Article Text
PDF
Statistics from Altmetric.com
Editor,—In a recent article, Lagergrenet al () reported no relation between body mass and gastro-oesophageal reflux in a Swedish population and concluded that reflux symptoms occur independently of body mass index. As the authors point out, the evidence on this subject is conflicting. A large recent US cross sectional study1 reported a strong positive association between body mass index and the prevalence of reflux symptoms (table 1). One possible explanation for the difference between the two studies is the younger age distribution of the US cohort. The prevalence of overweight has increased dramatically throughout Europe and North America in recent decades.2 As a consequence, the younger US cohort is likely to have accumulated more person years of overweight by any given age and the risk of reflux symptoms may be related to both the magnitude and years of overweight exposure. The authors also concluded, in the light of their findings, that weight reduction may not be justifiable as an antireflux therapy. Even if overweight is a poor predictor of reflux symptoms, this does not necessarily imply that weight reduction will not be of benefit in providing symptom relief. A significant beneficial effect of weight loss on symptoms of gastro-oesophageal reflux in overweight patients has recently been reported in a small study involving 34 patients.3 In addition, the degree of weight loss was directly correlated with improvement in symptom score. Elsewhere, strong and independent associations have been reported between both overweight and reflux symptoms and oesophageal adenocarcinoma.4 ,5 The evidence suggests that an overweight individual with reflux symptoms is at significantly increased risk of oesophageal adenocarcinoma. Further studies clarifying the role of weight loss in the management of reflux symptoms are clearly warranted.
Prevalence of reflux symptoms by body mass index (BMI)
Reply
Editor,—Like Maric and Cheng, we are also fascinated by the conflicting results in the literature on body weight and its possible association with gastro-oesophageal reflux. The critical question is whether the variation in results is explained by biologically meaningful differences (that is, if there is relevant effect modification) or if the discrepancies should be attributed to various biases, technical flaws, or simply to chance. Although it deserves to be pointed out that our data () primarily pertain to an elderly Swedish population (mean age 66 years), we believe that effect modification by age is a less likely explanation for the discrepancy between our results and those of Locke and colleagues.1-1Unpublished analyses stratified by age in our material showed the same absence of association both in relatively young (age <60 years) and old subjects. The combination of effects of age and nationality (as a proxy for early onset of overweight in the USA), as proposed by Maric and Cheng, seems in our view to be an even more improbable explanation. If the occurrence of reflux symptoms is dependent primarily on accumulated “exposure” to overweight, one would expect to find a stronger relation in our study, which involved older people, presumably with longer exposure time. One would then also expect that the prevalence of reflux symptoms would increase with age. But Locke and colleagues1-1 found, if anything, a tendency towards falling rates with age. An increasing prevalence with age was found in only a minority1-2 of numerous similar epidemiological studies. Moreover, if Maric and Cheng's hypothesis on the importance of accumulated time with overweight is correct, our study, which assessed body mass index at different points in time in the past should have had a greater chance of ascertaining any true association compared with the study of Locke et al which dealt with only current body mass index.
Non-differential misclassification of both exposure (body mass index) and outcome (reflux symptoms) due to imperfect recollection in our study may have attenuated our estimates.1-3 But prevalence rates for reflux symptoms well in agreement with the previous literature, and the strong association that we observed between these measures and the risk of oesophageal adenocarcinoma1-4somewhat allays this concern. A further possibility that may explain the conflicting results is if the relation between body mass index and reflux propensity is non-linear with a definite trend only in the very high end of the body mass index distribution, and hence the range of body mass index values in the negative studies was insufficient to detect it. The proportion (15%) of obese subjects (body mass index >30) in our sample was considerably lower than that in the Mayo study (23%), and few subjects (n=17) had ever had a body mass index greater than 35. The data of Locke et al were not however consistent with such a threshold effect, and although statistical precision was poor, we did not see any important tendency towards a positive relation, even in the very highest end of our body mass index distribution (unpublished data).
Thus the variation in results remains unexplained. Given that there is no clear geographical pattern among positive and negative studies, it appears that genetic differences between populations is an unlikely explanation. While uncontrolled non-randomised intervention studies, like the one cited by Maric and Cheng, contribute relatively little to our understanding of the importance of body weight (patients who manage to lose weight may differ from those who fail to do so in several important aspects), more in depth clinical and epidemiological studies are needed to resolve the apparent enigma.