Limb Positioning and Magnitude of Postural Tremor Ratings: A Study of 100 Essential Tremor Cases

Abstract

Background:

Essential tremor (ET) is a highly prevalent movement disorder. As the diagnosis is purely clinical, it is important for clinicians to develop a reliable taxonomy of signs and symptoms of ET. Here, we investigate two common clinical maneuvers for determining the presence and severity of postural tremor in ET cases: the evaluation of tremor observed when arms are held in the outstretched and wing posture positions.

Methods:

Two movement disorders neurologists assessed tremor severity from videotaped neurological evaluations of 100 ET cases displaying outstretched and wing posture positions. These raters also completed a forced choice measure indicating whether upper limb tremor was more severe in the outstretched position, the wing posture position, or equivalent in the two positions.

Results:

Our raters judged tremor observed in the wing posture position to be more severe than tremor observed in the outstretched position (average ratings of 1.17 and 1.03, respectively, p’s <0.05). This finding replicated across the two independent neurologists’ ratings of tremor severity. Analyses of the forced choice measure revealed parallel results, with tremor observed in the wing posture position more often identified as more severe than tremor observed in the outstretched position (37.5% versus 14.5%, respectively, p <0.05).

Conclusion:

Our data reveal that ET cases generally display more severe upper limb tremor in the wing posture position than in the outstretched position. These findings having potential clinical and research implications, suggesting that postural tremor in ET may be more easily identified from observations of the wing posture than the outstretched position.

Introduction

Essential tremor (ET) is a movement disorder estimated to affect 1.33% of the entire human population (i.e., all ages) and more than 5% of persons aged 65 and older.¹ The diagnosis can be challenging and diagnostic errors are common.^2,3 Refined clinical phenotypic data could aid in the evaluation and diagnosis of ET.

ET is diagnosed based on the presence and severity of both postural and kinetic tremors. Postural tremor is commonly assessed by observing patients hold their arms in either an outstretched position or a bent (“wing posture”) position.⁴ Both of the more commonly employed clinical tremor rating scales - the Fahn-Tolosa-Marin Clinical Rating Scale and the Essential Tremor Rating Assessment Scale (TETRAS) - assess postural tremor in the outstretched position, whereas only the TETRAS assesses postural tremor in the wing posture position.^5,6 This suggests that the wing posture position may be less commonly utilized; however, we know of no published data assessing which position is more frequently used in clinical practice.

Few studies have assessed the relative severity of postural tremor when evaluated in different positions. In one study of 26 ET cases,⁷ investigators used accelerometry to measure tremor severity in five extended arm positions in the sagittal plane, two in the horizontal plane, and two with elbow flexion. Differences in ET across positions were modest and did not reach statistical significance (see Figure 3A in⁷). In contrast, a study of 76 ET cases, which used the TETRAS, reported that the wing posture position produced greater tremor than the outstretched position.⁶ A smaller study of 12 ET cases found similar results.⁸ In sum, the number of studies is small, sample sizes are modest (n = 12, 26, and 76) and results across all studies are not in complete agreement.

Perceptions among physicians may also be mixed. For example, we surveyed the nine faculty in the Movement Disorders Section of our Department, querying them about the relative severity of tremor in these two positions in ET. Five (55.5%) endorsed the wing posture position as the one typically characterized by the greater severity, one (11.1%) endorsed the outstretched position, and the remainder (33.3%) indicated that tremor would be equal in the two positions.

Our primary question was whether ratings of postural tremor in ET would be greater in the outstretched or in the wing posture position. Additionally, we evaluated whether observed differences would replicate across raters, across dominant and non-dominant arms, or across cases with different baseline levels of tremor.

Methods

Study Sample

Cases enrolled between July 2017 and March 2020 in a National Institutes of Health (NIH)-funded observational clinical research study of ET at Yale University, where one of the authors (E.D.L.) was on faculty. All cases signed informed consent approved by the university ethics committee. These cases were recruited either from movement disorder clinics at the university practice or through advertisements through the International Essential Tremor Foundation. Each completed a battery of self-report medical and tremor assessments and a detailed videotaped neurological examination at the time of enrollment. Based on these data, a movement disorders neurologist with special interest in tremor (E.D.L.) assigned Washington Heights-Inwood Genetic Study of Essential Tremor (WHIGET)-based diagnosis of ET.⁹

For the current analysis, we began with a sample of 191 videotaped ET cases. We excluded cases with (1) an additional diagnosis of Parkinson’s Disease (n = 9), (2) an additional diagnosis of dystonia (n = 13), or (3) a zero rating for postural tremor in both arms assessed in the outstretched and wing posture positions (n = 42). Given the fact that many of our results would be reported as percentages, values based on a final sample size of 100 were viewed as particularly straightforward. Hence, from the remaining 127 cases, we selected 100 based on date of enrollment (earliest selected first). The mean age of these 100 cases was 69.2 ± 11.0 years and 52.0% were male.

Procedure for Rating Postural Tremor Severity

Two movement disorders neurologists (E.D.L. and M.S.) re-reviewed videotaped neurological examinations of the 100 ET cases and assigned clinical ratings of postural tremor severity. During the neurological examination, trained study personnel videotaped participants holding their arms first in the outstretched position (both arms extended forward) and then in the wing posture position (both arms held at shoulder height with elbows pointing away from the midline with both pronated hands pointing towards the midline). For each position, the fingers were splayed slightly but not with high tension. Each participants’ hands and arms were recorded in each position for ten seconds. E.D.L. and M.S., independent of one another, evaluated postural tremor severity for each position in the dominant and again in the non-dominant arm, yielding a total of 400 evaluations (100 cases × 2 positions × 2 arms) per rater. Tremor severity was assessed using the WHIGET scale, which is both reliable and valid.^10,11 Response options were absolutely no visible tremor (0); very low amplitude tremor and almost never present (0.5); low amplitude tremor or intermittent tremor (1); moderate amplitude and clearly oscillatory tremor, but only sometimes of moderate amplitude (1.5); moderate amplitude [1–2cm] and clearly oscillatory tremor, and usually of moderate amplitude (2); and large amplitude (3).⁹ Inter-rater reliability analyses demonstrated significant agreement between our two raters’ assessments of tremor severity. Specifically, intraclass r’s ranged from 0.88 to 0.91, all p’s < 0.001, documenting considerable consistency in judgments across observers. Raters additionally were asked to judge whether tremor severity was greatest in the outstretched position, greatest in the wing posture position, or equivalent. This forced choice question was completed separately for the dominant and non-dominant arms. This forced choice question was added because there are instances in which a rater might assign the same ordinal category for tremor in each of the two positions (e.g., a rating of 2 for both the outstretched and wing posture positions) yet also observe that the tremor is slightly worse in one of the two positions. The forced choice question captures that additional information. For the forced choice measure, percent agreement between raters ranged from 53.6% to 68.0%, all p’s < 0.001.

Statistical Analyses

Because the WHIGET is not an interval level scale, we used the Wilcoxon signed-rank test to analyze tremor severity. These analyses were conducted separately for each rater and for the dominant and non-dominant arms. The forced choice measures were analyzed using chi-square tests.

Additionally, we stratified cases based on the severity of their tremor. Postural tremor severity ratings ranged from a low of 0 to a high of 3. Cases receiving a rating of 3 on at least one of the four assessments (outstretched dominant, outstretched non-dominant, wing posture dominant, wing posture non-dominant) from at least one of the two raters were assigned to the higher severity group (n = 22). Cases receiving no ratings of 3 were assigned to the lower severity group (n = 78).

Results

Case Characteristics

The sample ranged in age from 30 to 87 years (mean age = 69.2 ± 11.0 years), and mean age of reported tremor onset 42.5 ± 21.5 years. Most participants were non-Hispanic white (95.0%) and right-handed (87.0%). About one-half was male (52.0%). Fifty-seven cases used tremor medication (57.0%). Tremor severity did not differ between those using and not using tremor medication, M’s = 1.12 and 1.16 respectively, p = 0.45.

Tremor Ratings as a Function of Wing posture vs Outstretched Position

In general, cases displayed more severe tremor ratings in the wing posture than in the outstretched position (Table 1). More specifically, Wilcoxon signed-rank tests revealed that this difference was significant for tremor severity ratings of the non-dominant arm provided by Rater 1 and Rater 2, as well as for Rater 2’s ratings of the dominant arm (Table 1). For Rater 1, the mean value of dominant arm tremor was 1.06 in the wing posture position and 0.97 in the outstretched position, representing a modest (i.e., 9.3%) difference that was in the same direction as the other differences; however this 9.3% difference did not reach conventional levels of statistical significance. Overall, tremor observed in the wing posture position was more severe than tremor observed in the outstretched position (average ratings of 1.17 and 1.03, respectively, Wilcoxon Z = 3.74, p’s <0.05).

Table 1.

Tremor Severity as a Function of Position

		Outstretched Position	Wing posture Position	pa
Rater 1
	Dominant Arm	0.97 ± 0.61	1.06 ± 0.69	0.14
	Nondominant Arm	0.88 ± 0.60	1.06 ± 0.69	0.001
Rater 2
	Dominant Arm	1.19 ± 0.73	1.43 ± 0.76	0.002
	Nondominant Arm	1.11 ± 0.76	1.38 ± 0.77	0.001

^aWilcoxon signed-rank test

Values represent mean ± SD. Bolded values are significant at p <0.05.

Chi-square analyses of the forced choice data revealed a consistent pattern (Table 2). When asked to identify the position in which greater tremor was displayed, the wing posture position was chosen more often than the outstretched position overall (37.5% versus 14.5%, respectively, p <0.05). This pattern was observed for both Rater 1 and Rater 2.

Table 2.

Forced Choice Measure of Tremor Severity as a Function of Position

	Response Option:	Outstretched	Wing posture	Same	pa
Rater 1
	Dominant Arm	16a (16.0)	31b (31.0)	53c (53.0)	0.001
	Nondominant Arm	14a (14.0)	30b (30.0)	56c (56.0)	0.001
Rater 2
	Dominant Arm	14a (14.0)	44b (44.0)	42b(42.0)	0.001
	Nondominant Arm	14a (14.0)	45b (45.0)	41b (41.0)	0.001

Values are numbers and (row percentages) of cases identified as displaying greater severity in outstretched position, greater severity in wing posture position, and same level of severity in both positions.

^aChi Square test.

Bolded numbers indicate significant p values at p <0.05.

Percentages within a given row with different subscripts are significantly different at p<0.05.

Given a significant overall p value, subscripts identify which specific pairs of percentages within a given row are significantly different. Values that share a common subscript do not differ at p < 0.05. Conversely, values that do not share a common subscript differ at p < 0.05.

Tremor Severity Scores as a Function of Tremor Severity Classification

To determine whether the observed differences were consistent across cases displaying different severities of tremor, participants were further stratified by the previously described tremor severity categories. These analyses were conducted separately for low (n = 78) and high severity cases (n = 22). Results for the low and high severity groups paralleled those observed for the entire sample (Tables 3 and 4). In five of the eight comparisons displayed in Table 3, cases received significantly higher tremor scores in the wing posture than in the outstretched position, all p’s <0.04. Although the three remaining comparisons did not attain statistical significance, in each, the outstretched position score was lower than the wing posture score and in two, the p value was marginal (≤ 0.09).

Table 3.

Tremor Severity Scores as a Function of Position and Tremor Severity

		Outstretched Position	Wing posture Position	p a
Rater 1
Lower Severity Tremorb
	Dominant Arm	0.86 ± 0.48	0.88 ± 0.51	0.76
	Nondominant Arm	0.76 ± 0.46	0.86 ± 0.48	0.06
Higher Severity Tremorc
	Dominant Arm	1.34 ± 0.84	1.71 ± 0.84	0.09
	Nondominant Arm	1.32 ± 0.82	1.75 ± 0.87	0.01
Rater 2
Lower Severity Tremorb
	Dominant Arm	1.01 ± 0.54	1.16 ± 0.47	0.02
	Nondominant Arm	0.92 ± 0.54	1.10 ± 0.41	0.01
Higher Severity Tremorc
	Dominant Arm	1.82 ± 0.96	2.37 ± 0.82	0.04
	Nondominant Arm	1.78 ± 1.04	2.36 ± 0.94	0.02

Values represent mean ± standard deviation.

Bolded values are significant at p <0.05.

^aWilcoxon signed-rank test

^bN=78; Cases receiving a rating of 3 on at least one of the four assessments (outstretched dominant, outstretched non-dominant, wing posture dominant, wing posture non-dominant) from at least one of the two raters were assigned to the higher tremor severity group.

^cN=22; Cases receiving no ratings of 3 were assigned to the lower tremor severity group (n = 78).

Table 4.

Forced Choice Measure as a Function of Position and Tremor Severity

	Response Option:	Outstretched	Wing posture	Same	pa
Rater 1
Lower Severity Tremorb
	Dominant Arm	13a (16.7)	19a (24.4)	46b (59.0)	0.001
	Nondominant Arm	11a (14.1)	19a (24.4)	46b (61.5)	0.001
Higher Severity Tremorc
	Dominant Arm	3a (13.6)	12b (54.5)	7ab (31.8)	0.06
	Nondominant Arm	3a (13.6)	11b (50.0)	8ab (36.4)	0.11
Rater 2
Lower Severity Tremorb
	Dominant Arm	9a (11.5)	30b (38.5)	39b (50.0)	0.001
	Nondominant Arm	12a (15.4)	30b (38.5)	36b (46.2)	0.002
Higher Severity Tremorc
	Dominant Arm	5a (22.7)	14b (63.6)	3a (13.6)	0.01
	Nondominant Arm	2a (9.1)	15b (68.2)	5a (22.7)	0.002

Values represent numbers (row percentages) of cases identified as displaying greater severity in outstretched position, greater severity in wing posture position, and same level of severity in both positions.

^aChi Square test; Bolded values are significant at p <0.05.

^bN=78; Cases receiving a rating of 3 on at least one of the four assessments (outstretched dominant, outstretched non-dominant, wing posture dominant, wing posture non-dominant) from at least one of the two raters were assigned to the higher tremor severity group.

^cN=22; Cases receiving no ratings of 3 were assigned to the lower tremor severity group (n = 78).

Values within a given row with different subscripts are significantly different at p<0.05.

Given a significant overall p value, subscripts identify which specific pairs of percentages within a given row are significantly different. Values that share a common subscript do not differ at p < 0.05. Conversely, values that do not share a common subscript differ at p < 0.05.

Analyses of the forced choice measure conducted separately for low severity and high severity cases revealed some interesting results (Table 4). First, the wing posture position was chosen as displaying greater tremor more often than the outstretched position across all comparisons. However, cases in the low severity group were most often described as displaying the same level in tremor across comparisons, whereas cases in the high severity group were most often described as displaying the greatest level of tremor in the wing posture position.

Tremor Severity as a Function of Type of Posture – Additional Analyses

Finally, to compare the relative ability of each position to reveal tremor, we identified those cases in which tremor was undetectable in one position (i.e., received a rating of 0) but was visible in the other position (i.e., received a rating greater than 0). Rater 1’s data included ten such cases, and Rater 2’s data contained 21 such cases. In the vast majority of these, tremor was observed in the wing posture but not the outstretched position [7/10 (70%), and 19/21 (90.5%), respectively, for Raters 1 and 2], suggesting that the wing posture position may provide an especially sensitive assessment of postural tremor.

Discussion

Our goal was to assess whether ratings of postural tremor would be higher when ET cases held the wing posture position or the outstretched position. Evaluations provided by two raters reveal that tremor observed in the wing posture position is more severe than that observed in the outstretched position. Furthermore, analyses of the forced choice measure revealed that a sizable number of ET cases were categorized as displaying an equivalent severity of tremor in the outstretched and wing posture positions (i.e., they were assigned the same ordinal 0 – 3 rating). However, in these forced choice analyses, among cases in whom severity differed between positions, the wing posture position was more likely to be identified as displaying the more severe tremor. In addition to replicating across two raters, similar results were observed for the dominant and non-dominant arms and for low and high severity tremor cases. In fact, the tendency for raters to evaluate the wing posture position as most severe was most pronounced for high severity cases.

It may be difficult to diagnose ET early in the disease because tremor can be mild or intermittent. As such, the ability to detect tremors of low amplitude is of value. The finding that tremors not detectable in the outstretched position may be identified in the wing posture position (but not vice versa) suggests that the latter may be more sensitive to mild tremor. This further suggests that the evaluation of cases in the wing posture position may be of particular clinical use.

This study brings a clinical perspective to the study of tremor severity in ET. A previous study used accelerometers to measure tremor severity.⁷ Although these measures may be more precise, they require additional technologies.

Our results differ somewhat from one of those published previously.⁷ That study used accelerometry to measure tremor severity across limb positions. Differences in ET across positions were modest, with a weak signal that elbow-flexed positions evinced slightly more tremor than other positions, but standard deviations overlapped, and no differences reached statistical significance. We found that ratings of tremor severity differed significantly across these two positions. Our findings align with other reports from studies using the TETRAS that the wing posture position produces more severe tremor than does the outstretched position.^6,8

Our findings beg the question of what mechanisms underlie the observed positional differences in tremor severity ratings. Perhaps muscle fatigue affected the muscle groups for wing posture tremor more than for outstretched tremor, although the position only had to be held for ten seconds, so this was not likely to have been a major contributor for most cases. Alternatively, position complexity (e.g., number of muscles used, or number of joints bent) may impact tremor severity.

A limitation of this study may be that videotaped neurological examinations limit the accuracy of the severity ratings. Although poor video quality and various filming techniques by the researcher conducting the neurological assessment could both hinder assessment of tremor amplitude,¹² having videotaped evaluations allows for multiple assessments of the same case at that point in time. Second, in assessing postural tremor, the WHIGET scale assigns ratings for tremor severity; tremor severity can be a function of tremor amplitude as well as tremor persistence. For assessing postural tremor, the TETRAS scale assigns ratings that are more purely based on tremor amplitude. Despite these differences, data from our study and from prior studies using the TETRAS largely agree. Third, cases in our study were first examined in the outstretched position and then in the wing posture position. Future studies may wish to randomize the order in order to minimize any temporal effects. Fourth, we assessed each posture once. It is conceivable that repeated assessments several times could have yielded different results. Finally, rating tremor with both arms raised simultaneously can pose a challenge for raters, however, when ratings are assigned based on a videotaped examination, the rater can separately focus on each arm, provided the video is replayed. Rating tremor while both arms are raised simultaneously allows the rater to directly compare the amplitudes of tremor during the same time epoch. The strengths of this study include our use of two independent raters and a standardized protocol and tremor rating scale.

Our findings easily translate into clinical practice. As mentioned previously, the outstretched position may be the more commonly used diagnostic tool for assessing postural tremor in ET. However, the wing posture position was more effective in detecting such tremors. Clinicians and researchers should apply these findings to assess patients more critically by incorporating the wing posture position in neurological assessments.

• We investigate two maneuvers for determining severity of postural tremor in ET.

• Two movement disorders neurologists assessed tremor in 100 ET cases.

• Cases displayed more severe tremor in the wing posture than outstretched position.

• Our findings easily translate into clinical practice.

• Clinicians should incorporate both positions in neurological assessments.