An early referral to a specialist brachial plexus multidisciplinary clinic is encouraged (Borschel and Clarke 2009). The diagnosis of obstetrical brachial plexus palsy is primarily one of clinical acumen; not infrequently a child may need to return for repeated assessment if compliance was lacking at the initial consultation. The clinician must exclude diagnoses that may potentially mimic obstetrical brachial plexus palsy such as underlying intracranial pathology (including cerebral palsy and brain tumors), a radial nerve palsy, or a fracture or dislocation involving the shoulder girdle or upper extremity (“pseudoparalysis ”) (Alsubhi et al. 2011). Plain radiographs should be requested if there is suspicion of an associated bony injury or joint dislocation. Soft tissue contractures tend not to develop until some months after birth; thus any limitation in the passive range of motion of the upper extremity during neonatal assessment should raise the suspicion of underlying musculoskeletal pathology. For example, injury to the proximal humeral epiphyses should be suspected in a child when both active and passive motion of the arm are equally restricted; a plain radiograph should confirm the diagnosis (Adler and Patterson 1967). A thorough birth history including predisposing risk factors and postnatal complications (including respiratory distress, associated fractures, torticollis, or Horner’s syndrome) is taken as detailed in chapter “Supracondylar Humerus Fracture.​”

Clinical assessment is best performed with the infant placed on a clean sheet on a padded mat at floor level (or low examination bench) as this facilitates efficient examination in a safe environment. Clothing above the child’s waist should be removed and particular care taken to note the head posture; children with obstetrical brachial plexus palsy tend to look to the contralateral side. The presence of torticollis should be noted; this may result from birth trauma or as a secondary phenomenon due to positioning of the child (invariably, but incorrectly, nursed with the affected extremity uppermost). The lower extremities must also be examined in order to exclude a hemiparesis or neonatal tetraplegia.

The posture of the upper extremity gives invaluable information as to the level of the plexus lesion. Shoulder adduction and internal rotation with the elbow in extension, the forearm pronated, and wrist/fingers held in flexion (the “waiter’s tip” position) is highly suggestive of an Erb-Duchenne upper plexus (C5, C6 ± C7) injury. Recent evidence indicates that the diagnostic specificity of the “waiter’s tip” posture in infants is less than was originally assumed, as the only absolute requirement to produce a “waiter’s tip” posture is an intact T1 nerve root (Fattah et al. 2012). A flail limb, with the absence of any motor activity, implies a total plexus palsy (C5-C8, T1). Less common phenotypes include a Klumpke lower plexus (C8, T1) palsy (a flaccid hand in an otherwise functioning upper extremity), an intermediate (C7, C8, T1) palsy (with an abducted shoulder, flexed elbow, and flaccid hand), and an isolated C7 palsy (with the elbow held in a flexed posture).

Physical signs that would support the diagnosis of an upper plexus injury include winging of the scapula (due to paralysis of the long thoracic nerve of Bell; root value C5-C7), although assessment of the scapulae in neonates is not possible. In addition, paradoxical breathing between the chest and abdomen could be the result of an associated phrenic nerve injury (root value C3-C5; although the phrenic nerve typically contributes to C5) though this is impossible to detect clinically. An ipsilateral Horner syndrome (characterized by ptosis, miosis, anhidrosis, and enophthalmos) can be seen on routine examination and indicates disruption of T1 proximal to the sympathetic rami communicantes (i.e., a preganglionic avulsion injury) (Fig. 2).

A thorough motor examination of the upper extremity is central to confirming the diagnosis, prognosis, and subsequent treatment plan and probably represents the most challenging element of the examination. Traditionally, the Medical Research Council muscle grading scale (Table 2) has been used (Medical Research Council 1943); however, its use in infants is limited on account of their inability to fully cooperate with the examination, and thus the authors favor the validated Active Movement Scale developed at The Hospital for Sick Children (Table 3) (Clarke and Curtis 1995; Curtis et al. 2002). Fifteen cardinal motions of the upper extremity are assessed, including those of the shoulder (abduction, flexion, internal rotation, and external rotation), elbow (flexion and extension), forearm (pronation and supination), and wrist, fingers, and thumb (flexion and extension). The Active Movement Scale measures motion within the range of joint motion, not voluntary strength. It incorporates gravity as a standard; movements are tested both against gravity and with gravity eliminated in order to document early recovery with precision. The Mallet scale (Fig. 3), which was developed to provide a quantifiable assessment of shoulder function in obstetrical brachial plexus palsy, is also widely used although it demands a degree of patient compliance that negates its routine use in children under the age 3 years (Mallet 1972).

It is generally accepted that in those children who meet the criteria for surgical intervention, treatment should not be delayed, as this is likely to prejudice the maximal benefit that could otherwise be obtained. The challenge facing brachial plexus surgeons has been to evolve the ability to statistically prognosticate which children are most likely to benefit from surgical intervention versus those who are best managed expectantly. Clearly, surgical intervention should only be warranted if the outcome after intervention is likely to result in enhanced long-term function compared to nonoperative management; such a judgment requires a sound understanding of the natural history of obstetrical brachial plexus palsy. Identification is straightforward in the scenario of an infant with a flail limb and an associated Horner syndrome; equally, surgery is contraindicated in a child with mild paralysis who recovers substantial motor function within the first month as near complete functional recovery would be anticipated (Al-Qattan et al. 2000). Clinical evidence of an isolated T1 avulsion (i.e., a limp hand resting in the intrinsic minus position) is a reliable indicator for operative intervention within the first 3 months. Despite accepting that the majority of children with obstetrical brachial plexus palsy will not require surgical exploration, there is currently no universal consensus as to how to ascertain which children will – with reasonable certainty – require surgery.

Tassin noted that complete recovery of shoulder function at the age of 5 years (Grade V in the Mallet classification system) was only seen in those children with contraction of biceps and deltoid muscles in the first month and normal contraction of both muscles by the age of 3 months (Tassin 1983). Gilbert and coworkers found testing the deltoid problematic in neonates (due to difficulty in isolating its action from pectoralis major activity) and therefore generated three modified criteria for surgical intervention: a flail arm with Horner syndrome, a complete C5–C6 palsy without muscle contractility by 3 months and a negative EMG (thus suggestive of a root avulsion), and a C5–C6 palsy with no biceps activity at 3 months (Gilbert et al. 1988). Failure of spontaneous return of elbow flexion at 3 months of age is now the most widely accepted criteria for surgical intervention in infants with an upper plexus lesion (Gilbert and Whitaker 1991; Gilbert 1995; Eng et al. 1996; Gilbert et al. 2006).

Waters confirmed the findings of Tassin and Gilbert that infants in whom biceps function recovered by 3 months of age progressed to achieve normal neurologic function, while those whose biceps recovery occurred late (i.e., from 4 to 5 months of age) had a worse long-term recovery according to the Mallet classification (Waters 1999). Those infants who had microsurgical brachial plexus reconstruction for absent biceps recovery by 6 months of life had superior function to those who experienced late spontaneous biceps recovery, but their function was not as good as those in whom elbow flexion recovered early.

However, Michelow et al. determined that overreliance on a single functional parameter alone at the age of 3 months would incorrectly predict a poor recovery in 12.8 % of patients, additionally implying that some children were being reconstructed unnecessarily. This false-positive rate could be reduced to 5.2 % if elbow flexion and elbow, wrist, thumb, and finger extension were incorporated into a combined test score. The utility of adopting multiple muscle assessments as an indicator for surgical intervention is becoming increasingly accepted and should lead to a reduction in unnecessary surgery, although a more conservative threshold may result in some children who would benefit from surgery ultimately being managed nonoperatively (Laurent et al. 1993).

The importance of repeat assessment has been promoted by Clarke and Curtis as a small proportion of infants demonstrating evidence of return of elbow function at 3 months may then fail to progress considerably with an ultimately suboptimal recovery, thus endorsing the concept of reassessment at 3-monthly intervals (Clarke and Curtis 1995). The algorithm used for the assessment and management of infants presenting with obstetrical brachial plexus palsy at The Hospital for Sick Children is detailed in Fig. 4. An initial assessment is undertaken at 3 months using the Active Movement Scale (Table 3); a converted test score (the sum of the five scores for elbow flexion and extension of the elbow, wrist, fingers and thumb after conversion according to the values in Table 4) less than 3.5 is strongly predictive of a poor functional outcome and surgical reconstruction is therefore recommended. Those children with a test score of 3.5 or greater, with an intact T1 and no evidence of a Horner syndrome, are observed and undergo daily physical therapy to maintain passive range of motion; they are subsequently reassessed a 3-month interval. In the scenario of a child who initially demonstrates biceps function at the 3-month assessment, but then fails to meaningfully progress by the 6-month assessment (i.e., similar test scores are obtained), then it is likely that this heralds a poor functional outcome and surgical intervention should be considered.

At the 9-month assessment, the child undertakes the “cookie test ” in an attempt to identify those in whom a good early recovery (i.e., a test score of 3.5 or greater) failed to predict adequate elbow flexion by the end of the first year (Clarke and Curtis 1995; Curtis et al. 2002). If the child fails to achieve greater than half the range of elbow flexion against gravity (i.e., an Active Movement Score of at least 6), then exploration of the brachial plexus is advocated (Fig. 5). In practice, the test is performed with the child either seated or standing with the elbow held adducted against the trunk by the examiner to negate the trumpet sign; a small cookie is placed into the hand of the child, and they are encouraged to eat the cookie by actively flexing their elbow against gravity. The test is considered void if there is excessive neck flexion (greater than 45°). Application of the test score should never supplant sound clinical judgment; it is intended as a guide to early treatment.

In those patients who require surgical exploration, a preoperative diaphragmatic ultrasound is performed to document phrenic nerve function (Borschel and Clarke 2009). A chest radiograph is recommended if considering extra-plexal neurotization from the intercostal nerves (Slooff et al. 2001). Additionally, a computed tomographic myelography is undertaken to determine the probability of a root avulsion injury: the presence of a pseudomeningocele without demonstrable ventral rootlets indicates (with a specificity of 0.98) that the nerve root is probably avulsed (Chow et al. 2000). Pseudomeningoceles indicate an underlying dural injury, and in isolation, they are not absolute proof of a root avulsion injury – discontinuity of the rootlets should be identified in both axial and coronal planes to confirm the diagnosis of a preganglionic injury. The necessity for general anesthesia, intrathecal instillation of contrast medium, and exposure to ionizing radiation has driven the application of magnetic resonance imaging (MRI) in the radiologic assessment of the infant brachial plexus as it is noninvasive, nonionizing, and can be performed under sedation alone. However, MRI currently lacks the higher spatial resolution of CT myelography in identifying individual nerve roots (Caranci et al. 2013), and the quality of the images appears to vary widely across centers. Many specialist units do not routinely image the brachial plexus save for specific indications such as a suspected upper root avulsion injury following a breech delivery (Gilbert 1995; Al-Qattan 2004; Shenaq et al. 2004).

Electrodiagnostic studies have been used to provide information on the location, severity, and extent of the obstetrical brachial plexus palsy. Although more challenging to perform in infants, nerve conduction studies (NCS) may help to differentiate between preganglionic and postganglionic plexus injuries. Motor and sensory nerve responses (conduction velocities and distal latencies) vary significantly with age and in neonates are approximately 50 % the magnitude of those seen in adults (Miller and Kuntz 1986). Electromyography (EMG) is not routinely used in the assessment of obstetrical brachial plexus palsy; Gilbert et al. found that EMG findings correlated poorly with prognosis, although evidence of fibrillation on the background of absent voluntary motor unit activity was suggestive of a root avulsion injury (Gilbert et al. 1988). Overall, there is reasonable agreement that electrodiagnostic studies are overly optimistic in the setting of obstetrical brachial plexus palsy and are therefore not used at The Hospital for Sick Children, Toronto (Vredeveld et al. 1999).