An association between overweight and gastro-oesophageal reflux disease (GORD) has long been suspected and observed in case series. Recently, this association has become established in large-scale analytical epidemiological research.^1,2 In a recent meta-analysis based on the available literature, it was concluded that there is a dose-dependent association between increasing body mass index (BMI) and GORD, usually assessed as typical reflux symptoms—that is, heartburn or regurgitation.³ The precise strength of this association is difficult to assess, but it seems that persons with a BMI >30 kg/m² are at an about twofold increased risk of GORD than persons of normal weight. Interestingly, it seems that any increase in BMI, even within the range of what is considered normal BMI, is associated with a gradually increased risk of reflux symptoms.² A similar finding has been observed with regard to the relation between increased BMI and adenocarcinoma of the oesophagus,⁴ a cancer closely linked with GORD.⁵

In the present issue of Gut, Corley et al⁶ (see page 756) present a large cohort study that not only evaluates BMI in relation to GORD, but also evaluates the role of objectively measured abdominal diameter, with and without adjustment for BMI. Moreover, the study addresses the possible ethnic differences with regard to body composition and GORD. Intriguingly, the association between high BMI and symptoms of GORD was particularly strong among whites than among blacks and Asians. Moreover, after adjustment for BMI, the abdominal diameter correlated with symptoms of GORD in whites only, whereas no such correlation was identified in blacks or Asians. This is a new and possibly valid finding, but it needs to be confirmed by future studies before it can be considered true. Some potential methodological sources of error in the study might contribute to explain the findings. As pointed out by the authors, a particular problem might have been the differential misclassification of the outcome (GORD) between ethnic groups, which could distort effects of the true relation between measures of body composition with regard to reflux. Moreover, the cross-sectional design makes it impossible to reliably assess the timing of the relation between the exposure (body measures) and the outcome (GORD) under study. Yet, the reported prevalence of GORD seems to vary greatly in different geographical areas and between ethnic groups according to the available literature, and the results of the current study further indicate the possibility of ethnic biological differences in the development of GORD.

If a true ethnic difference in the aetiology of GORD is confirmed by future research, it might open a discussion of the occurrence of GORD-related diseases, including oesophageal adenocarcinoma.⁵ Adenocarcinoma of the oesophagus is considerably more common in whites than in other ethnic groups.^7,8 Thus, ethnic differences in the development of GORD might explain or at least contribute to the ethnic differences in the incidence of oesophageal adenocarcinoma. In this context, it should be stressed that the strong link between GORD and oesophageal adenocarcinoma should encourage further research that focuses on the aetiology of GORD. Such research might be the key to understanding the aetiology of oesophageal adenocarcinoma, a highly lethal cancer with a rapidly increasing incidence in Western populations, particularly among white males.^7,8

The current study by Corley et al⁶ also suggests a lack of difference between sexes with regard to the relation between increased BMI and GORD. This result is in contrast with some population-based studies, where a substantially stronger association between BMI and reflux was found among females than males.^1,9 Some data indicate that exposure to oestrogen might have a role in promoting reflux among women.^1,9 It has also been hypothesised that oestrogen might relax the lower oesophageal sphincter and thereby facilitate reflux.¹⁰ As data are conflicting, there is a continued need to assess the possibility of sex differences in the occurrence of GORD in future studies.

In the current issue of Gut, El-Serag et al¹¹ (see page 749) present an interesting study that evaluates the role of body measures in relation to GORD. They objectively measured various anthropometric variables, including waist circumference, in relation to objectively measured 24 h pH measurements of oesophageal acid exposure. The cross-sectional design and the risk of selection bias should not materially hamper this study, as reverse causality and selection due to obesity is unlikely. The study provides objective data regarding both the exposure (body measures) and the outcome (GORD). Moreover, the study includes waist and hip circumference, which has rarely been evaluated. Much of the previous research has been based on self-reported weight and height, and subjective symptoms of GORD. The current study by El-Serag et al¹¹ provides further evidence for an association between obesity and GORD, as the acid exposure was higher in obese persons, a finding in line with the previous literature. An interesting finding was that the obesity effect on oesophageal acid exposure was attenuated by adjustment for waist circumference, indicating that the association between obesity and GORD might be mediated by waist circumference. This information gives some clues to the biological mechanism of the association between BMI and GORD.

In the study by El-Serag et al, however, no differences were found between sexes or ethnic groups with regard to body measures and acid exposure of the oesophagus. Thus, this study does not provide much support for the ethnic differences found in the article by Corley et al.⁶

Intensified epidemiological research of the risk factors for GORD, such as those exemplified by the studies of Corley et al⁶ and El-Serag et al¹¹ should be encouraged. The physiology of GORD has been extensively studied and the physiological mechanisms have been well defined, but the actual underlying factors that actually cause GORD deserve much more attention. It is surprising that the available literature in this field is so sparse. Therefore, more detailed studies on various possible risk factors for GORD are warranted. As GORD is usually a chronic or recurrent disease, the available epidemiological research has suffered from the problems of cross-sectional design. A further development of the aetiological research of GORD would therefore be to assess incident cases of GORD. Such research would supply us with a reliable estimation of the incidence of GORD and also increase the validity of the aetiological research of GORD.