Surgery

At surgery, patients were placed in the lawn- chair position, with a transverse shoulder roll. The head and neck were abducted away from the side of surgery. The entire supraclavicular area was prepared and draped with a thyroid sheet. The skin incision was created one fingerbreadth posterior and parallel to the clavicle. The incision was sinusoidal and extended 6 to 8 cm. lateral to the palpated lateral clavicular border of the sternocleidomastoid muscle. Dissection proceeded through the platysma muscle, taking care to protect the underlying supraclavicular nerves. The omohyoid muscle was resected to allow access to the scalene fat pad and to remove a potential compressive structure of the brachial plexus. The scalene fat pad was elevated from inferior to superior, revealing the upper brachial plexus. Great care must be taken to identify the suprascapular branch of the upper trunk, as it tends to travel within the middle layers of the scalene fat pad, and is theoretically prone to iatrogenic injury at this point.

Once the scalene fat pad was elevated, the upper trunk and its trifurcation into the anterior and posterior divisions and the suprascapular nerve was explored. Typically, epineurial scarring was evident at this point, and external neurolysis with microsurgical instruments and technique was performed. Anterior scalene resection was also performed at this time, although this was generally partial, and only sufficient to release the most superficial fibers compressing the upper trunk. This typically amounted to 15 or 20% of the thickness of the anterior scalene muscle.

The long thoracic nerve was then exposed laterally and posteriorly to the upper trunk. Disa and colleagues point out the underappreciated anatomy of the long thoracic nerve in the supraclavicular area, and discuss the inadequacy of standard gross anatomic descriptions showing the nerve to be more medial. It is worth noting that the long thoracic nerve is delicate and no more than 2 to 3 millimeters in diameter in this location. Its lack of substance in relation to the bulk of the serratus anterior muscle teleologically predisposes the neuromuscular unit to dysfunction. The situation is made worse by the passage of the nerve through the thick middle scalene muscle.

Once the nerve was isolated, it was then neurolysed using microsurgical instruments and the operating microscope. This was necessary because of the delicate nature of the nerve and to decrease surgical scar formation within the operative field. As with the anterior scalenectomy, the middle scalene was resected sufficiently to decompress the long thoracic nerve as it traversed and exited the muscle. In 6 of 22 surgical procedures (27%), a demarcated area of compression within the nerve was clearly noted, more so toward the exit point of the nerve from the muscle. This took the form of a narrowing and rubor of the epineurium, perhaps representing neovascularization at the site of compression. In one case, a complete resection of the middle scalene was required, and in the other 21 cases, a partial release of 15% or slightly more was accomplished. This was suitable to expose the long thoracic nerve and at least remove the circumferential muscle fibers of the middle scalene.

Direct electrical stimulation of the long thoracic nerve and upper trunk was performed in all cases, with a Radionics ™ intraoperative nerve stimulator. Current of up to 10 milliamps was used to stimulate contraction of the serratus anterior and the muscles supplied by the upper trunk. It was interesting that the contractions of the serratus anterior appeared uniformly to improve following decompression and neurolysis. It is a point of speculation as to the importance of this electrical "overstimulation" of the nerves in assisting recovery of the paralyzed muscles. It is compelling to note that even in cases where no clear anatomic point of epineurial compression could be noted, recovery of serratus function was excellent. Further, in fully 50% of cases, recovery was noted within 24 hours, an unusual circumstance in most situations of nerve decompression.

Prior to closure, an examination of the superior- most and inferior margins of the surgical wound was performed to identify and release compressive fascial bands potentially capable of causing compression of the upper trunk and the long thoracic nerve. This was accomplished sharply under direct vision and with blunt digital dissection into the recesses of the wound.

Wounds were closed in three layers with reconstruction of the platysma and two skin layers. No drains were used. Postoperative management consisted of immediate active range of motion at the shoulder and neck. By the third postoperative day, patients were to have a full range of motion at preoperative levels or beyond, where capable.