Botulinum toxin as a potential adjunct therapy in stiff person syndrome spectrum disorders

Abstract

Background:

Stiff person syndrome spectrum disorders (SPSD) are a disabling group of immune-mediated disorders that most commonly cause progressive rigidity and painful spasms. Botulinum neurotoxin (BoNT) has anecdotally improved SPSD symptoms, though evidence regarding treatment strategy and clinical efficacy is scarce.

Objectives:

To characterize the location, frequency, dosage, and clinical response to BoNT injections in patients with SPSD.

Design:

We conducted a retrospective observational cohort study.

Methods:

SPSD patients who received BoNT treatments between August 2018 and February 2024 were included. Detailed information about the injections (formulation, dose, muscle), subjective patient-reported response, and concurrent therapies was recorded and compared between the first, third, and eighth BoNT visits.

Results:

Thirty-seven SPSD patients were included. The majority had classic SPS (83.8%), were female (67.7%), white (78.4%), and on immune therapies (70.3%). The paraspinal muscles, hip flexors, distal leg flexors, and shoulder girdle muscles were most frequently injected. Supramaximal total doses up to 980 units of BoNT were used safely. The most common side effect was transient worsening of pain/spasms, which resolved with peak dose effect. Subjective clinical response was positive, with a median patient-reported 5-point Likert rating of 4, 5, and 5 after visits 1, 3, and 8.

Conclusion:

BoNT may be an effective and durable adjunctive symptomatic therapy for people with SPSD with targeted muscle selection based on specific symptomatology. Injections into multiple body regions and use of supramaximal dosages may be required for adequate symptom control in this patient population. As our data lacked objective measures and relied on semiqualitative self-reported patient responses, conclusions about the utility of BoNT are limited and randomized placebo-controlled trials are needed to evaluate the impact of BoNT on improving quality of life, mobility, and burden of systemic symptomatic treatment.

Plain language summary

Use of botulinum toxin injections to treat stiffness and pain in stiff person syndrome spectrum disorders

Stiff person syndrome spectrum disorders (SPSD) are rare autoimmune conditions that cause muscle stiffness and painful spasms, making walking and activities of daily life difficult. Many oral medications used to treat these symptoms are limited by sedation and other side effects. Botulinum toxin (BoNT), a local injectable treatment commonly used for other muscle disorders, has shown promise for relieving SPSD symptoms, but research on its effectiveness and best use is limited. This study followed 37 SPSD patients who received BoNT treatments over several visits between 2018 and 2024. The researchers described which muscles were targeted for treatment, and these included the back, hips, legs, and shoulders, most often. People in this study received higher-than-typical doses of BoNT safely. Some patients reported improvements in symptoms after their treatments. By the eighth visit, the average patient satisfaction was rated as ‘very good’ or ‘excellent’. Side effects were mild, mainly temporary increases in pain or spasms that resolved as the treatment took full effect. Weakness and flu-like symptoms were rarely reported. Some patients were able to reduce their reliance on oral medications for symptom control. The findings suggest BoNT can be a helpful and long-lasting addition to standard SPSD treatments. However, it may require injections in multiple areas and tailored dosing specific for each patient. Barriers to obtaining BoNT include expertise of injectors and cost/potential lack of insurance coverage. Larger studies are needed to confirm these results and explore how BoNT could further improve patients’ quality of life, mobility, and reliance on other therapies.

Introduction

Stiff person syndrome spectrum disorders (SPSD) are immune-mediated disorders with varied presentations but most notably, chronic stiffness, and muscle spasms triggered by various stimuli, which can lead to significant disability.^1–3 SPSD are most commonly associated with antibodies against glutamic acid decarboxylase-65 (GAD65) but can also be associated with other autoantibodies. ⁴ The autoantibodies ultimately result in the impairment of inhibitory GABAergic pathways, resulting in hyperexcitability of the central nervous system (CNS). ⁵ This can manifest as increased tone, muscle spasms, hyperreflexia, exaggerated startle response, and anxiety. ⁶

The spectrum of SPSD has evolved over time, and some experts include several phenotypes under this expanding spectrum realizing the underlying pathogenesis could differ; classic SPS (stiffness and spasms primarily affecting the trunk and lower extremities), stiff limb syndrome (asymmetric stiffness of a single limb), SPS-plus (classic features combined with cerebellar and/or brainstem deficits), and progressive encephalomyelitis with progressive rigidity (PERM).^4,7–9 The cornerstone of treatment in SPSD is symptomatic therapy for treatment of muscle stiffness and spasms (e.g., benzodiazepines, baclofen, tizanidine). At times, these treatment options may be poorly tolerated or dose-limited due to side effects and may be associated with development of tolerance over time. Despite optimizing oral spasticity treatment, symptoms often persist along with disability. Therefore, a great unmet need exists for therapies that minimize the musculoskeletal symptoms of SPSD and have less systemic side effects.^6,10

While ample evidence exists for efficacy of botulinum neurotoxin (BoNT) in dystonia, spasticity, and other neurological disorders, evidence for use in SPSD is scarce in the medical literature. In 1993, Davis and Jabbari published the first blinded, randomized, placebo-controlled evaluation of BoNT in SPS with a sample size of two patients. Following just one BoNT injection, stiffness, pain, and ambulation improved.^11,12 Other than this singular study, case reports have been published confirming anecdotes of positive response to BoNT injected into targeted muscles of people with SPS who had axial,^12–14 extremity,^15–17 and cervicofacial ¹⁸ spasticity.

While it remains unclear if BoNT affects the underlying impaired GABAergic inhibition in SPSD, it may work through both peripheral and central factors, with the latter likely being most relevant to SPSD. Recent evidence suggests that BoNT may exert effects through retrograde axonal transport to the CNS. This transport allows BoNT to reach spinal cord neurons, where it can modulate synaptic activity and reduce hyperexcitability.

The current retrospective, observational semiqualitative study aimed to characterize the location, frequency, dosage, and clinical response to BoNT injections in a cohort of SPSD. We hypothesized that patients would report improvement in stiffness following BoNT and may experience a longer duration of effect with repeated injections over time. Additionally, based on prospective noninterventional data as well as weight-based BoNT safety data in pediatric populations, we conjectured that data from our cohort of adults with SPSD would confirm tolerability of supramaximal BoNT doses.^11,19–23

Methods

This retrospective study was performed at the Johns Hopkins SPS Center and approved by the Johns Hopkins Institutional Review Board (Stiff Person Syndrome: A Review of Clinical Characteristics IRB00154798; participants provided written informed consents as per IRB guidelines). Adults (⩾18 years-old) with a diagnosis of classic SPS, SPS-plus, or PERM treated by SDN and who received ⩾1 BoNT injection by Emilw Sami Moukheiber (ESM) between August 2018 and February 2024 were included. Diagnoses were determined by SDN based on a combination of clinical and paraclinical assessments. Patients later found to have an alternative diagnosis were excluded.

The primary focus was on patients who received ⩾3 BoNT treatments. However, we also included data on patients with 1–2 BoNT injections for comparison.

Clinical notes from BoNT treatment visits were reviewed in the patients’ medical record. Demographics and clinical characteristics including phenotype, antibody seropositivity and titer, BoNT formulation, and doses were obtained from the medical record and our longitudinal SPS database. All serum GAD65 antibody titers were completed by ELISA assay and resulted in IU/ml, although the lab and respective cutoffs had some variability.

Botulinum toxin type A, either onabotulinumtoxinA (BOTOX^®) or incobotulinumtoxinA (Xeomin^®), was administered by a single movement disorder specialist (ESM), with muscle selection based on patient phenomenology after a thorough history and exam revealed the individual patients’ pattern of spasms, spasticity, and/or stiffness. Handheld electromyography (EMG) guidance was used for muscle targeting, as randomized controlled trials and systematic reviews indicate that instrumented guidance, including EMG, improves the effectiveness of botulinum toxin injections compared to manual placement alone.^24–28

Of note, there are three established methods to injecting the psoas muscle: (1) The posterior approach; (2) posterior anterior approach (PAA, which requires anesthesia); and (3) distal anterior approach (DAA). The approach we opted to use is the DAA which is performed using anatomical landmarks, notably two fingerbreadths lateral to the femoral artery and one fingerbreadth below the inguinal ligament. It is a more tendinous portion and a bit further away from the optimal Motor Endplate (MEP) of the muscle. Ultrasound along with EMG was used to confirm needle placement away from the neurovascular bundle and sartorius. The advantages of this approach are avoidance of general anesthesia or sedation involved in the PAA approach, as well as safety from lack of abdominal penetration. It does also encompass the confluence of both psoas and iliacus portions of the muscles. The pitfall is possibly the requirement of higher doses, but the typical 2–4 cm spread/travel of toxin in botulinum toxin injections is probably enough for an effect.

BoNT dose and muscles injected at visits 1, 3, and 8 were recorded for all patients where this data were applicable. Visit 3 was selected because it typically takes multiple BoNT visits to optimize the location and dose of BoNT for each patient. Visit 8 was secondarily added to assess long-term tolerability (e.g., 8 visits represent approximately 2 years of regularly scheduled BoNT injections). Data on patients with only 1 visit (subjective response, side effects, reason for stopping treatment) was interpreted from the Hopkins SPS Center clinic notes, notes from local providers, or via patient portal messages, when available.

At the following visit, patients were asked about the time to effect, duration of effect, side effects, and overall subjective treatment response. As all patients were evaluated by the same provider, documentation was relatively consistent, though descriptions varied in response to stiffness versus spasms or response by body region. We used a patient-reported outcome approach similar to the Patient Global Impression of Change (PGIC), a widely used single-item Likert-type measure.^29,30 Patients were asked to reflect on their peak effect phase (typically ~4 weeks postinjection) and describe overall improvement (e.g., “excellent,” “moderate,” “slight”). These qualitative responses, or patient-reported percentages (from 0% to 100%, if given by the patient), were then converted to a 5-point Likert scale (1 = no response (0%), 2 = slight response (1%–25%), 3 = mild response (26%–50%), 4 = moderate response (51%–75%), 5 = excellent response (76%–100%)). This approach mirrors how PGIC responses such as “much improved” are used as meaningful indicators of treatment benefit in BoNT studies for other conditions.

Statistics

Summary statistics are shown where appropriate. Patients were grouped by number of BoNT visits (<3, 3–7, or ⩾8) and compared on demographics (age, gender, race, ethnicity), antibody seropositivity, GAD65 titer, immune therapies, BoNT formulation, total BoNT dose, and Likert rating. Statistical analyses were completed using the Fisher exact test for comparison of categorical variables and the Mann–Whitney U for comparison of continuous variables, as appropriate. p-Values were set at 0.05. Secondary analysis was done using Spearman’s correlation to evaluate whether an association existed between GAD65 titer and Likert rating. The statistical analysis was performed with SPSS version 29.0.

Results

Participant demographics and BoNT disposition

A total of 37 patients with SPSD seen at the Johns Hopkins SPS Center during the study period received at least one BoNT treatment and were eligible for inclusion. Of these patients, 14 had 1–2 visits, 10 had between 3 and 7 visits, and 13 had ⩾8 visits. Treatments were generally scheduled every 3 months, with some disruptions during the COVID-19 pandemic.

Full demographic information by treatment group is outlined in Table 1. Among all patients receiving BoNT injections, the average age at time of SPSD diagnosis was 49.8 years, 67.6% were female, and 78.4% were white. The most common autoantibody identified was anti-GAD65 (83.8%), with 3 patients (8.1%) being seronegative, 1 patient with amphiphysin antibodies (2.7%), and 2 patients with glycine receptor antibodies (8.1%). The patient with amphiphysin antibodies was the only paraneoplastic associated SPS patient in our cohort. The range of GAD65 titers were from 28 to 2,707,290 IU/ml (Table 2). Classic SPS was the most common SPSD phenotype (78.4%) with three patients each with SPS-plus and PERM, respectively (8.1%). There were no statistical differences among age, race, or ethnicity among patients who were seen for 1–2 visits as compared to those who completed 3 or more visits over time (p > 0.05). However, patients with ⩾3 visits were more likely to be on immune therapy and patients with ⩾8 visits were more likely to have classic SPS phenotype and GAD65 seropositivity when compared to patients with 1–2 visits (p < 0.05).

Table 1.

Participant demographics and clinical characteristics.

	<3 visits n = 14	3–7 visits n = 10	⩾8 visits n = 13	Total n = 37
Sex
Female	9 (64.3%)	6 (60.0%)	10 (76.9%)	25 (67.7%)
Male	4 (28.6%)	4 (40.0%)	3 (23.1%)	12 (32.4%)
Age at diagnosis (years)
Median (IQR)	48.1 (4.3)	55.0 (7.5)	49.8 (4.8)	49.8 (5.5)
Race
White	13 (92.9%)	7 (70.0%)	9 (69.2%)	29 (78.4%)
Black	1 (7.1%)	2 (20.0%)	2 (15.4%)	5 (13.5%)
Asian	0 (0.0%)	0 (0.0%)	1 (7.7%)	1 (2.7%)
Unknown/other	0 (0.0%)	1 (10.0%)	1 (7.7%)	2 (5.4%)
Ethnicity
Hispanic	0 (0.0%)	0 (0.0%)	1 (7.7%)	1 (2.7%)
Non-Hispanic	14 (100.0%)	10 (100.0%)	12 (92.3%)	36 (97.3%)
SPSD antibody
Serum GAD65 Ab	9 (64.3%)	9 (90.0%)	12 (92.3%)	31 (83.8%)
GAD65 titer range (IU/ml), min–max	84–61,696	28–1,339,400	105–2,707,290	28–2,707,290
GAD65 titer >1000 IU/ml	6/9 (66.7)	5/9 (55.6%)	4/12 (33.3%)	15/31 (48.4%)
CSF GAD65, n/#tested	5/10	4/5	3/6	12/21
Glycine receptor Ab	2 (14.3%)	0 (0.0%)	0 (0.0%)	2 (5.4%)
Amphiphysin Ab	0 (0.0%)	1* (10.0%)	0 (0.0%)	1 (2.7%)
Seronegative	3 (8.1%)	0 (0.0%)	0 (0.0%)	3 (8.1%)
SPSD phenotype
Classic SPS	11 (78.6%)	7* (70.0%)	13 (100.5%)	32 (86.5%)
SPS plus	2 (14.3%)	1 (10.0%)	0 (0.0%)	3 (8.1%)
PERM	1 (7.1%)	2 (20.0%)	0 (0.0%)	3 (8.1%)
Toxin
IncobotulinumtoxinA	8 (57.1%)	9 (90.0%)	8 (61.5%)	25 (67.6%)
OnabotulinumtoxinA	6 (42.9%)	1 (10.0%)	5 (38.5%)	12 (32.4%)

All data are shown as n (%) unless indicated. There was one patient with amphiphysin paraneoplastic SPSD, indicated with a *.

GAD65, glutamic acid decarboxylase-65; PERM, progressive encephalomyelitis with progressive rigidity; SPSD, Stiff person syndrome spectrum disorders.

Table 2.

Individual patient clinical characteristics, including sex, race, SPSD phenotype, seropositivity and titer, total doses of BoNT for each visit, and Likert rating for each visit.

No. visits	Sex	Race	Diagnosis	Antibody	Highest titer	CSF GAD65	Immune therapy at visit 1	Visit 1 total BoNT dose	Visit 1 Likert	Visit 3 total BoNT dose	Visit 3 Likert	Interim change in immune therapy	Visit 8 total BoNT dose	Visit 8 Likert	Interim change in immune therapy
1	F	White	Classic	None	n/a	–		195	n/a
1	M	White	Classic	None	n/a	–	IVIG	160	5
1	F	White	Classic	None	n/a	n/a		150	1
1	F	Black	Classic	GAD65	41,557	+		80	n/a
1	F	White	Classic	GAD65	196.3	–	IVIG	255	n/a
1	F	White	PERM	GAD65	89,856	n/a	IVIG	150	n/a
1	M	White	Classic	GAD65	74,501	+		90	n/a	n/a
1	M	White	Classic	Glycine	17	n/a		165	n/a
1	M	White	Classic	GAD65	464	–		330	n/a
2	F	White	Classic	GAD65	1,16,650	+	IVIG	195	n/a
2	F	White	SPS-plus	GAD65	84	n/a	IVIG, Ritux	145	4
2	F	White	Classic	GAD65	11,90,100	+		45	n/a				n/a
2	M	White	SPS-plus	Glycine	146	–	IVIG	250	5
2	F	White	Classic	GAD65	7,90,000	+		157.5	4
3	F	White	Classic	GAD65	28	n/a		250	5	320	n/a	+IVIG
3	M	Black	Classic	GAD65	13,39,400	+	IVIG	185	2	250	5
3	F	White	PERM	GAD65	218.9	+	IVIG	343	5	192.5	5
4	M	White	PERM	GAD65	187	n/a	IVIG	190	4	300	5
4	F	Other	Classic	Amphiphysin	1:7680	n/a	IVIG	200	1	600	1
5	F	White	Classic	GAD65	7252	–	Ritux	275	1	495	n/a	+SCIG
5	F	Black	SPS-plus	GAD65	4,27,900	n/a		155	3	285	4	+IVIG
6	M	White	Classic	GAD65	12,02,900	+	PLEX	295	3	600	4	+Ritux
7	F	White	Classic	GAD65	1,16,425	+	IVIG	200	5	260	5
7	M	White	Classic	GAD65	800	n/a	IVIG	300	4	640	5
8	F	Other*	Classic	GAD65	25,00,000	+	IVIG	265	4	295	5		340	5
8	F	White	Classic	GAD65	488	–	IVIG	165	5	380	5		550	5
8	F	White	Classic	GAD65	2,53,050	n/a	Ritux	250	4	380	5		695	5	+PLEX
9	M	Black	Classic	GAD65	8298	n/a	IVIG	390	4	570	5		620	5
9	F	White	Classic	GAD65	640	+	Ritux	100	4	200	5		395	5
11	F	Asian	Classic	GAD65	196	–	Ritux	300	5	400	4		800	5
11	F	White	Classic	GAD65	338	n/a	IVIG	200	5	500	5		800	5
13	F	Black	Classic	GAD65	27,07,290	+		65	4	100	4		200	4	+IVIG
14	F	White	Classic	GAD65	105	n/a	IVIG, Ritux	395	5	695	4		795	5	IVIG → SCIG
14	F	White	Classic	GAD65	330	–	IVIG, MMF	400	3	600	4	+Ritux	685	5	−MMF
14	M	White	Classic	GAD65	463.2	n/a	IVIG, Ritux, PLEX	465	3	795	5		980	4
16	F	White	Classic	GAD65	194	n/a	IVIG, MMF	300	3	495	5		600	5
20	M	White	Classic	GAD65	2161.4	n/a	IVIG	155	5	285	5		397.5	5	+PLEX

Titers are reported in IU/ml unless specified. Patients with positive GAD65 antibody in CSF are indicated with +, those that tested negative with −, and n/a if not tested. Likert scores that were not able to be calculated are listed as “unknown”—one patient in the 3–7 visit group could not recall how much improvement they had, and the other patients were lost to follow-up. Changes to immune therapy are indicated with a + (addition) or − (discontinuation). There was one paraneoplastic SPSD, the patient with amphiphysin antibodies. Additionally, all patients were non-Hispanic, except for one, which is annotated with as asterisk (*). n/a indicates data not available due to patients completing a fewer number of visits.

BoNT, botulinum toxin; GAD65, glutamic acid decarboxylase-65; IVIG, intravenous immune globulin; MMF, mycophenolate mofetil; PERM, progressive encephalomyelitis with progressive rigidity; Plex, plasmapheresis; Ritux, rituximab; SPSD, stiff person syndrome spectrum disorders.

Technique and muscles injected

Injected muscles and range of dosing are found in Table 3. Patients with ⩾8 visits received significantly higher total BoNT doses at visit 1 when compared to patients with 1–2 visits. Additionally, dose per muscle and total dose tended to increase over time (Figure 1).

Table 3.

Muscles injected per visit with corresponding BoNT dosing.

Body region	Muscle	Visit 1			Visit 3			Visit 8
		n = 37			n = 23			n = 13
		No. patients (%)	Unilateral muscles	Median units (range)	No. patients (%)	Unilateral muscles	Median units (range)	No. patients (%)	Unilateral muscles	Median units (range)
Head and neck	Frontalis	0 (0.0%)	0	–	1 (4.3%)	1	10 (10–10)	3 (23.1%)	3	25 (10–25)
	Temporalis	1 (2.7%)	2	15 (15–15)	2 (8.7%)	4	7.5 (5–10)	1 (7.7%)	2	15 (15–15)
	Masseter	2 (5.4%)	4	12.5 (10–15)	3 (13.0%)	6	22.5 (10–30)	3 (23.1%)	8	20 (15–35)
	Corrugator	2 (5.4%)	4	6.25 (5–7.5)	3 (13.0%)	6	7.5 (5–7.5)	4 (30.8%)	8	6.25 (2.5–10)
	Procerus	2 (5.4%)	2	6.25 (5–7.5)	3 (13.0%)	3	5 (5–7.5)	4 (30.8%)	4	5 (5–5)
	Orbicularis	1 (2.7%)	2	10 (10–10)	0 (0.0%)	0	–	2 (15.4%)	3	10 (10–10)
	Splenius capitus	1 (2.7%)	1	25 (25–25)	1 (4.3%)	2	42.5 (20–65)	2 (15.4%)	4	10 (10–15)
	Sternocleidomastoid	5 (13.5%)	7	25 (15–35)	4 (17.4%)	5	35 (20–40)	2 (15.4%)	4	35 (20–55)
	Levator scapulae	8 (21.6%)	12	25 (15–40)	8 (34.8%)	13	35 (15–60)	9 (69.2%)	16	35 (15–85)
Torso	Semispinalis	4 (10.8%)	6	15 (10–40)	2 (8.7%)	3	10 (10–10)	3 (23.1%)	4	15 (15–15)
	Trapezius	9 (24.3%)	14	25 (15–40)	11 (47.8%)	17	25 (10–60)	10 (76.9%)	18	30 (15–60)
	Deltoid	2 (5.4%)	2	27.5 (25–30)	2 (8.7%)	2	25 (25–25)	0 (0.0%)	0	–
	Rhomboid	2 (5.4%)	3	15 (15–20)	2 (8.7%)	4	25 (15–45)	2 (15.4%)	4	45 (40–45)
	Pectoralis major	4 (10.8%)	5	20 (15–30)	3 (13.0%)	4	25 (25–50)	0 (0.0%)	0	–
	Thoracic paraspinal	17 (45.9%)	34	30 (15–60)	11 (47.8%)	22	47.5 (30–100)	6 (46.2%)	14	45 (15–100)
	Lumbar paraspinal	26 (70.3%)	51	30 (15–90)	19 (82.6%)	38	50 (25–100)	12 (92.3%)	24	62.5 (15–130)
	Rectus abdominus	1 (2.7%)	2	40 (40–40)	1 (4.3%)	2	70 (70–70)	2 (15.4%)	4	27.5 (20–35)
	Transverse abdominus	0 (0.0%)	0	–	1 (4.3%)	2	20 (20–20)	0 (0.0%)	0	–
	Oblique	2 (5.4%)	4	17.5 (15–30)	3 (13.0%)	6	25 (25–40)	5 (38.5%)	9	30 (25–40)
	Latissimus	2 (5.4%)	3	15 (15–25)	1 (4.3%)	1	25 (25–25)	0 (0.0%)	0	–
	Longissimus	1 (2.7%)	1	20 (20–20)	1 (4.3%)	1	30 (30–30)	1 (7.7%)	1	20 (20–20)
Arm	Biceps brachii	3 (8.1%)	3	25 (25–40)	2 (8.7%)	2	45 (40–50)	1 (7.7%)	2	37.5 (25–50)
	Brachialis	2 (5.4%)	2	20 (20–20)	1 (4.3%)	1	40 (40–40)	0 (0.0%)	0	–
	Subscapularis	0 (0.0%)	0	–	1 (4.3%)	1	30 (30–30)	0 (0.0%)	0	–
	Brachioradialis	2 (5.4%)	2	25 (20–30)	1 (4.3%)	1	25 (25–25)	0 (0.0%)	0	–
	Intrinsic hand muscles	1 (2.7%)	1	30 (30–30)	2 (8.7%)	2	20 (10–30)	1 (7.7%)	2	3.75 (2.5–5)
Hips	Iliopsoas	18 (48.6)	27	30 (20–40)	11 (47.8%)	21	30 (10–50)	8 (61.5%)	12	32.5 (20–65)
	Tensor fasciae latae	17 (45.9%)	26	30 (20–40)	13 (56.5%)	24	35 (15–70)	10 (76.9%)	18	30 (10–80)
	Adductor	4 (10.8%)	4	25 (20–25)	0 (0.0%)	0	–	0 (0.0%)	0	–
	Gluteus medius	2 (5.4%)	4	32.5 (25–45)	0 (0.0%)	0	–	0 (0.0%)	0	–
	Gluteus maximus	0 (0.0%)	0	–	2 (8.7%)	3	35 (15–35)	1 (7.7%)	2	45 (45–45)
	Quadriceps	8 (21.6%)			8 (34.8%)			6 (42.6%)
	Rectus femoris	5 (13.5%)	6	25 (20–35)	3 (13.0%)	4	42.5 (25–45)	3 (23.1%)	5	40 (20–75)
	Vastus medialis	2 (5.4%)	2	15 (15–15)	3 (13.0%)	5	20 (15–25)	2 (15.4%)	3	35 (20–65)
	Vastus lateralis	1 (2.7%)	1	15 (15–15)	2 (8.7%)	3	15 (15–25)	1 (7.7%)	2	40 (20–60)
	Hamstring	6 (16.2%)			6 (26.1%)			10 (76.9%)
	Biceps femoris	2 (5.4%)	3	30 (25–35)	2 (8.7%)	4	30 (20–40)	5 (38.5%)	12	25 (20–45)
	Semitendinosus	4 (10.8%)	8	25 (20–25)	4 (17.4%)	7	35 (20–50)	5 (38.5%)	4	25 (15–30)
Legs	Posterior tibialis	18 (48.6%)	29	25 (17.5–60)	9 (39.1%)	15	50 (20–100)	7 (53.8%)	11	50 (20–90)
	Tibialis anterior	2 (5.4%)	2	37.5 (25–50)	2 (8.7%)	3	50 (35–100)	2 (15.4%)	4	15 (10–15)
	Gastrocnemius	0 (0.0%)	0	–	2 (8.7%)	3	25 (25–30)	1 (7.7%)	2	25 (25–25)
	EHL	1 (2.7%)	1	15 (15–15)	3 (13.0%)	3	15 (15–20)	3 (23.1%)	4	20 (15–20)
	EDL	1 (2.7%)	1	15 (15–15)	0 (0.0%)	0	–	0 (0.0%)	0	–
	FHL	2 (5.4%)	2	17.5 (15–20)	0 (0.0%)	0	–	0 (0.0%)	0	–
	FDL	9 (24.3%)	16	25 (20–40)	8 (34.8%)	15	35 (15–80)	3 (23.1%)	5	25 (15–35)
	Intrinsic foot	2 (5.4%)	3	20 (15–25)	1 (4.3%)	3	30 (15–30)	0 (0.0%)	0	–

The most frequently injected muscles (>20%) are bolded.

EDL, extensor digitorum longus; EHL, extensor hallicus longus; FDL, flexor digitorum longus; FHL, flexor hallicus longus.

Figure 1.

Total dose of BoNT administered at each visit. Patients with ⩾8 visits were administered a statistically higher BoNT dose at the first treatment visit compared to patients who had 1–2 visits (p < 0.05).

BoNT, botulinum neurotoxin.

For patients with classic SPS, we found a common phenomenological pattern of spasms and stiffness in specific regions. Spasms typically began with involvement of paraspinal complex muscles with concomitant contraction of antagonist hip flexor muscles, causing prominent hyperlordosis with associated hip flexion and occasionally internal rotation. For many patients, there was also spread of spasms caudally with partial extension/flexion at the knee, plantarflexion of the toes, and inversion with plantarflexion of the foot, commonly with some degree of asymmetry between left and right. For those with cervical musculature involvement retrocollis and mild anterocaput was predominant. A portion of those with abdominal muscle involvement experienced shortness of breath during spasms with a higher respiratory rate and forced expiratory breathing pattern, which was not associated with reduced oxygen saturation. When this symptom was reported, the external oblique muscles were targeted.

The trunk (axial) musculature was the most common location injected, particularly the thoracic and lumbar paraspinal muscles. The thoracic paraspinal muscles were injected in 45.9%, 47.8%, and 46.2% of patients, and the lumbar paraspinal muscles were injected in 70.3%, 82.6%, and 92.3% at visits 1, 3, and 8, respectively. Other muscles injected frequently included the hip flexor muscles including iliopsoas (48.6%, 47.8%, 61.5%) and tensor fasciae latae (45.9%, 56.5%, 76.9%), as well as cervical muscles including trapezius (24.3%, 47.8%, 76.9%) and levator scapulae (21.6%, 34.8%, 69.2%); parentheses percentages represent visits 1, 3, and 8, respectively. Posterior tibialis and flexor digitorum longus were injected frequently as well. Individual muscle data are shown in Table 3, and the most commonly injected muscles are summarized in Table 4.

Table 4.

Summary of the most commonly injected muscles and dosage among our SPSD cohort.

Muscle	Median dose	Range
Levator scapulae	30	15–85
Trapezius	25	10–60
Thoracic paraspinal	40	15–100
Lumbar paraspinal	40	15–130
Iliopsoas	30	10–65
Tensor fasciae latae	30	10–80
Posterior tibialis	35	17.5–100
Flexor digitorum longus	25	15–80

SPSD, stiff person syndrome spectrum disorders.

Treatment response

Excluding the 9 patients who only completed 1 visit and for whom subjective response is unavailable, 67.9% of the remaining 28 patients who received treatment at visit 1 reported a positive subjective response (Likert 4 or 5) following the first injection. Among the 23 patients who completed ⩾3 visits, Likert 4 or 5 was reported by 95.0% after visit 3. By visit 8, all of the 13 patients who were still scheduling treatments reported a Likert 4 or 5 (Tables 2 and 5). Patients with respiratory symptoms from abdominal spasms reported good response with targeting of the external oblique muscles. Many patients reported improvement in ambulation and ability to complete activities of daily living (ADLs), although this data were not captured objectively during visits.

Table 5.

Immune therapies and clinical response to BoNT.

	All	<3 visits	3–7 visits	⩾8 visits
	n = 37	n = 14	n = 10	n = 13
Immune therapy at Visit 1
IVIG	21 (56.8%)	6 (42.9%)	6 (60.0%)	9 (69.2%)
Rituximab	7 (18.9%)	1 (7.1%)	1 (10.0%)	5 (38.5%)
PLEX	2 (5.4%)	0 (0.0%)	1 (10.0%)	1 (7.7%)
MMF	2 (5.4%)	0 (0.0%)	0 (0.0%)	2 (15.4%)
None	11 (29.7%)	8 (57.1%)	2 (20.0%)	1 (7.7%)
Symptomatic therapy at Visit 1
Clonazepam	21 (56.8%)	8 (57.1%)	5 (50.0%)	8 (61.5%)
Diazepam	32 (86.5%)	10 (71.4%)	9 (90.0%)	13 (100.0%)
Baclofen	22 (59.5%)	8 (57.1%)	9 (90.0%)	5 (38.5%)
Robaxin	5 (13.5%)	4 (28.6%)	0 (0.0%)	1 (7.7%)
Tizanidine	4 (10.8%)	1 (7.1%)	2 (20.0%)	1 (7.7%)
Flexeril	0 (0.0%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
∆ Immune therapy—Visit 3
Increase	4 (17.4%)	n/a	3 (30.0%)	1 (7.7%)
Decrease	0 (0.0%)	n/a	0 (0.0%)	0 (0.0%)
∆ Symptomatic therapy—Visit 3
Increase	10 (43.5%)	n/a	2 (20.0%)	8 (61.5%)
Decrease	4 (17.4%)	n/a	3 (30.0)	1 (7.7%)
∆ Immune therapy—Visit 8
Increase	4 (30.8%)	n/a	n/a	4 (30.8%)
Decrease	0 (0.0%)	n/a	n/a	0 (0.0%)
∆ Symptomatic therapy—Visit 8
Increase	10 (76.9%)	n/a	n/a	10 (76.9%)
Decrease	2 (15.4%)	n/a	n/a	2 (15.4%)
Subjective response to BoNT, Likert rating
Visit 1	3.8 (1.3)	3.8 (1.6)	3.3 (1.6)	4.2 (0.8)
Visit 3	4.7 (0.9)	n/a	4.3 (1.4)	4.7 (0.5)
Visit 8	4.8 (0.4)	n/a	n/a	4.8 (0.4)
Duration of BoNT effect, weeks
Visit 1	8.8 (3.5)	10*	8.5 (4.2)	8.9 (3.5)
Visit 3	10.5 (1.5)	n/a	10.7 (1.2)	10.4 (1.6)
Visit 8	10.4 (1.6)	n/a	n/a	10.4 (1.6)
Reported side effects
Visit 1	Spasms—8 (21.6%) Flu-like—2 (5.4%) Weakness—2 (5.4%)	Spasms—2 (14.2%) Flu-like—2 (14.2%) Weakness—1 (7.1%)	Spasms—3 (30.0%) Weakness—1 (10.0%)	Spasms—3 (23.1%)
Visit 3	Spasms—5 (21.7%) Weakness—1 (4.3%)	n/a	Spasms—1 (10.0%) Weakness—1 (10.0%)	Spasms—4 (30.8%)
Visit 8	None	n/a	n/a	None

Values are presented as n (%) or average (standard deviation). Regarding side effects, spasms denotes transiently worsening pain and spasms in the days following injection, weakness refers specifically to leg weakness, flu-like refers to two patients where upper respiratory or flu-like symptoms were reported. In the column showing data for all patients, visit 3 data included patients in the 3–7 visit and ⩾8 visit cohorts, n = 23. Visit 8 data included only ⩾8 visit patients, n = 13. n/a indicates data not available due to patients completing a fewer number of visits.

^*Only 1 patient in the <3 visit cohort provided duration effect estimate.

BoNT, botulinum neurotoxin; IVIG, intravenous immunoglobulin; MMF, mycophenolate mofetil; PLEX, plasmapheresis.

Nine patients completed one injection visit and did not return for follow-up, so no subjective response was available. Five patients completed 2 visits, one of whom reported no improvement, while the other 4 reported Likert scores of 4 or 5, average Likert 3.8 (Table 5). The reason for treatment cessation at our center in those with 1–2 visits was unknown for three patients. For the remaining 11 patients, 4 had a positive response and continued to receive BoNT injections by local providers but were lost to follow-up for the purpose of this study, 3 stopped due to costs of travel and BoNT, 3 cited side effects as the main reason to stop treatment, and 1 felt the duration of relief (reported 2 weeks) was too short to warrant continued treatment.

Duration of efficacy varied among patients, but the average patient-reported effect duration was 10.5 weeks for the third BoNT injection and 10.4 weeks for the eighth. This is in keeping with duration of effect with other conditions. ³¹

Safety

The maximum FDA recommended dose of both incobotulinumtoxinA and onabotulinumtoxinA is no more than 400 units every 12 weeks. Per standard clinical practice of ESM, dosage of BoNT was appropriately optimized over time based on location and response. Across all patients, the median dose at visit 1 was 200 units (range 45–465), at visit 3 was 380 units (range 100–795), and at visit 8 was 620 units (range 200–980; Figure 1).

Reporting of side effects for patients with 1–2 visits was limited as many did not return to BoNT clinic. Nonetheless, among all patients in this cohort, the most common side effect was transient worsening of pain and spasms up to 1 week following the injections, which typically abated after several days (Table 5). Eight patients (21.6%) reported this side effect after the first visit. Four patients (17.4%) reported this side effect following the third visit. No patients reported this after visit 8, which included all 4 of the patients who experienced it after visit 3, suggesting this may resolve with time. Additional side effects were reported only after the first visit and included two patients with leg weakness (5.4%) and two patients with flu-like symptoms following injection (5.4%).

None of the patients in this cohort experienced clinical resistance to BoNT, even with supra-maximal doses. There were six patients (26.1%) at visit 3 and eight patients (61.5%) at visit 8 who received a total cumulative dose of 600 units or greater. Notably, as spasms in SPSD are often triggered by muscle activation as well as external stimuli, patients in this cohort had ongoing spasms during BoNT injection.

Change in immune and symptomatic therapy

At the time of the third BoNT treatment for the 23 patients with ⩾3 visits, 3 (13.0%) patients had escalated immune therapy while no patients had reduced immune therapy (Table 2). Ten patients (43.5%) had increased the frequency or dosage of symptomatic medications, while 4 patients (17.4%) had reduced symptomatic therapies. Reduction included both dosage (two patients) and discontinuation of a medication (two patients) without changes to immune therapy. For the 13 patients who had ⩾8 BoNT visits, between the third and eighth BoNT visit, 4 (30.8%) had escalation in immune therapy, 10 (76.9%) had increase in symptomatic medications, and 2 (15.4%) had reduction in symptomatic medications. No reduction in immune therapy was observed.

Discussion

SPSD generally results in severe disability from spasms and stiffness that are often inadequately treated with oral or intrathecal therapies. These symptomatic therapies are associated with side effects and risk of tolerance. BoNT offers an alternative symptomatic treatment that may improve quality of life and physical function.

Among 37 patients treated with BoNT, there was an overall positive subjective clinical response. Over 94% of the 23 patients who completed ⩾3 visits reported a very good or excellent response (Likert 4 or 5) after the third BoNT treatment and all 10 patients with ⩾8 visits did so after the eighth treatment. These findings demonstrate the durability of BoNT in SPSD. The mean duration of BoNT effect was approximately 10.5 weeks, with wearing off for some patients before the standard 12-week dosing interval. The patients who experience this wearing off phenomena have amplification of their baseline symptoms just prior to their next BoNT treatment.

The majority of our patients were injected with incobotulinumtoxinA. We did not observe attenuation of effect with repeat injections or evidence of clinical resistance, although longer-term follow-up may evaluate the risk of immunogenicity with increased dosing. ³² Injections in our cohort frequently targeted paraspinal muscles, hip flexors, plantar flexors, and shoulder girdle muscles. For patients with dyspnea during spasms, external oblique targeting was beneficial. FDG-PET/CT uptake patterns may correlate with spasm distribution and warrant further study as a diagnostic or targeting tool. ³³

Reported benefit may increase over time as a result of optimization of dose and muscle targeting based on individual needs. However, patients experiencing the greatest benefit may also be more inclined to return for ongoing treatment, introducing a likely selection bias. It is also important to note that response was assessed using a retrospective Likert scale derived from clinical notes, a subjective and psychometric tool that lacks objectivity, limiting interpretability. Placebo effects, which can be particularly prominent in SPSD populations, cannot be excluded.

To evaluate whether certain characteristics were associated with a positive and sustained response to BoNT, we compared patients who had only 1–2 treatment visits with those who completed ⩾8 visits. Patients with ⩾8 visits were noted to have higher BoNT doses at their first treatment and were more likely to be GAD65 seropositive, on immune therapy at baseline, and have a classic SPS phenotype compared to those with only 1–2 visits. The latter observation is interesting given our anecdotal experience that patients with a classic SPS phenotype appear to have the best response to BoNT. We found no association between GAD65 titer and response to BoNT.

We observed four patients reduced oral symptomatic therapies after three BoNT injections without changes to their immune therapy. However, a greater number required escalation of symptomatic therapies, and several patients required escalation in immune therapy over the study period, while no reduction in immune therapy was observed. This pattern may reflect the underlying progressive course of SPSD. BoNT is not a disease-modifying therapy, and these results highlight the frequent need for multiple symptomatic interventions alongside immune therapies in SPSD.

BoNT dosages used in clinical practice are typically higher than doses published in randomized trials and official prescribing information.^11,23,34 Extrapolating from prospective noninterventional studies and pediatric weight-based BoNT safety data, adults may tolerate doses more than 400 units, which our data support.^20,23 In our cohort, supramaximal dosing of BoNT was used with doses gradually increasing at subsequent visits, to a maximum 980 units by visit 8 for one patient. Notably, median dosing per muscle was not high compared to dosing ranges in the literature; however, the total dose injected per session was high given the extent of muscle involvement and number of injections required for adequate symptom control. ²³ These higher doses were generally well tolerated without significant side effects, including weakness.

There was a relatively infrequent but unique phenomenon seen in our cohort with paradoxical worsening of stiffness, pain, and spasms in the week after injections. We did not find a clear pattern between SPSD phenotype, antibody positivity, antibody subtype, or titers and occurrence of this side effect. However, despite this side effect, patients continued to report benefit from BoNT, and it was not reported by the 8th visit. We hypothesize that this may have been due to local muscle irritation in the setting of general hyperexcitable nervous system. ⁶ Patients with SPSD undergoing BoNT injections should be counseled about this potential early side effect and the expectation that it will likely improve over time. Adjustment of anti-spasmodic medication or the addition of over-the-counter analgesics could be suggested to offset this transient symptom.

Spasms may be triggered at the time of BoNT injection. In our experience, patients benefit from prolonged visit times, and spasms may be reduced by premedication with benzodiazepines and/or application of a vibration stimuli to the affected muscle. The use of EMG to guide injections may improve accuracy of muscle localization by providing real-time confirmation of needle placement, reduce the risk of injecting nontarget muscles, muscles with overlapping anatomy, minimize side effects, and potentially allow for lower toxin doses by ensuring precise delivery.^25–28

Strengths of this study include a relatively large SPSD cohort treated by consistent providers (e.g., all immune and symptomatic therapies were managed by SDN and all BoNT injections were performed by ESM), limiting variability in clinical management.

Limitations include the retrospective design, lack of objective outcome measures (e.g., Modified Ashworth score, timed 25-foot walk), reliance on subjective assessment, and variability in injection timing and targeting. Only half of the patients in our cohort had GAD65 titers >1000, and 8% were seronegative, introducing phenotype heterogeneity and possible diagnostic uncertainty. BoNT is not routinely used in many SPS clinics due to its cost, need for expertise, and perceived limited benefit, and prior literature has questioned long-term efficacy. ³⁵

Our cohort is inherently biased as patients who self-select to continue injections are more likely to be those who derive the greatest benefit from BoNT and patients who derive little benefit or experience intolerable side effects are unlikely to return for repeat injections; although, in clinical practice, this is likely also the case with other conditions for which BoNT is used. In treating patients with SPSD, there may be a placebo effect as well as a conditioning effect which may significantly impact patient-reported outcomes and would be best addressed with a double-blind randomized, placebo-controlled study design.^36,37

Conclusion

BoNT may provide subjective symptomatic benefit for some patients with SPSD, particularly with localized painful spasms and stiffness in accessible areas. This study describes the distribution, dose, and subjective response to BoNT injections in patients with SPSD, based on experience from one center, with the goal of informing providers administering BoNT to SPSD patients. Our findings suggest possible value in select patients but should be interpreted cautiously. Future prospective studies should attempt to utilize a placebo-controlled design and incorporate objective measures of clinical improvement to explore the efficacy and role of BoNT in SPSD management.