Early results of the femoral neck system in fixation of femoral neck fractures

INTRODUCTION

Femoral neck fractures can present in elderly patients following low-energy mechanism or in younger patients following high-energy mechanism.[1] In the elderly population, arthroplasty is recommended for displaced fractures, while fixation is recommended for undisplaced fractures.[2] The choice of fixation device is either multiple cancellous screw (MCS) or fixed-angle device (FAD), guided by the fracture orientation as per Pauwels classification.[3] In the younger population, internal fixation after closed or open reduction is done to preserve the joint, but the optimal implant is unknown.[3] Multiple cancellous screw (MCS) is a simple and minimally invasive method, but it is not able to withstand the shear force in vertically oriented fractures.[3] On the other hand, FAD can withstand greater shear force, but may be technically more challenging to insert and is associated with more blood loss.[1]

The femoral neck system (FNS) (DePuySynthes, Zuchwil, Switzerland) has been introduced as a device that provides superior angular and rotatory stability compared to MCS,[4] with the benefit of minimally invasive instrumentation compared to conventional FAD.[5] Data on FNS are limited to biomechanical studies,[4,6] small clinical studies[7,8] and those of unstable configuration only.[9] We aimed to investigate the early outcomes of femoral neck fractures stabilised with the FNS. We hypothesised that FNS fixation will be performed quickly and have a low failure rate. This study will aid our understanding of this new implant’s utility and allow surgeons to have an informed choice of implants.

METHODS

This was a single-institution retrospective study of all femoral neck fracture fixations with the FNS performed between July 2020 and June 2021. An institutional change in practice was implemented from July 2020, whereby all femoral neck fractures were fixed with the FNS. Ethical approval was obtained before the initiation of the study.

Study subjects included young patients with undisplaced and displaced isolated femur neck fractures and elderly patients (≥60 years) with type I or II femur neck fractures according to Garden’s classification.[10] The clinical records of the subjects were reviewed to obtain data on demographics, clinical parameters including Charlson comorbidity index (CCI),[11] energy level of trauma, operative duration, consumption of alcohol, steroids and cigarettes, and complications.

Fixation

All surgeries were either performed or supervised by consultants on the traction table. Under the guidance of image intensifier anteroposterior (AP) and horizontal beam lateral images, alignment of the fractures was assessed by the consultants. For young patients with displaced fractures, reduction of the fractures was performed before fixation. For elderly patients, no reduction attempt was made regardless of the alignment. The FNS was then instrumented according to its surgical technique guide,[5] via a lateral approach. Typically, a one-hole FNS plate was used to avoid inserting a screw in the subtrochanteric region, which may be a stress riser. Two-hole FNS plates were used only when the fractures were deemed basicervical, Pauwels type III or comminuted. Postoperatively, the weight-bearing status was dependent on the stability of the fracture and the surgeon’s preference.

Radiographic analysis

Preoperative pelvis AP and lateral radiographs of the affected hip were used for radiographic analysis. On the AP view, fractures were assessed to be either undisplaced or displaced using simplified Garden’s classification.[10] Garden’s alignment index (GAI) was measured.[12] Pauwels angle[13] was also assessed and categorised into type I (<30°), type II (30°–50°) or type III (≥50°).

On the lateral view, the posterior tilt was measured according to the method described by Palm et al.[14] This was described as the angle between the line of the mid-column of the distal neck and the line from the centre of the head to the intersection of the drawn caput circle and the mid-column line.[14] This was stratified into <20° or ≥20°, as a posterior tilt of ≥20° was known to be associated with higher risk of failure of fixation of femoral neck fractures in the elderly.[15] If fracture reduction was performed intraoperatively, the immediate post-fixation films were used for measurement of the above parameters. These were done as the quality of reduction has been known to influence the outcome of the fixation.[16]

Pauwels angle, posterior tilt and GAI were separately measured again by the same author and another author to ascertain their intra- and interobserver reliability. The average of the three readings was used. Patients were reviewed at postoperative intervals of 1 month, 2 months, 3 months, 6 months and 1 year.

Statistical analysis

The Shapiro–Wilk test was performed to determine whether the data was parametric or nonparametric. Its central values and spread were reported accordingly. To assess interobserver reliability, Cohen’s kappa test was performed for categorical data and intraclass correlation coefficient (ICC) was calculated for continuous data. All analyses were performed using IBM SPSS Statistics version 22.0 (IBM Corp, Armonk, NY, USA). A P value of < 0.05 was deemed statistically significant.

RESULTS

The FNS was used in the fixation of femoral neck fractures in 36 subjects. Table 1 shows the demographics, radiological and surgical parameters. Overall, 31 (86.1%) patients sustained low-energy trauma, while four (11.1%) and one (2.8%) patients had high-energy trauma and atraumatic fractures, respectively, with 15 (41.6%) having either alcohol, steroids or cigarette consumption. The surgeries were performed by six consultants, two associate consultants, one staff surgeon and three residents. Postoperatively, ten (27.8%) patients were allowed full weight bearing immediately. Sixteen (44.4%) patients were allowed some form of partial weight bearing, and six (16.7%) patients had progressive weight bearing that lasted between 1 and 3 months. Only four (11.1%) patients were not allowed weight bearing postoperatively. The duration of partial, progressive and non-weight bearing ranged from 2 weeks to 3 months.

Table 1.

Demographics, radiological and surgical parameters of the subjects.

Parameter	Overall (n=36)	Undisplaced (n=31)	Displaced (n=5)	P*
Agea (yr)	77 (44, 89)	51 (47, 60)	0.001
Gender
Female	19	18	1	0.136
Male	17	13	4
Body mass indexb (kg/m2)	22.1±4.5	21.2±3.8	27.4±5.5	0.112
Charlson comorbidity index
≤2	11	6	5	0.001
≥3	25	25	0
Pauwels/GAIb (°)				0.001
Type I/GAI	6/179±11	6/179±11	0/–	I: –
Type II/GAI	26/173±7	24/173±8	2/168±3	II: 0.498
Type III/GAI	4/150±10	1/159(–)	3/147±7	III: 0.500
Posterior tilt (°)
<20	32	27	5	0.534
≥20	4	4	0
Time from injury to operationa (day)	2 (1, 8)	3 (1, 8)	1 (1, 2)	0.005
Time from admission to operationa (h)	49 (16, 182)	50 (15, 182)	20 (17, 24)	<0.001
Operative timeb (min)	71±18	69±18	81±14	0.600
Complications	5	4	1	0.549

^aData presented as median (range). ^bData presented as mean±standard deviation. *Between undisplaced and displaced groups. GAI: Garden’s alignment index

There were statistically significant differences between the undisplaced and displaced groups in the distribution of age, CCI, Pauwels type, median time from injury to operation, and mean time from admission to operation. All reduction of the displaced fractures was performed closed.

The intra- and interobserver ICC values for Pauwels angle were 0.82 (95% confidence interval [CI] 0.68–0.90) and 0.88 (95% CI 0.78–0.93), respectively, which are regarded as ‘good reliability’.[17] The interobserver kappa value for Pauwels classification was 0.77 (P < 0.001), which is regarded as ‘substantial’.[18] The intra- and interobserver ICC values for posterior tilt were 0.91 (95% CI 0.83–0.95) and 0.93 (95% CI 0.86–0.96), respectively, which are regarded as ‘excellent reliability’.[17] The interobserver kappa value for stratifying the posterior tilt to <20° or ≥20° was 0.87 (P < 0.001), which is regarded as ‘almost perfect agreement’.[18] The intra- and interobserver ICC values for GAI were 0.94 (95% CI 0.88–0.97) and 0.90 (95% CI 0.81–0.95), respectively, which are regarded as ‘excellent reliability’.[17]

The mean follow-up duration was 19.9 ± 9.7 weeks. One patient was lost to follow-up after 6 months due to demise. One patient was lost to follow-up after 3 months, and three more patients were lost to follow-up immediately after discharge, 1 month and 2 months postoperatively each. None reported complications at their last review. All, except the patient lost to follow-up immediately after discharge, were full weight bearing with no complaints.

Four (11.1%) patients developed complications. The first patient was an 89-year-old woman with an undisplaced, Pauwels type II fracture with a posterior tilt of 19.6°. She was allowed full weight bearing postoperatively and was noted to have implant cut-out at the femoral head at 2 months after surgery. This was revised with total hip arthroplasty. Of note, she had a significant history of 30 pack-years of cigarette consumption.

The second patient was a 79-year-old woman with an undisplaced, Pauwels type II fracture with a posterior tilt of 23.0°. She was allowed full weight bearing postoperatively, and she reported persistent pain at 5 weeks post operation. Computed tomography of the hip at 2 months after surgery showed partial union, and non-weight bearing was advised for a month. At 3 months post operation, she was allowed full weight bearing, as the radiograph showed stability and her pain had improved. She returned at 22 weeks after surgery with acute pain, and radiographs showed a broken locking screw at the junction of the screw head and screw shaft [Figure 1a–d], which was revised with hemiarthroplasty.

Figure 1.

Immediate postoperative (a) anteroposterior (AP) and (b) lateral radiographs of subject. The average posterior tilt angle reading was 23°. (c) AP and (d) lateral radiographs at 22 weeks postoperation show the broken locking screw.

The third patient was a 75-year-old woman with an undisplaced, Pauwels type II fracture with a posterior tilt of <20°. The surgery was completed uneventfully in 68 min. She had partial weight bearing for 2 weeks before progressing to full weight bearing. She was noted to have radiographic signs of avascular necrosis of the femoral head with a slight collapse of the head at 6 months postoperation. The FNS implant was intact. Her medical history had no significant risk factors for avascular necrosis. She remained asymptomatic and was planned for close observation.

The fourth patient was a 69-year-old man who presented 4 days after an initial injury with an undisplaced, Pauwels type III fracture that extended to the basicervical region inferiorly. In anticipation of higher shear stress, a two-hole FNS plate was used [Figure 2a–e]. He refractured at 7 weeks after surgery due to a near-fall incident, during which he stepped hard on the operated limb and felt a crack. He was asymptomatic before the incident. Radiographs of the hip showed a refracture, with the anti-rotation screw and the bolt cutting out superiorly through the fracture site. Radiograph also showed that the fracture had collapsed into a varus position. It was revised with total hip arthroplasty.

Figure 2.

(a) The AP radiograph of subject 36, showing the near-vertical orientation of the fracture line from the subcapital region superiorly to the basicervical region inferiorly (arrows). (b) and (c) show the AP and lateral radiographs, respectively, immediately post-fixation. (d) and (e) are the respective AP and lateral radiographs following his near-fall incident, showing a refracture. AP: anteroposterior

No other complication, including infection, was reported.

DISCUSSION

The FNS was introduced as a device for femoral neck fracture fixation that incorporates the advantages of the minimally invasive insertion technique of MCS and the increased biomechanical properties of an FAD.[4] Its initial biomechanical study in a 70°-oriented femoral neck fracture found the FNS to be superior to the three cancellous screw (3CS) fixation but comparable to the dynamic hip screw (DHS) system in withstanding shortening of the fracture.[4] In another biomechanical study investigating a 40°-oriented femoral neck fracture with 15° posterior bone wedge loss, the FNS was able to withstand varus deformation for a significantly higher loading cycle compared to fixation with Hansson pins. In addition, evidence based on its design showed that the FNS provided more rotational stability than a 3CS fixation.[6] Clinical papers comparing the FNS to other fixation devices have a small sample size[7,8] or are specific to young, unstable fracture type.[9] To the best of our knowledge, this present clinical study provided the largest number of FNS fixations.

The minimally invasive instrumentation of the FNS should reduce the operative time. Once the guide wire is in place, subsequent steps are performed via its jig including the antirotation screw, which would require an additional step in a DHS system. Our mean operative time of 71 ± 18 min was longer than what had been reported in the literature, which was 43 ± 10,[7] 42 ± 13[8] and 42 ± 4 min.[9] The learning curve to the new implant technique was the likely cause as we had 12 surgeons of various levels sharing the operative load. In comparison to other studies, the fixations were performed by a single surgeon in one study[8] and the ‘same group of surgeons’ in another study.[9]

Our study reported four (11.1%) cases of reoperation. The reoperation rate in our study is lower than that of an international, mutlicentre randomised controlled trial showing 20% and 22% of reoperation after sliding hip screw and cancellous screws fixations, respectively.[19] Zhou et al.[9] reported a case of delayed union and a case of refracture from a subsequent fall among a group of 30 young patients with displaced, Pauwels type III fractures that were fixed with the FNS. All their patients achieved satisfactory reduction and were non-weight bearing for 6 weeks. Vazquez et al.[7] reported neither failures nor reoperation in their study of 15 fixations of geriatric Garden-I and -II femoral neck fractures with <20° of posterior tilt using the FNS. Nibe et al.[8] reported union in all 25 cases with the FNS, with none requiring revision surgery; the Pauwels classification of these patients was not known. Nibe et al. attributed their good outcome to strict patient selection criteria for fixation;[8] this could have also accounted for the success in Vazquez et al.’s[7] study. In Zhou et al.’s[9] study, the achievement of satisfactory reduction in a timely manner could have resulted in their success in fixation with the FNS.

Among the four patients who had complications in our study, one patient had no apparent cause for development of avascular necrosis, apart from the initial trauma. In the other three patients, complications likely arose due to various other causes — the high posterior tilt angle could have contributed to the cut-out and the screw breakage. Of note, the cut-out had occurred despite the blunt-tip bolt design of the FNS. Additionally, the promise of greater biomechanical strength of the FNS could have led to false confidence in accepting these parameters. In our patient with fracture displacement, the force of the new trauma in its early postoperative stage was still greater than the strength of the fixation, despite the use of a two-hole plate. In light of these factors, the FNS cannot be used to compensate for an unfavourable fracture alignment. Protected weight bearing may still be needed in certain cases.

The FNS has a smaller implant footprint and allows for a controlled collapse of 15 mm with no lateral protrusion.[5] These should reduce the risk of soft tissue irritation on the lateral hip. None of our patients reported symptoms in the lateral hip, although it may be due to a short follow-up. Other potential benefits in comparison to the DHS include not requiring to tap the femoral head, which may cause a rotational loss of reduction due the high torque generated while tapping, especially in young hard bone, even in the presence of an anti-rotation wire.

One of the strengths of our study was the institutional implementation of the FNS, which reduced bias in implant selection and allowed for heterogeneity of surgeons. Another strength is our detailed description of the circumstances surrounding the complications, highlighting the potential risk factors that contributed to their failure. Additionally, our initial broad inclusion of patients enabled us to uncover the limits of the FNS.

There are limitations to the present study. First, our sample size was modest; nonetheless, the small number still allowed for complications to surface. Second, the short follow-up period limited our appreciation of any potential long-term complications, such as avascular necrosis or hardware-related trochanteric pain.[7] Nonetheless, in testing a new device that is designed to provide superior stability to conventional devices, early reports are useful for detecting any early failure or safety issues, such as those reported in this study, although subgroup analysis was not performed to draw out specific causes. Finally, we did not report on functional outcomes or have a comparative group in our study; these areas should be addressed in future studies.

In conclusion, the FNS is a new device that is reasonably fast and simple to instrument. Varying level of familiarisation may be required with the device. Failures can still occur with the FNS; multiple factors, including patient and fracture pattern factors, need to be considered to determine the cause of failure. Surgeons’ discretion is needed in patient selection, keeping in mind the need for satisfactory radiological parameters and appropriate postoperative weight-bearing status.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.