Efficacy and Safety of Acupuncture against Chemotherapy-Induced Peripheral Neuropathy: A Systematic Review and Meta-Analysis

Objective To determine the effectiveness and safety of acupuncture for chemotherapy-induced peripheral neuropathy. The review has been registered on the “PROSPERO” website; the registration number is CRD42020151654. Methods A comprehensive literature search was performed on 7 electronic databases from the time of inception to March 2020. RCTs studies on acupuncture for CIPN compared with medication or sham acupuncture were included. Statistical analysis was carried out using RevMan 5.3. Results In total, 19 RCTs covering 1174 patients were enrolled. The results showed that acupuncture significantly increased the effective rate of CIPN compared with medicine and sham acupuncture. And acupuncture had a good effect on the recovery of nerve conduction velocity and improving pain. Among the acupoints involved in the treatment of CIPN, LI4, LI11, ST36, EX10 (Bafeng), and EX-UE 9 (Baxie) were the most commonly used. Conclusion The use of acupuncture in the management of CIPN is safe and effective. The most used acupoints for CIPN are LI4, LI11, ST36, EX10 (Bafeng), and EX-UE 9 (Baxie).


Introduction
Chemotherapy-induced peripheral neuropathy (CIPN) has a severe dose-limiting side effect for cancer patients treated with taxanes, vinca alkaloids, platinum drugs, and proteasome inhibitor bortezomib [1]. It is caused by damage or impairment in the dorsal root ganglia (DRG); that is to say, most of the neuropathy is sensory neuropathy, less involving motor and autonomic nerves [2]. It can manifest as numbness, tingling, and burning pain and mostly present in a stocking-and-glove distribution which reduces patients' quality of life and leads to patients' chemotherapy dose reductions or treatment discontinuation [3][4][5]. A recent meta-analysis [6] demonstrated a CIPN prevalence of 68% in the first month after chemotherapy and 30% at 6 months or more. Every third patient (37%) with neuropathic symptoms ranked them as the most troublesome symptom [7].
Currently, the clinical prevention and treatment of CIPN has no recognized specific means. e effectiveness of drugs used to treat other types of neuropathy may not extend to CIPN, such as gabapentin [8], tricyclic antidepressant amitriptyline [9], and opioids [10]. e only successful CIPN trial [11] shows that the antidepressant duloxetine at 30 mg daily for 1 week followed by 60 mg daily for 4 more weeks produced a significant reduction in BPI-SF average pain compared with the placebo arm (mean change score, 1.06 vs. 0.34; P � 0.03). But it also has grades 2 to 3 adverse events (mainly fatigue and nausea), and 12 patients discontinued treatment due to toxicity. erefore, acupuncture, a common adjuvant therapy, has been used encouragingly to treat peripheral neuropathy associated with diabetes [12] and HIV [13] and has attracted interest. Acupuncture is widely used in cancer-related diseases such as hot flashes [14],cancer-related fatigue [15],cancer-related insomnia [16],and leukopenia [17], and it can also reduce pain in arthritis [18][19][20],sciatica [21,22], and gout [23]. Although several studies [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42] have suggested that acupuncture was effective for treating CIPN, few systematic reviews have been published. We performed a systematic review and metaanalysis to assess the strength of the current evidence to support the efficacy and safety of acupuncture and electroacupuncture to treat CIPN, which might be a complementary therapy for CIPN.

Eligibility Criteria.
e included studies met the following eligibility criteria. (1) Studies were designed up to randomized controlled trial (RCT) standards. (2) e patients who had malignant tumor received chemotherapy with peripheral neurotoxic drugs and had symptoms related to peripheral nerve injury. (3) e control group was treated with conventional methods, such as the use of neurotransmitters and antiepileptic drugs, or sham acupuncture. e exclusion criteria were as follows: (1) Studies were nonrandomized controlled trials. (2) e experimental group was treated with pharmacoacupuncture, auricular acupuncture, laser acupuncture, or other uncommon forms of acupuncture. (3) Studies described interventions combined with other TCM therapies such as Chinese herbal medicine, acupoint injection, or herbal extracts. (4) Patients with original nervous system disease or peripheral neuropathy caused by other diseases.

Data Sources and Search Strategy.
e following seven electronic databases were searched to identify eligible trials published from inception to March 15, 2020: Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, PubMed, EMBASE, Chinese Biomedical Literature Database, Chinese National Knowledge Infrastructure, and the Wanfang database. e 2 key concepts used in the search, peripheral neuropathy and acupuncture, included their synonyms. Two reviewers (JYX and ZJY) independently screened the titles and abstracts for eligibility and examined the full text of the articles. Any discrepancies were resolved by consensus or after consulting a third party (ZQ). e search keywords used were "Chemotherapy-induced peripheral neuropathy," "Peripheral neuropathy," "Neurotoxicity," "Acupuncture," "Acupoints," "Needle," as well as "random," "clinical" and there was no restriction on language. More details have been provided on the "PROSPERO" website, and the registration number is CRD42020151654.

Data Extraction.
Two reviewers (JYX and WYX) independently extracted data using an extraction sheet. e extracted data included general trial characteristics (authors and year); baseline patient and disease data (sample size, age, and gender); and interventions and outcomes. We contacted the authors via e-mail or telephone to acquire further information when the information in the literature was incomplete.

Quality Assessment and Data
Analysis. Two reviewers (JYX and XX) independently assessed the methodological quality of the RCTs using the Cochrane Collaboration Risk of Bias tool. e risk of bias was assessed according to the Cochrane Handbook, which consists of six items: random sequence generation (i.e., selection bias), allocation concealment (i.e., selection bias), blinding of participants and personnel (i.e., performance bias), blinding of outcomes assessment (i.e., detection bias), incomplete outcomes data (i.e., attrition bias), selective reporting (i.e., reporting bias), and other biases. Discrepancies in this interpretation were resolved by consensus or after discussion with a third party (ZQ). e data were analyzed using Review Manager 5.3 software (Cochrane Collaboration, Oxford, UK). e results are presented as the risk ratios (RR) or mean differences (MD) with the 95% confidence interval (95% CI). I 2 statistics were used to assess heterogeneity.

Characteristics of Included Trials. 19
RCTs met the inclusion criteria: 1 study was conducted in Germany [42], 2 in the United States [36,41], 1 in Brazil [40], and the other 16 studies in China. In total, 1174 patients (588 in the acupuncture or electroacupuncture group, 586 in the control group) participated. e characteristics of the included studies are shown in Table 1.

Risk of Bias and Methodological Quality.
e assessment of risk of bias is summarized in Figure 2. All studies were described as "randomized" and 10 studies [24,27,28,31,33,34,36,38,40,42] described appropriate methods for sequence generation. e remaining RCTs did not report specific random methods. One study [36] conducted allocation concealment.

Total Effective Rate.
Use the nimodipine scoring method to determine whether it is markedly effective, effective, or ineffective: markedly effective-the assessment of peripheral neurotoxicity is reduced by 2 levels or higher; effective-reduced by 1 level; and ineffective-the peripheral neurotoxicity is not reduced or increased; total effective rate � markedly effective rate + effective rate. Although all studies used the scale to evaluate patients, the results of some scales are expressed in the "I/II/III;/IV;" levels, which is not convenient to compare the scores as a continuous variable. erefore, the change of patients' grades is used to show whether the treatment is effective, and by comparing the effective rate, the improvement degree of the disease is reflected.    Figure 3, there is a significant difference between the experimental groups and the control groups in total effective rate (RR, 1.63; 95% CI, 1.38 to 1.93; P < 0.00001; and I 2 � 0%). In the subgroup analysis, we found that acupuncture improved the sensory function significantly (RR, 1.62; 95% CI, 1.37 to 1.91; P < 0.00001; and I 2 � 0%), while the improvement of motor function was not obvious (RR, 1.82; 95% CI, 0.76 to 4.34; P � 0.18; and I 2 � 0%).
As shown in Figure 4, 3 studies [28,33,34] compared the improvement of sensory and motor nerve disorders. e results showed that the recovery of sensory function was better than that of motor function (RR, 4.07; 95% CI, 2.41 to 6.86; P < 0.00001; and I 2 � 7%).

Adverse Events.
One trial [36] reported mild adverse events which were discomfort, minor swelling, and bruising after acupuncture needle withdrawal. ree trials reported [24,33,41] that no adverse events occurred. e remaining trials did not mention any adverse events.

Analysis of Efficacy.
is systematic review shows that the overall effectiveness of acupuncture treatment of CIPN is superior to that of neurotrophic drugs (Figure 3). Meanwhile, the recovery of sensory function is significantly better than motor function (Figure 4). It is interesting to see that many non-CIPN symptoms [26,28,31,38] caused by chemotherapy, like nausea, appetite change, and drowsiness, had also improved,suggesting that acupuncture for CIPN can impact on a wider range of symptoms. However, on the other hand, the results [26,40] show that acupuncture for mild and moderate limb numbness, pain, and other sensory disorders is significantly effective, while the curative effect for severe CIPN patients was poor. And measurable results     Evidence-Based Complementary and Alternative Medicine can only be found after a longer period of treatment (up to 10 weeks) [43]. While the mechanism of action of acupuncture is not completely understood,several animal studies [44,45] have reported that acupuncture can stimulate C nerve fibers which can be damaged from chemotherapy and enhance spinal/central GABA receptors, serotoninergic, and adrenergic neurotransmission [46]. Other hypotheses including "de qi" sensation,the principal of acupuncture,can increase blood flow in the capillaries,release local opioid peptides, stimulate specific areas of the brain [47], and participate in signal integration of central nervous system [48].

upper extremity
Test for overall effect: Z = 3.64 (P = 0.0003) Heterogeneity: chi 2 = 11.11, df = 2 (P = 0.004); I 2 =82%  e result shows that acupuncture has a good effect on the recovery of nerve conduction velocity ( Figures 5 and 6), and the recovery of motor nerve function is better than that of sensory nerve function (Figure 7), but this is contrary to the results of the above questionnaire survey (Figure 4). is may be related to the fact that the motor nerve recovery is covered up by the decrease of myodynamia due to reduced exercise. e evidence shows that long-term aerobic exercise training can prevent the onset or modify the natural history of diabetic peripheral neuropathy [49]. is conclusion has also been confirmed in an animal experiment [50]. More favorable outcomes of peripheral neuropathy and psychosocial functioning are also reported in ovarian cancer survivors who were meeting physical activity guidelines [51]. ese articles suggest that moderate exercise may play an important role in the treatment of CIPN. On the other hand, questionnaires and self-reports are the preferred measure of CIPN [52], but they can be highly susceptible to subjective factors. Significant statistical difference is found in the change of mean values of the NCI-CTCAE Scale in Alessandro et al.'s experiment [40], and some literatures [53] think that the assessment modalities for neuropathy may overestimate the presence of motor neuropathy and misdiagnose CIPN. So, current scales may not be appropriate to measure CIPN in a valid way. More objective outcome measurements and a combination of scales for the standardization of CIPN are advisable.

Analysis of FACT-NTX Score.
We can see the positive effect of acupuncture in the analysis of FACT-NTX score (Figure 8), but the results are highly heterogeneous. It may be due to the fact that FACT-NTX score is a subjective index, and results are easily affected by artificial errors. e different therapeutic effects caused by different acupoints may also be the cause of differences. We found that the acupoints in Yangming Meridian are generally used for the treatment of CIPN because "treatment for flaccidity aims at Yangming meridian" is a therapeutic principle of acupuncture in traditional Chinese medicine, which means that the acupoints on Yangming Meridian like LI4, LI11, and ST36 are often used in the treatment of muscle atrophy and peripheral neuropathy, such as CIPN, diabetic peripheral neuropathy, stroke, and myasthenia gravis. Yangming Meridian is also closely related to the gastrointestinal system, so it can explain that in the treatment of CIPN, gastrointestinal related symptoms like nausea, loss of appetite, and fatigue are also improved. Another characteristic of the treatment of CIPN is using local points on the extremes like EX10 (Bafeng) and EX-UE 9 (Baxie) to stimulate the peripheral nerves and improve the local symptoms of the limbs.

Analysis of BPI Score.
Although the result shows that acupuncture has a certain effect on improving pain, the heterogeneity of the results is high. In Rostock et al.'s study [42], there is no therapeutic advantage of electroacupuncture over an orally administered placebo control.
ey attribute it to a floor effect that their study design is based on patients with more severe CIPN symptoms and higher pain scores,so there are no large margins for many patients to improve considerably. Alessandro et al. and Yan et al.'s studies [26,40] also show that the benefits of acupuncture for pain are moderate and appear to be more significant for patients with recent neuropathy symptoms. In Greenlee et al.'s study [36],the pain symptoms become worse after the postintervention in the electroacupuncture arm,which is similarly reported in Chen et al.'s RCT [54] of electroacupuncture in reducing pancreatic cancer pain. e mechanism is not clear, but we found that both Rostock et al. and Greenlee et al.'s studies use electroacupuncture. e strong stimulation of electroacupuncture may aggravate the pain. erefore, electroacupuncture may not be as effective and safe as traditional acupuncture in CIPN.

Limitations of Study.
We only searched the studies from the English and Chinese databases, so some studies may be omitted, which may lead to the sampling bias. e sample size included in the study is small and because there is no generally agreed-upon placebo control, the control group is inconsistent, which may affect the accuracy of the effectiveness evaluation. Lack of blinding of both subjects and investigators is a common methodological flaw of the selected studies, which means positive expectation may amplify acupuncture's effect. erefore, the findings of this study should be interpreted cautiously.

Conclusion
In conclusion, our analysis indicates that the use of acupuncture in the management of CIPN is safe and effective. e most used acupoints for CIPN are LI4, LI11,  Evidence-Based Complementary and Alternative Medicine ST36, EX10 (Bafeng), and EX-UE 9 (Baxie). However, considering the methodological limitation of the studies, caution needs to be taken in applying the conclusions of this review. e existing studies lack long-term observation. Furthermore, none of the included studies compared acupuncture with duloxetine [3], a drug which is currently the most effective treatment in CIPN. us, we would recommend that future studies use this medication as a comparator and observe the long-term curative effect of acupuncture.

Data Availability
All the data were taken from the published studies.

Conflicts of Interest
e authors declare that they have no conflicts of interest regarding the publication of this paper.