Vascular access recirculation rate (AR) monitoring is fundamental to guarantee treatment adequacy and to detect access failure early. We have evaluated the GIT2 test to measure AR unaffected by cardio-pulmonary recirculation (CPR), based on a short glucose infusion in place of the bolus and on a two-operator sampling, differently from the classical glucose infusion test (GIT). The GIT2 test is based on four steps: 1) basal (B) glucose arterial sample; 2) 10% glucose infusion for 1 min, by infusion pump at 600 ml/hr; (or 20% at 300 ml/hr); 3) simultaneous sampling at arterial (A) and venous (V) ports, after 35-40 sec from starting the infusion, taking care to avoid blood pump stop during the test; 4) AR=100*((A-B)/(V-B)). In vitro tests by dialysis on a 40 L tank containing a urea solution, with AR volumetrically simulated at 0, 5, 10, 20%, and in vivo comparison of GIT, GIT2 with stop-flow (SF) urea method. Our results have shown in vitro an almost perfect correspondence of SF urea method and a better reliability of GIT2 than GIT. The methylene-blue test has shown that a single color bolus in V reaches the A port after variable time, depending on blood flow and AR, while the continuous infusion determines a steady gradient after about 30". In vivo tests (n=24) show good correspondence between GIT2 (4.37 +/- 3.36) and SF (4.51 +/- 3.62), while GIT data (1.01 +/- 0.51) are significantly underestimated. In conclusion, our preliminary results have evidenced a good reliability of the new test, the continuous infusion causing a steady gradient in V and A that more precisely reflects the AR rate.