1 Introduction

Less than 1% of paediatric fractures are femoral fractures, making them an uncommon occurrence¹,². Children's strong bones and thick periosteum may be responsible for this low occurrence.³ The majority are brought on by high-energy trauma from car crashes and falls from great heights. Despite being an uncommon fracture, there is a considerable risk of long-term morbidity and consequences.⁴ These might result from children's osseous structure and precarious blood supply. Avascular necrosis of the femoral head can occur in between 0% and 92% of cases.⁵ Furthermore, issues including nonunion, coxa vara, and delayed union, particularly in situations when internal fixation is not used.⁶ The hips of children and adults have different anatomical features. Therefore, in order to execute appropriate care, a solid understanding of anatomy is essential.⁷ Neck femur fractures were classified according to Dlbet classification into four types: Type I (trans-epiphyseal) was the least common type accounting for less than 10%, type II (trans-cervical) (40%–50%), type III (cervical-trochanteric) (25%–30%), and Type IV (intertrochanteric) (6%–15%).⁶ The incidence of avascular necrosis regarding each type was as the following: Type I had the highest incidence of 38%, type II at 28%, type I at II 18% and Type 5%. Moon et al. performed a meta-analysis and found that the more the proximity of the fracture, the higher incidence of avascular necrosis.⁸ Alkhatib et al. found that type I and II accounted for the highest risk OR = 3.8 for type I and OR = 2.4 for type II.⁹ An accepted protocol of treatment is not available due to the low incidence of these fractures. Till the 1970s undisplaced fractures were treated by cast immobilization.¹⁰ Hip spica casts and closed reduction were used to treat displaced cervicotrochanteric and transcervical discs. But according to a Hall et al. examination of the muscular forces, fixation using a spica cast alone was unable to offset the muscle forces surrounding the fracture.¹¹ Smooth or cancelous screws are used to fix the displaced fractures after either an open or closed reduction is applied.¹² However, for basal fractures, side plate fixation and compression screws are recommended.¹³ For younger children, immobilization with a Spica cast should be used in addition to fixation. Our goal in doing this meta-analysis is to assess the results following both closed reduction and internal fixation (CRIF) and open reduction internal fixation (ORIF) for children who have fractured their neck femur.

2 Methods

3 Results

3.1

3.1 Results of the literature search

At the end of our literature search, we had 641 studies to be screened which turned into 389 studies after discarding duplicates. We included only 20 studies after title/abstract screening, and we reached a final of 13 papers after the full-text screening that met our eligibility criteria. All the included studies were retrospective.^3–15Fig. 1 shows the PRISMA chart of the literature search. In the included studies, 294 patients underwent ORIF procedures and 266 patients underwent CRIF procedures. The age of patients ranged from two years up to 18 years. Follow-up durations were mostly more than 1 year. Most patients had displaced fractures of the neck femur. The summary of baseline characteristics of included studies is shown in Table 1. Fractures of the femoral neck were classified according to Delbet's criteria into four types. The majority of patients in the included studies had type Ⅱ fractures, to a lesser extent type Ⅲ and type Ⅳ, while only nine patients had type Ⅰ. Table 2 shows details of fracture types among included studies.

Close

Fig. 1 shows a PRISMA flow diagram of our literature search.

Export to PPT

Table 1 Baseline characteristics of included studies.

Study	Number	Country	Time of the study	Follow-up duration (Years)	Age (years)	Gender (Male/female)	Injury to treat time (days)	Fracture displacement (N./Total)
Dendane2010		Maroc	1999–2006	1.1
ORIF	13				13.2 ± 1.75	11/2	5.23 ± 5	13/13
CRIF	8				10.3 ± 2.25	7/1	4.56 ± 3.63	8/8
Bali2011		India	1998–2007	2.7 (1.1–9.5)
ORIF	18				10 ± 3.25	20/16	NR	17/18
CRIF	13							10/13
Song2010		Korea	1989–2007	2.8 (1.3–13)
ORIF	15				9.7 ± 2.5	11/16	<1 (in 24 cases)	15/15
CRIF	12				10.1 ± 2.75			12/12
Dhammi2005		India	1992–2002	42–9
ORIF	9				11.67 ± 3.4	6/3	<10	NR
CRIF	17				10.38 ± 4.22	8/9
Ju20016		China	2005–2014	2.9 (1.4–5.1)
ORIF	37				9.21 ± 3.14	26/11	4.9 ± 2.47	37/37
CRIF	21				8.85 ± 3.79	14/7	4.3 ± 2.1	21/21
Wang2019		China	2010–2017	1.63 ± 1.06
ORIF	138				10 ± 3.9	136/105	3.7 ± 2.7	NR
CRIF	103
Yerli2021		Turkey	NR	2.05 ± 1.1
ORIF	16				9.5 ± 5.06	20/15	<1	NR
CRIF	9
Chaudhary2021		India	2017–2019	1.15 ± 0.17
ORIF	12				11 ± 3.18	13/8	<1 (in 9 cases) > 1 (in11 cases)	NR
CRIF	9
Lin2012		China	2006–2010	1.1 (1–3)
ORIF	19				8.1 ± 1.3	11/8	NR	19/19
CRIF	15				7.9 ± 1.5	9/6		15/15
Stone2015		USA	2003–2012	2.1 (1–8.7)
ORIF	6				11.5 ± 3.2	13/9	<1 (in 18), >1 (in 4)	6/6
CRIF	16							16/16
Flynn2001		USA	1980–2000	7.61 ± 4.37
ORIF	5				7.6 ± 2.65	2/3	<1	5/5
CRIF	13				8.15 ± 3.18	8/5		13/13
Midrad2001		KSA	1992–1999	3.28 ± 2.18
ORIF	3				9.1 ± 3.8	6/8	NR	9/14
CRIF	10
Togrul2005		Turkey	1978–1994	1.21 ± 0.41
ORIF	3				11.67 ± 1.25	3/0	<2 (in all except 5 cases)	3/3
CRIF	20				9.45 ± 3.25	7/13		20/20

Table 2 Fracture types according to Delbet's criteria.

Study	Type of fracture (Delbet's criteria) (Number)
	Type Ⅰ	Type Ⅱ	Type Ⅲ	Type Ⅳ
Dendane2010
ORIF		4	7	2
CRIF		4	4	0
Bali2011
ORIF		12	3	3
CRIF		3	5	5
Song2010
ORIF		8	7
CRIF		7	5
Dhammi2005
ORIF		3	6
CRIF		7	10
Ju20016
ORIF		19	14	4
CRIF		11	7	3
Wang2019	5	163	72	1
ORIF
CRIF
Yerli2021		15	5	15
ORIF
CRIF
Chaudhary2021	4	13	4
ORIF
CRIF
Lin2012		25	21	7
OR
CRIF
Stone2015		13	8	1
OR
CRIF
Flynn2001
OR		2	3
CRIF		6	5	1
Midrad2001		6	4	3
OR
CRIF
Togrul2005
OR		2	1
CRIF		7	8	4

3.2

3.2 Results of outcomes

3.2.1

3.2.1 1: avascular necrosis

All the included studies reported avascular necrosis outcomes.^3–15 The overall analysis revealed no significant difference between either groups (RR = 0.84, [95% confidence interval (CI) = 0.60, 1.18], P = 0.32). The data were homogeneous (P = 0.22, I² = 27%). Fig. 2 shows the results of analysis.

Close

Fig. 2 shows avascular necrosis outcome.

Export to PPT

3.2.2

3.2.2 2: coxa vara

Only eight studies reported Coxa vara outcome.^{4–7,9–11,15} There was no significant difference between either groups (RR = 0.69, [95% CI = 0.30, 1.58], P = 0.38). The data were homogeneous (P = 0.22, I² = 27%). Fig. 3 shows the results of the analysis.

Close

Fig. 3 shows the coxa vara outcome.

Export to PPT

3.2.3

3.2.3 3: non-union

Six studies reported non-union outcomes.^4–7,12,13 There was no significant difference between the ORIF group and the CRIF group (RR = 0.45, [95% CI = 0.16, 1.14], P = 0.12). The data were homogeneous (P = 0.49, I² = 0%) as shown in Fig. 4.

Close

Fig. 4 shows the non-union outcome.

Export to PPT

4 Discussion

Femoral neck fractures are rare in children and account for 1% of whole fractures. A road traffic accident is the main aetiology of femoral neck fracture as it is usually high-energy trauma.⁴ The neck of the femur in children is denser than in adults which explains the lower incidence and the required higher energy.¹⁶ However, fractures can also happen by low energy traumas when the bone is diseased as in bone cysts or fibrous dysplasia which is called a pathological fracture.¹⁷ Among all types of femoral fractures in children, fracture neck femur contributes to only 7% in patients aged less than two years and around 12.8% in patients aged from 13 to 18 years.¹⁸ It is associated with a high incidence of complications such as avascular necrosis, coxa vara, and non-union¹⁹; therefore internal fixation is the choice to achieve stability and proper anatomical reduction.¹⁹ The evidence on the proper treatment strategy is still unclear due to the lower incidence of this type of fracture.²⁰ Therefore, we performed this systematic review and meta-analysis to investigate the proper method of reduction to decrease the complications.

We included 13 studies with a total population of 560 children: 294 underwent ORIF and 266 underwent CRIF. We observed no significant difference between both two methods of reduction in decreasing the incidence of avascular necrosis, coxa vara, or non-union. Also, the literature failed to favour a special method over another. Alkhatib et al. conducted a meta-analysis on 231 patients and found a non-significant difference between ORIF and CRIF in avascular necrosis risk which supported our result.²¹ While Moon et al. found that surgical interference increased avascular necrosis risk; they explained that as in these cases who required surgery, the blood supply had been interrupted by the action of trauma which was more likely to be severe, and repeated trials of reduction or fixation devices could interrupt the blood supply to the head of the femur.^22,23 It was the same in Yeranosian et al. – a meta-analysis that included 935 patients – which found that open reduction increased avascular necrosis risk.²⁴ On the contrary, Chen et al. found that open reduction significantly decreased the risk of all complications regarding avascular necrosis, coxa vera, and non-union.²⁵

Avascular necrosis is considered the long-term outcome of a fractured neck of the femur, and it is likely to appear in the first year of trauma.^8,26 It is almost the commonest complication of fracture neck femur with incidence ranging from 17% to 47%.¹⁹ It is caused by multiple mechanisms such as interruption of blood supply by severe trauma, injury of blood vessels during reduction or fixation,²⁷ or intra-capsular hematoma that increases pressure inside the joint and prevents blood arrival to the head of the femur. Many factors affect avascular necrosis as age, displacement, type of fracture, and method of reduction.⁸ Regarding the method of reduction, there is a debate in the literature about its effect on decreasing avascular necrosis risk. Bali et al. reported When it is possible, they advise internal fixation for paediatric femoral neck fractures because conservative treatment has a significant failure rate for reduction. They contend that anatomical reduction should be the goal of aggressive surgical procedures, and ORIF should always be preferred to CRIF.⁴ Besides, they considered that avascular necrosis was not related to the used reduction method. However, it was related to other factors such as fracture type, age, displacement, fixation method, and surgery time.²³ This was in line with Wang et al. who found that avascular necrosis incidence significantly increased with older ages, displacement, poor reduction quality, and closed reduction. Also, in open reduction, the use of screws and plates was significantly better than Kirschner wire or cannulated screws.⁸ Moreover, a systematic review and meta-analysis found a significant reduction in avascular necrosis by using ORIF over CRIF; however, a small number of studies were included in this study only six studies with a total population of 198 patients.²⁵ On the contrary, another systematic review and meta-analysis found that the rates of avascular necrosis decreased significantly by using CRIF; however, this result was uncertain as there was a predominance of ORIF in the type three Dlbet classification.²⁴

Also, the time of reduction had a role in lowering the avascular necrosis incidence as it could reduce the time of ischemia and regain circulation.^12,19 In Yeranosian et al. the avascular necrosis risk increased by 4.2% when the treatment was delayed more than 24 h.²⁴ However, Alkhatib et al. found no significant difference between fractures managed early before 24 h and fractures managed after 24 h.²¹ In addition, age affected avascular necrosis as its risk increased by 1.14% per year²³ which supported Wang et al. results.⁸ The variety of neck fracture types chosen for proper management is still difficult and unclear.

Coxa vara is a less common outcome. It accounts for 20%–30% of the complications of femoral neck fractures. It is usually asymptomatic and realignment osteotomy is the treatment choice (21,30). Many reasons contribute to coxa vara such as improper reduction, absence of reduction, osteonecrosis, delayed union, and premature closing of proximal femoral physis (20,30). Moreover, coxa vara increases the incidence of hip arthritis and affects lower limb growth.²⁴ Internal fixation can reduce coxa vara.¹⁹ However, the proper method of reduction is still unclear. Also, its incidence is the highest in Type I in the Delbet classification and is the lowest in Type IV.²⁴ Although Chen et al. found that ORIF significantly decreased coxa vara's risk,²⁵ Yeranosian et al. found no difference between the two methods which is consistent with our analysis.²⁴

Non-union occurs in mid or basal cervical fractures because of mal-reduction or improper configuration of the fracture.²⁰ Non-union treatment should not be delayed on contrary to avascular necrosis and coxa vara. Then, internal fixation is used.¹⁹ Reciprocally with coxa vara and avascular necrosis, Delbet type II fracture accounts for the highest incidence of mal-union.²⁴ Yeranosian et al. found that the method of reduction did not affect non-union which agreed with our analysis²⁴; however, Chen et al. found a significant reduction of its risk by using ORIF.²⁵

5 Strengths and limitations

Our systematic review and meta-analysis included 13 studies with a total of 560 patients which is a proper number in such a rare condition. Also, all the statistical results were homogenous. The main limitation of our study was that all included studies were retrospective which increased the risk of selection bias.

The difference in management protocols between the included studies could be a bias as no specific management was done in similar patients' conditions. The majority of the studies changed the closed reduction to open reduction when the closed reduction failed which could affect the outcomes of open reduction negatively as this failure consumed time and led to an increased incidence of complications.

Another risk of bias could be the difference in clinical skills, diagnostic skills, and proper assessment of patients' conditions between studies' conductors. One example of this risk could be the wrong interpretation of Dlbet classification between physicians which could affect the most proper decision of reduction that could affect the risk of complications.

The variety of treatment options and factors affecting the healing of the fracture, besides the rare incidence of the disease, made it difficult to prefer one method over another. The majority of included studies did not report separate data on each type of fracture, age group, or presence of displacement at the beginning. All these factors could affect healing and complications risk. This made us unable to perform statistical analysis to know the most proper method in all these conditions. The rarity of the femur's neck fracture in children made it difficult to perform studies with a huge number of patients with subgrouping according to these factors. Therefore, the upcoming primary studies should properly study the risk factors, especially the fracture type and displacement of the fracture and should investigate the efficacy of each method in decreasing the complications of fracture neck of the femur on each type of fracture whether displaced or not.

6 Conclusion

At the end of our study, We found no difference between ORIF and CRIF in decreasing its risk. Other factors could contribute to it as fracture type as the incidence was increased in type II compared to the other types. Despite we found non-significant differences between ORIF and CRIF in decreasing the risk of complications. Further investigations of other factors are mandatory to approve this.

Registration and protocol

The paper is not registered and the protocol is not prepared.

Funding

No funding was received for this study.

Data availability statement

Data will be available to any researcher who contact the corresponding author.

Ethical statement

Ethical approval was obtained from the Ethical Research Committee of Cairo University, Egypt.

Funding statement

No funds were received.

Guardian/patient's consent

Not applicable.

CRediT authorship contribution statement

Mohamed Abdel-Wahed: Conceptualization, Methodology, Software. Mohammed Elmahdy Ismail: Data curation, Writing – original draft. Ahmed Amin Galal: Visualization, Investigation. Asmaa Younis Elsary: Software, Validation. Hisham Abdel-Ghani: Writing – review & editing. Ebeed Yasin: Supervision.