Background Manipulation under anesthesia has been reported to improve range of motion when other rehabilitative efforts fail to obtain adequate motion after total knee arthroplasty. The purpose of this study was to evaluate the effects of the timing of the manipulation on knee range of motion and clinical outcomes. Methods All 2128 total knee arthroplasties performed at our institution from 2005 to 2011 were reviewed to determine the number of patients who had undergone manipulation under anesthesia. A total of 144 manipulations in eighty-eight women and forty-five men were reviewed. Manipulations under anesthesia that were performed within the first twelve weeks after total knee arthroplasty were considered early and those after that period were considered late. Patients were further substratified according to the timing of the manipulation: Group I included those who had the manipulation within six weeks; Group II, at seven to twelve weeks; Group III, at thirteen to twenty-six weeks; and Group IV, after twenty-six weeks. Outcomes evaluated included gains in flexion and final range of motion, and Knee Society objective and function scores between early and late manipulation, using various adjusted multivariable regression models and at a mean follow-up of fifty-one months (range, twelve to eighty-one months). Mediation analysis was used to investigate whether gains in range of motion from the manipulations under anesthesia alone had mediated the effect between the timing of the manipulation and the clinical outcomes. Results Patients who underwent early manipulation had a significantly higher mean gain in flexion (36.5° versus 17°), higher final range of motion (119° versus 95°), and higher Knee Society objective (89 versus 84 points) and function scores (88 versus 83 points) than those who had late manipulation under anesthesia. There were no significant differences in the outcomes of Groups I and II. Manipulations after twenty-six weeks resulted in unsatisfactory clinical outcomes. Multivariable regression analyses confirmed significantly better clinical outcomes with early manipulation. Mediation analysis showed that the timing of manipulation independently had significantly contributed to the outcomes. Conclusions Orthopaedic surgeons should have a low threshold for performing early manipulations with the patient under anesthesia within twelve weeks after an arthroplasty, to achieve higher knee range of motion and improved clinical outcomes. Level of Evidence Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.