Abstract Objective Approximately one in four total knee replacement patients develop persistent pain. Identification of those at higher risk could help inform optimal management. Methods We searched MEDLINE, EMBASE, CINAHL, AMED, SPORTDiscus, and PsycINFO for observational studies that explored the association between risk factors and persistent pain (≥3 months) after total knee replacement. We pooled estimates of association for all independent variables reported by >1 study. Results Thirty studies (26,517 patients) reported the association of 151 independent variables with persistent pain after knee replacement. High certainty evidence demonstrated an increased risk of persistent pain with pain catastrophizing (absolute risk increase [ARI] 23%, 95% confidence interval [CI] 12 to 35), younger age (ARI for every 10-year decrement from age 80, 4%, 95% CI 2 to 6), and moderate-to-severe acute post-operative pain (ARI 30%, 95% CI 20 to 39). Moderate certainty evidence suggested an association with female sex (ARI 7%, 95% CI 3 to 11) and higher pre-operative pain (ARI 35%, 95% CI 7 to 58). Studies did not adjust for both peri-operative pain severity and pain catastrophizing, which are unlikely to be independent. High to moderate certainty evidence demonstrated no association with pre-operative range of motion, body mass index, bilateral or unilateral knee replacement, and American Society of Anesthesiologists score. Conclusions Rigorously conducted observational studies are required to establish the relative importance of higher levels of peri-operative pain and pain catastrophizing with persistent pain after knee replacement surgery.