Validity of published outcome data concerning Anatomic Graduated Component total knee arthroplasty: a structured literature review including arthroplasty register dataSchuh, R., Dorninger, G., Agreiter, M. et al.
Total knee arthroplasty (TKA) as a treatment for end-stage osteoarthritis of the knee shows good results in terms of patient satisfaction. For the assessment of outcome and revision rate after total joint arthroplasty, there are two major data sources: clinical studies and national arthroplasty registers. The purpose of this study was to analyse the outcome of Anatomic Graduated Component (AGC) TKA reported in clinical studies and to perform a comparison with the outcome reported by national arthroplasty registers.
A systematic literature review was performed using standardised methodology in order to determine the outcome and revision rate of AGC TKA. In a comprehensive meta-analysis of clinical studies and worldwide register results we examined the quality of the basic data and the occurrence and influence of potential bias factors. Confidence intervals were calculated to determine the statistical significance of differences.
We found significant differences as regards the revision rate measured in revisions per 100 observed component years. Compared to worldwide register data it turned out to be significantly lower in clinical studies published by the implant development team. Actually, they reported a revision rate of 0.18 revisions per 100 observed component years, whereas annual reports of national arthroplasty registers report 0.74 revisions per 100 observed component years. A comparison of the results from national arthroplasty registers of different countries revealed a significantly higher revision rate for Denmark in relation to worldwide register data.
A conventional meta-analysis of clinical studies is affected by the influence of the development team and therefore subject to bias. For the assessment of outcome arthroplasty register data should be rated as superior and, being used as reference data for the detection of potential bias factors in the clinical literature, could make an essential contribution to the quality of scientific meta-analysis.