The objectives of this study are to assess perioperative opioid use in patients undergoing knee surgery and to examine the relationship between preoperative opioid use and 2-year postoperative patient-reported outcomes (PROs). We hypothesized that preoperative opioid use and, more specifically, higher quantities of preoperative opioid use would be associated with worse PROs in knee surgery patients. We studied 192 patients undergoing knee surgery at a single urban institution. Patients completed multiple PRO measures preoperatively and 2-year postoperatively, including six patient-reported outcomes measurement information system (PROMIS) domains; the International Knee Documentation Committee (IKDC) questionnaire, numeric pain scale (NPS) scores for the operative knee and the rest of the body, Marx’s knee activity rating scale, Tegner’s activity scale, International Physical Activity Questionnaire, as well as measures of met expectations, overall improvement, and overall satisfaction. Total morphine equivalents (TMEs) were calculated from a regional prescription monitoring program. Eighty patients (41.7%) filled an opioid prescription preoperatively, and refill TMEs were significantly higher in this subpopulation. Opioid use was associated with unemployment, government insurance, smoking, depression, history of prior surgery, higher body mass index, greater comorbidities, and lower treatment expectations. Preoperative opioid use was associated with significantly worse 2-year scores on most PROs, including PROMIS physical function, pain interference, fatigue, social satisfaction, IKDC, NPS for the knee and rest of the body, and Marx’s and Tegner’s scales. There was a significant dose-dependent association between greater preoperative TMEs and worse scores for PROMIS physical function, pain interference, fatigue, social satisfaction, NPS body, and Marx’s and Tegner’s scales. Multivariable analysis confirmed that any preoperative opioid use, but not quantity of TMEs, was an independent predictor of worse 2-year scores for function, activity, and knee pain. Preoperative opioid use and TMEs were neither independent predictors of met expectations, satisfaction, patient-perceived improvement, nor improvement on any PROs. Our findings demonstrate that preoperative opioid use is associated with clinically relevant worse patient-reported knee function and pain after knee surgery.Thieme Medical Publishers 333 Seventh Avenue, New York, NY 10001, USA.