Surgical outcomes for spinal deformity in osteogenesis imperfecta

IRB approved retrospective review of the the SHOnet (Shriners Health Outcomes Network).

Osteogenesis imperfecta (OI) is the most prevalent genetic disorder of bone and connective tissue in the pediatric population, with an incidence of 1/15,000-20,000. Scoliosis has been reported to be present in 39-100% of OI patients and may continue to progress into adulthood but there is little information on the perioperative outcomes, rates of revision surgery and complications in this fragile population. The purpose of this study is to examine the prevalence of scoliosis in the OI population across a tertiary care multihospital system. The rate at which these patients progress to surgery, the perioperative complication rate and if there is an association between complications and age at surgical intervention, pre-operative Cobb angle, number of fractures, and type of OI. Incidence of scoliosis in OI may be lower than previously reported, many may never require surgical intervention and those that do require surgical intervention have an 8% reoperation rate at an average of 3.3 years postop.

IRB approved retrospective review of the the SHOnet (Shriners Health Outcomes Network), electronic health record data warehouse from January 1, 2011 and December 31, 2017. Inclusion criteria included International Classification of Diseases (ICD) code for osteogenesis imperfecta and scoliosis. The data warehouse was queried for age, gender, presence of vertebral body fractures, proximal junctional kyphosis, basilar invagination, bisphosphonate use, and perioperative complications including postoperative infection.

There were 2372 patients with osteogenesis imperfecta, 429 or 18.1% also had a diagnosis of scoliosis, while 81.9% did not (see Table 1). Only 74 patients (17.2%) of the patients that had scoliosis underwent spine surgery, 12 of which had staged surgery with an average preop thoracic Cobb 58.18 (range 7-115), and thoracolumbar Cobb 59.83 (range 5-145). Six patients (8%) required revision spine surgery. Average time to revision was 3.88 years (6.9-69mo). Bisphosphonate use was present in 35.5% of patients that did not require surgery and in 40.5% that did.

With over 2300 patients, this is the largest study to date on scoliosis in patients with OI. We found that contrary to prior studies which had indicated the incidence of scoliosis ranges from 39 to 100% that it was only present in 18% of our patients. This indicates that the incidence of scoliosis in OI is potentially lower than previously reported, though due to the nature of this study it may be underreported. Furthermore, only 17% of those patients in our study with scoliosis eventually underwent surgery. It is important to note that many patients treated non-operatively had deformity within the operative range, which may indicate that they were not considered good candidates for surgical intervention. Lastly, we found that bisphosphonate use did not appear to affect the likelihood of progression to surgery in this group of patients, though this may be due to later initiation of the bisphosphonate use.