It has been estimated that up to 54% of the variance in postoperative pain experience may be predicted with preoperative pain responses to experimental stimuli, with suprathreshold heat pain as the most consistent test modality. We aimed to explore if 2 heat test paradigms could predict postoperative pain after total knee arthroplasty (TKA). Patients scheduled for elective, unilateral, primary TKA under spinal anesthesia were consecutively included in this prospective, observational study. Perioperative analgesia was standardized for all patients. Outcomes were postoperative pain during walk: From 6-24hrs (primary), from postoperative day (POD) 1-7 (secondary), and from POD14-30 (tertiary). Two preoperative tonic heat stimuli with 47°C were used; short (5sec) and long (7min) stimulation upon which patients rated their pain response on an electronic VAS. Multivariate stepwise linear and logistic regressions analyses were carried out including 8 potential preoperative explanatory variables (among these anxiety, depression, preoperative pain and pain catastrophizing) to assess pain response to preoperative heat pain stimulation as independent predictor for postoperative pain. 100 patients were included, and 3 were later excluded. A weak correlation [rho(95% CI);p-value] was observed between pain from POD1-7 and pain response to short [rho=0.25(0.04; 0.44);p=0.02] and to long [rho=0.27(0.07; 0.46);p=0.01] heat pain stimulation. However, these positive correlations were not supported by the linear and logistic regression analyses, where only anxiety, preoperative pain and pain catastrophizing were significant explanatory variables (but with low R-Squares;0.05-0.08). Pain responses to 2 types of preoperative heat stimuli were not independent clinical relevant predictors for postoperative pain after TKA.