The data from the Leeds HELP study should be placed in context with other studies that have assessed the prevalence and incidence of reflux symptoms. We therefore conducted a Medline literature search from 1966 to January 2005 (terms used available from the corresponding author) and performed a recursive search of the bibliographies from eligible papers.
We identified 31 papers,8–37 evaluating 77 671 subjects that assessed the period prevalence of heartburn symptoms in the community (Table 1). Eleven studies administered the questionnaire by post, five by telephone, and 15 by direct interview with a researcher (Table 1). In 15 papers, the questionnaire had been validated (Table 1). The period over which heartburn were assessed varied from enquiring about current symptoms to that about symptoms over a lifetime. The most common period of assessment was 1 year. None of these factors influenced the overall period prevalence; so despite these different methods of assessment, all studies were included.
Period prevalence of heartburn in Western populations
Studies assessed heartburn at different frequencies and these could be categorized as at least once a month, more than once per week and every day. These three frequencies of heartburn are therefore considered separately. Severe heartburn as well as predominant heartburn are also considered separately.
Twenty papers,8–11, 14–15, 19–23, 25, 28, 30–32, 35–37 assessing 51 866 subjects evaluated the period prevalence of heartburn occurring at least monthly (Table 1). There was significant heterogeneity between studies [Q = 806, degrees of freedom (d.f.) = 19, P < 0.0001], with the period prevalence varying between 13% and 36%. The data were synthesized using a random-effects model38 and the pooled period prevalence was 25% (95% CI 22–27%) (Figure 1). Three of these studies23, 30, 36 also reported heartburn and/or regurgitation as a definition of reflux symptoms. This increased the period prevalence of reflux symptoms by 3% (23) to 8% (36) compared with heartburn alone.
Figure 1. Forest plot of Western studies evaluating heartburn that occurred at least monthly. Explanation of Forest plot: each square represents an individual study. The size of the square is a measure of the size of the study and the line through the square gives a graphical representation of the 95% confidence interval of that study. The prevalence of heartburn in that study is given on the left-hand side of the graph with the 95% confidence intervals in brackets. The diamond at the bottom of the graph gives the pooled prevalence and the line gives the pooled 95% confidence interval.
Download figure to PowerPoint
Twelve papers10, 11, 15, 16, 19, 21, 22, 33, 34, 36, 37 reporting on 20 102 subjects described the period prevalence of heartburn symptoms occurring at least once per week (Table 2). There was less variability than heartburn occurring on a monthly basis but the heterogeneity was still statistically significant (Q = 72, d.f. = 11, P < 0.0001) (Figure 2). The period prevalence varied between 8% and 18% with a pooled rate of 12% (95% CI 11–14%) in a random-effects model38 (Figure 2).
Nine studies10, 15, 16, 19, 21, 22, 36, 37 assessing 17 006 subjects reported on the period prevalence of heartburn occurring on a daily basis. The period prevalence varied between 2% and 9% with statistically significant heterogeneity between studies (Q = 56, d.f. = 8, P < 0.0001) (Figure 3). The pooled period prevalence in a random-effects model38 was 5% (95% CI 4–6%) (Figure 3).
Only four studies23, 30, 39, including this one, evaluated the period prevalence of heartburn in terms of severity rather than frequency of symptoms. The proportion of subjects with severe heartburn varied between 1% and 7%. Three studies20, 39, including this one, described period prevalence in terms of heartburn as the predominant symptom and all reported 13% of subjects having GERD according to this definition.
Period prevalence of heartburn in East Asian populations
We identified six papers17, 18, 24, 26, 27, 29 reporting on 18 393 subjects that evaluated heartburn in East Asian populations. Three17, 24, 27 were from China, and one each from Korea,18 Singapore26 and Malaysia.29 Four studies administered the questionnaire by face-to-face interview while two used the telephone. Four studies used a validated questionnaire.
All six studies reported the period prevalence of heartburn at least once a month. The proportion varied from 3% to 11% with significant heterogeneity between studies (Q = 64, d.f. = 5, P < 0.0001). The pooled period prevalence was 7% (95% CI 6–9%) in a random-effects model (Figure 4).38 Only one study17 reported heartburn occurring more frequently than once per month. This study17 evaluated 2532 subjects and reported that heartburn occurred at least once per month in 11%, at least once a week in 4% and daily in 1.7% of cases.
Influence of gender on period prevalence of heartburn
We identified 13 studies10, 11, 13, 14, 19–21, 23, 24, 27, 28, 33 that reported the period prevalence of heartburn in men and women. A total of 48 231 subjects were studied and all used a definition of heartburn occurring at least once a month. Overall, 28% of men had heartburn at least once per month compared with 24% of women. There was significant heterogeneity between study results (heterogeneity χ2 = 276, d.f. = 12, P < 0.001) with no statistically significant effect of gender on the period prevalence of heartburn (OR of heartburn in men 1.14; 95% CI 0.90–1.44; P = 0.28) (Figure 5).
Variations of period prevalence of heartburn over time
We assessed whether there was any change in the period prevalence of heartburn in papers from almost 30 years ago to the present. A definition of heartburn occurring at least once a month was chosen to increase the number of papers available. Twenty papers were included8–11, 14–15, 19–23, 25, 28, 30–32, 35–37 and there was no trend towards an increase in the prevalence of heartburn, with time (Figure 6) although only three papers9, 15, 37 reported on the prevalence of heartburn before 1990.