Six hundred twenty-two primary total knee arthroplasties were studied prospectively in 512 patients. A group with no pain and one with severe pain at 5 years were statistically compared. The following were significant predictors of poor pain outcomes: age below 60 (17%) compared with above 60 (7%, P < .05). The first knee was most likely to be in the poor outcome group (13%) compared with the second knee (6%). In contrast, patients who underwent simultaneous bilateral arthroplasty faired better (2%, P < .01). Performing lateral release and sacrificing the posterior cruciate ligament also significantly predicted for poor pain outcomes but may have been influenced by selection bias. We conclude that avoiding surgery in patients younger than 60 and choosing a simultaneous approach to bilateral disease reduce the chance of poor pain outcomes.