Reliability of RUST and modified RUST scores for evaluation of union in pediatric and adult femoral shaft fractures.

Objective: This study aimed to determine the reliability levels of the radiographic union scale for tibial fractures (RUST) and the modified version of the system, mRUST, for femoral shaft fractures in pediatric and adult patients and to evaluate the value of the scores for total and each cortex in the decision making on fracture union.

Methods: A total of 15 orthopedic surgeons scored the radiographs of 24 pediatric and 24 adult patients with femoral shaft fractures that were obtained at 0, 4, 8, 12, and 16 postoperative weeks treated with elastic stable intramedullary nail in pediatric patients and locked intramedullary nail in adult patients using the RUST and mRUST scores. Intra-class correlation coefficient (ICC) was used in the evaluation of reliability of the RUST and mRUST scores. The Fleiss kappa (k) coefficient was used in the agreement between evaluators regarding union decision (united or non-united). The thresholds for RUST and mRUST for radiographic union decision were determined. Receiver operating curves were created to evaluate the contribution of total and individual cortical scores in the decision of united or non-united.

Results: Intra- and inter-rater reliabilities of mRUST (ICC: 0.92 and 0.86, respectively) were slightly higher than those of RUST (ICC: 0.81 and 0.77, respectively) with perfect intra- and inter-rater reliabilities for RUST (ICC: 0.92 and 0.90, respectively) and mRUST (ICC: 0.88 and 0.83, respectively) in pediatric patients and substantial intra- and inter-rater reliabilities in adult patients (ICC: 0.80 and 0.76, respectively, for mRUST, and 0.76 and 0.71, respectively, for RUST). At each time point, the mean mRUST and RUST scores were higher for pediatric fractures (p<0.001). The Fleiss k coefficient for union decision was perfect for pediatric fractures (0.88) and substantial for adult fractures (0.79). The total mRUST score had a higher predictive value of union than the total RUST score (area under the curve: 0.984 vs. 0.922 in adult fractures and 0.990 vs. 0.943 in pediatric fractures). A RUST score of ≥10 and mRUST score of ≥12 were excellent predictors of fracture union.

Conclusion: Fracture union of simple two-part pediatric and adult femoral shaft fractures treated with intramedullary fixation can be reliably assessed using the RUST and mRUST scores. The diagnostic value of the mRUST score is more evident in adult fractures.

Level Of Evidence: Level II, Diagnostic Study.