Gabapentin has shown acute postoperative analgesic effects, but the optimal dose and procedure-specific benefits vs harm have not been clarified. In this randomized, double-blind, placebo-controlled dose-finding study, 300 opioid-naive patients scheduled for total knee arthroplasty were randomized (1:1:1) to either gabapentin 1300 mg/d (group A), gabapentin 900 mg/d (group B), or placebo (group C) daily from 2 hours preoperatively to postoperative day 6 in addition to a standardized multimodal analgesic regime. The primary outcome was pain upon ambulation 24 hours after surgery, and the secondary outcome was sedation 6 hours after surgery. Other outcomes were overall pain during well-defined mobilizations and at rest and sedation during the first 48 hours and from days 2-6, morphine use, anxiety, depression, sleep quality, and nausea, vomiting, dizziness, concentration difficulty, headache, visual disturbances, and adverse reactions. Pain upon ambulation (visual analog scale, mean [95% confidence interval]) 24 hours after surgery in group A vs B vs C was as follows: 41 [37-46] vs 41 [36-45] vs 42 [37-47], P = 0.93. Sedation (numeric rating scale, median [range]) 6 hours after surgery was as follows: 3.2 [0-10] vs 2.6 [0-9] vs 2.3 [0-9], the mean difference A vs C being 0.9 [0.2-1.7], P = 0.046. No between-group differences were observed in overall pain or morphine use the first 48 hours and from days 2-6. Sleep quality was better during the first 2 nights in group A and B vs C. Dizziness was more pronounced from days 2-6 in A vs C. More severe adverse reactions were observed in group A vs B and C. In conclusion, gabapentin may have a limited if any role in acute postoperative pain management of opioid-naive patients undergoing total knee arthroplasty and should not be recommended as a standard of care.