Page 226 - Reconstructive Techniques in Musculoskeletal Tumor Surgery

Chapter 11
surgical procedure is dependent on the right indication and a precise technique of execution.
General considerations
The vast majority of clinical articles in orthopaedic journals are single center observational case series on a surgical technique128, leading to a substantial risk of selection bias and heterogeneity. A systematic review demonstrated that 92% of studies published on surgical management of lower extremity bone tumors are level IV or V studies129. The overall quality of reporting is generally poor, and studies are therefore prone to confounding bias, sampling bias and recall bias129. Furthermore, studies on surgical techniques often report single-center results from a highly specialized center – commonly one that was involved in the development of the technique – and thus may overestimate clinical outcome. Reasons for the lack of higher level of evidence studies include the rareness of diseases, heterogeneity in presentation and surgical approaches, loss of follow- up due to patient mortality, and ethical considerations. International cooperation is key to obtaining su cient patient numbers, although di erences in expertise, treatment protocols and surgeon preferences may introduce other types of bias. In that regard, it is essential that uniform de nitions are employed and that standard reporting guidelines, such as the STROBE statement, are applied as much as possible130.
The classi cation of failure modes as described by Henderson et al was one of the rst widely supported classi cation systems that aimed to stimulate uniform reporting9. Although the authors must be applauded for their initiative, there are a number of aws in the classi cation. First, the system only classi es failures, not complications. As a result, isolated revision of the bushing is counted as a failure – while many authors consider this routine maintenance57, 58. On the other hand, servicing procedures result in secondary deep infection in approximately 5% of cases58 and we therefore encourage striving for an implant system that is free of the need of maintenance. Second, to distinguish early from late infections, the Henderson classi cation system uses a cuto point of two years for endoprostheses, and six months for biological reconstructions. Rather than the time from primary surgery to the onset of symptoms, a classi cation system should distinguish infections with an acute onset from delayed or chronic infections; this dictates the treatment strategy and the probability of being able to retain the implant40, 131. Third, the Henderson classi cation did not include massive bone resorption
224