Search strategy and selection criteria

References for this review were identified from searches of Pubmed (from 1966 to 2005) and the Cochrane Library (2005, Issue 2) with the search terms “multiple sclerosis”, “tremor”, “ataxia”, “disability”, “prevalence”, “surgery”, “thalamotomy”, “deep brain stimulation” and “treatment”. Articles were also identified through searches of the reference tables of identified papers. Furthermore we searched the ISI Science Citation Index for relevant articles citing identified papers. Only articles in English and German were included in the review; publications in abstract form were not considered.


Tremor is such a common problem in multiple sclerosis (MS) that the famous French neurologist Charcot (1825–1893) described it as a part of his triad of characteristic symptoms in MS, together with nystagmus, and scanning speech.

Tremor of the upper limbs can be very disabling and seriously impair many activities of daily living and quality of life. In practice the treatment of tremor in MS is often frustrating. In this review we discuss medical, surgical and other treatment options.


Although the epidemiology of symptoms in MS has been studied extensively, it remains difficult to ascertain the prevalence of tremor in historic study populations, mostly because Kurtzke’s functional systems scale for cerebellar function, which is most widely used in clinical studies in MS, does not record tremor [55]. In a large long-term follow-up study, Weinshenker and colleagues found functionally relevant cerebellar deficits in one third of 259 patients [94]. In another study, disability and dependency of patients with MS were shown to correlate with the severity of ataxia [95].

Two newer studies assessed the prevalence of tremor in MS. One study by Alusi and coworkers examined 100 patients randomly selected from a London MS-clinic and found tremor in 58% of patients. The tremor was minimal in 27%, mild in 16% and moderate to severe in 15% of all patients [6]. As patients were selected from a specialist clinic, the relatively high prevalence and severity of tremor might be due to selection bias.

In a prevalence cohort study by Pittock and colleagues, 200 MS patients living in Olmsted County, Minnesota, USA were assessed for tremor and measures of disability. Tremor was found in 25.5% and severe tremor in 3% of the study population [74]. Probably the community based population in the latter study gives a more realistic estimate of tremor prevalence in MS. In both studies, however, tremor was associated with greater disability as measured on the expanded disability status scale (EDSS). In the Olmsted County population, patients with tremor of any severity were more likely to be unemployed or retired early because of disability.

Tremor subtypes

Tremor in MS can involve the head, neck, vocal cords, trunk and limbs, whereas involvement of the tongue, jaw or palate has not been reported [5].

The different types of tremor are currently classified according to a working consensus of the Movement Disorder Society [29]. In MS, the two most prevalent tremor forms are postural tremor (tremor present whilst voluntarily maintaining a position against gravity) and intention tremor (tremor occurring during target directed movement where tremor amplitude increases during visually guided movements towards the target). True rest tremor (tremor present in a body part that is not voluntarily activated and is completely supported against gravity) is unusual in patients with MS, and Holmes (or “rubral”) tremor is also very uncommon.

In the two main prevalence studies, [6, 74] rest tremor was observed only in the Olmsted County survey (1% of patients), whereas Holmes tremor was observed in neither study. An overview of the affected body parts observed in both studies is given in Table 1. In both studies tremor was most commonly found in the arms. In the London study population, 36% of patients suffered from bilateral arm tremor, making this the most common pattern of limb involvement.

Table 1 Details of the two main prevalence studies on tremor in MS

Pathophysiology of tremor

The pathophysiology of tremor in MS is a difficult area of investigation, partly because MS is by definition a multifocal disease, so that tremor occurrence cannot easily be linked to a single neuroanatomical site. Systematic postmortem studies on the link between lesion site and the clinical phenomenon of tremor have never been undertaken.

The predominance of action tremors (postural and intention) in patients with MS point to the cerebellum and its connections as the most likely source of tremor production, whereas the rarity of rest tremor argues against an involvement of the basal ganglia. The common occurrence of bilateral tremor might indicate that damage to the cerebellum and its connections is often multifocal.

Another link to the cerebellum in the pathophysiology of tremor in MS is the effect of peripheral cooling on intention tremor. Intention tremor is thought to be modulated through increased long latency stretch reflexes [30]. Cooling has been shown to decrease the sensitivity of muscle spindles [64] and the velocity of peripheral nerve conduction [56]. In two experimental studies, cooling of the arms markedly reduced intention tremor severity in patients with MS [77, 35]. The authors argue that this effect might have been partly due to decreased muscle spindle function and decreased nerve conduction velocity, which in turn resulted in decreased input into tremor producing cerebellar circuits.

Alusi and coworkers drew attention to the placement site of deep brain stimulation (DBS) electrodes to help understand the neuronal circuits involved in tremor production. They stated that neurosurgeons increasingly chose the nucleus ventralis oralis posterior (VOP) of the thalamus rather than the classic target, the nucleus ventralis intermedius (VIM). This is interesting because the VOP is the basal ganglia output nucleus of the thalamus [6], suggesting that the cerebellar tremors seen in MS might actually be generated by the basal ganglia. The standard electrode placement site reported in the literature, however, is the VIM and no studies comparing placement sites have been published. Whittle and coworkers comment on the difficulty of electrode placement in MS patients: in most patients there are major brain distortions due to demyelination, plaque formation and e vacuo hydrocephalus, and it is therefore uncertain whether the anatomy of these patients conforms to standard stereotactic atlasses [100]. Keeping these comments in mind, it seems unwise to base pathophysiological theories purely on DBS electrode placement site.

A range of animal experiments have been undertaken to indentify the anatomical structures involved in tremor production (for review [101]). It has been shown that damage or removal of the cerebellar cortex does not induce intention tremor in monkeys, [31, 76] whereas partial or complete cerebellectomy leads to tremor during movement and posture [75, 42, 40]. A slow, 3–5 Hz tremor during target directed movements can also be induced by reversible cooling of the dentate nucleus (the origin of most cerebellar efferents) in monkeys [22].

Most cerebellar efferents project from the dentate nucleus via the superior cerebellar peduncle (brachium conjunctivum) to the red nucleus and the thalamus. According to some studies, transection of the superior cerebellar peduncle causes intention tremor in monkeys [34, 93], whereas other authors do not mention this effect.

Interestingly, Carpenter and colleagues report that tremor induced by transection of the superior cerebellar peduncle can be alleviated by a second lesion in the lateroventral or centromedian thalamus. [19, 20] The results of these animal studies suggest that damage to cerebellar efferents (through lesions of the dentate nucleus or superior cerebellar peduncle) may cause disinhibition of thalamic nuclei which are the main producers of intention tremor.

Although this is an interesting pathophysiologic model of intention tremor production, it remains uncertain whether the results of animal studies can be generalized to patients with MS.

In summary, clinical observation, animal studies and some experimental evidence in humans favour the cerebellum and the thalamic nuclei connected to it as the major locus of intention tremor production, but more research is needed to evaluate the role of the basal ganglia and other systems in tremor production in MS.

Assessment of tremor

Depending on the subtype of tremor, several methods for the assessment of tremor severity and its impact on the lives of patients have been developed. Rest tremor is often assessed with the tremor subscale of the Unified Parkinson’s Disease Rating Scale (UPDRS). Fahn and colleagues devised the most comprehensive tremor scale for non-parkinsonian tremor in 1984 [33]. This instrument measures tremor in nine body parts at rest, while maintaining a posture and during goal directed movements. It also includes an assessment of arm tremor while writing and pouring water as well as a subscale for functional disability (interference of tremor with dressing, writing, eating etc.). Some studies included in this review use a modified and heavily abbreviated version of this scale, but most studies assess tremor by clinical examination only (e.g. by finger-to-nose testing, drinking from a cup, nine-hole-peg-test or writing and drawing tasks) or use a simple ordinal severity scale, often classifying tremor as absent, mild, moderate or severe. A simple 0–10 tremor severity scale devised by Bain and coworkers [11] has been shown to be a valid and reliable measurement tool in patients with MS [7], but has so far only been used in few clinical studies [6, 4, 15].

Accelerometry and polarized light goniometry are neurophysiologic methods of tremor assessment. While these methods offer an objective measurement of tremor severity, they can only measure one aspect of an often complex movement problem at a time, and e.g. cannot measure the ataxia which often complicates tremor in MS.

Matsumoto and colleagues devised a more complex “quantitative movement analysis technique”, which records the patients’ goal directed movements in three dimensions using an electromagnetic tracking device, but the complexity of this method [62] as well as the computer aided tracking tasks employed by Aisen and coworkers [2] prohibit their widespread use.

One of the most important aspects of tremor in MS is its impact on the daily life of the patients. Functional status in MS-patients is often measured by asking the patients to complete questionnaires assessing activities of daily living such as writing, eating or dressing. Often these scales are self-devised and not validated. Validated scales used in the studies reviewed in this article include the Frenchay Activities Index [85] and the quality of life subscale of the Functional Assessment of Multiple Sclerosis Scale (FAMS) [62].

Medical treatment

A summary of the published studies on medical treatment of tremor in MS is given in Table 2. Most of the published literature on medical treatment consists of case reports and uncontrolled open label studies. The few randomized controlled trials comprised small numbers of patients and very likely lacked the power to reveal small treatment effects.

Table 2 Studies on medical treatment of tremor in MS

Beneficial effects have been reported for a variety of drugs: case reports and small uncontrolled open label studies claim a positive effect of primidone [47], gluthetimide [2], intrathecal baclofen [96] and isoniazid. [80, 32, 67, 38] Controlled clinical trials have been published on the use of propranolol [52], ethanol [52], isoniazid, [52, 44, 16] carbamazepine [86], ondansetron [78] and dolasetron [65].

Koller evaluated the effect of propranolol, isoniazid and ethanol on tremor in three tremulous MS patients in a double blind crossover trial and found no treatment effect for any of the treatments [52]. To the best of our knowledge, no further trials with propranolol have been published, but it is interesting to note that Alusi and coworkers excluded two tremulous MS patients from a thalamotomy trial because they had achieved functional improvement after propranolol use [4].

In an open study the hypnotic-sedative drug glutethimide exhibited visible functional benefit in six of eight MS patients with tremor [2]. Apparently, a controlled trial to confirm these findings was never undertaken.

Two double-blind placebo-controlled trials using isoniazid were published. In the first study reduction of tremor occurred in six of eight patients [16], with functional improvement in four patients. In the second study all six patients had measurable tremor reduction after treatment with isoniazid but this did not lead to functional improvement [44]. Doses of isoniazid used to treat MS-related tremor were very high (up to 1200 mg a day), and treatment was in some patients associated with anorexia and nausea [38] or with a combination of drowsiness, dysphagia and increased bronchial secretion. [32, 67] Other reported side effects were abnormal liver function tests, [80, 32] fatigue [32] and increased weakness [67]. Isoniazid inhibits GABA aminotransferase activity and increases GABA in the central nervous system, but no correlation was found between the degree of GABA elevation in the cerebrospinal fluid and clinical response [16].

Sechi and coworkers published a small single-blind placebo-controlled trial with carbamazepine [86]. They reported improvement of tremor as assessed by clinical examination and accelerometry in all seven included patients, but failed to report whether this translated into functional improvement.

A placebo controlled, double-blind, crossover study, suggested that a single intravenous dose of the 5-HT3 receptor antagonist ondansetron led to tremor reduction in twelve, and to functional improvement in nine of the sixteen included tremulous MS patients [78]. A subsequent open label study by Gbadamosi and colleagues which used the same intervention found no significant treatment effect [39]. Likewise, a small clinical trial with dolasetron, another 5-HT3 receptor antagonist, showed no significant treatment effect on cerebellar ataxia [65].

Weiss and coworkers reported a positive effect of intrathecal baclofen in a single patient with bilateral arm tremor [96]. The tremor amplitude decreased almost linearly as the dosage increased and tremor was abolished at a dosage of 250 µg per day. This possible treatment option has, however, not been observed or evaluated in other studies.

In recent years, there has been growing interest in cannabis as a possible therapeutic in MS. In the 1980’s a case report [63] and a small uncontrolled study [21] on the beneficial effects of cannabis on tremor and spasticity in MS were published and a survey taken among MS patients revealed that many patients experienced positive effects of smoked cannabis on MS-related symptoms [23].

Baker and colleagues reported a decrease of tremor and spasticity in an animal model of MS after treatment with Δ9-tetrahydrocannabinol, the active ingredient of cannabis [12]. All this furthered hopes of cannabis as a possible new treatment option for tremor in MS, but much to the disappointment of tremulous MS patients, several well conducted randomized controlled trials did not show a significant effect of orally administered cannabis extracts [103, 37, 92] or oral Δ9-tetrahydrocannabinol [103] on tremor.

Surgical treatment

The surgical treatment options for tremor in MS are stereotactic thalamotomy and DBS. An overview of the published studies on surgical treatment is given in Tables 3 and 4. Most of the studies are small observational retrospective studies. When reviewing the literature on surgical treatment, it is surprising as well as disappointing that the majority of studies are remarkably imprecise in providing basic information on the length of follow-up, on adverse effects and-most importantly-on the effect on functional status and tremor associated disability.

Table 3 Studies on stereotactic surgery for the treatment of tremor in MS
Table 4 Studies on DBS for the treatment of tremor in MS

The first study on thalamotomy for tremor in MS was published by Cooper in 1960 [25]. Brice and colleagues were the first to report improvement of tremor through continuous thalamic DBS in 1980 [17].

Strategies for patient and treatment site selection

The earlier studies on thalamotomy used the thalamic nucleus ventralis lateralis (VL) as the target, whereas the nucleus VIM was chosen in most DBS studies and most later thalamotomy studies. This preference for the VIM is probably due to the experience with this thalamic nucleus in the treatment of tremor in Parkinson’s disease (PD) and essential tremor (ET). Research groups in Oxford and London used the nucleus VOP and the zona incerta (ZI) as targets for both lesional surgery and DBS, but the results yielded with these targets [4, 69, 15] were not different from those reported for VIM or VL thalamotomy or stimulation.

The same researchers advocate the use of tremor frequency analysis during movement tasks as a method to identify patients likely to benefit from surgery [60]. This may be a valuable tool for patient selection, although it has only been validated in a small number of patients and studies in which this technique is used [4, 69, 15] do not report better results than studies without this selection method.

The site for lesional surgery or electrode placement is classically chosen relative to the site of the anterior and posterior commissures using a standardized atlas. More recently, surgeons tried to refine this placement strategy with microelectrode recording within the target area. Neurons discharging synchronous to peripheral tremor are identified and the treatment site is placed in an area where the most tremor related neuronal activity is found. [73, 59, 57, 58]

This method of treatment site selection is often used in studies on tremulous patients with PD and ET, but only a small number of MS-patients were included in these studies. [87, 13, 66, 62, 84] Since this method offers a theoretical advantage over the classical methods of treatment site selection, it should be further evaluated.

Outcome after surgical treatment

Because of the many shortcomings of the published studies, the results need to be interpreted with great caution. It does, however, seem as if almost all patients experience tremor reduction immediately after thalamotomy, and roughly 70% of patients continue to benefit from thalamotomy beyond a follow-up period of one year (Table 3). In the three studies on thalamic DBS with a follow-up period longer than one year, 69% to 100% of the patients experienced reduced tremor [41, 84, 102]. Functional improvement after both thalamotomy and DBS is much more variable and unfortunately not reported in many studies (Tables 3 and 4).

Niranjan published an interesting study on gamma-knife thalamotomy as a possible alternative to neurosurgery. In this study all three patients with MS-related tremor experienced marked improvement of tremor after radiosurgery. Unfortunately, no further studies have been published to evaluate this interesting non-invasive treatment option [72].

There are two trials in which thalamotomy and DBS were compared in patients with MS. In a randomized controlled trial conducted by Schuurman and colleagues [85] patients with ET, tremor due to PD and tremulous MS patients were randomized to undergo either VIM thalamotomy or DBS. Five tremulous MS patients were randomized to each group. After a short follow-up period of six months, there were no significant differences between thalamotomy and DBS in functional outcome for the MS subgroup. In the same study, fewer adverse effects were observed in the DBS group, except for the subgroup of MS patients, where adverse effects were about equal for both interventions (Tables 3 and 4). Bittar and colleagues compared cumulative tremor scores for intention and postural tremor between patients undergoing VOP/ZI thalamotomy or DBS in a non-randomized study [15]. Ten patients were in each group. Interestingly, after a mean follow-up period of 15 to 16 months, patients in the thalamotomy group had a better outcome (tremor reduction: 78% for postural tremor and 72% for intention tremor) than patients in the DBS group (tremor reduction: 64% for postural tremor and 36% for intention tremor). However, as there were more adverse effects in the thalamotomy group, no clear recommendations could be given.

Adverse events reported for the neurosurgical interventions comprised increase of hemiparesis, dysarthria, dysphasia, mental changes, depression, seizures, intracerebral haemorrhage, subdural haematoma, wound infection and MS relapse. Thalamotomy was associated with a higher risk of adverse events than DBS. Bilateral thalamotomy carries such a high risk of adverse effects, that it is no longer recommended. If bilateral treatment is necessary, either bilateral DBS or unilateral thalamotomy followed by contralateral DBS are possible treatment options.

Other treatment options

Electromagnetic fields, limb cooling, physiotherapy, weight bracelets, orthoses and specialized software have been advocated as additional treatment options.

Sandyk and Dann reported a reduction of intention as well as postural tremor in three tremulous MS patients treated with pulsed electromagnetic fields [82], but these interesting findings have not been substantiated in a larger trial.

Albrecht and coworkers published a small clinical trial on the effect of arm cooling on intention tremor [3]. In their study, patients achieved significantly better results on a clinical testing battery after immersion of the tremulous arm in ice water. As this effect lasted for about 45 minutes the authors recommend limb cooling to achieve transient tremor control for activities such as working with a PC, signing a document or self-catherisation. In a similar study, Feys and colleagues report tremor reduction lasting for about 30 minutes after limb cooling with a special cooling device [35].

Weighted wrist cuffs are a mechanical tool to reduce tremor amplitude, and one article reported their beneficial effects in three MS patients [28]. Although wearing bracelets decrease tremor amplitude and therefore may offer some benefit to tremulous patients, their effect on intention tremor is small and their use is therefore unlikely to yield important functional improvement. A more advanced computer aided tremor reducing orthosis provided functional benefit in a small case series including patients with MS [1].

Physiotherapy aimed at improving ataxia in MS was evaluated by Armutlu and coworkers. In their small pilot study they found that rehabilitative physiotherapy using Johnstone pressure splints was superior to physiotherapy alone [9]. Unfortunately no larger trials followed this pilot study, so that it remains uncertain which patients may benefit from which form of physiotherapy.

The use of a mouse driven computer system is a special challenge for tremulous MS patients. Feys and colleagues published a study on the use of specialized software developed to aid computer use in 36 tremulous MS patients and found significant improvement in the time needed to complete some basic mouse driven computer operations [36].


The treatment of tremor remains a great challenge for everyone caring for patients with MS. Drug treatment with currently available medication is unsuccessful in most cases and much more research on the pathophysiology and biochemistry of tremor production in MS will be necessary before an efficient medical treatment can be developed. Stereotactic surgery can be an effective means to treat severe tremor, but it is currently uncertain whether lesional surgery or DBS is the treatment of choice. Larger clinical trials comparing both interventions are needed. Other treatment options, including physiotherapy, tremor reducing orthoses, and limb cooling can lead to valuable improvements in activities of daily living.