Systematic review and meta-analysis of the anti-obesity effect of cupping therapy

Background:

Cupping therapy is a common practice in Korean medicine. Despite developments in this clinical and research area, the current knowledge is insufficient to identify the effects of cupping therapy on obesity. We aimed to assess the effects and safety of cupping therapy on obesity by performing a systematic review and meta-analysis of the effects of cupping therapy.

Methods:

A systematic search of databases was conducted, including MEDLINE/PubMed, EMBASE, Cochrane Central Register of Controlled Trials, Chinese National Knowledge Infrastructure, Citation Information by the National Institute of Informatics, KoreaMed, Oriental Medicine Advanced Searching Integrated System, and ScienceON, for full-text randomized controlled trials (RCTs) published through January 14, 2023, with no language restrictions. The experimental groups received cupping therapy combined with traditional Chinese medicine (TCM) and conventional therapy. The control groups received no treatment, conventional therapy, or TCM treatments alone. The experimental and control groups were compared in terms of body weight (BW), body mass index (BMI), hip circumference (HC), waist circumference (WC), waist-hip ratio (WHR), and body fat percentage (BFP). We evaluated the risk of bias using the 7 domains stipulated by the Cochrane Collaboration Group and performed a meta-analysis using Cochrane Collaboration software (Review Manager Software Version 5.3).

Results:

A total of 21 RCTs were included in this systematic review and meta-analysis. The analysis revealed improvements in BW (P < .001), BMI (P < .001), HC (P = .03), and WC (P < .001). However, there were no clinically significant changes in WHR (P = .65) or BFP (P = .90), both of which had very low certainty of evidence. No adverse events were reported.

Conclusion:

Overall, our results show that cupping therapy can be used to treat obesity in terms of BW, BMI, HC, and WC and is a safe intervention for the treatment of obesity. However, the conclusions of this review should be interpreted with caution in clinical practice because of the uncertain quality of the included studies.

1. Introduction

Obesity is a chronic, progressive metabolic disease that impairs an individual’s health and quality of life, reduces lifespan, and is increasing globally in the adult population.^[1,2] The development and persistence of obesity are influenced by genetics and environmental factors, such as^[3] lack of physical activity owing to the development of technology and overconsumption of food.^[4] Obesity is more than just the state of being overweight because it can lead to many other fatal diseases, such as hypertension,^[5] cardiovascular disease,^[6] diabetes mellitus,^[7] asthma,^[8] and cancer.^[9]

Many treatments have been developed for obesity. Initial treatments include comprehensive lifestyle modifications,^[10] dietary therapy,^[11] exercise,^[12] and behavioral modifications. For diet therapy, various types of diets have been suggested by healthcare researchers, such as low-calorie or low-carbohydrate diets,^[13] whereas exercise can be effective in maintaining weight loss by increasing energy expenditure.^[13] Additionally, less sedentary time and increased physical activity should be combined with calorie-restricted diets to be more effective.^[13] If treatments such as lifestyle therapy are not effective, medications prescribed by doctors may be considered a subsequent treatment.^[14] However, these dietary and exercise regimens should be conducted simultaneously when applying pharmacological therapies. Even with a combination of therapies, the success rate of obesity treatment is unsatisfactory owing to its multifactorial nature.^[15]

Cupping therapy is a common medical practice in Korean medicine.^[16] Recently, there has been growing evidence regarding the potential benefits of cupping for various diseases, including musculoskeletal,^[17] dermatologic,^[18] and respiratory conditions,^[19] and in East Asian countries, it is a widely used treatment in clinical situations for overweight. Despite developments in the clinical and research areas of cupping therapy, identification of the effects of cupping therapy on obesity remains insufficient. Therefore, this study presents a systematic review and meta-analysis of cupping therapy for obesity to investigate its effect.

2. Materials and methods

2.1. Study registration

The protocol for this systematic review has been registered in the Open Science Framework (osf.io/hx89s) and follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (see Supplemental Digital Content 1, http://links.lww.com/MD/J145, Table of PRISMA checklist).^[20]

2.2. Eligibility criteria for the study selection

2.2.1. Types of studies.

Only randomized clinical trials (RCTs) were analyzed to ensure a higher quality of evidence. Non-RCTs, retrospective chart reviews, observational studies, and case studies were excluded. There were no language limitations in the selection process.

2.2.2. Types of participants.

This study focused on patients diagnosed with obesity. There were no restrictions on sex, age, race, disease duration, or disease severity.

2.2.3. Types of interventions.

The experimental group treated with cupping therapy was included in this study. Blood-letting and moving cupping therapies were considered for analysis because they contain cupping constituents. The control groups were sham, no treatment, or standard conventional treatment. They were subjected to no other restrictions under the assumption that the same treatments were applied to the intervention group. Accompanying therapies included drugs, lifestyle interventions, and other types of interventions from traditional Chinese medicine (TCM).

2.2.4. Types of outcome measures.

Body weight (BW) was the primary outcome because it is the most essential outcome of obesity. Other measurements such as body mass index (BMI), waist-hip ratio (WHR), waist circumference (WC), hip circumference (HC), body fat mass (BFM), body fat percentage (BFP), and adverse events were the secondary outcomes.

2.3. Search strategy

2.3.1. Electronic data.

Three Korean databases (KoreaMed, Oriental Medicine Advanced Searching Integrated System, and ScienceON), 1 Japanese database (Citation Information by the National Institute of Informatics), 1 Chinese database (Chinese National Knowledge Infrastructure), MEDLINE/PubMed, EMBASE, and the Cochrane Central Register of Controlled Trials were searched from their inception to January 14, 2023. The search strategy for PubMed is shown in Supplemental Digital Content 2, http://links.lww.com/MD/J146.

2.3.2. Search for other resources.

The reference lists of the selected studies were read and referenced for a wider literature review. The researchers also manually searched for offline data. To identify the relevant literature that was not captured by these search methods, the references of the papers were screened by titles and abstracts. The World Health Organization International Clinical Trials Registry Platform was used for unpublished trials. Furthermore, literature that could not be searched online (i.e., hard copies) was manually searched. We also contacted the researchers of ongoing studies to verify the information when necessary.

2.4. Data collection and analysis

2.4.1. Study selection.

Following a singular guideline, 2 reviewers independently searched databases and other sources. Duplicates were removed by investigating their titles. Titles and abstracts were read to select relevant studies for a full-text review. The 2 reviewers independently read the filtered studies to extract the necessary data. Disagreements between the 2 researchers were discussed with a third researcher, and the conclusion of the third reviewer was the final decision on the issue. Specific content with examples was provided in the guidelines to ensure better agreement. A flow diagram of the search and processes is shown in Figure 1 (data extraction and management).

Figure 1.

Flow diagram of the study.

2.4.2. Data extraction and management.

An Excel spreadsheet was shared by the 2 reviewers for data extraction. The items included the exact information of the authors/groups; number of participants; and details of treatments, such as the type, frequency, and duration; measurements; data of results; and adverse events.

2.4.3. Assessment of the risk of bias and quality of the included studies.

The risk of bias was evaluated by 2 reviewers using the 7 domains stipulated by the Cochrane Collaboration Group.^[21]

The 7 domains included random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other sources of bias. The risk of bias was graded as high-risk, low-risk, or unclear for each domain. Discrepancies between the 2 reviewers were resolved by discussion with third reviewers. Researchers assessed bias as “high risk,” “low risk,” or “uncertain risk.”

2.4.4. Assessment of the effect of treatment.

Continuous data are presented as mean differences (MDs) and 95% confidence intervals.

2.4.5. Management of missing data.

If data were missing in the selected papers, the corresponding author of the study was contacted for additional information. If sufficient data were not provided, they were excluded.

2.4.6. Data synthesis.

Weighted MDs were adopted when the same scale was used, whereas standardized MDs were adopted when different indicators were used to measure the outcomes of the included studies. We conducted a meta-analysis to estimate the differences between the groups in primary and secondary outcomes using the Cochrane Collaboration software (Review Manager Software Version 5.3). Depending on the level of heterogeneity among the included studies, fixed- or random-effects models were used. A random-effects model was used when I² > 50%, whereas a fixed-effects model was used when I² < 50%. A subgroup analysis was conducted when heterogeneity was considered high, and the criteria designed by the researchers were applied when dividing the groups. The potential publication bias was assessed using funnel plots.

2.4.7. Subgroup analysis.

The criteria for the subgroup analysis were based on the type of treatment administered to the experimental group.

2.4.8. Sensitivity analysis.

Sensitivity analyses were executed to analyze sources of heterogeneity when necessary.

2.4.9. Ethics and dissemination.

The aforementioned protocol was used for systematic review and meta-analysis. Patient information was not required in this study. As this study assessed completed studies, it did not require approval from the institutional review board. The results and conclusions of this review will be submitted to a worldwide journal for the dissemination of the findings.

The Materials and Methods section should contain sufficient details so that all procedures can be repeated.

3. Results

3.1. Included studies

The search and selection procedures were performed according to the PRISMA guidelines (Fig. 1). A total of 244 studies were initially identified. Two reviewers screened the studies, and 13 duplicate records were removed. The titles and abstracts of the remaining studies were further examined, eliminating 82 records because the intervention was not cupping, the condition was not obesity or simple obesity, or the study design was not an RCT. Subsequently, 126 studies were removed because the retrieved reports included patients who were not obese, lacked an intervention, were not RCTs, did not involve humans, or lacked appropriate control groups. Full-text studies were assessed for eligibility, and 2 records were eliminated because they had no outcome of interest. Finally, 223 studies were excluded, and 21 RCTs^[22–42] involving 1563 patients were included in the systematic review and meta-analysis. The reasons for exclusion and the selection flow are presented in Figure 1.

3.2. Study characteristics

The characteristics of the included studies^[22–42] are presented in detail in Table 1. These studies were published between 2007 and 2020 and were all published in Chinese. Sample sizes ranged from 41 to 177. All patients in these studies underwent cupping therapy for obesity.

Table 1.

Characteristics of the included studies.

Study ID	Total sample size (T/C)	Intervention group	Control group	Treatment frequency/period	F/U	Outcome measurement	Results	Adverse events	Origin of the study
Bo[33]	54 (27/27)	Cupping + Exercise and diet	Exercise and diet	Twice a week/2 mo	Nr	1) BW 2) BMI 3) WHR 4) WC 5) HC	1) T < C, P = .27 2) T < C, P = .02 3) T > C, P = .51 4) T < C, P = .38 5) T < C, P = .05	Nr	China
Danna et al[24]	76 (38/38)	Cupping + HM	HM	Once a week/3 mo	Nr	1) BMI	1) T < C, P = .0007	Nr	China
Wang[34]	107 (55/52)	Cupping + HM	HM	1 day in 10 d/42 d	Nr	1) BW 2) BMI 3) BFP	1) T < C, P = .04 2) T < C, P = .0007 3) T < C, P = .50	Nr	China
Xiaoling[38]	60 (30/30)	Cupping + MA	MA	3 days a week/3 mo	Nr	1) BW 2) BMI 3) WHR 4) WC 5) HC	1) T < C, P = .0008 2) T < C, P < .00001 3) T < C, P = .05 4) T < C, P < .0001 5) T < C, P = .19	Nr	China
Chen[40]	60 (30.30)	Cupping + MA	MA	5 days a week/30 d	Nr	1) BW 2) WC	1) T < C, P = .0008 2) T < C, P < .0001	Nr	China
Bu et al[22]	120 (63/57)	Cupping + MA	MA	Every other day/60 d	Nr	1) BW 2) WC	1) T > C, P = .0008 2) T < C, P < .0001	Nr	China
Chen[39]	63 (33/30)	Cupping + MA	MA	3 days a week/4 wk	4, 8 weeks	1) BW 2) BMI 3) WC	1) T < C, P = .0008 2) T < C, P < .00001 3) T < C, P < .0001	Nr	China
Ma[25]	60 (30/30)	Cupping + MA	MA	Nr/3 wk	Nr	1) BW 2) BMI 3) WC 4) HC	1) T < C, P = .0008 2) T < C, P < .00001 3) T < C, P < .0001 4) T < C, P = .19	Nr	China
Xu[30]	100 (50/50)	Cupping + MA	MA	Every other day/90 d	Nr	1) BMI 2) BFP	1) T < C, P < .00001 2) T < C, P < .00001	Nr	China
Haochun[29]	177 (86/91)	Cupping + MA	MA	Every other day/47 d	Nr	1) BW 2) BMI 3) WHR 4) BFP	1) T < C, P = .0008 2) T < C, P < .00001 3) T < C, P = .05 4) T < C, P < .00001	Nr	China
Bu et al[22]	57 (32/25)	Cupping + MA	MA	Every other day/47 d	Nr	1) BW 2) BMI 3) BFP	1) T < C, P = .0008 2) T < C, P < .00001 3) T < C, P < .00001	Nr	China
Liang[23]	41 (21/20)	Cupping	None	3 days a week/8 wk	Nr	1) WC	1) T < C, P = .0002	Nr	China
Chen[41]	60 (30/30)	Cupping + EA	EA	Every other day/1 mo	Nr	1) BW 2) BMI 3) WHR 4) WC 5) HC	1) T < C, P = .0002 2) T < C, P < .00001 3) T > C, P = .16 4) T > C, P = .15 5) T < C, P = .44	Nr	China
Huang[42]	73 (37/36)	Cupping + EA	EA	Every other day/3 mo	Nr	1) BW 2) BMI 3) WC 4) BFP	1) T < C, P = .0002 2) T < C, P < .00001 3) T < C, P = .15 4) T < C, P = .002	Nr	China
Shengqi[36]	66 (33/33)	Cupping + EA	EA	Every other day/3 mo	Nr	1) BW 2) BMI 3) WC 4) BFP	1) T < C, P = .0002 2) T < C, P < .00001 3) T < C, P = .15 4) T < C, P = .002	Nausea, local hematoma	China
Miaojun[27]	92 (44/48)	Cupping + WA	WA	Every other day/3 mo	Nr	1) BW 2) BMI 3) WC	1) T < C, P = .04 2) T < C, P = .002 3) T < C, P = .03	Nr	China
Binbin[37]	60 (30/30)	Cupping + ET	ET	1 in 10 days/2 mo	Nr	1) BW 2) BMI	1) T < C, P < .00001 2) T < C, P < .0001	Nr	China
Dan[31]	57 (29/28)	Cupping + ET	ET	Once a week/3 mo	Nr	1) BW 2) BMI 3) WC	1) T < C, P < .00001 2) T < C, P < .0001 3) T < C, P < .00001	Nr	China
Lin[28]	60 (30/30)	Cupping + ET	ET	Nr/2 mo	Nr	1) BW 2) BMI	1) T < C, P < .00001 2) T < C, P < .0001	Nr	China
Junfang[35]	60 (30/30)	Cupping + ET	ET	Nr/12 wk	4, 8 wk	1) BW 2) BMI 3) WC	1) T < C, P < .00001 2) T < C, P < .0001 3) T < C, P < .00001	Nr	China
Yi[26]	60 (30/30)	Cupping + GT	GT	Once a week/3 mo	Nr	1) BW 2) BMI 3) WHR 4) WC 5) HC 6) BFP	1) T < C, P = .05 2) T < C, P = .31 3) T > C, P = .19 4) T > C, P = .59 5) T < C, P = .30 6) T > C, P < .00001	Nr	China

BFP = body fat percentage, BMI = body mass index, BW = body weight, C = control group, EA = electroacupuncture, ET = embedding therapy, GT = guasha therapy, HC = hip circumference, HM = herbal medicine, MA = manual acupuncture, Nr = not reported, T = treatment group, WA = warm acupuncture, WC = waist circumference, WHR = waist-hip ratio.

The experimental group received TCM and conventional therapies. Two studies used cupping with herbal medicines.^[24,33] One study used cupping with exercise and diet.^[34] In 8 studies, patients were treated with cupping and manual acupuncture,^{[22,25,29,30,32,38–40]} whereas in 3 studies,^[36,41,42] patients were treated with cupping and electroacupuncture. In one study, patients were treated with cupping and warm acupuncture,^[27] whereas in 4 studies, patients were treated with cupping and embedding therapy.^{[28,31,35,37]} One study used cupping with Guasha therapy,^[26] whereas another study used cupping alone.^[23]

The duration of treatment in the experimental groups ranged from 3 weeks to 3 months. The details of the cupping therapy are presented in Table 2. Cupping was performed from an acupoint based on TCM theory, including moving cupping therapy for meridians or specific sites in the body. Cupping therapy is performed at acupoints in the abdomen, and moving cupping is mainly used in the governor vessel, conception vessel (CV), bladder meridian, spleen meridian, and stomach meridian (ST). The 2 most frequently used acupoints were CV12 and ST25, both of which were located in the abdomen. Moving cupping has also been widely used on the back, abdomen, and upper and lower limbs. The cupping therapies were applied for 5 to 20 minutes.

Table 2.

Details of cupping therapy in different studies.

First author	Cupping rationale	Cupping points	Cupping duration
Bo[33]	TCM	Cupping (ST25, GB26, CV4, CV12) + Moving cupping (abdomen)	Cupping (10 min), moving cupping (5 min)
Danna et al[24]	TCM	Moving cupping (CV3, SP15, SP16, ST25, ST28)	Until the treatment area is red
Wang[34]	TCM	Moving cupping (SP, ST, KI, BL, CV, GV, LI, SI, TE)	10 min
Xiaoling[38]	TCM	ST26, ST28, ST30, KI12, KI14, and moving cupping (abdomen)	Cupping (5 min), Moving cupping (10–15 min)
Chen[40]	TCM	CV4, ST25, SP15, and moving cupping (abdomen)	5 min
Bu et al[22]	TCM	Abdomen, upper limb, lower limb	20–25 min
Chen[39]	TCM	CV4, CV6, CV9, CV12, ST25, SP15, GB26	5 min
Ma[25]	TCM	CV4, CV12, ST25	Nr
Xu[30]	TCM	BL, ST	10 min
Haochun[29]	TCM	CV6, CV12, ST25, GB26, LR10, SP11	10 min
Bu et al[22]	TCM	Moving cupping (BL)	Nr
Liang[23]	TCM	CV6, CV12, ST21, ST25, ST28, GB26,	5 min
Jiaoqi[41]	TCM	CV4, CV6, CV12, ST25, ST28, SP15	Nr
Mong[42]	TCM	Moving cupping (BL, GV, CV, SP, ST)	15 min
Shengqi[36]	TCM	Moving cupping (BL, GV, CV, SP, ST)	15 min
Miaojun[27]	TCM	Moving cupping (CV, SP, ST, KI)	Nr
Binbin[37]	TCM	ST24, ST25, ST36, CV12, SP15, GB26, BL20, BL23, and moving cupping (abdomen, low back, thigh, upper limb)	Nr
Dan[31]	TCM	ST25, CV12, ST24, ST26, CV9, GB26	10 min
Lin[28]	TCM	Abdomen, upper limb, thigh, low back	Nr
Junfang[35]	TCM	Cupping and moving cupping (CV9, CV12, ST25, ST26, ST28, SP15, GB26)	Cupping (8–10 min), moving cupping (20 min)
Yi[26]	TCM	Back, abdomen, upper limb, lower limb	5–15 min

BL = bladder meridian, CV = conception vessel meridian, GB = gallbladder meridian, GV = governor vessel meridian, KI = kidney meridian, LI = large intestine meridian, N.r = not reported, SI = small intestine meridian, SP = spleen meridian, ST = stomach meridian, TCM = traditional Chinese medicine, TE = triple energizer meridian.

Patients in the control group underwent manual acupuncture, including embedding therapy, Guasha therapy, pharmacological treatments, exercise, and diet. The same treatments were applied to all the experimental groups. One control group did not receive any treatment.

Eighteen studies^{[22,25–29,31–42]} used BW as a primary indicator to measure the effects of cupping. BMI, WHR, WC, HC, BFM, BFP, and adverse events were considered secondary outcomes. Overall, 18,^{[22,24–31,33–39,41,42]} 5,^{[25,26,34,38,41]} 14,^{[23,25–27,31,32,34–36,38–42]} 5,^{[26,29,34,38,41]} and 7^{[22,26,29,30,33,36,42]} studies reported BMI, HC, WC, WHR, and BFP data, respectively. One study^[36] reported adverse events such as nausea and local hematoma after TCM treatment, which did not significantly differ between the experimental and control groups.

3.3. Risk of bias in the included studies

The risk of bias is shown in Figures 2 and 3. For the assessment of sequence generation, 9 studies^{[22,24,28,34,36,38–40,42]} were considered to have a low risk of bias, whereas 3 studies^[25,29,37] were considered to have a high risk of bias. The remaining 9 studies^{[23,26,27,30–33,35,41]} were considered to have an uncertain risk of bias because the authors stated that they used a randomization method, but specific details were missing.

Figure 2.

Risk of bias graph.

Figure 3.

Summary of risk of bias.

In the assessment of allocation concealment, one study^[38] adequately performed allocation concealment using opaque closed-letter envelopes, whereas other studies^{[22–37,39–42]} did not describe their approaches to allocation concealment in detail. Only one study^[38] was evaluated as low-risk based on participant and personnel blinding because it was difficult to exclude performance bias in administering TCM treatments.

Blinding of participants could not be achieved in all studies because a placebo cupping model was not possible. Eighteen studies^{[22,23,25–37,39,41,42]} were evaluated as high-risk. The other 3 studies^[24,38,40] did not include participant blinding.

For blinding the outcome assessment, 11 studies^{[22,23,26,27,31,33,34,36,37,41,42]} were assessed as having a high risk of bias. Other studies^{[24,25,28–30,32,35,38–40]} did not clarify the specific methods used to prevent detection bias.

When evaluating incomplete outcome data, 10 RCTs^{[25,26,28,30–35,41]} were ranked as having a low risk of bias because the reasons for participant dropout were reported. Five RCTs^{[23,36–38,42]} reported the absence of dropouts. The remaining 6 studies^{[22,24,27,29,39,40]} did not provide this information and were evaluated as having an uncertain risk of bias.

In 6 studies,^{[27,28,30,32,35,37]} selective reporting was not decided; therefore, reporting was ambiguous. Thirteen studies^{[22–26,29,33,34,36,38,39,41,42]} published protocols, and all pre-specified outcomes were reported in the results. Two other studies^[31,40] reported different evaluation time points in the materials and methods and results sections.

Other forms of bias were evaluated as having an uncertain risk in all studies.^[22–42] None of the included studies clarified a significant difference between the general characteristics of the groups.

3.4. Effects of interventions

Twenty-one studies^[22–42] involving 1563 participants were included in this meta-analysis. The meta-analysis was based on 8 outcomes: BW, BMI, WHR, WC, HC, BFM, BFP, and adverse events. Standardized MDs was used for the meta-analysis. A subgroup analysis was conducted for all outcomes because considerable heterogeneity was revealed in the analysis of the pooled effect.

3.4.1. Primary outcomes.

BW was the primary outcome, which was used to evaluate the effects of cupping. A meta-analysis of 1280 participants^{[22,25–29,31–42]} revealed that cupping showed a significant improvement in participants with obesity, with low heterogeneity (MD = −3.38, 95% Cl = −4.11 to −2.65, I² = 0%; Fig. 4).

Figure 4.

Body weight showing subgroup analysis by treatment types. EA = electroacupuncture, ET = embedding therapy, GT = Guasha therapy, HM = herbal medicine, MA = manual acupuncture, WA = warm acupuncture.

The subgroup analysis was based on the type of TCM treatment. Most TCM treatment types, including cupping plus manual acupuncture (MD = −2.70, 95% Cl = −4.27 to −1.13, I² = 0%),^{[22,25,29,30,32,38–40]} cupping plus herbal medicine (MD = −3.45, 95% Cl = −6.74 to −0.16),^[24,33] cupping plus warm acupuncture (MD = −3.60, 95% Cl = −6.95 to −0.25),^[27] cupping plus electroacupuncture (MD = −4.91, 95% Cl = −7.52 to −2.29, I² = 0%),^[36,41,42] and cupping plus embedding therapy (MD = −3.82, 95% Cl = −4.89 to −2.75, I² = 0%),^{[28,31,35,37]} showed better effects than the control group excluding cupping plus Guasha therapy (MD = −1.98, 95% Cl = −3.98 to 0.02)^[26] and cupping plus exercise and diet (MD = −3.43, 95% Cl = −9.47 to 2.61).^[34]

3.4.2. Secondary outcomes.

3.4.2.1. Body mass index.

In the meta-analysis of 18 studies^{[22,24–31,33–39,41,42]} involving 1342 patients, the cupping group showed significantly better improvement in BMI than the control group (MD = −1.23, 95% Cl = −1.48 to −0.99, I² = 57.3%; Fig. 5).

Figure 5.

Forest plot for body mass index showing subgroup analysis by treatment types. EA = electroacupuncture, ET = embedding therapy, GT = Guasha therapy, HM = herbal medicine, MA = manual acupuncture, WA = warm acupuncture.

In the subgroup analysis, the cupping groups showed better BMI than the control group: cupping plus herbal medicine (MD = −1.05, 95% Cl = −1.65 to −0.44, I² = 0%),^[24,33] cupping plus manual acupuncture (MD = −1.47, 95% Cl = −1.88 to −1.06, I² = 0%),^{[22,25,29,30,32,38–40]} cupping plus electroacupuncture (MD = −1.47, 95% Cl = −2.10 to −0.85, I² = 0%),^[36,41,42] cupping plus warm acupuncture (MD = −2.30, 95% Cl = −3.73 to −0.86),^[27] cupping plus embedding therapy (MD = −1.38, 95% Cl = −2.07 to −0.69, I² = 26%),^{[28,31,35,37]} and cupping plus exercise and diet (MD = −1.76, 95% Cl = −3.28 to −0.24).^[34] However, one cupping group showed no significant difference between the intervention and control groups: cupping plus Guasha therapy (MD = −0.31, 95% Cl = −0.91 to 0.29).^[26]

3.4.2.2. Hip circumference.

In a meta-analysis of 5 studies^{[25,26,34,38,41]} involving 294 patients, the cupping group showed better improvement in HC than the control group (MD = −1.25, 95% Cl = −2.36 to −0.14, I² = 0%; Fig. 6). However, all 4 subgroups showed no significant difference between the intervention and control groups: cupping plus manual acupuncture (MD = −1.87, 95% Cl = −4.64 to 0.90, I² = 0%),^[25,38] cupping plus electroacupuncture (MD = −2.42, 95% Cl = −8.60 to 3.76),^[41] cupping plus Guasha therapy (MD = −0.71, 95% Cl = −2.06 to −0.64),^[26] and cupping plus exercise and diet (MD = −3.11, 95% Cl = −6.23 to 0.01).^[34]

Figure 6.

Forest plot for hip circumference (subgroup analyzed by treatment types). EA = electroacupuncture, ET = embedding therapy, GT = Guasha therapy, MA = manual acupuncture, WA = warm acupuncture.

3.4.2.3. Waist circumference.

In a meta-analysis of 14 studies^{[23,25–27,31,32,34–36,38–42]} involving 926 patients, the cupping group showed significantly better improvement in WC than the control group (MD = −3.19, 95% Cl = −3.81 to −2.56, I² = 88%; Fig. 7). The cupping groups showed significant improvement in WC: cupping alone (MD = −10.64, 95% Cl = −16.14 to −5.14),^[23] cupping plus manual acupuncture (MD = −3.93, 95% Cl = −5.79 to −2.07, I² = 0%),^{[25,32,38–40]} cupping plus embedding therapy (MD = −6.28, 95% Cl= −7.30 to −5.25, I² = 0%),^[31,35] and cupping plus warm acupuncture (MD = −2.98, 95% Cl = −5.63 to −0.33).^[27] However, 3 subgroups showed that there was no significant difference between the intervention and control groups: cupping plus electroacupuncture (MD = −1.26, 95% Cl = −2.96 to 0.45, I² = 69%),^[36,41,42] cupping plus Guasha therapy (MD = 0.32, 95% Cl = −0.83 to 1.47),^[26] and cupping plus exercise and diet (MD = −1.92, 95% Cl = −6.22 to 2.38).^[34]

Figure 7.

Forest plot for waist circumference (subgroup analyzed by treatment types). EA = electroacupuncture, ET = embedding therapy, GT = Guasha therapy, MA = manual acupuncture, WA = warm acupuncture.

3.4.2.4. Waist hip ratio.

In a meta-analysis of 5 studies^{[26,29,34,38,41]} involving 411 patients, the cupping group showed no improvement in WHR (MD = −0.01, 95% Cl = −0.03 to 0.02, I² = 50%; Fig. 8). All 4 subgroups showed no significant difference between the intervention and control groups: cupping plus manual acupuncture (MD = −0.03, 95% Cl = −0.07 to 0.00, I² = 0%),^[29] cupping plus electroacupuncture (MD = 0.08, 95% Cl = −0.03 to 0.19),^[41] cupping plus Guasha therapy (MD = 0.96, 95% Cl = −0.48 to 2.40),^[26] and cupping plus exercise and diet (MD = 0.01, 95% Cl = −0.02 to 0.04).^[34]

Figure 8.

Forest plot for waist hip ratio (subgroup analyzed by treatment types). EA = electroacupuncture, ET = embedding therapy, GT = Guasha therapy, MA = manual acupuncture.

3.4.2.5. Body fat percentage.

In the meta-analysis of 7 studies involving 640 patients,^{[22,26,29,30,33,36,42]} the cupping groups showed no improvement in BFP (MD = −0.02, 95% Cl = −0.39 to 0.34, I² = 98%; Fig. 9). Two subgroups of cupping showed significant improvement in BFP: cupping plus manual acupuncture (MD = −1.18, 95% Cl = −1.65 to −0.72, I² = 0%)^[22,29,30] and cupping plus electroacupuncture (MD= −1.11, 95% Cl = −1.83 to −0.39, I² = 48%).^[36,42] However, the other 2 subgroups showed no significant difference between the intervention and control groups: cupping plus herbal medicine (MD = −1.04, 95% Cl = −4.05 to 1.97)^[33] and cupping plus Guasha therapy (MD = 10.64, 95% Cl = 9.45–11.83).^[26]

Figure 9.

Forest plot for body fat percentage (subgroup analyzed by treatment types). EA = electroacupuncture, GT = Guasha therapy, HM = herbal medicine, MA = manual acupuncture.

3.5. Adverse events

Only one study^[36] reported side effects in the cupping and control groups. Zhang et al^[36] confirmed a case of local hematoma and nausea. Since symptoms emerged in the intervention and control groups, adverse events in the cupping groups remain uncertain.

3.6. Publication bias

Publication bias in studies was investigated using BW^{[22,25–29,31–42]} and BMI^{[22,24–31,33–39,41,42]} using funnel plots. They showed a relatively asymmetrical distribution, indicating high-risk publications (Figs. 10 and 11).

Figure 10.

Funnel plot for the outcomes of body weight. EA = electroacupuncture, ET = embedding therapy, GT = Guasha therapy, HM = herbal medicine, MA = manual acupuncture, WA = warm acupuncture.

Figure 11.

Funnel plot for the outcomes of body mass index. EA = electroacupuncture, ET = embedding therapy, GT = Guasha therapy, HM = herbal medicine, MA = manual acupuncture, WA = warm acupuncture.

4. Discussion

We searched a wide range of databases for studies investigating the effects of cupping therapy on obesity. A total of 21 RCTs^[22–42] were included in the systematic review and meta-analysis, from which we concluded that cupping therapy was effective for obesity. BW,^{[22,25–29,31–42]} BMI,^{[22,24–31,33–39,41,42]} WHR,^{[26,29,34,38,41]} WC,^{[23,25–27,31,32,34–36,38–42]} HC,^{[25,26,34,38,41]} BFM, and BFP^{[22,26,29,30,33,36,42]} were used to evaluate the obesity status. However, none of these studies used BFM as an indicator of the effects of obesity treatment.

Currently, there are several TCM treatments for obesity, including herbal medicine, manual acupuncture, electroacupuncture, warm acupuncture, embedding therapy, and Guasha therapy. Conventional therapies include exercise and diet. The effects of cupping therapy on obesity are unknown to the public, although there is growing evidence for its potential benefits in the treatment of various musculoskeletal, dermatological, and pulmonary diseases. Therefore, this systematic review and meta-analysis aimed to identify the clinical effects of cupping therapy on obesity.

This meta-analysis found that cupping therapy combined with other TCM or conventional therapies was effective for the treatment of obesity. Furthermore, compared with the no-treatment group, only the cupping group showed an improvement in obesity. BW,^{[22,25–29,31–42]} which was the primary outcome of this review, decreased in the cupping groups, excluding the cupping plus Guasha therapy^[26] and cupping plus exercise and diet groups.^[34] Among the indicators used for the comprehensive evaluation of obesity, BMI,^{[22,24–31,33–39,41,42]} HC,^{[25,26,34,38,41]} and WC^{[23,25–27,31,32,34–36,38–42]} decreased in the cupping groups based on the meta-analysis. Owing to the considerable heterogeneity involved, we conducted a subgroup analysis. During the subgroup analysis of WHR^{[26,29,34,38,41]} and BFP^{[22,26,29,30,33,36,42]} based on the treatment type, cupping plus Guasha therapy^[26] showed no improvement in WHR or BFP. However, cupping plus herbal medicine^[24,33] and manual acupuncture^{[22,25,29,30,38,39]} resulted in significant improvements in BW and BMI with low heterogeneity, suggesting that more diverse TCM treatments tend to be effective for obesity.

The WHR used to evaluate the effects of cupping was presented in 5 studies.^{[26,29,34,38,41]} All studies revealed no difference between the intervention and control groups. The results of the meta-analysis revealed no significant improvements. Moreover, BFP was calculated in 7 studies.^{[22,26,30,33,36,42]} Cupping plus manual acupuncture^[22,29,30] and electroacupuncture^[36,42] showed significant effects of cupping. However, the results of the meta-analysis showed no significant improvement.

The most frequently used acupoints for cupping were CV12 and ST25 (Table 2), which are located in the abdomen. Moreover, all acupoints used for cupping therapy were located within the abdomen.

One trial^[36] reported adverse events, including nausea and local hematoma, after TCM treatment. However, this was observed in both the intervention and control groups. Furthermore, the symptoms resolved spontaneously without any life-threatening complications. This systematic review and meta-analysis showed that cupping therapy combined with other treatments or conventional therapies for obesity is relatively safe. However, more studies are needed to validate our findings because only a few of the included RCTs reported adverse events.

The risk of bias across studies^[22–42] was relatively high with respect to the blinding of participants and personnel. The experimental groups were treated with cupping therapy, whereas the control groups were treated with other TCM treatments, exercise, or diet. Therefore, it was impossible to blind the participants completely. Allocation concealment^{[22–37,39–42]} has not been clearly reported in most studies. However, selective reporting has been reported in many studies.^{[22–26,29,33,34,36,38,39,41,42]} Other biases regarding inclusion and exclusion criteria^[22–42] have not been properly stated.

The quality of evidence was downgraded by 1 level for the risk of bias in the subjective outcomes because it seemed impossible to blind the participants completely to the TCM treatments (Table 3). The quality of evidence was also downgraded because of inconsistencies in the results owing to considerable heterogeneity. Indirectness was not proven because all studies directly compared interventions. There was no evidence of imprecision because all studies used sufficient sample sizes (>100 participants). However, the MD was 0. Therefore, the quality of the evidence was downgraded by one. The ratings of the quality of evidence for the overall outcomes ranged from moderate to very low because of performance bias and heterogeneity. The evidence quality ratings were as follows: moderate for BW, moderate for BMI, moderate for HC, very low for WC, very low for WHR, and very low for BFP.

Table 3.

Summary of findings.

Outcome	Anticipated absolute effects (95% CI)	Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
BW	MD 3.38 lower (4.11 lower to 2.65 lower)	–	1346 (18 RCTs)	⨁⨁⨁◯ MODERATE*	Patients who received cupping therapy had a lower BW with only moderate certainty due to a high risk of performance bias.
BMI	MD 1.23 lower (1.48 lower to 0.99 lower)	–	1342 (18 RCTs)	⨁⨁⨁◯ MODERATE*	Patients who received cupping therapy had a lower BMI with moderate certainty due to a high risk of performance bias.
HC	MD 1.25 lower (2.36 lower to 0.14 lower)	–	294 (5 RCTs)	⨁⨁⨁◯ MODERATE*	Patients who received cupping therapy had a lower HC with moderate certainty due to a high risk of performance bias.
WC	MD 3.19 lower (3.81 lower to 2.56 lower)	–	926 (14 RCTs)	⨁◯◯◯ VERY LOW*,†	Patients who received cupping therapy had a lower WC with uncertainty due to a high risk of performance bias and considerable heterogeneity.
WHR	MD 0.01 lower (0.03 lower to 0.02 higher)	–	411 (5 RCTs)	⨁◯◯◯ VERY LOW*,‡,§	There was no significant difference in the WHR between the intervention and control groups with uncertainty due to performance bias, heterogeneity, and confidence interval.
BFP	MD 0.02 lower (0.39 lower to 0.34 higher)	–	640 (7 RCTs)	⨁◯◯◯ VERY LOW*,†,§	There was no significant difference in the BFP between intervention groups and control groups with uncertainty due to performance bias, considerable heterogeneity, and confidence interval.

BFP = body fat percentage, BMI = body mass index, BW = body weight, HC = hip circumference, MD = mean difference, RCT = randomized controlled trial, WC = waist circumference, WHR = waist-hip ratio.

^*Blinding of the treatment and researchers conducting the study is impractical given the nature of cupping.

^†If I square > 75%, researchers decide the study as “very serious.”

^‡Since the I square is 50% with an opposed direction of effect, researchers decide the study as “serious.”

^§Although the total participants are over 100 people, confidence interval includes 0.

Funnel plots, a reference for the judgement of publication bias, were drawn for the 2 meta-analyses of BW^{[22,25–29,31–42]} and BMI.^{[22,24–31,33–39,41,42]} These seemed asymmetrical, suggesting a high risk of publication bias.

This study has several limitations. First, owing to inconsistencies in the outcome measurements, the meta-analysis was restricted to a limited number of included studies. The small number of studies has led to insufficient evidence supporting this effect. For high-quality meta-analyses, a larger number of well-designed RCTs with consistent outcome measures are needed. Second, several studies with a high risk of bias were included, which should be considered when interpreting the results of this study. Third, the heterogeneity in the included studies was relatively high, owing to differences in regimen, treatment duration, treatment site, and trial design. Therefore, the results of this meta-analysis should be interpreted with caution, and well-designed RCTs with high-quality methodologies should be conducted in the future. Fourth, all the included studies were conducted and published in China. Since TCM treatments are commonly performed in China, cultural customs may have introduced bias in this review. More importantly, the findings reflect only East Asian populations; therefore, the conclusions of this review should be extrapolated to Western populations with caution. Consequently, more diverse RCTs, especially on the Western population, are required. Despite these limitations, this study is the first to analyze RCTs that used cupping therapy as an intervention for obesity.

Notably, cupping therapy can affect obesity by decreasing BW,^{[22,25–29,31–42]} BMI,^{[22,24–31,33–39,41,42]} HC,^{[25,26,34,38,41]} and WC.^{[23,25–27,31,32,34–36,38–42]} Despite these promising results, the influence of the evidence is weak, and no clear conclusions can be drawn, given that the overall quality of the studies was relatively low and the heterogeneity was not resolved by subgroup analysis. However, this study is the first systematic review and meta-analysis of the effects of cupping therapy on obesity and will be useful for future research on cupping therapy for obesity.

5. Conclusions

This study systematically reviewed and conducted a meta-analysis of the effects of cupping on obesity and concluded that cupping therapy combined with other TCM treatments or conventional therapies had a positive effect on obesity without any serious adverse events. Cupping therapy alone improved obesity. However, the conclusions of this review should be interpreted cautiously in clinical practice because of the high heterogeneity of the included studies and the potential uncertainty of the analyses. Further studies with robust designs and larger numbers of participants should be conducted in the future.

Author contributions

Conceptualization: Doyoung Kang.

Funding acquisition: Won-Seok Chung.

Investigation: Doyoung Kang.

Methodology: Sungha Kim, Jae-Heung Cho.

Project administration: Won-Seok Chung.

Resources: Sungha Kim.

Software: Taeoh Kim.

Supervision: Jae-Heung Cho.

Validation: Hyungsuk Kim, Mi-Yeon Song.

Visualization: Woo-Chul Shin, Mi-Yeon Song.

Writing – original draft: Doyoung Kang, Taeoh Kim.

Writing – review & editing: Woo-Chul Shin, Hyungsuk Kim.

Supplementary Material

medi-102-e34039-s001.pdf ^{(46KB, pdf)}

medi-102-e34039-s002.pdf ^{(201KB, pdf)}