Comparison of Magnetic Resonance Imaging to

FOOT & ANKLE INTERNATIONAL Copyright  2011 by the American Orthopaedic Foot & Ankle Society DOI: 10.3113/FAI.2011.1110

Comparison of Magnetic Resonance Imaging to Physical Examination for Syndesmotic Injury After Lateral Ankle Sprain ´ Paulo César de César, MD; Eduardo Muller Avila, MD; Marcelo Rodrigues de Abreu, MD Porto Alegre, Brazil

Level of Evidence: II, Prospective Comparative Study

ABSTRACT

Key Words: Magnetic Resonance Imagery; Lateral Ankle Sprain; Physical Examination

Background: Clinical assessment of syndesmotic injury usually consists of two tests: the ankle external rotation test and squeeze test. This study sought to determine the sensitivity and specificity of both for syndesmotic injury secondary to lateral ankle sprain. Methods: Fifty-six patients with sprained ankles underwent clinical examination for syndesmotic injury with the aforementioned tests. Clinical findings were compared against magnetic resonance imaging (MRI) of the ankle. Sprains were graded on anatomical and functional classification scales, and correlation and agreement between both scales were assessed. Results: The MRI prevalence of syndesmotic injury in patients with lateral ankle sprains was 17.8%. Sensitivity and specificity were 30% and 93.5% for the squeeze test, and 20% and 84.8% for the external rotation test, respectively. Using the anatomical scale for sprain grading, 40% of syndesmotic injuries occurred in Grade I, 40% in Grade II, and 20% in Grade III sprains. Ten percent of patients with syndesmotic injury had no lateral ligament injury on MRI, 70% had injury of the anterior talofibular (ATFL) ligament, and 20% had injury to the ATFL and calcaneofibular (CFL). Conclusion: The sensitivity of the squeeze test and external rotation test was low, suggesting that physical examination often fails to diagnose syndesmotic injury. Conversely, specificity was very high; nearly all patients with a positive test actually had syndesmotic injury. Severity of ankle sprain was not associated with prevalence of syndesmotic injury.

INTRODUCTION

Syndesmotic injury occurs in 10% to 20% of ankle sprains,6,10 and is associated with distinct treatment approaches and increased time away from physical activity and sports involvement.5 The term “high ankle sprain” is commonly used to refer to isolated syndesmotic injury. In this study; however, we will focus on patients with lateral ankle sprains and concomitant syndesmotic injury, which poses a distinct diagnostic and therapeutic challenge. The stability of the tibiofibular syndesmosis is dependent on its bone structure and ligaments. Bony stability is provided by the fibula, located at the groove formed by the anterior and posterior tibial processes.7 The stabilizing ligaments of the syndesmosis are the anterior inferior tibiofibular ligament (AITFL), the posterior inferior tibiofibular ligament (PITFL), the inferior transverse ligament (ITL) and the interosseous ligament (IOL).7,22 Physical tests available for clinical examination of the syndesmosis include the squeeze test, the ankle external rotation test, the posterior fibular translation test (commonly known as the drawer test), the Cotton test, and the crossedleg test.4,13 We chose to evaluate two of these in the present study, the ankle external rotation test and the squeeze test, because positivity of these tests is associated with delayed return to pre-injury levels of activity.1,14 Ankle injuries account for approximately 10% of traumarelated emergency room visits; most of these injuries are sprains.8,9 Due to the high prevalence of these injuries, it is very important that the reliability of physical examination for diagnosis of syndesmotic injury in patients with ankle sprains be determined. In an investigation of syndesmotic injury, Takao et al.20 compared magnetic resonance imaging (MRI) with arthroscopic findings and showed that MRI has a sensitivity and specificity of 100% and 93%, respectively, for

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Corresponding Author: Paulo César de César, MD Hospital Mãe de Deus de Porto Alegre, RS, Brazil Department of Orthopedy and Traumatology Marques do Pombal, 250/402 Moinhos de Vento Porto Alegre Rio Grande do Sul 90540000 Brazil E-mail: [email protected] For information on pricings and availability of reprints, email [email protected] or call 410-494-4994, x232.

1110

Downloaded from fai.sagepub.com at UNIV CALIFORNIA SAN DIEGO on January 6, 2015

Foot & Ankle International/Vol. 32, No. 12/December 2011

AITFL injury and 100% sensitivity and specificity for PITFL injury. Using these findings as a starting point, we conducted a study to evaluate the sensitivity and specificity of physical examination (ankle external rotation test and squeeze test), as compared to MRI, for syndesmotic injury. To the best of our knowledge, this is the first attempt to assess the sensitivity and specificity of these two clinical tests for detection of syndesmotic injury secondary to acute ankle sprain. MATERIALS AND METHODS

All patients assessed in our outpatient clinic for ankle sprains in 2008 and 2009 were potential participants. Because ours is a reference clinic, all patients referred to us had already been seen at emergency departments. Initially, plain radiographs of the foot (AP and lateral views) and ankle (AP, lateral, and mortise view) were obtained. Patients with ankle or foot fractures were excluded from the study. Assessment of the syndesmosis on plain ankle radiographs consisted of measurement of the tibiofibular clear space and tibiofibular overlap (measured 1 cm proximal to the tibial plafond), with the tibiofibular clear space considered normal when it was less than 6 mm on AP view and tibiofibular overlap considered normal when it was more than 6 mm (or at least 42% of fibular width) on AP view or less than 1 mm on mortise view.12 Patients with diastasis of the syndesmosis visible on initial X-ray were excluded from the sample; therefore, all participants had normal ankle films. A total of 56 patients with ankle sprain were included in the study. Mean patient age was 32 ± 13 (range, 18 to 66) years. Mean time elapsed between onset of injury and physical examination was 6.6 ± 2.3 days. Patients were assessed for ankle sprains by the lead investigator (PCC). Sprains were first graded according to the standard anatomical classification17 and on a functional scale. On anatomical classification, sprains were considered Grade I when the anterior talofibular ligament (ATFL) was affected, Grade II when the ATFL and calcaneofibular ligament (CFL) were affected, and Grade III when the ATFL, CFL, and posterior talofibular ligament (PTFL) were affected; injury was diagnosed on the basis of tenderness to palpation of the affected ligament. Sprains were functionally classified into Grade I (mild injury, immediate pain followed by temporary relief, patient not forced to terminate activity, worsening of pain after rest, minor edema), Grade II (moderate injury, immediate, continuous pain, very difficult or impossible to resume activity, moderate edema, later onset of mild-tomoderate bruising), or Grade III (severe injury, immediate, continuous pain, patient immediately forced to terminate activity, inability to bear weight, major edema, extensive bruising). Specific clinical tests were then performed. The external rotation test consisted of stabilization of the leg, placement of the ankle in neutral position, and external rotation of the foot; the test was considered positive when the latter maneuver

EVALUATION OF THE SENSIBILITY ANKLE

1111

elicited anterolateral ankle pain (Figure 1).1 The squeeze test consisted of lateromedial compression at the transition between the middle and distal third of the leg (Figure 2). This test was considered positive when it elicited pain at the distal syndesmosis. No anesthesia was administered prior to clinical assessment of syndesmotic integrity. Patients were asked to describe the mechanism of injury, but due to the low reliability of responses and patient uncertainty concerning the circumstances surrounding their injury, we chose not to correlate mechanism of injury with the presence or absence of syndesmotic lesions. After physical examination, patients underwent MRI of the ankle. Scans were read by a radiologist (M.A.) specializing in musculoskeletal MRI. The radiologist was not made aware of the physical examination findings; in case of doubt, a second

Fig. 1: External rotation stress test of the foot.

Fig. 2: Squeeze test.

Copyright  2011 by the American Orthopaedic Foot & Ankle Society


1112

´ DE CESAR ET AL


specialist was asked to read the scan and both reached a consensus of their impressions. MR imaging was performed on a high-field (1.5T) scanner (GE Excite HD, Milwaukee, WI), with a T2-weighted fat-saturated sequence in the axial, coronal, and sagittal planes (TE, 40 to 60 ms; TR, 2500 to 2800 ms) and a T1-weighted sequence in the axial and sagittal planes (TE, 9 to 12 ms; TR, 400 to 550 ms). The syndesmotic, lateral collateral ligaments of the ankle and deltoid ligament were assessed. The criteria used diagnosis of the ligament injury were T1 and/or T2 signal changes in the area of any ligament, poorly defined ligament borders, or frank discontinuity (Figure 3). For statistical analysis, quantitative variables were expressed as means and standard deviations and categorical variables, as absolute and relative (percentage) frequencies. The Spearman correlation coefficient was used to test for associations between quantitative variables, and the kappa coefficient was used to assess agreement between the two injury grading scales. Fisher’s exact test was used for analysis

A

B

Fig. 3: Patient MRI.

of categorical variables. Measures of test performance were calculated and reported. The significance level was set at 0.05.

RESULTS

On the anatomical classification scale, Grade I sprains accounted for 18 patients (32.1%), Grade II for 24 patients (42.8%) and Grade III for 14 patients (25%). Divided by functional classification, nine patients (16%) were Grade I, 26 patients (46.4%) were Grade II, and 21 patients (37.5%) were Grade III. Agreement between the two classifications was weak (κ = 0.40, p < 0.001), but a strong correlation was identified (rs = 0.721, p < 0.001). Overall, 51 patients (91.1%) had lateral ligament injury (that is, injury to at least one of the three lateral collateral ligaments) confirmed on MRI. Five patients (8.9%) had no visible injury to the lateral ligaments on MRI, 13 (23.2%) had evidence of injury to one lateral ligament, 34 (60.7%) had injury to two lateral ligaments, and four (7.1%) had injury to all three lateral ligaments. The ATFL was affected in all cases of lateral ligament injury. The prevalence of syndesmotic injury was 17.8%; of the 56 patients, ten had evidence of syndesmotic injury on MRI. Of the 56 patients, six (10.7%) had a positive squeeze test and nine (16%) had a positive ankle external rotation test. Isolated assessment of the squeeze test showed a sensitivity of 30% and a specificity of 93.5%, versus 20% and 84.8%, respectively, for the ankle external rotation test. Assessment of both tests performed in combination, with physical examination considered positive for syndesmotic injury when at least one of the two tests was positive, yielded a sensitivity of 40% and a specificity of 84.8%. Analysis of the presence of syndesmotic injury on MRI according to anatomic classification of the sprain showed that 40% of injuries (n = 4) occurred in Grade I sprains, 40% (n = 4) in Grade II sprains and 20% (n = 2) in Grade III sprains. Comparison between MRI findings and functional classification showed that 10% of syndesmotic injuries (n = 1) were associated with Grade I sprains, 60% (n = 6) with Grade II sprains and 30% (n = 3) with Grade III sprains. Comparison of MRI findings of syndesmotic injury with presence of lateral ligament injury (also on MRI) showed that 10% of patients with syndesmotic injury (n = 1) had no lateral ligament injury, 70% (n = 7) had injury to one lateral ligament (the ATFL), 20% (n = 2) had injury of two ligaments and no patients with syndesmotic injury had lesions affecting all three lateral ligaments of the ankle. Of the ten patients with syndesmotic injury on MRI, two (20%) had MRI evidence of deltoid ligament injury, whereas six (13%) of the 46 patients with no syndesmotic injury on MRI had MRI evidence of deltoid injury. The difference was not statistically significant (p = 0.623).



Foot & Ankle International/Vol. 32, No. 12/December 2011 DISCUSSION

In the normal tibiofibular syndesmosis, the intermalleolar distance increases 1 to 1.25 mm with dorsiflexion of the talus, the fibula rotates externally by 2 degrees, translates distally 2.4 mm and is displaced 0.2 to 0.4 mm posteriorly.7,16,18 According to Ogilvie-Harris et al.,15 the AITFL, PITFL, ITL, and IOL account for 35%, 9%, 33% and 22%, respectively, of the total stability of the distal syndesmosis; therefore, rupture of two ligaments is enough to produce major mechanical instability. In a cadaver study, Bachmann et al.2 performed selective division of the syndesmotic injuries and attempted to forcefully open the articulation with a bone spreader. The authors were only able to induce frank syndesmotic diastasis after division of the IOL. Diagnosis of syndesmotic injury after an ankle sprain appears to be an important prognostic factor. In a study carried out at the United States Military Academy, Gerber et al.10 found syndesmotic involvement to be the single strongest predictive factor of chronic ankle dysfunction 6 months after a sprain. Boytim et al.,5 assessing the outcomes of 98 ankle sprains in a sample of professional athletes, found that 18 patients (18%) had syndesmotic injury, and noted that presence of syndesmotic injury was associated with an approximate six-fold increase in time to return to sports activities as compared to simple ankle sprains. The term “high ankle sprain” is often used to refer to a diagnosis of isolated syndesmotic injury. The present study assessed the presence or absence of syndesmotic injury in 56 patients with physical examination findings consistent with injury to the lateral ligaments of the ankle and normal plain radiographs (that is, no radiographic evidence of fractures or syndesmotic diastasis). MRI revealed that 17.8% of these patients had indeed sustained a syndesmotic injury. Some patients with sprained ankles go on to develop chronic symptoms, such as pain or subjective instability, after initial treatment. In a followup study of various sports injuries in high school athletes, Swenson et al.19 found the ankle to be the most common site in which chronic symptoms develop after acute injury. The chronic complaints that afflict some patients after acute ankle sprain may be due to misdiagnosed syndesmotic injury. In a study of chronic syndesmotic injury, Han et al.11 found that only 15% of patients with an arthroscopically confirmed diagnosis of syndesmotic injury had a positive external rotation test, and a mere 10% had a positive squeeze test. In our sample, which consisted of patients with lateral ankle injury, the prevalence of syndesmotic injury in association with ankle sprain, as diagnosed by the current gold standard method (MRI), was 17.8%. The ankle external rotation test and the squeeze test had sensitivities of 20% and 30% and specificities of 84.8% and 93.5%, respectively, for detection of syndesmotic injury. Joint assessment of both tests yielded a sensitivity and specificity of 40% and 84.8%, respectively. Physical examination requires sensitive clinical

EVALUATION OF THE SENSIBILITY ANKLE

1113

tests, that is, tests that are able to arouse clinical suspicion of injury in most cases in which injury is actually present. In this respect, both tests, whether alone or in combination, were not very sensitive. Clinical use of low-sensitivity tests most likely leads to underdiagnosis of syndesmotic injury in the urgent care and emergency room setting. Nevertheless, the specificity of both tests was high, meaning that patients are very likely to actually have a syndesmotic injury when one or both tests are positive. In our sample, syndesmotic injury was not more common in severe sprains. Syndesmotic injury occurred secondary to a Grade I or II sprain in 80% of cases using an anatomical classification or 70% of cases by functional classification. MRI assessment of the number of lateral ligaments injured showed that 70% of syndesmotic injuries occurred in patients with isolated ATFL injury, 20% occurred in patients with injury to two ligaments, and none occurred in patients with injury of all three lateral collateral ligaments. Uys and Rijke21 obtained similar results in their MRI-based study of syndesmotic injury, that is, syndesmotic injury was most commonly associated with milder sprains. This may be explained by the fact that, although both injuries often occur simultaneously, they are due to distinct mechanisms of trauma. Lateral collateral ligament injury occurs when the plantarflexed foot is forcefully inverted, injuring the ligaments in the anteroposterior direction. Conversely, syndesmotic injury is due to the action of external rotation forces on the foot more properly, when there is internal rotation of the leg over the fixed foot. Only one of the 10 patients with syndesmotic injury had no evidence of lateral ligament injury on MRI. Despite physical examination findings suggestive of a mild lateral ankle sprain, this patient had actually sustained a high ankle sprain (isolated syndesmotic injury). Therefore, 90% of patients in this sample with MRI evidence of syndesmotic injury also had some degree of injury of the lateral collateral ligaments, as confirmed by MRI. One potential point of criticism in this study is that stress radiographs of the ankle were not obtained. We chose not to perform stress tests based on a cadaver study by Beumer et al.,3 in which stress radiographs were unable to demonstrate syndesmotic separation after surgical division of the syndesmosis; the authors thus concluded that stress radiographs are not a reliable method for diagnosis of syndesmotic instability. The low sensitivity of clinical tests for syndesmotic injury was a striking finding of this study, and it implies that syndesmotic injury is being underdiagnosed in the urgent care setting. The question remains whether positive MRI findings after a missed clinical diagnosis of syndesmotic injury are associated with a poorer prognosis than isolated injury of the lateral ligaments. As MRI is an extremely sensitive method, it may sometimes reveal syndesmotic injury that is not severe enough to have an impact on clinical prognosis. Comparison between the anatomical and functional sprain grading scales showed high correlation (that is, increasing or



1114

´ DE CESAR ET AL


decreasing grades on one scale correlate with increasing or decreasing grade on the other) but poor agreement; despite this strong correlation, injuries are usually graded higher on the functional scale than on the anatomical classification. It should be noted that, although physical examination findings suggested lateral collateral ligament injury in all 56 patients in the sample, MRI showed that five patients (8.9%) had no such injury. One potential limitation of this study is that pain elicited at the level of the syndesmosis by the squeeze test or ankle external rotation test may be due to lateral ligament injury rather than to syndesmotic injury per se. This raises the possibility of positive physical examination findings with no true syndesmotic injury. The authors believe that, due to the mechanics of the test, the squeeze test is less likely to elicit pain caused by lateral ankle sprain than the ankle external rotation test. The present study shows that the ankle external rotation test and the squeeze test have low sensitivity and high specificity for diagnosis of syndesmotic injury in the presence of a lateral ligament injury. Our next research direction will be a followup study of patients with positive MRI findings to assess whether positivity or negativity of initial physical examination, with later treatment following a single protocol, has any effect on functional outcome. REFERENCES 1. Alonso, A; Khoury, L; Adams, R: Clinical tests for ankle syndesmosis injury: realiability and prediction of return to functional. J Orthop Sports Phys Ther. 27:276 – 284, 1998. 2. Bachmann, L; Seifert, C; Zwipp, H: Experimental and clinical diagnosis of ankle injuries with the syndesmosis spreader. In: Schmidt R, Benesch S, Lipke K, editors. Chronic ankle instability. Ulm (Germany): Libri: 235 – 238, 2000. 3. Beumer, A; Valtstar, ER; Garling, EH: External rotation stress imaging in syndesmotic injuries of the ankle: comparasion of lateral radiography and radiostereometry in a cadaveric model. Acta Orthop Scand. 74:201 – 205, 2003. http://dx.doi.org/10.1080/0001647031001 3969 4. Beumer, A; Van Hemert, WL; Swierstra, BA; Jasper, LE; Belkoff, SM: A biomechanical evaluation of clinical stress tests for syndesmotic ankle instability. Foot Ankle Int. 24:358 – 363, 2003. 5. Boytim, MJ; Fischer, DA; Neumann, L: Syndesmotic ankle sprains. Am J Sports Med. 19:294 – 298, 1991. http://dx.doi.org/10.1177/036354 659101900315

6. Cedell, CA: Ankle Lesions. Acta Orthop Scand. 46:425 – 445, 1975. 7. Close, JR: Some applications of functional anatomy of the ankle joint. J Bone Joint Surg Am. 38:761 – 781, 1956. 8. Cox, JS: Surgical and nonsurgical treatment of acute ankle sprains. Clin Orthop. 198:118 – 126, 1985. 9. Dunlop, MG; Beattle, TF; White, GK: Guidelines for selective radiological assessment of inversion ankle injuries. Br Med J. 293:603 – 605, 1986. http://dx.doi.org/10.1136/bmj.293.6547.603 10. Gerber, JP; Williams, GN; Scoville, CR; Arciero, RA; Taylor, DC: Persistent disability associated with ankle sprains: a prospective examination of na athletic population. Foot Ankle Int. 19:653 – 660, 1998. 11. Han, SH; Lee, JW; Kim, S; Suh, JS; Choi, YR: Chronic tibiofibular syndesmosis injury: the diagnostic efficiency of magnetic resonance imaging and comparative analysis of operative treatment. Foot Ankle Int. 28:336 – 342, 2007. http://dx.doi.org/10.3113/FAI.2007.0336 12. Harper, MC; Keller, TS: A radiographic evaluation of the tibiofibular syndesmosis. Foot Ankle. 10:156 – 160, 1989. 13. Kiter, E; Bozkurt, M: The crossed-leg test for examination of ankle syndesmosis injuries. Foot Ankle Int. 26:187 – 188, 2005. 14. Nussbaum, ED; Hosea, TM; Sieler, SD; Incremona, BR; Kessler, DE: Prospective evaluation of syndesmotic ankle sprains without diastasis. Am J Sports Med. 29:31 – 35, 2001. 15. Ogilvie-Harris, DJ; Reed, SC; Hedman, TP: Disruption of the ankle syndesmosis: biomechanical study of the ligamentous restraints. Artroscopy. 10(5):558 – 560, 1994. http://dx.doi.org/10.1016/S07498063(05)80014-3 16. Peter, RE; Harrington, RM; Henley, MB: Biomechanical effects of internal fixation of the distal tibiofibular syndesmotic joint: comparison of two fixation techniques. J Orthop Trauma. 8(3):215 – 219, 1994. http://dx.doi.org/10.1097/00005131-199406000-00006 17. Sammarco, G; Russo-Alesi, F: Ankle sprain: four treatment algorithms. Orthopedic Special Edition. New York, McMahon Group. 26, 1996. 18. Scranton, PE, Jr; McMaster, JG; Kelly, E: Dynamic fibular function: a new concept. Clin Orthop Relat Res. 118:76 – 81, 1976. 19. Swenson, D; Yard, E; Fields, S; Comstock, R: Patterns of recurrent injuries among US high school athletes, 2005-2008. Am J Sports Med. 37(8):1586 – 1593, 2009. http://dx.doi.org/10.1177/0363546509332500 20. Takao, M; Ochi, M; Oae, K; Naito, K; Uchio, Y: Diagnosis of a tear of the tibiofibular syndesmosis: the role of arthroscopy of the ankle. J Bone Joint Surg Br. 85:324 – 329, 2003. http://dx.doi.org/10.1302/ 0301-620X.85B3.13174 21. Uys, H; Rijke, AM: Clinical association of ankle lateral ankle sprain with syndesmotic involvement. Am J Sports Med. 30:816 – 823, 2002. 22. Xenos, JS; Hopkinson, WJ; Mulligan, ME; Olson, EJ; Popovic, NA: The tibiofibular syndesmosis: evaluation of the ligamentous structures, methods of fixation and radiographic assessment. J Bone Joint Surg Am. 77:847 – 856, 1995.

