Clinical Neurology and Neurosurgery 195 (2020) 105854

Contents lists available at ScienceDirect

Clinical Neurology and Neurosurgery

journal homepage: www.elsevier.com/locate/clineuro

A comparison of efficacy among different radiofrequency ablation T

techniques for the treatment of lumbar facet joint and sacroiliac joint pain: A
systematic review and meta-analysis
Chia-Lung Shiha, Po-Chih Shena, Cheng-Chang Lub,c, Zi-Miao Liuc, Yin-Chun Tiena,c,
Peng-Ju Huanga,c, Shih-Hsiang Choua,*
a
Department of Orthopedics, Kaohsiung Medical University Hospital, Kaohsiung Medical University, Kaohsiung, Taiwan
b
Department of Orthopedics, Kaohsiung Municipal Siaogang Hospital, Kaohsiung, Taiwan
c
College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan

A R T I C LE I N FO A B S T R A C T

Keywords: Objective: To compare the efficacy of different radiofrequency techniques (thermal, pulsed, and cooled radio-
Cooled radiofrequency frequency) for treating lumbar facet joint (LFJ) or sacroiliac joint (SIJ) pain.
Thermal radiofrequency Patients and methods: The inclusion criteria were as follows: (1) age > 18 years; (2) patients suffering from LFJ
Pulsed radiofrequency or SIJ pain; and (3) patients receiving radiofrequency treatments. Four electronic databases, including Pubmed,
Lumbar facet joint
Embase, Cochrane Library, and ISI Web of Knowledge were systematically searched from inception until
Sacroiliac joint
December 2019 for relevant articles. The search was conducted on 2 January 2020. When the outcomes among
articles showed heterogeneity, then a random-effects model was adopted to calculate the effect size; otherwise, a
fixed-effects model was adopted.
Results: All the three techniques showed significant improvements in LFJ or SIJ pain for up to 12 months
compared with the baseline level. However, no significant differences among the three techniques were observed
at any follow-up visits except for possibly a trend for variance in efficacy. For treating LFJ pain, cooled radio-
frequency was the most effective, followed by thermal radiofrequency and then pulsed radiofrequency as the
least respectively for the follow-up visit at 6 months. No serious complications were reported after receiving
treatment using the three techniques.
Conclusion: Sequentially, cooled radiofrequency followed by thermal radiofrequency and then pulsed radio-
frequency for treating LFJ pain were identified as most to least effective at the 6-month follow-up.

1. Introduction treating chronic back pain in 1975 [13]. It is a minimally invasive

procedure and is operated under light intravenous sedation or local
Chronic low back pain is one of the most common health problems anesthesia when necessary. Radiofrequency energy is delivered to the
and it can cause disability [1]. Low back pain occurs in 70 %–85 % of target nerves through an insulated needle, and this energy heats and
the general population, and sacroiliac joint (SIJ) pain (40 %) and denatures the nerve for the purpose of pain relief. The radiofrequency
lumbar facet joint (LFJ) pain (15 %–40 %) are the major causes of this techniques include pulsed, thermal, and cooled radiofrequency. Pulsed
disease [2–4]. SIJ pain is an inflammatory condition of SIJ and can radiofrequency uses less energy and lower temperature compared with
extend down the legs. Osteoarthritis is considered to be the main cause thermal radiofrequency, whereas cooled radiofrequency adopts in-
of LFJ pain [5]. Degenerative changes in LFJs can cause abnormal strain ternally cooled radiofrequency probes to increase lesion size, and it can
and stress and increase load on the LFJs [6]. Interventional treatments increase the chance of complete denervation [14]; therefore, the effi-
for SIJ or LFJ pain include intra-articular joint injections [7,8], radio- cacy of the three radiofrequency techniques for treating low back pain
frequency ablation [9,10], extracorporeal shockwave therapy [11] and might be different.
surgical treatment [12]. Previous meta-analysis studies have investigated the efficacy of
Radiofrequency ablation was first described in the literature for radiofrequency in the treatment of low back pain. For example, thermal

⁎
Corresponding author at: Department of Orthopedics, Kaohsiung Medical University Hospital, No.100, Tzyou 1st Road Kaohsiung 807, Taiwan.
E-mail address: [email protected] (S.-H. Chou).

https://doi.org/10.1016/j.clineuro.2020.105854
Received 10 February 2020; Received in revised form 9 April 2020; Accepted 14 April 2020
Available online 19 April 2020
0303-8467/ © 2020 Elsevier B.V. All rights reserved.

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

radiofrequency was reported to have significant reductions in LFJ pain 2.4. Data extraction
up to 12 months [15]. A previous meta-analysis study has demonstrated
that cooled radiofrequency can significantly relieve pain and disability The main characteristics of included articles were independently
without severe complications in patients with SIJ pain [16]. Only one extracted by the two authors, including the first author’s name, pub-
previous study has compared the efficacy of pulsed radiofrequency lication year, sample size, mean age of patients, type of radiofrequency
versus thermal radiofrequency in the treatment of LFJ pain, and the ablation techniques, types of outcomes, and follow-up times. If detailed
results showed that pulsed radiofrequency was less effective than outcomes were not available from these articles, we contacted the au-
thermal radiofrequency in regard to pain relief and return of func- thors via e-mail. Any discrepancies were rechecked by the two authors
tionality [17]; however, a limited number of trials (three trials) were until the consistency was reached.
included in that systematic review, and the conclusion was not derived
from meta-analysis results [17]. Up-to-date, a comparison of the effi- 2.5. Assessment of risk of bias
cacy among the three radiofrequency techniques in the treatment of
low back pain has not been investigated using a meta-analysis. We assessed the risk of bias for each article using the Cochrane
Due to the limited number of randomized controlled trials com- Collaboration risk-of-bias tool. The following items regarding the risk of
paring the efficacy of different radiofrequency techniques for low back bias were adopted for assessment: Random sequence generation, allo-
pain, we systematically searched articles investigating the efficacy of cation concealment, blinding of participants and personnel, blinding of
the radiofrequency techniques and compared these techniques by using outcome assessment, incomplete outcome data, selective reporting and
the improvement in clinical outcomes from baseline to follow-up visits other bias. These items were independently assessed by the two au-
in which this method had been suggested by a previous meta-analysis thors. Any discrepancies were discussed until a consensus was reached.
[18]. These articles could provide sufficient data for us to compare the
efficacy among the three radiofrequency techniques using a meta-ana- 2.6. Methodological quality assessment
lysis.
The aim of this study was to compare the efficacy of different The Quality Appraisal of Reliability Studies (QAREL) criteria were
radiofrequency techniques for the management of low back pain arising used to assess methodological quality of prevalence or diagnostic ac-
from LFJ or SIJ using a meta-analysis. curacy articles. Scores of 8–12 indicated high quality, 4–6 indicated
moderate quality, and less than 4 indicated poor quality. These QAREL
2. Patients and methods criteria were independently assessed by the two authors. Any dis-
crepancies were discussed until a consensus was reached.
2.1. Search strategy
2.7. Statistical analysis
Four electronic databases, including Pubmed, Embase, Cochrane
Library, and ISI Web of Knowledge, were independently searched by All statistical analyses were performed using Review Manager
two authors (C.L.S. and S.H.C) from inception until December 2019 for (version 5.3; Cochrane collaboration). The data were sub-grouped
relevant articles that conducted radiofrequency ablation in the treat- based on follow-up time and type of treatment techniques. The follow-
ment of low back pain arising from LFJ and SIJ. Boolean logic with up visits were grouped into 1 month (3–6 weeks), 3 months (3−4
search terms was adopted to search relevant articles from the databases: months), 6 months, and 12 months for assessment of the efficacy of
(“radiofrequency”) AND (“sacroiliac joint” OR “facet joint”). The re- radiofrequency at different follow-up periods. For subgroup analysis,
ferences cited by reviews or relevant articles were searched for addi- mean difference (MD) and 95 % confidence interval (CI) were calcu-
tional eligible articles, with all possible articles independently searched lated to assess the improvement in pain from baseline to follow-up
by the two authors. First, the title and abstract of each article were visits. Heterogeneity among included articles was detected using X2 test
reviewed for excluding non-related articles, and then a full-test analysis and I2 statistic. When a p-value of X2 test < 0.05 or I2 statistic > 50 %,
was conducted for these remaining articles. Any discrepancies were the outcomes among articles showed heterogeneity and a random-ef-
discussed until mutual agreement was achieved. fects model was adopted to calculate the effect size; otherwise a fixed-
effects model was adopted.
2.2. Inclusion and exclusion criteria
2.8. Ethical approval
Articles were included if they met the following inclusion criteria:
patients were older than 18 years; had been diagnosed with low back There were no issues regarding participant privacy, so ethical ap-
pain arising from LFJ or SIJ; and had been treated by pulsed, thermal, proval was not necessary for this study.
or cooled radiofrequency technique, with either Visual analog scale
(VAS) or numeric rating scale (NSR) being recorded to assess patient 3. Results
pain level before and after treatment. Additionally, articles had to have
been published in English. Articles were rejected if they met the fol- 3.1. Literature review
lowing exclusion criteria: articles were editorials, letters, reviews,
conference abstracts, proceedings, personal communications and case The flowchart of the literature review is shown in Fig. 1. A total of
reports; were animal studies; and no VAS or NSR scores had been re- 1411 articles were retrieved from the four databases by the initial
corded. screening. After removing duplicates, the number of remaining articles
was 961. After further screening through title or abstract, 188 articles
2.3. Types of outcomes met the inclusion criteria followed by full-text analysis, which excluded
167 articles for a variety of reasons (e.g., no outcomes of interest and
VAS and NSR were adopted in our meta-analysis to assess the effi- the lack of mean or SD values of the outcomes). A total of 21 eligible
cacy of radiofrequency ablation in the treatment of low back pain. The articles were used for our meta-analysis [19–39].
two scores are self-reported measures of patients’ pain levels, ranging
from 0 to 10 where 0 shows no pain and 10 shows the worst possible 3.2. Main characteristics
pain; thus, VAS and NSR were considered as the same type of outcome
for the use of our meta-analysis. The main characteristics of the included articles are summarized in

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

Fig. 1. Flow chat of article searching procedure.

Table 1. These articles were published from 2005 to 2019 and enrolled 3.3.2. Blinding
1129 patients. The number of subgroups for thermal, pulsed, and Seven articles were considered to be at low risk of performance bias
cooled radiofrequency techniques were 14, 13, and 5 respectively. because they were double-blind trials (Fig. 2); six articles did not de-
Fourteen treatment subgroups suffered from LIJ pain, and the other 18 scribe the design of trials; and the remaining eight articles were con-
subgroups suffered from SIJ pain. The mean age of each treatment sidered to be at high risk of bias because they were not double-blind
subgroup ranged from 41.3 to 70.0 years of age. For assessing the pa- trials.
tients’ pain level, 10 articles used VAS and the other 11 articles used Nine articles blinded the assessors for treatment groups and were
NRS. The extracted follow-up time ranged from 3 weeks to 12 months. considered to be at low risk of bias (Fig. 2); six articles did not describe
the information about blinding of outcome assessment; while the re-
maining six articles did not blind the assessors for treatment groups and
3.3. Risk of bias assessment were considered to be at high risk of bias.

3.3.1. Allocation
3.3.3. Incomplete outcome data
Only 4 articles described the methods to generate random sequence,
Seventeen articles were considered to be at high risk of bias because
including a computerized random [19,23,39], and a random table [25]
follow-up was completed by less than 80 % of participants, but the
(Fig. 2). Eleven articles did not provide the methods, while the re-
remaining four articles were considered to be at low risk of bias because
maining six articles did not use generation random sequence; only 4
follow-up was completed by at least 80 % of participants (Fig. 2).
articles described the methods of allocation concealment, including
opaque envelope [19] and sealed envelope [20,26,34]; 11 articles did
not describe the methods of allocation concealment; and the 6 re- 3.3.4. Selective reporting
maining articles did not use allocation concealment. The 18 articles were considered to be at low risk of bias but the
remaining three articles were considered to be at high risk of bias be-
cause they did not report complications (Fig. 2).

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

Table 1
Main characteristics of the included articles for the meta-analysis.
OBS Author Type of radiofrequency Origin of pain Sample size Mean age (year) Sex (M/F) Type of outcome Follow-up time

1 McCormick 2019 (group1) Thermal LFJ 18 58.4 8/10 NRS 6 months

McCormick 2019 (group2) Cooled LFJ 21 53.6 8/13
2 Dutta 2018 Pulsed SIJ 15 43.1 4/11 NRS 1, 3, and 6 months
3 Mehta 2018 Thermal SIJ 11 NA NA NRS 3 months
4 Çetin 2018 (group1) Thermal LFJ 43 53.39 14/29 VAS 1, 3, 6, and 12 months
Çetin 2018 (group2) Pulsed LFJ 75 53.9 32/43
5 Juch 2017 (group1) Thermal LFJ 125 52.98 60/65 NRS 1.5, 3, 6, and 12 months
Juch 2017 (group2) Thermal SIJ 116 51.58 29/87
6 Stelzer 2017 (group1) Cooled LFJ 16 NA NA VAS 1 and 6 months
Stelzer 2017 (group2) Cooled LFJ 27 NA NA 1 and 12 months
Stelzer 2017 (group3) Cooled SIJ 53 NA NA 1 and 6 months
Stelzer 2017 (group4) Cooled SIJ 56 NA NA 1 and 12 months
7 Do 2017 Pulsed LFJ 30 66.9 12/18 NRS 1, 3, and 6 months
8 van Tilburg 2016 Thermal SIJ 30 NA 5/25 NRS 1 month
9 Reddy 2016 Thermal SIJ 16 45.4 7/9 NRS 6 and 12 months
10 Patel 2016 Cooled SIJ 25 NA NA NRS 3, 6, and 12 months
12 Martínez 2016 (group1) Thermal SIJ 20 53.2 5/15 VAS 1, 3, and 12 months
Martínez 2016 (group2) Thermal SIJ 20 54.1 7/13
11 Romero 2015 Thermal SIJ 32 58.3 18/14 NRS 1, 6, and 12 months
13 Zheng 2014 Thermal SIJ 82 41.3 59/23 VAS 3 and 6 months
14 Stelzer 2013 (group1) Cooled SIJ 26 66 4/22 VAS 3−4 weeks
Stelzer 2013 (group2) Cooled SIJ 45 67 16/29
Stelzer 2013 (group3) Cooled SIJ 34 70 9/25
15 Ho 2013 Cooled SIJ 20 54.3 6/14 NRS 1, 3, and 12 months
16 Ma 2011 Thermal LFJ 22 72.4 6/16 VAS 1 month
17 Kapural 2008 Cooled SIJ 26 61 6/20 VAS 3−4 months
18 Cohen 2008 Cooled SIJ 14 51.9 5/9 NRS 1, 3, and 6 months
19 Kroll 2008 (group1) Thermal LFJ 13 59.5 7/6 VAS 3 months
Kroll 2008 (group2) Pulsed LFJ 13 57 5/8
20 Tekin 2007 (group1) Thermal LFJ 20 60.5 45/55 VAS 6 and 12 months
Tekin 2007 (group2) Pulsed LFJ 20 59.6 40/60
21 Dobrogowski 2005 Pulsed LFJ 45 66.4 23/22 VAS 1, 3, and 6 months

SIJ: sacroiliac joint; LFJ: lumbar facet joint; NRS: numeric rating scale; VAS: visual analog scale; NA: not available.

3.3.5. Other potential sources of bias 3.4. Methodological quality assessment

Eight articles did not have other potential sources of bias and were
considered to be at low risk of other bias. The remaining 13 articles After assessing the methodological quality of each article, all of the
were considered to be at high risk of other bias because they had small 21 included articles achieved the inclusion criteria (Table S1). Nineteen
sample size (n < 30). articles were considered to be moderate quality, and the other two
articles were assessed as poor quality.

Fig. 2. Risk of bias for each included article (A) and the risk-of-bias summary across all included articles (B).

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

Fig. 3. The workflow of the meta-analysis.

3.5. Meta-analysis borderline statistical significance in pain relief (P = 0.06), although

efficacy of cooled radiofrequency was the most, thermal radiofrequency
3.5.1. LFJ pain was the second, and pulsed radiofrequency was the least. However,
The workflow of this study is shown in Fig. 3. The patients had only two radiofrequency techniques (thermal and pulsed) had sufficient
significant improvements in pain for the follow-up visit at 1 month for data for the follow-up visit at 12 months for meta-analysis. Both tech-
all the three radiofrequency techniques compared with the baseline niques showed significant improvements in pain at this follow-up
level (random-effects model: 3 trials, MD = 3.50, 95 % CI = 1.66–5.33, compared with the baseline level (random-effects model: 3 trials,
P = 0.0002 for thermal radiofrequency; random-effects model: 3 trials, MD = 3.72, 95 % CI = 2.54–4.89, P < 0.00001 for thermal radio-
MD = 3.34, 95 % CI = 2.26–4.43, P < 0.00001 for pulsed radio- frequency; random-effects model: 2 trials, MD = 3.83, 95 %
frequency; random-effects model: 2 trials, MD = 3.34, 95 % CI = 2.55–5.10, P < 0.00001 for pulsed radiofrequency) (Fig. 7). The
CI = 2.61–4.06, P < 0.00001 for cooled radiofrequency) (Fig. 4). The efficacy of both techniques was similar (P = 0.90).
efficacy among the three techniques did not show significant difference
(p = 0.99). Only two radiofrequency techniques (thermal and pulsed)
had sufficient data (≥2 trials) at 3-month follow-up for meta-analysis. 3.5.2. SIJ pain
The patients had significant improvements in pain at this follow-up for Only two radiofrequency techniques (thermal and cooled) had suf-
two radiofrequency techniques compared with the baseline level ficient data for the follow-up visit at 1 month for meta-analysis. Patients
(random-effects model: 3 trials, MD = 3.24, 95 % CI = 0.93–5.56, had significant improvements in pain at this follow-up for the two
P = 0.006 for thermal radiofrequency; random-effects model: 4 trials, radiofrequency techniques compared with the baseline level (random-
MD = 2.94, 95 % CI = 2.05–3.84, P < 0.00001 for pulsed radio- effects model: 5 trials, MD = 3.05, 95 % CI = 1.83–4.28, P < 0.00001
frequency) (Fig. 5). The two techniques did not show significant dif- for thermal radiofrequency; random-effects model: 7 trials, MD = 5.06,
ference (P = 0.81) but possibly revealed a trend for variance in efficacy 95 % CI = 3.94–6.19, P < 0.00001 for cooled radiofrequency) (Fig. 8).
where thermal radiofrequency appeared better than pulsed radio- The efficacy of cooled radiofrequency was better than that of thermal
frequency. The patients had significant improvements in pain for the radiofrequency (P = 0.02). The patients had significant improvements
follow-up visit at 6 months for all three radiofrequency ablation tech- in pain for the follow-up visit at 3 months for all three radiofrequency
niques compared with the baseline level (random-effects model: 4 techniques compared with the baseline level (random-effects model: 4
trials, MD = 3.64, 95 % CI = 2.34–4.93, P < 0.00001 for thermal trials, MD = 3.36, 95 % CI = 2.69–4.02, P < 0.00001 for thermal
radiofrequency; random-effects model: 4 trials, MD = 2.77, 95 % radiofrequency; random-effects model: 2 trials, MD = 4.16, 95 %
CI = 2.24–3.30, P < 0.00001 for pulsed radiofrequency; random-ef- CI = 3.55–4.76, P < 0.00001 for pulsed radiofrequency; random-ef-
fects model: 2 trials, MD = 4.01, 95 % CI = 3.06–4.97, P < 0.00001 fects model: 4 trials, MD = 3.68, 95 % CI = 2.66–4.69, P < 0.00001
for cooled radiofrequency) (Fig. 6). The three techniques reached for cooled radiofrequency) (Fig. 9). The three techniques did not show
significant difference in pain relief (P = 0.21) but possibly a trend for

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

Fig. 4. Meta-analysis of efficacy in pain scores using radiofrequency to treat LFJ pain for the follow-up visit at 1 month.

variance in efficacy where pulsed radiofrequency was more effective visit at 6 months reached borderline statistical significance (P = 0.06)
than the other two techniques. Only two radiofrequency techniques using the random effects model but become significantly different
(thermal and cooled) had sufficient data for the follow-up visit at 6 (P < 0.00001) when using the fixed-effects model. Other results esti-
months for meta-analysis. The patients had significant improvements in mated by random-effects and fixed-effect models appeared similar.
pain at this follow-up for the two radiofrequency techniques compared
with the baseline level (random-effects model: 4 trials, MD = 3.69, 95 3.6. Complications and adverse effects
% CI = 2.81–4.57, P < 0.00001 for thermal radiofrequency; random-
effects model: 3 trials, MD = 3.57, 95 % CI = 2.62–4.52, P < 0.00001 No serious complications were reported after receiving treatment
for cooled radiofrequency) (Fig. 10). The efficacy of the two techniques using the three radiofrequency techniques [25,27,28,30,31,33,38], and
was similar (P = 0.85). Only two radiofrequency techniques (thermal only minor complications were reported, such as pain, hemorrhage, and
and cooled) had sufficient data at the 12-month follow-up for meta- infection [19,21,23,29,34,39].
analysis. The patients had significant improvements in pain at this
follow-up for the two radiofrequency techniques compared with the 4. Discussion
baseline level (random-effects model: 5 trials, MD = 3.38, 95 %
CI = 2.44–4.32, P < 0.00001 for thermal radiofrequency; random-ef- A comparison of the efficacy among three radiofrequency techni-
fects model: 3 trials, MD = 3.53, 95 % CI = 2.64–4.42, P < 0.00001 ques in the treatment of low back pain has not been well investigated.
for cooled radiofrequency) (Fig. 11). The efficacy of the two techniques We endeavored to compare the efficacy of the three techniques using a
was similar (P = 0.82). meta-analysis. All three techniques showed significant improvements in
LFJ or SIJ pain for up to 12 months when compared with the baseline
3.5.3. Methodological uncertainty level. Previous meta-analysis studies also concluded that thermal
To assess the methodological uncertainty, we investigated if the radiofrequency has significant reductions in LFJ pain up to 12 months
results changed in the analyses or comparisons using fixed-effects [15], and cooled radiofrequency can significantly relieve pain in pa-
model or random-effects model (Table S2). There were changes in one tients with SIJ pain [16]. For treating LFJ or SIJ pain, the three tech-
comparison. For treating LFJ, the three techniques for the follow-up niques did not show significant differences in pain relief for the follow-

Fig. 5. Meta-analysis of efficacy in pain scores using radiofrequency to treat LFJ pain for the follow-up visit at 3 months.

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

Fig. 6. Meta-analysis of efficacy in pain scores using radiofrequency to treat LFJ pain for the follow-up visit at 6 months.

up visits from 1 to 12 months but possibly a trend for variance in ef- that its efficacy may be better than thermal and pulsed radiofrequency
ficacy. For treating LFJ pain, cooled radiofrequency was the most ef- techniques. Our results show that the efficacy of cooled radiofrequency
fective, thermal radiofrequency was the second, and pulsed radio- for treating FIJ pain was more significant than that of pulsed and
frequency was the least effective only for the follow-up visit at 6 thermal radiofrequency techniques only for the follow-up visit at 6
months. The efficacy of thermal radiofrequency was slightly more ef- months and reached borderline statistical significance (P = 0.06). For
fective than pulsed radiofrequency for the follow-up visits from 1 to 6 treating SIJ pain, our results show that cooled radiofrequency had
months. For treating SIJ pain, the patients treated with cooled radio- significantly better improvement in pain only for the follow-up visit at 1
frequency had significantly better improvement in pain than thermal month compared with thermal radiofrequency. The current evidence
radiofrequency only for the follow-up visit at 1 month. The efficacy of seems to support that the efficacy of cooled radiofrequency for treating
the two techniques was similar at the other follow-up visits. No serious low back pain is better than that of thermal and pulsed radiofrequency
complications were reported after receiving treatment using the three techniques at some follow-up visits.
techniques. We found that the efficacy of pulsed radiofrequency in the treatment
To avoid neuronal tissue damage, pulsed radiofrequency uses less of SIJ pain was slightly larger than thermal and cooled radiofrequency
energy and lower temperature than thermal radiofrequency. This im- techniques for the follow-up visit at 3 months (Fig. 9). In our meta-
plies that the efficacy of pulsed radiofrequency may be less than that analysis, only two included articles used pulsed radiofrequency for
using thermal radiofrequency. Our results showed that the efficacy of treating SIJ pain at 3-month follow-up [20,21]. Small sample sizes
pulsed radiofrequency for treating LFJ pain was slightly smaller than (n = 15 and 11) seem to be the primary limitation in the two articles
thermal radiofrequency for the follow-up visits at 1, 3, and 6 months [20,21], possibly mitigating reliable conclusions. In addition, pulsed
(Figs. 4–6). However, the efficacy of pulsed radiofrequency for treating radiofrequency for treating SIJ only for the follow-up visit at 3 months
SIJ pain was slightly higher than thermal radiofrequency for the follow- had sufficient data for conducting a meta-analysis. We could not de-
up visit at 3 months. A systematic review also concluded that pulsed termine if the efficacy of pulsed radiofrequency for treating SIJ pain
radiofrequency was less effective than thermal radiofrequency re- was also higher than thermal and cooled radiofrequency techniques at
garding pain relief and return of functionality in patients with FIJ pain other follow-up visits.
[17]. The current evidence could support that the efficacy of thermal The efficacy of radiofrequency ablation for treating low back pain is
radiofrequency is slightly better than pulsed radiofrequency in the considered to decrease with time due to the natural process of nerve
treatment of low back pain from 1- to 6-month follow-ups. regeneration [40]. For treating LFJ pain, pain relief reduced with time
Cooled radiofrequency is a new technique in the treatment of low was not observed for the three techniques based on our meta-analysis.
back pain, and it was designed to increase lesion size [14], indicating However, this event was found in the treatment of SIJ pain using cooled

Fig. 7. Meta-analysis of efficacy in pain scores using radiofrequency to treat LFJ pain for the follow-up visit at 12 months.

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

Fig. 8. Meta-analysis of efficacy in pain scores using radiofrequency to treat SIJ pain for the follow-up visit at 1 month.

radiofrequency. We inferred that cooled radiofrequency could provide treatment of LFJ and SIJ pain for the follow-up visits at 1, 3, 6, and 12
higher efficacy for treating SIJ pain at the initial follow-up period, and months were first compared using a meta-analysis. Our results support
thus pain relief reduced with time tends to be observed. that the three radiofrequency techniques were effective for treating LFJ
There are some possible limitations in this meta-analysis. Firstly, or SIJ pain for up to 12 months. However, the efficacy of the three
although this study performed the efficacy of radiofrequency at dif- techniques did not reach statistical significance but possibly revealed a
ferent follow-up periods, pulsed radiofrequency for treating SIJ pain trend for variance in efficacy. For treating LFJ pain, the efficacy of
and cooled radiofrequency for treating FIJ pain lacked sufficient data at cooled radiofrequency was the most effective, thermal radiofrequency
some follow-up periods for conducting a meta-analysis. The current was the second, and pulsed radiofrequency was the least effective only
results could not provide detailed results regarding the differences in for the follow-up visit at 6 months. Our meta-analysis shows hetero-
the efficacy of the three radiofrequency techniques. Secondly, this geneity for all subgroup analyses, and the results were not consistent at
meta-analysis used clinical outcomes between baseline and follow-up some follow-up visits. Further high-quality trials should be conducted
visits to compare the efficacy among different techniques, and our re- to compare the efficacy of the three radiofrequency techniques in the
sults lacked a high level of evidence. More high-quality trials should be treatment of low back pain.
further conducted to provide a high level of evidence, such as rando-
mized controlled trials. Thirdly, no standards of the three radio- Funding
frequency techniques have been established to treat low back pain [40].
This could be the major cause of heterogeneity in the meta-analysis, The authors did not receive any financial support for the research.
and our results might not be reliable. Finally, a limited number of ar-
ticles were included in some subgroup analyses, and this might weaken
CRediT authorship contribution statement
a more persuasive conclusion.
Chia-Lung Shih: Methodology, Software, Formal analysis, Writing -
5. Conclusion original draft. Po-Chih Shen: Visualization. Cheng-Chang Lu:
Visualization. Zi-Miao Liu: Visualization. Yin-Chun Tien:
Thermal, pulsed, and cooled radiofrequency techniques in the Visualization. Peng-Ju Huang: Visualization. Shih-Hsiang Chou:

Fig. 9. Meta-analysis of efficacy in pain scores using radiofrequency to treat SIJ pain for the follow-up visit at 3 months.

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

Fig. 10. Meta-analysis of efficacy in pain scores using radiofrequency to treat SIJ pain for the follow-up visit at 6 months.

Fig. 11. Meta-analysis of efficacy in pain scores using radiofrequency to treat SIJ pain for the follow-up visit at 12 months.

Conceptualization, Resources, Writing - review & editing. medial branch block for the treatment of lower lumbar facet joint pain: a retro-
spective comparative study, Medicine 96 (16) (2017) e6655.
[9] C.H. Chen, P.W. Weng, L.C. Wu, Y.F. Chiang, C.J. Chiang, Radiofrequency neu-
Declaration of Competing Interest rotomy in chronic lumbar and sacroiliac joint pain: a meta-analysis, Medicine 98
(26) (2019) e16230.
[10] A. Cetin, A. Yektas, Evaluation of the short- and long-term effectiveness of pulsed
The authors declared no conflicts of interest with respect to the radiofrequency and conventional radiofrequency performed for medial branch
research, authorship, and publication of this article. block in patients with lumbar facet joint pain, Pain Res. Manag. 2018 (2018)
7492753.
[11] Y.E. Moon, H. Seok, S.H. Kim, S.Y. Lee, J.H. Yeo, Extracorporeal shock wave
Appendix A. Supplementary data therapy for sacroiliac joint pain: a prospective, randomized, sham-controlled short-
term trial, J. Back Musculoskelet. Rehabil. 30 (4) (2017) 779–784.
Supplementary material related to this article can be found, in the [12] S.C. Yson, J.N. Sembrano, D.W. Polly Jr., Sacroiliac joint fusion: approaches and
recent outcomes, PM R 11 (Suppl 1) (2019) S114–s117.
online version, at doi:https://doi.org/10.1016/j.clineuro.2020.105854.
[13] C.N. Shealy, Percutaneous radiofrequency denervation of spinal facets. Treatment
for chronic back pain and sciatica, J. Neurosurg. 43 (4) (1975) 448–451.
References [14] I. Watanabe, R. Masaki, N. Min, N. Oshikawa, K. Okubo, H. Sugimura, T. Kojima,
S. Saito, Y. Ozawa, K. Kanmatsuse, Cooled-tip ablation results in increased radio-
frequency power delivery and lesion size in the canine heart: importance of ca-
[1] R. Vekaria, R. Bhatt, D.R. Ellard, N. Henschke, M. Underwood, H. Sandhu, Intra- theter-tip temperature monitoring for prevention of popping and impedance rise, J.
articular facet joint injections for low back pain: a systematic review, Eur. Spine J. Interv. Card. Electrophysiol. 6 (1) (2002) 9–16.
25 (4) (2016) 1266–1281. [15] C.H. Lee, C.K. Chung, C.H. Kim, The efficacy of conventional radiofrequency de-
[2] J.K. Freburger, G.M. Holmes, R.P. Agans, A.M. Jackman, J.D. Darter, A.S. Wallace, nervation in patients with chronic low back pain originating from the facet joints: a
L.D. Castel, W.D. Kalsbeek, T.S. Carey, The rising prevalence of chronic low back meta-analysis of randomized controlled trials, Spine J. 17 (11) (2017) 1770–1780.
pain, Arch. Intern. Med. 169 (3) (2009) 251–258. [16] H.H. Sun, S.Y. Zhuang, X. Hong, X.H. Xie, L. Zhu, X.T. Wu, The efficacy and safety
[3] A.C. Schwarzer, S.C. Wang, N. Bogduk, P.J. Mcnaught, R. Laurent, Prevalence and of using cooled radiofrequency in treating chronic sacroiliac joint pain A PRISMA-
clinical-features of lumbar zygapophysial joint pain – a study in an Australian po- compliant meta-analysis, Medicine 97 (6) (2018).
pulation with chronic low-back-pain, Ann. Rheum. Dis. 54 (2) (1995) 100–106. [17] W.O.C. Lopez, P.A. Navarro, M.D. Vargas, E. Alape, P.A.C. Lopez, Pulsed radio-
[4] A.C. Schwarzer, C.N. Aprill, R. Derby, J. Fortin, G. Kine, N. Bogduk, The false- frequency versus continuous radiofrequency for facet joint low back pain: a sys-
positive rate of uncontrolled diagnostic blocks of the lumbar zygapophysial joints, tematic review, World Neurosurg. 122 (2019) 390–396.
Pain 58 (2) (1994) 195–200. [18] L. Andriolo, S.A. Altamura, D. Reale, C. Candrian, S. Zaffagnini, G. Filardo,
[5] J.D. Eubanks, M.J. Lee, E. Cassinelli, N.U. Ahn, Prevalence of lumbar facet arthrosis Nonsurgical treatments of patellar tendinopathy: multiple injections of platelet-rich
and its relationship to age, sex, and race – an anatomic study of cadaveric speci- plasma are a suitable option: a systematic review and meta-analysis, Am J Sport
mens, Spine 32 (19) (2007) 2058–2062. Med 47 (4) (2019) 1001–1018.
[6] A.C. Gellhorn, J.N. Katz, P. Suri, Osteoarthritis of the spine: the facet joints, Nat. [19] Z.L. McCormick, H. Choi, R. Reddy, R.H. Syed, M. Bhave, M.C. Kendall, D. Khan,
Rev. Rheumatol. 9 (4) (2013) 216–224. G. Nagpal, M. Teramoto, D.R. Walega, Randomized prospective trial of cooled
[7] J. Sanapati, L. Manchikanti, S. Atluri, S. Jordan, S.L. Albers, M.A. Pappolla, versus traditional radiofrequency ablation of the medial branch nerves for the
A.D. Kaye, K.D. Candido, V. Pampati, J.A. Hirsch, Do regenerative medicine treatment of lumbar facet joint pain, Reg. Anesth. Pain Med. 44 (3) (2019)
therapies provide long-term relief in chronic low back pain: a systematic review and 389–397.
metaanalysis, Pain Physician 21 (6) (2018) 515–540. [20] K. Dutta, S. Dey, P. Bhattacharyya, S. Agarwal, P. Dev, Comparison of efficacy of
[8] S.H. Han, K.D. Park, K.R. Cho, Y. Park, Ultrasound versus fluoroscopy-guided

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.
C.-L. Shih, et al. Clinical Neurology and Neurosurgery 195 (2020) 105854

lateral branch pulsed radiofrequency denervation and intraarticular depot me- [30] W. Stelzer, M. Aiglesberger, D. Stelzer, V. Stelzer, Use of cooled radiofrequency
thylprednisolone injection for sacroiliac joint pain, Pain Physician 21 (5) (2018) lateral branch neurotomy for the treatment of sacroiliac joint-mediated low back
489–496. pain: a large case series, Pain Med. 14 (1) (2013) 29–35.
[21] V. Mehta, K. Poply, M. Husband, S. Anwar, R. Langford, The effects of radio- [31] K.Y. Ho, M.A. Hadi, K. Pasutharnchat, K.H. Tan, Cooled radiofrequency denervation
frequency neurotomy using a strip-lesioning device on patients with sacroiliac joint for treatment of sacroiliac joint pain: two-year results from 20 cases, J. Pain Res. 6
pain: results from a single-center, randomized, sham-controlled trial, Pain Physician (2013) 505–511.
21 (6) (2018) 607–618. [32] K. Ma, M. Yiqun, T. Wu, W. Wang, X. Liu, X. Huang, Y. Wang, Efficacy of diclofenac
[22] A. Cetin, A. Yektas, Evaluation of the short- and long-term effectiveness of pulsed sodium in pain relief after conventional radiofrequency denervation for chronic
radiofrequency and conventional radiofrequency performed for medial branch facet joint pain: a double-blind randomized controlled trial, Pain Med. 12 (1)
block in patients with lumbar facet joint pain, Pain Res. Manag. (2018). (2011) 27–35.
[23] J.N.S. Juch, E.T. Maas, R. Ostelo, J.G. Groeneweg, J.W. Kallewaard, B.W. Koes, [33] L. Kapural, F. Nageeb, M. Kapural, J.P. Cata, S. Narouze, N. Mekhail, Cooled
A.P. Verhagen, J.M. van Dongen, F. Huygen, M.W. van Tulder, Effect of radio- radiofrequency system for the treatment of chronic pain from sacroiliitis: the first
frequency denervation on pain intensity among patients with chronic low back case-series, Pain Pract. 8 (5) (2008) 348–354.
pain: the mint randomized clinical trials, Jama 318 (1) (2017) 68–81. [34] S.P. Cohen, R.W. Hurley, C.C. Buckenmaier 3rd, C. Kurihara, B. Morlando,
[24] W. Stelzer, V. Stelzer, D. Stelzer, M. Braune, C. Duller, Influence of BMI, gender, and A. Dragovich, Randomized placebo-controlled study evaluating lateral branch
sports on pain decrease and medication usage after facet-medial branch neurotomy radiofrequency denervation for sacroiliac joint pain, Anesthesiology 109 (2) (2008)
or SI joint lateral branch cooled RF-neurotomy in case of low back pain: original 279–288.
research in the Austrian population, J. Pain Res. 10 (2017) 183–190. [35] H.R. Kroll, D. Kim, M.J. Danic, S.S. Sankey, M. Gariwala, M. Brown, A randomized,
[25] K.H. Do, S.H. Ahn, Y.W. Cho, M.C. Chang, Comparison of intra-articular lumbar double-blind, prospective study comparing the efficacy of continuous versus pulsed
facet joint pulsed radiofrequency and intra-articular lumbar facet joint corticos- radiofrequency in the treatment of lumbar facet syndrome, J. Clin. Anesth. 20 (7)
teroid injection for management of lumbar facet joint pain: a randomized controlled (2008) 534–537.
trial, Medicine 96 (13) (2017) e6524. [36] I. Tekin, H. Mirzai, G. Ok, K. Erbuyun, D. Vatansever, A comparison of conventional
[26] C.W. van Tilburg, F.A. Schuurmans, D.L. Stronks, J.G. Groeneweg, F.J. Huygen, and pulsed radiofrequency denervation in the treatment of chronic facet joint pain,
Randomized sham-controlled double-blind multicenter clinical trial to ascertain the Clin. J. Pain 23 (6) (2007) 524–529.
effect of percutaneous radiofrequency treatment for sacroiliac joint pain: three- [37] N. Patel, Twelve-month follow-up of a randomized trial assessing cooled radio-
month results, Clin. J. Pain 32 (11) (2016) 921–926. frequency denervation as a treatment for sacroiliac region pain, Pain Pract. 16 (2)
[27] V.S. Anjana Reddy, C. Sharma, K.Y. Chang, V. Mehta, ‘Simplicity’ radiofrequency (2016) 154–167.
neurotomy of sacroiliac joint: a real life 1-year follow-up UK data, Br. J. Pain 10 (2) [38] J. Dobrogowski, A. Wrzosek, J. Wordliczek, Radiofrequency denervation with or
(2016) 90–99. without addition of pentoxifylline or methylprednisolone for chronic lumbar zy-
[28] F.R. Romero, R.B. Vital, M.A. Zanini, L.G. Ducati, R.C. Gabarra, Long-term follow- gapophysial joint pain, Pharmacol. Rep. 57 (4) (2005) 475–480.
up in sacroiliac joint pain patients treated with radiofrequency ablative therapy, [39] L. Canovas Martinez, J. Orduna Valls, E. Parames Mosquera, L. Lamelas Rodriguez,
Arq. Neuropsiquiatr. 73 (6) (2015) 476–479. S. Rojas Gil, M. Dominguez Garcia, Sacroiliac joint pain: Prospective, randomised,
[29] Y. Zheng, M. Gu, D. Shi, M. Li, L. Ye, X. Wang, Tomography-guided palisade sa- experimental and comparative study of thermal radiofrequency with sacroiliac joint
croiliac joint radiofrequency neurotomy versus celecoxib for ankylosing spondylitis: block, Rev. Esp. Anestesiol. Reanim. 63 (5) (2016) 267–272.
a open-label, randomized, and controlled trial, Rheumatol. Int. 34 (9) (2014) [40] S.M. Aydin, C.G. Gharibo, M. Mehnert, T.P. Stitik, The role of radiofrequency ab-
1195–1202. lation for sacroiliac joint pain: a meta-analysis, PM R 2 (9) (2010) 842–851.

10

Downloaded for Anonymous User (n/a) at Comunidad de Madrid Consejeria de Sanidad from ClinicalKey.com by Elsevier on March 23, 2021.
For personal use only. No other uses without permission. Copyright ©2021. Elsevier Inc. All rights reserved.