Hemodialysis (HD) prescription is usually based on the periodical measurement of the Kt/V achieved in a midweek dialysis session. The purpose of the study was to assess the duration of each HD session to achieve a target dose of dialysis. This allowed to determine whether a given dialysis session may be considered representative of the other HD sessions. Seventy-two HD sessions were studied in 4 stable patients, who were randomly dialyzed during 3 consecutive periods, each lasting 2 weeks, using a different blood flow rate (Qb) in each period: 400, 300 or 200 ml/min. All HD were prolonged to achieve an on-line dialysate urea monitor (UM) Kt/V of 1.2. The UM Kt/V was compared with the Kt/V calculated using pre-HD, post-HD and rebound (45 min post-HD) plasma water urea concentrations. Comparison of the duration of the second midweek dialysis session with the length of the other HD showed 95% concordance intervals (±2 SD) of ±21.08 min for Qb 400, ±26.88 min for Qb 300 and ±37.02 min for Qb 200 ml/min. The 95% concordance intervals for whole body urea clearance were ±32.0, ±20.36 and ±15.62 ml/min for Qb 400, 300 and 200 ml/min, respectively. No differences were observed between UM Kt/V and blood-based double-pool Kt/V obtained by the second-generation Daugirdas (1.18 ± 0.08) and Garred (1.19 ± 0.08) Kt/V formulas. In conclusion, a great variability was observed between different HD sessions with regard to the whole body urea clearance and the time required to attain a target Kt/V even when the HD characteristics remained constant. The length of every HD required to achieve a target dose of dialysis can be assessed by on-line dialysate urea monitoring.