How Normal Is ‘Normal'?Mild Tremor in a Multiethnic Cohort of Normal Subjects

Background  While many "normal" subjects exhibit mild clinically detectable tremor, the extent to which this tremor is present has received little attention.

Objective  To characterize the prevalence and clinical characteristics of mild, clinically detectable tremor in a multiethnic cohort of normal subjects.

Methods  Normal control subjects (n=36) and their relatives (n=67) were enrolled in a community-based case-control study of the familial aggregation of essential tremor. Subjects underwent a tremor interview and videotaped tremor examination. Two neurologists independently rated the severity of tremor during different postures and tasks, and a total tremor score (maximum score, 36) was calculated for each subject. Eight subjects were randomly selected to undergo quantitative computerized tremor analysis.

Results  In 103 normal subjects (36 control subjects and 67 relatives of control subjects) the mean total tremor score was 4.8 (range, 0-12.5). Ninety-nine (96%) of 103 subjects had tremor, as defined by a total tremor score of 0.5 or higher. For most tasks, the mean tremor score was greater in the nondominant than in the dominant arm. The total tremor score correlated highly with age (r=0.28; P=.004). There were no sex or ethnic differences in the mean total tremor scores. On tremor analysis, the amplitude and frequency of the tremor differed from that in a group of similarly studied subjects with essential tremor.

Conclusions  Normal subjects almost uniformly have a clinically detectable tremor that is mild and age dependent. Characterization of this tremor helps to establish standards for normal tremor. These standards are crucial for accurate diagnostic classification in population-based studies of essential tremor.

IN POPULATION surveys (ie, prevalence and incidence studies) and genetic linkage studies of essential tremor (ET), diagnostic misclassification often occurs because of the presence of what has been labeled "normal tremor," "physiological tremor," or "enhanced physiological tremor."^1-3 There are several reasons why the potential for diagnostic misclassification is high. First, there are no biological markers for ET; therefore, the distinction between ET and these normal forms of tremor is based on clinical criteria rather than results of neuroimaging, serological analysis, or pathological findings. Second, the large majority of ET cases ascertained from the population have extremely mild tremor; as many as 99.5% of individuals with ET living in the community have tremor that is so mild that it does not result in a visit to a physician.^4,5 Finally, clinically detectable normal tremor is an entire order of magnitude more prevalent in the general population than ET itself; one estimate of the prevalence of clinically detectable normal tremor was 25%,⁶ compared with 0.4% to 3.9% for ET.^7,8

While it is generally accepted that many "normal" subjects exhibit some degree of clinically detectable yet normal tremor,⁹ the prevalence, clinical characteristics, and severity of this tremor have, with few exceptions,¹⁰ rarely been the subject of systematic scrutiny. One result of this is that the selection of control subjects for studies of ET is now carried out without any standards for the severity of clinically detectable normal tremor.¹ As part of a population-based familial aggregation study of ET,^1,11 we examined a multiethnic cohort of normal control subjects spanning a considerable age range. The purpose of this study was to characterize the prevalence, clinical characteristics, and severity of clinically detectable tremor in these "normal" subjects.

Two thousand one hundred seventeen subjects aged 65 years and older who were residents of Washington Heights-Inwood, northern Manhattan, NY, were enrolled in a longitudinal, community-based study of health issues in the elderly, the Northern Manhattan Aging Project (NMAP).⁷ Subjects underwent a 90-minute medical interview and a standardized medical and neurological examination conducted by a neurologist. Eighty-three subjects with ET were identified.⁷ The following subjects were then enrolled in a second study, the Washington Heights-Inwood Genetic Study of Essential Tremor (WHIGET)^1,10: the 83 subjects with ET, 83 control subjects, and first- and second-degree relatives of the control subjects. The control group was composed of elderly subjects matched by age, sex, and ethnicity to the 83 subjects with ET. All 83 control subjects had undergone a medical interview, and all but 5 also underwent a standardized medical and neurological examination as part of NMAP.

Participation in WHIGET was solicited first by letter and then by telephone. If they participated, subjects underwent a 10- to 30-minute semistructured tremor interview and a 10-minute videotaped tremor examination. The tremor interview and examination were conducted in person by a study physician, either in the subject's home or at the Ambulatory Research Unit at Columbia-Presbyterian Medical Center, New York, NY.^1,11

To date, 53 of the 83 control subjects from NMAP have been selected at random to be evaluated in WHIGET. After undergoing the tremor interview, videotaped examination, and diagnostic protocol outlined below, 2 (3.8%) of these 53 were diagnosed by both of the neurologists (B.F. and E.D.L.) as having probable or definite ET. This figure (3.8%) is similar to the prevalence of ET previously reported in this community (3.6%).⁷ An additional 15 (28.3%) of the 53 were diagnosed by 1 or 2 of the neurologists as having possible ET. These 15 were excluded from the present analyses because our goal was to study subjects who were unambiguously normal according to both of the neurologists. The 36 remaining control subjects were diagnosed as normal by both of the neurologists.

To date, 78 first- and second-degree relatives of the remaining 36 control subjects have been enrolled in WHIGET. After the tremor interview, videotaped examination, and diagnostic protocol outlined below, 2 (2.6%) of these 78 were diagnosed by both of the neurologists as having probable or definite ET. An additional 9 (11.5%) of the 78 were diagnosed by 1 or 2 of the neurologists as having possible ET. These 9 were excluded from the present analyses because our goal was to study subjects who were unambiguously normal according to both of the neurologists. The 67 remaining relatives were diagnosed as normal by both of the neurologists.

Hence, including the 36 normal control subjects and their 67 normal relatives, we analyzed data on 103 normal subjects.

For ethnic group classification, we used the format suggested by the 1990 US Census Bureau,¹² which identified Hispanics as a cultural group, reporting this population as a proportion of the total. For this study, we used the categories black, white, Hispanic, and other (Asian American).

The 84-item, 10- to 30-minute, semistructured, physician-administered tremor interview included 12 questions designed to screen for ET in the general population, as published elsewhere.^1,11 In addition, the interviewer collected demographic information and information on concurrent medical conditions, medication use, distribution and severity of tremor and change in these parameters over time, effects of alcohol, cigarettes, and caffeine, effectiveness of tremor medications, change in tremor with activity or rest, and specific functional impairments resulting from tremor.^1,11

The 26-item, 10-minute tremor examination was designed to elicit tremor during 2 different postures, 5 different tasks, and 2 different positions at rest.^1,11 Tasks included pouring water between 2 cups, drinking water from a cup, using a spoon to drink water, finger-to-nose movements, and drawing spirals.

Each task was first performed with the dominant arm and then performed with the nondominant arm. The examination, conducted by a study physician immediately following the tremor interview, was videotaped using a manually operated videocamera recorder (Sony CCD-TR700, Sony, Park Ridge, NJ). Hi-8mm videotapes were used to increase resolution.

As described elsewhere,^1,11 2 neurologists specializing in movement disorders who were shielded from the individual's status as a case, control subject, or relative randomly and independently reviewed data collected during each subject's tremor interview and reviewed the videotaped tremor examination. This review process required 10 to 20 minutes per subject. The reviewers rated the severity of tremor as observed during different postures and tasks. Ratings were based on an ordinal scale (from 0 to +3), similar to those of Fahn et al.¹³ The ratings were as follows: 0 (no visible tremor), +1 (low amplitude, barely perceivable tremor or intermittent tremor), +2 (tremor of moderate amplitude and usually present; clearly oscillatory), and +3 (large amplitude, violent, jerky tremor resulting in difficulty completing the task due to spilling or inability to hold a pen to paper).^1,11 A total tremor score (maximum score, 36) was calculated for each subject by addition of each of the postural and task-specific scores.

Each reviewer independently assigned a diagnosis of ET (definite, probable, or possible) or normalcy based on their ratings and their review of the data collected during the tremor interview.^1,11 This diagnostic protocol is highly reliable. Two neurologists used this protocol to assign diagnoses to 100 subjects (40 individuals with ET and 60 control subjects) and demonstrated excellent interrater reliability (weighted κ, 0.85).¹¹ Diagnostic criteria^1,11 even for possible ET were stringent. For example, subjects were diagnosed as having possible ET only when a +2 tremor (tremor that is clearly oscillatory, of moderate amplitude, and usually present) was observed during a minimum of 3 separate tasks. The entire process (interview, videotaped examination, data review, and diagnosis) required 30 to 60 minutes per subject.

Eight subjects were selected at random to undergo quantitative computerized tremor analysis in the Motor Neurophysiology Laboratory at Columbia-Presbyterian Medical Center to further characterize the electrophysiological characteristics of a representative sample of subjects.

The tremor analysis involved the use of ultralight piezoresistive miniature accelerometers (±25g and 500 mg) with linear sensitivities of approximately 4.5 mV/g in the physiologic range that were attached over the dorsum of each hand at the distal end of the middle metacarpal bone. Silver and/or silver chloride electromyographic surface electrodes were used to record the activity of the flexor carpi radialis and extensor carpi radialis muscles along with the accelerometry.

Accelerometric and electromyographic signals were digitized at 500 Hz using a 15-microsecond 16-bit analog-to-digital system and stored in eight 4-second trials during 3 conditions: with the arms at rest, with arms extended, and during finger-to-nose movements. Tremor was sampled during a 1-hour period to record variation over time. Rest measurements were done with the subject's arms flexed 90° and kept stationary at the elbow to prevent transmitted upper arm movement into the forearm and hand. Posture and action measurements were performed with the arms extended and the patient freely able to touch finger to nose as previously described.^14,15

Tremor amplitudes were derived offline by double integration of wrist accelerometric data after filtering out low-frequency drift (<2 Hz) and averaging. Tremor frequencies were calculated using a fast Fourier transform algorithm to generate autocorrelation spectra. Electromyograms were full-wave rectified, integrated, and processed with the accelerometric data as described previously.^14,15

The tremor score for each subject was the mean of the 2 raters' tremor scores. A 2-tailed Student t test or its standard normal approximation (z) and analysis of variance were used in the analysis of continuous variables, and r, the correlation coefficient, was used to assess correlation between 2 continuous variables.^16,17

Including the 36 normal control subjects and their 67 normal relatives, there were 103 normal subjects (Table 1). The total tremor score for these normal subjects ranged from 0 to 12.5 (mean score, 4.8; median score, 4.5) (Figure 1). Ninety-nine (96%) of 103 subjects had tremor, as defined by a total tremor score of 0.5 or higher (Figure 1). One individual answered yes to 2 of the 12 screening questions for ET; however, her total tremor score was only 3.5. All the remaining 102 subjects answered no to each of the 12 screening questions for ET. The mean tremor score for each posture and task was almost always greater in the nondominant arm than the dominant arm, with 1 exception—pouring water (Table 2). Twenty subjects were rated by both neurologists as having a tremor of +2 severity on at least 1 task; 9 additional subjects were rated as having a tremor of +2 severity on 2 tasks. A tremor rating of +2 in severity most commonly occurred when subjects used a spoon to drink water or during finger-to-nose movements (accounting for 28 [75.7] of the 37 tremor scores of +2 severity). Of the 37 tremors rated as +2 severity, 26 (70.3%) occurred in the nondominant arm.

Individual subjects had varying degrees of tremor, depending on the task. Several tasks (eg, using a spoon with the nondominant arm and finger-to-nose movements with the nondominant arm) were more commonly tremulous than others (eg, pouring water between cups) (Table 2).

The total tremor score correlated highly with age (r=0.28; t=2.95; P=.004). The mean total tremor score did not differ between men (4.7) and women (4.9) (t=0.20; P=.84). The mean total tremor score did not differ among ethnic groups: Hispanic (5.1), black (4.4), and white (4.6) (F=0.51; P=.60). There were no significant age differences between ethnic groups or between men and women.

Nine subjects had a history of hyperthyroidism or were currently taking levothyroxine sodium. These 9 had a mean total tremor score of 5.5 (range, 1.5-11.0; median, 5.5), which was not different from the mean total tremor score (4.8) of the remaining 94 without hyperthyroidism or use of levothyroxine (t=0.68; P=.50). Nine subjects were taking a tremorogenic medication (eg, valproate sodium, prednisone, or an oral hypoglycemic agent) (mean total tremor score, 3.0), 17 were taking a tremor-suppressing medication (eg, calcium channel blocking agent or β-adrenergic blocking agent) (mean total tremor score, 5.1), and 4 were taking both (mean total tremor score, 3.9) (F=0.53; P=.60).

The tremor interview included the questions "Do you feel depressed?" and "Are you anxious?" Subjects who reported being both anxious and depressed (n=16) during the interview had a similar mean total tremor score (5.7) compared with those who reported either anxiety or depression (4.9) and those who reported neither (4.9) (F=0.34; P=.71).

Eight subjects were randomly selected to undergo quantitative computerized tremor analysis to define the electrophysiological features of their tremor and to further exclude the possibility that they had ET. One of these subjects was excluded because she also had chorea gravidarum. The mean amplitude and frequency of postural arm tremor (including both right and left arms) in 2 subjects younger than 50 years were 0.38 mm and 10.3 Hz, respectively. The mean amplitude and frequency of postural arm tremor (including both right and left arms) in 5 subjects older than 50 years were 0.45 mm and 10.2 Hz, respectively.

In studies of ET, diagnostic accuracy is seriously confounded by the presence of what has been labeled normal tremor, physiological tremor, or enhanced physiological tremor. The etiology of these nonpathological tremors has been studied extensively, and different mechanisms have been proposed, including a cortical origin, origin in a self-regulating mechanism of a servoloop in the reflex arc, a result of neuromuscular activity, or a result of cardioballistic impulses.^2,3,18-20 While there is a sizable literature on these tremors,^6,21-29 it focuses on the frequency spectrum of the tremor as studied electrophysiologically rather than on the extent to which the tremor is detectable by the clinician. One study reported that 56 (15.7%) of 356 individuals aged 65 years and older exhibited a low-amplitude high-frequency tremor present only with posture.¹⁰ While it is common knowledge that most normal subjects exhibit some clinically detectable tremor,⁹ the prevalence and severity of this tremor have received little attention.

We examined a cohort of normal subjects spanning a considerable age range as part of a multiethnic, population-based, familial aggregation study of ET,^1,11 and characterized the extent of tremor that was clinically detectable in these subjects. We have defined this tremor as normal in an epidemiological sense (ie, in conformity with the average pattern of a large group). This normal tremor probably encompasses the physiological terms physiological tremor and enhanced physiological tremor. Ninety-nine (96%) of 103 subjects had clinically detectable tremor, as defined by a total tremor score of 0.5 or higher. In general, this tremor was a mild +1 tremor (either low amplitude or intermittent tremor) that was present on several tasks, although 29 subjects (28%) had a +2 tremor (tremor that is clearly oscillatory, of moderate amplitude, and usually present) on 1 and sometimes 2 tasks.

There are conflicting results in the literature regarding the association between the amplitude of normal tremor and advancing age. Marshall,²⁶ in a study of 347 control subjects, and Van Buskirk and Fink,²⁸ in a study of 161 control subjects, demonstrated in adults an inverse correlation between tremor frequency and age. Because frequency is inversely related to amplitude,²⁷ this implies that there is a positive correlation between tremor amplitude and age. However, other studies have found no association between tremor amplitude and age.^6,23 We found a highly significant correlation between clinically detectable tremor severity (ie, total tremor score) and subject age. It is possible that both ET³⁰ and normal tremor increase in amplitude with age.

There were no associations between tremor severity and sex or ethnicity. In contrast, ET may be more prevalent among whites than blacks,³¹ with Hispanics intermediate.⁷

In this study, we have shown that tremor can be clinically detected in almost all individuals at every age, even when subjects are unaware of having a tremor. We have presented clinical and electrophysiological data that allow a comparison between this normal tremor and ET in a cohort of subjects who underwent a rigorous research protocol. The nature of this normal tremor and its relation to ET are not clear. Physiological forms of tremor vary in severity and probably form a continuum with ET, making some distinctions arbitrary. In some cases, formes frustes of ET or early cases of ET may appear as normal physiological tremor. Distinguishing between mild ET and normal tremor will require deeper insight into the mechanisms of tremor, as well as long-term follow-up data from individuals with normal tremor. Our criteria for ET were stringent and therefore it is possible that some of our normal subjects may have had ET. However, several features of our normal control group deserve highlighting. First, these subjects uniformly exhibited mild tremor, with a mean total tremor score of only 4.8. This was significantly different from the mean total tremor score (22.8) observed in the 22 subjects enrolled in WHIGET to date who have been diagnosed as having probable or definite ET (Figure 1) (t=21.43; P<.001). Second, with only 1 exception, none of the normal subjects reported a tremor during an extensive 12-item screen for tremor, compared with 18 (82%) of the 22 individuals with ET enrolled in WHIGET to date who have been diagnosed as having probable or definite ET. This is because the majority of our normal subjects exhibited mild, low-amplitude tremors that were more apparent in their nondominant arms. Third, a random sample of those with normal tremor was studied electrophysiologically, and their mean tremor amplitudes were lower and mean tremor frequencies higher than in a group of 20 ET cases seen in the same laboratory.³² In a previously published cohort of 20 patients with ET who underwent quantitative computerized tremor analysis in the same laboratory,³² the mean amplitude and frequency of arm extension tremor in those younger than 50 years was 4.1 mm and 7.9 Hz, respectively, compared with 0.38 mm and 10.3 Hz, respectively, in the 2 normal subjects who were younger than 50 years in the present series. Similarly, the mean amplitude and frequency of arm extension tremor in those older than 50 years was 7.0 mm and 5.8 Hz, respectively,³² compared with 0.45 mm and 10.2 Hz in the 5 normal subjects older than 50 years in the present series. Hence, the mean amplitude of our normal subjects' tremors was 10 to 15 times less than the mean amplitude of the tremors in those with ET. Finally, if all our subjects with tremor had ET, this would imply that the population prevalence of ET was 96%. All 19 of the published prevalence estimates for ET are considerably lower than this, with most estimates in the range of 0.4% to 3.9%.^7,8

One limitation of this study is that it was difficult in several cases to distinguish normal tremor from mild ET, and a number of subjects were diagnosed by 1 or 2 of the neurologists as having possible ET. However, the goal of this study was not to distinguish all cases of ET from normal tremor, but rather to document mild tremor in subjects who did not fulfill our criteria for ET. One could also argue that tremor commonly increases under conditions of stress or anxiety,²⁰ and that our normal subjects were anxious and therefore exhibited more tremor. This is certainly a possibility; however, the majority of subjects were examined in their homes rather than in the hospital, and the examination was performed after a 10- to 30-minute interview rather than immediately on initiation of the evaluation. In addition, those subjects who reported being anxious during the interview had tremor scores that were not different from those who reported not being anxious.

In summary, normal control subjects almost uniformly (96%) have a clinically detectable tremor that is mild; 28% have a clearly oscillatory tremor of moderate amplitude that is usually present during maintenance of a posture or performance of one task. The tremor is more severe in the nondominant arm, and the severity of tremor is associated with advancing age. The etiology of this tremor is uncertain, but its high prevalence (96%) in the population suggests that it is normal (ie, in conformity with the average pattern of a large group) rather than pathological. Characterization of this tremor will help to further establish standards for normal tremor. These standards are crucial for accurate diagnostic classification in population-based studies of ET.

Accepted for publication July 7, 1997.

This study was supported by federal grant NIH NS01863 from the National Institutes of Health, Bethesda, Md, and the Paul Beeson Physician Faculty Scholars in Aging Research Award, presented by the American Federation for Aging Research, New York and the Alliance for Aging Research, Washington, DC (Dr Louis).

Reprints: Elan D. Louis, MD, MS, Unit #198, Neurological Institute, 710 W 168th St, New York, NY 10032.