The low-flow method has been shown as a reliable evaluation of access recirculation. Few data is available on temporary central catheter blood recirculation; results of 2% and 4% have been reported in subclavian, 10% in 24 cm long femoral, and 18% in 15 cm long femoral catheters, mostly in indwelling catheters for chronic hemodialysis. The purpose of this prospective study was to evaluate blood recirculation in a larger number of recently inserted temporary intravenous catheters for acute hemodialysis, comparing subclavian and femoral sites. Fifty blood recirculation measurements were performed in 38 different temporary central venous dialysis catheters inserted in thirty-one critically ill patients from medical and surgical intensive care units presenting acute renal failure supported by intermittent hemodialysis. All the catheters used were well-functioning 11.5 French dual lumen Quinton of 13.5 or 19.5 cm length. Catheters presenting mechanical dysfunction, which did not allow a blood flow rate of 300 ml/min or for which lines had to be reversed were excluded from the analysis. Access blood recirculation was measured shortly after catheter insertion according to the low flow method applied after the first 30 minutes of hemodialysis at a blood flow rate of 300 ml/min. Mean blood recirculation for the 50 measurements was 10.3 +/- 9.2%. It was significantly higher in the 26 femoral catheters than in the 24 subclavian catheters, reaching respective means of 16.1 +/- 9.1% and 4.1 +/- 3.6% (p = 0.0001). Blood recirculation rate was not different between 13.5 cm and 19.5 cm long subclavian catheters (3.0 +/- 2.6%, n = 13, versus 5.4 +/- 4.3%, n = 11, respectively), but was significantly higher in 13.5 cm long femoral catheters (22.8 +/- 9.1%, n = 9, versus 12.6 +/- 6.9%, n = 17) (p = 0.004). Blood recirculation was measured on two separate occasions in 12 catheters randomly selected (5 femoral and 7 subclavian catheters); the obtained results were reproducible with a mean difference of only 2.1 +/- 1.8% between the two measurements and a correlation of 0.96. The mean time elapsed between catheter insertion and recirculation assessment was 2.2 +/- 3.1 days and was similar for femoral and subclavian catheters. No correlation was found between the percentage of recirculation and the arterial and venous resistances recorded during dialysis session or with the time from catheter insertion. Mean urea reduction ratio (URR) for the 50 dialysis sessions was 57.8 +/- 13.0%. It was significantly higher for sessions performed with subclavian than with femoral catheters (62.5 +/- 10.9%, n = 24, versus 54.5 +/- 14.2%, n = 26) (p = 0.03). In conclusion, the expected blood recirculation in well-functioning and recently inserted temporary dialysis catheters is under 5% for subclavian, over 12% in 19.5 cm femoral, and over 22% in shorter 13.5 cm femoral catheters at a blood flow rate of 300 ml/min. The consequently reduced dialysis efficiency with femoral catheters is another factor to be considered in the choice of a site for temporary dialysis catheter insertion in acute renal failure patients, particularly when dialysis dose delivery is a priority, such as intoxication cases treated by extracorporeal therapy.