Article Text
PDF
Abstract
Objectives To determine and identify the characteristics and circumstances of femur fractures in ambulatory young children.
Design and setting Retrospective review of 203 ambulatory children, between 1 and 5 years old, presenting with femur fractures to an urban paediatric hospital over a 10-year period. χ2 And Student's t test were employed for statistical analysis.
Results The mean age was 36.6 months, with 155 (76.2%) being male. The most frequent mechanism of injury was fall from a height (n=62, 30.5%). The highest number of injuries occurred in 2–3-year-olds. The most common history in 1–2-year-olds was stumbling on/over something causing a fall. For 4–5 year olds it was road traffic accidents. Other additional physical findings were infrequent (14.3%) and not suspicious of inflicted injury. Child protective services concluded three of the cases to be likely non-accidental, and four cases were inconclusive but requiring close follow-up. Of these seven children, six occurred in 1–2-year-olds. No distinguishing feature was noted in fracture type or location.
Conclusions Femur fractures can occur with low velocity injury whether from a short fall or twisting/stumbling injury in young healthy ambulatory children.
- paediatrics
- accidental
- non accidental injury
- paediatric injury
- fractures and dislocations
Statistics from Altmetric.com
Introduction
In non-ambulatory children, the occurrence of a femur fracture appropriately raises suspicion of possible non-accidental injury and requires careful evaluation. Femur fractures have been documented in 12%–29% of physically abused children.1 In 50%–60% of these cases, the femur fracture was the only injury noted. Distinguishing injuries due to accidents from those due to child abuse in young children can be difficult because of the inability of the young child to provide a history.
The incidence of abuse in children presenting with fractures increases with decreasing age of the child and appears to be related to whether a child is ambulatory,2–,6 Schwend et al found that the failure of a child to have reached walking age was the strongest predictor of abuse in infants and young children with femur fractures.7 They found a 42% risk of abuse in non-walking children compared with 2.6% in walking children. In a systematic review of patterns of skeletal fractures in child abuse, Kemp et al, report the probability of abuse in femoral fractures to be between 0.28 (CI 0.15 to 0.44) and 0.43 (CI 0.32 to 0.54) depending on the definition of abuse used and the developmental stage of the child.8 Developmental level was noted to be an important discriminator. Other studies have shown percentages for non-accidental causation of femur fracture ranging from 30% to 36% for children aged 3 years and younger.5 ,9 This increases to a rate between 39% and 93% for children less than 12 months of age.10 ,11 In a Canadian population, Hui et al reported femur fractures to be associated with non-accidental injury in 11% of cases overall, increasing to 17% in patients under the age of 12 months.12 These figures are similar to figures from two large studies from the USA.13 ,14
It is our hypothesis that ambulatory toddlers may sustain femur fractures during falls typical of usual childhood activity. Although the possibility of non-accidental injury should be carefully evaluated in a young child with a femur fracture, it is important to understand the characteristics of accidental femur fractures to prevent misdiagnosis of abuse.
Methods
The clinical records of all children older than 1 year and less than 5 years of age who were admitted to The Hospital for Sick Children with femur fracture from 1995 through 2004 were reviewed. Non-ambulatory children and children with bone disease were excluded. Patients were identified through a health records search using the International Classification of Diseases 9th revision, and International Classification of Diseases 10th revision diagnostic codes, for a diagnosis of femur fracture. The Suspected Child Abuse and Neglect (SCAN) multidisciplinary team database and orthopaedic database were utilised to ensure no cases were missed. The project was approved by the hospital's research ethics board. Individual patient or family consent was not required for this retrospective protocol.
Information obtained from the records included age, gender, relevant developmental history with regards to onset of walking, and history of injury including place, mechanism of injury, and whether the traumatic event was witnessed. Symptoms present when medical attention was sought, and the type and location of the fracture, were recorded. We extracted from the chart whether the injury was determined to be consistent with the history, results of skeletal survey if done, whether child protective services were notified, their adjudication and the need for further follow-up. Radiographs were not reviewed.
Each child had a full history and physical examination from the admitting service (usually orthopaedics), and each child had also been evaluated by a paediatrician (seen in our emergency department) or general practitioner (if referred). Consultation with our SCAN team was initiated by either the orthopaedic service, the emergency department or the outside physician if presentation was felt to be possibly consistent with inflicted injury. The threshold for such consultation has been characteristically low.
χ2 And Student's t test were employed for statistical analysis where applicable. The 95% CI, (p<0.05) was used to identify statistical significance.
Results
We reviewed 217 charts. Fourteen cases were excluded either because history was inadequate to ascertain whether the child was walking independently at the time of injury, or the fracture was felt to be pathological. The mean age at presentation of the remaining 203 children was 36.6 months (median 35, range 15–59, see table 1). There were 155 males (76.3%, CI 0.7 to 0.82 for the proportion of males >50%) The age at onset of walking was only specifically documented in 18/203 (8.9%).
Cases identified for inclusion
The most frequent mechanisms of injury, according to the provided history were fall from a height (n=62, 30.5%), fall when stumbled on/over something (n=44, 21.7%), and fall while running (n=24, 11.8%, see Table 2). Males were more likely to have fractures from falls. The majority of falls from a height were less than 2 feet (57.5%) with only 20% higher than 4 feet. In 20 cases (9.9%), the fracture occurred when a person and/or object fell on the child. Motor vehicle collision was the causative event in 44 cases (21.7%), including 15 cases (7.4%) where the child was a pedestrian. In five cases (2.5%), the leg was grabbed by a caretaker to prevent a fall or during play activity. In four cases (2%) there was no history provided to explain the injury.
Reported mechanism of injury
The SCAN team was consulted in 44 cases (21.7%, age range 15–51 months, average age 27.6 months). In 23 of these 44 cases, child protective services were not notified as the multidisciplinary team's evaluation concluded that the fracture was accidental. The remaining 21 cases (47.7%) were referred to child protective services for further assessment. Child protective services had also been notified by the referring hospital in three additional cases, but these three injuries were felt to be non-abusive by our orthopaedics service, and the SCAN team was not notified (see table 3).
Evaluations for possible child abuse
The data were analysed by age groups: 1–2 years old (n=39), 2.1–3 (n=72), 3.1–4 (n=50) and 4.1–5 (n=42). Falls were the most common mode of injury in the age groups between 1 and 4 years old cumulatively (p<0.001). In the 4.1–5 year-old group, road traffic accidents were the most common cause of fracture in this age group. The most frequent cause of fracture was a fall from a height for children aged 2.1–3 and 3.1–4 years old. The most common fracture cause in the 1–2 year-old group was stumbling on/over something causing a fall. Injuries reported to have occurred at home represented 60.2% of fractures with 39.8% occurring in a public place. The events leading to 95 fractures (46.8%) were witnessed.
Clinical symptoms included irritability (92.6%), refusal to weight bear (73.4%), swelling of the leg (38.4%), and a ‘crack’ heard by an adult (3.9%). The symptoms were not clearly documented in all charts. In one case, the femur fracture was noted by an ultrasound technician while doing an abdominal ultrasound for presumed abdominal pain. In four cases, medical attention was delayed because the child reportedly slept prior to attending hospital. Physical findings (14.3%) included ecchymosis of the body, abrasion, swelling, periocular ecchymosis, an erythematous toe, scar, epistaxis, laceration and lip injury. There was no indication that any of the additional physical findings were suspicious or serious, apart from those in a motor vehicle accident, although 11 of the 43 cases of motor vehicle accidents had no other additional physical injury.
A previous fracture was reported in seven cases (3.4%). In 176 cases (86.7%), the past medical history was documented as non-contributory or negative. In 20 cases (9.9%), past medical history was not documented in the chart.
The type of femur fracture was most frequently oblique/spiral (70.4%) or transverse (19.7%). Two were buckle fractures (1%). No metaphyseal fracture was noted. Other types of femur fractures (combinations, eg, oblique and transverse or intertrochanteric neck of femur) were associated with motor vehicle accidents. Sixty-five percent of all femur fractures involved the left leg. Ninety-four percent were diaphyseal.
Of the 24 children reviewed by child protective services, three cases were adjudicated as abuse and four cases as inconclusive. The other 17 were adjudicated as accidental. In each of the three cases of abuse, the child was taken into the care of The Child Protection Services. Of the seven cases (four boys, three girls), six occurred in the 1–2- year age group (p<0.0001), all between 18 and 23 months of age. The remaining child was in the 4.1–5-year-old age group who gave a history of inflicted injury, causing the fracture. Fractures were oblique in two of the children apprehended, and spiral in the other child apprehended, and the four inconclusive cases. Skeletal surveys were performed in six of these seven cases, all of which were normal. No skeletal survey was done in the child in the 4.1–5-year-old age group. Two children had previous femur fractures, 3 and 5 months prior to presentation, one from the apprehended group, and one from the inconclusive group. In the accidental group, five children had previous fractures, but 9.9% of charts had no documentation if a previous fracture had occurred or not. One child who was apprehended had a sibling who died from a cot death 10 years previously. The family of one child apprehended was previously known to child protective services. Six of the seven children presented to hospital within 5 h of symptoms becoming apparent. The remaining child presented more than 12 h after symptoms. In two of the children apprehended, the presenting history changed significantly on subsequent questioning: who was present with the child at the time of the fall (in one), and falsely claiming review by a primary care physician following injury (in the other). In addition, rarely visiting the child while in hospital was noticed in one case apprehended, and not at all in the other.
Discussion
Our study is one of the largest to our knowledge, that specifically looks at femur fractures in healthy ambulatory children in a Canadian setting. Our findings are in keeping with other studies in the literature. We found, allowing for limitations of the study, that approximately 1.5% of young ambulatory children with femur fractures may have had an inflicted injury (CI 0.003 to 0.04). An additional 2% an inflicted injury (CI 0.005 to 0.04), and 96.5% did not appear to have an inflicted injury (CI 0.9 to 0.98).
Leventhal et al describe that abuse was diagnosed in 60% of infants less than 1 year of age, and in no child over 23 months of age. Thomas et al reported femur fractures in ambulatory toddlers to frequently result from injuries sustained when the running child trips or falls during normal childhood activities.5 They looked at 25 femur fractures of which 14 were considered accidental, nine secondary to abuse, and two cases inconclusive. Sixty percent of femur fractures in children younger than 1 year of age were due to abuse, but only 20% were from abuse in the 1–2 year-old age group. Similarly, Blackmore et al demonstrated that ambulatory young children can generate sufficient force to fracture the femur during typical childhood falls from a low height or from running.6 They reported a low rate of abuse with femoral shaft fractures: one out of 42 fractures, between 1 and 5 years of age.
We found a higher number of boys presenting with femur fractures, with a ratio of greater than three boys to one girl. This is consistent with the general finding that boys predominate in childhood injuries.7 ,15–,18 The largest proportion of femur fractures occurred in the 2.1–3 year-old age group, which is developmental when children are adventurous, challenging and exploring their surroundings. Children in this age group have typically started climbing, which may explain the observation in this study that the majority of injuries were from falls from a height. Gross and Stranger state that as children acquire walking ability, they also acquire the capacity to encounter many other potential mechanisms of injury.19
Farrell et al20 in their review of symptoms and time to medical care in children with accidental extremity fractures (of which 8.7% were femur fractures), showed that although some children did not manifest all expected responses, no child with an accidental fracture was asymptomatic.
In keeping with other reports, the majority of the falls in this study were reported to occur from low heights: less than 2 feet (57.5%), between 2 and 4 feet (22.5%), and greater than 4 feet (20%).6 ,12 ,13 ,18 As shown in table 1, mechanisms of injury included falls while running or walking. In total, falls represented 64% of the cases. Haney et al reported that impacted transverse fractures of the distal femoral metadiaphysis may occur as a result of an accidental short fall in young children.21
Paediatricians can play a key role in educating and providing support for parents and children in injury prevention.22 ,23 Injuries, such as those due to falls from windows, seatbelt use in cars, stair gates and objects not well secured to walls, thus falling on children, are just a few examples of cases noted in this study, which may be preventable. Fractures may be indicative of a high-risk environment in which trauma is likely to occur and, therefore, adequate supervision of children's activities may also play a role.
The majority of fractures were oblique/spiral in nature (70.5%) while two cases had a buckle fracture. Previous studies have indicated that fracture type alone cannot distinguish abusive from non-abusive trauma8 ,10 In a systematic review of skeletal fractures in child abuse, Kemp et al found no difference in the distribution of transverse, spiral or oblique fractures between the abused and non-abused children. One study found that spiral fracture was the most common abusive femoral fracture under 15 months old, but not significantly so, after15 months of age3 Metaphyseal fractures have been reported in a greater proportion of abused children.
A limitation of this study is that not every child was reviewed by the SCAN team or the child protection team. This may lead to ascertainment bias. As some of our cases occurred over a decade ago, it is possible that thresholds for reporting to child protective services may have been different, relative to current thresholds (likely less involvement). Referral to SCAN should not be seen as a stigma or statement of high suspicion of abuse, but a request for consultation, and represents good practice. Each child had a full history and physical examination from the admitting service (usually orthopaedics), and each child had also been evaluated by a paediatrician (seen in our emergency department) or general practitioner (if referred).
However, in a paper by Ravichandiran et al24 they noted that approximately 20% of abusive fractures were missed during the initial medical visit to a primary care or a non-paediatric emergency department with an extremity fracture, especially in boys.
Recognition of paediatric fractures and screening for possible abuse-related fractures from accidental trauma, can be difficult, but very important in emergency department and orthopaedic clinics.25 ,26 This highlights the importance of ensuring ongoing education/quality improvement programmes of emergency staff/orthopaediatric staff/primary care staff and paediatricians on child protection issues.
Another limitation of our study is its retrospective design. In only 18/203 (9%) cases, did the doctor document the age that the child started to walk. We had hypothesised that a child who had just started walking may be at increased risk of fracture, but this could not be assessed. A study by Pandya et al,27 in their review of child abuse and orthopaedic injury patterns, found that beyond 18 months of age, presentation with a long bone fracture was more likely to be related to accidental trauma than child abuse. The difficulty, however, in establishing the diagnosis of abuse, with certainty, is reflected in many studies;6 ,7 ,23 our data suggests that femur fractures can occur with an accidental low velocity injury, including short falls or tripping injury, in young, healthy, ambulatory children between 1 and 5 years of age. In our cases, factors that raised concern for possible abuse, resulting in the child either being apprehended or requiring child protective services to follow-up (inconclusive cases), were inconsistent histories, delay in presentation, previous fractures in a short period of time, lack of explanation, allegation by one child of inflicted injury, prior involvement with child protective services, and absence of parent figure during the child's admission to hospital. This highlights that a key portion of the medical evaluation is actually in the medical history and observation of family dynamics.
Prospective studies are required to detail more clearly the characteristics of accidental versus non-accidental femur fractures, to further improve our diagnostic sensitivity and specificity.
Acknowledgments
The authors thank Derek Stephens for assistance with the statistical analysis for this study.
References
Footnotes
-
Contributors LC wrote the manuscript with input from all authors.
-
Competing interests None.
-
Ethics approval The Hospital for Sick Children research Ethics Board.
-
Provenance and peer review Not commissioned; externally peer reviewed.