Background: Implant fractures are complications that have a great impact on the patient’s quality of life after total hip arthroplasty. Nevertheless their occurrence is often considered as rare in clinical practice. We compared incidences of implant fractures in various datasets in order to calculate the risk of a fracture and assess the quality of these datasets for such evaluations.

Methods: In a structured literature analysis based on a standardised methodology the incidence of reoperations was evaluated comparing clinical studies published in Medline-listed journals and annual reports of National Arthroplasty Registers worldwide. Case reports and experimental studies were not considered.

Results: The majority of clinical studies are monocentre trials. The publications comprise a cumulative number of 72,571 stems with 234 stem fractures, 73,743 cups with 191 component fractures, and 16,381 ceramic heads with 44 fractures. A survey among the members of the American Association of Hip and Knee Surgeons covered 64,483 primary operations, hence including a similar number of primary cases as all monocentre studies together. This dataset involves 355 implant fractures.

By contrast, worldwide Register data refer to 733,000 primary operations, i.e. approximately 10 times as many as sample-based datasets.

In general, sample-based datasets present higher revision rates than register data. The deviations are high, with a maximum factor of 64 for hip stems. Whereas the AAHKS survey exhibits lower deviations than the monocentre trials, they are still too high for this data collection tool being considered as reliable and safe to provide valid data for general conclusions.

The incidence of implant fractures after total hip arthroplasty in pooled worldwide arthroplasty register datasets is 304 fractures per 100.000 implants. In other words, one out of 323 patients has to undergo revision surgery due to an implant fracture after THA in their lifetime.

Conclusion: For general assessments in the context of implant fractures, register data have the highest value. Clinical studies, which often focus on a particular implant, are of very limited value for global conclusions. Structured surveys produce more reliable data than clinical studies and are superior to monocentre trials. However, the AAHKS survey presents data that also show considerable differences to data from registers.

For the detection of rare, but severe complications like implant fractures sample-based studies achieve the goal of providing accurate figures only to a very limited extent, even if the samples are large. Here, too, comprehensive national arthroplasty registers are the most suitable tool to identify such incidents and calculate reliable figures.

Contrary to the prevalent opinion, implant fractures still are a relevant problem in arthroplasty.

Background: Within the scope of the EU project EUPHORIC a methodology for direct comparison of different datasets was developed and applied on a sample of implants, among them the Oxford Unicompartmental Knee Arthroplasty (Oxford Uni). The aim was to identify potential bias factors inherent in the datasets and evaluate the outcome achieved with this implant.

Materials and Methods: A structured comparison was performed of data published on the revision rate of the Oxford Unicompartmental prosthesis. Both clinical follow-up studies published in Medline-listed journals and worldwide Register data were included. The data were stratified with regard to potential influence factors like the individual research groups or the geographical origin of the papers.

Results: A major proportion of the published data, between 50 and 75%, depending on the method of calculation, comes from studies including the developing institution in Oxford. The results published by this group deviates statistically significantly from the reference datasets from Register data or independent research groups. Data from the developing hospital show mean revision rates that are 4.4 times lower than those based on worldwide Register data, and 2.74 times lower than in independent studies. As opposed to this, independent studies on average publish data that are reproducible in Register data.

Conclusion: A conventional meta-analysis of clinical studies is significantly affected through the influence of the developing institution and is therefore subject to a bias. Neither through arthroplasty Register outcome data nor by other research groups that have disclosed outcome information on the Oxford Uni can the excellent results be reproduced that were published by the inventors.

Compared to other implants for unicompartmental knee arthroplasty in worldwide arthroplasty Registers, the Oxford Uni shows good results.

For the assessment of the outcome of implants, register data are to be rated superior and, in terms of reference data for the detection of potential bias factors in the clinical literature, can provide an essential contribution for scientific meta-analyses.

In a systematic review, reports from national registers and clinical studies were identified and analysed with respect to revision rates after joint replacement, which were calculated as revisions per 100 observed component years.

After primary hip replacement, a mean of 1.29 revisions per 100 observed component years was seen. The results after primary total knee replacement are 1.26 revisions per 100 observed component years, and 1.53 after medial unicompartmental replacement. After total ankle replacement a mean of 3.29 revisions per 100 observed component years was seen.

The outcomes of total hip and knee replacement are almost identical. Revision rates of about 6% after five years and 12% after ten years are to be expected.

Introduction: Clinical follow-up studies are sample based, in contrast to arthroplasty register data, which refer to the entire population treated. Aim of this study is to assess the differences in revision rate to quantify bias-factors in published literature.

Materials and Methods: A structured literature review of Medline-listed peer reviewed journals on examples has been performed concerning implants with sufficient material in both data sources available. Products with inferior outcome were subsumed in a subgroup.

Results: The number of cases presented in peer reviewed journals are relatively low in general and show a high variability.

The average revision rate in peer reviewed literature is significantly lower than in arthroplasty register data-sets.

Studies published by the inventor of an implant tend to show superior outcome compared to independent publications and Arthroplasty Register data. Factors of 4 to more than 10 have been found, which has a significant impact for the results of Metaanalyses.

When an implant is taken from the market or replaced by a successor there is a significant decrease in publications, which limits the detection of failure mechanisms such as PE wear or insufficient locking mechanisms.

The final statement made about the product under investigation seem to follow a certain mainstream.

Discussion and Conclusion: Arthroplasty Register datasets are superior to Metaanalyses of peer reviewed literature concerning revision rate and the detection of failure mechanisms. Combined reviews could reduce bias factors and thereby raise the quality of reports.