Background and purpose - Prior to primary total knee arthroplasty (pTKA), 6-34% of patients have undergone surgical procedure(s) of their knee. We investigated whether history of previous surgeries influences the risk of revision of pTKA, the risk according to the type of previous surgery, and how previous surgery influences specific causes of revision and the time of revision.Patients and methods - This is a prospective cohort study from the Geneva Arthroplasty Registry. All pTKA between 2000 and 2016 were included and followed until December 31, 2019. Outcomes were risk of revision, evaluated using Kaplan-Meier survival and Cox and competing risks regression, the specific causes, and time of revision.Results - Of 3,945 pTKA included (mean age 71 years, 68% women), 21% had a history of previous surgery, with 8.3% revisions vs. 4.3%, at 3-20 years' follow-up (mean 8.6). 5- and 10-year cumulative failure by previous surgery (yes vs. no) were 6.6% (95% CI 5.1-8.5) vs. 3.3% (CI 2.7-4.0), and 8.4% (CI 6.6-10.6) vs. 4.5% (CI 3.8-5.4). Baseline differences explained only part of the higher risk (adjusted HR 1.5, CI 1.1-2.1). The risk of failure was higher for all causes of revision considered. Patients in the previous surgery group had a higher risk of an early revision.Interpretation - A history of previous surgery adversely affected the outcome with a 1.5 times higher cumulative risk of all-cause revision over the course of up to 20 years after index surgery. The increased risk was seen for all causes of revision and was highest in the first years.