NEUROLOGY 1996;46:747-752
Primary orthostatic tremor (OT), a rapid (14 to 16 Hz), regular lower limb tremor causing unsteadiness upon standing, [1,2] is synchronous between homologous muscles and may be associated with a rapid upper limb postural tremor. [3-5] There has been debate whether OT is a variant of essential tremor (ET) or a separate entity. [5-7] OT and ET share a number of characteristics such as variable age of onset, the absence of a resting component, and EMG patterns that show either alternating or synchronous cocontraction of antagonist muscles. However OT, unlike ET, shows a high degree of synchrony between different muscles and cannot be reset by a peripheral nerve stimulus. [5] OT rarely responds to therapeutic agents (such as propanolol and alcohol) that have proven benefit in ET, but can be dramatically eased by clonazepam. [5] In addition, a negative family history is usual in OT and there is a small female preponderance of cases. These factors all suggest that OT and ET have separate etiologies.
PET with H215 O is a minimally invasive technique that can be used to measure regional cerebral bloodflow (rCBF) repeatedly in subjects while they perform different tasks. Areas of normal and abnormal cerebral activation may be delineated. In previous studies, we have shown that ET is associated with abnormal bilateral overactivity of cerebellar connections whereas voluntary wrist oscillation only activates the ipsilateral cerebellum. [8]
In this study, we examined the abnormal patterns of cerebral activation induced by the postural upper limb tremor of OT and contrasted this with previous findings in other tremulous disorders. The 10.65-cm axial field of view of our camera included the whole brainstem to ascertain the role of inferior structures in the generation of tremor. The cerebral cortex was only partially visualized. PET images of individual subjects were also co-registered with MRI to obtain accurate localization of cerebral activation. [9] All subjects had electrophysiologic analysis of their tremors.
Methods.
Patient Selection.
We studied four patients with OT (mean age 62, range 50 to 75). All patients presented with unsteadiness that occurred within seconds of standing and was relieved by walking or sitting. The symptoms of unsteadiness coincided with the appearance of a 14- to 16-Hz tremor of the trunk and legs. All patients also had a rapid postural upper limb tremor on arm extension. There were no other neurologic abnormalities except that patients were unable to tandem walk slowly whereas their performance improved markedly at a rapid pace. All patients had been medication-free for at least 1 week prior to commencement of the PET study (clinical details are given in Table 1).
Table 1. Primary orthostatic tremor group
Before PET scanning all four OT patients underwent EMG analysis of their tremors. This revealed a highly synchronous 14- to 16-Hertz leg tremor that was present only on standing. All subjects also had upper limb recordings on arm extension while standing that confirmed the presence of a synchronous 14- to 16-Hz tremor Figure 1. Electromyographic findings while supine and maintaining a posture of the outstretched arm revealed a background activity of 12 Hz with occasional bursts of faster (14 to 16 Hz) tremor. There was no evidence of coincidental ET in any of the subjects studied.
Figure 1. Surface EMG recordings from right deltoid, paraspinal, quadriceps, and anterior tibial muscles when standing with arms outstretched. A synchronous 16-Hz tremor is evident in all muscle groups.
All subjects gave informed written consent. The studies were approved by the Hammersmith Hospital Medical Ethics Committee. Permission to administer radioactive material was obtained from the Administration of Radioactive Substances Advisory Committee of the Department of Health of the United Kingdom (ARSAC).
Experimental design.
All subjects had 12 sequential PET measurements of rCBF during a single session. Subjects were scanned supine, during maintained posture of the right arm leading to involuntary tremor (condition A), and at rest (condition B). The two conditions were performed in a sequence designed to counteract time and order effects (ABABBAABBAAB). Electrophysiologic recording was not undertaken during the scanning session. The rCBF measurements were performed six times for each of the two conditions in order to detect statistically significant focal flow differences in single subjects.
Data acquisition.
Regional cerebral blood flow was measured with PET by recording the distribution of cerebral radioactivity following the intravenous bolus injection of H215 O. Any increase in rCBF is associated with an increase in the amount of15 O radioactivity recorded from that region. [10] PET was performed using a CTI 953B PET scanner (CTI Inc, Knoxville, TN). The scanner collects data from 16 rings of crystal detectors with an axial field of view of 10.65 cm. Emission data were corrected for the effects of tissue attenuation by using measured data from a transmission scan acquired with external rotating68 Ge rods, prior to the activations. The corrected data were reconstructed as 31 axial planes with a Hanning (0.5) filter giving a reconstructed resolution of 8.5 times 8.5 times 4.3 mm at full width half maximum (FWHM) of the resulting images. Each image plane was displayed in a 128 times 128 format with a pixel size of 2.00 times 2.00 mm. [11]
Before each rCBF measurement a background scan was performed lasting 30 seconds. Immediately after completion of the background scan a 20-second bolus of H215 O was injected into the subject's left cubital vein. This was followed by a 20-second flush of normal saline. A second 90-second scan was started 30 seconds after the background scan. Five seconds before the start of this scan the subjects commenced the motor task. The experiment was designed so that the entry of radioactive water into the brain coincided with the initiation of the particular activating task. This scanning protocol has been previously described by Silbersweig et al. [12] The subjects continued the activating task until the termination of the second scan. This process was performed 12 times in each subject with 10 minutes between the start of each scan to allow radiation to decay to background levels.
Data analysis.
Calculations and image manipulation were carried out on Sun Sparc2 workstations (Sun Computers Europe Inc, Surrey, UK) using ANALYZE version 6.0 image display software [13] (BRU, Mayo Foundation, Rochester, MN) and PROMATLAB (Mathworks Inc, Natick, MA). Analysis of images was undertaken with statistical parametric mapping using SPM software (MRC Cyclotron Unit, London, UK).
The 31 original tomographic planes were interpolated to 43 planes to produce images with approximately cubic 2-mm voxels. Correction was made for head movement between scans using automated image realignment software. [14] A mean image was then constructed for each subject from the 12 realigned scans and the intercommissural (AC-PC) line identified. All images from all subjects were transformed into standard anatomic space as defined by the atlas of Talairach and Tournoux, [15] resulting in 26 planes at 8-mm intervals parallel to the AC-PC line. [16] The resultant images were then ``smoothed'' with a gaussian filter (FWHM 20 times 20 times 12 mm) to increase signal to noise by compensating for intersubject gyral variability and attenuating high-frequency noise. This process also tends to minimize functional variability.
The data for individual subject MRI to PET coregistration were subjected to the same process except no anatomic transformation to stereotactic space was undertaken. PET images were smoothed with a narrower gaussian filter of 10 times 10 times 6 mm. Less smoothing was necessary in statistical analysis of single-subject data because of the absence of gyral variability between scans.
Statistical analysis.
Group.
In order to separate global from focal cerebral blood flow changes associated with the presence of tremor, scans of individual subjects were subjected to a pixel-by-pixel analysis of covariance (ANCOVA) with global flow as the confounding variable. This is performed as a randomized (split plot) blocked design where subject or block effects are explicitly excluded. [17] This analysis normalizes global CBF to 50 ml/100 g/minute and generates a statistical parametric map (SPM sub [t]) of group mean rCBF for each pixel with associated error variance for each task. Because only rises in rCBF were being considered, one-tailed t tests were then applied with a significance threshold set at p less than 0.05 following correction for multiple nonindependent comparisons. Phantom studies have confirmed that this Bonferroni correction at the p less than 0.05 level gives protection against false-positive results in individual transaxial sections. [18] Pixels exceeding this threshold were displayed on coronal, sagittal, and transverse projection maps of the brain. The stereotactic coordinates of significant sites of activation were mapped to anatomic areas by using the atlas of Talairach and Tournoux. [15] Normalized percentage rCBF increases were calculated for all areas of significant change. One withingroup comparison was then made contrasting involuntary upper limb postural tremor (A) against rest (B).
Within our primary OT group there were two patients (patients 2 and 4) who had unusual features (a family history of tremor and a response to alcohol and primidone). In order to circumvent the potential problem of subject heterogeneity the data for the two subgroups with typical and atypical features were separately analyzed by comparing tremor (A) versus rest (B) in the respective groups. This was performed at a significance level of p less than 0.01 without a correction for multiple nonindependent comparisons.
In addition an SPM comparison was made of resting rCBF between our four OT cases and four age- and sexmatched normal subjects. This was done by assuming a normalized global CBF of 50 ml/100 g/minute in both groups and identifying any differences in regional profiles with a chi-squared test. In a previous PET activation study using quantitative methods to measure CBF Jenkins et al. [8] demonstrated no resting differences in global CBF between ET patients and normal volunteers. [8] We therefore assumed that there would be no global differences between OT and normal subjects. A less conservative statistical threshold was chosen for the current study of p less than 0.01 without a correction for multiple nonindependent comparisons.
Individual.
MRI to PET coregistration was performed with automated image realignment software. [14] A ``mean'' PET image was constructed by averaging the 12 realigned rCBF scans from an individual. Reorientation parameters were calculated, saved, and then used subsequently to coregister the statistical parametric maps of significant rCBF change with the subject's own MRI scan. [9]
MRI scans were obtained with a 1-tesla Picker HPQ Vista system using a radiofrequency (RF) spoiled volume acquisition that was relatively spin-lattice relaxation time (T1) weighted to give good grey-white matter contrast and anatomic resolution. After reconstruction the resultant images were aligned parallel to the AC-PC line and interpolated to give cubic voxels of 1 times 1 times 1 mm. The resulting images were then used to coregister the PET images.
Subsequent ANCOVA analysis of significant focal flow changes was similar to the group procedure except that because our hypothesis regarding brain activations was restricted to brain stem, cerebellum, and thalamus, a less conservative analysis was performed and statistical parametric maps were generated at an omnibus significance level of p less than 0.001 during tremor, corresponding to a Z score of 3.09. This approach has been justified in hypothesis-led experiments. [19] A secondary filter of 8 mm FWHM was applied to the SPM sub [t] to counteract the effect of increased statistical noise in single subjects.
Results.
The data set extended from 40 millimeters below to 32 millimeters above the intercommissural line and visualized medullary to thalamic structures. Cortical structures were incompletely visualized by the low field of view and will not be considered further.
Group analysis (four OT patients).
Involuntary postural tremor of the extended right arm when compared with rest was associated with bilateral cerebellar hemisphere activation. There was also significant activation of the cerebellar vermis, left lentiform nucleus and left thalamus Figure 2. No activation of the inferior olives or medulla was evident. Table 2 shows the extent of areas activated, location of peaks of change, and percentage increase in rCBF.
Figure 2. Orthostatic tremor group. SPM projections of the comparison of postural tremor of the right arm with rest. (c, cerebellar hemisphere; v, cerebellar vermis; t, thalamus/lentiform nucleus; VAC, vertical line through anterior commissure; VPC, vertical line through posterior commissure; arrow, superior extent of the PET data set.
Table 2. Extent of areas activated, location of peaks of change, and percentage increase in rCBF
Subgroup analysis (two typical and two atypical OT patients).
Involuntary postural tremor of the extended right arm when compared with rest was associated with bilateral cerebellar hemisphere activation in both subgroups. There was also significant activation of the cerebellar vermis, left lentiform nucleus, and left thalamus. The extent of areas activated, location of peaks of change, and percentage increase in rCBF were similar in both subgroups. The axial extent of vermal activation was slightly greater in the two atypical patients.
Between-group resting comparison.
A between-group resting comparison for the four OT and four control subjects demonstrated a significant increase in both cerebellar hemispheres and vermis in the absence of tremor in the OT group Table 2.
Individual analysis.
Significant activation was seen in both cerebellar hemispheres in all four OT cases when tremor was compared with rest. This was mainly localized in the cerebellar cortex although some additional activation was also seen in deeper structures. In addition, activation was also seen in the cerebellar vermis, contralateral thalamus, and lentiform nucleus. This pattern of activation was seen in all four subjects. There was no relationship between the extent or magnitude of the activations and the degree of drug responsiveness. No increased blood flow was evident in the medulla or olives Figure 3.
Figure 3. MRI/PET coregistration. Data from a single subject with orthostatic tremor. The MR and SPM sub [t] images have been coregistered and superimposed. Axial sections through the medulla, pons, and thalamus have been depicted and are parallel with the AC-PC plane. The SPM sub [t] images share a common color scale where all pixels with a Z score of greater than 3.09 have been depicted. The image depicts the coregistered SPM sub [t] of involuntary postural tremor of the right arm against rest (the image right is on the reader's right).
Discussion.
In a previous study of essential and writing tremor, [20] we demonstrated abnormally increased cerebellar blood flow bilaterally when tremor was compared with rest or posture without tremor. Our current study also showed abnormal bilateral cerebellar activation in OT subjects when involuntary postural arm tremor is compared with rest. In addition, we have demonstrated activation in the contralateral thalamus and lentiform nucleus and abnormally raised rCBF in bilateral cerebellar hemisphere and vermis in OT at rest, in the absence of tremor, suggesting that the cerebellum is overactive.
Our subjects did not have EMG recordings during the PET study. However, recording was undertaken on a separate occasion under conditions identical to those imposed in the experimental paradigm (supine with arms extended) and demonstrated bursts of 14-to 16-Hz tremor in the arms. We are thus confident that the upper limb postural tremor in our OT cases did not represent coincidental ET, which has a characteristic frequency of 4 to 8 Hz.
In contrast to ET, there are no animal models of OT. However harmaline or ibogaline can generate a relatively rapid tremor of 8 to 12 Hz when injected into intact monkeys. [21] This tremor, like OT, exhibits marked synchronicity in antagonist muscles and between muscles of different limbs. The degree of synchronicity may, therefore, not be a characteristic feature of OT but simply be related to the rapidity of the underlying tremor frequency. This issue remains to be resolved.
Other workers have shown that essential, neuropathic, and parkinsonian rest tremors are associated with abnormal bilateral cerebellar activation. [22-24] Voluntary wrist oscillation leads only to significant ipsilateral cerebellar activation. [8] Our findings in OT provide further evidence that abnormal bilateral activation of cerebellar connections is a common feature of all pathologic tremors. As a consequence, we cannot exclude that OT is a variant of ET (representing a second harmonic). The question arises whether the abnormal cerebellar activation in tremor results from abnormal input (from the olives or peripheral afferent-efferent traffic mismatch leading to oscillation) or a primary disorder of the cerebellum, which is in itself responsible in some way for the generation of tremor. At present we are unable to distinguish these mechanisms. Experimental observations on monkeys have shown that cooling of the dentate nucleus of the cerebellum can induce a postural tremor. [25] Clinical observations have shown that an ipsilateral cerebellar stroke can abolish essential tremor. [26] The cerebellum, therefore, may be the primary site of tremorgenesis. This is supported by cerebellar overactivity in ET patients when they are at rest, without tremor. [8]
In previous PET studies we demonstrated that maintaining a posture of the outstretched arm in normal subjects (with concomitant physiological tremor) causes only ipsilateral cerebellar activation. [8] In contrast, our OT subjects demonstrated abnormal bilateral cerebellar activation. Their postural upper limb tremor cannot strictly be defined as OT because it was present while they were supine and showed variable electrophysiologic features. However, the underlying frequency was 12 Hz with short bursts of 14- to 16-Hz activity typical of OT and there was no evidence of concomitant 4- to 8-Hz ET. These findings are in accord with our hypothesis that all pathologic tremors are associated with abnormal bilateral cerebellar activation.
We also demonstrated abnormal activation in the contralateral thalamus and lentiform nucleus. There are reciprocal connections between the pallidum and thalamus mediated by a number of neurotransmitters such as GABA, acetylcholine, and various neuropeptides, [27] suggesting the possibility of novel pharmacologic approaches to ameliorate OT. Thalamic stimulation is useful in a number of tremulous disorders [28] and our findings suggest that this therapy might have a useful role in OT if it were disabling.
In summary, the upper limb postural tremor of primary orthostatic tremor is associated with abnormal bilateral cerebellar activation similar to other tremors. This suggests that all tremors may be associated with overactivity of a common circuit involving cerebellothalamic connections and explains why thalamotomy is an effective therapy for tremor irrespective of etiology.
Acknowledgments
We are grateful to the radiographers at the MRC Cyclotron Unit (Mrs. A. Williams, Mr. G. Lewington, Mr. A. Blyth) for their assistance with scanning. We are also extremely grateful to Mr. P. Astelman and Dr. J. Rothwell for undertaking detailed electrophysiological analysis of our patients.
Footnotes
-
.AB.-Primary orthostatic tremor (OT), a clinical syndrome in which a rapid (14 to 16 Hz), regular lower limb tremor causes unsteadiness on standing, may be associated with a postural upper limb tremor of similar frequency. We used H sub 2 sup 15 O PET to analyze the abnormal pattern of cerebral activation associated with the postural upper limb tremor in four patients with primary OT. Patients had regional cerebral bloodflow (rCBF) measured during involuntary tremor while maintaining a posture with their outstretched right upper limb and again at rest. Tremor was associated with abnormal bilateral cerebellar and contralateral lentiform and thalamic activation. These findings were evident on group analysis of pooled PET data after transformation into standard stereotactic space and in single subjects when PET images were coregistered with structural MRI of the brain. At rest, cerebellar blood flow was significantly increased bilaterally in OT when compared with age- and sex-matched controls. We have previously demonstrated similar abnormal bilateral cerebellar activation in essential and writing tremors and conclude that abnormal bilateral overactivity of cerebellar connections is a common feature of tremulous disorders.|
- Copyright 1996 by the Advanstar Communication Inc.
REFERENCES
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.