This study identified the mechanisms of failure and the variables associated with failure after revision knee arthroplasty. Five hundred sixty-six index revision knee arthroplasties were studied. Of index revisions, 12.0% failed at an average of 40.1 months. Predominant revision failure modes included infection (46%), aseptic loosening (19%), and instability (13%). Only 4.3% of knees revised for aseptic loosening required rerevision as compared to 21% of knees revised for infection. Revision knee arthroplasty was more likely to fail in younger patients and in those who underwent polyethylene exchanges. Mechanisms of failure for revision arthroplasties are different than for primary knee arthroplasties. Revisions for infection are 4 times more likely to fail than revisions for aseptic loosening. The survivorship for the entire cohort, with revision for any reason as an end point, was 82% at 12 years.