Introduction

Radial head fractures are very common. They comprise of up to 5% of all fractures and up to one-third of elbow fractures. Radial head and neck fractures have been estimated to occur in nearly 20% of all elbow trauma.¹ The incidence of radial head and neck fractures in the entire population has been shown to be 2.5 per 10 000 per year and they account for 0.2% of all visits to accident and emergency departments.²

Radial head fractures occur more often than radial neck fractures in a ratio of 7 : 1. The average age of patients at the time of fracture is 45 years old. Although male : female ratio is approximately equal, in general men are older at the time of fracture (48 vs 41 years old) and sustain more severe types of injury.³

Because of the high incidence of proximal radial fractures it is important for any clinician treating these injuries to have a good knowledge of its function. The role of the proximal radius and its relationship and interaction with other structures at both the elbow and forearm has become increasingly clear in recent years. Anatomically, the proximal radius is formed by the radial neck and radial head. The radial neck is about 13 mm long⁴ and assures both the position of the radial head onto the capitellum and aligns the radiohumeral joint during forearm rotation. The radial neck is not straight on the radial shaft, but forms an average angle of 16.8°. This angle is variable, ranging from 6° to 28°.⁴ This plays a role in forearm rotation, as the rotation axis passes through the radial neck and radial head.⁵ The radial head articulates with the capitellum on the distal humerus, and also with the sigmoid notch (radial notch) on the proximal ulna, forming the radiohumeral and proximal radio-ulnar joints respectively. The radial head is therefore an important anatomical structure for both flexion and extension of the elbow as well as pronation and supination of the forearm. There is also some longitudinal motion between the radius and the ulna during rotation of the forearm. The radius moves about 1 mm proximally relative to the ulna when the forearm is rotated from supination to pronation.⁵

Besides motion, the proximal radius has biomechanical functions. The radial head plays a key role in load transfer between the forearm and the humerus. It is assumed that up to 60% of the load on the forearm is transferred to the humerus via the radiohumeral joint.⁶ Both the outer shape of the radial head and the articulating inner dish are elliptical in shape;⁷ this is an important feature for efficient transfer of forces as different areas of the dish will be in contact with different positions of the elbow and forearm. Load through the proximal radius depends on the position of the elbow and forearm and is greater with the elbow in extension and the forearm pronated.⁸ Varus or valgus stress also alters the load transfer between the forearm and the humerus; varus stress can decrease the portion of the load through the radiohumeral joint to less than 10%.⁹

Besides playing a role in mobility and load transfer, the proximal radius also acts as a stabilizing structure. The stabilizing role of the radial head has been studied extensively. The radial head has been shown to be a secondary stabilizer to valgus stress, with the medial collateral ligament as the primary stabilizing structure.¹⁰ Together with the lateral ulnar collateral ligament (LUCL) and the coronoid, the radial head has recently also been shown to be important in posterolateral rotatory (in) stability (PLRI).¹¹

Background/aetiology

Proximal radial fractures typically occur from a fall on the outstretched hand, with the forearm in pronation. The impact is transferred through the radiocapitellar joint and the radial head and/or neck when it impacts onto the capitellum. There is also palmar translation of the radius in pronation. This is why most commonly the radial head fracture fragment is located on the anterolateral side of the radial head. Concomitant displaced capitellar fractures only occur in about 2%;³ however, smaller osteochondral lesions of the capitellum are ten times more common and capitellar bone bruising occurs in almost all patients.¹² The radial head has been shown to fracture with the elbow flexed in a range between 0° and 80°.¹³ Finally, direct impact to the elbow may also cause the proximal radius to fracture, although this occurs much less frequently.

Associated lesions frequently complicate proximal radius fractures. The likelihood of associated lesions increases with the severity of the fracture. This increases from about 10% of patients with a non-displaced fracture, to 50% of patients with a displaced fracture and up to as much as 75% of patients with a comminuted fracture. Generally, approximately one-third of patients have one or more additional lesions adjacent to the proximal radius fracture. Nearly 25% of patients have ligamentous rupture or one or more additional fractures to the elbow. Clinically significant ligamentous lesions that require some form of treatment occur in 10% of fractures, where either the lateral collateral ligament (LCL), medial collateral ligament (MCL) or both are ruptured.³ If an MRI investigation is performed after the fracture, ligamentous injuries are found much more frequently: up to 80% of cases.¹² Fortunately only a small percentage of these will need to be addressed surgically.

Associated articular elbow fractures are found in 15% of cases, with coronoid fractures being the most common (10%). In addition, between 15% and 20% of all radial head fractures are associated with a dislocation of the elbow.^3,14 Many patients in this group need secondary treatment and will develop chronic problems such as stiffness or residual instability and require further treatment.

Distal radio-ulnar joint dissociation (Essex-Lopresti injury) is diagnosed acutely in only 0.3% of patients following a radial head fracture.³ This may, however, be an underestimation of the number of patients that have an Essex-Lopresti injury, as longitudinal instability can easily be missed at the initial presentation. Despite the small number of patients, it is extremely important to be vigilant for this type of injury as it can result in severe long-term disability if undiagnosed. Nine out of ten patients operated for longitudinal radio-ulnar dissociation at the Mayo Clinic between 1997 and 2002 were treated for chronic lesions as a consequence of failure of previous treatment.³

Radial head fractures are historically classified according to the Mason classification.¹⁵ Recently, it has become increasingly clear that the outcome of the treatment of radial head fractures does not depend solely on the type of fracture but also on associated lesions. As was stated above, clinically significant associated lesions occur in a large percentage of patients, and this does depend on the type of fracture. Several changes have therefore been made to the original Mason classification; the most recent adaptation was based on a demographic study of over 300 radial head fractures and their associated lesions.^1,16

The proximal radial fracture itself is classified according to the original Mason classification, from types I to III. A type I fracture is non-displaced. The fragments are displaced more than 2 mm in a type II fracture, and a comminuted, non-reconstructible radial head fracture constitutes a type III. The Mayo extended classification then adds a suffix to the fracture type to show any associated lesions.¹ A ‘c’ is added for coronoid fractures and an ‘o’ for olecranon fractures. Ligamentous lesions are represented by ‘m’ for medial collateral ligament (MCL) lesions, ‘l’ for the lateral collateral ligament (LCL) complex and ‘d’ for longitudinal distal radio-ulnar joint (DRUJ) dissociation (Fig. 21.1). This classification, as well as guiding treatment, was extended to provide information after treatment. The classification is adapted post surgery depending on the structures that needed surgical care. Capital letters are used if surgery was necessary. For example, ‘type IImL’ means that both the medial and lateral ligaments were ruptured in a displaced fracture, but that only the LCL was repaired. Finally, all options for treatment of the radial head fracture have been included as well; ‘X’ (excision), ‘F’ (fixed), ‘P’ (prosthesis). Results of treatment of radial head fractures can be compared more easily if associated lesions are taken into account using this classification system.

Figure 21.1 van Riet and Morrey,¹⁶ Mayo Classification¹ of radial head fractures. Type I fractures are non- or minimally displaced fractures that can be treated conservatively. Type II fractures are displaced fractures that probably need surgical reduction and fixation. Type III fractures are non-reconstructible, comminuted radial head fractures. These are best treated with radial head resection or prosthetic replacement. Associated lesions are classified by adding a suffix to the description of the fracture type. m: MCL; l: LCL; d: DRUJ; c: coronoid process; o: olecranon. Upper case is used if the associated lesion needed treatment. A second suffix is added illustrating radial head treatment. X: excision; F: fixation; P: prosthesis.

Adapted with permission from: van Riet RP, Van Glabbeek F, Morrey BF. Radial head fracture. In: Morrey BF, ed. The elbow and its disorders. 4th edn. Philadelphia, PA: Saunders Elsevier; 2009:359–381.

Presentation, investigation and treatment options

An isolated radial head fracture itself may not be very painful. Immediately following the trauma there is often only a deep ache and the range of motion may be normal. Within a few minutes, however, a haemarthrosis occurs and the increased pressure within the capsule results in pain. Thereafter, the range of motion decreases due to both the pain and mechanically as a result of the swelling.

Patients with a proximal radial fracture present at the accident and emergency department with a painful elbow. The history will usually reveal a fall on the outstretched hand, but other mechanisms such as direct trauma are possible. The elbow is inspected for bruising or obvious deformity, indeed dislocation of the elbow. A haematoma on the medial side may be a sign of an MCL rupture (Fig. 21.2). A visible and palpable swelling over the lateral ‘soft spot’ is indicative of a haemarthrosis. The elbow capsule can contain nearly 25 mL of fluid.¹⁷ Flexion and extension motion decreases nearly 2° per millilitre of intracapsular fluid.¹⁸

Figure 21.2 A haematoma is seen on the medial side of the elbow of this patient. This is indicative of an additional medial injury and may be a sign of an MCL lesion.

Rotation of the forearm may be nearly normal in non-displaced fractures, being limited and painful only when the fragments are displaced. Crepitus may also be heard or palpated when the elbow or forearm is moved. The patient may be guarding the elbow. Intra-articular pressure is least at about 70° of elbow flexion¹⁷ and, as pressure of the capsule is one of the most important mediators of pain in these patients, this will therefore be the least painful position for the patient.

Stability of the elbow is examined carefully if pain allows. Valgus stress may displace the fracture fragments and should only be applied very carefully. If the pain is predominantly on the lateral side with valgus stress, one should not persist in testing medial stability.

Palpation of the joint will often reveal the haemarthrosis mentioned above. This will especially be the case if the capsule remains intact. If there is a large capsular defect, however, and the haemarthrosis dissipates into the soft tissues, a boggy, more diffuse swelling of the elbow will be present. Pain with palpation of the radial head may not be immediately obvious, but will become clear when the elbow is pronated and supinated while mild pressure is applied to the radial head. Associated lesions should be suspected if palpation of other surrounding structures is painful. Palpation of the forearm may reveal pain in the interosseous membrane. The wrist always needs to be formally examined and radiographs of the wrist are taken if there is any suspicion of a fracture or DRUJ injury.

Plain anteroposterior (AP) and lateral radiographs of the elbow are taken and analysed. Specific radial head views may reveal a fracture that is not visible in other views. Care should be taken not to miss subtle associated injuries, such as a small coronoid fragment. Non-displaced radial head fractures may not be visible on plain radiographs. In this case a posterior fat pad sign (Fig. 21.3) is almost pathognomonic for a radial head fracture and treatment should be directed as such. There should be a low threshold for a CT scan of the elbow in cases of minimally displaced fractures, as this may show associated lesions (Fig. 21.4) and displacement of the fracture can be more easily assessed (Fig. 21.5). Fragments can be displaced more than one would suspect from the plain radiographs only. Both associated lesions and a correct measurement of the displacement may change the course of treatment.

Figure 21.3 Plain radiographs of the elbow do not show a clear fracture. There is, however, a posterior fat pad sign (F) on the lateral radiographs, which is indicative of a haemarthrosis. Combined with the clinical examination, the diagnosis of a non-displaced radial head fracture is made.

Figure 21.4 CT scans of the elbow can be made to better judge the displacement of fractures and to diagnose the presence of associated injuries. Example of a CT image showing a displaced radial head fracture.

Figure 21.5 Coronoid process fracture associated with a non-displaced radial head fracture.

Other investigations could include MRI or ultrasound if ligamentous lesions are suspected. Ultrasound of the forearm can reveal an interosseous membrane lesion. Radiographs of the wrist are almost standard in the investigation of radial head fractures, and should certainly be ordered if there are any clinical signs of wrist pathology. Scaphoid fractures can be associated with radial head fractures and specific scaphoid radiographs should also be ordered if there is a clinical suspicion of a fracture.

The treatment of radial head fractures depends on the type of fracture as well as any associated lesions. The latter are treated on their own merit and will be discussed in detail in other chapters.

Non- or minimally displaced Mayo type I fractures are treated conservatively. Pain relief is, however, important to allow early mobilization of the elbow. The best way to decrease the pain acutely is to aspirate the joint. A local anaesthetic can then be injected into the joint if further pain relief is necessary. It is extremely important that the aspiration is undertaken using a strict aseptic technique. The elbow is mobilized immediately, although a sling can be used temporarily for comfort. If immobilization in a plaster cast is used, it should not be prolonged longer than a few days. Oral analgesia is prescribed to the patient and ice can be applied regularly to diminish swelling and to decrease pain. Non-steroidal antiinflammatory drugs can be prescribed if there are no medical contraindications to its use. The patient is instructed to mobilize the elbow in all directions as pain allows. As residual haemarthrosis dissolves, range of motion will increase.

The patient is not allowed to grip forcefully with the ipsilateral hand or to support themselves with their hand for up to 6 weeks after the injury, as secondary displacement could potentially occur. Repeat radiographs are taken at 2 weeks to assess for displacement. At this point the patient is instructed to further mobilize the arm and discontinue the sling completely. A physiotherapy programme can be initiated if necessary, with passive and active range of motion exercises. Strengthening exercises are usually not indicated.

Unless there is a contraindication for surgery or an acute infection of the elbow, displaced Mayo type II fractures are treated by surgical reduction and fixation of the fracture fragments. Small displaced fragments that do not block motion may be treated conservatively depending on the patient’s needs. However, it has been shown that it is probably advantageous to fix smaller radial head fragments as well.¹⁹

Reduction and fixation of radial head fractures may be undertaken arthroscopically²⁰ or open.²¹ Fixation can be undertaken using different types of screws. It is important to place these screws in the non-articulating portion of the radial head or to use headless screws that can be buried beneath the articulating cartilage. Bioabsorbable pins or screws have also been shown to be effective, seemingly without any adverse effects from the material.²² The fixation of radial neck fractures is evolving as well. Plate fixation has been the treatment of choice for a long time and still is for specific types of radial neck fracture. Low-profile plates have been developed in response to the bulk of earlier, bigger plates which resulted in problems, especially rotation of the forearm. Other low-profile techniques have been described and more recently screw fixation of the radial head onto the neck has been shown to achieve reasonably stable fixation.²³

Whichever material or fixation type is used, however, in displaced radial head and neck fractures, the aim of the treatment is to achieve a near-anatomical reduction with stability of the fragments and elbow, so that early mobilization can be initiated.

Mayo type III fractures are, by definition, comminuted fractures and treatment is somewhat controversial. Reduction and fixation of type III proximal radius fractures can be technically challenging and results reported in clinical follow-up studies show that it may not be in the best interest of the patient to do everything possible to fix these fractures, if there are more than three fragments.^24,25 As a consequence, it may be necessary to alter treatment during surgery from reduction and internal fixation to resection or radial head replacement.

Resection has been the gold standard for the treatment of comminuted radial head and neck fractures in the past. In recent years, however, prosthetic replacement of the radial head has become the treatment of choice for many surgeons treating these injuries. This is for several reasons. First, some of the complications following radial head resection are extremely difficult to treat. A small degree of cubitus valgus and up to 2 mm of proximal migration of the radius are usually not symptomatic. It has been postulated that this is a physiological phenomenon.²⁶ However, symptomatic proximal migration of the radius relative to the ulna following resection of the radial head has been shown to occur in up to 25% of patients, when the radial head was excised without consideration of associated lesions.²⁷ Patients mainly complain of wrist pain and decreased range of motion. Ulnar shortening and Sauvé–Kapandji procedures can be indicated for wrist pain, but some patients may even end up needing a so-called ‘one-boned forearm’ procedure, with reduced function. Results of delayed placement of a radial head prosthesis following resection of the radial head in these patients are generally poor when compared to acute placement. Delayed placement has also resulted in erosion of the capitellum as a result of an inability to reduce the forearm once proximal migration has occurred.²⁸ The second reason why the use of radial head prostheses to treat comminuted fractures of the proximal radius has become increasingly more common is the evolution of prosthetic design. Increased knowledge of the anatomy of the proximal radius and biomechanics of the elbow has led to changes in radial head prosthesis design. The first radial head prosthesis widely used was the silastic radial head developed by Swanson. Because of material problems such as foreign body synovitis and breakage of the prostheses, this type of prosthesis is no longer used. Secondary to these complications, and perhaps more importantly, the biomechanical and material properties of the silicone radial head prostheses were not able to reproduce the normal stability and kinematics of the native radial head. Metal radial head prostheses, on the other hand, are better equipped to improve stability. Metal radial head prostheses also replicate the normal biomechanics more closely as they are able to withstand higher loads²⁹ than silicone prostheses. In addition, these modern radial head prostheses are also designed to replicate the normal anatomy of the elbow. Most designs are modular and allow for the shaft and the head to be placed in several different combinations. In this way the size of the radial head can be adapted to the actual size of the fractured head, while the stem fits separately into the radial intramedullary canal. The head and stem components are usually assembled outside the patient, and the prosthesis is then placed in the injured elbow. Another feature of the native radial head that has been addressed in prosthetic design is the shape of the radial head itself. As was mentioned earlier, the radial head is not round but has an elliptical shape.⁷ Recently this shape has also been incorporated into radial head prosthetic design in order to improve load transfer and radiocapitellar alignment.

One of the potential problems that has not completely been solved, however, by these more rigid, anatomically correct radial head prostheses is malalignment of the radius onto the capitellum. If, despite optimal placement of the rigid radial head prosthesis, a malalignment persists between the proximal radius and the capitellum, this may lead to subluxation and even dislocation of the radiohumeral joint. Bipolar types of radial head prosthesis have been designed to allow for this malalignment. In these designs there is an articulation between the fixed stem and the radial head component of the prosthesis. The radial head component is therefore able to adjust to the capitellum and improve radiocapitellar contact.

While bipolar radial head prostheses have been shown not to be as stable as a rigid prosthetic design in optimal laboratory circumstances, they are more stable than radial head resection in the MCL-deficient elbow³⁰ and will intuitively be better than a subluxed rigid prosthesis.

Surgical techniques and rehabilitation

Aspiration of the joint is the treatment of choice for Mayo type I radial head fractures. The patient can be either seated or lying supine. The latter may be preferable as the pain may result in a vaso-vagal reaction in some patients. The soft spot is determined by palpation of the radial head, the tip of the olecranon and the lateral epicondyle. They form a triangle when the elbow is flexed and the soft spot is located at the centre of this triangle. In the case of a haemarthrosis there is an obvious swelling in this area and aspiration is easy. The skin surrounding the soft spot is prepared with an antiseptic and sterile gloves should be worn. A 16 or 18 gauge needle and a 20 mL syringe are used to aspirate the joint. The needle punctures the skin perpendicular to the soft spot and blood is evacuated from the elbow. Some of the blood is applied to a gauze swab and examined for fatty droplets, indicative of a fracture. It is not necessary to completely drain the elbow as it may be very uncomfortable to the patient when an attempt is made to draw the final few millilitres. Usually about 15–20 mL can be evacuated comfortably. The needle can then be left in place to inject a small amount of local anaesthetic into the joint. Unless there is a medical contraindication or infection in the elbow, displaced proximal radial fractures are treated surgically. Surgery is preferably performed within 1 week of injury, as longer delays are associated with poorer results.

Different techniques are available for the surgical treatment of proximal radial fractures. Positioning of the patient depends on the technique used, as prone, supine and lateral decubitus are all possible.

Arthroscopic reduction and fixation can be used for selected acute fractures. The ideal fracture has one large fragment, with minimal impaction of the fracture onto the radial neck. If an arthroscopic approach is used, the patient is positioned to the side of the operating table in either the lateral decubitus or prone position. A tourniquet is used and the arm is placed over a roll. The humerus can be fixed onto the roll or left free. The forearm and elbow are free to move in any direction. The arm, hand and fingers are prepped and draped in a standard fashion. We use a stockinet to cover the hand and part of the forearm. The ulnar nerve is palpated prior to surgery. Anatomical landmarks can be drawn onto the patient’s arm. Usually the radial head, olecranon, lateral and medial epicondyle and ulnar nerve are marked. A needle is placed into the joint through the soft spot. The capsule is distended using approximately 25 mL of saline. A standard medial viewing portal is made and a trochar is used to pierce the capsule. A lateral portal is then made using a needle to determine the optimal position. Care should be taken not to place the lateral portal too anteriorly as this may put the radial nerve at risk. A small cannula can then be placed in the lateral portal, as instruments need to be passed through the lateral portal in order to reduce and fix the head. The elbow is irrigated thoroughly and the fracture site visualized and debrided as one would do using an open technique. The reduction is performed under direct vision through the arthroscope. Thereafter, the fragments are temporarily fixed using a K-wire or small guide pin, depending on the system used. The correct length of screw is then measured. This is an important step as the screws should not pierce the adjacent cortex. This is slightly more difficult than in the open technique, as is it not easy to rotate the forearm without bending the pin, once the pin is placed. Fluoroscopy may also help in determining the correct length of screw. Typically the length will be between 18 and 24 mm depending on the gender and size of the patient. If necessary, the pin can then be drilled through the head into the lesser sigmoid notch for stability. In most cannulated systems, the screw hole is pre-drilled over the pin and the screw is inserted. A second, percutaneous pin can then be drilled through the fragment for added stability. Another option is to use a more proximal radial portal, through which an instrument can be placed to stabilize the fragment. The pins are then removed and the stability of the fixation is assessed. Placement of the screws can be checked directly, by rotating the forearm once the K-wire or pin is removed. Once stable fixation is achieved, the posterior compartment can be viewed. The scope is directed into the lateral gutter and the posterior part of the radial head and especially the capitellum are inspected for any loose osteochondral lesions.

For the open approach, the patient is usually positioned supine on the operating table, although lateral decubitus is an option. Depending on the approach used, the arm can be placed on a hand table or positioned over the chest of the patient when a posterior incision is used. A posterior incision is used if medial structures, such as a coronoid fracture, need to be surgically addressed. The treatment of associated fractures will not be discussed further in this chapter. The incision is centred over the olecranon but may be slightly lateral to this to avoid subsequent wound problems. Full-thickness flaps are developed laterally and medially, allowing standard approaches to be performed. The lateral Kocher approach will be described later. A lateral or posterolateral incision can be used for simple radial head fractures. In this case the arm of the patient can again be supported on a hand table, but also pulled over the chest of the patient. The incision is centred over the lateral epicondyle. Kocher’s interval is indicated by a ‘fatty streak’ laterally (Fig. 21.6). The interval is entered by incising the fascia between the extensor carpi ulnaris and the anconeus muscle. Alternatively, Kaplan’s interval between the extensor carpi radialis and extensor digitorum communis may be used. The annular ligament is then divided longitudinally (Fig. 21.7). Stay sutures can be placed on both sides of the incision, so that the ligament can be closed anatomically at the end of the procedure. Care should be taken not to injure the LCL complex, including the LUCL. Sometimes, however, the LCL complex is torn, together with the common extensor tendon, and the approach to the radial head is then directed by the tear in the tendon and ligaments (Fig. 21.8). The proximal radius fracture is then visualized and the fracture site debrided.

Figure 21.6 Once a lateral skin incision is made, Kocher’s interval is easily found by following the ‘yellow brick road’, a fatty streak underneath the fascia between the extensor carpi ulnaris and anconeus muscles.

Figure 21.7 The annular ligament is divided sharply in a longitudinal fashion. This provides an excellent view onto the proximal radius, without disturbing the LUCL, which is posterior to the incision in the annular ligament.

Figure 21.8 Rupture of the common extensor tendon and LCL complex guided the approach to the radial head in this patient. (A) Posterolateral rotatory instability is clearly demonstrated by simple rotation of the forearm. (B) The radial head was replaced by a prosthesis. (C) The LCL complex and common extensor tendon were securely repaired. The elbow was completely stable following the radial head replacement and repair of soft tissues. Ca, capitellum; Co, coronoid process; R, radial neck.

Screws, pins and plates should be available in the operating room, as well as a metal radial head prosthesis.

Once the fracture site is debrided, any haematoma is evacuated and the area is irrigated thoroughly. Thereafter, radial head and neck fixation is carried out using the necessary hardware. Pins or various types of screws can be used to fix radial head fractures. Radial neck fractures are traditionally fixed using plates and screws (Fig. 21.9), but stable fixation can be achieved by screws alone.²³ In complex radial head and neck fractures it may be useful to reconstruct the head on the operating table before fixing it onto the neck. Rotation of the forearm after fixation of the radial head and neck may result in impingement of the hardware into the lesser sigmoid notch. To avoid this complication any hardware should be placed under the surface of the articular cartilage and preferably in the non-articulating portion of the radial head.³¹

Figure 21.9 Plate fixation of a radial head and neck fracture. In this case a plate was used to fix the fracture, but low-profile fixation using screws may also be possible for radial neck fractures.

If it is not possible to reduce and fix the fragments, or if there are more than three fragments,²⁴ the radial head is excised. Stability of the elbow and forearm is then examined. Valgus instability, PLRI or longitudinal instability are clear indications for a radial head replacement (Fig. 21.8). Fluoroscopy can be used to check for medial gapping of the ulnohumeral joint, while a valgus load is applied to the elbow. Longitudinal stability can be tested using the ‘radial pull test’, described by Smith et al.³² Proximal translation of the radius of more than 3 mm is indicative of an interosseous membrane lesion. Fluoroscopic imaging of the wrist during stress testing may also reveal instability, indicative of acute longitudinal radio-ulnar dissociation. If the elbow and forearm remain stable following resection of the radial head fracture fragments, radial head resection alone should be considered as the definitive treatment. The majority of type III radial head fractures are, however, complex injuries with associated bone or ligamentous lesions, and in these patients radial head replacement will be the treatment of choice.

There are many different types of radial head prostheses and the technique of insertion depends on the prosthesis used. In general, the goal of any radial head replacement is to re-create the anatomy and biomechanical properties of the proximal radius as closely as possible, so that good clinical function of the elbow is obtained and maintained in the long term.

The native radial head is elliptical in shape,⁷ which influences the load transfer through the proximal radius.^8,33 If an elliptical prosthesis is to be used, care should be taken to orient the radial head properly.

A second important technical issue in replacing the radial head with a prosthesis is the correct length of the reconstruction.^21,34,35 Overstuffing has been linked to loss of motion^36,37 and capitellar wear.²⁸ Conversely, if the reconstructed radius is too short it does not provide stability to the elbow. Anatomical guidelines are now available to make sure the length is reconstructed properly.

After the radial head is fixed, resected or replaced, the lateral soft tissues are repaired. It may be necessary to repair the LCL complex or reattach this to the lateral epicondyle with an anchor. The annular ligament is sutured. Sometimes, however, it is not possible to completely close the annular ligament owing to the bulk of the hardware used. The annular ligament can be lengthened and approximated in these cases to provide stability to the proximal radio-ulnar joint. Kocher’s interval is closed tightly, and the common extensor tendon is reattached if necessary (Fig. 21.8). Stability is tested again following closure of the lateral soft tissues and a decision is made to repair the MCL if necessary. Unless there is clear instability, the MCL is usually not repaired or reconstructed, as this may lead to excessive stiffness of the elbow. Depending on the strength of fixation of other associated injuries such as a distal humerus fracture, or if repair of the soft tissues needs to be protected temporarily, a hinged external fixator or brace can be used to protect the repair/fixation. A dynamic external fixator is only rarely necessary following the surgical treatment of simple radial head fractures.

After surgery a protective and compressive bandage is placed on the elbow and forearm to reduce swelling. Range-of-motion exercises are typically started on the first day after surgery, with the elbow in a hinged brace if necessary. Alternatively, the arm can be placed in a long arm cast for 3–4 days. This provides some pain relief and will probably not influence the final outcome in terms of stiffness.³⁹

Active flexion and extension exercises are started as soon as possible, with the help of a physiotherapist. Pronation and supination are permitted as pain allows. The patient is instructed not to support their body weight with their hand, nor to lift any weights or do any forceful activities with their hand. The patient returns to the outpatient clinic 2 weeks after surgery. Stitches are removed and radiographs are taken. Passive stretching can be initiated from this point onwards but strengthening and support are still not allowed. A second postoperative follow-up appointment is made at 6 weeks. Radiographs are taken again and the positions of the fragments and hardware evaluated. Radiographs are also checked for signs of consolidation of the fracture, although this is usually very hard to determine. Depending on the clinic and evolution of associated lesions, the patient can start more strenuous exercises and use the elbow ‘as normal’ during activities of daily living. Passive stretching with dynamic braces can be indicated if the range of motion is not progressing as expected.

Outcome including literature review

Good long-term results can be expected from conservative treatment of Mayo type I radial head fractures.⁴⁰ Range of motion exercises should be initiated as soon as pain allows. A delay of a few days does not negatively influence early outcome.³⁹ Although good results have been published from conservative treatment of selected fractures with moderate displacement (2–5 mm),⁴¹ surgical reduction and internal fixation seem to yield better results. Residual pain was found in 42% of conservatively treated patients, compared to 32% of patients that were treated surgically. Good to excellent Mayo Elbow Performance Scores were found in 52% and 88% respectively.⁴²

Results of open reduction and internal fixation (ORIF) of radial head fractures depend significantly on associated lesions and type of fracture.^24,25 Several reports have shown that the results of osteosynthesis for comminuted fractures can be disappointing.^24,25 Up to 93% good to excellent results are reported with Mayo type II fractures.²⁴ However, direct comparison of radial head replacement to ORIF in type III fractures reveals that the results for replacement are superior, with 12.5% good or excellent results in the ORIF group, compared to 93% in the radial head replacement group.⁴³

Radial head resection is somewhat controversial as results of resection are typically good in the long term.^44,45 However, in the short term and in small patient cohorts, resection has been shown to be inferior to osteosynthesis^46,47 or metal radial head replacement.⁴⁸ Furthermore, and as was stated previously, some of the complications of radial head resection may be disastrous for the patient.²⁷

The results of radial head prosthesis have been summarized by Morrey. In general, about 80% of patients are satisfied after a mean follow-up of nearly 4 years. This increases to a little over 90% if only acute fractures are considered.⁴⁹ Although good results have been reported in chronic conditions,^50,51 in general satisfactory results are only found in approximately 50% for delayed radial head replacement.⁴⁹

Complications of treatment

The most common complication of treatment of radial head fractures is degeneration of the radiohumeral articulation. This has been reported following conservative treatment, osteosynthesis, radial head resection and prosthetic replacement.^40,51–53 Another common complication of the treatment of a type I radial head fracture is loss of extension of the elbow.¹⁵ This is more commonly seen following long periods of immobilization but can still occur following immediate mobilization. Non-displaced fractures are less prone to an extension deficit than mildly displaced fractures.^40,53 Range of motion can increase for a long time after injury and conservative treatment should be continued as long as there is improvement. Physiotherapy, including both passive stretches and active range-of-motion exercises, are indicated early in patients in whom range of motion does not increase soon after the fracture. Static and dynamic bracing may also increase range of motion. Other potential complications of Mayo type I radial head fractures include non-union and secondary displacement of the fracture. Non-union is rare and usually asymptomatic.⁵⁴ Secondary displacement is also very infrequent. If it does occur, it will happen early and should be anticipated. Radiographs should therefore be repeated at 2 weeks following the fracture. If repeat radiographs show secondary displacement, a CT scan should be ordered to show the extent of displacement. A type I radial head fracture could become a type II and should then be treated accordingly.

Septic arthritis of the elbow is a potentially devastating complication. It can obviously occur after surgical fixation of radial head fractures or radial head replacement with a prosthesis, but it could also occur following aspiration and intra-articular infiltration of the joint in seemingly harmless type I fractures. Prevention is key and aspiration and infiltration should therefore be undertaken under strict asepsis. If a deep infection does occur, this should be treated promptly with arthroscopic or open lavage and debridement, followed by a prolonged period of intravenous antibiotics. If necessary, the debridement should be repeated until clinical and haematological signs of infection clear. If the infection occurs early following an operation the hardware may be retained, but removal of all hardware is indicated in chronic or delayed infections.

As in conservative treatment of radial head fractures, some limitation of range of motion can be expected following ORIF of the radial head, with reported motion from an average of 10° extension to 130° of flexion in type II radial head fractures.²⁴ Loss of reduction and failure of hardware can also complicate the treatment of type II radial head fractures. Especially when plate fixation is performed, hardware removal is sometimes indicated in an attempt to improve range of motion.

Limited data are available on complications of radial head replacement. Range of motion following radial head replacement is often decreased^37,51,52 and has been shown to improve following removal of the prosthesis.^37,52 Radiolucent lines are frequently seen with intentionally loose radial head prostheses but this does not seem to correlate with functional results or pain.⁵⁵ Temporary nerve dysfunction has been described following radial head replacement surgery of the ulnar, radial and posterior interosseous nerves.⁵² This usually resolves spontaneously. The modularity of modern radial head replacements has introduced a unique complication, in that they can dissociate.⁵⁶ Heterotopic ossification⁵⁷ and capitellar erosion²⁸ have also been described as case reports or as part of a larger series. Pain, stiffness and instability are the most common clinical causes of failure of a radial head replacement, necessitating removal or revision of the implant. Loosening, over-lengthening and radial head subluxation are common radiographic findings in these patients.⁵⁸

Conclusions/personal view

Radial head fractures are extremely common. Luckily, the vast majority of fractures are Mayo type I fractures without displacement or associated injuries. These fractures should be treated functionally. From personal experience, I know that a non-displaced radial head fracture in itself is not very painful. Immediately following the trauma there is merely a deep ache. Range of motion is full at this time. This actually gave me a false sense of relief, although this did not last long. Within a few minutes a haemarthrosis occurs and the increased pressure within the capsule causes pain. Range of motion decreases due to pain from the haemarthrosis but mechanically the haemarthrosis also limits movement. A type I fracture should be treated by pain relief and this is best achieved by aspiration of the joint with the appropriate injection of a small amount of local anaesthetic, but this is not always necessary.

Once the diagnosis is made, treatment depends on both the type of fracture and associated lesions. An elbow dislocation, for example, should be reduced immediately. It is our preference to do this in the operating room so the patient can be sedated and the elbow examined using fluoroscopy immediately following the reduction. A displaced fracture–dislocation is almost always an indication for surgery.

Unless there is no displacement whatsoever we routinely obtain CT scans of the elbow, with three-dimensional rendering of the images (Fig. 21.10). The decision to operate is based on these X-ray and CT findings, taking the patient’s age and comorbidity into consideration. The size of the fragment is almost of secondary importance, as small fragments can have large biomechanical consequences. Arthroscopic reduction is performed in acute fractures (<1 week), where there is one large fragment with little or no impaction. The majority of displaced fractures are treated by open reduction. The radial head is typically fixed with one or two headless cannulated screws. Radial neck fractures are fixed using screws with the low-profile method, or with a specifically designed radial head plate and screws.

Figure 21.10 Three-dimensional rendering of the proximal radius, showing a displaced radial head fracture.

Personally, for comminuted proximal radial fractures, I rarely perform a radial head resection without replacing the radial head with a metal prosthesis. I use an anatomical modular system, as I believe that long-term results and function must improve, if the normal anatomy of the elbow is reproduced more closely. Bipolar designs are indicated in subacute and chronic cases, as alignment of the radius relative to the capitellum may be compromised, and it is extremely difficult to judge how the radiocapitellar joint will behave after surgery.

Although the results of treatment for radial head fractures may be dictated by the associated lesions, in my opinion the most important goal of treatment of all types of fractures is to be able to initiate range-of-motion rehabilitation as soon as possible as this seems to improve results dramatically.