Purpose of the study: Spinal fractures in patients with ankylosing spondylitis or idiopathic skeletal hyperostosis can raise difficult diagnostic and therapeutic problems. Spinal fracture is well known in ankylosing spondylitis but exceptional in diffuse idiopathic skeletal hyperostosis. The purpose of the present work was to identify clinical and radiological features in patients with ankylosing spondylitis, to determine whether similar risks and clinical expression are observed in patients with diffuse idiopathic skeletal hyperostosis, and to present a radiological classification of these fractures. We did not assess therapeutic methods in the present study.
Material and methods: Forty-eight fractures in 48 patients were observed over a period of 17 years. Twenty patients (mean age 62 years) had ankylosing spondylitis and 28 patients (mean age 81 years) had diffuse idiopathic skeletal hyperostosis. A fall was the immediate cause of the fracture in more than half of the patients. No notion of trauma could be identified in six patients. The radiological classification was established as follows; type I open-wedge anterior fracture, type II "sawtooth" fracture, type III occult or radiologically invisible fracture, type IV non-specific fractures comparable to other spinal fractures. A computed tomography was obtained in all patients seen after 1992 and magnetic resonance imaging was performed in case of suspected extradural hematoma. The ASIA classification (as modified by Frankel) was used for cord injuries. Clinical course and complications were noted.
Results: Diagnosis was established the day of fracture in 32 patients (12 spondylitis and 20 hyperostosis) and between day 2 and 30 for 16 (8 spondylitis and 8 diffuse idiopathic skeletal hyperostosis). The radiological classification was: type I n=30, type II n=4, type III n=8, type IV n=6 (one odontoid fracture, five compression fractures). Three patients had extradural hematomas (2 spondylitis and 1 hyperostosis). Thirty-four patients (11 spondylitis and 23 hyperostosis) had cord injuries, including 16 with a symptom-free interval. The ASIA classification was: type A n=4, type B n=6, type C n=20, type D n=4. Thirty-two patients died within the first three months after spinal fracture (10 spondylitis and 22 hyperostosis), due to bed rest related complications in 30. One patient died after rupture of an aortic aneurysm.
Discussion: Spinal fractures in patients with ankylosing spondylitis or diffuse idiopathic skeletal hyperostosis generally occur spontaneously or after low-energy trauma. Subsequent complications have serious consequences. Late diagnosis either results from missing a radiologically visible fracture or from the presence of an occult "paper thin" fracture. We do not have experience with diagnostic scintigraphy or magnetic resonance imaging. In our opinion, repeating standard x-rays the second and third weeks and use of a spiral scan or multiple spiral scan could provide early diagnosis.
Conclusion: The possible diagnosis of spinal fracture should be explored very extensively in patients with a symptomatic ankylosed spine who present symptoms compatible with spinal fracture, with or without trauma.