Neonatal upper limb fractures – a narrative overview of the literature

Clavicle fracture is the most frequent birth related fracture with an incidence of 0.5-11.2/1000 live births [9‐14, 32, 42, 43] while the humerus fracture is the second most common of the long bones with an incidence of 0.04-0.2/1000 live births [19, 22, 25, 44, 45].

There are several risk factors described for these injuries: advanced maternal age, maternal overweight or obesity, short stature, increasing birth weight, malpresentation, Type 1 diabetes, gestational diabetes, shoulder dystocia, vacuum delivery, use of oxytocin (need for labor – high risk pregnancies) and pain relief during labor [1, 9, 10, 12‐14]. Of the risk factors described, macrosomia seems to be present in 20-50% of fetuses [9, 10, 13‐15]. It’s also interesting to see that high birth weight is an important feature closely related with other risk factors such as pre-pregnancy obesity, diabetes, induction of labor, vacuum-assisted delivery and shoulder dystocia [17, 25, 32, 46‐54]. Shoulder dystocia is also an important risk factor for humerus fracture and like high birth weight, may also increase the likelihood of concomitant brachial plexus injuries [11, 43, 50, 55]. In patients with perinatal brachial plexus injury, studies report humerus fracture rates ranging from 2 to 11% [56, 57].

The association with instrumented vaginal deliveries is frequent in the literature and the majority of these are performed due to maternal distress or suspected fetal asphyxia or distress [11, 34]. Högberg states that assisted vaginal delivery by vacuum extraction was associated with clavicle fracture in 21.8% of the cases, while shoulder dystocia was present in only 4.3% of clavicle fractures [10]. The most frequent mechanism of injury during delivery occurs when maneuvers are performed to reduce the chest circumference in shoulder dystocia or by compression of the fetal anterior shoulder of the fetus being compressed against the maternal pubic symphysis leading to fracture [10]. Nevertheless, Kekki in his study states that although rare, there is an incidence of 0.46/1000 live births of clavicle fractures associated with cesarean sections, 66% of which occurred in unplanned procedures [11]. This association is also observed in fractures of the humerus, which although historically associated with breech maneuvers during vaginal delivery, there has been a relevant increase associated to caesarean deliveries that accompanies the growing popularity of this technique. Although caesarean deliveries avoid the risk of head entrapment when the baby is in breech position, the breech maneuvers necessary to extract the baby are similar to those performed in vaginal deliveries and may lead to fractures of the long bones [19, 58]. Despite of all this risk factors, approximately one quarter of the patients with clavicle fracture have no identified risk factor [11]. According to Madsen, within humeral fractures, transverse midshaft are the most common, followed by proximal humeral fractures and distal physeal fractures [22, 45]. Regarding epiphyseal separations of the distal humerus, these are very rare and are defined as epiphyseal fractures (Salter-Harris I or II), for which we must have a high diagnostic suspicion, since they often go undetected [19, 22, 45, 59].

The clinical presentation of upper limb fractures in neonates is variable according to the fracture pattern and deviation. Displaced fractures are usually identified immediately after delivery due to their greater clinical exuberance. Local deformity, crepitus, oedema, pseudoparalysis of the affected limb and crying aroused by passive mobility of the limb may be present. Non-displaced fractures may only be diagnosed at a later stage (days or weeks), when local swelling due to bone callus formation is perceptible upon inspection or palpation. In the case of humerus fracture the Moro reflex is consistently asymmetric, whereas in non-displaced clavicle fracture there may not be an obvious asymmetry. Fracture of the clavicle is usually located in the middle third, while fracture of the humerus also occurs most frequently in the middle third, followed by the proximal third. Regarding the distal third of the humerus, this should be given special attention, as even after the clinical examination and initial complementary diagnostic study with X-rays, the diagnosis of tranphyseal fracture is not always unequivocal, justifying high suspicion index and other complementary means. A careful inspection of the whole limb is mandatory. The fracture pattern is usually simple (not comminuted) and with a transverse or oblique fracture line. Due to the possibility of associated perinatal brachial plexus injury mainly to clavicle and proximal humerus fracture, it is also important to document the motor function of the elbow, wrist and hand (spontaneous and after stimulation) and to maintain surveillance after the expected period of consolidation, in order to assess the evolution of the active mobility and function of the affected limb (Table 1).

When in the presence of a clinical suspicion of fracture, the diagnosis of neonatal humeral diaphyseal or clavicle fracture is usually simple and almost always made by means of a radiograph, allowing differential diagnosis with other birth-related traumatic pathologies, such as perinatal brachial plexus injury or glenohumeral dislocation (Figs. 1 and 2) [60, 61]. Clavicle and humerus fractures may be associated with other traumatic bone or neurological injuries, thus the radiographic study should be amplified and not only focused on the suspected location of the fracture (consider requesting a chest and full upper limb X-ray). Neonatal humeral shaft fractures may present with neuropraxia of the radial nerve, an injury that, although usually belated diagnosed, should be identified and distinguished from perinatal brachial plexus injury. In these injuries, patients do not show active extension of the wrist or fingers but preserve an active abduction of the shoulder and a flexion and extension of the elbow that is similar to the contra-lateral side, although this is often only fully identified after immobilization is removed and resolution of the pseudoparalysis of the limb by the fracture (Table 2). This differential diagnosis is important as these cases have a good prognosis with 72% of patients making a full recovery after 2 months and 100% after 6 months [62, 63]. In addition, closed injuries of the radial nerve and median nerve in children are generally associated with favorable spontaneous recoveries, while lesions of the ulnar nerve are associated with a worse prognosis and have a more uncertain prognosis [64]. For this reason, in the presence of a radial nerve injury in the context of a humeral shaft fracture, the posture to adopt should be very conservative and expectant, since a complete resolution of the clinical picture is to be expected, in a self-limiting manner.

With this concept in mind most of the birth-related fractures are diagnosed on the first day of life or in the early neonatal period (between day 1 and day 7). Högberg in his study states that only 5.8% of clavicle fractures are diagnosed after this period (4.7% between day 8 and 28, and 1.1% after day 28) [10]. Still it is known that some fractures are undiagnosed, and there are several cases of fractures that are diagnosed incidentally by X-ray [9]. It is also important to note that transphyseal fractures of the distal humerus may be particularly difficult to identify and therefore justify a more cautious approach by requesting other complementary means of diagnosis, whenever there is suspicion of fracture [34]. In these cases the diagnosis is usually late, sometimes only made after accidental observation of callus formation on the X-ray or palpation of a rigid mass in the elbow region [45, 61, 65]. This results from the fact that the distal and proximal physis of the humerus are not yet ossified at birth and are therefore not clearly visualized in the X-rays [22, 45, 66, 67]. Jacobsen reinforces this idea, stating that most of his patients with distal humeral transphyseal injury were diagnosed between 9 and 30 days of life [59].

Furthermore, these injuries as well as proximal humeral physeal fractures, when observed by inexperienced doctors, are often confused with joint dislocations, and may lead to inadequate and inconsequent attempts of joint reduction [22, 65, 68]. For this reason, it is necessary to keep these entities in mind and carry out other complementary studies such as ultrasound, MRI or even arthrography, in some specific cases (Table 3) [65, 69‐90]. In his article, Sherr-Lurrie mentions that of the 5 patients with distal or proximal epiphyseal fracture of the humerus identified, these injuries were not diagnosed by X-ray, but by ultrasound and in one of the cases MRI was even necessary for further clarification, in a fracture of the proximal physis of the humerus [22].

Ultrasound, although operator-dependent, is an excellent option as it is readily accessible, non-invasive, radiation-free, inexpensive and does not require sedation or general anaesthesia, unlike MRI or arthrography [22, 65, 69, 70, 78, 91]. These advantages are also particularly important in the context of patients hospitalised in NICUs where the fragility of their condition or the fact that they are less comfortable to mobilise makes the portability and non-invasiveness of the ultrasound even more relevant [22]. Ultrasound can also be used for the diagnosis of humeral shaft or clavicle fractures and is a diagnostic alternative that should be taken into account for different types of fractures. One thing to be pointed out is that this examination implies an anatomical pressure on the injured segment and a painful stimulus for the baby and that this situation also disturbs the execution and quality of the exam and therefore it must be done by experienced doctors who reduce as much as possible the time the child is in pain caused by the ultrasound probe. Whenever available, a high-resolution ultrasound should be used [71, 72].

Neonatal fractures of the humerus and clavicle are usually treated conservatively. There are different options, and immobilization is usually done using clothing, traction, splints or casts [19, 22, 45, 61, 66]. Reduction maneuvers are not usually necessary, although in cases of humeral fractures with severe displacement, it is possible to realign the fracture during immobilization [22].

One option for greater comfort for the baby, and to limit the mobility of the affected limb, involves a soft immobilization, wearing a long-sleeved shirt and placing the affected limb with the sleeve attached to the chest with the elbow at 90º flexion. This immobilization, although less relevant in cases of clavicle fracture, it is of greater interest in mid-shaft humerus fractures due to the more significant prevention of secondary rotational displacement. Sherr-Lurie reports a similar approach for proximal humerus fractures (simple swaddling) while for the fractures of the shaft and distal humerus, the preferred method was a closed gentle manipulation and an above-elbow plaster cast for 2 weeks, with the upper limb held against the body by the baby's shirt [22]. The author also states that neonates do not show any apparent discomfort with the immobilization performed [22].

Nevertheless it is likely that in the first 7 days, the patient may feel more discomfort/pain, in which case oral or rectal analgesic medication (paracetamol) may be administered. The presence of a visible and painless swelling around 7-10 days usually indicates the presence of an adequate consolidation process, without the need for additional X-rays or greater exposure to ionizing radiation. The beginning of active mobility of the limb appears around 2-3 weeks, corresponding to bone healing [22, 87]. Parents are instructed of the natural evolution and it is explained to them that they should take care of a gentle mobilization when it is necessary to dress or wash their children.

If the expected evolution is not observed, the radiographs may be repeated at around 4 weeks and a high index of suspicion should be maintained for other associated lesions, such as perinatal brachial plexus injury. Neurovascular injuries due to neonate humerus diaphyseal fractures are rare [92].

Parents should be warned that although the X-rays may show angular displacement, this is not expected to cause any functional impairment of the limb or lead to any future limitation, as it is expected that the child's growth will allow for complete bone remodelling and realignment.

In some cases of distal transphyseal fractures of the humerus, particularly displaced and unstable ones, surgical treatment options have been described, preferably by closed reduction and percutaneous fixation with K-wires [20, 34, 87, 91, 93].

Neonatal fractures of the upper limb have a good prognosis and no long-term sequelae are expected. Non-operative treatment (short-term splinting or immobilization or observation) is the option in the vast majority of the fractures and is associated with excellent results, with good healing, full range of motion, adequate remodeling without obvious deformity, neurologic impairment or functional implications [19, 22, 45, 59, 61].

In specific and rare cases, such as transphyseal fractures of the humerus, the complications described in the literature include reduced range of motion, cubitus varus/valgus or the need for secondary surgical procedures. Transphyseal fractures of the distal humerus are those that can most commonly associate these coronal deviations, as the distal physis is the one that contributes least to humeral growth and one of those with the least capacity for remodelling in the whole skeleton [59, 65, 76, 87]. Within these complications, decreased range of motion and the cubitus varus are the most frequent [65, 76, 80, 87].

Nevertheless, the literature mentions that even in neonatal transphyseal fractures and regardless of the treatment option, the prognosis is good and the long-term results (although scarcely reported) are equally favourable [59, 65, 75, 76, 93, 94].