Clinical study

Prospective study of neuropsychological and psychosocial
outcome following surgical excision of intracerebral
arteriovenous malformations

Geoffrey A. Marshall

1

MSSC, DIP CLIN PSYCH

MSSC, DIP CLIN PSYCH

, Benjamin P. Jonker

2

MBBS

MBBS

, Michael K. Morgan

2

MD

MD

,

FRACS

FRACS

,

Alan J. Taylor

3

PHD

PHD

1

Department of Neurology, Royal North Shore Hospital, Sydney, Australia,

2

North and West Cerebrovascular Unit, Department of Surgery, University of Sydney,

Sydney, Australia,

3

Department of Psychology, Macquarie University, Sydney, Australia

Summary In this prospective study the neuropsychological and psychosocial function of 64 patients undergoing surgical resection of cerebral
arteriovenous malformations was examined prior to surgery (T1), one month post-surgery (T2) and one year post-surgery (T3). A mild but
widespread cognitive decline was observed pre-operatively. There was a trend toward decreased neuropsychological function at T2. All
neuropsychological tests showed a trend toward improvement at T3 compared with both pre-operative (generally not statistically significant)
and early post-operative values (generally significant). Patients were assessed for change between testing times. At T2 patients were more
likely to have deteriorated than improved, whereas at T3 the group which had altered from baseline were more likely to have improved than
deteriorated. Deterioration in some verbal/language tasks was more common for left sided AVMs. Outcome did not differ significantly for
patients presenting with haemorrhage. Psychosocial function was unchanged at late follow-up for the majority of patients.
s

c 2002 Published by Elsevier Science Ltd.

Keywords: cerebral arteriovenous malformation, neuropsychology, cognitive, psychosocial

INTRODUCTION

Patients

harboring

intracranial

arteriovenous

malformations

(AVMs) present with a diverse range of clinical features which
include haemorrhage, seizures, headaches, focal neurological
deficits, and Ômental changesÕ.

1–5

These Ômental changesÕ have

been well documented in a number of case reports,

6–10

but overt

cognitive impairment appears to account for only a small pro-
portion of overall presentations.

4

Decisions regarding treatment of these lesions are necessarily

complex, and involve balancing the risks and benefits of various
treatment options such as surgical excision, embolization, radio-
surgery, and conservative management. Whilst sufficient data is
available concerning mortality and neurological morbidity,

11–13

cognitive and psychosocial outcome have been reported less fre-
quently.

14;15

Knowledge of the likely cognitive and psychosocial

outcome following surgery could assist clinicians in their decision
making.

In this prospective study a large, relatively unselected group of

patients underwent comprehensive neuropsychological and psy-
chosocial evaluation both prior to and following surgery in order
to determine the effects of AVM excision on cognitive function-
ing.

MATERIALS AND METHODS

Patients

Between July 1992 and March 1996 110 patients underwent
radical surgical excision of supratentorial AVMs of the brain by

the neurosurgeon (MKM) and were considered for eligibility by
the following criteria:

1. aged between 16 and 75 years.
2. fluent English speakers.
3. no history of previous neurological or psychological disease

unrelated to the AVM.

4. cerebral AVM was confirmed angiographically prior to

surgery.

5. sufficiently alert and attentive to undertake neuropsychological

testing.

6. initial surgery performed electively or urgently, but not as an

emergency.

Forty-six patients were ineligible, with the most common reasons
being insufficient alertness (15 patients), not fluent English
speakers (11 patients), or initial surgery performed as an emer-
gency (8 patients). The clinical characteristics of the 64 eligible
patients are presented in Table 1.

Neuropsychological assessment

Patients were assessed on 3 occasions: prior to surgery (T1),
within 4–6 weeks post-surgery (T2), and at 12 months post-sur-
gery (T3) (range: 9–24 months). Patients generally underwent
surgery within one month of diagnosis.

As a result of clinical and practical considerations not all pa-

tients received all assessments or tests. One patient was unas-
sessable at T2. Final neuropsychological follow-up (T3) was not
performed in 12 patients. Of those twelve, 2 had died, 6 had left
the area, and 4 declined to return for neuropsychological assess-
ment.

Neurocognitive functioning was evaluated using a variety of

tests, outlined in Table 2. To assess the patientsÕ general ability
prior to the current presentation the National Adult Reading Test

16

with adjustment for education

17

was included to estimate Ôpre-

morbidÕ cognitive levels and gauge any longstanding disruption to
cognitive functions. Psychosocial assessment was undertaken on

Journal of Clinical Neuroscience (2003) 10(1), 42–47
ª 2002 Published by Elsevier Science Ltd.
DOI:10.1016/S0967-5868(02)00217-5

Received 28 December 2001

Accepted 15 April 2002

Correspondence to: Professor Michael K. Morgan, MD FRACS, North and

West Cerebrovascular Unit, Department of Surgery, University of Sydney,

Suite 40-41, Level 8, 193 Macquarie Street, Sydney, NSW 2000, Australia.

Tel.: [phone]; Fax: [phone];

E-mail:

[e-mail]

42

the basis of interview and measured level of drive, independence,
occupational status, and affective disturbance. Each measure was
rated on a 3 point scale where 1 represented intact function, 2
represented a mild decline, and 3 represented a marked decline in
the area concerned. All assessments were performed by a single
neuropsychologist (GAM).

Statistical analysis

Repeated measures analysis of variance was used to test the sta-
tistical significance of changes in neuropsychological test scores
over time. Since there were 14 tests, an alpha of 0.05/14

¼ 0.0036

was used for each test, and a sequential Bonferroni test procedure
was used for tests of overall effects.

18

Significant overall results

were followed up by paired comparisons, with alpha set at 0.0036/
3

¼ 0.0012. The sequential Bonferroni test procedure was used for

each of the three sets of comparisons (T2 vs T1, T3 vs T1 and T3
vs T2).

Since practice effects were likely to play a part, the overall re-

sults were examined in the light of effects reported for some of the
measures by other investigators.

19;20

Reliable change indices (RCIs)

were also calculated using previously described methods.

21;22

Since it is probable that there were real changes in neuropsy-

chological function for some of the patients who underwent sur-
gery, regardless of the significance of overall changes, RCIs

21;22

were calculated for scores at T2 and T3 vs those at T1. (Since the
size of the effect of practice on differences between T2 and T3
was impossible to gauge, this comparison was not considered. It
seems unlikely that the effect of practice on scores obtained at the
third sitting of a test was as great again as the effect at the second
sitting, especially when T3 came a year after T2.)

The calculation of the RCIs was based on the standard devi-

ation of scores at T1 and the relevant test-retest correlations and
practice effects.

19;20

Ninety percent RCIs were derived.

Analyses of covariance were carried out to see whether

changes in test scores were related to patient characteristics such
as sex, age, hand, and the site, size and grade of the AVM.

The psychosocial measures were grouped into two categories:

intact function and mild or marked decline (there were relatively
few patients in the marked category). Changes in the proportion of
patients in each category over the three test occasions were as-
sessed using CochranÕs Q test and significant overall results were
followed up with McNemar tests. Since there were four psycho-
social measures, an alpha of 0.05/4

¼ 0.0125 was used for overall

tests, and 0.0125/3 was used for the follow-up tests. Analyses
were based on 51 patients who had data for all three times. An-
alyses to see whether changes over time were related to patient
characteristics were carried out with the SAS CATMOD proce-
dure.

23

For this purpose continuous variables such as age were

recoded to make them into categorical variables.

RESULTS

Study population

Demographic data for the 64 eligible patients is provided in Table
1, along with the equivalent information for the 51 patients who
were included in statistical analysis, and the eleven patients ex-
cluded from analysis because of incomplete data (excluding the 2
patients who died). Patients with data on all 3 occasions did not
differ significantly with those who were excluded. As can be seen
in Table 1, the patient sample who underwent statistical analysis
in this study is representative of the study population, and there is
no evidence that patients unavailable for neuropsychological
testing were in any way different from those who were available
to undertake neuropsychological testing.

Premorbid cognitive function

Comparison of estimated IQ with measured FSIQ, VIQ, and PIQ
revealed a small but statistically significant difference, each actual
IQ measurement lower than the estimated IQ (Table 3). Separate
analysis on the basis of AVM grade, site, and size did not show

Table 1

Clinical characteristics of 64 patients

N = 64

N = 51

N = 11

Age

35.6

36

34.9

Sex

Male

33/64 (52%)

27/51 (53%)

5/11 (45%)

Female

31/64 (48%)

24/51 (47%)

6/11 (55%)

AVM Grade

I

8/64 (13%)

6/51 (12%)

2/11 (18%)

II

29/64 (45%)

24/51 (47%)

5/11 (45%)

III

18/64 (28%)

15/51 (29%)

2/11 (18%)

IV

8/64 (13%)

5/51 (9.8%) 2/11 (18%)

V

1/64 (1.6%)

1/51 (2%)

0/11 (0%)

Side

Left

28/64 (44%)

23/51 (45%)

5/11 (45%)

Right

36/64 (56%)

28/51 (55%)

6/11 (55%)

Size (cm)

0–2

19/64 (30%)

14/51 (27%)

5/11 (45%)

>

2–4

27/64 (42%)

23/51 (45%)

3/11 (27%)

>

4

18/64 (28%)

14/51 (27%)

3/11 (27%)

Location

Anterior

23/64 (36%)

20/51 (41%)

3/11 (27%)

Posterior

41/64 (64%)

31/51 (61%)

8/11 (73%)

History

Haemorrhage

32/64 (50%)

21/51 (41%)

3/11 (27%)

Seizures

36/64 (56%)

31/51 (61%)

5/11 (45%)

Preoperative embolization

13/64 (20%)

9/51 (18%)

3/11 (27%)

Associated aneurysms

11/64 (17%)

10/51 (20%)

0/11 (0%)

General neurological

outcome-Modified Rankin
Scale (MRS)

0

43/64 (67%)

35/51 (69%)

8/11 (73%)

1

12/64 (12%)

11/51 (22%)

1/11 (9%)

2

6/64 (6%)

4/51 (7.8%) 2/11 (18%)

3

0/64 (0%)

0/51 (0%)

0/11 (0%)

4

1/64 (1%)

1/51 (2%)

0/11 (0%)

5

0/64 (0%)

0/51 (0%)

0/11 (0%)

6

2/64 (2%)

0/51 (0%)

0/11 (0%)

Note: Sixty-four patients were eligible for the study. Of these there were 51
patients who had data for all 3 occasions (T1, T2, T3). Of the remaining 13
patients: 2 died (one of pancreatitis, one of unrelated septicemia 3 months
post-op), 1 was unassessable at early follow-up (T2), and 10 did not attend
late follow-up (T3) because they had left the area (6 patients) or declined to
return for testing (4 patients). Therefore: N

¼ 64 represents the entire study

population; N

¼ 51 represents those patients included in statistical analysis;

N

¼ 11 represents those patients for whom data was incomplete (excluding

the 2 patients which died).

Table 2

Neuropsychological measures

Function

Neuropsychological test

General cognitive ability

WAIS-R Full Scale IQ

44

WAIS-R Verbal IQ

44

WAIS-R Performance IQ

44

Attention

Digit Span

44

Mental processing speed

Digit Symbol

44

Visual perception

Rey Complex Figure Test

45

Memory and learning

Logical Memory I&II

46

Visual reproduction I&II

46

Rey Auditory Verbal Learning Test (Total)

45

Naming

Boston Naming Test

47

Verbal fluency

Controlled Oral Word Association Test

48

Executive ability

Mazes

49

Neuropsychological outcome of AVM surgery

43

ª

2002 Published by Elsevier Science Ltd.

Journal of Clinical Neuroscience (2003) 10(1), 42–47

any of these factors to be associated with larger differences be-
tween estimated and measured IQ.

Neuropsychological status prior to surgery (T1), at
early follow-up (T2), and at late follow-up (T3)

Two different patterns over time are evident in Table 4. The IQ,
D.Sp, D.Sy, RCFT, Rey_T, COWAT and MAZE scores all show a
decrease from T1 to T2 and an increase at T3 to a level higher
than that at T1. The WMS-R memory scores (LMII, LMII, VRI
and VRII) and the BNTscore, on the other hand, increase from T1
to T2 and from T2 to T3. In the first group of measures, only the
COWAT showed a significant decrease from T1 to T2 (results of
comparisons are shown in the last three columns of Table 4). The
increase from T2 to T3 was significant for all of the IQ measures
and also for the D.Sp, D.Sy and COWATand Maze scores, while
the difference between T1 and T3 was significant for FSIQ and
PIQ. In the second group of measures, the increase from T1 to T2
was significant for LMI and LMII, as was the difference between
T1 and T3 scores.

Practice effects

Non-zero practice effects for WAIS IQ scores

19

and LMI and

LMII scores

20

were included in recalculations of the T2 vs T1 and

T3 vs T1 comparisons for these measures shown in Table 4 (the
rounded practice effect given in the papers was added to the ob-
tained difference and the result was divided by the standard error
for the original analysis). Although the decrease in IQ scores
between T1 and T2 became more marked (in the original analysis,

IQ scores decreased despite a presumed practice effect), none
reached significance. In addition, none of the adjusted differences
between T1 and T3 was significant. The only difference for the
LMI scores which remained significant was that between T1 and
T3 for LMII (t(48)

¼ 4.6, p < 0:0005).

Correlates of change

There was no relationship between age, sex, years of education,
hand, size, and grade of the AVM and the nature of the event
preceding diagnosis (seizure, haemorrhage or other) and the
magnitude and direction of change in neuropsychological score
over time. However the interaction between time and site was
significant for REY_Tscores

ðp < 0:0036Þ and almost so for VIQ

scores (p

¼ 0:005Þ. In both cases the decrease in the score at T2

(and the recovery from T2 to T3) was much more marked when
the AVM was on the left side than when it was on the right.

Reliable changes

The percentage of patients who showed a reliable change in either
a positive or negative direction for each test is shown in Table 5.
At T2 up to 25% of patients showed a decline in scores. For FSIQ,
VIQ, D.Sy and Rey_T, the number of patients whose scores had
changed for the worse outnumbered those whose scores had in-
creased. At T3, patients were more likely to show an increased
score relative to that at T1. For FSIQ, PIQ D.Sy, LMI, LMII, VRI
and VRII the number whose score increased outnumbered those
whose score decreased. The percentage of patients who showed a
reliable decrease in scores between T1 and T3 was greater than
10% for only the VIQ and D.Sp scores. Very few patients whose
scores had not decreased reliably from T1 to T2 showed a reliable
decrease at T3 (1/37 for PIQ, 1/34 for D.Sp and 2/37 for Rey_T).
However, some patients who had shown a reliable decrease at T2
also showed a reliable decrease at T3. The largest numbers oc-
curred for VIQ (6/9), D.Sp (6/9), D.Sy (4/12), PIQ (3/7) and FSIQ
(2/9). Analyses showed that patients whose scores decreased re-
liably from both T1 to T2 and T1 to T3 did not differ on any of the
demographic or AVM variables listed earlier. There was also no
relationship between reliable change from either T1 to T2 or T1 to
T3 and the same patient characteristics. The nearest approach
ðp ¼ 0:016Þ occurred for the change in the Rey_Tscore from T1
to T3: the four patients who showed a reliable decrease tended to

Table 3

Estimated and measured general cognitive ability prior to surgery

(T1)

Mean

SD

Difference from

Est IQ

Significance

Est. IQ

100.7

12.4

–

–

FSIQ

96.1

14.5

)4.7

p < 0:001

VIQ

95.7

14.5

)5.1

p < 0:001

PIQ

97.7

15.1

)3.1

p < 0:05

Est IQ: Estimated IQ, FSIQ: WAIS-R Full Scale IQ, VIQ: WAIS-R Verbal IQ,
PIQ: WAIS-R Performance IQ.

Table 4

Mean scores on neuropsychological tests prior to surgery (T1), early post-surgery (T2) and at long-term follow-up (T3)

Modality

Test

N

T1

T2

T3

Comparisons

Mean

SD

Mean

SD

Mean

SD

2v1

3v1

3v2

General cognitive ability

FSIQ

48

97.2

15.4

96.5

14.5

101.0

15.2

+

+

VIQ

48

97.0

15.1

96.0

12.4

98.6

13.3

+

PIQ

49

98.3

16.2

97.9

18.2

103.6

17.2

+

+

Attention

D.Sp

50

10.0

2.7

9.5

2.8

10.4

2.5

+

Mental processing speed

D.Sym

48

9.3

3.1

8.4

3.4

9.8

3.4

+

Visual perception

RCFT

48

33.0

4.6

32.1

5.4

33.2

4.0

Memory and learning

LMI

49

26.4

7.7

29.5

7.8

30.8

6.9

+

+

LMII

49

20.5

9.3

25.0

9.7

27.2

8.8

+

+

VRI

50

32.4

8.4

33.2

6.5

34.2

6.2

VRII

50

29.5

10.2

31.0

9.1

32.3

8.5

REY T

45

48.6

9.5

46.7

11.0

49.6

9.8

Naming

BNT

30

51.5

8.8

51.9

8.9

53.3

9.2

+

+

Verbal fluency

COWAT

43

33.4

15.3

27.7

13.1

34.8

16.4

)

+

Executive ability

MAZE

44

25.1

5.4

23.2

5.7

26.0

4.3

+

Note: + or

) Significant difference between times at p < 0:05 (sequential Bonferroni-adjusted for three comparisons per test). + indicates that the mean score was

higher at the later time (e.g., T2 in T2 vs T1);

) indicates that the mean score was lower for the later test.

FSIQ: WAIS-R Full Scale IQ, VIQ: WAIS-R Verbal IQ, PIQ: WAIS-R Performance IQ, D.Sp: Digit Span, D.Sym: Digit Symbol, RCFT: Rey Complex Figure Test,
LMI: Logical Memory (Immediate), LMII: Logical Memory (Delayed), VRI: Visual Reproduction (Immediate), VRII: Visual Reproduction (Delayed), REY T: Rey
Auditory Verbal Learning Test (Total), BNT: Boston Naming Test, COWAT: Controlled Oral Word Association Test, MAZE: Mazes.

Significant overall time difference at p < 0:05 (sequential Bonferroni-adjusted for 14 tests).

44 Marshall et al.

Journal of Clinical Neuroscience (2003) 10(1), 42–47

ª

2002 Published by Elsevier Science Ltd.

be older than the 34 patients whose scores did not change sig-
nificantly (mean age 53.0 [12.6] and 33.5 [11.1]).

Psychosocial measures

The proportion of patients who were judged to have a deficit on
each of the psychosocial measures at each time is shown in Table
6. The pattern was similar for all measures: there was a rise in the
proportion showing a deficit at T2 which declined at T3, but not to
the pre-operative level. The overall change over time was sig-
nificant for Drive, Occupational status and Independence. Follow-
up tests showed a significant increase from T1 to T2 for all three
measures. Only one variable was related to changes in psycho-
social measures. It was found that the increase in deteroration of
Occupational status from T1 to T2 was much greater for patients
with 11or more years of education than it was for those with up to
10 years of education: the deterioration occurred for only one of
the 26 patients in the lower education group and for 11 of the 25
with in the higher education group. Further investigation showed
that of the 18 patients who were in unskilled or semi-skilled oc-
cupations, 17 showed no deterioration in Occupational status from
T1 to T2. Thirteen of these subjects were in the lower educational
group.

DISCUSSION

Preoperative cognitive and psychosocial status

Preoperatively, our patients showed significant reductions in
general cognitive abilities compared with expected levels. Al-

though the overall differences were of small magnitude, they were
highly significant for two of the three indices. These results
suggest that some degree of cognitive impairment is common in
patients harboring AVMs. This is consistent with other contem-
porary studies in which neuropsychological evaluation has been
undertaken.

14;24–27

The current study confirms a mild but wide-

spread cognitive deterioration prior to surgery.

The effect of cognitive decline on an individual patientÕs level

of independence and occupational status depends on the extent
and nature of the decline as well as the specific situation of the
patient. In this series the widespread cognitive decline which was
witnessed was not associated with a consequent reduction in in-
dependence or occupational status in the majority of patients.

The basis of impaired cognition may lie in the effects of

haemorrhage, compression of surrounding tissue, or chronic ce-
rebral hypoperfusion.

14;26;28

Haemorrhage and tissue compression

cause neurologic dysfunction by well-established mechanisms.
The absence of capillaries and consequent low resistance results in
shunting of blood away from viable brain tissue into the AVM.
This produces both regional arterial hypotension and venous hy-
pertension resulting in chronic cerebral hypoperfusion.

29–35

This

may in turn be associated with altered neuronal structure and
function

36;37

and neurological or neuropsychological deteriora-

tion.

2;38–42

Post-operative cognitive outcome (early)

At early post-operative follow-up there was a trend towards re-
duced scores in most areas of assessment. The exception was in
the area of memory and learning in which scores tended to in-
crease. With the incorporation of adjustment for practice effect the
trend toward reduced scores became more marked, and the sig-
nificant increase in some areas of memory and learning was no
longer seen. The only statistically significant result after inclusion
of the practice effect was a decrease in verbal fluency as measured
by COWAT. These results confirm that in the short term (4–6
weeks post-operative) this group of patients showed a trend to-
wards cognitive decline, although the magnitude was small.

The percentage of patients with a reliable change in neuro-

psychological scores was calculated, and when positive implies
the onset of a neuropsychological deficit or improvement in the
given area. Our results demonstrate that for each given neuro-
psychological variable most patients did not show a significant
change from pre-operative status at one month after surgery.
However for the proportion that did show a change (approxi-
mately 25% for most variables), deterioration was more common
than improvement. Patients can be warned that at 4–6 weeks
following surgery they will not yet have improved maximally, and
they should allow up to a full year for cognitive recovery.

Post-operative cognitive outcome (late)

When patients underwent neuropsychological evaluation at late
follow-up (T3) the trend towards cognitive decline witnessed for
most tests at T2 was no longer apparent. Instead, a trend towards
improved neuropsychological function was seen in all tests,
reaching significance in 5/14. The trend towards improved cog-
nitive function remained after adjustment for practice effect, and
was statistically significant for LMII. When outcome data from T3
was compared with T2 this trend was even more marked, reaching
significance in 8/14 tests.

Potential causes, other than practice effect, of improved neu-

ropsychological function following AVM resection include
evacuation of a hematoma, natural recovery following haemor-
rhage, or improvement in cerebral hemodynamics with normali-

Table 5

The percentage of patients who showed a reliable change in

neuropsychological scores from T1 to T2 and T1 to T3

N

T1

! T2

T1

! T3

%

%

%

%

%

%

)

0

+

)

0

+

FSIQ

48

19

75

6

6

77

17

VIQ

48

19

71

10

13

73

15

PIQ

49

14

76

10

8

76

16

D.Sp

50

18

68

14

14

68

18

D.Sy

48

25

71

4

8

75

17

LMI

49

8

82

10

2

92

6

LMII

49

8

80

12

0

82

18

VRI

50

2

96

2

2

88

10

VRII

50

2

96

2

2

88

10

Rey_T

45

13

82

4

9

78

13

FSIQ: WAIS-R Full Scale IQ, VIQ: WAIS-R Verbal IQ, PIQ: WAIS-R
Performance IQ, D.Sp: Digit Span, D.Sym: Digit Symbol, LMI: Logical Memory
(Immediate), LMII: Logical Memory (Delayed), VRI: Visual Reproduction
(Immediate), VRII: Visual Reproduction (Delayed), REY T: Rey Auditory
Verbal Learning Test (Total).

Table 6

Percentage of patients with deterioration of psychosocial

condition prior to surgery (T1), in the early post-operative period (T2) and
at long term follow-up (T3)

Measure

N

T1 (%) T2 (%) T3 (%)

Comparisons

2 vs 1 3 vs 1 3 vs 2

Drive

51

3.9

17.6

9.8

+

Affect

51

15.7

27.5

19.6

Occupational status

51

2.0

25.5

13.7

+

Independence

51

2.0

15.7

9.8

+

Note: + or

) Significant difference between times at p < 0:05 (sequential

Bonferroni-adjusted for three comparisons per test). + indicates that the
proportion of patients with a psychosocial deficit was higher for the later time
(e.g., T2 in T2 vs T1).

Significant overall time difference at p < 0:05 (sequential Bonferroni-

adjusted for 4 tests).

Neuropsychological outcome of AVM surgery

45

ª

2002 Published by Elsevier Science Ltd.

Journal of Clinical Neuroscience (2003) 10(1), 42–47

sation of cerebral perfusion. This is consistent with other studies
which suggest that neuropsychological impairment is at least
partially reversible with surgery.

7;9;15;43

Effect of correlates of change

A significant difference was found in performance of an auditory
learning task (Rey T) over time, patients with left sided AVMs
showing a more marked deterioration, particularly at early follow-
up. In addition the Verbal IQ scores showed a similar trend, nearly
reaching significance. Since language is predominantly a function
of the left cerebral hemisphere, the poor performance in these
tests, especially at T2 may represent a local effect from resection.
Whilst the localisation of language is widely recognized, most
neuropsychological tests do not represent a single function
or correspond to a single brain locus, and it is therefore not sur-
prising that other correlations between site and test scores were
not found.

There was no significant effect of any of the above factors on

the likelihood of sustaining a reliable change in neuropsycho-
logical function, although there was a trend towards poorer out-
come at T3 in older patients for at least one function (Rey T).

The potentially at risk groups with larger AVMs, higher

Spetzler–Martin grades, and hemorrhagic presentations did not
fare worse than other patients. It must be noted, however, that the
absolute numbers of high grade AVMs were low, and increased
numbers may be required to obtain the statistical power to detect
such an effect. In addition there were a number of patients who
were excluded from our study for insufficient alertness, or re-
quired emergency surgery because of a deterioration. These pa-
tients were almost exclusively patients with haemorrhage. Thus
the haemorrhage patients in our study population must be con-
sidered to represent the more benign side of the spectrum of
hemorrhagic AVM disease.

Post-operative psychosocial outcome

There was a significant increase in the number of patients showing
impairment in psychosocial function in the areas of drive, occu-
pational status, and independence at early follow-up. These sig-
nificant deteriorations had resolved by late follow-up, but not to
pre-operative levels. In some patients psychosocial measures were
improved by surgery, but this was unusual. Our results indicate
that whilst some patients had a permanent deterioration in psy-
chosocial function this was unusual.

One sensitive indicator of the effect of AVM resection on

patients ultimate outcome is their ability to return to work in the
same capacity as they left. Our study shows that the deterioration
in occupational status was more profound for patients in the
higher educational group and those in skilled occupations. This
effect was significant at early follow-up, though it was no longer
so at late follow-up. Data regarding the natural history of occu-
pational status in conservatively treated patients is scant. How-
ever, in one report of AVM patients treated conservatively 6/25
(24%) showed reduction in occupational status at a mean follow-
up of 10.6 years.

27

By contrast, at long term follow-up 13.7% of

our patients had some reduction in ability to undertake the same
level of employment.

CONCLUSIONS

We conclude that AVM surgery has a relatively benign effect on
cognitive and psychosocial functioning in the vast majority of
patients. At one year most patients perform similarly to pre-op-
erative levels on most cognitive tasks. Of the small group of pa-

tients showing a change in cognitive function at 1 year,
improvement is slightly more common than deterioration. As
expected, deterioration in some verbal and language tasks is more
common with surgery for left sided AVMs. There is no increased
cognitive morbidity of surgery for patients with hemorrhagic
presentations. Most patients return to a similar level of psycho-
social functioning post-operatively, but a small group deteriorate
in their potential to undertake employment. Surgery continues to
be a highly effective method of AVM obliteration, and our data
suggests that with the use of careful microsurgical techniques
significant cognitive morbidity is unusual.

REFERENCES

1.

Mast H, Mohr JP, Osipov A et al. ÔStealÕ is an unestablished mechanism for the
clinical presentation of cerebral arteriovenous malformations. Stroke 1995; 26:
1215–1220.

2.

Olivecrona H, Riives J. Arteriovenous aneurysms of the brain. Arch Neurol
Psychiatry 1948; 59: 567–603.

3.

Paterson JH, McKissock W. A clinical survey of intracranial angiomas with
special reference to their mode of progression and surgical treatment: A report
of 110 cases. Brain 1956; 79: 233–267.

4.

Perret G, Nishioka H. Report on the cooperative study of intracranial aneurysms
and subarachnoid haemorrhage. Section VI. Arteriovenous malformations and
fistulae reported to the cooperative study. J Neurosurg 1966; 25: 467–490.

5.

Wyburn-Mason R. Arteriovenous aneurysm of mid-brain and retina, facial
naevi and mental changes. Brain 1943; 66: 163–189.

6.

Wertz RT, Messert B, Collins M, Rosenbek JC, Kao CC. Right hemisphere
language dominance in a case of left hemisphere arteriovenous malformation. J
Speech Hear Disord 1977; 42: 106–112.

7.

v Overbeeke JJ, Bosma NJ, Verdonck AFMM, von Huffelen AC. Higher
cortical disorders: An unusual presentation of an arteriovenous malformation.
Neurosurgery 1987; 21: 839–842.

8.

Koffler SP, Roberts EV, Mancall EL. Replacement of dominant temporal lobe
by arteriovenous malformation with minimal neuropsychological impairment:
Report of a case. Neuropsychology 1990; 4: 201–213.

9.

Carter PL, Morgan M, Urrea D. Psychological improvement following
arteriovenous malformation excision. J Neurosurg 1975; 42: 452–456.

10.

Conley F, Moses JA, Helle TL. Deficits of higher cortical functioning in two
patients with posterior parietal arteriovenous malformations: Use of a
standardized Luria–Nebraska neuropsychological battery for pre- and
postoperative assessment. Neurosurgery 1980; 7: 230–237.

11.

Morgan MK, Sekhon LH, Finfer S, Grinnell V. Delayed neurological
deterioration following resection of arteriovenous malformations of the brain. J
Neurosurg 1999; 90: 695–701.

12.

Morgan MK, Johnston IH, Hallinan JM, Weber NC. Complications of surgery
for arteriovenous malformations of the brain. J Neurosurg 1993; 78: 176–182.

13.

Heros RC, Korosue K, Diebold PM. Surgical excision of cerebral arteriovenous
malformations: Late results. Neurosurgery 1990; 26: 570–578.

14.

Stabell KE, Nornes H. Prospective neuropsychological investigation of patients
with supratentorial arteriovenous malformations. Acta Neurochir (Wien) 1994;
131: 32–44.

15.

Mahalick DM, Ruff RM, Heary RF, U HS. Preoperative versus postoperative
neuropsychological sequelae of arteriovenous malformations. Neurosurgery
1993; 33: 563–570.

16.

Nelson. The National Adult Reading Test: Test Manual. 1982 NFER-Nelson,
Windsor 1982.

17.

Willshire D, Kinsella G, Prior M. Estimating the WAIS-R IQ from the National
Adult Reading Test: A cross-validation. J Clin Exper Neuropsych 1991; 13.

18.

Holm S. A simple sequentially rejective multiple test procedure. Scand J Stat
1979; 6: 65–70.

19.

Hermann B, Seidenberg M, Schoenfeld J, Peterson J, Leveroni C, Wyler A.
Empirical techniques for determining the reliability, magnitude, and pattern of
neuropsychological change after epilepsy surgery. Epilepsia 1996; 37: 942–950.

20.

Sawrie S, Chelune G, Naugle R, Luders H. Empirical methods for assessing
meaningful neuropsychological change following epilepsy surgery. J Int
Neuropsych Soc 1996; 2: 556–564.

21.

Chelune G, Naugle R, Luders H, Sedlak J, Awad I. Individual change after
epilepsy surgery: Practice effects and base-rate information. Neuropsychology
1993; 7: 41–52.

22.

Jacobson N, Truax P. Clinical significance: A statistical approach to defining
meaningful change in psychotherapy research. J Consult Clin Psychol 1991; 59:
12–19.

23.

Stanish W, Koch G. The use of CATMOD for repeated measurement analysis
for categorical data. In: Proceedings of the Ninth Annual SAS Users group
International SUGI Conference. SAS Institute, Cary, NC 1984; 761–770.

46 Marshall et al.

Journal of Clinical Neuroscience (2003) 10(1), 42–47

ª

2002 Published by Elsevier Science Ltd.

24.

Wenz F, Steinvorth S, Wildermuth S et al. Assessment of neuropsychological
changes in patients with arteriovenous malformation (AVM) after radiosurgery.
Int J Radiat Oncol Biol Phys 1998; 42: 995–999.

25.

Waltimo O, Putkonen A-R. Intellectual performance of patients with
intracranial arteriovenous malformations. Brain 1974; 97: 511–520.

26.

Mahalick DM, Ruff RM, U HS. Neuropsychological sequelae of arteriovenous
malformations. Neurosurgery 1991; 29: 351–357.

27.

Andersen EB, Petersen J, Mortensen EL, Udesen H. Conservatively treated
patients with arteriovenous malformation: Mental and physical outcome. J
Neurol Neurosurg Psychiatry 1988; 51: 1208–1212.

28.

Nornes H, Grip A. Hemodynamic aspects of cerebral arteriovenous
malformations. J Neurosurg 1980; 53: 454–456.

29.

Weisberg LA, Pierce JF, Jabbari B. Intracranial hypertension resulting from a
cerebrovascular malformation. South Med J 1977; 70: 624–626.

30.

Vassilouthis J. Cerebral arteriovenous malformation with intracranial
hypertension. Surg Neurol 1979; 11: 402–404.

31.

Obrador S, Soto M, Silvela J. Clinical syndromes of arteriovenous
malformations of the transverse-sigmoid sinus. J Neurol Neurosurg Psychiatry
1975; 38: 436–451.

32.

Lasjaunias P, Chiu M, Ter Brugge K, Tolia A, Hurth M, Bernstein M.
Neurological manifestations of intracranial dural arteriovenous malformations.
J Neurosurg 1986; 64: 724–730.

33.

Hurst RW, Hackney DB, Goldberg HI, Davis RA. Reversible arteriovenous
malformation-induced venous hypertension as a cause of neurological deficits.
Neurosurgery 1992; 30: 422–425.

34.

Barrow DL. Unruptured cerebral arteriovenous malformations presenting with
intracranial hypertension. Neurosurgery 1988; 23: 484–490.

35.

Barnett GH, Little JR, Ebrahim ZY, Jones SC, Friel HT. Cerebral circulation
during arteriovenous malformation operation. Neurosurgery 1987; 20: 836–842.

36.

Sekhon LHS, Morgan MK, Spence I, Weber NC. Chronic cerebral hypoperfusion
and impaired neuronal function in rats. Stroke 1994; 25: 1022–1027.

37.

Sekhon LHS, Morgan MK, Spence I, Weber NC. Chronic cerebral
hypoperfusion: Pathological and behavioural consequences. Neurosurgery
1997; 40: 548–556.

38.

Fedoruk S, Feindel W. Measurement of brain circulation time by radioactive
iodinated albumin. Can J Surg 1960; 3: 312–318.

39.

Feindel W, Yamamoto YL, Hodge CP. Red cerebral veins and the cerebral steal
syndrome. Evidence from of fluoroscein angiography and microregional blood
flow by radioisotopes during excision of an angioma. J Neurosurg 1971; 35:
167–179.

40.

Homan RW, Devous MD, Stokely EM, Bonte FJ. Quantification of
intracerebral steal in patients with arteriovenous malformation. Arch Neurol
1986; 43: 779–785.

41.

Okabe T, Meyer JS, Okayasu H et al. Xenon-enhanced CT CBF measurements
in cerebral AVMs before and after excision: Contribution to pathogenesis and
treatment. J Neurosurg 1983; 59: 21–31.

42.

Tanaka K, Yonekawa Y, Kaku Y, Kazekawa K. Arteriovenous malformations
and diaschisis. Acta Neurochir (Wien) 1994; 131: 32–44.

43.

Guidetti B, Delitala A. Intracranial arteriovenous malformations. J Neurosurg
1980; 53: 149–152.

44.

Wechsler D. Wechsler Adult Intelligence Scale-Revised. 1981 The
Psychological Corporation, San Antonio, SC 1981.

45.

Lezak M. Neuropsychological Assessment. 1995 Oxford University Press, New
York 1995.

46.

Wechsler D. Wechsler Memory Scale-Revised. 1987 The Psychological
Corporation, San Antonio, SC 1987.

47.

Kaplan EF, Goodglass H, Weintraub S. The Boston Naming Test. 1983 Lea and
Febiger, Philadelphia, PA 1983.

48.

Spreen O, Strauss E. Compendium of Neuropsychological Tests. 1991
University Press, Oxford 1991.

49.

Wechsler D. Wechsler Intelligence Scale for Children-Revised. 1974 The
Psychological Corporation, San Antonio, SC 1974.

Neuropsychological outcome of AVM surgery

47

ª

2002 Published by Elsevier Science Ltd.

Journal of Clinical Neuroscience (2003) 10(1), 42–47