[Detection of inferior products in arthroplasty and implementation of findings: a retrospective analysis of the Boneloc incident].

Aim: By a retrospective assessment of the Boneloc incident, a bone cement which had an inferior outcome in terms of survival rate, the value of published datasets for the detection of inferior outcomes was evaluated.

Method: A structured literature review of English and German peer reviewed journals was conducted. The articles were assessed with respect to revision rate and statements about the product. In a standardised methodology, adjusted for number of cases and follow-up period, the revision rate was calculated. Main goal was to assess the agreement of published information from different datasets.

Results: In the first 4 years after Boneloc had been brought on the market exclusively experimental studies were published, most of which were in favour of the product. In 1995, clinical studies, migration analyses and register-based articles were published. Most of them reported about inferior results, in the same year Boneloc was taken from the market worldwide. Sample-based clinical follow-up studies were not able to contribute to the decision-making process, they were published with a delay of several years and were underpowered from a statistical point of view. All of them published critical statements--after the product had no longer been available on the market for many years. The average revision rate in sample-based studies exceeded the reference value in the Norwegian Arthroplasty Register 7.35-fold. When the inferior results with Boneloc were published, the product had already disappeared from the national markets in Scandinavian countries' operating registers. The central position of orthopaedic scientific societies in the entire outcome monitoring system in these countries seems to be a key factor for success and rapid reaction to identified problems.

Conclusion: Arthroplasty registers and migration analyses have the highest value for the rapid and reliable detection of inferior outcomes in comparative analyses of published articles. Experimental studies did not agree with the performance of the product in a retrospective view, the data cannot be transferred from the estimation of future clinical outcome like survival rates. The involvement of scientific societies in the assessment and dissemination of the results is a key factor to realise potential benefit by an advanced quality monitoring project like arthroplasty registers.