We have shown in 62 251 primary THRs the relative performance of implanted constructs compared to a contemporary best practice benchmark. There is substantial variation in the performance of THR constructs. A noninferiority approach conveys distinct advantages as opposed to component/years in the NZJR annual reports, standard Kaplan-Meier analyses in the NZJR reports, or categorical grades provided by organisations such as ODEP. There was also, as found by Deere et al., 2019; heterogeneity in constructs used in females and males. We were unable to stratify by both gender and age given insufficient numbers at risk.
Our study also shows that the best practice benchmarks were predominantly hybrid constructs with a dual taper polished cemented stem and metal-on-polyethylene bearing and is in accordance with the findings of Gwynne-Jones et al.18. In the overall comparisons, the MS30 paired with the Fitmore cup metal on polyethylene was the reference construct at both 3 and 5 years. The Exeter V40 cemented stem paired with the Trilogy (7 years) and Trident (10 years) was the benchmark construct. This study also showed that the Exeter V40/Trident metal on polyethylene was either noninferior or the benchmark at 7 and 10 years whilst the Exeter V40/Trident ceramic on ceramic was noninferior at 3 and 5 years. This strongly suggests that the Exeter V40/Trident combination could appropriately be used as default options for the majority of patients. This is particularly relevant for inexperienced surgical teams, as they can focus training on, and become expert with, a single prosthesis construct19. Potentially this may reduce the risk of technical error, to be cost saving through bulk purchasing arrangements and via a reduction in failure rates. The absolute level of failure of commonly used constructs is relatively low, and < 5% in many instances. Interestingly uncemented constructs compared favorably to hybrid constructs. The Corail/Pinnacle ceramic on ceramic, Accolade/Trident ceramic on ceramic and CLS/Morscher ceramic on polyethylene were noninferior at 10 years compared to the Exeter V40/Trident metal on polyethylene benchmark.
Whilst the Exeter V40 cemented stem had outstanding performance in the NJR study by Deere et al.5, in our study when paired with the Tritanium cup metal on polyethylene at 3 years and Continuum cup metal on polyethylene at 5 years there was twice the revision rate compared to the reference. There was therefore great variation in construct performance with the Exeter V40 depending on which cup it was used with. This illustrates the need to benchmark constructs as opposed to individual implants which make up prosthesis constructs, which has the potential to provide false reassurance in terms of efficacy as the individual elements of a construct are not independent.
This study has a number of strengths. Firstly, we illustrate the need to compare implant constructs as opposed to individual implant components. Secondly, the unambiguous presentation of data allows surgeons, patients and policy makers to directly compare commonly used prosthesis constructs to a reference construct. The constant application of benchmarking methodology and observed trends across both the New Zealand and English and Welsh National Joint Replacement Registers suggests results are generalizable and will be useful to both patients, surgeons, and policy makers.
Our study has a number of limitations; case-mix adjustment by stratification is difficult to account for. Despite efforts to restrict confounding factors, residual confounding factors may be present. The ability to interpret analyses from a causal perspective is limited. It is also known that revision rate is influenced by factors such as the primary indication and the severity of preoperative hip disease.
The results from this study have implications for the way both practicing surgeons, purchasers and patients approach total hip replacement. The transparent presentation of data empowers all parties to have a considered approach to either purchasing a hip replacement or deciding to have surgery with a particular construct and the likely chances of experiencing a revision. With specific reference to New Zealand policy of subsidization of prosthesis constructs used in public hospitals, it is essential that purchasers have access to all local and globally relevant and independent sources of data to ensure public money is used in the most cost-effective manner possible, ensuring as many patients will be treated as possible within the available budget constraints.
For new surgeons, or surgeons looking to optimize the care of their patients, they now have an independent and detailed source of data which compares a wide variety of prosthesis constructs using clinically relevant strata. This will ensure they can pick prostheses that match their surgical competencies or reflect on their need to seek further training, for example in the use of particular prostheses, to ensure they can use implants with a strong track record of performance. Lastly, we hope detailed data will be made available to patients in order to facilitate the shared decision-making process required to inform patients of the risk of revision before deciding to undergo surgery.