There is no difference in the disease severity of gastro-oesophageal reflux disease between patients infected and not infected with Helicobacter pylori

Summary

Background : The role of Helicobacter pylori in gastro-oesophageal reflux disease (GERD) is controversial.

Aim : To compare the severity of GERD in infected vs. non-infected patients, as part of an ongoing randomized controlled trial that examines the impact of H. pylori eradication on GERD-related outcomes.

Methods : Consecutive GERD patients underwent urea breath testing and completed validated GERD symptom severity, and quality of life questionnaires as well as, 24-h pH-metry. These parameters, as well as demographics and endoscopic findings were assessed in double-blinded fashion and compared between H. pylori-infected and non-infected subjects.

Results : Helicobacter pylori-infected GERD patients (n = 50) were significantly older and less educated than non-infected patients (n = 51). They also used proton pump inhibitors less often but had no difference in symptoms (as measured with both the Spechler's Activity Index and the Gastrointestinal Symptom Rating Scale), quality of life, endoscopic findings or 24-h pH-metry findings.

Conclusion : This prospective, double-blind study demonstrates, using excellent GERD quantifying measures including validated symptom severity scores, endoscopy, and 24-h pH-metry, that there exist no clinically significant differences in clinical or laboratory-related GERD manifestations between H. pylori-infected and non-infected GERD patients.

Introduction

Helicobacter pylori infection is quite common, affecting approximately 30% of the population from Western countries.¹ Although the relationship of H. pylori to peptic ulcer disease, MALT lymphoma and gastric adenocarcinoma is clearly defined, its role in gastro-oesophageal reflux disease (GERD) remains controversial. GERD is also a very common condition affecting 25–40% of the population.² It has been suggested that the prevalence of H. pylori infection in GERD patients is lower than in controls.^{1, 3, 4} In addition, some studies have suggested that eradication of this organism in certain patients can lead to the development of oesophagitis or GERD-related symptoms.^{5, 6} It has been postulated that the main mechanism by which H. pylori may reduce gastric acid secretion is that of proximal gastritis, hence affecting parietal or non-parietal cell secretion.^6–8

Other studies have however failed to demonstrate an increase in GERD with H. pylori eradication.^9–12 Furthermore, some studies assessing the effect of eradication in GERD have shown improvement^{13, 14} or no change¹⁵ in GERD. Recently, a well-designed study did not demonstrate any increase in the relapse rate of moderate to severe GERD symptoms in GERD patients post-H. pylori eradication,¹⁶ although the investigators did not perform 24 h pH-metry.

A controversy, therefore, still exists on the effect of H. pylori on GERD. Comparing infected with non-infected GERD patients may help clarify this controversy. Three comparisons using 24 h pH-metry have been previously performed and have found no difference, but these studies were either small,¹⁷ included patients with atypical GERD¹⁸ or patients not clearly diagnosed with GERD¹⁹ and none included validated symptom severity and quality of life assessments.

The purpose of this study was to determine whether the severity of symptoms and acid reflux quantification in GERD patients infected with H. pylori are different than non-infected GERD patients using validated GERD and quality of life questionnaires as well as endoscopy and 24 h pH-metry. These were assessed in a double-blind fashion, and compared in 101 GERD patients.

Methods

Patient population

Consecutive patients aged 18–90 referred to participating gastroenterologists at the McGill University Health Centre between August 2000 and March 2003 with a clinical diagnosis of GERD were prospectively offered participation in a double-blind randomized controlled trial (RCT) designed to evaluate the impact of H. pylori eradication on GERD-related outcomes. The results presented relate to the baseline analyses of patient characteristics stratified according to H. pylori status assessed in a double-blind fashion. In order to be included in this study, the patients had to identify their predominant symptom as a burning feeling rising from the stomach or lower chest up towards the neck, i.e. heartburn.^{20, 21} They were required to be off acid suppression (proton pump inhibitors or H-2 receptor antagonists) for the 2 weeks prior to the administration of the questionnaires, breath test and 24 h pH-metry. Patients were excluded from the study if they had previous oesophago-gastric surgery, Barrett's oesophagus or oesophageal carcinoma. This study was approved by the Institution's Ethics Review Committee and written informed consent was obtained from all subjects.

Determination of H. pylori status

Determination of H. pylori status was performed by C-13 urea breath test (UBT) (Dianatec ISO Inc., Montreal, Canada), after a 3-h fast. This particular UBT has excellent sensitivity (99%) and specificity (97.5%) in the diagnosis of this infection^{22, 23} and has been used previously in clinical trials.²⁴ Patients and investigators were blinded to UBT results throughout the study.

Data abstraction and administration of validated questionnaires

Extensive data collection was performed including patient characteristics (gender, age, race, education, smoking, alcohol and coffee intake, weight and height), use of acid suppressive medications in the month prior to study entry, number of days in which heartburn was experienced in the week prior, presence of other gastrointestinal symptoms (regurgitation, epigastric pain, bloating, dysphagia, odynophagia) and presence of possible extra-gastrointestinal manifestations of GERD (asthma, cough, sore throat, laryngitis).

Validated questionnaires were also completed by all the subjects including the Carlsson Dent questionnaire, used to confirm the diagnosis of GERD.²⁵ The Spechler Disease Activity Index, a recognized validated method to quantify GERD severity, was also calculated for all subjects.^{26, 27} The Gastrointestinal Symptom Rating Scale (GSRS), which discriminates between different levels of GERD severity, was also administered. The GSRS has been validated in GERD patients^{28, 29} and is sensitive enough to detect changes in GERD-related outcomes attributed to different GERD therapies.^{30, 31} Quality of life was also assessed with the Psychological General Well-Being (PGWB) index, which measures well-being in six dimensions (anxiety, depressed mood, positive well-being, self-control, general health and vitality). It has been validated in GERD patients.^{29, 32}

Endoscopy, motility and 24 h pH-metry

The grade of oesophagitis was determined using the L.A. Classification.^{33, 34} If the patient had previously undergone an endoscopy, the most recent one prior to the study was used. An endoscopy was not repeated solely for the purpose of this study for ethical reasons. Motility studies were performed at study entry, as were 24 h pH-metry examinations with calculation of mean percentage of the time pH <4, number of reflux episodes, longest reflux duration and DeMeester Score.^{35, 36}

Statistical analysis

Patients were stratified into two groups based on the presence or absence of H. pylori. Continuous variables were expressed as mean ± standard deviation. Ninety-five per cent confidence intervals (95% CI) were calculated using the standard normal approximation of the binomial distribution. Between-group comparisons were performed using the Fisher exact test or the Student's t-test where appropriate. P < 0.05 was considered statistically significant. A similar analysis was carried out only on subjects with scores of ≥4 on the Carlsson Dent questionnaire, as a score of 4 or above on this questionnaire is felt to be confirmatory for a diagnosis of GERD.²⁵

Recruitment for this study was ended when a sufficient number of patients in each group were obtained based on an a priori determined power calculation. A pilot study estimated a total time with pH <4 of 8% in H. pylori-infected GERD patients and 17% in those not infected as well as a DeMeester score of 27 and 63 in those infected and non-infected respectively.³⁷ For total time with pH <4, assuming a range for the difference of 4–25% with a standard deviation of 6% and a minimal difference of 5% deemed clinically relevant, 44 subjects per group were required with an accuracy of 2.5%. For the DeMeester scores (range of 0–100, s.d. = 25, minimal difference of 25 deemed clinically relevant), 48 subjects were required per group with accuracy of 20. Hence, when 50 H. pylori non-infected patients were recruited, only H. pylori-infected patients were subsequently enrolled until 50 subjects were present in each group.

Results

Patient population

A total of 101 patients were included in this study. The H. pylori-negative patients were much easier to identify, taking just over  years to recruit 50 (only half of the H. pylori-infected GERD patients were enrolled at this time) whereas an additional 15 months were required to enrol the remaining H. pylori-infected patients.

All 101 subjects completed demographic, medication used, Carlsson Dent, Spechler Disease Activity, GSRS and PGWB questionnaires. Endoscopy results were available in 84 (83%). Motility studies and 24 h pH-metry examinations were successful in 92 (91%) and 87 (86%) respectively. Reasons for unsuccessful studies were patient refusal and withdrawal (six), unsuccessful intubation (five), patient intolerance to the pH catheter (two) and a kinked 24 h pH-metry catheter (one).

The mean (±s.d.) age of the patients was 52.8 ± 13.3 years and ranged from 21 to 80 years, with 49% being male. The majority (83%) were Caucasian and 48% had a post-secondary education. Twenty per cent were smokers, 69% coffee drinkers, and their mean weight was 76 ± 14 kg with an average height of 164 ± 15 cm. Seventy per cent of the patients used proton pump inhibitors, 15% used H2 antagonists, 34% used antacids and 3% used motility agents (not exclusive categories). Other symptoms included regurgitation in 78%, sore throat in 9% and laryngitis (reduced voice) in 4%. A Carlsson Dent score of ≥4 (consistent with GERD) was found in 91% (92/101). Endoscopic oesophagitis was present in 27% (23/84) with 61% of these having Grade A (LA Classification) oesophagitis. A hiatal hernia was present in 31%. The mean Spechler activity index was 133 ± 31 which represents a relatively severe rating of GERD.¹⁶ The GSRS and PGWB mean scores were 44.7 ± 15.2 and 91.2 ± 20.7 respectively. The mean score of the Reflux domain of the GSRS was 4.2 ± 1.6. These scores are reflective of slightly more severe GERD in our patient population than previously reported in other studies using these questionnaires in patients with upper gastrointestinal diseases.^{28, 29} Oesophageal motility demonstrated normal peristalsis and lower oesophageal sphincter (LES) relaxation in all cases with a mean LES pressure of 15 ± 10 mmHg. The pH-metry documented a ≥4% total time pH <4 in 64% (56/87) of subjects. The mean percentage of time pH <4 was 10.21 ± 14.87% with 71 ± 68 reflux episodes, 5 ± 11 reflux episodes greater than 5 min and the mean longest reflux episode duration was 25.5 ± 40.6 min. The Demeester score was 31.7 ± 37.3.

Differences in patient characteristics

Patient characteristics stratified according to H. pylori status are shown in Table 1. Helicobacter pylori-infected GERD patients were significantly older and less educated (less likely to have achieved a post-secondary education level) than non-infected patients. There was a non-significant trend suggesting that the infected patients were more likely to be non-Caucasian and abstinent from alcohol when compared with non-infected patients. No differences were present for other characteristics.

Table 1.  Differences in patient characteristics amongst Helicobacter pylori-non-infected and infected gastro-oesophageal reflux disease patients*

Characteristic	H. pylori-non-infected	H. pylori-infected	P-value
* 95% confidence intervals are shown in [ ], otherwise standard deviation shown; N.S., not significant.
n	51	50
Male gender (%)	23 (45%); [31–60]	26 (52%); [37–66]	N.S.
Age	49.2 ± 12.8	56.5 ± 12.9	<0.005
Weight (kg)	75.2 ± 10.9	76.6 ± 14.8	N.S.
Height (cm)	165 ± 11	163 ± 16	N.S.
Race
Caucasian	46 (90%); [79–97]	38 (76%); [62–87]	0.067
African Canadian	3 (6%); [1–16]	6 (12%); [5–24]
Middle-Eastern	2 (4%); [0–13]	2 (4%); [0–14]
Hispanic	0	2 (4%); [0–14]
Asian	0	1 (2%); [0–11]
Native Indian	0	1 (2%); [0–11]
Education
Primary	1 (2%); [0–10]	4 (8%); [2–19]
Secondary	18 (35%); [22–50]	30 (60%); [45–74]
Post-secondary	32 (63%); [48–76]	16 (32%); [20–47]	0.003
Smokers	12 (24%); [13–38]	8 (16%); [7–29]	N.S.
Packs per day	0.77 ± 0.53	0.57 ± 0.54	N.S.
Alcohol drinkers	25 (49%); [35–63]	15 (30%); [18–45]	0.067
Grams per day	1.72 ± 2.7	0.71 ± 0.84	N.S.
Coffee drinkers	33 (65%); [50–78]	37 (74%); [60–85]	N.S.
Cups per day	2.5 ± 2.0	2.3 ± 1.3	N.S.
Time since diagnosis of GERD (years)	5.6 ± 4.8	7.3 ± 7.7	N.S.

Differences in symptoms

Heartburn was present on a daily basis during the week prior to entering the study (off acid suppression) in the overwhelming majority of patients in both groups. The Carlsson Dent scores were consistent with GERD (i.e. score ≥4) in 96% (49/51) of patients not infected with H. pylori compared with 86% (43/50) of H. pylori-infected patients (N.S.). No significant differences were found between the two groups, but there existed a trend suggesting that the H. pylori non-infected patients were suffering from more frequent GERD-associated symptoms, including regurgitation (82% vs. 74% respectively), odynophagia (27% vs. 24%), asthma (8% vs. 6%), chronic cough (10% vs. 4%) and sore throat (14% vs. 4%). No differences were present in the rates of laryngitis (4% in both groups) or dysphagia (16% in both groups). The bloating and epigastric pain rates were also similar in both groups (59% vs. 70% and 57% vs. 52% for H. pylori non-infected vs. infected patients respectively).

As shown in Table 2, H. pylori-infected patients used the most potent acid suppressive medications (i.e. proton pump inhibitors) significantly less often than uninfected subjects (58% on PPI vs. 82%, respectively, P = 0.009). No differences between the two groups were found in the Spechler's Activity index (137.80 ± 27.84 in uninfected patients compared with 128.35 ± 32.69 in infected patients, P = N.S.), the GSRS (mean score 45.8 ± 14.0 vs. 43.7 ± 16.3 respectively, P = N.S.), or the GSRS Reflux domain (4.3 ± 1.3 vs. 4.1 ± 1.7 respectively, P = N.S.) which consists of heartburn (4.8 ± 1.4 vs. 4.4 ± 1.9 respectively, P = N.S.) and acid reflux (3.9 ± 1.9 vs. 3.9 ± 1.9 respectively, P = N.S.). With regard to quality of life, the PGWB was also not significantly different between the two groups (90.0 ± 19.4 in non-infected vs. 92.5 ± 22.1 in infected patients, P = N.S.). No differences were found in the responses to individual questions for either the GSRS or PGWB (data not shown), although the self-control domain of the PGWB was slightly different between the two groups (Table 2).

Table 2.  Differences in symptom severity between Helicobacter pylori-non-infected and H. pylori-infected gastro-oesophageal reflux disease patients*

	H. pylori-non-infected	H. pylori-infected	P-value
* 95% confidence intervals are shown in [ ], otherwise standard deviation shown; N.S., not significant; PPI, proton pump inhibitor; GSRS, Gastrointestinal Symptom Rating Scale; PGWB, Psychological General Well-Being (the six dimensions are listed).
n	51	50
Symptoms
Heartburn daily	44 (86%); [74–94]	48 (96%; [86–100]	N.S.
Regurgitation	42 (82%); [69–92]	37 (74%); [60–85]	N.S.
Dysphagia	8 (16%); [7–29]	8 (16%); [7–29]	N.S.
Odynophagia	14 (27%); [16–42]	12 (24%); [13–38]	N.S.
Asthma	4 (8%); [2–19]	3 (6%); [1–17]	N.S.
Cough	5 (10%); [3–21]	2 (4%); [0–14]	N.S.
Sore throat	7 (14%); [6–26]	2 (4%); [0–14]	N.S.
Laryngitis	2 (4%); [0–13]	2 (4%); [0–14]	N.S.
Carlsson Dent Score	9.9 ± 3.4	9.0 ± 4.0	N.S.
Symptom severity
Medication required
PPI used	42 (82%); [69–92]	29 (58%); [43–72]	<0.01
H2 blockers used	7 (14%); [6–26]	8 (16%); [7–29]
Antacids used	13 (25%) [14–40]	21(42%); [28–57]
Motility agents used	3 (6%); [1–16]	0
Spechlers Activity Index	138 ± 29	128 ± 33	N.S.
GSRS global score	45.8 ± 14	43.7 ± 16.3	N.S.
Reflux domain	4.3 ± 1.3	4.1 ± 1.7	N.S.
Quality of life
PGWB Global Score	90.0 ± 19.4	92.5 ± 22.1	N.S.
Anxiety	3.8 ± 0.7	4.0 ± 0.9	N.S.
Depressed mood	4.3 ± 0.7	4.3 ± 1.6	N.S.
Positive well-being	3.65 ± 0.4	3.7 ± 0.9	N.S.
Self-control	3.0 ± 0.9	3.8 ± 1.4	<0.01
General health	3.2 ± 0.8	3.1 ± 0.04	N.S.
Vitality	3.2 ± 0.6	3.4 ± 1.0	N.S.

Endoscopy results

Endoscopy results were available in 84% (43/51) of H. pylori-negative and 82% (41/50) of H. pylori-positive patients. Oesophagitis was present in 28% (12/47, 95% CI:15–44) of H. pylori-negative vs. 27% (11/41, 95% CI: 14–43) of H. pylori-positive subjects (P = N.S.). Amongst those with oesophagitis, the H. pylori non-infected had more grade A oesophagitis (83%, 10/12, 95%CI: 52–98 vs. 36%, 4/11, 95% CI: 11–69, P < 0.05) than infected patients. There were no patients with grade C or D oesophagitis in the study population. A hiatal hernia was present in approximately 30% of patients in each group.

Motility and 24 h pH results

All patients had normal peristalsis and swallow-induced LES relaxation. No significant difference was noted in the LES pressure between the two groups (14.3 ± 10.9 mmHg in H. pylori-non-infected patients vs. 16.6 ± 9.8 in H. pylori-infected patients, P = N.S.).

The 24 h pH results are shown in Table 3. No differences were found between the 44 non-infected and 43 infected GERD patients who successfully completed the 24 h pH-metry testing.

Table 3.  Differences in ambulatory 24 h pH results amongst Helicobacter pylori-non-infected and H. pylori-infected gastro-oesophageal reflux disease patients*

Characteristics	H. pylori-non-infected	H. pylori-infected	P-valueδ
* 95% Confidence intervals are shown in [ ], otherwise standard deviation shown; N.S., not significant. δ None of the P-values were ≤0.10.
n	44	43
Number (%) with pH <4 for ≥4% of the time	30 (68%); [52–81]	26 (60%); [44–75]	N.S.
Mean time pH <4 (%)	8.0 ± 9.2	12.5 ± 18.9	N.S.
Mean no. of reflux episodes	82.0 ± 79.8	60.2 ± 51.0	N.S.
Mean no. of reflux episodes >5 min	5.7 ± 14.4	4.5 ± 5.5	N.S.
Mean duration of longest reflux (min)	19.3 ± 23.0	31.9 ± 52.5	N.S.
DeMeester Score	28.9 ± 32.6	37.6 ± 45	N.S.

Planned subgroup analysis for patients with a Carlsson Dent questionnaire suggesting GERD

A planned sub-group analysis was also performed only for those patients who had Carlsson Dent scores of ≥4 given that such a score is considered diagnostic of GERD.²⁵ Findings were very similar. Helicobacter pylori infected GERD patients were significantly older (57 vs. 49 years), less likely to be Caucasian (72% vs. 90%), less often schooled to the post-secondary level (35% vs. 63%) and less often alcohol drinkers (26% vs. 51%). Proton pump inhibitors were used less frequently (53.5% vs. 81.6%, P < 0.01) in infected patients but no differences were found in the Spechler disease activity index, GSRS or PGWB global scores. No difference was noted in the presence or severity (grade) of oesophagitis in this analysis. On motility studies, it was found that the LES pressure was significantly greater (16.6 ± 9.9 mmHg) in the H. pylori-infected patients than in the uninfected patients (12.9 ± 5.0 mmHg, P < 0.05). The 24 h pH results, as in the original analysis, were no different between H. pylori-infected and non-infected groups (data not shown).

Discussion

It has been postulated that H. pylori infection, if in a body-predominant pattern, would lead to reduced gastric acid,^6–8 resulting in reduced acid content in any potential refluxate. In fact, it has been shown that eradication of H. pylori can lead to an increase in acid secretion.³⁸ The question thus has arisen as to whether H. pylori-infected GERD patients are any different than non-infected GERD patients.

Our prospective, double-blind study has clearly shown, using thorough GERD quantifying measures including validated symptom severity scores, endoscopy, and 24 h pH-metry, that any differences existing between H. pylori-infected and non-infected GERD patients would not be clinically significant. Although trends for more frequent GERD-associated symptoms (i.e. regurgitation, odynophagia, asthma, cough) were found in the non-infected group, in keeping with the original postulate, these differences were minor and did not achieve statistical significance. We did, however, find that the non-infected group used stronger acid suppressive medications suggesting more severe GERD symptoms; this was not attributable to longer duration of disease in the H. pylori-negative patents (Table 1). A difference in oesophageal acid sensitivity could explain this but we see no reason to expect such a difference in the two groups. Validated symptom scores including the Spechler's disease activity index and the GSRS failed to demonstrate any difference between the two groups. Individual reflux questions and the reflux domain of the GSRS were also no different between the two groups. Quality of life as per the PGWB was also similar with only the self-control domain being different for unexplained reasons. It is unlikely that this difference bears clinical significance.

Additional objective measures such as endoscopy also failed to demonstrate any clinically significant difference in the presence or absence of oesophagitis between infected and non-infected GERD patients. In fact, in patients with oesophagitis, those not infected with H. pylori had more grade A oesophagitis, a finding contrary to the postulate of less severe GERD with infection. Most convincingly, 24 h pH-metry demonstrated that the total time of pH <4 was similar in both groups. Indeed, H. pylori-infected individuals had a higher, although non-significant difference in mean total time of 12.5 ± 18.9% vs. 8.0 ± 9.2% in non-infected patients (Table 3). No differences were found for any of the other indices of 24 h pH-metry, including the DeMeester Score. When the analysis was performed only in individuals with a Carlsson Dent Score of ≥4, similar findings were noted. These findings are consistent with Zentilin et al.¹⁸ as well as other similar 24 h pH-metry comparisons between H. pylori-infected and non-infected GERD patients.^{17, 19}

An explanation for a possible increase in GERD post-H. pylori eradication is the effect of H. pylori-induced hypergastrinemia on raising the LES pressure.⁶ We did find a significantly higher LES pressure in H. pylori-infected patients compared with non-infected patients when limiting the analysis to those with Carlsson Dent scores ≥4 (16.6 ± 9.9 mmHg vs. 12.9 ± 5.0 mmHg, respectively; P < 0.05), but no differences in the LES pressure when the analysis was performed on all subjects (16.6 ± 9.8 mmHg vs. 14.3 ± 10.9 mmHg, N.S.).

Our study confirmed the previously demonstrated association between H. pylori infection and older age, non-Caucasian race and lower level of education (Table 1).

The current results are compatible with the findings by Moayeddi et al.¹⁶ In that well-designed study, H. pylori-infected GERD patients did not experience any increase in relapse of moderate to severe GERD symptoms post-eradication. Tefera et al.¹⁵ also found no change in total time pH <4 in GERD patients 12 weeks after eradication of this infection, and in fact, Schwizer et al.¹³ found that eradication positively affected GERD relapse. We are currently completing an RCT to determine whether eradication alters long-term (1 year) 24 h pH-metry results in H. pylori-infected GERD patients. It is unclear whether the findings in GERD patients can be extrapolated to non-GERD patients such as those with peptic ulcer disease or non-ulcer dyspepsia.

We have previously reported that eradication of H. pylori led to an increase in GERD in duodenal ulcer patients.⁶ Although not the same population, the present results would lead to an opposite conclusion. Given that the previous report was a retrospective analysis, and given the findings of more recent prospective studies in duodenal ulcer patients,^{11, 14} as well as a post hoc analysis of eight studies¹² which did not show an increase in GERD post-H. pylori eradication, we tend to favour the current conclusion that no clinically significant effect of H. pylori on GERD appears to be present.

One deficiency of this study is that we did not determine the pattern of gastritis in these patients, which given the potential influence of H. pylori on GERD via proximal gastritis, would have been interesting. It was chosen not to do so because of the difficulty in performing and interpreting these results. However, given that we found no difference between the H. pylori-infected and non-infected GERD patients, the findings would likely not have altered our conclusions. Furthermore, Bowrey et al.³⁹ examined gastric patterns in GERD patients and found gastritis of the cardia commonly regardless of whether patients were infected or not with H. pylori.

In conclusion, there is no clinically significant difference between H. pylori-infected and non-infected GERD patients with regard to the subjective or objective measures of disease severity, as demonstrated by this prospective double-blind study using thorough measures of GERD including validated questionnaires, endoscopy, motility and 24 h pH-metry.

Acknowledgements

This work was supported by the Fonds de la Recherche en Santé du Québec. Drs. Fallone and Barkun are Chercheurs Cliniciens Boursiers of the Fonds de la Recherche en Santé du Québec.