ACR Appropriateness Criteria®
Clinical Condition: Acute Trauma to the Knee
Variant 1: Patient any age (excluding infants); fall or twisting injury, no focal tenderness, no effusion; able to walk. First study.
Radiologic Procedure |
Rating |
Comments |
RRL* |
X-ray knee |
2 |
|
|
MRI knee without contrast |
2 |
|
O |
MRI knee without and with contrast |
1 |
|
O |
US knee |
1 |
|
O |
MRA knee without and with contrast |
1 |
|
O |
MRA knee without contrast |
1 |
|
O |
CT knee without contrast |
1 |
The RRL for the adult procedure is . |
|
CT knee with contrast |
1 |
The RRL for the adult procedure is . |
|
CT knee without and with contrast |
1 |
The RRL for the adult procedure is . |
|
Tc-99m bone scan with SPECT lower extremity |
1 |
|
|
Rating Scale: 1,2,3 Usually not appropriate; 4,5,6 May be appropriate; 7,8,9 Usually appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the table are listed at the end of the "Major Recommendations" field.
Variant 2: Patient any age (excluding infants); fall or twisting injury, with one or more of following: focal tenderness, effusion, inability to bear weight. First study.
Radiologic Procedure |
Rating |
Comments |
RRL* |
X-ray knee |
9 |
|
|
MRI knee without contrast |
5 |
|
O |
US knee |
2 |
|
O |
CT knee without contrast |
2 |
The RRL for the adult procedure is . |
|
Tc-99m bone scan with SPECT lower extremity |
2 |
|
|
MRI knee without and with contrast |
1 |
|
O |
MRA knee without and with contrast |
1 |
|
O |
MRA knee without contrast |
1 |
|
O |
CT knee with contrast |
1 |
The RRL for the adult procedure is . |
|
CT knee without and with contrast |
1 |
The RRL for the adult procedure is . |
|
Rating Scale: 1,2,3 Usually not appropriate; 4,5,6 May be appropriate; 7,8,9 Usually appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 3: Patient any age (excluding infants); fall or twisting injury with either no fracture or a Segond fracture seen on a radiograph, with one or more of the following: focal tenderness, effusion, inability to bear weight. Next study.
Radiologic Procedure |
Rating |
Comments |
RRL* |
MRI knee without contrast |
9 |
|
O |
CT knee without contrast |
5 |
The RRL for the adult procedure is . |
|
US knee |
1 |
|
O |
MRI knee without and with contrast |
1 |
|
O |
MRA knee without and with contrast |
1 |
|
O |
MRA knee without contrast |
1 |
|
O |
CT knee with contrast |
1 |
The RRL for the adult procedure is . |
|
CT knee without and with contrast |
1 |
The RRL for the adult procedure is . |
|
Tc-99m bone scan with SPECT lower extremity |
1 |
|
|
Rating Scale: 1,2,3 Usually not appropriate; 4,5,6 May be appropriate; 7,8,9 Usually appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 4: Patient any age (excluding infants); fall or twisting injury with a tibial plateau fracture on a radiograph, with one or more of the following: focal tenderness, effusion, inability to bear weight. Next study.
Radiologic Procedure |
Rating |
Comments |
RRL* |
CT knee without contrast |
9 |
May be helpful for treatment planning or prognosis. The RRL for the adult procedure is . |
|
MRI knee without contrast |
7 |
|
O |
US knee |
1 |
|
O |
MRI knee without and with contrast |
1 |
|
O |
MRA knee without and with contrast |
1 |
|
O |
MRA knee without contrast |
1 |
|
O |
CT knee with contrast |
1 |
The RRL for the adult procedure is . |
|
CT knee without and with contrast |
1 |
The RRL for the adult procedure is . |
|
Tc-99m bone scan with SPECT lower extremity |
1 |
|
|
Rating Scale: 1,2,3 Usually not appropriate; 4,5,6 May be appropriate; 7,8,9 Usually appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 5: Patient any age (excluding infants). Injury to knee 2 days ago, mechanism unknown. Focal patellar tenderness, effusion, able to walk. First study.
Radiologic Procedure |
Rating |
Comments |
RRL* |
X-ray knee |
9 |
|
|
MRI knee without contrast |
5 |
|
O |
US knee |
2 |
|
O |
CT knee without contrast |
2 |
The RRL for the adult procedure is . |
|
Tc-99m bone scan with SPECT lower extremity |
2 |
|
|
MRI knee without and with contrast |
1 |
|
O |
MRA knee without and with contrast |
1 |
|
O |
MRA knee without contrast |
1 |
|
O |
CT knee with contrast |
1 |
The RRL for the adult procedure is . |
|
CT knee without and with contrast |
1 |
The RRL for the adult procedure is . |
|
Rating Scale: 1,2,3 Usually not appropriate; 4,5,6 May be appropriate; 7,8,9 Usually appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Variant 6: Patient any age (excluding infants). Significant trauma to knee from motor vehicle accident, suspect posterior knee dislocation. First study.
Radiologic Procedure |
Rating |
Comments |
RRL* |
X-ray knee |
9 |
Initial examination to assess overall injury. |
|
MRI knee without contrast |
9 |
Necessary to evaluate extent of damage to ligament and other support structure. |
O |
MRA knee without and with contrast |
7 |
Performed in conjunction with MRI of knee. See statement regarding contrast in text under "Anticipated Exceptions." |
O |
Arteriography lower extremity |
7 |
The RRL for the adult procedure is . |
|
CTA lower extremity with contrast |
7 |
Performed in conjunction with trauma CT imaging. The RRL for the adult procedure is . |
|
MRA knee without contrast |
3 |
Performed in conjunction with MRI of knee. |
O |
US knee |
2 |
|
O |
CT knee without contrast |
2 |
The RRL for the adult procedure is . |
|
Tc-99m bone scan with SPECT lower extremity |
2 |
|
|
MRI knee without and with contrast |
1 |
|
O |
CT knee with contrast |
1 |
The RRL for the adult procedure is . |
|
CT knee without and with contrast |
1 |
The RRL for the adult procedure is . |
|
Rating Scale: 1,2,3 Usually not appropriate; 4,5,6 May be appropriate; 7,8,9 Usually appropriate |
*Relative Radiation Level |
Note: Abbreviations used in the tables are listed at the end of the "Major Recommendations" field.
Summary of Literature Review
A 2001 report stated that there are 1.3 million annual visits to United States emergency departments because of acute knee trauma, and over $1 billion is spent on radiographs of the knee. The knee radiograph is the most common radiograph performed for trauma in the emergency room and has the lowest yield for diagnosing clinically significant fractures. A retrospective review of 1,967 patients with acute knee injuries revealed that 74.1% of patients had radiographs and only 5.2% had fractures. One study concluded that radiographs obtained for acute trauma do not reliably depict all important injuries and that the findings in 25% of knee radiographs obtained for acute trauma do not correlate with the clinical findings.
Radiography
A prospective survey of the judgment and attitudes of experienced clinicians in the use of knee radiography in 1,040 patients with acute knee injuries showed that, despite its inability to accurately predict the probability of fracture and to discriminate between fracture and nonfracture cases, radiographs were usually ordered. The proportion of patients referred for knee radiographs varied from 65.9% to 84.6%. According to the physicians, radiographs were ordered for the following reasons: 1) patients expected it and would otherwise be dissatisfied; 2) the physician lacked confidence in the clinical examination, or the orthopedic surgeon considered the radiograph routine; and 3) possible medicolegal repercussions. These reasons and the patient's demand for imaging are recognized as the reasons that implementation of ordering guidelines was not overwhelming.
Caution should be used when relying on clinical examination for diagnosing certain knee injuries. One study reported that the correct diagnosis of bilateral quadriceps tendon rupture was established in only 61% (17/28) of cases by history and clinical examination alone. Another study reported that fractures missed on clinical examination included fractures of the patella, tibial spine, and fibular head.
Clinical decision rules for the acutely injured knee suggest that radiographic examination of the knee following acute injury can be eliminated in most instances by applying specific clinical guidelines. A prospective and retrospective study of 334 patients reported that patients between 12 and 50 years of age suffering a fall or blunt trauma and unable to ambulate or those who sustained multiple trauma should be radiographed. These authors reported 92% sensitivity and 79% specificity for identifying clinically significant fractures. Their study also reported that applying the clinical decision rules could reduce the number of radiographs taken in the emergency room by 78%.
A group of researchers concluded that a clinically significant fracture can be excluded in patients older than 18 years who can walk without limping or if there was a twisting injury to the knee and no joint effusion. If an effusion was present on physical examination, the odds of a fracture were 7.5 times greater. Using this clinical decision rule, the sensitivity for detecting a knee fracture was 100%, and specificity was sufficient to eliminate the need for 29% of knee radiographs ordered in the emergency room.
Another group applied a clinical decision rule (Ottawa Knee Rule) using parameters based on age, palpable tenderness, and function. Under the rule patients with acute knee pain and one or more of the following parameters should have a radiographic examination if they:
- Are 55 years of age or older
- Have palpable tenderness over the head of the fibula
- Have isolated patellar tenderness
- Cannot flex the knee to 90 degrees
- Cannot bear weight immediately following the injury, or
- Cannot walk in the emergency room (after taking four steps)
This rule was applied prospectively in 1,047 adults with acute knee injuries, and it was determined that its application would result in a 28% relative reduction in the number of radiographs ordered, a decrease from 68.6% to 49.4%.
A later study was performed to validate the Ottawa Knee Rule, and prospective validation analyzing 1,096 patients found it to be 100% sensitive for identifying knee fractures. The decision rule was interpreted correctly 96% of the time, and when applied, the probability of missing a fracture was zero. The decision rule was 100% sensitive for identifying a fracture in patients older than 18 years who were not referred from another hospital, returned for reassessment, had a knee injury for seven days, or had isolated skin lesions. The potential relative reduction in use of radiography was estimated to be 28% (from 74% to 53%).
In a pooled analysis of data from six studies, it was concluded that a negative result using the Ottawa Knee Rule accurately excluded knee fracture after acute knee injury. A meta-analysis to determine the role of radiography in evaluating knee fractures concluded that among the five decision rules evaluated, the Ottawa Knee Rule had the strongest supporting evidence. Further prospective analysis of the Ottawa Knee Rule showed that it allowed a decrease in the number of radiographs performed after knee trauma by 35%, with a sensitivity of knee fracture detection of 100%.
Another study compared the implementation of the Ottawa Knee Rule by triage nurses and emergency medicine physicians. No fracture was missed by either group, but triage nurses were found to order 3.6 times more radiographs than emergency physicians, maintaining sensitivity at the expense of specificity and cost savings. An additional study evaluated the use of the Ottawa Knee Rule when applied by users with different levels of clinical training, including medical students and surgical residents, and found sensitivity and negative predictive value of 1.0 for both groups and a reduced radiography rate of 25% with application of the rule.
In a study of 214 patients, it was determined that the use of radiographs in the setting of acute trauma could be further reduced by obtaining a single lateral view. It was reported that the probability of not having a fracture if the lateral view was normal was 100%, thus reducing the need for additional radiographs by 67%.
With regard to mechanism of injury, history and physical examination are key elements for determining the indication for radiographs and the application of a decision rule. The most common mechanisms for knee injury are a direct blow, a fall, or a twisting injury. Twisting injuries are responsible for three-fourths of all knee injuries; however, 86% of all knee fractures result from blunt trauma. The risk of fracture also increases with age; fracture is four times more likely in patients older than 50 years, presumably secondary to osteoporosis, increased frequency of blunt injury, and inability to protect the knee during a fall.
Absence of immediate swelling, ecchymosis, effusion, deformity, increased warmth, and abrasion/laceration are significant predictors of a normal radiograph. It was generally agreed that radiographs should be obtained and that the clinical decision rule should not be applied for patients with gross deformity, a palpable mass, a penetrating injury, prosthetic hardware, unreliable clinical history or physical examination secondary to multiple injuries, altered mental status (e.g., head injury, drug or alcohol use, dementia), neuropathy (e.g., paraplegia, diabetes), or a history suggesting increased risk of fracture. The physician's judgment and common sense, however, should supersede clinical guidelines.
Transient patellar dislocation is unsuspected clinically in 45% to 73% of patients with evidence of dislocation subsequently seen on magnetic resonance imaging (MRI). Radiographs may demonstrate a fracture of the medial patella or lateral trochlear, and can also show anatomic features that predispose to dislocation such as a decreased sulcus angle, patella alta, patellar tilt, or patellar subluxation. MRI is more sensitive than radiographs for detecting lateral patellar dislocation, including injury to the medial patellofemoral ligament, bone contusions and osteochondral injuries.
Magnetic Resonance Imaging
In addition to clinically significant fractures, other injuries must be considered. Most patients (93.5%) who present with acute knee injuries in the emergency room have soft-tissue rather than osseous injuries. Even in patients with fractures, concomitant soft-tissue injuries frequently are present. A study found that in 90% of patients with otherwise nonoperative tibial plateau fractures there were significant soft-tissue injuries diagnosed by MRI, including ligament and meniscal tears. Another study reported unstable meniscal tears in 36% of patients with tibial plateau fractures. An accurate clinical examination is essential to identify patients at high risk for delayed function recovery due to major soft-tissue injuries. However, using MRI, another study showed that the first clinical examination after acute knee trauma has a low diagnostic value and that the incidence of anterior cruciate ligament (ACL) injuries is higher than previously described. It is recognized that MRI is the optimal imaging modality for identifying soft-tissue, cartilaginous surface, and bone injuries around the knee.
To image internal knee derangement, MRI has been the technique of choice since the 1990s. The accuracy and reliability of MRI depend on experience and training. Nonetheless, numerous studies have shown that MRI has a high diagnostic accuracy in identifying traumatic intra-articular knee lesions. This is particularly true when strict diagnostic criteria are used, and this applies to both spin-echo imaging and fast spin-echo imaging, as well as imaging at both low and high field strength. MRI has been shown to demonstrate minor meniscocapsular tears when performed with understanding of anatomy. Characteristic findings on MRI, including specific bone marrow edema patterns and osteochondral defects, can allow accurate diagnosis of injuries such as transient dislocation of the patella that cannot be detected by radiographs.
MRI is a valuable tool in the decision-making process, altering the treatment plan in 18% of patients with meniscal or chondral surface injuries and allowing earlier surgical intervention because of the more accurate diagnosis obtained. Multiple authors and studies have validated that unnecessary diagnostic arthroscopy can be avoided because of the high predictive value of a negative MRI. One study found MRI to have a positive predictive value twice that of clinical examination for meniscal tears. It also found that MRI would decrease negative diagnostic arthroscopy to 5% and would help reduce the need for a second therapeutic arthroscopic procedure. Another study reported MRI accuracy to be approximately 94%, showing that it can effectively replace diagnostic arthroscopy for evaluating meniscal and ligament tears. Yet another study reported that when the clinical examination is equivocal within 6 weeks of sudden trauma with a hemarthrosis present, MRI could have prevented diagnostic arthroscopy in 22% of patients. In randomized studies of patients with knee injuries, MRI findings have been shown to shorten the time to completion of diagnostic workup, reduce the number of additional diagnostic procedures, improve quality of life in the first 6 weeks, and potentially reduce costs associated with lost productivity. MRI should be read with caution in trauma patients without mechanical signs who have osteoarthritis.
ACL rupture is responsible for more than 70% of all acute hemarthrosis in young athletes and 17% in a mixed sedentary and athletic population. Locking, the presence of a loose body on radiographs, and hemarthrosis within 12 hours of injury have previously been reported as indications for arthroscopy instead of MRI. However, it was found that in 48% of patients presenting with an acutely locked knee, management was changed from surgical to conservative based on MRI findings.
Single Photon Emission Computed Tomography
In addition to MRI, single photon emission computed tomography (SPECT) has been proposed for diagnosing meniscus injuries. A specific crescentic pattern of uptake on the transaxial view has been described as having a sensitivity of 77% and specificity of 74%. With the additional criterion of increased equilibrium activity in the adjacent femoral condyles, these values increase to 90% and 84%, respectively. Considerable concordance has been shown between SPECT results and those of other modalities for assessing meniscal tears and the bone contusions from an ACL tear in acute knee trauma.
Ultrasound
Sonography has been reported to be 91% sensitive and 100% specific for diagnosing an acute ACL tear within 10 weeks of an acute hemarthrosis when there is no prior trauma and no bone abnormalities. Sonography can be used both for initial detection and confirmation of this injury and for follow-up. Furthermore, a comparison of sonography and radiography using lipohemarthrosis as a criterion of acute intra-articular fracture yielded a sensitivity and specificity of 94% for sonographic detection of such fractures. One study showed that the presence of an effusion at sonography in the acutely injured knee has a 91% positive predictive value for internal derangement. However, intra-articular knee sonography should only be performed and interpreted by personnel with the appropriate expertise in its application.
Computed Tomography
Computed tomography (CT) with three-dimensional reconstruction has been shown to reflect the severity of tibial plateau fractures more accurately than radiography in 43% of cases and to modify the surgical plan in 59% of operative cases. In severely injured patients, diagnostically sufficient radiographs are sometimes difficult to obtain, and therefore a negative radiograph is not reliable in ruling out a fracture. In these patients, multidetector CT is a fast and accurate examination for evaluating tibial plateau fractures and other complex knee injuries. In a 2007 study, researchers concluded that in the acute setting, CT offers 80% sensitivity and 98% specificity for depicting osseous avulsions and a high negative predictive value for excluding ligament injury.
Patellar Dislocation
Transient patellar dislocation is unsuspected clinically in 45% to 73% of patients with evidence of dislocation subsequently seen on MRI. Radiographs may demonstrate a fracture of the medial patella or lateral trochlear, and can also show anatomic features that predispose to dislocation such as a decreased sulcus angle, patella alta, patellar tilt, or patellar subluxation. MRI is more sensitive than radiographs for imaging findings of lateral patellar dislocation, including injury to the medial patellofemoral ligament, bone contusions, and osteochondral injuries.
Knee Dislocation
Dislocation of the knee results from a fall from a height, a motor vehicle accident, a vehicle striking a pedestrian, or contact sports. This injury, which often reduces spontaneously, constitutes a true orthopedic emergency because of possible nerve or arterial damage. Vascular injury may be found in one-third of patients following posterior knee dislocation. Physical signs of clinically significant vascular injury are the absence of pulses, ischemia, active bleeding, and bruit/thrill. These signs have been reported to have 100% accuracy for determining the need for surgical exploration. Although one study concluded that angiography is unnecessary in the routine evaluation of patients with blunt lower-extremity trauma who present with a normal neurovascular examination, a systematic review suggested that the isolated presence of abnormal pedal pulses on initial examination following knee dislocation is not sensitive enough to detect a vascular injury that necessitates surgery and that the workup after knee dislocation should include angiography. Computed tomography angiography (CTA) may be used as an alternative to conventional angiography in these patients.
A group of researchers endorsed the use of MRI when evidence of an acute popliteal artery injury is absent, but in the presence of ischemia or lack of pulses to the lower extremity, surgical exploration is suggested. An MRI should also be performed to identify ligamentous injuries and associated pathology. Another analysis found excellent correlation between MRI findings and surgical findings in patients with knee dislocation. Furthermore, these authors reported 100% correlation between magnetic resonance angiography (MRA) findings and conventional angiography findings in multiple-ligament injured knees, including knee dislocations.
Summary
Clinical decision rules for evaluating the acutely injured knee have been studied by various investigators, who determined that their application can considerably reduce the number of radiographs ordered without missing a clinically significant fracture. Although different parameters and definitions were used for the various decision rules, there were sufficient similarities between the investigations to allow usable conclusions to be drawn.
In patients of any age except for infants, the clinical parameters used for not requiring radiographs following knee trauma are as follows:
- Patient is able to walk without a limp
- Patient had a twisting injury and there is no effusion
The clinical parameters for ordering knee radiographs in this population following trauma are as follows:
- Joint effusion within 24 hours of direct blow or fall
- Palpable tenderness over the fibular head or patella
- Inability to walk (four steps) or bear weight immediately or in the emergency room or within a week of the trauma
- Inability to flex knee to 90 degrees
- Altered mental status
It has also been reported that a fracture can be excluded if a single lateral view of the knee is normal, eliminating the need for additional radiographic views.
In general, these studies excluded patients with superficial skin injuries, gross deformity, a palpable mass, a penetrating injury, prosthetic hardware, altered consciousness (from alcohol and/or drug use), multiple injuries, decreased limb sensation, or a history indicating an elevated risk of fracture. They also excluded pregnant patients, patients returning for reassessment, and patients whose injury occurred more than 7 days prior to initial evaluation.
Soft-tissue injuries (meniscal injuries, chondral surface injuries, and ligamentous disruption) are best evaluated by MRI. Although lateral patellar dislocation may be reduced at the time of presentation in the emergency room, characteristic findings on MRI, including specific bone marrow edema patterns and osteochondral defects, can allow accurate diagnosis.
Knee dislocation, even if spontaneously reduced, constitutes a potential threat to the popliteal nerve or artery. A systematic review has suggested that the isolated presence of abnormal pedal pulses on initial examination following knee dislocation is not sensitive enough to detect a vascular injury that necessitates surgery, and that the workup should include angiography. One study has shown a 100% correlation between MRA findings and conventional angiography findings in multiple-ligament injured knees, including knee dislocations. An MRI should also be performed to identify ligamentous injuries and associated pathology.
Abbreviations
- CT, computed tomography
- CTA, computed tomography angiography
- MRA, magnetic resonance angiography
- MRI, magnetic resonance imaging
- SPECT, single photon emission computed tomography
- Tc-99m, technetium-99 metastable
- US, ultrasound
Relative Radiation Level Designations
Relative Radiation Level* |
Adult Effective Dose Estimate Range |
Pediatric Effective Dose Estimate Range |
O |
0 mSv |
0 mSv |
|
<0.1 mSv |
<0.03 mSv |
|
0.1-1 mSv |
0.03-0.3 mSv |
|
1-10 mSv |
0.3-3 mSv |
|
10-30 mSv |
3-10 mSv |
|
30-100 mSv |
10-30 mSv |
*RRL assignments for some of the examinations cannot be made, because the actual patient doses in these procedures vary as a function of a number of factors (e.g., region of the body exposed to ionizing radiation, the imaging guidance that is used). The RRLs for these examinations are designated as "Varies". |