Despite the standardization of two-stage knee revision protocols, a high percentage of failures still occurs. Identifying the predictors of failure is necessary to determine appropriate management and counsel for patients with a periprosthetic knee infection. This study aimed to identify risk factors predicting the failure, to describe implant survival, and to report the mid-term clinical outcomes of patients undergoing two-stage revision for periprosthetic knee infection. Data of patients who underwent two-stage knee revision from 2012 to 2016 were analyzed, and 108 patients were included. The mean age was 66.6 ± 9.2years. The mean follow-up was 52.9 ± 15.6months. Logistic regression was conducted to identify predictors of treatment failure. Kaplan-Meier curves were generated to assess implant survival. Preoperative functional outcomes were compared to those registered at the final follow-up. Difficult-to-treat infections (OR = 4.2, 95% CI 1.2-14.5, p = 0.025), the number of previous surgeries (OR = 1.8, 95% CI 1.2-2.6, p = 0.005), and the level of tibial bone defect (OR = 2.3, 95% CI 1.1-4.7, p = 0.027) significantly predicted the failure of two-stage knee revision. Survivorship of implants was significantly lower for patients presenting these risk factors (p < 0.05). Mean Knee Society Score improved from 49.0 ± 12.0 to 80.2 ± 13.6 (p < 0.001). Mean Oxford Knee Score improved from 22.2 ± 4.9 to 36.1 ± 6.0 points (p < 0.001). Difficult-to-treat pathogens, the number of previous surgeries, and the level of tibial bone defect were independent risk factors of two-stage knee revision failure. Overall, the two-stage protocol provided a good survival rate and functional outcome.