Abstract
Over 800 total hip replacement (THR) constructs were implanted in the UK in 2017. To ensure reliable implants are used, a NICE revision benchmark of 5% after 10 years exists. Surgeons are guided in choice by organisations such as the Orthopaedic Data Evaluation Panel (ODEP). Currently, ODEP publishes ratings for stem and cup separately and not for constructs. We used NJR data to investigate whether revision estimates of an individual stem (with all cups) is an accurate indicator of survival of all constructs using that stem.
The dataset comprised 234,289 THRs using the most frequently implanted stem between 2004 and 2017. Crude ten-year revision estimates were calculated using Kaplan-Meier for all THRs and for the five most implanted constructs. Adjusted comparisons between individual constructs and the overall stem revision estimate were made using flexible parametric survival analysis.
The 10-year crude, revision estimate for all THRs was 2.3% (95% CI 2.2, 2.4). Only four of the most frequently used constructs had long enough follow-up to analyse. 10-year estimates for these constructs ranged from 1.8% (95% CI 1.5, 2.1) to 3.7% (95% CI 3.2, 4.1), a log-rank test revealed strong evidence against the null hypothesis that revision estimates were the same for all constructs (p<0.001). Adjusted for age, sex and ASA, three of the four constructs showed a difference in 10-year revision estimates compared to this stem with all cups (P=0.03, P<0.001, P<0.001).
This study suggests 10-year revision estimates for all THRs using the most implanted stem in the NJR are not representative of all constructs involving that stem in crude or adjusted analyses. Current benchmarking systems report survival for the stem in combination with all cups and not for constructs. We suggest that benchmarking ratings basing on revision estimates for THR constructs would provide more accurate information, enabling informed construct decisions.
