Subscribe to eTOC


Intraspinal injection treatments for back and sciatica complaints with various volumes and substances, especially local anaesthetics, have been reported for over a century. The discovery of cortisone in 1936 was followed in 1951 by evidence that local injection of hydrocortisone could produce palliative pain relief in histologically controlled inflammatory rheumatic joint disease.


Absent inhibitions or controls, by mid-century, the challenge of treating various neurological diseases with intrathecal and epidural steroid infusion was enthusiastically engaged. During the 1960s and 1970s, intrathecal steroids became a popular fad for treatment of multiple sclerosis and for treatment and prevention of the arachnoiditis complication of myelographic contrast agents.

For multiple sclerosis, the enthusiasm soon waned as the illness persisted and serious complications occurred, such as arachnoiditis, myelopathy, and aseptic meningitis. No controlled studies were ever reported. It became evident that contrast agent protection by methylprednisolone acetate also failed. A few months ago, however, purposeful intrathecal steroid injection resurfaced as the proclaimed cure for post-herpetic neuralgia.


Dr. Landau is Professor of Neurology at the Washington University School of Medicine in St. Louis, MO.


Spinal pain, especially in the lower back and sciatica, remain the most popular rationale for epidural steroid treatment. Usual clinical diagnoses include nonspecific back pain, failed back syndrome, spinal stenosis, herniated disk, bulging disk, foraminal arthrosis, facet disorders, and radicular compression. The central therapeutic presumption is that progressive degenerative backbone disease evokes an inflammatory condition like that of rheumatoid arthritis. But no convincing evidence of such inflammation has yet been cited in the literature.

There is no directly pertinent animal research; however, hydrocortisone has been found to reduce talc-induced arachnoiditis. Half a dozen randomized and double-blind placebo-controlled clinical studies of epidural MPA demonstrated no objective or symptomatic long-term advantage of steroids. These results do not eliminate the need for surgery as a treatment option.

In addition to placebo effect, transient improvement of symptoms may be attributed to the proved blockade or destruction of nociceptive nerve terminals by methylprednisolone acetate, its preservatives, the local anesthetic, or its hypertonicity and local anesthetic mixture. The most serious reported complications have been aseptic meningitis, myelopathy, encephalopathy, and painful arachnoiditis, especially with repeated injections.


There is general acceptance that intrathecal injection is more risky than epidural placement. Nevertheless, purposeful epidural needle infusion may be mistakenly subarachnoid in a significant percentage of cases, perhaps as many as 10 to 25 percent. Furthermore, rapid venous transport from epidural to intraneural veins has been readily shown in animals.

A popular nostrum for the back syndromes is a “dose pack” of steroid medication by mouth, usually 10 to 14 days at moderate dosage. Here, too, there does not seem to be any published evidence of efficacy for this treatment.

From 1979 to 1989, the Physicians Desk Reference advised that Depo-Medrol was not recommended or was contraindicated for intrathecal administration. In addition, in many clinical case reports of complications, especially from Australia, animal studies have shown that intrathecal injection of hydrocortisone produced pleocytosis and increased protein values. Furthermore, MPA as well as its vehicle preservatives produced histological lesions, demyelination, and slowed impulse conduction in several nerve structures. As noted above, less severe nonspecific neurotoxic effects might contribute to transient symptomatic improvement.


In the present managed care era when technical procedures are most highly rewarded, it has been estimated, absent documentation, that up to a quarter of the income to some pain treatment centers derives from epidural steroid infusions. Some HMOs are reported to require a trial of epidural steroid treatment before elective surgery can be approved.

More careful investigation of both efficacy and injury in animals and human subjects is needed. If, indeed, a transient neurotoxic effect may be of value for short-term clinical management, it might be worthwhile to try purposeful controlled neurotoxic blocks with dilute phenol or alcohol.

A review of the present database concerning intraspinal steroid therapy leads to the conclusion that there is no evidence of a significant inflammatory pathological component of the common back syndromes. It is certain that epidural infusion is not securely confined to the epidural space and that steroid treatment is no more effective than placebo control. It is certain that serious neurotoxic complications may be due to MPA, its preservatives, or the hypertonic combination with local anesthetic. There are no accurate data to define the percentage risk of serious complications, and the risk is not trivial.

Informed consent from patients should require advice that the treatment offers no long-term efficacy. Further they should be informed that serious complications, including persistent pain and paralysis, are a rare but certain risk.