Surgery in the management of gastrooesophageal reflux disease (GORD) is controversial; its availability and frequency show variability between, and even within, countries. GORD is a common chronic condition, with some 20 per cent of the population in Western countries experiencing symptoms of reflux at least once a week1. When gastric juice flows back into the oesophagus, it causes heartburn and acid regurgitation (so-called ‘typical’ symptoms) and/or a variety of symptoms often referred to as ‘atypical’. Atypical symptoms include cough, hoarseness, dental erosions and angina-like chest pain.The regurgitated gastric contents may damage the oesophageal mucosa, inducing erosions, ulcers and even metaplastic change (Barrett’s oesophagus). Most patients with GORD are treated effectively by non-prescription medication or acid-suppressing drugs (proton pump inhibitors (PPIs)), or a combination of the two. Although effective2, surgery has a relatively marginal role. It is generally reserved for those few patients who are unwilling to take drugs or whose condition is considered ‘refractory’ to such treatments (patients whose symptoms or mucosal damage persist or are cured only partially by PPIs)3. When faced with a patientwith refractory disease it falls to the surgeon to decide whether or not an antireflux procedure will confer benefit. As will be shown below, the present difficulty in defining the role of surgery in the management of refractory GORD stems partly from the currently accepted Montreal definition of GORD as ‘a condition that develops when the reflux of stomach contents causes troublesome symptoms and/or syndromes’4. According to this symptom-focused definition, all patients with typical or atypical symptoms related to acid or nonacid reflux have the disease. These patients may be prescribed a trial of medical therapy, even if their overall oesophageal exposure to gastric contents is normal. Four groups of refractory patients who might request surgery can be identified. The first group consists of poorly compliant patients, or those having an inappropriate PPI consumption because of poor timing and/or frequency of dosing. This is probably the largest group and these patients are the simplest to treat, either by correcting themedical treatment (enforcing compliance, modifying the timing and/or frequency of dosage) or, if necessary, by antireflux surgery. The second group contains patients whose refractory symptoms are different to their initial symptoms. PPIs often resolve the severe heartburn felt in the initial stages of GORD but fail to reduce volume reflux and regurgitation while stooping or straining. Cough (especially at night), episodes of wakening and choking, and poor sleep pattern are common refractory symptoms that prompt a patient to seek surgical intervention. The third group of patients have symptoms that are unresponsive to PPIs. Themain reason is a persistence of alkaline or weakly acidic reflux of duodenal contents (bile and pancreatic juice). In this situation PPIs have reduced gastric acid output, but are ineffective with respect to the alkaline component of the refluxate. The final group comprises those whose refractory symptoms are due to an altered sensitivity of the oesophageal mucosa. This results in an abnormal perception of otherwise normal events, often described as ‘acid-sensitive oesophagus’ or ‘functional heartburn’. Although no exact figure is available, it has been estimated that these last two groups account for 10–20 per cent of refractory GORD. Without refractory symptoms, there is little evidence to promote antireflux surgery, except for Barrett’s oesophagus with severe reflux requiring progressively increasing doses of PPI, or when there is an associated large paraoesophageal hernia (to prevent future strangulation or respiratory complications). For a surgeon to operate, something more than a purely symptombased definition ofGORDis required; a trial of treatment by operation is not acceptable if the chance of successful outcome is minimal. An accurate definition of disease before surgery requires the measurement of acid reflux (by 24-h pH monitoring after suspension of PPI therapy) and exclusion of dysmotility disorders (achalasia, diffuse oesophageal spasm). If acid exposure is abnormal, GORD may be regarded as present and surgery as justified, if symptoms warrant it. This diagnostic approach, however, does not deal satisfactorily with the last two of the four groups described above. In