Quality of Outcome Data in Total Hip Arthroplasty: Comparison of Registry Data and Worldwide Non-Registry Studies from 5 Decades

Abstract

Purpose This systematic review assessed evidence on outcome (revision rate for all reasons) following hip arthroplasty from its beginning 5 decades ago. Methods We evaluated all studies from all current hip implants since their market introduction in 1962 regarding “revision rate per 100 observed component years”. Data were compared with arthroplasty registries. Results A total of 54 different hip implants were included: for 81% (44 of 54) data is either absent or poor; for 30% (16 of 54) not a single publication could be found. For 52% (28 of 54) less than 100 revisions for all reasons are published in non-registry studies. The remaining 10 implants (19%) comprise 92638 primary implants with 4473 revisions. Control group were the same implants with 111658 primary cases and 3029 revisions from arthroplasty registries. A systematic developer bias as in knee arthroplasty could not be found but several independent authors were found to significantly bias the literature. The overall revision rates per 100 observed component years from non-registry studies (and joint registries) are 0.4 (0.5) for stems, 0.7 (0.7) for cups and 1.4 (2.1) for resurfacing systems. Conclusions For 81% of all hip implants assessed limited evidence exists from non-registry studies regarding outcome (revision rate) even 5 decades after market introduction. For the remaining 19% of implants no systematic developer bias could be found but several individual authors significantly biased results of single implants. We therefore ask for a more active publication of new implants.