Dolph Houben

65 Combined massive allograft and intramedullary vascularized fibula 3 Union Sixteen studies stated their primary union rates and described cases of nonunion or delayed union. Primary union rates between 66.7 and 100 percent were found. Based on these rates a meta-analysis was performed (Fig. 4). One publication [29] was not included in our meta-analyses since actual numbers were unknown. After pooling the data, a primary union rate of 86.5 (95% CI: 79.6 to 92.2) percent was calculated. Although the data could be pooled, there was a statistically significant difference in primary union rates between the studies (p=0.003). Manfrini, Bindiganavile, Say, Colangeliet al. [25] performed a case-control study comparing the outcomes of pedicled or free vascularized fibular grafts in combination with a massive allograft. This study showed no significant difference between the groups concerning complications or union. Figure 4: Meta-analyses on primary union rates: calculated means are represented by the blue squares; each square differs in size since the size of the square is representation of the numbers of patients included in the study. The blue error bares represent the 95% confidence interval. The blue Rhombus indicates the outcome of the quantitative statistical analyses for both fixed and random effect models. For this review we choose the random effect model (86.5, 95% CI: 79.6 to 92.2) since this is the right effect model if there is a lot of variation within the data. Function Eight studies used the Musculoskeletal Tumor Society Scoring System (MSTS-score) to measure their functional outcome after oncologic resection around the knee [13, 15, 16, 25-27, 30] . This scoring system ranges from either 0 to 30 points with 0 being a poor functional outcome and 30 point a good functional outcome. These studies showed an average MSTS score of 26.3 ranging from 24 to 29 (Fig. 5). Capanna et. al. [18] reported 72% excellent, 20% good, 5 % fair and 3% poor MSTS-scores. The remaining studies used other functional outcome rating systems such as the Mankin evaluation score [14, 22] and the Enneking system [19] . These data were not comparable.