Approximately 15 to 20% of children presenting to the emergency department (ED) with an acute ankle injury have an ankle fracture, and the diagnosis of an occult ankle fracture in a child can be challenging.1–4 The incidence of a distal fibula fracture with no fracture noted on the ED ankle radiograph (occult fracture) has been reported to occur in up to 1/4 of children with an ankle fracture.5–7

Children are evaluated on their initial ED visit by weight bearing, tenderness over the physis, swelling, and ecchymosis. If non–weight bearing, physicians sometimes presume an occult fracture to be present, especially if there is tenderness over the physis.1,3,8,9 Tenderness at the site of injury has been used as a factor to predict the probability of an occult fracture with some success,1,10–14 but radiograph-measured magnitude of acute ankle swelling has not to our knowledge been studied to predict an occult fracture.

Objective

The objective of this study was to determine whether an initial ED visit assessment of magnitude of ankle swelling on a radiograph with or without added assessment of weight bearing helps in prediction of an ankle fracture in children without a fracture detected on initial ED ankle radiographs.

Methods

This was a retrospective chart review study of Children's Hospital, New Orleans, ED visits for a suspected lateral malleolus injury (2021 to 2022). IRB approval was obtained from the facility. Soft-tissue measurements of ankle swelling were made using AP ankle radiographs that were obtained at the time of initial presentation, and an assessment as to weight bearing or non–weight bearing were made.

Swelling measurements were obtained by measuring the distance between the lateral malleolus and the skin projection.

Patients were followed over time, and repeat radiographs were performed identifying any new bone formation or callus as evidence of an occult fracture when the initial radiographs were read as normal.

Statistical Methods

Cross-tabulation was used to estimate the sensitivity and specificity of weight-bearing status when used to identify a true fracture. Fisher exact test was used to assess statistical significance.

Beta regression was used to estimate the cumulative distribution of the probability of an occult fracture as a function of weight-bearing status and mm of swelling. That estimated cumulative distribution was then used to construct smooth curves for the probability of an occult fracture, level of false positives, and level of false negatives as a function of mm of swelling (Figures 1 and 2). All statistical analyses were done using Stata (StataCorp LLC).

Figure 1:

Graph showing the probability of an occult fracture calculated as a function of mm swelling and mm swelling with weight-bearing status.

Figure 2:

Graph showing occult fracture false-negative rate and false-positive rate plotted as a function of the degree of ankle swelling.

Results

A total of 61 children presented to Children's Hospital, New Orleans, with an acute ankle injury during the time frame studied. Of these, 32 had initial ankle radiographs negative for fracture. Of these 32 patients, 8 (25%) had an occult fracture (lateral malleolus) diagnosed on follow-up radiograph as evidenced by new bone growth.

Non–weight bearing had a nonstatistically significant (P = 0.220) 75% sensitivity and 58% specificity while degree of ankle swelling was statistically significant (P < 0.0001) to predict an occult fracture.

Of the follow-up 24 non-fractures, 10 (42%) were non–weight bearing. The mean degree of ankle swelling was 6.48 mm (SD 2.14). Six of the eight occult fractures were non–weight bearing. The mean degree of ankle swelling was 11.01 mm (SD 1.964).

Using the computer-generated predictive model, the probability of an occult fracture was calculated as a function of mm swelling and then also calculated for mm swelling with weight-bearing status (Figure 1). To some extent, non–weight bearing or weight bearing did contribute to the predictive model, but the strongest predictor of occult fracture was mm ankle swelling.

From the negative radiographs of the initial visit, the occult fracture false-negative rate (FNR) and false-positive rate (FPR) were plotted as a function of the degree of ankle swelling (Figure 2). For example, from this model, a patient with a measured 4-mm radiograph ankle swelling would have a FNR of < 5% and FPR of approximately 85% for diagnosis of an occult fracture. Similarly, a patient with a measured 11-mm radiograph ankle swelling would have a FNR of 50% and FPR of < 10%.

Discussion

Approximately 1/5 of pediatric patients presenting to the ED with an acute ankle injury have an ankle fracture. The incidence of a distal fibula fracture with no fracture noted on the initial ankle radiograph (occult fracture) may occur in up to ¼ of children.

Presently, children are evaluated by tenderness at the site of injury or proximity to a physis, as well as to weight bearing or non–weight bearing, to help determine likelihood of an occult ankle fracture. The radiograph-measured magnitude of acute ankle swelling has, to our knowledge, not been used clinically to help predict likelihood of an occult fracture.

This retrospective study evaluated children presenting to our facility with an acute ankle injury. We measured ankle radiograph swelling in those with no evidence of an ankle fracture. The degree of swelling was found to be statistically significant for differentiating an occult fracture from non-fracture, whereas non–weight bearing vs weight bearing was not.

A computer model was written to predict the probability of an occult fracture based on degree of ankle swelling. While non–weight bearing vs weight bearing was not statistically significant, it did contribute to some extent in the model. However, lack of statistical significance of weight-bearing status may not remain true when the relationships can be estimated using a larger sample size.

Of note, there was a subpopulation of this study of nondisplaced lateral malleoli fractures in patients approaching skeletal maturity. We found that adolescence with closed tibial physis and open fibular physis populated our study group of ankle injuries initially with negative radiographs than any other population. Care must be taken in evaluating this subpopulation of the study group, especially with the abovementioned parameters of tissue swelling. Farther studies with larger cohorts may be needed to further assess the transitioning population.

The identification of fracture is important because the presence of a fracture may increase the recovery period compared with the absence of a fracture, and patients with a fracture may need more rigid immobilization compared with patients with an ankle sprain. Although an MRI scan is a viable option to differentiate between a sprain and a fracture, it comes with the associated costs and added time for diagnosis.15 Furthermore, an MRI is not readily available in all community hospitals, so it would not be feasible to perform on a routine basis. In comparison, our study highlights a potential quick and simple way to identify fracture patients.

We reviewed our non-fracture patients for weight and false-positive findings. We found no correlation between weight and false positives, but this could be a focus of research in the future. We understand this study has a small sample size, and we acknowledge it as one of the many limitations of this study.

Conclusion

In this study, when pediatric patients presented with an acute ankle injury and no fracture on initial radiograph, degree of acute ankle swelling on radiograph was predictive of an occult ankle fracture. Weight-bearing status did contribute somewhat to a predictive model but was not statistically significant. Probability curves of an occult fracture as a function of ankle swelling, and ankle swelling along with weight-bearing status were plotted. In addition, false-negative and false-positive prediction rates of an occult fracture based only on ankle swelling were generated.

Based on this study, radiograph measurement of the degree of ankle swelling at the time of initial ED presentation seems to be statistically significant for prediction of an occult ankle fracture. Weight-bearing status was not statistically significant to predict an occult fracture, but may help improve the predictive value of measured ankle swelling.

References 1. Sujitkumar P, Hadfield JM, Yates DW: Sprain or fracture? An analysis of 2000 ankle injuries. Arch Emerg Med 1986;3:101-106. 2. Brooks SC, Potter BT, Rainey JB: Inversion injuries of the ankle: Clinical assessment and radiographic review. Br Med J 1981;282:607-608. 3. Vargish T, Clarke WR, Young RA, Jensen A: The ankle injury – indications for the selective use of X-rays. Injury 1983;14:507-512. 4. Montague AP, McQuillan RF: Clinical assessment of apparently sprained ankle and detection of fracture. Injury 1985;16:545-546. 5. Sankar WN, Chen J, Kay RM, Skaggs DL: Incidence of occult fracture in children with acute ankle injuries. J Pediatr Orthop 2008;28:500-501. 6. Boutis K, Plint A, Stimec J, et al.: Radiograph-negative lateral ankle injuries in children: Occult growth plate fracture or sprain? JAMA Pediatr 2016;170:e154114. 7. Launay F, Barrau K, Petit P, Jouve JL, Auquier P, Bollini G: Ankle injuries without fracture in children: Prospective study with magnetic resonance in 116 patient. Rev Chir Orthop Reparatrice Appar Mot 2008;94:427-433. 8. Stiell IG, Greenberg GH, McKnight RD, Nair RC, McDowell I, Worthington JR: A study to develop clinical decision rules for the use of radiography in acute ankle injuries. Ann Emerg Med 1992;21:384-390. 9. Clark KD, Tanner S: Evaluation of the Ottawa ankle rules in children. Pediatr Emerg Care 2003;19:73-78. 10. Gravel J, Hedrei P, Grimard G, Gouin S: Prospective Validation and Head-to-Head comparison of 3 ankle rules in a pediatric population. Ann Emerg Med 2009;54:534-540.e1. 11. Dowling SK, Wishart I: Use of the Ottawa ankle rules in children: A survey of physicians’ practice patterns. Can J Emerg Med 2011;13:333. 12. MacLellan J, Smith T, Baserman J, Dowling S: Accuracy of the Ottawa Ankle Rules applied by non-physician providers in a pediatric emergency department. Can J Emerg Med 2018;20:746-752. 13. Clark TWI, Janzen DL, Ho K, Grunfeld A, Connell DG: Detection of radiographically occult ankle fractures following acute trauma: Positive predictive value of an ankle effusion. AJR Am J Roentgenology 1995;164:1185-1189. 14. Stiell IG, Greenberg GH, McKnight RD, Nair RC, McDowell I, Worthington JR: A study to develop clinical decision rules for the use of radiography in acute ankle injuries. Ann Emerg Med 1992;21:384-390. 15. Boutis K, Narayanan UG, Dong FF, et al.: Magnetic resonance imaging of clinically suspected Salter-Harris I fracture of the distal fibula. Injury 2010;41:852-856.