Fractures of the femoral neck in children are rare injuries that typically result from high-energy trauma and account for fewer than 1% of all pediatric fractures.1 These severe injuries are associated with a high rate of complications, including osteonecrosis of the femoral head,1–9 malunion with the coxa vara,2,5 nonunion,2 and premature physeal growth arrest.1 The most serious and common complication after femoral neck fractures in children is osteonecrosis. Its occurrence is highly variable, with reported incidences ranging from 0% to 92%, depending on various prognostic and treatment factors.1–9 A variety of factors, including severity of displacement,4,5,10,11 fracture type,5,6,12 age,6,13 timing,6,14 and quality of reduction,13,15,16 have been implicated in the development of osteonecrosis. The best treatment strategy to reduce osteonecrosis remains a matter of debate. Some studies suggest that open reduction and internal fixation (ORIF) reduces the risk of osteonecrosis,8,15 whereas others report an increased incidence of osteonecrosis associated with open treatment.9,17 There is limited direct comparative evidence in the literature on open vs closed reduction.8
This study compared pediatric patients with fully displaced femoral neck fractures treated with either ORIF or closed reduction and internal fixation (CRIF). Factors evaluated included: (1) the rate of osteonecrosis of the femoral head; (2) the quality of radiographic reduction and fracture healing; (3) the total occurrence and severity of complications; and (4) the final clinical and radiographic outcomes, as assessed by the Ratliff criteria.1
Materials and Methods
Through an International Classification of Diseases, Ninth Revision, code query that was approved by the institutional review board, the authors identified 53 patients who were treated for femoral neck fractures at their institution between 2003 and 2012. Inclusion criteria were as follows: (1) diagnosis of a fully displaced femoral neck fracture with no anatomic cortical contact; (2) age of at least 4 years at the time of treatment; and (3) clinical/radiographic follow-up of at least 1 year. Excluded were 16 patients who had nondisplaced or partially displaced (partial anatomic cortical contact) fractures, 4 who were outside of the eligible age range, and 1 who had insufficient radiographs. In addition, 10 patients (19%) were excluded because of insufficient follow-up. Medical records, including operative reports and all radiographs, were retrospectively analyzed. Demographics, clinical characteristics, injury mechanism and type, treatment details, complications, and subsequent surgeries were recorded.
A total of 22 patients/hips were included in the study (Table 1). There were 13 boys and 9 girls, with an average age of 11.5 years (range, 4.5–17.4 years) at the time of treatment. Total mean follow-up was 2.1 years (range, 1–8.7 years). In 6 patients, ORIF was performed; CRIF was performed in 16 patients. There was no difference (P=.856) in age at treatment, body mass index (P=.657), preoperative fracture classification (P=.195), or duration of follow-up (P=.112) between the 2 groups (Table 2). Only 50% of the ORIF group and 94% of the CRIF group underwent treatment within the first 24 hours after injury (P=.046). All patients treated with ORIF were male, so there was a significant difference (P=.046) in sex distribution between groups.
Preoperative Variables by Treatment Group
Open reduction was performed with an anterolateral Watson Jones18 approach. Briefly, an incision was made about 2.5 cm distal and lateral to the anterior superior iliac spine and curved distally on the lateral aspect of the proximal femur. The fascia was incised and the interval between the tensor fascia lata and the gluteus medius was developed, exposing the hip capsule. The capsule was opened longitudinally, in line with the femoral neck, and then around the labrum in a T fashion. The fracture was exposed, carefully debrided, and reduced gently. Provisional fixation was achieved with K-wires. Definitive fixation was then introduced from the lateral cortex of the femur from the base of the greater trochanter. The fracture was fixed with 3 cannulated 6.5-mm screws (Synthes, Inc, West Chester, Pennsylvania) in 2 cases, a pediatric compression hip screw and plate (Smith & Nephew, Inc, Memphis, Tennessee) in 3 cases, and a cannulated screw (Synthes, Inc) and compression hip screw (Smith & Nephew) in 1 case. For closed reduction, the patient was placed on a fracture table, with traction applied to the leg that was manipulated, with internal rotation and variable degrees of abduction. Internal fixation was achieved percutaneously during closed treatment with cannulated 6.5-mm screws (Synthes) in 8 cases and a sliding compression hip screw (Smith & Nephew) in 8 cases. All surgical procedures were performed by an orthopedic surgeon who specialized in pediatric orthopedics.
Preoperative anteroposterior and cross-table or frog leg lateral radiographs were assessed to classify fractures according to the Delbet system described by Colonna19: type I (transepiphyseal), type II (transcervical), type III (basocervical), and type IV (intertrochanteric). Postoperative radiographs were assessed to establish evidence of osteonecrosis of the femoral head. The extent of osteonecrosis was graded with the Ratliff criteria1: type I (total involvement of the femoral head), type II (partial involvement of the femoral head), and type III (involvement of only the femoral neck, from the physis to the fracture line). Immediate postoperative anteroposterior and frog leg lateral radiographs were used to analyze the accuracy of fracture reduction. Reduction quality was graded based on the degree of residual angulation in any plane and the residual displacement of the fracture fragments. The reduction was graded as anatomic when there was no step-off and no angulation, as partial anatomic when there was less than 5 mm cortical step-off and less than 10° angulation, and as nonanatomic when there was cortical displacement of greater than 5 mm and angulation of more than 10°. The most recent radiographs were used to assess fracture healing, which was determined by bone bridging the fracture site and the presence of sequelae, including malunion, premature physeal closure, and nonunion. A pediatric orthopedic fellow (J.D.S.) who was not involved in the clinical care of the patients assessed all radiographs.
The severity of complications was defined with a modified Clavien-Dindo20 complication scale that was validated for hip surgery.21 According to this system, complications are graded from I to V in severity, based on long-term morbidity and treatment required to manage the complication. Grade I complications, which are considered minor, require no treatment and have no clinical relevance. Higher grades of severity involve deviation from the expected postoperative course. Grade II complications, also considered minor, require outpatient treatment. Grade III complications are considered major and require invasive surgical or radiologic intervention or unplanned hospital admission. Grade IV complications, also considered major, are potentially life-threatening, have high long-term morbidity, or require total hip replacement. Grade V complications, also considered major, result in death.
Final outcomes in pain, movement, activity, and radiographic appearance were assessed by a professional research assistant (M.K.H.) according to the Ratliff criteria.1 Clinical and radiographic outcomes were graded as good (no pain, full motion, normal activities or avoidance of sports, and normal radiographs or minimal deformity), fair (occasional pain, hip motion of at least 50%, normal activities or avoidance of sports, and severe deformity of the neck with osteonecrosis), and poor (disabling pain, less than 50% hip motion, restricted activities, and severe deformity of the femoral neck, with degenerative arthritis, arthrodesis, or total hip replacement).
Statistical analysis was conducted with SAS version 9.3 software (SAS Institute Inc, Cary, North Carolina). The 2 treatment groups were compared with Fisher's exact test and Wilcoxon rank-sum test for categorical and continuous outcome variables, respectively.
There was no evidence of osteonecrosis in hips treated with ORIF, whereas 50% of hips treated with CRIF had osteonecrosis. This difference was marginally statistically significant (P=.051). Of the 8 cases of osteonecrosis, 5 were Ratliff type I, 2 were type II, and 1 was type III (Figure 1).
Anteroposterior radiograph of the left hip of a 9-year-old boy who had a fully displaced Delbet type III fracture after a fall (case 12) (A). Anteroposterior radiograph of the left hip 1 year after closed reduction and internal fixation showing osteonecrosis of the femoral head (B). Anteroposterior radiograph of the left hip after a secondary procedure. The patient underwent a surgical dislocation approach with a trapdoor procedure for bone grafting of the femoral head (C).
On radiographs obtained immediately postoperatively, 5 of 6 hips (83%) in the ORIF group and 2 of 16 hips (13%) in the CRIF group had anatomic reduction (P=.009). All fractures treated with ORIF and 7 of 16 (44%) fractures treated with CRIF healed anatomically (P=.046). Of the remaining cases treated with CRIF, 2 (13%) were complicated by malunion, 2 (13%) by nonunion, and 5 (31%) by residual deformity (Figure 2).
A 17-year-old boy had a fully displaced Delbet type II fracture after a fall (case 10) (A). Radiograph obtained immediately after closed reduction and internal fixation showing nonanatomic reduction (B). At 3.5 months after closed reduction and internal fixation, the fracture began to drift into varus alignment (C). Radiograph obtained 1 year after closed reduction and internal fixation and 8 months after valgus intertrochanteric osteotomy (D).
The overall complication rate was higher in the CRIF group (P=.009), with complications developing in 15 hips (94%) treated with CRIF vs only 2 hips (33%) treated with ORIF. The ORIF group and the CRIF group had a mean of 0.3 and 2.9 complications per case, respectively (P=.001). Hips treated with ORIF were marginally less likely (P=.054) to have minor complications (Clavien-Dindo grades I and II) than those treated with CRIF. Additionally, major complications (Clavien-Dindo grades IIII and IV) occurred in significantly fewer (P=.015) hips after ORIF than after CRIF. Excluding implant removal, 1 or more secondary procedures were required in 8 cases (50%) treated with CRIF and in none of the cases treated with ORIF (P=.051).
There was no significant difference (P=.477) between groups, according to the Ratliff assessment of final clinical and radiographic outcomes. All hips treated with ORIF achieved a good result, according to the Ratliff criteria (Figure 3). Eleven hips (69%) treated with CRIF were rated as good, 1 (6%) as fair, and 4 (25%) as poor. When the Ratliff assessment was analyzed for osteonecrosis, all 14 (100%) hips without osteonecrosis and only 3 hips (38%) with osteonecrosis achieved a good result (P=.002).
A 10-year-old boy had a fully displaced Delbet type II fracture that occurred while he was skiing (case 2) (A). Radiograph obtained 1 year after open reduction and internal fixation showing the patient before hardware removal (B). A good result with anatomic union is seen at final follow-up 2 years after surgery (C).
Management of pediatric femoral neck fractures is controversial, particularly with regard to associated risk factors and strategies to prevent osteonecrosis. Many possible complications have been reported after fracture treatment. The most severe is osteonecrosis, which often has a poor prognosis. Although multiple studies have compared surgical and nonsurgical management of femoral neck fractures, far fewer have evaluated the clinical and radiographic results of ORIF vs CRIF. The current study compared open vs closed reduction of fully displaced pediatric femoral fractures with regard to the following: (1) the rate of osteonecrosis of the femoral head; (2) the quality of radiographic reduction and fracture healing; (3) the total occurrence and severity of complications; and (4) the final clinical and radiographic outcomes, as assessed by the Ratliff criteria.
The authors observed a lower incidence of osteonecrosis after ORIF (0%) than after CRIF (50%). In a comparable review of displaced femoral neck fractures, Dhammi et al22 reported no osteonecrosis in patients treated with ORIF and a 24% incidence of osteonecrosis after CRIF. Likewise, Song8 found a lower rate of osteonecrosis after ORIF (0%) vs CRIF (17%) for the treatment of displaced fractures. In contrast, a recent systematic review and meta-analysis by Yeranosian et al9 evaluated 30 pediatric femoral neck studies (935 patients) and concluded that patients undergoing CRIF had a lower rate of osteonecrosis than those undergoing ORIF. However, these findings were based on the integrated results of all fracture types, and the authors reported that patients with more severe fractures usually underwent ORIF. Additionally, although Yeranosian et al9 reported a fourfold increase in the rate of osteonecrosis when definitive treatment was delayed for more than 24 hours, the current findings showed a lower incidence of osteonecrosis after ORIF, despite delayed treatment, compared with the CRIF group. A possible reason for this finding could be that decompression of the increased intracapsular pressure secondary to the hematoma could lead to a tamponade effect and impaired blood flow to the femoral head. Whether capsular decompression protects against osteonecrosis is unclear.3,6,9,14,15
In this study, anatomic reduction was achieved in 83% of patients treated with ORIF vs only 13% of patients treated with CRIF. These findings are in line with a previous retrospective study by Song8 that compared open vs closed reduction of totally and partially displaced femoral neck fractures in children. Of the fractures treated with ORIF, 93% were anatomically reduced compared with only 25% in the CRIF group. Anatomic reduction is preferred in the treatment of pediatric femoral neck fractures because it has been associated with a decreased risk of osteonecrosis.13
In the authors' series, 1 or more complications occurred in 17 of 22 hips (77%). There was a greater occurrence of both minor (grades I and II) and major (grades III and IV) complications after CRIF than after ORIF. Femoral neck fractures in children are known to have a high rate of complications.4,5,8,11,14,23–25 As shown in the current series, the method of treatment may affect the occurrence of complications. Song8 also reported a higher rate of complications in patients treated with CRIF than in those treated with ORIF. In Song's8 series, 7 complications occurred in 5 patients treated with CRIF. Moreover, ORIF has been reported to reduce the rate of femoral deformity, including coxa vara, after neck fractures.8,9,13,15 There was no evidence of femoral deformity after ORIF in the current series.
Although the Ratliff clinical and radiographic outcomes did not statistically differ between groups, the current findings suggest that good results can be expected after ORIF in pediatric femoral neck fractures. In addition, the authors found that a good outcome was more likely in the absence of osteonecrosis. This finding is in line with previous studies that reported osteonecrosis as the most important single cause of poor outcomes.2,4,6,7,9,13
The authors acknowledge several limitations of this study. First, the small sample size prevented more robust statistical analysis, including multivariate analysis to identify factors predictive of osteonecrosis. However, the relative rarity of this pediatric injury precluded obtaining a large number of subjects at a single institution. The authors believe that restricting their sample to severely displaced cases reduced the confounding bias and allowed for more comparable treatment groups. Second, this study had inherent biases and limitations because of the retrospective nature of this cohort. There was also the potential for assessment bias because radiographic assessment and grading of complications and clinical outcomes were not performed in duplicate. Finally, follow-up was relatively short. However, complications, including osteonecrosis, typically develop within a year after pediatric femoral neck fractures.1,2,7
The authors reported a higher quality of fracture reduction, with fewer postoperative complications, including osteonecrosis of the femoral head, in fully displaced pediatric femoral neck fractures treated with ORIF vs CRIF. However, the findings were based on a limited sample size. The authors believe that further large multicenter studies are needed to identify factors predictive of osteonecrosis and to determine whether pediatric femoral neck fractures should be treated with ORIF or CRIF.
- Ratliff AH. Fractures of the neck of the femur in children. J Bone Joint Surg Br. 1962; 44:528–542.
- Canale ST, Bourland WL. Fracture of the neck and intertrochanteric region of the femur in children. J Bone Joint Surg Am. 1977; 59(4):431–443.
- Cheng JC, Tang N. Decompression and stable internal fixation of femoral neck fractures in children can affect the outcome. J Pediatr Orthop. 1999; 19(3):338–343. doi:10.1097/01241398-199905000-00010 [CrossRef]
- Forlin E, Guille JT, Kumar SJ, Rhee KJ. Complications associated with fracture of the neck of the femur in children. J Pediatr Orthop. 1992; 12(4):503–509. doi:10.1097/01241398-199207000-00017 [CrossRef]
- Hughes LO, Beaty JH. Fractures of the head and neck of the femur in children. J Bone Joint Surg Am. 1994; 76(2):283–292.
- Moon ES, Mehlman CT. Risk factors for avascular necrosis after femoral neck fractures in children: 25 Cincinnati cases and meta-analysis of 360 cases. J Orthop Trauma. 2006; 20(5):323–329. doi:10.1097/00005131-200605000-00005 [CrossRef]
- Ng GP, Cole WG. Effect of early hip decompression on the frequency of avascular necrosis in children with fractures of the neck of the femur. Injury. 1996; 27(6):419–421. doi:10.1016/0020-1383(96)00025-3 [CrossRef]
- Song KS. Displaced fracture of the femoral neck in children: open versus closed reduction. J Bone Joint Surg Br. 2010; 92(8):1148–1151. doi:10.1302/0301-620X.92B8.24482 [CrossRef]
- Yeranosian M, Horneff JG, Baldwin K, Hosalkar HS. Factors affecting the outcome of fractures of the femoral neck in children and adolescents: a systematic review. Bone Joint J. 2013; 95-B(1):135–142. doi:10.1302/0301-620X.95B1.30161 [CrossRef]
- Canale ST, Tolo VT. Fractures of the femur in children. Instr Course Lect. 1995; 44:255–273.
- Heiser JM, Oppenheim WL. Fractures of the hip in children: a review of forty cases. Clin Orthop Relat Res. 1980; (149):177–184.
- Beaty JH. Fractures of the hip in children. Orthop Clin North Am. 2006; 37(2):223–232. doi:10.1016/j.ocl.2005.11.003 [CrossRef]
- Shrader MW, Jacofsky DJ, Stans AA, Shaughnessy WJ, Haidukewych GJ. Femoral neck fractures in pediatric patients: 30 years experience at a level 1 trauma center. Clin Orthop Relat Res. 2007; 454:169–173. doi:10.1097/01.blo.0000238794.82466.3d [CrossRef]
- Flynn JM, Wong KL, Yeh GL, Meyer JS, Davidson RS. Displaced fractures of the hip in children: management by early operation and immobilisation in a hip spica cast. J Bone Joint Surg Br. 2002; 84(1):108–112. doi:10.1302/0301-620X.84B1.11972 [CrossRef]
- Song KS, Kim YS, Sohn SW, Ogden JA. Arthrotomy and open reduction of the displaced fracture of the femoral neck in children. J Pediatr Orthop B. 2001; 10(3):205–210.
- Togrul E, Bayram H, Gulsen M, Kalaci A, Ozbarlas S. Fractures of the femoral neck in children: long-term follow-up in 62 hip fractures. Injury. 2005; 36(1):123–130. doi:10.1016/j.injury.2004.04.010 [CrossRef]
- Dendane MA, Amrani A, El Alami ZF, El Medhi T, Gourinda H. Displaced femoral neck fractures in children: are complications predictable?Orthop Traumatol Surg Res. 2010; 96(2):161–165. doi:10.1016/j.otsr.2009.11.007 [CrossRef]
- Jones RW. Fractures of the neck of the femur. Br J Surg. 1936; 23(92):787–808. doi:10.1002/bjs.1800239213 [CrossRef]
- Colonna PC. Fracture of the neck of the femur in children. Am J Surg. 1929; 6(6):793–797. doi:10.1016/S0002-9610(29)90726-1 [CrossRef]
- Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg. 2004; 240(2):205–213. doi:10.1097/01.sla.0000133083.54934.ae [CrossRef]
- Sink EL, Leunig M, Zaltz I, Gilbert JC, Clohisy JAcademic Network for Conservational Hip Outcomes Research Group. Reliability of a complication classification system for orthopaedic surgery. Clin Orthop Relat Res. 2012; 470(8):2220–2226. doi:10.1007/s11999-012-2343-2 [CrossRef]
- Dhammi IK, Singh S, Jain AK. Displaced femoral neck fracture in children and adolescents: closed versus open reduction. A preliminary study. J Orthop Sci. 2005; 10(2):173–179. doi:10.1007/s00776-004-0875-3 [CrossRef]
- Morsy HA. Complications of fracture of the neck of the femur in children: a long-term follow-up study. Injury. 2001; 32(1):45–51. doi:10.1016/S0020-1383(00)00109-1 [CrossRef]
- Canale ST. Fractures of the hip in children and adolescents. Orthop Clin North Am. 1990; 21(2):341–352.
- Leung PC, Lam SF. Long-term follow-up of children with femoral neck fractures. J Bone Joint Surg Br. 1986; 68(4):537–540.
|Case No./Sex/Age, y||Delbet Type||Treatment Type||Time to Treatment, h||Mechanism||Reduction Quality||Follow-up, y||Minor Complications (Grades I and II)||Major Complications (Grades III and IV)||Secondary Procedures||Ratliff Result|
|3/M/12.6||III||ORIF||>24||Fall from height||Anatomic||1||Complex regional pain disorder||Good|
|5/M/12.2||III||ORIF||>24||Fall from scooter||Anatomic||1||Good|
|6/M/12.1||III||ORIF||>24||Fall from height||Nonanatomic||8.7||Good|
|7/M/17||IV||CRIF||<24||Fall from height||Partial anatomic||1.7||Prolonged antibiotic course||Deep infection||Irrigation and debridement||Good|
|8/M/7.6||III||CRIF||<24||Snowboarding||Nonanatomic||1||Osteonecrosis type II||Good|
|9/F/9.5||II||CRIF||>24||Fall from height||Nonanatomic||1||Premature physeal closure+residual deformity||Osteonecrosis type III||Good|
|10/M/17.4||II||CRIF||<24||Fall||Nonanatomic||1||Leg length discrepancy||Nonunion, osteonecrosis type II, osteoarthritis||Valgus intertrochanteric osteotomy||Good|
|11/M/16.2||III||CRIF||<24||Assault||Nonanatomic||1.9||Superficial infection, prolonged course of antibiotics||Loss of fixation twice, nonunion, deep infection, osteomyelitis, peri-implant fracture, osteonecrosis type I, osteoarthritis, chronic pain||Revision of fixation, surgical hip dislocation+grafting to femoral head, irrigation and debridement, total hip replacement||Poor|
|12/M/9||III||CRIF||<24||Fall||Partial anatomic||4.1||Leg length discrepancy||Osteonecrosis type I, osteoarthritis, subluxation||Surgical hip dislocation+bone grafting, recommended total hip replacement||Poor|
|13/M/10.5||II||CRIF||<24||Traffic accident||Anatomic||1||Peri-implant fracture||ORIF||Good|
|15/F/8.4||III||CRIF||<24||Traffic accident as a pedestrian||Partial anatomic||1.3||Leg length discrepancy, premature physeal closure||Good|
|16/F/14.2||II||CRIF||<24||Horseback riding||Nonanatomic||5.8||Leg length discrepancy, piriformis syndrome with bursitis||Osteonecrosis type I, osteoarthritis||Fibular autograft, total hip replacement||Poor|
|17/F/13.8||II||CRIF||<24||Ice skating||Nonanatomic||1.6||Premature physeal closure+residual deformity||Good|
|18/M/12.5||II||CRIF||<24||All-terrain vehicle rollover||Nonanatomic||2.3||Premature physeal closure+residual deformity||Osteonecrosis type I, osteoarthritis||Fibular autograft||Poor|
|20/F/10.3||II||CRIF||<24||Horseback riding||Partial anatomic||1.6||Premature physeal closure+residual deformity||Good|
|21/F/10.5||II||CRIF||<24||Fall from height||Nonanatomic||2.1||Premature physeal closure+residual deformity||Good|
|22/F/6.9||II||CRIF||<24||Fall from height||Partial anatomic||1.1||Leg length discrepancy||Loss of fixation, malunion, osteonecrosis type I||Anterior cheilectomy||Fair|
Preoperative Variables by Treatment Group
|Variable||Open Reduction and Internal Fixation (n=6)||Closed Reduction and Internal Fixation (n=16)||P|
| Male||6 (100%)||7 (44%)|
| Female||0||9 (56%)|
|Age, median (interquartile range), y||12.2 (10.7–12.6)||10.7 (9.3–14)||.856b|
|Body mass index, median (interquartile range), kg/m2||22 (13.6–26.7)||19.2 (16.8–20.8)||.657b|
|Time to treatment, No.||.046a|
| ≤24 h||3 (50%)||15 (94%)|
| >24 h||3 (50%)||1 (6%)|
|Fracture type, No.||.195a|
| II||2 (33%)||11 (69%)|
| III||4 (67%)||4 (25%)|
| IV||0||1 (6%)|
|Follow-up, median (interquartile range), y||1 (1–1.5)||1.6 (1.1–2.2)||.112b|