a standardized structured questionnaire, which was completed by all patients immediately before endoscopy, and included information on the type, duration and frequency of symptoms. Patients on antisecretory and/or prokinetic drugs stopped all treatment at least 3 weeks before endoscopy. None of the patients or controls in the study population had a history of gastrointestinal surgery, except for appendectomy. Written informed consent was obtained from all individuals and the study protocol was approved by the Ethics Committee of University Campus Bio Medico of Rome.

Analysis of the data

A reflux episode was defined as a pH decrease below 4 pH units at the distal oesophageal sensor lasting ≥ 4 s. If the pH decreased to below 4 pH units in the middle oesophagus or both the middle and proximal oesophagus, simultaneously with a similar pH decrease in the distal oesophagus, the reflux episode was defined as propagated (proximal reflux). The duration of each reflux episode was assessed at the three oesophageal sensors.

The acid exposure time was defined as pathological if the percentage of time during which pH < 4 exceeded the upper limits of normal values in the total recording time (5%) at the level of the distal oesophagus.20 The acid clearance time was calculated by dividing the total acid exposure duration in minutes by the number of reflux episodes at each site (mean duration). Reflux episodes were classified as symptom related if they occurred ≤2 min before the onset of the symptom.21 Heartburn and/or acid regurgitation were considered in the analysis of symptoms. The Symptom Index and Symptom Association Probability Index were calculated at each oesophageal level (Software Medtronic, Minneapolis, MN, USA) according to the formulae described elsewhere.22, 23

Statistical analysis

Data are expressed as the mean and 95% confidence interval (CI). Analysis of variance (anova) and chi-squared test were used to compare the characteristics of reflux events between the groups. The relationships between the acid exposure time, symptoms and proximal extent were analysed using logistic regression models. As the risk of the proximal propagation of acid reflux and of symptoms was not linearly related to the acid exposure duration, reflux events were classified according to the acid exposure duration at the distal oesophageal level. As reflux events were intra-patient 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. 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 unity.24 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, USA).


Spatio-temporal characteristics of reflux episodes in patients with NERD, patients with erosive oesophagitis and healthy controls

Of the 45 patients with NERD, 25 (55.5%) (six males, 19 females; mean age, 43 years) showed pathological acid exposure time at the distal oesophagus (mean, 15%; range, 5.1–27.2%), and were therefore defined as NERD pH-positive; 20 patients (44.5%) (six males, 14 females; mean age, 39 years) showed a normal pH-metric profile (NERD pH-negative). Of the 20 patients with erosive oesophagitis, 18 (90%) (eight males, 10 females; mean age, 42 years) showed pathological acid exposure time (mean, 20%; range, 5.7–35.5%) at the distal oesophagus, whilst two patients (both males; mean age, 35 years) showed a normal pH-metric profile.

A total of 5890 reflux events were detected at the distal oesophageal level (mean [range]: 31 [17–82] in healthy controls; 36 [1–154] in the NERD pH-negative group; 119 [24–317] in the NERD pH-positive group; 86.7 [25–176] in patients with erosive oesophagitis; P < 0.001). The spatio-temporal characteristics of the reflux events in the NERD pH-negative and pH-positive groups, patients with erosive oesophagitis and healthy controls are shown in Table 1. The patient groups showed a higher percentage of proximal reflux when compared with healthy controls. Considering overall reflux events, the percentage of proximal reflux was 10% in the healthy control group, 15% in patients with NERD and 25% in the erosive oesophagitis group (P < 0.001). Taking into consideration acid refluxes of comparable duration in the various groups at the distal oesophageal level, the highest percentage of proximal reflux was observed in the NERD pH-negative group, which was significantly higher than that in the erosive oesophagitis group (P < 0.001, Figure 1). Contrary to patients with erosive oesophagitis, the mean acid exposure in patients with NERD in the middle and proximal oesophagus was higher than that in the distal oesophagus (Tables 1 and 2). In patients with NERD, irrespective of the acid exposure time and unlike the erosive oesophagitis group, no direct relationship was observed between the duration of reflux episodes at the distal oesophageal pH sensor and their proximal extent.

Table 1.  Dynamic characteristics of total reflux events in healthy controls, patients with erosive oesophagitis and patients with non-erosive gastro-oesophageal reflux disease (NERD) (pH-negative and pH-positive)
Mean duration
min (CI)
Mean duration
min (CI)
Mean duration
min (CI)
  1. CI, confidence interval.

  2. The mean duration of reflux events was higher in patients with erosive oesophagitis and in NERD pH-positive patients vs. healthy controls (*P < 0.001). In NERD pH-negative patients, a significant difference was observed only at the proximal oesophagus.

Healthy controls4390.7 (0.4–1)1.1 (0.4–1.9)0.6 (0.2–0.9)
All NERD37201.5 (1.3–1.7)*2.9 (2.2–3.4)*2.4 (1.8–3)*
NERD pH-negative7230.6 (0.5–0.7)1.3 (0.9–1.6)1.1 (0.7–1.5)*
NERD pH-positive29971.7 (1.5–2)*3.5 (2.6–4.3)*2.8 (2–3.4)*
Erosive oesophagitis17333.3 (2.7–3.8)*4.1 (3.4–4.8)*3.2 (1.6–4.8)*
Figure 1.

Proximal acid reflux (%), according to reflux duration at the distal oesophagus, in patients with non-erosive gastro-oesophageal reflux disease (NERD) (both pH-negative and pH-positive) and erosive oesophagitis and in healthy controls. In all patients, the proximal spread of reflux was significantly greater than that in healthy controls (*P < 0.001). NERD pH-negative patients showed the highest percentage of proximal reflux (#P < 0.001 vs. patients with erosive oesophagitis).

Table 2.  Characteristics of symptom- and non-symptom-associated refluxes. The percentage of symptom-related reflux events increased from the distal to proximal oesophageal level, maximally in patients with non-erosive gastro-oesophageal reflux disease (NERD)
 NERD patientsErosive oesophagitis patients
Reflux events (n)Mean duration (min) (95% CI)Reflux events (n)Mean duration (min) (95% CI)
  1. CI, confidence interval; LOS, lower oesophageal sphincter; NSAR, non-symptom-associated reflux; SAR, symptom-associated reflux; UOS, upper oesophageal sphincter.

5 cm above LOS184 (5)3563 (95)2.1 (1.7–2.6)1.5 (1.3–1.6)86 (5)1647 (95)6.7 (3.8–9.6)3.1 (2.5–3.7)
10 cm below UOS120 (23)393 (77)2.3 (1.7–2.7)3 (2.2–3.8)59 (13)387 (87)4.6 (2.7–6.5)4.0 (3.2–4.8)
3 cm below UOS67 (37)112 (63)1.7 (1.1–2.2)2.7 (1.9–3.4)26 (13)174 (87)1.4 (0.5–2)3.5 (1.7–5.4)

Perception of acid reflux events

During pH monitoring, of the 297 symptoms reported by patients with NERD (mean, 7.3; range, 0–38), 185 (62%) were associated with an acid reflux episode [138/184 (75%) in pH-positive patients and 47/113 (42%) in pH-negative patients]. Of the 119 symptoms reported by patients with erosive oesophagitis (mean, 5.9; range, 0–39), 85 (72%) were associated with an acid reflux event (P < 0.01). No symptoms were reported by three patients with NERD and six patients with erosive oesophagitis during the pH test.

The characteristics of symptomatic and asymptomatic refluxes are shown in Table 2. In all patients with NERD and erosive oesophagitis, most of the symptomatic reflux events reached the middle oesophagus, and approximately one-third of these reached the proximal oesophagus; in contrast, the vast majority of non-symptom-related refluxes did not reach the proximal sensors. On the other hand, only 5% of distal reflux events were perceived by all patients, whereas 30% and 10% of proximal refluxes were perceived by patients with NERD and erosive oesophagitis, respectively. The logistic regression model showed that, controlling for reflux duration, the overall OR of symptoms in the proximal vs. distal events was 7.6 (95% CI, 4.6–12.5; P < 0.001). The OR of symptoms in patients with NERD was greater (OR, 11.1; 95% CI, 5.5–22.4), although not significantly, than that in patients with erosive oesphagitis (OR, 7.1; 95% CI, 4.4–11.3; Table 3). Figure 2 shows the relationship between the perception of acid reflux and its proximal extent according to the acid exposure duration at the distal oesophagus. When considering the duration of refluxes, patients with NERD were significantly more sensitive than patients with erosive oesophagitis to proximal acid reflux (P < 0.001, Figure 2). The Symptom Association Probability Index at the distal oesophagus was positive (> 95%) in 12 of 20 NERD pH-negative and in 12 of 25 NERD pH-positive patients. The Symptom Index was positive (≥ 50%) in nine NERD pH-negative and 19 NERD pH-positive patients. The proportion of positive patients was higher when the Symptom Association Probability Index was determined at both the distal and middle oesophagus than at the distal oesophagus only (70% vs. 51%, respectively). On the other hand, 30% of patients with a positive Symptom Index at the distal oesophagus were negative at the middle oesophagus.

Table 3.  Odds ratio of reflux perception in proximal refluxes with respect to distal refluxes
 Odds ratio95% confidence interval
  1. NERD, non-erosive gastro-oesophageal reflux disease.

All NERD11.15.5–22.4
NERD pH-negative10.85.2–22.3
NERD pH-positive7.21.7–30.3
Erosive oesophagitis7.14.4–11.3
All patients7.64.6–12.5
Figure 2.

Risk estimation (odds ratio) of reflux perception, according to acid exposure duration at the distal oesophagus and proximal extent, in patients with non-erosive gastro-oesophageal reflux disease (NERD) and erosive oesophagitis. In all patients, proximal reflux was perceived more than distal reflux, independent of reflux duration. Patients with NERD were significantly more sensitive than those with erosive oesophagitis to proximal reflux (*P < 0.001).


The present study focused on patients with recurrent and typical GERD symptoms in the absence of mucosal injury at endoscopy. This large group encompasses patients with excess acid exposure of the distal oesophagus as well as those with normal acid exposure, with and without a significant temporal relationship between symptoms and reflux episodes.10–12 The absence of a pathological pH test and disturbances of visceral perception mean that patients in the latter group are included in the population with functional gastrointestinal disorders (functional heartburn); these patients have been studied in terms of the psychopathological profile,25 and a less favourable response to standard-dose acid-suppressive therapy has been reported.26, 27

In the present study, efforts have been made to select true NERD patients, i.e. those in whom previous endoscopic findings and/or anti-reflux treatment should have excluded a history of erosive disease. Patients with normal pH values were included in the population of reflux patients, as the aim of the investigation was to establish the role of acid reflux in symptom generation. Spatio-temporal data concerning reflux events in our healthy subjects confirm those of others18, 19 in terms of the number, duration and proximal extent, at comparable pH recording levels, and show that about 10% of events reach the proximal oesophagus. Our results show, for the first time, that patients with NERD are characterized, in comparison with healthy controls, by a significantly higher proximal spread of acid reflux episodes, irrespective of their duration at the distal oesophagus. In NERD pH-positive and pH-negative patients, a similar trend of proximal extent was observed, as refluxes lasting for 1–2 min above the lower oesophageal sphincter were more propagated in an oral direction; in contrast, in the erosive oesophagitis group, in agreement with previous reports,18 the proximal extent of reflux was directly related to its duration above the oesophago-gastric junction (Figure 1).

A prolonged acid clearance time has frequently been reported in patients with GERD11, 28 and, in patients with pathological acid exposure time, most of whom present with erosive disease, a longer duration and a greater proximal extent of reflux have been shown when compared with healthy controls.18 In our study, the clearance time, derived from the mean acid duration at multiple oesophageal levels, was significantly delayed, as expected, in patients with erosive oesophagitis and NERD with a pathological acid exposure time at the distal oesophageal level, whereas a higher proximal extent was observed in all patients, and maximally in NERD pH-negative patients, within each class of reflux duration (Figure 1).

The relationship between symptoms and gastro-oesophageal reflux episodes is far from being fully elucidated, particularly in patients in whom endoscopy and pH monitoring prove to be normal. Data from the present study show that, in patients with NERD, over 60% (75% in NERD pH-positive patients and 42% in NERD pH-negative patients) of the typical symptoms are related to a decrease in the distal oesophageal pH to below 4 pH units. Similar results have recently been reported in another study in which, as here, NERD patients were not selected on the basis of pH monitoring or symptom–reflux association index results.29 The comparison with the erosive oesophagitis group presented here, in agreement with data in the literature,22 confirms that patients with normal endoscopy have a lower, albeit marked, symptom–reflux association. Results from NERD pH-negative patients also indicate that physiological acid exposure is sufficient to elicit symptoms, and this finding is consistent with studies showing ‘hypersensitivity’ to intra-oesophageal balloon distension,30 acid perfusion tests31 and refluxed acid32 in symptomatic patients presenting with a normal 24-h pH profile. Nevertheless, in our study, more than one-third of typical symptoms were not related to a decrease in pH below 4 pH units, and only 5% of the acid events detected at the distal oesophagus, and 30% at the proximal oesophagus, were perceived by NERD patients themselves, further confirming that other factors should be taken into consideration to explain the onset of symptoms. In the past, attempts to detect minor acid and/or non-acid refluxes in patients with erosive and non-erosive disease have failed to detect any significant differences with respect to healthy subjects;7 furthermore, no significant relationship with symptoms was found.9

The present data concerning the relationship between the spatio-temporal characteristics of reflux events and their perception show that most of the symptom-associated refluxes reached the middle oesophagus, and some 40% of these the upper oesophagus. Similar results were found in patients with non-erosive and erosive disease, thus confirming the pathophysiological and clinical significance of proximal reflux in GERD (Table 3). Weusten et al. demonstrated that, in patients with GERD presenting with pathological acid exposure time and a significant association between symptoms and reflux events, both the duration of acid exposure above the lower oesophageal sphincter and the proximal extent of the refluxate were determinants in the perception of reflux.33 In that study, the assessment of the ascending velocities of acid reflux was possible by means of ion-sensitive field-effect transistor pH electrodes, but this technique failed to reveal any difference between asymptomatic and symptomatic reflux events.33

The most surprising finding emerging from the present study is that, taking into account the duration of reflux events, patients with NERD are significantly more sensitive than those with erosive oesophagitis to short proximal refluxes (Figure 2). The highest association between proximal acid reflux and its perception was observed in NERD pH-negative patients (Table 3). Moreover, the higher proportion of patients with positive Symptom Association Probability Indices, when assessed at both the distal and middle oesophageal levels, supports the important role of the proximal spread of the refluxate in producing symptoms. Of the 45 patients with NERD, 11 would have been classified as Symptom Association Probability Index-negative patients using traditional pH-metric variables, thus underestimating the role of acid in the disease. It is generally accepted that, in GERD, mechanisms leading to mucosal injury differ from those responsible for the development of symptoms.34 Our results show that, in patients with NERD, the proximal extent of acid could be the key event in symptom production, playing a far more important role than its contact time with the mucosa; this is in keeping with studies showing hypersensitivity to acid in both the proximal and distal oesophagus in these patients,35 and could be explained by the different distribution and/or enhanced activation of chemoreceptors in the oesophageal body.36, 37 Interestingly, in a randomized, double-blind trial vs. placebo, doubling the standard dose of omeprazole resulted in a significant improvement of symptoms in patients with NERD presenting with normal pH monitoring; the best results, comparable with those obtained in patients with erosive oesophagitis, were observed in patients with a positive Symptom Index.26 This study supports the finding that a sub-group of patients with NERD has a ‘hypersensitive oesophagus’, requiring a marked acid suppression for symptom relief, and, indeed, these patients would greatly benefit from drugs aimed at modulating sensitivity to acid. Hopefully, in the not too distant future, drugs with these characteristics will be developed.

Changes in the volume of gastric refluxate and/or the pressure gradient between the stomach and oesophagus may account for the higher spread of reflux in NERD.38 A significant overlap between NERD and functional dyspepsia has been widely demonstrated,39, 40 in agreement with the common findings of visceral hypersensitivity, impaired fundic accommodation41 and delayed gastric emptying42 in patients with NERD. It is also tempting to suggest that alterations in the oesophageal motor response to reflux episodes may be responsible for the intra-oesophageal distribution of acid reflux, as suggested by the more frequent oesophageal motility disorders reported to occur in patients with NERD,42 and by the recent reports of a close temporal association between heartburn and sustained oesophageal contraction detected at high-frequency intraluminal ultrasound.43 The assessment of the motor function of the proximal stomach and of transient lower oesophageal sphincter relaxations may better explain the pathophysiological mechanisms of reflux distribution in patients with NERD.

In conclusion, independent of the acid exposure duration of the distal oesophagus, patients with NERD are characterized by a higher intra-oesophageal proximal spread of gastric acid when compared with healthy controls. Unlike patients with erosive oesophagitis, those with NERD have a higher proportion of proximal short-lived refluxes and are highly sensitive to such short reflux events. The distribution of acid proximally within the oesophagus possibly represents the main determinant of symptom production in all patients with GERD and plays a dominant role in patients with NERD and a normal pH monitoring. Assessment of the intra-oesophageal distribution of the acid refluxate and of the symptom–reflux association at the different oesophageal levels may better identify those patients presenting with normal diagnostic tests whose symptoms are due to acid perception.


The authors are grateful to Dr Stanley Lemeshow (Epidemiology and Biometrics, College of Mathematics and Applied Sciences, Columbus, OH, USA) for useful criticisms and advice in the statistical elaboration of the data and to Mrs Marian Shields for help with the English.