Mail converted by MHonArc
2.6.10
| Search or browse years of Parkinsn Email list messages | |||||||
|
| |||||||
|
Mail converted by MHonArc
2.6.10
|
Parkinsn Current Topics
A Study of Hereditary Essential Tremor Bain et al.
Many, who later in life are diagnosed with Parkinson's disease first appear with tremor. Although no association has been discovered between essential tremor and Parkinson's disease, similarities exist.
Essential tremor has a hereditary component. Offspring of an essential tremor parent have a 46% chance of developing the disease.
Bain's study documents that onset predominately occurrs in the twenties and is progressive to include all by the age of 65. There is no evidence of the disease skipping a generation.
Tremor generally starts in the hands and progresses upward. Bain concludes, "Physicians probably underestimate the effect of tremor on patients' lives and overestimate the practical benefits of the medical treatments available."
The following study is presented for research purposes.
A study of hereditary essential tremor
P.B. Bain,[1] L.J. Findley,[2] P D. Thompson,[1] M.A. Gresty,[1] J.C. Rothwell,[1] A.E. Harding and C.D. Marsden[1]
[1] MRC Human Movement and Balance Unit, Institute of Neurology, London and [2] The Regional Centre for Neurology and Neurosurgery, Oldchurch Hospital, Romford, UK
Correspondence to: Dr P.G. Bain. MRC Human Movement and
Balance Unit, Institute of Neurology, Queen Square. London WCIN 3BG, UK
Summary
Twenty index patients with hereditary essential tremor and their kindreds were studied to define the phenotype of this condition. Ninety-three first degree and 38 more distant relatives were examined; 53 definite and 18 possible secondary cases were identified. The age of tremor onset was bimodally distributed with a median at ~15 years. Segregation analysis indicated autosomal dominant inheritance and penetrance was virtually complete by the age of 65 years. There were no examples of the disease skipping a generation. Men and women were affected in equal proportions. About 50% of cases were alcohol responsive. In the majority of families alcohol responsiveness was either consistently present or did not occur but in 20% of kindreds definite heterogeneity of responsiveness was encountered within each family. The typical phenotype was a mild symmetrical postural tremor of the upper limbs. Tremor of the legs, head, facial, voice, jaw and tongue occurred but never in isolation and rest, task specific (e.g. primary writing tremor) and orthostatic tremors were not found. Head tremor was invariably mild and 75% was of a 'no-no' type. Dystonia (torticollis and writers cramp) were not encountered, a finding which strongly suggests that many previous studies of 'essential tremor' were contaminated by cases of idiopathic or hereditary torsion dystonia. No association with Parkinson's disease was found but classical migraine occurred in ~26% of cases and co-segregated with tremor. The severity of arm tremor (assessed using a clinical rating scale and by scoring tremor in Archimedes spirals) and disability increased with advancing age and increasing tremor duration. but there was no correlation between age at tremor onset and either tremor severity or disability. Men and women were affected with equal severity. The sex of the affected parent had no influence on the severity of tremor or the degree of disability experienced by an affected child. Disability commenced in the second decade and progressively increased. All the index patients and 59% of the definite secondary cases had tremor induced disabilities. Eighty five percent of index patients and 38% of secondary cases also reported some degree of social handicap. Twenty-five percent of index patients and 12% of secondary cases had been compelled to change jobs or retire. Biological fitness was normal.
Key Words: hereditary essential tremor; phenotype; segregation analysis
Introduction
The concept of hereditary essential tremor evolved from the observations that tremor could be familial (Most, 1836), present in adolescence and persist throughout life (Sinkler, 1886). Many detailed reviews by Critchley (1949, 1972), Hassler (1953), Jager and King (1955), Larsson and Sjogren (1960), Marshall (1962), Murray (1981), Larsen and Calne (1983). Findley (1986), Hubble (1989), Lou and Jankovic (1991) Koller et al (1992) have established that the impact of essential tremor falls predominantly on the upper limbs and that it may produce tremulousness of the head, legs, trunk, voice, jaw, and facial muscles. Typical essential tremor manifests as a postural tremor (of the arms) but kinetic. intention and resting components have been reported (Critchley, 1949, 1972; Davis and Kunkle, 1951; Marshall, 1962; Findley, 1987). At first tremor may appear intermittently during periods of excitement (Critchley. 1949; Marshall. 1962) but it usually progresses to become permanent, although the amplitude can fluctuate. and remissions have been described (Kreiss, 1912). Tremor is inevitably worsened by emotion as well as hunger, fatigue and extremes of temperature (Critchley, 1949, 1972). Conversely, essential tremor is under some degree of voluntary control and can be suppressed during the performance of skilled manual tasks, at least for short periods of time (Critchley, 1949; Jager and King, 1955; Bain et al., l993b).
Several neurologists have considered a 'yes-yes' type of head tremor (tremblement affirmatif) to be characteristic of the condition (Findley, 1984; Lou and Jankovic, 1991), whilst others report that a 'no-no' tremor (tremblement negatif) was more common (Jager and King, 1955; Critchley, 1972) and complex (oblique) head tremors have also been described (Critchtey, 1949; Biary and Koller, 1985). There is further disagreement concerning the usual mode of onset and pattern of spread. Some authors have described an asymmetric onset in one or other hand, as the norm (Critchley, 1949) and Findley (1987) considered that the retention of this asymmetric pattern was inevitable; whilst other accounts have indicated that a symmetrical onset and picture were more usual (Larsson and Sjogren, 1960; Marshall, 1962). Most authoritative accounts have agreed that, once the upper limbs were affected, 'upward' spread to the head, face, tongue or jaw was more common than involvement of the legs (Critchley, 1949, 1972; Larsson and Sjogren, 1960; Marshall, 1962; Lou and Jankovic, 1991). However, hemitremulous states in which tremor appears in an arm and the ipsitateral leg have been documented, albeit rarely (Critchley. 1949; Larsson and Sjogren, 1960).
Inheritance
Hereditary essential tremor is believed to be caused by an autosomal dominant mutation (Critchley, 1949; Davis and Kunkle. 1951; Larsson and Sjogren. 1960) but the extent of penetrance is contentious. Larsson and Sjogren (1960) and Rautakorpi (1978) concluded from their own studies that complete penetrance had occurred 'by the age of 70 years or shortly thereafter' but others suggested that penetrance is incomplete (Marshall, 1962; Critchley, 1972; Findley, 1984).
Sporadic cases of essential tremor have also been widely reported (Critchley, 1949, 1972; Marshall. 1962; Hubble et al., 1989; Lou and Jankovic. 1991; Koller et al., 1992) and have always been considered to be the same entity as the hereditary form of the disease, an assumption that has never been formally questioned. The proportion of patients with essential tremor reporting that at least one other relative was affected has varied in different studies from 17 to 70% (Marshall. 1962; Hornabrook and Nagurney, 1976; Rautokorpi, 1978; Aiyesiloju et al., 1984; Rajput et al., 1984). In part these diverse figures are the result of relying upon patients' histories rather than an examination of their relatives.
Age of onset
Although essential tremor can occur at any age (Critchley 1949; Findley, 1984) the peak age or onset has not been well established. Critchley (1949) considered 'adolescence or early adult life' to be the most usual age of onset. Gerstenbrand et al. (1982) and Lou and Jankovic (1991) found a bimodal distribution with peaks in the second and fifth decades, while Larsson end Sjogren (1960) concluded that the disease seldom begins in youth or at an advanced age but often commenced at about the age of 50', an observation which in broad agreement with the mean ages of onset (37 years) obtained from the data of other studies (Critchley 1972; Koller et al., 1992). Critchley (1949). having studied the family trees in several early papers, pointed out that the phenomenon of 'anticipation', the tendency for tremor to present at an earlier age in successive generations was evident in these kindreds. No one else has found evidence for this observation (Hubble et al., 1989) and, although Larsen and Calne (1983) thought it might have been an artefact, this has recently become a controversial issue.
Epidemiology
Essential tremor has been established as a multi-racial disorder following the epidemiological surveys carried in Scandinavia (Larsson and Sjogren. 1960; Rautakorpi et al., 1982), the USA (Haerer et al., 1982; Rajput et al., 1984 Africa (Longe, 1985), Papua New Guinea (Hornabrook and Nagurney, 1976) and amongst the Parsi community in India (Bharucha et al., 1988). The overall prevalence of essential tremor within these populations ranged from 305 (Rajput et al., 1984) to 1700 (Larsson and Sjogren, 1960) per 100 and increased with age. A prevalence of between 0 (Haerer et al., 1982) and 6.7% (Bharucha et al., 1988)) has been found amongst people over 40 years old and between 8.37% (Larsson and Sjogren. 1960) and 12.6% (Rautaokorpi et al., 1982) in those over 70. The age-specific prevalence, increases with advancing age and the prevalence among people under 30 years old is reported to be less. (Rautakorpi et al., 1982). Only 14% of the cases of Larsson and Sjogren, (1960) and 24% of those of Rautakorni (1978) had an age of onset before 30 years. Similarly. the age-specific incidence is reported to increase after the age of 49 years and reaches a maximum (84 per 100 000) in the ninth decade (Rajput et al., 1984).
There is no consensus about the sex distribution of the disorder and in this regard it is interesting that sex chromosome abnormalities have been discovered in some patients (Baughman et al., 1973). The Swedish (Larsson and Sjogren 1960) and Finnish studies (Rautakorpi, 1978) produced female to male ratios of 0.5 and 0.71. respectively, but the reverse was found by Huerer et al., 1982) in the USA ( 1 for white and 1.24 for black populations) and Hornabrook and Nagurney (1976) in Papus New Guinea (2.06). Rajput et al. (1984) and Bharucha et al. (1988) did not detect any sex differences amongst their respective American and Indian populations.
Relationship to other conditions
Rigidity is widely acknowledged to be a useful sign differentiating essential tremor from Parkinson's disease, it is one of the three cardinal signs of the latter. However several authors have reported finding rigidity in patients with essential tremor' (Larsson and Sjogren, 1960; Salisachs. 1978; Salisachs and Findley. 1984) and a 'pill rolling' tremor (thought to be pathognomonic of parkinsonism) has also been described (Larsson and Sjogren. 1960; Hornabrook and Nagurney. 1976).
The possibility that essential tremor could be a forme fruste of Parkinson's disease or that the two disorders formed the extremes of a continuum were investigated by Cleeves et al. (1988), but no relationship between the two diseases was discovered, a view supported by the findings of Marttila: et al. (1984). However, other workers (Hornabrook and Nagurney, 1976; Barbeau and Pourcher, 1982; Roy et al., 1983; Geraghty et al., 1985; Lou and Jankovic. 1991) provide provocative evidence to suggest that the two conditions are related; a debate which is far from resolved, although the weight of current evidence suggests that the two conditions are unrelated (Pahwa and Koller, 1993).
Similarly, there have been numerous reports of an association between essential tremor and spasmodic torticollis or other forms of dystonia (Critchley, 1949. 1972; Couch 1976; Marsden, 1976; Baxter and Lal. 1979; Jankovic and Ford, 1983: Lou and Jankovic, 1991). Conversely. Koller et al. (1992) excluded patients with tremor and signs of dystonia (e.g. torticollis) from their studies of essential tremor an approach which was also taken by Larsson and Sjogren (1960) in their epidemiological survey. The remaining major clinical studies of essential tremor have all included patients with signs of dystonia, either intentionally or otherwise, and this problem is compounded by the observation that some patients with idiopathic torsion dystonia may exhibit tremor as their only clinical abnormality (Fletcher et al., 1990, 1991). There is one notable exception: Jager and King's (1955) informative description of a single large family in Utah with true hereditary essential tremor.
Tremors resembling essential tremor have also been documented in a variety of neuropathic conditions including IgM paraproteinaemic neuropathies (Smith et al., 1983,1984; Dalakas et al., 1984; Leger et al., 1992), types I and II hereditary motor and sensory neuropathies (Shahani et al., 1973; Dyck, 1975; Harding and Thomas, 1980), acute and chronic idiopathic demyelinating polyneuropathies (Thomas et al., 1969; Matthews et al., 1970; Shahani and Young, 1978; Dalakas and Engel, 1981) as well as a variety of other types off neuropathy (Said et al., 1982) and diseases of the anterior horn cells (Thomas, 1975). The mechanisms of tremor in these conditions are poorly understood and there has been controversy about whether or not there is a genetic association between the dominant gene for hereditary motor and sensory neuropathies and that for essential tremor (Shahani et al., 1973; Dyck, 1975; Shahani, 1984)
Focal, site or task-specific tremors
Another area of controversy is whether isolated tremors affecting parts of the body other than the arms are formes fruste of hereditary essential tremor. Instances of 'isolated' head(Larsson and Sjogren. 1960; Marshall, 1962; Critchley. 1972; Lou and Jankovic, 1991), tongue (Biary and Koller, 1987), voice (Hachinski, 1975; Massey and Paulson, 1985) and jaw (Frey, 1930; Critchley, 1949; Grossman, 1957) tremor have been described. In addition, some task-specific tremors have been considered to be formes fruste of hereditary essential tremor. For example, primary writing tremor has been deemed by some authors to be a variant of essential tremor (Rothwell et al., 1979; Kachi et al., 1985; Koller and Martyn, 1986; Rosenbaum and Jankovic, 1988) and by others to be a variant of writer's cramp (Ravits et al., 1985: Rosenbaum and Jankovic, 1988; Elbe et al., 1990). In addition, it is debated whether primary orthostatic tremor is or is not a separate entity to essential tremor (Thompson et al., 1986; Rothwell, 1989; Britton et al., 1992a). These issues await resolution.
Pharmacology
The responsiveness of essential tremor to alcohol is a characteristic but not unique feature of the condition (Critchley, 1949,1972; Davis and Kunkle, 1951; Ashenhurst, 1973; Sutherland et al., 1975; Findley, 1987; Koller et al., 1992) and the percentage of patients reported to respond has varied from 42% (Findley, 1987) to 75% (Koller et al., 1992). However, relief is temporary and tremor rebounds in an exaggerated form, so that the need for another drink arises (Critchley, 1949). Several reports have claimed that there is an increased incidence of alcoholism in patients with essential tremor (Massey and Paulson, 1978; Nasralla et al., 1982), but a prospective study concluded that this was not significantly different from other chronic neurological diseases or tremulous conditions (Koller, 1983). Injections of alcohol into the brachial artery have no ameliorating effect on tremor of that arm (Growdon et al., 1975).
Beta-adrenergic receptor blocking agents (Marshall, 1968: Sevitt, 1971; Winkler and Young, 1971) and primidone (O'Brien et al., 1981. Findley and Calzetti, 1982; Findley et al., 1985) are established treatments for essential tremor, having been subjected to randomized double-blind trials, but these drugs are only partially effective and are associated with significant side-effects (Koller et al,. 1986; Findley. 1987). Phenobarbitone has been found to be significantly better than placebo is reducing tremor amplitude but not at improving tests of motor performance or patients' self-assessments of disability (Findley and Cleeves, 1985). Similarly, some studies have suggested that propranolol does not improve the functional capabilities of patients (Foster et al., 1973; Sweet et al., 1974; Baruzzi et al., 1983) whereas primidone does (Chakrabarti and Pearce, 1981) Koller et al., (1986) concluded that both propranalol and primidone facilitated eating, drinking and writing but had no beneficial effect on the extent of embarrassment, fine manipulative skills and motor performances during pegboard and tapping tasks. The mechanisms of action of primidone and phenobarbitone are not known, but are presumed to be within the central nervous system. The response to beta blocking drugs was considered, in a detailed review by Findley (1987). to be predominantly mediated by way of peripheral beta2adrenoreceptors. However, effects within the central nervous system or a less accessible peripheral compartment or involving beta1-receptor sites may also be important (Young et al., 1975; Abila et al., 1983, 1985a,b; Findley, 1987).
Pathophysiology
Post-mortem studies have failed to demonstrate a consistent pathological substrate for essential tremor (Cestan, 1899; Bergamasco 1907; Hassler, 1939; Mylle and van Bogaert, 1940, 1948; Herskovitz and Blackwood, 1969; Lapresle et al., 1974; Rajput et al., 1991). Ipsilateral hemiparesis (Young, 1986), cerebellar stroke (Dupuis et al., 1989) and stereotactic lesions of the contralateral thalamus (Laitinen, 1965; Blacker et al., 1968; Hirai et al., 1983) can abolish essential tremor in man. C15O2 positron emission activation studies have demonstrated that in patients with essential tremor there is bilateral overactivity of the cerebellar circuitry even during rest, when no tremor is apparent (Colebatch et al., 1990; Jenkins et al., 1993). Furthermore, hypermetabolism of glucose within the medulla has been detected using [18F]2-deoxyglucose and was believed to represent overactivity of the inferior olive (Dubinsky and Hallett, 1987).
The reported frequencies of the postural component of essential tremor to the upper limbs ranged from -4 to 12 Hz in different subjects and are by no means diagnostic (Critchley. 1949; Marshall, 1962; Hubble et al., 1989; Calzetti et al., 1987; Koller et al., 1992).
Electromyographic studies have shown that essential tremor is generated by segregation of muscle activity in bursts and different patterns have been detected in agonist/antagonist muscle pairs. Both co-contracting and alternating patterns have been recorded, the former more frequently (Shahani and Young, 1976; Rothwell et al., 1987). In addition 'skipping' between these two patterns or involvement the antigravity (agonist) muscles alone are well-established pictures (Deuschl et al., 1987; Koller et al., 1992).
Subclassification of essential tremor has been attempted using clinical, electrophysiological and pharmacological criteria (Findley and Gresty, 1981; Marsden et al., 1983, Deuschl et al., 1987) but more recent studies have found support for these proposals (Lou and Jankovic. 1991; Koller et al., 1992).
The pathophysiology of essential tremor has been widely debated but the specific neurophysiological fault remains elusive (Elbe and Koller, 1990; Bain, 1993). The stretch reflexes, which are of normal size, latency and duration, may be followed by an underdamped oscillation (Rothwell et al., 1987). However, as the severity of essential tremor worsens the phase of tremor becomes increasingly difficult to reset by peripheral mechanical perturbations (Britton et al., 1992b) In addition, accelerometric studies have shown that as tremor severity increases the averaged spectra (which display mean squared acceleration of the frequency components against frequency) changes from multiple small peaks of comparable magnitude towards a single dominant peak of greater size, which may be accompanied by harmonics. As tremor becomes more symptomatic, the stability of the accelerometric spectral peak frequency, measured during different manual tasks, increases and the 4-5 Hz frequency changes seen in mild cases were reduced to ~1 Hz or less (Bain et al., 1993b) Elble (1989) measured the attractor dimension of essential tremor and demonstrated that it fluctuated between limit-cycle and chaotic modes oscillation, when tremor was respectively, symptomatic and relatively quiescent. He concluded that essential tremor resulted from a reduction in the functional degrees of freedom within the involved neural pathways.
Clearly, several basic but important facts about essential tremor have not been established or are subject to controversy. The assumption that hereditary essential tremor and the sporadic essential tremors are the same entity has arisen insidiously within the literature but may not be true. Isolated tremors, other than that of the arms. and task-specific tremors may or may not be manifestations of essential tremor. Dystonia poses another problem because it causes tremor that is similar to that seen in essential tremor, and could and has contaminated many studies of the latter. This study has concentrated on defining the core condition of hereditary essential tremor to clarify these issues through detailed personal observation of the clinical phenotype of inherited essential tremor.
Patients and methods
The various components of tremor were as defined by ad hoc committee of the Tremor Investigation Group of the International Tremor Foundation in the following way (Findley et al., 1993): (i) real tremor. i.e. tremor occurring when the muscles are not voluntarily activated and the relevant body part is supported against gravity; (ii) tremor, i.e. tremor present whilst voluntarily maintaining a position against gravity; (iii) kinetic tremor, i.e. tremor during any form of movement; (iv) task-specific tremor. i.e. appearance of kinetic tremor during the performance of highly specific skilled movements; (v) intention tremor, i.e. the pronounced exacerbation of kinetic tremor towards the end of a goal directed movement.
Index cases
Individuals with hereditary essential tremor were recruited from the clinics of the National Hospital for Neurology and Neurosurgery, London, and the Havering Hospital Trust Essex. Index cases had a family history of tremor involving at least three generations and fulfilled the clinical diagnostic criteria agreed by the Tremor Investigation Group above. (Findley et al., 1993).
Inclusion criteria
(i) The presence of visible and persistent postural tremor involving the hands or forearms which may or may not be accompanied by kinetic tremor. The postural upper limb tremor can be asymmetric and tremor may affect other parts of the body. (ii) Tremor must have persisted for at least 5 years, albeit with some fluctuation in severity, but need not produce disability.
Exclusion criteria
(i) The presence of other abnormal neurological signs with the exception of 'cogwheeling' without rigidity and Froment's sign which is cogwheeling induced when the contralateral limb performed voluntary repetitive movements, a derivation of the 'signe de Froment' (Froment and Gardere, 1926).(ii) The existence of known causes of enhanced physiological tremor (e.g. hyperthyroidism). (iii) Concurrent or recent exposure to tremorgenic drugs or the presence of a drug withdrawal state. (iv) A history of neurological trauma in the 3 months prior to the onset of tremor. (v) Clinical evidence for a psychogenic origin of tremor. (vi) Tremor of sudden onset.
Comprehensive histories were obtained from the index patients who were re-examined and recorded on videotape using a Panasonic VHS NV-MSlB video-recorder. Each patient completed a standard disability questionnaire (Appendix I) and handicap assessment form (Appendix 2) (Bain and Findley, 1993). Neurophysiological studies were performed to exclude large fibre peripheral neuropathies and in addition the tremors of the index patients were studied by surface polymyography and accelerometry (Britton et al., l992a; Bain et al., l993a).
Secondary cases
Nixety-three first degree relatives (74.4% of those resident it the British Isles) and 38 more distant relatives of the index relatives were visited, interviewed and examined by one of the authors (P.G.B.). Definite or possibly affected relatives were videotaped and specimens of handwriting and a drawing of an Archimedes spiral obtained from each family member. The relatives of the index cases were then classified into three distinct categories: (i) definitely affected, i.e. symptomatic with obvious postural tremor and a tremulous Archimedes spiral, fulfilling the clinical diagnostic criteria above, except that duration of tremor could be <5 years but >2 years in some cases; (ii) possibly affected, i.e. either asymptomatic with definite signs (abnormal postural tremor or a tremulous spiral) or symptomatic without definite signs; (iii) normal.
The severity of postural tremor and that apparent in Archimedes spirals were scored using a 0-10 clinical rating scale which had previously been assessed for both inter- and intra-rater reliability and been shown to provide valid indices of tremor induced disability (Bain et al., 1993a). The tremor evident in the spirals collected during this study were graded by three 'blind' raters and the median of their scores used in the results. The k coefficients for the inter-rater reliability of the scores varied from 0.63 to 0.85 (substantial to almost perfect agreement). The severity of the postural tremors seen in this study was scored by one trained rater (P-G.D.). It had been hoped that the tremor apparent on each person's videotape would be scored by three independent raters but various technical limitations to the use of videotape arose. Consequently no quantitative information could be obtained from them. These difficulties and a critique of the various techniques for assessing tremor severity have been discussed elsewhere (Bain and Findley, 1993; Bain et al., l993a).
There was still one major problem to be overcome, namely the question of how to differentiate between essential tremor as it presents in its early milder stages and the tremors (physiological and enhanced physiological) which may also be seen in healthy individuals. Even with modern neurophysiological techniques, there is no accepted method of making this distinction and thus the solution was inevitably pragmatic. The scores obtained by rating postural tremor and the tremor in spirals were used to separate essential tremor from the tremors seen in normal individuals. In order to be considered abnormal the severity of postural tremor or the tremor indicated in a spiral had to be at least twice that of the 95th percentile of that seen in healthy controls. These thresholds were determined prior to classification by examining the postural tremor and spirals of 100 healthy control subjects (age range 3-80 years). The 95th percentile was found to be -0.1 for postural tremor and 1.0 for the tremor visible in spirals.
Children under the age of 10 were examined but some of them could not cooperate with drawing, writing and various aspects of the examination. Consequently, they were classified as (i) definitely affected (the mother stated that the child was tremulous and tremor was evident on examination), (ii) possibly affected (mother considered them to be tremulous but there were no abnormal signs or tremor was evident on examination but the mother was not aware that it was abnormal) or (iii) normal. Children under the age of 15 years have been excluded from the data on handedness. alcohol responsiveness and treatment. None of them had been treated.
The presence and extent of tremor-induced disability and handicap were obtained by asking the index patietsts and the definitely and possibly affected secondary cases (aged over 15 years) to complete disability (Appendix 1) and handicap questionnaires (Appendix 2) (Bain and Findley, 1993).
Statistics
Segregation ratios were calculated for the siblings of the index patients (excluding the index) and the offspring of definitely affected individuals (including those of the index patients and their siblings). Segregation ratios were calculated only for those relatives examined personally by P.G.B., but confidence limits were based on all relatives, including those not seen. The lower limit assumed that all of those stated to be normal but not examined were normal, whilst the upperlimit assumed that this group contained the same proportion of definite and possible cases as was found in the examined kin but with the possible cases counted as affected (based upon the method used by Fletcher et al., 1990).
The risks for a currently unaffected child of a parent with hereditary essential tremor developing the disease were calculated using Bayesian statistics with the actual segregation ratios found in this study.
Results
Clinical features
Index patients
Twenty index patients. 12 males and eight females, were studied. The sex difference was not significant (X2 = 0.8, P> 0.05). Seventeen were right-handed and three lefthanded. The distributions of age, age at tremor onset (the age at which patients were first noticed to be tremulous) and the duration of tremor are shown in Fig. 1. The distribution of age at tremor onset was bimodal [F(18) = 110.6, p < 0.001). The medians and ranges of patients' ages, age at tremor onset and tremor duration are shown in Table 1 and their cumulative ages of onset plotted in Fig. 2.
In every index patient, tremor presented in the arms; in 15 (75%) the onset was symmetrical and in five (25%) tremor was first noticed in the dominant hand. In two patients (10%) the upper limbs remained the only affected site but in the majority (n = 18,90%) tremor spread to affect the legs (n = 9, 45%), head (n = 7, 35%), voice (n = 6, 30%), tongue (n = 4, 20%), facial muscles (those supplied by the seventh cranial nerve) (n = 3, 15%) and jaw (n = 1, 5%). Four of the patients' voice tremors and three of the head tremors were intermittent.
The most common sequence of spread was from the upper limbs to the legs (n = 7, 35%) but in six patients (30%) tremor spread firstly from the hands to the head, in two (10%) from the hands to the tongue, in one (5%) from the hands to the jaw, in one (5%) to the facial muscles (5%) and in one other (5%) to the voice.
All but one of the index patients had bilateral postural upper limb tremors which were highly symmetrical (rated using scores for postural tremor; correlation coefficient r = 0.62, P < 0.01). Four patients (20%) had significant end of movement accentuation of tremor during the finger-nosefinger test (an intention component) and five (25%) had a mid-movement component of kinetic tremor of a similar magnitude to their postural tremor, but in the majority (n = l5, 75%) tremor magnitude was diminished by movement. None of the patients had a rest tremor when completely) relaxed. The frequencies of the postural upper limb tremor, ranged from 4.5 to 10 Hz (mean 6.55 Hz). The electromyogram pattern in the wrist flexors and extensors was alternating in 11 (55%), co-contracting its two (10%), varied between alternating and co-contracting in three (15%) and was segregated only in the wrist extensors in four (20%).
Leg tremor was always postural and symmetrical and except in two patients, trivial. The frequency of leg tremor varied from 8 to 10Hz. One of the patients with symptomatic leg tremor had difficulty using the pedals of his car and another was greatly disabled: on standing up. her leg tremor would gradually increase in amplitude so that her leg eventually gave way, preventing her from washing up and compelling her to use a wheelchair whenever she out. This leg tremor was not relieved by walking and eenhanccd by fatigue. It had a frequency of 8 Hz.
Various types of postural head tremor were seen. the most common being a 'no-no' variety which occurred in six cases (30%). This was intermittent in three patients ( l5%); continuous in three others (15%). One patient had a 'yes-yes' tremor. The intermittent head tremors were often induced by talking or writing. The frequency of head tremor varied front 5 to 8 Hz.
The facial muscles involved were orbicularis oculi (n = 2, 10%), orbicularis oris (n = 1, 5%) and mentalis (n = 1, 5%) and were often activated by talking, smiling or grimacing. Voice tremor [present in six patients (30%)] was never associated vith dysphonia or dysarthria.
Cogwheeling was detected at the wrist in three patients 15%) but could be made to disappear in every case by getting the patient to relax completely; in four other patients (20%) Froment's sign was present. None of the index cases had diminished movement of the arms whilst walking and Wartenberg's sign was invariably negative (normal passive arm swing on rocking the patient's shoulders)
Secondary cases (affected relatives)
One hundred and thirty-one relatives were examined, of whom 53 (40.5%) were definitely and 18 (13.7%) possibly affected. Fifty-five of the definite and possibly affected relatives were right-handed, eight were left-handed and two
-----------------------------------------------------------
Table 1 Medians and ranges for age, age at tremor onset and tremor duration for index patients and affected relatives.
Age at study Age of onset Tremor duration
(years) (years) (years)
Median Range Median Range Median Range
index patients (n = 20) 54.0 17-78 15.0 5-52 26.5 5-58
Affected relatives
Definite cases (n = 53) 45.0 16-77 14.5 2-65 19.5 2-72
Definite and possible cases (n = 71) 39.0 12-77 15.0 2-65 15.0 2-72
-----------------------------------------------------------
were ambidextrous. Ninety-three were first degree relatives of whom 43 were definitely and 10 possibly affected. Of the 38 more distant relatives 10 were definite and eight possible cases. Amongst 14 clsildren under 15 years of age there vere six possible cases. There were no unaffected obligate gene carriers and no consanguinous marriages. The sex ratios (male:female) of the secondary cases, inclusive and exclusive of possible cases were 0.71 and 0.83, respectively. There were no significant differences between the numbars of affected men and women irrespective of whether possible cases were included (X2 = 0.69, P > 0.05) or omitted (X2 = 1.53, P > 0.05). The distributions of age, age at tremor onset and tremor duration for the secondary cases are shown in Fig. 3 and the medians and ranges of these details are shown in Table 1. The distribution of age at tremor onset was bimodal [F(48) = 173, P <0.001]. The cumulative ages at tremor onset are plotted in Fig 2.
Adults. Tremor presented in the upper limbs of every single affected relative. The onset was symmetrical in 44 (83%) of the definite cases but in eight of the remaining nine definite cases (15.1%), tremor was first noticed in the dominant hand. Two of these cases switched to writing with the non-dominant hand and then swapped back to the original hand when some years later their tremor became more symmetrical. Only one right-handed patient first noticed tremor in his non-dominant arm.
In the majority of affected relatives (66% excluding and 74,7% including possible cases) the upper limbs remained the only parts of the body affected by tremor. In a minority tremor was also detected in the legs (30,2% excluding and 23.9% including possible cases), the head (17% excluding and 12.7% including possible cases), jaw (5.7%), tongue (3.8%), facial muscles (3.3%) and voice (5.7%). The sequence of spread was most commonly from the arms to the legs(26.4%) but less often (7.6%) tremor went directly from the hands to the head.
There was a high degree of symmetry for the scores for upper limb postural tremor, irrespective of whether or not possible cases were included (correlation coefficient (r = 0.93, P <0.001) or excluded (r = 0.92, 0<0.001 None of the affected relatives had rest tremor but a kinetic component (assessed in mid-movement) of comparable severity to upper limb postural tremor was detected in 11.8%, of definite cases and end of movement accentuation (intention tremor) in 5,7% of those definitely aftected.
Leg tremor when present was symmetrical and usually trivial, but in three cases (5.7%) it was problematic. To of these individuals had marked enhancement of leg tremor standing (with frequencies of 7.5 and 11 Hz) and when walking, particularly when tired, In one other case amplitude of leg tremor increased substantially when skilled manual tasks (e,g. writing) were performed.
Head tremor was more often intermittent (11,3%) than not (5.7%). The intermittent tremors were invariably of the 'no-no' variety and could often be induced when the patient talked or concentrated on writing. The more persistent head tremors were present when sitting or standing, but not where the head was completely supported (rest), and were of a complex type with both 'yes-yes' and 'no-no' component (either could be predominant).
Tremor of the jaw, voice, tongue and face were never seen in isolation and were only evident in patients with quite severe upper limb tremors. The likelihood of tremor affecting the cranial musculature increased with the severity of upper limb tremor. Tremor of osbicularis oculi was most obvious whilst the eyes were closed. Tremulousness of orbicularis oris, mentalis and masseter were most visible during normal conversation.
None of the secondary cases had evidence of parkinsonian dystonia or other neurological disease. Cogwheeling could be palpated in four cases (7.6%) but would disappear when the subjects relaxed fully and Froment's' sign was noted in one other patient (1.9%). Rigidity was never apparent and arm-swing invariably normal. Wartenberg's test was negative in every case.
Children. Amongst 14 children aged under 15 years there were six possible cases (age range 2-13 years). Four of the remaining two had symptoms produced by tremor
Natural history
It was typical for individuals to state that they had felt shakes 'inside' for several months prior to developing overt essential tremor. Later this would progres to cause an intermittant action tremor which appeared when the affected person was excited or fatigued, a situation that in many instances led to that person being labelled 'nervous'. Subsequently, the tremor would become more persistent but could be voluntarily suppressed for short periods of time, so that little disability occured. At this stage the postural tremor often appeared to be jerky, with tremor arriving in flurries that lasted for no more than a few seconds. In the younger, mild cases, tremor was most evident for an instant when a new posture was adopted. It could also be seen at an early stage to interrupt slow voluntary pronation-supination movements of the wrist. Subsequently, as tremor increased in amplitude it became more autonomous and oscillatory. It was then more difficult to control and caused disability. However, even with the most severe upper limb tremors, considerable fluctuations in amplitude were evident and at times tremor would virtually cease, even when a posture was maintained, only to return a few seconds later. Tremor amplitude was also significantly altered by changes in posture, task and the general state of activity of the individual, but could still be voluntarily suppressed, to some extent, for short periods of time. Tremor frequency was also observed to vary with different tasks and positions of the arms, as has been reported elsewhere (Bain et al., 1993b). Hunger, emotion, fatigue and temperature (e.g. hot baths) exacerbated tremulousness.
Tremor severity in index and secondary cases
The means and ranges for the severity of head, upper and lower limb postural tremors are shown in Table 2, from which it can be seen that the brunt of hereditary essential tremor falls on the upper limbs. The severity of upper limb tremor (as assessed by spirography and the clinical rating scale scores) worsened with age and the duration of tremor (Table 3). However, there was no relationship between tremor severity measured in these ways and the age at which tremor was first noticed (age of onset) (Table 3).
The severity of leg tremor amongst the index and secondary cases was highly correlated with that in the upper limbs (corrrelation coefficient = 0.43, P <0.001).
Severity of upper limb postural tremor and tremor evident in spirals did not significantly differ between male and female cases or amongst those cases with a maternal rather than a paternal mode of inheritance (Wilcoxon's rank sum test).
Disability and handicap in index and secondary cases
All the index patients and 81% of the definite secondary cases completed and returned the disability and handicap questionaires. All of the index patients and the majority (79%) of those secondary cases who returned their questionaires documented some disability (64% of all definite secondary cases). The relationships between the extent of disability and the ages of the affected cases, the age at onset and duration of tremor are shown in Table 3. Decade specific disability is shown for all cases (inclusive and exclusive of possible cases) in Fig. 4A. Hereditary essential tremor never produced
-----------------------------------------------------------
Table 2 The severity of tremor in Archimedies spirals rated from 0 (none) to 10 (most severe) and the postural tremors of the upper limbs, lower limbs and head which were also scored from 0 (none) to 10 (most severe) using a clinical rating scale
Index Affected relatives
patients Excluding Including
possible possible
cases cases
(i)Spiral score (0-10)
Mean 3.8 2.6 2.3
Range 0-6.5 0-5 0-5
(ii)Right arm (0-l0)
Mean 1.7 1.1 1.0
Range <4.0 0.2-4.0 0.1-4.0
(iii)Left arm (0-10)
Mean 1.5 1.0 0.9
Range 0.3-4.0 0.2-4.0 0.1-4
(iv)Right leg (0-10)
Mean 0.3 0.2 0.2
Range <1.7 <1.0 <1.0
(v)Left leg (0-10)
Mean 0.3 0.2 0.2
Range <1.7 <1.3 <1.3
(vi)Head (0-10)
Mean Negligible Negligible Negligible
Range <0.5 <1.0 <1.0
----------------------------------------------------------
Table 3 Correlation coefficients (r-values) for the relationships between (i) tremor severity in the right anm (graded clinically from 0 to 10); (ii) tremor in spirals (graded from 0 to 10); (iii) disability (%) and the patients' ages, ages at tremor onset and tremor deviations
Index and definite Spiral score Right arm tremor Disability secondary cases severity
Age 0.49* 0.60* 0.36* Age or tremor onset 0.05 0.23 0.07 Duration of tremor 0.49* 0.3l** 0.28***
__________________________________________________________
Significant at: *P < 0.001; **P < 0.01; ***P < 0.05. The significance levels were not altered by the inclusion of possible cases.
---------------------------------------------------------
disability before the age of 15 years. The frequency with which difficulties were documented with each item on the disability questionnaire is shown in Appendix I. The most commonly affected tasks were those normally performed with the dominant hand (Appendix 1). Perhaps surprisingly several patients reported that bathing was difficult, explaining that they would become more tremulous in a hot environment. and 30% of index patients and 9% of secondary cases had some difficulty with stairs. There was no significant difference in the degree of disability found in men compared with women or among those cases who had paternal rather than maternal inheritance (Wilcoxon's rank sum test).
The extent of social handicap and the various situations in which this was experienced are shown in Table 4, Sixtie (80%) of the index cases and 47% of those secondary cases who returned their questionnaires (38% of all definite secondary cases) reported that they were handicapped some way by their tremor; generally this was because of both physical impairment and embarrassment rather than either factor alone. Five (25%) of the index patients and 12% of the definite secondary cases had been forced to change jobs or take early retirement because of tremor. Furthermore 12 (60%) of the index patients and 19% of the definite secondary cases had not applied for a job or promotion because of tremulousness. The majority (65%) of the index patients and 33% of the definite secondary cases no longer dined out. Decade-specific handicap is shown for all cases (including and excluding possibles) in Fig. 4B. Fifty percent of cases were handicapped by the fifth decade and 80% by the sixth. The subsequent decline is probably caused by the social consequences of retirement.
Clinical associations
The conditions affecting two or more index and definite secondary cases are reported in Table 5A and the number of families in which co-segregation was observed between hereditary essential tremor and either classical migraine (migraine associated with transient focal neurological symptoms) or the restless legs syndrome or hypertension are shown in Table 5B.
Alcohol responsiveness and effect of treatment
Eleven of the index patients (55%) and 25 (47.2%) of these (definite) secondary cases reported that drinking between
-----------------------------------------------------------
Table 4 The impact of hereditary essential tremor upon livelihood and social behaviour (handicap)
Index Reason for handicap Secondary Reason for handicap cases cases*
E P E+P E P E+P
-- -- - --- -- - ---
(a) Percentage forced to change job/retire
25 - - - 12 - - 12
(b) Percentage not to apply for a job/promotion
60 20 5 35 19 5 5 9
(c) Percentage no longer shopping alone
55 15 5 35 7 0 2 5
(d) Percentage no longer eating out
65 10 0 55 33 16 5 12
(e) Percentage no longer attending parties
50 0 0 50 23 12 2 9
(f) Percentage no longer able to practise their favourite
sport/hobby
30 5 15 10 19 7 7 5
(g) Percentage no longer travelling by public transport
35 10 15 10 7 0 0 7
(h) Percentage no longer driving a car
20 0 10 10 7 0 2 5
(i) Percentage no longer going on holiday
35 10 5 20 7 5 0 2
*Of those definite secondary canes who returned completed questionnaires (81.1% of all definite secondary cases). Handicap caused by: P = physical impairment; E = embarrassment; P+E = both physical impairment and embarrassment.
----------------------------------------------------------
-----------------------------------------------------------
Table 5 Possible clinical associations with hereditary essential tremor
(A) Conditions affecting two or more index or definite secondary cases
No. Percentage
affected affected
classic migraine 19 26.0
Left handedness 9 12.3
Hypertension 8 11.0
Non-insulin dependent diabetes mellites 4 5.5
Restless legs syndrome 3 4.1
Osteoarthritis 3 4.1
Cataracts 2 2.7
Duodenal ulcers 2 2.7
(B) Conditions co-segregating with essential tremor within one
or more kindreds
No. of families Co-segregation ratio
co-segregating
Affected* Unaffected^
Classical migraine 3 0.67(10/15) 0.0 (0/8) Restless leg syndrome 1 1.0 (2/2) 0.0 (0/6) Hypertension 1 1.0 (3/3) 0.0 (0/0)
*The ratio of the number of cases with either migraine, restless leg syndrome or hypertension who also have tremor divided by (be total number of kin definitely affected by tremor (affected); ^ the ratio of thc number of cases who have one of these three conditions but do not have tremor divided by the total number of atremulous kin (unaffected).
----------------------------------------------------------- and four units of alcohol would virtually abolish their tremor for -2-4h but the tremor would rebound and be worse the following day. In spite and because of the transient nature response two index patients (10%) and two secondary cases (4%) admitted to being dependent upon alcohol and would not go out of the house without a drink. Four other index cases and 10 secondary cases did not benefit from alcohol and five further index cases and 18 secondary cases did not drink.
The degree of homogeneity or heterogeneity of the responses to alcohol within the kindreds was variable. (i) amongst the families of the 11 index patients who were alcohol responsive, all affected relatives were alcohol responsive in nine of the kindreds; in one kindred the effect of alcohol on the other affected members of the family was not known; and in one family responders and non-responders coexisted; (ii) amongst the families of the four unresponsive index patients, two kindreds were homogeneously unresponsive and two families exhibited heterogeneity; (iii) amongst the families of the five index patients in whom the response to alcohol was not known, three were homogeneously responsive, one unresponsive and one included both responders and non-responders.
Thus, in summary, 15 of the families showed homogeneity and four heterogeneity of responsiveness to alcohol (for the remaining family the facts are not known).
Twelve index patients had been prescribed propranolol of whom seven (58%) had experienced some benefit. which varied from modest to substantial. Three other patients (15%) were taking primidone with good results.
Only four affected relatives (8%) had received any treatment for their tremor-two responding to propranolol and two to primidone.
Adverse effects of drugs
Three cases with long-standing postural tremor had been prescribed prochlorperazine for unrelated symptoms. In each case tremor severity was greatly exacerbated and this effect was evident within days of taking the drug. Two of these cases also developed a marked upper limb rest component to their tremor; this disappeared when prochlorperazine was withdrawn but in both cases a degree of cogwheeling was evident 1 year after the drug had been stopped.
Biological fitness
The mean number of children born to the index patients was 2.10 (SD +/-2.17) and was 2.63 (SD +/-2.12) if only index patients over 35 years old were included. These figures are similar to general population data in the United Kingdom (Office of Population Censuses and Surveys 1984,1987).
Segregation analysis
The proportions of affected to unaffected siblings of index patients and offspring of index patients and their affected relatives are shown in Table 6. The null hypothesis that the actual proportions of affected to unaffected relatives would not differ from the expected ratio of 1:1 was tested using the formula given by Roberts and Fensbrey (1978) which includes Yates' correction. The actual proportions did not significantly differ from 1:1, irrespective of whether possible cases were considered to be affected (X2 = 0.387, P > 0.05) or unaffected (X2 = 0.688, P > 0.05).
A plot of the ages at tremor onset for the index patients
versus that of their (definitely) affected relatives showed that
both young and late onset index cases had relatives with both
young and late onset tremors (Fig. 5). There was a slight trend
for individuals within the same family to have a similar age of
onset but this did not reach statistical significance {Spearman
correlation coefficients: 0.22 (excluding) and 0.17 (including
possible cases), P > 0.05].
Genetic counselling
The risks of a currently unaffected child of a parent with hereditary essential tremor developing essential tremor disability or handicap are shown in Table 7. The risk of developing hereditary essential tremor after the ages of 10 and 50 years are <20 and <6%, respectively.
Discussion
The purpose of this study was to delineate the clinical features of hereditary essential tremor. We approached this by selecting 20 index patients with incontrovertible hereditary essential tremor and then examining their relatives so as to define the phenotype of the gene or genes involved.
Relationship to other neurological diseases
We did not find any dystonia amongst our subjects. Several previous studies have included patients with abnormal dystonic postures under an expansive definition of 'essential tremor' (Critchley, 1949, 1972; Hornabrook and Nagurney, 1976)
-----------------------------------------------------------
Table 7 Data for giving generic counselling to an apparently unaffected child who has a parent with hereditary essential tremor
Age Risk of Approximate percentage risk of developing a of developing tremor (T),disability(D)or handicap (H) child tremor (%) by the ages of (years)
20 years 30 years 40 years 60 years
T D H T D H T D H T D H
0 46 30 18 9 33 19 13 40 34 17 45 44 38 5 41-42 24 14 7 27 16 11 35 30 15 40 40 34 10 38-39 20 12 6 24 14 11 32 28 13 37 37 32 15 28-29 8 5 2 12 7 5 21 19 9 27 27 23 20 22-23 4 3 2 14 10 6 20 20 18 25 19-20 11 12 5 17 17 15 30 18-19 11 12 5 17 17 15 35 15-17 7 8 3 13 13 11 40 9-1l 7 7 6 45 7-9 6 6 5 50 5-6 3 3 3 55 3-5 0 0 0 60 0
For example, at birth (0 years), (the risk of ever developing hereditary essential tremor is 46%; the risks of that individual having tremor, being disabled by it, or handicapped by it, by the age of 20 years are 30,18 and 9%, respectively.
-----------------------------------------------------------
Rajput et al., 1984; Martinelli et al., 1987; Lou and Jakovic, 1991). In one of these studies, Lou and Jankovic (1991) reported that amongst 350 patients diagnosed as having 'essential tremor', 47% had dystonic posturing, which was equally distributed amongst their familial and sporadic cases. However, in the light of our findings and the observations that patients with idiopathic and hereditary torsion dystonia often exhibit tremor (Zeman et al., 1959; Marsden and Harrison, 1974; Couch, 1976; Baxter and Lal, 1979; Sheehy and Marsden, 1982; Fletcher et al., 1990, 1991) it would appear that these accounts of 'essential tremor', in fact, describe at least two separate diseases. Similarly, the absence of nystagmus or other cerebellar signs in our subjects suggests that papers including patients with these features (Nettleship, 1911; van Bogaert and de Savitsch. 1937; Critchley, 1949, 1972) were not describing a single entity. Furthermore, there were no cases of Parkinson's disease or signs of peripheral neuropathy amongst our kindreds. Evidently, much of the previous literature on 'essential tremor' was describing several diseases in which postural tremor occurred and resembled that seen in hereditary essential tremor.
The clinical characteristics of hereditary essential tremor
The median age of onset was 15 years for our index patients and between 14.5 and 15 years for the secondary cases (Table 1). which is in accordance with the descriptive findings of Critchley (1949) and Jager and King (1955). However, the actual age of onset is likely to precede the reported date which would tend to shift the true cumulative age of onset plot to the left (Fig. 2). The distributions of ages at tremor onset of the index patients and secondary cases were bimodal (Figs I and 3, respectively), as has been found in a previous study (Gerstenbrand et al., 1982), which might suggest that two abnormal genes might be involved. However, the 'late onset' (over 35 years) index cases had affected relatives with early onset tremors and there was no tendency for late onset cases to cluster within particular families (Fig. 5). These results do not support a two (distinct) gene hypothesis (Harris and Smith, 1947) but indicate an individual temporal variability to expressing the phenotype.
Examination of the pedigrees and segregation ratios indicated an autosomal dominant disorder and there were no detectable instances of reduced penetrance. Penetrance was complete by the age of 65 years. These findings are in agreement with the observations of Jager and King (1955), who documented a single large pedigree in Utah through Larsson and Sjogren (1960) in their epidemiological study of northern Sweden, except that the latter found a mean age at tremor onset of ~50 years perhaps because the information on ~50% of their 210 'cases' was obtained after the patients had died.
Our data (Figs 2 and 4) can be operated on with Bayesian statistics and be used for genetic counseling (Table 7). Thus, the risk of a clinically normal 20-year-old offspring of an affected individual developing tremor is ~23%. The likelihood of this offspring developing a disability or handicap due to hereditary essential tremor before the age of 60 years (during working life) are ~20 and 18%, respectively.
The proportions of affected men and women were not significantly different for our cases, which is in accordance with Jager and King's (1955) description and the epidemiological surveys conducted by Hornabrook and Nagumey (1976), Rajput et al. (1984) and Bharucha et al. (1988) on 'essential tremor' in their respective communities. The biological fitness of the index patients who were over the age of 35 years and thus considered to have completed their families was not diminished.
Biary and Koller (1985) suggested that the incidence of left handedness in patients with essential tremor (17%) was greater than in a normal control population (8%). Our figures of 15% for index and 11.3% for definite secondary cases provide some support for this notion.
Tremor invariably commenced in one or both of the upper limbs of our subjects. No examples of isolated head, tongue. voice, jaw, trunk and leg tremors were seen either in our (selected) index patients or the secondary cases. These entities have all been described in the literature as manifestations of essential tremor (see Introduction), so either these 'isolated' tremors represent rare forms of 'essential tremor' or more likely result from other disease processes. Dystonia can produce tremor alone (Fletcher et al., 1990). The most common type of focal dystonia is spasmodic torticollis (Marsden and Fahn, 1982) and isolated head (and trunk) tremor may occur as a manifestation of that condition (Rivest and Marsden, 1990). Tremors of the jaw, face, tongue and voice can occur in oromandibular dystonia, lingual dystonia and spasmodic dysphonia and at least some cases of isolated tremor at these sites may be produced by dystonia. Isolated leg tremor suggests the diagnosis of Parkinson's disease, particularly if a rest component is present, or primary orthostatic tremor if it occurs on standing with a frequency of between 14 and 16 Hz (although transient epochs of period doubling have been recorded in this condition (Thompson et al., 1986; Rothwell, 1989; Britton et al., l992a). Indeed, most patients with isolated leg tremor of frequencies between 4 and 6 Hz have a reduced [18F]dopa uptake into the putamen as seen on PET studies, a finding characteristic of Parkinson's disease (Brooks et al., 1992). It is also conceivable that isolated leg tremor occurs in dystonia, but our findings suggest that hereditary essential tremor is nor a likely culprit.
The onset of tremor was reported to be in 'both hands' by 80% of the index and 83% of the secondary cases. In the remainder, with one exception. tremor commenced in the dominant hand. However, in all but one mild case both hands became affected and there were no examples of prominent unilateral tremor. A severe persistant postural tremor of one arm is more likely to be the result of dystonia, parkinsonism or a structural lesion, than hereditary essential tremor. A 'hemiparetic' pattern of tremulousness was never evident in this study.
In 10% of our index patients and 66% of the definite secondary cases the upper limbs were the only sites affected, but initial spread from the arms to the legs (35% of index and 26.4% of secondary cases) was more commonly reported than spread directly from the arms to the head (30% of index and 7.6% of secondary cases). This is contrary to the opinions expressed in several previous accounts (Critchley, 1949. 1972; Larsson and Sjogren, 1960; Marshall, 1962; Lou and Jankovic, 1991) but most of these studies included patients with torticollis.
In descending order of frequency, tremor of the legs, head, voice, tongue, face and jaw were seen and these occurred more often in individuals with severe postural tremor of the arms. No true isolated truncal tremors were observed, those encountered being invariably due to transmission of leg tremor.
No rest tremors were detected whilst the subjects were completely relaxed and in those cases where 'cogwheeling' was palpable, it would disappear when the subjects' voluntary muscle activation declined. Rigidity at the wrists and elbows or reduced shoulder swing (Wartenberg's sign) were not encountered and no 'pill-rolling' tremors were seen. We suspect that previous reports of