Author statement

Dr Vidyasagar Chandankere was involved in study design, data collection, data analysis and writing manuscript.

Dr Hitesh Shah was involved in study design, data collection, data analysis and writing manuscript.

1 Introduction

The pediatric neck of femur fractures are rare and account for about 1% of pediatric injuries. They are usually caused by high-energy trauma like RTA and fall from height. Children do have anatomical differences, different healing responses. The frequency of complications (AVN, Coxa vara, Non-union, and Premature physeal closure) are very high and often irreversible (AVN, PPC). Various factors influencing management and complications include age, the severity of the trauma, fracture type, displacement, timing of intervention, reduction quality, implant choice, immobilization, and joint decompression. Among them, only the last 5 are modifiable through intervention. Optimal management of acute fractures with an effort to reduce complications is important at the time of index surgery. Hence begs the questions like does early intervention reduce complications?, what is the acceptable quality of reduction?, is physeal fixation necessary? the best implant for different fracture types and the usefulness of joint decompression. This rare injury with so many confounding factors, lack of uniform protocol, and variations in a modality of treatment may further confuse a young surgeon. Most of the literature evidence is from retrospective observational case series.

Close

Fig. 1 Flow chart of literature search and selection of articles according to PRISMA guidelines.

Export to PPT

Close

Fig. 2 5 years old boys had a fall from height and sustained right side Delbet type Ia fracture (a), he presented after 24 h, he was managed with Closed reduction and K wire fixation with joint aspiration and hip spica (b), fracture united in 6 weeks and K wires were removed (c). At 9 years of age, the radiograph shows Premature physeal closure and Coxa magna, Coxa breva (d).

Export to PPT

Close

Fig. 3 12-year-old child presented with bilateral displaced femoral neck fractures (Delbet type II) (a), he underwent closed reduction and osteosynthesis with screw fixation with joint aspiration, Right side anatomical reduction and left non-anatomical reduction (b), Fracture united at 3 months (c), the boy developed the left side avascular necrosis after 8 months (d). He underwent implant removal at 9 months on the left side (e).

Export to PPT

Close

Fig. 4 14-year-boy met with RTA and sustained left side Delbet type IV fracture (a). He underwent closed reductions and Locked Plate fixation (b). The fracture healed uneventfully in 3 months (c).

Export to PPT

Close

Fig. 5 2.5-year-boy had fallen from a height and sustained right femur neck Delbet type IV fracture (a). He was managed within 24 h by closed reduction and K wire fixation with hip spica application. 2 Threaded 2.5 mm K wire fixation stopping short of physis and a single 2 mm K wire transfixing the physis to avoid secondary SCFE (b). After 6 weeks' implants were removed and the fracture united uneventfully (c).

Export to PPT

2 Material and methods

We followed the PRISMA guidelines for this research paper (Fig. 1). We searched English literature in the Pubmed database without date restrictions in August 2020. It showed results from January 1977 to August 2020. The keywords and medical subject headings(MeSH) terms used were “pediatric neck femur fracture, proximal femur fracture, complications, management”. We screened the reference list of included articles to identify any further relevant studies. Flow chart attached. (Fig. 1)

3 Results

We included 15 studies for final analysis, which resulted in a total number of 616 pediatric femur neck fractures (range 25–108). The children's age ranged from 6months to 18years(average 10.2yrs). The range of the follow-up period was 1 to 26.2 years (3.9yrs average). The common mechanism was high-velocity injuries in 75% of cases, which included mostly road traffic accidents and fall for height. There were 538(87.3%) displaced and 78(12.6%) undisplaced fractures. The numbers according to Delbet types are; Type I:25(4%), Type II:315(51.1%), Type III:218(35.3%), Type IV:58(9.4%) (Fig. 2-5). Time of intervention was mentioned clearly in only 12 studies, which included 518 cases, among them 214 (41%)cases were fixed within the first 24 h from injury. Most studies attempted closed reduction before opening the fracture for better reduction except Song et al., who did primary open reduction for a certain period. Watson & Jones approach was commonly used for open reduction, followed by the anterior approach. In the total of 616 fractures, 236 underwent open reduction and internal fixation, 321 were managed by closed reduction and internal fixation, 20 with closed reductions and hip spica, 3 with in-situ fixation and hip spica, 1 in-situ fixation, 6 hip spica without reduction, 5 pelvis to foot plaster cast and in 1 case there was no treatment given except bed rest. About 95% of fractures were managed by surgical fixation. Cannulated cancellous screws were the most commonly used implants. Quality of reduction as per 5 studies was noted as excellent, acceptable, and unacceptable in 167,70,15 cases respectively. In 1 case no reduction was done and 2 cases status unknown. Only 4 studies looked into physeal penetration by implant and half of them used smooth pins to cross the physis, the rest 11 studies did not provide any data. Joint decompression was mentioned in 6 studies only, out of which only 3 studies routinely performed it in all cases. Immobilization to protect inadvertent weight-bearing after surgical fixation was done by 10 studies at an average of 6 weeks but the protocol was not uniformly followed in most of them. A total of 259 overall major complications (42%) were noted, they included 144 osteonecrosis (23%), 10 non-union, 43 coxa vara(6.9%), 31 premature physeal closure(5%), 23 limb length discrepancy, 3 deep infections and 2 slipped capital physis. Implant removal was mentioned by only 1 study and they performed it at the mean of 2 years interval.

4 Mechanism and anatomical differences

The inherent resilience in the femoral neck of children and greater osteogenic potential is very good for bone strength and fracture healing.

High energy trauma can disrupt its already precarious blood supply. Damage to the growth plate may further contribute to the shortening and deformation of the femoral head. Seizures and electrocution can stimulate strong muscle contractions and cause this fracture. When associated with trivial trauma, pathological cause (NAI, Osteogenesis imperfecta, bone tumors, metabolic bone disease) must be suspected.¹

In the femoral head, the physis acts as a watershed and separates the blood supply of metaphysis and epiphysis. The predominant blood supply to the femoral head varies according to age.^2–5 The medial circumflex artery is a branch of the femoral artery and predominantly supplies the femoral head via lateral epiphyseal artery from 3 to 4 years onwards. The supply from the artery of ligamentum teres (a branch of the obturator artery) gradually reduces from birth to age 4 months and later increases from age 8 years to provide a peak of 20% of total supply to the femoral head in early adulthood.^4,5

5 Classification

Delbet classification is an anatomical classification based on the location of the fracture. It's relatively simple guides the management, and carries a significant prognosis related to complications (AVN & PPC). Canale et al. noted a high incidence of AVN: Delbet Type I (100%), type II(50%), Type III(27%), and Type IV(14%).⁶ In a 2013 meta-analysis involving 935 cases, AVN incidence was noted as follows: Type I (40%), type II (27%), type III (20%), and type IV (5%).⁷ However, the same has been challenged too.⁸ This classification does not consider displacement, comminution, or fracture configuration. Few studies do not consider Delbet type 4 as a true neck femur fracture. AO classification does consider comminution but does not include physeal separation.

6 Reduction manoeuvres which one to use?

Commonly used manoeuvres are Leadbetter,⁹ Whitman¹⁰ and Flynn¹¹ reduction manoeuvres. Su et al.¹² has also described a percutaneous manipulation technique. Most fractures are managed on a traction table. When open reduction is planned, a radiolucent table with a bump under the affected hip is useful.

7 Early vs delayed fixation

Associated injuries may delay the management.¹³ The pathophysiology after femur neck fracture due to kinking of vessels may be reversible when reduction is done early and fracture is stabilized. However, the damage to vascularity may happen during trauma.¹⁴ AVN incidence (6–10%) was significantly lower with early definitive management.^15,16 Various time limit from 12 to 48 hours following trauma has been suggested as optimal.^17,18 Lack of level I study, RCT, studies with small numbers, delayed presentation make a strong conclusion difficult. Reversibility of pathophysiological events related to the blood supply of the femoral head have been reported.^15,16,19 Bukva et al. reported a reduced incidence of AVN with early intervention.¹⁷ A systematic review showed AVN was 4.2 times higher in the late fixation group as compared to the early treatment group.⁷ Some studies fail to prove any association between time to surgery and AVN.^20,21 Spence et al. noted higher AVN with early intervention but they noted severe displacement and more proximal fractures went to surgery early.¹⁴ Overall complications are higher with delayed fixation.^22,23

8 Conservative vs operative management

Undisplaced fractures are defined as a fracture with cortical continuity in both views.²⁴ Earlier hip spica, Thomas splint, or fixation frame were used to treat undisplaced fractures.²⁵ Without internal fixation coxa vara, delayed union, or nonunion were common.^25,26 Children treated by a spica cast alone had a greater incidence of coxa vara.⁶ The incidence of displacement in undisplaced fracture was very high (50%) with a high rate of complications.²⁷ Feigenberg however managed 7 displaced fractures conservatively without any evidence of AVN.²⁸ Togrul managed stable undisplaced fractures with Pauwel's angle less than 50* conservatively and more than 50* with in-situ fixation.²⁹ Good results were also noted with the protocol of conservative management of undisplaced fractures and prompt reduction and internal fixation for displaced fractures.³⁰ Yeranosian noted a 2.5 times higher rate of AVN in the operated group than the conservative group. They had a high number of Delbet type1 & 2 which required operative intervention.⁷ It can be argued that displaced fractures which mostly need operative intervention, could have suffered higher energy trauma at the time of injury.

9 Reduction quality: what is acceptable and effect on complications

Only 5 studies mentioned reduction quality.^21,33–36 They all used either Shrader et al.³¹ or Song et al.³³ criteria for reduction quality.

Even smaller translations as proposed by Shrader et al.³¹ in younger femur neck may be significantly more than that in older children with bigger diameter neck. As proposed by Song the acceptable 20* angulation limit seems to be very high for a person with an already low neck-shaft angle. Greater angulation or displacement may be less tolerated proximally than distally.³³ It must be borne in mind that displacement may be under-estimated on radiographs. Several studies noted that poor quality reduction could significantly increase the incidence of AVN.^{31,33,34,37,38}

Canale and Bourland found that 96% of fractures in their series which developed avascular necrosis were displaced.⁶ Ratliff reported that 71% of displaced fractures developed complications.²⁵ Baysal et al. noted AVN in 9 (50%) of 18 displaced fractures compared with 1 (11%) of 9 undisplaced fractures.²⁴ In a recent Multicentric study by Wang et al. on 108 children, they noted a rise in the incidence of AVN with the severity of initial displacement. Certain fracture characteristics like posterior translation compared to anterior translation and medial comminution were significantly associated with higher AVN rates. (Fig. 3b)They attributed it to instability and damage to the deep branch of the medial circumflex artery during injury.³⁵ However, 86% of patients were treated after 24hrs, and rotation of distal fracture fragments was not considered.

Non-anatomic reduction with either inferior offset or varus angulation was the strongest predictor for re-displacement after fixation in a study by Weinrobe et al.³⁹ Many others have shown no correlation between the quality of reduction and complications.²¹

10 Closed vs open reduction

Anatomical reduction is known to reduce complications and it is one of the surgeon's modifiable factors. Closed reduction was attempted first and if the reduction is unacceptable or failed then open reduction performed.^{6,15,18,23,24,29,40,41} These studies were not uniform in mentioning the number of attempts before proceeding to open reduction. Forceful manipulation may increase comminution and jeopardize the blood supply further.¹¹ Bimmel et al. cautioned against more than two reduction attempts, to prevent greater displacement and damage to the blood supply.⁴² Ju et al. on the other hand attempted closed reduction 5 times before proceeding for open reduction.³⁴ Only 3 studies have demonstrated, lesser AVN rates with improved quality of reduction and outcomes in the open reduction group.^33,34,38 Ju et al. managed late presenting fractures (more than 24hrs) with fixation and noted lower AVN and coxa vara in open reduction group.³⁴

Song et al. noted an anatomical reduction in 93.3% and 25% cases managed with the open and closed reduction respectively.³³ Yeranosian systematic review had high AVN rates in open reduction and closed reduction. A significantly higher number of open reductions in Delbet type I&II could have swayed the results.⁷ In Delbet type II & III fractures, Dedane noted a higher incidence of AVN (87.33%) in the open reduction group.¹⁸ Baysal recommended more meticulous surgical exposure during open reduction to prevent AVN.²⁴ Riley et al. noted a trend towards higher AVN in open reduction group.²¹ No RCT compared closed reduction with fixation and open reduction.

11 Implant type, size, and number

Age, weight, fracture location, and diameter of femur neck must be considered when choosing the perfect implant and its size. Surgeon preference and availability of implants are also factors affecting this selection. Minimum implants of appropriate size which can maintain stable anatomical reduction are preferred (Fig. 5). When screws or pins are used, placing them in a parallel fashion helps to achieve compression even during healing.(Fig. 2b) Shrader et al. noted higher AVN in the older generation of implants (Smith-Peterson nail, Canakis pins, and Gouffon pins) when compared to the new generation implant.³¹ CC screws are implants commonly used for fixation^17–19 and can be placed short of physis to allow growth while achieving compression. K wires, Knowles pins, and Moores pins on the other hand need placement unto subchondral bone and may cause distraction at fracture site.⁶ When possible 3 largest possible diameter CC screws are used, a however smaller neck may only accommodate fewer numbers or smaller diameter CC screws. 3 Screws placed in a triangular configuration had a higher peak, higher ultimate load, less displacement, and more energy absorption than other configurations (Selvan 2004). Jintao et al. noted inverted triangular fixation demonstrated less stress concentration, was advantageous mechanically, and less likely to cut out.⁴⁶(Fig. 3)

Titanium screws are stronger and have better osteointegration, however, removal may be challenging when compared to SS screws.

One study from Austria documented 22 children with Delbet type II & III, treated with CC screw fixation and 12 of them had a loss of reduction. A higher number of Delbet type III fractures and greater Pauwel angle could be the reason for the failure of reduction in our opinion, as bio-mechanically a fixed angle construct is superior in these scenarios. They did not mention immobilization or early weight-bearing.⁴⁷

Bio-mechanically, it has been shown that a sliding hip screw device is stronger than three parallel cancellous screws for the treatment of basicervical fractures, the orientation of which is similar to but more distal than that of Pauwel's type-3 fractures.^48,49 Aminian et al. showed the construct stiffness of fixed-angle devices to be superior to that of cannulated screws alone for the fixation of a Pauwel's type-3 fracture in cadaveric femora.⁵⁰ Pauwel's Type III fractures experience a higher shear force and demonstrate a tendency to varus, shortening, and non-union.⁵¹ It is important to achieve adequate compression at the fracture site before inserting locking screws while using LCP (Fig. 4). Spence et al. did not find any correlation between the type of fixation and AVN.¹⁴ No significant difference in results was noted on comparing various surgical techniques or implants.¹⁷

12 Physis crossing: is it safe?

Proximal femoral physis contributes about 15% of the femur length and it grows about 3 mm per year. Physeal spanning screws may hamper growth but good union with LLD, or mild deformity is easier to deal with than nonunion, AVN, and coxa vara associated with poor fixation. Premature physeal closure is the most common complication with incidence of 6.5–63.6%.^{6,7,23,25,52,53} Incidence of PPC was higher in the group where pins penetrated the physis when compared to physeal sparing pins.^6,24,29,37 Physis spanning fixation is necessary for all Delbet type 1 and high Delbet type 2 fractures for achieving a stable fixation.⁶ They also have a higher incidence of PPC.⁷ Ng & Cole used smooth pins or screws with distal threads removed while crossing the physis.⁴⁰ Dai et al. noted a significant risk of PPC when epiphysis is penetrated in children younger than 10 yrs.¹³ PPC in a younger child will have a higher risk of shortening or deformity at the proximal femur. Avoiding multiple physeal penetrations and using smooth pins for shorter durations may help prevent PPC. Only 4^24,32,34,40 out of 15 studies in this review mentioned physeal penetration and half of them used smooth pins only.^34,40

13 Decompression of joint

AVN occurring in undisplaced fractures lead some to believe that tamponade in the capsule may be the reason,⁶ hence advocating decompression. Presumed integrity of neck capsule in undisplaced fractures can cause significant elevation of intraarticular pressure, leading to AVN.⁵³ It is usually done after surgical fixation of fracture to be beneficial, as manipulation during reduction practice is there^16,17,19,40 but the level of evidence is not high and others have refuted any and fixation can cause further bleeding and tamponade. The literature evidence supporting this association with AVN.^{7,14,15,18,31} Joint aspiration, capsulotomy, and arthrotomy are various methods to reduce intraarticular pressure, the latter two being more effective than the former.⁵⁴ No difference was found between the efficacy of needle aspiration or arthrotomy in preventing AVN.¹⁷

In a recent study using an intraoperative pressure probe, six patients did not have perfusion to the femoral epiphysis after reduction, and percutaneous capsular decompression was performed, resulting in the return of perfusion in all six patients.⁵⁴ It may become a very useful tool in the future for Intra-operative monitoring of perfusion of the femoral head after reduction and fixation.

In this review only 6 studies mentioned decompression^{17,21,32,34,36,40} and half of them did it in all cases.^32,34,36

14 Post-op protocol: when to brace?

Hip spica being used as primary management for undisplaced fractures is well known but its role as a supplement, to protect primary fixation is poorly mentioned or studied.

Frankel (1959) noted that bone absorbed 75% of load applied to femoral head and appliance took remaining 25% load after osteosynthesis of experimentally created femoral neck fractures.⁵⁵ It highlights the importance of anatomical reduction and protection till fracture union. Younger and active children may not understand instructions and their compliance to the strict non-weight bearing is poor. 11 studies^{20,21,24,30,32–34,36,40,56,57} in this review used immobilization with spica cast or brace or traction in Thomas splint, for a period of 6–8weeks. Half of them used it regularly in all cases^{32,34,36,40,57} and the other half used it variably in a select few.

Flynn immobilized children irrespective of age, with additional hip spica for 6 weeks to protect fracture fixation and prevent inadvertent weight bearing.¹⁵ Singh et al. immobilized children under 6years of age in hip spica for 6 weeks.³² Dedane et al. used long leg casts for 6 months in all children less than 10yrs age.¹⁸ Varshney used 6 weeks of hip spica in all children post-operatively.⁴¹ Inan et al. recommended Hip spica for less than 7yrs age children.²⁰ Dai immobilized all children until callus formation.¹³

15 Implant removal: timing and reasoning

There is no clear consensus or studies to guide the appropriate time for implant removal in children. It is mostly based on institutional practice, surgeon preference, age, symptoms, location of the implant, and complications related to index surgery. In this review, only Baysal et al. mentioned implant removal at the mean of 2 years.²⁴ Implant irritation or prominence, implant failure, non-union, re-fracture with the implant in situ, implant migration, and infection are common indications for implant removal. These are invariably symptomatic. Symptomatic implant removal is justified even with the risk of the additional procedure, however when asymptomatic the guidelines are not clear.

The proximal femur is an area of high stress mechanically and these differ greatly over short anatomic distance and with growth.⁵⁸ The problems associated with the presence of an implant in the proximal femur are stress riser, stress shielding, growth disturbances, and these pose a greater risk in children participating in contact sports. Stress shielding from plates and stress riser from screw holes are well known to increase the risk of fracture. David et al. reviewed implant removal in 801 children with 1223 implants and found 4 clinical predictors, which when a present increased risk of complication to 14.3 times. These 4 predictors are complications after insertion of the initial implant or with a non-elective indication for removal, a neuromuscular disease associated with a seizure disorder, or a neuromuscular disease in those unable to walk. ⁵⁹

16 Conclusion-

Urbanization and increasing road traffic accidents have significantly increased the incidence of pediatric neck femur fracture than in the past. Delayed presentation is common in developing countries. Mechanism of injury, type, and displacement of fracture is only under nature's control. Prompt anatomical reduction with stable fixation must be accomplished. Decompression of joint is a fairly easy and benign procedure without many ill effects, hence it must be performed. Protection of fixation by hip spica or non-weight bearing is necessary until signs of healing are noted. Complications continue to happen despite the best efforts and a longer follow-up is important. Parents must be counselled regarding the frequency of all complications.

This study was done with equal contributions from both authors. They did not receive any financial aid or sponsorship and all expenses were borne by them while preparing this manuscript. This study is not registered anywhere.