Gastro-oespohageal reflux is common and has been the subject of numerous clinical trials of both medical and surgical treatments. However, few of these trials have directly compared surgery with non-surgical treatment. Before the recently published REFLUX trial,¹ only four studies had compared surgery with medical treatment. The REFLUX trial originally recruited 810 participants who had had gastro-oesophageal reflux disease for longer than 12 months at baseline. Self-reported quality of life (QoL) was better at five years on several measures in patients who underwent surgery than in those who received drugs, whereas complications and reoperations after surgery were few.

Patients and surgeons often have strong preferences for surgery or non-surgical treatment that are based on values and perceptions of risk rather than evidence, which makes randomisation a challenge. The inclusion of parallel cohort studies of patients who refused randomisation was therefore a great strength of the REFLUX study. The data from these cohorts reinforced the message that surgery provides more effective long-term relief from symptoms of reflux, and the worse the symptoms, the greater the improvement.

The study also dealt well with some other concerns that typically cause problems in randomised trials of surgical techniques. To ensure optimisation of the surgical intervention, participating surgeons had to have performed 50 anti-reflux procedures before trial entry. The study lacked any assessment of surgical quality, such as manometry or operative videos, but the low rate of reoperation for complications or recurrence suggests that quality was at least as good as in other trials of this operation.

The study reports a low rate of serious complications. Complications of surgery were the patients’ biggest fear and a previous (LOTUS) trial reported a high rate of long-term post-operative morbidity, which impacted on QoL. If this had been reproduced in the long-term outcomes of the REFLUX trial it might have eliminated the benefits found for the surgical strategy. Why long-term morbidity was so low in REFLUX is a matter for informed speculation. The double opportunity for self (and clinician) selection provided by the REFLUX protocol may have excluded some patients prone to side-effects; the choice of centres and high experience bar may have ensured more expert surgery; or, as the authors suggest, the fact that 50% of patients had partial wraps, compared with all having complete (360°) wraps in LOTUS, may have led to the reduced morbidity.

This trial makes it clear that surgery is better than drugs for relieving the symptoms of gastro-oesophageal reflux in the medium-to-long term, but only if the surgical complication rate is low. Patients are more likely to benefit if they are younger and fitter, and if they have worse reflux to start with. When given enough pause for thought, a third of patients allocated to surgery changed their minds, and they may have been right to do so. Perhaps one of the most important lessons is that, when designing trials, more attention needs to be paid to the difficulties encountered when trying to synthesise scientific evidence and patient values while choosing between an operation and a pill.

Reference

1. Grant AM, Cotton SC, Boachie C et al. Minimal access surgery compared with medical management for gastro-oesophageal reflux disease: five year follow-up of a randomised controlled trial