With an increasing number of patients on maintenance haemodialysis treatment, providing satisfactory, long-lasting vascular access becomes critically important. There are a large number of reports of the relative success rates of various forms of vascular access, but there is general agreement that the Brescia-Cimino wrist fistula ( I ) is the most satisfactory. For the patient whose wrist fistula has failed, the surgeon has a choice of reconstructing the fistula when possible, constructing another fistula in the contralateral wrist, a brachial fistula (2) or inserting a graft of biological or synthetic origin. Grafts can be autologous (3) or allogeneic vein (4), umbilical vein ( 5 ) , bovine carotid artery (6) or synthetics such as polytetrafluoroethylene (PTFE) (7) and Dacron (8). Experience has shown that long term results are poor for vein loops (9, 10) and bovine heterografts ( 1 I), although with aggressive treatment of the frequent complications of bovine grafts, it is possible to achieve a 2-year patency rate of 77 per cent (12). Umbilical vein is relatively untried, though a I-year patency rate of 50 per cent has been reported (13). Patency rates for PTFE vary from 62 per cent at I year (14) to 83 per cent (15). Reports of long term patency for PTFE are harder to find; a 2-year patency rate of 76 per cent has been reported (10) and a 3-year patency rate of 63 per cent (10).The difficulty in drawing any conclusions from these studies is that the selection criteria are unknown-grafts have a greater patency rate if inserted as a primary procedure rather than in patients whose access sites are nearing exhaustion. The first aim of the present study was to examine the need for more complex procedures than simple arteriovenous fistulas in a population without previous surgery. Thus, all patients undergoing firsttime fistula operations have been studied, rather than all fistulas, regardless of whether they were first, second or subsequent procedures. The second aim of the study was to attempt to correlate fistula failure with possible aetiological factors, in the hope of predicting those likely to have subsequent thrombotic problems.
Patients and methodsBetween January 1976 and May 1981 157 consecutive patients in chronic renal failure requiring access procedures were accepted on to the dialysis programme. Six patients were excluded from analysis because of previous fistula operations elsewhere. One hundred and forty-eight patients had a first-time wrist fistula constructed, 2 patients had an exploration of the wrist revealing no suitable vessels and 1 patient had a primary brachial fistula due to bilateral cephalic cutdowns in the past. Five patients with functioning fistulas were excluded due to inadequate data; 3 were transferred to other units and 2 died. Using actuarial analysis (16). fistula patency curves were calculated for the remaining 145 patients undergoing an attempt at a first-time wrist fistula. This included the two patients undergoing an exploration of the wrist to avoid biasing the results favo...
