Childhood Onset of Stiff-Man Syndrome

Abstract

IMPORTANCE

Reports of pediatric-onset stiff-man syndrome (SMS) are rare. This may be an underrecognized disorder in child neurology practice.

OBJECTIVE

To describe patients with disorders in the SMS spectrum beginning in childhood.

DESIGN, SETTING, AND PARTICIPANTS

This study was a medical record review and serological evaluation conducted at child and adult neurology clinics at the Mayo Clinic, Rochester, Minnesota. Systematic review of the literature was conducted of patients who presented from 1984-2012 with onset of symptomatic SMS occurring at age 18 years or younger.

MAIN OUTCOMES AND MEASURES

Response to symptomatic and immunotherapies, patient and physician reported, including modified Rankin scale.

RESULTS

We identified 8 patients with childhood-onset SMS, representing 5% of patients with SMS evaluated at Mayo Clinic during a period of 29 years (4 were girls). The median age at symptom onset was 11 years (range, 1-14 years). The diagnosis in 3 patients was not established until adulthood (median symptom duration at diagnosis, 14 years; range, 0-46 years). The phenotypes encountered were: classic SMS (n = 5, involving the low back and lower extremities), variant SMS (n = 2, limited to 1 limb [with dystonic posture] or back), and progressive encephalomyelitis with rigidity and myoclonus (n = 1). Initial misdiagnoses included functional movement disorder (n = 2), generalized dystonia and parkinsonism (n = 1), and hereditary spastic paraparesis (n = 1). Six patients had 1 or more coexisting autoimmune disorders: type 1 diabetes mellitus (n = 4), thyroid disease (n = 2), and vitiligo (n = 2). Serologic study results revealed glutamic acid decarboxylase 65–IgG in all cases (median value, 754 nmol/L; range, 0.06-3847 nmol/L; normal value, ≤0.02 nmol/L) and glycine receptor antibody in 3 cases. Improvements were noted with symptomatic therapy (diazepam, 6 of 6 patients treated, and oral baclofen, 3 of 3 treated) and immunotherapy (intravenous immune globulin, 3 of 4 treated and plasmapheresis, 3 of 4 treated). The 3 patients with glycine receptor antibody all improved with immunotherapy. At last follow-up, 4 patients had mild or no symptoms, but 4 had moderate or severe residual symptoms and required maintenance symptomatic therapy (n = 5) and immunotherapy (n = 4). Ten of 12 pediatric SMS cases identified by literature review had a severe whole-body phenotype resembling progressive encephalomyelitis with rigidity and myoclonus.

CONCLUSIONS AND RELEVANCE

Childhood-onset SMS is a rare but underrecognized and treatable disorder. Serological and electrophysiological testing aid diagnosis.

Historically, stiff-man syndrome (SMS)¹ has been an underrecognized disorder among adults and is frequently misdiagnosed as another neurological disorder or as a psychogenic disorder.² The paucity of literature pertaining to childhood-onset SMS suggests this also may be so in child neurology practice. Patients with classic SMS (adults and children) may present with stiffness and spasms of the lumbar region and lower extremities, but some patients have a more anatomically restricted form of the disorder (SMS variants including stiff limb and stiff trunk), and rare cases present with whole-body stiffness as a component of a severe autoimmune encephalomyelitis (known as progressive encephalomyelitis with rigidity and myoclonus [PERM]).^3-6

Patients within this SMS spectrum typically have coexisting autoimmune diseases, most commonly type 1 diabetes mellitus or thyroid dysfunction, and serological evidence of thyrogastric autoimmunity.³ An IgG specific for the 65-kDa isoform of glutamic acid decarboxylase (GAD65-IgG) is detected in 80% of SMS cases.^7,8 Other autoantibody markers aiding the diagnosis of the SMS spectrum in adults include amphiphysin-IgG detected in patients with paraneoplastic neurological autoimmunity associated with breast adenocarcinoma or small-cell lung carcinoma,^9,10 and glycine receptor α1 subunit (GlyRα1)–IgG, which is usually encountered in a nonparaneoplastic context.^11,12

Five patients in a recently reported cohort of 99 patients with SMS evaluated during a period of 25 years³ had symptom onset before age 18 years; 2 of those were seropositive for both GlyRα1-IgG and GAD65-IgG.¹² This prompted us to interrogate our medical record index system for all pediatric patients assigned a diagnosis of SMS and to review the literature for pediatric cases.

Methods

This study was approved by the Mayo Clinic institutional review board (IRB 08-00807). We searched the Mayo Clinic’s computerized diagnostic index for patients with a first appointment at age 18 years or younger (January 1984-June 2012) who at some point had one of the following considered in the differential diagnosis: stiff man, stiff person, stiff limb, stiff child, or PERM. We reviewed 28 medical records, eliminating duplicate results and patients without these diagnoses. The results of this search were cross-referenced and combined with the known patients from our previous comprehensive study of 79 patients with SMS at the Mayo Clinic, 5 of whom had symptom onset prior to the age of 18.3 years. A total of 8 patients were identified as having onset of SMS prior to age 18 years. Informed consent for medical record review for research purposes had been obtained for all patients at the first clinical evaluation.

Of the 8 patients, 4 were girls (Table). There were 7 white patients and 1 African American patient. The median age at symptom onset was 11 years (range, 1-14 years). Three patients did not receive a diagnosis until adulthood (median age at diagnosis, 16 years; range, 8-51 years). The symptom duration at diagnosis was 14 years (range, 0-46 years). The median follow-up duration from diagnosis at the Mayo Clinic to the last follow-up visit was 9 months (range, 0-300 months).

Table. Features of 8 Patients With Pediatric-Onset SMS Spectrum Treated at Mayo Clinic.

	Patient 1	Patient 2	Patient 33a	Patient 43a	Patient 53a	Patient 63,12a	Patient 73,12a	Patient 8
Age at onset, y	11	12	7	12	5	14	5	1
Age at diagnosis, y	13	13	8	26	49	17	51	14
Onset to diagnosis, y	2	0	0	14	44	3	46	13
Sex	Male	Female	Female	Male	Male	Male	Female	Female
Race/ethnicity	White	White	African American	White	White	White	White	White
SMS phenotype	Classic	PERM	Classic	Classic	Classic	Variant stiff trunk	Classic	Variant stiff leg
Distribution of symptoms/signs
Upper limbs	−	+	−	−	+	−	−	−
Lower limbs	−	+	+	+	+	−	+	+
Trunk	+	+	−	+	−	+	+	−
Head and/or neck	−	+	−	+	−	−	+	−
Whole-body  spasms	+	+	−	−	−	−	−	−
Respiratory	−	−	−	−	+	−	−	−
Deep tendon reflexes	Normal	Brisk	Reduced	Brisk	Reduced (in context of acute neuromyopathy at onset of DM1)	Normal	Normal	Brisk
Autoimmunity	No	DM1	DM1, vitiligo	Vitiligo	DM1; Graves disease with subsequent hypothyroidism; premature gray hair	No	No	DM1, Hashimoto thyroiditis
Family history of autoimmunity	Both parents: hypothyroid	Mother: hyperthyroid, father: vitiligo	Both grand- mothers: DM1, also 1 with hypothyroid- ism; cousin: DM1	Grandparent: DM1	Thyroid disease, premature graying	Grandmother: hypothyroid, RA	2 sisters: autoimmune thyroid disease; 1 sister: DM1	Uncle: thyroid disease
GAD65-IgG, nmol/L	0	373	285	734	27.7	0.06	12.3	3847
GlyRα1-IgG	Negative in serum, CSF unavailable	Negative in serum, CSF unavailable	Positive in serum, CSF unavailable	Negative in serum, CSF unavailable	Negative in serum, CSF unavailable	Negative in serum, positive in CSF	Negative in serum, positive in CSF	Negative in serum, nega- tive in CSF
Electrophysiology	Not done	EMG (2009): normal; movement laboratory (2009): hyperekplexia	EMG (2005): normal	Not done	EMG (2005): mild axonal sensorimotor neuropathy and diffuse myopathy	EMG (2003): nonspecific in- ability to relax; movement labo- ratory (2003): paraspinal spasmb	EMG (2007): normal; movement laboratory (2007): normalb	EMG (2009): normal
Treatments with good clinical response	PLEX, prednisone	Diazepam	Diazepam, baclofen, IVIg, PLEX (dramatic but short lived, thus mycophe- nolate recommended)	Diazepam (initially got benefit, but this waned after 2 y; intrathecal baclofen and steroids recommended)	None tried	Diazepam, baclofen, dantrolene, botulinum toxin, PLEX	Diazepam, baclofen, IVIg, azathioprine (initial good responses with IVIg with mRS of 1, but then regressed)	Diazepam, IVIg
Treatments with poor clinical response	Narcotics, hydroxyzine, cyclobenzaprine	IVIg, PLEX, rituximab	NA	NA	None tried	NA	NA	NA
mRS score
Initial	4	3	2	3	2	3	3	3
Last follow-up	0	3	1	3	2	1	3	1
Duration of follow-up, mo	300	7	3	24	1	7	24	36

Abbreviations: CSF, cerebrospinal fluid; DM1, type 1 diabetes mellitus; EMG, electromyography; GAD65, glutamic acid decarboxylase 65-kDa isoform; GlyRα1, glycine receptor α1 subunit; IVIg, intravenous immunoglobulin; mRS, modified Rankin Scale; NA, not applicable; PLEX, plasma exchange; PERM, progressive encephalomyelitis with rigidity and myoclonus; RA, rheumatoid arthritis; SMS, stiff-man syndrome.

^aPreviously reported.

^bTaking baclofen and/or diazepam at the time of movement laboratory study.

We included patients with the final diagnosis of SMS, a limited form of the disorder, or PERM. Patients were classified as having classic SMS if lower-extremity and lumbar stiffness and spasms were present, and variant SMS if stiffness was restricted to axial, lower-extremity, or upper-body musculature. Patients were classified as having PERM if stiffness and spasms involved the whole body and were accompanied by a rapidly progressive encephalomyelopathy.

For all patients, we documented a detailed history and findings from physical examination and serological and electrophysiological studies at the time of initial evaluation. Using medical record review and telephone interviews, we documented extended follow-up data where available. The level of disability was assigned at last point of surveillance using the modified Rankin score as follows: 0 = no symptoms, 1 = symptoms but no disability, 2 = slight disability but able to look after own affairs, 3 = moderate disability with a need for help but able to walk without assistance), 4 = unable to walk without assistance, and 5 = bedridden.

Serum samples from all patients were tested in the Mayo Clinic Neuroimmunology Laboratory by radioimmunoprecipitation assays for GAD65 antibody and other autoantibody markers of type 1 diabetes mellitus (insulin and IA-2) and for paraneoplastic autoantibodies, as previously described.³ Testing for GlyRα1-IgG was performed at the University of Barcelona by immunofluorescence assay on GlyRα1-transfected cells, as previously described.¹² Electrophysiological studies in selected cases consisted of multichannel surface electromyographic recordings over the right orbicularis oculi, sternocleidomastoid, biceps, abductor pollicis brevis, thoracic paraspinals, lumbar paraspinals, tibialis anterior, gastrocnemius, and soleus muscles. Auditory startle reflexes (pattern and habituation of motor responses) were evaluated using binaural 105-dB intensity stimuli, 1 minute apart, for 3 to 5 trials.¹³ Exteroceptive responses were determined by electrically stimulating the medioplantar nerve.¹⁴ Concentric-needle studies were performed on the lumbar paraspinal muscles.

Review of the literature to identify reports or case series of pediatric-onset SMS or related disorders consisted of searching PubMed English-language publications with descriptions of individual patients using combinations from the 2 groups of following terms: group 1 = stiff person, stiff man, stiff limb, and progressive encephalomyelitis; group 2 = boy, girl, child, and pediatric. We did not specify any starting or ending dates in the search to be as inclusive as possible. The search was completed on February 5, 2013; 11 reports were identified. The references section of each report was reviewed to ensure that no additional reports were referenced that were not identified in the PubMed search.

Results

Presenting Symptoms and Neurological and Serological Findings

Lower-extremity stiffness or spasms causing gait difficulties were the most common initial symptoms (5 patients either recalled this from childhood or reported this at presentation). Patient 1 had his first episode of severe paraspinal muscle spasm resulting in marked lumbar lordosis at age 11 years and also experienced left-leg stiffness at that time. His symptoms resolved without specific treatment but then recurred 10 months later and persisted until treated with immunotherapy. Three patients had exaggerated startle. One of these (patient 4) reported exaggerated startle, and limb symptoms were brought on by being asked to speak in front of the class at school. This frequently resulted in falling over “stiff as a board” without protective reflexes, resulting in facial injury. An additional 3 patients reported that emotional stress or anxiety precipitated symptoms. Patient 5 reported limb muscle cramps and breathing difficulties when exercising during childhood and adolescence. This was recognized at age 49 years, at the time of diagnosis, as SMS-related limb and diaphragmatic spasms. The same patient also developed a seizure disorder at age 14 years and subsequently neuropathy and myopathy at age 48 years, when diagnosed as having diabetes mellitus and SMS. Patient 6 presented with intrascapular muscle spasms and mid-thoracic back pain. Patient 8, who walked pigeon toed, was prone to falls and had a monomelic stiff-limb phenotype.

Stiff-man syndrome phenotypes encountered at diagnosis were: classic (n = 5; involving low back and lower extremities), variant (n = 2; limited to 1 limb [with dystonic posture] or back) and PERM (n = 1). Prior diagnoses included psychogenic movement disorder (n = 2), idiopathic generalized dystonia and parkinsonism (n = 1), and hereditary spastic paraparesis (n = 1).

Coexisting Autoimmunity

Five patients had a coexisting autoimmune disorder. Four patients had type 1 diabetes mellitus, 2 patients had thyroid disease, and 2 patients had vitiligo. Of the 4 patients with diabetes mellitus, the diagnosis of diabetes preceded the onset of SMS symptoms in 2 patients; in 2 patients, diabetes was diagnosed after the first onset of SMS. Serological testing results revealed GAD65-IgG in 7 patients (median value, 754 nmol/L; range, 0.06-3847 nmol/L; normal value, 0.00-0.02 nmol/L). Glycine receptor α1 subunit–IgG was detected in 3 of 8 patients (1 of 7 was positive in serum, 2 of 3 were positive in cerebrospinal fluid). One or more other autoantibody specificities were detected in 3 patients: antinuclear antibody (n = 2), thyroid peroxidase antibodies (n = 2), and insulin antibody (n = 1). Cancer was not identified in any patient during the short surveillance period (median, 9 months).

Neurophysiological Findings

Of 6 patients who underwent conventional electromyographic testing, results for 4 were interpreted as normal; 1 patient was unable to relax consistent with a hyperexcitability disorder and patient 5 demonstrated a mixed large-fiber, axonal sensorimotor peripheral neuropathy, as well as diffuse myopathy (the patient had type 1 diabetes mellitus). Of 3 patients who underwent movement electrophysiologic studies, results for 2 were abnormal (exaggerated acoustic startle and exteroceptive responses, n = 1; and cervical and thoracic paraspinal muscle spasms, n = 1). Results for 1 were interpreted to be normal (although some measurements acquired were suggestive of excessive startle); the patient was only modestly symptomatic and was taking baclofen and diazepam treatment at the time of the study, which may have masked abnormalities.

Treatments and Outcomes

Improvements in stiffness and spasms were reported among all 6 patients treated with diazepam and all 3 who received baclofen. Of the 4 patients for whom symptomatic treatment was undertaken during childhood, the median total daily dose of diazepam used was 14.0 mg (range, 7.5-90.0 mg) and the median daily oral baclofen dose was 60 mg (range, 60-80 mg). Immunotherapies that improved symptoms were intravenous immunoglobulin (IVIg; 3 of 4 treated), plasmapheresis (3 of 4 treated; 5-7 treatments given over 10-14 days), prednisone (1 patient treated), and azathioprine (1 patient treated). Among those who received IVIg as children, dosing ranged from 0.5 to 1.0 g/kg per infusion, with frequency of administration ranging from daily infusions (for 3 days only) to once weekly infusions (for several months). A combination of symptomatic and immunological therapies were used to control symptoms in 5 patients. Patient 8 with stiff limb ultimately had orthopedic surgery to correct the foot deformity, with significant improvement in functionality and pain. All 3 patients seropositive for GlyRα1-IgG had improvements with immunotherapy.

At last follow-up of all 8 patients, 4 had mild or no symptoms and 4 had moderate or severe residual symptoms and required maintenance with symptomatic therapy (n = 3) and immunotherapy (n = 2). At last follow-up, 1 patient had a modified Rankin Scale score of 0, 3 patients had a score of 1, 1 patient had a score of 2, and 3 patients had a score of 3.

Literature Review

We identified 12 cases in the literature reported as childhood-onset SMS phenotypes (eTable in Supplement).^15-26 Six of the cases were published before SMS was first recognized as an autoimmune disease.⁷ The SMS phenotypes reported were 2 patients with classic SMS and 10 patients with generalized limb and truncal disorders including 5 patients with craniocervical involvement. Episodic whole-body spasms superimposed on stiffness were reported in 8 patients, 1 of whom died during such an episode. Glutamic acid decarboxylase 65–IgG was detected in 2 of 4 patients tested; GlyRα1-IgG was detected in the single patient tested who had a PERM phenotype. Just 1 of the 12 patients was reported to have a coexisting autoimmune disease (diabetes mellitus). Improvements with symptomatic therapies were reported in 8 of 10 patients treated with benzodiazepines and 4 of 7 treated with oral baclofen. Immunotherapies that proved beneficial were IVIg (all 4 patients treated) and corticosteroids (4 of 5 patients treated). Improvement occurred with a combination of immunotherapy and symptomatic treatment in 4 patients (diazepam, baclofen, valproic acid, and steroids [n = 1]; diazepam, IVIg, and steroids [n = 1]; diazepam, clonazepam, IVIg, and rituximab [n = 1]; and IVIg, steroids, and levetiracetam, [n = 1]). One patient was reported to have spontaneous remission or improvement.

Discussion

Disorders of the SMS spectrum are rare. Approximately 4 cases per year are identified at the Mayo Clinic, Rochester.³ Pediatric cases are even rarer and account for just 5% of patients with SMS spectrum diagnosis. Recognition is nonetheless important because these are treatable disorders, usually responding to 1 or more of symptomatic therapy and immunotherapy. Classic (n = 5), variant (n = 2), and PERM (n = 1) forms were encountered; all had a personal or family history of autoimmunity. Electrophysiology and antibody marker testing (GAD65-IgG and GlyRα1-IgG) helped facilitate patient identification. Patients treated with diazepam and baclofen universally improved; responses to immunotherapy were variable but frequently complemented symptomatic treatments.

Of the 8 pediatric SMS cases identified at the Mayo Clinic over 29 years, 3 did not receive this diagnosis until adulthood. Thus, the diagnosis of SMS in childhood was presumptive in those 3 patients because it was based on patient report alone. The average duration before diagnosis of SMS was 14 years, and 2 patients reported receiving a diagnosis of a psychogenic cause for their symptoms. The rarity of pediatric SMS, and some of the clinical features of the disorder, may hinder recognition. Features that could potentially bias the examiner to a psychogenic diagnosis include motor signs provoked by patient anxiety,² the common occurrence of psychiatric comorbidities,²⁷ and the distressed appearance of patients with SMS resulting from stiffness, spasms, and fear of injurious falls.

It is also important to differentiate pediatric SMS from other organic neurological disorders encountered in childhood, including inherited hyperekplexia, or stiff-baby syndrome²⁸; sporadic and inherited dystonias^29,30; hereditary spastic paraplegia,³¹ a hereditary disorder resembling SMS responsive to diazepam³²; and muscle rigidity in a newborn due to continuous peripheral nervehyperactivity.³³ In addition to the characteristic neurological features described, the coexistence of a nonneurological autoimmune disease in 4 patients, diagnostic electrophysiological abnormalities in 3 patients, and serological findings in all proved critical to the diagnosis. Among the Mayoseries, the SMS diagnosis for patient 1 was initially less certain because of a lack of definitive autoimmune history, antibody markers (testing negative), and electrophysiological findings (testing unavailable). However, the patient had a classic phenotype and symptoms improved dramatically with plasma exchange.

Glutamic acid decarboxylase 65–IgG was detected in all but 1 of our patients. Among all patients, adults and children, we have previously evaluated, approximately 20% of patients with an SMS spectrum disorder were GAD65-IgG seronegative.¹² Ten percent of that same cohort of patients with SMS spectrum were GlyRα1-IgG positive including 25% of GAD65-IgG seronegative patients.¹² The spectrum of GlyRα1-IgG–associated disorders in childhood includes classic SMS, variant SMS (both reported here), and PERM.²⁶ It was not surprising that none had amphiphysin-IgG, a marker of paraneoplastic neurological autoimmunity accompanied by breast adenocarcinoma or small-cell carcinoma in adults.¹⁰

Reports of pediatric SMS in the literature (12 total) were also rare. The patients described in those reports differed from our own patients in some respects. First, the literature review patients had severe phenotypes more frequently including whole-body symptoms and signs. In many instances, the whole-body presentation depicted closely resembled descriptions of PERM, although that term was used for only 1 of the 12 cases reported.²⁶ In contrast, we identified only 2 children in the literature who had the classic lumbar and proximal lower-extremity–restricted SMS presentation.^21,24 This observation may reflect a bias for reporting more severely affected patients as single case reports, or, alternatively, may be consistent with underrecognition of different SMS phenotypes as childhood diagnoses. Second, results of serological testing were reported in just 5 patients^21-26; those 5 reports^15-20 antedated recognition of an autoimmune etiology for SMS and the discovery of diagnostic antibody biomarkers.^7,34

Treatment responses were difficult to quantify both for our own cases and the literature cases. Factors hindering interpretation included the retrospective study design, the small number of patients with this rare disorder, the heterogeneity and combinations of the rapies used, and the subjectivity of most outcome measures. Nonetheless, symptomatic improvements were almost universally reported among children treated with diazepam alone or in combination with oral baclofen. Gradual dose escalation under close medical supervision was usually required to maximize the benefit. One or more modalities of immunotherapy (corticosteroids, IVIg, plasma exchange, rituximab, and azathioprine), usually administered in addition to symptomatic therapy, also proved beneficial in most patients. Response rates to immunotherapy are variable.³ This variability may reflect a delay in implementing immunotherapy or different pathophysiological mechanisms in individual patients. Glutamic acid decarboxylase 65–IgG recognizes a cytoplasmic synaptic vesicle protein and is considered a surrogate marker of an inflammatory organ-specific autoimmune disorder mediated by effector T cells, a hypothesis supported by autopsy findings.³⁵ These patients may be less responsive to immunotherapy than those with a pathogenic effector antibody targeting a synaptic plasma membrane antigen such as GlyRα1. Consistent with our observations among the Mayo Clinic children with SMS and with the findings reported for 1 pediatric PERM case,²⁶ GlyRα1-IgG positivity may predict immunotherapy responsiveness.¹²

Stiff-man syndrome phenotypes are rarely encountered in childhood. Characteristic symptoms and signs, an autoimmune background, and electrophysiology and IgG biomarker testing should facilitate the identification of affected patients for whom treatment with gamma-aminobutyric acidergic and immunotherapeutic drugs are likely to be beneficial.