Brachial plexus birth injury: classification,natural history and shoulder external rotation

Introduction

Although the incidence of brachial plexus birth injury (BPBI) has fallen in some countries with changes in obstetric practice, it remains a significant cause of disability globally. There is much interest in research into the pathology, natural history and treatment of BPBI, with many submissions to the Journal of Hand Surgery (European Volume). This article comments on a few recent papers published in the Journal to highlight updates and new research.

Classification

Classification of BPBI is important for predicting and comparing outcomes after operative and nonoperative treatment. Currently, BPBI is commonly classified using grouping of the extent of nerve root involvement described by Narakas (1987), although some specialists simply describe Erb’s, extended Erb’s or global/complete palsy. The Narakas classification is fairly imprecise, being based on clinical examination after any transient element of the injury has had time to resolve. There is variability in the underlying nerve lesion and therefore the course of recovery. In addition, the classification description has been altered over time (Al Qattan et al., 2009), with various inconsistencies, for example, the importance of Horner’s syndrome.

Zlotolow et al. (2025) have presented a new classification, based on findings in 162 surgical explorations of the brachial plexus, aimed at more precisely describing the nerve lesion. There are four injury types depending on the number of injured nerve roots: A (C5–6), B (C5–7), C (C5–8) and D (C5–T1). These categories are similar to the Narakas groups I–IV, with an additional numeral indicating the number of complete root avulsions in each category, making a total of 18 possible subtypes.

Clinical correlation was with the Active Movement Scale (AMS) for motor function. Although the AMS is used in some units, variation remains in outcome measures used for BPBI, with other studies recommending the Mallet shoulder score, active and passive range of movement, and patient-reported measures (Brown et al., 2024; Pondaag et al., 2018). The AMS has the advantage of assessing most movements, but it is a non-linear scale, with only scores of 4 or higher being functionally useful. The new classification appears to correlate well with total pre- and postoperative AMS scores.

The new classification is likely to be useful for reporting outcomes of nerve repair for BPBI, enabling more accurate comparisons. However, while the number of injured roots is assessed together with avulsions, the severity of postganglionic ruptures is not fully classified. Also, the classification is applicable mainly in cases of brachial plexus exploration, this being necessary for only a minority of children. Clinical classification is likely to remain less precise, although improvements in MRI may help to define the pathology of the injuries without operation (Hu et al., 2025).

As well as the classification, the authors have reported a substantial series of outcomes of nerve repairs for BPBI, noting that the improvement in AMS scores after operation was greatest for type D, the most severe injury. This is difficult to interpret, as some spontaneous recovery occurs in most cases of BPBI, and the evidence for the benefit of nerve repair remains limited (Pondaag and Malessy, 2021). Observation of the outcomes for individual movements (table 4, Zlotolow et al., 2025) leaves uncertainty. Some of the movements in which improvement was seen, e.g. shoulder adduction, flexion and internal rotation, and elbow flexion and extension, are those in which spontaneous recovery is most likely to have occurred. Unfortunately, median scores for shoulder abduction and external rotation remained poor at 4 or less, and for digital movement in type D cases. While postoperative median AMS scores improved compared with preoperative scores, it is notable that scores were lower in some cases (figure S2, Zlotolow et al., 2025).

The study appears to provide substantial evidence that, while Horner’s syndrome is associated with worse preoperative function, it does not predict poorer postoperative recovery. Hence, its inclusion in the classification of BPBI remains doubtful.

Zlotolow et al. (2025) have made a substantial new contribution to our thinking on the classification of BPBI, which should help in reporting and comparing outcomes. Acceptance of any new classification would take time and validation by the wider surgical community (Lam, 2025). It will become evident over time if the classification is widely applied.

Natural history

Brachial plexus birth injury is a condition which generally improves spontaneously, but normal function is not achieved except in cases of transient paralysis, which recover fully within the first month of life. For cases with incomplete recovery, the shoulder is almost invariably affected. An understanding of the likely extent of spontaneous improvement helps to inform decisions for surgical intervention. There remains little consensus about the length of follow-up or at what time points (de Joode et al., 2024). Hyttinen et al. (2025a) studied recovery of shoulder abduction in 68 children in Finland who had not regained 90° active shoulder abduction at 3 months of age and did not undergo exploration and repair of the brachial plexus. Ninety-six per cent of cases regained 90° abduction, with most recovery by 14 months of age, with 47% achieving full recovery (170°) of abduction by 37 months. Overall, there appears to be good chance of recovery of a satisfactory range of shoulder abduction. The authors recommend that surgery to improve shoulder abduction should not be considered until age 1 year. Also, it is worth noting that the authors reported that active elbow flexion at 4 months (often taken as an important prognostic indicator of recovery) was not associated with recovery of shoulder abduction.

The findings relate to a selected group of children who did not have extensive injuries or were recovering well, so that surgical treatment was not undertaken. However, the indications for brachial plexus exploration applied at the treating unit are not entirely clear. Hence, the generalizability of the recommendations is uncertain. The lack of association of shoulder abduction with age of recovery of elbow flexion is interesting since other studies have found elbow flexion to be an indicator of a number of aspects of upper limb function (Hems, 2022; Johansson et al., 2019; Michelow et al., 1994; Smith et al., 2004; Waters, 1999), including shoulder abduction. Indeed, in a separate study from Hyttinen et al. (2025b), recovery of shoulder external rotation was found to be significantly associated with early recovery of elbow flexion. Early prognostic indicators of recovery remain an important field of research in BPBI.

Shoulder external rotation

External rotation of the shoulder is an important movement in placing the upper limb and hand in space but often recovers poorly after BPBI. Hyttinen et al. (2025b) studied spontaneous recovery of external rotation (ER) in 103 children who had not gained active ER at age 3 months. Only 51% of cases recovered ER, with most improvement occurring by age 2 years. The reason ER tends to recover worse than other movements, including shoulder abduction and elbow flexion, innervated by the same nerve roots (C5 and C6) remains unclear. Even when the infraspinatus is innervated, active ER may be ineffective, suggesting other factors at play, such as failure of central activation of the movement and co-contraction with the internal rotators. The amount of ER observed is also an interaction between muscle activity and the range of passive movement at the shoulder, with internal rotation contracture often developing.

In a recent article, Heitz et al. (2025) reported outcomes of suprascapular nerve reconstruction in 33 children with C5–6 or C5–7 BPBI. Repair was either with nerve grafts from the C5 root or transfer of the accessory nerve. The authors are to be congratulated for achieving long-term follow-up (minimum 3 years, mean 10.5 years) and for a thorough assessment of shoulder function. The primary outcome was the range of ER with the shoulder adducted, with the measurement method clearly described. Results for patients who underwent secondary shoulder surgery were analysed separately. Overall, active ER recovered in only 23 of the 33 cases, emphasising the difficulty of restoring this movement with nerve reconstructions. There was no significant difference in active ER between those who had nerve grafts or nerve transfers. The range of active ER appears to have been consistently less than the passive range (by approximately 40°), indicating the influence of muscle weakness.

The authors discuss their results extensively in comparison with other publications, including a meta-analysis by Nickel et al. (2023), which suggests better results for nerve transfer. However, possible differences in inclusion criteria and in the methods of ER assessment are identified. It therefore looks as though there is no clear evidence to favour nerve grafting or transfer for suprascapular nerve reconstruction. Other factors may be considered in making the choice. Nerve grafting usually involves excision of an upper trunk neuroma and repair for elbow flexion as well as shoulder movement. It is questionable whether reconstruction of elbow flexion is necessary, as this movement almost invariably recovers spontaneously (Hems et al., 2017; Narakas, 1987; Pondaag and Malessy, 2021). For accessory to suprascapular nerve transfer, the possibility of donor nerve morbidity still needs to be considered (Leblebicioğlu and Pondaag, 2024). While symptoms related to weakness of the lower trapezius have not been reported after transfer for BPBI, this has been observed after reconstruction for traumatic brachial plexus injury (Hems, 2011).

Brachial plexus birth injury is characterised by complex interactions between nerve injury and regeneration, with effects on skeletal and muscle growth and development of the central nervous system. There is still much room for further studies to define the natural history of this condition and provide evidence for treatment strategies. These articles are helpful in taking us one step further towards providing optimal care for these patients.