A Systematic Review and Meta-Analysis of Moxibustion Treatment for Knee Osteoarthritis

Jang Mi Park; Cham Kyul Lee; Kyung Ho Kim; Eun Jung Kim; Chan Yung Jung; Byung-Kwan Seo; Bon Hyuk Goo; Yong Hyeon Baek; Eun Yong Lee

doi:10.13045/jar.2020.00108

J Acupunct Res > Volume 37(3); 2020 > Article

Park, Lee, Kim, Kim, Jung, Seo, Goo, Baek, and Lee: A Systematic Review and Meta-Analysis of Moxibustion Treatment for Knee Osteoarthritis

Review Article

Journal of Acupuncture Research 2020; 37(3): 137-150.

Published online: August 24, 2020

DOI: https://doi.org/10.13045/jar.2020.00108

A Systematic Review and Meta-Analysis of Moxibustion Treatment for Knee Osteoarthritis

Jang Mi Park¹

, Cham Kyul Lee¹

, Kyung Ho Kim²

, Eun Jung Kim³

, Chan Yung Jung⁴

, Byung-Kwan Seo⁵

, Bon Hyuk Goo⁵

, Yong Hyeon Baek⁵

, Eun Yong Lee^1,^*

¹Department of Acupuncture and Moxibustion Medicine, Chungju Hospital of Korean Medicine, Semyung University, Chungju, Korea

²Department of Acupuncture and Moxibustion Medicine, Dongguk University Ilsan Oriental Medicine Hospital, Goyang, Korea

³Department of Acupuncture and Moxibustion Medicine, Dongguk University Bundang Oriental Medicine Hospital, Seongnam, Korea

⁴Institute of Oriental Medicine, College of Korean Medicine, Dongguk University, Goyang, Korea

⁵Department of Acupuncture and Moxibustion, Kyung Hee University, Korean Medicine Hospital, Gangdong, Seoul, Korea

^*Corresponding author. Eun Yong Lee, Department of Acupuncture and Moxibustion Medicine, Chungju Hospital of Traditional Korean Medicine, Semyung University, Chungju, Korea, E-mail: acupley@naver.com

Received April 24, 2020 Revised June 24, 2020 Accepted June 29, 2020

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Abstract

The purpose of this study was to evaluate the evidence supporting the effectiveness of moxibustion treatment for osteoarthritis. There were 9 databases used to retrieve randomized controlled trials (RCTs) that used moxibustion as treatment for osteoarthritis. The quality of methodology for the RCTs was assessed using the Cochrane Risk of Bias tool [Review Manager (RevMan) Version 5.3 Windows, The Nordic Cochrane Centre, Copenhagen, Denmark]. The inclusion criteria for this review was met by 27 RCTs. All studies were conducted in China. A 4-week moxibustion treatment period was the most common. EX-LE4 and SP10 and GB34 acupoints were most frequently selected in the treatment of osteoarthritis. The most commonly used evaluation index was the visual analog scale (VAS). All studies, including a meta-analysis showed that moxibustion treatments were statistically significantly effective at treating knee osteoarthritis. However, well-designed randomized RCTs without a high risk of bias should be designed in the future.

Key Words: knee osteoarthritis; meta-analysis; moxibustion; randomized controlled trial; systematic review.

Introduction

Osteoarthritis is a degenerative disease characterized by the erosion of joint cartilage, hardening of the subchondral bone, bone hypertrophy, and joint deformation due to changes in cartilage metabolism of the joint and increased physical pressure [1]. Osteoarthritis is most common in knee joints that are directly subject to weight load [2]. Osteoarthritis causes chronic pain, stiffness, joint instability, and limited range of motion [3]. The prevalence of osteoarthritis is higher in older age, and in Korea, about 80% of patients over 55 years old, and almost all elderly people (65 years and older) have knee osteoarthritis [4]. As an aging society, the prevalence of osteoarthritis is expected to increase.

Osteoarthritis treatment primarily focuses on conservative treatments such as exercise, weight loss, heat therapy, non-steroidal anti-inflammatory drugs, and intra-articular injections of hyaluronic acid [5]. Surgical treatment may be performed if there is severe pain, joint instability, and movement limitations due to severe degenerative changes [5]. In traditional Korean medicine (which is a representative conservative treatment), various treatments such as moxibustion, acupuncture, pharmacopuncture, and herbal medicine are used to treat osteoarthritis. In Korea, fire needling treatment of osteoarthritis, and Bee venom treatment of osteoarthritis have been reported [6,7]. However, there are no systematic reviews or meta-analyses of moxibustion treatment of osteoarthritis in Korea.

Moxibustion is a representative Korean medicine treatment that promotes blood circulation by burning wormwood in acupuncture points, and is used in various musculoskeletal disorders including osteoarthritis [8].

The purpose of this study was to investigate the effects of moxibustion treatment of knee osteoarthritis through a systematic review and meta-analysis to understand the effectiveness of moxibustion treatment. Randomized controlled trials (RCTs) of moxibustion treatment for osteoarthritis were retrieved using national and international databases, and analyzed to review research trends and effectiveness of treatment.

Materials and Methods

Search strategy

PubMed, EMBASE, the Cochrane library, Citation Information by NII (CiNii), China National Knowledge Infrastructure (CNKI), Korean Medical database (KMBASE), Korean studies Information Service System (KISS), National Digital Science Library (NDSL), the Korean database Oriental medicine Advanced Searching Integrated System (OASIS) were searched to retrieve related articles until the 10^th August, 2019.

The subject search terms were based on “Medium Subject Headings (MeSH)” including “arthritis,” “osteoarthritis,” “knee joint,” “degenerative arthritis,” “knee,” “moxibustion,” “moxa,” and “randomized controlled trial.” The actual search term was modified according to the language of each database.

Eligibility criteria

Inclusion criteria

The purpose of this study was to determine whether moxibustion treatment was an effective treatment for knee osteoarthritis. This study included RCTs applying moxibustion treatment to osteoarthritis. There were no restrictions on the language or the year of publication of the papers. In order to confirm the effect of the thermal stimulation of moxibustion treatment, no limitation was placed on the type of moxibustion, and the method of treatment. Although comparative interventions included treatments other than moxibustion, studies that analyzed the effectiveness of moxibustion treatments were included in this study.

Exclusion criteria

Literature reviews, case reports, observational research, animal experiment research, management studies after surgery (e.g. knee replacement), studies on treatment methods that are not commonly practiced in Korean medicine, studies conducted in children and adolescents were excluded. Moxibustion studies in both the intervention group and the control group, or studies comparing moxibustion types were excluded because the effectiveness of moxibustion treatments could not be compared. In addition, studies in which only abstracts were available, but not the original article were excluded.

Data collection and risk of bias

Studies that met the selection criteria and 2 reviewers independently assessed and extracted the data. If there was any discrepancy, it was re-evaluated by a third party with similar qualifications as the 2 reviewers. To assess the bias risk of the RCTs, the Cochrane risk of bias tool (RoB) was used.

Statistical analysis

To summarize the effects of moxibustion treatments for knee osteoarthritis in each study, the Cochrane Collaboration software [Review Manager (RevMan) Version 5.3 for Windows, The Nordic Cochrane Centre, Copenhagen, Denmark] was used. The relative risk, the standardized mean difference, the mean difference and the 95% confidence interval were calculated. The results of the included studies were synthesized using a fixed effect model.

Results

Study selection

As a result of the database search, a total of 869 articles were retrieved (PUBMED 114, EMBASE 150, Cochrane 169, CiNii 84, CNKI 217, KMBASE 16, KISS 16, NDSL 44, OASIS 15, and 44 additional records identified through other sources). Of these, 260 duplicated articles were removed, and the primary screening was conducted on 609 articles to exclude studies that were not related to knee osteoarthritis, non-clinical studies, and non-RCTs which was performed by reading the titles and abstracts leaving 511 studies. As a result of checking full texts, 98 articles remained, from which a total of 69 articles were excluded, including 17 articles that did not contain the original text, 1 article that did not meet the selection criteria for the intervention and control group, and 51 articles that set improper groups. Finally, a total of 27 articles were selected from this systematic literature review, and meta-analysis were conducted. The selection process is illustrated using the flow chart of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA; Fig. 1).

Year of publication

There were 6 studies conducted in 2015, 4 studies each in 2011 and 2017, 3 studies in 2008, 2 studies in 2009, 2010, 2013, and 2014, and 1 study in 2016 and 2018.

Sample size of study

There was a total of 2,175 patients included in the RCTs in the 27 articles included in this review. In each study, the total number of patients in both intervention and control groups was at least 31 and at most 185. There was 1 study with less than 50, 19 studies with more than 50 and less than 100, 6 studies with more than 100 and less than 150, and 1 study with more than 150.

Duration and frequency of treatment

Ten studies had a treatment duration of less than 4 weeks and 16 studies had more than 4 weeks. There was 1 study with 10 weeks of treatment [9]. One study did not report the exact duration of treatment [10].

The sessions of treatment were to some extent proportional to the duration of treatment. There were 2 studies with less than 10 sessions of treatment, 12 studies with more than 10 and less than 20 sessions of treatment, and 11 studies with more than 20 sessions of treatment. There were 2 studies that did not report the exact number of treatments [10,11].

Frequency of acupoints

As a result of analyzing selected acupoints in the intervention group, EX-LE4 was the most frequently selected (16 times in the 27 studies). In addition, acupoints such as SP10, GB34, ST35, and ashi point were frequently used.

Evaluation index

The most commonly used evaluation index was the visual analog scale (VAS) which was used as an evaluation scale in 12 studies. Effective rate was used as an evaluation index for 11 studies. Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) was used as an evaluation index in 8 studies, and Short Form 36 (SF-36) was used in 2 studies. In addition to these evaluation indexes, numerical rating scale (NRS), and knee-joint function scores (IAC Lennox, Lysholm) were used. The results are summarized in Table 1 [9–35].

Analysis of treatment effect

In most studies, the intervention group showed better results after treatment than in the control group. However, Zhang et al [14] showed good treatment effects using VAS and total effective rate, but there was no mention of p value, which made it difficult to determine statistical significance. In the study of Zhou et al [23], the effective rate immediately after treatment and 2 months after treatment was significantly higher in the intervention group than in the control group (p < 0.05). However, the pain scores measured by VAS did not differ significantly between the 2 groups. In the study of Cheng et al [10] using the NRS as an evaluation index, both the intervention group and the control group showed significant pain reduction, but there was no statistically significant difference. In the study of Wu and Xiong [25], there was no difference in the effective rates of the interventional group (heat-sensitive acupoint moxibustion) and control group (intra-articular injection), but heat-sensitive acupoint moxibustion showed a faster treatment effect. In all other studies except for the study by Zhou et al [15] and Wu et al [16], which did not mention the p value, the intervention group showed a statistically significant improvement over the control group.

Analysis of side effect

As a result of analyzing adverse events in the intervention group, adverse events were not reported in 19 studies. In 1 study, adverse events were reported (skin flushing), but the patients’ symptoms disappeared within 3 days, and without any treatment [27]. The remaining 7 studies reported adverse events but none were recorded.

Risk of bias

For the 27 RCTs included in this review, 2 researchers evaluated the risk of bias by applying the Cochrane risk assessment tool [risk of bias (RoB) by Cochrane collaboration] to each study. Evaluation of all items was accepted only when the contents of the selected study were clearly stated. In case of disagreement between evaluators, there was sufficient discussion and other researchers’ opinions were reflected upon. The results of the bias risk assessment are plotted using the Revman program (Figs. 2 and 3 ).

Random sequence generation

There were 18 Low-Risk studies in which random assignments were made using random number tables and lottery tickets. The remaining 8 studies were classified as Unclear Risk because only the expression “random assignment” existed, and there was no mention of the process of random assignment. Only 1 study [21] had a High Risk of bias. This was because odd numbers were assigned to the intervention group and even numbers were assigned to the control group and odd and even numbers were assigned in the order of the visit.

Allocation concealment

In 5 studies [16,24,27,29,32], the risk of bias was Low Risk. The opaque sealed envelope was assigned a serial number and the assignment order was kept. The remaining studies did not mention the allocation order concealment, so they were classified as an Unclear Risk.

Blinding of participants and personnel

Of the 27 studies, 22 were classified as High Risk. This was because of the nature of moxibustion treatment was difficult to blind. There were 5 studies [12,16,24,27,29] which had a Low Risk of bias due to double blinding.

Blinding of outcome assessment

There were 3 studies [24,27,29] that had a Low Risk of bias because the study was blinded to the reviewer, but the other studies had an Unclear Risk of bias because there were no remarks on the related contents.

Incomplete outcome data

All 27 studies were classified as Low Risk because they had no missing values because the total number of participants in the intervention group and the control group matched before and after the study.

Selective reporting

All 27 studies had a protocol for the study and the risk of bias was Low Risk because the study handled the results in a predetermined way.

Other bias

Since there was no risk of additional bias, all 27 studies were classified as Low Risk.

Meta-analysis result

Moxibustion versus sham moxibustion

Meta-analysis of all 27 studies comparing moxibustion with sham moxibustion showed statistically significant results when using the WOMAC evaluation index (p < 0.0001), WOMAC index for stiffness (p = 0.02), and WOMAC index for function (p = 0.007; Fig. 4).

Moxibustion versus conventional Western medicine

The meta-analysis of studies comparing moxibustion with conventional Western medicine (analgesic, glucosamine, injection, ointment) showed that moxibustion treatment was statistically significantly more effective when using the VAS score for pain (p < 0.00001), the WOMAC index total (p = 0.25), WOMAC index for pain (p < 0.0001), WOMAC index for stiffness (p < 0.0001), WOMAC index for function (p = 0.002), and the effective rate of the moxibustion treatment (p < 0.00001) compared with conventional Western medicine (Figs. 5 –7).

Moxibustion with conventional Western medicine versus conventional Western medicine alone

The meta-analysis of studies of moxibustion treatment with conventional Western medicine comparing with studies of conventional Western medicine alone, showed that the intervention group had no statistically significant effect compared with the control group in pain reduction (VAS; p = 0.68), but moxibustion treatment was statistically significantly more effective when the effective rate (p = 0.001) was used compared with conventional Western medicine (Fig. 8).

Moxibustion with conventional Korean medicine versus conventional Korean medicine without moxibustion

The meta-analysis of studies of moxibustion treatment with conventional Korean medicine compared with studies of conventional Korean medicine without moxibustion showed that both pain reduction (VAS; p = 0.0004) and the effective rate of treatment (p = 0.04) showed statistically significant improvement when moxibustion was used for treatment (Fig. 9).

Discussion

The purpose of this study was to evaluate the effect of moxibustion treatment for osteoarthritis of the knee. There were 27 RCT studies selected for this review based on a preselected selection and exclusion criteria, using 9 online databases. All 27 selected studies were conducted in China.

According to the purpose of the study, there were 21 studies comparing moxibustion with Western medicine. There were 4 studies comparing moxibustion with a placebo control, and 2 studies comparing moxibustion with Korean medicine. Celecoxib, diclofenac, ibuprofen, glucosamine, and sodium hyaluronate were used as the Western medicine group. Among them, diclofenac was used in 10 studies.

All the treatments used in the Korean medicine group were acupuncture. EX-LE4, ST35, GB34, and ST35 were the most selected acupoints. In addition, most of the acupoints around the knee were commonly used.

VAS, effective rate, WOMAC, SF-36, NRS, and Knee Function Score were used as the evaluation index, and VAS (the measure of pain), was used the most. Typically, the intervention group in most studies showed improvement over the control group after treatment. Zhang et al [14] showed good treatment effects using both the VAS and the total effective rate: 5.1 ± 1.33 in the intervention group, 5 ± 1.63 in the control group, and 83.3% in the intervention group, 86.6% in the control group. However, it was difficult to judge statistical significance because there was no mention of p value. In the study of Zhou et al [23], the markedly effective rates immediately after treatment and 2 months after treatment showed a statistically significant difference of 58% in the intervention group, and 27.1% in the control group (p < 0.05). However, the pain score measured by the VAS was 5.53 ± 1.46 in the intervention group, and 5.31 ± 1.55 in the control group. There was no significant difference between the 2 groups. In the study by Cheng et al [10] using the NRS as an evaluation index, the score changed from 6.42 ± 2.32 to 1.13 ± 0.87 in the intervention group, and 6.31 ± 2.41 to 1.28 ± 0.85 in the control group which showed significant pain reduction after treatment compared with before treatment. However, there was no statistically significant difference. The total effective rate of the study by Wu and Xiong [25] was 95.8% in the intervention group (heat-sensitive acupoint moxibustion) and 100% in the control group (intra-articular injection). There was no significant difference between the 2 groups, but heat-sensitive acupoint moxibustion had a faster treatment effect.

One of the 27 studies showed side effects (skin flushing) in the moxibustion treatment group. The remaining 19 studies did not mention measurement of adverse events or side effects. The remaining 7 studies reported no adverse reactions after moxibustion. However, moxibustion treatment is a therapy that applies heat stimulation, so there was a risk of burns which always requires the attention of the clinician.

Analyzes from the Cochran bias risk assessment tool “RoB” by Cochrane collaboration determined that the majority of the studies used a random number table or a lottery table to assign the order of placement. Thus, the risk of bias was Low Risk in most studies. Most studies, however, did not mention allocation concealment and blinding of outcome assessment. In addition, due to the nature of moxibustion as interventional treatment tool, there was a limit to blinding of participants and personnel.

According to the study design, RCTs could be classified into 4 types: RCTs comparing moxibustion with sham moxibustion, RCTs comparing moxibustion with conventional Western medicine, RCTs comparing the intervention group that combined moxibustion and conventional Western medicine with the control group that performed only conventional Western medicine, and RCTs comparing the intervention group that combined moxibustion with conventional Korean medicine with the control group that performed conventional Korean medicine without moxibustion.

As a result of meta-analysis, moxibustion treatment compared with sham moxibustion showed statistically significant results in the WOMAC index (pain, stiffness, function). In addition, moxibustion treatment compared with conventional Western medicine showed statistically significant results in the VAS, WOMAC index (pain, stiffness, function), and effective rate. The combination of moxibustion treatment and Western medicine showed a significant effect in the effective rate measure compared with the control group using only Western medicine, but did not have a significant effect in the pain reduction scale (VAS score). RCTs combined with moxibustion treatment and conventional Korean medicine showed statistically significant effects on both the effective rate and the VAS score compared with the control group that performed conventional Korean medicine without moxibustion.

This study confirmed that moxibustion treatment had a statistically significant effect on knee osteoarthritis. However, research from other databases and unpublished research are not included in this review. All the selected studies for this review were conducted in China which may limit the validity of moxibustion efficacy through objective meta-analysis. In addition, because there were many studies using subjective evaluation indicators and lack of common evaluation indicators, there were limitations in comparing and interpreting the results. In the future, standardized evaluation indicators are needed. In the results related to the risk of bias, there were items which were not mentioned, and there was a limit in the capacity to blind the study participants. So, in order to confirm the effectiveness of moxibustion treatment further research is needed into research design to minimize the risk of bias.

In conclusion, based on the results of this systematic review and meta-analysis, moxibustion can be considered as a treatment for knee osteoarthritis.

Conflict of Interest

Conflicts of Interest

The authors have no conflicts of interest to declare.

Fig. 1

A flow diagram according to the PRISMA.

PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses; RCT, randomized controlled trial.

Fig. 2

Risk of bias summary for randomized controlled trials.

Fig. 3

Risk of bias for randomized controlled trials.

Fig. 4

The meta-analysis of moxibustion versus sham moxibustion (WOMAC pain, stiffness, function).

WOMAC, Western Ontario and McMaster Universities Arthritis Index.

Fig. 5

The metaanalysis of moxibustion versus conventional Western medicine (VAS, WOMAC total).

VAS, visual analog scale; WOMAC, Western Ontario and McMaster Universities Arthritis Index.

Fig. 6

The meta-analysis of moxibustion versus conventional Western medicine (WOMAC pain, stiffness, function).

WOMAC, Western Ontario and McMaster Universities Arthritis Index.

Fig. 7

The meta-analysis of moxibustion versus conventional Western medicine (effective rate).

Fig. 8

The meta-analysis of moxibustion with conventional Western medicine versus conventional Western medicine (VAS, effective rate).

VAS, visual analog scale.

Fig. 9

The metaanalysis of moxibustion with conventional Korean medicine versus conventional Korean medicine without moxibustion (VAS, Effective rate).

VAS, visual analog scale.

Table 1

Data of Included RCTs.

First author (y) [ref]	Sample size total (I:C)	Intervention group	Control group	Outcome measurements	Summary of results
Wang (2015) [9]	60 (30:30)	Ginger moxibustion (n = 30)	Glucosamine sulfate potassium capsule 0.75 g/d + intra-articular injection (sodium hyaluronate, 2 mL/week) (n = 30)	WOMAC (total)	(I vs C, baseline→immediate post treatment) WOMAC total: 69.13 ± 6.2→35.33 ± 17.27 (p = 0) vs 69.3 ± 5.58→45.67 ± 13.22 (p = 0). IG vs CG (p = 0.012)
Cheng (2008) [10]	120 (60:60)	Moxibustion (n = 60)	Diclofenac sodium 75 mg/d (n = 60)	NRS	(I vs C, baseline→immediate post treatment) NRS: 6.42 ± 2.32→1.13 ± 0.87 (p < 0.01) vs 6.31 ± 2.41→1.28 ± 0.85 (p < 0.05). IG vs CG (ns)
Zhang (2015) [11]	185 (48:46:43:48)	IG: Moxibustion (n = 48)	CG1: Rehabilitation therapy (n = 46) CG2: Moxibustion + rehabilitation therapy (n = 43) CG3: Celecoxib 200 mg/d (n = 48)	1) WOMAC score 2) VAS	(I vs C1 vs C2 vs C3, Baseline→ immediate post treatment) 1) WOMAC total: 32.47 ± 3.91 vs 41.23 ± 8.85 vs 44.2 ± 7.24 vs 29.36 ± 5.28. IG vs CG1 (p < 0.05), IG vs CG2 (p < 0.05), IG vs CG3 (NS). 2) VAS: 2.47 ± 0.67 vs 4.62 ± 0.92 vs 4.44 ± 0.83 vs 3.35 ± 0.81. IG vs CG1 (p < 0.05), IG vs CG2 (p < 0.05), IG vs CG3 (p < 0.05)
Ren (2011) [12]	59 (31:28)	Moxibustion (n = 31)	Placebo moxibustion (n = 28)	WOMAC score	(I vs C, Baseline→12 wk after treatment) 1) WOMAC pain: 32.42 ± 16.97→12.67 ± 7.51 (p < 0.01) vs 26.94 ± 13.13→22.47 ± 15.41. IG vs CG (p < 0.01) 2) WOMAC function: 32.07 ± 16.7→15.39 ± 12.56 (p < 0.01) vs 28.35 ± 15.06→23.73 ± 18.07. IG vs CG (p < 0.05) 3)WOMAC stiffness: 37.15 ± 22.73→15.39 ± 11.74 (p < 0.01) vs 33.95 ± 20.49→26.18 ± 22.6. IG vs CG (p < 0.05)
Zhao (2013) [13]	87 (35:33:19)	IG1: moxibustion (n = 35) IG2: EA (n = 33)	Celebrex 0.2 g/d (n = 19)	1) VAS 2) knee-joint function score (IAC Lennox + Lysholm)	(I1 vs I2 vs C, immediate post treatment) 1) VAS: 3.41 ± 1.73 vs 3.43 ± 1.89 vs 2.56 ± 1.47. IG1 vs CG (p < 0.05), IG2 vs CG (p < 0.05) 2) knee-joint function score: 89.11 ± 10.47 vs 88.7 ± 11.54 vs 71.63 ± 18.69. IG1 vs CG (p < 0.05), IG2 vs CG (p < 0.05)
Zhang (2011) [14]	60 (30:30)	Moxibustion (n = 30)	Celecoxib 200 mg/d (n = 30)	1) VAS 2) Effective rate	(I vs C, immediate post treatment) 1) VAS: 5.1 ± 1.33 vs 5 ± 1.63 2) Effective rate: 25/30 (83.3%) vs 26/30 (86.6%)
Zhou (2014) [15]	105 (39:44:22)	Moxibustion (n = 39)	CG1: EA (n = 44) CG2: Celebrex 0.2 g/d (n = 22)	VAS	(I vs C1 vs C2, immediate post treatment) VAS: 2.76 ± 1.95 vs 3.07 ± 1.62 vs 3.65 ± 1.42
Wu (2011) [16]	31 (15:16)	Laser moxibustion	Sham laser moxibustion	SF-36	(I vs C, Baseline→8 wk after treatment) 1) SF-36 PF (physical): 60.67 ± 16.78→75 ± 13.09 (p < 0.05) vs 60.31 ± 15.65→65.31 ± 13.96 2) SF-36 VT (physical & mental): 50.33 ± 21.25→55 ± 21.04 vs 49.38 ± 16.62→50 ± 15.38 3) SF-36 MH (mental): 58.4 ± 22.52→60.27 ± 14.61 vs 69.5 ± 15.03→64.5 ± 11.94 4) SF-36 SF (physical & mental): 67.5 ± 19.93→75 ± 15.67 vs 73.44 ± 12.81→75.78 ± 9.64 5) SF-36 BP (physical): 54.5 ± 14.52→63.83 ± 16.58 vs 64.22 ± 10.98→64.38 ± 9.77 6) SF-36 GH (physical & mental): 40 ± 21.88→46 ± 19.01 vs 43.75 ± 9.22→43.75 ± 9.04
Sun (2008) [17]	56 (29:27)	Aconite cake-separated moxibustion (n = 29)	Sodium diclofenate slow-released tablet 75 mg/d (n = 27)	Effective rate	(I vs C, immediate post treatment, 10 wk after treatment) 1) effective rate: 39/41 vs 35/39. IG vs CG (NS) 2) markedly effective rate: 26/41 (63.4%) vs 19/39(48.7%). IG vs CG (p < 0.05)
Deng (2015) [18]	70 (35:35)	Salt moxibustion (n = 35)	Ibuprofen 0.6 g/d (n = 35)	1) WOMAC (total) 2) Effective rate	(I vs C, immediate post treatment) 1) WOMAC total: 22.63 ± 9.54 vs 31.22 ± 8.62. IG vs CG (p < 0.05) 2) effective rate: 32/35 (91.43%) vs 26/35 (74.29%). IG vs CG (p < 0.05)
Yuan (2015) [19]	148 (74:74)	Heat-sensitive point Thunder-fire moxibustion (n = 74)	Medication (diclofenac sodium enteric-coated tablet, 100 mg/d)	1) VAS 2) WOMAC (total)	(I vs C, immediate post treatment) 1) VAS: 24.07 ± 2.83 vs 36.22 ± 2.32. IG vs CG (p < 0.01) 2) WOMAC total: 14.15 ± 4.81 vs 15.12 ± 4.82. IG vs CG (p < 0.01)
Deng (2015) [20]	80 (40:40)	Salt moxibustion (n = 40)	Ibuprofen 0.6 g/d (n = 40)	VAS	(I vs C, baseline→immediate post treatment) VAS: 6.89 ± 1.28→0.98 ± 0.63 (p < 0.05) vs 7.08 ± 2.06→ 1.01 ± 0.86 (p < 0.05). IG vs CG (p < 0.05)
Yang (2008) [21]	64 (33:31)	Notoginseng cake-separated moxibustion (n = 33)	Diclofenac sodium sustained release tablets 75 mg/d (n = 31)	Effective rate	(I vs C, 5 session→2 months after treatment) Effective rate: IG vs CG (p < 0.05)→IG vs CG (p < 0.01)
Song (2013) [22]	80 (40:40)	Notoginseng cake-separated moxibustion (n = 40)	Diclofenac sodium sustained release tablets 75 mg/d (n = 40)	WOMAC (pain, stiffness, function)	(I vs C, baseline→2 months after treatment) 1) WOMAC pain: 15.49 ± 1.22→7.48 ± 2.24 (p < 0.05) vs 15.39 ± 1.21→7.53 ± 2.25 (p < 0.05). IG vs CG (ns) 2) WOMAC stiffness: 4.21 ± 1.18→1.26 ± 0.76 (p < 0.05) vs 4.16 ± 1.2→3.09 ± 1.08 (p < 0.05). IG vs CG (p < 0.05) 3) WOMAC function: 48.51 ± 3.68→28.94 ± 3.77 (p < 0.05) vs 48.26 ± 34.08→34.08 ± 4.7024 (p < 0.05). IG vs CG (p < 0.05)
Zhou (2010) [23]	70 (35:35)	Notoginseng cake-separated moxibustion (n = 35)	Diclofenac sodium sustained release tablets 75 mg/d (n = 35)	1) Effective rate 2) VAS (pain)	(I vs C, 2 months after treatment) 1) Effective rate: 45/50 (90%) vs 39/48 (81.3%). IG vs CG (ns) 2) Markedly effective rate: 29/50 (58%) vs 18/48 (27.1%). IG vs CG (p < 0.05) 3) VAS: 5.53 ± 1.46 vs 5.31 ± 1.55. IG vs CG (ns)
Ren (2015) [24]	136 (69:67)	Moxibustion (n = 69)	Sham moxibustion (n = 67)	SF-36	(I vs C, Baseline→12 wk after treatment) 1) SF-36 PF (physical): 57.83 ± 16.88→63.99 ± 18.24 (p < 0.01) vs 60.15 ± 17.21→66.04 ± 14.42 (p < 0.01) 2) SF-36 RP (physical): 33.7 ± 37.33→57.49 ± 43.48 vs 42.91 ± 41.7→45.15 ± 40.88 3) SF-36 RE (mental): 49.76 ± 44.14→57.49 ± 43.87 vs 56.22 ± 41.52→50.25 ± 44.71 4) SF-36 VT (physical & mental): 53.55 ± 18.19 (p < 0.01)→59.42 ± 15.87 vs 52.99 ± 18.55→56.49 ± 3.09 5) SF-36 MH (mental): 78.62 ± 18.7→79.17 ± 13.33 vs 68.78 ± 15.84→66.99 ± 14.43 6) SF-36 SF (physical & mental): 58.55 ± 12.82→67.54 ± 14.45 vs 75 ± 16.71→78.54 ± 15.05 7) SF-36 BP (physical): 6.73 ± 2.35→3.03 ± 2.33 (p < 0.001) vs 57.69 ± 18.49→65.93 ± 13.53 (p < 0.01) 8) SF-36 GH (physical & mental): 50.29 ± 15.31→54.42 ± 15.92 (p < 0.05) vs 47.99 ± 15.88→48.43 ± 12.01
Wu (2011) [25]	50 (24:26)	Heat-sensitive acupoint moxibustion (n = 24)	Intra-articular injection (sodium hyaluronate 2 mL/week) (n = 26)	Effective rate	(I vs C, baseline→immediate post treatment) Effective rate: 23/24 (95.8%) vs 26/26 (100%). IG vs CG (ns)
Chi (2016) [26]	120 (40:40:40)	Heat-senitization acupoint moxibustion (n = 40)	CG1: Moxibustion (n = 40) CG2: intra-articular sodium hyaluronate injection 2 mL/6 d (n = 40)	Effective rate	(I vs C1 vs C2, 6 months after the last moxibustion session) Effective rate: 50% vs 27.5% vs 30%. IG vs CG1 vs CG2 (p < 0.05)
Zhao (2014) [27]	110 (55:55)	Moxibustion (n = 55)	Sham moxibustion (n = 55)	WOMAC score	(I vs C, Baseline→24 wk after treatment) 1) WOMAC pain: 6.73 ± 2.35→3.03 ± 2.33 vs 6.29 ± 2.7→4.56 ± 3.09. IG vs CG (p < 0.001) 2) WOMAC function: 33.47 ± 15.37→16.43 ± 12.16 vs 33.99 ± 17.82→21.7 ± 16.53. IG vs CG (p = 0.015)
Bu (2018) [28]	62 (31:31)	Heat-sensitive acupoint moxibustion (n = 31)	Diclofenac sodium gel (n = 31)	VAS	(I vs C, baseline→immediate post treatment) VAS: 8.34 ± 0.98→2.01 ± 0.28 (p < 0.05) vs 8.12 ± 1.05→5.76 ± 0.71 (p < 0.05). IG vs CG (p < 0.05)
Zhou (2017) [29]	60 (30:30)	Moxibustion + placebo gel 12 g/d (n = 30)	Placebo moxibustion + diclofenac sodium gel 12 g/d (n = 30)	1) WOMAC 2) VAS	(I vs C, baseline→4 wk after treatment) 1) WOMAC total: 35.47 ± 15.45→19.37 ± 10.53 (p < 0.05) vs 35.73 ± 7.68→24.57 ± 5.32 (p < 0.05). IG vs CG (p = 0.043) 2) WOMAC pain: 8.07 ± 3.36→4.07 ± 2.16 (p < 0.05) vs 8.41 ± 2.29→4.37 ± 1.4 (p < 0.05). IG vs CG (p = 0.32) 3) WOMAC stiffness: 2.03 ± 1.99→1.07 ± 1.23 (p < 0.05) vs 2.07 ± 1.26→1.3 ± 0.84 (p < 0.05). IG vs CG (p = 0.195) 4) WOMAC function: 24.37 ± 11.43→14.23 ± 7.84 (p < 0.05) vs 25.27 ± 5.13→18.9 ± 3.93 (p < 0.05). IG vs CG (p = 0.036) 5) VAS: 3.98 ± 1.2→2.42 ± 0.86 (p < 0.05) vs 4.45 ± 1.01→2.67 ± 0.61 (p < 0.05). IG vs CG (p = 0.183)
Zhang (2009) [30]	60 (30:30)	Moxibustion (n = 30)	Placebo moxibustion + diclofenac sodium gel 12 g/d (n = 30)	Effective rate	(I vs C, immediate post treatment) Effective rate: 83.33% vs 80%. IG vs CG (p < 0.01)
Zheng (2017) [31]	60 (30:30)	Acupuncture + ginger moxibustion (n = 30)	Acupucntrue (n = 30)	VAS	(I vs C, baseline→immediate post treatment) VAS: 7.37 ± 1.56→2.27 ± 1.26 (p < 0.05) vs 7.1 ± 1.69→3.13 ± 1.53 (p < 0.05). IG vs CG (p < 0.05)
Zhang (2010) [32]	62 (32:30)	Non-scarring moxibustion + acupuncture (n = 32)	Acupuncture (n = 30)	1) Effective rate 2) VAS	(I vs C, immediate post treatment) 1) Effective rate: 30/32 (93.8%) vs 26/30 (86.7%). IG vs CG (p < 0.05) 2) VAS: 0.63 ± 0.35 vs 1.23 ± 1.18. IG vs CG (p < 0.01)
He (2009) [33]	60 (30:30)	Moxibustion + diclofenac sodium 75 mg/d (n = 30)	Diclofenac sodium 25 mg/d (n = 30)	Effective rate	(I vs C, immediate post treatment) Effective rate: 27/30 (90%) vs 20/30 (66.67%). IG vs CG (p < 0.01)
Tian (2017) [34]	60 (30:30)	External application of indirect moxibustion + glucosamine 0.942 g/d (n = 30)	Glucosamine 0.942 g/d (n = 30)	Effective rate	(I vs C, immediate post treatment) Effective rate: 56.67% vs 33.33%. IG vs CG (p < 0.05)
Xiao (2017) [35]	60 (30:30)	Moxibustion + intra-articular injection (sodium hyaluronate 2 mL/wk) (n = 30)	Intra-articular injection (sodium hyaluronate 2 mL/wk) (n = 30)	VAS	(I vs C, 1 month after treatment) VAS: 3.81 ± 1.18 vs 3.98 ± 1.92. IG vs CG (p < 0.05)

C, Control; CG, control group; EA, electroacupuncture; I, intervention; IG, intervention group; NRS, numeric rating scale; NS, not significant; RCT, randomized controlled trial; SF-36, medical outcomes study 36-item short form health survey; VAS, visual analog scale; WOMAC, Western Ontario and McMaster Universities Arthritis Index.

References

1. Korean Acupuncture & Moxibustion Society. The Acupuncutre and Moxibustion Medicine. Seoul (Korea): Jibmoondang. 2012:559–564, 573–577.

2. Ettinger WH, Afable RF. Physical disability from knee osteoarthritis: The role of exercise as on intervention. Med Sci Sports Exerc. 1994;26:1435–1440.

3. The Korean Orthopedic Association. Orthopedics. 7th ed. Seoul (Korea): The newest medical Publishers CO. 2013:317–327.

4. Rehabilitation medicine of Korean Medicine. Oriental Rehabilitation Medicine. 3rd ed. Seoul (Korea): Koonja publishing. 2011:97.

5. The Korean Orthopedic Association. Orthopedics. 7th ed. Seoul (Korea): The newest medical Publishers CO. 2013:994–1071.

6. Ko HJ, Yoo JH, Shin JC. A Systematic Review and Meta-Analysis of Fire Needling Treatment for Knee Osteoarthritis: Focused on Comparative Studies with Manual Acupuncture Treatment during Recent Five Years. J Acupunct Res. 2019;36:104–114. [in Korean].

7. Gu JH, Kim ES, Park YC, Jung I, Lee EJ. A Systematic Review of Bee Venom Acupuncture for Knee Osteoarthritis. J Korean Med Rehabil. 2017;27:47–60. [in Korean].

8. Korean Acupuncture and Moxibustion Society. The Acupuncutre and Moxibustion Medicine. Seoul (Korea): Jibmoondang. 2012:327–328.

9. Wang W. Dissertation. Clinical study of the ginger moxibustion in treating yang deficiency type knee osteoarthritis. Yunnan (China): Yunnan University of Traditional Chinese Medicine, 2015. [in Chinese].

10. Cheng HL, Han W, Hu PJ, Yang J. A clinical study on the treatment of primary osteoarthritis of the knee with moxibustion. Clin J Tradit Chin Med. 2008;20:114–116. [in Chinese].

11. Zhang J, Hen Y, Huang G. Clinical observation of moxibustion combined with knee joint rehabilitation for the treatment of knee osteoarthritis. J Clin Acupunct Moxibustion. 2015;31:7–9. [in Chinese].

12. Ren XM, Cao JJ, Shen XY, Wang LZ, Zhao L, Wu F, et al. Knee osteoarthritis treated with moxibustion: a randomized controlled trial. Zhongguo Zhen Jiu. 2011;31:1057–1061. [in Chinese].

13. Zhao HY, Kong JJ, Lu W, Zhang Y, Zhang J, Xu J, et al. Clinical effect of acupuncture and moxibustion in treating knee osteoarthritis. J Shanghai Univ Tradit Chin Med. 2013;27:45–47. [in Chinese].

14. Zhang QJ, Cao LH, Li ZD, Wang SC, Ma YH, Su JC, et al. The clinical effects and safety of moxibustion and cerecoxib for osteoarthritis of knee. Chinese J Tradit Chin Orthop. 2011;19:113–115. [in Chinese].

15. Zhou YL, Li J, Hou WG, Bao CL, Zhang Q, Wang SS, et al. Clinical observation of moxibustion in treatment of knee osteoarthritis. Shanghai J Acupunct. 2014;33:1086–1088. [in Chinese].

16. Wu F, Zhang HM, Wang LZ, Zhao L, Deng HP, Cheng K, et al. Effects of CO2 laser moxibustion on quality of life in patients with knee osteoarthritis: A double-blind randomized controlled trial. J Clin Rehabil Tissue Eng Res. 2011;15:4885–4890. [in Chinese].

17. Sun K, Yang J, Shen DK. Clinical observation on treatment of primary knee osteoarthritis of liver and kidney deficiency type with Aconite cake-separated moxibustion. Zhongguo Zhen Jiu. 2008;28:87–90. [in Chinese].

18. Deng JM, Chen Y, Wang SX. Comparative study of the efficacies of warm needling versus salt moxibustion in treating knee osteoarthritis. Shanghai J Acupunct Moxibustion. 2015;34:243–245. [in Chinese].

19. Yuan QD, Guo X, Han YC, Zhang JQ, Feng XD. Observations on the therapeutic effect of heat-sensitive point Thunder-fire moxibustion on knee osteoarthritis. Shanghai J Acupunct Moxibustion. 2015;34:665–668. [in Chinese].

20. Deng JM, Chen Y, Wang SX. Salt insulation moxibustion in the treatment of knee osteoarthritis: Analysis of 35 cases. J Clin Acupunct Moxibustion. 2015;31:14–17. [in Chinese].

21. Yang YH, Sun K, Su GH, Zhou ZL. Clinical Study of Gesan Qibing Moxibustion for Treatment of Primary Osteoarthritis of Knee of Qi Stagnation and Blood Stasis Type. Clin J Tradit Chin Med. 2008;20:53–55. [in Chinese].

22. Song YC, Liu DC, Zhu JC. Clinical Research of Treating Knee Osteoarthritis by Gesan Qibing Moxibustion. J Clin Acupunct. 2013;29:40–42. [in Chinese].

23. Zhou ZL, Sun K, Cheng HL, Liu DC, Yang J. Observations on the Therapeutic Effect of Herbal Cake-separated Moxibustion on Knee Osteoarthritis of Blood Stasis Type. Shanghai J Acupunct. 2010;29:45–47. [in Chinese].

24. Ren X, Yao C, Wu F, Li Z, Xing J, Zhang H. Effectiveness of moxibustion treatment in quality of life in patients with knee osteoarthritis: a randomized, double-blinded, placebo-controlled trial. Evid Based Complement Alternat Med. 2015;2015:569523.

25. Wu F, Xiong P. Moxibustion on heat-sensitive acupoints in the treatment of KOA (swelling type). J Clin Acupunct. 2011;27:1–4. [in Chinese].

26. Chi ZH, Xiong J, Jiao L, Zhang L, Zhang B, Xie DY, et al. Moxibustion location heat-sensitive rule of heat-sensitive moxibustion in the treatment of knee osteoarthritis: A exploratory RCT research. Chin Med Modern Distance Edu China. 2016;14:104–106. [in Chinese].

27. Zhao L, Cheng K, Wang L, Wu F, Deng H, Tan M, et al. Effectiveness of moxibustion treatment as adjunctive therapy in osteoarthritis of the knee: A randomized, double-blinded, placebo-controlled clinical trial. Arthritis Res Ther. 2014;16:R133, [in Chinese].

28. Bu AX, Wang LG, Wiong J, He D, Hu XC, Guo RC. Observation of curative effect of Heat-sensitive acupoint moxibustion on knee joint osteoarthritis of Han Shi Yu Bi Type. Shizhen Tradit Chin Med. 2018;29:1114–1116. [in Chinese].

29. Zhou JY. Dissertation. Moxibustion for treating knee osteoarthritis: A double-blinded, double-dummy, randomized controlled trial. Chengdu (China): Chengdu University, 2017. [in Chinese].

30. Zhang S. Dissertation. The Clinical observation the therapy of knee Joint osteoarthritis using Moderate Moxibustion. Guangzhou (China): Guangzhou University, 2009. [in Chinese].

31. Zheng J, Zhang H, Liu J. Different acupuncture and moxibustion methods at Heding (EX-LE 2) for knee osteoarthritis with yang-deficiency and cold-stagnation syndrome. Zhongguo Zhen Jiu. 2017;37:594–598. [in Chinese].

32. Zhang QR, Fu WB. Osteoarthritis of knee joint treated with acupuncture and moxibustion. Zhongguo Zhen Jiu. 2010;30:375–378. [in Chinese].

33. He Y, Xu YP, Huo XT. Clinical research of moxibustion in treating knee joint osteoarthritis. Chin J Tradit Chin Orthop. 2009;17:38–39. [in Chinese].

34. Tian DD, Wan L, Huang CB, Wang SP, Xi X, Liu J. Effect of moxibustion on clinical efficacy and quality of life in patients with knee osteoarthritis. Henan Tradit Chin Med. 2017;37:2208–2210. [in Chinese].

35. Xiao YD. Effect of moxibustion assisted intravitreal injection of sodium hyaluronate on joint function in patients with knee osteoarthritis. Med Equipment. 2017;30:117–118. [in Chinese].