Nerve surgery now restores elbow flexion in roughly 95% of operated infants. Shoulder external rotation remains the unsolved half of this injury, and the most consequential finding of the last decade had nothing to do with technology.
Neonatal brachial plexus palsy occurs in about 1.5 per 1,000 total births, counting both transient and persistent impairment (Obstetrics & Gynecology, 2014). The research of the last decade has been unusually candid about which parts of the problem are solved and which are not, and the split runs between the elbow and the shoulder.
The most consequential finding here is not a device. When the American College of Obstetricians and Gynecologists convened its Task Force on Neonatal Brachial Plexus Palsy, the executive summary recorded that multiple peer-reviewed reports describe the injury occurring without any clinically recognizable shoulder dystocia, at both vaginal and cesarean delivery. That matters in two directions. It means the injury is not always produced by traction at the shoulder, and it means a delivery record with no noted dystocia does not by itself settle the question. Our overview of Erb’s palsy covers the injury itself.
What follows is ordered by what the evidence supports rather than by what is newest.
Elbow flexion is close to a solved problem. Shoulder external rotation is not, and the field says so in its own papers.
A meta-analysis in Child’s Nervous System pooled five observational studies covering 194 children to compare the two ways of rebuilding elbow flexion: grafting a length of donor nerve across the injured segment, or transferring a working nerve to power the target muscle directly. Functional recovery reached 95.2% after nerve transfer and 96.4% after grafting, a difference too small to matter, with no donor site problems reported in any included study. When two competing operations both work more than nine times in ten, the argument has effectively been won.
The shoulder tells the opposite story. A multicenter cohort of 145 infants published in The Journal of Bone and Joint Surgery compared nerve grafting to the suprascapular nerve against transferring the spinal accessory nerve into it, the two standard ways of restoring shoulder external rotation. Mean Active Movement Scale scores afterwards were 2.70 and 3.21 out of 7. Seventeen percent of the whole group reached a score above 5. Forty percent of the transfer group and 53% of the grafting group went on to a second shoulder operation. The authors’ own summary of that is the word disappointing.
The transfer did win on the margins that count: nearly five times as many infants reached functional external rotation, 24% against 5%, and the risk of needing further shoulder surgery was significantly lower. So there is a better option. There is not yet a good one.
Timing is the part still under argument. Surgeons watch for elbow flexion to return and treat its absence as the trigger for reconstruction, but the age at which that call gets made varies widely. Across the studies pooled in the elbow flexion meta-analysis, mean age at surgery ranged from 5.7 months to 11.9 months. A four-month spread in the mean, between centers publishing on the same operation, is itself the answer to how settled the question is.
Most children with this injury never reach an operating room, and for them the therapy program is the treatment rather than the preparation for one.
Passive range of motion is the work that runs underneath everything else. A shoulder that loses external rotation early develops an internal rotation contracture, and that contracture is what drives the joint damage described below, so protecting range is not a holding pattern while the nerve decides what it is going to do. It is what keeps the later options open. A surgeon offered a stiff shoulder at three years old has fewer choices than one offered a supple shoulder with the same nerve deficit. The detail sits in our guide to physical therapy for Erb’s palsy.
From the studies cited on this page:
Botulinum toxin injected into the muscles that pull the arm inward, chiefly subscapularis and pectoralis major, is used to buy range while the antagonists are strengthened, and splinting and serial positioning work the same territory. Both are common practice. Neither carries evidence anywhere near the strength of the surgical comparisons above, which is a reason to ask a specific question rather than a reason to refuse: what is this injection expected to change, measured how, and by when.
Read the study design before the conclusion. Nothing cited on this page is a randomized controlled trial, and that is a fact about the whole field rather than a gap in this page.
The elbow flexion meta-analysis pooled observational studies. The shoulder cohort is labeled therapeutic Level III, meaning a retrospective comparison rather than a trial with random assignment. That is close to the ceiling of evidence available in this condition, for two reasons that are not going away: the injury is uncommon enough that recruiting a large trial takes years, and surgeons are reluctant to randomize which operation an infant’s arm receives.
The practical consequence for a family is a reading habit. When a center describes a technique as promising, ask how many children, over how long, compared against what. Ask which outcome scale was used and what score counted as success. A paper reporting Active Movement Scale results at 18 to 36 months is telling you something; a paper reporting satisfaction at six weeks is not. Understanding what is being measured also helps in conversations about long-term effects.
The registry is the place to start, and a short list is what a family should expect to find. Specialist centers also run studies that never recruit publicly, so ask directly at the hospital doing the surgery rather than relying on a search alone.
Families interested in clinical trials can search the federal ClinicalTrials.gov registry for active Erb’s palsy and brachial plexus studies. Specialist pediatric brachial plexus centers (often at children’s hospitals) also recruit participants directly.
The advance that has changed decisions is imaging, and it changed them by finding a joint problem that used to be discovered too late to fix.
Incomplete nerve recovery leaves the muscles that rotate the arm inward stronger than the ones that rotate it out. The arm sits turned in, the humeral head drifts backward out of the socket, the glenoid remodels into retroversion, and the head itself flattens. That sequence is glenohumeral dysplasia, and until recently the humeral head part of it was described rather than measured.
Researchers at Cincinnati Children’s reviewed MRI scans of 32 children with this injury, aged 0.7 to 11.5 years, and produced a way to quantify the deformity as a ratio of the anterior to posterior area of the humeral head. On affected shoulders that ratio averaged 0.76, significantly different from the same child’s unaffected shoulder, and it tracked with both glenoid version and posterior subluxation of the humeral head. The finding that changed practice came next: after joint rebalancing surgery the humeral head remodeled measurably, so a flattened head, long treated as a reason not to operate, does not by itself rule the operation out.
Three scales now appear in most serious papers on this injury: the Active Movement Scale, which grades movement from 0 to 7 in an infant who cannot follow instructions, the Mallet classification for shoulder function in an older child, and the Narakas type for how much of the plexus was injured. Standardizing them is what makes one center’s series comparable to another’s, and it is the reason the two studies above could be read against each other at all.
Robotic rehabilitation and algorithm-driven treatment planning show up in a great deal of writing about this condition, including an earlier version of this page. Neither has produced evidence in brachial plexus birth injury of the kind cited above. Treat them as research directions worth watching rather than as care a family should expect to be offered, and be wary of any clinic that leads with them.
Weakness or paralysis of the arm caused by injury to the brachial plexus, the nerve bundle running from the neck into the shoulder, usually at birth. It affects about 1.5 per 1,000 total births counting transient and persistent cases together, and the classic presentation is passive range of motion in the arm that exceeds what the baby moves actively.
Clinically at first, by watching what the arm does rather than what it can be moved into, then followed over months with a standardized scale. Nerve conduction studies, EMG and MRI add detail where surgery is being considered. See our full diagnosis guide.
For elbow flexion, either nerve grafting or nerve transfer, which a meta-analysis of 194 children found equivalent at 96.4% and 95.2% functional recovery. For the shoulder, transferring the spinal accessory nerve to the suprascapular nerve, which in a 145-infant cohort produced functional external rotation in 24% against 5% for grafting and fewer second operations. Therapy to protect range runs alongside both.
Because losing shoulder range starts a chain that is hard to reverse. An internal rotation contracture pulls the humeral head backward out of the socket, the glenoid remodels, and the head flattens. Keeping the joint supple while the nerve recovers is what keeps the later surgical options open.
Failure of elbow flexion to recover is the usual trigger, but the age at which centers act on it is genuinely disputed. Across the studies in the elbow flexion meta-analysis, mean age at surgery ranged from 5.7 to 11.9 months. Ask a surgeon what threshold they use and why, because the answer varies between hospitals.
Two concrete ones. Imaging can now quantify humeral head deformity, and the same work showed the head remodels after joint rebalancing surgery, so a flattened head no longer rules out an operation it once ruled out. And standardized scales let one center’s results be compared with another’s, which is how the shoulder problem got named as a problem.
It depends on whether the injury resolves in the first year or leads to nerve reconstruction and then to secondary shoulder surgery, which 40% to 53% of one operated cohort required. Years of therapy sit on top of either path. Legal options can help families recover treatment costs when an injury was preventable.
Kelsey is an experienced surgical nurse who works extensively with pediatric patients and families navigating birth injuries and cerebral palsy diagnoses. She is licensed by the Missouri Board of Nursing and operates under the Nurse Licensure Compact, allowing her to practice across compact-member states. As both a mentor and patient advocate, Kelsey focuses on safety, communication, and compassionate care: helping families understand medical procedures, treatment options, and the realities of pediatric conditions.
View Kelsey Pabst, RN's full bio →
Our nurses, patient advocates and legal experts are solely focused on bringing you the latest cerebral palsy information, options for financial assistance and access to community support.
Editorial process →The information on this page is provided for general educational purposes and is not a substitute for professional medical or legal advice, diagnosis, or treatment. Always seek the guidance of a qualified physician or attorney with any questions about a medical condition or legal matter. Read our full disclaimer.
