Background and purpose - Obesity defined as increased BMI is commonly associated with higher revision rates following total knee arthroplasty (TKA). We examined the effect of BMI on the rate of revision after TKA, for both infection and other reasons, and analyzed weight and height separately to provide better understanding of the risk profile. Patients and methods - The Swedish national knee arthroplasty register was used to identify 107,228 patients operated with primary TKA for osteoarthritis between 2009 and 2017. Cox proportional hazards regression was used to calculate hazard ratios (HRs) with 95% confidence intervals (CIs) for BMI (categories: < 18.5, 18.5-24.9, 25-29.9, 30-34.9, 35-39.9, >= 40), weight (categories: < 65, 65-89, 90-114, >= 115 kg) and height (categories: < 160, 160-179, >= 180 cm Results - There were 2,503 revisions in the follow-up period; 1,036 for infection and 1,467 for other reasons. Higher BMI and weight categories were associated with a similar and statistically significantly increased risk of revision for all causes and for infection. The risk of revision for infection was almost twice in the highest BMI and highest weight group: HR = 3.4 (CI 2.3-4.7) and HR = 3.1 (CI 2.5-3.9) respectively. For BMI and weight categories there was no statistically significant association between revision for other reasons than infection, contrary to the tallest height category where it was statistically significant (HR = 1.3 [CI 1.1-1.5]). Interpretation - BMI, weight, and height may be associated with different types of risks for revision following TKA.