Because I have a kid with an unfused thoracic high curve, I've been paying attention over the years to patients who develop neck problems after fusion surgery. Looking through the literature, I had the sense that there were a larger number of these problems with the more recent techniques (pedicle screw).

A recent (2011) paper from a group of respected surgeons (including Betz) confirms the sense that I had, and also offers an explanation.

My summary: A prospective database was reviewed from 2003 to 2006 to find pediatric patients with all pedicle screw constructs and Lenke 1 and 2 (single and double thoracic) curves. Going into surgery, 6 of the 22 patients had cervical kyphosis (c-shaped necks) (around 13 degrees). The other 16 had normal (neural or lordotic - backwards c shape) cervical alignment going in. Postoperatively, all of the 6 who had neck misalignment going in still had the same misalignment, and 8 of the 16 who hadn't had neck misalignment previously now did - 6 of them with frank cervical kyphosis of 10 degrees or more. The problem? It looks like they're overcorrecting the (normal range) kyphosis in the backs of these scoliosis patients so that the normal curve in their thoracic spine is too small (hypokyphotic), and that's causing the kyphosis to move up into the neck.

Here's the conclusions section:
"The sagittal profile of the thoracic spine is related to that of the cervical spine. The surgical treatment of Lenke Type 1 and 2 curves by using all pedicle screw constructs has a significant hypokyphotic effect on thoracic sagittal plane alignment (19 [86%] of 22 patients). If postoperative thoracic kyphosis is excessively decreased (mean 25.6, p < 0.05), the cervical spine may decompensate into significant kyphosis."

http://www.ncbi.nlm.nih.gov/pubmed/21800955