179 patients undergoing total knee arthroplasty (TKA) were randomized to receive a four day perioperative regimen of either gabapentin or placebo as an adjunct to celecoxib. The primary intent of this trial was to determine whether gabapentin improves in-hospital rehabilitation and physical function 4 days, 6 weeks, and 3 months after surgery. No significant differences in favour of gabapentin were observed with respect to pain and physical function at any follow-up time point. However, gabapentin decreased morphine consumption in the first 24 hours after surgery, reduced the incidence of nausea, pruritus and dizziness, and increased in-hospital knee range of motion compared to placebo.