Circadian pattern of intragastric acidity in patients with non-erosive reflux disease (NERD)

Authors


Correspondence to: Professor V. Savarino, Dipartimento di Medicina Interna e Specialità Mediche (DIMISM) — Cattedra di Gastroenterologia, Università di Genova, Viale Benedetto XV, n. 6, 16132 Genova, Italy. E-mail: vsavarin@unige.it

Summary

Background : Most patients with gastro-oesophageal reflux disease have non-erosive reflux disease. Proton pump inhibitors are less effective than expected in these patients, but no previous study has measured their 24-h gastric pH values.

Aims : To evaluate whether there are differences in 24-h intragastric acidity between reflux patients with and without oesophagitis and controls. The influence of Helicobacter pylori on the gastric pH of reflux patients was also assessed.

Methods : Sixty-three consecutive patients with gastro-oesophageal reflux disease symptoms who agreed to undergo endoscopy and 24-h pH-metry were recruited. Twenty-five (39%) had erosive oesophagitis and 38 (61%) did not. H. pylori was diagnosed by CLO test, histology and 13C-urea breath test. Gastric pH was also measured in 30 controls without digestive symptoms.

Results : H. pylori was found in seven of the 25 (28%) patients with oesophagitis and 14 of the 38 (37%) patients with non-erosive reflux disease. Oesophageal pH-metry was abnormal in 21 of the 25 (84%) patients with oesophagitis and in 32 of the 38 (84%) patients with non-erosive reflux disease. The median gastric pH did not differ between patients with and without oesophagitis or between them and controls during the 24 h (P = 0.8) and other time intervals (P = 0.2–0.4). The gastric pH did not differ between infected and non-infected patients with oesophagitis (P = 0.2–0.4) or non-erosive reflux disease (P = 0.3–0.8).

Conclusions : The circadian pattern of intragastric acidity does not differ between patients with non-erosive reflux disease and oesophagitis. Moreover, the study confirms that H. pylori infection does not affect the gastric pH in either group of reflux patients.

Introduction

It is now clear that non-erosive reflux disease (NERD) may be present in up to 60–70% of individuals with gastro-oesophageal reflux in the general population.1–3 This means that the large body of knowledge that has been accumulated over many years on gastro-oesophageal reflux disease (GERD) has been derived almost exclusively from cases with erosive oesophagitis. If the true spectrum of GERD includes, for the most part, patients with NERD, it seems to be appropriate to re-evaluate all of these patients from a pathophysiological point of view. This is even more relevant in view of the fact that therapeutic trials in patients with NERD have demonstrated a lower efficacy of proton pump inhibitors in symptom control than that which has been reported previously for patients with erosive oesophagitis taking similar doses.4–6 As it is well known that GERD is a multifactorial disorder, and that acid represents the most aggressive factor favouring both symptoms and mucosal lesions,7, 8 it cannot be excluded that the poorer response to proton pump inhibitors in patients with NERD may be due to the presence of gastric hyperacidity.

Furthermore, recent observations suggesting that patients with GERD have more acid than non-acid reflux and their transient lower oesophageal sphincter relaxations do not differ from those of controls,9, 10 and that, after eating, highly acidic unbuffered gastric juice is present at the gastro-oesophageal squamocolumnar junction,11 demonstrate the pathogenetic role of gastric acid also in patients with NERD.

No previous study has investigated the pattern of gastric acidity in patients with NERD by means of continuous pH monitoring, which is the sole technique capable of describing the fluctuations of gastric pH over the entire circadian cycle and under nearly physiological conditions.12

In this investigation, we assessed the 24-h intragastric acidity in reflux patients with and without erosive oesophagitis and in normal subjects. Moreover, the influence of Helicobacter pylori infection on the gastric pH of these patients was evaluated.

Patients and methods

Sixty-three patients with typical and atypical symptoms of GERD, who agreed to undergo both upper gastrointestinal endoscopy and 24-h oesophagogastric pH monitoring, were consecutively recruited for this study. Typical symptoms included the presence of heartburn and/or acid regurgitation at least twice a week for the previous 4 weeks. Atypical symptoms included recurrent chest pain of more than 3 months' duration with a cardiologist's evaluation that the symptoms were not cardiac in origin, or recurrent respiratory symptoms for at least 3 weeks (chronic cough, asthma and hoarseness) with the absence of concurrent infections or allergic causes. No history of duodenal ulcer disease was present, and no concurrent gastric or duodenal ulcer was found in any patient. Reflux disease clearly related to scleroderma was also excluded. None of the patients had previously undergone any type of anti-reflux or gastric surgery. The presence of Barrett's oesophagus was shown on the basis of histological findings on four biopsy specimens taken from the distal oesophagus.

All patients underwent upper gastrointestinal endoscopy after an overnight fast, and were grouped as ‘H. pylori-positive’ or ‘H. pylori-negative’ on the basis of the concomitant results of at least two of the following three tests: CLO test (Delta West Ltd, Bentley, Australia), histology on one antral and one corpus biopsy (using haematoxylin and eosin with modified Giemsa) and 13C-urea breath test.13 Two additional biopsy specimens, one from the antrum and one from the corpus along the greater curvature, were taken to diagnose the presence or not of chronic gastritis, according to the updated Sydney classification.14

Furthermore, all patients underwent 24-h oesophagogastric pH-metry. Antisecretory drugs, if any, were discontinued at least 1 week (2 weeks for proton pump inhibitors) and prokinetic agents 72 h before testing. Oesophageal pH was measured by a glass minielectrode (Ingold 440 M3, Urdorf, Switzerland) positioned 5 cm above the upper border of the manometrically identified lower oesophageal sphincter. A second glass pH minielectrode (Ingold 440 M1.5, Urdorf, Switzerland) was positioned in the gastric corpus, about 10 cm below the cardia, for reading the gastric acidity. The electrodes were connected to a two-channel datalogger that permitted pH readings to be taken every 6 s (Proxima light 2, Synectics AB, Stockholm, Sweden). During the test day, meal times (breakfast at 08.00 h, lunch at noon and dinner at 18.00 h) and composition were standardized.15 All the experimental procedures have been reported previously.16

Thirty healthy subjects (18 males and 12 females; mean age, 53.2 ± 9.7 years), who were asymptomatic and had no history of gastroenterological or systemic diseases, represented the control group against which the basal gastric acidity of reflux patients was compared.

Data processing

Experimental data were off-loaded into a personal computer and then transferred to a calculator for further analysis. The gastric pH was expressed as the median pH obtained from the working files of 1440 data points per 24 h, and was calculated over three different time intervals: 24 h (17.00–16.59); night-time (20.00–07.59); and daytime (08.00–19.59). Data were compared by Wilcoxon test and corrected for multiple testing.17 Demographic and clinical variables were compared by chi-squared test or Student's t-test, where appropriate. A P value of less than 0.05 was considered to be statistically significant for all data analyses.

The reflux parameters were assessed according to Johnson and DeMeester.18 Only the percentage total time spent at pH < 4.0 during the whole 24-h period was evaluated in our study, because this has been shown to be the most useful discriminator between physiological and pathological reflux.19 It was considered to be abnormal at > 5.5% over the circadian period. Our normal values, derived from 115 healthy individuals (62 males and 53 females; mean age, 49 ± 16 years), are as follows: percentage total time pH < 4.0 = 1.84 ± 1.77; percentage upright time pH < 4.0 = 1.51 ± 1.52; percentage supine time pH < 4.0 = 0.32 ± 0.62; number of total episodes pH < 4.0 = 37.3 ± 35.8; number of episodes of > 5 min when pH < 4.0 = 0.92 ± 1.45; longest episode in minutes = 6.84 ± 7.89.

Results

There was no difference between the demographic and clinical variables of the reflux patients with and without oesophagitis, as reported in Table 1.

Table 1.  Demographic and clinical variables of patients with oesophagitis and non-erosive reflux disease (NERD)
 Patients with oesophagitis (n = 25)Patients with NERD (n = 38)
Age (mean ± s.d.) (years)52.3 ± 10.655.6 ± 12.2
Male/female16/923/15
Smokers (= 20 cigarettes/day)12 (48%)16 (42%)
Alcohol intake
 (0–40 g/day)23 (92%)34 (89%)
 (40–80 g/day)2 (8%)4 (11%)
Symptom duration (years)5.9 ± 7.25.5 ± 6.8
Barrett's oesophagus20
Hiatal hernia16 (64%)25 (66%)
H. pylori infection 7 (28%)14 (37%)
Abnormal oesophageal pH-metry21 (84%)32 (84%)

H. pylori infection was detected in seven of the 25 patients with oesophagitis (28%) and in 14 of the 38 patients with NERD (37%). Overall, it was present in 33% of our reflux patients. Oesophageal pH monitoring was abnormal in 53 of the 63 cases (84%), in particular in 21 of the 25 patients with oesophagitis and in 32 of the 38 patients with NERD.

The 24-h mean gastric pH profile of the controls is compared with those of the patients with NERD and oesophagitis in Figure 1. The curves pertaining to the three groups are very similar, thus confirming the almost identical gastric acidity in reflux subjects and controls.

Figure 1.

The 24-h median pH profiles of control subjects and endoscopy-positive and endoscopy-negative patients.

The individual and median gastric pH values over the three time intervals analysed are displayed in Figure 2. The gastric acidity of patients with NERD and oesophagitis does not differ significantly (P = 0.8–0.2) from that of the controls in any of the three time intervals.

Figure 2.

Individual and median pH values during three different time intervals in each population studied. GERD, gastro-oesophageal reflux disease; NERD, non-erosive reflux disease.

The influence of H. pylori infection on the gastric pH values of patients with NERD and oesophagitis is reported in Table 2. There was no significant difference between H. pylori-positive and H. pylori-negative patients with oesophagitis (P = 0.2–0.4) or NERD (P = 0.3–0.8).

Table 2.  Median (± semi-interquartiles) gastric pH values in H. pylori -positive and H. pylori -negative patients with oesophagitis or non-erosive reflux disease (NERD)
Time periodPatients with oesophagitisPatients with NERD
H. pylori -positive ( n  = 7) H. pylori -negative ( n  = 18) H. pylori -positive ( n  = 14) H. pylori -negative ( n  = 24)
17.00–16.592.4 ± 0.72.0 ± 0.32.0 ± 0.52.4 ± 0.7
20.00–07.592.0 ± 0.32.1 ± 0.41.9 ± 0.42.2 ± 0.6
08.00–19.592.2 ± 0.32.2 ± 0.62.9 ± 1.02.1 ± 0.7

Tables 3 and 4 report the degree of inflammation, activity and atrophy in both H. pylori -positive and H. pylori -negative patients with oesophagitis or NERD, respectively.

Table 3.  Pattern of inflammation, activity and atrophy in antral and corpus mucosa in oesophagitis patients with or without H. pylori infection
 CorpusAntrum
Score 0Score 1Score 2–3Score 0Score 1Score 2–3
Inflammation
H. pylori -positive  ( n  = 7) 043016
H. pylori -negative  ( n  = 18) 17101620
P  < 0.02    P  < 0.01
Activity
H. pylori -positive  ( n  = 7) 331124
H. pylori -negative  ( n  = 18) 18001800
P < 0.05
Atrophy
H. pylori -positive  ( n  = 7) 601601
H. pylori -negative  ( n  = 18) 18001710
Table 4.  Pattern of inflammation, activity and atrophy in antral and corpus mucosa in non-erosive reflux disease (NERD) patients with or without H. pylori infection
 CorpusAntrum
Score 0Score 1Score 2–3Score 0Score 1Score 2–3
Inflammation
H. pylori -positive  ( n  = 14) 194059
H. pylori -negative  ( n  = 24) 19501770
P < 0.01
Activity
H. pylori -positive  ( n  = 14) 6711112
H. pylori -negative  ( n  = 24) 22102121
P  < 0.01   P  < 0.01  
Atrophy
H. pylori -positive  ( n  = 14) 13101211
H. pylori -negative  ( n  = 24) 23102121

The degree of inflammation and activity was higher in H. pylori-positive than in H. pylori-negative patients with oesophagitis. The score of inflammation was predominantly mild (P < 0.02) in the corpus and predominantly severe (P < 0.01) in the antrum. The activity was more severe in the antrum (P < 0.05).

The degree of inflammation and activity was also higher in H. pylori-positive than in H. pylori-negative patients with NERD. The inflammation was more severe in the antrum (P < 0.01), whereas the activity was more frequently mild than severe in both the corpus (P < 0.01) and the antrum (P < 0.01).

In the whole group of patients with NERD and oesophagitis, corpus inflammation and activity were more frequently mild than severe (P < 0.01 and P < 0.001, respectively).

Finally, atrophy was detected in the gastric corpus in one patient with oesophagitis who was infected with H. pylori and in two patients with NERD (one H. pylori-positive and the other H. pylori-negative).

Discussion

Our study shows that the 24-h intragastric acidity of patients with NERD does not differ from that of patients with oesophagitis and controls. In a previous study carried out in patients with oesophagitis and Barrett's oesophagus, gastric hyperacidity was not a characteristic feature of reflux patients.20 Our results confirm this study, and add further information about the gastric pH in patients with NERD: this group does not show any increase in this parameter compared with a well-matched population of control subjects. Thus, gastric hyperacidity is not peculiar to patients with NERD, and this abnormality cannot be considered as a pathophysiological factor contributing to the development of endoscopy-negative reflux disease.

Two previous conventional secretory studies have shown the existence of gastric acid hypersecretion in patients with GERD and Barrett's oesophagus, but they examined a small group of patients in whom the concomitant presence of duodenal ulcer was not excluded.21, 22 Conversely, Hirschowitz did not find any difference in the gastric acid and pepsin output between a large series of patients with reflux oesophagitis and an appropriate group of controls, taking into account the co-existence of duodenal ulcer.23 Although the technique used to measure gastric acidity in our study, that is continuous pH monitoring, does not consider the volume of acid secretion, the information generated by this examination parallels in general that achieved with the traditional acid secretory studies using maximal stimulation with histamine or pentagastrin. This is confirmed, in particular, by the fact that studies performed with continuous gastric pH monitoring showed the same differences between healthy subjects and patients with gastric and duodenal ulcers as those reported in studies using the measurement of acid output.24, 25

The prevalence of H. pylori infection in our total group of patients with both NERD and oesophagitis is 33%, and this low rate reflects that found in many other studies carried out in reflux patients.26, 27 According to this low prevalence of H. pylori in GERD, it is very unlikely that the infection may play a significant role in the pathogenesis of reflux disease.

Our investigation also shows that the presence of H. pylori does not influence the gastric acidity in patients with NERD or oesophagitis. Other studies have already excluded any relevant effect of the infection on this physiological parameter, because the 24-h gastric pH pattern has been shown to be similar in H. pylori-positive and H. pylori-negative subjects.28, 29 In addition, several traditional secretory studies, using maximal pentagastrin stimulation, have confirmed that there is no difference in basal or peak acid output between non-ulcer H. pylori-positive and H. pylori-negative subjects.30, 31

Our study confirms that antrum and body gastritis are more frequent in H. pylori-positive than in H. pylori-negative patients with oesophagitis or NERD. However, the infection is responsible for mild rather than severe inflammation and activity in the gastric corpus of the whole group of patients with GERD. Moreover, gastric body atrophy is very rare: it was only detected in one patient with oesophagitis and two patients with NERD. Accordingly, this pattern of mild body gastritis with very few cases of atrophy cannot lead to a reduction in acid secretion, and this explains why the circadian gastric acidity of patients with oesophagitis or NERD does not differ from that of control subjects in our study. Other authors have already shown that gastric body gastritis is poorly represented in patients with GERD,32, 33 and therefore the ability to secrete acid by the oxyntic mucosa is maintained in this disease.

In our study, oesophageal pH monitoring was abnormal in 53 of the 63 cases (84%). It is well known that the pH test is not the gold standard for the diagnosis of GERD,34 and, in some studies, it has been shown to be normal in about 30–50% of patients with NERD.35, 36 This means that oesophageal acid exposure is lower in patients with NERD than in patients with oesophagitis.37 Therefore, other pathophysiological mechanisms must be involved in NERD patients with a normal pH test. In a number of these patients, there is a clear relationship between symptoms and reflux events, as shown by a high symptom index.2 Indeed, although gastric hyperacidity is not a characteristic feature of patients with NERD, it cannot be excluded that the oesophageal mucosa is highly sensitive to a rapid decrease in pH values, even though the total amount of acid refluxate over the whole 24-h period remains within normal limits. In other cases, the symptom index is negative, and therefore non-acid reflux, that is air or alkaline liquid, may be responsible for reflux symptoms, as recently demonstrated by studies using electrical impedance combined with oesophageal pH-metry.38 Indeed, acid is not the only stimulus capable of inducing heartburn, the most classical reflux symptom. Recently, it has been demonstrated that heartburn may be provoked in normal subjects by oesophageal balloon distension, in either the proximal or distal part of the oesophagus.39 This means that at least some patients with NERD, without any evidence of excess acid in their oesophagus, are highly sensitive to mechanical stimuli, such as the presence of intraluminal gas, or have an increased perception of normal reflux events.40

Finally, the remaining cases, in whom all parameters are normal and acid suppressive treatment is unsuccessful, are likely to demonstrate psychopathology and may be interpreted as a form of functional pathology which is close to dyspepsia. Through brain–gut interplay, stress seems to be a relevant factor in the generation of symptoms in patients with NERD.2

In conclusion, our study shows that the circadian gastric acidity of patients with NERD does not differ from that of patients with oesophagitis and controls. Moreover, the presence of H. pylori infection in the stomach of some patients with both forms of reflux disease does not affect this physiological parameter. Therefore, mechanisms other than acid seem to be responsible for the lower than expected response to powerful antisecretory drugs in patients with NERD.

Ancillary