A study of medial pallidotomy for Parkinson's disease: clinical outcome, MRI location and complications.

Abstract

We have studied the effects of unilateral ventral medial pallidotomy in 26 patients with medically intractable Parkinson's disease with marked drug-induced dyskinesias. Preoperatively, all patients were assessed during one 5-day admission according to the Core Assessment Programme for Intracerebral Transplantation (CAPIT) protocol, including rating in the 'practically defined off' and 'best on' states before and during a single-dose levodopa challenge. Motor performance was assessed with subset categories of the Unified Parkinson's Disease Rating Scale (UPDRS), timed motor tests and a standard dyskinesia rating scale. Pallidotomy was performed under stereotaxic CT guidance with intra-operative extracellular microelectrode recording made from the basal ganglia. All patients were re-assessed 3 months postoperatively and a subgroup (n = 9) have so far also been re-assessed after 1 year. Pre- and postoperative performance scores were compared in order to determine which categories of performance improved postoperatively. Significance was accepted at P < 0.005 in order to take into account the multiple number of comparisons performed. Patient medication was compared pre- and postoperatively and the morbidity associated with surgery was also recorded. The most significant improvement postoperatively was the diminution of 'on' dyskinesias contralaterally (67%, P = 0.0001); however, ipsilateral (45%, P = 0.0006) and axial (50%, P = 0.0008) dyskinesias also improved. Contralateral to pallidotomy, the median 'off' motor UPDRS score improved by 27% (P = 0.001) and a significant improvement was also observed in contralateral rigidity by 25% (P = 0.001). There were trends towards improvement in contralateral tremor (33%, P = 0.016) and bradykinesia (24%, P = 0.013) scores. Ipsilateral rigidity improved by 22% (P = 0.005), but other ipsilateral motor scores did not alter significantly. The 'off' gait/postural instability score and 'off' walking time showed marginally significant improvements by 7% (P = 0.007) and 29% (P = 0.014), respectively. On medication, no significant postoperative improvements in parkinsonism were detected. Anti-parkinsonian medication increased by 11% postoperatively. In the subgroup who were available for assessment 1 year postoperatively, responses were generally maintained. Two (7.7%) of the 26 patients had fatal complications (one cerebral haemorrhage and one haemorrhagic infarct) directly related to surgery. Among the remaining 24 patients, four (15.4% of the total 26) had major complications (two persisting and two transient). Ten patients (38.5%) had minor complications. The majority of the complications (major and minor) occurred in the earlier operated patients and the complication rate subsequently declined with increasing operative experience. The remaining 10 patients (38.5%) had no significant side-effects. One of these 10 patients died from an incidental malignant glioma 6 months postoperatively. These findings confirm that levodopa-induced dyskinesias are dramatically reduced following ventral medial pallidotomy and constitute the principal indication for pallidotomy. Improvements in underlying parkinsonism were of smaller magnitude. Pallidotomy may also offer some patients an opportunity to increase antiparkinsonian medication. Patient selection for medial pallidotomy should, therefore, be based largely on anticipated improvements in levodopa-induced dyskinesias, but this must be balanced against the associated morbidity and mortality.

Full text

Page 1

Brain (1998), 121, 59–75
A study of medial pallidotomy for Parkinson’s
disease: clinical outcome, MRI location and
complications
M. Samuel,1,2E. Caputo,2D. J. Brooks,1,2A. Schrag,2T. Scaravilli,2N. M. Branston,2J. C. Rothwell,2
C. D. Marsden,2D. G. T. Thomas,2A. J. Lees2and N. P. Quinn2
1MRC Cyclotron Unit, Hammersmith Hospital and
2Institute of Neurology, London, UK
Correspondence to: Dr M. Samuel, MRC Cyclotron Unit,
Hammersmith Hospital, London W12 0NN, UK
Summary
We have studied the effects of unilateral ventral medial
pallidotomy in 26 patients with medically intractable
Parkinson’s disease with marked drug-induced dyskinesias.
Preoperatively, all patients were assessed during one 5-day
admission according to the Core Assessment Programme for
Intracerebral Transplantation (CAPIT) protocol, including
rating in the ‘practically defined off’ and ‘best on’ states
before and during a single-dose levodopa challenge. Motor
performance was assessed with subset categories of the
Unified Parkinson’s Disease Rating Scale (UPDRS), timed
motor tests and a standard dyskinesia rating scale.
Pallidotomy was performed under stereotaxic CT guidance
with intra-operative extracellular microelectrode recording
made from the basal ganglia. All patients were re-assessed
3 months postoperatively and a subgroup (n ? 9) have so
far also been re-assessed after 1 year. Pre- and postoperative
performance scores were compared in order to determine
which categories of performance improved postoperatively.
Significance was accepted at P ? 0.005 in order to take into
account the multiple number of comparisons performed.
Patient medication was compared pre- and postoperatively
and the morbidity associated with surgery was also recorded.
The most significant improvement postoperatively was the
diminution of ‘on’ dyskinesias contralaterally (67%, P ?
0.0001); however, ipsilateral (45%, P ? 0.0006) and axial
(50%, P ? 0.0008) dyskinesias also improved. Contralateral
topallidotomy,themedian‘off’motorUPDRSscoreimproved
by 27% (P ? 0.001) and a significant improvement was also
observed in contralateral rigidity by 25% (P ? 0.001). There
were trends towards improvement in contralateral tremor
(33%, P ? 0.016) and bradykinesia (24%, P ? 0.013) scores.
Keywords: pallidotomy; Parkinson’s disease; clinical; MRI; complications
Abbreviations: AC–PC ? anterior commissure–posterior commissure (plane); CAPIT ? Core Assessment Programme for
Intracerebral Transplantation; DDCI ? levodopa/dopa decarboxylase inhibitor; UPDRS ? Unified Parkinson’s Disease
Rating Scale
© Oxford University Press 1998
Ipsilateral rigidity improved by 22% (P ? 0.005), but other
ipsilateral motor scores did not alter significantly. The ‘off’
gait/postural instability score and ‘off’ walking time showed
marginally significant improvements by 7% (P ? 0.007) and
29% (P ? 0.014), respectively. On medication, no significant
postoperative improvements in parkinsonism were detected.
Anti-parkinsonianmedication
postoperatively. In the subgroup who were available for
assessment 1 year postoperatively, responses were generally
maintained. Two (7.7%) of the 26 patients had fatal
complications (onecerebral
haemorrhagic infarct) directly related to surgery. Among the
remaining 24 patients, four (15.4% of the total 26) had major
complications (two persisting and two transient). Ten patients
(38.5%) had minor complications. The majority of the
complications (major and minor) occurred in the earlier
operated patients and the complication rate subsequently
declined with increasing operative experience. The remaining
10 patients (38.5%) had no significant side-effects. One of
these 10 patients died from an incidental malignant glioma 6
months postoperatively. These findings confirm that levodopa-
induced dyskinesias are dramatically reduced following
ventral medial pallidotomy and constitute the principal
indication for pallidotomy. Improvements in underlying
parkinsonism were of smaller magnitude. Pallidotomy may
also offer some patients an opportunity to increase anti-
parkinsonian medication. Patient selection for medial
pallidotomy should, therefore, be based largely on anticipated
improvements in levodopa-induced dyskinesias, but this must
be balanced against the associated morbidity and mortality.
increased by 11%
haemorrhage andone

Page 2

60
M. Samuel et al.
Introduction
James Parkinson noted in 1817 that the tremor of paralysis
agitans improved following an acute lesion of the motor
system which resulted in co-incident paralysis. By the
1930s, surgical section of the motor and premotor cortex
had been investigated as treatment for Parkinson’s disease
but hemiparesis occurred in conjunction with alleviation
of tremor (Bucy and Case, 1939). The early work of Russell
Meyers, however, demonstrated that surgical resection of
the head of the caudate nucleus and the anterior limb of
the internal capsule (Meyers, 1942) and of pallidofugal
fibres (Meyers, 1951) could improve rigidity and tremor
without inducing paresis. Concurrent evolution of the
stereotaxic technique (Spiegel et al., 1947) allowed human
neurosurgery to be performed with improved precision and
less hazard and stereotaxic chemopallidotomy using procaine
oil (Narabayashi and Okuma, 1953), alcohol (Cooper and
Bravo, 1958a) and electrocoagulation (Guiot and Brion,
1953; Guiot, 1958) were reported to improve parkinsonian
tremor and rigidity effectively. Svennilson et al. (1960)
were the first to report that when the posteroventral internal
pallidum was lesioned additional benefit to general motor
function (interpreted as corresponding to relief of akinesia)
could be obtained. At this time, stereotaxic lesions of
thalamic nuclei which directly receive pallidal efferents
(i.e. of ventral anterior and lateral thalamus) were generally
felt to provide more effective relief of parkinsonian tremor
and rigidity than pallidotomy (Cooper and Bravo, 1958a,
b) but the introduction of levodopa as effective anti-
akinetic therapy for Parkinson’s disease in 1967 led to a
major reduction in the use of pallidotomy and thalamotomy
in the treatment of parkinsonism.
However, the chronic administration of levodopa in
Parkinson’s disease is associated with the development of
fluctuations in motor response and with dystonic and choreic
involuntary movements (dyskinesias). Early in the course of
Parkinson’s disease, dyskinesia is generally associated with
peak plasma levels of levodopa and can usually be controlled
by decreasing the size of each dose and increasing the
frequency of dosing. Eventually, with disease progression, the
‘therapeutic window’ narrows and dyskinesias may develop,
either throughout any period of benefit or biphasically. At
this stage, medical management consists of balancing the
relief of bradykinesia against induction of dyskinesias, often
with neither being managed optimally. This unsatisfactory
state of affairs prompted the resurrection of surgical therapy
in Parkinson’s disease and recent reports of pallidotomy
(Dogali et al., 1995; Iacono et al., 1995; Laitinen, 1995;
Lozano et al., 1995; Sutton et al., 1995; Baron et al., 1996)
have demonstrated significant improvement of the cardinal
features of Parkinson’s disease (bradykinesia, tremor, rigidity
and gait disturbance) as well as a marked diminution of
levodopa-induced dyskinesias. We wished to study the
efficacy of medial pallidotomy in the treatment of Parkinson’s
disease and over the last 2 years we have performed unilateral
ventral medial pallidotomy in 26 patients with severe
Parkinson’s disease. We report here the surgical method
employed, lesion locations and the effect of the procedure
on motor performance, levodopa-induced dyskinesias, level
of dopaminergic medication required postoperatively and the
associated morbidity and mortality.
Methods
Between March 1995 and July 1996, 26 patients (17 male,
nine female, mean age 55.9 ? 9.1 years, mean duration
of Parkinson’s disease 15.4 ? 6.3 years) with medically
intractable idiopathic Parkinson’s
unilateral medial pallidotomy. All patients were selected by
at least two senior neurologists (N.P.Q., A.J.L. and C.D.M.)
and fulfilled the criteria for the diagnosis of idiopathic
Parkinson’s disease (Gibb and Lees, 1988). Thirteen of them
had young-onset Parkinson’s disease (Quinn et al., 1987).
Patients with ‘parkinsonian plus’ syndromes and other
neurologicaldisorderswereexcluded.Patientswereexamined
by a psychologist and underwent formal psychometric testing
in order to exclude those suffering from dementia. All patients
had asignificant responseto levodopaand otherdopaminergic
drugs, but in each case the response to medical therapy was
accompanied by the development of severe dyskinesias,
unpredictable on–offfluctuations
phenomena. In two subjects (Patients 3 and 5) levodopa-
induced dyskinesias were so disabling that they avoided
taking levodopa altogether preoperatively, preferring to
remain bradykinetic. The characteristics of the patients are
detailed in Table 1.
Two other patients had also been selected for pallidotomy
but did not undergo surgery. One of these, with severe ‘off’
period panic and anxiety declined to proceed on the morning
of surgery while ‘off’ medication. The other developed
respiratory embarrassment related to the flexed position of
her neck when the stereotaxic frame was fixed to the operating
table and the procedure was abandoned uneventfully after
the burr hole had been made. These two patients are not
considered elsewhere in this paper.
The study protocol was approved by the Joint Medical
Ethics Committee of the National Hospital for Neurology
and Neurosurgery and the Institute of Neurology. Informed
consent was obtained from each patient prior to surgery.
diseaseunderwent
and wearing-off
Preoperative assessments
Patients were admitted to hospital 1–3 months preoperatively
for assessment. Motor performance was assessed according
to the Core Assessment Programme for Intracerebral
Transplantation (CAPIT) (Langston et al., 1992). All patients
were initially assessed in the ‘practically defined off’ state.
From the Unified Parkinson’s Disease Rating Scale (UPDRS),
‘activity of daily living’ scores (items 5–17 of the UPDRS

Page 3

Pallidotomy in Parkinson’s disease
61
Table 1 Patient characteristics preoperatively
Patient SexAgeDuration of
Parkinson’s disease
(years)
Side of
operation
Modified ‘Off’
Hoehn and
Yahr
UPDRS
‘Off’
UPDRS
‘On’
Preoperative medication
1*
2*
3*
4*
5
6
7*
8
9
F
M
M
M
F
F
M
M
F
M
M
F
M
M
M
F
M
F
F
M
M
F
M
M
M
M
57
52
46
40
59
66
42
66
61
61
61
69
61
62
72
40
49
41
51
59
53
61
48
54
58
65
55.9
9.1
18
18
11
13
11
24
17
19
19
25
17
12
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
Right
Right
Left
Left
Right
Left
Right
Left
Left
Left
Left
Right
Left
Left
V
V
V
III
IV
IV
IV
IV
V
III
IV
IV
IV
III
V
V
III
III
III
II
II.5
III
IV
III
III
II
70
84
71
66
44
38
68
68
53
26
55
52
59
78
96
53
36
27
51
27
35
39
92
37
54
33
53.3
18.9
33
39
31
23
18
31
35
25
36
9
16
12
18
40
29
23
17
6
18
15
16
10
38
10
16
11
22.8
11.3
Ap, Cl, L, P
Ap, L, P
Ap, Cl, D, P
Bz, L, P
D, O
Cb, L
Am, Bz, D, L
D, L, P
Ap, Br, Cl, D, L, P
Am, D, L, O
Ap, D, L, P
L
D, L, P
Am, L
Bz, L, P
Bz, D, L
D, L
Ap, L
Ap, L, P
L, P
L, P
Ap, L
Bz, L
D, L, O, P
L, P
L, P
10*
11
12
13
14*
15
16*
17*
18*
19*
20
21
22
23*
24
25*
26
Mean
SD
8
31
16
7
18
9
19
15
6
5
15
12
24
12
15.4
6.3
*Young onset Parkinson’s disease. Am ? amantadine; Ap ? apomorphine; Br ? bromocriptine; Bz ? benzhexol; Cb ? carbergoline;
Cl ? clonazepam; D ? deprenyl; L ? levodopa; O ? orphenadrine; P ? pergolide.
scale, maximum ? 52) and total motor scores (items 18–31,
maximum?108)wereconsideredtoreflectoverallfunctional
severity of Parkinson’s disease. However, as we were
interested to determine the effects of pallidotomy on specific
aspects of motor control, we calculated separate subset
UPDRS scores for the following categories of motor
performancecontralaterallyandipsilaterally:hemibodymotor
score (items 20–26, maximum ? 36); hemibody tremor
(items 20–21, maximum ? 12); hemibody rigidity (items 22,
maximum ? 8) and hemibody bradykinesia (items 23–26,
maximum ? 16). Subset scores were also generated for gait
and postural instability (items 13–15 and 27–30, maximum ?
28), speech (items 5 and 8, maximum ? 8) and freezing
(item 14, maximum ? 4). Patients were also timed during
performance of standard motor tests which included repetitive
alternate pronation and supination of each forearm (20
repetitions), sequential finger to thumb opposition for each
hand (10 cycles) and repetitive hand/arm movements between
two points 30 cm apart (10 repetitions). To assess gait, a
timed test was performed in which the patients were required
to stand from sitting, walk 7 m, turn, walk a further 7 m and
sit again.
Patients were also tested preoperatively with an identical
battery of tests following the administration of 200 mg of
dispersiblelevodopa(plus
withdrawal of anti-parkinsonian medication for 12 h. The
three most severely affected subjects (Patients 1–3) were not
able to tolerate 12 h without medication and required a
nocturnal dose of apomorphine, but in each case this was
given at least 6 h prior to the clinical evaluation. Patients
were assessed repeatedly throughout the duration of action
of the levodopa dose and ‘best on’ responses, as agreed by
the patient and clinician, were recorded for each patient.
These responses usually occurred 30–90 min after the
administration of the test dose of levodopa.
All patients developed choreic and dystonic involuntary
movements during the duration of action of the test dose of
levodopa. Since the severity of dyskinesia varied with motor
actions, we rated dyskinesias according to the protocol of
Goetz et al. (1994) during rest and performance of the
following motor tasks: (i) walking; (ii) raising a glass to the
mouth with each hand and attempting to drink; and (iii)
putting on, buttoning, unbuttoning and then taking off a
jacket. For each patient, dyskinesia was assessed every 20
min for the duration of action of the test dose of levodopa.
Dyskinesia was rated for neck, trunk and each limb
50mgbenserazide)after

Page 4

62
M. Samuel et al.
individually and scores were assigned from 0 to 4. A score
of 0 was recorded for successful task performance without
dyskinesia, while a score of 4 was given to functionally
disabling dyskinesia which prevented the completion of a
motor task. Scores of 1, 2 and 3 were given to intermediate
levels of disability. Subset scores were calculated for (i)
maximum contralateral dyskinesia (arm and leg, maximum
score ? 8); (ii) maximum ipsilateral dyskinesia (arm and
leg, maximum score ? 8); and (iii) maximum axial dyskinesia
(neck and trunk, maximum score ? 8).
Preoperatively, all patients had their medication recorded
and overall daily ‘off’ time was scored according to part IV
of the UPDRS (item 36). All patients underwent routine MRI
brain scans (1.5-T GE Signa system, spoiled grass, TR ?
650 ms, TE ? 4.2 s, flip angle ? 20°, field of view ? 20
cm, 256 ? 224 matrix and 1.5-mm slice thickness). Patients
were also examined by a neuro-ophthalmologist and their
visual fields were objectively documented by Goldmann
perimetry or tangent screen examination.
Surgery
Twenty-one patients underwent unilateral left pallidotomy
and five underwent unilateral right pallidotomy. On the
morning of the operation, patients initially received a short
general anaesthetic for the fitting of a Cosman–Roberts–Wells
stereotaxicframe(Radionics,USA).Whilestillanaesthetized,
they were transferred to a CT scanning suite for anatomical
selection of the target which was identified as the
ventroposterior medial pallidum contralateral to the side
which showed theworse
subsequently returned to the operating theatre and allowed
to wake from the anaesthetic prior to proceeding.
Patients were awake throughout the procedure, which was
preceded by appropriate application of local anaesthetics. A
17-mm burr hole was made in the precoronal frontal bone. A
monopolar microelectrode for recording electrophysiological
responses from the pallidum was introduced into the brain
at an angle of 30° to the vertical, aiming posteriorly, in a
sagittal plane 20–22 mm lateral to the midline. This was
slowly advanced towards the anatomical target by a manually
operated hydraulicmicromanipulator
calibrated so that the distance of the electrode tip from the
target was known at all times and displayed. The electrode
consisted of a tungsten wire of diameter 0.12 mm, tapered
and insulated except at the tip and fixed inside an insulated
stainless steel tube of 0.6 mm diameter, so that 12 mm of
the tungsten wire projected from the tube. The tube, in turn,
passed through a 1.0-mm diameter steel tube whose tip was
fixed at a preset distance (10 or 15 mm) above the target
and which was itself supported by a 1.5-mm guide tube
higher up.
As the microelectrode tip was advanced towards the target,
extracellular action potentials from cells in lateral pallidum,
medial pallidum and intralaminar border zones were
amplified, appropriately filtered and recorded, displayed on
dyskinesia. Patients were
whichwas pre-
an oscilloscope screen and heard on a loudspeaker. Cell
activitywasrecordedwiththepatientsatrestand,onoccasion,
during contralateral passive or voluntary movements of
the wrist, elbow or ankle. The typical patterns and frequencies
of activity obtained in these different regions have been
documented elsewhere (Hutchinson et al., 1994) and were
easily identifiable. The ventral border of the medial pallidum
was clearly demarcated by the characteristic transition from
thehighfrequencycellactivity(?60Hz)tothelowfrequency
sparse activity characteristic of white matter. Below this
demarcation the microelecelctrode was advanced further into
white matter and massed neuronal activity recorded during
1-Hz stroboscopic visual stimulation in order to confirm the
proximity of the upper surface of the optic tract.
The positionalinformation
electrophysiological data was then correlated with the
appropriate sagittal brain maps of the Schaltenbrand and
Wahren (1977) neurosurgical atlas and, if necessary, a
correction made to the anticipated position of the anatomical
target. While a single track was sufficient to achieve this
correction in 10 of the 26 patients, it was necessary to make
at least one additional track at a different anterior–posterior
or lateral coordinate in the remainder. Three tracks were
made in six patients and four tracks in two patients. The
target correction averaged over all 26 patients was 2.4 mm
(range 0–5.0 mm).
Abipolarradio-frequencystimulatingelectrode(Radionics,
USA; diameter 2 mm) was then introduced into one of the
microelectrode tracks and advanced to the target. High
frequency stimulation was applied (300 Hz, 0.2-ms pulses,
intensity 0.1–1.5 mA) and thresholds for motor and visual
phenomena were identified. The electrode position and
thresholds were finalized when stimulation produced no
detrimental motor, visual or speech effects. Pallidotomy was
performed by thermocoagulation (70–72°C for 60–80 s) via
the same electrode in situ at the physiological target. Clinical
testing of the patients was performed repeatedly during
electrode stimulation and lesioning. Each pallidotomy
comprised two to four overlapping individual lesions and
typically occupied the space of a cylinder 2–4 mm in diameter
and 6–8 mm in length, with the distal end lying in the most
ventroposterior portion of the medial pallidum. Clinical
improvements in rigidity, tremor and bradykinesia were often
evident as soon as the lesion was made.
Following pallidotomy, anti-parkinsonian medication was
restarted on the same evening of the operation. We made
particular efforts to document any possible complications of
surgery. Usually, after one night in intensive care, patients
were returned to the neurological ward for a further 4–6 days
during which they were re-examined and questioned about
any postoperative changes. Unwanted effects were classified
as major if they were life-threatening or interfered with the
patient’s daily routine and minor if they caused minimal or
nointerference withthe
Complications were also classified as persisting (lasting ?3
obtained fromthese
patient’s dailyroutines.

Page 5

Pallidotomy in Parkinson’s disease
63
months) or transient (lasting ?3 months) and as delayed if
they developed only after discharge home.
Postoperative assessments
All patients underwent routine MRI 1–2 days postoperatively
to document lesion locations. Eleven patients were also
available for high resolution volume MRI 5–6 weeks
postoperatively (1.0-T Picker HPQ system, TR ? 24 ms,
TE ? 6 ms, voxel size 1.0 ? 1.0 ? 1.3 mm3at the
Hammersmith Hospital, London, UK) in order to determine
the location and size of pallidotomy more accurately.
Calculations were performed using Analyze version 7.0
image display software (BRU, Mayo Foundation, USA).
Three months postoperatively, patients were re-admitted
for repeat neurological examination in a manner identical to
their preoperative assessments. In order to reduce inter-rater
variability, each patient was assessed by the same clinician
who had performed the preoperative assessment. UPDRS
subset scores and CAPIT performance times were calculated
for the same categories as preoperatively (i.e. ‘practically
defined off’ and ‘best on’ contralateral, ipsilateral and axial
scores). Medication requirements and visual fields were
documented as preoperatively.
To date, it has also been possible to assess nine patients 1
year postoperatively. These assessments were carried out
in an identical manner to the preoperative and 3-month
postoperative assessments.
Statistical comparisons
Since UPDRS and dyskinesia scores were categorical data
and all of our patients had severe Parkinson’s disease, it was
unlikely that the distribution of our subset scores would be
normal. Therefore, we calculated median preoperative and
postoperative group scores for each category and compared
them using paired Wilcoxon sign rank tests. For the
contralateral, ipsilateral and sit–walk–stand timed tests, we
compared mean preoperative performance time with mean
postoperative performance time by paired Student’s t tests.
Patients who were not able to complete a timed test were
excluded from the analysis for that category. In order to take
into account the multiple number of comparisons performed,
a statistical threshold of P ? 0.005 was considered to be
sufficiently conservative. Levels of significance which did
not reach this level, but in which P ? 0.05, were considered
as marginally significant. For categories which showed a
significant change postoperatively,
percentagechangeinmedianscoreormeanperformancetime.
Preoperatively, patients were taking a variety of anti-
parkinsonianmedication
decarboxylase inhibitor (DDCI) formulations (n ? 24),
apomorphine (n ? 3 by continuous pump infusion, n ? 5
by regular injection), pergolide (n ? 15) and bromocriptine
(n ? 1). Although we attempted to leave postoperative
medication for each patient unchanged for at least 3 months
wecalculatedthe
includinglevodopa/dopa
after pallidotomy, this was not possible in 13 patients since
their postoperative clinical condition necessitated a change
in medication. We compared preoperative with postoperative
meandailydoses oflevodopa/DDCI,
apomorphine separately using Student’s t tests. In order to
assess changes in medication for the entire group, we
calculated preoperative and postoperative total equivalent
levodopa/DDCI doses (equivalents of 100 mg of levodopa
were 133 mg of sustained release levodopa, 10 mg
bromocriptine or 1 mg pergolide) for each patient and
compared these by a Student’s t test. An equivalent
apomorphineconversionfactorislackingandsoapomorphine
doses were excluded from this conversion. Furthermore,
those patients taking apomorphine were only considered for
the group analysis if their total daily doses of apomorphine
had remained unchanged postoperatively (n ? 5). Since the
number of comparisons relating to medication was four, we
accepted a level of P ? 0.05 for significant changes in
medication doses.
For theninepatientsso
postoperatively, the UPDRS scores, dyskinesia scores, motor
performance times and medications were compared with their
preoperative scores in an identical manner.
pergolideand
farre-assessed1 year
Results
Three months postoperatively, we were unable to carry out
a complete series of assessments on one patient and another
was recovering from a major complication (see below).
Additionally, two patients died unexpectedly in the immediate
postoperativeperiod.Wewere,therefore, onlyabletoperform
group analyses of subset scores from 22 out of the 26 patients
Lesion location
The routine 1.5-T MRI scans which were obtained 1–2 days
postoperatively confirmed that the lesion was situated in the
medial pallidum in all cases. In the first patient, in whom it
had been possible to record extracellular action potentials
from only one track, the lesion was seen to encroach on the
posterior limb of the internal capsule. In three other patients,
the lesion extended more dorsally and laterally and included
a portion of lateral pallidum. The preoperative MRI scan of
another patient had revealed that his posterior cerebral artery
was in close proximity to the ventral portion of the medial
pallidum and so the most ventral point of his lesion was
tailored to lie 2–3 mm more dorsally than usual, but was
still confined to the medial pallidum. This was confirmed on
his postoperative high-resolution volume MRI brain scan and
also at autopsy (see below).
In 11 patients, high resolution volume 1.0 Tesla MRI scans
were obtained 5–6 weeks postoperatively. These included
nine patients in the first half of the series and two patients
in the second half. Overall, the mean lesion volume was 145
? 49 mm3. The location of the centre of the lesion was
located 21 ? 2 mm lateral to the midline and 3 ? 2 mm