The results of J-FOCUS, a randomized, open-label study, showed that omeprazole (10 mg once daily) was significantly more effective than famotidine (10 mg twice daily), mosapride (5 mg three times daily) or teprenone (50 mg three times daily) for providing sufficient and complete relief of upper GI symptoms after 4 weeks of treatment in endoscopically uninvestigated, H. pylori-negative patients. Symptoms were assessed using a previously validated symptom scale (GOS [19]), which has already been used in clinical studies to measure symptoms and treatment success in patients with dyspepsia with and without H. pylori infection [15, 20, 21]. In the present study, 66.9% of patients treated with omeprazole reached the primary endpoint of sufficient overall relief of upper GI symptoms (GOS ≤ 2) after 4 weeks of treatment, as compared with 41.0%, 36.3% and 32.3% of patients treated with famotidine, mosapride and teprenone, respectively (all, p < 0.001 vs omeprazole). The proportion of patients with complete overall symptom relief (GOS = 1) after 4 weeks of treatment was about three-fold higher in the omeprazole group (27.8%) than in the other three treatment groups (9.8%, 8.0% and 6.2%, respectively; all, p < 0.001 vs omeprazole). Within the omeprazole group, the proportion of patients with sufficient overall symptom relief after 4 weeks of treatment was highest in patients with predominant GERD symptoms, followed by those with predominant dyspepsia symptoms. None of the included patients was prescribed a PPI within 1 month of entering the study. Because PPIs are not available over-the-counter in Japan, it is very unlikely that any of the included patients had used a PPI in the month before the study.
Our results support those of 512 Canadian patients with H. pylori-negative dyspepsia in the randomized, double-blind CADET–H. pylori Negative (HN) study, which showed that omeprazole was significantly more effective than ranitidine, cisapride or placebo for relieving upper GI symptoms after 4 weeks of treatment [15]. Accordingly, both studies support the use of PPIs as first-line therapy for the treatment of dyspepsia.
Patients in J-FOCUS presented with multiple symptoms: the mean number of upper GI symptoms reported by each patient at study entry was six. The prevalence and severity of upper GI symptoms in J-FOCUS were comparable with those described in Japanese patients with GERD who had reflux esophagitis [10, 23]. The mean number of upper GI symptoms per patient was 5.4 in the study by Adachi et al., which included patients with GERD who had reflux esophagitis, and a similarly high symptom load was observed in the CADET-PE study, in which 80% of patients with uninvestigated dyspepsia had at least six upper GI symptoms [10, 11]. We observed substantial coexistence of different symptom types, with 41.4% of patients in J-FOCUS having coexisting GERD, dyspepsia and/or other upper GI symptoms. Coexistence of GERD and dyspepsia was common, with 19.8% of patients reporting both GERD and dyspepsia as their predominant symptom types. Adachi et al. also observed substantial coexistence of symptoms in patients with GERD [10]. In that study, the most prevalent symptoms were heartburn and regurgitation, which were reported by 71% and 68% of patients, respectively, but these were closely followed by upper abdominal heaviness (63%), upper abdominal pain (54%) and upper abdominal fullness (53%). In our study, a lower number of upper GI symptoms at baseline was associated with treatment success, but symptom severity did not affect treatment success on logistic regression analysis.
Within each treatment group, no pronounced differences in symptom response were observed among the eight symptoms assessed by the GOS (Figure 5). Nevertheless, it cannot be ruled out that there may have been different effects on symptom subtypes consisting of epigastric pain, postprandial fullness or early satiety, as defined by the Rome III diagnostic criteria. However, we decided against conducting such analyses, because the marked coexistence of symptoms that we observed in this study would have made the data difficult to interpret, and thus, inappropriate to draw conclusions from.
The Rome III diagnostic criteria for functional dyspepsia exclude predominant symptoms of troublesome heartburn and/or regurgitation, which are diagnostic criteria for GERD, with or without reflux esophagitis [8, 9]. However, dyspepsia and GERD often coexist in the same patient, and underlying GERD is a potential cause of dyspeptic symptoms [24]. GERD develops when the reflux of stomach contents causes troublesome symptoms and/or complications [8]. Acid can also play a role in dyspepsia because acid infusions into the stomach and duodenum have been shown to increase the perception of upper abdominal symptoms and affect gastroduodenal motor function in healthy individuals [25–27]. Duodenal acid exposure is increased in patients with functional dyspepsia and nausea, and this is associated with greater symptom severity [28]. Dyspepsia has also been linked with increased perception of postprandial symptoms [29], changes in visceral sensory function [30], and hypersensitivity to gastric distension [31].
GI endoscopy is useful to identify the underlying structural causes of symptoms, but can only be used to diagnose a subgroup of patients with upper GI diseases and disorders. In the CADET-PE study, the prevalence of reflux esophagitis on endoscopy was 55% (215/393) among patients with predominant heartburn and/or regurgitation, and 38% (236/647) among patients with other predominant upper GI symptoms, while only about 5% of patients had peptic ulcer disease [11]. In clinical practice in Japan, endoscopy does not usually form part of the immediate management of patients who visit their medical practitioner because of upper abdominal symptoms, unless alarm symptoms such as GI bleeding, acute weight loss or severe anemia are present. Instead, symptom control is the first priority and patients are usually treated with GI drugs with different modes of action that are selected according to the type and severity of upper GI symptoms, either in monotherapy or in combination therapy. However, up-to-date management guidelines for GERD (with and without reflux esophagitis) and dyspepsia in Japan need to be developed [16, 17]. These will also need to take into account the requirement for higher doses of PPIs in view of the decreasing rate of H. pylori infection in Japan [32].
It is important to consider that the empirical use of PPIs for patients with uninvestigated GI symptoms should be done with care. For example, the National Institute of Health and Clinical Excellence in the United Kingdom [33] recommends that immediate referral for endoscopy (seen within 2 weeks) is indicated for progressive dysphagia, unintentional weight loss, an epigastric mass, suspicious barium meal, iron deficiency anemia or persistent vomiting. They also recommend that, in patients aged over 55 years, endoscopy should be considered if symptoms persist despite H. pylori testing and acid suppression therapy. Furthermore, they recommend that empirical full dose PPI therapy is offered for 1 month to patients with dyspepsia. In the present study, approximately one-third of patients treated with omeprazole and up to two-thirds of patients treated with the other drugs did not attain sufficient symptom relief. In these patients, further investigations, including endoscopy, may be necessary for diagnosis and to guide future treatments.
J-FOCUS was the first study to compare the efficacy of omeprazole, famotidine, mosapride and teprenone in Japanese patients with uninvestigated upper gastrointestinal symptoms. Its randomized, multicenter, parallel-group design further adds to the strength of this study. The present study has the following limitations: it was open label; information on over-the-counter drug use before study entry was not collected; the long-term effects of each drug beyond the 8-week trial period were not investigated; and some centers were unable to register the target number of patients meaning the allocation ratio (i.e., 2:2:2:1) was not fully met. In addition, the study did not have a placebo arm, but because we only enrolled patients with moderate to severe symptoms at inclusion made this difficult to justify from an ethical standpoint. Thus, whereas the study provided a comparison of symptom relief among active drugs with different modes of action, the therapeutic gain of each drug compared with placebo could not be determined. However, the CADET-HN study, which included a placebo arm, showed that symptom response after 4 weeks of treatment was 51% (95% CI: 43–60%) with omeprazole, 36% (95% CI: 28–39%) with ranitidine and 31% (95% CI: 22–39%) with cisapride, compared with 23% (95% CI: 16–31%) with placebo [15]. No treatment-related adverse events were reported at either follow-up visit. This may be related to the 4-week treatment duration, which was possibly too short for adverse events to occur. However, we must acknowledge the possibility that some patients who did not attend the follow-up visits may have discontinued treatment because of adverse events. It is also possible that discontinuing prior therapies 1 week before the study might have aggravated symptoms or led to the onset of new symptoms. However, we were unable to assess this possibility in the present study.
Some other limitations warrant mention. First, we used a urine antibody test to screen patients for evidence of H. pylori infection. Although this test shows high sensitivity and accuracy, it has not yet been recommended by guidelines as a screening test. It is therefore possible that some patients with H. pylori infection were included in this study. Although the use of endoscopic biopsy may have avoided this possibility, this procedure is uncomfortable for the patient, time consuming and costly. Urea breath testing, rapid urease tests or serology are well established alternatives but, considering their reported accuracies and specificities [34], they are unlikely to improve the true detection rate over that achieved with the urine antibody test in this study. Second, it is possible that the opportunity for the investigators to exclude patients based on their judgment of eligibility could have introduced some bias into the study, as the investigators may have preferentially included patients likely to respond favorably to treatment. However, we deemed this to be important and useful, as our exclusion criteria could not cover every eventuality.