Lumbar facet arthropathy accounts for a significant portion of cases of chronic axial low back pain (LBP), with prevalence rates estimated at between 10% and 15%.1 Not surprisingly, facet interventions are the second most commonly performed procedures in pain clinics throughout the United States.2 Among the various types of interventions, the strongest evidence for sustained benefit is for radiofrequency (RF) denervation of the medial branch nerves innervating the zygapophysial joints.1,3,4
RF exerts its beneficial effect by purportedly ablating the afferent nociceptive input from pain-generating structures. First used for lumbar zygaopophysial joint pain in the early 1970s,5 the technique is supported by several controlled and many uncontrolled studies.1,4,6–9 However, the concept has more recently been called into question after the publication of 2 negative controlled studies.10,11 Nevertheless, the ensuing controversy has led to an increased emphasis on refining selection criteria and technique.12–14
Similar to other nerves, there can be significant intra- and interperson variability in the locations of the medial branches of the primary dorsal rami.15,16 Because both the diameter of the target nerve (<2mm) and RF lesion radius (1.1 to 2.2 mm for conventional RF depending on the electrode size) are small,17 ensuring close proximity between the electrode and nervous tissue is essential to maximizing the likelihood of success. Although a few investigators have used contraction of the multifidus muscle as the primary criterion for lesioning,7,18 this approach may be limited by several confounding factors including differential nerve fiber innervation (i.e., the zygaopophysial joints are innervated by sensory fibers, whereas the multifidus muscle is innervated by larger motor neurons), multifidus muscle atrophy associated with chronic LBP, diminished twitch response at the lower lumbar levels most commonly targeted (e.g., L5), prolonged multifidus denervation after RF that may limit the efficacy of repeat procedures, and inability to observe a contraction in obese patients.19–22 On a similar note, empirically performing multiple lesions on the basis of “standard” anatomical relationships, though successful in several controlled studies,6,9 is unlikely to demonstrate utility in patients with aberrant anatomy, and is associated with greater tissue trauma and increased costs.
This combination of factors has led the majority of clinicians to use sensory stimulation to ensure adequate proximity to the target structure.1,8,10–12 However, sensory stimulation is essentially a surrogate measure for proximity, and may thus be influenced by numerous factors such as genetics, age, comorbid conditions (e.g., diabetes), medication usage, psychology, sedation, and underlying pathology. Not withstanding the possible effect some of these variables (e.g., genetics, obesity, and age) could have on anatomical pathology, these factors should otherwise have little direct relationship to the technical success rate of neuroablation.23 Despite the frequency with which sensory testing is used before RF denervation, no study has sought to correlate stimulation threshold with physical proximity to the target nerve. The purpose of this study was to determine whether there is any association between sensory stimulation threshold and outcome in patients undergoing lumbar facet RF denervation.
Permission to conduct this study was granted by the Internal Review Boards at Walter Reed Army Medical Center and Johns Hopkins Medical Institutions, as well as by all participants who provided informed written consent. All procedures and follow-up visits were performed between January 2007 and May 2010.
Inclusion criteria were age ≥18 years, predominantly axial LBP ≥3 months in duration, failure to respond to more conservative therapy, and paraspinal tenderness. Exclusion criteria were as follows: magnetic resonance imaging–confirmed specific etiology for LBP (e.g., significant spinal stenosis or grade II or III spondylolisthesis), focal neurological signs or symptoms, a positive response to previous nonzygaopophysial joint spine interventions (e.g., epidural steroid injections), previous facet interventions, lumbar spine fusion, untreated coagulopathy, and comorbid medical or psychiatric condition (e.g., untreated depression) likely to influence outcome.
Diagnostic Facet Blocks
Diagnostic medial branch blocks (MBB) were performed in an outpatient setting with the judicious use of superficial local anesthesia and no sedation. The targeted levels were chosen on the basis of fluoroscopic localization of paraspinal tenderness. Subjects with predominantly unilateral pain underwent 1-sided blocks, whereas those with midline or bilateral paraspinal pain underwent bilateral procedures. For MBB, the needletip was positioned midway between the superomedial aspect of the transverse process at its intersection with the superior articular process, and the site at which the mamilloaccessory ligament lies.24 For procedures aimed at the L5 dorsal ramus, the target point was the groove between the sacral ala and S1 articular process. Correct positioning for each block was confirmed using anteroposterior, oblique, and lateral fluoroscopy, negative aspiration, and real-time contrast injection. After the pattern of constrast flow was deemed adequate, 0.5 mL of 0.5% bupivacaine was injected at each site.
Pain Relief Rating
In the recovery area, all subjects were instructed to engage in their normal activities, disregard procedure-related pain, and maintain a written pain diary every 30 minutes for the ensuing 8 hours. In addition to containing serial 0 to 10 numerical pain rating scales (NRS), diaries were used to monitor postblock activities. To control for coexistent pain generators, a positive block was predesignated as ≥50% pain relief from baseline. However, 6 patients who reported slightly <50% pain relief but obtained significant functional benefit per pain diary activity logs and expressed satisfaction with the results also proceeded to RF. Failure to return the pain diary in a prompt and informative manner was grounds for exclusion.
Radio-Frequency Denervation and Electrical Stimulation
All denervation procedures were done within 30 days of the diagnostic block unless extenuating circumstances dictated otherwise. RF lesioning was conducted in accordance with our previous published technique using superficial local anesthetic and light sedation as necessary.25 Patients were instructed in the holding area beforehand regarding what sensation(s) they might experience during electrical stimulation (e.g., pressure, vibration, tingling) and the necessity of informing us as soon as they perceived anything different in the operative area. For medial branch nerves, 20-gauge curved RF needles with 10-mm active tips (BMC RF Cannula, Baylis Medical, Montreal, Quebec, Canada) were inserted in an oblique, cephalad coaxial view parallel to the target nerve until bone was contacted at the superomedial aspect of the transverse processes, and walked slightly cephalad along the lateral neck of the superior articular process. For procedures involving the L5 dorsal rami, the needletip was positioned in the groove between the lateral aspect of the S1 articular process and the sacral ala, with the convex surface of the electrode facing bone. Once needle placement was judged to be satisfactory, sensory electrostimulation was commenced at 0.1 V and slowly titrated up in increments of 0.1 V until the patient experienced a new sensation(s). At each target site, electrodes were incrementally adjusted in each direction to maximize sensory stimulation and to obtain the lowest stimulation threshold. Once obtained, the sensory stimulation threshold was confirmed by restimulating at each level. Although several patients failed to perceive sensory stimulation at <0.5 V, this was deemed acceptable only after at least 3 attempts to obtain lower stimulation failed. Mean stimulation threshold was derived from averaging the stimulation threshold for each nerve lesioned. The maximum stimulation threshold used in calculations of the mean for any nerve was 1.0 to mitigate the effect of outliers (i.e., a stimulation threshold of 2.0 was counted as 1.0). After optimal stimulation was obtained, multifidus muscle stimulation and the absence of leg contractions were verified with electrostimulation at 2 Hz. After satisfactory electrode placement, 0.5 mL of lidocaine 2% with 5 mg of depomethylprednisolone was injected through each cannula to minimize procedural pain, prevent neuritis, and enhance lesion size.14,26,27 The RF probe was then reinserted and a 90-second, 80°C lesion was made using an RF generator (Electrothermal 20S Spine System, Smith and Nephew, Andover, MA; Baylis Medical Pain Management Generator 115V, Baylis Medical).
Outcome Measures and Follow-Up
Baseline demographic and clinical data were recorded before the diagnostic MBB. Information collected included age, gender, duration of pain, active duty status for patients treated at military hospitals, opioid and other analgesic usage, previous decompression surgery, 0 to 10 NRS pain scores at rest and with activity for the week preceding the block, and Oswestry disability index (ODI) score (version 2.0, MODEMS, Des Plaines, IL). The ODI is one of the most commonly used instruments for assessing functional capacity in patients with LBP, having been validated in multiple languages and contexts. Scores of 0%–20% signify minimal disability, 21%–40% moderate disability, 41%–60% severe disability, and >60% indicates a patient who is crippled or bedbound because of back pain.28 The primary independent variable, mean sensory stimulation threshold, was calculated from data recorded for each targeted nerve at the time of the RF procedure. In the event that stimulation was not appreciated at <1.0 V, 1.0 V was predesignated to be recorded as the “default” threshold.
Follow-up visits were performed by a disinterested observer with no knowledge of stimulation records at 1 and 3 months, with no contact permitted between the patient and any investigator in the interim. A successful primary dichotomous outcome was predefined as a ≥50% decrease in either rest or activity pain lasting ≥3 months, coupled with a positive global perceived effect that precluded additional procedural interventions.25 All patients with an interval successful outcome at their 1-month follow-up visit were evaluated at 3 months, whereas subjects who failed to obtain adequate relief at their initial follow-up, or whose relief dissipated between visits, were considered treatment failures. With regard to 3-month follow-up, only patients with a positive outcome at their initial visit who elected to forego treatment had data recorded.
As per previous studies, “medication reduction” was predetermined to be either a ≥20% reduction in opioid use or complete cessation of a nonopioid analgesic.29 Global perceived effect, a measure of satisfaction, was predefined as an affirmative response to the following 2 questions:
My pain has improved/worsened/stayed the same since my last visit.
I am satisfied/not satisfied with the treatment I received and would recommend it to others.
Because this study was conducted at both military and civilian institutions, and service members are subject to unique physical (e.g., heavy load carriages) and psychosocial stressors, clinical variables and outcome measures were calculated separately for each cohort, as well as collectively.
Statistical analyses were done using Stata 11.1. Continuous variables were analyzed using the t test with unequal variances and the Wilcoxon ranked sum test. Categorical variables were assessed using the χ2 test or Fisher exact test as appropriate. Pairwise Pearson correlation coefficients with confidence intervals and scatter graphs were used to estimate the relationship between the mean sensory stimulation threshold during RF denervation and the pain relief obtained after the procedure. In addition, a receiver operating characteristic (ROC) curve was constructed to estimate sensitivity and 1 specificity for each cutoff value for sensory stimulation threshold. For the ROC curve, a successful outcome was used as the reference variable, reflecting the true state of the condition (i.e., if the person had facet disease that was responsive to treatment, success = 1), and the classification variable was the mean stimulation threshold used for the RF procedure. The area under the ROC curve and its 95% confidence interval (CI) were calculated. If the 95% CI included 0.5, the area under the ROC curve was judged not to be statistically significantly different from the area below the 45-degree reference line (i.e., test is noninformative). A multiple logistic regression model was also constructed to assess the relation between clinical and demographic characteristics of the study participants and successful outcome. All variables were entered into the model simultaneously, and all terms were retained in the model, regardless of statistical significance. A post hoc power analysis determined that 61 patients had 78% power to detect a small (r = 0.1) effect size, and 99% power for identifying a moderate (r = 0.3) correlation.
Baseline Demographic and Clinical Information
Informed consent was obtained from 143 individuals with suspected facet-mediated LBP. Seventy-three patients were excluded because they did not experience the requisite relief during their diagnostic MBB, 3 because they obtained sustained benefit that precluded the need for denervation, and 6 because they were lost to follow-up (n = 5) or their insurance company did not authorize the procedure (n = 1). This left 61 subjects eligible for inclusion in the study.
Demographic and clinical data sorted by military status are presented in Table 1. Baseline data were comparable between military (n = 26) and nonmilitary (n = 35) personnel except that service members were younger (mean age 40.0 years, SD 8.8 vs 58.7, SD 12.8; P < 0.001) and had lower disability scores than did civilians (30.7, SD 13.2 vs 37.4, SD 16.0; P = 0.08). In the overall cohort, the mean duration of pain was 6.5 years (SD 5.7). Fifty-nine percent of subjects were male, and 30% were receiving opioids. A substantial majority of persons (77%) had 2 levels (3 nerves) treated, with 61% receiving bilateral procedures.
Fifty-four percent of the study population experienced a positive dichotomous outcome, which did not differ significantly between military (58%) and nonmilitary persons (51%) (risk difference 0.06, 95% CI: −0.19 to 0.30, P = 0.63). Mean pain scores at rest and with activity declined to 2.0 (SD 1.7) at rest and 4.4 (SD 2.6) with activity at 1 month, and 2.5 (SD 2.0) at rest and 4.7 (SD 2.6) with activity at 3-month postprocedure, respectively. Functional capacity, as signified by mean improvement in ODI scores, decreased an average of 33.2% (SD 29.3) and 28.3% (SD 28.9) at 1- and 3-month follow-up visits, respectively.
Effect of Sensory Stimulation on Treatment Outcome
Treatment outcomes by categories of mean sensory stimulation threshold are shown in Table 2 and Figure 1. Among the 33 patients who experienced a positive dichotomous outcome, the mean stimulation threshold was 0.32 V (SD 0.14), vs 0.35 V (SD 0.18) in those who failed to achieve meaningful relief. Notably, none of the outcomes assessed was statistically significantly different among the various stimulation groups.
Two patients were unable to experience sensory stimulation at ≤1.0 V at 2 or more nerves. Neither of these patients had a positive outcome. When these patients were excluded from the analysis, the mean stimulation threshold was 0.32 (SD 0.12) in those with a negative outcome. In the 6 patients who underwent RF despite obtaining <50% pain relief during their diagnostic block, the mean stimulation threshold was 0.35 (SD 0.30). Only 1 of these patients experienced a positive outcome.
Table 3 and Figures 2 and 3 demonstrate the lack of a statistically significant relationship between the mean stimulation threshold and 3-month pain levels at rest (r = −0.01, 95% CI: −0.24 to 0.23, P = 0.97) and with activity (r = −0.17, 95% CI: −0.40 to 0.07, P = 0.20). In Figure 4, a receiver operating characteristic (ROC) curve revealed ≥0.17 V to be the “optimal” stimulation threshold, being 100% sensitive and 7.1% specific for a positive RF outcome, with an observed agreement (i.e., success rate) of 57.4%.
Factors Associated with Outcome
Table 4 shows factors associated with outcome in univariable and multivariable logistic regression. Neither preprocedure pain scores at rest (multivariable odds ratio [OR] 0.72, 95% CI: 0.38 to 1.36) nor with activity (OR 0.70, 95% CI: 0.38 to 1.29) were significantly associated with outcome. When confounding factors were controlled for, opioid use (OR 0.03, 95% CI: 0.002 to 0.69) and bilateral treatment (OR 0.14, 95% CI: 0.02 to 0.93) were associated with 97% and 86% increased likelihood of treatment failure, respectively. The mean sensory stimulation threshold was not statistically significantly associated with the outcome in either univariable or multiple regression models.
Facet joint interventions are the second most commonly performed pain intervention throughout the United States.2 Considering the significant economic costs associated with the management of chronic LBP and the frequency with which RF neurotomy is performed, there is significant interest in optimizing techniques to further enhance outcomes and control costs in the application of RF ablation. The main finding in this study is that no association was found between lower stimulation thresholds during medial branch sensory testing and positive treatment outcomes. This is somewhat surprising, because deductive reasoning suggests that if RF works via ablation of nerves supplying nociceptive input, closer proximity to those nerves increases the likelihood of technical success, and sensory threshold is a surrogate measure of proximity, then lower sensory thresholds should augur well for successful outcomes.
After the initial reports of excellent outcomes using RF neurotomy for the treatment of lumbar facet arthropathy by Shealy in the 1970s,5 subsequent anatomical studies demonstrated that commonly described electrode placement did not correlate well with the location and course of lumbar medial branches.30 A later study demonstrated that percutaneous RF lesions develop not distal to the tip of the electrode as previously thought, but along the long axis of the active tip.17 Collectively, these findings imply that other factors besides the proximity between electrode and nerve play a role in RF outcomes. Despite our improved understanding of lumbar spine anatomy and technical considerations for maximizing RF lesion size, low-amplitude sensory stimulation capture remains the most frequently used tool to ensure proximity of the RF needle to the lumbar medial branches.
The origin of the use of sensory stimulation before denervation stems from the original studies by Shealy,5 who described its use to “determine possible proximity of the …electrode tip to the spinal nerve root,” not to localize the targeted medial branches. In other words, sensory stimulation was originally used as a safety mechanism to prevent inadvertent lesioning of motor nerves. Notwithstanding the limited safety rationale for the use of sensory stimulation, it has nevertheless become commonplace.
Several studies noting the lack of consistent muscle stimulation have even called into question the utility of performing motor nerve stimulation to prevent accidental motor fiber ablation.31–33 Unfortunately, these studies were not designed to determine whether the subset of patients in whom motor stimulation failed to elicit the expected muscle contractions had a comorbidity that might have made their nerves less responsive to stimulation. Factors that might render a person less responsive to stimulation include male gender, older age, sedation, variations in genetics and phenotypic expression, ethnic and cultural backgrounds, medication usage, previous surgery, and diabetes and other systemic illnesses associated with neuropathy.32,34–36 Individual differences in sensory perception could translate into a sensory nerve stimulation threshold of 0.2 V in one person indicating equivalent proximity to a nerve as a stimulation threshold of 0.7 V in another person. Another confounding factor is that some patients with low perception thresholds may be actually experiencing local tissue stimulation rather than nerve stimulation. Some investigators have even proposed that nerve stimulation itself can produce a conduction block that may limit its utility (Table 5).37
Among randomized clinical trials, using sensory stimulation as the primary means to guide RF does not ostensibly seem to improve outcomes. Whereas the results of studies that have used sensory stimulation as the means to effect neuroablation have been decidedly mixed,8,10,11,38 those that have used other means to target medial branches (e.g., empirical lesioning based on anatomic landmarks, multifidus muscle contraction) have uniformly reported positive outcomes.6,7,9 However, direct comparisons among such studies are limited because those that have used nonsensory modalities to target nerves have tended to use stricter inclusion criteria, sounder study designs, and better technique.
There are several limitations to this study that should be considered. The main one is that all RF procedures were performed after optimizing sensory stimulation by repeated electrode adjustments. To place this in context with an example, this means that a mean sensory threshold of >0.5 V in an elderly male with diabetes might denote comparable proximity to the target nerves as a mean threshold of 0.2 V in a young, healthy female. The implications of this are that one cannot assume on the basis of our findings that a first-time stimulation threshold of >0.5 V should be accepted prima facie as evidence of sufficient proximity to the target to effect neuroablation, or will have a comparable success rate to denervation performed using a lower threshold. Instead, it should only be deemed acceptable if subsequent attempts to decrease the sensory stimulation threshold are unsuccessful.
A second flaw is that our sole independent variable was mean sensory stimulation threshold, which was derived by averaging the stimulation threshold of each nerve targeted. In our study, a 4-level stimulation sequence of 0.4 V, 0.4 V, 0.4 V, and 0.4 V would be indistinguishable from a sequence of 1.0 V, 0.4 V, 0.1 V, and 0.1 V. This may be justified because it is not possible to apportion pain relief by spinal level treated, nor can one know whether nerve ablation was successfully accomplished. Hence, one cannot distinguish between 50% pain relief that resulted from a 100% successful denervation in a patient in whom the lumbar zygapophysial joints account for 50% of their pain burden, or a 50% successful denervation in a patient in whom the facet joints are the sole source of back pain.
Finally, our results might have differed had we limited our inclusion criteria by treating only those patients who responded with ≥80% pain relief to confirmatory or placebo blocks, and excluding those at higher risk for treatment failure, such as people taking high-dose opioids, prior back surgery, secondary gain issues, and any psychiatric comorbidity. However, implementing these measures has been shown in clinical studies to be counterproductive (i.e., the increase in specificity is outweighed by the reduced “sensitivity.”12,25,39
In summary, our findings suggest that there is no meaningful correlation between mean sensory stimulation threshold and lumbar facet RF denervation outcomes, provided that sensory threshold was optimized by trial and error. Future studies should aim to determine whether using sensory stimulation, motor stimulation, or serial electrode placements based on standard anatomical landmarks is best for identifying RF lesion sites.
Name: Steven P. Cohen, MD.
Contribution: Study design, drafted protocol, patient recruitment, drafted manuscript, table and graph creation.
Name: Scott A. Strassels, PhD, PharmD.
Contribution: Drafted and edited manuscript, statistical analysis, table and graph creation.
Name: Connie Kurihara, RN.
Contribution: Chief research nurse, helped draft protocol, data collection and patient follow-up, edited manuscript.
Name: Ivan K. Lesnick, MD.
Contribution: Drafted and edited manuscript.
Name: Steven R. Hanling, MD.
Contribution: Drafted and edited manuscript.
Name: Scott R. Griffith, MD.
Contribution: Patient recruitment, edited manuscript.
Name: Chester C. Buckenmaier, III, MD.
Contribution: Medical monitor, funding, edited manuscript.
Name: Conner Nguyen, MD.
Contribution: Patient recruitment, edited manuscript.
This manuscript was handled by: Spencer S. Liu, MD.
1. Cohen SP, Raja SN. Pathogenesis, diagnosis and treatment of lumbar zygapophysial (facet) joint pain. Anesthesiology 2007;106:591–614
2. Manchikanti L. The growth of interventional pain management in the new millennium: a critical analysis of utilization in the Medicare population. Pain Physician 2004;7:465–82
3. Markman JD, Philip A. Interventional approaches to pain management. Anesthesiol Clin 2007;25:883–98
4. Bogduk N. Evidence-informed management of chronic low back pain with facet injections and radiofrequency neurotomy. Spine J 2008;8:56–64
5. Shealy CN. Percutaneous radiofrequency denervation of spinal facets: treatment for chronic back pain and sciatica. J Neurosurg 1975;43:448–51
6. Nath S, Nath CA, Pettersson K. Percutaneous lumbar zygapophysial (facet) joint neurotomy using radiofrequency current, in the management of chronic low back pain: a randomized double-blind trial. Spine 2008;33:1291–7
7. van Kleef M, Barendse GA, Kessels A, Voets HM, Weber WE, de Lange S. Randomized trial of radiofrequency lumbar facet denervation for chronic low back pain. Spine 1999;24:1937–42
8. Gallagher J, Petriccione di Vadi PL, Wedley JR, Hamann W, Ryan P, Chikanza I, Kirkham B, Price R, Watson MS, Grahame R, Wood S. Radiofrequency facet joint denervation in the treatment of low back pain: a prospective controlled double-blind study to assess its efficacy. Pain Clinic 1994;7:193–8
9. Lord SM, Barnsley L, Wallis BJ, McDonald GJ, Bogduk N. Percutaneous radio-frequency neurotomy for chronic cervical zygapophyseal-joint pain. N Engl J Med 1996;335:1721–6
10. Leclaire R, Fortin L, Lambert R, Bergeron YM, Rossignol M. Radiofrequency facet joint denervation in the treatment of low back pain. A placebo-controlled clinical trial to assess efficacy. Spine 2001;26:1411–7
11. van Wijk RM, Geurts JW, Wynne HJ, Hammink E, Buskens E, Lousberg R, Knape JT, Groen GJ. Radiofrequency denervation of lumbar facet joints in the treatment of chronic low back pain: a randomized, double-blind, sham lesion-controlled trial. Clin J Pain 2005;21:335–44
12. Cohen SP, Hurley RW, Christo PJ, Winkley J, Mohiuddin MM, Stojanovic MP. Clinical predictors of success and failure for lumbar facet radiofrequency denervation. Clin J Pain 2007;23:45–52
13. Lau P, Mercer S, Govind J, Bogduk N. The surgical anatomy of lumbar medial branch neurotomy (facet denervation). Pain Med 2004;5:289–98
14. Provenzano DA, Lasilla HC, Somers D. The effect of fluid injection on lesion size during radiofrequency treatment. Reg Anesth Pain Med 2010;35:338–42
15. Giles LG, Taylor JR. Human zygapophyseal joint capsule and synovial fold innervation. Br J Rheumatol 1987;26:93–8
16. Bogduk N, Wilson AS, Tynan W. The human lumbar dorsal rami. J Anat 1982;134:383–97
17. Bogduk N, Macintosh J, Marsland A. Technical limitations to the efficacy of radiofrequency neurotomy for spinal pain. Neurosurgery 1987;20:529–34
18. Dreyfuss P, Halbrook B, Pauza K, Joshi A, McLarty J, Bogduk N. Efficacy and validity of radiofrequency neurotomy for chronic lumbar zygapophysial joint pain. Spine 2000;25:1270–7
19. Kader D, Wardlaw D, Smith F. Correlation between the MRI changes in the lumbar multifidus muscles and leg pain. Clin Radiol 2000;55:145–9
20. Wallwork TL, Stanton WR, Freke M, Hides JA. The effect of chronic low back pain on size and contraction of the lumbar multifidus muscle. Manual Ther 2009;14:496–500
21. Windsor RE. Radiofrequency lumbar zygapophysial (facet) joint denervation: a preliminary report of a new concept. Pain Physician 2003;6:119–23
22. Dreyfuss P, Stout A, Aprill C, Pollei S, Johnson B, Bogduk N. The significance of multifidus atrophy after successful radiofrequency neurotomy for low back pain. PM R 2009;1:719–22
23. Siao P, Cros DP. Quantitative sensory testing. Phys Med Rehabil N Am 2003;14:261–86
24. Dreyfuss P, Schwarzer AC, Lau P, Bogduk N. Specificity of lumbar medial branch and L5 dorsal ramus blocks. A computed tomography study. Spine 1997;22:895–902
25. Cohen SP, Williams KA, Kurihara C, Nguyen C, Shields C, Kim P, Griffith SR, Larkin TM, Crooks M, Williams N, Strassels SA. Randomized, comparative cost-effectiveness study comparing 0, 1 and 2 medial branch blocks before lumbar facet radiofrequency denervation. Anesthesiology 2010;113:395–405
26. Dobrogowski J, Wrzosek A, Wordliczek J. Radiofrequency denervation with or without addition of pentoxifylline or methylprednisolone for chronic lumbar zygapophysial joint pain. Pharmacol Rep 2005;57:475–80
27. Bruners P, Müller H, Günther RW, Schmitz-Rode T, Mahnken AH. Fluid-modulated bipolar radiofrequency ablation: an ex-vivo evaluation study. Acta Radiol 2008;49:258–66
28. Fairbank JC, Pynsent PB. The Oswestry Disability Index. Spine 2000;25:2940–5
29. Cohen SP, Bogduk N, Buckenmaier CC 3rd, Dragovich A, Griffith S, Kurihara C, Raymond JL, Williams N, Yaksh TL. A double-blind, placebo-controlled, dose-response, and preclinical safety study evaluating transforaminal epidural etanercept for sciatica. Anesthesiology 2009;110:1116–26
30. Bogduk N, Long DM. The anatomy of the so-called “articular nerves” and their relationship to facet denervation and the treatment of low back pain. J Neurosurg 1979;51:172–7
31. Robards C, Hadzic A, Somasundaram L, Iwata T, Gadsden J, Xu D, Sala-Blanch X. Intraneural injection with low-current stimulation during popliteal sciatic nerve block. Anesth Analg 2009;109:673–7
32. Bigeleisen P, Moayeri N, Groen G. Extraneural versus intraneural stimulation thresholds during ultrasound-guided supraclavicular block. Anesthesiology 2009;110:1235–43
33. Tsai TP, Vuckovic I, Dilberovic F, Obhodzas M, Kapur E, Divanovic KA, Hadzic A. Intensity of the stimulating current may not be a reliable indicator of intraneural needle placement. Reg Anesth Pain Med 2008;33:207–10
34. Wiesenfeld-Hallin Z. Sex differences in pain perception. Gend Med 2005;2:137–45
35. Fillingim RB, King CD, Ribeiro-Dasilva MC, Rahim-Williams B, Riley JL 3rd. Sex, gender, and pain: a review of recent clinical and experimental findings. J Pain 2009;10:447–85
36. Tremblay J, Hamet P. Genetics of pain, opioids, and opioid responsiveness. Metabolism 2010;59(suppl 1):S5–8
37. Li J, Kong X, Gozani SN, Shi R, Borgens RB. Current-distance relationships for peripheral nerve stimulation localization. Anesth Analg 2011;112:236–41
38. Stovner LJ, Kolstad F, Helde G. Radiofrequency denervation of facet joints C2–C6 in cervicogenic headache: a randomized, double-blind, sham-controlled study. Cephalalgia 2004;24:821–30
39. Cohen SP, Stojanovic MP, Crooks M, Kim P, Shields CH, Schmidt RK, Hurley RW. Lumbar zygapophysial (facet) joint radiofrequency denervation success as a function of pain relief during diagnostic medial branch blocks: a multi-center analysis. Spine J 2008;8:498–504