OBJECTIVE To determine whether intrathecal baclofen is an effective adjunctive agent to decrease acute and chronic postoperative pain after total knee arthroplasty. METHODS Prospective, randomized, double-blind controlled trial. METHODS Operating room and inpatient units of a university hospital. METHODS 60 adult, ASA physical status I, II, and III patients presenting for total knee arthroplasty. METHODS Anesthesia was provided by spinal injection of 15 mg of 0.75% hyperbaric bupivacaine combined with either 100 mcg baclofen or saline. Sedation was provided with intravenous midazolam and propofol. METHODS Data were collected on adverse effects, opioid usage, and verbal pain scale (VPS) from 0 to 10. The study period was divided into six discrete time intervals that included the 1(st) 72-hour postoperative period and a three-month post-discharge follow-up telephone call. RESULTS The baclofen group used less morphine in the PACU than the control group (5 mg vs. 9.3 mg; P = 0.04). VPS were lower in the baclofen group than the treatment group, but significant differences could be demonstrated only in the time periods 48-72 hours and three months postoperatively. At three months, fewer patients in the baclofen group reported pain than the control group (8/27 vs. 19/29; P = 0.009). Regression analysis showed that the baclofen group was 4.5 times less likely to report pain at three months (95% CI: 1.5-16.6). CONCLUSIONS IT baclofen used as an adjuvant to spinal anesthesia for total knee arthroplasty allows for less postoperative opioid usage and less chronic pain at three months.