Moxibustion for the treatment of diabetic peripheral neuropathy

Supplemental Digital Content is available in the text


Introduction
A recent rapid increase in the number of diabetic patients has made diabetes a serious public-health concern. The latest edition of IDF Diabetes Atlas states that 463 million adults are currently living with diabetes worldwide and estimates that there will be 578 million adults with diabetes by 2030. [1] Peripheral neuropathy is one of the common chronic complications of diabetes mellitus, with the incidence increasing with the escalating number of diabetics. The prevalence of neuropathy in patients with diabetes is approximately 30% with up to 50% eventually developing neuropathy during their disease. [2] The most common presentation of diabetic peripheral neuropathy (DPN) is distal symmetric polyneuropathy that is characterized by pain, numbness, abnormal sensation, and weakness that affect the nerves in distal extremities [3] that can easily lead to conditions such as diabetic foot, foot ulcer, and amputation, adding further burden to the public healthcare. Controlling glycemia and cardiovascular risks are now considered to be vital in the management of DPN patients. [4] There are several measures available for addressing the painful symptoms including mecobalamin, tricyclic compounds, antioxidant alpha-lipoic acid, anticonvulsants, opiates, etc. However, few therapies are available for the improvement of painless symptoms. Moreover, the efficacy of western-medicine (WM) is poor and limited. For example, a study by Su et al [5] pointed out that a simple application of mecobalamin could not improve the ischemia and anoxemia status of nervous tissues.
For this reason, many physicians have started to explore what the traditional Chinese medicine (TCM) could offer to DPN therapy. TCM prevents and cures diseases based on its guidance that includes many therapeutic options (e.g., herbal medicine, Chinese patent medicine, acupuncture, moxibustion, manipulation, etc.). External therapies of TCM are widely used in clinical practice; in particular, acupuncture has been proven to be clinically effective and is being widely applied to treat DPN. [6,7] Besides acupuncture, moxibustion is also a representative external treatment in TCM that can regulate and harmonize qi and blood, warm meridians, and activate blood circulation, and thus that treats and prevents diseases. The clinical effectiveness of moxibustion in treating DPN has been widely recognized. Recent studies showed that moxibustion can increase serum superoxide dismutase concentration, [8] reduce free-radical production, prevent impairments of nerve tissues resulting from free-radical accumulation, and alleviate neuroinflammation possibly by inhibiting NF-kB and activating Nrf2. [9] The efficacy of moxibustion therapy has been claimed by many studies but there is still a lack of objective evaluation of its benefits in treating DPN. Therefore, the effectiveness of moxibustion therapy in DPN remains controversial and its application is limited.
We conducted this meta-analysis to assess the strength of the current evidence to support the efficacy and safety of moxibustion for the treatment of DPN that might be a novel treatment strategy for DPN.

Methods
Our systematic review was registered with PROSPERO in June 2019 (registration number CRD 42019138266). The methods of this meta-analysis were performed following the PRISMA guideline. [10] 2.

Data extraction
Two reviewers extracted the following data: study ID, sample size, average age, gender, duration of diabetes, diabetes types, TCM-syndrome types, DPN diagnostic criteria, interventions (Types and dosage of WM, a dose of moxibustion therapy), treatment duration, moxibustion treatment times, and outcome measures. Discrepancies were resolved through discussions with a third party (QN).

Risk-of-bias assessment
Based on the Cochrane Risk of Bias Tool, [12] 2 reviewers independently evaluated the methodological quality of the included studies. The following 7 elements were assessed: random-sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete-outcome data, selective reporting, and other bias. Any discrepancies were resolved by consensus.

Data analysis
Statistical analyses were performed using the RevMan 5.3 software. Dichotomous data were expressed as the risk ratio (RR), and continuous outcomes between groups as mean difference (MD), both with a 95% confidence interval (CI). We accessed heterogeneity by the x 2 test. When there was substantial heterogeneity (P < .10, I 2 >50%), we used the random effects model to analyze the data. Otherwise, a fixed-effect model was applied (i.e., when P > .1 or I 2 < 50%). [13] The sensitivity analysis and subgroup analysis would be performed to explore the possible sources of heterogeneity. Furthermore, publication bias was assessed using a funnel plot. [14] 3. Results

Description of selected studies
All 11 included RCTs were conducted in China, 1 was published in English. [22] The characteristics of the 11 studies are presented in Table 1. The sample size of individual studies varied from 39 to 132 participants. In 5 studies, enrolled patients suffered from type-2 diabetes, [15,16,19,20,23] and the other studies enrolled patients with no restriction to the type of diabetes. All included patients met the diagnostic criteria for DPN. Additionally, 4 trials reported the TCM syndromes of their participants. [19,20,22,25] Participants in both groups received hypoglycemic therapy. All treatment groups received a combination of moxibustion plus WM therapies, and control groups received only WM interventions. Six trials used mild moxibustion (i.e., the lighted Moxa sticks were pointed to the acupoint for several minutes until flush and hot feelings were detected). [15,18,20,[22][23][24] In 1 trial, [17] electronic moxibustion was used. One trial [25] did not provide a clear description of the moxibustion equipment. Other trials [16,19,21] were performed with moxibustion equipment (i.e., moxibustion-massage apparatus). WM treatments in most of the included trials used mecobalamin (8/11), other trials used vitamin B1, vitamin B12, vitamin B6, or a-lipoic acid. The studies' treatment duration ranged from 14 days to 3 months. The moxibustion treatment was administered from 14 to 90 times. All the specific moxibustion acupoints adopted in 11 studies are shown in Table 2.

Methodological quality within studies
The results of the bias-risk assessment are shown in Figures 2 and 3. Of the 11 studies, 4 RCTs adopted strict randomization and reported their methods of random-number-sequence generation in detail. Two studies were assessed as a "high risk" for the random-number-sequence generation. [17] Since insufficient information was available on the allocation concealment, blinding of participants and personnel, blinding of outcome assessment, and other bias in all 11 trials, the above items were judged to be "unclear." There was no attrition bias in the 11 studies since the outcome data were complete. None of the 11 studies was found to have reporting bias.
3.4. Meta-analysis results 3.4.1. Nerve-conduction velocity (NCV). Seven trials reported outcomes of NCV and those that did, examined different nerves. Seven trials provided the MNCV and SNCV of the median nerve, 6 trials provided the MNCV and SNCV of the peroneal nerve.
Two studies with ≥60 moxibustion treatments compared the median SNCV between treatment groups and control groups (MD = 9. 50, 95% CI 7.75-11.26, Z = 10.61, P < .00001) without heterogeneity (x 2 = 0.46, P = .5, I 2 = 0%). The other 5 trials that had less than 60 moxibustion treatments compared the median SNCV between treatment groups and control groups (MD = 5.45, 95% CI 1.34-9.57, Z = 2.60, P = .0009) with high heterogeneity (x 2 = 203.36, P < .00001, I 2 = 98%). Heterogeneity was reduced after studies by Han et al and Zheng et al were removed from the moxibustion-treatment times <60 subgroup (x 2 = 0.38, P = .83, I 2 = 0%). The heterogeneity may have been caused by the moxibustion type used; the Han et al study used electronic moxibustion and the Zheng et al study did not describe in enough detail the type of moxibustion that was employed.

Toronto clinical scoring system (TCSS). As shown in
3.4.4. Adverse events. Only 1 study mentioned adverse events, [16] however, there were no serious side effects reported for the treatment period in the treatment and control groups. The other 10 studies did not mention adverse events.

Publication bias
The outcomes of the total-effectiveness rate involving 8 studies were tested for publication bias. As indicated in Figure 10, the funnel-graph shape was visually imperfectly symmetrical indicating a potential publication bias.

Discussion
DPN is the most common complication of diabetes that may lead to the occurrence of diabetic foot ulcers and even to foot or limb amputations. At present, the main important strategy to prevent and treat DPN is to control hyperglycemia and keep the bloodglucose level stable. Other therapeutic approaches to control DPN including those evaluated in clinical trials have shown limited efficacy especially on painless symptoms. [9,26] According to the theory of TCM, DPN belongs to the "bi syndrome" of TCM and is related to the blockage of meridians. Moxibustion can warm Yang, eliminate cold, and dredge meridians, and is a representative external treatment in TCM. Based on meridians and acupoints, it can intervene in various diseases utilizing heat, light radiation, and drug effects. [27] Clinical effectiveness of moxibustion in DPN has been widely recognized. A study demonstrated that moxibustion could improve DPN-related neuroinflammation by restoring the balance between NF-kB and Nrf2 in rats and may thus be complementary to the current treatment of DPN. [9] Our study evaluated the efficacy and safety of moxibustion in treating DPN.
The study conducted a meta-analysis of 11 studies involving 927 patients. The data suggested that the efficacy of moxibustion in treating DPN was significantly better than that of the control groups. NCV is considered to be the most objective and reliable method in the diagnosis of DPN having 40% sensitivity and 100% specificity [28] and was the primary outcome criterion in our analysis; however, only 7 studies reported NCV results and the nerves examined by NCV were not identical. Therefore, it was necessary to standardize NCV evaluation to improve the reliability of data analysis. Moreover, we speculated that many factors would influence the results performed by subgroup and sensitivity analyses such as moxibustion type, the times of moxibustion treatment, the degree of peripheral neuropathy, and so on. Such factors will need to be controlled in future clinical trials involving moxibustion applied to DPN treatment. Only 2 studies reported outcomes based on the assessment of TCSS symptoms. TCSS is a relatively simple, comprehensive, and effective screening method that includes symptoms and signs. The examination is relatively objective, consistent with clinical examinations, and highly reliable. [29] Accordingly, it is essential to standardize the assessment of symptoms.
Only 1 study reported changes in blood-glucose level, [24] hence we could not evaluate the effect of moxibustion treatment on this parameter. Consequently, it remains unclear whether the improvement of outcomes such as NVC is related to an improvement of blood-glucose levels. Two studies have shown that moxibustion can improve hemorheological indexes [17,24] ; this effect may also be related to improvements in neurological function. More pharmacological and clinical studies are still required to verify the mechanism of moxibustion in treating DPN.
This meta-analysis has several limitations. All the clinical studies included in our analysis were performed in China; this may suggest that the reported positive results might have a likelihood of publication bias possibly resulting from high heterogeneity, an insufficient number of trials, and a small sample size. Further, no follow-up study and long-term effects of     moxibustion on DPN were reported. Additionally, the use of different therapeutic acupoints, treatment frequency, and moxibustion equipment would likely affect the result; we were not able to assess this due to the lack of such data in the included studies. Although the conclusion of our meta-analysis is limited, it may still provide some inspiration. For any follow-up study, establishing methodological quality is critical. For example, the studies we analyzed generally failed to ensure patient blinding; suitable devices for sham moxibustion treatment will be necessary for future studies. Moreover, attention must be paid to adverse events because moxibustion is not free of risks and generates heat, smoke, and tar that may present a risk of adverse events. The availability of a large amount of safety data will be necessary to standardize the moxibustion therapy.

Conclusions
Moxibustion therapy has been shown to have better clinical effects compared with control treatments and to be an effective and safe alternative for treating DPN patients. However, due to the poor methodological quality of the included trials, more rigorous RCTs are required to evaluate the efficacy and safety of moxibustion before definitive recommendations for the use of the procedure to treat DPN patients can be made.