Introduction

Supracondylar humeral fractures are the most common type of elbow fracture in children [1–3] and usually occur in children aged 5–10 [4]. These fractures are virtually unique to the paediatric population due to the remodelling of the distal humerus that occurs in this age group, resulting in a thinner supracondylar region, with a more slender cortex [5]. The vast majority of these fractures are an extension-type injury, resulting from a fall onto an outstretched hand. Flexion-type supracondylar fractures are much rarer and result from direct trauma with the elbow joint in flexion, or a fall onto a flexed elbow.

Displaced supracondylar fractures pose a significant risk of neurovascular injury due to the complex local anatomy, and for this reason, such fractures are traditionally considered a surgical emergency. More recently, the need for emergency surgery has been debated, especially amongst children with closed fractures and an intact neurovascular supply. A lack of consensus exists in such patients with some suggesting delayed intervention is appropriate [6–10], while others continue to favour emergency intervention [11–14].

This study was sought to analyse the timing of surgery and the influence of surgical delay on patient outcome in a large UK tertiary provider of children’s orthopaedic care.

Patients and methods

A retrospective case note analysis was performed amongst patients with Gartland Grades 2 and 3 supracondylar fractures managed at our UK tertiary paediatric trauma centre. Exclusion criteria included cases treated nonoperatively, those with an absent radial pulse on Doppler examination, open fractures and incomplete records. Patients were classified according to the Gartland classification [15] based upon their radiograph at presentation. The time of admission to the initial consulting hospital was considered as a proxy for the time of injury. The time of commencement of anaesthesia was considered as the time of surgical intervention. This allowed median time-to-surgery to be determined.

All patients were treated with closed reduction and with either crossed-pin or lateral-only pin fixation with Kirschner wires. Open reduction was performed if adequate closed reduction was not achieved.

Outcomes were analysed for the rate of iatrogenic nerve injury, pin-site infection, need for open reduction and development of compartment syndrome. These perioperative complications were identified from the inpatient records and outpatient follow-up clinic letters.

Statistical analysis was conducted using a nonparametric test of association (Fisher’s exact test), with a p value <0.05 considered significant.

Results

A total of 137 patients were presented with a displaced supracondylar fracture during the study period. Twenty-two patients were excluded [16 patients were managed nonoperatively (all Gartland 2a fractures)], one child sustained an open fracture, median time-to-surgery was unable to be obtained in four patients, and one child had an absent pulse on Doppler examination. There were no cases of compartment syndrome.

Thus, 115 patients were included. Patient demographics are summarised in Table 1. Patients were divided into those treated with early surgical fixation (<12 h) or delayed surgery (>12 h), and the results were analysed using Fisher’s exact test. The median time-to-surgery was 15 h and 30 min (range 2 h, 45 min–62 h, 50 min).

Table 1 Patient demographics
Full size table

In the early treatment group, three children developed a superficial pin-site infection, all of which grew Staphylococcus aureus and resolved with simple oral antibiotics. Closed reduction failed in four children who all proceeded to open reduction under the same anaesthetic; in three, soft tissue interposition between the bone ends was noted. Five children sustained an iatrogenic ulnar nerve injury, all of which had recovered by date of discharge. Two children required reoperation: one child had fracture rotation following initial reduction and fixation, and the other child was felt to have unsatisfactory position at 1-week follow-up.

Table 2 Complications
Full size table

In the delayed treatment group, one patient developed a superficial infection which grew pseudomonas. This was treated with wound debridement, 1 week of IV antibiotics followed by oral antibiotics for 8 weeks. Four children treated in a delayed manner required open fracture reduction, all of which were undertaken under the same anaesthetic. One child had a rotational deformity in the sagittal plane which could not be reduced by closed means, and one was noted to have the medial edge of the humerus button-holing through the medial joint capsule preventing reduction. Two children sustained an iatrogenic ulnar nerve injury, two sustained an iatrogenic nerve injury to both the median and ulnar nerves, and three children sustained iatrogenic ulnar nerve injuries. All cases of neuropraxia had resolved by discharge. Two children required reoperation: one child was noted to have fracture rotation at initial follow-up and the other child required a corrective de-rotational osteotomy for a rotational malunion.

Bivariate analysis of our data using Fisher’s exact t test revealed no statistically significant difference between early and delayed surgery groups with regard to infection rates (p = 0.10), need for open reduction (p = 0.43), iatrogenic nerve injury occurrence (p = 0.52) or need for reoperation (p = 0.60) (Table 2).

We further analysed our data by looking specifically at Grade 2 fractures and Grade 3 fractures independently. We again found no statistically significant differences between early and delayed surgery groups except that Grade 3 fractures treated within 12 h were more likely to have an infection (p = 0.03).

Discussion

The need for urgency in treating displaced supracondylar fractures is a controversial topic amongst paediatric orthopaedic surgeons. Our experience as a large UK paediatric trauma centre is that delayed surgery appears to be a safe option, although it is vitally important that each case is treated based on its individual merits, with neurovascular deficit, compartment syndrome and open fracture remaining indications for emergency intervention. In this regard, it is important that children are thoroughly assessed with respect to these indications for urgent intervention; assessment can prove challenging due to pain and the young age of these patients along with unfamiliarity of the often junior medical staff assessing these patients. To help ensure that adequate assessment is undertaken, we have previously reported on the development of an assessment proforma to ensure children with indications for emergency intervention are identified [16].

A significant proportion of displaced supracondylar fractures treated at our institution are referred from other hospitals in the area as they offered a limited paediatric orthopaedic service, with children often presenting to our institution some hours after the time of injury. This has resulted in our centre having a large number of patients treated in a delayed manner. Our results indicate that transfer to a specialist paediatric centre, often with consequent surgical delay, is a safe management option and helps ensure that surgery is undertaken by surgeons with suitable expertise in the management of these injuries.

Overall infection rate in our study was 3.5 %, comparable to other studies in the literature, [17] with no significant difference between those treated early or with delayed intervention. Likewise, rates of iatrogenic nerve injury were similar between the groups. Majority of patients in our series were treated with a crossed-pin configuration, and so it is perhaps unsurprising that the ulnar nerve was the most commonly injured nerve due to its vulnerability during placement of the medial pin.

Our results have actually shown higher proportions of complications in the early surgery group compared with the delayed surgery group, although this was not statistically significant. There are likely to be a number of factors explaining these findings: firstly, in our study, we had a higher percentage of Gartland Grade 3 fractures treated within 12 h; nevertheless, a large number of Grade 3 injuries were treated in a delayed manner, and whenever we independently analysed Grade 3 fractures, there was no significant difference in complication rates except that Grade 3 fractures treated within 12 h were more likely to develop a superficial infection (p = 0.03). Another explanatory factor was thought to be the fact that the fractures treated in an early manner were more likely to be treated by trainees, and also out-of-hours. It is well known that out-of-hours operating is known to heighten complication rates in all manner of surgical diseases due to factors such as unfamiliarity of the theatre staff and surgeon fatigue [18, 19].

In addition to the higher percentage of Gartland Grade 3 fractures treated within 12 h, other limitations of our study included the fact that a number of different surgeons were included, yet this arguably raises the generalisability of the findings, as well as the limitations of any retrospective review.

Conclusions

In conclusion, in our experience, delayed surgery appears to offer a safe management approach to the treatment of displaced supracondylar fractures. However, it is vitally important that the individual nature of each case is closely taken into account, and in this regard, it is crucial that a comprehensive evaluation is undertaken at presentation to ensure that children with indications for emergency surgery are identified. Safe delay therefore allows a safe window of transfer for children to travel to specialist centres for treatment and negates the obligation to carry out these procedures at night when complication rates are higher. The indications for emergent surgery remain vascular deficit, compartment syndrome and open fracture.