Introduction
Gastroesophageal reflux disease (GERD) is common in the general population and represents a frequent reason for visits in primary care practices [
1‐
4]. The spectrum of disease manifestations associated with GERD ranges from mild heartburn to erosive esophagitis and, less commonly, may generally include more serious medical conditions such as esophageal ulcer, esophageal stricture, and a pre-cancerous condition known as Barrett's esophagus [
1]. Effective treatment of GERD is necessary to heal the esophagus and to provide symptom relief. The frequency and severity of symptoms associated with GERD have been demonstrated to diminish patient health-related quality of life (HRQL) [
5‐
8]. Studies have shown that medical intervention in the treatment of GERD provides improvement in GERD-related symptoms and in turn, an improvement in well-being and functioning (i.e., HRQL) [
5,
9‐
14]. Revicki et al. have demonstrated an association between the complete resolution of heartburn symptoms and improvements in measures of HRQL [
15]. Few studies have been published relating the severity of GERD symptoms at the start of treatment with HRQL outcomes throughout the course of treatment. Of further interest is whether complete resolution of heartburn symptoms has the same effect on patients in terms of HRQL given more or less severe symptoms at the start of treatment.
The primary objectives of this study were to evaluate the association between baseline heartburn severity and endpoint HRQL in patients treated for heartburn; and to examine the relationship between complete resolution of heartburn symptoms, while controlling for baseline severity, and HRQL outcomes. Previous research indicates that baseline severity of disease and health status scores may be associated with changes in HRQL outcomes [
16,
17] where patients with very severe disease (and worse HRQL) may demonstrate differential impact of effective medical therapy than those with milder disease. Since the diagnosis of GERD and related treatment decisions are based on severity of symptoms and the reported interference of GERD symptoms with patient well-being [
18], clinicians treating patients with GERD are concerned about differential effects of treatment by initial disease severity. It is therefore useful to evaluate whether the association between symptom resolution and HRQL varies by baseline disease severity, since there may be differences in proportion of patients who experience symptom resolution, depending on pre-treatment disease severity. We performed a secondary analysis of three clinical trials comparing omeprazole and ranitidine to evaluate the relationship between severity of heartburn symptoms at the start of treatment and HRQL at subsequent weeks of treatment. A secondary objective was to evaluate whether complete resolution of heartburn symptoms, while controlling for baseline severity, results in improved well-being and functioning. The latter objective extends previous work that demonstrated complete resolution of these symptoms improves HRQL [
15].
Methods
Data sources
Data from three clinical trials were used to evaluate the relationship between baseline level of heartburn symptom severity and HRQL. Each of the three trials was originally designed to compare the effectiveness of omeprazole versus ranitidine in the treatment of patients with symptoms of GERD.
Clinical Trial 1
This two-phase multi-center trial was designed to compare omeprazole and ranitidine in the treatment of patients with at least moderate to severe heartburn symptoms and to evaluate HRQL in patients with poorly responsive symptomatic GERD [
5,
13]. Phase 1 (open label) enrolled 533 patients with a 6 month or longer history of heartburn, including moderate to severe heartburn during four out of the last 7 days prior to the start of the phase. Patients enrolled in the open label phase were started on ranitidine 150 mg twice daily for 6 weeks. Patients who experienced one or more episodes of moderate to severe heartburn during Week 6 and who had taken at least nine ranitidine tablets during Week 6 were classified as non-responsive. These patients were eligible for Phase 2 (double-blind) and were randomized to receive either omeprazole 20 mg once daily or ranitidine 150 mg twice daily for 8 weeks. Three hundred sixteen patients (59%) were randomized in the double blind phase of the trial; baseline and 8-week data were included in this secondary analysis.
Clinical Trial 2
The second study was a multi-center, double-blind, randomized clinical trial designed to compare omeprazole and ranitidine in the treatment of heartburn and included baseline and follow-up data (weeks 4, 8, 12, and 16) [
14]. Eligible subjects were drawn from four large managed care health plans and were under the care of a primary care physician. To be enrolled, patients had to be at least 18 years of age, have heartburn, and have their physician consider them appropriate candidates for treatment of suspected GERD. Heartburn was defined as a rising, uncomfortable burning feeling in the chest behind the breastbone. Patients were randomly assigned to receive either omeprazole 20 mg once daily or ranitidine 150 mg twice daily. Six hundred eighty-five patients were enrolled in this study.
Clinical Trial 3
This clinical trial was a prospective, multi-center, open-label, randomized trial designed to evaluate the effectiveness of omeprazole versus ranitidine in the initial treatment of patients with GERD [
8,
19]. Patients were enrolled over a 24-month period from five family practice clinics in the Charleston, West Virginia area. All enrolled patients were at least 18 years of age, were clinically diagnosed as having GERD, and were determined to need pharmaceutical intervention in the treatment of GERD based on frequency of heartburn / acid regurgitation. Patients were randomly assigned to receive either omeprazole 20 mg once daily or ranitidine 150 mg twice daily. This study population had a higher level of chronic illness than the other two study populations with 61% of the patients reporting one or more medical co-morbidities. Secondary analyses included data collected at baseline, 4-, 12-, and 24-week follow-up from 251 patients.
The Psychological General Well-Being Index (PGWB) was used to evaluate health-related quality of life in each of the three clinical trials. The Medical Outcomes Study (MOS) Sleep Disturbance scale was also included in Clinical Trial 1 and Clinical Trial 3.
The PGWB measures psychological well-being and distress [
20,
21]. The instrument is composed of 22-items which, when scored, constitute 6 subscales and one total score with higher scores indicating better health status and psychological well-being. Subscale scores include anxiety, depression, self control, positive well-being, general health, and vitality. Normal values for the total score are considered to fall in the range of 100–105, with women generally reporting lower well-being than men [
6]. The PGWB has good evidence supporting internal consistency, test-retest reliability and validity, and has been shown to be sensitive to gastrointestinal disease occurrences [
6,
7,
9,
13,
21‐
23]. For the purposes of this study only the PGWB total score was examined.
The MOS Sleep Disturbance scale [
24], included in Clinical Trials 1 and 3, is composed of four items pertaining to sleep initiation and maintenance. The sleep scale ranges from 0 to 100 with higher scores indicating better sleep quality (i.e., less sleep disturbance). The MOS Sleep Disturbance score has been evaluated and demonstrates good reliability and validity [
24].
Clinical Measures
To evaluate the gastrointestinal symptoms of patients, Clinical Trial 1 used both diary cards and the Gastrointestinal Symptom Rating Scale (GSRS). Diary cards were completed by patients each day for a 1-week period prior to both the baseline and the 8-week HRQL assessment [
13]. Ninety-seven percent of patients provided complete diary data during the study. The GSRS is a 15-item, disease specific instrument used to evaluate common gastrointestinal symptoms [
23,
25]. The items are measured on a 7-point Likert-type scale ranging from 1 = 'no discomfort' to 7 = 'very severe discomfort'. The GSRS was administered at baseline and 8-week follow-up. To assess the gastrointestinal symptoms in Clinical Trial 2, patients were asked about how much of the time during the past 7 days they were bothered by each of three symptoms: heartburn, regurgitation, and difficulty swallowing. These three questions were asked at baseline and follow-up weeks 4, 8, 12, and 16 and were measured on a 6-point response scale: from 1 = 'none of the time' to 6 = 'all of the time' [
14]. Clinical Trial 3 used the GSRS to assess gastrointestinal symptoms of patients at baseline and follow-up weeks 4, 12, and 24.
Baseline Heartburn Severity
Baseline heartburn severity (i.e., none/minor, mild, moderate, severe) was defined based on their level of heartburn symptoms. For Clinical Trial 1 and Clinical Trial 3 baseline severity was determined using the GSRS heartburn item: Have you been bothered by HEARTBURN during the past week? (by heartburn we mean a burning pain or discomfort behind the breastbone in your chest). For Clinical Trial 2 severity was determined based on the patients' indicated frequency (i.e., 'none of the time' to 'all of the time') of bothersome heartburn during the 7 days prior to baseline assessment.
Complete Symptom Resolution
For Clinical Trial 1 diary data were used to define whether or not patients were experiencing complete resolution of heartburn symptoms. Using these data, complete resolution of heartburn symptoms was defined for persons who specified no episodes of heartburn during the 7 days prior to the follow-up visit [
15]. For Clinical Trial 2, patients indicating that they were bothered by heartburn 'none of the time' during the 7 days prior to the follow-up visit were classified as completely resolved [
15]. For Clinical Trial 3, completely resolved patients were identified using the GSRS heartburn item. Complete resolution status was determined at each follow-up period across each of the three studies [
15].
Statistical Analyses
PGWB total and MOS Sleep Disturbance scores were compared between patients classified as having none/minor, mild, moderate, or severe heartburn symptoms at start of treatment. Comparisons were conducted for each of the three clinical trials and were performed independent of treatment group assignment. To evaluate whether baseline heartburn severity is associated with endpoint HRQL scores, we focused on the interaction between baseline symptom severity and complete resolution status at follow-up, based on an analysis of covariance (ANCOVA) model. This ANCOVA model included the interaction term and main effects for baseline symptom severity and complete resolution. The covariates included in the models were age, gender, relevant baseline HRQL measure. A separate ANCOVA model was fit for each follow-up assessment for each of the three trials. Patients with missing data for any of the variables included in the models were dropped from the analyses. A second set of ANCOVAs evaluated mean PGWB total scores and Sleep Disturbance scores by symptom resolution status, controlling for baseline symptom severity, age, gender and baseline score. Standard errors (SE) for the covariate adjusted follow-up scores are reported. All statistical tests were two-sided with p-values ≥ 0.05 considered to be statistically significant. Analyses were performed using SAS System Software Package, Version 8.0.
Discussion
This secondary analysis was designed to evaluate the relationship between baseline level of heartburn symptom severity, resolution of heartburn symptoms, and HRQL outcomes assessed at various time-points during pharmaceutical treatment for GERD. This study failed to find evidence supporting the interaction between baseline symptom severity and complete resolution of heartburn symptoms on PGWB total or Sleep Disturbance scores. However, baseline symptom severity was significantly associated with psychological well-being, as measured by the PGWB, and sleep problems. Those patients reporting greater severity in heartburn symptoms were more likely to report psychological distress and impaired well-being compared with those who reported no or milder symptoms. Heartburn severity was also associated with greater reports of sleep problems in these GERD patients.
We evaluated the relationship between complete resolution of heartburn symptoms and PGWB and Sleep Disturbance scores, after controlling for baseline levels of symptom severity. Baseline heartburn severity was significant in only one model for the PGWB and one model for Sleep Disturbance scores, both for Clinical Trial 3 at the 12 week follow-up. For most of the analyses, baseline symptom severity was not significantly associated with endpoint health outcomes. In general, complete resolution of heartburn symptoms resulted in reports of improved psychological well-being. Those patients reporting complete relief from their heartburn symptoms also reported better psychological well-being. The observed differences in mean PGWB total scores are clinically meaningful, since differences or changes of 4 to 5 points represent the minimally important difference for the PGWB [
22,
26]. Using a 4-point difference as the criteria for minimal important, 5 out of the 8 comparisons (63%) are clinically meaningful. No association between symptom resolution and Sleep Disturbance scores was observed in this secondary analysis. These results extend previously published results that demonstrated complete resolution of heartburn symptoms is associated with improved patient functioning and well-being [
15]. Even after adjusting for baseline symptom severity, complete resolution of heartburn symptoms remains associated with improved psychological well-being. These findings were consistent across all three studies.
Other published research has reported an association between GERD-related gastrointestinal symptoms (e.g., heartburn, epigastric pain, regurgitation) and patient functioning and well-being [
5‐
7,
10,
23]. An earlier study has shown that the complete resolution of these symptoms results in greater patient functioning and well-being compared to continued symptoms and to an overall improvement in functioning and well-being over the course of medical treatment [
15]. While these analyses did not demonstrate significant differences in HRQL by baseline severity of heartburn symptoms, they did provide further support for the association between complete resolution of these symptoms and improvement in HRQL.
Recently published work by Farup, et al. reported that measures of nocturnal GERD, such as discomfort with nocturnal GERD symptoms, frustration with sleep loss, and worry and concern about symptoms were associated with impaired HRQL [
27]. Earlier work by Revicki et al. demonstrated a negative and statistically significant correlation between both the continuous measures of number and severity of heartburn episodes and the sleep disturbance scores [
13]. Others have reported that there are no statistically significant associations between measures of symptom severity and the sleep disturbance scores [
8]. Analyses presented in this report failed to demonstrate that baseline severity of heartburn symptoms had an effect on measures of sleep disturbance during the course of medical treatment of GERD.
Limitations
Interpretation of the findings of this study should be moderated by several limitations. One limitation of the study is that the information used to generate the two primary study variables (i.e., baseline symptom severity, complete resolution of symptoms) was collected through patient self-report and that these self-reported outcomes were measured differently for each of the three studies. The patient diary reports used in Clinical Trial 1 collected information on symptom frequency and severity. Clinical Trials 2 and 3 collected information on discomfort or frequency of symptoms. Clinician assessments, patient diary data, and patient self-report data have been observed to be in general agreement when using the GSRS [
23]. The patient reports of heartburn symptoms are believed to be reliable and valid.
A second limitation is the means by which the baseline heartburn severity measure was defined across all three clinical trials. The severity measure was defined based on physician consultation following an examination of the distribution of responses to heartburn symptom items for each of the three studies. It is uncertain as to the extent that modifications in this measure would impact the results. However, examination of mean HRQL scores at baseline has demonstrated a clear association between symptom severity and HRQL scores.
In conclusion, this secondary analysis of clinical and HRQL data from three clinical trials demonstrated that complete resolution of heartburn symptoms is associated with improved psychological well-being in patients with GERD. No significant association was seen between complete resolution of heartburn symptoms and sleep problems. Adjusting for baseline heartburn severity did not affect the relationship between symptom resolution and psychological well-being. In general, the severity of heartburn symptoms at the start of medical treatment for GERD does not modify the association between complete symptom resolution and follow-up HRQL outcomes. Complete resolution of heartburn symptoms is associated with improvements in overall psychological well-being, but not sleep disturbance outcomes. Medical and surgical treatment of GERD is focused on relief and complete resolution of heartburn, and other, symptoms and in improving patient HRQL. These findings confirm that relief of GERD-related symptoms is associated with significant improvements in patient-reported psychological well-being.