Discussion

Tremor is a common feature among patients with AOPD, with arm and head tremor occurring in 30% and 40% of cases, respectively. Previous studies have reported tremor frequencies ranging from 7% to 70% in AOPD.2 ,8–10 Such a huge discrepancy between studies might be due to several factors, including small patient samples, with focus on only one type of focal dystonia, and use of medical chart reviews rather than clinical examination to detect tremor. Intriguingly, the largest study which specifically addressed this question in 429 AOPD patients, found postural or postural/kinetic tremor in 16.7% of the cases, while none had rest tremor.2 We do not believe that methodological issues can explain such discrepancy: similarly to Defazio et al, we have, in fact, clinically assessed consecutive AOPD patients, with the specific aim to look for tremor and its characteristics.

Dystonic tremor is a relatively new category in tremor classification.1 It is very likely under-recognised and might be mistaken as ET or even PD, depending on the tremor type and distribution. This relies on the lack of markers for the differential diagnosis of tremulous condition, and highlights the difficulty with the clinical diagnosis of dystonia. Among tremulous disorders, ET is undoubtedly overdiagnosed.11 In fact, a review of 350 patients supposed to have ‘pure ET’ has disclosed associated dystonia, including cervical dystonia, writer’s cramp, spasmodic dysphonia, and cranial dystonia, in half of them.12 The suggestion that dystonic patients are frequently misdiagnosed as ET has been further confirmed by others.13 This may also depend on the fact that tremulous dystonic patients may have such confounding clues as autosomal dominant familial history and alcohol responsiveness.6

With regard to head tremor, Agnew et al14 have recently shown that in ET cases it may represent a postural tremor that dissipates when patients lie down, whereas in patients with cervical dystonia (CD), head tremor more often persists. Such a phenomenon was not tested during our audit, but we think it may be unwise to consider ET in subjects with isolated head tremor, or in whom head tremor is more severe than their arm tremor.15 In fact, Louis et al studied the prevalence of ‘isolated head tremor’ among 583 ET patients, and did not identify any case of head tremor in complete absence of arm tremor, and only 2.7% of cases with head tremor and only mild arm tremor.16 This confirms previous findings by Bain et al17 among 20 families with three-generational pure ET. Patients with PD may seldom present with head tremor, but it is usually not a feature of PD and has been very rarely described.18 Moreover, we have demonstrated that head tremor is much more common in dystonic patients than in PD cases.5 All these evidences suggest that head tremor is rather suggestive of concomitant dystonia.

With a view to rest tremor, it is highly suggestive of PD, but here we show that it may occur in up to 40% of tremulous dystonic patients (meaning approximately 12% of all patients with AOPD), being unilateral in more than two-thirds of them. In the majority of them (92.9%), rest tremor was asymmetric, if not unilateral. This result is in line with our previous observations that a number of patients with unilateral rest tremor and preserved dopaminergic nigrostriatal pathway (ie, patients supposed to be affected with PD, but having scans without evidence of dopaminergic deficit—SWEDDs), have dystonic tremor as per the consensus classification4 and also by electrophysiology.5 Such tremor characteristics as frequency and amplitude seem to be not useful to differentiate between these two conditions (PD and dystonic patients may indeed both show a tremor at about 5–6 Hz).5 However, it is remarkable that none of our patients had re-emergent tremor (ie, a postural tremor that emerges after a variable latency of several seconds) while this phenomenon is commonly seen in patients with PD. On the other hand, rest tremor can be seldom seen in ET, and the existence of an ET-subgroup with associated rest tremor has been suggested.1 ,10 However, a recent study has shown that ET patients with associated rest tremor display a blink reflex recover cycle similar to that detected in dystonic tremor, thus indicating that they may have a dystonic tremor rather than a subtype of ET.19 Once again, this evidence suggests that tremulous dystonic patients are likely misdiagnosed. This may depend on the fact that dystonic features can be mild and involve body parts not affected with tremor, or even that have not yet appeared.15

When we compared patients with head versus arm tremor regardless of dystonia classification, we found similar age, disease duration and sex distribution. We also found that dystonic symptoms more frequently involved body regions affected by tremor, suggesting that these two phenomena often coexist. On the other hand, tremor involving body regions which were not affected by dystonic symptoms (ie, ‘tremor associated with dystonia’ as per consensus classification1), was less frequently observed.

In our cohort, we have found patients with segmental or multifocal dystonia to be more likely tremulous than patients with focal dystonia. This suggests a stronger tendency for spread of dystonic features in patients with tremor, also in line with previous reports.2 ,6 However, whether the presence of tremor in a body segment may anticipate the spread of dystonia in that segment needs to be elucidated in further prospective studies. Moreover, patients with segmental and multifocal dystonia had a longer disease duration (21 and 23 years, respectively), than patients with focal dystonia (15 years), even though this was not significant (p=0.09). On the other hand, tremulous patients had a longer disease duration regardless of the dystonia classification (data not shown), in line with existing literature.7 This may suggest that tremor can be a relatively late symptom in patients with AOPD. The long disease duration across the subgroups may also partly account for the discrepancy with previous observations which have included patients with shorter disease duration.7 However, we miss data regarding tremor versus dystonia onset. We therefore cannot totally exclude that some patients have developed dystonic symptoms after tremor onset. Moreover, the possibility exists that, at least in some patients, dystonia and tremor are unrelated phenomena, leading to an overestimation bias. However, we believe this is unlikely, given the large sample size.

Neurologists and other practitioners should be aware of the possibility of dystonic tremor. It is frequent in patients with AOPD, with the head being the most common affected site. Arm tremor occurs in one-third of AOPD patients, with or without associated head tremor: arm tremor usually features a postural component, but can be also characterised by kinetic and rest components in some cases. Rest tremor presents more frequently unilaterally (and it is usually asymmetric in patients with bilateral rest tremor), thus it would be wise to consider dystonia even in the differential diagnosis of PD. Our clinical experience would suggest that such features as position/task specificity, jerkiness, presence of tremor-flurries and thumb hyperextension are fairly typical for dystonic tremor.5 ,15

We acknowledge some limitations. First, we did not perform electrophysiological recordings, and it may be argued that rest tremor seen in our patients (which was assessed with the arms in semisupination/pronation on their lap) is actually postural tremor that continues when the limb is supported. However, this is a matter of further laboratory research and does not correspond to what is seen in clinical practice. It may be also argued that some patients, particularly those with head tremor, may adopt abnormal head postures in order to compensate or suppress their tremor. This might mislead to the diagnosis of dystonia, but we do not believe this is the case here. Our patients were in fact diagnosed according to established clinical criteria,7 and all our CD cases showed restricted head motion due to paradoxical muscle overactivity, which is suggestive of underlying dystonia. As a tertiary centre for movement disorders, we cannot exclude a referral bias. However, we do not believe that this may account for the whole discrepancy between ours and others studies. We would like to remark the importance of careful clinical examination when assessing tremulous patients, looking for dystonic symptoms also in non-tremulous body segments.