Page 174 - Reconstructive Techniques in Musculoskeletal Tumor Surgery

Chapter 8
The risk of nonunion did not di er signi cantly between patients who received chemotherapy (14/70, 20%) and those who did not (6/42, 14%) (p = 0.445).
Discussion
Nonunion is among the leading causes for failure of intercalary allografts. In this retrospective case series, we evaluated risk factors for nonunion and assessed whether cortical contact at the allograft-host junction results in a decreased likelihood of nonunion.
Our study had a number of limitations. First, we recognize the retrospective design of this study and the selection bias for the patients who were treated in two di erent countries by two di erent groups. We were not able to obtain the presence of other potential risk factors, such as smoking status. Second, because digital radiographs were not available before 2008 in one of our centers, we included patients who were treated at di erent periods in time. However, over the years, little has changed in our perioperative protocols. Third, the number of events was limited and therefore, we could not perform a multivariable analysis. Fourth, the group has some inherent heterogeneity, which could a ect the incidence of nonunion. To minimize the risk of bias, we chose to only include one-plane transverse osteotomies that were xed using one or more plate(s).
Sixteen percent of the osteotomies did not initially heal. Reported rates of nonunion in literature vary from 15 to 50%1-5, 7, 9, 10, 21, 22. However, as we noted previously, some studies assessed nonunion per patient, while others scored both osteotomies and therefore score more nonunions, but report a lower percentage of nonunion (table 3)1. In addition, previous authors used di erent de nitions of nonunion. Most large studies determined union radiographically3, 6, 9. Although some de ned nonunion as a lack of progressive healing at six months23, most large studies de ned nonunion as the lack of cortical continuity in three cortices after 12 months6, 9. Apart from that, previously reported incidences may have also included infected nonunions. To avoid bias, we chose to exclude patients with an infection from our study. Nevertheless, in clinical practice, the possibility of infection should always be excluded if a junction does not heal. Clinical workup should include physical examination, laboratory testing (including white blood cell count, C-reactive protein and sedimentation rate), a conventional radiograph or CT-scan and, in case of doubt, leukocyte scintigraphy.
172