Study design: Prospective radiographic analysis of vertebral compression fractures (VCFs) that underwent vertebroplasty.
Objective: Illustrate the variability in apparent magnitude of vertebral height restoration when this outcome is reported by four different methods commonly used in the vertebroplasty literature. Propose a consensus method for reporting vertebral height restoration.
Summary of background data: Measuring and reporting height restoration of fractured vertebrae presupposes a consensus of method that does not exist. Lack of consensus makes the interpretation of reports and comparison of outcomes of interventions that claim vertebral height restoration difficult.
Materials and methods: Preoperative and postoperative standing lateral radiographs of 65 VCFs in 41 patients were compared to assess operative vertebral height restoration. Restorations of vertebral height occurred in 23 instances and were reported by each of the following commonly used methods: (1) absolute restoration in millimeters; (2) percent restoration relative to initial fracture height; (3) percent restoration relative to lost vertebral height; and (4) percent restoration relative to referent vertebral height.
Results: Apparent magnitude of height restoration varied nearly four-fold depending on initial fracture severity and reporting method.
Conclusions: Substantial apparent variability in the reported magnitude of identical height restorations demonstrates the need for a consensus method for measuring, reporting, and interpreting this outcome. Rationale is presented to support the recommendation that reports of vertebral height restoration should: include all index vertebral height dimensions (posterior (Hp), middle (Hm) and anterior (Ha) vertebral height); include absolute measurements of all referent vertebral heights; be reported relative to a referent normative height; include a correction for inter-radiographic measurement error; take into consideration the dynamic mobility of some osteoporotic VCFs; and include the calculated precision error for all measurements.