Back Surgery Begets Back Surgery

F. Reed Murtagh

You probably last heard that old saying not very long ago when you and a colleague were reviewing the spinal MR images of some unfortunate patient with a history and pictorial evidence of multiple prior spine surgeries and the ominous history of “Failed Back Syndrome.” It didn’t matter if it was the cervical, thoracic, or lumbar spine. Everyone from the MR imaging technicians to the maintenance people knew exactly what this was all about: the patient had a simple starter laminectomy many years ago, which didn’t cure the symptoms and had had two or more spine surgeries since. That patient is now either addicted to painkillers, being considered for yet another surgery, or both. He may also have arachnoiditis. He probably gets frequent flier miles for his hospitalizations. He is never going to be pain-free.

Why is this “failed back” syndrome such a common entity? Although it is true that patients sometimes simply do not respond to the best, most targeted, and correct surgical procedure for their symptoms, most often the “failed back” patient didn’t get better, because the most appropriate surgical procedure may not have been the one he received. Or it wasn’t enough. In turn, this error may have been due to the original diagnosis not being precise enough. The problem goes all the way back to the multiple overlapping possible causes of back pain.

Back pain is such a protean condition that not any one simple etiology is usually the cause, unless the patient has a very specific radiculopathy due to a very specific extruded disk fragment seen at MR imaging to be sitting directly on that irritated nerve root. Then it can be reasonably expected that, if the fragment is removed, the radiculopathy will disappear and not recur unless the disk extrudes again in the same anatomic location. This is the appropriate course of action in this situation, emergently if there is severe neurologic deficit, and after conservative therapy fails in other situations. Clinically, this disk impingement causing radiculopathy is actually less frequently encountered than the scenario in which a patient has vague but real back pain, with or without radiculopathy, and a good contribution of facet pain, muscle spasm pain, and psychogenic overlay. How does the clinician ever figure out which anatomic component is responsible for which pain component, and how much? Is operating on the intervertebral disk going to relieve this patient’s pain?

Fortunately, more recent trends in spinal disease diagnosis and therapy place much more emphasis on obtaining the most precise anatomic diagnosis the first time around, before any surgery takes place at all. Recent trends also strongly favor conservative trials, such as physical therapy and anti-inflammatory medications before even considering surgery. Very often, nothing else is necessary and in the natural course of things large numbers of patients get over the acute phase and can exist symptom–free, or relatively so, without ever going under the knife. This course of action seems much more likely to ensure successful symptom relief for the suffering patient than trying to figure it out over many years and many trial-and-error surgeries in a futile effort to eliminate one possible cause after another until the patient has nothing left but enough titanium to set off the metal detectors at the airport.

Under this conservative scenario, surgery become a last choice, when conservative therapy fails. Outcomes are much better when surgery is thus targeted, more focused and selective, with the correct procedure now likely to be applied to the correct anatomic or biomechanical diagnosis the first time around. Not every case of back pain needs a laminectomy and diskectomy, and not everyone needs screws and plates.

Part of the this evolving better understanding of the individual but interdependent causes of spinal pain syndromes has come courtesy of the orthopedic spine surgeons, who emphasized biomechanics. Neither neuroradiologists nor neurosurgeons had much of any understanding of this concept before the orthopedics started talking about it. The first Interdisciplinary Spine Conference held in Snowbird, Utah in 1989 was the first time most participants ever heard anyone take the biomechanical theories of Punjabi and White seriously. These concepts make a great deal of sense, because the spine is not a static construct and is actually designed to function through a wide range of movement, and have subsequently led to increased considerations of biomechanics in every surgical procedure.

Most tellingly, it is now desirable to leave the postoperative spine patient with some muscle and tendon dorsally in the operative site to support the spine when laminectomies alone are performed. If wide decompression surgery is contemplated, some sort of fusion procedure will be necessary to lend support afterward. Hardware development for spinal fusion became a growth industry and companies that manufactured the devices enjoyed phenomenal initial public offerings in the late 1990s. Primary instrumentation and stabilization came to be considered beneficial in cases of spinal instability, with or without decompression. Disk space cages, interbody fusion devices, and, on the horizon, disk replacement devices were designed and helped many patients.

Perhaps even more important, practitioners recognized that not all spinal pain syndromes started and ended with the intervertebral disks, and facets and other tissues could be involved in pain production. Disks did not even need to protrude or extrude (or herniate) and could hurt all on their own without compressing neural tissue. It is easy to forget how revolutionary these concepts were not that long ago.

In this issue of the AJNR, Cyteval et al present a clinical approach to spine disease than we as neuroradiologists may be accustomed to. Interventional neuroradiologists have been at the forefront, with clinical outcomes articles that have gone a long way toward the ultimate acceptance of embolization techniques in vascular diseases of the brain and vertebroplasties in the spine. This article brings us further into an era in which the efficacy of spinal injection procedures may finally be subjected to necessarily rigorous clinical studies to determine their ultimate usefulness.

For as many nerve root sleeve blocks, facet blocks, and diskographys neuroradiologists perform, there are no well-designed long- or even short-term clinical studies in the neuroradiology literature to support the efficacy of these techniques. We write articles about techniques, but cannot seem to define the problems well enough to examine the efficacy. There are some studies to be found in the orthopedic, physical therapy and rehabilitaion, and anesthesia literatures, but these spinal injection procedures are still performed by those specialties (and by us) very much in a spirit of empiricism. Long-term studies are difficult for many reasons, not the least of which is the question of how to define “long-term relief.” Without any specific endpoint, some authors consider long-term relief to be up to and including 3 months. Others have no problem with long-term relief because they argue that the only reason to perform a nerve root block or facet block is to define the pain generator for the surgeon and then to ensure a better ultimate surgical outcome. This really is not a bad goal and may actually be the most important use of the procedures after all.