Background One of the most important goals of primary total knee arthroplasty (TKA) is to achieve a functional range of motion (ROM). However, up to 20% of patients fail to do so, which can impair activities of daily living. Questions/purposes The purpose of this study was to evaluate the effect of various (1) demographic factors;(2) comorbidities; and (3) knee-specific factors on the frequency of manipulation under anesthesia, which was used as an indicator of knee stiffness after a primary TKA. Methods We evaluated the registries of two high-volume centers and reviewed all 3182 TKAs that were performed between 2005 and 2011 to identify all patients who had undergone manipulation under anesthesia (MUA). A total of 156 knees in 133 patients underwent MUA after an index arthroplasty. These patients were compared in a one-to-four ratio with a group of patients with satisfactory ROM drawn from the same database who met prespecified criteria and who had not undergone MUA. Effects of various factors, including age, sex, body mass index, race, comorbidities, and the underlying cause of knee arthritis, were compared between these two cohorts using multivariable logistic regressions. Results After controlling for various confounding, nonwhite race was associated with an increase (odds ratio [OR], 2.01; p = 0.03), and age C 65 years (OR, 0.17; 95% confidence interval [CI], 0.04-0.74; p = 0.0179) was associated with a reduction in the incidence of MUA. In comorbidities, diabetes (OR, 1.72; 95% CI, 1.02-2.32; p = 0.03), high cholesterol levels (OR, 2.70; p = 0.03), and tobacco smoking (OR, 1.59; 95% CI, 1.03-2.47; p = 0.03) were associated with an increase in frequency of MUA. In knee-specific factors, preoperative knee ROM of less than 100°(OR, 0.80; p \ 0.0001) and knee osteonecrosis (p = 3.61; 95% CI, 1.29-10.1; p = 0.014) were associated with increased frequency of MUA. Conclusions We identified several demographic, medical, and knee-specific factors that were associated with poor