The findings of this study demonstrate that within the presented cohort, a temporizing cast immobilization was associated with no worse results compared to ex-fix. Even a trend toward a less favorable rate of soft-tissue-related complications was seen for ex-fix. Those temporized with a cast underwent definite fixation earlier than the patients with ex-fix, which resulted in a shorter length of stay. An important finding of this study was that posterior malleolar fragment size was the sole predictor of loss of reduction of all assessed variables. Therefore, casting of patients with a critical posterior malleolar fragment size must be discussed.
Soft-tissue problems associated with temporizing cast immobilization were broadly in keeping with the rates reported by other authors with SSI ranging between 1 to 20% and skin necrosis rate ranging between 4.9 to 8.9% [6, 12, 17, 24, 25, 29, 30]. Wawrose et al. compared 28 patients with temporizing plaster splint immobilization to 28 patients with ex-fix following ankle fracture dislocation . The authors found that splint immobilization was associated with a high skin necrosis rate of 17.8% when compared to a temporizing ex-fix which yielded 0% skin necrosis. The rather high necrosis rate of 17.8% and sSSI rate of 17.8% in the splint group and the absence of complications in the ex-fix group favorized the use of ex-fix as a temporizing fixation following ankle dislocation fractures. These remarkably good results following ex-fix could not be reproduced at our institution; yet, our necrosis rate within the cast group was substantially less. These differences between this study and our study might be a result of the high re-dislocation rate of 50% in the splint group seen by Wawrose et al.
Loss of reduction was contributed to the type of cast in the study by Baker et al. . A 50% (11/22 cases) loss of reduction was noticed when temporizing immobilization was performed with a plaster splint, whereas no re-dislocation was seen in the bivalved fiberglass cast group (0/17). In our study, the type of cast was not a predictor for loss of reduction. Only size of the posterior malleolar fragment was highly associated with loss of reduction. Although size of the posterior malleolar fragment does not correlate with clinical and radiological outcomes after ORIF , our findings support the idea that the posterior malleolus is an important indicator for fracture stability of the ankle. Given this result, the substantial difference between our moderate re-dislocation rate of 19.8% within the cast group and an unacceptable high re-dislocation rate of 50% reported by Wawrose et al. might be attributed to differences in posterior malleolar fragment sizes between the study groups . It must be noted that, in contrast to our study, patients included in the study by Wawrose et al. were discharged after splinting, which makes strict monitoring of the ankle and a prompt reaction to soft-tissue problems difficult. Moreover, the lack of control of patients’ behavior at home might lead to noncompliance with patients walking on the cast.
In another retrospective study Buyukkuscu et al. showed a higher rate of reduction loss and skin necrosis in the splint group (n = 69) compared to the external fixator group (n = 48) . The time to surgery was shorter in the external fixator group. This difference may be explained by inclusion of patients with poor soft tissue conditions only.
Regarding predictors, a ROC-analysis identified two thresholds for posterior malleolar fragment size to predict loss of reduction in a cast demonstrating the same tradeoff between sensitivity and specificity (Fig. 3). Whereas a 13.5% threshold would overcome a re-dislocation in a temporizing cast in the majority of cases, a substantial percentage of patients would be treated with an additional surgery that was not necessary due to the false-positive rate associated with a specificity of 57%. Since our results regarding complications with cast immobilization were promising, we prefer a more conservative approach. Therefore, the 22.5% threshold, which allows for some degree of re-dislocation but overcomes surgical overtreatment, was preferred as a cutoff for critical posterior malleolar fragment size to define ankle fracture-dislocations in which a primary temporizing ex-fix seems appropriate.
Our study is limited by its retrospective nature. The slightly skewed distribution between groups and the limited power are exemplary for this. The latter highlights the challenges associated with a monocentric study, even in a level-1 trauma center, in which small differences in complication rates are seen, needing a larger sample size to provide a sufficient power. Patients were treated by several surgeons, since this retrospective analysis covered 10 years of treatment in a teaching hospital. At the time, no clear indication to do a primary ex-fix was defined and depended on surgeon´s decision-making. This imposes a further selection bias of the study cohort. Because of the multiple variables that were needed for the analyses, a substantial number of cases had to be excluded due to incomplete data. A prospective assessment would have increased the sample size by assessing those variables in a standardized fashion. The short-term follow-up of a minimum of 6 months prohibited us to make any conclusions on long-term outcomes on the initial use of a cast versus ex-fix for ankle fracture-dislocation management. Although the majority of cases underwent a CT scan of the ankle, the size of the posterior malleolar fragment was only assessed on lateral radiographs. This measurement might be influenced by malrotated views, as commonly seen in the acute setting. Yet, radiographs are the primary images acquired in the acute assessment of an ankle fracture and hence makes the results of this study more applicable for quick decision making.