Gastroesophageal reflux disease (GERD) is a chronic digestive disorder resulting from the reflux of gastric contents into the esophagus that is often accompanied by troublesome symptoms of heartburn, acid regurgitation, or other extra-esophageal symptoms such as chest pain, chronic cough, hoarseness or globus.1 GERD imposes an important burden of illness worldwide with an estimated worldwide prevalence of between 8% and 33%.2 GERD is heterogeneous and presented with different phenotypes. Erosive reflux disease (ERD) is defined as mucosal defect by endoscopy and occurs in approximately 25% of patients with GERD symptoms.3 Non-erosive reflux disease (NERD) is present in approximately 70% of patients and is characterized by the presence of typical GERD symptoms associated with pathological acid reflux but the absence of demonstrable esophageal mucosal injury on endoscopy.4
Currently, acid suppressive therapy with proton pump inhibitors (PPIs) has proved to be the most effective treatment strategy for GERD patients and is recommended as a first-line treatment.5 PPIs have shown superiority over histamine H2-receptor antagonists for controlling symptoms as well as for healing erosions.6 Although symptomatic relief and acute healing of esophageal lesions can be achieved by short-term PPI treatments, up to 75% of patients with NERD and up to 90% of patients with ERD experience relapse within 6 months to 1 year after termination of initial treatments.7 Therefore, long-term continuous maintenance treatment with PPIs is required for the majority of patients with GERD to adequately control symptoms and to heal mucosal lesions. However, recent studies suggest that long-term PPI treatment can increase the risk of
This meta-analysis was conducted in accordance with the principles of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement.12 The need for approval from the institutional review board was waived since this study was performed by reviewing the literatures.
We searched all relevant studies that evaluated with efficacy of on-demand PPI maintenance and continuous PPI maintenance in GERD patients using PubMed, EMBASE, the Cochrane library, and KoreaMed from inception to December 2020. The following search terms were used: ([gastroesophageal reflux] or [reflux esophagitis] or [erosive esophagitis] or [heartburn]) and ([proton pump inhibitors] or [PPIs]) and ([maintenance] or [on-demand] or [continuous]). The detailed search strategies in each database and the search results are shown in Supplementary Table 1. Cited references in published studies were manually searched to identify other relevant studies.
The inclusion criteria in this study were as follows: (1) patients with GERD including uninvestigated GERD, NERD, and ERD who need PPI maintenance therapy; (2) comparison of on-demand PPI maintenance treatment with continuous PPI maintenance treatment; (3) randomized controlled trials (RCTs) with parallel design; (4) adults aged over 18 years; and (5) studies available in full-text form. The exclusion criteria were as follows: (1) abstract-only publications or unpublished studies, (2) case reports and narrative reviews, (3) systematic review/meta-analysis, (4) proceedings or study protocols, and (5) studies with insufficient data regarding treatment outcomes. In the first stage of study selection, irrelevant articles were excluded by reviewing the titles and abstracts of the studies retrieved by keyword search. Thereafter, the full texts of selected studies were reviewed in accordance with our inclusion and exclusion criteria.
All studies were assessed using Cochrane’s “Risk of Bias” tool, which includes the following domains: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data addressed over the short and long terms (attrition bias), selective reporting (reporting bias), and other biases.13 Two authors (S.J.K. and C.H.T.) independently assessed the methodological quality of all included studies, and any disagreements between the 2 evaluators were resolved by discussion and consensus. If no agreement could be reached, a third investigator (H.K.J.) was consulted.
The primary outcome of this meta-analysis was treatment failure with maintenance therapy, which means that patients discontinued the allocated maintenance therapy prematurely. The secondary endpoints were as follows: (1) proportion of patients who achieved successful symptom relief, (2) satisfaction with maintenance therapy, (3) amount of pill usage in each maintenance therapy, and (4) adverse events in both treatment groups. Using a data extraction form that had been developed in advance, 2 reviewers (S.J.K. and C.H.T.) independently extracted the following information: first author, year of publication, study design, country, study period, PPI maintenance dosage, symptom evaluation tools, rates of treatment failure, degree of satisfaction, pill usage during maintenance, and adverse events.
Meta-analyses were performed to calculate the pooled RRs with a 95% CI. The Mantel-Haenszel random-effects model was used for binary outcomes, and the inverse variance method was used for continuous outcomes. In addition, we evaluated subgroup analyses according to symptomatic GERD patients who did not undergo endoscopic evaluation, patients with NERD and patients with ERD. Heterogeneity was assessed using the
The flow diagram of the selection process, performed in accordance with PRISMA guidelines, is shown in Figure 1. The electronic database search identified a total of 630 records, of which 277 remained after removing duplicates. After screening the titles and abstracts, we excluded 246 irrelevant articles. The full texts of the 31 remaining articles were reviewed for eligibility. Of these, 11 studies met the inclusion criteria, and 20 studies were excluded from the final analysis.
The characteristics of the 11 included studies are summarized in Supplementary Table 1.14-24 The studies were RCTs published between 2004 and 2018, consisting of a total of 6138 GERD patients. The study design was a randomized, open-label study except for 1 study by van der Velden et al,24 which was a randomized, double-blind study. Two studies16,20 were conducted in Asian countries, and all other studies were performed in Western countries. Three studies by Hansen et al,17 Morgan et al,19 and Szucs et al22 included patients with GERD symptoms. Two studies by Bayerdorffer et al14 and Tsai et al23 included patients with NERD. The remaining 6 studies were performed in patients with ERD proven by endoscopy. The detailed GERD definition in each study is described in Supplementary Table 1. Except for 2 studies,16,23 the PPI doses for on-demand and continuous maintenance treatment groups were the same. In the study by Cho et al,16 40 mg esomeprazole was used in the on-demand group and 20 mg esomeprazole was used in the continuous group. In the study by Tsai et al,23 20 mg esomeprazole was used in the on-demand group and 15 mg lansoprazole was used in the continuous group. The duration of maintenance therapy in the 2 studies by Cho et al16 and van der Velden et al24 was 12 and 13 weeks, respectively, and the remaining studies were maintained for 6 months. The risk of bias graph and risk of bias summary of included studies are presented in Supplementary Figure 1 and Supplementary Figure 2, respectively.
A meta-analysis was performed on 8 RCTs that analyzed outcomes for treatment failure. The detailed outcome measurements of each study are summarized in Supplementary Table 2. Treatment failure was observed in 269 of 2958 patients (9.1%) in the on-demand PPI group, and 208 of 2838 patients (7.3%) in the continuous PPI group (Fig. 2). The RR with 95% CI was 1.26 (0.76-2.07) (
The proportion of patients who achieved successful symptom relief was higher in the continuous PPI group (RR, 1.16; 95% CI, 1.05-1.29) (Fig. 4). Funnel plot of 9 studies is shown in Supplementary Figure 6. No significant publication bias was detected from Egger’s linear regression method test (
Pill usage was significantly lower in the on-demand group as shown in Figure 5. In the on-demand group, pill usage per day was less than in the continuous group (risk difference, −0.52; 95% CI, −0.62-−0.42), which means that half of the PPI was used in the on-demand group compared to the continuous group. Patient satisfaction was measured in 6 RCTs. No significant difference in patient satisfaction was found between the 2 groups upon meta-analysis (RR, 0.97; 95% CI, 0.93-1.01) (Supplementary Fig. 11). The frequency of adverse events did not differ between the 2 treatment groups (RR, 1.02; 95% CI, 0.90-1.14) (Supplementary Fig. 12).
The meta-analysis of RCTs on maintenance treatment in GERD patients found no difference in treatment failure and satisfaction between the on-demand and continuous PPI maintenance groups. In terms of symptom relief, continuous therapy was more effective than on-demand therapy. However, meta-regression analysis demonstrated that the severity of esophagitis can affect the effectiveness of maintenance therapy. In NERD and mild ERD patients, treatment failure rate was similar in both therapies, and symptom relief was slightly higher in the continuous therapy group, although the clinical significance of this difference does not appear to be large (NNT = 14.3). In terms of treatment failure and symptom relief, continuous maintenance therapy was more effective in studies including severe ERD patients. However, in Asian patients where severe ERD patients are rare, there was no difference in the effect of symptom relief between both maintenance therapies. The amount of pill usage was significantly lower in the on-demand group. Therefore, in GERD patients with NERD and mild ERD patients or Asian patients, on-demand therapy may be preferentially recommended as an effective maintenance therapy.
Although many guidelines recommend on-demand therapy as a maintenance therapy for GERD, few meta-analysis articles have reported its effectiveness. Boghossian10 reviewed the efficacy of on-demand PPI therapy versus continuous PPI therapy in NERD or mild ERD (LA grade A or B). The authors concluded that on-demand PPI therapy may increase risk of “lack of symptom control” compared with continuous PPI use (RR, 1.71; 95% CI, 1.31-2.21), thereby favoring continuous PPI use. Khan et al11 also performed a meta-analysis comparing on-demand therapy with continuous PPI use in patients with NERD or mild ERD. Their conclusion that on-demand PPI was superior to continuous PPI (OR, 0.50; 95% CI, 0.35-0.72) in terms of ‘discontinuation of maintenance treatment’ is contrary to the results of the former study. In our study, when limited to NERD and mild ERD patients as in the 2 studies above, continuous maintenance treatment was slightly more effective in symptom relief (RR, 1.09; 95% CI, 1.01-1.18), as in the study of Boghossian, and there was no difference between the 2 groups in treatment failure (RR, 0.85; 95% CI, 0.43-1.66). This result was different from that of Khan’s study.11 This is because van der Velden’s study24 was added to our study due to the difference in inclusion criteria, and the results of this study have an influence on the outcome. However, in studies including moderate to severe esophagitis, continuous therapy showed superior efficacies in terms of treatment failure (RR, 4.24; 95% CI, 2.32-7.70) and symptom relief (RR, 1.37; 95% CI, 1.23-1.52). The American Gastroenterological Association guideline also reports that the use of long-term PPI has a greater benefit than harm in GERD patients with severe esophagitis or complications such as strictures.25
In subgroup analysis, 3 studies included all grades of esophagitis. However, grade C and D esophagitis patients accounted for only 6.0% (7/117) in the Nagahara’s study20 and 2.5% (2/80) in Cho’s study.16 In the Sjöstedt’s study,21 patients with grade C and grade D esophagitis accounted for 17.0% (80/470) and 5.0% (24/470) of all patients, respectively. Therefore, to confirm the effect of maintenance therapy in severe esophagitis such as LA grade C and D, the results of Sjöstedt’s study21 should be referenced. Treatment failure was significantly higher (RR, 4.24; 95% CI, 2.32-7.77), and symptom relief was significantly lower (RR, 1.37; 95% CI, 1.23-1.52) in the on-demand PPI group compared to the continuous therapy group. Furthermore, the authors also investigated the cumulative healing rate of esophagitis for up to 6 months. The cumulative proportion of patients in remission on the continuous PPI group was 81.0% compared with only 58.0% in the on-demand PPI group (
Among the 11 studies included in the analysis, 2 were conducted in Asia. Nagahara’s study20 was conducted in Japan and Cho’s study16 was conducted in Korea. Both studies included all grades of erosive esophagitis, but the proportion of patients with grade C and D esophagitis were only 6.0% in Nagahara’s study20 and 2.5% in Cho’s study.16 According to the results of epidemiologic studies, the proportion of severe erosive esophagitis in Asia is very low compared to that in Western countries. From a multicenter prospective study from Korea, out of 354 erosive esophagitis patients, only 12 patients were grade C and 4 were grade D, and the proportion of grade C and D among all erosive esophagitis was 4.5%.26 This figure is similar to the proportion of grade C and D patients in the 2 Asian studies included in the analysis, and much lower than the proportion (22.0%) found in the Western study by Sjöstedt et al.21 The effect of maintenance therapy in the Asian studies also shows different results from Western studies. When symptom relief was an outcome, continuous maintenance therapy was more effective than on-demand therapy in the entire study. However, the results of 2 studies from Asia showed no difference in symptom relief between the on-demand and continuous groups. Therefore, on-demand PPI therapy can be an effective maintenance treatment option in Asian patients with GERD.
Recent studies suggest that long-term PPI use may be associated with adverse events or complications, including enteric infection, bone fracture, kidney diseases, and micronutrient deficiency.8 A meta-analysis of 36 case-control studies and 14 cohort studies showed that
This study has several limitations. First, as described in the result section and Supplementary Table 1, 10 of the 11 studies included in analysis were open-label studies. On-demand therapy is expected to be difficult to conduct with a double-blind study. Only 1 study performed a double-blind study that showed that symptom control and quality of life were similar in the on-demand and continuous therapies. Since the open-label study is very similar to the actual clinical situation, the interpretation of the results is unlikely to be different significantly. Second, as a result of meta-analysis, the heterogeneity among the studies was high. The inclusion of patients with varying degrees of symptoms and esophagitis may have an effect on the heterogeneity as shown in the meta-regression analysis. Also, as a result of the subgroup analysis by the esophagitis grade, the heterogeneity within each group was significantly reduced. The diversity of outcome measurement also contribute to the high heterogeneity of this study. Third, no studies reported CYP2C19 polymorphism and their effect on PPI maintenance therapy. However, 2 studies used rabeprazole, which is not affected by CYP2C19 polymorphism, reported no difference between the maintenance therapies in terms of “treatment failure” and “symptom relief.” Forth, only 2 Asian studies were included in this analysis, and the results between Asian and Western studies show slightly different trends as shown in the analysis. It may be difficult to apply the results of this study to Asian patients. In this regard, more research is needed in the Asian region. Finally, intermittent therapy is another recommended maintenance therapy in some guidelines, and is also known to be an effective maintenance option. However, because of lack of RCTs on the intermittent therapy, its effect cannot be compared with the on-demand or continuous therapy.
In conclusion, the on-demand PPI therapy showed similar efficacies to continuous therapy in the outcome of treatment failure, symptom relief, and patient satisfaction in GERD patients, especially in NERD and mild ERD patients, and Asian patients. The severity of esophagitis significantly influenced the effect of the 2 maintenance therapies. In particular, continuous maintenance treatment seems to be more effective in patients with severe ERD. However, on-demand therapy has the advantage of reducing the amount of PPI usage by half compared to the continuous therapy. Therefore, the on-demand PPI therapy is recommended as an effective maintenance treatment modality in GERD patients unaccompanied by severe esophagitis.
This research was supported by a grant of the Korea Health Technology R&D Project through the National Evidence-based Healthcare Collaborating Agency (NECA), funded by the Ministry of Health & Welfare, Republic of Korea (Grant No. HI19C0481, HC19C0060).
Conceptualization: Hye-Kyung Jung and Kwang Jae Lee; data curation: Seung Joo Kang and Chung Hyun Tae; formal analysis: Seung Joo Kang, Chung Hyun Tae, and Seung Young Kim; fund acquisition: Hye-Kyung Jung and Kwang Jae Lee; methodology: Hye-Kyung Jung; supervision: Kwang Jae Lee; writing-original draft: Seung Joo Kang; and writing-review and editing: Chung Hyun Tae and Seung Young Kim.