Effect of baseline symptom severity on patient-reported outcomes in gastroesophageal reflux disease

Patient-reported outcome instruments are becoming increasingly important tools in clinical trials. We aimed to investigate the effect of baseline symptom severity on: (i) a global rating of change using the Overall Treatment Effect questionnaire; (ii) the change from baseline for some commonly used questionnaires, namely the Gastrointestinal Symptom Rating Scale and the Quality of Life in Reflux and Dyspepsia questionnaire; and (iii) treatment differences assessed by Gastrointestinal Symptom Rating Scale in patients with symptoms of gastroesophageal reflux disease.

We used pooled data from two 4-week studies of patients with gastroesophageal reflux disease (but without endoscopically verified esophagitis) randomized either to esomeprazole (20 or 40 mg) or to omeprazole 20 mg. We also used data from a study of omeprazole (10 or 20 mg) or ranitidine 300 mg in gastroesophageal reflux disease patients with or without esophagitis. The analysis is based on descriptive statistics.

Little impact of baseline symptoms/problems exists on the global rating of change (Overall Treatment Effect questionnaire). For example, the percentage of patients experiencing a large improvement with acid-suppressive therapy according to Overall Treatment Effect ranged from 63.2% for those with mild reflux at baseline to 70.3% for those with severe reflux. Change from baseline in the Gastrointestinal Symptom Rating Scale reflux dimension and quality of life in reflux and dyspepsia sleep disturbance and food/drink problems dimensions were associated with baseline symptoms/problems, with the largest change seen for patients with severe symptoms/problems at baseline. Baseline symptom severity was also associated with a clear effect on treatment differences, which tended to be larger for patients with severe symptoms than for patients with mild baseline symptoms.

In patients with gastroesophageal reflux disease, baseline symptom/problem severity had little association with the global rating of change, whereas a clear association with change in symptoms/problems and on treatment differences was apparent. The magnitude of change increased with increasing baseline severity of symptoms/problems.