Plate fixation versus flexible intramedullary nails for management of closed femoral shaft fractures in the pediatric population: A systematic review and meta-analysis of the adverse outcomes

Purpose: Fractures of the femoral diaphysis are associated with a risk of morbidity in children. Various fixation methods have been developed, but with only limited evidence to support their use. This systematic review assesses the evidence regarding clinical outcomes of closed femoral diaphyseal fractures in children treated with plate fixation or flexible intramedullary nails. Methods: A PROSPERO-registered, PRISMA-compliant systematic review and meta-analysis were conducted. MEDLINE, Embase, and Web of Science (WoS) databases were searched from inception to February 2023. Inclusion criteria included clinical studies reporting adverse outcomes following surgical treatment of pediatric closed femoral diaphyseal fractures using plate fixation and flexible intramedullary nails. The ROBINS-I and RoB 2 tools evaluated the risk of bias. Results: Thirteen papers (2 prospective randomized controlled trials and 11 retrospective cohorts) reported 805 closed diaphyseal femoral fractures in 801 children (559 males, 242 females). There were 360 plate fixations and 445 flexible intramedullary nails. Two cases of osteomyelitis and one nonunion were reported. Meta-analysis showed that plate fixation had a lower risk of soft tissue infection (relative risk 0.26 (95% confidence interval 0.07–0.92)). There was no difference in the following outcomes: malunion (relative risk 0.68 (95% confidence interval 0.32–1.44)); unplanned reoperation (relative risk 0.59 (95% confidence interval 0.31–1.14)), and leg-length difference (relative risk 1.58 (95% confidence interval 0.66–3.77)). The risk of bias was high in all studies. Conclusions: An analysis of 805 fractures with minimal differences in meta-analyses is considered high quality even when the quality of the evidence is low. The findings are limited by important flaws in the methodology in the published literature. Well-designed multicentre prospective studies using standardized core outcomes are required to advise treatment recommendations. Level of evidence: III.


Introduction
Fractures of the femoral diaphysis (shaft) are the most common major pediatric injury treated by orthopedic surgeons. 1,2Their reported incidence ranges from 5.82 to 16.4 per 100,000 children and fractures occur most frequently in the summer. 3,4They are the leading cause for hospitalization in pediatric trauma patients and a significant cause of morbidity. 5These injuries often require prolonged immobilization and/or surgery which contributes to significant psychological stresses on the patients and their parents/ guardian.By 14 years of age, males are 4.7 times more likely than females to have had a femoral shaft fracture. 6espite the impact of this injury, definitive treatment in patients aged 4-12 years and <50 kg continues to be a controversial area.Historically, there has been a preference toward a particular "in vogue" surgical treatment (flexible intramedullary nails (FIN) in the 1990s or plate fixation (PF) in the 2000s), demonstrated in scoping reviews. 5,7In the last 15 years, FINs have become popular again, despite the lack of high-quality evidence to support their use. 1,8mportantly, the lack of evidence has resulted in ongoing regional practice variation in the management of this injury. 3his systematic review aimed to answer the question: in children with closed diaphyseal femoral fractures, what is the risk of negative outcomes following fixation with plates (PF) versus flexible intramedullary nails (FIN)?Secondary aims were to inform practice surrounding the consenting of patients and research in the future.

Methods
We used the methodology outlined in the Cochrane Handbook for Systematic Review of Interventions. 9 The review is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement 10 and the meta-analysis of observational studies in epidemiology (MOOSE) guidelines. 11An a priori protocol was registered prospectively on the PROSPERO international register of systematic reviews (CRD42020193281). 12EDLINE, Embase, and Web of Science (WoS) were searched from the outset to February 2023 using MeSH terms and free-text strategies (Figure 1, Supplemental Appendix 1).Two authors (A.S. and W.B.) searched independently, with medical librarian support.No language limits were applied, and the reference lists of included articles were hand-searched to identify additional publications.The gray literature was searched using Google Scholar.To ensure the best possible direct comparison between the two groups, only studies reporting adverse outcomes following PF and FIN in children with a closed femoral fracture were included.The fracture was the unit of analysis.
The co-primary outcomes were malunion and nonunion.The definition of a malunion was a deformity of 15-30° in the sagittal plane and 10° in the frontal plane requiring operative intervention.Nonunion was an arrest or absence of healing on serial radiographs, 6 months after primary management.The use of the term (malunion/nonunion) within the included studies was deemed adequate.Secondary outcomes were infection (osteomyelitis (OM) and/or soft tissue infection), unplanned reoperation, leg-length difference (>1 cm), time to union (weeks), length of stay (days), operative duration (minutes), and blood loss (ml).
For patient demographics, treatments and outcome data, descriptive analyses were performed.When appropriate, individual study incidences of adverse events were pooled, and meta-analysis was performed using RevMan5 (The Cochrane Collaboration, Copenhagen, Denmark).Relative risks (RRs) and 95% confidence intervals (CIs) were calculated using the Cochran-Mantel-Haenszel test.A random-effects model was used because of the anti cipated study heterogeneity.Forest plots were used to display results.Publication bias was viewed with funnel plots. 13,14The I 2 statistic measured heterogeneity. 15
The risk estimates from all 13 papers were amalgamated for the meta-analysis.A 74% reduction in the risk of soft tissue infection (RR 0.26 (95% CI 0.07-0.92))was  As assessed by the ROBINS-I tool for non-randomized studies of intervention (Table 5), 28 and because of the retrospective observational nature of these studies, there were few pre-published protocols.All studies were at moderate to serious risk of bias due also to confounding between treatment groups, attrition, and bias in reporting.The RoB 2 tool was used for randomized studies (Table 6), 29 and revealed a high risk of bias due to underpowering, lack of prospective statistical analysis plan and missing outcome data.Most studies also lacked homogeneous definitions of outcomes (Tables 5 and 6).

Discussion
The anatomy of a child's femur is fundamentally different to that of an adult.The presence of a thicker and more vascular periosteum means that the bone has a high healing and remodeling capacity.The pediatric femur is able to tolerate up to 15° of malrotation and 25° of angulation in any plane. 30,31In these patients, various operative strategies have been advocated to reduce the adverse consequences be they physical, social, psychological, and/or financial asso ciated with lengthy periods of immobilization. 1 As younger children (≤ 4 years) have better healing potential, non-surgical treatment (traction or spica cast) is first-line. 32n older children (5-12 years), surgical treatment is preferred to aid an earlier return to weightbearing/activity and to prevent prolonged periods of absence from school. 1,2hildren > 12 years and/or weighing > 50 kg are treated with rigid intramedullary nails to balance mechanical stresses at the fracture site with diminishing remodeling potential. 2,8owever, definitive management of closed femoral shaft fractures in patients aged 4-12 years and <50 kg continues to be a controversial area with marked regional differences.It is argued that, biomechanically, length stable (transverse) fractures would be more suitable for FIN 18 whereas, PF may be superior for unstable (long oblique, comminuted) fractures. 16These variations in fracture patterns and the suitability of PF versus FIN in achieving healing without adverse outcomes continue to be debated.Notably, the American Academy of Orthopedic Surgeons (AAOS), 32 the United Kingdom's National Institute of Clinical Excellent (NICE) 33 and a recent Cochrane review 34 have all emphasized the poor quality of available evidence when making management recommendations.
To address this ongoing debate, this systematic review's methodology focuses on patients within a specific mean age range (5-12) and weight (<50 kg).It compares adverse outcomes following operative management (PF vs FIN) of closed pediatric femoral shaft fractures.Direct comparison meta-analysis of 805 fractures indicates a substantially reduced risk of soft tissue infection in the PF group, with no difference in malunion, unplanned reoperation and leg-length difference.
The literature already suggests that PF may confer advantages through earlier full weightbearing and time to fracture union however, the markedly reduced risk of infection in this group is an interesting finding.This could be secondary to advances in PF techniques which include stab incisions for submuscular plating compared to the large open approach. 2It could also be due to the fact that while FIN requires a smaller wound and offers a shorter hospital stay, 2 it has a higher incidence of metalwork prominence which may irritate/ulcerate the soft tissue leading to a wound infection.This is of value to clinicians when organizing their follow-up protocols.While the marked cost difference of implants in favor of FIN (FIN-£138 for 2; PF (locking compression plate, LCP)-£260) may lead to its preferential usage in some units, this benefit must be carefully balanced against the cost implications of unplanned hospital admission and/or reoperation, which continues to rise within healthcare systems. 35he disadvantages of both techniques are also wellreported: with PF, longer operative time and increased blood loss, difficulty in plate removal, re-fracture following removal; and with FIN, longer time to weightbearing, higher rates of malunion and leg-length difference. 1,2,5Our descriptive analysis clearly demonstrates these disadvantages (Tables 4a/b) and provides pooled incidence rates to inform the consenting process.The difference in the incidence of unplanned reoperation rates (PF (3.7%) and FIN (5.8%)), which is not clinically significant, may favor PF as deformity and nail migration necessitating early removal is secondary to relative stability conferred by FIN (Figure 3(a)).Whereas, the difference in leg-lengths favoring FIN (PF (3%) and FIN (2.6%)) may be secondary to the anatomical reduction in PF leading to overgrowth.In our review, time to union is longer in the PF group (PF 10.6 vs FIN 8.9 weeks) which is not in keeping with published literature.This could be because studies reporting union were only referring to radiological parameters and primary healing in PF does not produce as much callus as a secondary process in FIN and may be harder to see.Furthermore, time to union is often influenced by the timing of follow-up appointments so this difference is unlikely to be clinically significant.We hypothesize that the lack of difference in some outcomes between the groups means that either there really is no difference or the patient numbers are not large enough to reliably show a difference.

Strengths and limitations
The ROBINS-I and RoB 2 tools assessed the risk of bias for non-randomized and randomized interventional study designs, respectively.Using the ROBINS-I tool (Table 5), the quality overall of the research assessing surgical management of closed pediatric femoral diaphyseal fractures was very low.Reporting of definitions was inconsistent, and most studies were methodologically flawed, therefore, quantifying true risk in pooled analyses was challenging.This introduced a bias to real-world risk estimates and small sample sizes created wide confidence intervals in meta-analyses.As the risk of bias, predominantly from confounding within the observational, retrospective, non-randomized included studies was high, it was not possible to make robust treatment recommendations.Surgeon preference was an additional significant confounding factor.However, relative to research in this specific area, a pooled analysis of 805 fractures with minimal variance may be considered high quality.

Conclusion
This analysis suggests that a pediatric femoral diaphyseal fracture treated with PF is associated with a low risk of soft tissue infection.However, there is no difference between PF and FIN groups with regards to malunion, unplanned reoperation and leg-length difference.Our work further informs consenting practice and highlights that the operative treatment of pediatric femoral shaft fracture remains an area of clinical equipoise.Importantly, this paper goes beyond just "advocating" for prospective trials, it provides a robust appraisal of the available literature and the   statistical foundation to justify such a study.It also provides trialists with the necessary incidence information to design adequately powered prospective studies.

Table 1 .
Characteristics of included studies.

Table 2 .
Injury characteristics and management techniques.

Table 5 .
26udy definitions of outcome measure and risk of bias assessment (non-randomized studies).Ahmed et al.24NU not defined, term used as diagnosis, MU not defined, term used as diagnosis.SI term used but not defined, OM term not used, UR term not used, LLD term used but not defined.Caglar et al.26NU not defined, term used as diagnosis, MU not defined, term used as diagnosis.SI term used but not defined, OM term used but not defined, LLD term used but not defined, UR term not mentioned but assumed in hardware failure.

Table 6 .
27udy definitions of outcome measure and risk of bias assessment (randomized studies).James et al.25NU defined as lack of bridging callus on less than three cortices, MU term used but not defined.SI defined as per CDC SSI criteria published 2017, OM term not used: substituted for deep bone infection, LLD term used but not defined, UR defined as operation excluding elective implant removal.Hayat et al.27NU/MU not defined, terms bundled together as a diagnosis.SI term used, not defined, OM term not used, LLD/UR terms not used but used within the Flynn criteria.Bias domains: (1) Bias arising from randomization, (2) Bias due to deviation from intended intervention, (3) Bias due to missing outcome data, (4) Bias in measurement of the outcome,(5)Bias in selection of the reported result.