We would like to comment on the paper by Coughlanet al on the relationship between medical treatment for reflux oesophagitis and asthma control.1Gastro-oesophageal reflux (GOR) can cause dyspnoea in non-asthmatic patients with normal pulmonary function and bronchial reactivity that improves with antireflux therapy.2 ,3 Coughlanet al state that we included uncontrolled trials in our analysis.1 This is incorrect.

We identified 12 studies—three uncontrolled, one with an untreated control, and eight controlled.4 We felt, however, that these studies were not amenable to meta-analysis since outcomes varied, different classes and doses of antireflux medications were used, treatment periods ranged from 1 week to 6 months, different diagnostic criteria for GOR and asthma were used, asthma severity differed, and studies were done in different populations. We excluded the open studies and the paper with the untreated control group.5 In table 3 studies were categorised according to Sackett's criteria.6 In the abstract, materials and methods, figure legends, results, and discussion we clearly stated that the results of the controlled trials were analysed and presented.4 In addition to these eight controlled trials, Coughlan included one with an untreated control and three controlled trials published since our review.5 ,7-9 The small number of patients with GOR symptoms and its mild nature may explain the apparent lack of benefit reported by Boeree et al.7 The study by Levin et alonly included nine subjects.8 Kiljanderet al reported a trend in asthma symptom improvement that may have been significant had the study been properly powered.9

The effects of antireflux surgery on asthma have also been reported. Most studies were uncontrolled, did not document GOR or asthma objectively, and did not measure objective outcomes.10Both controlled studies reported that asthma symptoms, but not pulmonary function, improved, which is consistent with our hypothesis.11-13 An improvement in asthma symptoms was the most consistent change in both the medical and surgical antireflux therapy trials and may be an important clue to the nature of the relationship between GOR and asthma.4 ,10 We would caution clinicians not to dismiss GOR as an irritant in poorly controlled asthmatics, especially those with reflux associated respiratory symptoms. We agree with Coughlan that further properly controlled and powered studies are required to assess the effects of antireflux therapy on asthmatics with GOR.