Because my son has such a high curve, I've been paying attention to the occurrence of problems above the fusion after surgery. With older hardware, most of the problems after surgery were below the fusion. With the newer hardware, surgeons are seeing more problems above the fusion.

There was an interesting article on this topic. (Which, of course, I can't find). But here's a recent article from Lenke about the occurrence of PJK after surgery.

http://www.ncbi.nlm.nih.gov/pubmed/23232215

Abstract
STUDY DESIGN.: Retrospective comparative study. OBJECTIVE.: We aimed to examine the difference in clinical outcomes in proximal junctional kyphosis (PJK). SUMMARY OF BACKGROUND DATA.: To date, PJK has been primarily a radiographical finding. Inferior outcomes associated with PJK have not been reported. We performed an analysis of PJK in adult deformity patients to identify risk factors and to evaluate clinical outcomes. METHODS.: A total of 364 patients at a single institution from 2002 to 2007 with adult scoliosis, with an average 3.5 years' follow-up were analyzed. Inclusion criteria were age more than 18 years and fusion greater than 5 levels from any thoracic upper instrumented vertebrae to any lower instrumented vertebrae. Cobb measurements in the coronal and sagittal plane in addition to measurements of the PJK angle at postoperative time points were performed. Clinical assessment was performed using Scoliosis Research Society (SRS) scores and the Oswestry Disability Index. RESULTS.: The prevalence of PJK was 39.5% (144/364). The average age in the non-PJK group (n-PJK) was 48.9 versus 53.3 in the PJK group (PJK), and, specifically, age more than 60 years posed a higher prevalence. The prevalence of osteoporosis was 9.8% versus 20.4% in the n-PJK versus PJK groups, respectively. Sex, body mass index, revision surgery, and smoking status were not different between groups. Pain was prevalent in 0.9% versus 29.4% in n-PJK versus PJK, which resulted in lower composite SRS Pain scores (mean change +1.2 vs. +0.8), despite no differences seen in other SRS domains, total SRS score, or Oswestry Disability Index. On multivariate analysis, the presence of pain of the upper back was highly predictive of PJK (odds ratio, 12.5, 95% confidence interval, 2.5-63.2).Radiographically, no differences were seen between groups. However, increasing distance of the upper instrumented vertebrae to C7 plumb line had a higher prevalence of PJK. Instrumentation type, surgical approach, and crosslink use were not different between groups. CONCLUSION.: PJK results in worse clinical outcomes measured by the SRS Pain subscore. Our regression model suggests that pain in the upper back has a strong predictive value for PJK.Level of Evidence: 3.