Dolph Houben

60 CHAPTER 3 Exclusion criteria Upper extremity reconstructions were excluded. Case reports or case series describing fewer or equal to five cases were also excluded. We limited this study by language other than English or Dutch. No limitations for the follow-up period were made. When double data from the same author was suspected, we contacted the author. If the difference in the number of cases was fewer than five, the article with the smallest number was excluded. Animal studies and reviews were also excluded. Data extraction Data was independently extracted by two reviewers (RH and MR) on a preconceived data extraction form using Excel. The following data was extracted: first author, year of publication, level of evidence, blinding, population characteristics, underlying pathology, complications, union rates, graft characteristics, fixation method, immobilization method, time to full weight baring (FWB), re- intervention, function, and amputation. Delayed union was defined as failure to reach bone union after within 6 months after reconstruction. Whereas non-union is defined as failure to reach bone union after 6 months without surgical intervention. When data or articles could not be provided by the library an attempt to contact the author was made. After independent data collection, the reviewers compared the extracted data. Statistical analysis Statistical analysis and graphic design were done in collaboration with the Department of Epidemiology and Biostatistics using MedCalc, version 17.8.2. A Meta-analysis was performed on 16 of the included studies for primary union rate. The other data could not be used for meta- analysis because of the heterogeneity of the data (Fig. 1). Results Study selection A total of 1190 results were returned after a search in Pubmed, Embase, and Medline. The original articles (1989, 1991 and 1993) of the Capanna technique were not found by the performed search. One of the original Capanna articles was retrieved by the use of Google Scholar [21] . After removal of duplicates, screening on title and abstract and full-text assessment 17 articles were used for qualitative syntheses. Meta-analysis was performed on 16 included studies for primary union rate only. The other data could not be used for meta-analysis because of the heterogeneity of the data. A summary of our study selection is given in our flow diagram (Fig. 1). Study characteristics The included studies (n=17) were published between January 1997 and May 2017. The average follow-up was 58.5 months (15.0 -139.3). The included studies report their results of a massive allograft combined with a vascularized inlay for reconstruction of the lower extremity (femur + tibia) in 329 cases. The average age reported in all studies was 16.0 years (3-54). All studies presented results of reconstructions after oncologic resection, except for the study of Venkatramani et al.