Olfactory dysfunction in essential tremor

Methods.

We enrolled 37 ET cases and 37 control subjects. Cases were identified from patients billed at our center in the last three years. Those with PD, dementia, or other neurologic disorders were excluded. Potential control subjects were selected from the New York, New Jersey, and Connecticut area by random digit dialing and were frequency matched to cases on age, sex, and ethnicity. Cases and control subjects were screened for cognitive impairment using the 10-minute Telephone Interview for Cognitive Status.3 Individuals with cognitive impairment (score <30 of 41) were excluded. Because cigarette smoking causes olfactory dysfunction,4 we excluded current (within 1 year) cigarette smokers.

Data were collected on disease duration and smoking habits. A limitation of this study is that we did not assess second-hand smoke exposure, which could lead to olfactory dysfunction. Odor identification testing was performed with the University of Pennsylvania Smell Identification Test (UPSIT, Sensonics Inc., Haddon Heights, NJ).2,4⇓ The UPSIT was administered by a tester who released 40 standardized microencapsulated stimuli and placed them under the subjects’ nares. The UPSIT is highly reliable2,4⇓ and sensitive to a variety of olfactory deficits. The score ranges from 0 to 40 (normal). Scores may be converted to age- and sex-specific normative percentiles.4

A videotaped neurologic examination was reviewed by a researcher (E.L.) blinded to UPSIT scores, and a total tremor score (0–36)5 was assigned based on a 0 to 3 rating of action tremor during 12 activities. ET diagnostic criteria were moderate amplitude tremor during ≥3 activities or head tremor.5 No subjects had PD on videotape.

Based on UPSIT data on individuals >65 years old,2,4⇓ and assuming alpha =0.05 and power =80%, we needed 37 subjects per group to detect a 10% difference between cases and controls.

χ² And Student’s t-test were used. When analyzing UPSIT scores, nonparametric tests (Mann–Whitney U, Spearman’s r) were used or scores were transformed (log[41 − UPSIT]). For some analyses, subjects were stratified into five olfactory groups (normosmia, mild microsmia, moderate microsmia, severe microsmia, and anosmia) as described elsewhere.4 To compare differences between cases and control subjects in UPSIT scores while adjusting for potential confounders (age, sex, past cigarette smoker [yes/no], cigarette pack years) an analysis of covariance (ANACOVA) was performed. Analyses were performed in SPSS (Version 9.0, Chicago, IL)

Results.

Cases and control subjects were similar in age, sex, ethnicity, and education (table 1). None had acute or chronic rhinitis or sinusitis. Mean tremor duration in cases was 18.9 years; 23 (62.2%) had ET for >10 years.

We examined potential confounding variables among control subjects. The UPSIT score was negatively correlated with age (r = −0.56, p < 0.001) but was similar in men and women (31.4 ± 5.1 vs 32.4 ± 4.1, z = 0.38, p = 0.70), and in those who did and did not have a past history of cigarette smoking (32.5 ± 4.1 vs 31.5 ± 5.0, z = 0.43, p = 0.68). It was similar in ET patients who were vs ET patients who were not currently taking medication for their tremor (28.1 ± 4.4 vs 29.3 ± 6.7, z = 1.27, p = 0.22).

UPSIT scores ranged from 10 to 39 (UPSIT percentile = 5% to 83%, table 2). The mean UPSIT score was lower in cases than in control subjects (29.0 ± 6.1 vs 31.9 ± 4.6, z = 2.25, p = 0.02). The mean UPSIT percentile was 11.5% lower in cases than in control subjects (35.5% ± 20.2% vs 47.0% ± 20.3%, z = 2.49, p = 0.01). Subjects were stratified into five olfactory diagnostic groups, and cases and control subjects differed (table 3, χ² = 13.2, p = 0.01); 18.9% of cases vs 40.5% of control subjects were normosmic (χ² = 4.14, p = 0.04, see table 3), and 27% of cases vs 2.7% of control subjects had severe microsmia or anosmia (χ² = 8.65, p = 0.003, see table 3).

In an ANACOVA that included diagnostic group (case vs control) and several covariates (age, sex, past cigarette smoker [yes/no], cigarette pack years), the transformed UPSIT score differed by diagnostic group (F = 5.68, p = 0.02) and age (F = 17.93, p < 0.001) but not by other covariates.

In cases, there was no correlation between UPSIT score and total tremor score (r = −0.07, p = 0.68) or disease duration (r = 0.11, p = 0.53). Total tremor score in cases was stratified into quartiles, and there was no difference in mean UPSIT score between cases in the lowest vs the highest quartile (25.0 ± 10.0 vs 28.0 ± 5.2, z = 0.52, p = 0.60). Disease duration was stratified into quartiles, and there was no difference in mean UPSIT score between the lowest vs the highest quartile (28.7 ± 6.3 vs 31.0 ± 4.0, z = 1.11, p = 0.31). Ten cases had disease of short duration (<5 years). Their mean UPSIT percentile was 35.7%, and in eight of these, the UPSIT percentile was <50%.

Discussion.

A mild olfactory deficit was present in ET cases compared with matched control subjects. As in PD, the deficit is unrelated to disease duration and severity and seems to be present early in the disease.2

In a previous study,6 lower mean UPSIT scores in ET cases (36.3) did not differ (p = 0.19) from those of control subjects (37.2),6 although with 16 subjects per group, the study power, which was not reported, may not have been adequate. Also, their cases were younger than ours (mean age 50.6 years). Age-related deficits in olfactory dysfunction are well documented.4 A normal age-related deficit in olfaction in our older population of ET patients may have unmasked an underlying disease-related deficit.

The role of the UPSIT in distinguishing ET cases from PD requires further study. Although the mean UPSIT score is lower in PD, there is also overlap between ET cases and similarly aged patients with PD (see table 2). Better differentiation between the two groups could occur in certain age groups or differently in men and women. A larger data set is needed to address these issues.

The basis for the olfactory deficit in ET is not clear. Cerebellar pathology could be responsible. Cerebellar involvement in ET is apparent in fMRI studies.7 Study results also suggest that the cerebellum plays a role in human olfaction. In a fMRI study, odorants induced significant activation of the posterior and the lateral portions of the cerebellar hemispheres.8 Tumors in and around the cerebellum are reported to cause olfactory dysfunction.9 In the animal literature, the staggerer mutant mouse, with a reduction in cerebellar size and abnormal connections among Purkinje cells and parallel and climbing fibers, exhibits olfactory deficits.10 Alternatively, there may be involvement of the olfactory pathways in ET, as has been hypothesized in other neurologic disorders of late life such as PD.2 Regardless of the basis of the olfactory disorder associated with ET, its presence provides a similarity between this disease and other neurodegenerative disorders.