Introduction. Infection after total hip arthroplasty is a severe complication. Controversies still exist as to the use of cemented or uncemented implants in the revision of infected THAs. Based on the data in the Norwegian Arthroplasty Register (NAR) we have studied this topic. Material and Methods. During the period 2002–2008 45.724 primary THAs were reported to NAR. Out of these 459 were revised due to infection (1,0%). The survival of the revisions with uncemented prostheses were compared to revisions with cemented prostheses with antibiotic loaded cement and to cemented prostheses with plain cement. Only prostheses with the same fixation both in acetabulum and in femur were included in the study. Cox-estimated survival and relative revision risks were calculated with adjustments for differences among groups in gender, type of surgical procedure, type of prosthesis, and age at revision. Results. 92 (23%) of all the revisions were performed with uncemented prostheses, 286 (71%) with cemented prostheses with antibiotic loaded cement, and 25 (6%) with plain cement. Compared to uncemented prostheses and with all reasons for revision as endpoint in the Cox-analyses, prostheses fixed with antibiotic loaded cement had 3.0 (1.4–6.3) times increased risk for re-revision (p=0.004) and prostheses with plain cement 1.9 (0.4–9.3) times increased risk (p=0.44). With infection as endpoint, prostheses with antibiotic loaded cement had 2.8 (1.2–6.4) times increased risk for re-revision (p=0.02) and prostheses with plain cement 2.6 (0.5–13.7) times increased (p=0.26). 77% of the re-revisions (48 of 60) were performed due to infection. Conclusion. Data in the Norwegian Arthroplasty Registry indicate that uncemented prostheses should be used in the revision of infected total arthroplasties