Scapular winging is an important cause of functional disability. The morbidity associated with long thoracic nerve dysfunction is underappreciated and under treated. Ferry  argues that conservative treatment is an option based on inadequate data; we likewise favor a more aggressive approach involving surgery. Ten of our patients had significant functional deficits due to winging that had been present for more than 6 years; all 10 responded to surgery, often (70% of cases) within 1 week of surgery. The average time to improvement was 5 days (range 1 day to 3 months).
Our findings support nerve surgery as a treatment option in specific cases of supraclavicular injury to the long thoracic nerve and diminish the possibility that improvement was spontaneous. Risk factors for supraclavicular nerve injury include vigorous athletic maneuvers with the affected extremity, lifting heavy weights, and direct external pressure (as in deep massage). Injury to the upper trunk of the brachial plexus is also associated with the proposed stretch or compression mechanisms of injury. Upper-trunk disease reduces the reliability of shoulder examination because deltoid and spinati strength are reduced by long- standing scapular instability. However, all of our patients showed direct evidence of upper-trunk injury (BMG biceps score of M3 or M4).
The relatively delicate structure of the long thoracic nerve is contrasted with the densely composed upper trunk and predicts the consequences of trauma to each element: given similarly applied forces, the upper trunk shows less dysfunction than the long thoracic nerve. This explains why electrophysiologic examination of the upper trunk-supplied muscles of the affected extremity often reveals no clear abnormalities, the upper trunk injury being relatively minor [19, 26]. EMG of the serratus anterior, however, often reveals greater dysfunction than in the other muscles. Unfortunately, electrodiagnostic studies of the long thoracic nerve can be unreliable indicators of injury severity. The tendency toward normal results with serratus anterior testing in our population may then be ascribed to inadvertent testing of the latissimus dorsi, teres major, or other unaffected chest wall muscles, which can be difficult to isolate . The three main portions of the serratus anterior are preferentially innervated by branches from different roots of the long thoracic nerve , and both the stimulation point and measurement point in relationship to the nerve constriction site could all affect the EMG results . The fact that the long thoracic nerve was in continuity in all 50 cases might also have hampered EMG-based attempts to uncover subtle (but functionally significant) denervation abnormalities.
The observed rapid recovery of half of the patients might indicate a less severe nerve injury (no focal demyelination) that remains functionally debilitating. It is unlikely that the fast recovery is due to an ischemic mechanism, since muscle atrophy and denervation would be expected in cases persisting several years. Such injury may be classified as a Sunderland grade 0.5, wherein the myelin of the injured nerves is apparently intact and a different mechanism causes loss of function [30, 31]. One possibility is the concept of "silent synapses" – inhibited acetylcholine release at the motor end plate due to blocked axonal flow caused by mechanical forces. Transport is sufficient to sustain nerve health, but not to stimulate muscle contraction. By surgically removing the source of nerve constriction, normal muscle function could quickly resume .
This study expands the database of nerve-based management of scapular winging caused by supraclavicular long thoracic nerve injury. Muscle- and tendon-based procedures should remain important tools in cases of refractory, symptomatic scapular winging, but the effectiveness and low morbidity of long thoracic nerve decompression appear to justify its becoming the initial treatment of choice in appropriate cases, for which we suggest the following paradigm:
1. Clinical evaluation with particular attention to the cause of injury or any associated events. Patients with symptomatic scapular winging related to injury localized at the long thoracic nerve near the middle scalene are candidates for nerve surgery.
2. Direct physical examination, with particular attention to scapular movements and strength. Strength of the serratus anterior muscle can be measured in terms of maximal protrusion at the inferior scapular angle (normal is 0 cm; >5 cm signifies extreme loss of function).
3. Electrical studies to detect loss of nerve continuity, which could indicate the need for nerve grafting or transfer. The lack of abnormal findings with electrical testing should not obviate surgery when clinical findings suggest surgical intervention.
4. Observation of the following guidelines for determining surgical candidacy and time since onset of injury:
<7 years, candidate for nerve surgery in the absence of other contraindications such as established loss of nerve continuity for > 18 months
7–10 years, relative candidate (good outcomes are less predictable)
>10 years, tendon transfer becomes the primary treatment option (nerve surgery is a secondary option)
Pain and inflammation do not automatically improve even with successful flattening of the scapula, and they improve unpredictably in our experience. We find that patients are not always happy with the results of surgery when the primary presenting symptom is established pain, even when the winging is reversed. These cases, however, represent a minority of all patient experiences.
Postoperative management is generally restricted to gentle range-of-motion (ROM) therapy and electrical stimulation. The protocol for patients with longstanding winging is daily stretching for up to 1 year following surgery, after which strengthening work can begin. Patients with winging for less than 2 years begin strengthening after 3 months of ROM therapy, which helps prevent and treat adhesive capsulitis at various shoulder girdle joints.