Evidence Base for Destructive Procedures

Published on 26/03/2015 by admin

Filed under Neurosurgery

Last modified 26/03/2015

Print this page

rate 1 star rate 2 star rate 3 star rate 4 star rate 5 star
Your rating: none, Average: 0 (0 votes)

This article have been viewed 3013 times

CHAPTER 169 Evidence Base for Destructive Procedures

The surgical treatment of pain has been an integral part of neurosurgery since the early 20th century when Harvey Cushing pioneered ganglionectomy for trigeminal neuralgia (TN). Over the ensuing years as anatomic and physiologic knowledge of pain systems grew, new techniques aimed at new targets were developed for an array of pain conditions.14 Targets have included the anterolateral quadrant of the spinal cord, the sensory and sympathetic ganglia, the dorsal and ventral spinal roots, the dorsal root entry zone (DREZ), and decussating extralemniscal fibers, as well as the hypophysis, thalamus, and cingulate gyrus. Careful detailed descriptions of these techniques and their purported efficacy from pioneers in the field established the importance of surgery for the treatment of pain. The invasive nature of these techniques and associated comorbid conditions, as well as the high degree of technical skill required for their use, led to the development and widespread adoption of newer pharmaceuticals and nondestructive treatments. As a result, the use of destructive techniques for the surgical treatment of pain has dramatically declined over the past few decades. They are being or have been replaced with nondestructive techniques, such as spinal cord stimulators and intrathecal opiate pumps, newer medications, and high-dose opiates. However, it is becoming increasingly apparent that none of these techniques are free of their own set of detriments. For example, the chronic administration of opioids is associated with addiction, opioid-induced hyperalgesia, cognitive disorders, high cost, and suppression of the immune and reproductive systems.5,6 Furthermore, it has yet to be proved that the long-term use of opioids improves such outcomes as pain relief, functional capacity, and health-related quality of life.7,8

Neurostimulation and intrathecal drug administration are expensive and require ongoing battery and catheter maintenance and pump refills.9 Despite their widespread use and study, clinical efficacy has not been demonstrated according to modern standards of evidence-based medicine.10

Given that all three broad approaches to the treatment of pain (destructive, nondestructive, and medical) have “matured” and the initial wide-eyed enthusiasm of their introduction is over, it now seems timely to revisit the existing literature and see how the older destructive techniques compare. By so doing we can develop and implement the techniques that are effective and relegate to history those that are not.

In this chapter we review the published data on destructive procedures for both malignant and nonmalignant pain to determine the degree of evidence supporting continued use and to help define areas that warrant further study. Therefore, we reviewed human clinical studies that reported outcomes for destructive techniques used in the treatment of nonmalignant pain conditions. Reviewed studies were grouped according to the surgical target, beginning with peripheral or “first-order neuron” targets.11


In total, 149 papers on noncancer pain and 124 papers on cancer pain were reviewed; a few articles were included in both noncancer and cancer pain reviews. The large majority of studies constituted class III evidence. Only 18 class I or class II studies were found. Thirteen of these studies evaluated radiofrequency (RF) rhizotomy for different causes of pain such as lumbar facet syndrome, cervical facet pain, and type I (typical) TN, whereas two examined ganglionectomy for either lumbar or cervical pain. Two studies, one class I and one class II, evaluated and supported sympathectomy for visceral cancer pain. One evaluated the use of cordotomy in malignant cancer states.


Our database search and secondary review of references located 41 rhizotomy studies that in general dealt with spinal or facet pain and facial pain syndromes (imageTables 169-1 to 169-5), and 10 papers that addressed cancer pain (imageTable 169-6). The majority of the studies (29/41) reported data relevant to either TN (imageTable 169-1, 20 studies1425,2633) or lumbar facet syndrome (imageTable 169-2, 14 studies3447), whereas the remainder evaluated the effects of rhizotomy on a variety of truncal and extremity neuralgias (imageTable 169-3, 1 study48), cervical pain (imageTable 169-4, 9 studies4957), cluster headaches (imageTable 169-5, 3 studies5860), or cancer pain (imageTable 169-6, 10 studies44,6169) of various causes.

Rhizotomy for Truncal or Extremity Neuralgia and Lumbar Facet Syndrome

Initial success rates were around 60% but declined on long-term follow-up. Despite the variability in treated pain syndromes, number of sectioned roots, outcome measures, and follow-up, the results were sufficiently discouraging that the procedure was relatively abandoned and replaced with modified rhizotomies directed at facet denervation for the treatment of lumbar facet syndrome.44

Four randomized, blinded controlled trials compared RF rhizotomy of the medial branch of the dorsal root with sham treatment (3 studies) or intrafacetal versus extrafacetal rhizotomy (1 study). Outcome measures and patient selection differed sufficiently between the groups to preclude meta-analysis.70 Modest long-term improvement in pain scores in the treated group was observed in all three studies comparing rhizotomy with sham treatment. In the last study, intrafacetal rhizotomy was superior to extrafacetal procedures. Complications were minimal in all four studies. Of the remaining 14 papers, 4 were prospective. The proportion of patients with “good outcomes” varied from 41% to 75% in long-term follow-up. The remaining 7 series are classified as class III studies and reported long-term success rates of 40% to 60% (see imageTable 169-2, 14 studies3447).

Rhizotomy for Chronic Diskogenic Back Pain

One RCT (see imageTable 169-3) evaluated the effects of RF lesions of the ramus communicans for the treatment of single-level chronic diskogenic back pain.48 All selected patients had previously failed an intradiskal electrothermal procedure and had a good response to local nerve blockade of the ramus communicans. At 1 and 4 months of follow-up, the treated group showed significant improvement in visual analog scale (VAS) scores and the medical outcomes study short-form health survey (MOS 36).

Rhizotomy for Cervical Pain

RF rhizotomy was compared with sham procedures for the treatment of cervical pain or cervicogenic headache in three RCTs.50,56,57 One prospective class II clinical trial51 and four case series also addressed rhizotomy for cervical pain. The number of patients in each RCT was small, ranging from 12 to 24.56 Two of the RCTs demonstrated a significant treatment effect for RF lesions,56,57 although one did not (an underpowered study).50 The class II study also reported durable reduction in validated pain scores.50 Both studies with long-term follow-up demonstrated similar median time to recurrence51,56 (see imageTable 169-4, nine studies4957).

Rhizotomy and Neurectomy for Cluster Headache and Facial Pain

Two retrospective studies (19 patients)59,60 and one prospective study (18 patients) were reviewed. In the prospective study, 83% of patients had immediate pain relief but also demonstrated a 39% recurrence rate at long-term follow-up58 (see imageTable 169-5, three studies5860).

Rhizotomy for Cancer Pain

Ten case series were identified and the majority evaluated the effects of dorsal root rhizotomy for malignant pain of the extremities. The largest series contained 71 patients and was one of the earliest papers.44 Reported outcomes were generally favorable initially but faded quickly with time. The open nature of the procedure, the fading effect, and the presence of more effective procedures all contributed to abandonment of this procedure (see imageTable 169-6, 10 studies44,6169).


A total of 17 articles were found that met the selection criteria indicated earlier. All studies addressed benign pain conditions. Two of these studies were RCTs: one compared RF lumbar ganglionectomy with sham surgery for sciatica,72 and the other compared RF lesions made at two different temperatures.73 The remaining articles were case series ranging in size from 3 to 102 patients (imageTable 169-7, 17 studies41,7287).

Ganglionectomy for Lumbar and Cervical Radicular Pain

Geurts and colleagues in 2003 randomized 83 patients to either RF lesioning (45 patients) or needle placement without RF lesioning (38 patients) of selected lumbar ganglia.72 Evaluators and patients were blinded to the protocol used. Despite adequate power, no statistical difference was found between the groups.

Slappendel and associates in 1997 found no difference in outcomes between patients receiving an RF lesion at 60° C (group I) or 40° C (group II) for cervicobrachialgia.73 In a case series of 61 patients with sciatica,79 59% had “markedly reduced” postoperative pain at follow-up ranging between 1 month and 15 years. Definite conclusions concerning long-term outcomes are difficult given the variability in follow-up data. The remaining studies were small and inconsistent.

Ganglionectomy for Occipital Neuralgia

The largest case series evaluated the efficacy of C2 ganglionectomy for occipital neuralgia.75 These authors found that 80% of patients had significant relief from symptoms. Unfortunately, long-term follow-up data were unclear, thus rendering any conclusions difficult. The second largest series of patients with occipital neuralgia compared ganglionectomy performed on patients with occipital pain described as either “sharp, burning, jabbing, electrical, or exploding” (group I) or “dull, aching, throbbing, or pressure-like” (group II).76 Patients in group I tended to have a higher prevalence of a traumatic history (74%) and had the best response, nearly 80%. Overall, group II patients had a poor response. Follow-up was between 19 and 48 months. Other published series on occipital neuralgia were small and without standardized follow-up data.

One retrospective paper evaluated the effects of occipital neurectomy for the treatment of occipital neuralgia of various causes.87

Ganglionectomy for Other Causes

One study evaluated sphenopalatine ganglionectomy for cluster headache and found that patients with intermittent headache (group A) had a 67% response rate whereas chronic, continuous headache sufferers had only a 24% response.41 Mean follow-up was 24 and 29 months, respectively.

Two small series, one with 4 patients82 and another with 10 patients,86 reported patients treated by thoracic ganglionectomy for a variety of symptoms. Both claimed some success. Ganglionectomy articles are listed in image Table 169-7 (17 articles41,7287).

Dorsal Root Entry Zone Lesions

A total of 31 (26 noncancer pain, 5 cancer pain) case series were obtained for DREZ lesions. No class I or II studies were found. These reports included between 3 and 124 patients and had a wide range of follow-up time points. For noncancer pain, all studies reported greater than 50% relief of pain in the majority of patients, and these results tended to be durable.88 Patients with brachial plexus avulsion (BPA) injuries and traumatic spinal cord injury tended to have the best response to DREZ lesions in studies that compared different initial causes directly, whereas patients with postherpetic neuralgia and peripheral neuropathy had the worst response.89,90 Outcomes for patients with BPA injury were quite consistent across series, with 54% to 86% of patients having a greater than 50% reduction in reported pain levels. imageTable 169-8, 26 studies, is a summary of the articles obtained for DREZ lesions.88113

Only four (imageTables 169-9 and 169-10) papers were identified that addressed the use of DREZ lesions for cancer pain.114117 Two were case series that claimed long-lasting pain relief in patients with mixed cancer pain causes. Specific follow-up and outcome measures were not reported. For facial pain from craniofacial malignancies, three of the four papers were small case series that reported some benefit from DREZ lesions of the nucleus caudalis.114,116,117

Trigeminal Tractotomy for Facial Pain

Eight pertinent articles were found for tractotomy of noncancer pain, including the oldest article retrieved for this review.118 No class I data were found for this topic. Instead, all papers were case series. The largest case series was reported by Schvarcz in 1975, who found that excellent pain relief was achieved in 16 of 17 patients.119 Interestingly, all case series reported good to excellent results, even in patients with postherpetic neuralgia. In one series, for example, deep pain was eliminated or significantly decreased in 6 of 6 patients with postherpetic craniofacial dysesthesia.120 image Table 169-11, eight case series, presents an overview of the articles just reviewed.118125

Seven papers were found that involved tractotomy for cancer facial pain (imageTable 169-12).119,122,126130 All were cases series with one exception,126 an open label prospective study. In this prospective class II study, a standardized outcomes measure (VAS), Karnofsky Performance Scale, activities of daily living, and sleep duration were assessed at 6 months. The authors reported sustained relief of pain in 80% of patients at 6 months. Although the study was categorized as class II, the number of tractotomy patients included was limited (10 patients). We did not consider this sufficient to claim class II evidence for the efficacy of tractotomy in relieving facial cancer pain. The current literature supports only class III evidence for tractotomy for cancer pain.


A review of cordotomy for patients with noncancer pain produced 11 case series published between 1958 and 1990. No class I or II studies were found. The number of patients in these case series ranged from 13 to 72, and follow-up was between 1 month and 10 years. Outcome data from these studies generally demonstrated an overall improvement in patient-reported pain postoperatively over preoperative values.

In general, for studies with long-term follow-up, surgical benefits declined with time. However, determining significance was difficult because neither outcome nor follow-up was standardized for these patients. Furthermore, outcome measures were inconsistent among studies, thus complicating any attempts to make generalized recommendations. The articles that were included in this review are outlined in imageTable 169-13, 11 case series.100,131140

For cancer pain, cordotomy is the invasive procedure most often reported in the literature. Forty-seven papers qualified under our search criteria (imageTable 169-14).65,126,132,134136,141181 Many reports included more than 100 patients, many with long-term follow-up. A surge in publications followed the introduction of percutaneous cordotomy in the mid-1960s, followed by a gradual decline in published reports and patients in the late 1990s. Kanpolat introduced CT guidance for percutaneous cordotomy in the late 1980s, which was considered a major contribution to the field of pain surgery. However, intrathecal opioid use was also introduced at around the same time and gained widespread popularity.

One prospective trial was identified; this trial used a standardized outcome measure (VAS), Karnofsky Performance Scale, activities of daily living, and total sleeping hours. This prospective trial reported statistically significant improvement in all outcomes measures with respect to postprocedure and baseline pain levels. Three papers involved retrospective cohorts with survival analysis of pain relief of the whole cohort until death.64,158,159 Many other papers consisted of retrospective cohorts with large numbers (>100 patients) and 6 months’ follow-up.145,153,163,164,179 There were no RCTs.

Despite the heterogeneity of outcome measures, the vast majority reported excellent lasting relief within the context of cancer longevity. Reported cordotomy outcomes in cancer pain patients contrast with outcomes for noncancer pain patients, in whom pain relief was moderate, short-lived, and often complicated by dysesthesias.

Within the medical literature there is limited class II evidence for cordotomy as an effective treatment modality in cancer pain patients.


Three articles examining the use of myelotomy for noncancer pain were found. They were small case series (3 to 14 patients) with variable follow-up periods and causes of pain. The largest series reported that 64% of patients experienced a complete or marked reduction in deep “background” pain.182 imageTable 169-15, three case series, represents a summary of the articles.182184

For cancer pain, 19 papers were identified (imageTable 169-16).124,185202 These articles were all case series and included papers about commissural as well as extralemniscal myelotomy, both open and percutaneous. With the exception of 3 papers,198200 most reports involved a small number of patients (<20).

Outcome measures and follow-up periods were heterogeneous. In general, reported outcomes were less favorable when compared with cordotomy. However, a consistent observation was that myelotomy was most effective in treating midline visceral cancer pain syndromes (i.e., cervical, pancreatic, or gastric cancer).

The current literature supports only class III evidence for myelotomy for pain relief.


Using the criteria described earlier, nine relevant articles on the use of mesencephalotomy for noncancer pain were reviewed. All these articles were case series, and most had relatively small numbers of patients. The most recent publication reported the effects of lesions made in one of two locations in 27 patients with central pain after stroke.203 Long-term pain relief was achieved in 75% of patients with lesions created at the level of the superior colliculus, although significant ocular side effects occurred. These side effects were reduced with lesions adjacent to the inferior colliculus. Long-term pain relief was achieved in 58% of patients in this latter group. Follow-up ranged between 3 months and 5 years. In a separate study, good pain relief was reported in 23 patients (67%) with thalamic syndrome and tabes dorsalis monitored for 3 to 70 months.204 imageTable 169-17, nine case series, provides a summary of articles pertaining to mesencephalotomy.203211

Mesencephalotomy and pontine tractotomy for cancer pain were reported in seven papers (imageTables 169-18 and 169-19).207,212217 All were case series with variable outcome measures and follow-up. Outcome varied between extreme success (92% until death)215 to very poor success (1 of 12 had lasting pain relief).207 There was agreement on the frequency of side effects, specifically those related to ocular mobility.


All published articles describing thalamotomy were limited to small case series with heterogeneous patient populations. Similar to cingulotomy, the largest case series (N = 85) was not published in a peer-reviewed article but in book format218 and was not included here; however, a subset of patients was described earlier.219 Follow-up, outcome measures, target site, and cause of pain were inconsistent within or among series, thus making definite conclusions difficult. Dougherty and associates220 and Tasker221 have provided thorough discussions of thalamotomy. Please refer to image Table 169-20, 12 case series, for a summary of articles.133,205,219,222230

Sixteen papers on thalamotomy for cancer pain were found (imageTable 169-21)133,205,222,226,229,231241 (some articles are included in the noncancer pain tables). All were case series with variable outcome measures and follow-up. Most reported lasting benefits in more than 50% of patients. In some instances, bilateral pain relief was achieved with unilateral thalamotomy. Effects tended to fade with time and were often accompanied by persistent psychiatric complications.238


Buy Membership for Neurosurgery Category to continue reading. Learn more here