Introduction

Multi-ligament knee injuries (MLKI) are rare but potentially devastating injuries that can result in considerable functional limitation [27]. These injuries can occur following a variety of mechanisms including high-velocity accidents, sport-related injuries, and ultra-low velocity ground level falls [23, 33, 39]. When MLKI occur as a result of a high-velocity mechanism, nearly, one-third of patients present with a life-threatening injury to the head, chest, or abdomen [5]. In addition, 60% of patients sustain an associated fracture and 41% have multiple fractures [38]. Poly-traumatized patients are commonly assessed at presentation using Trauma Scoring Systems which have demonstrated correlation with both short-term mortality and functional recovery [2, 14, 35].

Surgical reconstruction of MLKIs can provide substantial functional improvement [31]. However, significant variability in outcomes remains amongst individual patients [25]. Over the past decade, efforts have focused on identifying potential factors that can predict a poor outcome. Patient factors including age [25], sex [20] and body mass index (BMI) [7] as well as surgical factors including repair versus reconstruction [22] and timing [28] have all been correlated with outcome. However, despite recognizing that MLKIs commonly occur as a component of a more severe poly-traumatic injury, no study to date has assessed the impact that these associated injuries have on a patient’s long-term function.

This study compares the post-reconstruction functional outcomes of knees with similar multi-ligament injury patterns that occurred as a component of poly-trauma to those that occurred in isolation. The cohorts were matched for age, sex, and KD classification score. The hypothesis was that poly-trauma patients would have lower functional scores despite restoration of similar knee range of motion and stability. This finding would highlight the influence concomitant injuries and psychosocial factors have on the functional outcomes of patients with MLKI sustained as a component of polytrauma.

Materials and methods

A prospective registry was created in 2007 dedicated to tracking patient outcomes after multi-ligament knee reconstruction. Patients who had injuries before 2007 were retrospectively added to the database. Three hundred and two consecutive patients who underwent surgical reconstruction for a multi-ligament knee injury between 1993 and 2014 were retrospectively reviewed. Injury patterns were evaluated using the new injury severity score (NISS). The NISS rating scale is from 0 to 75 points (most severe). A score of 16 or greater was classified as a poly-traumatic injury [9]. Forty-three (15.5%) patients reviewed met this criterion (Fig. 1).

Fig. 1
figure 1

Diagram detailing patient search and inclusion

Full size image

All patients underwent surgical intervention for their multi-ligament knee injury by one of the two senior surgeons. Ligamentous reconstruction and/or repair were performed at the discretion of the operating surgeon. Consistent tunnel placement, graft preparation, and graft fixation were performed [23]. Neurovascular status was assessed using previously described methods [13, 30].

The patients meeting criteria for poly-trauma (NISS > 15) were matched against patients with isolated knee injuries from our multi-ligament database by age, sex, and modified Schenck classification [32]. There were no major differences in regards to group characteristics (Table 1). In the poly-trauma cohort, four patients (13%) suffered complete peroneal nerve injuries and four patients (13%) suffered vascular injuries. One patient (3%) suffered a combined peroneal nerve and vascular injury. All patients who suffered a vascular injury underwent popliteal artery bypass grafting. Of the four patients with complete peroneal nerve injuries, one patient underwent a nerve transfer procedure, one patient underwent a nerve transfer with a posterior tibial tendon transfer, and two patients were prescribed an ankle–foot orthosis (AFO). In the control group, three patients (10%) suffered complete peroneal nerve injuries and three patients (10%) suffered vascular injuries. One patient (3%) suffered a combined peroneal nerve and vascular injury. All patients who suffered a vascular injury underwent popliteal artery bypass grafting. Of the three patients with complete peroneal nerve injuries, two patients underwent a nerve transfer procedure and one patient was prescribed an ankle–foot orthosis (AFO).

Table 1 Summary of poly-trauma and control cohorts
Full size table

The mechanism of injury in the poly-trauma group was variable: 25 (81%) motor vehicle collisions (MVC), 3 (9.5%) all-terrain vehicle (ATV) collisions, 1 (3.2%) industrial accident, 1 (3.2%) horse trampling, and 1 (3.2%) fall from a tree. Concomitant injuries ranged from tibial plafond fracture requiring amputation to aortic dissection requiring vascular repair. All patients were given a Glasgow Coma Score (GCS) upon admission to the emergency department (ED). Each patient’s chart was assessed for the types of concomitant injuries, intensive care unit (ICU) admission, duration of stay, and endotracheal intubation. Patients were also classified by NISS score then stratified into severe (16–25) and profound (> 25) [5]. Patient outcomes were assessed with physical exam by one of the two surgeons (MJS, BAL), International Knee Documentation Committee (IKDC) and Lysholm scores [4, 19]. This study was approved by the Mayo Clinic Institutional Review Board (IRB # 07-004018) prior to commencement.

Statistical analysis

Descriptive analyses of patient characteristics were performed using the mean and standard deviation for continuous variables along with the percentages and frequencies for dichotomous variables. Comparisons of patient characteristics between groups were conducted using the Wilcoxon-signed ranked sums for continuous variables and Chi-square or Fisher’s exact test for categorical variables. Patients in the poly-trauma group were randomly matched by age, sex, and KD grade to non poly-trauma patients using the SPSS software (Version 18.0; SPSS, Chicago, IL, USA). Final patient outcome scores (Lysholm, IKDC) were assessed using a paired t test. Risk factors for worse outcomes in the poly-trauma group were analyzed using Wilcoxon-Signed Ranked sums test. All statistical tests were two sided with a p value of < 0.05 being considered significant. Statistical analyses were completed employing the JMP software (Version 7, SAS 135 Institute, Inc., Cary, NC, USA).

Results

Thirty-one patients in the poly-trauma group had a mean age of 34.0 years (range 18–52) and were followed for a mean of 67 months (range 24–220). Mean NISS score was 40.9 ± 20.4. Associated abdominal injuries were most common followed by lower extremity fractures. Eighteen patients (58%) had injuries that were categorized as profound by NISS scale (> 25). Fifteen patients required admission to the intensive care unit (ICU) with a mean duration of stay of 2.8 days (range 1–15 days). Median GCS on arrival was 12 (range 3–15), with 13 patients (42%) requiring intubation on arrival in the ED (Table 2).

Table 2 Summary of associated injuries for poly-trauma group
Full size table

The mean IKDC and Lysholm scores in the poly-trauma group were 57.2 ± 21.9 and 62 ± 22, respectively. In comparison, the control cohort measured 71.1 ± 26.5 and 78 ± 23, respectively. Patients in the poly-trauma cohort had significantly lower IKDC and Lysholm scores (p = 0.01 and p = 0.003, respectively). On physical exam, there were no differences between final ROM between the two groups (Table 3). In addition, there were no major differences between the groups in objective tests of knee stability including: pivot shift, Lachman, posterior drawer test, varus stress at 0° and 30°, Valgus stress at 0° and 30°, or external rotation drawer test (Table 3).

Table 3 Functional outcomes of poly-trauma versus isolated knee injury cohort
Full size table

Complications

Three patients (9.6%) in the poly-trauma cohort suffered post-operative infections requiring arthroscopic debridement and intravenous antibiotics. Four patients (12.9%) developed arthrofibrosis requiring manipulation under anesthesia, and one patient (3.2%) had a failed surgery requiring revision 15 months after primary surgery. Two patients (6.4%) in the control cohort developing arthrofibrosis requiring MUA and one patient (3.2%) had a wound infection that required revision surgery.

Risk factor analysis

Risk factors for worse outcomes were analyzed in the poly-trauma group. Age, sex, KD grade, body mass index (BMI), NISS category, associated injuries, ICU stay, intubation requirement, and GCS score on admission were assessed. No significant risk factors were found to be predictive of worse patient reported functional scores.

Discussion

The most important finding of the present study was that the post-reconstruction functional outcomes of patients who sustained a MLKI as a result of poly-trauma were significantly lower than those that occurred in isolation. This occurred despite controlling for age, gender, and pattern of ligament injury. Furthermore, the post-operative knee stability examination and range of motion were similar between the two cohorts. This information highlights the influence that concomitant injuries and psychosocial factors can have on the functional outcomes following MLKI reconstruction in poly-traumatized patients.

The two cohorts matched according to age, gender, and knee dislocation classification were compared using IKDC and Lysholm scores as measures for functional outcome. The mean IKDC score was 57.2 points in the poly-trauma injured cohort in comparison with 71.2 (p = 0.01) for the isolated knee injury group. This difference is not only statistically significant, but more than doubles the minimal clinically important difference (MCID) for IKDC (6.3 points) [16]. A clinically and statistically significant difference was also found in regards to the Lysholm score with a 16-point difference (MCID = 10.1) [3] between the two cohorts (62 versus 78; p = 0.003).

It is well accepted that multi-ligament knee injuries commonly occur in conjunction with poly-trauma [10, 26]. Detailed treatment algorithms have been proposed to address both life-threatening and limb-threatening injuries [11, 18, 23, 36]. Poor outcomes in poly-trauma patients with a knee dislocation have been attributed to worse ligament injury patterns, the presence of a popliteal artery injury or increased rates neurologic injury [12, 37, 40]. However, the patients in this study were matched based on their ligament injury and there was no identifiable difference between the two groups regarding the rate of vascular or neurologic involvement. The absolute numbers demonstrated four patients in the poly-trauma cohort compared to three patients in the isolated knee injury cohort suffered complete common peroneal nerve palsy. Furthermore, it has previously been shown that common peroneal nerve palsy may not affect long-term functional outcomes following knee dislocation [21].

The timing of surgical reconstruction is an important consideration in multi-ligament reconstruction surgery. Acute single-stage reconstruction has been shown to lead to improved functional outcomes following multi-ligament knee injuries in comparison with delayed or staged procedures [10, 23, 24, 28]. The delay in knee ligament reconstruction because of associated life-threatening injuries has been blamed for poor outcomes in poly-trauma patients. In this study, there was no difference in the number of patients who underwent acute or staged reconstruction (Table 1) between the poly-trauma group (6 acute, 25 chronic) and the isolated knee injury group (4 acute, 27 chronic) (n.s.).

There was no difference in the pattern of ligament injury, the rate of neurologic injury, the rate of vascular injury, or the number of patients reconstructed on a chronic basis (> 3 weeks after injury). Physical examination demonstrated no difference in range of motion or ligament stability between the two groups. This information suggests that associated injuries unrelated to the knee influence functional outcomes in poly-trauma patients. This theory is supported by the trauma literature, where the Injury Severity Score or New Injury Severity Score is used to predict morbidity and mortality at the time of presentation to hospital [29].

The ISS was first described by Baker et al. [1] in 1974 and has since become a commonly used anatomic scoring method for assessing injury severity [2, 8, 15, 17]. This score is calculated using Abbreviated Injury Scale (AIS) scores in each of the three most critically injured body regions. This was later modified in 1997 by Osler et al. [29] to the New Injury Severity Score. This modification continues to use the AIS, but allows calculation using the three most critical injuries regardless of body region. As an example, if a patient sustained bilateral femur fractures and a knee dislocation, the ISS would only consider only the most severe of these three injuries when determining the lower extremity score. Alternatively, in the NISS, each of these injuries would be considered individually if they represent the three most severe injuries that the individual has sustained. The NISS was used for inclusion over the ISS as it allows extremity injuries to achieve greater influence when multiple severe injuries are preset. The NISS better predicts the functional outcomes in survivors of musculoskeletal trauma than the ISS [34, 35].

The current study demonstrated very severe injuries in the poly-trauma cohort. According to the NISS, 18 (58%) of patients had scores greater than 25 points representing profound injuries. In addition, 15 (48%) of patients required admission to the intensive care unit (ICU) and 13 (42%) required intubation during their admission. Associated injuries affected the head, chest wall, abdomen, solid organs, spine, pelvic ring, and upper and lower extremity fractures. Each of these factors was analyzed as a risk factor for a poor outcome; none reached statistical significance. This is likely the result of significant heterogeneity observed amongst the individual patients regarding injury location, patterns, and severity.

The following limitations should be considered when interpreting the results reported in this article. First, the reported data are a retrospective analysis of a prospectively collected database which introduces inherent bias and may limit the generalizability of the findings. Second, there is significant heterogeneity amongst the poly-trauma cohort regarding the location and severity of the injuries sustained. This did not allow determination if specific associated injuries were more likely to cause long-term functional deficits. Third, variability exists in all patients regarding associated meniscal injuries and cartilage damage. This was not assessed in either cohort of this study. Last, when time to reconstruction was assessed as a continuous variable, there was a significantly greater delay to reconstruction in the poly-trauma cohort when compared to the isolated knee injury cohort at 307.8 versus 107.8 days, respectively. The influence of surgical delay once a patient is being treated on a chronic basis is not clear. This is a potential bias influencing the difference in functional outcomes between the two groups. Last, this study did not assess for the influence of low self-efficacy, lower education level, poverty, or involvement in disability or compensation litigation that have been shown to affect outcomes [6].

This study also has several important strengths that should be noted. This is the first series to assess the influence of poly-trauma on long-term functional outcomes in patients with a knee dislocation. All patients in each of the matched cohorts were treated by a single surgical team and subsequently underwent a standardized rehabilitation protocol. Minimum 2-year follow-up was obtained for all patients.

There is currently considerable variability in functional outcomes following MLKI reconstruction in the literature. The independent influence that poly-trauma can have on functional outcomes may account for some of this variability. Future research should identify poly-traumatized patients using a validated Trauma Scoring System such as the NISS. Poly-traumatized patients should be assessed independently from patients with isolated knee injuries. In the isolated knee injury population, unbiased and more predictable outcomes will be reported. However, more importantly, selective evaluation of poly-traumatized patients will allow determination of which concomitant injuries have the greatest influence on functional outcomes. In addition, psychosocial barriers can be more easily identified and resources can be selectively allocated to this population. Multi-center collaboration will likely be required due to the challenges encountered in the management of these rare and heterogeneous injuries.

Conclusion

When comparing knees with similar multi-ligament and neurovascular injury patterns, patients who sustained their injury as a result of poly-trauma demonstrated significantly lower functional scores following reconstruction. This is despite restoration of similar knee stability and range of motion. The functional outcomes following multiple knee ligament reconstruction in poly-traumatized patients are influenced by factors other than the knee including concomitant injuries and psychosocial factors.