The evaluation of pediatric elbow radiographs in the setting of acute trauma may be challenging for many emergency department physicians, orthopedic surgeons, and radiologists. Diagnostic difficulties stem both from the complex developmental anatomy of the elbow and from significant differences between children and adults in the patterns of injury after elbow trauma.
According to Rabiner et al, ultrasonography is highly sensitive for elbow fractures, and a negative ultrasound may reduce the need for radiographs in children with elbow injuries. Of 130 patients (mean age, 7.5 yr), 43 (33%) had a radiograph result positive for fracture. A positive elbow ultrasound had a sensitivity of 98% and a specificity of 70%. 
Understanding the developmental anatomy of the pediatric elbow helps ensure that normal ossification centers are not misinterpreted as fracture fragments. It also aids recognition of an injury when the pattern is altered. For example, the medial epicondyle usually ossifies prior to the trochlea. Therefore, if the medial epicondyle is not seen in its expected location and a single ossicle is seen beneath the medial aspect of the distal humeral metaphysis, the ossicle should be interpreted as an avulsed medial epicondyle that is entrapped in the joint rather than a normal trochlea.
Knowledge of the mechanisms of injury, the range of skeletal and soft tissue findings in the different patterns of injury, and the proper indications for additional views all aid in the recognition of subtle fractures. [2, 3, 4]
The elbow is composed of 3 articulations. The ulna articulates with the humerus at the trochlea, which is the grooved and rounded medial articular portion of the distal humerus. The articular portion of the ulna is formed by the olecranon process proximally and by the coronoid process more distally. This humeroulnar or trochleoulnar joint is a hinged articulation that essentially permits motion in a single plane, allowing flexion and extension. The concave head of the radius articulates with the capitellum, which is the convex lateral articular surface of the distal humerus. This humeroradial or radiocapitellar joint permits the radius to flex and extend relative to the humerus and to rotate throughout elbow flexion and extension. This rotation allows for supination and pronation of the forearm and depends on proper motion of the proximal radioulnar joint (the third articulation of the elbow) and on the normal mobility of the forearm and wrist.
The distal humeral articular surface has several grooves and ridges that are important in determining anatomic stability after a fracture. Medially, the trochlear notch articulates with a corresponding ridge along the ulna. More laterally, the capitellotrochlear sulcus separates the humeral articular surface of the radius from that of the ulna. Between these grooves is the lateral crista of the trochlea, which provides lateral stability to the trochleoulnar joint.
A review of medical records of 462 children (median age, 6 yr) with elbow fractures identified the most common fractures as supracondylar (N=258, 56%), radial neck (N=80, 17%), and lateral condylar (N=69, 15%). On follow-up, additional fractures were seen in 32 of the children, and of these, 25 had a different type of fracture than that identified on the initial radiographs. The most common follow-up fractures were olecranon (N=23, 72%), coronoid process (N=4, 13%), and supracondylar (N=3, 9%). The frequency of olecranon fractures on follow-up may suggest the occult nature of these fractures. 
The American College of Radiology Appropriateness Criteria for chronic elbow pain includes the following  :
Initial evaluation of chronic elbow pain should begin with radiography.
Chondral and osteochondral abnormalities can be further evaluated with MRI or CT. The addition of arthrography is helpful, especially for detecting intra-articular bodies.
Radiographically occult bone abnormalities can be detected with MRI, CT, or bone scintigraphy.
Soft-tissue abnormalities (tendon, ligament, nerve, joint recess, and masses) are well-demonstrated with MRI or US.
Dynamic assessment with US is effective for diagnosing nerve or muscle subluxation.
A study of 62 elementary school baseball players (grades 4-6; ages 9-12 yr) for elbow injuries using MRI found positive findings in 26 (41.9%), all confined to the MCL. Screening was performed using low-magnetic-field (0.2-T) MRI. 
Education information resources
See the Medscape Reference article Salter-Harris Fracture Imaging for more information. For patient information resources, see eMedicineHealth's First Aid and Injuries Center, as well as Broken Elbow and Elbow Dislocation. For additional reading, see the articles by Barton et al and Skaggs et al. [8, 9]
Ossification of the elbow region is complex, but knowledge of it is essential in analyzing elbow trauma in children. The distal humerus has 4 secondary ossification centers: those for the capitellum and trochlea (which form the articular surfaces) and those for the medial and lateral epicondyles. The capitellar ossification center eventually extends beyond the capitellum into the lateral aspect of the trochlea and accounts for ossification of the lateral crista of the trochlea. Typically, none of these centers is ossified at birth.
Invariably, the capitellum is the first secondary center to ossify, usually followed by the medial epicondyle, the trochlea, and the lateral epicondyle. The age at which ossification centers are first seen varies considerably; maturation usually proceeds earlier in girls than in boys. With this in mind, the average age at which the centers are seen first in 50% of children is age 3 months for the capitellum, 5 years for the medial epicondyle, 8 years for the trochlea, and 10 years for the lateral epicondyle.
The corresponding ages at which the ossification centers of the proximal forearm bones appear are 4.5 years for the radial head and 9 years for the olecranon.  The acronym CRMTOL describes the usual order of appearance of all 6 elbow centers: capitellum, radial head, medial epicondyle, trochlea, olecranon, and lateral epicondyle. These ossification centers vary not only with regard to the age of the patient at the time of development but also with regard to their radiographic appearances.
The capitellum develops as a single smooth center, whereas trochlear ossification most often has a fragmented and irregular appearance. The medial epicondyle usually develops as a single center. The lateral epicondyle may arise as either a single elongated center or as multiple centers of ossification. The lateral epicondyle usually fuses to the distal humeral epiphysis (lateral condyle) before fusing to the metaphysis. Ossification of the lateral epicondyle begins peripherally and progresses toward the epiphysis and metaphysis. Initially this leaves a wide space between the lateral epicondyle ossification center, which typically has a linear pattern, and the lateral condyle, which can be misinterpreted as an avulsion fracture. The radial head ossification center is initially oval and subsequently becomes flattened and disk shaped. The olecranon is often ossified from 2 secondary centers that should not be confused with fracture fragments.
In addition to the findings in the multiple ossification centers described above, other normal findings may simulate pathology. A notchlike defect in the proximal radial metaphysis may be confused with a fracture (see the image below).
The proximal radius has normal angulation between the neck and shaft, with the neck angulated laterally and slightly anteriorly relative to the shaft, which should not be confused with a fracture. Additionally, it is the orientation of the neck rather than the shaft that should be used to evaluate radiocapitellar alignment.
Normal radiographic findings that may simulate nontraumatic pathology include a radial tuberosity that appears as a lytic lesion when viewed en face (see the image below) and the olecranon fossa of the distal humerus, which may be unusually large and lucent.
Standard radiographic evaluation of the elbow includes imaging in the anteroposterior (AP) and lateral views. Other views may also be helpful. In particular, for lateral condyle fractures, the internal oblique view is often best for demonstrating the fracture and determining the degree of displacement.  Because supracondylar fractures may be oriented obliquely on the lateral view, coursing proximally from anterior to posterior, an AP view with cephalad angulation of the x-ray beam may help to better demonstrate such a fracture. Although infrequently used, the capitellum and radiocapitellar joint are best seen on the radiocapitellar view, which is a lateral view with the beam angulated toward the shoulder to project the radiocapitellar joint away from the trochlea and coronoid processes, which otherwise overlap these structures on a standard lateral view.
Evaluation of soft tissues and joint effusions
Evaluation of the soft tissues is important in elbow trauma. Localized soft tissue swelling over the lateral aspect of the elbow should heighten the suspicion for a possible lateral condyle fracture. Similarly, medial epicondyle avulsion fractures are often accompanied by localized soft tissue swelling in the medial aspect of the elbow.
Displacement of the elbow fat pads is an important indicator of an elbow joint effusion (see below).
The elbow fat pads are situated external to the synovium and within the fibrous external joint capsule; the posterior fat pad is located in the olecranon fossa and the anterior fat pad in the coronoid fossa. When the elbow is flexed, the posterior fat pad lies deep within the olecranon fossa and is hidden from view by the medial and lateral condyles. Hence, depiction of the posterior fat pad indicates outward displacement by a joint effusion. With elbow extension, the olecranon moves into the olecranon fossa, displacing the fat pad outward, and hence visualization of the posterior fat pad as a sign of effusion is valid only if the elbow is flexed. The coronoid fossa is not as deep as the olecranon fossa, and hence the anterior fat pad may be seen just anterior to the distal humeral cortex even without an effusion. However, if the anterior fat pad is angled outward, it indicates an elbow effusion.
The presence of an elbow effusion alone does not indicate a fracture. In the setting of acute trauma, the presence of an elbow effusion strongly suggests a hemarthrosis, often from a fracture. In other settings, the etiology may be different. Joint effusions resulting from septic arthritis or juvenile rheumatoid arthritis displace the fat pads, and patients with hemophilia often have an elbow hemarthrosis in the absence of fracture or trauma. In addition to circumstances in which visualization of the fat pad may be due to processes other than fracture, there are also circumstances in which the fat pads are not visualized, even though a fracture is present. If the joint capsule is torn by the fracture fragments, blood may escape from the joint, preventing joint distension and fat pad displacement. Marked edema may also obscure the fat pads, preventing their recognition even though they are displaced. Extra-articular fractures, such as those of the medial epicondyle, also do not result in hemarthrosis.
Approximately 70-90% of children with an elbow joint effusion following trauma have a fracture as demonstrated on the initial or follow-up radiographs. In most cases, the fracture is identified initially. A more important question concerns the risk of an occult fracture in those patients with an elbow effusion and no fracture is seen on initial examination.
Using sclerosis or periosteal new bone to indicate a healing fracture, Morewood initially estimated the risk of occult fracture in this setting to be approximately 30%.  Subsequent studies have shown considerable variability in this risk. Skaggs et al found healing fractures in 76%, supporting the concept of considering the presence of an effusion to be essentially indicative of an occult fracture.  More recently, Donnelly et al reported that the incidence of healing fracture on follow up was only 17%, with the difference likely related to improved radiographic technique and interpretive criteria for recognizing fractures initially.  Using multidetector CT at the time of initial injury, Chapman et al found fractures in approximately 50% of children with a postramatic elbow effusion but no fracture seen radiographically. 
Hence, although the presence of a posttraumatic elbow effusion in children raises the possibility of an occult elbow fracture, this situation is not synonymous with an occult fracture, as it seems equally likely that a fracture is not present. The clinical decision as to which patients need to be treated for a presumed fracture must be made on the basis of clinical and radiographic findings and knowledge of the associated risks.
The absence of an effusion in children is strong evidence against an intra-articular fracture. Although a torn capsule or marked edema may produce a false-negative fat pad sign, in these circumstances an obvious fracture is usually present. However, in adults, a significant number of fractures are not associated with an identifiable effusion. Therefore, the negative predictability of this sign is less useful in adults than in children. 
Use of the anterior humeral and radiocapitellar lines
Two lines may be drawn on radiographs to help in evaluating elbow trauma: the anterior humeral line and the radiocapitellar line (see below).
On a true lateral view of the elbow, a line drawn along the anterior aspect of the distal humeral metaphysis should pass through the middle third of the capitellum, which is also part of the humerus. Since the anterior humeral line evaluates the relative positions of 2 parts of the same bone, malalignment is indicative of a fracture. Specifically, if the anterior humeral line passes either through the anterior third of the capitellum or anterior to the capitellum, it indicates that the capitellum is displaced posteriorly relative to the humeral metaphysis. This displacement most frequently results from an extension-type supracondylar fracture, although posterior displacement of the capitellum may also be seen in lateral condyle fractures.
Although this sign is useful in interpreting findings in most children, caution is needed in young children. When the capitellar ossification center is small, its anterior aspect may still be well within the middle third of the entire capitellum, which is still mostly cartilage. In this case, the anterior humeral line may pass anteriorly relative to the small capitellar ossification center even without a fracture. The exact stage of development at which the capitellum is sufficiently ossified for this sign to be reliable is not certain. Initially, 2.5 years had been suggested along with the indication that occasionally comparison with the position of the capitellum of contralateral elbow may be useful.  More recent data suggest that the anterior humeral line reliably passes through the middle third of the capitellum in children older than 4 years. 
The radiocapitellar line evaluates the relationship of the proximal radius to the capitellum. Because the radius usually bends in the region of the tuberosity, with the radial neck angled anteriorly and laterally relative to the orientation of the radial shaft, only the most proximal part of the radius should be used for drawing the radiocapitellar line. This line should intersect the capitellum on all views, although in young children, the capitellar ossification center may occupy an eccentric position within the largely cartilaginous capitellum. Since the radiocapitellar line compares the relative positions of 2 adjacent bones, malalignment indicates dislocation.
Fractures of the Distal Humerus
Fractures of the distal humerus include supracondylar fracture, lateral condyle fracture, medial epicondyle fracture, medial condyle fracture, and transphyseal (transcondylar fracture), and T-condylar fracture. [2, 11, 18, 19, 20, 21, 22, 23, 24]
Supracondylar fractures are the most common elbow fracture in children, accounting for 50-60% of all elbow fractures. The peak age is 5-7 years, and the nondominant arm is involved more frequently than the dominant arm. While previously supracondylar fractures were more frequent in boy than in girls, this discrepancy has diminished. The vast majority (98%) of supracondylar fractures are extension injuries that result from a fall on an outstretched arm. With the elbow fully extended, or hyperextended with relative ligamentous laxity during childhood, the olecranon acts as a fulcrum to transmit the load into a bending force on the distal humerus in the supracondylar region. With such bending, the joint capsule applies a tension force to the anterior cortex of the distal humerus, accounting for the frequent anterior position of the lucent fracture line. [18, 25]
In the coronal plane, the fracture line extends transversely across the metaphysis at the level of the olecranon fossa. In the sagittal plane, the fracture may be transverse of oblique, extending upward from anterior to posterior.
Most supracondylar fractures involve posterior displacement or angulation of the distal fragment. Often, medial displacement accompanies supracondylar fractures. With medial displacement or medial comminution, loss of support for the medial aspect of the distal fragment allows the distal fragment to shift into varus alignment. When significantly displaced, supracondylar fractures usually have clinically obvious deformity.
The much less common flexion-type supracondylar fracture is usually caused by a direct blow to the posterior aspect of the elbow, usually from a fall onto the elbow. These fractures usually have anterior displacement of the distal fragment.
Radiographic findings in supracondylar fracture
Both direct and indirect findings are helpful in the radiographic diagnosis of supracondylar fractures. The presence of a joint effusion does not specifically indicate that a fracture is present, but a joint effusion does signal that a fracture is likely; in such cases, a careful search is required. The anterior humeral line may be extremely useful in the diagnosis of supracondylar fracture. In 94% of supracondylar fractures, an abnormally posterior position of the capitellum is demonstrated by passage of the anterior humeral line anterior to the middle third of the capitellum.  See the images below.
Abnormality of the anterior humeral line indicates distal humeral deformity and, therefore, either an acute or previous fracture.
Supracondylar fractures may be complete or incomplete and have a wide range of severity. The Gartland classification as modified by Wilkins and expanded by Leitch defines extension supracondylar fractures as follows [26, 27] :
Type 1 - Fractures with no or minimal posterior displacement or angulation of the distal fragment such that the anterior humeral line still intersects part of the capitellum
Type 2 - Fractures with more posterior displacement or angulation, but with an intact posterior cortex; type 2 fractures have been divided into type 2A, with no rotation or translation, and type 2B, with some rotation or translation in addition to posterior displacement and angulation
Type 3 - Fractures with displacement and complete cortical disruption (see the image below)
Type 4 - Fractures with displacement, complete cortical disruption, and complete loss of the periosteal hinge anteriorly and posteriorly leading to multidirectional instabilitySupracondylar fracture, type 3. Anteroposterior (A) and lateral (B) views show significant lateral and posterior displacement of a distal fragment.
With complete fractures, the fracture line and displacement are obvious. Even incomplete fractures often have enough disruption in 1 of the cortices (usually the anterior cortex) to make diagnosis easy (see the image below).
However, in approximately 25% of cases, the fracture may be subtle. These cases include greenstick and plastic bowing fractures.  With greenstick fractures, cortical disruption is seen on the tensile side (usually the anterior cortex), and they may be accompanied by cortical buckling of the compression side (usually the posterior cortex). With plastic bowing, no discrete fracture line is present. Rather, only deformity is observed, as demonstrated by the anterior humeral line. Subtle cortical deformity also may be present medially or laterally, which may be associated with varus or valgus deformity.
In searching for subtle fractures, knowing their expected location is essential. On the frontal view, supracondylar fractures typically extend transversely through the metaphysis across the region of the olecranon fossa. With subtle fractures, the fracture line may be initially seen through only a portion of the metaphysis. With healing, sclerosis is demonstrated across the entire metaphysis, indicating the full extent of the fracture (see the image below).
In the lateral projection, the fracture is often transverse, but may be oblique, extending proximally from anterior to posterior. The orientation of the fracture line in the sagittal plane has both diagnostic and clinical implications. Diagnostically, oblique fractures may be demonstrated more easily by use of an AP view with cephalad angulation, which shows the fracture en face. Although not routinely acquired, this view may be useful when a fracture is highly suspected but is not found on standard views. Clinically, obliquity is important because rotation along an oblique fracture line leads to varus or valgus in addition to deformity.
Complications of supracondylar fracture
The 2 major complications of supracondylar fractures in children are cubitus varus (see images below), which is relatively common, and vascular injury, which is uncommon but has considerable morbidity when present.
Since relatively little growth occurs at the distal humerus, angular deformity in most cases is not due to growth disturbance, but rather malunion of varus deformity. In some cases, cubitus varus results from medial comminution and collapse. Angular deformity also results from rotation at an oblique fracture line. In supracondylar fractures with medial displacement of the distal fragment, there is often internal rotation, which results in varus if the fracture is oblique. Less often, the distal fragment is displaced laterally, and these fractures tend to have external rotation, producing valgus.
Cubitus varus has also been recognized to result from posttraumatic trochlear deformity, which is likely due to avascular necrosis of the trochlear ossific nuclei or ischemic injury of growth plate chondrocytes following distal humeral fractures, most commonly supracondylar fractures.  Cubitus varus may be evaluated with the use of the Baumann angle, which is determined by lines drawn along the axis of the humeral shaft and the physis for the capitellum.
Although the Baumann angle does not define the true carrying angle of the elbow, it uses radiographically identifiable landmarks and is useful in comparison with the contralateral elbow. Care must be taken to ensure a true AP view, as rotation changes the Baumann angle. Recently, it has been suggested that at least 7 cm of the humeral shaft should be visualized and that using the medial or lateral cortices of the distal humeral diaphysis rather than the center of the shaft to define the humeral axis led to a more reproducible measurement of Baumann angle.  Cubitus varus after supracondylar fractures is relatively common and had previously been considered to be primarily a cosmetic problem. However, additional morbidity includes a predisposition to subsequent lateral condyle fracture, pain, and late development of posterolateral elbow instability. In cases in which it is clinically indicated, cubitus varus may be corrected by valgus osteotomy.
Vascular injury may be a severe complication of supracondylar fractures, usually occurring with significant posterior displacement of the distal fragment, with the brachial artery injured by the sharp distal end of the proximal fracture fragment. Initial evaluation of vascular injury is clinical. In those cases in which vascular injury is recognized, reduction usually corrects the vascular abnormality, and hence reduction and pinning should not be delayed for arteriographic assessment.  If this does not adequately restore circulation, vascular repair, usually following arteriography, may be needed. See the image below.
Nerve injuries may complicate supracondylar fractures. In a meta-analysis of 5154 supracondylar fractures in children, nerve injury occurred in 11%.  For extension fractures, the anterior interosseous branch of the median nerve is most frequently injured, whereas with the rare flexion type supracondylar fractures, the ulnar nerve is most often involved. In most cases, neurological deficit recover in a few months. Data also indicate that ulnar nerve injury may result from placement of a medial pin, with series showing no ulnar nerve injury in patients treated only with lateral pins versus a 7.7% risk with cross-pinning.  Although this has not been shown in several other series, including prospective studies, in most cases supracondylar fractures are now treated with only lateral pins to avoid nerve injury. [18, 19]
In 166 pediatric patients (median age, 7 yr) with supracondylar fractures referred for nerve injury consultation, the most commonly affected nerves were the ulnar (43.4%), median (36.7%), and radial (19.9%). A nondegenerative injury was seen in 27.5%, and 67.9% were degenerative injuries. According to the authors, referral to a nerve specialist following supracondylar fractures is recommended in cases of complete nerve palsy, a positive Tinel sign, or neuropathic pain or vascular compromise. 
The rate of flexion-type fractures has been estimated to be 1.2%. In one study, 7 out of 606 supracondylar humeral fractures were flexion-type injuries. The mean annual incidence was 0.8 per 105. In long-term follow-up, mean carrying angle was 50% more in injured elbows (21°) than in uninjured elbows (14°). [34, 35]
Supracondylar fractures may be associated with ipsilateral fractures remote from the elbow, most frequently of the distal radius. See the image below.
Lateral condyle fracture
Fractures of the lateral condyle are the second most common elbow fracture in children, accounting for 15-20%. Although lateral soft tissue swelling may be prominent, clinically evident deformity is less common in lateral condyle fracture compared with supracondylar fracture. Many pediatricians and emergency physicians are not as familiar with these fractures as they are with supracondylar fractures, and some lateral condyle fractures may be subtle. As a result, accurate and timely radiographic interpretation is essential for alerting the clinical staff to the features of the fractures and the need for orthopedic treatment.
In most cases, lateral condyle fractures are distraction injuries from the forearm extensors, usually as a result of acute varus stress applied to an extended elbow. Less commonly, some may be due to axial force transmitted through the radius. The fracture originates in the lateral aspect of the distal humeral metaphysis and passes obliquely to the physis. In most cases, the fracture line then partially traverses the physis and then passes into the cartilaginous distal humeral epiphysis (see the image below).
Hence, lateral condyle fractures are Salter-Harris type IV injuries, even though they often have the radiographic appearance of a Salter-Harris type II injury. Less frequently (4 of 48 in Jakob's series), the fracture passes through the lateral aspect of the metaphysis, crosses the physis, and continues through the ossified capitellum, with the typical radiographic appearance of a Salter-Harris type IV fracture (see the image below). 
The stability of the distal fragment is partly determined by whether the fracture extends all the way to the articular surface or whether a cartilaginous hinge remains intact to help prevent motion of the fracture fragment.
The Milch classification scheme for lateral condylar fractures defines a type I fracture as one that passes through the distal humeral epiphysis lateral to the lateral crista of the trochlea, in most cases passing through the ossified capitellum. For these fractures, the lateral crista of the trochlea is intact, maintaining stability of the elbow joint. However, the adjacency of fracture margins for the metaphysis and capitellum poses the risk of focal physial closure. The more frequent Milch type II fracture follows dense collagenous fibers through the epiphyseal cartilage into the trochlea medial to the lateral crista. In this case, the lateral crista is part of the distal fracture fragment, leading to instability of the elbow joint. 
A staging system for displacement of lateral condyle fractures is as follows  :
Stage I fractures have an intact articular surface. These may have some angulation but no true displacement of the fracture fragment and no shift of the olecranon.
Stage II fractures extend through the articular surface, allowing for a small amount of displacement of the distal fragment and olecranon shift.
Stage III fractures have significant displacement, usually laterally and proximally, leading to translocation of the olecranon and radial head. In addition, traction from the common extensor muscles leads to rotation so that the cartilage-covered articular surface of the fractured lateral condyle is in contact with the metaphysis, leading to nonunion if not corrected.
Radiographic findings in lateral condyle fracture
The radiographic depiction of lateral condyle fractures depends on the degree of separation at the fracture site. If separation is significant, as shown below, recognition of the fracture is easy, although distinguishing these fractures from supracondylar fractures depends on knowing the characteristic course (see the image below).
When no displacement is present, findings indicating a lateral condyle fracture may be subtle. A joint effusion helps in suggesting a subtle fracture; lateral soft tissue swelling localizes the region to be examined most carefully. Nondisplaced or minimally displaced (< 2 mm) lateral condyle fractures, may be stratified according to their risk of subsequent displacement. Type A fractures have no or minimal gap at their lateral aspect and cannot be traced all of the way to the physis. Type B fractures are similar, other than the fracture line can be traced to the physis. With type C fractures, the fracture line remains is as wide medially as laterally. In one study, alltType A fractures were stable, whereas 17% of type B fractures and 42% of type C fractures showed subsequent displacement. 
Although posteriorly displaced lateral condyle fractures may show an abnormal relationship between the anterior humeral line and the capitellum, this finding is not as useful in lateral condyle fractures as in supracondylar fractures. These fractures often demonstrate only a subtle subcortical fracture line along the lateral aspect of the metaphysis, as shown below. Although only a very thin sliver of bone may be viewed, it represents the small ossified portion of the entire distal fragment that is mostly cartilage (see the image below).
Oblique views may be required to depict these fractures, since some are not apparent on AP views. In particular, the internal oblique view has been shown to be better than the AP view for showing the presence of lateral condyle fracture, the degree of displacement, and findings suggesting instability.  On lateral views, cortical disruption is usually seen posteriorly rather than anteriorly, as with supracondylar fractures (see the image below).
Because several secondary ossification centers exist in the elbow, a small flake of bone adjacent to the metaphysis may be misinterpreted as a developmental center, such as the lateral epicondyle. However, because the lateral epicondyle is the last center in the elbow to ossify, most pediatric patients with lateral condyle fractures have elbows that are too immature to have a lateral epicondyle ossification center. Therefore, the flake of bone must represent a fracture fragment. However, caution should be taken where there is partial overlap of the capitellum with the metaphysis. The double density caused by such overlap may simulate a flake of bone, with lucency of the physis simulating an adjacent fracture line.
Although the radiologic diagnosis of lateral condyle fracture depends on plain radiographic findings, MRI, arthrography, or ultrasonography (US) may be useful in the further evaluation of the fractures, particularly with regard to the course of the fracture through the cartilaginous epiphysis, as shown below. A role for 3-D kinematic analysis of CT and MRI imaging for evaluation of lateral condyle fracture with nonunion has also been proposed  .
Other injuries that may be confused with lateral condyle fractures include supracondylar fracture, true Salter-Harris type II fracture, and, in young infants, separation of the distal humeral epiphysis (transphyseal fracture, Salter-Harris type I). Supracondylar fractures usually extend transversely across the metaphysis, whereas lateral condyle fractures are oblique and more distal. The rare Salter-Harris type II may mimic lateral condyle fracture radiographically, but not clinically. These injuries are due to valgus rather than varus stress and distract the physis starting medially. The fracture then propagates through the physis and eventually passes into the metaphysis, producing a typical Salter-Harris appearance. However, these injuries have marked medial soft tissue swelling compared with the lateral soft tissue findings with lateral condyle fracture.  Distinction between lateral condyle fracture and transphyseal fracture is discussed in that section.
Complications of lateral condyle fracture
Unlike supracondylar fractures, vascular and neurologic complications are extremely rare with lateral condyle fractures. However, lateral condyle fractures may be complicated by instability (see the image below), avascular necrosis, and malunion or nonunion, which are more problematic for lateral condyle fractures than supracondylar fractures.
In the series by Jakob et al involving 48 patients with lateral condyle fractures, 20 patients had fractures that were minimally displaced; 28 patients had significant displacement that required surgical reduction and fixation.  Nonunion has been considered to be more of a problem in patients with minimally displaced fractures than in patients with significant displacement, presumably because the lack of surgical fixation allows a small amount of motion and because of the development of fibrocartilage between the fragments.  Nonunion often leads to valgus deformity from a lateral shift of the fracture fragment. Valgus may also result from malunion, and varus deformity may be caused by malunion or stimulation of growth of the lateral condylar physis. These deformities may cause posttraumatic arthritis with pain and diminished range of motion, which are often not correctable.
For comparison, with supracondylar fracture, which does not involve the articular surface, elbow motion is maintained and cubitus varus is correctable. Although acute nerve injury is rare, elbow deformity following lateral condyle fracture may lead to ulnar neuritis (tardy ulnar palsy), a late complication (average interval from injury = 22 years).  See the image below.
Since Milch II lateral condyle fractures separate the lateral crista of the trochlea (lateral trochlear ridge) from the rest of the trochlea, there may be accompanying elbow dislocation through loss of lateral support for the olecranon process (see the image below).
Lateral condyle fractures may be associated with other elbow fractures. Particularly common are those involving the olecranon (shown below), which occur with varus stress applied to a fully extended elbow with the olecranon locked in the olecranon fossa.
However, unlike supracondylar fractures, lateral condyle fractures are seldom associated with fractures remote from the elbow.
Medial epicondyle fracture
Fractures of the medial epicondyle account for 10-15% of elbow fractures in children. Medial epicondyle fractures are 3 times more common in boys than girls and tend to occur in older children more often than supracondylar or lateral condyle fractures, with a peak age of 11-12 years, although younger children may also be affected. Most medial epicondyle fractures are avulsion injuries caused by traction from the ulnar collateral ligament or the forearm flexor muscles that arise from the medial epicondyle. Such distraction injuries may arise from valgus stress applied to an extended elbow or muscular stress. Distraction may also result from the ulnar collateral ligament with elbow dislocation or subluxation, which accounts for approximately half of medial epicondyle fractures in children. Rarely, the medial epicondyle may also be fractured by direct trauma. 
Radiographic findings in medial epicondyle fracture
Medial epicondyle avulsions may include separation of the entire medial epicondyle from the metaphysis, avulsion of only part of the medial epicondyle (see the image below), or avulsion of the epicondyle together with a small portion of the adjacent metaphysis.
These injuries resemble Salter-Harris type I, III, and II fractures, respectively, though the Salter-Harris classification is usually applied to injuries of the epiphyses rather than those of the apophyses. For those injuries that include a small portion of the metaphysis, care must be taken to distinguish medial epicondyle fracture (usually an extraarticular injury) from medial condyle fracture, which extends to the articular surface.
Owing to traction from the forearm flexors, the medial epicondyle is displaced distally, and usually medially, from its anatomic position (see the image below).
Localized soft tissue swelling is usually present. In most patients, the medial epicondyle is extra-articular; therefore, a joint effusion is not present. In some cases, widening of the physis and displacement of the medial epicondyle may be quite subtle, and comparison views of the contralateral elbow may be useful. Evaluation of displacement of the medial epicondyle may also be aided by recent data regarding the position of the normal medial epicondyle relative to other distal humeral landmarks.  Radiographic evaluation of the amount of displacement is also known to be limited, with many cases showing substantially more displacement by CT than radiography. 
Complications of medial epicondyle fracture
The fractured medial epicondyle may become entrapped in the elbow joint, representing a major complication. With acute valgus stress, the medial side of the elbow joint is opened. When the medial epicondyle is pulled downward (distally) by the forearm flexor muscles, it may enter the medial joint space. When the valgus force is removed, the medial epicondyle may then become entrapped as the medial joint space closes.
Examples of entrapment of the medial epicondyle in a young child, before ossification of the trochlea occurs, and of entrapment in an older child, after trochlear ossification has occurred, are presented (see the images below).
Entrapment is particularly common after an elbow dislocation or subluxation. It is important that such entrapment be recognized; the diagnosis may be made on the basis of radiographic findings. To make the diagnosis, it is helpful that the radiologist be familiar with the normal developmental anatomy of the elbow. Because the entrapped medial epicondyle is positioned just distal to the medial side of the distal humeral metaphysis, it may be misinterpreted as the ossification center for the trochlea. However, the trochlea does not become ossified before the medial epicondyle. Therefore, the trochlea should not be seen unless the medial epicondyle is identified as well. In addition, usually, the trochlea initially appears as multiple fragmented ossification centers; by contrast, the medial epicondyle has a smooth and regular appearance.
Entrapment of the medial epicondyle may be difficult to detect on the frontal view; such entrapment is often better depicted on the lateral view. On the lateral view, a clue that is helpful in recognizing entrapment of the medial epicondyle is widening of the medial joint space. However, widening of the joint space may be difficult to evaluate in patients in whom the elbow is immature; in such cases, the largely cartilaginous trochlea makes the normal gap between the distal humerus and ulna appear quite wide.
In the radiographic evaluation of pediatric elbow trauma, it is important to assess the status of the medial epicondyle, particularly after an elbow dislocation. If the elbow is mature enough for ossification of the medial epicondyle to be expected, the position of the medial epicondyle should be verified. If the medial epicondyle is not seen in its normal anatomic position, it should be searched for elsewhere, including within the elbow joint.
Avulsion fractures of the medial epicondyle may occur before ossification, and they cannot be detected on plain radiographs. However, such an injury may be suggested by localized tenderness and soft tissue swelling and by the presence of a posterolateral elbow dislocation. Stress radiographs demonstrating widening of the medial joint space with valgus stress indicate either avulsion of the medial epicondyle or disruption of the ulnar collateral ligament. In children, the ligaments are generally stronger than the bone; therefore, avulsion fractures occur more frequently than ligamentous injury. MRI is useful in identifying medial epicondyle fractures prior to ossification of the medial epicondyle and for delineating the full extent of the cartilaginous fracture in children with a small medial epicondyle ossification center. See the image below.
Conventional, magnetic resonance, or CT arthrography may be helpful in searching for a cartilaginous entrapped medial epicondyle in patients in whom the medial epicondyle is intra-articular.
Medial condyle fracture
Fracture of the medial condyle is an uncommon injury in children. As with lateral condyle fractures, these are typically Salter-Harris type IV injuries. The fracture extends through the metaphysis and into the epiphysis, typically arising just above the medial epicondyle and extending to the trochlear groove, as shown in the image below.
In young patients with a nonossified or only partially ossified trochlea, the epiphyseal component of the fracture is not visible, and only the metaphyseal flake is identifiable. The medial epicondyle is included in the distal fragment. As with lateral condyle fractures, medial condyle fractures are often unstable and may be complicated by nonunion. Like lateral condyle fractures, medial condyle fractures may show marked rotation of the fracture fragment (see the image below). Although it is important to differentiate medial condyle fractures from medial epicondyle fractures, the distinction is not always easy to make with radiographs.
The presence of a metaphyseal flake fracture is not specific because some medial epicondyle avulsions extend into the metaphysis as a Salter-Harris type II fracture. In general, medial condyle fractures (Salter-Harris type IV injuries) have larger metaphyseal components than medial epicondyle fractures that involve the metaphysis have. Joint effusion is more likely to be present with medial condyle fractures, although joint effusions may be seen with medial epicondyle avulsion fractures. Clinical features that suggest a medial condyle fracture include instability and a limitation of elbow motion.
Transphyseal fracture (also called transcondylar fracture) is a fracture through the distal humeral physis that separates the entire distal humeral epiphysis from the metaphysis. In most patients, the fracture is a Salter-Harris type I injury, passing entirely through the growth plate. In some cases, the fracture may extend into the metaphysis, producing a Salter-Harris type II injury.
Transphyseal fractures most often occur in young children (< 2 y); they are reportedly associated with birth injury and child abuse. The mechanism of injury may be rotational shear.  It has also been suggested that extension force in infants may be more likely to cause a transphyseal fracture than supracondylar fracture. In infants, the distal humeral physis has a flat transverse configuration without the stability afforded by a V-shaped configuration in older children. The physial line is also located more proximally in infants, predisposing them to a physial fracture from a force that would have caused a supracondylar fracture in an older child.  Because the distal humerus has a broader base at the physis than in the region of the olecranon fossa where supracondylar fractures occur, there is more contact between the fragments, and hence less tilting.
In a transphyseal fracture, the epiphysis is usually medially displaced relative to the metaphysis (see the image below).
The proximal radius and ulna maintain a normal relationship with respect to the epiphysis; hence, the forearm bones are also displaced relative to the humeral metaphysis. In young children in whom the distal humeral epiphysis is not yet ossified, this malalignment of the forearm bones and the distal humeral metaphysis may be mistaken to indicate an elbow dislocation. Often, the capitellum has ossified; in such cases, it may serve as an important marker in the otherwise cartilaginous distal humeral epiphysis.
Demonstration of normal alignment between the proximal radius and the capitellum (radiocapitellar line) and normal alignment of the proximal radius and ulna with each other are the keys to differentiating transphyseal fracture from elbow dislocation. If the capitellum is not yet ossified and hence cannot be used to evaluate elbow alignment, the direction of displacement of the forearm bone relative to the distal humeral metaphysis may be useful in distinguishing transphyseal fracture from elbow dislocation. In transphyseal fracture, the distal humeral epiphysis and forearm bones are usually displaced medially, whereas in true elbow dislocations, the radius and ulna are dislocated either laterally and posteriorly (in children >2 y) or primarily posteriorly (in children < 2 y). MRI, US, or arthrography may be used to directly depict the relationship of the cartilaginous distal humeral epiphysis to the metaphysis (see the image below).
Some transphyseal fractures include a small portion of the metaphysis as shown in the image below; such a finding is helpful in recognizing that a fracture is present.
However, this finding may cause the injury to be confused with a lateral condyle fracture. Distinguishing between these fractures is important because lateral condyle fractures are often unstable and require operative fixation, which is frequently not necessary for transcondylar fractures, which are more stable following reduction.
Features that help in distinguishing between transphyseal and lateral condyle fractures include alignment of the radiocapitellar joint and the direction of displacement. In transphyseal fractures, radiocapitellar alignment remains normal, whereas in lateral condyle fractures, the distal fragment is often displaced or rotated, as described above, with alteration of the radiocapitellar alignment. Because the lateral crista of the trochlea is often included in the fracture fragment, the elbow joint loses lateral support in lateral condyle fractures. Thus, lateral displacement of the proximal forearm bones is seen in lateral condyle fracture, rather than medial displacement, which is typically seen in transphyseal fractures.
In most cases, patients with transphyseal fractures have a good prognosis, although correct diagnosis may be problematic. In some patients, impaction of the epiphysis on the medial aspect of the metaphysis may cause growth plate injury, leading to subsequent varus deformity (see the image below). It should be borne in mind that transphyseal fractures are associated with child abuse.
T-condylar fractures are uncommon in pediatric patients, particularly prior to skeletal maturity, although they may be misdiagnosed as other elbow injuries with clinical presentation often similar to supracondylar fracture and radiographs that may be confused with supracondylar, lateral condyle, or medial condyle fractures. In addition to a transverse or oblique component through the supracondylar region, the distinguishing aspect of T-condylar fracture is a sagittally oriented component that extends to the articular surface, splitting the medial and lateral condyles.
T-condylar fractures may result from flexion or extension injury, with the articular surface of the olecranon acting as a wedge to split the humeral condyles. Flexion injury is often from a fall on a flexed elbow, whereas extension injury is often from a fall on a slightly flexed and outstretched arm, with the coronoid process acting as the wedge. 
Other Elbow Injuries
Fractures of the proximal radius
Although the proximal radius is the most common site of elbow fracture in adults, it accounts for only 5% of elbow fractures in children. When proximal radial fractures occur in children, they primarily involve the radial neck. Fractures of the radial head epiphysis are uncommon in children.
Radiographic findings of proximal radius fractures
Most proximal radial fractures in children are either Salter-Harris type II injuries that extend through the growth plate and the lateral aspect of the metaphysis or metaphyseal fractures that extend across the neck near the growth plate but do not involve the growth plate directly. Rarely, a Salter-Harris type IV fracture extends vertically through the metaphysis and epiphysis, crossing the physis. With some proximal radial fractures, no displacement of the epiphysis occurs; detection of the fracture depends on the metaphyseal component, which may show only subtle abnormal angular deformity, as in the image below. This finding must be distinguished from the normal angulation that is usually present at the junction of the radial neck and shaft.
The radial head epiphysis may show displacement with varying amounts of shift and angulation that may lead to limitation of motion of the proximal radioulnar joint.  Some proximal radial fractures may result in abnormal articulation of the radial head and capitellum and therefore are fracture/dislocations. Displacement of the radial head may be marked, usually with the head displaced distally, and its articular surface may be rotated into the coronal plane posteriorly. However, the displacement may also be lateral, as shown in the image below.
Displaced proximal radial fractures may result from transient posterior elbow dislocation. When the proximal radius and ulna return to normal position, the capitellum may shear off the radial head, leaving it posteriorly displaced. During reduction of these completely displaced fractures, the radial head may become inverted, such that the physial fracture surface of the radial head articulates with the capitellum. Less often, as the proximal radius and ulna are dislocating posteriorly, the capitellum holds the radial head in position, causing the radial neck fracture and leaving the radial head displaced anteriorly and distally.
Complications of proximal radius fractures
A major complication of a radial neck fracture is limitation of motion at the proximal radioulnar joint, which mostly limits supination. This complication is usually caused by malalignment of the radial head and neck; more severe limitation of motion may result from radioulnar synostosis. Radial head displacement or injury to the proximal radial growth plate may cause growth arrest, leading to radial shortening that may affect alignment of the wrist. Proximal radial fractures in children are frequently associated with other injuries; such injuries most frequently involve the olecranon. The identification of a proximal radial fracture should alert the examiner to carefully search for other injuries.
Fractures of the proximal ulna
Fractures of the proximal ulna are uncommon in children, accounting for 6% of elbow fractures. In evaluating the proximal ulna in children, the normal olecranon apophysis must not be mistaken for a fracture fragment. The olecranon apophysis usually appears in children at approximately age 10 years, and it fuses by age 18 years. The normal apophysis may have separate ossifications centers near its tip.
The olecranon apophysis fuses in an anterior-to-posterior direction; radiographs may reveal a residual posterior cleftlike lucency with well-defined sclerotic margins. The characteristic location of the olecranon ossification centers, their smooth uninterrupted cortical margins, and the typical appearance of the partially fused physis help in distinguishing olecranon ossification from fractures at that site.
Radiographic findings of proximal ulna fractures
Distraction stress on the olecranon may occur from falling on an arm with the elbow partially flexed so that acute hyperflexion stress is applied against the triceps. Alternatively, it may result from excessive muscular activity, often in association with throwing. The incidence of distraction fractures is particularly high in patients with osteogenesis imperfecta, including patients with relatively normal-appearing bones and few fractures elsewhere (see the image below).
Some distraction fractures of the olecranon may be subtle, whereas others may have significant proximal displacement of the fracture fragment. These fractures are usually Salter-Harris type II injuries that include a metaphyseal fragment of variable size. Salter-Harris type I fractures that pass entirely through the physis of the olecranon apophysis may occur, but they are relatively uncommon. The detection of these fractures requires a high index of suspicion and comparison with the noninjured elbow.
When the elbow is fully extended, the olecranon becomes locked into the olecranon fossa, making it susceptible to fracture by varus or valgus stress. These fractures may be subtle and have only a linear lucent line through the trabecular region, as shown in the image below.
In other patients, the fracture is best seen at the proximal tip of the olecranon metaphysis, as depicted in the image below.
Complications of proximal ulna fractures
Valgus stress fractures may be associated with a compression fracture of the radial neck or avulsion of the medial epicondyle. Varus stress fractures may be associated with a lateral condyle fracture or a lateral dislocation of the radial head (type 3 Monteggia fracture/dislocation).
Olecranon fractures are often associated with other injuries. It is believed that the most common injuries found in association with olecranon fractures are fractures of the proximal radius. Olecranon fractures may be associated with lateral condyle fractures with varus stress or medial epicondyle fractures with valgus stress.
Fractures of the coronoid process are infrequent in children, but they may be seen with posterior elbow dislocation.
The elbow is the most frequently dislocated joint in children, whereas in adults, dislocations of the shoulder and interphalangeal joints of the fingers are more common. Elbow dislocation accounts for approximately 5% of elbow injuries in children. The mechanisms of dislocation include a fall on an outstretched arm with the elbow partially flexed and forced hyperextension, although both mechanisms more frequently result in fractures than in dislocations. The most common direction of displacement is posterior or posterolateral (see the images below), although lateral and anterior dislocations also occur.
Dislocations often are associated with fractures, most often involving the medial epicondyle and coronoid process of the ulna. Other fractures that may be associated with elbow dislocations include fractures of the proximal radius, particularly fractures in which the radial head is markedly displaced and rotated into the coronal plane; fractures of the lateral condyle; and remote fractures in the same extremity, most often the distal radius and ulna. Detection of an elbow dislocation should alert the radiologist to carefully search for the other injuries.
Radiographic findings of elbow dislocation
Elbow dislocations are usually readily apparent on radiographs. In young patients, alignment of the radiocapitellar joint is evaluated by using the radiocapitellar line, whereas in the more mature skeleton, articulating surfaces of the radial head and capitellum are revealed directly. The articular relations of the medial condyle and proximal ulna are not as easy to evaluate in the immature skeleton.
After spontaneous reduction, prior elbow dislocation may be suggested by the identification of the fractures described above. In patients younger than 2 years, elbow dislocations are exceedingly rare, and transphyseal fractures (distal humeral epiphyseal separation) are often mistaken for elbow dislocation. Radiographic findings that indicate transphyseal fracture rather than dislocation include maintenance of normal radiocapitellar relations and medial displacement of the forearm bones.
Complications of elbow dislocation
Complications of elbow dislocation in children include associated fractures, neurologic injury (usually involving the ulnar nerve or the anterior interosseous branch of the median nerve), joint contracture, and heterotopic ossification in the regions of the disrupted medial or lateral collateral ligaments. Vascular complications are less common than neurologic injury and are usually accompanied by severe injuries, often including open fractures. 
Monteggia fracture/dislocation involves dislocation of the radial head accompanied by fracture of the proximal or mid ulna, with the apex of the ulnar fracture pointing in the same direction as the radial head dislocation. Normal articulation of the medial condyle and proximal ulna is maintained. In 55-85% of patients, the radial head is anteriorly dislocated, with an associated apex anterior ulnar fracture (Monteggia type 1 injury). In the remainder of patients, fractures/dislocations are divided equally between posterior (Monteggia type 2 injury) and lateral (Monteggia type 3 injury) dislocation of the radial head. Lateral (Monteggia type 3) injuries most often occur in children 5-9 years of age (see the image below).
Simplistically, a Monteggia fracture/dislocation may be thought of as the result of a force that dislocates the radial head and simultaneously fractures the ulna in the same direction. However, in most patients, the injury is caused by a fall onto a pronated forearm, which forces the arm into hyperpronation. This motion causes the ulna to fracture and contact the proximal radius, forcing the radial head to become dislocated from the capitellum.
In children, an ulnar fracture often is manifested by plastic bowing without a discrete fracture line, as shown in the image below.
In some patients, the finding may be subtle; recognition of this injury requires a high index of suspicion and the use of comparison views of the contralateral forearm, when needed. Most cases of isolated radial head dislocation in children are likely to actually be Monteggia fracture/dislocation with a subtle ulnar bowing fracture. Conversely, ulnar fractures in a child are often accompanied by a radial fracture or dislocation, even if the ulnar fracture is a relatively subtle greenstick injury. If an associated radial fracture is not identified, a careful search should be made for a radiocapitellar dislocation or subluxation. The elbow should be well visualized in all patients who have an ulnar injury, with or without an associated radial fracture.
A Monteggia variant has fractures of the radius and ulna. The radial fracture is so close to the joint that the injury may superficially resemble a radial head dislocation. In these cases, only the radial head is still in alignment with the capitellum. The rest of the radius appears dislocated with respect to the capitellum; however, this is a displaced fracture rather than a dislocation (see the image below).
In cases in which the radial head is not yet ossified, this injury cannot be distinguished from a true Monteggia fracture/dislocation by use of plain radiographs.
A similar situation occurs in the wrist in children; that is, a fracture through the distal ulnar physis may occur in association with a distal radial diaphyseal fracture and result in a pseudo-Galeazzi injury (see the image below).
In fact, Monteggia variant and pseudo-Galeazzi injuries are forearm fractures involving both bones, with 1 of the fractures occurring so close to the joint that a dislocation is erroneously suggested.
A pulled elbow is a distraction injury. It is also called nursemaid's elbow and other names; it usually results from a sudden pull on the hand. In children younger than 5 years, the annular ligament is relatively loose, allowing the radial head to be pulled through it when acute traction is suddenly placed on a pronated forearm (which is the usual position of the forearm when a child is being pulled along by an adult). Although the annular ligament becomes transiently interposed between the radial head and capitellum, this movement does not cause recognizable widening of the radiocapitellar joint. Therefore, elbow radiographic findings are normal in a pulled elbow. MRI could demonstrate the abnormal relationship of the radial head and annular ligament, but such studies are seldom needed.