Zygapophysial (facet) joint pain is a significant cause of chronic axial low back pain, accounting for 10% to 15% of such cases.1 Lumbar medial branch radiofrequency denervation (LMBRFD) is a validated treatment for lumbar facet joint pain and is also commonly used to treat chronic axial lower back pain.2–4 Approaches to LMBRFD are considered to have anatomically sufficient precision and to be methodologically acceptable if radiofrequency (RF) cannulae are placed adjacent and parallel to the target medial branches.
Several RF techniques have been reported to improve the outcome of LMBRFD.5 Among them, the “tunnel vision” approach has been widely recommended for precise RF needle placement when performing LMBRFD.5,6 There is no doubt that the tunnel vision approach is easy to use for variable interventional procedures by making the target point, the needletip, and the hub of the cannula as a “dot.” However, special attention is always required when performing LMBRFD using the tunnel vision approach, because perception of the needle depth is difficult to ascertain before obtaining a lateral image, and the needle can traverse or slip off a landing point. A deeper positioning of the needle when performing LMBRFD using the tunnel vision approach may cause inadvertent contact with critical structures such as the dorsal root ganglion or the ventral ramus exiting anterior to the target point.7
It has been suggested that lateral branches that provide sensory distribution to the skin overlying the spinous processes lie close to the L1 to L4 medial branches,8 and injury to the lateral branch during LMBRFD could result in severe anesthesia dolorosa over the buttock.9 At our institution, we use an alternative technique involving a “distal approach” to the L1 to L4 medial branches in the oblique fluoroscopic view.10 To avoid injuries to the lateral branch and the ventral ramus, we altered the RF cannula placement method by inserting the cannula in a straight ipsilateral oblique view, landing the needletip distal to the target point and advancing the cannula toward the target point while maintaining electric stimulation until the onset of paravertebral contractions. No study has been conducted to validate our distal approach by comparing it with the conventional technique for LMBRFD.
Accordingly, the aim of this study was to describe and assess the distal approach for L1 to L4 MBRFD. The effectiveness of our distal approach was evaluated using numeric rating scale (NRS) pain scores and directly compared with the conventional tunnel vision approach in a prospective, randomized trial.
This study was approved by the IRB of our university-based hospital and was registered with Clinical Trials (Ref. NCT01300715). Each subject provided written, informed consent before participating in the study. Before the initiation of LMBRFDs, the sample size was calculated to detect a 25% difference in 0 to 10 NRS pain score between groups from a baseline NRS score of 6.1 and a SD of 2.0. We chose this method for calculating sample size taking into consideration the change of NRS, which was suggested by our previous data11 and another similar study.12 Two-tailed power analysis determined that a sample size of 27 per group would have an 80% power (β of 0.2) with a significance level (α) of 0.05 and an effect size (d) of 0.79. We adjusted the sample size for an estimated follow-up loss rate of 5% at the 1-month and 30% at the 6-month postprocedure time points, which resulted in 40 patients in each group.
All LMBRFD procedures were performed in an outpatient setting using local anesthesia. The inclusion criteria were as follows: age >18 years, predominantly axial lower back pain for ≥3 months, paraspinal tenderness overlying the L2 to L4 lumbar facet joints, failure to respond to conservative therapy such as physical therapy or pharmacotherapy, and concordant pain relief of >50% after a comparative local anesthetic block with 0.5 mL lidocaine 1% (≥1 hour) and levobupivacaine hydrochloride (Chirocaine®, Abbott Korea, Seoul, South Korea) 0.5% (≥3 hours) at the L1 to L4 medial branches. The exclusion criteria were as follows: any focal neurologic sign or symptoms, radiologic evidence of a symptomatic herniated disc, severe spinal stenosis or structural lumbar spinal deformity, a positive response to previous spinal interventions such as sacroiliac joint blocks or epidural steroid injection, discogenic pain verified by discography, lumbar spine fusion, untreated coagulopathy, or a concomitant medical or psychiatric condition likely to undermine the diagnostic workup or the assessment of the treatment response.
Randomization and Intervention
Eighty-two patients were randomly allocated to 1 of the 2 study groups using the envelope method (Fig. 1). In the tunnel vision approach group (n = 41), LMBRFD was performed under fluoroscopic guidance in an oblique “tunneled” view, as described by Bogduk,13 whereas a fluoroscopic distal approach was used for the L1 to L4 medial branches in the distal approach group (n = 41).10
The following procedures were performed by 1 of 2 board-certified pain specialists with extensive experience in both approaches. No sedation was used during the procedures. With the patient in the prone position under fluoroscopic guidance, a 20-gauge curved RF needle (Inomed Medizintechnik GmbH, Teningen, Germany) with a 10-mm exposed tip was advanced to the target nerve at the involved level as confirmed by previous medial branch blocks.
Our distal approach is demonstrated in Figure 2A. With the C-arm, the vertebral end plates were aligned parallel in an anteroposterior (AP) projection. The C-arm was then turned ipsilateral oblique 15° to 25°. The target point was each junction of the superior border of the transverse process and the superior articular process (SAP). We inserted an RF needle with a 10-mm active tip from the skin entry site, which was determined at the upper margin of the vertebral body below the target vertebra, and introduced it toward the landing point in the center of the pedicle where it made contact with the bone (Fig. 2A). The RF needle was then carefully advanced toward the target point, while maintaining contact with the bone and administering 2 Hz and <0.5 V stimulation. RF needle advancement was stopped when paravertebral rhythmic contractions occurred. Once obtained, the electrode tip was slightly withdrawn until muscle contractions disappeared. Lateral fluoroscopic views were taken to see how proximal the electrode tip was to the middle two-fourths of the neck of the SAP (Fig. 2B). Sensory electrostimulation was then performed at 50 Hz and <1.0 V to confirm the proximity of the electrode to the sensory fibers without signs of stimulating the nerve root. We deemed it acceptable if sensory stimulation was reported at >0.25 and <1.0 V. When sensory stimulation was obtained at <0.25 V, the cannula was withdrawn a few millimeters. When sensory stimulation was obtained at >1.0 V, the cannula was rotated in each direction or advanced a few millimeters until stimulation was obtained at 1.0 V. The fluoroscopic image of final needle position with the distal approach is shown in Figure 3A. Once the needle placement was judged to be satisfactory after sensory electrostimulation, lateral views were rechecked to ensure that the RF needletip was not advanced beyond the middle two-fourths of the base of the SAP (Fig. 3B). An AP image was also obtained. If paravertebral rhythmic contractions were not elicited despite needle placement, it was judged to be satisfactory, and lesioning was performed after electrical stimulation for sensory testing.
The tunnel vision approach used a steep caudocephalad axial tilt of the fluoroscopy beam.5,13 Briefly, with the C-arm intensifier positioned in an ipsilateral oblique and sharp, caudocephalad direction in an orientation parallel to the course of the target nerve, the RF needle was inserted using the tunnel vision approach until bone contact was achieved at the junction between the superior proximal edge of the transverse process and the SAP (Fig. 2C). After making bone contact, the RF needletip was then advanced 3 to 5 mm further until bone contact was lost, then the lateral image was taken (Fig. 2D). The fluoroscopic image of the final needle position in the sublaminar oblique view with the tunnel vision approach is shown in Figure 3C. In the lateral view, the RF needletip should reach the anterior two-thirds of the base of the SAP (Fig. 3D). The needle position was confirmed in the AP view. Electrical stimulation at 50 Hz was then applied for sensory testing and at 2 Hz for motor stimulation.
The major differences between the 2 approaches are highlighted in Table 1. If patients were unable to experience either motor or sensory stimulation, broader lesions were made based on the radiographic anatomy. However, we excluded patients who were unable to experience either sensory or motor stimulation at ≥2 nerves.
Before activating the RF generator (Neuro N50, Stryker Leibinger GmbH & Co. KG, Freiburg, Germany), 1% lidocaine (0.5 mL) was injected to enhance lesion size,14 and impedance was confirmed. The lesions were made by raising the temperature of the RF needletip to 80°C for 90 seconds. The curved tip of the RF needle was then rotated cephalad and caudad by rotating the RF needle hub, and second and third lesions were made using the same variables. After LMBRFD, 5 mg of triamcinolone acetonide suspension (Tamcelon®, Hanall, Seoul, South Korea) was injected per segment through the needle. The time required to find each medial branch was defined as time taken to find the target medial branch by electrical stimulation.
Postprocedurally, patients were observed for any adverse events in the recovery room, where procedure-related pain was immediately assessed by a specially trained nurse independent of this study. Procedure-related pain was determined using the following question: “During the procedure, you might have felt painful sensations which were different from your baseline low back pain. What was the degree of pain experienced during the procedure on a scale of 0 to 10, with 0 being no pain at all and 10 being the worst pain you could imagine?”
The primary end point in the study was a comparison of the mean difference in the change of NRS scores of low back pain from entry to the scores at 1 month (NRS at baseline—NRS at 1 month) and 6 months (NRS at baseline—NRS at 6 months) between the distal approach group and the tunnel vision approach group. The NRS has well-documented reliability, validity, and sensitivity to treatments that are expected to affect pain.15,16
The secondary end point was a change of NRS and the Oswestry Disability Index 2.0 (ODI)17 over periods of time in each group. The ODI, which measures the functional activities of daily life likely to be limited in people with low back pain, has been shown to be valid and reliable in patients with mechanical low back pain.11,15,16 The 9-item Korean version of the ODI18 was used, from which a section 8 question, involving sex life, was omitted from the original (ver. 2.0). ODI responses with 1 question without an answer were corrected as total score/(5 × number of questions answered) × 100%.
According to the protocol, the baseline data recorded at least 1 week after the completion of diagnostic blocks and before undergoing LMBRFD included age, gender, duration of symptoms, opioid use, and baseline NRS and ODI testing. After undergoing LMBRFD, NRS and ODI were assessed at 1 and 6 months postprocedure.
One month after undergoing LMBRFD, a 5-point satisfaction scale (5, very satisfied; 4, somewhat satisfied; 3, neither satisfied nor dissatisfied; 2, somewhat dissatisfied; 1, very dissatisfied) was used to evaluate satisfaction with treatment. We coded it into dichotomous levels: successful (5, very satisfied; or 4, somewhat satisfied) or failed (3, neither satisfied nor dissatisfied; 2, somewhat dissatisfied; or 1, very dissatisfied).
Postprocedural complication assessments were performed 1 week after LMBRFD and were evaluated during the 6-month follow-up period. Complications were categorized as: (a) localized pain at the RF site; (b) neuritic pain; (c) a new sensory or motor deficit; and (d) others. Patients were told to attend our pain clinic immediately if their previous pain recurred.
For the primary end point, an independent t test was used to compare the mean difference in changes of NRS scores from baseline at 2 time points (1 and 6 months) between groups. To minimize the chance of a type 1 error, the reported P-values and confidence intervals (CIs) have been Bonferroni corrected; a correct P < 0.025 and the 2-sided 97.5% CI were considered to be statistically significant for the primary end point.
For the secondary end point, NRS and the ODI scores were submitted to a repeated-measures analysis of variance with effects on the group, time, and group-by-time interaction. We used the Bonferroni adjustment procedure to do the follow-up analysis as it can control the type 1 error.
To compare the procedure-related variables between groups, we used an independent t test with P < 0.01 (2-sided 99% CI) as criterion for statistical significance to minimize the risk of type 1 error. The Student t test with unequal variances and in the time scale can be reasonably used when provided, n > 25; however, the modest increased risk of a type 1 error is recognized.19
Patient demographics at the baseline were compared by the t tests and χ2 tests for continuous and categoric variables, respectively. Statistical analysis was performed using the SPSS version 19.0 (SPSS Inc, Chicago, IL). All parametric data are presented as the mean (SD) and nonparametric data as numbers and proportions.
During the recruitment phase, 48 patients were excluded for various reasons, the most common of which was failure to have pain return after diagnostic lumbar medial branch blocks (n = 26). Forty-one patients in each group were eventually enrolled. After finishing LMBRFDs, 3 patients in the distal approach group and 2 patients in the tunnel vision approach group had broader lesions because the motor and sensory stimulation was not successful. Among them, 3 patients (2 in the distal and 1 in the tunnel vision approach group) were excluded because they did not experience either motor or sensory stimulation at ≥2 nerves. Eleven patients (5 in the distal and 6 in the tunnel vision approach group) were lost to follow-up. Therefore, 34 patients per group completed the follow-up schedules (Fig. 1). The randomized patients were similar in both groups, as shown in Table 2.
The primary end point results are shown in Table 3. The mean NRS pain score in the distal approach group decreased by 2.4 (1.5) at 1 month and 1.9 (1.7) at 6 months, respectively. In the tunnel vision approach group, NRS pain score decreased by 1.9 (1.7) at 1 month, and 2.0 (2.0) at 6 months. However, there were no statistically significant differences in the change of NRS scores between the groups at 1 month (corrected P = 0.19; 97.5% CI, −1.37 to 0.37) and 6 months (corrected P = 0.53; 97.5% CI, −1.36 to 0.77). We also confirmed the results by using analysis of covariance with our baseline NRS scores as covariates.a
Table 4 shows that the average NRS and ODI scores in both groups decreased over time (both results P < 0.0001 by repeated-measures analysis of variance). The group-by-time interaction was not significant in both NRS and ODI scores (P = 0.35 and P = 0.32, respectively), suggesting that the decline in the NRS and ODI scores over time was not statistically different between the distal approach group and the tunnel vision approach group. In multiple comparison tests by Bonferroni, a decrease in NRS and ODI scores from entry to scores at 1 month and 6 months was statistically significant; all resulting P < 0.0001 and 95% CI were 1.67 to 2.60 (NRS scores at baseline—1 month), 1.66 to 2.81 (NRS scores at baseline − 6 months), 5.05 to 8.01 (ODI scores at baseline—1 month), and 4.75 to 7.99 (ODI scores at baseline—6 months). However, there were no statistically significant differences in NRS and ODI scores from 1 to 6 months; all resulting P = 1.00 and 95% CI were −0.37 to 0.78 and −1.10 to 1.57, respectively.
In the procedure-related evaluation, the time required per level (minutes) was significantly longer in the distal approach group than in the tunnel vision approach group (P = 0.001; 99% CI, 9.75 to 73.66) as shown in Table 5. On the other hand, procedure-related pain scores were significantly lower in the distal approach group (P = 0.001; 99% CI, −2.00 to −0.23).
Five patients in the tunnel vision approach group and 1 patient in the distal approach group reported an RF-associated complication during the 6-month follow-up. Two patients in the tunnel vision and 1 patient in the distal approach groups experienced mild, localized pain at the RF lesion site lasting <1 month, but 3 patients in the tunnel vision approach group experienced the development of a new neuropathy-like pain that lasted longer than 3 months as follows: (1) itching sensation and burning pain in a lower buttock after bilateral L2 to L4 MBRFD; (2) itching and sunburn-like allodynia of the skin overlying the spinous processes and bilateral lower back after bilateral L1 to L4 MBRFD; or (3) hyperalgesic pain in the bilateral lower buttocks after bilateral L2 to L4 MBRFD. These symptoms were relieved by oral gabapentin or 5% lidocaine patches, and all resolved by 6 months. There were no adverse events that lasted longer than 1 month in the distal approach group. However, no significant differences in the incidence of complications between the groups (P = 0.24, Fisher exact test) was noted.
Patients in both the tunnel vision and distal approach groups experienced significant pain relief after LMBRFD, as demonstrated by reductions in NRS scores and ODIs (%). It took longer to perform LMBRFD at each level with the distal approach than with the tunnel vision approach. Procedure-related pain was significantly lower in the distal approach group. Although neuropathy-like complications lasting longer than 3 months were observed more frequently in the tunnel vision approach group (n = 3), it was not statistically significant.
In the study, the time required per level was significantly longer in the distal approach group than in the tunnel vision approach group whereas procedure-related pain scores were lower in the distal approach group. These results conflict with each other. Maintaining motor stimulation during cannula advance from the landing point to the target point seemed to be uncomfortable, but the patients who underwent LMBRFD by the distal approach reportedly experienced less procedure-related pain. Although detecting the difference in procedure-related time or pain between the groups was beyond this study and requires investigation in the future, we speculated on the possible reasons. During the procedure, patients often complained of back pain or discomfort mainly evoked by sensory electrostimulation. However, motor stimulation at 2 Hz and <0.5 V was performed for a few seconds, which was tolerable for patients. Also, a steep caudocephalad decline coupled with lateral rotation in the tunnel vision approach may result in a longer distance to the target site5 than that in the distal approach. A relatively shorter distance to the target site in the distal approach may be related to less procedure-related pain. Finally, although there was a significant statistical difference in required time per level between the groups, we cannot state with certainty that the mean difference in the required time between groups has any clinical implication because it was <1 minute at the most.
LMBRFD has been reported to have few associated complications.2,20 However, the overall incidence of complications in the study was a little higher (n = 5, 6.0%) than that (1.0%–4.0%) reported in other studies. The sample size in this study, 34 patients in each group, was insufficient to draw conclusions about the complication rate,20 which can be magnified by the more ambitious lesioning scheme used here or by the authors’ individual technique. Although there were no statistically significant differences in the incidence of complications between the groups, it is debatable why postprocedure neuritic pain was more prevalent with the tunnel vision approach in the study. Numbness or anesthesia dolorosa of the buttock might be caused by an injury to the cutaneous fibers of the L1 to L3 lateral branches or parent dorsal rami during rhizolysis.9 Similarly, the neuropathy-like low back pain encountered in the tunnel vision approach group was possibly caused by thermal injury to the L1 to L3 primary dorsal rami or lateral branches. We reviewed the lateral fluoroscopic radiographs of these patients and found that RF needles had been inserted nonparallel to the base of the SAP with respect to the upper endplate of the vertebra. The neuropathy-like complication may have occurred in the tunnel vision approach group because the angle between the endplate and the base of the SAP ranges from 20° to 40°5,21 and determines the slope of the medial branch as it crosses the neck of the SAP. However, in our pilot investigation, we measured the angle between the endplate and the base of the SAP in the lateral scout images of 12 patients in flexion, and the angle between the endplate and the base of the SAP was much more obtuse than expected (range, 30°–60°). During the tunnel vision approach to the medial branch, this angle became a standard for caudocephalad axial tilt of the fluoroscopy beam. If the angle of axial tilt was smaller than the actual value, it was possible to make bone contact with the electrode more proximally and deeply (Fig. 4A), which allowed the electrode tip to approach the dorsal primary ramus or lateral branch. With the distal approach, on the other hand, it could be somewhat easier to avoid the possibility of primary dorsal ramus or lateral branch thermal injury (Fig. 4B).
There are several limitations in this study. Our investigation could be criticized for the absence of a control group and lack of a blinding and crossover design. Psychiatric comorbidity influencing poor outcome in low back pain patients could have been investigated in each group.22 Moreover, we were unable to control for other treatments, such as physical therapy or analgesic use during the follow-up. However, although it is important to report overall clinical outcomes to suggest the general success of LMBRFD, our primary purpose was to show that the distal approach to LMBRFD improved with the tunnel vision approach.
In conclusion, patients who underwent LMBRFD with the tunnel vision or distal approaches experienced significant pain relief with respect to NRS and ODI scores during the 6-month follow-up. If performed correctly, LMBRFD should be largely free of complications.6,20 Our findings show that patients who underwent LMBRFD using the distal approach felt less periprocedural pain and experienced fewer complications after the procedure. Therefore, we suggest that the distal approach for LMBRFD can be an improved alternative for patients with L1 to L4 facet joint pain.
Name: Pyung Bok Lee, MD, PhD.
Contribution: This author helped design the study, conduct the study, and write the manuscript.
Attestation: Pyung Bok Lee has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Jee Youn Moon, MD, PhD.
Contribution: This author helped conduct the study, analyze the data, and write the manuscript.
Attestation: Jee Youn Moon has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Yong Chul Kim, MD, PhD.
Contribution: This author helped design the study, analyze the data, and write the manuscript.
Attestation: Yong Chul Kim has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Seung Pyo Choi, MD.
Contribution: This author helped conduct the study, analyze the data, and write the manuscript.
Attestation: Seung Pyo Choi has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
Name: Woo Seog Sim, MD, PhD.
Contribution: This author helped analyze the data and write the manuscript.
Attestation: Woo Seog Sim has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.
This manuscript was handled by: Spencer S. Liu, MD.
a Difference in change of NRS scores from baseline between groups was −0.82 (97.5% CI, −2.61 to 0.97; P = 0.30) at 1 month and −0.40 (97.5% CI, −2.56 to 1.77; P = 0.68) at 6 months in analysis of covariance. Cited Here...
1. Cohen SP, Raja SN. Pathogenesis, diagnosis, and treatment of lumbar zygapophysial (facet) joint pain. Anesthesiology. 2007;106:591–614
2. Cohen SP, Williams KA, Kurihara C, Nguyen C, Shields C, Kim P, Griffith SR, Larkin TM, Crooks M, Williams N, Morlando B, Strassels SA. Multicenter, randomized, comparative cost-effectiveness study comparing 0, 1, and 2 diagnostic medial branch (facet joint nerve) block treatment paradigms before lumbar facet radiofrequency denervation. Anesthesiology. 2010;113:395–405
3. 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
4. Richardson J. A (pain free) step in the right direction. Br J Anaesth. 2004;93:173–4
5. Gofeld M, Faclier G. Radiofrequency denervation of the lumbar zygapophysial joints–targeting the best practice. Pain Med. 2008;9:204–11
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. Sehgal A, Valentine JM. Lumbar radiculopathy after zygapophyseal joint injection. Br J Anaesth. 2007;99:412–4
8. Bogduk N, Wilson AS, Tynan W. The human lumbar dorsal rami. J Anat. 1982;134:383–97
9. Bogduk N. Lumbar lateral branch neuralgia: a complication of rhizolysis. Med J Aust. 1981;1:242–3
10. Kim YCKim DH, Kim KH, Kim YC. Medial branch block and radiofrequency lesioning. In: Minimally Invasive Percutaneous Spinal Techniques. 2011 Philadelphia Elsevier Sounders Co.:149–63
11. Moon J, Kim YC, Park SY, Lee SC, Choi SP, Nahm FS, Lee PB, Goo EK, Kang JM. Psychometric characteristics of the Korean version of the Roland-Morris Disability Questionnaire. J Korean Med Sci. 2011;26:1364–70
12. Cohen SP, Strassels SA, Kurihara C, Lesnick IK, Hanling SR, Griffith SR, Buckenmaier CC 3rd, Nguyen C. Does sensory stimulation threshold affect lumbar facet radiofrequency denervation outcomes? A prospective clinical correlational study. Anesth Analg. 2011;113:1233–41
13. Bogduk NInternational Spine Intervention Society. . Lumbar medial neurotomy. Practice Guidelines for Spinal Diagnostic and Treatment Procedures. 2004 San Francisco, CA International Spinal Intervention Society:188–218
14. Provenzano DA, Lassila HC, Somers D. The effect of fluid injection on lesion size during radiofrequency treatment. Reg Anesth Pain Med. 2010;35:338–42
15. Dworkin RH, Turk DC, Farrar JT, Haythornthwaite JA, Jensen MP, Katz NP, Kerns RD, Stucki G, Allen RR, Bellamy N, Carr DB, Chandler J, Cowan P, Dionne R, Galer BS, Hertz S, Jadad AR, Kramer LD, Manning DC, Martin S, McCormick CG, McDermott MP, McGrath P, Quessy S, Rappaport BA, Robbins W, Robinson JP, Rothman M, Royal MA, Simon L, Stauffer JW, Stein W, Tollett J, Wernicke J, Witter JIMMPACT. . Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain. 2005;113:9–19
16. Jensen MP, Karoly PTurk DC, Melzack R. Self-report scales and procedures for assessing pain in adults. In: Handbook of Pain Assessment. 2001 2nd ed. New York Guilford Press:15–34
17. Fairbank JC, Pynsent PB. The Oswestry disability index. Spine. 2000;25:2940–52
18. Kim DY, Lee SH, Lee HY, Lee HJ, Chang SB, Chung SK, Kim HJ. Validation of the Korean version of the oswestry disability index. Spine. 2005;30:E123–7
19. Zhou XH, Gao S, Hui SL. Methods for comparing the means of two independent log-normal samples. Biometrics. 1997;53:1129–35
20. Kornick C, Kramarich SS, Lamer TJ, Todd Sitzman B. Complications of lumbar facet radiofrequency denervation. Spine. 2004;29:1352–4
21. Calodney AKRaj PP. Lumbar facet joint blocks and neurotomy. In: Interventional Pain Management: Image-Guided Procedures. 2007 Philadelphia, PA Saunders Elsevier:368–81
22. Dunn KM, Jordan KP, Croft PR. Contributions of prognostic factors for poor outcome in primary care low back pain patients. Eur J Pain. 2011;15:313–9