Multidimensional Effectiveness of Botulinum Toxin in Neuropathic Pain: A Systematic Review of Randomized Clinical Trials

Although botulinum toxin (BoNT) has been suggested as a treatment to counter neuropathic pain, no previous systematic reviews investigated the multidimensional effects of BoNT on pain relief and Health-Related Quality of Life (HR-QoL). The aim of this systematic review is to summarize the current evidence on the effectiveness of BoNT treatment for neuropathic pain, and to characterize its multidimensional effectiveness in order to guide physicians in clinical practice. Five databases were systematically searched up to 4 April 2022, to identify randomized controlled trials satisfying the following criteria: adults suffering from neuropathic pain, BoNT administration, any comparator, multidimensional assessment of pain as primary outcome, HR-QoL, physical function, anxiety and depression, and sleep quality as secondary outcomes. Twelve studies were included. The multidimensional pain scales used were short-form McGill Pain Questionnaire, Neuropathic pain scale, Neuropathic Pain Symptom Inventory, International SCI Pain Basic Data Set, West Haven-Yale Multidimensional Pain Inventory, Brief Pain Inventory, and Douleur Neuropathique 4. These scales highlighted the positive effects of BoNT administration. According to the Jadad scale, all the RCTs included were high-quality studies. BoNT administration might be effectively introduced in the comprehensive management of neuropathic pain. Further research should focus on optimal and cost-effective therapeutic protocols.


Introduction
Pain is currently defined by the International Association for the Study of Pain (IASP) as 'an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage' [1]. Among the different pain types, neuropathic pain is characterized by increased pain sensitivity and/or spontaneous pain and is defined by the presence of neuropathy, a lesion or disease affecting the somatosensory nervous system [1]. It is currently considered a challenge in the clinical setting due to its chronic course and poor responsiveness to medications [2][3][4][5]. In further detail, the recent systematic review by van Hecke et al. [6] reported that approximately 6.9-10% of the European population suffer from neuropathic pain, with detrimental consequences in terms of physical and psychosocial wellbeing, health-related quality of life (HR-QoL) and economic burden [3,7,8].
Pathophysiology of neuropathic pain has been widely investigated and the present evidence underlines the role of independent mechanisms triggered by various damages to an afferent pathway [5,[9][10][11][12]. However, the exact pathophysiological mechanisms underpinning neuropathic pain are far from being fully understood [13]. In this context, the downregulation of sodium channels [14], the dysregulation of Transient Receptor Potential Vanilloid 1 (TRPV1) and Transient Receptor Potential Member 8 (TRPM8) receptors have been proposed to have a role in this complex framework [15][16][17]. Interestingly, it has recently been highlighted that abnormal sensory messages characterizing neuropathic pain might stimulate the cortex, promoting the excitation of neurons in the limbic areas related to anxiety, depression, and sleep problems, frequently accompanying neuropathic pain [5]. A deeper assessment of these pathological mechanisms might play a key role in optimizing a multidimensional treatment, selecting a precise pathophysiological pathway [3,18]. In this complex framework, multimodal therapeutic interventions targeting specific structures involved in the neuropathic pain circuits are crucial to promoting the optimal response in pain relief [5]. Furthermore, several pharmacological and non-pharmacological approaches have been proposed for the complex management of neuropathic pain and growing evidence recommends a comprehensive patient-centered approach in order to improve pain management and minimize the side effects of single therapies [3,[19][20][21][22].
In the last decade, botulinum toxin (BoNT) has been proposed as a therapeutic option to treat neuropathic pain [23], although the antinociceptive effects of BoNT have been widely ascribed to the muscle relaxation effects alone [24,25]. However, several studies reported positive results of BoNT treatment in the management of neuropathic pain [26][27][28]. Park et al. [28], in a neuropathic pain animal model, demonstrated the dissociation between the duration of muscle relaxation and pain relief after BoNT treatment, suggesting a pure antinociceptive role.
Despite these findings, the recent GRADE classification by Finnerup et al. [29] reported BoNT as a third-line pharmacological treatment in general neuropathic pain. They suggested gabapentin, pregabalin, SNRIs (duloxetine/venlafaxine), and tricyclic antidepressants as first-line treatments, followed by capsaicin patches, lidocaine patches and tramadol. Similarly, the guidelines for neuropathic pain published by Moisset et al. [30] in 2020 recommended the use of BoNT as second-line therapy for peripheral neuropathic pain, while lidocaine plasters or Transcutaneous Electrical Nerve Stimulation (TENS) therapy were the first-line interventions. On the other hand, SNRI drugs, gabapentin, or tricyclic antidepressant were recommended for the treatment of central neuropathic pain, while pregabalin, tramadol, or combination therapy were recommended as second-line therapy.
Although BoNT has been suggested as an effective treatment to counter neuropathic pain [29][30][31][32][33][34][35], evidence in the literature is mainly focused on the unidimensional evaluation of pain, with different systematic reviews assessing the Visual Analogue Scale (VAS) or Numerical Rating Scale (NRS). Conversely, given the psychosocial and functional burden of neuropathic pain, a multidimensional assessment of this condition is needed, in order to promote a patient-centered approach.
However, to the best of our knowledge, no previous systematic reviews investigated the multidimensional effectiveness of BoNT on pain relief and quality of life in patients suffering from neuropathic pain.
Therefore, this systematic review of randomized controlled trials (RCTs) aimed at summarizing the current evidence on the efficacy of BoNT treatment for neuropathic pain, characterizing the multidimensional effectiveness of BoNT related to different neuropathic pain etiologies to improve the complex management of this burdensome condition.

Results
The search strategy performed on 4 April 2022 identified 1688 records from the five databases and six records from the reference lists of the included studies. Figure 1 shows the PRISMA flow diagram of the search process. After duplication removal, 1269 studies were assessed for eligibility and screened for title and abstract. After the exclusion of 1187 records, 82 full-text records were assessed for eligibility. Due to inconsistency with the eligibility criteria, 70 articles were excluded (36 were not RCTs, 3 studies involved animals, 4 were in a language other than English, 1 was retired for plagiarism, 11 did not assess patients with neuropathic pain, 1 was an ongoing trial, 5 were congress abstracts, 5 were registered protocols not published, and 4 did not assess appropriate outcomes). The Supplementary Table S2 shows the lists of excluded studies assessed in full text and the reasons for exclusions. As a result, 12 studies were included in the present work [26,[36][37][38][39][40][41][42][43][44][45][46]. The studies included in this systematic review were published between 2006 [46] and 2020 [43]. Among these, two studies were conducted in the USA [37,46]; three studies were conducted in Iran [39,42,43]; one study was conducted in Taiwan [45]; one was conducted in Greece [36]; one was conducted in Canada [38]; one was conducted in South Korea [40]; one was conducted in France [41]; and one was conducted in China [44]. The remaining study was an international collaboration (France and Brazil, n = 1 [26]). A total of 522 subjects were included in the present systematic review (90 with postherpetic neuropathic pain [36,44], 48 with spinal cord injuries [37,40], 29 with posttraumatic/postoperative nerve lesion or post-herpetic neuropathy [41], 66 with peripheral nerve lesion [26], 231 with diabetic neuropathy [39,42,43,45], 38 with thoracic outlet syndrome [38], and 20 with carpal tunnel syndrome [46]). The mean age of the subjects included ranged from 36.8 ± 8.9 years [38] to 77.5 ± 8.2 years [36], while 230 patients were males and 272 females. However, it should be noted that Breuer et al. [46] did not report the mean age and sex of the study participants.
The time of follow-up varied somewhat among the studies included with five studies reporting a total duration of 24 weeks [26,36,38,41,45], one study of 20 weeks [37], two studies of 12 weeks [42,44], one study of 13 weeks [46], one study of 8 weeks [40], one study of 4 weeks [43], and one study of 3 weeks [39]. The sample characterization of each study included has been summarized in detail in Table 1.
Two studies assessed patients with postherpetic neuropathy with an amount of BoNT-A in the study protocols ranging between 100 and 200 units [36,44]. Both study protocols assessed the effects of BoNT-A injected subcutaneously in the painful area. In further detail, one of the two administration protocols was characterized by a chessboard distribution of 40 different injections, with a minimum distance of 1 cm between injection sites [36]. The other study assessed injections every 1.0-2.0 cm radius of skin [44].
Only one study protocol has been proposed in patients with peripheral nerve lesions and was characterized by two subcutaneous injections of up to 300 units of Onabotulinum-toxinA performed after 12 weeks [26]. The maximum number of injections was 60, all performed at a distance of 1.5-2 cm [26].
Similarly, the study by Ranoux et al., assessed the effects of subcutaneous injections (maximum 200 units) of OnabotulinumtoxinA in patients with post-traumatic/postoperative nerve lesion or postherpetic neuropathy [41]. The authors performed up to 40 different subcutaneous injections at a distance of 1.5 cm [41].
One study assessed the effects of 75 units of OnabotulinumtoxinA in the anterior and middle scalene muscles in patients with neuropathic pain due to thoracic outlet syndrome [38]. The intramuscular injections were performed under EMG guidance [38].
The effects of BoNT-A in patients with diabetic neuropathy have been studied by four different RCTs with different injection protocols. All the studies included administered BoNT-A subcutaneously. One study performed an injection of 50 units of Onabotulinum-toxinA [45] while two studies injected 100 units of AbobotulinumtoxinA [39,42]. BoNT-A was administered in the sole, dorsum or entire surface of the foot. The injections number ranged between 12-20 sites per patient [39,42,43,45].
In patients with neuropathic pain due to Spinal Cord Injury (SCI), two studies assessed different protocols of subcutaneous administration of BoNT-A in the painful area. The amount of BoNT-A ranged between 200 and 400 units [37,40]. One study performed 5 units per injection site [37] and the other did not report the single injection amount of BoNT-A [40].
Lastly, one study injected BoNT-B in patients with carpal tunnel syndrome [46], dividing 2500 units of RimabotulinumtoxinB into three intramuscular injections (one for each muscle), under EMG guidance to identify opponens digiti minimi and flexor digiti minimi, and anatomically locating palmaris brevis muscle [46].
All the BoNT administration protocols are summarized in detail in Table 2.
Interestingly, Chun et al. [37] assessed pain intensity with the International SCI Pain Basic Data Set. The authors reported that 33% of patients assessed showed a significant change in pain intensity at 8 and 12 weeks, 50% showed a decreased pain interference with daily activities at 2 and 4 weeks, 50% reported a reduced pain interference with mood at 2 weeks, 33% at 4 and 8 weeks, and 50% reported a reduced pain interference with sleep at 2 and 4 weeks, 17% at 8 and 12 weeks. On the other hand, it should be noted that the statistical analysis was based on a descriptive approach [37].
In contrast, Han et al. [40] investigated the effectiveness of BoNT-A administration with the short-form McGill Pain Questionnaire, reporting significant differences between groups at 4 weeks (p < 0.05) and 8 weeks (p < 0.05) [40]. The Douleur Neuropathique 4 question scale has been used to assess pain in patients with diabetic neuropathy, showing significant improvement in electric shocks (p = 0.01), burning (p < 0.01), pins and needles (p = 0.03) and brushing (p < 0.001) subitems [39].
Lastly, Breuer et al. [46] assessed pain intensity with the West Haven-Yale Multidimensional Pain Inventory, highlighting improvements in quality-of-life indicators, reaching significance in some of the different time points assessed (p < 0.05). However, no significant differences between groups were reported (p = NS) [46].

Secondary Outcomes
The secondary outcomes assessed in the present systematic review were HR-QoL, physical function, anxiety and depression, and sleep quality.
In particular, HR-QoL has been assessed in six studies [26,37,38,40,42,45]. SF-36 has been assessed in three RCTs [38,42,45] reporting controversial results. In further detail, in patients with diabetic neuropathy, the RCT by Salehi et al. [42] reported significant improvement of SF-36 (p = 0.007) [42], while Yuan et al. [45] did not demonstrate significant differences between groups (p = NS) [45]. On the other hand, the study by Finlayson et al. [38] did not report significant improvement in SF-36 (p = NS) after BoNT-A treatment [38]. The World Health Organization Quality of Life questionnaire (WHOQOL-BREF) has been proposed by Han et al. [40] to assess pain relief in patients with SCI undergoing BoNT-A treatment; however, no significant differences were reported in the four domains of WHOQOL-BREF after the BoNT-A intervention [40]. On the other hand, Attal et al. [26] assessed EQ5D VAS scale, without underlining significant differences between groups (p = NS). Lastly, Chun et al. [37] reported at least moderate improvements in QoL in 33% of patients assessed at 2, 4, and 12 weeks, 17% at 8 weeks. Unfortunately, descriptive statistics was performed, without assessing the significance of the reported results [37]. Interestingly, physical function was assessed by Finlayson et al. [38] by the Disabilities of the Arm, Shoulder, and Hand scale; however, the authors did not report significant improvement (p = NS) after BoNT-A treatment [38].
On the other hand, anxiety and depression have been assessed through the Hospital Anxiety and Depression Scale (HADS) in two studies [26,41]: there were no significant differences between the BoNT-A and the placebo group (p = NS) in both studies [26,41].
Sleep quality has been specifically studied in six RCTs [26,36,42,[44][45][46] assessing patients with herpetic neuropathy [36,44], peripheral nerve lesions [26], diabetic neuropathy [42,45] and carpal tunnel syndrome [46]. Thus, a high heterogeneity of the outcome measures was reported. The Pittsburgh Sleep Quality Index (PSQI) has been used in two studies [42,45] performed on patients with diabetic neuropathy. In further detail, the study by Salehi et al. [42] reported significant differences in the PSQI after the BoNT-A intervention compared to the placebo group [42]. On the contrary, in the RCT by Yuan et al. [45], the difference in the improvement in sleep quality between the BoNT-A group and the placebo group reached significance (1.72 ± 1.82 vs. −0.11 ± 2.78, p < 0.05) exclusively at 4 weeks after intervention. [45]. The Sleep Problem Index has been used by Attal et al. [26], reporting significant differences (Sleep Problem Index I, six items: 43.9 ± 21.4 vs. 40.6 ± 20.6; p = 0.02; Sleep Problem Index I, nine items: 45.3 ± 19.3 vs. 41.7 ± 19.6; p = 0.03) in the intergroup analysis [26]. Similarly, the study by Apalla et al. [36] assessed sleep quality with a five-item questionnaire in 14 patients with post-herpetic neuropathy treated with BoNT-A, showing significant improvements at week 2 (p < 0.001) and week 4 (p < 0.001), compared to placebo [36]. Sleep time has been also assessed by Xiao et al. [44], showing a significant improvement at day 7 and after 3 months from the BoNT-A treatment (p < 0.01) in patients with postherpetic neuropathy compared with lidocaine and placebo groups [44].
Lastly, the RCT by Breuer et al. [46] assessed sleep interference by pain in patients with carpal tunnel syndrome, reporting significant improvements (p < 0.05) in some of the time points assessed. On the other hand, the authors did not find a statistically significant difference between groups (p = NS) [46]. Table 1 reported further detail of the main results of the RTCs included in the present review.

Discussion
To date, there is a lack of consensus about the multidimensional effectiveness of BoNT in neuropathic pain and the optimal BoNT administration protocols are still debated [49][50][51][52].
Our findings showed a significant effect of BoNT administration in patients suffering from neuropathic pain due to postherpetic neuralgia [36,44], SCI [37,40], peripheral nerve lesion [26], diabetic neuropathy [39,42,43,45], post-traumatic/postoperative neuropathies [41], and carpal tunnel syndrome [46]. Our study results are in accordance with other evidence reporting positive effects of BoNT on pain management of non-specific neuropathic pain [51][52][53]. On the contrary, the etiological cause of neuropathic pain seems strictly related to the treatment effectiveness. In particular, Finlayson et al. [38] did not show a statistically significant improvement in patients with thoracic outlet syndrome pain [38]. Accordingly, Breuer et al. [46] did not reveal any significant differences between BoNT-B administration and placebo group in carpal tunnel syndrome, suggesting that BoNT might not provide additional benefits in the management of neuropathic pain with nerve compression etiology [38,46]. In contrast, most of the studies included highlighted positive results in the multidimensional management of neuropathic pain in several pathological conditions [26,36,37,[39][40][41][42][43][44][45]. However, it should be noted that all these studies [26,36,37,[39][40][41][42][43][44][45] assessed the effectiveness of BoNT-A administration, while the RCT by Breuer et al. [46] was the only study that assessed BoNT-B; therefore, the role of BoNT-B in neuropathic pain management is far from being fully characterized.
As a result, clinicians should be aware of the evidence supporting BoNT use in specific conditions and the therapeutic intervention should be based on a precise diagnosis in order to select the patients more suitable to achieve better pain relief. Interestingly, our data showed positive long terms results of BoNT compared to lidocaine injections [44] in patients with post-herpetic neuropathy. These controversial data might be related to the characteristics of neuropathic pain and the BoNT administration protocols that were often heterogeneous in the studies included in the present review [26,[36][37][38][39][40][41][42][43][44][45][46].
In this scenario, the current literature underlines a large gap of knowledge regarding the optimal BoNT therapeutic strategy, and this might be related to the lack of standardized BoNT administration protocols and injective techniques [54][55][56]. On the other hand, the French Recommendation for Neuropathic Pain of 2020 [30] provided a general indication of dosage from 50 to 300 units (onabotulinumtoxinA) every 3 months, without fully characterizing the intervention protocols or without suggesting any differences based on patients' characteristics. Our findings showed that although the maximum amount of BoNT-A injection might reach 400 units [37] and BoNT-B might reach 2500 units [46], specific subgroup analysis based on neuropathic pain should be considered and a wide difference in the dosage injected based on patient characteristics has been reported [26,[36][37][38][39][40][41][42][43][44][45]. Similarly, in the past few years, different narrative and systematic reviews assessed the effects of BoNT administration characterizing patients with different types of neuropathic pain [49,51,52].
In particular, the systematic review by Hary et al. [57] assessed the effectiveness of BoNT-A administration in terms of pain relief in patients with peripheral neuropathic pain [57]. The authors reported significant effects in VAS scores, reporting better results in patients with diabetic polyneuropathy compared to patients with postherpetic, posttraumatic, or postsurgical neuralgia at 1 and 3 months post injection [57]. However, the authors mainly focused on unidimensional pain assessment and sleep improvement [57].
On the other hand, it should be noted that unidimensional scales lack the ability to characterize pain as a complex personal experience: these measurements heavily weight not only patient treatment satisfaction but also physician decision making [58]. In this context, multidimensional pain scales might better characterize pain intensity, nature, and location, and its consequences in function or mood, producing a quantitative description aiming at becoming the preferential assessment in a holistic approach [58]. Therefore, to the best of our knowledge, the present work represents the first systematic review of RCTs summarizing the current evidence on specific BoNT administration protocols providing data about the multidimensional effectiveness based on etiological cause of neuropathic pain to guide physicians in effective and safe therapeutic interventions in clinical practice.
Moreover, the large heterogeneity of administration protocols raises questions about the need to identify the lowest effective dose, not only to minimize the risk of adverse events but also from a cost effectiveness standpoint [59][60][61]. In particular, given the high prevalence of neuropathic pain and the strictly related sanitary costs [6,62], cost-effective therapies are mandatory in large-scale interventions aiming at improving quality of life and well-being of these patients. None of the studies included in this systematic review provided a precise cost analysis to better address the critical issue of the sanitary costs of pain management [26,[36][37][38][39][40][41][42][43][44][45][46]. Therefore, future research is needed to better address this emerging issue in the clinical scenario.
Although unidimensional assessment has been proposed in the current literature to provide relevant quantitative data about BoNT efficacy [3,8,63], to date, clear evidence in multidimensional assessment for neuropathic pain is still lacking. In the present work, we sought to assess the effects of BoNT-induced pain relief in HR-QoL of patients with neuropathic pain. Unfortunately, our results revealed conflicting evidence regarding supporting BoNT efficacy on the overall well-being of patients suffering from neuropathic pain [26,[36][37][38][39][40][41][42][43][44][45][46]. Nevertheless, it should be noted that recent research is now focused on a multidisciplinary framework with growing evidence emphasizing the need for comprehen-sive and synergic treatments to improve outcomes of patients with chronic neuropathic pain [3,20,64,65].
In this complex scenario, several therapeutic approaches have been investigated and might be proposed to improve pain management in patients with chronic neuropathic pain, including mini-invasive interventions [22,[66][67][68], pharmacological drugs [67,69,70], nutraceuticals [71,72], physical exercise [21,73,74] and instrumental rehabilitative techniques [20,75]. In addition, recent advances in understanding the pathophysiological mechanisms underpinning pain chronification reveals that specific peripheral [11,12,76] and central circuits might be involved [12,77]. In further detail, it has been reported that after peripheral nerve damage, sodium channels increase in quantity in both the involved fibers and surrounding ones, which might lower the action potential threshold of the stimulus [14]. Hence, pain in the absence of an external stimulus might be due to an ectopic signal generated along this pathway [78][79][80]. In the hyperalgesia state, some receptors like TRPV1, involved in the noxious heat pathway [15], and the receptor TRPM8 involved in the cold pathway are upregulated [16,17]. Furthermore, it has been proposed that central sensitization mechanisms might affect continuous discharge of peripheral afferent fibers in the dorsal horn of the spinal cord inducing structural modifications in postsynaptic neurons [9,11]. Other contributors to pain hypersensitivity after a nerve lesion are inflammation, loss of inhibitory GABAergic interneurons in the spinal horn and enhanced sympathetic activity [9]. Even if the precise mechanism of action of BoNT is far from being understood in detail [77], recent research suggested a possible role in the nociceptive peripheral pathway, inflammation and even in central activities related to retrograde axonal transport to the spinal cord [81]. BoNT might implement HR-QoL in patients affected by neuropathic pain and might be considered as a part of a comprehensive management strategy including both pharmacological and non-pharmacological approaches [82,83]. However, to date, the effects of combined interventions to treat neuropathic pain have not been studied in the current literature; indeed, the role of BoNT injections in integrated multitarget interventions still remains unknown.
Taken together, the findings of this systematic review of RCTs might improve the knowledge about the possible role of BoNT treatment in chronic neuropathic pain [84]. In addition, the data reported by the RCTs included in the present review might support previous evidence suggesting positive effects of BoNT in patients with neuropathic pain [54][55][56], underlining the effectiveness of specific administration protocols tailored to patient characteristics.
Despite these considerations, we are aware that the present systematic review is not free from limitations. In particular, the lack of meta-analysis represents the main limitation of the present work. Unfortunately, the large heterogeneity of participants, intervention and outcomes assessed did not allow performance of a quantitative analysis, in accordance with the Cochrane Handbook for Systematic Review of Intervention (Ver, 6.1, 2020) [85]. Moreover, our search strategy might not include all records in these field and other sources have been searched in accordance with the PRISMA 2020 guidelines [86] in order to cover the relevant literature related to this topic. Lastly, 5 out of 12 studies included were directly supplied by pharmaceutical companies distributing the BoNT. Therefore, potential conflicts of interests in the studies included should be considered before making strong conclusions.

Conclusions
To date, the mechanisms underpinning the therapeutic role of BoNT in neuropathic pain are not completely understood, but the RCTs included in the present systematic review showed promising results in terms of pain relief, suggesting that BoNT-A might effectively improve symptoms in patients with neuropathic pain.
Although our findings provided evidence about the current BoNT protocols for specific neuropathic pain treatment, this systematic review of RCTs underlined the need for highquality studies to better elucidate the optimal and cost-effective therapeutic strategies of BoNT administration. Therefore, further evidence with standardized BoNT protocols in deeply characterized populations is needed to provide strong conclusions aiming to guide clinicians to implement precise and tailored treatment to improve the management of neuropathic pain.

Registration
This systematic review of randomized controlled trials (RCTs) was performed ethically in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement [86]. The study protocol was realized before study initiation and submitted to PROSPERO (https://www.crd.york.ac.uk/prospero; accessed on 4 April 2022) with registration number CRD42022299703.

Search Strategy
We systematically searched PubMed/Medline, Scopus, Cochrane Central Register of Controlled Trials (CENTRAL), Physiotherapy Evidence Database (PEDro), and Web of Science for RCTs published up to 9 December 2021. Two investigators independently searched the databases. The search strategy is reported in Supplementary Table S1.

Selection Criteria
In accordance with the PICO model [87], we considered eligible RCTs satisfying the following criteria: 1.
(O) Outcome: the primary outcome was self-reported pain relief in terms of multidimensional pain scales. The secondary outcomes were HR-QoL, physical function, anxiety and depression, and sleep quality.
We included RCTs published in peer-reviewed international journals in the English language. The exclusion criteria were the following: (i) studies involving animals; (ii) language other than English; (iii) participants with pregnancy; (iv) conference abstracts.
After duplication removal, two investigators independently reviewed the title and abstracts of retrieved articles to choose relevant articles. Any discordance was resolved by collegial discussion. If consensus was not achieved, a third reviewer was asked.

Data Extraction and Synthesis
All data were assessed and extracted by two authors independently from full-text documents into Microsoft Excel (IBM, New York, NY, USA). Missing data was directly requested from corresponding authors. Any disagreement between the two reviewers was solved by collegial discussion among the authors. In case of disagreement, a third author was asked.
The data extracted were summarized in tables. Subgroup analysis was performed based on neuropathic pain characteristics and by the BoNT administration modalities.

Quality Assessment and Risk of Bias
The quality of the studies included was assessed independently by two of the authors, according to the Jadad scale [47]. Discordances were solved by discussion between the authors or by asking a third reviewer. A Jadad score between 3 to 5 points was considered high quality.
The risk of bias was assessed by the Cochrane risk-of-bias tool for randomized trials (RoBv.2) [48]. Bias was classified as low, high, or unclear based on the item of RoBv.2. In particular, the domains assessed by RoBv.2 were: (i) random process; (ii) deviation from the intended interventions; (iii) missing outcome data; (iv) measurement of the outcome; and (v) selection of the reported result.  Funding: This research received no external funding.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board.

Informed Consent Statement: Not applicable.
Data Availability Statement: The datasets generated during the current study are available from the corresponding author on reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.