Two reviewers independently performed a comprehensive search of the following electronic databases: four English databases [PubMed, Embase, Cochrane Central Register of Controlled Trials database (CENTRAL), and Web of Science (WOS)] and four Chinese databases (China National Knowledge Infrastructure (CNKI), SinoMed Database (CBM), VIP Database, and WF Database). In addition, three clinical trial registration platforms were searched: the World Health Organization (WHO) International Clinical Trials Registry Platform,² the Chinese Clinical Trial Registry,³ and ClinicalTrials.gov⁴ on wet cupping for neurodermatitis from their inception to March 2024. The following key search terms were imposed: (a) clinical conditions: neurodermatitis, lichen simplex chronicus, etc.; (b) wet cupping therapy-related words: cupping therapy, wet cupping, cupping, bloodletting, bleeding cupping, etc.; and (c) trial type: an RCT. The terms “and” and “or” were combined between the search terms. The search strategies for these sources are shown in the Supplementary Material.

All investigators received professional, evidence-based medicine training to implement this systematic review. After excluding duplicate articles and uploading potentially eligible studies into Endnote V.20 software, two reviewers (LD and JC) independently screened the titles, abstracts, and keywords of all search items. They identified trials that met the above-mentioned inclusion criteria. Divergences between the two reviewers were resolved through discussion between the two. A third party also assisted in making the final decision. A PRISMA flowchart illustrating the study selection process is included in Figure 1.

Two reviewers (YZ and LD) independently performed the data collection using a standardized tabulation. The extracted elements were as follows: basic information (title, year of publication, first author, language, and country of implementation), participant characteristics at baseline (sample size, age, gender, and course of disease), study design (randomization method, allocation concealment, blinding procedures, diagnostic criteria, intervention frequency and duration, and acupoint composition), and outcomes. In the case of multi-arm RCTs, they were reclassified as dual-arm RCTs to ensure compatibility for result synthesis. The two reviewers conducted data extraction and cross-checking, and any discrepancies were resolved through consultation with the corresponding author.

Two independent reviewers (Z-FJ and XT) evaluated the quality of the included studies using the Cochrane Collaboration’s Risk of Bias Assessment tool (25). We evaluated the five items (bias in the randomization process, bias in deviations from the intended intervention, bias in missing outcome data, bias in the measurement of an outcome, and selection bias in the reported results). Each item was assessed and classified as high risk, low risk, or risk of some concern. The corresponding author were asked to resolve any divergences.

The statistical software ReviewManager V.5.4 was used for data analysis. The mean difference (MD) was used as the effect size for continuous variables, while the risk ratio (RR) was used for dichotomous variables. The 95% confidence interval (CI) was calculated. A significance level of p-value <0.05 was considered statistically significant. Statistical heterogeneity was assessed using the chi-squared tests and the Higgins I² test. If p-value ≥0.1 and I² ≤ 50%, indicating a high degree of homogeneity, a fixed-effects model was used for the pooled analysis. Conversely, if p-value <0.1 or I² > 50%, indicating a significant degree of heterogeneity, a random-effects model was used. If there was significant heterogeneity between studies and the data were available, subgroups of the different cupping therapies and intervention forms were constructed to explore the potential causes of heterogeneity.

If more than 10 studies were included in the analysis, the funnel plot method and Egger’s test were used to evaluate publication biases. A p-value of <0.05 was considered indicative of significant publication bias. The analysis for publication bias was performed using STATA V.17.0.

The Grading of Recommendations Assessment, Development, and Evaluation System (GRADEpro) evaluation tool⁵ was independently used by two reviewers (YZ and Z-FJ) to assess the quality of evidence in terms of five dimensions: risk of bias, inconsistency, indirectness, imprecision, and publication bias. The quality of evidence was classified as high, moderate, low, or very low. In the event of any disagreement, a third reviewer was consulted to facilitate consensus.

A total of 1,771 studies were initially identified by searching the databases using the keywords mentioned earlier. After stepwise screening, 19 RCTs with 1,505 participants were included in our meta-analysis (29–47). The flowchart of PRISMA is depicted in Figure 1.

The main characteristics of the included RCTs are shown in Table 1. All studies were conducted in China between 1998 and 2023. Only one study was published in English (34), and the remaining studies were published in Chinese. All included studies were single-center, two-arm RCTs, of which 3 were master’s degree theses (37, 38, 41) and 16 were journal articles. The average age of the 1,505 participants ranged from 29 to 53 years and did not involve any minors or elderly patients. The course of the disease varied from 1 week to 20 years.

The risk of bias assessment was conducted for each study, and the results are shown in Figure 2. All the included studies mentioned the term “randomized” or “random” in their methodology; 11 studies used random number tables, 1 study used statistical software to generate random numbers, and 1 study used the method of drawing lots for random allocation, while the other 6 studies did not mention specific randomization methods. Except for two studies (41, 42), none of the other 17 studies explicitly discussed the details of allocation concealment. Due to the nature of cupping therapy, it is unfeasible to blind the operators, so all studies received a “high risk” rating in the third item. Except for two studies (41, 42), the other 17 studies did not assess blinding in their outcomes, leading to “unclear” ratings. The outcome data of all the included studies were complete, and no studies were selectively reported. In terms of other biases, one study neither reported the total time patients received treatment nor assessed the time to treatment outcome (31), and another study had unbalanced baseline data (42), resulting in a high risk of bias.

Since only the TER meets the requirement that the number of studies reporting outcomes be greater than 10, publication bias was tested only for the total effective rate. Visually, the funnel plot showed an asymmetric distribution of included studies, with four studies falling outside the pseudo 95% confidence interval (Figure 8). Furthermore, Egger’s test was conducted (Figure 9), and the results showed a total effective rate with p = 0.007, indicating the existence of significant publication bias. See Figures 8, 9 for details

Table 2 lists the quality assessment of each outcome of wet cupping and control in the meta-analysis. The GRADEpro system was used to assess the quality of evidence based on different intervention measures and outcomes. The results revealed six studies with low-quality evidence, two with very low-quality evidence, and none with high-quality or moderate-quality evidence. The majority of the studies had limitations in their experimental design, such as inadequate description of random sequence generation, allocation concealment, and blinding. Consequently, they were downgraded. In addition, the presence of publication bias in the majority of the studies contributed to their downgrading.

Neurodermatitis is one of the most prevalent and persistent pruritic dermatoses (8). Nocturnal paroxysmal pruritus, a hallmark of the condition, disrupts sleep and impairs daily functioning, imposing substantial psychological strain on patients (8, 48). Given the limitations of conventional treatments, there is an urgent need for complementary interventions. Our meta-analysis revealed that wet cupping alone did not show a higher TER than high-potency steroids, suggesting that both are equally effective. When used in combination with medication or moxibustion, wet cupping demonstrated superior efficacy compared to medication alone. Regarding secondary outcomes, wet cupping alone or in combination with other therapies can significantly reduce the recurrence rate and the incidence of adverse events.

We conducted an overall search of relevant literature databases and noted the rare use of dry cupping in neurodermatitis. Despite its potential benefits, such as regulating skin immunity and stimulating the release of inflammatory mediators (49), we hypothesized that the non-invasive nature of dry cupping may limit its effectiveness in fully stimulating the deeper layers of hypertrophic lichenified skin. The above reasons influenced our decision to exclude dry cupping from the systematic review. In contrast, wet cupping combines bloodletting and cupping as a physical stimulation procedure. This physical stimulation can cause local tissue congestion and decompression, accelerate blood circulation, and inflammatory exudate discharge (47); it also causes the release of histamine, stimulates phagocytosis, regulates the body’s immune network, and promotes the self-repair of damaged skin tissue (50). Among the studies in our systematic review, 12 used plum-blossom needles, 2 used three-edged needles, and 1 used fire needles, all of which are traditional Chinese acupuncture instruments. On the one hand, acupuncture may alleviate itching by exerting antihistamine effects, reducing itch mediator production, and modulating the brain network involved in the central transmission of itch (7, 51, 52). On the other hand, acupuncture can bolster the body’s stress response, induce denaturation of diseased tissue proteins, and activate macrophages and leukocytes to phagocytose necrotic tissue, thereby ameliorating skin lichenification (53, 54).

As an important effector of the immune system, the release and recruitment of inflammatory factors actively participate in and mediate cellular inflammatory responses (55). IL-6 has been reported to promote the activation and infiltration of inflammatory cells, exacerbating the inflammatory state of the skin and leading to keratinization and desquamation of the skin (56, 57). TNF-α is a tumor necrosis factor that induces the activation of macrophages and T cells, which, in turn, produce pro-inflammatory cytokines, etc. (58). The expression levels of IL-1β, IL-4, IL-6, and TNF-α in peripheral blood positively correlate with the severity of dermatitis (59) and may contribute to disease recurrence (60). Scholars have proposed that wet cupping reduces the aggregation of inflammatory factors by accelerating the metabolism of skin tissues and promoting the skin to repair the damaged tissues and the absorption of inflammatory substances to benefit neurodermatitis patients (50, 61). Our meta-analysis results indicated that, compared to the control group, wet cupping can significantly reduce the levels of inflammatory factors (TNF-α, IL-1β, and IL-6). However, at present, the specific mechanisms of wet cupping to reduce the serum inflammatory factor need to be further confirmed.

Our review has important limitations that should be considered with caution. First, high heterogeneity appeared in the comparison of IL-6 and in the subgroup analysis of wet cupping alone versus high-potency steroids. The variations in study designs, intervention frequencies, outcome measurement intervals, skin lesion sites, or unaccounted confounding factors across comparisons may introduce significant clinical heterogeneity. This may lead to the unreliability of our pooled analysis results. Second, all the studies included in this meta-analysis were completed in China and multicenter trials were lacking, which may affect the quality of evidence in this analysis. Third, due to the nature of cupping therapy, blinding of cupping operators and participants was not feasible, which may have affected the reporting of treatment results in favor of wet cupping. Fourth, 4 studies did not provide information on random sequence generation, 2 used incorrect randomization methods, and 17 studies did not clarify whether allocation concealment was applied. In addition, significant publication bias was evident. Consequently, the overall quality of evidence in these studies may not be sufficient to extrapolate the findings to broader contexts. Fifth, the primary outcome used in the included studies was the total effective rate, which is not an internationally recognized standard. Furthermore, trial details were not comprehensively reported under standard reporting guidelines such as CONSORT. Sixth, the variability in total treatment time and frequency of cupping in each study is often determined by the cupping practitioner’s individual experience, resulting in varying efficacy and outcomes within the same intervention framework.

Our data show that wet cupping is effective and that patients may gain additional benefits by adding wet cupping to corticosteroid or antihistamine therapy. This finding offers patients more treatment options and aids physicians make informed decisions. For patients who have long relied on medication to manage their condition and suffer from its side effects, wet cupping could be a viable alternative. Our research results suggest that wet cupping is relatively safe; however, it is important to adhere to operating specifications to prevent skin trauma or infection. We recommend the establishment of standardized guidelines for cupping therapy to facilitate its clinical promotion and application, despite being somewhat challenging.

The neuro-anxiety state plays a key role in the pathogenesis of neurodermatitis, with a demonstrated correlation between patients’ anxiety levels and the severity of pruritus (48). However, the included studies failed to incorporate relevant indicators to assess patients’ mental state and improvement. Future studies should consider incorporating the Hamilton Anxiety Scale (HAMA) and the Quality of Life Questionnaire (QOL) as outcomes to evaluate an individual’s mental status and overall quality of life. This would help define the impact of mental state on neurodermatitis patients so that appropriate interventions or research protocols can be prescribed.

Overall, the mechanism of action of wet cupping therapy on neurodermatitis remains unclear, necessitating further basic research to complement the clinical evidence. Given the limited quality of evidence in the studies reviewed, we anticipate more SRs or RCTs with reasonable designs and appropriate methods to validate our findings.