Essential tremor cases were recruited through the essential tremor centralized brain repository (ETCBR), which serves as a centralized repository for the prospective collection and study of brains of ET cases throughout the United States (10). ET cases were recruited as future brain donors through advertisements on patient advocacy group websites and the ETCBR website (www.essentialtremor.us). Each recruited case underwent an in-person baseline evaluation. This was followed by regular follow-up telephone assessments, the purpose of which was to update clinical information. Enrollment of the wave of cases reported here began in March of 2009 and ended in April of 2011. Follow-up has continued to present.

Once enrolled, ET cases were visited in their homes by a trained research assistant, who performed a detailed baseline evaluation. Each case signed an informed consent form approved by the Internal Review Board of Columbia University. The research assistant administered a series of semistructured clinical questionnaires eliciting demographic and medical information. Medical comorbidity was evaluated using the Cumulative Illness Rating Scale (CIRS, range 0–42) (11). To briefly assess cognition, the Folstein Mini-Mental State Examination (MMSE, range 0–30) was administered (12). Additionally, each case drew four standardized, hand-drawn Archimedes spirals (two right, two left), each on an 8.5″ × 11″ sheet of paper.

The research assistant obtained a standardized, videotaped neurological examination, including a detailed assessment of tremor (13). The videotaped examination included assessments of postural tremor (two positions), kinetic tremor (five activities with each arm), and intention tremor of the arms, as well as head tremor. Tremor was rated by a senior movement disorder neurologist (Elan D. Louis) using a reliable (14) and valid (15) clinical rating scale, which included ratings from 0 to 3 for each of 12 items. A total tremor score (range 0–36) (6), which is a measure of action tremor, was calculated based on these ratings. The videotaped examination also included the motor portion of the Unified Parkinson’s Disease Rating Scale (16).

Essential tremor diagnoses were carefully assigned and then confirmed using the following methods. First, the vast majority of cases were diagnosed clinically with ET by their treating physician (either by their primary health physicians or neurologists); the remaining few were self-diagnosed cases and included individuals with a strong family history of ET. Second, as noted above, cases were asked to complete a series of semistructured clinical questionnaires (demographic data, general medical data including medications, and tremor-specific data), which included data on age of onset and family history information. As noted above, Archimedes spirals were also drawn. These data were supplemented with additional clinical information (from clinical records, treating physicians, and family members). ET diagnoses were then confirmed by a senior neurologist specializing in movement disorders (Elan D. Louis) who used the following criteria: (1) moderate or greater amplitude arm tremor (rating ≥2) in at least one of the submitted Archimedes spirals; (2) no history of Parkinson’s disease or dystonia; and (3) no other etiology for tremor (e.g., medications, hyperthyroidism). Third, as noted above, ET cases underwent a standardized, videotaped neurological examination. Each videotape was reviewed (Elan D. Louis), and based on the questionnaire and videotape data, the diagnosis of ET was re-examined in each case using published diagnostic criteria (moderate or greater amplitude kinetic tremor (tremor rating ≥2) during three or more videotaped activities or a head tremor in the absence of Parkinson’s disease or other known causes) (14, 15, 17).

Cases received regular follow-up evaluations by telephone. The protocol was approved by the Internal Review Boards of Columbia and Yale Universities. During the initial 3 years of the study, these evaluations were performed at 6-month intervals. After that point, the interval was lengthened to 9 months due to concerns about participant fatigue.

During each follow-up evaluation, the research assistant assessed progression of symptoms. Thus, cases answered the question, “has your ET worsened since our last call?” If cases answered “Yes” to this question, the research assistant asked them, “how has your ET worsened?” Cases were also asked about changes in medications and ethanol consumption, and they submitted four new standardized Archimedes spirals (13). Finally, they were asked a series of screening questions for Parkinson’s disease and dystonia. If any screening question was positive for Parkinson’s disease or dystonia or if a spiral showed signs of micrographia, the research assistant revisited the case at home to obtain a follow-up videotaped neurological examination.

Medications were recorded at baseline and each follow-up telephone interview. In order to consider the potential effects of medication changes on the reported change in tremor symptoms, a “medication change score” was created. The medication list of each case was reviewed to identify tremorogenic and tremor-reducing medications. A medication was deemed tremorogenic based on previously published reviews (18) and included lithium, β-adrenergic agonists, thyroxine, and other agents. Similarly, a medication was labeled as tremor reducing based on published data (19). Each tremorogenic medication was assigned a value of 1; also, each tremor-reducing medication, a value of −1. A baseline score was calculated as the sum of all these values at baseline, and a follow-up score was calculated as the sum of all these values at the most recent follow-up evaluation. The final “medication change score” was calculated as the difference between the baseline medication score and the most recent follow-up medication score; a negative value indicated greater use of tremor-reducing medication at follow-up.

Similarly, to consider potential effects of changes in daily ethanol intake on reported change in tremor symptoms, a “change in ethanol intake score” was created. This was based on the average number of drinks per week that each case reported consuming at each evaluation. The change in ethanol intake score was calculated as the change in number of drinks per week reported between baseline and the most recent follow-up.

Finally, to obtain a more objective measure of change in tremor severity with time, two trained research assistants (Jesús Gutierrez and Jemin Park) rated all of the baseline and most recent follow-up spirals using the Bain and Findley 10-point scale (0 = no detectable tremor to 9 = severe tremor) (20). This rating scale has been validated in previous studies assessing tremor severity (21–23). The two research assistants were trained by independently rating 50 Archimedes spirals, and their agreement with those of the senior movement disorders neurologist was substantial (24) [weighted kappa (κ) = 0.63 and 0.64]. To assess the change of spiral ratings over time, a “change in total spiral score” was calculated. Each case submitted two right hand spirals and two left hand spirals. The 0–9 ratings of each spiral were averaged. The change in total spiral score was calculated as the difference between the baseline and the most recent follow-up scores.

Of the 177 cases who had a baseline evaluation, 13 (7.3%) were excluded due to lack of follow-up data (5 deaths, 1 moved outside of the United States, 4 withdrew due to loss of interest, 1 co-diagnosed with dystonia, and 2 miscellaneous). The final sample for analysis (164 ET cases) was similar to the 13 excluded cases in terms of baseline total tremor score (23.8 ± 6.2 vs. 22.6 ± 9.4, p = 0.56) and age at onset of tremor (42.6 ± −23.0 vs. 38.4 ± 20.7 years, p = 0.60). They did not differ by education (15.0 ± 3.2 vs. 13.4 ± 3.7 years, p = 0.18) or gender [62 (37.8%) women vs. 7 (53.8%) women, p = 0.25]. The two groups differed by age (83.3 ± 5.6 vs. 88.7 ± 6.3 years, p < 0.01).

Statistical analyses were performed using SPSS (version 21.0; Chicago, IL, USA). The main variable of interest, self-reported worsening was not normally distributed (Kolmogorov–Smirnov test = 0.13, p < 0.01). Therefore, to explore the demographic and clinical correlates of the cases’ self-reported worsening, we used Spearman’s correlation coefficients and the Mann–Whitney test. Subsequently, we distributed cases into quartiles based on the percentage of time that they reported worsening. We compared quartiles by demographic and clinical variables using chi-square tests and one-way analysis of variance (ANOVA) when variables were normally distributed. For measures that were not normally distributed, non-parametric approaches were used (e.g., Kruskal–Wallis test). The Jonckheere–Terpstra test was used when we tested hypotheses that included a priori ordering.

When we explored the correlation between percentage of times cases reported worsening and change in total spiral scores, we also stratified our analysis by whether or not cases were taking tremor-reducing medications.

The mean baseline age of ET cases was 83.3 years, and the mean age at tremor onset was 42.6 years (Table 1). On average, ET cases completed 6.8 follow-up assessments (median = 7, range = 1–10) spanning an average time interval of 3.7 years (median = 4.3, range = 0.5–5.25 years).

Of the 164 cases, 44 (26.8%) reported worsening of symptoms at every follow-up interval, 104 (63.4%) reported worsening during one-half or more of the follow-up intervals, and 145 (88.4%) reported worsening at one or more follow-up intervals. Only 19 (11.6%) never reported worsening. Data are shown (Figure 1). We obtained similar results when we restricted our analysis to 32 cases who were never on any medications for the treatment of tremor.

When cases reported worsening of symptoms, they were asked “how has your ET worsened?” Their open-ended responses were collapsed into the following basic categories (Table 2): I am shaking more (34.0%), certain activities are harder to accomplish (33.1%), my medicine is not working as well (6.0%), balance issues (5.9%), my tremor is harder to control (3.3%), and other/miscellaneous (17.7%).

To assess the clinical correlates of self-reported worsening, we performed several analyses. First, we assessed the correlation between percentage of times that each case reported worsening and a range of demographic and clinical variables (Table 3). There was no association between worsening of symptoms and baseline age or gender (Table 3). There was a marginally significant negative correlation between years of education and reported worsening (Spearman’s rho = −0.14, p = 0.08). Similarly, medication change score and reported worsening were marginally correlated (Spearman’s rho = −0.12, p = 0.12). There was also a marginally significant positive correlation between CIRS scores and reported worsening (Spearman’s rho = 0.12, p = 0.12) and MMSE scores and reported worsening (Spearman’s rho = 0.13, p = 0.10). There was no correlation between reported worsening and any of the other clinical features (e.g., tremor duration, age at tremor onset, and total tremor score, Table 3). Finally, we divided our cases into two groups: early tremor onset (defined by tremor onset at the age of ≤40 years) and late tremor onset (tremor onset at the age of >40 years) (25). There were no differences between these two groups when we compared the percentage of times that each case reported worsening (64.3 ± 31.4 vs. 60.5 ± 36.3, p = 0.67).

Second, ET cases were divided into quartiles based on the percentage of times they had indicated worsening of symptoms during follow-ups (Table 4). The analysis showed that the quartiles were similar in terms of basic demographic variables (baseline age, gender, education, family history, Table 4). There were no significant differences amongst the quartiles when we compared them by tremor duration, age at tremor onset, and total tremor score (Table 4). The quartiles did not differ by CIRS scores. The change in ethanol intake score differed marginally across quartiles but not in an ordinal manner. There was a marginally significant decrease in medication change score with increasing quartile; in other words, greater self-reported worsening was associated with greater use of tremor-reducing medication from baseline to follow-up. There was no progression of MMSE scores by quartile (Table 4).

We also assessed the correlation between subjective worsening and an objective measure of worsening (i.e., change in spiral scores). As described above, two of the authors rated all of the spirals obtained at baseline and at the most recent follow-up evaluations using the Bain and Findley 10-point scale (20) and calculated a change in total spiral score for each case. We then explored the correlation between the change in total spiral score and the percentage of times our cases reported worsening. There was no correlation (Spearman’s r = 0.08, p = 0.35).

In order to consider the potential impact of ET treatments on these analyses, we performed sensitivity analyses. First, we restricted the analysis to those cases (N = 32) who had never taken tremor-reducing medications, and this produced similar results (Spearman’s r = 0.19, p = 0.37, i.e., no correlation between percentage of times our cases reported worsening and change in total spiral score). We repeated this analysis with cases (N = 116) who were always on at least one tremor-reducing medication. Again, there was no correlation between self-reported worsening and change in total spiral score (Spearman’s r = 0.07, p = 0.48).