Possible role of the botulinum toxin in the management of neurogenic thoracic outlet syndrome: a systematic review

Abstract

INTRODUCTION

Thoracic outlet syndrome (TOS) is related to the compression and/or the traction of the upper-limb neurovascular bundle, responsible for a chronic painful impairment. Neurogenic TOS (NTOS) is the most common manifestation. It remains a challenging diagnosis and its treatment is also difficult. Botulinum toxin (BTX) has been described to help both the diagnosis and the symptoms improvement.

EVIDENCE ACQUISITION

A systematic literature research was performed using PubMed, ScienceDirect, and Embase databases to collect studies reporting the use of BTX in NTOS management. We followed the PRISMA guidelines, and the included studies were evaluated using the GRADE approach.

EVIDENCE SYNTHESIS

We included 10 original articles representing 555 patients. Various outcomes were considered, and results varied from a study to another. Symptoms relief varied from an absence of BTX effectiveness to 84.1% of improvement; relief duration was also reported from none to 88 days. BTX injections were debatable predictors of surgical procedure successes due to low evidence. There was a huge gap between the studies concerning side-effects of the BTX procedures, from none to 100% of the patients.

CONCLUSIONS

There is no evidence for considering BTX injection as a validated tool for the management of NTOS. There might be a slight effect on symptoms, but outcomes are very variable, which prevents further interpretations. The use of BTX should be evaluated in larger prospective cohorts with more standardized outcomes.

Introduction

Thoracic outlet syndrome (TOS) is a rather rare condition,^{1, 2} which is related to the compression or sometimes to the traction of the upper-limb neurovascular bundle, responsible for a chronic painful impairment.^3-5 There are three main potential compression areas:^6-9 the interscalene triangle, delimited by the anterior and middle scalene muscles and the first rib; the costo-clavicular space, delimited posteriorly and inferiorly by the first rib, and anteriorly by the sub-clavian muscle and the inferior aspect of the clavicle; and the sub-coracoid space delimited anteriorly by the tendon of pectoralis minor inserting onto the coracoid process. Three distinct forms are usually described: neurogenic TOS (NTOS), venous TOS and arterial TOS.^{6, 10} NTOS is the most common manifestation in more than 90% of the cases.^{4, 6} It is usually due to intermittent compression of the brachial plexus at the supraclavicular scalene triangle and between the first rib and the clavicle, or at the sub-coracoid space levels.¹⁰ NTOS induces functional impairment and has social consequences.^{3, 11, 12} Patients usually describe upper-limb pain, paresthesia, weakness and strength loss.^13-15 It remains a challenging diagnosis despite helpful guidelines provided by the Consortium for Outcomes Research and Education on Thoracic Outlet Syndrome (CORE-TOS),¹⁶ because pathognomonic imaging and electrodiagnosis are still controversial.^17-20 NTOS management is also difficult as rehabilitation is usually the first step of the treatment but surgery remains sometimes necessary.^21-23 So, other strategies have been developed, such as the use of botulinum toxin (BTX).^24-27

BTX is produced by the bacteria Clostridium botulinum, and may induce muscle weakness, nervous system dysfunction, and respiratory distress in case of intoxication.²⁶ The Botulinum toxin prevents the release of acetylcholine which is necessary for muscle contraction.²⁵ So, BTX can be used in medical therapy to decrease muscle overactivity and also the resulting pain due to this overactivity.²⁸ In NTOS, muscle anesthetic blocks can help both the diagnosis and the prediction of surgery success,^{26, 29-31} so, some authors have also suggested that BTX could help clinicians for these indications.^{32, 33} For instance, BTX might help decrease the overactivity of the scalene muscles or the pectoralis minor, and decrease the compression on the neurovascular bundle.²⁶ In the same way, BTX could be of therapeutic use thanks to its prolonged duration of action that can theoretically be several months,³⁴ and possibly make physiotherapy easier. Yet, studies seem heterogenous due to their different aims, diagnosis, prediction, or treatment, but also concerning their findings. Because of this heterogeneity, we aimed to perform a systematic review to assess the interest of BTX in the management of NTOS, both in conservative and surgical strategies.

Evidence acquisition

Literature search

This systematic review was registered with PROSPERO under the registration number CRD42022375861. We searched articles in different medical databases: PubMed, ScienceDirect, and Embase in November 2022. Multiple searches were performed using the following MeSH terms: (“Thoracic outlet syndrome”) AND (“Botulinum toxin” OR “Botox” OR “BTX” OR “Dysport” OR “Xeomin”). The search was performed independently by two authors (A.F.C., G.P.) to assess titles and abstracts of potentially relevant articles, and then the full-text articles were retrieved. In case of doubt, a third assessor’s advice was asked (PD). All relevant articles were read independently in full text by the two researchers (GP, AFC) to assess if the articles met the inclusion criteria. After identification of key articles, their references and citation lists were also hand searched for further information sources.

Eligibility criteria and data extraction

The inclusion criteria were the studies concerning neurogenic thoracic outlet syndrome treated with botulinum toxin, in surgical and non-surgical studies. We excluded case reports and studies on arterial TOS or venous TOS. Only studies in English language were selected. Data of the included studies were extracted and summarized in Microsoft Excel 2013 (Microsoft Corp., Redmond, WA, USA): study design, publication year, management, and outcomes.

Quality analysis

We followed the PRISMA guidelines for this systematic review.^{35, 36} The included studies were evaluated using the GRADE approach in order to assess their quality of evidence.³⁷ This approach classifies the quality of evidence in high, moderate, low, or very low, according to the study design but also to the limitations and the biases of the study.

Evidence synthesis

Study selection

Our research found 127 results. After removing duplicates and reading abstracts, articles were assessed for full-text reading. We finally included 10 original articles representing 555 patients. A PRISMA flowchart summarized the search strategy for this systematic review (Figure 1). The articles included from 13 to 157 patients with NTOS treated with botulinum toxin injections. All studies were monocentric, and only one had a control group not treated with BTX.³⁸ In this late study, the control group received a placebo, consisting in an injection of 0.75 cc of preservative-free normal saline solution.

Figure 1.

—Flow chart of the included studies.

Demographic data and NTOS diagnosis

Finally, our review assessed 555 patients (Supplementary Digital Material 1: Supplementary Table I).^{33, 38-46} Mean age ranged from 31.4 to 41.0 years old, but 4 out of the 10 studies did not provide the age of their patients.^{33, 39-41} Only six studies out of the 10 mentioned the sex of their patients with 72% of females.^{38, 42-46} Mean symptoms evolution varied from at least 6.0 months to 51.3 months (information lacking in four studies).^{39, 40, 42, 46} Quality analysis of the included studies is reported in Supplementary Digital Material 2 (Supplementary Table II).^{33, 38-46} Nine out the 10 included studies had low quality of evidence, eight were retrospective case series^{33, 39-41, 43-46} and one a prospective study.⁴² Only one study was a randomized controlled trial with high quality of evidence.³⁸ In two studies, the diagnosis of NTOS was made thanks to clinical exams and to the use of anesthetic blocks into the scalenes, which should induce a symptom relief of at least 50%.^{33, 41} In five studies, positive diagnosis of NTOS was exclusively made on clinical exams and history.^{38-40, 42, 44} Imaging (X-rays, CT or MRI), electromyography³⁸ and duplex scanning^{40, 42} were only made to exclude differential diagnoses. On the contrary, Torriani et al. performed systematically a computed tomography (CT) and a magnetic resonance imaging (MRI) prior to BTX injection, to assess abnormalities responsible for plexus compression.⁴³ They reported abnormalities such as post-surgical changes and scarring from previous NTOS surgery, narrowing of the scalene triangle, osseous narrowing of the costoclavicular space, enlarged subclavius muscle, cervical rib, enlarged C7 transverse process or soft tissue band in the scalene triangle, in 49% of the patients. Martinez del Carmen et al. also used CT to assess anatomical particularities linked to the pectoralis minor syndrome, they also performed electromyography and doppler scanning to support the diagnosis. Fereydooni et al.⁴⁶ performed the diagnosis of NTOS according to the clinical guidelines of the Society of Vascular Surgery.¹⁰ They also systematically prescribed cervical X-rays and MRI, and electrodiagnostic testing.

Botulinum toxin injection procedures

Methods of BTX injection varied from a study to another. Indeed, in 2000, Jordan et al.⁴¹ used electromyographic identification of the anterior scalene and confirmed with fluoroscopy (instillation of iodinated contrast) to perform the BTX injection into the anterior and the middle scalene muscles, and in the trapezius. In 2007, the same team used similar technic in 68.6% of the patients but the other 31.4% had injection after confirmation by a combination of ultrasonography and electromyography.³³ There was no difference between the two methods concerning outcomes (P=0.32). In this last study, more muscles were involved because the authors also performed injections in the subclavius muscle and the pectoralis minor muscle, and also used BTX in case of dystonic neck. Two studies used CT-guided anterior scalene injections with 20 UI of BTX.^{39, 42} Four studied had an ultrasound-guided approach.^43-46 Torriani et al. included different sites of injection according to the patients’ symptoms,⁴³ whereas Donahue et al. systematically performed injections both in the anterior scalene and the pectoralis minor.⁴⁴ Martinez del Carmen et al. and Fereydooni et al. only focused on one site, the pectoralis minor and the anterior scalene, respectively.^{45, 46} The number of toxin units varied depending on the injection site and the studies.

Indication of botulinum toxin use

In these included studies, the use of BTX had different indications for the management of NTOS. Indeed, BTX was indicated for pain symptoms in the initial management of patients with NTOS to alleviate symptoms in order to facilitate physiotherapy or pending a possible surgery.^{33, 38, 41, 42} Other authors, considered BTX injection useful to confirm the diagnosis of NTOS, Torriani et al. used it to evaluate the diagnosis in case of suspected NTOS and allow symptom relief,⁴³ in the same way, Martinez del Carmen et al. proposed BTX as a diagnostic tool to confirm clinical diagnosis of pectoralis minor syndrome, and considered also BTX as a therapeutic option.⁴⁵ The BTX injections were considered by some teams as a prerequisite before surgery. In fact, Rochlin et al. used BTX in the management of NTOS before first rib resection and scalenectomy but also as a therapeutic tool in case of post-surgical symptoms recurrence.³⁹ Fereydooni et al. used it in the management strategy to propose surgical procedure if the injection provided symptoms relief.⁴⁶

Two other studies did not clearly provide their usual strategies as they proposed clinical studies on retrospective series. Lum et al. evaluated the ability to predict surgical outcomes by BTX injections prior to trans-axillary decompression,⁴⁰ and Donahue et al. compared the response to pre-surgical sonographically-guided BTX injections of the pectoralis minor and the anterior scalene to surgical outcomes in patients with NTOS.⁴⁴

Outcomes

Various outcomes were considered in the included studies Supplementary Digital Material 3 (Supplementary Table III).^{33, 38-46} In 2000, Jordan et al. reported 14 patients out of 22 (64%) who had a symptoms relief >50% (pain, numbness, and fatigue) at 1 month after the procedure, with an improvement mean duration of 88 days (from 30 to 180 days).⁴¹ In 2007, the same team found that 206 procedures out of 245 (84.1%) in 109 patients were successful on brachial pain and paresthesia with a symptoms relief duration of 4.7±1.7 months.³³

Christo et al. assessed at 1, 2 and 3 months post injection, 27 patients with the short-form of the McGill Pain questionnaire (SF-MPQ). This score is composed of several sub-scores: sensory, affective, VAS, present pain intensity and total components. Authors reported a significant improvement at the different time-points compared to pre-BTX values for sensory (P=0.02), VAS (P=0.04) and total (P=0.05) components. Yet, the percentage of improvement declined throughout the follow-up. There was no improvement concerning the affective component (P=0.52) and only a trend for the present pain intensity (P=0.06).

Torriani et al. reported their experience on 41 patients and 55 BTX procedures.⁴³ They used a binary (yes/no) self-assessment regarding improvement after the procedure. In case of improvement, patients had to rate their pain levels before and after BTX, using a VAS from 0 to 10. Thirty-eight procedures out of 55 provided an improvement (69%). Yet, VAS results were available for only 14 procedures (25%) showing a pain decrease of 4.0±1.4 points. The duration of symptoms improvement was provided for only 38% of the BTX procedures with a mean duration of 31 days.

Finlayson et al. performed a double-blind randomized controlled study to evaluate the effectiveness of BTX injection in the scalene muscles on pain (VAS), paresthesia (VAS) and function (Disability of the Arm, Shoulder, and Hand (DASH) questionnaire and SF-36 Health Survey physical and mental sum-scores).³⁸ They compared 20 patients having BTX injections to 18 patients receiving placebos at 6 weeks, 3 months, and 6 months. There was no statistically or clinically significant improvement for pain (P=0.36, 0.98, 0.42 at the different times of the follow-up). The other parameters were also not different between the two groups.

Rochlin et al. used a subjective percent of symptom improvement in 45 BTX injection procedures, without clearly defining the type of symptoms.³⁹ Success was considered if the improvement rate was superior to 50%, which was the case for 23 patients (51%). Yet, the authors did not provide any information on the duration of the symptom relief.

In the study by Lum et al., BTX blocs were successful in only 20 patients out of 44 (45%) considering a symptom relief >50% over 2 weeks after injection, moreover, they found that BTX injection prior to surgery was not predictive of symptom relief after trans-axillary decompression.⁴⁰ Yet, the nature of the symptoms was not provided in this study.

Donahue et al. assessed the relationship between responses to BTX injection prior to NTOS surgery and surgical outcomes in 157 patients.⁴⁴ They assessed the percentage of improvement of patients’ main symptoms, which were divided into four categories: A) best (>67% of improvement), B) moderate (34-66%), C) mild (1-33%) and D) unchanged symptoms. They considered responders by grouping A, B and C categories versus group D – no responders). Using this method, they reported 63% of successful procedures out of 178. They reported a significant correlation between responses to BTX injections and surgical responses (P=0.03) but did not provide the coefficient of correlation to assess the degree of association. A high proportion of patients responding to BTX responded to surgical procedures (positive predictive value of 99%). BTX injections were moderately sensitive and accurate to determine surgical responses, 66 and 67% respectively. Besides, it had a low negative predictive value of 14%. The type of symptoms considered in this study was not provided in the patients’ description.

Fereydooni et al. also used the percentage of the overall improvement of patients’ symptoms (focal tenderness at potential sites of compression, radicular pain, paresthesia, and weakness) using the same categories and grouping them as A, B and C categories versus category D.⁴⁶ According to these groups, the authors found 60 successful procedures out of 77 (77.9%). The mean quickDASH of the whole group of patients is supposed to have decreased but the value was not provided. Female sex and positive physical examination findings were predictors of symptoms relied after BTX injections (P=0.049 and 0.03, respectively). No information was provided concerning the duration of relief after the procedure.

Concerning pectoralis minor syndrome, Martinez del Carmen et al., used a yes/no symptom improvement.⁴⁵ In case of yes, patients were asked to fulfil a pain VAS and a paresthesia VAS. The authors reported 11 positive procedures out of 17 (64.7%). The mean paresthesia VAS decreased of -38.4 mm and -25.7 mm at 1 and 3 months respectively, and the mean pain VAS decreased of -40.7 mm and -19.5 mm at 1 and 3 months respectively. The average clinical effect duration was of 65.3 days (22-109 days).

Adverse events

Three studies did not provide information about the occurrence of any side-effects.^{39, 40, 46} Out of 55 procedures, Torriani et al. did not report any adverse-effects.⁴³ Christo et al. reported that the most frequent side-effect in their study was neck weakness.⁴² Yet, they did not provide its occurrence rate. In 2000, Jordan et al. found two dysphagia out of 22 procedures (complication rate of 9%) and only two dysphagia and a shoulder weakness out of 245 procedures (complication rate of 1.2%) in their study of 2007. In the study of Finlayson et al., the group which had BTX injections reported fewer adverse events than the group with placebos, 21 versus 40. It was mainly pain increase, weakness, pain at the injection site or dysphagia. One study had a 14.6% higher rate of complications (25/178 procedures) even if they were considered minor: muscle soreness, rash/itching or flu-like symptoms.⁴⁴ Martinez del Carmen reported side-effects in 100% of their 13 patients, nine had localized pain at the puncture site, three had a hematoma and one a fainting.

Discussion

In this systematic review, we aimed to assess all the clinical studies dealing with botulinum toxin (BTX) used for the management of NTOS, especially because BTX has been considered a therapeutic option in the reporting standards of the Society for Vascular Surgery for NTOS in 2016.¹⁰ Moreover, theoretically its use appears particularly attractive because BTX could allow via a chemo-denervation, a relaxation of the muscles involved in the compression of the brachial plexus.^{25, 47} Yet, we found particularly discordant results concerning the effectiveness of BTX injections, from excellent results to an absence of results, despite including studies with patients having the usual characteristics of NTOS, a female rate of 72% and an age around 40 years old (from 31.4 to 41.0).^{3, 48-50} Several reasons can be proposed to explain such results.

Firstly, it could be due to methodological issues or biases. Interestingly, the unique randomized controlled trial has negative results³⁸ whereas the others with positive results are mainly retrospective case series, excepted those by Christo et al.⁴² Furthermore, even if these studies included mainly patients without previous surgery of TOS, two of them had patients who had already been operated, and for whom surgery failed to improve their symptoms.^{39, 43} Differences could also be due to the differences in diagnostic criteria of NTOS. Indeed, the diagnosis of NTOS was completed with variations among studies, clinical criteria only or with imaging or electrodiagnosis. Yet, the Consortium for Outcomes Research and Education on Thoracic Outlet Syndrome (CORE-TOS) only retained history and clinical examination for the positive diagnosis of NTOS, complementary imaging are only considered to exclude differential diagnosis.^{16, 51} Moreover, these guidelines have been first published in 2013⁵² and then up-dated in 2021,¹⁶ but the majority of our included studies were either published before this recommendation or used retrospective cohorts prior to these guidelines. For example, Fereydooni et al. reported to have used the clinical guidelines of the Society for Vascular Surgery.⁴⁶ Yet, their inclusions started in 2011 and the cited reference has been published in 2016.¹⁰ Furthermore, two studies considered the diagnosis of NTOS thanks to doppler ultrasound,^{45, 46} even if it has been considered not useful for the diagnosis of NTOS, with the risk of considering vascular forms.^{17, 18}

Secondly, these findings may be due to a heterogeneity concerning the clinical evaluation of the patients. Indeed, the clinical parameters vary from a study to another with very few standardized outcomes. Some studies used subjective symptoms relief without using a visual analogue scale which could reliably assess pain or other symptoms. Moreover, due to the variety of the symptoms it could be more useful to describe clearly the parameters assessed rather than using “symptom relief,” which is particularly imprecise. Concerning, the efficiency of BTX injections some evaluations appeared debatable. Indeed, Fereydonni et al.⁴⁶ and Donahue et al.⁴⁴ used 4 categories: A) best (>67% of improvement), B) moderate (34-66%), C) mild (1-33%) and D) unchanged symptoms, and considered responders by grouping A), B) and C) categories. So, we can really discuss the clinical pertinence of improvements from 1 to 33%, which may be only due to a placebo effect, and are closed to no responders. Unfortunately, none of these studies provided a detailed repartition of patients per sub-groups, which might have helped us analyse these results more accurately. This point is also particularly important for the study of Donahue et al. who tried to correlate BTX injections to surgical results. Indeed, including the category C) may have induced misinterpretation concerning the ability of BTX injections to predict surgical outcomes. Especially regarding the study of Lum et al. which found the BTX injection results not predictive of the symptom relief after trans-axillary decompression.⁴⁰ So, the use of BTX as a diagnostic tool or as a pre-surgical prerequisite seems particularly debatable because of the available studies and the lack of prospective and systematic evaluation.^{39, 43, 45, 46} Furthermore, impairment is also a major concern for patients with NTOS.^{11, 13} However, Only, two studies clearly assessed this parameter, one using the DASH questionnaire and the other one, its short version, the QuickDASH. This is all the more lacking as the use of the QuickDASH is easy and has already been used routinely in the evaluation of the NTOS.^{13, 15, 50, 53-55} Concerning the pectoralis minor syndrome, the exact cause of the symptoms appearance is not clearly assessed, between structural shortness and overactivity.⁴⁵ Indeed, we can assume that BTX effect would be limited in the case of structural shortness, and rather effective in case of muscle overactivity. More generally, NTOS development due to high muscle tone should be differentiated from structural abnormalities. In fact, Clinicians should try to differentiate such entities to better define the use of BTX. High muscle tone can be diagnosed with clinical history and examination whereas structural abnormalities can be recognized based on imaging methods.

Thirdly, there were various methods of BTX injections either ultrasound or electrodiagnosis or fluoroscopy or CT or a mixture of several technics. Only one study compared fluoroscopy and electromyography versus ultrasound and electromyography; and found no difference.³³ However, the main difference may concern the quantity of BTX injected and the muscles selection. Indeed, four studies selected only anterior scalene,^{39, 40, 42, 46} and one only pectoralis minor,⁴⁵ whereas the others considered from two to five different muscular structures.^{33, 38, 41, 43, 44} Particularly, Jordan et al. used BTX for dystonic neck, meaning that their patients with NTOS had also concurrent cervical dystonia.³³ In this study, the results should be carefully interpreted because BTX has been confirmed to treat cervical dystonia effectively.⁵⁶ Concerning, the dosage, scalene could have been injected with 12 UI to 50 UI, and pectoralis minor from 20 to 50 UI of BTX according to the studies. So, these different strategic choices also show the risk of having included a very heterogeneous population regarding the origin of their NTOS.

Fourthly, the results may also be biased insofar as most of them are series from retrospective clinical cohorts managed by experts, which are common authors in 7 out of the 10 selected articles.^{33, 39-44} This may constitute a bias linked to local strategies even if these teams are well-known for their expertise. Indeed, there is a need for multicentric prospective studies to standardize the practices, to limit also the consequences of retrospective series such as missing⁴³ or not reported data.^{33, 39-41, 44, 46}

Finally, there was a huge gap between the studies concerning side-effects of the BTX procedures, from none⁴³ to 100% of the patients.⁴⁵ The complete absence of complications as well as systematic complications can both be surprising. Indeed, studies about cervical dystonia reported from 6% to 34% of complications, mainly around 30%.⁵⁷ We can suggest that there were deep differences to define what a complication was. In fact, Martinez Del Carmen et al. considered pain at the injection site in 69.2% of the cases, which could have been considered a “normal” consequence for some clinicians and possibly not reported. Standardized check-lists of searched side-effects should be provided by the authors to limit this issue, especially because 3 studies did not mention looking for adverse-effects.^{39, 40, 46}

Limitations of the study

We included all the studies which mentioned the use of BTX injections, as a diagnostic test, a pre-surgical evaluation, a specific treatment or in case of surgical failure. That leads to more heterogeneity, but it might enhance a potential specific BTX-effect, regardless of the indication. The publication bias seems rather high in our study, given that we have included only one study with negative results, especially as it is the one with the highest level of proof. So, there could be more unpublished studies with negative findings, leading to an under-estimation of the absence of effects of BTX injections in NTOS management. We could not conduct a meta-analysis because of the heterogeneity of the methods, the diversity of the outcomes, the limited quality of most of the studies, and due to the absence of control groups in 9 studies out of 10. Ultimately, the number and quality of the publications used in this review assessing BTX as a treatment for NTOS do not make it possible to reach a strong conclusion on this subject. Researchers and clinicians who will work on this subject should start their research with a high standard work plan. Yet, our work enables clinicians to have an overview of the possible role of BTX in the management of NTOS.

Conclusions

The current studies available avoid considering the use of botulinum toxin injection as a validated tool in clinical practice for the management of neurogenic outlet syndrome, especially because only one high quality study exists. There might be a slight effect on symptoms, but outcomes are very variable according to the studies, which prevents general interpretations. So, the use of botulinum toxin in neurogenic outlet syndrome should be evaluated in larger prospective cohorts with standardized outcomes to better define its place.

Supplementary Digital Material 1

Supplementary Table I

Demographic data and botulinum toxin injection procedures of the included studies.^{33, 38-46}

Supplementary Digital Material 2

Supplementary Table II

Evaluation of the level of evidence using GRADE approach.^{33, 38-46}

Supplementary Digital Material 3

Supplementary Table III

Outcomes and side-effects of the botulinum toxin injections.^{33, 38-46}