Brain/awq009 PDF
Brain/awq009 PDF
Brain/awq009 PDF
BRAIN
A JOURNAL OF NEUROLOGY
Limb-shaking is a specific clinical feature of transient ischaemic attacks that has been associated with a high-grade stenosis or
occlusion of the internal carotid artery. The aim of this study was to describe the clinical characteristics of limb-shaking in
patients with internal carotid artery occlusion and to investigate whether patients with limb-shaking have a worse haemody-
namic state of the brain than patients with internal carotid artery occlusion without limb-shaking. We included 34 patients
(mean age 62 7 years, 82% male) with limb-shaking associated with internal carotid artery occlusion and 68 sex- and
age-matched controls with cerebral transient ischaemic attack or minor disabling ischaemic stroke associated with internal
carotid artery occlusion, but without limb-shaking. We investigated clinical characteristics, collateral pathways on contrast
angiograms and carbon dioxide-reactivity measured by transcranial Doppler. The results showed that limb-shaking usually
lasted less than 5 min and was often accompanied by paresis of the involved limb. Compared with controls, patients with
limb-shaking more frequently had symptoms precipitated by rising or exercise (odds ratio 14.2, 95% confidence interval
4.2–47.9), more frequently had recurrent ischaemic deficits after documented internal carotid artery occlusion (but before
inclusion in the study) (odds ratio 8.2, 95% confidence interval 2.3–29.3), more often had leptomeningeal collaterals (odds
ratio 6.8, 95% confidence interval 2.0–22.7), and tended to have a lower carbon dioxide-reactivity (mean 5% 16 versus
12% 17; odds ratio 0.97 per 1% increase in carbon dioxide-reactivity, 95% confidence interval 0.94–1.00). In conclusion,
limb-shaking transient ischaemic attacks in patients with internal carotid artery occlusion can be recognized by their short
duration, are often accompanied by paresis and precipitated by rising or exercise and are indicative of an impaired haemody-
namic state of the brain.
Received October 28, 2009. Revised December 15, 2009. Accepted December 21, 2009. Advance Access publication February 15, 2010
ß The Author (2010). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.
For Permissions, please email: journals.permissions@oxfordjournals.org
916 | Brain 2010: 133; 915–922 S. Persoon et al.
The protocol for the current study was prepared after collection of
Introduction all data, but before the analysis was performed. For each patient with
limb-shaking we randomly selected two controls, matched for sex and
Several case reports have described limb-shaking as a rare clinical
age, who had presented during the same period with cerebral TIA
feature of transient ischaemic attacks (TIAs) (Firlik et al., 1996;
(lasting524 h) or moderately disabling ischaemic stroke (modified
Leira et al., 1997; Niehaus et al., 1998; Zaidat et al., 1999; Rankin scale 4 3) associated with ICA occlusion, but who had not
Klempen et al., 2002; Schulz and Rothwell, 2002; Cheshire and reported limb-shaking. The ICA occlusion was demonstrated by the
Meschia, 2006; Kiechl et al., 2007). Limb-shaking has been char- absence of filling of the ICA by contrast angiography, or in one patient
acterized by brief, jerky, coarse, involuntary movements involving by absence of flow in the ICA on magnetic resonance angiography.
an arm or leg (Fisher, 1962; Baquis et al., 1985) and has been The degree of an additional stenosis in the contralateral ICA, ipsilateral
associated with high-grade stenosis or occlusion of the internal external carotid artery or vertebral artery was measured according to
carotid artery (ICA). Small observational studies have shown the North American Symptomatic Carotid Endarterectomy Trial criteria
(Fox, 1993). Patients with an ICA occlusion caused by a dissection or a
impaired cerebral blood flow or cerebrovascular reserve capacity
radiation-vasculopathy were not included. In patients and controls, we
in patients with limb-shaking and ICA stenosis or occlusion in
investigated the presence of vascular risk factors as listed in Table 2.
comparison with normal controls (Yanagihara et al., 1985;
The mean arterial blood pressure was calculated by two times the
with TIAs without limb-shaking (excluding control patients with limb-shaking, 28 (82%) of the 34 patients noticed a transient
ischaemic stroke). We used logistic regression analysis to study the paresis of their arm or leg. In 14 (41%) patients, limb-shaking
effect of the time interval between the patient’s last ischaemic symp- occurred subsequent to precipitating factors such as rising, exer-
toms and the CO2-reactivity measurement on the association between cise or coughing (Table 1). Table 2 shows the characteristics of the
limb-shaking TIAs and CO2-reactivity and expressed this adjusted
patients with limb-shaking and of controls. The presence of vas-
association as OR per 1% increase in CO2-reactivity. Finally, we
cular risk factors was similar in patients with limb-shaking and
assessed the relationship between CO2-reactivity and leptomeningeal
collaterals by a multivariable regression model. The study was
controls, except for a history of hypertension that we found
approved by the institutional review board of the University Medical more often in patients with, than in those without, limb-shaking
Centre Utrecht. TIAs (OR 4.3, 95% CI 1.5–12.5). All patients with limb-shaking
also reported symptoms other than limb-shaking; 27 (79%)
patients had additional TIAs without limb-shaking (reported symp-
Results toms, isolated or combined, were paresis of a limb in 24, sensory
symptoms in 12, dysphasia in 6 and dysarthria in 2 patients), and
Of the 313 patients with symptomatic ICA occlusion, 34 (11%) seven (21%) patients had additional permanent deficit caused by
reported limb-shaking. The characteristics of limb-shaking are a minor ischaemic stroke characterized by isolated or combined
shown in Table 1. The duration of limb-shaking was shorter symptoms of paresis of a limb in six, sensory symptoms in one,
than 5 min in the majority of patients. Most patients reported and dysphasia in four patients. In the control group, 22 (32%)
multiple episodes of limb-shaking. The arm was more frequently patients had presented with cerebral TIA and 46 (68%) patients
involved than the leg. In almost one-third of patients the arm and with ischaemic stroke. Compared with controls, patients with
leg shook simultaneously. Most patients demonstrated shaking of limb-shaking more frequently presented with TIAs than with
their whole limb and not just the hand or foot. During or following ischaemic stroke (OR 8.1, 95% CI 3.0–21.4), and more
918 | Brain 2010: 133; 915–922 S. Persoon et al.
Table 1 Clinical characteristics of limb-shaking events 520% ipsilateral to the ICA occlusion in 24 (83%) patients with
among 34 patients with occlusion of the ICA limb-shaking and in 43 (68%) patients without limb-shaking (OR
2.2, 95% CI 0.7–6.7). On average, CO2-reactivity in patients with
Number
of cases limb-shaking (mean 5% 16) tended to be lower than in those
without limb-shaking (mean 12% 17; OR 0.97 per 1% increase
Duration
in CO2-reactivity, 95% CI 0.94–1.00). The median time interval
51 min 13
1–5 min 15
between the most recent symptom and the CO2-reactivity mea-
45 min 3 surement was 18 (range 0–140) days in patients with limb-shaking
Unknown 3 and 57 (range 0–206) days in patients without limb-shaking. After
Frequency of limb-shaking TIAs adjustment of the OR for this time interval, the finding of a lower
1 episode 5 CO2-reactivity in patients with limb-shaking TIAs lost significance
2–5 episodes in a month 18 in comparison with controls (adjusted OR 0.98 per 1% increase in
45 episodes in a month 11 CO2-reactivity, 95% CI 0.95–1.01).
Side of limb-shaking In the subgroup analysis of patients with limb-shaking TIAs in
Right 21
a Precipitating factors were present in 14 patients. Four patients had two and one
patient had three precipitating factors.
Discussion
This study shows that limb-shaking in patients with ICA occlusion
usually lasts less than 5 min, is often accompanied by paresis of
often had additional retinal ischaemic symptoms (OR 3.6, 95% the involved limb and is often, but not necessarily precipitated by
CI 1.2–10.6). Precipitating factors that may compromise cerebral activities that may compromise cerebral perfusion such as rising,
perfusion, such as rising or exercise, were more often—but not exercise or coughing. In comparison with patients with ICA occlu-
always—present in patients with limb-shaking than in controls (OR sion without limb-shaking, patients with limb-shaking are about
14.2, 95% CI 4.2–47.9). Patients with limb-shaking TIAs more seven times more often dependent on leptomeningeal collaterals
frequently had recurrent ischaemic deficits after documented ICA and tended to have a lower CO2-reactivity. Compared with con-
occlusion (but before inclusion in the study) than control patients trols with TIAs without limb-shaking and no ischaemic stroke, the
without limb-shaking (OR 8.2, 95% CI 2.3–29.3). Patients with patients with limb-shaking TIAs had a significantly lower mean
limb-shaking less often had an infarct on their MRI or CT com- CO2-reactivity. The time period between the last symptoms and
pared to controls (OR 0.2, 95% CI 0.1–0.6) and more often a the CO2-reactivity measurement was shorter in patients with than
stenosis or occlusion in one of the vertebral arteries (OR 4.0, 95% in patients without limb-shaking, probably because patients with
CI 1.5–10.4). Patients with limb-shaking were seven times more limb-shaking TIAs more frequently had recurrent ischaemic deficits
often dependent on leptomeningeal collaterals than controls (OR after documentation of the ICA occlusion, which contributed to
6.8, 95% CI 2.0–22.7, Table 3). We found a CO2-reactivity their relatively low CO2-reactivity.
Limb-shaking TIAs Brain 2010: 133; 915–922 | 919
Table 2 Characteristics of patients with symptomatic ICA occlusion with (n = 34) and without (n = 68) limb-shaking,
matched for age and sex
a Defined as a blood pressure 4160/95 mmHg or the current use of anti-hypertensive medication.
b Comparison patients with and without limb-shaking, P50.05.
c Defined as patients with either a history of hyperlipidaemia, patients on drugs because of hyperlipidaemia or patients with levels of cholesterol, triglycerides, or high
density lipoprotein cholesterol beyond the normal ranges.
d In 14 patients the limb-shaking was precipitated by activities that may compromise cerebral perfusion, in two patients only additional TIAs without limb-shaking were
precipitated by rising. In the control-group symptoms were precipitated by rising in two patients and by exercise in two patients.
e 94 patients had an MRI scan of their brain, seven patients a CT scan and in one patient a recent CT or MRI scan could not be performed.
ECA = external carotid artery.
Compared with previous small case series of 5–12 patients hyperkinetic movements such as hemidystonia and hemichorea-
(Baquis et al., 1985; Yanagihara et al., 1985; Baumgartner and hemiballism have also been described in relation to TIA or stroke
Baumgartner, 1998), we were able to identify a relatively large (Ghika-Schmid et al., 1997; Shimizu et al., 2001; Kim, 2001; Salah
group of patients with limb-shaking. We confirmed that Uddin, 2004), but they are exceedingly rare with a prevalence of
limb-shaking TIAs occur in about 10% of patients with occlusion 1% in acute stroke (Ghika-Schmid et al., 1997). In addition, those
of the ICA (Bogousslavsky and Regli, 1986). The underlying mech- hyperkinetic movement disorders seem to be related to an ischae-
anism of limb-shaking is unclear, but most studies suggest that the mic lesion in the basal ganglia or thalamic nuclei in the majority of
shaking movements are caused by transient focal cerebral ischae- patients (Ghika-Schmid et al., 1997; Kim, 2001), whereas a spe-
mia (Baquis et al., 1985; Yanagihara et al., 1985; Tatemichi et al., cific location of cerebral ischaemia in patients with limb-shaking
1990; Firlik et al., 1996; Baumgartner and Baumgartner, 1998; has not been found thus far.
Salah Uddin, 2004). Limb-shaking TIAs may resemble epileptic Previous studies concluded that limb-shaking TIAs are likely to
seizures but can be distinguished by a normal level of conscious- be caused by a low flow state of the brain and not by emboli,
ness, precipitation of symptoms by specific circumstances that may based on diminished vasomotor reactivity by TCD (Tatemichi
lower cerebral blood flow in patients with ICA occlusion, such as et al., 1990; Baumgartner and Baumgartner, 1998; Niehaus
rising or exercise, the absence of tonic contractions or a march of et al., 1998) and cerebral blood flow by Xenon inhalation
symptoms, no involvement of the face or trunk, and no epileptic (Yanagihara et al., 1985; Tatemichi et al., 1990). However, of
discharges on an EEG (Yanagihara et al., 1985; Baumgartner and patients with ICA occlusion in general, 12% had an exhausted
Baumgartner, 1998; Schulz and Rothwell, 2002). Various other and 29% a diminished CO2-reactivity when investigated by TCD
920 | Brain 2010: 133; 915–922 S. Persoon et al.
Table 3 Comparison of collateral blood flow pathways and TCD CO2-reactivity between patients with ICA occlusion with
(n = 34) and without (n = 68) limb-shaking
Limb-shaking group (n = 34) Control group (n = 68) Odds ratio (95% CI)
Collateral flow via anterior communicating artery 18/28 (64%) 44/56 (79%) 0.5 (0.2–1.3)
Collateral flow via posterior communicating artery 21/24 (88%) 30/44 (68%) 3.3 (0.8–12.8)
Collateral flow via ophthalmic artery 20/30 (67%) 23/52 (44%) 2.5 (0.99–6.4)
Leptomeningeal vessels 21/25 (84%) 21/48 (44%) 6.8 (2.0–22.7)
CO2-reactivity in %a (mean SD) 5 16 12 17 0.97 (0.94–1.00)b
a CO2 -reactivity on the symptomatic side could be measured in 29 patients with limb-shaking and in 63 patients without limb-shaking. In three patients TCD could not be
performed and we had to exclude six patients because of an absent temporal bone window and one control patient with ischaemic stroke as an outlier with a CO2-reactivity
of 127%.
b CO2 -reactivity is expressed as OR per 1% increase in CO2-reactivity.
we only classified the symptomatic infarcts, whereas patients with perfusion in patients with carotid occlusion. Stroke 1999b; 30:
1025–32.
ICA occlusion often have asymptomatic ischaemic lesions in the
Firlik AD, Firlik KS, Yonas H. Physiological diagnosis and surgical treat-
deep watershed area. ment of recurrent limb shaking: case report. Neurosurgery 1996; 39:
Fourth, the current study inevitably had some element of retro- 607–11.
spective ascertainment of the data, as it was not yet designed at Fisher CM. Concerning recurrent transient cerebral ischemic attacks. Can
the beginning of the prospectively collected series of patients with Med Assoc J 1962; 86: 1091–9.
a symptomatic ICA occlusion. Fifth, we did not perform an Fox AJ. How to measure carotid stenosis. Radiology 1993; 186: 316–8.
Garrett MC, Komotar RJ, Starke RM, Merkow MB, Otten ML,
inter-observer study with respect to the clinical diagnosis of Sciacca RL, et al. The efficacy of direct extracranial-intracranial
limb-shaking. This might have influenced the frequency of bypass in the treatment of symptomatic hemodynamic failure second-
limb-shaking in our series, although we strictly defined the criteria ary to athero-occlusive disease: a systematic review. Clin Neurol
for this diagnosis before inclusion of patients. Finally, we did not Neurosurg 2009; 111: 319–26.
perform EEG in all patients to confirm the absence of epileptic Ghika-Schmid F, Ghika J, Regli F, Bogousslavsky J. Hyperkinetic move-
ment disorders during and after acute stroke: the Lausanne Stroke
discharges at the time of limb-shaking.
Registry. J Neurol Sci 1997; 146: 109–16.
In conclusion, we have further characterized the clinical features Grubb RL Jr, Derdeyn CP, Fritsch SM, Carpenter DA, Yundt KD,
of limb-shaking TIAs that may improve their recognition by clini-
Smith HA, Thompson-Dobkin J, Yonas H, Flint E. Correlation of xenon- of stroke in patients with symptomatic carotid artery occlusive disease.
enhanced computed tomography-defined cerebral blood flow reactiv- J Vasc Surg 1995; 21: 338–44.
ity and collateral flow patterns. Stroke 1994; 25: 1784–7. Widder B, Kleiser B, Krapf H. Course of cerebrovascular reactivity in
Tatemichi TK, Young WL, Prohovnik I, Gitelman DR, Correll JW, patients with carotid artery occlusions. Stroke 1994; 25: 1963–7.
Mohr JP. Perfusion insufficiency in limb-shaking transient ischemic Yamauchi H, Kudoh T, Sugimoto K, Takahashi M, Kishibe Y,
attacks. Stroke 1990; 21: 341–7. Okazawa H. Pattern of collaterals, type of infarcts, and haemodynamic
The EC/IC Bypass Study Group. Failure of extracranial-intracranial arterial impairment in carotid artery occlusion. J Neurol Neurosurg Psychiatry
bypass to reduce the risk of ischemic stroke. Results of an international 2004; 75: 1697–1701.
randomized trial. N Engl J Med 1985; 313: 1191–1200. Yanagihara T, Piepgras DG, Klass DW. Repetitive involuntary movement
Vernieri F, Pasqualetti P, Passarelli F, Rossini PM, Silvestrini M. Outcome associated with episodic cerebral ischemia. Ann Neurol 1985; 18:
of carotid artery occlusion is predicted by cerebrovascular reactivity. 244–50.
Stroke 1999; 30: 593–8. Zaidat OO, Werz MA, Landis DM, Selman W. Orthostatic limb shaking
Webster MW, Makaroun MS, Steed DL, Smith HA, Johnson DW, from carotid hypoperfusion. Neurology 1999; 53: 650–1.
Yonas H. Compromised cerebral blood flow reactivity is a predictor