(v) The management of intracapsular fracture of the femoral neck

Epidemiology 

Reduced bone mineral density is perhaps the strongest predictor of future hip fracture and in the vast majority of cases this is due to osteoporosis. A reduction of bone mass at the hip of one standard deviation doubles the risk of hip fracture. After peak bone mass is reached, there is a gradual reduction in bone mineral density with increasing age. Women have a lower peak bone mass than men and the rate of bone loss increases after the menopause. Furthermore, women live longer than men and so are at increased risk of osteoporosis and have a higher lifetime risk of hip fracture compared to men.

There are racial variations in the incidence of hip fractures, which are most common in the white populations of Europe and North America. Other specific risk factors for hip fracture include a Body Mass Index (BMI) of less than 18.5, low recreational activity, a previous osteoporotic fracture after the age of 50 years, smoking, a history of maternal hip fracture and corticosteroid treatment. Having sustained a hip fracture there is a 10% risk of sustaining a further hip fracture in women and a 5% risk in men. An increase in length of the femoral neck has been shown to be associated with an increased risk that is independent of bone mineral density and age. In younger patients other predisposing factors should be considered, such as a history of alcohol abuse, steroid use, rheumatoid arthritis, renal failure and endocrine disease, which are associated with a reduced bone mineral density.

The risk of falls increases with age, as many of the specific risk factors are age related. These include poor visual acuity, neurological disease, medications (hypnotics, sedatives, anti-hypertensives), muscle weakness, gait abnormality and poor balance.1 The other group of patients who sustain femoral neck fractures are young patients who have sustained high energy trauma. In such instances the patient may have sustained multiple injuries and clearly any life threatening injuries should be treated first.

Mechanism of injury 

A fall onto the side results in force transmission through the greater trochanter to the femoral neck. An alternative mechanism is external rotation of the leg resulting in tension of the anterior femoral neck and the anterior capsuloligamentous structures. This mechanism may also account for the posterior neck comminution frequently encountered in these fractures.

Preoperative traction 

Some of the proposed benefits of preoperative traction include relief of pain, reduced soft tissue injury, maintenance of fracture reduction and ease of operative reduction. However, these perceived benefits have not been supported by randomized studies comparing traction with no traction. Preoperative traction for intracapsular fractures is therefore not recommended.

Imaging and other diagnostic studies 

The majority of hip factures can be diagnosed on good quality anterioposterior and lateral radiographs. In approximately 2% of cases a nondisplaced femoral neck fracture may not be visible on plain radiographs, so alternative imaging techniques may be required. A bone scan was formerly the imaging of choice to identify occult femoral neck fractures, but was associated with a risk of a false negative result if carried out within the first 72 h of injury. Both CT and MR scanning are more accurate imaging modalities. MRI is the current additional imaging of choice, since it is more accurate than a bone scan soon after fracture, and can identify pathology in adjacent soft tissues in the absence of a fracture. If osteoporosis is suspected then a dual energy X-ray absorbiometry (DEXA) scan may be appropriate, but in patients over 70 osteoporosis is presumed, so prophylactic treatment can be commenced without DEXA.

Garden classification 

The Garden classification is based on the degree of fracture displacement.2 Four fracture subgroups are described. Type 1 and 2 fractures are undisplaced, whilst type 3 and 4 fractures are displaced. Type 1 fractures are impacted in valgus, with a fracture of the lateral cortex. Type 2 fractures are complete fractures of the femoral neck with no displacement. Type 3 fractures are displaced, but there remains some contact between the femoral head and neck, which is in varus, such that there is angulation between the trabecular lines of the acetabulum and the femoral head. In type 4 fractures there is complete displacement of the fracture and the head is free to return to a neutral position in the acetabulum so the trabecular lines in the femoral head realign with those in the acetabulum.

Although the Garden classification is widely used, a criticism has been the poor levels of inter- and intraobserver agreement between the four subgroups. Surgeons appear to be better at distinguishing between undisplaced (Garden types 1 and 2) and displaced fractures (Garden types 3 and 4), where levels of agreement are better than those between individual subgroups.

Pauwels classification 

The Pauwels classification grades the orientation of the fracture line with reference to the horizontal on an anterioposterior radiograph of the hip, and provides a measure of how vertically the fracture is orientated. In type 1 fractures the angle is 30° or less. In type 2 fractures the angle is 30–50° and in type 3 fractures it is greater than 50°.

More vertical fracture lines (esp. Pauwels type 3) have greater shear forces and are a more unstable pattern. This classification has not been shown to be particularly useful for the majority of hip fractures that occur in elderly patients. In these patients the fracture plane is seldom vertical. However the classification may be of more use in younger patients where vertical fracture lines are more common. There is some evidence that use of a fixed angle device in the Pauwels type 3 fracture is associated with a lower risk of fixation failure and nonunion than the cannulated screw fixation which is generally used to fix subcapital hip fractures.

AO/OTA classification 

The AO/OTA comprehensive classification of long bone fractures classifies fractures based on the bone involved, location of the fracture and morphology of the fracture. Fractures of the femoral neck are designated 31B. 31B1 are minimally displaced fractures of the femoral neck. The B2 group defines transcervical fractures and the B3 group comprises displaced subcapital fractures. Each group is then divided into subgroups to further define fracture morphology. This classification, although comprehensive is complex and therefore not suited to routine clinical use. Surgeons are able to divide fractures into the major groups with consistency, but levels of agreement between the subgroups are poor. The key factor in determining outcome is whether the fracture is displaced or nondisplaced, and surgeons are generally good at distinguishing between these simple groups. Available classification systems have not been shown to be better than such simple categorizations. Most clinical research is based on the simple division of fractures into undisplaced and displaced type without specifying subdivisions of more detailed classifications systems.

Skeletal anatomy 

The relationship between the femoral neck and the femoral shaft is characterized by the neck–shaft angle and femoral anteversion. In the normal hip the neck–shaft angle is between 130 and 135°. Femoral neck anteversion is defined as the angle between the neck and the transcondylar axis and is between 15 and 25°. Other measurements include the hip axis length and the femoral neck width. The hip axis length is the distance between the lateral aspect of the trochantric region and the inner table of the pelvis (Figure 1).

View Large Image Download to PowerPoint

Figure 1 Hip axis length, distance from lateral aspect of the greater trochanter to the inner table of the pelvis depicted by blue line.

An increased risk of hip fracture is associated with higher neck widths, longer hip axis lengths and lower neck–shaft angles.3 In Caucasian females, average hip axis lengths have been shown to be longer than in Asian and black populations, which may contribute to an increased susceptibility to femoral neck fractures in this population.4

The calcar femorale reinforces the posterioinferior portion of the femoral neck. It is a dense vertical plate of bone that passes from the posteriomedial aspect of the proximal femoral shaft under the lesser trochanter and extends to the greater trochanter.5

Blood supply of the femoral head 

The medial and lateral circumflex arteries and the artery of the ligamentum teres contribute to the blood supply of the femoral head. The medial and lateral circumflex femoral arteries originate from the profunda femoris artery in 79% of cases and in 20% of cases one or more of these vessels arise from the femoral artery. In 1%, both arise from the femoral artery.6 The medial femoral circumflex artery passes behind the femoral neck and the lateral femoral circumflex artery passes anterior to the femoral neck. Together they form an extracapsular anastomotic ring at the base of the femoral neck. From this ring arise the retinacular vessels, of which there are four groups (superior, inferior, anterior and posterior). They pierce the hip capsule and pass onto the neck under the synovium and ascend on the femoral neck in a medial direction. At the junction of the articular surface of the femoral head and the neck they form a second ring, the subsynovial intra-articular ring, from which arise vessels that penetrate the femoral head near the junction of the articular cartilage of the femoral head and femoral neck.

The superior group of retinacular vessels, which are derived from the deep branch of the medial circumflex femoral artery, is the most important. They supply the majority of the blood supply to the weight bearing area of the femoral head. A contribution from the lateral femoral circumflex artery is less important. The artery of the ligamentum teres provides a small and variable amount of blood supply to the femoral head. It is derived from the obturator or the medial femoral circumflex artery.

Displaced fractures of the femoral neck can result in damage to the retinacular vessels, which can disrupt blood supply to the femoral head. If the important superior group of retinacular vessels is injured the risk of avascular necrosis is greatly increased. Although the supply to the femoral head through the ligamentum teres may be maintained following fracture, its contribution is often not sufficient to maintain viability of the femoral head.

Sensory supply 

Sensation to the hip is supplied from the obturator, femoral, sciatic and superior gluteal nerves. The obturator nerve supplies the anteriomedial portion of the joint. The femoral nerve supplies the anterior capsule and the posterior joint is supplied by the sciatic nerve. The superior gluteal nerve supplies a portion of the posteriolateral capsule. This is of relevance when considering nerve blockade for analgesia. Femoral nerve block alone will provide incomplete analgesia yet is often effective, as it has an additional effect by reducing muscle spasm.

Common surgical approaches 

The direct lateral (Hardinge) approach is a popular approach for hip arthroplasty. It is associated with a low dislocation rate when used for hip replacement. This is an advantage, as the risk of dislocation in the context of femoral neck fracture is higher than that when hip replacement is performed for arthritis. The gluteus medius tendon is divided close to its insertion on the greater trochanter and elevated as an anterior flap with gluteus minimus and vastus lateralis to expose the hip capsule. Proximally the muscle is split in line with its fibres for a short distance. The inferior branch of the superior gluteal nerve is at risk if the split is carried too far proximally. A safe distance of 3–5 cm has been reported in the literature but the position of the nerve is highly variable.7

The posterior approach provides good exposure of the acetabulum. Piriformis, obturator internus and the gemelli are divided close to their insertion on the greater trochanter and are reflected posteriorly to protect the sciatic nerve. Incision of the posterior capsule provides access to the hip joint. When used for total hip replacement after fracture it is associated with a higher dislocation rate than direct lateral or anteriolateral approaches.

The anteriolateral approach of Watson Jones is the recommended approach to use when open reduction and internal fixation of femoral neck fractures are being considered. It utilizes the intermuscular plane between the gluteal muscles and tensor fascia lata to expose the anterior capsule.

The anterior approach (Smith Petersen) utilizes the intermuscular plane between tensor fascia lata and sartorius and then at a deeper level, gluteus medius and rectus femoris to gain access to the anterior hip capsule. This approach may be used to reduce and fix fractures of the femoral head but is not routinely used to treat femoral neck fractures.

Treatment of undisplaced femoral neck fractures 

Nonoperative treatment: fifteen percent of intracapsular fractures are undisplaced. Fractures that are undisplaced can be treated nonoperatively, with protected weight bearing using crutches for approximately 6 weeks. This treatment has the advantage of avoiding surgery, but there is significant risk of displacement. Cserhati et al compared the nonoperative treatment of undisplaced femoral neck fractures with fixation and found a 20% rate of displacement in the group treated nonoperatively with no failures in the fixation group.8 Furthermore, fixation was associated with a shorter hospital stay and earlier full weight bearing. Other studies have shown displacement rates of up to 46%. In view of the high complication rate, the indications for nonoperative treatment of undisplaced fractures are limited. Nonoperative treatment, can, however be considered in patients with significant medical comorbidity or those who present late after fracture.

Fixation of undisplaced intracapsular femoral neck fractures: the most common devices used in contemporary orthopaedic practice are cannulated screws or a sliding hip screw device with a short plate (Figure 2, Figure 3). Numerous variations of these implants have been used but meta-analyses of randomized trials9, 10 have failed to show superiority of any particular device. In the elderly population, the degree of osteoporosis and effect on bone quality are probably the factors influencing success after fracture fixation. For the majority of undisplaced femoral neck fractures cannulated screws are now the most popular option and are associated with less blood loss and pain as the surgical approach is less invasive (Figure 2).

View Large Image Download to PowerPoint

Figure 2 a Undisplaced femoral neck fracture treated with three cannulated screws. b Complicated by subtrochanteric fracture 4 weeks after fixation.

View Large Image Download to PowerPoint

Figure 3 a Undisplaced right femoral neck fracture (Pauwels type 2). b Fixation of undisplaced fracture with sliding hip screw and short 2 hole plate.

Fractures which are less vertical are more stable and tend to compress when the hip is loaded, and so can be treated adequately with cannulated hip screws. Fractures which are more vertical are less stable. Shear forces are high and the fracture may not be adequately stabilized by cannulated screws alone. A fixed angle device such as a blade plate may be a better choice for these fractures.

Overall, the outcome after the treatment of undisplaced femoral neck fractures with fixation is good, with most individuals returning to their pre-injury level of function. The main failure modes are nonunion, failure of fixation and avascular necrosis (Table 1). Despite the fact that these fractures are nondispaced there is a rate of nonunion of around 7%. Avascular necrosis occurs in 4–22% of cases but is not always symptomatic and may take up to 2 years before it becomes evident.

Table 1. Results of fixation of undisplaced femoral neck fractures by fixation

	Author	Ref	No of patients	Union	Nonunion	Avascular necrosis
	Phillips and Christie	31	72	94.4%	5.6%	22.5%
	Chen	32	37	95%	5%	11%
	Chiu	33	250	90.4%	6%	7.2%
	Stappaerts	34	33	97%	3%	6%
	Bjorgul	35	225	87%	9%	4%
	Conn	36	375	90%	6.4%	4%
	Stromqvist	37	85	95%	3%	1.5%
	Total		1077	953/1056(90%)	68/1056(6.4%)	59/1024(6%)

Treatment of displaced femoral neck fractures 

The majority of femoral neck fractures are displaced. These fractures are not amenable to nonoperative treatment, but there are a few situations where nonoperative treatment can be considered. Patients with severe medical comorbidities with a very limited life expectancy or demented nonambulatory patients who present late may be treated without surgery. When possible, salvage of the patients' own femoral head is preferable and studies have shown that most surgeons prefer to treat displaced fractures in patients under the age of 60 with fixation. In patients over the age of 80, arthroplasty, which provides predictable results, is most often the treatment of choice. It is in patients between 60 and 80 that there is marked variability in treatment preference. In this age range reduction and fixation, hemiarthroplasty and total hip replacement are used with varying frequency. In the fit elderly patient, salvage of the femoral head with reduction and fixation may seem an attractive option but recent studies have shown that there is a high failure rate due to fixation failure, nonunion and avascular necrosis in this group.

Reduction and fixation: if fixation is the treatment of choice, closed reduction is usually achieved by applying gentle traction and internal rotation to the leg. A common error is to apply excessive traction and internal rotation. If closed reduction cannot be achieved then open reduction can be performed. This is probably only appropriate in younger patients where a good closed reduction cannot be obtained and reduction and fixation are considered essential. The Watson Jones approach which conserves the blood supply to the femoral head is most commonly used.

Comminution of the femoral neck is relatively common and may prevent anatomical reduction. Reduction with valgus deformity is inherently more stable than reduction with varus deformity. The limits of angulation that can be considered acceptable are debatable, however varus angulation of 20° is associated with a 55% risk of failure.11 Posterior angulation of 20° or more should have also been shown to increase risk of failure. If a satisfactory closed reduction cannot be achieved then some form of arthroplasty should be considered. As with undisplaced fractures, the most common choice is cannulated screws, but a sliding hip screw with short plate can be used as an alternative.

The risk of failure after fixation of displaced femoral neck fractures is much higher than for nondisplaced fractures (Table 2). Randomized trials performed over the last decade have compared fixation with various types of arthroplasty and have shown reduced complication rates and better functional scores in patients treated with arthroplasty. Reduction and fixation in these trials have been associated with a requirement for revision surgery in 40–45% of patients.12 The level of the fracture in the femoral neck has not been shown to influence failure rate nor has the type of fixation used. Although avascular necrosis is a well recognized complication of these fractures, revision surgery is most commonly required for early fixation failure or nonunion.

Table 2. Results of reduction and fixation of displaced femoral neck fractures

	Author	Ref	No of patients	Union	Nonunion	Avascular necrosis
	Stappaerts	34	85	68%	32%	33%
	Parker et al	38	568	70%	30%	Not recorded
	Tidermark	39	66	64%	27%	9%
	Stromqvist	37	215	74%	25%	9%
	Bjorgul et al	35	241	77%	20%	4%
	Asnis	12	91	94%	6%	20%
	Total		1266	867/1181(73%)	301/1181(25%)	64/586(11%)

Ancillary measures to improve the success rate of reduction and fixation have been studied. Intracapsular fractures are associated with a haemarthrosis that causes an increase in intracapsular pressure. It has been proposed that aspirating the haemarthrosis or releasing it through a capsulotomy should reduce intracapsular pressure and improve the femoral head blood supply. Open reduction has also been proposed to reduce intracapsular pressure, improve reduction quality and thus improve blood supply to the femoral head. There is no convincing evidence from published studies that such measures ultimately improve the outcome in clinical practice.

There is controversy regarding the optimal time to fixation following femoral neck fracture, with conflicting evidence in the literature. Cellular changes can be seen in the femoral head as early as 6 h following fracture, but osteocyte death occurs slowly and is not complete until 2–3 weeks. Therefore it seems reasonable to consider fixation even in patients who present late. This is supported in a study by Barnes et al who found that fixation up to 7 days following fracture had no effect on the rate of nonunion and avascular necrosis.13 However, Jain et al compared early fixation, before 12 h, and late fixation, after 12 h, and found a 16% rate of osteonecrosis in the delayed group but no osteonecrosis in the early treatment group.12 It seems logical that if reduction and fixation are chosen then in general it is preferable to carry out the procedure as soon as possible after the fracture occurs.

Total hip replacement for failed fixation: although total hip replacement remains an option if fixation fails, there is evidence from one study that the outcome is superior when hip replacement is performed as a primary procedure. McKinley and Robinson compared the outcome of age and sex matched patients who had undergone either primary or secondary total hip replacement following displaced intracapsular fractures.14 Rates of infection, dislocation and loosening were higher in the group that underwent total hip replacement as a secondary procedure.

Unipolar hemiarthroplasty 

First generation prostheses, such as the Austin Moore and Thomson prostheses, remain popular despite the availability of more modern implants (Figure 4). The Austin Moore prosthesis is associated with high rates of subsidence and thigh pain and should be reserved for elderly, infirm patients who are immobile or ambulate minimally.

View Large Image Download to PowerPoint

Figure 4 Unipolar hemiarthroplasty prostheses. From left to right; Austin Moore uncemented prosthesis, Exeter trauma system (ETS), Thompsons hemiarthroplasty.

The large heads of hemiarthroplasty implants confer increased stability and hence dislocation rates are low, typically of the order of 2–3%. Osteoporosis is common in this treatment group and the use of uncemented implants is associated with an increased risk of femoral fracture at the time of insertion of the implant, which occurs in up to 12% of patients. Limited data are available on the long-term outcome of these prostheses because the life expectancy in such patients is relatively low. One study showed a 94 survival at 5 years and an 83% survival at 12 years.15

Bipolar hemiarthroplasty 

It has been proposed that the dual articulation of bipolar designs reduces acetabular erosion. However, in some designs there is evidence that the articulation between the small head and outer shell stops functioning so that in essence it behaves as a unipolar prosthesis The modular nature of some bipolar designs is appealing to surgeons and has an advantage over older designs of unipolar implants by allowing offset and length to be tailored to an individual patient. However modular unipolar implants are now available.

Several comparative studies of unipolar and bipolar designs at short- to mid-term follow-up have found no difference in functional outcome or complication rates between the two groups. There is however some evidence that at long-term follow-up bipolar implants have a lower reoperation rate and a very low rate of acetabular erosion. A study has shown 10 year survivorship of up to 93.6%, which is comparable to that achieved with total joint arthroplasty for arthritis.16

Although the rate of dislocation is similar between bipolar and unipolar prostheses, dislocation of the former is more likely to require open reduction because of the mobile bipolar articulation which makes closed reduction difficult. In addition, there is a risk of dissociation of the bipolar head components during manipulation, which makes open reduction mandatory.

Cemented vs uncemented hemiarthroplasty 

Cement provides immediate stability to the femoral prosthesis, which can be difficult to achieve when older uncemented designs are used in elderly patients with poor bone stock and capacious femoral canals. Comparative studies of older uncemented implants and cemented implants have shown better clinical results with cemented hemiarthroplasty.17 Newer uncemented implants with improved geometry achieve better fit and fill of the proximal femur and studies comparing cemented prosthesis and modern cementless prosthesis have shown similar clinical results.

Pressurization of cement in the femoral canal carries the risk of delivering fat and debris from within the medullary canal into the pulmonary circulation. The risk of sudden death during cemented hemiarthroplasty is approximately 1 in 500. To reduce the risk of embolization of femoral canal debris, lavage and drying of the canal prior to cement pressurization are recommended.

Ipsilateral femoral neck and shaft fractures 

Ipsilateral hip fractures occur in 2–6% of femoral fractures. This injury combination usually occurs in young adults exposed to high energy trauma. If the femoral neck fracture is undisplaced or radiographs are inadequate the femoral neck fracture may go unrecognized. Femoral neck fractures may also be iatrogenic, occurring during placement of an antegrade femoral nail through an incorrect proximal entry point. The options for treating ipsilateral femoral neck and femoral shaft fractures include fracture stabilization with a second generation antegrade nail that allows fixation of the femoral neck through the nail. An alternative is to fix the femoral neck fracture and then use a retrograde nail or plate to stabilize the femoral shaft fracture. For fractures of the mid-shaft or distal femur with a typical subcapital femoral neck fracture a retrograde nail can be used to fix the distal fracture whilst cannulated screws can be used for the neck fracture, treating each fracture with its own implant. Subtrochanteric fractures combined with a femoral neck fracture, particularly if it is a Pauwels type 3, should be treated with a cephalomedullary nail.

Stress fractures 

Stress fractures may be fatigue fractures or insufficiency fractures. Fatigue fractures occur as a result of repetitive cyclical loading of normal bone, whereas insufficiency fractures occur when normal physiological loads are applied to bone that is weakened by a reduced bone density. Fatigue fracture in the femoral neck usually occurs in young adults who participate in regular physical activity. A 3–8% incidence has been reported in military recruits. Female athletes with the triad of eating disorder, amenorrhoea and osteoporosis comprise a group at increased risk of stress fracture.

Insufficiency fractures, on the other hand, are more common in elderly patients with osteoporosis. Stress fractures are often incomplete. Compression fractures involving the inferior cortex are more common and are usually stable, so can be treated initially with 6–8 weeks of protected weight bearing. If the fracture fails to unite then internal fixation can be performed. Fractures that involve the superior aspect of the neck are tension fractures, which have a higher risk of displacement and therefore should be treated with internal fixation (Figure 5).

View Large Image Download to PowerPoint

Figure 5 a Tension type stress fracture of left femoral neck in 89-year-old female. b Displacement of stress fracture.

Femoral neck fractures in Paget's disease 

Fixation of femoral neck fractures in bone affected by Paget's disease is associated with high failure rates. Nonunion rates of up to 75% have been reported. In addition, remodelling of the femoral neck resulting in coxa vara deformity increases the risk of fixation failure.

Hip arthroplasty is therefore the preferred treatment in these circumstances, but the technical aspects of this procedure in patients with Pagetic bone need to be considered. Pagetic bone is sclerotic and may be hypervascular, which can make hip arthroplasty surgery challenging. If the acetabulum is involved in the disease process then total hip replacement is preferable to hemiarthroplasty.

Mortality 

The mortality rate following femoral neck fracture is considerable. The in-hospital mortality rate is approximately 15% and the 1 year mortality rate is up to 30%.21 If there is cognitive impairment then the mortality rate rises to 50% at 1 year. Men have a higher mortality rate than women and comcommitant cardiorespiratory disease increases this risk. Impaired renal function is associated with a doubling of the 1 year mortality. For nondisplaced intracapsular fractures the mortality rate is lower and has been reported to be 15% at 1 year following fixation.

Delay to surgery 

The effect of delay to surgery on outcome is a controversial subject. Dolk found no difference in mortality between patients treated before and after 48 h.22 Similarly Orosz et al found no difference in the mortality or complications when surgery was performed within or after 24 h.23 There is, however, evidence that longer delays are associated with a less favourable outcome. Moran et al and Doruk et al found a significant increase in mortality between patients having immediate surgery and those having surgery after 4 and 5 days respectively.24, 25 Apart from mortality there is some evidence that longer delays are associated with increased morbidity and therefore, if patients are fit for anaesthesia, the aim is to carry out surgery within 48 h of injury. In a recent review of the literature on this topic Khan et al concluded that surgery within 48 h was associated with a reduced hospital stay and a probable reduction in mortality and complication rates.26 Based on current evidence therefore patients should have surgery carried out within this timeframe if possible.

Infection 

Infection of metalwork after fixation is generally less problematic than infection after arthroplasty. Metalwork can be removed after the fracture has healed.

Infection following arthroplasty is a major problem if it occurs. The majority of patients are of advanced age with associated medical comorbidities and are poor candidates for complex revision surgery. Single or two-stage revision arthroplasty is an option in fitter patients. In more frail patients excision arthroplasty may be a safer option but the functional outcome after this procedure is generally very poor. Alternatively, in this group, if the infecting organism is sensitive, suppression with long-term antibiotics may suffice. In some instances a sinus can be deliberately created to facilitate drainage from the lower part of the wound.

Deep venous thrombosis and pulmonary embolism 

Thromboembolic complications can occur following femoral neck fracture. Gillespie et al reported a radiographically proven DVT rate of 45% and a clinical DVT rate of 7%.27 There was a nonfatal pulmonary embolus rate of 8% and fatal pulmonary embolus occurred in 4%. There is evidence in the current literature supporting the use of asprin, heparin, fondaparinux and mechanical thromboprophylaxis to reduce the risk of deep venous thrombosis. The evidence that any of these prevent nonfatal or fatal pulmonary embolism is inconclusive. The best choice based on available evidence for prophylaxis of thromboembolic complications is for use of a low molecular weight heparin or fondaparinux augmented by some form of mechanical prophylaxis (if the latter is tolerated).

Fixation failure and nonunion 

Fixation failure and nonunion are more common in displaced than nondisplaced fractures. For nondisplaced fractures the nonunion rate is approximately 7%. In displaced fractures the rate of revision surgery for nonunion or fixation failure is 30–50% (Tables 1 and 2).

In older patients conversion to arthroplasty is the preferred treatment. Compared to primary hip replacement, hip replacement after failure of fixation is associated with poorer functional results and higher rates of infections and dislocation.

Nonunion in young patients: in young patients, particularly those under 40, salvage of the femoral head is a priority. Options for treatment of nonunion include proximal femoral valgus osteotomy, vascularized bone graft and revision of fixation. Meyers described the use of vascularized quadratus femoris pedicle graft to treat avascular necrosis of the femoral head.28 This pedicle graft is also used to treat nonunion of the femoral neck and has a reported success rate of 95%.

Proximal femoral osteotomy: proximal femoral valgus osteotomy can be used to treat nonunion of the femoral neck. This osteotomy changes the orientation of the fracture making it more horizontal so that compression rather than shear forces is transmitted across the fracture to promote union. This osteotomy is particularly useful if the femoral neck has displaced into varus. Marti et al studied a group of young patients with nonunion treated with this technique.29 Fourteen percent of patients failed and went on to have arthroplasty. Of those remaining there was radiographic evidence of avascular necrosis in 22 (44%), most of which were asymptomatic.

Avascular necrosis 

Reported rates of avascular necrosis vary widely. More recent studies show lower rates of avascular necrosis than historical series did, suggestive of a falling incidence of this complication. Lu-Yao performed a meta-analysis of hip fracture outcomes and reported a 16% rate of avascular necrosis.17 In a more recent meta-analysis by Bhandhari et al30 the rate of avascular necrosis was 6.9%. A possible explanation is that surgeons have become more selective with regards to treatment and are fixing displaced fractures less frequently because of the high failure rates, preferring arthroplasty as an alternative.

If avascular necrosis occurs, this does not necessarily mean that outcome will be poor. Barnes et al reported a series of patients who developed avascular necrosis after femoral neck fracture.13 Twenty-four percent of patients were asymptomatic and only 29% of patients had significant disability. Symptomatic patients with segmental collapse are usually treated with total hip replacement, as there is often associated degenerative change of the acetabulum. Hemiarthroplasty is a reasonable alternative if the acetabular cartilage is normal in elderly patients with lower functional demands.

Prosthesis dislocation 

Dislocation is the most common complication of hip arthroplasty. The dislocation rate for hemiarthroplasty is 2–3%. The dislocation rate for total hip replacement was reported as 6% in a meta-analysis by Bhandari et al, which is lower than the 11% rate reported by Lu-Yao et al performed 9 years previously.17, 30 The risk of dislocation after total hip arthroplasty can be reduced by careful patient selection, use of a direct lateral approach, and other factors such as use of femoral and acetabular components with larger head diameters.

Pain 

Studies have reported better clinical results with modern designs of hemiarthroplasty when compared to the results of fixation. Lu-Yao et al, in a meta-analysis, reported that at 2 years 90% of patients undergoing total hip replacement had no pain compared to 86% of patients undergoing bipolar hemiarthroplasty and 71% of patients undergoing hip fixation.17 The use of the uncemented Austin Moore prosthesis has been associated with a higher incidence of pain when compared to hip fixation and cemented total hip replacement.

Mobility 

There are a number of factors which influence mobility after the treatment of femoral neck fracture. Patients who are elderly, or who have cognitive impairment, have a poorer prognosis. In addition, patients who have impaired mobility prior to fracture may fare less well. Those with undisplaced fractures treated with fixation generally have a good prognosis, with many returning to their preoperative level of function. When fractures are displaced, the type of treatment can influence mobility. Overall, the proportion of patients undergoing fixation or arthroplasty who reach their previous level of mobility is 46%.21 Greater levels of mobility have been reported in patients undergoing arthroplasty for displaced fractures when compared to fixation.

Economic outcome 

Few studies have examined the cost of hip fracture treatment. Although the implant costs for cannulated screws are relatively small, studies have shown that when longer term follow-up is considered, it is a more expensive option than arthroplasty. The main economic burden is incurred not by implant cost but by duration of hospital stay. Reduction and fixation are associated with a similar length of initial stay to arthroplasty but the higher requirement for revision surgery renders this treatment option more expensive over longer term follow-up.

Intracapsular hip fracture treatment algorithm 

View Large Image Download to PowerPoint

Research directions 

Trials to compare performance of newer uncemented hemiarthroplasty prosthesis with cemented designs.

Further study to compare bipolar and unipolar hemiarthroplasty prosthesis.