Use of botulinum toxin in musculoskeletal pain

Search strategy for literature synthesis

In this evidence-based literature review, I performed a focused systematic review of published non-randomized (prospective and retrospective cohort studies and case series) and randomized studies of botulinum toxin for arthritis and musculoskeletal pain by performing two PubMed searches: one using the terms "botulinum toxin" and "musculoskeletal pain" and another using the terms "botulinum toxin" and "arthritis". In addition to summarizing the key studies identified from this search, I also reviewed the bibliography of key reviews to identify any studies that may have been missed using the search strategy, including our recent review of the use of botulinum toxin in osteoarticular pain⁶¹. Inclusion criteria included the use of botulinum toxin, randomized or non-randomized study that included adults with musculoskeletal pain and reported clinical outcomes. Studies of myofascial pain syndromes were not included in this review, since they have been the focus of previous reviews^62,63. I summarized case reports only when they provided new evidence for a particular use of botulinum toxin. In the sections below, I summarize data that provides initial evidence for the anti-nociceptive action of botulinum toxin, followed by existing evidence related to the use of botulinum toxin in various musculoskeletal pain conditions including refractory arthritis.

Refractory osteoarticular joint pain

In this section, I summarize the data from non-randomized studies followed by data from randomized controlled trials (RCTs) for the efficacy of botulinum toxin’s efficacy in refractory shoulder joint pain (1 randomized study) and refractory knee joint pain (2 randomized studies) (Table 1).

Non-randomized studies in patients with refractory joint pain

There are three non-randomized studies that provide evidence for the efficacy of BoNT in cases of refractory joint pain^64–66. In a retrospective study, Mahowald et al. reported the results of intra-articular injections of botulinum toxin type A (IA-BoNT/A) in 15 shoulder and lower extremity joints in 11 patients experiencing refractory pain⁶⁴. All patients were receiving analgesics/NSAIDs and none had experienced relief with intra-articular corticosteroids. Patients received 25–100 units of IA-BoNT/A. There were 9 men and 2 women with an age range of 42–82 years. Joint pain was measured on a 0–10 numeric rating scale (NRS). Pain relief began within 2 weeks in most patients. A single intra-articular injection of botulinum toxin was associated with a clinically and statistically significant decrease in pain severity and improvement in function compared to baseline: lower extremity and shoulder pain decreased from 7 to 2.7 and 8.2 to 2.4, respectively; shoulder flexion improved from 68 to 113 degrees and shoulder abduction improved from 50 to 74 degrees, respectively. Timed stand tests (the time taken to stand up ten times from a sitting position) improved from 36 to 23 seconds. The duration of relief/improvement lasted 3–10 months. None of the patients reported any adverse events and no motor-sensory deficits were found on lower extremity examination after the IA-BoNT/A injection.

Another report described the long-term follow-up of these 11 patients with 15 treated joints, of which ten joints were re-injected with 30–150 units of BoNT/A intra-articularly (Botox, Allergan) at the patient’s request, when pain returned⁶⁵. Nine of the ten re-injections were associated with pain reduction, as seen with the first injection. Pain severity decreased from 6.6 (SD, 1.2) to 3.3 (SD, 2.7) on a 0–10 NRS, which was statistically significant (p=0.003). Pain relief lasted 3–17 months. None of the patients reported any local or systemic adverse events related to BoNT/A. One patient experienced increased joint swelling with no increase in joint pain 3 weeks after BoNT/A injection. Another patient experienced a continued increase in joint pain after BoNT/A injection, which was relieved with a subsequent BoNT/A injection.

In another case series, 11 adults with refractory pain underwent injection of botulinum toxin type A (25–100 units; Botox, Allergan) or type B (5,000 units; Myobloc, Solstice Neurosciences) into the sacroiliac, cervical/lumbar facet or sternoclavicular joints, C-2 roots and lumbar disc⁶⁶. This case series comprised 9 women and 2 men with a mean age of 48 years (SD, 10 years; range, 32–68 years). Median pain scores decreased significantly after BoNT injections, with a median decrease of 3 on a 0–10 NRS pain scale (range for pain NRS reduction, 0–5; p=0.008). Three patients experienced no change in pain severity after botulinum toxin injections, while eight experienced a decrease. No patients reported any increase in pain severity after BoNT injection. Pain reduction began within 3–5 days. Five patients received repeat injections. There was no evidence of decreasing efficacy of pain relief with repeated BoNT injections; in fact, the duration of pain relief increased with each successive treatment for 4 of the 5 patients with multiple treatments. All patients that experienced pain reduction with BoNT injections also noted improved function in activities of daily life and the range of motion in their joints. The median duration of pain relief with BoNT injections was 1.6 months longer than that seen with previous corticosteroid injections. There were no adverse events reported by the patients.

Refractory shoulder joint pain: 1 RCT comparing IA-BoNT/A injections with IA-placebo

Singh et al. performed a double-blind RCT of BoNT/A (Allergan, Inc.) into glenohumeral (shoulder) joints of patients with refractory, chronic shoulder joint pain due to osteoarthritis or rheumatoid arthritis of the shoulder joint, who had all failed conservative management⁶⁷ (Table 1). Patients were randomized to one of the two treatment groups - a single injection of 100 units of botulinum toxin type A (IA-BoNT/A) reconstituted in 1cc of saline plus 1 ml of 2% lidocaine (treatment group; n=21) or a single injection of IA-saline plus 1 ml of 2% lidocaine (placebo group; n=22). Patients were enrolled in the study if they had chronic knee pain of 4.5 or more on a 0–10 NRS pain scale for at least 6-months, evidence of radiographic OA, had a failed response to oral analgesics/anti-inflammatory drugs and IA-corticosteroid injections, and were not candidates for shoulder joint replacement surgery. Patients were excluded if they were currently using aminoglycoside antibiotics, curare-like agents, or other agents that might interfere with neuromuscular function; had shoulder joint malignancy, a prosthetic shoulder joint or planned shoulder joint surgery in the next 6 months; prior botulinum toxin injection into the index shoulder joint; had disorders of neuromuscular function including myasthenia gravis, Eaton-Lambert syndrome, amyotrophic lateral sclerosis; known allergy/sensitivity to study medications; a history of recent or ongoing alcohol or drug abuse, uncontrolled systemic disease; or were pregnant, breast feeding or planning pregnancy during the study period. The study had >80% power to detect 1.5 unit difference in pain NRS scale between intervention and placebo.

At a 1-month follow-up or later, if patients did not experience any reduction in pain severity (due to inefficacy of the treatment), they were given the option to drop out of the randomized phase of the study, and request the second unblinded injection with 100 units of IA-BoNT/A, which led to the beginning of an open-label phase. The published paper reported the results from the 1-month double-blind portion of the study⁶⁷.

36 patients with 43 painful shoulder joints were randomized. Mean age was >70 years, 95% were men, >85% had OA as the underlying diagnosis, mean shoulder pain duration was 8–11 years, comorbidity index was high, at least 50% had been treated each with narcotics or non-steroidal anti-inflammatory drugs (NSAIDs), and all had failed previous IA-corticosteroid injections.

Reductions in shoulder pain severity were significantly greater in the IA BoNT/A group compared to the IA placebo group at 1 month (-2.4 vs. -0.8 unit reductions NRS respectively, p=0.014; primary outcome). 61% of patients in the IA BoNT/A group experienced clinically meaningful pain relief (defined as a 30% or 2-point reduction previously^68,69) at 1 month compared with 36% in the placebo group (p=0.22).

Several secondary outcomes were significantly better in the IA-BoNT/A versus the IA-placebo group. Quality-of-life improvements were significantly greater in the BoNT/A group versus the placebo group for five of the eight short form-36 (SF-36) subscales at 1-month, namely bodily pain, physical role functioning, vitality, emotional role functioning and mental health (p-values ranging from 0.04 to 0.001). Improvements in quality of life exceeded the clinically meaningful threshold of 10-points for all 5 SF-36 subscales in IA-BoNT/A group versus as opposed to only 1 subscale in the IA-placebo group. McGill affective pain dimension improved significantly more in BoNT/A compared to placebo at 1-month (p=0.047). Other secondary measures showed a trend in improvements, but were not statistically significantly different between treatment groups. The Shoulder Pain and Disability Index (SPADI) showed a trend towards greater improvement in BoNT/A versus placebo group (p=0.0826). Treatment-failure after 1 month, defined as a drop-out from the blinded phase due to inefficacy and request for an active treatment injection, was 2.5-times higher in the placebo vs. the BoNT/A group; 45% (10/22) vs. 19% (4/21; p=0.128) respectively. 61% in BoNT/A group experienced clinically meaningful pain relief at 1 month (a 2-point or 30% reduction in pain severity on the visual analogue pain scale (VAS)) compared to 36% in the placebo group (p=0.22).

The change in pain severity was influenced by baseline SPADI disability. In those patients with a higher SPADI disability score (>61.3), the decrease in VAS pain ratings at 1 month was significantly greater in the BoNT/A group (a 3.6 reduction) compared with the placebo (0.5; p=0.015). On the other hand, in patients with lower SPADI disability score (≤61.3), no significant difference was observed in VAS pain severity reduction between BoNT/A (a 1.9 reduction) and placebo groups (a 1.6 reduction; p=0.73).

Overall adverse events were similar in IA-BoNT/A and IA-placebo groups. No significant difference in serious adverse events was noted. Common adverse events such as dry mouth, flu-like symptoms, dizziness etc. were not statistically significantly different between treatment groups in this small study, powered for efficacy, but not safety outcomes.

Refractory knee joint pain: 2 RCTs of intra-articular BoNT/A (IA-BoNT/A)

Comparison of IA-BoNT/A to IA-corticosteroid. Boon et al. compared IA-BoNT/A to IA-corticosteroids in patients with radiographic and clinical knee OA in an RCT⁷⁰ (Table 1). In a single-center, prospective double-blind RCT, patients were randomized to one of three treatments - a single injection of 100 units (2 standard doses) of IA-BoNT/A (20 patients), 200 units (20 patients) or an IA-corticosteroid injection with 40 mg of methylprednisolone acetate (20 patients), into the painful knee joint. Patients were eligible for the study if they were adults ≥40 years that had both symptomatic knee OA with knee pain severity of ≥6 on a 10-point NRS scale that interfered with function most days of the week and Kellgren grade II or III tibiofemoral knee OA identified by plain radiographs. Patients were excluded if they had Kellgren grade I or IV tibiofemoral knee OA, inflammatory arthritis (such as rheumatoid arthritis, gout or pseudogout), reconstructive surgery on the affected knee, body mass index >35 kg/m², recent IA-corticosteroid or hyaluronic acid injection (last 3-months), a clinically unstable medical or psychiatric condition or history of neuromuscular disease, aminoglycoside or curare-like agents use or severe peripheral neuropathy.

The primary outcome was VAS (0–10) pain scores at 8-weeks post-injection (rated using a 10cm line). Secondary outcomes included lower extremity pain and function assessment using the Western Ontario McMaster Universities Arthritis Index (WOMAC), quality of life assessment using the SF-36, patient global assessment and time taken for a 40-meter self-paced walk (in seconds). The study was powered to detect a 2 cm reduction in a VAS pain rating within each group compared to baseline scores. Non-responders were allowed to request a second injection at 8-weeks during face-to-face clinic visit after the blinded injection. Additional follow-up assessments were done at 12 and 26 weeks using mailed questionnaires.

Baseline characteristics were similar in the three groups. Patients had a mean age of 61–64 years, body mass index 28–31 kg/m², symptom duration ranging from 6–10 years and 50–65% were women. Patients had tried multiple modalities for knee pain relief in the past, including: 70–80% had tried IA-corticosteroid injections, 35–50% had tried IA-hyaluronic acid injections, 55–65% had tried acetaminophen, 75–90% had tried corticosteroids and 45–80% had tried physical therapy. All 60 patients (35 women and 25 men) completed the 8-week follow-up.

VAS pain scores decreased from 6.4±1.8cm to 5.4±2.3cm (p=0.15; 22% reduction) in the IA-corticosteroid group. In the 100-unit IA-BoNT/A (low dose) group, pain severity decreased significantly from 6.6±1.9cm to 4.5±2.2cm (p=0.01; 28% reduction); and in the 200 units (high-dose) IA-BoNT/A group decreased from 6.6±1.4cm to 5.9±2.4cm (p=0.15; 26% reduction). A 2cm reduction in VAS pain scores, which is considered a clinically meaningful reduction in pain^68,69, was reported by 42% in the IA-corticosteroid group, 60% in 100-unit IA-BoNT/A group and 26% in the 200-unit IA-BoNT/A group (p=0.10 for comparison between three groups; differences between 100 and 200 units not significant). Secondary outcomes including WOMAC pain scores, function and stiffness and self-paced 40 meter walk time improved significantly within each of the three groups at 8-weeks, compared to baseline (p<0.05 for each; Table 1). However, there were no significant differences for WOMAC scores between groups at 8-weeks. SF-36 scales did not improve significantly in any group, except in the mental health subscale score at 8-weeks in the IA-BoNT/A 200 unit group (Table 1). At the 8-week follow-up 65% in the IA-corticosteroid group, 75% in 100-unit IA-BoNT/A group and 70% in 200-unit IA-BoNT/A group said that they would have the treatment again. Seventeen patients withdrew from the study at the 8-week follow-up requesting re-injection and an additional 11 patients did not complete the 6-month follow-up questionnaire. For the patients remaining in the trial, the reduction in VAS pain was sustained at 12-weeks for both 100- and 200-unit IA-BoNT/A and at 12 and 26 weeks in the 100-unit IA-BoNT/A group, but not in the IA-corticosteroid group. No deaths, anaphylactic reaction or septic arthritis were reported in any group, strength testing did not reveal any significant changes at any time-point in treatment groups and none of the common adverse events were significantly different between groups in this small study, powered for efficacy, but not safety assessments.

Comparison of IA-BoNT/A to IA-placebo. In a recent review article, Mahowald et al. summarized the results of an RCT comparing a single injection of 100 units of BoNT/A reconstituted in 1 ml of saline plus 1 ml of 2% lidocaine (treatment group) to a single injection of IA-saline plus 1 ml of 2% lidocaine (placebo group) in patients with refractory knee pain due to knee OA⁷¹ (Table 1). Patients were enrolled in the study if they had chronic knee pain of 4.5 or more on the NRS pain scale for at least 3 months, evidence of radiographic OA, had received no benefit from oral analgesics/anti-inflammatory drugs, IA-corticosteroid or IA-hyaluronic acid injections, and were not candidates for joint replacement surgery. Exclusion criteria were the same as those stated in the shoulder RCT described in the section earlier⁷². We considered the patient’s report of knee pain relief 5–10 minutes after the injection and/or aspiration of joint fluid as a surrogate for an accurate intra-articular placement of the needle.

Forty-two patients were randomized to either a IA-BoNT/A or IA-placebo group. At 1-month, McGill pain scores decreased significantly in both the IA-BoNT/A (4.7, p=0.048) and IA-placebo groups (5.6, p=0.02). However, at 3 months the decrease in pain was significant only in the IA-BoNT/A (4.2, p=0.002), but not in the placebo group (4.6, p=0.09).

Due to a differential treatment effect evident on scatter plots, patients were stratified by baseline NRS pain scale scores into moderate (4.5–6.9) or severe (7 or higher) pain for exploratory analyses. In the severe knee pain group, significant changes in McGill sensory, affective and total pain scores were noted in the IA-BoNT/A, but not in the placebo group (Table 1).

In summary, three RCTs and three non-randomized studies indicate that a single intra-articular injection of botulinum toxin is associated with a clinically meaningful reduction of pain and an improvement of function in patients with refractory joint pain due to osteoarthritis or other underlying arthritic-conditions. However, larger studies are needed to assess the most effective dose, the long-term safety of intra-articular injections and to define as to which patients might be the best candidates for this treatment option.

Non-randomized studies

In an open-label case series of 14 patients with "treatment-resistant" tennis elbow, Moore and colleagues injected 20–40 units of BoNT/A into the extensor digitorum communis and found >50% pain relief in 9/14 and complete pain relief in 4/14 patients during the 6–8 month follow-up⁷³. Pain relief began by 2 weeks in 10 patients, 3 weeks in one and after 1 month in two patients.

Three RCTs comparing BoNT to placebo or surgery

Hayton et al. studied 40 patients with refractory tennis elbow pain with a duration of over 6 months, who had experienced no pain relief from ≥1 corticosteroid injections and a full course of physiotherapy⁷⁴ (Table 2). Patients were randomized to either 50 units of botulinum toxin type A (Allergan; n=19) or normal saline placebo (2 ml; n=21) injected 5 cm distal to the area of maximal tenderness at the lateral epicondyle⁷⁴ (Table 2). There were no statistically significant differences in pain between the botulinum toxin and placebo groups at 3 months after the injection (difference of 1.1 between groups, p=0.54), grip strength (difference of 0.57 kg between groups, p=0.90), or quality of life measured by the Short Form-12 physical (difference of 6.24 points between groups, p=0.16) and mental (difference of 4.26 points between groups, p=0.42) component summary scores. Twelve of the 18 patients in the BoNT group had a transient extensors lag of the long finger at 1-weeks that disappeared 3-months after the injection; none reported this in the placebo group. The 1-point difference in pain scores between the BoNT and placebo group in this study is similar to that noted in studies of joint pain^67,70,75, however, the larger standard deviation likely led to this difference being non-significant in this case.

In another study by Wong et al. 60 patients experiencing tennis elbow pain for 3 months or longer were randomized to either a single injection of 60 units of botulinum toxin A (Dysport; Ipsen) or saline placebo injections into soft tissue and muscle 1 cm from the lateral epicondyle⁷⁶ (Table 2). Patients were treatment-naive with no prior local injections. Patients were followed for 12 weeks in a double-blinded multicenter study. The mean age was 45 years, 49 were women and symptom duration was between 12 and 19 months in the two groups. Pain on a VAS scale (0–100mm) decreased significantly more in the active treatment group (65.5mm at baseline to 25.3mm at week 4 and 23.5mm at week 12) than placebo (66.2mm at baseline to 50.5mm at week 4 and 43.5mm at week 12). The differences between groups were statistically significant at both week 4 (p<0.001) and week 12 (p=0.006). Mild weakness in finger extension at 4 weeks was seen in 10 patients in the BoNT/A group versus 6 patients in the saline group. Four patients in the Botulinum group had paresis of the fingers at 4 weeks (in one patient this persisted to week 12) compared to none in the placebo group, but grip strength was similar in both groups at the two time points.

Another randomized study compared 1 to 2 injections of 30–40 units of Botulinum toxin type A (Allergan Inc.) into the wrist extensor with the surgical release of the extensor origin of the extensor carpi radialis brevis tendon⁷⁷ (Table 2). Forty patients with refractory chronic tennis elbow pain (average duration of symptoms ~10 months) were randomized, 20 to each treatment. The mean age of the patients was 43 years, average symptom duration was 11 months (range 6–48 months), and 21 were women. They found no differences between the groups with regards to pain and grip strength up to 2 years of follow-up, while minor differences were noted at shorter follow-up periods. At 1 year, 65% of patients in the Botulinum toxin group and 75% in the operative group had good to excellent results (based on a validated composite scale with pain and patient satisfaction items). Limitations of the study included the lack of description of outcomes and the lack of a priori designation of outcomes as primary versus secondary.

In summary, results from three RCTs of patients with tennis elbow indicate that botulinum toxin may be effective (one of the three RCTs were positive; two RCTs were negative but had a small sample size), in at least some patients with tennis elbow. A small sample and short follow up period are limitations of these studies. These studies differed in patient population (patients with non-refractory vs. refractory disease), duration of disease (>3 months vs. >6 months vs. 10 months), site of injection (1cm vs. 5cm from the lateral epicondyle or direct into the wrist extensor) and the dosage of preparation used (60 units of Dysport vs. 50 units of Botox vs. 30–40 unit injections of Botox) in the three RCTs respectively.

Non-randomized studies

In an open-label study of 41 patients with painful hyperactivity of the masticatory muscles, 200 units of BoNT/A (Dysport) was injected intramuscularly, with 8 conducted under electromyographic guidance)⁷⁸. The patients in this study had not gained pain relief from conservative treatment after 3 to 12 months of symptom duration. In 29 cases, only one treatment was administered. The majority of the injections were administered intraorally. Patients were observed for up to 12 months. 80% of patients reported an improvement in pain. The mean pain severity decreased from 6.4 to 3.5 on the 0–10cm VAS scale. Thirteen patients experienced a "major improvement" as evident by the disappearance of pain. Relief lasted 3–12 months during the observation period for most individuals; only 7 patients requested re-injection. One patient experienced temporary speech impairment and swallowing difficulty post-injection, which had completely reversed after 2–5 weeks.

In an open-label study, 46 patients who had experienced TMJ pain for a median duration of 96 months were injected with 50 units of botulinum toxin A in each masseter muscle and 25 units in each temporalis muscles under electromyographic guidance⁷⁹. Patients were followed up to 8 weeks. The mean age was 41 years and 39 were women. Comparing pre-injection to 8-week post-injection assessments, a significant improvement was noted in the pain VAS from a median of 8 to 5, functional disability index scores dropped from 5.3 to 3.9mm, tenderness to palpation reduced from 15.5 to 6mm and jaw opening extent improved from 29.5 to 34.5 mm (p<0.05 for assessments). No subjects reported worsening of their condition or any side effects after the injection.

In another study, 15 patients with temporomandibular disorder received 150 units of BoNT/A with 50 units in each masseter and 25 units in each temporalis muscle bilaterally⁸⁰. The mean age was 39 years and 13 were women with a mean symptom duration of 10 years. Compared to pre-injection, at 8 weeks post-injection, a significant improvement was noted in patients; mean pain VAS scores dropped from 7.3 to 4, functional disability index scores reduced from 5.5 to 3.1, tenderness to palpation scores improved from 17 to 7.6 and jaw opening extent increase from 27 to 34mm (p<0.05 for assessments).

One RCT in patients with facial pain and temporomandibular disorder

In a single-blinded randomized study, 90 patients with chronic facial pain caused by hyperactivity of the masticatory muscles, received intramuscular injections of 35 mouse units (based on the amount required for a lethal dose in mice) of botulinum toxin A (Allergan; n=60) or placebo (n=30) in both masticatory muscle⁸¹ (Table 3). The patients in the study had experienced no relief from conservative treatment after 3 to 34 months of use. A significantly greater reduction in 0–10 VAS pain scores was noted in the Botulinum toxin (a 3.2 unit decrease) compared with the placebo group (a 0.4 unit decrease) (p<0.01) was found at a follow-up conducted between 1–3 months. A greater proportion of patients in the BoNT/A group had ≥2-point improvement in VAS pain scores during the follow-up. One patient experienced temporary paralysis of facial expression muscles and swallowing difficulty post-injection, which resolved after 4 weeks. No speech impairment or systemic botulism was reported.

In summary, intramuscular injections of botulinum toxin seem to induce short-term pain relief in cases of temporomandibular disease. The evidence for this is based on non-randomized studies and one single blind study where the nature of the blinding (patient- or physician- blinded) as well as the time-point for outcome measurement were not described. The results of this study are also limited by the short follow-up duration.

Non-randomized study

Ney et al. reported the results of injection of 200–500 units of botulinum toxin A intramuscularly in up to 4–5 trigger points per each side of the back from L2 to S1 in 60 patients with chronic low back pain in an open-label prospective study⁸². The mean age was 47 years, mean disease duration was of 9 years, 18 were women and 62% had concurrent radicular pain. A positive response was defined by the occurrence of 2 out of the following 3 criteria: a ≥50% improvement in pain VAS scores, an improvement of 2 or more grades in the pain and functional subsets of the Oswestry Low Back Pain Questionnaire, and a ≥30% increase in the number of pain-free days from baseline. 58% of patients responded positively at 2 months and of these 17% still reported improvement at 4 months of follow-up. Most patients with improvements at 2 months reported that the beneficial effects weaned off by 4 months. 18 of the 19 patients that received re-injection reported beneficial response at a 2 month follow up. Two patients reported a mild, flu-like reaction lasting 3–5 days. None of the patients reported any muscle weakness.

One RCT in patients with low back pain

Foster and colleagues compared 200 units of Botulinum toxin A (Allergan) injected intramuscular paravertebrally from L1-5/L2-S1 (n=15) to placebo (n=16) in 31 patients with chronic back pain of ≥6 month duration⁸³ (Table 3). 40 units/site were injected at five lumbar paravertebral levels on the side of maximum discomfort. The mean age was 46 years, 16 were women and the mean pain duration was 6 years (with a range of 0.5–30 years). The number of individuals experiencing greater than 50% reduction in VAS pain scores from baseline was significantly higher in the botulinum toxin group compared with placebo: 73% vs. 25% at 3 weeks (p=0.012), and 60% vs. 13% at 8-weeks (p=0.009). The Oswestry Low Back Pain Questionnaire score improved significantly more patients in patients receiving Botox (67%) compared with those receiving placebo (19%, p=0.011). No patient reported worsening of pain or function after the BoNT injection, but two patients reported worsening of pain after placebo injections.

In summary, intramuscular injections of botulinum toxin in patients with chronic back pain seems to provide short-term benefits of pain reduction. This is based on one non-randomized study and one randomized study. More evidence is needed.

Non-randomized study

Placzek et al. reported a case series of 9 patients with chronic plantar fasciitis, who were injected with 200 units of botulinum toxin A (Dysport) in the plantar fascia⁸⁴. Patients were followed up to one year after the injection. Statistically significant pain relief began two weeks after the injection (from 4.2 to 1.9 on a 0–10 VAS pain scale, p=0.012) and persisted for the 52 weeks of follow up (from 4.2 to 0.4, p=0.043). Muscle weakness or systemic effects were not seen.

One RCT in patients with heel pain and plantar fasciitis

Babcock and colleagues compared 70 units of Botulinum toxin A injected into the plantar fascia at two sites per foot to a placebo in 27 patients (43 feet in total) who had experienced chronic refractory plantar fasciitis for 6 months or more, and who had failed to respond to conventional therapies except surgery or extracorporeal shock therapy⁸⁵ (Table 3). In this study, 22 feet were randomized to BoNT/A and 21 to placebo. The median age was 44 years and 18 patients (12 with bilateral and six with unilateral plantar fascia) were women. Of the sixteen bilateral patients (male and female), 12 improved on all measures in the BoNT/A-treated foot and only one improved in the placebo-treated foot. Both pain and pain relief were measured on separate 0–10cm VAS scales. Compared to the placebo group, the Botulinum toxin group had significantly improved VAS pain relief scores (4.75cm vs. 0.6cm at 3 weeks and 4.95cm vs. 1.2cm at 8 weeks, p<0.005 for both), significantly lower VAS pain scores (2.7cm vs. 4.7cm at 3 weeks and 1.6cm vs. 4.4cm at 8 weeks, p<0.005 for both), significantly better foot function as assessed by the Maryland Foot Score (100-point scale) (72 vs. 49 at 3 weeks and 81 vs. 54 at 8 weeks, p<0.001 for both) and less muscle tenderness at plantar fascia insertion as assessed by a pressure algometry response (2.7 vs. 1.8 at 3 weeks and 2.8 vs. 1.8 at 8 weeks, p<0.003 for both). No complications were reported.