O2138: Has the Femur Shortened Too Much? A Reliability Study of Femoral Shortening in Femur Fractures in Children Less Than 4 Years of Age

Background: Closed reduction and spica casting is the treatment of choice for young children with femoral shaft fractures. Femoral shortening is an important metric in clinician decision making for treatment and gauging the success or failure of closed treatment of these fractures. In spite of the importance of this metric, no studies exist examining the reliability of the assessment of femoral shortening both pre- and post- casting.

Methods: Our group selected 50 consecutive femoral shaft fractures treated in spica casts in children younger than 4 years of age who presented to our level 1 trauma center. Four raters, two trained pediatric orthopedists with more than 10 years’ experience and two with less than 5 years’ experience, measured radiographic shortening on the AP and lateral views pre- and post- casting using medical imaging software. Following a 4-week washout period, the same reviewers assessed the x-rays again for shortening. We then assessed agreement using intraclass correlation coefficient with an alpha model mixed two-way analysis for absolute agreement and both inter-rater reliability and test retest (intra-rater reliability). We used Cohen's Kappa to assess agreement on whether a fracture was shortened more than 2 cm. Mean difference was used to quantify average difference to the mean rating.

Results: Inter-rater reliability on injury films was excellent for round 1 (0.706 for AP and 0.708 for lateral), and on follow-up casted x-rays (0.809 for AP and 0.811 for lateral). Similarly for round 2, inter-rater reliability was excellent for injury films, (0.745 for AP and 0.777 for lateral) and for follow up casted films (0.688 AP and 0.749 lateral). Intra-rater (test retest) reliability ranged between 0.711 and 0.911. There was no significant difference between raters of less than 5 years' experience and more than 10 years' experience. Agreement between raters on whether or not a fracture was shortened more than 2 cm was marginal (0.514). Mean difference in measurement of each rater to the mean was 9 ± 5 mm.

Conclusion: Our results suggest that although absolute agreement is good for measurements of femoral shortening after spica casting, there is a concerning amount of deviation about the mean measurement. Additionally, there is substantially worse agreement on whether there is more or less than 2 cm of shortening. Since 2 cm of shortening is considered the borderline of acceptability in non-operatively treated femoral fractures and the mean difference is 9 mm, our results suggest standardization is warranted in measurement technique.