Data collection from the included patients showed that the mean age of children with PHF was 8.6 years old with 56% of these fractures occurring in females and 44% in males. The age of the cohort ranged from one month to 15 years old.

The most common cause of injury was falls (70%, n=29). There were two sports-related injuries (5%), both of which were in male patients. There was one fracture from a non-accidental injury (2.4%) and one suspected pathological fracture in a patient with global developmental delay and suspected disuse osteoporosis (2.4%). It should be noted that 20% of the cases did not have a recorded mechanism of injury.

Choice of treatment was non-surgical in 85% of cases (n=35). These cases were managed in either a poly-sling or collar and cuff. Only one patient had an elastic stable intramedullary nail (2.4%) and two had an ORIF with K wire fixations (5%). There were three cases where fy the treatment method (7%) was not specified which were excluded from the analysis.

The overall mean time to the radiological union from the date of injury was 21.8 days with a range of nine to 78 days. The time to discharge due to satisfactory clinical assessment was a mean of 36.2 days from the date of injury ranging from five to 161 days. The earliest time to discharge was five days, at which time, the patient was clinically assessed and it was deemed appropriate to continue with non-operative management. The same patient who was from overseas subsequently travelled home after this assessment (Table 2).

Non-operatively managed cases had a mean time to discharge of 45.51 days ranging from five to 161 days. The operatively managed cases had a longer mean discharge time of 56.67 days which ranged from 22 to 113 days.

The proximal humerus forms from three ossification centres. It is completely cartilaginous until four to six months of age when the primary ossification centre forms. The greater and lesser tuberosities form at three years and five years respectively, with the fusion of the nuclei at around seven years old. The growth plate of the humerus usually closes at 17 to 18 years of age [1, 16, 21].

The glenohumeral joint is encompassed in a joint capsule. The attachments of this capsule may help explain the Salter-Harris type II configuration of fractures of the proximal humerus in skeletally immature patients. Namely, the medial capsular structures extend from the glenoid and labrum superiorly onto the medial aspect of the metaphysis inferiorly. Additionally, the lateral capsular attachment to the anatomical neck of the humerus also exerts a potential force through the physis in trauma [5].

Structures that are closely related to the proximal humerus can complicate the treatment of the fracture. The axillary nerve wraps around the humeral neck and may be injured. As the tendon of the long head of biceps passes down along the bicipital groove, it may become incarcerated in the fracture site [7, 8, 11, 12, 19, 22]. An open approach may be necessary if surgical reduction of such fractures is attempted. The rotator cuff, pectoralis major and latissimus dorsi muscles act as distracting forces due to their insertions and direction of action [5, 22, 23].

Therefore, anatomical structures should be considered when reducing fractures and immobilising the arm, as this may affect the functional outcome and complicate the reduction of the fracture.

Complications associated with PHF include a varus or valgus arm, a shortened arm and non-union. The goal of treating PHF is to restore the range of motion, full function, arm length and alleviate pain. Within our cohort, there was one incidence of complication during non-operative treatment. The patient had a subsequent fall after immobilisation which displaced the fracture, leading to a longer healing time. The patient was discharged from clinic 21 weeks following the subsequent injury with no report of pain, sensation change or functional deficits.

Although the techniques used in surgical intervention vary, there is a general consensus regarding the indications leading to operative management. These include an angulation of greater than 50 degrees in 10 year-olds or younger, 20 degrees or translation of greater than 50% in 11 year-olds or above, and failure of closed reduction of the fracture within the parameters mentioned above [1, 22, 24].

Most professionals agree that the three main factors which govern the decision to operate children with PHF are age, angulation and skeletal maturity of the patient [5, 7]. This has given rise to three distinct treatment groups within the paediatric population, namely patients of ages under 10, older than 13 and between 10 and 13.

In the first group (under 10 years of age), a non-operative approach is preferable (see Fig. 1 for non-operative approach X-Rays). As seen in our case series, even a severely displaced fracture of 21mm of translation and 49 degrees of AP angulation was deemed acceptable and treated with a collar and cuff. The outcome was successful with the restoration of full limb function. Although satisfactory clinical progress took longer than the average in our cohort, the risk of undergoing unnecessary surgery was avoided.

Fig. (1). X-Ray images showing the progression from injury (a) to 7 days after (b) a non-operative approach.

The second group (greater than 13 years of age) are more skeletally mature. Naturally, the remodelling potential is less than in younger patients. Studies show that surgical intervention is more common in this group [4, 22] typically with successful outcomes and in certain cases, a worse outcome without surgery [6, 8, 19]. Reasons for this include the fact that older children usually present with high energy impact injuries potentially leading to extensively displaced and angulated fractures. More aggressive management may be needed to reduce and stabilise the fracture and encourage union.

Notably, there were only three surgically treated patients in our cohort, of which two had buried K wires and one patient underwent stabilisation with ESIN (Fig. 2). The preference of buried over-exposed K wires is due to the increased risk of infection due to exposed wires [5, 25]. Literature suggests that ESINs are used more frequently in severely displaced fractures and favoured over K wires due to their ability to allow earlier mobilisation post-operatively [1, 13]. However, the utilisation of ESIN requires extensive experience to yield a positive outcome. Therefore, many surgeons prefer to use K wires [1].

There remains debate regarding the optimal management of the third group (10 to 13 years of age). These patients may undergo both non-operative and operative treatment methods. This is because the biological age of the patient does not always truly reflect skeletal age and maturity. Therefore, more flexibility in management decisions is advisable [1, 5, 22]. In our cohort of 10 to 13 year-olds, the number of non-surgically treated patients was higher than surgically treated patients.

Non-operative treatment was favoured in this case series, in agreement with current literature. The remodelling potential of children is extraordinary, providing a basis for excellent outcomes without surgical intervention. However, as previously discussed, the decision to operate should include a consideration of age, fracture angulation and skeletal maturity.

Fig. (2). X-Ray images showing the progressive treatment using K-wires of a skeletally immature patient. Injury (a), K-wire insertion (b), outcome after removal (c), Radiograph at final follow-up (d).

The focus of this study was the time to discharge from orthopaedic services due to clinical union. In 73% (n=30) of cases, the discharge time and radiological/clinical union date were the same. In cases when the discharge date was later than the radiological/clinical union, this was due to subsequent injuries, or other fractures. The majority of patients in the cohort would be deemed to have achieved a radiological and clinical union on the same date due to the time intervals between outpatient appointments. Furthermore, with non-displaced fractures in younger children, a repeat radiograph was not carried out as it was clinically unnecessary and did not justify the radiation exposure.