Classifications in Brief: The Hawkins Classification for Talus Fractures.

Abstract

History As reported by Coltart [5], Fabricius von Hilden first described talus injuries in 1608, and in 1919, Anderson et al. [1] reported a series of talus fracture dislocations in which they emphasized dorsiflexion as the mechanism of injury and coined the term “aviator's astragalus” (astragalus being another term for talus) because of the impressive frequency with which the mechanism—aircraft accidents—was associated with the injury in question. In 1952, Coltart [5] reported on 228 talus injuries treated by surgeons in the British Royal Air Force in an attempt to describe the variety of injuries that occur to the talus and surrounding joints. He described various talus fracture patterns, fracture-dislocation combinations, and isolated peritalar dislocations [5]. Others have presented case series of talus injuries and associated rarely used classification schemes [15, 19], but all have noted high complication rates with injuries to the talus [5, 11, 15, 19]. In 1970, proposed treatment strategies for vertical fractures of the talar neck ranged from nonoperative management to urgent open reduction internal fixation, bone graft augmentation, fusion of the dorsal talus to the distal tibia, and even complete talectomy [11]. Although the importance of early, anatomic reduction of these injuries was noted at that time, the incidence of avascular necrosis (AVN) was not. In 1970 Hawkins [11] described injury patterns and AVN rates in a series of 57 vertical talar neck fractures in 55 patients from three different institutions. All 55 patients sustained forced dorsiflexion injuries, similar to the mechanism described by Anderson et al. [1] in pilots. Purpose At the time of Hawkins’ report in 1970, the incidence of AVN after vertical talus neck fracture was not known. The purpose of his publication was to describe vertical fractures of the neck of the talus, their common correlation with adjacent joint dislocations, and report the incidence of AVN in the body of the talus after these injuries [11]. Blood is supplied to the talus from the posterior tibial, anterior tibial, and peroneal arteries. The posterior tibial artery is the dominant arterial supply to the talar body via the arteries of the tarsal canal and the deltoid artery. The anterior tibial artery supplies the head and neck regions of the talus in combination with the perforating peroneal artery via the tarsal sinus. Hawkins recognized the importance of the blood supply to the talus and expected perfusion disruption with different injury patterns as responsible for AVN [11]. He did not report on comminuted fractures or those involving the body of the talus, as he believed these to be more problematic injuries and outside the scope of his article [11]. Description of Hawkins Classification The original Hawkins classification includes three fracture types: Type I through Type III. The modern version includes Type IV, which was added by Canale and Kelly in 1978 [2] (Fig. 1).Fig. 1A-D: Fractures of the neck of the talus based on the modified Hawkins classification are shown. A (A) Type I fracture is a talar neck fracture, (B) Type II is a talar neck fracture with subtalar dislocation, (C) Type III is a talar neck fracture with subtalar and tibiotalar dislocations, and (D) Type IV is a talar neck fracture with subtalar, tibiotalar, and talonavicular dislocations. (Reprinted with permission, University of Washington Creative Services© 2014. All Rights Reserved).Type I injuries include vertical fractures of the neck of the talus with a minimally displaced fracture line entering the subtalar joint between the middle and posterior facets. The talus retains its anatomic position in the ankle and the subtalar joint remains reduced (six of 57 fractures) [11]. Hawkins suggested that only one of the three blood vessels to the talus is disrupted with these injuries and reported no cases of postinjury AVN [11]. Type II fractures are vertical, displaced, with the subtalar joint subluxated or dislocated, but the talus remains reduced in the ankle mortise (24 of 57 fractures) [11]. Fractures commonly extend into the body of the talus and the posterior facet, and, in Hawkins’ series, resulted in posterior dislocations of the subtalar joint in 10 of 24 fractures [11]. Union was achieved in all 24 fractures but AVN was seen in 42% (10 of 24) and Hawkins noted a likely disruption of two of the three main sources of perfusion to the talus as the cause [11]. Type III fractures are similar in characteristics to Type II, with the addition of a dislocation of the ankle (tibiotalar articulation) (27 of 57 fractures) [11]. Hawkins reported a common posteromedial dislocation pattern with the extruded body located between the posterior tibial and Achilles tendon [11]. The head of the talus remained reduced to the navicular bone and a 91% AVN rate was reported (20 of 22 fractures) [11], likely owing to disruption of all three major blood supply vessels to the injured talus. In 1978, Canale and Kelly [2] added a fourth group not previously described by Hawkins (Type IV), in which the fracture of the talar neck was associated with dislocation of the body from the ankle or subtalar joint with additional dislocation or subluxation of the head of the talus from the talonavicular joint [2]. Canale and Kelly had a series of patients with 71 fractures with an average 13-year followup (range, 2.5-32 years) and an AVN rate of 52% (37 of 71) [2]. Validation In his initial description of the classification, Hawkins described three groups of patients with differing fracture patterns and adjacent joint disruptions based on plain radiographic appearance at the time of injury. The primary goal of his work was to correlate these injury types with AVN development. In 2013, Halvorson et al. [10] validated the rates of AVN based on the modified Hawkins injury classification [2]. They performed a large retrospective review of 848 talus fractures and reported an overall rate of AVN of 33.3% (282 of 848): 5.7% with Type I (nine of 159), 18.4% with Type II (58 of 314), 44.7% with Type III (102 of 228), and 12.1% with rare Type IV injuries (four of 33) [10]. It is possible that modern surgical implants, modern operative techniques, surgical timing, availability and implementation of advanced imaging modalities, and larger cohorts of patients are responsible for the discrepancy in rates of AVN between Hawkins’ cohort [11] and that of Halvorson et al. [10]. Hawkins was the sole author on his landmark 1970 paper [11]; as such, he was the only observer to classify the injury patterns he described, therefore he could not provide interobserver reliability data, and he also did not provide intraobserver reliability. The differentiating factor between fracture types described by Hawkins is the presence of adjacent joint disruption observed on plain radiographs but this is open to observer interpretation and leads to variability. To evaluate the reproducibility of the Hawkins classification with the modification by Canale and Kelly [2], Drummond Filho et al. [7] had 36 physicians evaluate the injury radiographs of 20 talar neck fractures on two separate occasions and classify these injuries. They reported kappa scores of 0.627 and 0.668, improving as level of medical training increased, concluding the classification to be reproducible [4, 7, 14]. Evaluators ranged from first-year orthopaedic and radiology trainees to well-experienced doctors in the same fields. Participants knew they would be evaluated twice and were allowed to study the classification criteria before each session, possibly explaining the elevated kappa values reported. Haapamaki et al. [9] cited the known difficulty in understanding the complex anatomy of the talus observed on plain radiographs as the cause for challenges classifying these fracture dislocations. With the advent and near-ubiquitous implementation of advanced imaging, CT now is commonly used in the diagnosis and preoperative planning for patients with talus fractures [8, 9, 12, 13, 20]. CT was not available at the time of Hawkins original description, leading Williams et al. [21] to evaluate the inter- and intraobserver reliabilities of the Hawkins classification using radiographs and CT scans. Four providers evaluated 39 talus fractures and found the inter- and intraobserver reliabilities to improve with CT evaluation of talus fractures for most parameters studied. Interobserver reliability averaged κ = 0.356 with radiographs compared with κ = 0.590 with CT scans. Intraobserver reliability also improved from an average κ = 0.477 with radiographs to κ = 0.756 with CT scans [21]. Based on additional detail provided by advanced imaging, Williams et al. proposed differentiating the talar neck from talar body fractures by defining body fractures as those involving the tibiotalar joint and the subtalar joint; neck fractures are anterior to the joints [21]. Limitations When Hawkins reviewed his cohort of patients with talar neck fractures, validated, reliable and responsive outcomes measures were a common aspect of orthopaedic literature. Although Hawkins briefly reported his patients’ outcomes (excellent, good, fair, poor), the primary goal of his work was to correlate vertical fractures of the talus and adjacent joint dislocations with the development of AVN [11]. Hawkins recognized Type II injury patterns were challenging, noting the physician “cannot predict whether avascular necrosis will occur” with these injuries [11]. Canale and Kelly also were unable to identify a way to predict which Type II injuries would progress to have AVN develop [2]. In 2014, Vallier et al. [18] proposed that the extent of subtalar displacement at the time of injury is responsible for the variable rates of AVN after a Hawkins Type II injury. They suggested further subdividing Type II injuries into those in which the subtalar joint is subluxated (Type IIA) and those in which the subtalar joint is dislocated (Type IIB), and hypothesized that AVN rates would increase with greater subtalar displacement [17]. In reviewing 80 patients with 52 talar neck fractures, they identified 19 Type IIA and 16 Type IIB injuries treated with open reduction internal fixation and 12 months followup. None of the patients with Type I or IIA injuries had AVN develop whereas 25% (four of 16) with Type IIB injuries and 41% (11 of 27) with Type III injuries had AVN develop [18]. This observation supports the hypothesis that the deltoid artery is preserved when the subtalar joint is subluxated (Type IIA) but is disrupted with complete subtalar dislocation (Type IIB). Additionally, at the time of Hawkins’ work, CT and MRI were not available. Hawkins relied on observation of the presence of subchondral atrophy on the AP radiograph between 6 and 8 weeks after injury (known as the Hawkins sign [6]) to predict perfusion of the talus. He found it was a positive prognostic indicator in 15 of his patients [11]. Canale and Kelly confirmed the utility of the sign but noted that the absence of subchondral atrophy was not a universally reliable indicator of the development of AVN [2]. Tezval et al. [16] reviewed 41 patients with talus fractures and found the Hawkins sign to have 100% sensitivity for talar viability; however, the specificity of the sign to rule out the possibility of talar AVN was limited—only 57.7% [16]. In 2014, Chen et al. [3] found the Hawkins sign to be a reliable predictor of AVN development and suggested that MRI at 12 weeks might be valuable to identify early AVN in patients who have a negative Hawkins sign. Although osteonecrosis is the most common complication after a talar neck fracture, other complications frequently occur that affect patient outcomes and were not reported by Hawkins. Traumatic arthritis of the ankle and subtalar joint were found by Canale and Kelly to be a significant problem after talus fracture, occurring in more than 1/2 of their patients, 1/3 of whom had a malunion [11]. Vallier et al. [17] reviewed 100 talus fractures and found other complications, including osteonecrosis with collapse (31%), ankle osteoarthritis (18%), and subtalar osteoarthritis (15%). Operative intervention was complicated by superficial (3.3%) and deep infection (5%), wound dehiscence (3.3%), delayed union (1.7%), and nonunion (3.3%) [17]. They also found a substantial association between talar neck comminution and osteonecrosis, which resulted in worse functional outcomes as measured by mean musculoskeletal function scores [17]. The overall rate of posttraumatic arthritis was 54%; the subtalar joint was involved 38% of the time, and the overall rate of postfracture AVN was 25%. Conclusions and Uses The Hawkins classification, and its subsequent modification by Canale and Kelly, frequently is used to describe talus fractures [2, 11]. The injury type has been found to correlate with the development of posttraumatic AVN, with more severe injuries resulting in higher AVN rates. The modified Hawkins classification is based on plain injury radiographs, is simple, and reliable and can be used to predict the development of posttraumatic AVN after talar neck fracture. The Hawkins sign is a dependable indicator of talar perfusion and a positive prognostic sign.