Abstract
Background: Paroxysmal kinesigenic dyskinesia (PKD) is a rare disorder characterized by short episodes of involuntary movement attacks triggered by sudden voluntary movements. Although a genetic basis is suspected in idiopathic cases, the gene has not been discovered. Establishing strict diagnostic criteria will help genetic studies.
Methods: The authors reviewed the clinical features of 121 affected individuals, who were referred for genetic study with a presumptive diagnosis of idiopathic PKD.
Results: The majority (79%) of affected subjects had a distinctive homogeneous phenotype. The authors propose the following diagnostic criteria for idiopathic PKD based on this phenotype: identified trigger for the attacks (sudden movements), short duration of attacks (<1 minute), lack of loss of consciousness or pain during attacks, antiepileptic drug responsiveness, exclusion of other organic diseases, and age at onset between 1 and 20 years if there is no family history (age at onset may be applied less stringently in those with family history). In comparing familial and sporadic cases, sporadic cases were more frequently male, and infantile convulsions were more common in the familial kindreds. Females had a higher remission rate than males. An infantile-onset group with a different set of characteristics was identified. A clear kinesigenic trigger was not elicited in all cases, antiepileptic response was not universal, and some infants had attacks while asleep.
Conclusions: The diagnosis of idiopathic paroxysmal kinesigenic dyskinesia (PKD) can be made based on historical features. The correct diagnosis has implications for treatment and prognosis, and the diagnostic scheme may allow better focus in the search for the PKD gene(s).
Paroxysmal kinesigenic dyskinesia (PKD) is the most common type of paroxysmal dyskinesia, although the precise prevalence is unknown. This condition is characterized by attacks triggered by initiation of voluntary movements.1–3⇓⇓ Other characteristic features as previously defined include brief duration of attacks (typically <5 minutes), preserved consciousness between attacks, and clinical response to antiepileptic drugs (AEDs) based on anecdotal reports.
PKD can further be divided into idiopathic or secondary cases. The most common underlying etiology for secondary cases comprises multiple sclerosis, cerebral vascular insufficiency, metabolic derangements (hypo- or hyperglycemia, hypocalcemia), or previous trauma.4 In idiopathic cases, patients usually have normal neurologic examination between attacks. Family history is commonly noted in idiopathic PKD, but sporadic cases are also reported. There is a wide interest in the genetics of PKD. In 1997, the first linkage analysis for infantile convulsion and choreoathetosis (ICCA) syndrome was reported to the pericentric region of chromosome 16.5 Many affected members in this family had infantile convulsions (benign seizure syndrome with seizures only during infancy), but it was noted that some family members later developed attacks of hyperkinetic movement disorders (although the clinical features were not described in detail). Subsequently, multiple families with either PKD or ICCA syndrome from multiple populations were linked to chromosome 16, within or near the original ICCA linkage.6–10⇓⇓⇓⇓ PKD and ICCA are now believed to be the same disorder, based on clinical similarity and the overlap of the genetic loci. Many patients or nonaffected family members in PKD kindreds have infantile convulsions, and infantile convulsions are now thought to represent a variable expression of PKD. A different locus, termed EKD2, also on chromosome 16 was reported in an Eastern Indian kindred in 2000.11 Recently, a British family with PKD, not linked to either of these two loci, was reported, and a third locus is suspected.12
Based on its paroxysmal nature, some have postulated that paroxysmal dyskinesia is a channelopathy.13 Episodic ataxia-1, a potassium channel disorder with responsible gene (KCNA1) on chromosome 12p13, manifests PKD in some patients.14,15⇓ Both PKD and EA-1 start in childhood to early adolescence. Attacks are triggered by kinesigenic stimuli, and the attacks are brief, lasting a few seconds to a minute. However, to date, no PKD gene has been identified.
In patients with PKD, attacks are intermittent, and typically there are no physical findings on examination between attacks. Therefore, diagnosis is based solely on history. Furthermore, the physiologic mechanism that leads to the attacks is unknown, and there are no known objective studies that would confirm the diagnosis. Yet, accuracy of diagnosis, that is, phenotyping, is crucial in genetic studies. Improved diagnostic criteria for idiopathic PKD or a subset thereof will be useful for further genetic, physiologic, and therapeutic studies.
Most clinical analyses of PKD are either case reports, that is, sporadic cases, or analyses of families accompanying genetic studies. Here, we report the clinical characteristics of a large series of idiopathic PKD, directly interviewed by a single neurologist. Based on this broad database, we propose refined diagnostic criteria for idiopathic PKD.
Materials and methods.
We reviewed the clinical features of 121 affected individuals from 73 kindreds, all of whom were referred to us for genetic studies from neurologists, pediatric neurologists, or movement disorder specialists, with the clinical diagnosis of idiopathic PKD. All cases included here had “movement-induced” paroxysmal attacks of involuntary movements without loss of consciousness and no evidence of other neurologic disease based on workup by the referring physician.
One hundred six patients, 142 at-risk members, and 17 who married into their families were directly interviewed by the same neurologist using a standardized protocol. We obtained surrogate history on 16 other affected individuals and 57 at-risk members. Surrogate history was used only when the affected or at-risk member could not directly provide a history (due to being a child or deceased). In addition, 17 patients’ attacks were observed by the same neurologist either in person or by videotape documentation.
The evaluations documented the following: age at onset, trigger and precipitant of attacks (“trigger” was defined as an event that could cause attacks directly, whereas “precipitant” was defined as an event that might lower the threshold of an attack induced by a trigger), presence and description of aura, duration and phenomenology of attack, pain during attacks, alteration of consciousness, frequency of attacks, course of illness, medications tried and response, any other associated neurologic disorders, and family history.
For each data field, we plotted a histogram to see if there was any clustering for that particular data field. For example, the figure is a histogram showing the distribution of age at onset (of involuntary movement attacks, and not infantile convulsions). There were two distinctive clusters and outliers. The majority of patients had attack onset between ages 4 and 20, whereas 10 patients had attack onset before age 1. In further reviewing these 10 patients who had their attack onset within the first year of life, they had a slightly different clinical profile compared with the majority of cases (see below). We repeated this process for each data field described above to extract the criteria that would define a homogeneous group. We reviewed the outliers for each data field to see if they would fit better to an alternative diagnosis and to see if they were outliers for other data fields.
Figure. Histogram for age at attack onset in all probands initially referred as paroxysmal kinesigenic dyskinesia.
Results.
Clustering was observed in the following categories: age at onset, trigger (kinesigenic attack 96%), duration of attack (<1 minute in 95%), absence of pain during attacks (97%), absence of loss of consciousness (98%), and clinical response to medication (86% to either carbamazepine or phenytoin). Based on this, we created working criteria for idiopathic PKD as follows: age at onset between 1 and 20 years, identified kinesigenic trigger for the attacks, short duration of attacks (<1 minute), no loss of consciousness or pain during attacks, and control of attacks with phenytoin or carbamazepine. This group was further divided into whether the patient had a family history of similar disorders or not.
The infantile-onset group was defined as patients who had their attack onset within the first year of life. The outlier group included patients who did not meet the criteria for either the idiopathic or the infantile-onset group. Patients in the outlier group were further evaluated to establish an alternative diagnosis. Special attention was paid to see how many of these patients still had PKD as the most likely diagnosis to determine if we needed to modify the criteria for the idiopathic group.
Idiopathic PKD group.
Demographics.
There were 95 patients in this category (51 men and 44 women) (table 1). Sixty-four had a family history (23 kindreds), and 31 were sporadic. A majority of these patients had their initial attacks between ages 6 and 15 years (88%). The average age at onset was 11.6 ± 3.5 years, and attack onset typically coincided with puberty.
Table 1 Clinical characteristics of patients with classic paroxysmal kinesigenic dyskinesia
Trigger/precipitant.
The kinesigenic trigger was typically a whole-body activity such as initiation of standing, walking, or running. Turning in bed and jumping into a pool were also reported triggers. One patient eloquently stated, “It is when preparing to move very quickly, as if I am waiting for the gun to start the race, yet, the timing of that signal is random.” Some patients reported focal movement, such as stretching their arm, to be a trigger. Startle was a rare trigger. However, some patients reported that startle in combination with a subsequent “jolt of adrenaline” could be a trigger.
Anxiety seemed to lower the threshold in many patients; 62% reported that they noticed the attacks were more frequent when they were stressed, anxious, or in public. A common scenario for an attack was being called upon during class to come up to the blackboard or up to answer the telephone. Caffeine and physical fatigue were less frequent as precipitants. Alcohol itself was not reported to be a precipitant, although some patients reported that the relaxation caused by alcohol could possibly reduce attacks. Yet other patients reported that being “hung-over” the next day might worsen attacks. Other rare precipitants included cannabis, menstruation, cold weather, humidity, hunger, and external motion such as being in a fast car.
Aura.
A premonitory sensation of imminent attack was reported by 82% of patients. This was typically a general feeling that they could not explain: “upset stomach feeling when you are called by school principal,” “a feeling that a butterfly is in my stomach,” or “an electricity in my head.” Some patients felt focal numbness or a tingling sensation in the periphery (lower extremities, finger tips), where the involuntary movement began. The sequence of events was as follows: sudden quick movements, premonitory sensation, then attack. Sixty-eight percent reported that they could possibly minimize their attack if they stopped their movement when they first felt the premonitory sensation.
Attack characteristics.
Attack duration was typically very short, with 93% reporting their attacks as <30 seconds and 100% <1 minute as defined by the criteria. The phenomenology of the attacks, deduced from descriptions, indicated dystonia (57%), chorea (6%), ballismus (1%), not-otherwise-specified hyperkinesias (3%), and combinations of different hyperkinesias (33%). Attacks were unilateral in 36%, unilateral but alternating sides in 12%, bilateral in 35%, and either unilateral or bilateral in 18%.
AED responsiveness.
Sixty-seven percent of the patients received pharmacologic treatment. Carbamazepine and phenytoin were the most common medications used (52% for both among those who took medication). Other medications prescribed to these patients included benzodiazepines or other anticonvulsant medications such as valproate, gabapentin, lamotrigine, and levetiracetam. They reported moderate responses, and all eventually switched to either carbamazepine or phenytoin, either because of side effects or because of suboptimal response. One patient had complete resolution of the attacks on acetazolamide.
Natural history and medical evaluation.
Among patients who were older than 20 years (n = 81), 27% reported complete remission, and an additional 25% reported marked (>50% decrease in attack frequency by self-report) spontaneous improvement. Six patients (7%) reported worsening in adulthood. One of these patients had remission in his 30s, but the attacks came back in his 40s. The most common age at remission was in their 20s, but some patients had remission as late as in their mid-30s.
Thirteen of the affected women had a history of pregnancy; seven of them (54%) reported improvement during pregnancy, and one reported that her attacks were more severe and frequent during pregnancy. The others either had no notable changes or could not recall what happened to their attacks during their pregnancies.
The times spent from first seeking medical attention to obtaining the correct diagnosis varied, ranging from the first visit to 20 years (average 4.8 ± 6.0 years). Alternative diagnoses given during this period included seizures, pseudoseizures, psychogenic disorders, tics, subclavian steal syndrome, neuropathy, Sydenham chorea (St. Vitus’s dance), anxiety, and malingering. Thirty-six percent of familial patients never sought medical attention either because their attacks were mild or they were familiar with symptoms from seeing other family members. We identified five family members who admitted having attacks but who were previously reported to be unaffected by other family members. There were two female asymptomatic gene carriers in the familial group when we assumed autosomal dominant transmission.
Associated intermittent neurologic disorder.
History of infantile convulsions was common, as previously reported (table 2). As there were many first- and second-degree relatives of probands with infantile convulsions, including children with a younger age at onset than idiopathic PKD, we analyzed whether there was anybody in the kindred with intermittent neurologic disorders (see table 2, right columns). A history of infantile convulsions within the kindred was significantly more common in the familial group. Other associated neurologic disorders either in patients or in family members included migraine, writer’s cramp, and essential tremor.
Table 2 Associated history of other intermittent neurologic disorders in idiopathic paroxysmal kinesgenic dyskinesia
Comparison between familial and sporadic patients.
Statistical differences (p < 0.05, χ2 test) between familial and sporadic patients were observed in male-to-female ratio and attack frequency. Statistical differences were also found in the frequency of infantile convulsions between the familial and sporadic kindreds.
Male-to-female differences.
Females had a much higher rate of remission than males (p < 0.05, χ2 test) (table 3).
Table 3 Male and female differences in prognosis among patients over age 20
Infantile-onset group.
There were 12 patients (10 kindreds) in this group (table 4). Two of them had a family history of PKD. The earliest onset was 7 days of age. This girl was diagnosed with a seizure disorder by her pediatrician and was treated with phenobarbital. Her father, who had PKD, commented, “What she had was exactly what I had. She did not look ‘out of it,’ and she was playing right after she had this attack.” These infants tended to have their attacks while crawling, walking, sitting in a car seat, playing, when they were excited, and during sleep. Parents reported no loss of consciousness during attacks. The attacks were usually <1 minute in duration and sometimes only a few seconds. They had variable response to AEDs. Attacks remitted in three cases: two cases before age 5 and at age 18 in one case.
Table 4 Clinical characteristics of infantile-onset paroxysmal dyskinesia
Outlier group.
There were 14 patients in this group not meeting the criteria for idiopathic or infant groups (table 5). Two patients who started to have attacks at ages 27 and 25 were otherwise typical for the idiopathic group and had a family history that met the criteria for the idiopathic group. Therefore, we diagnosed these patients as having PKD. The remaining 12 patients had possible alternative diagnoses. Diagnosis of psychogenic movement disorder was made when the movement disorder was inconsistent over time or was incongruent with a classic movement disorder.16 In addition, the diagnosis was supported by the presence of other psychogenic neurologic signs, multiple somatization in the past, or an obvious psychiatric disturbance. A diagnosis of paroxysmal nonkinesigenic dyskinesia (PNKD) was made when the attacks occurred spontaneously without any specific precipitant.17 The diagnosis was supported by presence of family history, attack precipitation with caffeine, tobacco, or fatigue, typical duration of attack (few minutes to 4 hours), and lack of AED efficacy.18 Paroxysmal exertion-induced dyskinesia diagnosis was made when attacks were induced by prolonged exercise.17
Table 5 Clinical characteristics of outlier patients
Even when they reported kinesigenic trigger, on further questioning, some patients had attacks spontaneously in addition to those induced by a kinesigenic trigger. The kinesigenic triggers in some were task specific and differed from velocity-dependent gross movements involving the entire body axis, which was more typical in the idiopathic PKD group. For example, one patient with psychogenic movement disorder had attacks with running, moving one limb, tracking with eyes, and writing cursive, but not with writing in print or standing up from a seated position. Age at onset, attack duration, and medication responses were variable in this group.
Candidate genes.
We have ruled out approximately 50 genes by sequencing, including ATP2A1, SRCAP (Snf2-related CBP activator protein), STX4A, GPT2 (glutamic pyruvate transaminase 2), and those previously published.7,9,10⇓⇓ However, mutations in promotors or intronic sequences of these genes or intragenic regions might have been missed.
Discussion.
We report a large clinical study of patients with idiopathic (familial and sporadic) PKD directly interviewed by a single neurologist. Some of the issues previously unknown (natural history, gender differences, influence of pregnancy, prevalence of associated neurologic disease, and differences between familial and sporadic PKD) were analyzed in our study. Furthermore, we were able to extract the crucial factors defining a homogeneous subset of patients with idiopathic PKD.
In our series, 27% of patients had complete remission and another 25% had marked improvement, confirming the previous clinical impression.19–21⇓⇓ There was a gender difference in prognosis: woman having a better prognosis and higher chance of complete remission. Many women also reported improvement of attacks during pregnancy, suggesting that there may be some hormonal influences in this condition.
The clinical characteristics between familial and sporadic cases were similar. The male-to-female ratio was higher in sporadic cases, as reported previously in the literature, but we did not find this difference in familial cases. The attack frequency was higher in the sporadic cases, but this probably represents ascertainment bias because clinically milder subjects were included in familial cases.
In addition, familial kindreds had a higher rate of infantile convulsion in both patients themselves as well as many first- and second-degree relatives. These relatives included children younger than the onset of typical PKD, suggesting that they are gene carriers (or that infantile convulsion is an alternate phenotype, or both). However, we took a conservative approach and considered PKD to be in the familial group only when they had additional family members with PKD attacks.
Migraine, potentially a channelopathy in at least some cases, is a very prevalent disorder (estimates ranging from 3 to 35% of the population), and thus it is difficult to prove a higher association in PKD.22,23⇓ The number of patients or family members with writer’s cramp in our study is too small to make any statements, but considering the low prevalence of focal dystonia (estimated to be 30 per 100,000 population), there may be an association.24 A family with paroxysmal exercise-induced dystonia, writer’s cramp, and rolandic epilepsy linked to chromosome 16 and potentially allelic to the location of the PKD locus was previously reported.25
Diagnosis of intermittent disease is difficult, and our results show that the average time spent to establish a diagnosis in these patients was on the order of years. Nevertheless, we believe that the diagnosis of PKD can be made more rapidly on detailed clinical history. In examining the outlier group, we could identify an alternative diagnosis for most patients.
The only two patients we diagnosed as having PKD despite being in the outlier group had onset of their attacks at ages 25 and 27 but otherwise met the criteria and had family history of PKD. In the literature, there is a report of a patient with age at onset at around 40 years.21 The diagnosis of idiopathic PKD does not require family history, but if there is a family history of PKD, we can be more confident about the diagnosis. Stringent adherence to age criteria may not be as important in the familial cases. However, the age requirement is crucial in sporadic cases. PKD is unlikely to start after age 20 (nearly 90% of cases had onset younger than 15 years), and the patient is much more likely to have an alternative diagnosis. Secondary cases were already excluded in our series, but in clinical practice, secondary causes need to be pursued in patients older than 20.4 Neuroimaging and laboratory studies may be necessary to rule out multiple sclerosis, cerebral vascular insufficiency, or metabolic derangements (glucose, calcium). Previous history of trauma needs to be sought.
Based on our results, the clinical criteria we propose are as follows:
Identified kinesigenic trigger for the attacks
Short duration of attacks (<1 minute)
No loss of consciousness or pain during attacks
Exclusion of other organic diseases and normal neurologic examination
Control of attacks with phenytoin or carbamazepine, if tried
Age at onset between 1 and 20 years, if no family history of PKD
Our diagnostic criteria fit well to the previously reported series of PKD.1,5–8,11,12,26–28⇓⇓⇓⇓⇓⇓⇓⇓⇓ Nearly 100% of familial PKD patients met the criteria we propose, emphasizing the clinical homogeneity of familial (genetic) cases, whereas some sporadic cases did not. For example, Case 22 in one report28 had an attack duration lasting 10 seconds to 7 minutes, poor clinical response to carbamazepine, and attacks during sleep.
We have identified a small group of patients who present in infancy. Infantile-onset PKD has been previously reported.29 The infantile-onset group was less uniform in clinical manifestation and did not meet all of the criteria proposed for idiopathic PKD. The kinesigenicity of the attack trigger was less clear. In some cases, parents observed attacks while they were sleeping. Clinical responses to carbamazepine were not as uniform. Differential diagnosis includes PNKD, shuddering attacks, or benign myoclonus of early infancy. PNKD commonly starts in infancy; attacks are triggered spontaneously and do not respond to carbamazepine.18 Both shuddering attacks and benign myoclonus of early infancy are characterized by very short duration of hyperkinetic attacks lasting a few seconds.30,31⇓ Two of our infantile-onset cases had a family history of PKD, and we suspect that at least these two cases have the same genetic alteration as idiopathic PKD, thereby expanding the phenotype of PKD. Infantile- onset paroxysmal dyskinesia may be a heterogeneous group, but gene discovery will help clarify if each infantile- onset paroxysmal dyskinesia case represents PKD, PNKD, or a separate entity.
There may be few biologic explanations for patients who have attacks deviating from our diagnostic criteria. First, the genetic defect may have variable expressivity, and some patients may demonstrate alternative clinical phenotype. Second, they may have different genetic basis. Third, they may be secondary cases, where the underlying pathology has not yet been identified, and fourth, they may be psychogenic cases. Genetic studies will help clarify this point.
Note added in proof.
We have subsequently cloned a gene causing PNKD that is predicted to encode a protein in a stress response pathway. 32
Acknowledgments
Supported in part by NIH grant RO1 NS43533 (Y.-H.F., L.J.P.) and a Sandler neurogenetics grant (Y.-H.F.). L.J.P. is an investigator of the Howard Hughes Medical Institute.
The authors thank the families for their participation and Drs. Mary Kay Bowen, Mitchell Brin, Francis Filloux, Johnathan Flanzbaum, Bennett Lavenstein, Nicholas Lenn, Deborah Raymond, Shehla Mohammed, and Samuel Tucker for patient referral. They also thank Devee Schoenberg for editing the manuscript.
- Received April 1, 2004.
- Accepted in final form September 1, 2004.
References
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.