Incidence of and Risk Factors for Knee Collateral Ligament Injuries With Proximal Tibia Fractures: A Study of 32,441 Patients

Materials and Methods

The NTDB Research Data Set for admission years 2011 and 2012 was used for this study. The NTDB is the largest national database of trauma patients, containing more than 1.6 million cases from more than 900 level I to IV trauma centers during years 2011 and 2012.¹² Although all data in the NTDB are de-identified, the authors obtained a waiver for this study from the human investigations committee of their institution.

All patients with proximal tibia fractures were identified in the NTDB using International Classification of Diseases, Ninth Revision (ICD-9) diagnosis codes 823.00, 823.01, 823.10, and 823.12, representing closed fractures of the proximal tibia. Open fractures were excluded from this study because they may inaccurately skew the data for length of stay (LOS) and in-house complications, given that patients may need multiple trips to the operating room for serial debridement, wound coverage, or infections. The status of patients' collateral ligaments was then determined using the presence vs absence of ICD-9 diagnosis codes (844.1—MCL; 844.0—LCL; 836.50–836.69—complete knee dislocation). Pediatric patients (younger than 18 years) were also excluded from the study population.

Associations of collateral ligament injures with sex, age, Charlson Comorbidity Index (CCI), and 8 specific comorbidities (alcoholism, cancer, chronic respiratory disease, congestive heart failure, coronary artery disease, diabetes mellitus, hypertension requiring medication, and obesity) were analyzed in this study. These comorbidities were identified using NTDB-specific data elements, obtained via chart abstraction by local trauma data registrars. Mechanism of injury was determined using ICD-9 e-codes reported in the NTDB. Injury Severity Score (ISS) is directly reported in the NTDB. Associated injuries were also determined based on ICD-9 codes (Table 1).

Table 1: Codes Used for Characterization of Associated Injuries and Mechanisms of Injuries

A modified CCI¹³ was calculated to summarize patient comorbidities; however, this was not used in the multivariate analysis because it correlated with age and the individual comorbidities being studied in this analysis. Computation of the modified CCI is detailed in Table 2.

Table 2: Comparison of Modified Charlson Comorbidity Index With Original

Inpatient adverse events are reported in the NTDB using specific chart-reviewed variables. Aggregate surrogate variables of serious inpatient adverse events (SAEs), minor inpatient adverse events (MAEs), and any inpatient adverse events (AAEs) were then created. Serious inpatient adverse events include acute respiratory distress syndrome, cardiac arrest, death, myocardial infarction, severe sepsis, stroke, thromboembolic events, and unplanned return to the operating room. Minor inpatient adverse events include acute kidney injury, alcohol or drug withdrawal, bloodstream infection, compartment syndrome, decubitus ulcer, osteomyelitis, pneumonia, surgical site infection, unplanned intubation, unplanned return to the intensive care unit, and urinary tract infection. Total hospital LOS was also analyzed as an outcome variable.

Sex, age, CCI, comorbidities, mechanism of injury, ISS, and associated injuries were compared between patients with and patients without MCL or LCL injuries using Pearson's chi-square test. Multivariate logistic regression was used to identify associated patient characteristics for those with collateral ligament injury and proximal tibia fractures. Obesity was the only comorbidity associated with collateral ligament injuries on bivariate analysis. The following variables were included in the multivariate analysis: sex, age, obesity, mechanism of injury, ISS, and specific other lower extremity fractures, including proximal, midshaft, and distal femur fractures; midshaft tibia/fibula fractures; and distal tibia/fibula fractures. Continuous variables, age and ISS, were converted into discrete categorical variables for this analysis. Finally, multivariate logistic regression was used to determine the association of collateral ligament injury with SAEs, MAEs, AAEs, and hospital LOS after controlling for both CCI and ISS.

All statistical analyses were conducted using Stata version 13.0 statistical software (StataCorp LP, College Station, Texas). All tests were 2-tailed, and the Bonferroni correction was used to account for the multiple statistical tests performed. Based on 18 different hypotheses tested, the level of statistical significance was set at α=0.005.

Results

A total of 32,441 patients with proximal tibia fractures were identified and met inclusion criteria. Of the included patients, 1445 (4.5%) had collateral ligament injuries recorded (Table 3). Injuries to both the MCL and the LCL were most common (2.4%), followed by MCL only (1.4%) and LCL only (0.6%). Of the patients with plateau fractures, 2.2% had documented knee dislocation diagnoses. Of the 1445 patients with collateral ligament injuries, 21% underwent a collateral ligament repair during their initial hospitalization. Although ICD-9 coding does not allow for anterior cruciate ligament injuries to be distinguished from posterior cruciate ligament injuries, as a comparison, the authors found that 3.2% of all tibial plateau fractures had a cruciate ligament injury; 1.8% of the patients without collateral ligament injuries and 30.5% of those with collateral ligament injuries also sustained a cruciate ligament injury.

Table 3: Incidence of Collateral Ligament Injuries With Proximal Tibia Fractures (n=32,441)

Characteristics differed between patients with and patients without collateral ligament injuries (Table 4). Specifically, patients with collateral ligament injuries were more likely to be male, younger (mean age, 41 vs 50 years), and obese. They also had a lower CCI (1.1 vs 2.0). The difference between 1.1 and 2.0 on the CCI scale is nearly equivalent to a single additional comorbidity such as congestive heart failure, myocardial infarction, or chronic obstructive pulmonary disease or 10 years in age. This is clinically relevant in this study, where approximately two-thirds of the population had a CCI of 0 to 2. With increasing age, patients were less likely to suffer collateral ligament injury.

Table 4: Comparison of Patient Demographics and Comorbidities

Injury characteristics also differed between patients with and patients without collateral ligament injuries (Table 5). Patients with collateral ligament injuries were more likely to have higher-energy mechanisms of injury (P<.001). The most common mechanism of injury for both patients with and patients without collateral ligament injuries was motor vehicle accident. Patients with collateral ligament injuries were also more likely to have a higher overall ISS (mean, 14 vs 11; P<.001) and higher rates of all broad categories of associated injuries (P<.001).

Table 5: Comparison of Injury Characteristics

On multivariate analysis, increasing age was associated with decreasing risk of collateral ligament injury (P≤.001; Table 6). Obesity (odds ratio, 1.6; 95% confidence interval, 1.3–1.9) and ISS of 20 or higher (odds ratio, 1.4; 95% confidence interval, 1.2–1.7) were associated with increased risk of collateral ligament injury. Of the associated lower extremity fractures analyzed, only distal femur fracture was significantly associated with greater risk of collateral ligament injury (odds ratio, 2.1; 95% confidence interval, 1.8–2.5). Pedestrians struck by a motor vehicle (odds ratio, 2.0; 95% confidence interval, 1.7–2.3) and motorcycle accidents (odds ratio, 1.5; 95% confidence interval, 1.3–1.8) were more likely to have collateral ligament injuries.

Table 6: Multivariate Analysis of Risk Factors for Collateral Ligament Injuries With Proximal Tibia Fracture

Inpatient outcomes were different between patients with and patients without collateral ligament injuries (Table 7). After controlling for CCI and ISS, simultaneous injury to both collateral ligaments was associated with increased odds of SAEs (odds ratio, 1.38; P=.015), MAEs (odds ratio, 1.58; P<.001), and AAEs (odds ratio, 1.51; P<.001) and an increased inpatient LOS of 2.27 days (P<.001). Isolated LCL injury was associated with a mean increased inpatient LOS of 1.95 days (P=.009). Isolated MCL injury was associated with lower odds of AAEs (odds ratio, 0.66; P=.007).

Table 7: Association of Collateral Ligament Injuries With Short-term Outcomes

Discussion

Proximal tibia fractures are common high-energy injuries that have been associated with significant damage to local soft tissue structures, including the collateral ligaments of the knee.^1,5,6 These ligament injuries are sometimes repaired or reconstructed in the acute setting.¹⁴ Early operative repair can improve patient-reported outcomes and may reduce the chance of subsequent knee instability.¹² These injuries can often also be missed in the acute setting, as it is difficult to perform an effective ligamentous examination on a patient in pain, and they may not be identified in complex polytrauma patients. When these injuries are missed, patients may require additional surgical intervention at a later date because of knee instability.^9,10 Therefore, identification of patients at greatest risk for collateral ligament injuries with proximal tibia fractures is necessary in the acute setting. The current authors identified the incidence, injury characteristics, risk factors, and inpatient outcomes of patients with proximal tibia fractures and concurrent collateral ligament injuries.

Of the 32,441 patients diagnosed with proximal tibia fractures, 1445 had concurrent collateral ligament injuries. This incidence of 4.5% measured in a large patient cohort is similar to some previously reported rates of collateral ligament injury.^1,5,6,15,16 Other studies that have reported significantly higher rates of collateral ligament injury, up to 70%, have either included partial ligament tears⁷ or investigated smaller cohorts.^1,5 This finding may indicate that a significant number of ligamentous injuries are missed in the initial hospitalization, as many of the studies with significantly higher rates of collateral ligament injuries also performed magnetic resonance imaging for every patient.

Two high-energy mechanisms of injury were identified as significant risk factors for collateral ligament injury: motorcycle accidents and pedestrians struck by motor vehicles. This analysis also identified obese patients as being more likely to have concurrent collateral ligament injuries with proximal tibia fractures. Low-energy falls in the obese have previously been described as a unique mechanism of knee dislocation.¹⁷ One study reported 15 of 17 patients with complete disruption of one or both collateral ligaments.¹⁷ Higher mechanical loading of the joint at baseline and during injury likely contributes to the increased deforming forces on the ligaments. In addition, systemic inflammation, as seen in obesity and metabolic syndrome,^18–20 may also result in weakened ligaments predisposed to rupture.

Even after controlling for ISS and mechanism of injury, increasing age was associated with decreased risk of collateral ligament injury. One previous study reported that the tensile strength of bone decreases with age in comparison with the soft tissues, shifting the injury sequelae to bony structures rather than the ligaments.⁸

The finding that concurrent proximal tibia with distal femur fractures were associated with collateral ligament injuries was surprising. To the authors' knowledge, this relationship has not been described in previous studies. With more than twice the odds of a collateral ligament injury, patients with concurrent proximal tibia and distal femur fractures would benefit from a thorough evaluation of collateral ligaments in the acute setting. In terms of physical examination for collateral ligament injuries with proximal tibia fractures, the authors suggest that the bony injuries be fixed first in order to perform a ligamentous examination once the bony structures are out to length and stable. In the emergency department, physical examination of the collateral ligaments is often limited by bony instability and pain. Physical examination may be performed intraoperatively under fluoroscopic guidance when the patient is under anesthesia and pain is no longer an issue. Varus and valgus stress examinations at 0° and 30° of knee flexion are the most accurate way to test the collateral ligaments and can be compared with examinations of the contralateral side.²¹ Patients should also be reexamined 6 weeks postoperatively to assess stability of cruciate and collateral knee ligaments.⁴

Patients who had injury to both collateral ligaments were found to have increased rates of inpatient adverse events and a longer LOS, even after controlling for injury severity and comorbidities. Patients with LCL-only injuries, which are often repaired or reconstructed acutely, were at increased risk of longer LOS, whereas those with MCL-only injuries, which are often managed nonoperatively, did not have longer LOS. Twenty-one percent of patients with any collateral ligament injury received surgery. The LOS may be driven, in part, by the need for operative treatment. It also may be affected by the ability to mobilize patients, the need for rehabilitation placement, and pain control.

The current study had several strengths compared with previous studies of the concurrence of collateral ligament injuries and proximal tibia fractures. This study represents the largest cohort of proximal tibia fracture patients with concurrent collateral ligament injuries reported. Previous studies included cohorts of up to 100 patients with concurrent tibial plateau and collateral ligament injuries by performing magnetic resonance imaging on a series of plateau injuries.⁴ Another study identified risk factors such as young age, obesity, concurrent distal femur fracture, and high-energy mechanism of injury (motorcycle accident, pedestrian struck by motor vehicle) for concurrent collateral ligament injuries with proximal tibia fractures.⁷ This can help guide providers in the acute setting in making sure that these patients are evaluated for potential collateral ligament injuries and avoiding missing these injuries, which can lead to poor outcomes later on.

A limitation of this study was the limited specificity of ICD-9 diagnosis codes. Within the ICD-9 system, further detail regarding the specific proximal tibia fracture types (eg, Schatzker or Moore classifications) cannot be described.²² Gardner et al² showed that partial collateral ligament tears are more common than complete tears in tibial plateau fractures, except for the MCL in Schatzker IV, V, and VI fractures. In addition, LCL tears rarely occur in isolation from injuries to the other posterolateral corner structures of the knee, such as the popliteus tendon, popliteofibular ligament, iliotibial band, and biceps femoris tendon.²³ However, injuries to these other structures are not identified through ICD-9 codes. Because of the association between LCL injury and posterolateral corner injuries, the authors can assume that identifying the LCL injury includes these other injuries as well, although this might not always be the case. There is also no record of whether ligamentous injuries were diagnosed through magnetic resonance imaging or examination under anesthesia; thus, there is no reference to grade and degree of ligament injury. A prospective study with a larger sample of proximal tibia fractures including the grade and degree of ligamentous injuries and the types of fractures would help elucidate injury patterns and their incidence, risk factors for the different patterns, surgical interventions, and, ultimately, outcomes. Although a prospective study like this is warranted to better characterize proximal tibia fractures with concurrent collateral ligament injuries, the current study provides a large retrospective analysis of these injuries on which prospective studies can be based. Finally, the authors did not include cruciate ligament injuries, as these are often managed in a delayed fashion and not in the acute setting. Additionally, it is not possible to distinguish between anterior cruciate ligament injuries and posterior cruciate ligament injuries in the database. The number of generic cruciate ligament injuries for plateau fracture with collateral vs without collateral was discussed as a general comparison.

Conclusion

Risk factors such as younger age, obesity, concurrent distal femur fracture, and high-energy mechanism (pedestrian struck by motor vehicle, motorcycle accident) can be used to identify patients with proximal tibia fractures who may warrant more careful and thorough evaluation of their knee collateral ligaments. Collateral ligament injuries, particularly LCL injury with proximal tibia fractures, are associated with increased LOS when controlling for other variables such as ISS, CCI, and age. Many collateral ligament injuries may be missed during initial hospitalization, as the authors' incidence was lower than that of some studies that included routine magnetic resonance imaging. Additional work is needed to determine the best method of evaluation and the best method of treatment of proximal tibia fractures to ensure that concurrent collateral ligament injury is identified and treated appropriately.