Occult Spinous Process Fractures Associated With Interspinous Process Spacers

Spine 2011;36:E1080–E1085. DOI: 10.1097/BRS.0b013e318204066a

Biomechanical cadaveric studies have suggested adequate spinous process strength to support placement of interspinous process spacers (IPS). Postoperative spinous process fractures have been reported in one%—to 5.8% of patients in previous series based on routine biplanar radiographic evaluation. However, most fractures occur between the base and midportion of the spinous process in an area that is typically diffi cult to visualize on plain radiographs due to device design.

Methods. All patients underwent preoperative biplanar plain radiographs and computed tomography (CT) of the lumbar spine to confirm anatomy favorable for IPS placement and rule out fracture or spondylolysis. Postoperatively, all patients underwent repeat CT imaging within six months of surgery, biplanar radiographs at two weeks, six weeks, three months, six months, and one year. All studies were reviewed independently by a neuroradiologist and two orthopedic spine surgeons.

Results. Fifty implants (38 L4–5, 12 L3–4) were placed in 38 patients who completed follow-up and were included in final analysis. Three IPS designs were included (34 Medtronic X-STOP titanium, 8 X-STOP PEEK, 8 Lanx Aspen). Postoperative CT revealed 11 nondisplaced spinous process fractures in 11 patients (28.9% of patients, 22% of levels). Five fractures were associated with mild to moderate lumbar back pain and six fractures were asymptomatic. No patient reported a traumatic incident. No fracture was identifiable on plain radiographs. One fracture displaced during follow-up evaluation. Three patients underwent IPS removal and laminectomy. Three fractures healed by CT in one year. Overall, patients with fractures tended toward poorer outcomes by Zurich Claudication Questionnaire (ZCQ) (28.5% vs. 34.8% improvement in symptom severity, P=  0.496; 21.4% vs. 30.7% improvement in physical function, P = 0.199) and tended toward lower satisfaction rates (50% vs. 73.7%, P = 0.24) at one year compared to patients without fracture.

Conclusion. Interspinous process spacer surgery appears associated with a higher rate of early postoperative spinous process fracture than previously reported. In all cases, in this series, plain radiographs were inadequate to identify fractures because all fractures were initially minimal or nondisplaced, many patients were osteopenic, and the metallic wings of the devices often obscured fractures. Moreover, in most patients, fractures were associated with mild or no acute localized pain. This study suggests that unrecognized spinous process fracture may be responsible for a signifi cant number of patients who experience unsatisfactory outcome after IPS surgery. CT imaging is required to identify the vast majority of such fractures.