We have retrospectively analysed the usefulness of a subcutaneous tunnel in patients undergoing haemodialysis through a double lumen subclavian catheter; 194 catheters were used in 145 patients. In 105 patients a subcutaneous tunnel was created. Entry site infection was significantly higher in the no tunnel group (30%; P < 0.02). There was no statistically significant difference in the incidence of septicaemia in the two groups. Eighty five percent of episodes of septicaemia and 67% of entry site infections were due to Staphylococcus aureus. All responded to removal of the catheter and to antibiotics. The average duration the catheter was in place was 20.39 and 21.94 days in the groups with and without tunnel respectively. The average number of dialyses was 9.13 and 9.33 per catheter in the tunnel and no tunnel groups respectively. Three patients had pneumothorax. Subclavian vein thrombosis was suspected clinically in 3 cases. There was no catheter related mortality. We concluded that while entry site infection occurred more frequently in the no tunnel group, the overall incidence of septicaemia was not different in the two groups. Creation of a subcutaneous tunnel has no added advantage.