This study attempted to define the minimal dose of dialysis needed to produce a target KT/V in continuous peritoneal dialysis (CPD). In a training set of 143 clearance studies performed in 92 CPD patients, logistic regression identified low urine volume (UV) and low dialysate drain volume normalized by body water (DV/V) as predictors of weekly KT/V urea < or = 1.70. Solution of the regression equation with UV fixed at 0.00 1/24 h and at different probabilities of low KT/V provided a series of minimal DV/V values consistent with weekly KT/V > or = 1.70 in anuria. The accuracy of the logistic regression model and of the DV/V cut-offs was tested in a validation set (VS) of 189 urea kinetic studies performed in another 102 CPD patients. In the VS, the area under the Receiver Operating Characteristic curve generated by the regression model was 0.832 (95% Confidence Interval: 0.798-0.866). The DV/V cut-off value of 0.301 per 24 h, calculated by solving the regression model at p = 0.442 and with UV = 0, identified studies with weekly KT/V < 1.70 with a sensitivity of 89.3% and a specificity of 78.1% in anuric VS subjects (n = 60). Use of only the first urea kinetic study from each patient did not modify the predictors of KT/V or the cut-off values derived from solution of the regression model. The DV/V cut-off of 0.324 per 24 h, derived from the logistic regression model predicting KT/V < or = 1.90, identified KT/V < 1.90 in VS anuric subjects with a sensitivity of 94.3% and a specificity of 81.0%. Low UV and DV/V predict low KT/V urea in CPD. Prescribed 24 h exchange volume in anuric CPD subjects should be calculated to produce DV/V values exceeding 0.301 1/24 h per 1 body water for a KT/V of 1.70 and 0.324 1/24 h per 1 body water for a target weekly KT/V of 1.90.