Isolated syndesmotic injury in acute ankle Trauma: comparison of plain film radiography with 3T MRI

Abstract

OBJECTIVES: To determine cut-off values and the accuracy of plain film measurements for the detection of isolated syndesmotic injury after acute ankle trauma and to investigate MRI findings of concomitant ankle injury.

METHODS: Eighty-four consecutive patients with absent fracture in plain film radiographs were prospectively evaluated for isolated syndesmotic injury after acute ankle trauma. The tibiofibular clear space (TFCS), the tibiofibular overlap (TFO), and the medial clear space (MCS) were independently assessed in plain radiographs by two readers. MRI performed at 3T within 24h served as the reference standard. MRI was evaluated for syndesmotic injury, using a four-scale grading system (0=normal syndesmosis, 1a=periligamentous edema, 1b=intraligamentous edema, 2=partial rupture, 3=complete rupture), and for concomitant ankle injury. Inter-observer variability for x-ray measurements was assessed using Bland-Altman diagrams. ROC analyses were performed to determine cut-off values and sensitivity and specificity for TFCS, TFO, and MCS.

RESULTS: Eleven of 84 patients (13.1%) revealed syndesmotic injury (Grade 2 or 3) according to MRI. Between patients with and without syndesmotic injury significantly different measurements were obtained for TFCS (p=0.003) and MCS (p=0.04). ROC derived cut-off values were 5.3mm for TFCS, 2.8mm for TFO, and 2.8mm for MCS. Sensitivity and specificity was 82% and 75% for TFCS, 36% and 78% for TFO, and 73% and 59% for MCS. The bias and limits of agreement were -0.04 mm and [-1.54; 1.53] for TFCS, 0.8mm and [-2.5; 2.5] for TFO, and 0.05 mm and [-1.42; 1.43] for MCS. Patients with syndesmotic injury had a 5-fold increased risk of concomitant ankle injury (p=0.07).

CONCLUSIONS: The determined cut-off values aid in the evaluation of syndesmotic integrity in patients with absent fracture in plain radiographs. In case of increased distances MRI is recommended to assess severity of SI and to reveal associated ankle injuries.

Bibliografische Daten

OriginalspracheEnglisch
ISSN0720-048X
DOIs
StatusVeröffentlicht - 01.10.2014
PubMed 25043987