SJS) and the Austin Diagnostic Clinic, Austin, Texas (G.A.B.)Recognizing and treating the failed and failing vascular access is a subject of growing interest and concern among nephrologists. Even though a large percentage of vascular access problems are handled in the outpatient setting, access complications still account for 21% of all hospitalizations for chroinc dialysis patients and are a growing source of morbidity and cost (1-4). The increasing use of tunneled superior vena cava catheters (4, 5) and more recently, inferior vena cava catheters (6,7,8) as well as the attention being paid to construction of exotic tertiary access (9) all attest to the increasing difficulty of maintaining vascular access patency in an aging population. Lost of all vascular access sites, while still relatively rare, is a feared and devastating complication, particularly for the patient unable to receive peritoneal dialysis.The current thrust of investigative efforts in this area is aimed at prolonging the functional life of vascular accesses (10). Long term, there may be clinically important progress in areas such as the development of direct inhibitors of the stenotic process. Up to this point, however, the most useful clinical technique has been to use post-dialyzer venous drip chamber pressures (''venous pressure'') and recirculation measurements to screen patients for functionally significant stenotic lesions (11-15) which then can be treated by surgery or angioplasty to decrease the risk of clotting and loss of the access site (16).The measurement of recirculation using urea based methods has been fraught with difficulty, largely a result of problems in obtaining the ''systemic'' sample of blood necessary for making this measurement (13, 18). It is now well accepted that use of peripheral venous blood for the systemic sample overestimates recirculation substantially because the BUN from this site exceeds that in arterial blood as a result of arteriovenous disequilibrium (due to cardiopulmonary recirculation) (19-22) and venovenous disequilibrium (due to regional blood flow inequalities) (23-26). Two needle methods in which the systemic sample is drawn from the arterial line have the potential advantage of eliminating both of these causes of a falsely high BUN while also eliminating an unnecessary venipuncture.Recent advances in measuring recirculation using non urea based methods have shown that recirculation is absent (0%) in a properly cannulated, well functioning access (27-32). However, two needle methods have typically yielded results significantly greater than 0% in most patients. This is partially explained by the normal laboratory variability in urea measurement (33) but, probably of greater consequence, is the typical 30 second (or more) delay in sampling after slowing or stopping the blood pump (15). It now appears that the BUN in arterial line blood will begin to rebound within about 15 seconds of slowing or stopping dialysis, an interval shorter than previously appreciated (34). As a result, many of the two need...