Comparing Essential Tremor with and without Soft Dystonic Signs and Tremor Combined with Dystonia: The TITAN Study

Abstract

Background

Tremor disorders remain as clinical diagnoses and the rate of misdiagnosis between the commonest non‐parkinsonian tremors is relatively high.

Objectives

To compare the clinical features of Essential Tremor without other features (pure ET), ET plus soft dystonic signs (ET + DS), and tremor combined with dystonia (TwD).

Methods

We compared the clinical features of patients with pure ET, ET + DS, and TwD enrolled in The ITAlian tremor Network (TITAN). Linear regression models were performed to determine factors associated with health status and quality of life.

Results

Three‐hundred‐eighty‐three patients were included. Sex distribution was significantly different between the groups with males being more represented in pure ET and females in TwD. The initial site of tremor was different between the groups with about 40% of TwD having head tremor and ET + DS unilateral upper limb tremor at onset. This pattern mirrored the distribution of overt dystonia and soft dystonic signs at examination. Sensory trick, task‐specificity, and position‐dependence were more common, but not exclusive, to TwD. Pure ET patients showed the lowest degree of alcohol responsiveness and ET + DS the highest. Midline tremor was more commonly encountered and more severe in TwD than in the other groups. Regression analyses demonstrated that tremor severity, sex, age, and to a lesser degree the variable “group”, independently predicted health status and quality of life, suggesting the existence of other determinants beyond tremor.

Conclusions

Pure ET and TwD manifest with a phenotypic overlap, which calls for the identification of diagnostic biomarkers. ET + DS shared features with both syndromes, suggesting intra‐group heterogeneity.

Tremor disorders remain as clinical diagnoses, their definition having been revised in 2018 by the International Parkinson and Movement Disorder Society (IPMDS). ¹ The 2018 IPMDS tremor classification proposed new tremor entities previously subsumed within Essential Tremor (ET), for example focal tremors of the head or voice, with the aim of improving phenotypic homogeneity of ET and similar disorders. ¹ , ² In fact, misdiagnosis rates have been reported to occur in up to 50% of “ET” cases ³ , ⁴ and instead, the importance of correct diagnoses is supported by the fact that evidence‐based treatment options are deemed to differ between ET and other conditions which were previously lumped together with ET, including tremor combined with dystonia (TwD). ⁵ , ⁶ , ⁷ , ⁸ The over‐diagnosis of ET has been often advocated as one of the reasons explaining some of the early challenges in understanding disease pathophysiology and pathology ⁹ and this calls for the need of identifying phenotypic features that might aid the differential diagnosis between ET and similar disorders, given the absence of diagnostic biomarkers. Although some phenotypic features have been proposed to be suggestive of dystonic tremor, ¹⁰ , ¹¹ , ¹² , ¹³ there are, to our knowledge, no large studies that have directly compared patients with ET and TwD.

Moreover, complicating this issue, there has been the recent proposal of ET‐plus for those ET patients who, in addition to the core phenotype of action tremor, also exhibit soft signs of uncertain relationship to the tremor including subtle dystonia. ¹ It is unclear whether such patients represent a subgroup of ET, reflect a more advanced ET stage, ¹⁴ a forme fruste of TwD, or a different entity altogether. Preliminary evidence has been produced that they might not have ET ¹⁵ and that they would have gray matter brain structural alterations with the involvement of the basal ganglia‐cortical loop similar to what is observed in TwD. ¹⁶ However, there are no clinical studies comparing patients with ET plus dystonic soft signs (ET + DS) with both ET and TwD.

The ITAlian tremor Network (TITAN) is a multicenter data collection platform, ¹⁷ aiming to prospectively assess the phenomenology and natural history of tremor syndromes and to serve as a basis for future etiological, pathophysiological, and therapeutic research. The objective of the current study was to compare demographic and clinical features of patients with ET, ET + DS and TwD and to this aim, we extracted the data of these patients and present the cross‐sectional analysis of their baseline data.

Methods

The protocol of the TITAN study has been published elsewhere ¹⁷ and is available in the supplemental material. For the current work, we extracted from the TITAN platform data of patients with “pure” ET (eg, without any additional features including soft signs), ¹⁸ those with ET + DS, therefore excluding patients with ET+ with alternative soft signs, and patients with TwD with tremor in the upper limbs (UL) (eg, we excluded TwD patients without UL tremor, for instance with head tremor only, because of the focus of the current work to compare TwD with patients with ET/ET+ who, by definition, have UL tremor). TwD comprises both patients with “dystonic tremor” (eg, dystonia and tremor present in the UL) and “tremor associated with dystonia” (eg, TAWD due to the dystonia being present elsewhere). ¹

All patients were assessed by neurologists with clinical/research focus on movement disorders using a standardized protocol ¹⁷ that included a structured interview to gather demographic and historical information about the tremor syndrome as detailed in the supplemental material; Data S1.

Patients were then assessed with The Essential Tremor Rating Assessment Scale (TETRAS) ¹⁹ which is composed of the performance subscale (TETRAS_P) and of the activity of daily living (ADL) subscale (TETRAS_ADL). Using individual items of the TETRAS_P we computed a score for “postural tremor”, “kinetic tremor”, “postural‐kinetic index” and “asymmetry index” as detailed in the supplemental material; Data S1. Patients were further assessed with the Scale for the assessment and rating of ataxia (SARA) ²⁰ and were asked to report their perceived level of quality of life (QoL) in general and in terms of their health (HR‐QoL).

After the release of the first protocol and the beginning of the research activities, ¹⁷ an amendment was approved to gather information about alcohol responsiveness that patients had to report on a scale from 0 (no response/not available) to 10 (best imaginable response). As such, this information was available only for a subset of the recruited sample (205/383; 53.5%).

The study has been approved by the ethic committee of the coordinator centre (University of Salerno; study approval n.33_r.p.s.o._02/10/2020) and all subjects were requested to provide a written consent form to participate.

Descriptive and non‐parametric analyses have been performed as detailed in the supplemental material using STATA (StataCorp. 2023. Stata Statistical Software: Release 18. College Station, TX: StataCorp LLC).

Results

A total of 383 patients (251 with ET, 46 with ET + DS, and 86 TwD) were extracted from the TITAN database for the current work. ET + DS represented 12.95% of the entire ET+ population (N = 355), whereas the patients with TwD included in this analysis represented 66.67% of the entire TwD population (N = 129), the remaining having been excluded because of the absence of UL tremor. The body distribution of soft dystonic signs in ET+ was found to be significantly different than that of overt dystonia in patients with TwD (x ² = 42.39, P < 0.001, Fig. 1). Of the subjects with TwD, 27 patients had dystonic tremor (eg, dystonia in the UL) and 59 patients had TAWD (eg, dystonia being present elsewhere). Patients with TAWD were more likely to be female, to have tremor onset in the head and greater involvement and severity of head tremor than patients with dystonic tremor (Table S1); otherwise, the two subgroups were not different in terms of other demographics, of any features of upper limb tremor, of the total TETRAS and its subscales, and of the SARA scale (Table S1).

Figure 1.

Distribution of the soft dystonic signs in ET + DS and of dystonia in TwD. ET + DS, essential tremor plus soft dystonic signs; TwD, tremor combined with dystonia.

The three main groups were similar in terms of age, AAO and disease duration (Table 1), thus excluding these factors as potential confounders for subsequent analyses. Sex distribution was found to be significantly different with males being more represented in the ET group, females in the TwD group, and ET+ exhibiting a male/female ratio of 1/1 (χ ² = 7.53, P < 0.05, Table 1). About one‐third of patients with TwD reported a positive FH for tremor, in contrast to over 40% in ET and ET + DS (χ ² = 6.74, P < 0.05, Table 1).

TABLE 1.

Demographic and historical features of the groups

	ET (N = 251)	ET + DS (N = 46)	TwD (N = 86)	P
Sex (M:F)	1.3:1	1:1	1:1.5	0.023
Age [y, median (IQR)]	68 (15)	71 (12)	66.5 (88)	0.349
FH for tremor [N (%)]	113 (45.02%)	19 (41.30%)	25 (29.07%)	0.034
Onset [y, median (IQR)]	54 (40)	56 (47)	55 (25)	0.996
Disease duration [y, median (IQR)]	10 (25)	11.5 (31)	10 (19)	0.609

Abbreviations: ET, essential tremor; ET + DS, ET plus soft distonic signs; TwD, tremor combined with dystonia; IQR, interquartile ranges; FH, family history.

The body region at onset was different between the group with the UL being the initial site of tremor onset in over 90% of ET and ET + DS patients and about half of the patients with TwD (χ ² = 87.93; P < 0.001). A significantly higher proportion of ET + DS patients reported the UL tremor to be initially unilateral (43.48%) as compared to about 25% of ET and TwD patients (χ ² = 6.78, P < 0.05, Table 2). Task‐specificity at onset was reported to occur in 20.83% of TwD, but also in 13.86% of ET and in only 2.83% of ET + DS (χ ² = 6.06; P < 0.05). Presence of sensory trick was significantly higher in TwD (13.75%) than in ET/ET + DS (less than 1%; χ ² = 23.47; P < 0.001) whereas the presence of a position‐dependent component only reached a statistical trend (P = 0.055; Table 2). Response to alcohol was different between the groups (Kruskal‐Wallis χ ² = 6.79; P < 0.05): post‐hoc test demonstrated this result was driven by ET + DS reporting the highest response as compared to ET [median (IQR): 5.5 (3.25) vs. 3 (3), respectively, Mann–Whitney z = −2.89; P = 0.039], whereas no difference was found between ET and TwD (Table 2).

TABLE 2.

Clinical features of the three groups

	ET (N = 251)	ET + DS (N = 46)	TwD (N = 86)	P
Onset site (N, %) ç				<0.001
‐ Head	13 (5.18%)	3 (6.52%)	34 (39.53%)
‐ Face	1 (0.40%)	0 (−)	3 (3.49%)
‐ Voice	1 (0.40%)	1 (2.17%)	1 (1.16%)
‐ Upper limbs	234 (93.6%)	42 (91.30%)	48 (55.81%)
‐ Lower limbs	1 (0.40%)	0 (−)	0 (−)
Initial upper limbs involvement				0.029
Unilateral [N (%)]	63 (25.10%)	20 (43.48%)	28 (32.55%)
Bilateral [N (%)]	188 (74.90%)	26 (56.52%)	58 (67.44%)
Task‐specificity at onset (N, %) §	28 (13.86%)	1 (2.63%)	10 (20.83%)	0.143
Sensory trick (N, %) ^	1 (0.55%)	0 (−)	11 (13.75%)	<0.001
Position‐dependence (N, %) #	13 (7.34%)	4 (10.81%)	13 (17.57%)	0.055
Response to alcohol [median (IQR)] @	3 (3) a	5.5 (3.25) c	4 (2.5)	0.033
Midline tremor (N, %)	95 (37.85%)	22 (47.83%)	64 (74.42%)	<0.001
‐ Head (N, %)	57 (22.7%)	16 (34.7%)	52 (60.4%)	<0.001
‐ Face (N, %)	37 (14.7%)	12 (26.1%)	27 (31.4%)	<0.001
‐ Voice (N, %)	58 (23.1%)	15 (32.6%)	39 (45.3%)	<0.001
Head tremor severity [median (IQR)]	0 (0) a , b	0 (1) b , c	1 (2) a , c	<0.001
Face tremor severity [median (IQR)]	0 (0) a , b	0 (1) c	0 (1) c	<0.001
Voice tremor severity [median (IQR)]	0 (0) b	0 (1)	0 (1) c	<0.001
Upper limb action tremor
Postural [N(%)]	247 (98.41%)	46 (100%)	76 (88.37%)	<0.001
Kinetic [N(%)]	248 (98.80%)	45 (97.83%)	79 (91.86%)	0.005
Postural tremor
Severity [median (IQR)]	5.5 (3) a	6 (1.5) b , c	5 (3.5) a	<0.001
Asymmetry index [median (IQR)]	0 (0.75)	0.125 (1)	0 (0.25)	0.906
Kinetic tremor
Severity [median (IQR)]	6 (4)	6.5 (3.5)	6 (4.5)	0.152
Asymmetry index [median (IQR)]	0 (0.75)	0.125 (0.75)	0 (0.25)	0.636
Postural‐kinetic index [median (IQR)]	−0.5 (2)	0 (3)	−1 (3)	0.245
Upper limbs rest tremor
N (%)	‐	18 (39.13%)	39 (45.35%)	0.492
Unilateral/bilateral [%]	‐	61.11%/38.89%	30.77%/69.23%	0.029
Asymmetry index [median (IQR)]	‐	0 (0.1)	0 (0.2)	0.812
Leg tremor	94 (37.45%)	12 (26.09%)	24 (27.91%)	0.133
TETRAS
Total	30 (19)	36.5 (22.5)	28 (23.5)	0.052
Performance subscale	18 (9.5)	20 (10)	16 (12.5)	0.135
ADL subscale	12 (10)	14.5 (15) b	9.5 (15) a	0.037
SARA [median (IQR)]	1 (2.5)	1 (3.5)	1 (3.5)	0.895
HR‐QoL	75 (20) b	70 (20)	70 (30) c	0.019
QoL	80 (30) a	70 (30) c	70 (30) c	0.018

Abbreviations: ET, essential tremor; ET + DS, ET plus soft dystonic signs; TwD, tremor combined with dystonia; IQR, interquartile ranges; FH, family history; TETRAS, the ET rating assessment scale; SARA, scale for the assessment and rating of ataxia; QoL, quality of life; HR‐QoL, health related QoL.

^a Different from ET + DS (post‐hoc corrected P < 0.05).

^b Different from TwD (post‐hoc corrected P < 0.05).

^c Different from ET (post‐hoc corrected P < 0.05).

^ç 2 missing values (1 ET, 1 TwD).

^§ 57 missing values (49 ET, 8 ET + DS).

^{^} 84 missing values (69 ET, 9 ET + DS, 6 TwD).

^# 95 missing values (74 ET, 9 ET + DS, 12 TwD).

^@ 178 missing values (117 ET, 20 ET + DS, 41 TwD).

Almost the entire sample of ET and ET + DS patients manifested with both postural and kinetic tremor, whereas this was not the case for TwD (97.2%, 97.8%, and 81.4%, respectively, P < 0.01), but there were no differences between the groups in terms of tremor severity and asymmetry index of either component and postural‐kinetic ratio. Rest tremor was found in about 40% and 45% of ET + DS and TwD, respectively (χ ² = 0.472; P = 0.492), but it was more frequently unilateral in the former than in the latter (χ ² = 4.711; P < 0.01; Table 2). Midline tremor was more frequent in TwD (~75%) than in ET and ET + DS (~38% and ~48%, respectively; χ ² = 34.37; P < 0.001; Table 2) and was significantly different in terms of severity with ET patients manifesting the lowest scores of head (Kruskal‐Wallis χ ² = 46.19, P < 0.001; post‐hoc Mann–Whitney as follows: ET vs ET + DS: z = −2.016, P = 0.043; ET versus TwD: z = −6.82, P < 0.001; ET + DS versus TwD: z = −2.561, P = 0.011; Table 2), face (Kruskal‐Wallis χ ² = 14.16, P < 0.001; post‐hoc Mann–Whitney as follows: ET versus ET + DS: z = −2.00, P = 0.045; ET versus TwD: z = −3.69, P < 0.001; ET + DS versus TwD: z = −0.821, P = 0.411; Table 2) and voice tremor (Kruskal‐Wallis χ ² = 15.01; P < 0.001; post‐hoc Mann–Whitney as follows: ET versus ET + DS: z = −1.52, P = 0.129; ET versus TwD: z = −3.87, P < 0.001; ET + DS versus TwD: z = −1.15, P = 0.249; Table 2). No differences were found in terms of leg involvement.

Overall, only a statistical trend was found for the total score of the TETRAS (P = 0.055; Table 2), whereas the TETRAS_ADL was significantly different between the groups (Kruskal‐Wallis χ ² = 6.56; P < 0.05) with ET+ exhibiting the highest and TwD the lowest impact on ADL [15.5 (15) vs. 9.5 (15), median (IQR), respectively, z = 2.383, P = 0.01; Table 2), the direct comparison between ET and ET + DS only approaching a statistical trend (z = −1.83, P = 0.06). However, both HR‐QoL (Kruskal‐Wallis χ ² = 8.77, P < 0.01; post‐hoc Mann–Whitney as follows: ET vs. ET + DS: z = 1.62, P = 0.105; ET vs. TwD: z = 2.53, P = 0.01; ET + DS vs. TwD: z = 0.59, P = 0.555; Table 2) and QoL (Kruskal‐Wallis χ ² = 10.07, P < 0.01; post‐hoc Mann–Whitney as follows: ET vs. ET + DS: z = 2.45, P = 0.014; ET vs. TwD: z = 2.43, P = 0.016; ET + DS vs. TwD: z = −0.57, P = 0.563; Table 2) were found to be significantly different between the groups with TwD and ET + DS displaying worse scores than ET (Table 2).

Secondary analyses to compare ET/ET + DS with patients with dystonic tremor and with TAWD grossly confirmed the main results obtained in the comparisons with the group of TwD taken as a whole (Tables S2 and S3).

Regression analyses showed that the total TETRAS score, age, female sex, and the variable “group” (eg, ET/ET + DS/TwD) were significantly associated with worse HR‐QoL (F = 10.85; P < 0.001; Table 3). The same variables but age, which only approached the statistical threshold (P = 0.067) and “group”, were associated with worse QoL in general (F = 12.17; P < 0.001; Table 3).

TABLE 3.

Results of the linear regression models

	Coefficient	SE	P	[95% CI]
(A) Dependent variable: HR‐QoL
Sex	6.05	1.94	0.002	2.23/9.87
Age	−0.18	0.07	0.017	−0.34/−0.03
TETRAS total score	−0.46	0.12	<0.001	−0.73/−0.21
Group	−0.73	5.31	<0.001	78.45/99.37
(B) Dependent variable: QoL
Sex	9.47	0.13	<0.001	5.42/13.52
Age	−0.15	0.81	0.067	−0.31/0.01
TETRAS total score	−0.49	0.13	<0.001	−0.76/−0.23
Group	−0.45	0.31	0.149	−1.07/0.16

Abbreviations: SE, standard error; CI, confidence intervals; HR‐QoL, health related quality of life; QoL, quality of life.

Discussion

To the best of our knowledge, this is the first study providing a direct comparison of ET and TwD, further including the recently proposed entity of ET + DS according to the 2018 IPMDS classification. ¹ Our results demonstrate that ET and TwD display some differences that might aid the differential diagnosis, but also show that there are several phenotypic features that are shared between the two groups of patients, which hence calls for the identification of robust diagnostic biomarkers to be used in clinical practice. Furthermore, we have here shown that patients with ET + DS have some similarities with both ET and TwD and therefore we cannot conclusively claim whether the presence of soft dystonic signs is prodromal to the development of overt dystonia ¹⁵ , ¹⁶ or instead represent an unspecific finding that is unrelated to the tremor syndrome. ²¹ , ²² , ²³ Importantly, both AAO and disease duration were similar between the groups and therefore these features cannot explain the differences we found between the groups.

FH was found to be lower in TwD than in ET/ET + DS. Of note, FH only reflected familial aggregation of tremor (and not dystonia) and therefore this result supports the concepts that: (1) a dystonia gene might give rise to a predominant, or even isolated, tremor syndrome; and (2) tremor is associated with familial clustering of dystonia. ²⁴ Our results further support the concept of a strong genetic predisposition to ET, in our sample being present in almost 50% of cases, which calls for an effort to discover the genetic basis of the condition.

The initial site of the tremor was found different between the groups, being the head in about 40% of TwD and the UL but unilaterally in about 40% of ET + DS. Intriguingly, this pattern mirrored the distribution of dystonia and soft dystonic signs (Fig. 1) suggesting that the two phenomena tend to co‐exist, which might indirectly support the concept that the identified “soft signs” are indeed dystonic in nature. This suggestion is further supported by the results obtained in the subgroups of patients with TAWD and dystonic tremor. This information might be further useful in the differential diagnosis as head tremor and unilateral UL tremor at onset were observed in a minority of “pure” ET patients (e.g., ~5% and ~25%, respectively).

Interestingly, some phenotypic features that have been suggested to be suggestive of dystonia ¹⁰ , ¹¹ , ¹² were indeed found to be commoner in TwD: this was especially the case for the presence of a sensory trick, but also for task‐specificity at onset whereas only a trend was observed for the presence of a position‐dependent component. We acknowledged that the main analyses have been performed between ET/ET + DS and TwD taken as a whole, because our focus was on the tremor of the upper limbs and no differences were disclosed in terms of the latter between the subgroups of patients with dystonic tremor and TAWD (Tables S2 and S3). Instead, the two subgroups were found to be significantly different for other variables (eg, sex and presence/severity of head tremor) that are in keeping with what is known about the clinical entities of cervical dystonia and of upper limb dystonia. ¹¹ , ²⁵ Not surprisingly, presence of sensory trick was higher in patients with the former (eg, with TAWD) and task‐specificity in patients with the latter (eg, with dystonic tremor of the UL). It should be noted, however, that task‐specificity was reported in a non‐negligible proportion of ET patients and therefore it cannot be considered specific of dystonia. Of note, these data support one of the novelties of the 2018 IPMDS classification, which is the possibility of patients changing diagnostic allocation across time: indeed at least a proportion of these patients would have been labeled as “task‐specific tremor” at onset before developing the criteria for either ET or TwD. Longitudinal studies might reveal if there are phenotypic features at onset predicting future conversion into either syndrome.

Conversely, a feature that has been formerly considered typical for ET (eg, alcohol responsiveness) was found to be significantly different between the groups with ET showing the lowest degree of responsiveness. Although the results need to be confirmed because this variable was available only for about 50% of the included patients, this evidence might support the proposal of the 2018 IPMDS to discard alcohol responsiveness as a supportive criterion for the diagnosis of ET. ¹ On the other hand, our results demonstrate that a considerable proportion of patients with ET/ET + DS and TwD do show a benefit from alcohol ingestion and such an information is important to collect in view of possible future umbrella trials testing sodium oxybate ²⁶ or long‐chain alcohols ²⁷ in alcohol‐responsive tremors, regardless of the diagnostic allocation.

Regarding tremor distribution, action tremor of the UL was found to be similar between the groups whereas a difference was demonstrated in terms of rest tremor that, mirroring the unilateral presentation in ET + DS, was found to be more commonly unilateral in this group than in TwD (about 60% vs. 30%, respectively). Somehow similarly and following the concept that head tremor is the initial presentation in about 40% of TwD, midline tremor was found at examination to be commoner in TwD than in both ET and ET + DS. This finding is in line with previous literature showing that head tremor is the commonest type of tremor in adult‐onset idiopathic dystonia ¹¹ and might suggest an additional/preferential involvement of the cerebellar vermis/fastigial nuclei loop in dystonia as compared to ET/ET + DS. ²⁸ , ²⁹ Head tremor at onset has been in fact suggested to be a marker of an “ataxic” phenotype in cervical dystonia, ³⁰ but we failed to observe differences between the group in terms of SARA scores. This might be due to the fact that the SARA scale is more weighted on appendicular items reflecting the involvement of alternative cerebellar loops, which are also implicated in the generation of action tremor of the UL. ²⁸ Despite being intriguing, this speculation needs to be tested experimentally. Lastly, no differences were found in terms of leg tremor. This would contrast with previous evidence showing that leg tremor would be commoner in ET+ than in ET. ³¹ We note, however, that we only included in the current work ET + DS (representing about 13% of the entire ET+ population) and this might explain the apparent discrepancy. However, the latter indirectly supports the concept that deeper phenotyping is required in ET+ ³² since taken as a whole, it might provide spurious results.

Finally, significant differences were found in terms of tremor impact on ADL and HR‐QoL/QoL. TwD had the lowest impact on ADL but the worse HR‐QoL and QoL in general, which the regression analysis confirmed to be driven also by the tremor severity. This apparent discrepancy might be likely explained by the considerable presence of head tremor in TwD which might not directly impact ADL but can be a significant source of disability because it might be more visible than the action tremor of the UL and cause social stigma. ³³ Interestingly, the regression analysis demonstrated the variable “group” to be independently associated with worse HR‐QoL, beyond higher tremor severity, which implies that additional, paraphs non‐motor, features of these tremor syndromes have a detrimental impact on the perceived health status. Of note, the female sex was associated with worse HR‐QoL and QoL in general. The latter results, which are in line with recent evidence on ET+, ³⁴ , ³⁵ call for the need for a dedicated research effort since sex‐related differences in the field of tremor research have been mostly overlooked.

We acknowledge that our study has some limitations. We cannot entirely exclude a recruitment bias that is inherent to studies without a population‐based design. However, the involvement of both secondary and tertiary movement disorder centers in the TITAN study ¹⁷ might have, at least in part, attenuated this risk, and therefore our account of ET, ET + DS, and TwD should be more realistic than those obtained in single‐center studies. Furthermore, we note that the diagnosis of tremor syndromes is made on a clinical basis as per consensus and this might carry the risk of misdiagnosis possibly resulting in false discoveries. This might be particularly true for the diagnosis of ET, given that current criteria appear to be sensitive but not specific, as well as for the identification of dystonia and/or soft dystonic signs. ²¹ , ³⁶ However, the use of current consensus criteria is the only viable approach at present to ensure the external validity of any research and the comparability across studies ³² , ³⁷ in the global effort of deep‐phenotyping these tremor syndromes and in the search of their diagnostic biomarkers.

Obviously, in the absence of diagnostic gold standards one should consider that the diagnostic agreement between raters might be variable, also because it is influenced by the examiners’ clinical/research focus. ³⁶ The risk of disagreement is inherent to any multi‐center study adopting clinical criteria and we acknowledge that inter‐rater agreement analyses have not been performed, but it might be possible that the inclusion of multiple centers with different expertise might have attenuated the risk of imbalance in terms of diagnostic allocations. One might envision to confirm the current results on a subset of patients in whom the diagnosis is confirmed by multiple raters, but most importantly they have to be replicated in other cohorts.

Moreover, we cannot exclude a certain degree of recall bias regarding the presence of FH. However, it has been demonstrated that the sensitivity of self‐reported FH in patients with ET reaches about 85%. ³⁸ Similarly, self‐reported AAO might also have suffered from a recall bias. However, prior studies have indicated that it is reliably reported by ET patients. ³⁹ We also acknowledge unequal sample sizes between the groups which might represent an issue mostly when performing an analysis of variance, which was not the case here. Finally, we acknowledge that for some features there were variable rates of missing data; however, within each group the comparisons of the main variables (eg, age, sex, TETRAS total score and its subscales, QoL, etc) between patients with and without missing data did not disclose any differences (data not shown), thus excluding significant biases.

In summary, we have here provided a detailed comparison of the demographic and clinical features of ET, ET + DS and TwD, demonstrating some differences that might be useful in the differential diagnosis but also a considerable degree of phenotypic overlap, which calls for the need for diagnostic biomarkers. It might be argued that we did not gather some phenotypic features that have been suggested to possibly differentiate between ET and TwD such as, for instance, tremor irregularities/flurries. ¹³ We note, however, that: (1) there are no operational criteria for these phenotypic features; (2) they have also been described in a proportion of ET cases ⁴⁰ ; and (3) they are not omnipresent in TwD patients, some of whom are reported to have a regular, symmetric tremor that is clinically undistinguishable from ET. ⁴¹ , ⁴² This further calls for the use of additional tool, for instance electrophysiology, and machine learning techniques ⁴³ to gather tremor features which cannot be collected by the naked eye. Finally, we have also shown here that ET + DS has some phenotypic overlap with both “pure” ET and TwD. It might be possible that, as it stands, it represents a heterogeneous entity including different patients, deep phenotyping being therefore requested to clarify the exact nature of their tremor. ³² The analysis of the longitudinal phase of the TITAN study might reveal whether there are baseline features predicting the development of over dystonia and their final diagnostic allocation.

Author Roles

(1) Conception and design of the study, or acquisition of data, or analysis and interpretation of data; (2) Drafting the article or revising it critically for important intellectual content; (3) Final approval of the version to be submitted.

R.E.: 1, 2, 3

G.L.:1, 3

C.T.: 1, 3

G.P.: 1, 3

A.F.G.: 1, 3

R.D.M.: 1, 3

L.M.: 1, 3

F.D.B.: 1, 3

F.V.: 1, 3

V.M.: 1, 3

A.P.: 1, 3

M.E.: 1, 3

E.O.: 1, 3

M.C.M.: 1, 3

R.C.: 1, 3

C.D.: 1, 3

F.S.: 1, 3

A.N.: 1, 3

A.D.R.: 1, 3

R.D.G.: 1, 3

C.S.: 1, 3

A.P.: 1, 3

M.C.A.: 1, 3

C.P.: 1, 3

R.M.: 1, 3

E.C.: 1, 3

A.T.: 1, 3

S.M.: 1, 3

M.B.: 1, 3

V.R.: 1, 3

G.F.: 1, 3

P.B.: 1, 3

Disclosures

Ethical Compliance Statement: This was work has been approved by the ethic committee of the coordinator centre (University of Salerno; study approval n.33_r.p.s.o._02/10/2020) and written informed consent was obtained from the patients prior to their enrolment. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Funding Sources and Conflicts of Interest: This study did not receive any funding nor was performed as part of the employment of the authors. The authors state explicitly that there are no conflicts of interest in connection with this article.

Financial Disclosures for Previous 12 Months: RE receives royalties from publication of Case Studies in Movement Disorders—Common and Uncommon Presentations (Cambridge University Press, 2017) and of Paroxysmal Movement Disorders (Springer, 2020). He has received consultancies from Ipsen and honoraria for speaking from the International Parkinson's Disease and Movement Disorders Society. PB received consultancies as a member of the advisory board for Zambon, Lundbeck, UCB, Chiesi, Abbvie, and Acorda. A received grant support from Ministry of Education, Research and University (MIUR) and IMI H2020 Initiative (IDEA‐FAST project‐MI2‐2018‐15‐06), received research support from Zambon SrL Italy and Bial italy; he received speaker honoraria from Abbvie, Biomarin, Bial and Zambon Pharmaceuticals. AP received grant support from Ministry of Health (MINSAL) and Ministry of Education, Research and University (MIUR), from CARIPLO Foundation; personal compensation as a consultant/scientific advisory board member for Biogen 2019–2021 Roche 2019–2020 Nutricia 2020–2021 General Healthcare (GE) 2019; he received honoraria for lectures at meeting ADPD2020 from Roche, Lecture at meeting of the Italian society of Neurology 2020 from Biogen and from Roche, Lecture at meeting AIP 2020 and 2021 from Biogen and from Nutricia, Educational Consulting 2019–2021 from Biogen.

All other authors have nothing to disclose.

Supporting information

Data S1. Supporting information.

TITAN study group

Luca Angelini; Department of Human Neurosciences, Sapienza University of Rome, Rome, Italy

Simone Aramini; Department of Advanced Medical and Surgical Sciences, Università della Campania “Luigi Vanvitelli”, Napoli, Italy

Ruggero Bacchin; Clinical Unit of Neurology, Department of Emergency, Santa Chiara Hospital, Azienda Provinciale per i Servizi Sanitari (APSS), Trento, Italy

Salvatore Bertino; Department of Clinical and Experimental Medicine, University of Messina, Messina, Italy

Tiziana Benzi Markushi; Department of Neuroscience (DINOGMI), Clinical Neurology, University of Genoa, Genoa, Italy

Marta Bianchi; Neurology Unit, ASST Vallecamonica, Esine, Brescia

Anna Castagna; IRCCS Fondazione Don Carlo Gnocchi Onlus Milan Italy

Federico Colombatto; Department of Translational Medicine, Section of Neurology, University of Piemonte Orientale, Novara, Italy

Eleonora Del Prete; Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy

Lazzaro di Biase; Unit of Neurology, Neurophysiology, Neurobiology, Department of Medicine, Università Campus Bio‐Medico di Roma, Rome, Italy

Alessio Di Fonzo; Neurology Unit, Department of Neuroscience, Dino Ferrari Center, Fondazione IRCCS Ca′ Granda Ospedale Maggiore Policlinico, Milan, Italy

Giovanni Di Maggio; Department of Neurology, Antonio Perrino's Hospital, Brindisi, Italy

Augusta Giglio; Department of Neurosciences and Reproductive and Odontostomatological Sciences, Federico II University, Naples, Italy

Caterina Giordano; Department of Medicine, Surgery and Dentistry “Scuola Medica Salernitana”, Neuroscience Section, University of Salerno, Baronissi (SA), Italy

Nicoletta Manzo; Neurology Unit, San Filippo Neri Hospital ASL Roma1, Rome, Italy

Nicola Modugno; Neuromed Institute IRCCS, Pozzilli (IS), Italy

Giovanni Mostile; University of Catania, Department “G.F. Ingrassia”, Section of Neurosciences, Catania, Italy

Laura Maria Raglione; Unit of Neurology of Florence, Central Tuscany Local Health Authority, Florence, Italy

Maria Russo; Department of Medicine, Surgery and Dentistry “Scuola Medica Salernitana”, Neuroscience Section, University of Salerno, Baronissi (SA), Italy

Giovanna Savorgnan; Neurology Unit, Department of Medical Specialist Area, ASST Pavia, Italy

Assunta Trinchillo; Department of Neurosciences and Reproductive and Odontostomatological Sciences, Federico II University, Naples, Italy

Giampiero Volpe; Neurology Unit San Luca Hospital, Vallo della Lucania (SA), Italy

Stefano Zoccolella; Neurosensory Department, Neurology Unit, San Paolo Hospital, ASL Bari, Italy

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.