Gastro-oesophageal reflux is more prevalent in Western dyspeptics: a prospective comparison of British and South-East Asian patients with dyspepsia


Prof. P. Moayyedi, Gastroenterology Unit, McMaster University, Canada.


Background : There is a paucity of data directly comparing dyspepsia in Western and Eastern populations.

Aim : To compare clinical symptoms, epidemiological factors and endoscopic diagnoses in two sample populations with dyspepsia from the United Kingdom and South-East Asia in a cross-sectional study.

Methods : Patients with uncomplicated dyspepsia attending endoscopy units in Leeds, UK, and Kuala Lumpur, Malaysia, were prospectively interviewed and underwent subsequent endoscopy.

Results : A total of 1003 Malaysian patients (January 2002 to August 2003) and 597 Caucasian British patients (January 2000 to October 2002) were studied. The mean age was 48.7 ± 15.8 and 47.5 ± 13.8 years for the Malaysian and British patients respectively (P = NS). There was a higher proportion of cigarette smoking (35.7% vs. 12.4%, P < 0.0001) and alcohol consumption (34.4% vs. 2.0%, P < 0.0001) amongst British patients, but no difference in non-steroidal anti-inflammatory drug use nor having Helicobacter pylori infection. Gastro-oesophageal reflux disease (GERD) symptoms were more common in British compared with South-East Asian patients [heartburn (72% vs. 41%), regurgitation (66% vs. 29.8%) and dysphagia (21.1% vs. 7.3%), P < 0.0001]. This correlated with an increased endoscopic finding of oesophagitis (26.8% vs. 5.8%) and columnar-lined oesophagus (4.4% vs. 0.9%) amongst British patients (P < 0.001). A logistic regression model revealed that British Caucasian race (OR 9.7; 95% CI = 5.0–18.8), male gender (OR 2.0; 95% CI = 1.4–2.9) and not having H. pylori infection (OR 0.5; 95% CI = 0.3–0.7) were independent predictors for oesophagitis.

Conclusion : GERD is more common in British compared with South-East Asian dyspeptic patients suggesting that race and/or western lifestyle are important risk factors.


Dyspepsia is a global concern, although most of the published data have arisen from Western countries. It is assumed that dyspepsia in populations from ‘developing’ countries is mostly organic in nature, whilst functional dyspepsia is more prevalent in Western nations.1 The type of dyspeptic symptoms is also thought to vary in different populations. For example, symptoms associated with gastro-oesophageal reflux disease (GERD) are known to occur in some 20–25% of Western dyspeptics and is increasingly frequent in industrialized countries.1–3 However, several population studies from Asia have shown that GERD is not common in the East, particularly in people of Chinese ethnicity.4, 5

Although differences between dyspeptic populations from various parts of the world have been inferred, there is a paucity of data directly comparing different countries. Studying the differences in dyspepsia between various populations will enhance global understanding of the condition, facilitate better international research and allow more appropriate clinical management strategies at a local level in non-Western nations.

One of the authors (S.M.) had the opportunity to work in two centres across the globe with interests in the study of dyspepsia. Both hospitals where the study was conducted were secondary/tertiary referral centres and had direct access for upper gastrointestinal endoscopy from primary care doctors. The population in Malaysia is fairly typical of the South-East Asian region, consisting mainly of a racial mixture of indigenous Malays, Southern Chinese and South Indian emigrants of several generations. The present study was designed to examine the differences in clinical symptoms, epidemiological factors and endoscopic findings between dyspeptic populations in Leeds, UK, and in Kuala Lumpur, Malaysia.


Two consecutive series of patients who were referred with uninvestigated dyspepsia to the endoscopy units in both Leeds and Kuala Lumpur were prospectively evaluated. Dyspepsia was defined as chronic upper abdominal discomfort for at least 4 weeks with any associated symptom. The main races from Malaysia, namely Malay, Chinese and Indian, were grouped as South-East Asian for the purposes of this study, whilst all British patients were of Caucasian (white) ethnicity as non-white patients were excluded from this study. Referrals for endoscopy came from either Gastroenterology clinics or Primary Care Physicians as both units had an Open Access system. In both countries, uncomplicated dyspepsia in the community was usually treated empirically. Patients would be referred on for further evaluation if they failed to respond to this initial approach, or if they required repeat prescriptions for maintenance.

All patients were interviewed by medical personnel in both centres prior to endoscopy. Information obtained included basic demography, socioeconomic status (based on the age patients left full-time education), history of smoking (none, ex-smoker, current), amount of alcohol intake, history of non-steroidal anti-inflammatory drug (NSAID) ingestion within the last 4 weeks, use of anti-ulcer medication and clinical symptoms. Alcohol intake was defined as none, occasional (less than one drink per month) or regular (daily or weekly intake). In the UK, a validated instrument,6 based on the Leeds Dyspepsia Questionnaire, was used to obtain symptoms. In Malaysia, a self-designed questionnaire, also based on the locally validated Leeds Dyspepsia Questionnaire,7 was used to elicit the presence of epigastric pain, early satiety, nausea, abdominal bloating/distension, belching, weight loss, heartburn or regurgitation. Heartburn was defined as retrosternal discomfort or ‘burning sensation’ in both countries.

Endoscopists performing the procedure in this study were either specialist trainees or consultants in Gastroenterology in both centres. Standard diagnostic upper gastrointestinal endoscopy was performed in both endoscopy units with video endoscopes (GIF 130 or 140; Olympus Optical Co., Ltd, Tokyo, Japan). Endoscopic pathology was described similarly, with the use of the Los Angeles classification for oesophagitis in both units.8 The severity of oesophagitis in both populations was not compared due to inter-observer variation within the grades of oesophagitis.9

A rapid urease test of gastric biopsies was utilized by both centres as well to detect the presence of Helicobacter pylori infection. All patients in both centres had biopsies taken from the antrum and body of stomach for the purpose of this test.


Data analysis was performed using a standard software package (SPSS version 10, Chicago, IL, USA). Univariate comparisons between groups were performed using chi-square test, for all data except age which was analysed using the Student's t-test. All P-values were two-tailed with the level of significance defined at 0.05. The primary analysis was the association between ethnic background and risk of oesophagitis. The evaluation of ethnicity as an independent predictor of oesophagitis was determined by logistic regression adjusting for age, gender, educational level, smoking, alcohol intake and H. pylori status. These factors were decided a priori.


A total of 1003 Malaysian patients (January 2002 to August 2003) and 597 British patients (January 2000 to October 2002) who underwent consecutive upper gastrointestinal endoscopy for uncomplicated dyspepsia were analysed for the purpose of this study. None of the patients had declined having the procedure during the period of study and all of them had been conducted successfully. The majority of patients in both centres had been referred by primary care doctors in those countries, with a similar referral policy as highlighted in the previous section. Within the Malaysian (South-East Asian) population, the ethnic mixture was as follows: 257 (26%) Malays, 433 (43%) Chinese and 313 (31%) Indians. Table 1 shows the demographic details of both groups of patients with dyspepsia. The mean ages were similar, but there was a higher proportion of females (61.5% vs. 54.2%) in the South-East Asian group. Although there was a higher proportion of South-East Asians without any formal education, there were no other differences in attainment of higher education between the two groups of patients (Table 1).

Table 1.  Demographic details in dyspepsia patient groups
 South-East Asian (n = 1003)Caucasian (n = 597)P-value
  1. Values within parenthesis are expressed as percentage.

Age (years) (mean ± SD)48.9 ± 15.848.8 ± 14.60.91
Proportion male387 (38.5)274 (48.2)0.005
RaceMalay 257 (25.6)Caucasian 597 (100) 
Chinese 433 (43.2)
Indian 313 (31.2)
Time left full education
 None92 (9.2)00.05
 Left before minimum age253 (25.2)303 (50.8) 
 Left after minimum age (no degree)496 (49.5)210 (35.2) 
 Degree holder162 (16.2)84 (14.0) 
Alcohol consumption (occasional/regular)20 (2.0)183 (34.4)<0.0001
Smoking (>5 cigarettes/day)124 (12.4)190 (35.7)<0.0001
NSAID usage (up to 1 week before)226 (22.5)120 (22.6)0.97
H. pylori infection221 (22.0)127 (24.3)0.31

With regard to other parameters, British patients had a significantly increased frequency of alcohol intake (34.4% vs. 2.0%, P < 0.0001) and a higher rate of smoking (35.7% vs. 12.4%, P < 0.0001) compared with their South-East Asian counterparts. However, there was no difference in NSAID usage (British 22.6% vs. South-East Asian 22.5%) or patients with H. pylori infection (British 24.3% vs. South-East Asian 22%). In the UK, all patients had been instructed to discontinue any anti-ulcer/anti-secretory medication at least 2 weeks prior to endoscopy. In the Malaysian cohort, 314 (31.2%) patients were still taking anti-ulcer medication (78% H2 receptor antagonists, 22% proton pump inhibitors) at the time of endoscopy.

Within the South-East Asian races, the frequency of H. pylori infection was as follows: Malays 12.1%, Chinese 21.7% and Indians 30.7%.

Differences in reflux symptoms between Caucasians and South-East Asians

The variations in symptoms of dyspepsia are illustrated in Figure 1. Apart from the South-East Asian patients describing more belching, the main difference between the two populations related to symptoms of GERD. British patients with dyspepsia had more heartburn (72% vs. 41%, P < 0.0001), regurgitation (66% vs. 29.8%, P < 0.0001) and dysphagia (21.1% vs. 7.3%, P < 0.0001) compared with the South-East Asian races.

Figure 1.

Differences in symptoms of dyspepsia between South-East Asian and Caucasian patients with dyspepsia.

Differences in prevalence of oesophagitis between British and South-East Asians

The differences in symptoms have also been borne out in the endoscopic findings (Table 2). British patients had a significantly higher rate of endoscopic oesophagitis (26.8% vs. 5.8%, P < 0.0001) and columnar-lined oesophagus (4.4% vs. 0.9%, P < 0.001) when compared with the South-East Asian races. In addition, there were eight cases of benign oesophageal strictures due to GERD amongst British, but none amongst the South-East Asian patients (data not shown).

Table 2.  Endoscopic diagnoses in South-East Asian and Caucasian patients with dyspepsia
 South-East Asian (n = 1003)Caucasian (n = 597)
  1. Values within parenthesis are expressed as percentage.

Gastro-oesophageal tumour12 (1.2)3 (0.5)
Gastric ulcer41 (4.1)15 (2.5)
Duodenal ulcer38 (3.8)14 (2.3)
Reflux ooesophagitis58 (5.8)160 (26.8)
Columnar-lined oesophagus9 (0.9)26 (4.4)
Gastritis222 (22.1)37 (6.2)
Duodenitis22 (2.0)36 (6.0)
Hiatus hernia195 (19.4)149 (24.9)
Normal579 (57.7)278 (46.6)
Others13 (1.3)22 (3.7)

Other significant differences in the endoscopic diagnoses included a higher prevalence of gastritis, peptic ulcer disease and gastro-oesophageal malignancy amongst the South-East Asian patients with dyspepsia (Table 2). There were 10 gastric and two oesophageal cancer cases in the South-East Asian patients, whilst the British cases of malignancy consisted of two oesophageal and one gastric cancer. All cases of malignancy in the South-East Asian cohort were of Chinese ethnicity, and all above the age of 50 years. Gastritis had been defined endoscopically only (for the purpose of this study) as erythema, area gastricae, clefts and nodularity. However, we are well aware that this method has a low interobserver agreement, and have not placed much emphasis on the observed differences.

Differences in the reporting of reflux symptoms were also noted between the two populations studied. Exactly 58.6% of South-East Asian patients with endoscopic oesophagitis reported symptoms of reflux, in contrast to 80.6% of British Caucasians with oesophagitis. Similarly, symptoms of regurgitation were present in 39.7% of South-East Asian patients with oesophagitis compared with 69.4% of British patients with oesophagitis.

To investigate the differences in prevalence of oesophagitis that we had observed, we initially performed a univariate analysis for risk factors for oesophagitis in this sample population. This revealed that British Caucasian race, male sex, alcohol consumption, regular smoking and H. pylori infection were significantly associated with oesophagitis. After logistic regression, British Caucasian race (OR 9.67; 95% CI = 4.97–18.82), male sex (OR 2.01; 95% CI = 1.38–2.93) and not having H. pylori infection (OR 0.45; 95% CI = 0.28–0.72) remained independent significant risk factors for oesophagitis (Table 3). Within the South-East Asian races, Indian patients with dyspepsia appeared to have the highest prevalence of oesophagitis, that is 9% (OR 2.22; 95% CI = 1.11–4.45) compared with the Malays (5%) and Chinese (4%), but their risk magnitude was still lower compared with British Caucasians (27%). When all Malaysian races were considered together, British Caucasian race remained a strong risk factor for oesophagitis in the logistic regression model (OR 7.32; 95% CI = 4.74–11.29).

Table 3.  Logistic regression analysis for independent risk factor for oesophagitis
VariableNo. with oesophagitis (%)Unadjusted odds ratio95% CIAdjusted odds ratio95% CIP-value
  1. Values within parenthesis are expressed as percentage.

Age category
 <45 years86/627 (14)1.00 1.00  
 >45 years132/972 (14)0.990.74–1.330.960.66–1.390.81
 Female101/939 (11)1.00 1.00  
 Male117/658 (18)1.791.35–2.392.011.38–2.93<0.001
 Malay13/257 (5)1.00 1.00  
 Chinese17/433 (4)0.770.37–1.610.890.42–1.880.76
 Indian28/313 (9)1.840.93–3.642.221.11–4.450.02
 Caucasian160/596 (27)6.893.80–12.49.674.97–18.82<0.001
Age left education
 Degree holder22/217 (10)1.00 1.00  
 After minimum age78/633 (12)1.240.76–2.011.770.59–5.260.31
 Before minimum age69/452 (15)1.600.96–2.661.060.58–1.950.84
 None5/92 (3)0.510.19–1.391.360.76–2.430.29
 No/ex-smoker133/1220 (11)1.00 1.00  
 Yes71/314 (23)2.391.73–3.291.420.95–2.150.09
 No147/1331 (11)1.00 1.00  
 Yes57/203 (28)3.142.21–4.471.200.74–1.990.46
NSAID usage
 No159/1187 (13)1.00 1.00  
 Yes45/346 (13)0.970.68–1.380.920.60–1.420.37
H. pylori
 No166/1176 (14)1.00 1.00  
 Yes35/348 (10)0.680.46–1.000.450.28–0.72<0.001

In the Malaysian cohort with oesophagitis, 20 (34.5%) were taking anti-secretory medication at the time of endoscopy. In patients without oesophagitis, only 294 (31.1%) were on anti-secretory medication. None of the British patients were on anti-secretory medication at the time of endoscopy, in accordance with the local policy for open-access endoscopy.


We have shown that significant differences in dyspepsia exist between representative urban populations from both Europe (the West) and South-East Asia (the East). The clinical setting was in secondary care for both centres, but the majority of referrals for endoscopy had been from primary care doctors. No differences in referral criteria for endoscopy existed between primary care doctors in both countries at the time this study was conducted. Both sample populations were reasonably matched demographically, although differences in the overall standard of living still existed (e.g. higher economic status in Britain vs. Malaysia). However, a limitation of this study is that endoscopic reporting was performed by various personnel, leading to some degree of inter-observer variation with endoscopic diagnoses. Nevertheless, we have conducted a large study and the magnitude of the effect of race on risk of oesophagitis is unlikely to be completely explained by bias or confounding factors.

British patients with dyspepsia had a significantly higher proportion of GERD symptoms and endoscopic oesophagitis when compared with South-East Asian dyspeptics. Amongst Western patients, heartburn and acid regurgitation are known to be specific for GERD,10 and non-obstructive dysphagia has been reported in up to 46% of GERD patients as well.11 It is possible that the Asian patients under-reported symptoms of GERD as other symptoms of dyspepsia, such as belching, were more prominent. This is unlikely to explain the lower prevalence of oesophagitis in South-East Asian patients suggesting there is a genuinely lower prevalence of GERD in Asians compared with British dyspeptics. Furthermore, the use of anti-secretory medication amongst the Malaysian patients with dyspepsia is insufficient to account for the difference observed, as only a minority of patients without oesophagitis were taking them at the time of endoscopy (31.1%). Even if all of the Malaysian patients without oesophagitis on anti-secretory medication were assumed to have reflux disease that had been masked, our calculations indicate that this cannot make up the four times magnitude difference with the British cohort (26.8% vs. 5.8%).

The reason for the low prevalence of GERD in Asians has previously been postulated to be multi-factorial.12 Environmental factors such as dietary fat (shown to increase the frequency of transient lower oesophageal sphincter relaxation),13 smoking (transient reduction in lower oesophageal sphincter pressure, reducing salivary output and increasing acid clearance time)14 and alcohol consumption (inducing oesophageal motor dysfunction)15 are reportedly higher in the West than in the East. In this study, a significantly increased level of smoking and alcohol consumption were observed in British dyspeptics, but these were not independent risk factors for oesophagitis. Other factors including an increased usage of drugs including NSAIDs (can precipitate reflux symptoms)16 and an ageing population (reduced lower oesophageal sphincter pressure and polypharmacy)17 have also been linked with more GERD in the West. Once again, data from this study have shown no differences in NSAID usage or mean age disparities to account for the differences in GERD between British and South-East Asians.

Apart from these however, there are various other external elements that we have not been able to study which could have accounted for the observed differences in GERD. A Western standard of living, in particular its more sedentary lifestyle, when compared with the East, may certainly be relevant. Although vehicle usage as part of daily life in Malaysia and most of South-East Asia is on a steep incline, definite differences clearly exist with that of more industrialized nations like that of the UK. The nature of differing occupations and energy expenditure in both populations may also be contributory factors.

Another possibility for the variation in GERD between Eastern and Western patients with dyspepsia could lie in their intrinsic differences. In this study, we have demonstrated that the British Caucasian race alone was the most significant independent predictor of oesophagitis. Risk factors for GERD in the West have been shown to include a high body mass index (BMI),18 family history of reflux symptoms19 and the presence of a hiatus hernia.20 In a similar study to ours, but on a smaller scale, Kang and Ho demonstrated that an increased BMI, hiatal hernia and oesophagitis were more prevalent in English patients compared with Oriental Singaporeans.20 All of the endoscopy in this study was performed by the same individual, which excluded any inter-observer variation. We suspect under-reporting of hiatus hernia for the lack of difference of in our study, rather than the data being a true reflection. Nevertheless, there also remains some doubt whether hiatus hernia is an underlying cause of GERD or a result of chronic acid reflux.21

Additionally, variation in maximal acid output between races has been described before. Higher rates of maximal and basal acid output have been reported in Europeans and Americans compared with Indians and Chinese,22, 23 even when the lower BMI in Asians had been taken into account. All of these suggest a genetic predisposition to GERD, which could explain the difference in ethnicity that we have observed. A recent study implicating a significant genetic contribution to the aetiology of GERD, further showed that monozygotic twins were more likely to be concordant for GERD symptoms than dizygotic twins.24

Other independent risk factors for oesophagitis in this study have included male sex and not having H. pylori infection. Previous epidemiological studies have not demonstrated a relationship between sex and GERD.12, 18 However, it is well recognized that GERD symptoms are an important risk factor for Barrett's oesophagus25 and oesophageal adenocarcinoma,26 both of which are more prevalent in the Caucasian male population. More epidemiological studies will be required to examine this issue further.

The overall prevalence of H. pylori in Malaysian dyspeptics is generally lower than Eastern Asian countries due to the low prevalence amongst ethnic Malays.27 The finding that H. pylori infection is protective for GERD has also been described in East Asia.28 However, worldwide epidemiological data have not been consistent with the role of H. pylori in the aetiology of GERD29 and our data will, at present, add to the continuing body of evidence.

A recent systematic review examining the issue of ethnic and geographical differences in GERD between populations noted that most of the published data from different parts of the world were not consistent nor easily comparable due to varying methodologies.30 In this study, we have clearly described a higher prevalence of GERD in British compared with South-East Asian dyspeptics in socio-demographically comparable urban populations, using a similar dyspepsia questionnaire which had been validated in local settings and the same endoscopic criteria for diagnosis of GERD (i.e. the L.A. classification). We conclude that the cause for this difference is probably both a result of intrinsic ethnic differences and environmental influences that are more prevalent in the West.


This study was partly funded by the University Malaya Postgraduate Medical Grant.