By Michael Rubin, MD

Professor of Clinical Neurology, Weill Cornell Medical College

Dr. Rubin reports he is a consultant for Merck Sharp & Dohme Corp.

SYNOPSIS: Arachnoiditis, a feared complication of various spinal surgeries and procedures, shows a variety of nonspecific abnormalities on magnetic resonance imaging investigation, but there are no specific imaging features that predict prognosis.

SOURCE: Parenti V, Huda F, Richardson PK, et al. Lumbar arachnoiditis: Does imaging associate with clinical features? Clin Neurol Neurosurg 2020;192:105717.

Arachnoiditis, the inflammation of the exterior dura and interior arachnoid membranes, may occur anywhere along the spine, but appears most often in the lumbosacral region. Inflammation may be initiated by chemical insults, mechanical insults, or infectious agents, with spinal surgery and myelograms comprising the most common etiologies. Typically, symptoms include non-dermatomal neuropathic burning pain in the back, legs, and feet, with tingling, hypoesthesia, and weakness, and occasionally sphincteric issues, including constipation, incontinence, dysuria, and sexual dysfunction. Magnetic resonance imaging (MRI) of the lumbar spine is the diagnostic procedure of choice, with myelogram or computed tomography (CT) performed if MRI is contraindicated. Is there a correlation between the clinical features of lumbar arachnoiditis and MRI findings?

Using the term “arachnoiditis,” the investigators searched the electronic medical record of The George Washington University Hospital in Washington, DC, for reports of lumbar MRI scans performed between 2012 and 2018 on patients 18 years of age or older, which had a high likelihood of lumbar arachnoiditis. Reports that contained language that made arachnoiditis unlikely were excluded. Thirty reports satisfied the criteria, but two were subsequently excluded when a review of the MRI images failed to show any features suggestive of arachnoiditis. Twenty-eight scans remained in the study population and were reviewed by a radiologist to assess nerve root contour, nerve root thickness (nodular, thickened, or both), adhesion location (right, left, circumferential, central, or multiple), level of involvement (single or multiple), enhancement (yes or no), and confounding pathology. In addition, patients were assigned to a Delamarter group, which categorizes lumbar arachnoiditis into Group 1, showing conglomerations of adherent nerve roots residing centrally within the thecal sac; Group 2, demonstrating nerve roots adherent peripherally to the meninges, giving rise to an “empty sac” appearance; and Group 3, which shows a soft-tissue mass replacing the subarachnoid space. Clinical findings culled from the medical record included suspected etiology (post-procedural, post-operative, post-traumatic, post-infectious, neoplastic, or idiopathic), symptom dynamics (static or progressive), time course, pain (back, leg, or both), motor or sensory symptoms or both, sphincteric dysfunction, radiculopathy, and symptom location (right or left leg, back, or multiple). Statistical analysis employed Fisher’s exact test, with significance set at 0.05.

Patients ranged in age from 37 to 85 years, with an average age of 61.8 years, and a male:female ratio of 10:18, with static symptoms in 53.6% and progressive in 39.3%. Back or leg pain was present in 21.4% each, while another 50% had both back and leg pain. Motor symptoms were present in 25%, sensory symptoms in 17.9%, and 10.7% had both. Abnormal cauda equina contour was seen in 85.7%, with thickened nerves in 60.7% and nodular nerves in 25%. Adhesions were variable in location, and, among those who received intravenous contrast, 61.1% had nerve root enhancement. Degenerative disc disease was seen in 25%, with confounding pathology present in 50%, including laminectomy, spinal canal stenosis, discitis, epidural abscess, or tumor. Overall, the imaging findings did not correlate with clinical features. Adhesion location was not associated with symptom location, enhancement was not associated with any symptom category, localizable clinical features did not associate with location of adhesions, and Delamarter group was not associated with any grouped clinical findings.

However, nodular and thickened nerve roots, nerve root enhancement, and lack of adhesions were associated with a postoperative etiology, nerve root contour with motor and sensory symptoms, and confounding lumbar pathology with symptom progression. MRI findings are poor predictors of clinical presentation in lumbar arachnoiditis.

COMMENTARY

Arachnoiditis is a therapeutic challenge, requiring a multimodal approach, with effort directed at alleviating pain, improving quality of life, and managing symptoms. Nonsteroidal anti-inflammatories, opioids, and various adjuncts including duloxetine, gabapentin, pregabalin, and muscle relaxants, such as baclofen, may be helpful. Physical therapy, cognitive behavioral therapy, biofeedback, acupuncture, positive imagery, and psychotherapy may be offered. Surgical interventions, including shunt, cyst fenestration, myelotomy, duraplasty, lysis of adhesions, and laminectomy, have been reported, but long-term prognosis after surgery remains poor, with immediate improvement often followed by relapse and worsening symptoms.