Introduction

Intertrochanteric fractures are common, and data show that nearly half of the fractures around the hip are intertrochanteric. Intramedullary nail with a cephalomedullary screw or sliding hip screw–plate construct is the standard surgical treatment options chosen by most surgeons, and wide variability on fixation methods and outcomes has been observed. Fracture pattern, bone quality, fixation techniques and few other factors are important to achieve ideal results. In this article, we review various fixation methods in detail.

Dynamic hip screw

This is the most commonly used implant and extensively studied implant for extra-capsular hip fractures (Table 1) [120]. It is an extra-medullary fixation device that works on the concept of stabilizing the fracture but allowing controlled collapse of the fracture by allowing the screw to slide in the barrel. Proposed advantages of the device are ease of use, low cost, low blood loss, less reoperation rate and good functional outcome. Surgeons electing to use this implant for patients with intertrochanteric fracture should take into consideration various implant-related and patient-related factors.

Table 1 DHS studies

Design rationale

Dynamic hip screw constructs have a barrel plate, a lag screw that gains fixation in the femoral head and the cortical screws that fix the plate to the proximal femur. The barrel length varies among different manufactures but in general range from 25 mm (short barrel) to 38 mm (long barrel). DHS plates are also available in various barrel angles ranging from 130° to 150°, but the most commonly used is the 130° barrel plate. The femoral head lag screws have shaft diameter of 8 mm, distal thread length of 22 mm and thread diameter of 12.5 mm. The come in various lengths ranging from 50 to 145 mm based on the manufacturers. The lag screw slides in the barrel, thereby allowing dynamic compression at the fracture site when inserted perpendicular to the fracture line.

Biomechanical data

Biomechanical studies have shown good stability of DHS with static and dynamic loading conditions in cadaveric and surrogate bone models [2129]. A normal patient weighing 70 kg would place 2.5 times their body weight on their hip while walking amounting to around 2000 newton of loading and 3700 newton while standing. They undergo at least 10,000 cycles of load on the implant in the first 4–6 weeks after surgery. Most biomechanical tests were performed taking into consideration these aspects of postoperative mobilization (Table 1). Wiser et al. in a recent study on 12 cadavers noted that DHS can withstand static load up to 2778 newton before it fails and was able to withstand cyclical loading of 1400 newton up to 10,000 cycles at 2 Hz. Another study of DHS showed importance of proper DHS screw placement, and it showed that decentralized position leads to rotational failure in the specimen [21].

Clinical data

Biomechanical studies often reveal that intramedullary fixation is more rigid than DHS. However, clinical studies show contradictory results in favor of DHS, especially with stable intertrochanteric fractures. An often-reported clinical advantage of DHS over nailing is the low incidence of femoral fracture and low reoperations. DHS showed clear advantage over Gamma nail in terms of these complications; however, similar favorable results were not seen with DHS over newer nails (PFNA and InterTAN). A recent meta-analysis by Yu et al. [30] showed that DHS has lower cutout, less intraoperative and postoperative fractures and less reoperations compared to Gamma nail. However, DHS failed to show those advantages over non-Gamma nail (PFNA), which was also similar to the finding in other reports [31, 32]. However, Bhandari et al. [33] noted in a meta-analysis in 2008 that risk of femoral fracture is not a concern with third-generation nails, but their meta-analysis revealed a relative risk of 1.8 times for femoral fracture with newer Gamma nails compared to 4.5 times of risk with older nails. Even though there is significant reduction in the risk of femoral fracture compared to older nails, it was still unacceptably high. Various RCTs have shown similar favorable results with DHS compared to Gamma nail, especially with stable intertrochanteric fractures. Hence, with the available evidence, nailing cannot be routinely recommended for stable intertrochanteric fractures and DHS should be the standard of care for stable intertrochanteric fractures with reasonable bone quality.

Factors predicting failure of DHS

DHS fails by various mechanisms including cutout, implant breakage, nonunion and rarely fracture at the tip of plates. Various factors have been shown to increase the risk of DHS failure.

Patient factor: bone mineral density

Osteoporosis has been shown to be an important factor predisposing to cutout in various biomechanical and clinical studies. Windolf et al. showed high failure of DHS implants in specimen with BMD values around 290 mg/cm3 [34]. Patient’s age, which is a surrogate measure of osteoporosis, has been shown to be a risk factor in a study analyzing 63 cutouts in a cohort of 937 DHS fixation.

Fracture factors

Fracture patterns

DHS has produced consistent good results in stable intertrochanteric fractures. However, few authors have raised concerns about high failure rate associated with DHS use in unstable situation. Sadowski noted a failure rate of 35 % when DHS was used for unstable fractures.

Haidukewych et al. [35] in 2001 reported a failure rate of 56 % when using DHS in patients with reverse oblique fractures. Madsen et al. reported a failure rate of around 34 % with DHS when used for unstable fractures. Biomechanical studies have shown less rigidity when unstable fractures were fixed with DHS and constructs failed earlier with cyclical loading. A recent report in 2014 by Evidence-Based Working Group in Trauma after analyzing all the evidence concluded that failure rates of treatment of unstable trochanteric fractures (AO/OTA 31A3) with a sliding hip screw are too high to recommend its use [36, 37].

Fracture reduction

Haidukewych [38] in his report in 2009 emphasized the importance of proper fracture reduction. He reported that no implant, however strong it may be, can sustain the undue forces on the implant in the setting of poor reduction and nonunion. He also emphasized the importance of bone contact and avoiding over distraction. Carr et al. described various techniques that can be used to achieve good intraoperative reduction. Few authors have emphasized the importance of valgus reduction to avoid varus collapse and screw cutout. Pervez et al. [39] noted significant increase in cutout risk with varus fracture reduction in AP plane. De Brujin [40] showed that poor reduction according to Baumgartner criteria is associated with higher risk of cutout (odds ratio 5.19).

Surgeon factors

The surgeon factors include tip-apex distance (TAD), Parker’s ratio, Cleveland zones and Baumgartner fracture-reduction grade.

Baumgartner highlighted the importance of screw position and tip-apex distance [41, 42]. It is the sum of the distance of the lag screw tip from the center in AP and lateral views, and he suggested that risk of cutout increased manifold if the TAD was more than 25 mm. Various authors confirmed their findings. A recent report by Andruszkow [43] specifically looking at the predictors of cutout showed TAD as the most important predictor of cutout, and TAD more than 25 mm is associated with 24 times increase in risk of cutout irrespective of the fracture type. Pervez et al. [39] in their study revealed the Tad was more important than the lag screw position or the fracture reduction. Hsueh et al. [44] in their multivariate analysis of cutout following DHS showed that TAD is more important predictor of cutout than screw position, fracture pattern or fracture reduction.

Parker’s ratio has also been shown to be an important predictor of screw cutout. A Parker’s ratio of less than 40 % and peripheral placement of the DHS screw have been show to increase the risk of cutout by various authors [40]. Pervez showed that abnormal screw from center ratio (similar to Parker’s ratio) is associated with higher cutout risk [39].

Cleveland introduced the concept of screw placement zones. According to Cleveland, the head can be divided into 9 zones based on AP and lateral radiographs. Since then various authors have studied the importance of Cleveland zones in biomechanical setup and clinical setup [25, 27, 40]. Biomechanical studies show that C–C position and I–C position are associated with low cutout risk. Luo et al. [25] in their biomechanics study showed less displacement with CI position. Goffin et al. [27] showed better result with inferior–center position in finite element model even when TAD was more than 25, thereby questioning the practice of over-reliance on TAD.

Hsueh et al. [44] in their clinical study showed that the risk of cutout was highest in the superior 3 zones and lowest in the central zone followed by central–inferior zone. Pervez et al. [39] showed high cutout when the lag screw was anterior in the lateral but failed to mention the zone. De Brujin et al. [40] showed that anterior–inferior and central–inferior positions had favorable results (odds ratio 0.11).

DHS versus DHS blade/DHHS

In order to reduce the risk of failure in osteoporotic patients, a helical blade was designed in the place of DHS lag screw. Helical blade has the proposed theoretical advantages of minimal bone loss, insertion by impaction and better cutout resistance. Various authors have compared the biomechanical properties of helical blade plate to conventional DHS and noted favorable results [19, 21, 24, 29, 34, 4549]. Sommers et al. in their study on surrogate models showed that helical blade constructs showed better results with 1,00,000 cycles of loading at all test loads (0.8 KN, 1 KN, 1.2 KN and 1.4 KN) [24]. Similar favorable biomechanical results of helical blade design over DHS screw were shown by other authors as well [25, 29, 47, 48]. Fang showed comparable clinical results in his recent matched propensity study involving 355 patients (177 DHS blade and 177 DHS) [19]. They had only two failures with the blade plate group compared to 13 failures in the conventional DHS group. These results are in contrast to other studies comparing blade plate to DHS as most other studies failed to show clinical superiority. Fang et al. attribute their excellent results to the blade plate design they used in patients (DHS blade, Synthes), and they state that the less-than-favorable results in other studies [45, 50] may be due to the different blade design with tapering tip (DHHS, Synthes) used by the previous authors.

Cost

DHS has been shown to be the most cost-effective option for stable AO/OTA type 1 fractures and better option for type 2 fracture and least cost-effective for type 3 because of reoperation rate [51].

Cephalomedullary nails

Cephalomedullary nails are the most commonly used fracture fixation device for intertrochanteric fractures in North America [52]. European studies have also shown increasing trend of cephalomedullary nail usage [18]. Considering the design diversity of nails and conflicting data on their complication rate, it is worth looking at them individually.

Gamma nail

Gamma nail (Stryker Howmedica) was first introduced in 1988 [26]. Since then various design modifications have been done and biomechanical and clinical outcome has been reported [21, 28, 5365] (Table 2). It has been used in over million patients, and its results vary between different generations of nails.

Table 2 Gamma nail studies

First-generation Gamma nail

This nail has a proximal diameter of 17 mm, valgus angle of 10°, lag screw diameter of 12 mm and distal static locking screw diameter of 6.5 mm. This has been studied extensively.

Biomechanical studies

Sommers et al. [24] showed that Gamma nail screws (12 mm) sustained more loading cycles before cutout/bending compared to DHS for all dynamic loads tested in polyurethane foam model. Similar results were shown for Gamma nail by Curtis et al. [28] in 20 cadaveric specimens. Rosenblum showed Gamma nail fixation to be a very rigid construct with essentially no strain on the femur with both stable and unstable fractures [65].

Clinical studies

Despite excellent biomechanical results shown by first-generation Gamma nail, various authors have raised concerns about its clinical performance because of high fracture rate [14, 16, 66]. Even though some authors have reported high fracture rate and cutout, most other authors reported good outcome following the use of first-generation Gamma nail, especially in unstable trochanteric fractures (Table 2).

Adverse events

Cutout of the femoral screw is the most common adverse event associated with the use of Gamma nail. The reported cutout was 0–16 %. Next common complication that is more specific for first-generation Gamma nail was femoral fracture reported at 0–5 %. In one of the largest series till date, the reported cutout rate and postoperative fracture rates for first-generation Gamma nail were 1.8 and 0.5 %, respectively [64]. Various authors have reported nail breakage, and it is often quoted to be due to poor reduction and distraction at fracture site causing undue stress on the implant around the proximal lag screw hole [63, 67].

Second-generation Gamma nail

The second-generation Gamma nail proximal diameter was reduced to 16.5, proximal lag screw diameter was reduced to 10 mm, valgus angle was reduced to 4°, and distal static locking screw was changed to 5 mm. Second-generation Gamma nail was reported to have less fracture rate compared to first-generation Gamma nail. Bojan in his large series of 3066 Gamma nails including 933 second-generation Gamma nails noted only one periprosthetic fracture (0.1 %) with the second-generation Gamma nail. Gerri et al. and Efstathopoulos reported 0 % fracture with second Generation gamma nail (TGN, Stryker) in 66 patients and 56 patients, respectively [68, 69].

Third-generation Gamma nail

The third-generation Gamma nail was introduced in 2003, and the major design change is that a dynamic distal locking screw option was added to the nail. Buecking et al. reported only one fracture and one cutout in a series of 80 followed-up patients [70]. Winnock de Grave reported 0 % fracture rate and 0.01 % cutout rate in a cohort of 61 Gamma 3 nail [54]. However, authors like D’Arrigo et al. and Wu et al. reported continuing concern of femoral fracture (2.1 and 5.7 %, respectively) and cutout (5 and 8 %, respectively) with Gamma nail [71, 72].

Proximal femoral nail, Synthes (cephalomedullary nails with two 6.5-mm proximal lag screw)

These second-generation Gamma nails differ from Gamma nail by having two proximal lag screws. The most extensively studied nail of this group is the proximal femoral nail [55, 7380] (Synthes). Other similar nails of this group are Targon nail, Endovis nail, TSN SAN. Schipper et al. in a biomechanical study studied PFN and noted that design modification could lead to low cutout [81]. Ozkan in a biomechanical study compared PFN to locking plate and noted PFN to have higher failure to load with axial load [76]. Morihara et al. in a comparative study showed PFN to have low complications (no cutout, no Z-effect, no fracture) compared to Gamma nail [79]. Ozkan et al. also reported excellent results with no complications and 100 % healing in a cohort of 15 patients [82]. Herrera et al. have shown excellent results with no periprosthetic fracture risk with PFN compared to Gamma nail [55]. Similar excellent results with very low periprosthetic fracture risk were shown by various other authors Norris et al. [83] in a meta-analysis including 18 studies showed that the periprosthetic risk of PFN is less compared to that of Gamma nail and PFNA. One unique problem with this group of implants has been the Z-effect [84, 85]. Koyuncu report 17.7 % complication in this series of 152 patients with 3 Z-effect, 2 reverse Z-effect and 4 screw cutouts [78].

PFNA

These third-generation Gamma nails differ from Gamma nail by having a proximal blade instead of lag screw. The blade was inserted by impaction, thereby avoiding rotation of femoral head and compaction of bone around the blade and reducing the risk of cutout. Hwang et al. [47] showed that PFNA has much resistance to axial load with correct implant position in the center–center or inferior–center position with load to failure of 4175–4462 newton. Konstantinidis showed that in osteoporotic bone PFN withstood 400 cycles of load at 2100 N and in healthy bone it withstood the same load for 20,000 cycles [86]. Various authors have reported the clinical outcome of PFNA [2, 13, 31, 32, 47, 49, 57, 72, 87103]. Few authors have reported excellent result with no cutouts even when used for unstable fracture patterns. Periprosthetic rates have also been reported to be less than compared to Gamma nail. Simmermacher in his study on 315 patients with unstable trochanteric fractures treated with PFNA, largest cohort reported until date, had cutout in 6 patients and periprosthetic fracture in 7 patients, which is much less than the complications encountered with DHS in unstable situations [49]. Various meta-analyses have also shown the superiority of PFNA over DHS and Gamma nail with respect to blood loss, operative time and fluoroscopy time. Hence, PFNA has biomechanical advantage over DHS and also has shown favorable clinical results compared to DHS and Gamma nail in unstable fractures and may be a safer alternative to these devices. Hence, surgeons can consider using these newer-generation nails in unstable or potentially unstable fractures expecting a more favorable result.

InterTAN

InterTAN is the fourth-generation nail with trapezoidal proximal nail geometry giving rotational stability and has two proximal lags screws, and it interlocks with each other, thereby achieving primary compression at fracture site at the time of insertion. Few authors have reported the outcome of this nail [71, 90, 104]. Biomechanical studies have shown that these implants are almost twice as strong as contemporary nails with load to failure noted at around 8000 newton by few authors with the central position and 6000 newton for decentralized position. The reported torque resistance was also high at around 3.8 newton/m. Clinical data on this nail are limited. Matre et al. [18] studied this nail in one of the largest RCT series to date from Norway involving 341 patients. They included 191 unstable fractures and noted cutout [105, 106] in 13 patients and periprosthetic fracture in five patients. Overall, reoperation rate was 8 % (28 reoperations). Wu et al. [71] showed similar good results in a cohort of 87 patients with cutout in 1 patient and periprosthetic fracture in 1 patient. With the available biomechanical and clinical evidence, it is safe to say that InterTAN shows promising early results and may prove to be a valuable addition to surgeon’s armamentarium.

Hemiarthroplasty for intertrochanteric fracture (Table 3)

Hemiarthroplasty has been proposed as an alternative to internal fixation by few authors for unstable intertrochanteric fracture is frail in elderly patients [97, 105, 107116]. Most authors recommend it in a select subset of patients with ipsilateral hip osteoarthritis, ipsilateral AVN of the femoral head, inflammatory arthritis, unstable fracture pattern with poor bone quality, neglected fractures and failed internal fixations.

Table 3 Hemiarthroplasty for IT fracture

Few randomized controlled trails comparing hemiarthroplasty to internal fixation failed to show any no significant difference in functional outcomes, hospital stay, and time to weight bearing or general complication. Emami et al. [108] compared dynamic hip screw (DHS) and bipolar hemiarthroplasty with 30 patients in each group. The average follow-up is 16.5 weeks with bipolar having better functional status and hip range of movements at final follow-up. There was no significant difference in pain severity between these groups. Stappaerts et al. [107] showed similar results in their comparative study of DHS and hemiarthroplasty and noted that there was no significant difference between two groups in the operating time, wound complication, mortality rate and functional outcome. However, they noticed that blood transfusions rates were higher in the hemiarthroplasty group. Similar favorable result for hemiarthroplasty was shown by other authors in nonracist and prospective studies. However, most studies have follow-up of not more than 2 years. In contrast to these studies with favorable outcomes, Kim et al. [105] in a RCT compared proximal femoral nail and long-stem cementless calcar-replacement prosthesis with 29 patients in each group and noted that patients treated with a proximal femoral nail had a shorter operative time, less blood loss, fewer units of blood transfusion, a lower mortality rate and lower hospital costs compared to those treated with the long-stem cementless calcar-replacement prosthesis. There was no significant difference in the functional outcomes, hospital stay, and time to weight bear and general complications. They concluded that primary hemiarthroplasty for unstable intertrochanteric fracture is associated with higher complication rates and proximal femoral nail provides superior clinical outcomes.

Hence, in the absence of concrete evidence, hemiarthroplasty should be undertaken with caution in carefully selected patient with reduced life expectancy and surgeon should be aware of the increased complexity of doing the hemiarthroplasty in these frail patients because of increased blood loss, poor bone quality, absence of calcar or deficient lateral wall, need for abductor repair and higher incidence of dislocation.

Conclusion

Surgeon dealing with intertrochanteric fractures should be aware of contemporary fixation devices and should select a fixation device taking into consideration patient factors, implant factors and fracture factor in a logical evidence-based manner (Fig. 1).

Fig. 1
figure 1

Flow chart IT