Usefulness of an early MRI-based classification system for predicting vertebral collapse and pseudoarthrosis after osteoporotic vertebral fractures.

STUDY

DESIGN

Retrospective review.

OBJECTIVE

To clarify whether an early magnetic resonance imaging-based classification predicts pseudoarthrosis and final vertebral collapse in osteoporotic vertebral fractures.

SUMMARY OF

BACKGROUND

Initial therapy for osteoporotic vertebral fractures involves bed rest, orthotic use, and plaster casts. However, in some cases, pain persists because of progressive vertebral collapse or pseudoarthrosis. Prediction of these complications immediately after fractures can facilitate early proactive treatment despite the early prognosis being generally poor.

METHODS

A total of 109 patients (129 fractured vertebrae, 88 females, 21 males, and average age 79 y) followed up over 6 months after conservative treatment for thoracolumbar vertebral fractures were included. Early midsagittal T1-weighted and T2-weighted magnetic resonance images were analyzed. The incidence of final vertebral body collapse, pseudoarthrosis conversion, and delayed spinal cord paralysis were examined retrospectively for each vertebral fracture type.

RESULTS

According to the T1-weighted image-based classification, 74 of the vertebrae (57%) had total-type fractures. The final vertebral body collapse rate was significantly higher in this type than in others. Pseudoarthrosis was observed in 20 total-type fractures in 20 patients (18.3%); pseudoarthrosis conversion rate was significantly higher in these patients than in others. Delayed spinal cord paralysis occurred in only 1 patient (0.9%) with total-type fracture. According to the T2-weighted image-based classification, 69 vertebrae had the hyperintense wide-type fractures, which was the most common fracture type (53%). Hypointense wide-type fractures were associated with a significantly higher incidence of final vertebral body collapse, pseudoarthrosis, and delayed spinal cord paralysis. When total-type fractures of the T1-weighted image-based classification were subclassified according to the T2-weighted image-based classification, a significantly higher pseudoarthrosis conversion rate was observed in hypointense wide-type fractures.

CONCLUSIONS

Our results suggest that the radiologic prognosis can be estimated to a limited extent by determining the degree and extent of osteoporotic vertebral fractures using an early magnetic resonance imaging-based classification.