MR Imaging Studies after Epidural Injections: What Are We Really Imaging?

Although epidural anesthesia is generally considered safe, severe complications may rarely occur. With the increasing use of epidural injections for pain management, the number of cases with complications has steadily increased. Complications reported in the literature have been noted in different compartments of the spine, including the vertebrae and intervertebral disk spaces, the epidural space, the intradural extramedullary space, and the cord itself. Infectious, inflammatory, and vascular causes have been implicated as etiologies (1). Examples of these complications reported in the literature include diskitis and vertebral osteomyelityis, subarachnoid cysts and irregularities of the surface of the cord consistent with arachnoiditis, spinal cord lesions such as syrinxes, epidural or subdural hematomas, and finally, spinal epidural abscesses (SEAs).

Several hypotheses have been postulated as to the mechanism for the above complications. Spinal cord abnormalities may be secondary to ischemia, infarction, or edema. This ischemia may be related to venous stagnation due to the injection of anesthetic into the epidural space, thus interfering with flow in the epidural veins. This may be aggravated by lumbar stenosis in which there is a restricted epidural space. In addition, in patients with coincidental dural arteriovenous fistulas, further venous engorgement may produce spinal cord hypoxia.

In light of the serious consequences of the above complications, it becomes very important to diagnose these clinical complications in a timely fashion, because early diagnosis may change the outcome in many cases. The clinical symptoms of spinal infectious, inflammatory, and vascular processes may be nonspecific, especially in the early stages. However, MR imaging findings for infections, hematomas, and arachnoiditis have been well described in the literature, and application of these findings may help in narrowing the differential diagnosis. Nonetheless, one factor that may contribute to some uncertainty in the MR imaging diagnosis is the lack of scientific documentation of normal MR imaging findings following uneventful spinal injections. The article by Ikushima et al in this issue of the AJNR assesses the spinal MR findings associated with continuous epidural anesthesia in five patients with clinically uneventful spinal injections. Posterior epidural lesions were identified in all five cases similar to those in patients with epidural abscesses. In three of the patients, laboratory results ruled out infection. In the other two patients who did not undergo microbiological tests, the presence of infection was ruled out by their clinical course.

The authors of this article attempt to characterize the disease of the catheter-related lesions, but no pathologic specimens of these posterior epidural space lesions were obtained in this series. Several reports have described inflammatory mass lesions at the tip of intraspinal drug administration catheters (especially after infusion of high doses of morphine) in patients with long-term therapy. Surgical specimens have revealed noninfectious chronic inflammation, granuloma formation, and fibrosis or necrosis. Ikushima et al claim that their lesions are probably highly vascularized granulation tissue with increased water content because their lesions were CSF equivalent on T2-weighted images and granulomas are usually not as hyperintense as CSF. The authors also compare their lesions with SEAs, because MR imaging is very useful for the diagnosis of SEA (2). In the literature, catheter-related SEAs have been located in the posterior epidural space at the site of catheter tip insertion. The location, shape, and enhancement pattern of the cases by Ikushima et al were similar to those of chronic-phase catheter-related SEAs, but the SEAs usually do not have CSF-like high T2-weighted signal intensity. It is important to look for these differences in evaluating patients who have received continuous epidural anesthesia, because management would be completely different between sterile collections and SEAs.

The findings described by Ikushima et al are of great value to radiologists interpreting MR studies obtained in patients receiving continuous epidural anesthesia. It is important to remember, however, that some epidural injections are not continuous, especially those given for pain management where only a few milliliters are injected at one time. Findings here may be quite different. In our own experience, these epidural injections for lumbar back pain have produced subtle MR imaging findings. Prospective controlled studies with a larger series of patients are needed in the future to establish a normal baseline of MR imaging findings following uneventful epidural injections. Also, in the future, diffusion-weighted imaging may be used more frequently in examining patients with a question of infection. Until recently, only a few published reports have described the use of diffusion-weighted imaging to evaluate disease processes of the spine, but as experience with diffusion-weighted imaging of the spine increases, information about the findings of common spinal abnormalities such as infections will be more widely available. Eastwood et al (3) reported findings in spinal epidural abscess similar to those of abscess cavities in the brain. These further studies may aid in increasing the sensitivity and accuracy of diagnosing epidural catheter-induced lesions.

• Copyright © American Society of Neuroradiology