Anhar Hassan, Farrah J. Mateen, Elizabeth A. Coon, J. Eric Ahlskog. Painful Legs and Moving Toes SyndromeA 76-Patient Case Series. Arch Neurol. 2012;69(8):1032–1038. doi:10.1001/archneurol.2012.161
Author Affiliations: Departments of Neurology, Mayo Clinic, Rochester, Minnesota (Drs Hassan, Coon, and Ahlskog), and The Johns Hopkins Hospital, Baltimore, Maryland (Dr Mateen).
Objective To better characterize the clinical features, electrophysiologic features, and treatment outcomes of painful legs and moving toes (PLMT) syndrome.
Design Large case series.
Setting Neurology outpatient clinic at a tertiary referral center, 1983-2011.
Patients All cases of PLMT seen at our institution during an 18-year period were identified using our medical record linkage system.
Main Outcome Measures Key demographic, clinical, imaging, and electrophysiologic features of PLMT. Treatment outcomes and long-term follow-up are also reported.
Results Of 76 cases identified (including 50 women [66%]), the mean age at onset was 58 years (range, 24-86 years) and at neurologic evaluation was 63 years (range, 26-88 years). Pure lower limb involvement was most common (69 patients [91%]), and 44 cases (58%) were bilateral. The most frequently diagnosed causes were peripheral neuropathy (21 cases [28%]), previous trauma (8 [11%]), and radiculopathy (7 [9%]); 32 cases (42%) were cryptogenic. Electromyography consistently showed irregular 50-millisecond to 1-second bursts of normal motor unit potential firing at 2 to 200 Hz accompanying the movements. Pain occurred first in nearly all cases and was more distressing to patients than the movements. Both components were difficult to treat, with no consistent benefit from a variety of drugs and therapeutic modalities. The syndrome persisted in most patients (83%) during the mean follow-up of 4.6 years, suggesting low likelihood of spontaneous resolution.
Conclusions Painful legs and moving toes syndrome is a debilitating clinical syndrome, not because of the movements but rather because of the pain, which often is refractory to treatment. Segmental lower limb involvement is most common, and neurophysiologic findings support a pathophysiologic process localizing to a central generator at the spinal cord or brainstem level.
Painful legs and moving toes” (PLMT) is an uncommon movement disorder in most clinics; it is characterized by pain in 1 or more limbs and accompanied by repetitive, nonrhythmic digit movements. It was first described in the lower limbs by Spillane et al in 1971.1 “Painful arms and moving fingers” was recognized in 1985,2 followed by a single case of “painful mouth and moving tongue.”3 The lower limbs appear to be more commonly affected than the upper limbs.4 Pain severity ranges widely in patients, from constant discomfort to unbearable pain, and occasional painless variants have been described.1,5 Thus far, case reports and relatively small case series (<20 cases) constitute the existing literature.6,7
Case reports have suggested that an underlying lesion may be identified at the level of the peripheral or central nervous system, such as peripheral neuropathy, radiculopathy, or spinal cord lesions.1,6- 8 However, these disorders are fairly common in the general population, while PLMT is rare. Thus, it is curious as to why some patients manifest PLMT. An underlying genetic predisposition seems unlikely because family history of PLMT is almost unheard of.9 The exact pathophysiologic mechanism remains unclear, and the natural history outcome is not yet reported. Recommended treatment options are also unknown for the pain or for the involuntary movements and consist largely of anecdotal case reports.10
Therefore, we sought to better characterize PLMT by review of the large number of cases evaluated at the neurology outpatient clinic of our tertiary referral center during nearly 2 decades. We depict the baseline demographic and clinical features of the PLMT syndrome, with a particular focus on neurophysiologic findings, treatment response, and long-term outcomes.
This study was approved by the Mayo Clinic Institutional Review Board. The electronic medical record database at Mayo Clinic, Rochester, Minnesota, was queried to identify all cases with the search terms painful legs and moving toes, painful arms and moving fingers, and their variants between January 1, 1983, and April 1, 2011. Case records were reviewed to extract demographic, clinical, neuroimaging, and neurophysiologic data. Inclusion criteria included at least 1 evaluation by a neurologist who rendered the diagnosis of “painful legs moving toes” or its variants and complete medical records with sufficient data for abstraction.
Statistical analysis was performed using basic descriptive terms, including mean, median, and range. Tests of 2 proportions were performed using the Wilcoxon rank sum test, and comparison of 2 means in the case of small sample sizes was performed using the 2-tailed unpaired t test. Regression models to identify clinical and demographic predictors of response to pharmacologic treatment and improvement at the time of follow-up were tested. All tests were performed under 2-sided alternatives with a predetermined level of significance of .05. The statistical program STATA (version 11.1; StataCorp) was used to perform all analyses.
During the 18-year period, 76 cases were identified (of which 50 patients [66%] were women), with a mean age at onset of 58 years (range, 24 to 86 years) and a mean age at evaluation of 63 years (range, 26 to 88 years) (Table 1). Cases were most commonly diagnosed in a movement disorders clinic (28 cases [37%]), followed by a general neurology clinic (19 [25%]), a neuromuscular clinic (15 [20%]), and a peripheral nerve clinic (7 [9%]). Lower limb involvement accounted for most cases (87%), followed by both upper and lower limb involvement (13%). None had isolated upper limb involvement. Symptom onset was more often left-sided (31 patients [41%]) compared with right-sided (24%) or bilateral (33%). Symptoms eventually became bilateral in 44 cases (58%).
Pain was noted at presentation in 70 of 74 cases (95%) and was more distressing than movements in all but 1 case. Four cases were pain free. Pain information was not available for 2 cases. The pain description varied widely (Table 2) but was overall constant within an individual patient. Exacerbating features for pain were identified in 44 cases (58%), including various positions (sitting, walking/weight bearing, or bending), diurnal variation (night or day), cold temperature, external pressure (blankets or shoes), and the Valsalva maneuver. These same factors could alleviate pain in 29 of 76 cases (38%). None of these affected movement severity. One patient had pain and movements exacerbated or relieved by a full or empty bladder, respectively, suggesting a central integration mechanism with autonomic nervous system involvement.
The various movement descriptions are listed in Table 2, but movements tended to be constant, to wax and wane in severity, and to be stereotypical to a particular individual. For a sample of the movements, see the video. In 1 case, pain and movements started in one leg, then moved to the other side and resolved entirely in the original limb. Another patient sustained trauma to one leg but developed PLMT only on the contralateral side. Maneuvers to alter pain did not seem to affect digit movements, which could usually be voluntarily suppressed for a short period or by applying pressure to the sole of the foot.
Peripheral neuropathy accounted for the most common identifiable cause of the pain (in 21 of 76 cases [28%]) (Table 3), of which approximately one-third were small-fiber neuropathy (6 of 21 cases) followed by limb trauma (8 of 76 [11%]) and radiculopathy (8 of 76 [11%]). Limb trauma ranged from mild to severe, including minor falls, compartment syndrome, and a limb/pelvic fracture. Two cases were attributed to common peroneal neuropathy and another to severe ischemic peroneal and tibial neuropathy following lower limb embolus. There were single cases temporally linked to procedures: epidural injection, myelography, or toe surgery. Medication withdrawal was identified in 2 patients—one after cessation of neuroleptic therapy and the other after cessation of anxiolytic therapy (buspirone hydrochloride and hydroxyzine pamoate) with coincidental small-fiber neuropathy. One patient reported that a family member had similar movements (not observed). In 32 of the 76 cases (42%), the cause of PLMT was not identified (cryptogenic) (Table 3). When a cause could be identified, the pain usually commenced at that time or within a few days of the putative underlying cause (range, 0-24 months). Movements typically accompanied pain at onset but could emerge after months or even years (range, 0-20 years). Approximately half the cases (35 of 76 [46%]) had an identifiable neurogenic cause of pain.
The medical history included cancer (12 cases), hypothyroidism (11), type 2 diabetes mellitus (7), depression (6), hysterectomy and/or bilateral oophorectomy (6), low vitamin B12 level (5), gastric bypass (3), thrombosis (3), anemia (3), and suicidal history (2). However, no patients attempted suicide after symptom onset. No patients had a history of multiple sclerosis or previous chronic pain syndromes. Of note, almost one-quarter (18 of 76 patients [24%]) had an autoimmune disease, including hypothyroidism, vitamin B12 deficiency, idiopathic thrombocytopenia, primary ovarian failure, and rheumatoid arthritis.
Neurologic examination results were otherwise normal (apart from the movements) in 16 cases (21%). In the remaining 60 patients, findings were mild, and almost half (25 of 60) were markedly asymmetric. Neurologic conditions identified were sensory peripheral neuropathy (33 patients, of which 16 [48%] had reduced vibration sense), reduced or absent reflexes (typically ankle jerks) (28 of 60 patients), mild distal limb weakness (12 of 60), hyperreflexia (5 of 60), and impaired rapid alternating movements (5 of 60). Coexistent movement disorders, such as postural tremor (3 cases) or rest upper limb tremor (1), were rarely present.
Most patients had normal routine blood test results apart from 5 cases reporting a low vitamin B12 level. Neuroimaging results of the brain and spinal cord were essentially normal in all cases (44 of the 76 cases included imaging results). Two patients underwent measurement of somatosensory evoked potentials (SSEPs), and both had abnormal results despite normal spine magnetic resonance images. In the first patient, the tibial SSEP suggested a lesion in the central sensory pathways above the lumbar level. In the second, the left tibial lumbar potential showed low amplitude with normal latency, and the N20 waveform was dispersed and prolonged, indicating a lesion in the central proprioceptive pathways rostral to the cervical cord.
Routine nerve conduction studies and needle electromyography (EMG) were performed in 63 cases. The nerve conduction study results were normal in 33 of the 63 cases and the remainder showed mild abnormalities, including chronic radiculopathy (usually at L5) (14 of 63 cases), peripheral neuropathy (9), cramps and fasciculations (4), isolated fasciculations/fibrillations (5), mononeuropathy (3), and evidence of prior poliomyelitis (1).
The neurophysiologic features of the movements were assessed by EMG or in the movement laboratories in most cases. The digit movements consisted of irregular (not truly rhythmic) phasic bursts of varying duration (range, 50 milliseconds to 1 second) associated with cocontraction and correlating with digit or distal limb movements. There were broad frequencies of these repetitive movements, ranging from 2 to 200 Hz. The reporting neurologist frequently commented that the observed movements were characteristic for PLMT. One case involved associated myoclonus recorded on surface EMG in the ipsilateral lower limb.
Pain, rather than the digit movements, was the primary patient-reported symptom. Despite targeting this pain with a vast array of drugs, few patients had gratifying pain relief (Table 4). One patient experienced partial pain relief with a combination of gabapentin enacarbil, levodopa, and lorazepam. A few benefited from localized therapy (epidural corticosteroid or local tibial nerve blockade) or physical therapies (transcutaneous electrical nerve stimulation, massage, cold soaks, chiropractor-delivered therapy, or acupuncture). Surgical intervention did not help 5 patients, one of whom underwent multiple spinal laminectomies and fusions, and significantly worsened the pain in another patient.
Patients typically reported that the digit movements were overshadowed by the pain except for complaints related to rubbing on footwear. In only a few cases were the digit movements targeted. These movements were attenuated by treatment with clonazepam, pramipexole dihydrochloride, or ropinirole hydrochloride in a few cases (Table 4); however, treatment with these drugs did nothelp the pain. Botulinum toxin injections were rarely helpful, although success may have been limited by the failure of patients to return for repeated injections and escalating doses.
There appeared to be no evidence that this syndrome spontaneously resolved over time. The syndrome persisted in most patients (63 of 76 [83%]) during the mean follow-up time of 4.6 years, from symptom onset to last follow-up. Symptoms were mostly stable (31 of the 76 cases), worse (1), or better (7) and were not clearly determined in 21 cases.
Overall, one-third of patients (25 [33%]) responded to pharmacologic treatment for movements or pain (Table 5). There were no differences observed between treatment responders and nonresponders on key demographic variables, including sex, age at onset, presence of pain, and documentation of normal neurologic examination results. There was also no significant difference on these variables between patients whose clinical course worsened vs those whose course stabilized or improved.
This series represents nearly 2 decades of PLMT ascertainment (76 patients), and several themes become apparent. First, the majority of patients with PLMT were female (66%) and most were middle-aged. The predominant lower limb involvement reflects a similar observation in the literature.4 Although the movement aspect of PLMT tended to attract attention, patients were much more distressed by the pain. This syndrome appears to be a heterogeneous disorder. The pain often was neuropathic but with frequent exceptions. Thirty-two patients (42%) had no identifiable cause for the pain. This was a novel problem for these patients because no history of chronic pain syndrome or concern for a psychogenic cause emerged. Other movement disorders were largely absent. Finally, in this large series, when multiple treatment modalities were tried, they were generally inadequate. Strategies to relieve the pain rarely did more than produce modest benefit, whereas treatments directed at the movements were effective in a few cases. The movements followed the pain in almost all patients. The pain triggering the movements was mostly severe.
The clinical picture in these patients, as well as in previous reports,11- 14 suggests a potential origin in the peripheral nervous system, that is, a neurogenic origin. Peripheral injury to provoke movement disorders is controversial but has been previously described in hemifacial spasm,15 “jumpy stump” syndrome after limb amputation,16 belly dancers' dyskinesias,17 and complex regional pain syndrome with dystonia.18,19 We note that the most commonly identified causes were peripheral neuropathy or antecedent trauma in our cases, with a temporal relationship to PLMT, and the subjective description of the pain (Table 2) was often characterized by neurogenic terms (eg, tingling, burning, and shooting/lancinating). This raises suspicion as a potential mechanism. In this series as well, the condition in some patients was triggered by musculoskeletal injury or was cryptogenic.
However, the evolution of the disorder suggests central nervous system integration to perpetuate the pain and mediate the moving digits. Several lines of evidence support this. Although the initial disorder started in one limb, it eventually became bilateral and segmental in most cases (Table 1). In 2 cases, PLMT involved only the limb on the opposite side of the lesion. Most cases had marked asymmetry of neurologic examination deficits despite bilateral PLMT symptoms. In addition, abnormal SSEPs in 2 cases despite normal imaging demonstrate involvement at the spinal cord or perhaps brainstem level. The occurrence of myoclonus in 1 case also suggests a central lesion. One may speculate that these observations support a process of neuromodulation with central “hardwiring” underlying PLMT. Thus, our study findings provide impetus to the next step of studying central generators of disease.
The pain and movements were somewhat dissociated. Pain reduction in some cases did not reduce the movements in parallel. Conversely, botulinum toxin injection abolished movements, but pain persisted in a different case. Responses to medications directed at the pain were unfortunately never dramatically beneficial. The few responders appeared to have no distinguishing features (Table 5). That almost one-quarter of patients had an identifiable autoimmune disorder provokes interest because this could suggest an immune-mediated process, with precedents in the literature stating that dystonia may be an immune-mediated condition.20 Apart from one patient who failed a trial of intravenous immunoglobulin therapy, immunosuppressive therapies were not tried. This may present an avenue to explore future therapies.
The relationship of PLMT to certain dystonic conditions deserves comment. Dystonia is proposed to result from a combination of sensorimotor system neuroplasticity, abnormal sensory processing, and impaired motor inhibition, leading to distorted motor outflow.21 Some of our patients had dystonic toe posturing and aching limbs and could voluntarily suppress movements consciously or by pressure on the foot, similar to dystonia. The PLMT syndrome specifically overlaps with the dystonia of complex regional pain syndrome (also called reflex sympathetic dystrophy). However, unlike that disorder, there were no recorded skin changes or altered temperature.22 Although posttraumatic dystonia was recently proposed to be psychogenic,23 PLMT is extremely unlikely to be so. The dynamic movements of PLMT are almost impossible to perform voluntarily, either because of the pattern of recruitment or the excessive voluntary motor unit firing rate on EMG. None of the cases had red flags for psychogenic disorders, such as multiple somatic complaints, variability in presentation, history of physical or sexual abuse, or pseudoseizures.24
The limb injuries and neuropathic processes of PLMT did not appear to have any unique features that should have predisposed to such a disorder. Many of the provocative conditions or injuries are fairly common but rarely lead to PLMT syndrome. There does not seem to be a genetic predisposition because a family history of PLMT was absent in these cases and present in only a single previous case report.9
We found no single electrophysiologic signature. The EMG findings were similar to those reported by others, namely, that motor unit discharges generally fire in time with the digit movements as irregular or semirhythmic phasic bursts with cocontraction. The burst duration lasted 50 milliseconds up to 1 second, within the range others have reported (from 10 milliseconds up to 2 seconds).14,25 However, rates of motor unit discharge in our cohort ranged as high as 200 Hz compared with previously reported slower ranges of 1.5 to 30 Hz.1,8,12,14,25
This article represents the largest reported PLMT case series to date, to our knowledge, all evaluated at a single institution, which allowed a comprehensive overview of the condition. Painful legs and moving toes syndrome is predominantly a debilitating pain syndrome, and most patients require pain management rather than treatment of the movement of their affected digits. Treatment results for the pain and movements are overall disappointing, in stark contrast to anecdotal successes in the literature. The persistence of pain on long-term follow-up in many cases provides speculation for central remodeling and hardwiring of pain, making the condition very difficult to treat. This central hardwiring also presumably drives the moving digits via a generator at the spinal cord or brainstem level. The combination of electrophysiologic findings and normal imaging results musters support for an underlying central segmental pathophysiologic process that propagates the pain and movements.
Correspondence: Anhar Hassan, MBBCh, FRACP, Department of Neurology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (email@example.com).
Accepted for Publication: January 31, 2012.
Published Online: April 9, 2012. doi:10.1001/archneurol.2012.161
Author Contributions:Study concept and design: Hassan and Mateen. Acquisition of data: Hassan and Coon. Analysis and interpretation of data: Hassan, Mateen, and Ahlskog. Drafting of the manuscript: Hassan and Mateen. Critical revision of the manuscript for important intellectual content: Mateen, Coon, and Ahlskog. Statistical analysis: Mateen.
Financial Disclosure: None reported.
Online-Only Material: The videos are available here.