Ninety-two patients with severe gastro-oesophageal reflux submitted to fundoplication from 1979 to 1984 have been studied. Severity of symptoms pre- and postoperatively have been assessed using a symptom score. The mean pre-operative score was 5.39 out of 9. A standard procedure for the fundoplication was used, including a long (5 cm) wrap leaving the wrap in an intrathoracic position when it could not be brought completely into the abdomen. Vagotomy was added in 53 patients. Posterior gastropexy was used in 54 patients. There was a zero incidence of damage to the spleen and a zero mortality. The mean symptom score on follow up was 0.41 out of 9 with 90.5% patients having absent or minimal symptoms. However, only 68% remained satisfied with their overall results. The incidence of sequelae related to the procedure itself including gas bloat (19.6%), dumping (7.6%) diarrhoea (6.5%) and development of gastric ulcer (2.2%) explained this discrepancy. The addition of vagotomy did not improve the results but added its complications especially dumping and diarrhoea. There were no differences in clinical results whether the fundoplication had been left in the chest or in the abdomen but there were two hazardous complications of the intrathoracic fundoplication including a perforated gastric fundus and a gastric ulcer in the thoracic part of the stomach. Posterior gastropexy conferred no benefit to the results. Measures which might improve results include: avoidance of vagotomy, intrathoracic fundoplication and gastropexy; shortening the wrap; and the use of a 50-60 F dilator in the oesophagus during the wrap.