Stroke, ranked as the third leading contributor to global mortality and disability, imposes substantial socioeconomic burdens worldwide (1, 2). Of particular concern, approximately 50% of stroke survivors develop into shoulder-hand syndrome (SHS), with 50–60% progressing to chronic upper extremity pain disorder and motor impairment (UE-PDMI) that persists as long-term sequelae (3–5). Stage I SHS, clinically equivalent to complex regional pain syndrome type I (CRPS-I), presents with characteristic symptoms including localized hyperalgesia, dystonic spasticity, and functional deficits in motor coordination/dexterity (6, 7). This condition poses significant clinical challenges in halting its progression from stage I to terminal stages marked by irreversible muscular atrophy and joint dysfunction (8). Current therapeutic strategies for UE-PDMI in SHS encompass non-pharmacological and pharmacological interventions (9, 10). Rehabilitation training (RT), while remaining a cornerstone non-pharmacological approach with recent technological advancements, carries an inherent limitation: the induction of procedure-related pain that frequently compromises treatment adherence (11, 12). Emerging adjunctive therapies such as robot-assisted rehabilitation and virtual reality-based programs, though mechanistically promising, face implementation barriers due to prohibitive costs and unresolved technical constraints (13). Neurostimulation, encompassing both invasive and non-invasive neuromodulation techniques, has gained traction as an alternative UE-PDMI intervention (9). Contemporary high-grade randomized controlled trials (RCTs) substantiate that neurostimulation-enhanced RT protocols yield superior functional recovery outcomes (14–16). Nevertheless, invasive modalities carry inherent risks of adverse events and cost inefficiency (17, 18). Adjunctive pharmaceutical treatments involve employing a variety of western medications (WMs) for the alleviation of pain and spasticity resulting from stroke-induced hemiplegia. Non-steroidal anti-inflammatory drugs (NSAIDs) are prescribed for pain relief and muscle relaxants are utilized to decrease muscle hypertonicity, whereas the effects of these treatments are temporary, and with the potential side effect (7). Therefore, additional research is urgently needed to examine other safe, efficacious and affordable therapeutic options for UE-PDMI due to the lack of satisfactory solutions.

Moxibustion is an essential component of traditional Chinese Medicine (TCM) that has been widely utilized for over 2,500 years for both disease prevention and treatment (19). Various types of moxibustion treatment exist, including direct and indirect approaches (19). Within the indirect moxibustion category, additional subcategories also exist, such as suspended moxibustion (SM), ginger-partitioned moxibustion (GPM), herbal-partitioned moxibustion (HPM), heat-sensitive moxibustion (HSM) and warm needling moxibustion (WNM), which, respectively, involve ignited moxa floss being suspended over acupoints at a short distance, with the use of insulating materials such as air, ginger slices, herbal preparations, and acupuncture needles. The direct moxibustion category also includes subcategories such as thunder-fire moxibustion (TFM), herbal thread moxibustion (HTM) and wheat-grain moxibustion (WGM), in which ignited moxa floss in the shape of sticks, threads or cones is placed in direct contact with acupoints and then removed. Detailed information regarding these techniques is provided in Supplementary Table 1. Moxibustion is administered for a wide range of indications, owing to its clinical effect, convenience and cost-effectiveness (20). Most recently, it has even become endorsed in guidelines by a major medical society that now recommends its use for the treatment of cancer-related fatigue (21). In the setting of UE-PDMI, moxibustion appears to be effective for both alleviating pain and improving motor function, wisth its underlying mechanism of action potentially being related to its ability to induce capillary expansion, activate blood and lymphatic circulation, alleviate circulatory congestion and induce analgesia through thermal effects, infrared radiation and the biochemical actions of its chemical constituents (19, 22).

In the course of our literature search, we have found only one existing systematic review (SR) published in a Chinese medical journal in 2021 evaluated the efficacy of moxibustion for the treatment of post-stroke SHS (23). From a methodological perspective, this study exhibits the following critical limitations. The literature search strategies for both Chinese and English databases were insufficiently detailed, potentially missing relevant studies, which raises concerns about the comprehensiveness of evidence retrieval and possible publication bias. Also, the methodology failed to explicitly state adherence to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. For instance, the PRISMA flowchart lacks specific details on the number of excluded studies and reasons for exclusion (e.g., vague criteria such as “non-randomized controlled trials” without clarifying screening procedures). Meanwhile, no adverse events (e.g., burns, allergic reactions) were reported across the included studies, potentially underestimating treatment risks and compromising safety assessments. Moreover, the sensitivity analysis merely described the removal of high-heterogeneity studies without quantitatively detailing effect size changes (e.g., omitting numerical comparisons of MD [mean difference] or CI [confidence interval] pre- and post-exclusion). Furthermore, the sources of heterogeneity (e.g., variations in intervention protocols, patient demographics) were not rigorously investigated through subgroup analyses or meta-regression. Poor methodological transparency (e.g., ambiguous exclusion criteria, incomplete reporting of search strategies) and unresolved heterogeneity compromised the credibility of the findings. Given that additional relevant studies have been newly published since this review was performed, there is a need for researchers to conduct a novel SR and meta-analysis with more rigorous methodological design in order to update the global understanding of the safety and clinical efficacy of moxibustion in the context of UE-PDMI in stage I SHS patients.

All data was extracted from previously published studies, thus neither Institutional Review Board (IRB) approval nor patient consent was required to be obtained for this analysis. This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020 statement (shown in Supplementary Table 2) (24), and was pre-registered in the International Prospective Register of Systematic Reviews (PROSPERO) with the registration identifier number CRD42024601605.

Overall, 32 RCTs involving a total of 2,814 patients were included for the final meta-analysis. The pooled data indicated that moxibustion serving as a monotherapy or as an adjuvant therapy could be beneficial for alleviating pain as measured by VAS scores and the independent performance of ADLs as measured by scores on the FMA-UE scale, MBI, BI, NIHSS and TERs for UE-PDMI recovery in patients with stage I post-stroke SHS. Furthermore, no significant differences in efficacy were observed during subgroup analysis. There was also no significant difference noted in the incidence of AEs when comparing experimental and control groups. Unfortunately, the certainty of evidence was low to very low because of poor methodological quality in trial design and substantial heterogeneity. However, this systematic review still provides an updated synthesis of the existing evidence from RCTs examining moxibustion for UE-PDMI.

Owing to its high rate of morbidity, UE-PDMI significantly jeopardizes post-stroke SHS survivors’ health (1). The potential pathophysiological mechanisms of these conditions have still yet to be fully elucidated, but may involve aberrant immune and inflammatory responses, degeneration within the central and peripheral nervous systems, vasomotor dysfunction, psychological factors and even genetically driven variations in the response to injury (10, 68). To date, the management of this condition mainly focuses on providing analgesia through pharmacological means, interventional pain management and rehabilitation, with the cumulative effect of these treatments unfortunately remaining unsatisfactory (10). Moxibustion, a cornerstone of TCM for millennia, has yet to be thoroughly studied regarding the breadth of the potential benefit that it may provide as a potential treatment for UE-PDMI (69). With moxibustion’s record of non-invasiveness, convenience of delivery, cost-effectiveness and efficacy for a broad range of related clinical indications, our research may provide novel insights into the potential clinical applications of moxibustion for this condition. A dearth of high-quality RCTs has necessitated the inclusion of RCTs involving low-quality methodology and reporting practices in this review. Nevertheless, even the less robust evidence gleaned from our review in support of moxibustion as an effective monotherapy or adjunctive therapy in treating UE-PDMI provides some propaedeutic insight that may guide and inspire future research.

Improvement was observed in the primary outcome measures of VAS pain scores and FMA-UE scores of motor function. Additionally, improvement was also observed in the secondary outcome measures of MBI, BI and NIHSS, which is especially noteworthy since a deterioration of upper extremity function tends to adversely affect the independent performance of ADLs, and consequently, any intervention that can be demonstrated to preserve this function would be invaluable for SHS patients suffering from UE-PDMI. All pooled results in our study have so far demonstrated multi-dimensional beneficial effects of moxibustion on UE-PDMI. However, as persistent pain and motor dysfunction progress into the chronic phase, psychological disorders such as elevated anxiety and depressive symptoms can often unavoidably occur (70). Our present review is unable to draw any conclusions regarding whether moxibustion may help alleviate such psychological disorders, because the currently available RCTs have yet to focus on evaluating this dimension of clinical care.

A detailed analysis of the included RCTs also reveals a great diversity of acupoints that were selected for the application of moxibustion. Half of the top 10 most frequently applied acupoints in these RCTs are located on Hand Yangming meridian, which is believed to play an essential role in activating the blood circulation and facilitating Qi circulation. This pattern of acupoint selection is consistent with TCM theory. Given that UE-PDMI is a chronically degenerative disease, most of the included RCTs favored a longer duration of intervention with more frequent treatment sessions and longer treatment courses to achieve and maintain desired clinical effects. With rising concerns about the risk of burn injury and the safety of smoke released during the process of administering moxibustion, we have paid special attention to the reporting of related AEs in these RCTs. We are pleased to note that the incidence of severe burns from burning moxa was reported to be rare, even with direct moxibustion. Additionally, the incidence of moxa smoke-related respiratory issues requiring medical aid was similarly reported to be very low. It is lamentable that the quality of moxa floss utilized, which directly correlates with the magnitude of positive clinical outcomes, was not mentioned in any of the included RCTs (71). Furthermore, because evaluating the differential effects of various forms of moxibustion delivery was not the focus of this review, RCTs involving direct comparisons of different styles of moxibustion were excluded. Some of these nuances of moxibustion delivery could potentially serve as fruitful avenues for research in future studies.

Because of obvious heterogeneities observed among the included studies, a RE model was noted to be frequently utilized. Heterogeneity remained high even with subgroup analysis taking the style of moxibustion as a stratified factor. As mentioned in the introduction part, there are seven different approaches of moxibustion included in this SR with direct and indirect patterns. Firstly, the seven distinct moxibustion techniques differ entirely in their operational procedures. Variations in moxa cone size specifications, differences in the quality of moxa wool, discrepancies in the distance from the skin during application, and diverse degrees of thermal stimulation to the local skin area collectively contribute to the high heterogeneity observed. Additional factors including the combination with acupuncture needles, various acupoints selected, and notably, certain techniques even incorporate herbal components as a conductive medium, further amplify procedural variations. These multidimensional differences in material composition, operational parameters, and adjunct therapeutic elements possibly constitute the principal sources of clinical heterogeneity in moxibustion interventions. Secondly, substantial variability exists in the demographic characteristics of enrolled patients across studies. Although all patients were diagnosed with either ischemic or hemorrhagic stroke, the mean age ranged from 46.27 to 68.40 years, representing an age span exceeding two decades. Younger patients generally exhibit superior neurorehabilitation outcomes compared to elder individuals due to age-related differences in neuroplasticity. Furthermore, marked discrepancies were observed in disease duration among studies, with mean post-stroke intervals varying from 18.2 days (earliest intervention) to 120 days (latest intervention). Earlier therapeutic intervention is clinically associated with more pronounced neurological improvement. Critically, moxibustion administration parameters demonstrated significant heterogeneity: single-session duration spanned 15 to 60 min; treatment frequency ranged from once weekly to daily applications; and total intervention cycles differed from 15 days (shortest) to 12 weeks (longest). The absence of authoritative expert guidelines or consensus has permitted this substantial divergence in core therapeutic variables, including temporal parameters, dosage intensity, and treatment scheduling, which ultimately constitutes a potential primary source of outcome heterogeneity in clinical evaluations of moxibustion efficacy for UE-PDMI.

This systematic review demonstrates notable methodological merits, foremost in advancing the current evidence base through contemporary evaluation of moxibustion’s safety-efficacy profile for UE-PDMI management, while simultaneously identifying critical knowledge gaps that necessitate prioritized investigation through hypothesis-driven clinical trials. A particular scholarly contribution lies in its development of a standardized multidimensional assessment framework incorporating validated biomarkers, functional metrics, and patient-reported outcomes – enabling systematic evaluation of moxibustion’s therapeutic benefits within neurorehabilitation contexts. The analytical approach was further strengthened through predefined subgroup analyses (stratified by intervention protocols and stroke subtypes) and methodological sensitivity analyses (assessing outcome stability across risk-of-bias tiers), which substantially strengthened the validity of key inferences. Nevertheless, several methodological constraints warrant acknowledgement. First and foremost, even with 32 RCTs eventually identified for inclusion, the methodological flaws and lack of transparency in study design were obvious, with a number of studies failing to provide important information or demonstrating ambiguities about the randomization, allocation, blinding and evaluation practices that were utilized. Only VAS pain scores and FMA-UE scores of motor function were utilized as the primary outcome measures in these RCTs included in this systematic review. Regretfully, the numerical rating scale (NRS), action research arm test (ARAT) and Wolf Motor Function Test were not applied, of which are more reliable and proper to measure the pain and motor dysfunction of upper limbs (72–74). All of these shortcomings introduce some potential bias, and the lack of high-quality RCTs unfortunately precludes our ability to draw definitive conclusions about the efficacy of moxibustion in this setting.

Secondly, we have been unable to formulate a definitive explanation for the considerable heterogeneity that has been observed to exist among the included RCTs. Subgroup analysis was initially conducted in an attempt to reduce heterogeneity by differentiating trials according to the various styles of moxibustion that they employed. Meanwhile, sensitivity analysis was also conducted in an attempt to label the potential cause(s) of the heterogeneity. Given that moxibustion is complex form of treatment, key factors may exist in the clinical nuances of how this modality is delivered, such as the quality of moxa floss, the size of moxa cones and sticks that are utilized, the combination of acupoints that are selected for treatment, and the frequency and duration of treatments, all of which varied considerably among the different RCTs. Additionally, with UE-PDMI in SHS being a comorbidity that commonly occurs after a stroke, the intricacy of its incompletely understood pathogenesis and pathophysiological mechanisms, its progression through various phases of chronicity, and the impact that the routine delivery of standard medical care may contribute to the healing process could also constitute factors that may be driving the significant heterogeneity that was observed in these studies.

Finally, our search was limited only to research articles that have been published in the English and Chinese languages, which could potentially have led to the exclusion of related literature that may have been published in other languages. All of the included RCTs were conducted in China, with the clinical rationale guiding the choice of moxibustion therapy being directly informed by TCM theory, and with the selection of acupoints being mainly based on regional proximity to the affected area and consideration of its overlapping meridian pathways. This geographic limitation may also restrict the consideration of other global perspectives on how this treatment could be effectively applied for the treatment of UE-PDMI in SHS, and may also in turn hinder the global recognition of this study’s contribution to the existing body of literature on this topic as well.

The current evidence base remains insufficiently robust to endorse moxibustion as a standard therapeutic intervention for UE-PDMI in stage I post-stroke SHS. This systematic review reveals that diverse moxibustion modalities exhibit therapeutic potential in mitigating neuropathic pain components and facilitating sensorimotor recovery, while demonstrating favorable safety profiles. Notably, clinical outcomes appear independent of moxa delivery mechanisms, suggesting comparable efficacy across administration modalities. Collectively, these findings mandate methodologically rigorous RCTs featuring prospective registration, adequately powered sample sizes, and standardized outcome metrics to achieve three critical objectives: (1) Expand the evidence base through multi-center validation studies; (2) Generate definitive safety-efficacy data through biomarker-integrated trials; (3) Establish evidence-based clinical practice guidelines to optimize moxibustion protocol standardization (including dose–response parameters, treatment frequency, and duration). Such efforts would ultimately enable precision application of this TCM modality to optimize patient-centered care delivery in this neurologically vulnerable population.