We evaluated the effects of adding clonidine for continuous peripheral nerve infusions. Sixty patients undergoing total knee arthroplasty under combined single-injection sciatic block and continuous femoral infusion were randomly allocated to three groups: block induction with 0.75% ropivacaine followed by 0.2% ropivacaine (group control; n = 20); block induction with 0.75% ropivacaine and 1 microg/kg clonidine followed by 0.2% ropivacaine (group cloni-bolus; n = 20), and block induction with 0.75% ropivacaine and 1 microg/kg clonidine followed by 0.2% ropivacaine with 1 microg/mL clonidine (group cloni-infusion; n = 20). After surgery, continuous femoral infusion was provided with a patient-controlled infusion pump (basal infusion rate, 6 mL/h; incremental dose, 2 mL; lockout time, 15 min). The median (range) onset time of surgical block was 15 min (5-30 min) in group control, 10 min (5-35 min) in group cloni-bolus, and 10 min (5-30 min) in group cloni-infusion (P = 0.07). No differences were reported among groups in the degree of pain measured with the visual analog scale. The total consumption of local anesthetic solution after a 24-h infusion was 170 mL (144-220 mL) in group control, 169 mL (144-260 mL) in group cloni-bolus, and 164 mL (144-248 mL) in group cloni-infusion (P = 0.51); after the second day of infusion, total consumption was 168 mL (144-200 mL) in group control, 156 mL (144-288 mL) in group cloni-bolus, and 150 mL (144-210 mL) in group cloni-infusion (P = 0.48). Hemodynamic profiles and sedation were similar in the three groups. Motor function impairment after 48 h of infusion was observed in 27% of cloni-infusion patients but in only 6% of both the control and cloni-bolus groups (P = 0.05). We conclude that adding clonidine 1 microg/mL to local anesthetic for continuous femoral nerve block does not improve the quality of pain relief but has the potential for delaying recovery of motor function.