Written by B. Lambden, MD | 06 August 2009

Radiofrequency neurotomy is a form of nerve ablation (destruction of the nerve). These procedures have been utilized by neurosurgeons on many of the major nerves and ganglions over the past century with marginal success. This includes ablation of portions of the spinal cord, sensory and sympathetic ganglia, spinal roots as well as peripheral nerves and cranial nerves. Despite the long history and initial reports of success by some physicians, the use of destructive techniques has generally declined in modern practice, except for a few notable exceptions (Cites 2008) since their efficacy has not been well established by modern research standards.

One of the most difficult areas of medicine to study is chronic pain. This is due to the fact that chronic pain is a complex phenomenon, which is associated not only with possible peripheral damage or hypersensitivity, but also over time, with emotional factors and central plasticity. I am not going to discuss the complex role emotions play in chronic pain and simply suggest that neurologically, chronic pain is exceedingly complex. The neuronal signals are significantly modulated at multiple sites in the neural axis, resulting in changes in sensation, affect and function. In the case of medial branch neurolysis, the destruction of the nerve results in complex neuronal changes, not only in the nerve but also centrally in the spinal cord and brain which are very complex.

Given the above complexities, most physicians understand the necessity for extremely high level study design when performing research on pain. This is due not only to the complexities of pain, but also the fact that pain in and of itself is subjective in nature and is frequently, by definition, not even associated with underlying significant tissue damage. Consequently, pain studies are prone to significant confounding variables such as placebo effect, patient expectation, observer biases, as well as difficulty with technical aspects of research design, such as patient selection, diagnostic criteria and outcome measures. Given these obstacles, all physicians, when analyzing medical literature must be very selective regarding the literature reviewed.

As it relates to Z-joint pain, the obstacles to overcome for valid research are significant and perhaps even more difficult than other pain research, given that there are no known specific markers of facet mediated pain, including clinical examination or diagnostic studies. The diagnosis of Z-joint pain at this time relies completely on a patient's subjective reports of pain, utilizing differential blocks of the medial branch nerves. While these blocks are better than relying on clinical examination, they are not, by any means, fool proof and in my opinion have not been fully validated. They still rely on patient's response and can be complicated by technical difficulties, placebo response, patient expectations, as well as the known fact that pain reduction can outlast the theoretical duration of the local anesthetic, all of which makes these blocks less useful than most practitioners realize. Consequently, if one cannot be confident in the underlying diagnosis of Z-joint pain, research is meaningless. Obviously, a significant obstacle and challenge is proper patient selection for these types of procedures.

A practicing physician can only rely on the body of medical literature to make informed decisions as to the efficacy of radiofrequency neurotomy. As with all musculoskeletal conditions, again, one must be very selective reading the literature reviewed which opens up a whole new set of complications and difficulties given the volume of research to review. As mentioned previously, given the complexities of pain, treating physicians should rely only on high quality research meaning double blinded, randomized clinical trials. Fortunately, rather than waiting through large volumes of poor quality studies, there are several sources to sort out quality studies. One is the Cochrane Collaboration Back Review Group, although one may also utilize the Agency for Healthcare Research and Quality (AHRQ), a system to rate the strength of scientific evidence when assessing individual publications. Unfortunately, as we will see, all of the studies, as it relates to Z-joint pain have been limited by significant methodological short-comings. It is of interest that in 2003, Niemisto et al headed research for the Cochrane Database of Systematic Reviews, as it relates to radiofrequency denervation for neck and back pain. They found 661 citations for radiofrequency procedures, which when assessed by their stringent criteria, resulted in only 9 articles deemed to be valid and reliable. Of these, there was only one study on cervical medial branch radiofrequency neurolysis. The Niemisto study was updated and republished in August, 2008, however there were no additional studies added to the review. Their conclusion was that there was limited evidence that radiofrequency neurolysis offers short-term relief of chronic neck pain of facet origin. The study quoted by the Cochrane group was by Lord in 1996, published in the New England Journal of Medicine. This was a randomized, controlled trial with a population of 24, involving auto accident victims with chronic neck pain. Patients were identified by double-blinded, placebo controlled, local anesthetic procedures, of which the diagnosis was confirmed only if symptoms completely resolved with anesthesia and no change with normal saline injection. The control group was treated with Sham electrodes without lesioning. The data showed that 75% of the treated patients were pain free at Day 0, although 50% of the control patients were also pain free at Day 0. More significantly, 7/12 patients were pain free at 27 weeks, whereas only 1/12 control patients were pain free at 27 weeks. The average time to return to pre-operative level pain was about 263 days in the treated group and 8 days in the control group. The reviewers felt there was limited evidence for short-term benefit utilizing medial branch radiofrequency neurolysis. The problems were that this study had a small patient population and there was lack of replication of the study in another setting. Unfortunately there have been no other good studies of this procedure in the cervical spine, which is amazing given the number of radiofrequency procedures being performed today

Given the absence of high quality cervical studies, I did review lumbar studies, hoping for further validation of the procedure. In 2005, there was a randomized, double-blind Sham study by Vanwijk of lumbar radiofrequency neurolysis. This study found no difference in success between the patients that were treated with radiofrequency versus Sham procedures. Unfortunately, there were some methodological errors in the study in that they did not utilize diagnostic blocks, and the needle placement was less than optimal as it was placed according to Dutch techniques instead of ISIS recommended technique, resulting in less likelihood of significant nerve ablation. Nevertheless, it is interesting in this study that although there was no difference between the Sham and radiofrequency group, both groups reported reductions in pain. It was found that the placebo success rate was 29%, again emphasizing the complexities involved in pain research and the necessity for randomized, double-blinded controlled studies of large numbers.

In the absence of numerous high quality studies, we must, by necessity rely on non-randomized observation studies which are less than optimal. Nevertheless, some of these studies were of reasonable design and do show very significant responses to radiofrequency medial branch neurolysis. Dreyfuss reported 60% improvement in 80% of his patients at one year and McDonald reported 219 days of relief and 80% reduction in headaches in cervical, radiofrequency neurolysis patients. Boswell in Pain Physician 2007, provided a good review of the medical literature. He had previously authored a report in January, 2005 concluding that facet injections were variably effective for short term and long term relief of facet pain. His 2007 study is an update via a systematic review of randomized controlled trials, however there were no new controlled trials involving cervical radiofrequency neurolysis. He, however concluded that based on Lords 1996 study and McDonald's 1999 non-randomized study as well as Barnes 2005 nonrandomized study of the lumbar spine, that if one follows their procedures of multiple lesions of each facet nerve, that the evidence is strong for short and long-term relief with cervical medial branch neurotomy.

Gofeld performed a 10 year prospective audit of his lumbar radiofrequency neurolysis procedures. Again, this is not a randomized, controlled trial or even Sham controlled, but did have a population of 174. He reported 68% of patients had 50% decreased pain, lasting 6 to 24 months.

In summary, all of the high quality randomized clinical trials show limited evidence for pain relief, with medial branch radiofrequency neurolysis. Unfortunately, there is only one study to rely on. The study by Lord, as well as Vanwijk, which showed no difference between radiofrequency and Sham, were discounted by Dr. Nicoli Bogduk ( a pioneer in Z-joint pain and RF Procedures) based on his opinion that inaccurate surgical techniques used (Medical Journal of Australia, 2004 181 (1): 55-56). He clarifies his statement by stating that zygapophyseal joint pain can be relieved by radiofrequency neurotomy provided that correct techniques are used. He references VanKleef in Spine 1999 and Dreyfuss in Spine 2000, showing good treatment outcomes, although both of these studies are of poor quality.

As is often the case in pain medicine, we are left with very few high quality studies which show limited efficacy and a large number of poor quality studies, typically performed by physicians with a vested interest in the procedure, showing moderate to good outcomes. We are basically functioning in a gray area of medicine and there is no consensus, one way or another, regarding the efficacy of radiofrequency neurolysis for chronic Z-joint pain. The best we can do is work within the constraints of our limited knowledge. Dr. Bogduk, in the Spine Journal 2008, in his article titled Evidence Informed Management of Chronic Low Back Pain with Facet Injections and Radiofrequency Neurotomy provides some reasonably good suggestions. His article provides a very good historical description of Z-joint pathology and treatment. He points out the importance of utilizing diagnostic blocks with evaluation of the patient two hours after the block, or until relief ceases, whichever occurs first. He looks for preferably complete pain relief with the diagnostic block, or at the very least, 80%. He notes that 50% relief of pain is contentious and is at best a spurious response and cannot be held as a positive response to the diagnostic block. He also points out the fact that a series of three blocks are mandatory and that a single diagnostic block is not valid and carries a false positive rate between 25 and 41%. He stresses that the blocks are done to protect normal patients from undergoing radiofrequency neurolysis that do not need it. He also points out that proper electrode placement; according to ISIS technique is mandatory.

GENERAL RECOMMENDATIONS:

There is clearly a need for multiple modern, blinded, controlled randomized clinical trials, utilizing proper patient selection/electrode placement with larger populations to determine the efficacy of radiofrequency neurolysis and help define appropriate indications as well as methodologies. In the meantime, based on one somewhat adequate controlled clinical trial and multiple observational studies, it appears that radiofrequency neurolysis may be useful in some patients but this remains controversial. Therefore, given the lack of data suggesting long-term efficacy, one should be cautious and only consider the procedure after failed multi-disciplinary treatment, including aggressive medical and pharmacologic management, physical treatment and non-destructive interventional treatments. At the very least, treatment should include six months of multi-disciplinary treatment under the direction of a physician skilled in delayed recovery.

Once the patient is deemed to be a radiofrequency neurolysis candidate, they should be fully informed prior to the procedure of the risks and complications, including:

A. The patient should be made aware that although this is a generally safe, low risk procedure, there are still risks including improper electrode placement, resulting in weakness and sensory changes, as well as potential future complications such as neuritis.

B. The patient should be made aware that currently this is not a permanent solution and that typically pain lasts only 3 to 12 months and can require several additional procedures.

C. Long-term complications are unknown, but could include accelerated facet joint degeneration, neuroma formation and possibly permanent medial branch nerve hypersensitivity.

D. The patient should be made well aware that nerve ablation results in permanent changes to the nerve. Furthermore, most nerve ablation procedures performed over the past 50 years have ultimately been abandoned, due to lack of clinical efficacy.

As can be seen in the above narrative, the decision as to whether or not to pursue radiofrequency neurolysis is not simple and straight-forward. There is significant confusion regarding patient selection as well as whether or not the technique is truly efficacious. Given the above analysis and confusion, I believe that the recommendations from Dr. Bogduk should be followed closely, including requirements for a series of three diagnostic blocks, requiring at least 80% resolution of pain, and then following ISIS procedures closely. In addition, the procedure should only be performed after failing aggressive multi-disciplinary treatment.