Session III - Femur/Knee

Heterotopic Ossification after Knee Dislocation: The Predictive Value of the Injury Severity Score

William J. Mills, MD; Nirmal Tejwani, MD, Harborview Medical Center, Seattle, WA

Purpose: We assessed the utility of the Injury Severity Score (ISS) in predicting the formation of clinically significant heterotopic ossification (HO) after knee dislocation.

Methods: Thirty-five patients with knee dislocation admitted in a 24-month period were enrolled in a prospective treatment protocol that included examination under anesthesia, evaluation with radiographs and MRI, multiple ligament reconstruction as soon as the soft tissue envelope safely allowed, postoperative physical therapy, and protected weight bearing. The ISS was calculated at the time of admission, and all additional injuries documented at initial presentation and secondary patient examination. Dislocations resulted from a motor vehicle accident (18), an automobile-pedestrian accident (11), a fall from more than three stories above ground (5), and a crush injury (1). Ten injuries were isolated and three were open. All patients enrolled in postoperative physical therapy.

Results: Ankylosing HO developed about the involved knee in six of thirty-five patients (17 %) whose ISS ranged from 26 to 50 (Group A). Group A patients had GCS scores ranging from 3T to 15. The ISS for the remaining 29 patients (Group B) ranged from 4 to 26. No Group-B patient experienced clinically significant HO. The average knee flexion arc for Group B was 123 at one year. All but one patient of Group B had a Glasgow Coma Scale (GCS) score of 15. The patient from Group A with an ISS of 26 and a GCS score of 9T had a significant closed head injury, with bifrontal contusion and a dural tear. The patient from Group B with an ISS of 26 had a GCS score of 15 and a normal CT scan of the head. HO formation among Group-A patients was evident clinically by marked loss of knee motion within 3 weeks of the surgical procedure and radiographically within 6 weeks. HO progression was relentless in all six patients, and no form of physical therapy influenced this progression.

Discussion: An ISS of 26 appears to be a fairly distinct tidemark for determining the patients at risk for HO formation after knee dislocation and ligament reconstruction; the sensitivity of an ISS of 26 is 100%, the specificity 97%, and the positive predictive value 86%. Although we describe a relatively small number of patients, it appears that an ISS of 26 represents a distinct line above which patients are at high risk of HO formation and below which it is very unlikely.

Conclusion: These findings have changed our practice management. Although knee dislocations are generally treated with open reconstruction as soon as local soft tissue status allows, those multiply injured patients with ISS >26 are treated with initial immobilization for soft tissue protection, then passive motion if co-existing injuries such as plateau fracture or patellar dislocation are not present. Delayed reconstruction is then planned more than 6 weeks after injury, a period after which it is generally apparent whether or not HO will form. Grossly unstable knees in which a congruent reduction is not possible without surgical stabilization have recently been treated with knee-spanning external fixation and prophylactic irradiation.