The Effects of PPIs on Patient-reported Severity of Dyspepsia
The Effects of PPIs on Patient-reported Severity of Dyspepsia
COGENT was an international, randomised, double-blind, double-dummy, placebo-controlled, parallel-group, phase 3 study of the efficacy and safety of CGT-2168, a fixed-dose combination of clopidogrel (75 mg) and omeprazole (20 mg), as compared with clopidogrel alone, administered on background aspirin therapy. Randomisation was stratified according to baseline H. pylori serology status and the concomitant use of nonsteroidal anti-inflammatory drugs (NSAIDs). The study methodology has been previously described in detail.
Patients aged 21 years or older on clopidogrel therapy with concomitant aspirin anticipated for at least the next 12 months were eligible for inclusion. Major exclusion criteria included delayed hospital discharge, clinical indication for use of antacids; pre-existing erosive oesophagitis or oesophageal or gastricvariceal disease; previous non-endoscopic gastricsurgery; receipt of clopidogrel or another thienopyridine for more than 21 days before randomisation; receipt of oral anti-coagulation, or recent fibrinolysis.
The primary gastrointestinal efficacy endpoint in the COGENT study was the time from randomisation to the first occurrence of a composite of upper gastrointestinal clinical events: overt bleeding of gastroduodenal origin (confirmed by means of upper endoscopy or radiography), overt upper gastrointestinal bleeding of unknown origin, bleeding of pre-sumed occult gastrointestinal origin with a documented decrease in hemoglobin of 2 g per decilitre or more or in the hematocrit by 10% or more from the baseline value, symptomatic uncomplicated gastroduodenal ulcer (confirmed by means of endoscopy or radiography), persistent pain of presumed gastrointestinal origin with a duration of 3 days or more and with five or more gastroduodenal erosions (confirmed by means of endoscopy), obstruction or perforation. The pre-specified primary cardiovascular safety endpoint was the composite of death from cardiovascular causes, nonfatal myocardial infarction, coronary revascularisation, or ischemic stroke.
We assessed the SODA scores at 4 and 24 weeks and the change in SODA scores from baseline to 4 weeks (early response) and from baseline to 24 weeks (late response). The SODA assesses symptoms in three dimensions: pain intensity (six items that ask patients to rate their abdominal discomfort over the past 7 days; range 2–47), nonpain related symptoms (seven items that ask patients on common symptoms of dyspepsia in the last 7 days; range 7–35) and satisfaction with dyspepsia-related health (four items that assess the degree to which patients are satisfied with their levels of abdominal discomfort; range 2–23).
To assess the impact on clinical dyspepsia, we primarily used a definition of an increase greater or equal to 10 points on the pain intensity component of the SODA instrument from baseline to 4 or 24 weeks. To test the robustness of this clinical assumption, we assessed the impact of using different thresholds for the definition of dyspepsia on the outcome.
All randomised subjects were included in this analysis. Baseline characteristics and baseline SODA scores were compared between the two treatment groups using χ tests for discrete variables and t-tests for continuous variables.
SODA scores and change from baseline in SODA scores at the two follow-up time-points were compared between the two treatment groups using a mixed effects linear model. In view of observed trend towards lower pain intensity SODA scores in the omeprazole group at baseline, adjustment of follow-up SODA scores to baseline SODA scores was performed to account for the impact of differences in baseline SODA scores on follow-up scores. Baseline scores, treatment, time (operationalised as week 4 and week 24) and their interaction were included in the model as fixed effects. A random effect was included for the intercept. Treatment effect at week 4 and week 24 was estimated using estimate statements.
The potential modifying effect of baseline H. pylori status was assessed at 4 and 24 weeks visits by estimating the treatment for each baseline H. pylori status using a random intercept mixed effects model with time, treatment, baseline H. pylori status and all interactions among these predictors as fixed effects.
Sensitivity analyses were conducted to assess the robustness of our results to missingness. The vast majority of missing data was due to early termination of the trial with only a very small proportion due to other reasons. The sensitivity analyses were carried out by utilising the last observation carried forward (LOCF) approach and multiple imputations. With multiple imputations, each missing value in the response was replaced with a set of 10 imputed values. Multiple imputations were carried out separately for each time point (4 and 24 weeks), with baseline characteristics (listed in Table 1) included in each model.
We evaluated predictors for dyspepsia by using a logistic regression model and for the composite GI endpoint by using a Cox proportional regression model. SODA scores at baseline and other patient baseline characteristics (listed in Table 1) were selected in a step-wise approach using entry and stay levels of 0.2.
All statistical analyses were performed with sas version 9.1.3 (SAS Institute Inc., Cary, NC, USA).
Methods
COGENT was an international, randomised, double-blind, double-dummy, placebo-controlled, parallel-group, phase 3 study of the efficacy and safety of CGT-2168, a fixed-dose combination of clopidogrel (75 mg) and omeprazole (20 mg), as compared with clopidogrel alone, administered on background aspirin therapy. Randomisation was stratified according to baseline H. pylori serology status and the concomitant use of nonsteroidal anti-inflammatory drugs (NSAIDs). The study methodology has been previously described in detail.
Patients
Patients aged 21 years or older on clopidogrel therapy with concomitant aspirin anticipated for at least the next 12 months were eligible for inclusion. Major exclusion criteria included delayed hospital discharge, clinical indication for use of antacids; pre-existing erosive oesophagitis or oesophageal or gastricvariceal disease; previous non-endoscopic gastricsurgery; receipt of clopidogrel or another thienopyridine for more than 21 days before randomisation; receipt of oral anti-coagulation, or recent fibrinolysis.
Endpoints
The primary gastrointestinal efficacy endpoint in the COGENT study was the time from randomisation to the first occurrence of a composite of upper gastrointestinal clinical events: overt bleeding of gastroduodenal origin (confirmed by means of upper endoscopy or radiography), overt upper gastrointestinal bleeding of unknown origin, bleeding of pre-sumed occult gastrointestinal origin with a documented decrease in hemoglobin of 2 g per decilitre or more or in the hematocrit by 10% or more from the baseline value, symptomatic uncomplicated gastroduodenal ulcer (confirmed by means of endoscopy or radiography), persistent pain of presumed gastrointestinal origin with a duration of 3 days or more and with five or more gastroduodenal erosions (confirmed by means of endoscopy), obstruction or perforation. The pre-specified primary cardiovascular safety endpoint was the composite of death from cardiovascular causes, nonfatal myocardial infarction, coronary revascularisation, or ischemic stroke.
We assessed the SODA scores at 4 and 24 weeks and the change in SODA scores from baseline to 4 weeks (early response) and from baseline to 24 weeks (late response). The SODA assesses symptoms in three dimensions: pain intensity (six items that ask patients to rate their abdominal discomfort over the past 7 days; range 2–47), nonpain related symptoms (seven items that ask patients on common symptoms of dyspepsia in the last 7 days; range 7–35) and satisfaction with dyspepsia-related health (four items that assess the degree to which patients are satisfied with their levels of abdominal discomfort; range 2–23).
To assess the impact on clinical dyspepsia, we primarily used a definition of an increase greater or equal to 10 points on the pain intensity component of the SODA instrument from baseline to 4 or 24 weeks. To test the robustness of this clinical assumption, we assessed the impact of using different thresholds for the definition of dyspepsia on the outcome.
Statistical Analysis
All randomised subjects were included in this analysis. Baseline characteristics and baseline SODA scores were compared between the two treatment groups using χ tests for discrete variables and t-tests for continuous variables.
SODA scores and change from baseline in SODA scores at the two follow-up time-points were compared between the two treatment groups using a mixed effects linear model. In view of observed trend towards lower pain intensity SODA scores in the omeprazole group at baseline, adjustment of follow-up SODA scores to baseline SODA scores was performed to account for the impact of differences in baseline SODA scores on follow-up scores. Baseline scores, treatment, time (operationalised as week 4 and week 24) and their interaction were included in the model as fixed effects. A random effect was included for the intercept. Treatment effect at week 4 and week 24 was estimated using estimate statements.
The potential modifying effect of baseline H. pylori status was assessed at 4 and 24 weeks visits by estimating the treatment for each baseline H. pylori status using a random intercept mixed effects model with time, treatment, baseline H. pylori status and all interactions among these predictors as fixed effects.
Sensitivity analyses were conducted to assess the robustness of our results to missingness. The vast majority of missing data was due to early termination of the trial with only a very small proportion due to other reasons. The sensitivity analyses were carried out by utilising the last observation carried forward (LOCF) approach and multiple imputations. With multiple imputations, each missing value in the response was replaced with a set of 10 imputed values. Multiple imputations were carried out separately for each time point (4 and 24 weeks), with baseline characteristics (listed in Table 1) included in each model.
We evaluated predictors for dyspepsia by using a logistic regression model and for the composite GI endpoint by using a Cox proportional regression model. SODA scores at baseline and other patient baseline characteristics (listed in Table 1) were selected in a step-wise approach using entry and stay levels of 0.2.
All statistical analyses were performed with sas version 9.1.3 (SAS Institute Inc., Cary, NC, USA).