BACKGROUND: Intra-articular bupivacaine hydrochloride (HCl) infusion catheters and periarticular injections of liposomal bupivacaine are often used as postoperative local anesthetics. The purpose of this study was to compare the efficacies of these local anesthetics following total knee arthroplasty. METHODS: This study was a superiority trial with a randomized, controlled, double-blinded design. Patients were randomly assigned to either delivery of bupivacaine HCl by the ON-Q* Pain Relief System pump (n = 96) or by an injection of Exparel (liposomal bupivacaine) (n = 104). The primary outcome of this study was cumulative narcotic consumption on postoperative days 0 through 3. Narcotic consumption data were collected retrospectively from in-hospital records while patients were in the hospital. Following discharge, narcotic consumption data were gathered from patient surveys, as were secondary outcomes measures. RESULTS: We did not identify greater narcotic use in the ON-Q* group compared with the Exparel group (p = 0.641). The mean difference between the groups was 0.5 morphine equivalent (95% confidence interval [CI] = -1.7 to +2.8), with the ON-Q* group consuming an average 10.4 morphine equivalents (95% CI = 8.7 to 12.0) compared with 10.9 (95% CI = 9.3 to 12.5) in the Exparel group. There were no significant differences between groups with regard to any of the secondary measures of pain with the exception of pain while walking and pain with physical therapy (p = 0.019 and p = 0.010, respectively), both of which showed an approximately 1-point difference in favor of the ON-Q* group on a visual analog scale (VAS). There were also no differences in the postoperative side effects, including nausea, constipation, or vomiting, or in the rates of study-related complications, patient satisfaction, or length of hospital stay. CONCLUSIONS: Exparel did not have superior efficacy compared with the ON-Q* Pain Relief System as reflected by narcotic consumption, our primary outcome. There were small significant differences, in favor of the ON-Q* group, in 2 secondary measures of pain during activity, but these approximately 1-point VAS differences are unlikely to be clinically relevant. The choice of a local anesthetic modality should be based on a combination of safety, convenience, and cost considerations. LEVEL OF EVIDENCE: Therapeutic Level I. See Instructions for Authors for a complete description of levels of evidence.