Characteristics affecting health-related quality of life (HRQOL) in Japanese patients with reflux oesophagitis and the effect of lansoprazole on HRQOL

Abstract

Objectives: A survey of health-related quality of life (HRQOL) in the Japanese general population has suggested that HRQOL is affected by sex and age. We investigated whether there is any effect of baseline patients' characteristics on HRQOL of patients with reflux oesophagitis (RE), and its changes with lansoprazole (LPZ) treatment.

Research design and methods: This was a post hoc analysis of an open-label, multicentre, post-marketing, observational study which investigated the effect of LPZ (15 or 30 mg/day for 8 weeks) on HRQOL in patients with RE. At baseline, and after 4 and 8 weeks of treatment, HRQOL was assessed using a Japanese version of the 8-item Short-Form Health Survey (SF-8) and a newly-developed RE-specific questionnaire (RESQ). Stratified analysis of changes in HRQOL scores according to baseline patient characteristics such as sex, age and baseline severity of typical RE symptoms was performed.

Results: A total of 8,757 patients were included in the efficacy analysis. At baseline, physical component summary scores from the SF-8 were lower in older patients (≥60 years) than in younger patients (<60 years), and mental component summary scores were lower in younger patients than in older patients. These scores significantly improved after LPZ treatment regardless of sex and age. HRQOL scores were lower in patients with more severe typical RE symptoms at baseline. After LPZ treatment, HRQOL scores significantly improved to similar levels irrespective of the severity of the typical RE symptom at baseline.

Conclusions: This post hoc analysis suggests that sex, age and severity of typical RE symptoms affect HRQOL in patients with RE in Japan, and such factors do not affect the improvement of HRQOL with LPZ treatment. The present study suggests a positive impact of LPZ; however, it is difficult to estimate the degree of improvement in HRQOL that may have arisen from symptomatic relief as part of the natural history of the disease from the study. Our results may overestimate the efficacy of LPZ. Further well-controlled clinical studies are needed to confirm the efficacy of LPZ on HRQOL in patients with RE.

Key words::

• lansoprazole
• proton pump inhibitor
• quality of life
• reflux oesophagitis

Introduction

Reflux oesophagitis (RE) is one of the most common conditions encountered in general practice. It occurs in an estimated 10–20% of Western, and 2.5–6.7% of Asian populations^1,2, and its prevalence appears to have increased in recent years³. Clinical studies conducted in Western countries have revealed that health-related quality of life (HRQOL) in patients with RE is impaired in physical, psychological and social aspects, and that HRQOL can be improved by treatment with proton pump inhibitors (PPIs)^4,5. Patients with RE seen in general practice in Japan also complain of impaired HRQOL and report improvement in HRQOL with antisecretory therapy. Since no quantitative analysis of HRQOL has been reported in a large cohort of Asian patients with RE, an open-label, multicentre, post-marketing observational study was conducted – the REQUEST (Re-assessment of Quality of Life on Reflux Esophagitis Treatment) study – to determine HRQOL in patients with RE in Japan and to investigate the impact of antisecretory therapy with lansoprazole (LPZ) on HRQOL in a clinical practice setting⁶. This study demonstrated that the demographic characteristics of Japanese patients with RE differed from a Western population of patients with RE, having a greater proportion of older females. Furthermore, RE had a marked negative impact on HRQOL in Japanese patients and this was significantly improved after LPZ treatment.

An earlier HRQOL survey in the Japanese general population found that scores related to physical aspects of HRQOL were lower in older subjects compared with younger individuals, whereas scores related to mental aspects of HRQOL were lower in the younger population⁷. Furthermore, differences between male and female subjects were also reported. Age and sex may affect the HRQOL of patients with RE; however, the influence of these characteristics has not been well investigated. The aim of this post hoc analysis based on the REQUEST study was to investigate whether similar differences in HRQOL occurred in a large cohort of Japanese patients with RE as was observed in the general population. In addition, whether any baseline patient characteristics, such as age or sex, affected HRQOL scores in patients with RE treated with LPZ was analysed.

Patients and methods

Patients and treatment

A total of 1,809 hospitals and clinics throughout Japan participated in this study. Patients enrolled into the study included those with newly-diagnosed RE, those who had experienced a relapse of RE following previous antisecretory therapy, and those who agreed to switch their current RE medication to LPZ (mainly because of insufficient efficacy of previous treatment). Endoscopy at the time of study enrolment was not mandatory, but if performed, physicians were allowed to enrol patients with confirmed mucosal breaks prior to the study. Endoscopy findings were evaluated according to the Los Angeles classification system with Japanese modifications (grade N: normal mucosa, M: minimal changes to the mucosa, such as erythema and/or whitish turbidity, A: non-confluent mucosal breaks <5 mm in length, B: non-confluent mucosal breaks >5 mm in length, C: confluent mucosal breaks <75% circumferential, or D: confluent mucosal breaks >75% circumferential) ^8,9.

Patients received LPZ 15 or 30 mg/day for 8 weeks. The dosage was left to the discretion of the individual physicians based upon their usual prescribing habits in everyday clinical practice. The use of additional medications, including over-the-counter medicine, other than PPIs was not restricted during the treatment period.

This study was performed in compliance with the Good Post-marketing Study Practice (GPSP), a ministerial ordinance concerning the standards of the implementation of the post-marketing survey of drugs in Japan¹⁰. Prior to enrolment, the purpose and methodology of the study were explained to all patients. Examination by an institutional ethics committee in each institution and written informed consent were not mandatory because GPSP does not require these for non-invasive observational studies.

HRQOL and treatment satisfaction

At baseline and after 4 and 8 weeks of treatment, HRQOL was assessed by self-administered questionnaires – the Japanese version of the 8-item Short-Form Health Survey (SF-8) and a newly-developed questionnaire for RE-specific HRQOL (RESQ). The SF-8 is a generic questionnaire that was developed from the longer 36-item Short-Form Health Survey (SF-36) to estimate HRQOL based on scores from eight domains. HRQOL scores obtained from the SF-8 show a high correlation with scores from the SF-36⁷. The Japanese version of the SF-8 was developed using a cross-cultural adaptation method which requires translation of the original English version into Japanese followed by back-translation into English, and its validity was confirmed using the Japanese general population⁷. The SF-36 has been used to evaluate HRQOL in patients with RE included in clinical trials conducted in Western countries^4,5. However, we used the SF-8 in the present study for the convenience of both patients and physicians in primary care since it is easier, quicker and more convenient to use in general practice. Patients assessed their condition for eight domains (‘general health’, ‘physical functioning’, ‘role: physical’, ‘bodily pain’, ‘vitality’, ‘social functioning’, ‘mental health’, and ‘role: emotional’) during a 1-week recall period. Scores for the eight domains and for the physical (PCS) and mental (MCS) component summaries were calculated according to the manual for the Japanese version of the SF-8⁷. A score of 50 is the mean for the Japanese general population across the eight domains and two summary scores, higher scores indicate a better HRQOL.

While eating and sleeping conditions are known to be impaired in patients with RE, the SF-8 does not include questions which directly address these. To evaluate eating and sleeping, various RE-specific HRQOL questionnaires have been developed¹¹. However, no standard form has been established and no Japanese version of these questionnaires were available at the start of this study. As such, the RESQ was developed to enable HRQOL to be assessed more specifically in patients with RE¹². Five questions were selected for the RESQ which patients with RE often complain of in general practice because the generic condition was assessed by the SF-8 in the present study and a small number of questions might be convenient for use in general practice. The grades provided by patients for ‘difficulty in daily life’, ‘dissatisfaction with limitations in amount able to eat’, ‘dissatisfaction having to avoid favourite food’, ‘dissatisfaction with disturbances in falling asleep’, and ‘dissatisfaction with interrupted sleep’ during a 1-week recall period were assessed using a 5-point Likert scale (5 = not at all, 4 = slightly, 3 = moderately, 2 = quite a lot, 1 = extremely). A five-item response was selected for use with the RESQ because patients answer the SF-8 and the RESQ in series and a similar format may be favourable to avoid confusion. The score for ‘difficulty in daily life’ was used as an indicator of the ‘daily life’ domain, while the mean score of ‘dissatisfaction with limitations in amount able to eat’ and ‘dissatisfaction having to avoid favourite food’ were used as an indicators of the ‘eating’ domain. The mean score of ‘dissatisfaction with difficulty in falling asleep’ and ‘dissatisfaction with interrupted sleep’ was used for the ‘sleep’ domain. The mean score of these three domains was defined as the RE-specific summary (RES) score. The score for each domain and the RES ranged from 1 to 5, with higher scores indicative of better HRQOL. The validity of the RESQ was assessed using data from the current study¹². Cronbach's alpha for each domain and RES score ranged from 0.841 to 0.917, indicating high internal consistency. Spearman's correlation coefficients between the SF-8 and RESQ ranged from 0.406 to 0.601 for each domain and summary score.

After 4 and 8 weeks of treatment, each patient also answered recall questions that assessed ‘chest/abdomen discomfort’, ‘comfort in daily life’, ‘eating condition’, and ‘sleep condition’ as compared with baseline (much better than before, better than before, same as before, worse than before, much worse than before).

Patients' overall satisfaction with treatment was assessed using a five-grade scale (very satisfied, satisfied, neither satisfied nor dissatisfied, dissatisfied, very dissatisfied) after 4 and 8 weeks of treatment. Patients who had been treated with other medications prior to receiving LPZ were asked the question ‘which drug did you prefer?’ (preferred LPZ, no difference between treatments, preferred the prior drug).

Subjective symptoms

At baseline, and after 4 and 8 weeks of treatment, physicians assessed the severity of typical RE symptoms (heartburn and regurgitation) and other symptoms commonly observed in patient with RE (chest/epigastric pain, nausea/vomiting, and dysphagia) on a four-grade scale (none, mild/occasionally or slightly symptomatic, moderate/considerably symptomatic, severe/intolerable). This scale has been used traditionally in clinical trials designed to investigate the efficacy of PPIs in patients with RE^13,14.

Safety

Adverse events were monitored throughout the 8-week treatment period. Only those events for which a causal relationship with LPZ therapy could not be ruled out were considered as adverse drug reactions.

Statistical analyses

Paired t-tests were used to analyse the changes in HRQOL scores from baseline. The t-test or one-way analysis of variance were used to compare HRQOL scores stratified by patients' characteristics. Values of p < 0.05 (two-sided) were considered to be statistically significant. Correlations between RESQ and recall questions were determined by Spearman's coefficient analysis. As an indicator of clinically meaningful change, effect sizes were calculated by dividing the mean change score between baseline and week 8 by the standard deviation at baseline¹⁵. Effect sizes of at least 0.2 were considered small, around 0.5 were considered medium, and 0.8 or greater were considered a large clinically meaningful change¹⁶.

Results

Patient characteristics

Of the 10,933 patients enrolled in the study, 8,757 were included in the efficacy analysis. Among the 2,176 patients excluded from the analysis, 1,620 did not complete the HRQOL questionnaire at both weeks 4 and 8, 551 did not complete the HRQOL questionnaire at baseline, and five did not receive LPZ therapy. Patient characteristics of the study population are summarised in Table 1. At study enrolment, endoscopy was performed in 3,263 patients: 28.9% of these had a Los Angeles classification grade N or M, 56.1% had grade A or B, and 15.0% had grade C or D. The initial dose of LPZ was 30 mg in 57.5% of the study population and 15 mg in the remaining 42.5%. During the week prior to LPZ administration, 1,869 patients received treatment with H₂-receptor antagonists (H2RAs) and 598 with other PPIs. Prokinetic agents were used concomitantly in 1,308 (14.9%) patients, 42.9% of whom were receiving these agents before study enrolment.

HRQOL and treatment satisfaction

Of the 8,757 patients included in the efficacy analysis, HRQOL questionnaires were recovered from 92.1% of patients at week 4 and 85.5% at week 8. HRQOL scores were summarised in Table 2. Scores for all eight domains of the SF-8 were below the level of the Japanese general population (score 50) at baseline. After 8 weeks of LPZ treatment, mean scores for all domains were significantly increased to the level of the general population. Effect size of general health was 1.1 indicating large clinically meaningful change. Those of other seven domains ranged from 0.4 to 0.7, and 0.6 for PCS and 0.5 for MCS indicating moderate clinical meaningful change. Mean scores of the RESQ were also significantly increased after 8 weeks of LPZ treatment. Effect sizes of daily life, eating, and RES were greater than 0.8 indicating large clinically meaningful change.

With regard to the recall questions at week 8, 89.1% of patients reported improvement in chest/abdomen discomfort, while 84.2% indicated that their daily life was more comfortable after LPZ treatment. In addition, 75.2% of patients reported improvement in eating condition, and 58.0% reported improvement in sleep condition. The difference in RESQ scores between baseline and week 8 was larger when the answer given to the recall question was ‘better’ (e.g., 1.16 in ‘much better than before, 0.77 in ‘better than before’, 0.27 in ‘same as before’, –0.32 in ‘worse than before’, and –1.42 in ‘much worse than before’ in the ‘eating’ domain). Spearman's correlation coefficients between each RESQ domain and recall question ranged from 0.340 to 0.371. In terms of treatment satisfaction at week 8, 36.0% of patients responded as being ‘very satisfied’, 51.0% as ‘satisfied’, 11.2% as ‘neither’, 1.6% as ‘dissatisfied’, and 0.2% as ‘very dissatisfied’.

HRQOL scores and treatment satisfaction stratified by baseline patient characteristics

At baseline, PCS scores were significantly lower in older patients (≥60 years) than in younger patients (<60 years) and MCS scores were significantly lower in younger patients than in older patients. This applied to both males and females. These scores significantly increased after 8 weeks of treatment; however, PCS was still lower in older patients and MCS remained lower in younger patients (Figure 1). At baseline, HRQOL scores were lower in patients with more severe typical RE symptoms compared to patients with less severe patients. These scores significantly increased to levels similar to the total RE cohort following treatment with LPZ, irrespective of the severity of the typical RE symptom at baseline (Figure 2). In respect of treatment satisfaction at 8 weeks, the proportion of ‘very satisfied’ was higher in patients who had more severe typical RE symptoms at baseline (Figure 3).

In 6,290 patients who had not received acid suppressant drugs during the week prior to LPZ administration, typical RE symptoms at the start of LPZ treatment were severe in 14.4% of cases, moderate in 49.8%, and mild in 30.7%. Significant improvement in HRQOL scores was observed after 8 weeks of LPZ treatment (PCS 44.7–49.6, MCS 46.4–50.5, and RES 3.52–4.44). In terms of treatment satisfaction at week 8, 88.0% of patients responded as being ‘very satisfied’ or ‘satisfied’.

Of the 1,869 patients treated with H2RAs prior to LPZ administration, 850 received full dosages (cimetidine 800 mg/day, ranitidine 300 mg/day, famotidine 40 mg/day, roxatidine 150 mg/day, nizatidine 300 mg/day, raftidine 20 mg/day) and 58.6% of these had been receiving treatment for ≥2 months. The remaining 1,019 patients had been treated with half dosages of H2RAs, 62.6% of whom had been receiving treatment for ≥2 months. In this subgroup of patients previously treated with H2RAs, typical RE symptoms at the start of LPZ treatment were severe in 11.7% of cases, moderate in 48.6%, and mild in 33.4%. SF-8 scores at the start of LPZ treatment in this patient subgroup were lower than those in the Japanese general population but significantly increased to the level of the general population after treatment with LPZ (PCS 44.3–48.7 and MCS 46.9–50.3). The score for each RESQ domain also significantly increased (RES 3.56–4.42). In terms of treatment satisfaction at week 8, 85.9% of patients responded as being ‘very satisfied’ or ‘satisfied’. At week 8, 81.7% of patients expressed a preference for LPZ treatment. At baseline, 43 patients treated of this subgroup did not report any reflux/dyspepsia symptoms. Difference in scores of RESQ between baseline and week 8 of these patients was 0.10 for the ‘daily life’ domain, 0.08 for the ‘eating’ domain, and –0.04 for the ‘sleep’ domain. Nevertheless, during the recall questions at week 8, 48.8% reported that their chest/abdomen discomfort had improved. Improvements in daily life, eating condition and sleep condition were also reported in 41.9%, 41.9% and 27.9% of patients, respectively.

Safety

A total of 10,928 patients treated with LPZ were included in the safety analysis. Adverse drug reactions possibly related to LPZ treatment were reported in 183 patients (1.7%) with the most common events being diarrhoea (0.4%) and constipation (0.2%).

Discussion

In the present study, baseline PCS and MCS scores of the SF-8 in patients with RE were 44.6 and 46.6, respectively. In other disease settings in Japan, PCS and MCS scores were reported to be 47.1 and 48.1 in patients with neck/shoulder stiffness, 50.4 and 48.4 in those with cedar pollinosis (pollen allergy), and 35 and 47 in those with active rheumatoid arthritis, respectively^17–19. These results suggest that patients with RE in Japan suffer from a significant impairment of HRQOL. However, an HRQOL survey using the SF-8 in the Japanese general population reported PCS in an older population was lower compared with younger individuals (e.g., 50.2 for ages 30–39 years and 46.6 for ages 70–75 years), whereas MCS was lower in the younger population (e.g., 49.2 for ages 30–39 years and 51.3 for ages 70–75 years)⁷. Therefore, the influence of age on the interpretation of HRQOL scores should be taken into account. In the present study, a stratified analysis by age and sex revealed that PCS scores were lower in older patients than in younger patients, particularly in female subjects. While PCS scores significantly increased after LPZ treatment, the score for older patients remained lower than for younger patients. In contrast, MCS scores in younger patients were lower than in older patients at baseline. Although MCS scores increased after LPZ treatment, the mean score for younger patients remained lower than that for older patients. This suggests that physical HRQOL in older patients and mental HRQOL in younger patients with RE are basically impaired by factors unrelated to RE, possibly based on age-related limited physical mobility in older patients or increased psychoemotional stress in younger patients, and RE aggravates their burden. Since a significant improvement in HRQOL scores after LPZ treatment was recorded, irrespective of age and sex, LPZ treatment was beneficial in minimising the impairment of HRQOL associated with RE.

At baseline, HRQOL scores were lower in patients with more severe typical RE symptoms. After 8 weeks of LPZ treatment these scores significantly increased to levels comparable to those of the total patient cohort, irrespective of the baseline severity of the typical RE symptoms. Interestingly, the proportion of patients who responded ‘very satisfied’ after LPZ treatment was higher in patients with severe symptoms at baseline than in those with mild or moderate symptoms. Treatment satisfaction may reflect the changes in HRQOL scores from baseline.

With regard to treatment prior to LPZ treatment, a significant improvement in HRQOL scores after LPZ treatment was observed in patients treated with H2RAs prior to LPZ as well as those who had not received acid suppressant drugs prior to LPZ treatment. This result suggests that switching treatment from an H2RA to LPZ is one of the treatment options for improving HRQOL when prior treatment with H2RA has been insufficient. In clinical practice it is not uncommon to encounter patients who report improved well-being after PPI treatment despite having been asymptomatic at the start of PPI therapy. It is speculated that some patients may become inured to symptoms over time (possibly making their symptoms appear less problematic) and that there might be a subset of patients who assume that no further clinical benefit can be obtained after partial improvement in symptoms has been achieved by prior treatment. In such patients, prospective assessment of HRQOL using RESQ might underestimate improvement in HRQOL. For this reason, recall questions were also used in the present study to evaluate the perceived degree of improvement in HRQOL from baseline. Among patients previously treated with H2RAs in the present study, 43 patients did not complain of any reflux/dyspepsia symptoms at the start of LPZ treatment. After LPZ treatment, approximately half of them reported improvement in their HRQOL in recall questions despite the small difference in scores of RESQ between baseline and week 8. Although it is unclear whether RESQ underestimated improvement of HRQOL or that recall questions overestimated it, there may be some patients in whom switching from prior treatment to PPI is beneficial even if they did not complain of reflux symptoms during treatment prior to PPI.

The present study suggests a positive impact of LPZ treatment on HRQOL in patients with RE. However, our findings need to be interpreted with caution. The study was designed to investigate HRQOL in a large number of patients with RE being cared for in a clinical practice setting; as such it did not include a control group. Consequently, it is difficult to estimate the degree of improvement in HRQOL that may have arisen from symptomatic relief as part of the natural history of the disease. Moreover, patients' assessment of HRQOL after treatment might be influenced by the patient–physician relationship. There might be some patients who overrated HRQOL after LPZ treatment to please physicians. Our results may therefore overestimate the efficacy of LPZ on HRQOL in patients with RE. To confirm the efficacy of LPZ in this patient population, well-controlled clinical studies, such as comparison with H2RAs, would be required.

With regard to safety of LPZ treatment, the most common events were diarrhoea (0.4%) and constipation (0.2%), and most events were mild. Incidence of diarrhoea and constipation has been reported to be approximately 4% and 1%, respectively, in previous clinical studies of LPZ for RE^20,21. Our results confirm that LPZ treatment is well-tolerated.

Conclusion

This post hoc analysis suggests that sex, age and severity of typical RE symptoms affect the HRQOL of patients with RE in Japan, and such factors do not affect the improvement of HRQOL with LPZ treatment. However, additional well-controlled clinical studies are required to confirm the efficacy of LPZ on HRQOL in patients with RE in Japan.

Figure 1. HRQOL scores at baseline and 8 weeks of LPZ treatment stratified by sex and age.

Figure 2. HRQOL scores at baseline and 8 weeks of LPZ treatment stratified by severity of typical reflux oesophagitis symptom at baseline.

Figure 3. Treatment satisfaction after 8 weeks of LPZ treatment stratified by baseline patient characteristic.

Table 1. Baseline demographic characteristics.

Characteristics	n (%)
Sex
Male	3,531 (40.3)
Female	5,226 (59.7)
Age
Male (median)	62 years
<60 years	1,592 (45.1)
≥60 years	1,939 (54.9)
Female (median)	70 years
<60 years	1,289 (24.7)
≥60 years	3,937 (75.3)
Symptoms
Heartburn	8,020 (91.6)
Regurgitation	5,567 (61.3)
Chest/epigastric pain	5,249 (60.0)
Nausea/vomiting	4,194 (47.8)
Dysphagia	4,359 (49.8)
Severity of typical reflux oesophagitis symptom
None	471 (5.4)
Mild	2,790 (31.9)
Moderate	4,270 (48.8)
Severe	1,181 (13.5)
Not known	45 (0.5)
Prior episode of reflux oesophagitis
Yes	5,659 (25.6)
No	2,239 (64.6)
Unknown	859 (9.8)
Duration between diagnosis and enrolment in the study
<1 month	4,086 (46.7)
1–6 months	1,453 (16.6)
7–12 months	506 (5.8)
13–24 months	615 (7.0)
≥25 months	1,236 (14.1)
Not known	861 (9.8)

Table 2. HRQOL scores at baseline, 4 weeks and 8 weeks after treatment with lansoprazole. Values expressed as mean ± standard deviation (95% confidence interval).

	Baseline (n=8,757)	Week 4 (n=8,065)	Week 8 (n=7,487)
SF-8
General health	43.8 ± 7.5 (43.6, 43.9)	50.3 ± 6.8* (50.2, 50.5)	52.0 ± 7.0* (51.8, 52.1)
Physical functioning	46.2 ± 7.7 (46.0, 46.3)	48.4 ± 6.3 * (48.3, 48.6)	49.2 ± 6.0* (49.0, 49.3)
Role: physical	45.9 ± 8.4 (45.7, 46.1)	48.8 ± 6.6 * (48.7, 49.0)	49.5 ± 6.3* (49.4, 49.6)
Bodily pain	48.0 ± 9.2 (47.8, 48.2)	52.3 ± 8.1 * (52.1, 52.4)	53.3 ± 8.0* (53.1, 53.4)
Vitality	46.3 ± 7.8 (46.2, 46.5)	50.6 ± 6.7* (50.4, 50.7)	51.8 ± 6.7* (51.7, 52.0)
Social functioning	45.2 ± 9.5 (45.0, 45.4)	48.3 ± 8.0* (48.2, 48.5)	49.3 ± 7.7* (49.1, 49.5)
Mental health	47.5 ± 7.5 (47.4, 47.7)	51.0 ± 6.4* (50.8, 51.1)	52.0 ± 6.2* (51.8, 52.1)
Role: emotional	47.0 ± 8.1 (46.8, 47.2)	49.8 ± 6.1* (49.7, 50.0)	50.4 ± 5.9* (50.3, 50.6)
Physical component summary	44.6 ± 7.5 (44.4, 44.8)	48.4 ± 6.4* (48.2, 48.5)	49.3 ± 6.3* (49.2, 49.5)
Mental component summary	46.6 ± 7.8 (46.4, 46.7)	49.6 ± 6.3* (49.4, 49.7)	50.5 ± 6.1* (50.3, 50.6)
RESQ
Daily life	3.20 ± 1.10 (3.17, 3.21)	4.20 ± 0.90* (4.20, 4.24)	4.40 ± 0.80* (4.40, 4.43)
Eating	3.66 ± 1.12 (3.64, 3.69)	4.40 ± 0.79* (4.39, 4.42)	4.55 ± 0.71* (4.54, 4.57)
Sleep	3.75 ± 1.08 (3.73, 3.78)	4.22 ± 0.89* (4.20, 4.24)	4.33 ± 0.83* (4.31, 4.35)
Reflux oesophagitis specific summary	3.54 ± 0.89 (3.52, 3.55)	4.28 ± 0.69* (4.27, 4.30)	4.43 ± 0.64* (4.42, 4.45)

*p<0.001 (vs. baseline).

Acknowledgments

Declaration of interest: The REQUEST study was sponsored by Takeda Pharmaceutical Company Limited, Osaka, Japan. M. Hongo, Y. Kinoshita, H. Miwa, and K. Ashida were medical advisors for the present study.

M. Hongo contributed to the study design and drafted the manuscript. Y. Kinoshita, H. Miwa, and K. Ashida contributed to the study design and critical revision of the manuscript for important intellectual content. Editorial assistance was kindly provided by Dr Steve Clissold, ContentEdNet, and Takeda Pharmaceutical Company reviewed the manuscript. All authors approved the final draft for submission and take full and final responsibility for the contents of the article.

We thank the physicians at the 1,809 institutions that participated in the REQUEST study who made this work possible.