Treatment of chronic gastro-oesophageal reflux disease

Symptoms of gastro-oesophageal reflux are common—about 25% of the community experience heartburn and 5% have symptoms daily.1 Proton pump inhibitors are the mainstay of treatment for moderate to severe symptoms because randomised controlled trials (RCTs) show they are the most effective medical treatment.2 When these drugs are discontinued symptoms often relapse, so most patients take them long term.3 This is inconvenient for patients and expensive for the health service. Concerns also exist over the long term safety of this treatment—continued acid suppression is associated with increased risk of infective diarrhoea and hip fracture,4 5 although whether these associations are causal is unclear.

The alternative to long term treatment with proton pump inhibitors is surgery. In the linked study (doi:10.1136/bmj.b2576),6 Epstein and colleagues report the results of a cost effectiveness analysis of a large RCT by Grant and colleagues comparing laparoscopic fundoplication with proton pump inhibitors.7 The advent of this type of surgery has reduced the morbidity and cost of surgery and made it a viable alternative to drugs.8 Three earlier RCTs compared laparoscopic antireflux surgery with proton pump inhibitors in a total of 875 patients with gastro-oesophageal reflux, with conflicting results.9 10 11 Two RCTs reported that surgery was more effective than proton pump inhibitors in relieving reflux symptoms,9 10 whereas the largest trial indicated that they had equal efficacy.10

Grant and colleagues’ pragmatic trial conducted in secondary and tertiary care recruited more than 800 subjects with longer than a 12 month history of gastro-oesophageal reflux disease, confirmed by endoscopy or 24 hour pH monitoring.7 Almost half were randomised to receive laparoscopic fundoplication or optimal medical treatment. The primary end point was improvement in the patient reported reflux quality of life score, and at first glance the data seem compelling, with significant improvements in scores in those assigned to surgery, despite more severe symptom scores at baseline, and a reduction in subsequent use of acid suppressants. Perioperative morbidity and conversion rates to open procedure were low, and later complications of surgery were rare. These data suggest that people with gastro-oesophageal reflux disease who need long term acid suppression would benefit from laparoscopic fundoplication.

However, it may be premature to incorporate the results into clinical practice. Although randomisation will have minimised the possibility of confounding factors influencing the results, the nature of the surgical intervention precludes the trial from being double blind. Participants could therefore be biased in their assessment of symptoms, and those who had surgery may have, consciously or unconsciously, exaggerated their symptoms at baseline and minimised symptoms during follow-up.

Interestingly, half way through the trial the data monitoring committee requested that participants be informed of the group that they were randomised to after completing the baseline questionnaire because they were concerned that “this could potentially affect their response.” This change would have reduced possible bias at baseline, but not during follow-up. Of course, all trials of surgery for gastro-oesophageal reflux will have this flaw, and this is not a criticism of the investigators, but it means that results based on subjective symptom assessments need to be interpreted with caution. Trials of laparoscopic antireflux surgery understandably focus on subsequent improvement in reflux symptoms, but the procedure can have non-reflux related complications, such as dysphagia and bloating. Current evidence suggests these adverse sequelae are uncommon,7 8 9 10 11 but an evaluation of patient reported overall quality of life would be the most holistic assessment of the effect of surgery. Grant and colleagues measured this using the EQ-5D and found no significant difference between the two groups, which is consistent with another trial.10

Finally, an important consideration is whether surgery is cost effective. Epstein and colleagues extrapolated data from Grant and colleagues’ trial over a lifetime using a Markov model.6 Their results suggested that surgery was likely to be cost effective, but as with most economic modelling studies this was dependent on key assumptions. In particular, the base case model assumed that gains in quality adjusted life years would persist for the lifetime of the individual, and that the cost of proton pump inhibitors would stay the same. If both of these assumptions were inaccurate the authors concluded that it would be much more uncertain whether surgery or medical management was more cost effective. But as proton pump inhibitors become generic their price will probably fall dramatically, and RCT data on the long term benefits of surgery are sparse. We therefore need more data on long term benefits before we can be confident that surgery is cost effective. This is consistent with the modelling of another UK trial, which suggested that patients will need to be followed up for eight years before surgery becomes cost effective,12 and this duration may increase as generic proton pump inhibitors become cheaper.

Laparoscopic antireflux surgery is not for everyone. The fact that only 62% of patients randomised to surgery in this trial underwent the procedure shows that patients need to be selected carefully. Nevertheless, surgery is an important option to consider in patients taking long term proton pump inhibitors. Pragmatic RCTs of this type are difficult to conduct, as the failure to meet the power calculation in the present trial shows. However, initial results are encouraging, and more information will emerge with further follow-up.