BACKGROUND: Nerve blocks provide analgesia after surgery. The authors tested whether nerve blocks have antiinflammatory effects. METHODS: Patients had combined sciatic (single-shot) and continuous femoral block (48 h) (block group) or morphine patient-controlled analgesia after total knee arthroplasty. Pain at rest and upon movement was monitored at 1 (D1), 4 (D4), and 7 days (D7) and 1 (M1) and 3 months (M3) after surgery. Knee inflammation was evaluated (skin temperature, knee circumference) before surgery and at D1, D4, D7, M1, and M3. Plasma cytokine concentrations (interleukin [IL]-6, IL-1beta, tumor necrosis factor [TNF], IL-10, soluble receptor 1 of TNF [sTNF-R1]) were measured before surgery and at 4 h, D1, D4, and D7 after surgery. Capsule and synovial membrane cytokines were measured (IL-6, TNF, IL-1, IL-10). Knee flexion was evaluated before surgery and at D1, D4, D7, M1, and M3. Morphine use and recovery time to autonomy were monitored. RESULTS: Pain at rest and upon movement was lower in the block group than in patient-controlled analgesia patients between D1 and D7 (analysis of variance, P < 0.005). Knee flexion was improved in the block group for D1 to M1 (analysis of variance, P < 0.0001). Block group patients recovered nonassisted mobilization (t test, P = 0.04) and toilet use (t test, P = 0.03) more rapidly. Knee circumference and skin temperature were lower in the block group between D1 and D7 (analysis of variance, P < 0.05). Synovial membrane IL-1 (P < 0.05) and IL-10 (P < 0.01) increased, and plasma IL-6 and sTNF-R1 peaked at 24 h, with no difference between groups. CONCLUSION: Nerve blocks inhibited clinical inflammation after total knee arthroplasty, with no change in tissue and plasma cytokine concentrations.