Effectiveness of moxibustion alone on lumbar disc herniation: a Meta-analysis of randomized controlled trials

Abstract

OBJECTIVE:

To evaluate the available evidence from randomized controlled trials (RCTs) of moxibustion alone for lumbar disc herniation (LDH) treatment.

METHODS:

A systematic search of 10 databases (until August 30, 2021) was used to identify studies that reported the response rate, visual analogue scale (VAS) score, Japanese Orthopedic Association (JOA) score, and Oswestry Disability Index (ODI) score. Study selection and data extraction were independently performed by two reviewers. Cochrane criteria for risk of bias were used to assess the methodological quality of the trials. The Grading of Recommendations Assessment, Development, and Evaluation Methodology (GRADE) were also used to test the quality of the result evidence.

RESULTS:

Nineteen RCTs, including 1888 patients, met the inclusion criteria. Five studies showed no difference between moxibustion and acupuncture on response rate [risk ratio (RR) = 1.07, 95%CI (0.98, 1.16), P = 0.11]. Meanwhile, six studies suggested that there is no significant difference between moxibustion and acupuncture on VAS score [mean difference (MD) = －0.43, 95% CI (－0.91, 0.05), P = 0.08]. Eight studies implied that there is no significant difference between moxibustion and acupuncture on JOA score [MD = 0.84, 95% CI (－1.27, 2.96), P = 0.44]. Two studies indicated that moxibustion may have equivalent effects for treating LDH in the VAS score in comparison with drug therapy [MD = －1.16, 95%CI (－2.63, 0.31), P = 0.12]. The evidence level of results was determined to be very low to low.

CONCLUSIONS:

Based on the existing evidence, moxibustion may not be suitable for treating LDH alone, but it may be applied as an adjuvant treatment. Furthermore, well-designed RCTs with high quality and larger samples are still needed to evaluate the efficacy and safety of moxibustion alone for LDH treatment.

1. INTRODUCTION

Lumbar disc herniation (LDH), which is characterized by low back pain, numbness, and pain of lower limbs, is a common disease in orthopedics. At present, the incidence of LDH is 20%-30%.¹ Considering changes in lifestyle, the prevalence of LDH has increased dramatically to the extent in which young people are experiencing it as well.² LDH has also affected the quality of life of patients and increased the economic burden of the social health care system.³ At present, the majority of patients with LDH undergo non-surgical treatment,⁴ and 80%-90% receive mild rehabilitation after conservative treatment (non-surgical treatment).⁵

Notably, non-surgical treatment in Western Medicine primarily uses oral non-steroidal anti-inflammatory drugs and epidural steroid injection, which may be only marginally effective for short-term relief of pain with potentially long-term side effects, such as hepatic and renal toxicity and gastrointestinal discomfort.⁶ In addition, the cost-effectiveness of this type of treatment is unclear.⁷ Meanwhile, Traditional Chinese Medicine (TCM) has attracted increasing attention because of its superior effect and minimal side effects.⁸ As a TCM external treatment, moxibustion is considered as a major TCM therapy together with decoction and acupuncture, and it is widely used because of its convenience, efficacy, and cheapness. Moxibustion has become an important method for the treatment of LDH, and its curative effect is remarkable.⁹

In particularly, it is important to systematically assess and meta-analyze the effect of moxibustion for LDH. In evaluating the efficacy and safety of moxibustion for LDH treatment, a systematic review¹⁰ has included trials using moxibustion combined with other western or TCM therapies as the intervention method in the experimental group. However, it fails to assess the efficacy of moxibustion alone for treating LDH. With the widespread use of moxibustion, its effect has also received attention. However, a systematic review or meta-analysis that assesses the efficacy of moxibustion alone on LDH is lacking. Therefore, this study aims to critically assess the efficacy and safety of moxibustion without combining with other therapies in treating LDH.

2. METHODS

This Meta-analysis was performed on the basis of the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines.¹¹

2.1. Data sources

A systematic literature search was used in the following databases from their inception to August 30, 2021, with no restrictions on language and publication: the Cochrane Library, Web of Science, PubMed, EMBASE, EBSCO, the China National Knowledge Infrastructure (CNKI), Wanfang data, Oriental Medicine Advanced Searching Integrated System (OASIS), National Discovery for Science Leaders (NDSL), and Japan Science and Technology Information Aggregator, Electronic (J-STAGE). These searches were performed using the following keywords: “lumber disc herniation,” “lumbar disc herniation,” “lumbar disc protrusion,” “LDH,” “intervertebral disc displacement,” “moxibustion,” “moxibustion therapy,” “moxibustion,” “moxa,” “random,” “randomly,” “randomized controlled trial,” and “placebo.” The search strategies for the abovementioned 10 databases are shown in Appendix S1. For a comprehensive search, the references of included articles and relative reviews were manually searched. Two investigators independently searched and imported the identified literature into the EndNote software to delete duplications, and they selected potential articles by reviewing the titles and abstracts. The full texts of the selected articles were assessed on the basis of the inclusion and exclusion criteria.

2.2. Inclusion criteria

The included studies were published randomized controlled trials with parallel design, including those in the form of dissertations and conference papers.

P (population): Patients with LDH regardless of gender, age, ethnicity, education, and economic status who meet the diagnostic criteria were included in the study.^{12 -14}

I (intervention): trials that compared moxibustion with any other treatment were included. Studies of moxibustion combined with other therapies were excluded.

C (comparison): trials that compared moxibustion with any intervention method in control group were included.

O (outcomes): the outcome included the total response rate, which was based on diagnostic and therapeutic criteria for diseases and syndrome patterns of TCM or the guiding principle of clinical research on new drugs of TCM.^12,13 In addition, the visual analog scale (VAS) score served as an outcome measurement in this review,¹⁵ which was used for a total of 0-10 points, where 0-10 indicates different degrees of pain, 0 indicates no pain, and 10 indicates maximum pain. Meanwhile, the Japanese Orthopedic Association (JOA) back pain evaluation questionnaire was included.¹⁶ It is used to assess pain, the ability to conduct daily life and work, functional impairment, and particular clinical examinations. In addition, the other outcome measures included the Oswestry Disability Index (ODI),¹⁷ which is a questionnaire survey and assessment method for each of the nine aspects, including lifting, walking, sitting, standing, pain intensity, self-care, sleep disturbance, social life, and travel.

2.3. Data extraction

Two reviewers (CHEN Fengqin and YAO Yao) independently performed data extraction based on predefined criteria, including the authors, publication date, sample size, population information, intervention regimens, follow-up duration, outcomes, and adverse events. In the case of insufficient data, we would send an e-mail to the author requesting for more information.

2.4. Risk of bias assessment

The risk of bias was determined by two independent reviewers (ZHAO Zhenni and YAO Yao) in accordance with the Cochrane Handbook for Systematic Reviews of Interventions,^18,19 and it included the following items: (a) random sequence generation (selection bias), (b) allocation concealment (selection bias), (c) blinding of participants and personnel (performance bias), (d) blinding of outcome assessment (detection bias), (e) incomplete outcome data (attrition bias), (f) selective reporting (reporting bias), (g) and other bias. The evaluation of these items was rated as “low,” “high,” or “unclear.” Divergence was resolved through discussion. If two reviewers were unable to reach an agreement, then the third and fourth reviewers (SUN Zhiling and XU Xiao) were consulted for a final decision.

2.5. Statistical analysis

Meta-analysis was performed using RevMan 5.3. Dichotomous outcomes were reported as risk ratio (RR) and continuous data as weighted mean difference (WMD). Notably, χ² test and I² test were utilized to mea-sure the heterogeneity among studies. A fixed-effect model was adopted if I² < 50% and P > 0.1; otherwise, a random-effects model was used. When the clinical heterogeneity between the two studies was substantial, only descriptive analysis was performed. The potential publication bias was tested by using an inverted funnel chart when the number of RCTs in meta-analysis was more than 10. In addition, the causes of heterogeneity were identified, and the leave-one-out method was used for sensitivity analysis.¹⁹ If high heterogeneity still could not be explained, then the results should be carefully interpreted. Sensitivity analysis was performed using STATA 16.0 software (Stata Corp. 2019, College Station, TX, USA).

The quality of the evidence for each outcome was ass-essed on the basis of the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) software. Detailed GRADE guidance was used to evaluate the risk of bias, imprecision, inconsistency, indirectness, and publication bias, thereby downgrading the level of evidence.^20,21 The level of evidence can be upgraded in case of a large effect size, evidence of dose-response gradient, or all plausible confounding factors reducing an apparent effect. Two reviewers (LENG and PANG) separately conducted assessment. Any dis-agreement was resolved through discussion and con-sultation with a third author (SUN) until a consensus was reached. We described the quality of evidence as follows:

High-quality evidence: consistent findings were obtained among at least 75% of trials with no limitations of study design, consistency, direct and precise data, and known or suspected publication biases. We are confident that the real result is close to the estimation of the effect.

Moderate-quality evidence: one of the domains is not met. We have some confidence in the estimation of the effect. The true effect is likely close to the estimation of the effect, but it may also be substantially different.

Low-quality evidence: two of the domains are not met. Our confidence in the effect estimate is limited. The true effect may be substantially different from the estimate of the effect.

Very-low-quality evidence: three of the domains are not met. We have very little confidence in the effect estimate. The true effect is likely to be substantially different from the estimate of effect.

No evidence: no trials were identified that addressed this outcome.

3. RESULTS

3.1. Study selection

A total of 1422 potential studies were identified through an initial database search. After removing 645 duplicates and 680 articles by screening title and abstract, 97 full-text articles were downloaded and assessed. Of these articles, 78 were excluded for failing to satisfy the inclusion criteria. Finally, 19 RCTs with 1888 participants were included for systematic review.^{22 -40} A flow chart of the study selection process is presented in Figure 1.

Figure 1. Flowchart of the study selection process.

TCM: Traditional Chinese Medicine; RCT: randomized controlled trial.

3.2. Study characteristics

All included trials were conducted in China. The sample sizes ranged from 40 to 456. Four RCTs adopted a three-arm parallel group design,^25,33,37,38 and 15 trials used a two-arm parallel group design.

Intervention is different from group. The experimental groups were treated with moxibustion alone. In the control groups, interventions included drug th-erapy,^{25,26,37,38,40} conventional therapy,^22,23 acu-puncture,^{25,27,⇓,⇓,⇓,⇓,⇓,⇓,⇓,⇓-36} acupuncture plus TDP irradiation,³⁹ and fixed-point rotation reduction.²⁴ The characteristics of the participants, intervention details, adverse events, and outcome measures are presented in Table 1.

Table 1.

Summary of the 19 included studies

Study	Sample size	Mean age(years)		Duration(months)		Intervention					Follow up			Adverse events			Outcomes
						Experimental group			Control group
Chen RX et al 201438	456	E1: 45.5±10.6E2: 47.3±11.2C: 46.6±10.5		n.r.		(E1) Suspended moxa (2 times daily in the first 4 d, one time daily in remaining 10 d, 2 weeks, n = 152)(E2) Suspended moxa (1session = 45 min, 2 times daily in the first 4 d, 1 time daily in the remaining 10 d, 2 weeks, n = 152)			Drug therapy (20% mannitol, Voltaren tablets in the first 3 d. Voltaren tablets in subsequent 11 d.) Plus acupuncture (1session = 30 min, 1/d, 2 weeks, n = 152)		2 weeks, 6 months			None related to moxa.			JOA
Fu Y et al 201437	180	E1: 54.55±6.24E2: 57.22±6.37C: 56.21±5.27		E1: 5.45±3.46E2: 5.00±2.12C: 5.32±3.18		(E1) Suspended moxa (2 times daily in the first four days, 1 time daily in the remaining 10 d, 2 weeks, n = 60)(E2) Suspended moxa (1 session = 30 min, 2 times daily in the first 4 d, 1 time daily in the remaining 10 d, 2 weeks, n = 60)			Drug therapy(20% mannitol, 1/d, fortalin tablets, in the first 3 d, only taken fortalin tablets in subsequent 11 d.) Plus acupuncture, (1 session = 30 min, once daily, 2 weeks, n = 60)		2 weeks, 6 months			n.r.			JOA
Hu XW et al 201925	90	E: 40.19±5.16C1: 38.32±6.71C2: 40.32±7.92		E: 10.58±2.33C1: 8.90±1.80C2: 9.68±1.74		Indirect moxa (once a week, 1 month, n = 31)			(C1) Acupuncture (1 session = 30 min, 1/d, 10 times as a course, rest for 2 d between courses, 1 month, n = 28)(C2) Drug therapy(Diclofenac sodium tablets, 25 mg, 3 times a day, 1 month, n = 31)		1 month			n.r.			Response rate; VAS; ODI;
Li H et al 201231	60	E: 41.7C: 41.9		E: 7.4C: 7.2		Suspended moxa (2 times daily in the first 4 d, 1 time daily in the remaining 10 d, 2 weeks, n = 30)			Acupuncture (1 session = 30 min, 1/d, 6 times/ week, 8 weeks, n = 30)		2 weeks			n.r.			Response rate
Li L 201634	60	E: 39.62±5.26C: 40.80±4.68		E (yr): 3.60±1.09C(yr): 3.57±1.12		Indirect moxa (1 session = 1.5 h, 2 times/week, 8 weeks, n = 30)			Acupuncture (1 session = 30 min, 1/d, 6 times/ week, 8 weeks, n = 30)		8 weeks			n.r.			VAS; JOA
Lin WP 201236	60	E: 42.73±11.77C: 41.73±10.65		E(d): 18.77±19.58C(d): 20.53±23.56		Indirect moxa (1 session = 50 min, once daily, 10 d, n = 30)			Acupuncture (1session = 30 min, plus cup, 1 session = 30 min, 1/d, 10 d, n = 30)		10 d			n.r.			Response rate
Liu Q et al 201239	80	E: 43.20±10.26C: 42.77±10.50		E (yr): 4.81±2.96C (yr): 4.67±3.16		Indirect moxa (once every other one day, 3 times/week, 3 weeks, n = 40)			Acupuncture plus TDP irradiation (1 session = 30 min, 1/d, 5 times/ week, 3 weeks, n = 40)		3 weeks			n.r.			Response rate
Liu CY et al 202140	60	E: 52.42±3.17C: 52.38±3.25		n.r.		Suspended moxa (1 session = 15 min, 1/d, 6 d/course, rest for 1 d between courses, 4 weeks, n = 30)			Drug therapy (Citicoline sodium injection 0.5 g, salvia miltiorrhiza injection 20 mL, respectively, added to 5% of the glucose solution 250 mL intravenous injection，1/d, 4 weeks, n = 30)		4 weeks			n.r.			VAS; JOA
Lu YJ 201535	60	E: 52.20±7.14C: 52.53±7.53		E(yr): 5.63±2.94C(yr): 5.57±2.87		Indirect moxa (1 session = 1.5 h, 2 times/week, 8 weeks, n = 30)			Acupuncture (1 session = 25 min, 1/d, 6 times/week, 8 weeks, n = 30)		8 weeks			n.r.			VAS; JOA
Ma S et al 201033	120	E1: 41.5±12.8E2: 38.3±11.3C: 37.9±11.2		E1(d): 261.2±47.7E2(d): 267.3±34.5C(d): 274.8±50.2		(E1) Indirect moxa (once daily, 3 weeks, n = 30)(E2) Indirect moxa plus Acupuncture, (3 weeks, n = 60)			Acupuncture (1session = 30 min, 1/d, 3 weeks, n = 30)		3 weeks			n.r.
Ma S et al 201033	120	E1: 41.5±12.8E2: 38.3±11.3C: 37.9±11.2	E1(d): 261.2±47.7E2(d): 267.3±34.5C(d): 274.8±50.2		(E1) Indirect moxa (once daily, 3 weeks, n = 30)(E2) Indirect moxa plus Acupuncture, (3 weeks, n = 60)		Acupuncture (1session = 30 min, 1/d, 3 weeks, n = 30)			3 weeks		n.r.			JOA
Mai ZM, Tang JD201729	94	E: 42.81±9.31C: 43.17±9.63		E: 13.45±6.41C: 13.63±6.55		Suspended moxa (1 session = 40 min, 1/d, 5 d/course, rest for 2 d between courses, 4 weeks, n = 47)		Acupuncture(1 session = 30 min, 1/d, 5 d/course, rest for 2 d between courses, 4 weeks, n = 47)		4 weeks			n.r.			VAS; JOA
Mao LF 201726	60	E: 51.53±8.66C: 51.93±11.30		n.r.		Suspended moxa (5 d/course, rest for 2 d between courses of treatment, n = 30)		Drug therapy (Yaotongning capsule, 4-6 pieces, 3/d, 10 d, n = 30)		10 d			5 patients developed dry and thirst, normal after hydration			JOA
Song LJ et al 201622	60	E: 27.76±3.67C: 28.83±3.33		E: 15 d-3 yrC: 20 d-2.5 yr		Suspended moxa (1session = 15-20 min, 1/d,5 d/course, 10 d, n = 30)		Conventional therapy (10 d, n = 30)		10 d			n.r.			ODI
Wang XY 200828	72	E: 48.8±11.8C: 46.8±13.7		n.r.		Indirect moxa (once every other one day, 3 weeks, n = 36)		Acupuncture (1 session = 30 min, once every other one day, 3 weeks, n = 36)		1/2/3 week(s), 1 month			None related to moxa.			Response rate; JOA
Xu JF et al 201227	60	E: 41.1±11.6C: 40.1±10.1		E: 29.2±8.4C: 32.3±8.2		Suspended moxa (1 session = 1 h, 1/d, 3 weeks, n = 30)		Acupuncture (1 session = 30 min, 1/d,3 weeks, n = 30)		3 weeks			n.r.			VAS; JOA
Yang DL et al 201432	40	E: 42±13C: 45±12		E: 29.2±8.4C: 32.3±8.2		Suspended moxa (1 session = 1 h, 1/d, 3 weeks, n = 20)		Acupuncture (1 session = 30 min, 1/d, 3 weeks, n = 20)		3 weeks			n.r.			JOA
Yi GQ et al 201530	70	E: 47.09±9.06C: 45.49±10.48		E: 14.40±16.83C: 11.49±17.70		Suspended moxa (1 session = 30 min, 1/d, 5 d/course, rest for 2 d between courses, 3 weeks, n = 35)		Acupuncture (1 session = 30 min, 1/d,5 d/course, rest for 2 d between courses, 3 weeks, n = 35)		3 weeks			n.r.			Response rate; VAS;JOA
Zhang LZ 201823	106	E: 47.58±10.30C: 47.15±10.07		E (yr): 6.08±3.64C(yr): 6.34±3.51		Suspended moxa (1 session = 20-30 min, 1/d, n = 53), plus Conventional therapy		Conventional therapy (n = 53)		14-21 d			n.r.			VAS
Zheng CH et al 201924	100	E: 54.26±8.24C: 54.49±8.19		E: 3.79±2.88C: 3.86±2.91		Suspended moxa (5 times every week, 2 weeks, n = 50)		Fixed-point rotation reduction (once every other one day, 2 weeks, n = 50)		2 weeks			n.r.			Response rate; VAS; JOA; ODI

Notes: E: experimental group; C: control group; d: day; yr: year; n.r.: not reported; moxa: moxibustion; VAS: Visual Analogue Scale; JOA: Japanese Orthopaedic Association; ODI: Oswestry Disability Index.

3.3. Risk of bias

Based on the Cochrane criteria, the risk of bias assessment is presented in Figure 2. Among 19 studies, 15 trials utilized a random table for randomization,^{24 -28,30,32-40} and the other four trials did not provide the methods of sequence generation.^22,23,29,31 Only one reported trial used a central randomization system to conduct allocation concealment,³⁸ whereas the remainder did not provide concealment methods. One trial mentioned single blindness, but it did not specify who was blinded,³¹ and one trial used the blindness of the assessor.³⁸ Two studies mentioned dropouts without detailed information of handling.^31,36

Figure 2. Risk of bias graph and summary.

A: risk of bias analysis of each included studies; B: overall risk of bias analysis of included studies. “?”: unclear risk of bias; “+”: low risk of bias; “-”: high risk of bias.

3.4. Subgroup analysis

Of the 19 studies, seven trials used response rate; nine studies used the VAS score; 13 trials used the JOA score, and three trials used ODI to evaluate the effects of moxibustion for LDH.

3.5. Response rate

Seven RCTs reported response rates for moxibustion as the outcome measurement.^{24,25,28,30,31,36,39} Given the clinical heterogeneity, subgroup analysis was performed on the basis of different control interventions. In addition, five trials compared moxibustion with acupuncture and demonstrated no significant difference between these two interventions [RR = 1.07, 95%CI (0.98, 1.16), P = 0.11, I² = 0%].^{25,28,30,31,36} A single trial implied that no statistical significance of moxibustion was observed in improving the response rate compared with drug therapy [RR = 1.26, 95% CI (1.00, 1.58), P = 0.05].²⁵ One trial suggested that moxibustion had no statistically significant effects on the response rate compared with acupuncture plus TDP irradiation [RR = 1.08, 95% CI (0.97, 1.21), P = 0.17].³⁹ Another single trial indicated that moxibustion had no statistical significance on the response rate compared with fixed-point rotation reduction [RR = 1.22, 95% CI (1.00, 1.49), P = 0.05].²⁴ By contrast, the overall results suggested that moxibustion had a marked beneficial effect in improving the response rate when compared with four different treatments [RR = 1.11, 95% CI (1.04, 1.19), P = 0.001, I² = 0%] without high heterogeneity (Figure 3A).

Figure 3. Results of the Meta-analysis.

A: forest plot of moxibustion on response rate (1); B: forest plot of moxibustion on response rate (2); C: forest plot of moxibustion on VAS score; D: forest plot of moxibustion on JOA score; E: forest plot of moxibustion on ODI score. VAS: Visual Analogue Scale; JOA: Japanese Orthopaedic Association; ODI: Oswestry Disability Index.

Furthermore, subgroup analysis was conducted on the basis of the definition of the response rate. Here, six RCTs utilized the diagnostic and therapeutic criteria for diseases and syndrome patterns of TCM to evaluate the response rate.^{24,28,30,31,36,39} In comparison to the control interventions, subgroup analysis results revealed that moxibustion had significantly favorable effects in improving the response rate [RR = 1.10, 95% CI (1.02, 1.18), P = 0.01, I² = 0%]. Meanwhile, one trial with two groups utilized the guiding principle of clinical research on new drugs of TCM to assess the response rate.²⁵ Compared with the control interventions, moxibustion had a significantly better overall effective rate [RR = 1.17, 95% CI (1.02, 1.35), P = 0.03, I² = 2%]. In general, meta-analysis indicated no significant differences between these two criteria in evaluating the response rate (Figure 3B).

3.6. VAS score

Nine RCTs utilized the VAS score as a measure of the effects for LDH treatment.^{23 -25,27,29,30,34,35,40} Meanwhile, six studies reported equivalent effects in reducing pain between moxibustion and acupuncture [MD = －0.43, 95% CI (－0.91, 0.05), P = 0.08, I² = 91%].^{25,27,29,30,34,35} Two trials exhibited equivalent effects of moxibustion compared with drug therapy for the VAS score [MD = －1.16, 95% CI (－2.63, 0.31), P = 0.12, I² = 96%].^25,40 Notably, one RCT showed superior effects of moxi-bustion on pain reduction compared with conventional therapy [MD = －0.64, 95% CI (－0.87, －0.41), P < 0.000 01].²³ In addition, a single trial indicated that moxibustion had a significant pain relief effect on LDH compared with fixed-point rotation reduction [MD = －1.05, 95% CI (－1.22, －0.88), P < 0.000 01].²⁴ The overall results indicated that moxibustion can sig-nificantly relieve pain when compared with four different treatments [MD = －0.70, 95% CI (－0.99, －0.41), P < 0.000 01, I² = 90%] with severe heterogeneity (Figure 3C).

3.7. JOA score

Thirteen RCTs adopted the JOA score as a measure to assess the effects of a particular therapy on improving LDH,^{24,26 -30,32-35,37,38,40} and the heterogeneity among these studies was also high. Subgroup analysis sug-gested that moxibustion had no significantly favorable effects with regard to the JOA score compared with acupuncture [MD = 0.84, 95% CI (－1.27, 2.96), P = 0.44, I² = 90%]. In comparison with drug therapy, moxibustion had no statistical significance in improving the JOA score [MD = 4.15, 95% CI (0.05, 8.26), P = 0.05, I² = 96%]. Moreover, a single trial indicated that moxi-bustion had statistically significant favorable effects on the JOA score in comparison with fixed-point rotation reduction [MD = 8.69, 95% CI (5.08, 12.30), P < 0.000 01].²⁴ The pooled results indicated that moxibustion had no statistically significant favorable effects on the JOA score when compared with the three different treatments [MD = 2.08, 95% CI (－0.08, 4.24), P = 0.06] (Figure 3D).

In particular, two trials were not included in Meta-analysis, whose outcomes were mea-sured using a modified edition of the JOA score.^37,38 One trial demonstrated that mox-ibustion had signi-ficantly favorable effe-cts in improving the JOA score compared with drug therapy plus acupuncture.³⁸ Another study also illustrated that moxibustion imm-ensely improved the status of patients suff-ering from LDH com-pared with drug therapy plus acupuncture.³⁷

3.8. ODI

Three RCTs adopted ODI as an out-come.^22,24,25 In com-parison with acup-uncture, the subgroup analysis results reve-aled equivalent effects of moxibustion in improving the ODI score [MD = －1.16, 95% CI (－3.75, 1.43), P = 0.38]. In addition, moxibustion had better efficacy in relieving dysfunction caused by LDH compared with drug therapy, conven-tional therapy, and fixed-point rotation reduction. In general, the Meta-analysis re-sults significantly dem-onstrated the superior effect of moxi-bustion on the ODI score compared with control inter-ventions with high heterogeneity [MD = －8.78, 95% CI (－12.73, －4.83), P < 0.0001, I² = 94%] (Figure 3E).

3.9. Adverse effects

Of the 19 included studies, three studies reported adverse events, whereas 16 did not mention any adverse events. Two trials stated that moxibustion had no adverse events.^28,38 One trial reported five patients who developed dry and thirst in the moxibustion group but returned to normal after hydration.²⁶

3.10. Sensitivity analysis

Sensitivity analysis was conducted to evaluate the robustness of Meta-analysis by using STATA 16.0 (Stata Corp. 2019, College Station, TX, USA). Given the heterogeneity of intervention comparison, we conducted sensitivity analysis on the results of each subgroup but not on the subgroups with only 1 to 2 studies. Sensitivity analysis was presented in the form of forest maps (Figure 4). All studies were within reasonable intervals, and sensitivity was acceptable.

Figure 4. Sensitivity analysis.

A: sensitivity analysis of the response rate; B: sensitivity analysis of the VAS score; C: sensitivity analysis of the JOA score. VAS: Visual Analogue Scale; JOA: Japanese Orthopaedic Association.

3.11. Publication bias

Based on the response rate, we used a funnel plot to have qualitative analysis of publication bias. Based on the visual assessment, no significant evidence of publication bias was identified. This result may be related to the fact that fewer than 10 studies were included. The publication bias of studies is depicted in Appendix S2.

3.12. Quality of evidence

The evidence level of the results from the 19 studies that measured moxibustion was determined to be very low to low (Table 2). The level of evidence was downgraded because of the study design and risk of bias (the gen-eration of random sequence, blindness, and allocation concealment are not clear), inconsistency of results (unreasonable heterogeneity), and the imprecision of the included studies (the number of the participants was too small).

Table 2.

Summary of findings and quality of evidence for all outcomes included in this review

Item		Effect size (95% CI)	No. of participants (studies)	Quality of the evidence(GRADE)
Response rate	Moxibustion vs Acupuncture	RR 1.07 higher(0.98 higher to 1.16 higher)	321(5 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Drug therapy	RR 1.26 higher(1.00 higher to 1.58 higher)	62(1 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Acupuncture plus TDP irradiation	RR 1.08 higher(0.97 higher to 1.21 higher)	80(1 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Fixed-point rotation reduction	RR 1.22 higher(1.00 higher to 1.49 higher)	100(1 RCTs)	⊕○○○VERY LOW
	The diagnostic and therapeutic criteria for diseases and syndrome patterns of TCM	RR 1.10 higher(1.02 higher to 1.18 higher)	442(6 RCTs)	⊕⊕○○LOW
	The guiding principle of clinical research on new drugs of TCM	RR 1.17 higher(1.02 higher to 1.35 higher)	121(1 RCTs)	⊕○○○VERY LOW
VAS score	Moxibustion vs Acupuncture	MD 0.43 lower(0.91 lower to 0.05 higher)	403(6 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Drug therapy	MD 1.16 lower(2.63 lower to 0.31 higher)	122(2 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Conventional therapy	MD 0.64 lower(0.87 lower to 0.41 lower)	106(1 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Fixed-point rotation reduction	MD 1.05 lower(1.22 lower to 0.88 lower)	100(1 RCTs)	⊕○○○VERY LOW
JOA score	Moxibustion vs Acupuncture	MD 0.84 higher(1.27 lower to 2.96 higher)	516(8 RCTs)	⊕⊕○○LOW
	Moxibustion vs Drug therapy	MD 4.15 higher (0.05 higher to 8.26 higher)	120(2 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Fixed-point rotation reduction	MD 8.69 higher (5.08 higher to 12.30 higher)	100(1 RCTs)	⊕○○○VERY LOW
ODI score	Moxibustion vs Acupuncture	MD 1.16 lower(3.75 lower to 1.43 higher)	59(1 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Drug therapy	MD 16.03 lower(20.04 lower to 12.02 lower)	62(1 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Conventional therapy	MD 10.67 lower(12.72 lower to 8.62 lower)	60(1 RCTs)	⊕○○○VERY LOW
	Moxibustion vs Fixed-point rotation reduction	MD 8.19 lower(8.78 lower to 7.60 lower)	100(1 RCTs)	⊕○○○VERY LOW

Notes: CL: confidence interval; GRADE: Grading of Recommendations, Assessment, Development and Evaluations; RR: risk rate; RCT: randomized controlled trial; MD: mean difference; VAS: visual analogue scale; JOA: japanese orthopaedic association; ODI: oswestry disability index.

4. DISCUSSION

4.1. Principal findings

This systematic review and Meta-analysis evaluated the response rate, pain management, and functional improvement of moxibustion in the treatment of LDH. We evaluated 19 RCTs involving 1888 patients. All studies utilized moxibustion alone in experimental groups. Meanwhile, the control group intervention included acupuncture, drug therapy, conventional therapy, acupuncture plus drug therapy, acupuncture plus TDP irradiation, and fixed-point rotation reduction. In addition, four trials did not disclose the details of random sequence generation, and another fifteen RCTs used random tables for randomization. Notably, only one trial was reported to have used allocation concealment and blinding of outcome assessment. Two studies mentioned dropouts without reporting detail information.

Based on the abovementioned observations, the findings of our review indicated that moxibustion may have equivalent effects for treating LDH on all of the four outcome measurements when compared with acupuncture. When moxibustion was compared with drug therapy, the ODI score indicated that moxibustion had more favorable effects, but the VAS score suggested no difference between these two therapies. Moreover, when moxibustion was compared with acupuncture combined with TDP irradiation, no significant difference was identified in the response rate. However, compared with conventional therapy, moxibustion had significantly favorable effects with regard to its VAS score and ODI score. When compared with fixed-point rotation reduction, moxibustion had considerably favorable effects in improving the VAS score, JOA score, and ODI score. Notably, only one study was found in some subgroup analyses, which may lead to inaccurate results.

Sensitivity analyses showed that our results were robust. The funnel plot suggested no significant publication bias. Therefore, these findings require an expanded sample. In addition, the evidence level of these results was determined to be very low to low, which indicated that these positive results should be interpreted with caution. Moreover, the safety of moxibustion for LDH treatment remains unclear because of the lack of studies that provide sufficient details of the adverse events.

Meta-analysis evaluating the efficacy of moxibustion on LDH treatment concluded that moxibustion is more effective than acupuncture,⁴¹ which is different from our results. We found that they included trials that used acupuncture and moxibustion as the experimental group. Moreover, they did not adopt subgroup analysis based on different interventions in control groups, which may increase the unreliability of the results. Furthermore, our review used four outcome measurements to evaluate the effect of moxibustion for LDH, which was not available in previous systematic reviews.

Subgroup analysis was conducted to settle the heterogeneity among studies. Nevertheless, the results of subgroup analysis failed to explain the heterogeneity, which may be due to the experimental design. During literature screening, moxibustion therapies varied in implementation forms (such as moxibustion methods, treatment duration, selection of moxibustion points, and sessions) and lack of standards, which may limit the application and spread of moxibustion in clinical practice. Previous studies that focused on moxibustion for LDH, low back pian, and rheumatoid arthritis treatment also had high heterogeneity.^10,42,43 Based on these findings, further studies must focus on factors affecting the therapeutic effect of moxibustion.

In our review, the results demonstrated that all the outcomes had no statistical difference between moxibustion and acupuncture. Therefore, the therapeutic effect of moxibustion on LDH may be comparable to acupuncture. As an important means of TCM, acupuncture has been widely recognized. Recently, a meta-analysis has identified the superior effect of acupuncture on LDH,² which indicated that acupuncture had a more favorable effect in treating LDH than lumbar traction, ibuprofen, diclofenac sodium, meloxicam, dexamethasone plus mannitol and mecobalamin, mannitol plus dexamethasone, loxoprofen, etc. Therefore, as a non-invasive treatment, moxibustion may be more acceptable to patients than acupuncture.

With the increasingly more extensive application of moxibustion, its mechanism has gained increasing attention. Notably, modern medicine believes that moxibustion can prevent and treat diseases with the function of anti-inflammation and immune imp-rovement.⁴⁴ Based on a study of a LDH rat model,⁴⁵ moxibustion can effectively decrease the neurological function score of model rats and improve lower limb function while reducing the level of serum inflammatory cytokine IL-1β. It can also inhibit the infiltration of inflammatory cells in nerve root tissues and surrounding tissues and reduce the damage of chemical inflammation to nerve roots. One hypothesis emphasizes the fact that moxibustion is a result of warm stimulation, which stimulates and regulates the function of the nervous system by stimulating skin receptors.⁴⁶ In particular, heat produced by moxibustion activates local specific receptors, heat-sensitive immune cells, and heat shock proteins to initiate the warming and dredging function and produce various total effects. The warming stimulation signals and subsequent effects through nerve and body fluid pathways induce the effect of further specific target organs and body systems.⁴⁷ However, no systematic moxibustion treatment mechanism is identified at present, and more research on moxibustion treatment mechanisms is needed in the future.

4.2. Strengths and limitations

Our study was the first Meta-analysis that evaluated the efficacy of moxibustion alone for LDH treatment. We reviewed related literature in English, Chinese, Japanese, and Korean databases. In addition, four outcome measurements were used to evaluate the efficacy of moxibustion alone for LDH treatment, which was not available in previous systematic reviews. The GRADE profile was also adopted to test the quality of result evidence, which made the results more reliable. However, this study also had several important limitations that need to be discussed. First, the majority of the included studies were of low quality. Notably, most of the included studies failed to provide the details of sequence generation, allocation concealment, blinding methods, adverse events, and withdrawals or dropouts. In addition, considering that few trials required a long-term follow-up, analyzing the long-term effects of this treatment was impossible. Therefore, this study only presented the short-term curative effect of moxibustion on LDH treatment.

Based on the limitations, this study indicated that well-designed, multi-center randomized controlled trials were warranted to establish the efficacy of moxibustion on LDH treatment. Moreover, further research should be conducted to determine the long-term effects of moxibustion.

In conclusion, based on the existing evidence, moxibustion may not be suitable for treating LDH alone, but it may be applied as an adjuvant treatment. Furthermore, well-designed RCTs with high-quality and large samples are necessary to evaluate the efficacy and safety of moxibustion alone for LDH treatment.

5. ACKNOWLEDGEMENTS

Thank you to all the participants and clinical researchers involved in the publications cited in this review. Thank you to all the peer reviewers who contributed to the continuous improvement of this article. The authors are also immensely grateful for the financial support provided by the National Natural Science Foundation of China and the Nanjing University of Chinese Medicine.

S1 Appendix search strategies.

PubMed

#1. ((((lumber disc herniation[Title/Abstract]) OR (lumbar disc herniation[Title/Abstract])) OR (lumbar disc protrusion[Title/Abstract])) OR (LDH[Title/Abstract])) OR (intervertebral disk displacement[Title/Abstract])

#2. Moxibustion[Mesh]

#3. (((moxibustion[Title/Abstract]) OR (moxabustion[Title/Abstract])) OR (moxibustion therapy[Title/Abstract])) OR (moxa[Title/Abstract])

#4. #2 OR #3

#5. (((randomized controlled trial[Title/Abstract]) OR random[Title/Abstract]) OR randomly[Title/Abstract]) OR placebo[Title/Abstract]

#6. #4 AND #5

Embase

#1. 'lumber disc herniation':ti,ab,kw OR 'lumbar disc herniation':ti,ab,kw OR 'lumbar disc protrusion':ti,ab,kw OR ldh:ti,ab,kw OR 'intervertebral disk displacement':ti,ab,kw

#2. moxibustion:ti,ab,kw OR moxabustion:ti,ab,kw OR 'moxibustion therapy':ti,ab,kw OR moxa:ti,ab,kw

#3. randomized controlled trial (Topic) or random (Topic) or randomly (Topic) or placebo (Topic)

#4. #1 AND #2 AND #3

Cochrane Library

#1. (lumber disc herniation):ti,ab,kw OR (lumbar disc herniation):ti,ab,kw OR (lumbar disc protrusion):ti,ab,kw OR (LDH):ti,ab,kw OR (intervertebral disk displacement):ti,ab,kw

#2. MeSH descriptor: [Moxibustion] explode all trees

#3. (moxibustion):ti,ab,kw OR (moxabustion):ti,ab,kw OR (moxibustion therapy):ti,ab,kw OR (moxa):ti,ab,kw

#4. #2 OR #3

#5. (randomized controlled trial):ti,ab,kw OR (random):ti,ab,kw OR (randomly):ti,ab,kw OR (placebo):ti,ab,kw

#6. #4 AND #5

Web of Science

#1. lumber disc herniation (Topic) or lumbar disc herniation (Topic) or lumbar disc protrusion (Topic) or LDH (Topic) or intervertebral disk displacement (Topic)

#2. moxibustion (Topic) or moxabustion (Topic) or moxibustion therapy (Topic) or moxa (Topic)

#3. randomized controlled trial (Topic) or random (Topic) or randomly (Topic) or placebo (Topic)

#4. #1 AND #2 AND #3

EBSCO#1. SU lumber disc herniation OR SU lumbar disc herniation OR SU lumbar disc protrusion OR SU LDH OR SU intervertebral disk displacement

#2. SU moxibustion OR SU moxabustion OR SU moxibustion therapy OR SU moxa#3. SU randomized controlled trial OR SU random OR SU randomly OR SU placebo

#4. #1 AND #2 AND #3

the China National Knowledge Infrastructure

#1. SU=‘艾灸’ or ‘铺灸’ or ‘雷火灸’ or ‘热敏灸’ or ‘麦粒灸’ or ‘火龙灸’’ or ‘长蛇灸’or ‘艾柱灸’ or ‘艾条灸’ or ‘灸法’ or ‘灸疗’ or ‘（隔 /PREV 2 灸）’

#2. SU=‘腰椎间盘突出’ or ‘腰椎间盘膨出’

#1. 뜸

#2. 허리 디스크

#3. #1 OR #2

National Discovery for Science Leaders

#1. 뜸

#2. 허리 디스크

#3. #1 OR #2

Japan Science & Technology Information Aggregator, Electronic

#1. 灸

#2. 椎間板ヘルニア

#3. #1 OR #2

Figure S3. Funnel plot of publication bias.