The placebo effect is a relevant factor in evaluating effectiveness of therapies in functional gastrointestinal disorders

We read with great interest the paper by Kinoshita et al. [1] on the positive therapeutic effect of lansoprazole on dyspepsia and reflux symptoms in a very large group of Japanese patients with gastroesophageal reflux disease (GERD). We must congratulate the authors, since they were able to enroll a very large number of patients and to provide novel interesting data. However, we believe that their results must be interpreted with caution. Recently, Xiao et al. [2] prospectively evaluated 24-h pH monitoring and response to rabeprazole, 10 mg twice daily, in 186 functional dyspepsia (FD) patients identified by means of Rome III criteria [3], and found that those complaining of epigastric burning had the highest prevalence of abnormal reflux and the greatest response to anti-secretory therapy. These results were in keeping with those we observed in a study [4] evaluating the overlap between FD and Non-Erosive Reflux Disease (NERD) in 200 consecutive endoscopy-negative patients investigated using impedance-pH monitoring and a validated dyspepsia questionnaire. We observed that NERD patients with abnormal esophageal acid exposure more frequently reported epigastric pain and/or burning (Epigastric Pain Syndrome, EPS), while patients with functional heartburn (normal pH and negative symptom-reflux association = FH) referred more frequently to nausea, early satiety, postprandial fullness and bloating (Post-prandial Distress Syndrome, PDS). Moreover, patients with abnormal pH and overlapping EPS responded better to anti-secretory therapy (72 %) than patients with hypersensitive esophagus (normal pH and positive symptom-reflux association) and FH (53 and 29 %, respectively). In contrast, Kinoshita et al. [1] found that lansoprazole was equally effective in relieving EPS and PDS symptoms. This difference could in part be explained by the short-time clinical assessment after PPI administration (at 2 and 4 weeks), since it is well known that a large placebo effect (up to 50 %) is present in patients with functional gastrointestinal disorders [5, 6] and that this effect tends to decrease in a time-dependent manner. Moreover, the same explanation may be advocated considering the very high percentage of relief of dyspepsia symptoms recorded by Kinoshita et al., whereas in previous studies and more generally in clinical practice, the response of dyspepsia to anti-secretory therapy is less evident [7]. Furthermore, Kinoshita et al. [1] found that the therapeutic effect of both low and high doses of lansoprazole on dyspepsia, as well as on reflux symptoms, was higher in erosive esophagitis (EE) than in NERD. These data are in contrast with recent findings [8], and highlight an important limitation of this study; that is, the lack of functional tests, which are at present the only objective method to classify endoscopy-negative patients with typical reflux symptoms [9]. Indeed, recent functional investigations performed with impedance-pH technique have clearly shown that endoscopy-negative patients with typical reflux symptoms are greatly heterogeneous from a pathophysiological and histological point of view, and can be subdivided into several well-defined subgroups [10–14]. These new findings are relevant because they have contributed to change our current management of NERD [4]. For instance, patients with abnormal acid exposure respond to PPIs as well as those with EE [12, 13], but those with FH are usually refractory to PPIs and the use of antidepressants should be adopted [4]. Thus, it is quite possible that the difference found by Kinoshita et al. [1] in terms of clinical response between patients with EE and those with NERD is due to the incorrect inclusion in the latter group of patients with FH, who, based on their characteristics, no longer pertain to the realm of GERD.