Journal of Internal Medicine 2001; 249: 413±422
Aphasia in acute stroke and relation to outcome
A. C. LASKA1, A. HELLBLOM2, V. MURRAY1, T. KAHAN1 & M. VON ARBIN1
From the 1Division of Internal Medicine, Karolinska Institutet, Danderyd Hospital, and 2Department of Speech Pathology, Danderyd Hospital,
Danderyd, Sweden
Abstract. Laska AC, Hellblom A, Murray V, Kahan T,
von Arbin M (Danderyd Hospital, Danderyd, Swe-
den). Aphasia in acute stroke and relation to
outcome. J Intern Med 2001; 249: 413±422.
Objectives. The natural course of aphasia in unse-
lected, consecutive stroke patients is not well estab-
lished. We investigated morbidity, mortality and
recovery for different types of aphasia in consecutive
unselected aphasic patients with acute stroke.
Setting and subjects. In 119 aphasic patients, the
type and degree of aphasia were assessed acutely
and at 3, 6 and 18 months after stroke onset, using
Reinvang's `Grunntest for afasi' and Amsterdam-
Nijmegen-Everyday-Language-Test.
Results. About one-third of patients with acute
2 of type and degree of aphasia, great improvements
stroke had presented with aphasia. Mortality among
the aphasic patients during the 18-month follow-up
was twice that in non-aphasics (36 vs. 16%).
Presence of atrial ®brillation was associated with
poorer prognosis. At 18 months, 24% of the 119
aphasic patients had recovered completely, 43% still
had signi®cant aphasia, and 21% had died. The
Introduction
Aphasia is present in 21±38% of acute stroke
patients [1±3]. Ferro and Madureira [4] found that
non-¯uent aphasia predominated in young patients,
suggesting an anterior lesion, while in elderly
patients posterior infarcts were more frequent.
Knepper et al. [5] found cardiac emboli in 40% of
patients with cerebral infarction and Wernicke's
aphasia. Furthermore, Bogousslavsky et al. [6]
showed an association between a potential cardiac
source of embolism and Wernicke's aphasia, as well
as an involvement of some speci®c cerebral territor-
ies such as the posterior division of the middle
Ó 2001 Blackwell Science Ltd
proportion with global aphasia decreased from
almost 25% acutely to a few per cent after 18
months, that with Wernicke's aphasia from 25% to
less than 10%, whereas conduction aphasia
increased from 13 to 23% during follow-up. Among
those with initial mild aphasia, 70% recovered
completely. Great improvement was observed in
patients with initial low degree of speech function.
Younger patients recovered to a greater extent than
older patients.
Conclusion. The high long-term mortality among
aphasics may be seen as an indirect sign of advanced
cardiovascular disease. A combination of different
and adjusted aphasia tests provided the possibility to
assess almost all acute aphasic patients. Irrespective
were seen in almost all aphasic patients. Even
patients with severe speech impairment have a
considerable potential for recovery, particularly in
the ®rst 3 months after stroke.
Keywords: acute stroke, aphasia, atrial ®brillation,
mortality, recovery, types of aphasia.
cerebral artery. Atrial ®brillation is present in about
20% of patients with acute stroke, and mortality
among these patients is 1.5±3.0 times higher than
among stroke patients in sinus rhythm [7].
In most studies [8, 9], outcome for aphasic
patients refers to recovery, not to morbidity or
mortality. Factors such as age, gender, comorbidity
and severity of the stroke may in¯uence outcome,
but the results of earlier studies are not consistent
[9±12]. Recovery differs with the types of aphasia. It
is well documented that the rate of spontaneous
recovery in aphasia in stroke patients is highest
during the ®rst 3 months after stroke onset [8, 13].
However, most studies describing types of aphasia
413
414 A . C . L A S K A et al.
and recovery refer to selected groups, often small
and with the aphasia in a chronic stage [9,14,15].
Adequate planning of rehabilitation for stroke
patients with aphasia requires a good estimate of the
prognosis regarding morbidity and mortality, as well
as functional outcome. Studies of consecutive unse-
lected aphasic stroke patients in which types and
degrees of aphasia are tested are still lacking. The
aims of this study were thus the following: ®rst, to
examine the short- and long-term prognoses in a
group of well diagnosed, unselected, acute stroke
patients with and without aphasia; second, to
describe the different types of aphasia initially and
during follow-up; and third, to study the evolution of
aphasia and the pattern of recovery for different
types of aphasia.
Study design and patients
All acute stroke patients admitted to the stroke unit
at Danderyd Hospital from 1 June 1993 to 30
September 1994 were screened prospectively for
presence of aphasia. During this period, every fourth
consecutive acute stroke patient admitted to the
emergency care unit was strictly assigned directly to
the stroke unit. In this way, 106 consecutive,
unselected patients with acute stroke were included.
Among these, 36 patients with aphasia were iden-
ti®ed. Most of the remaining three-quarters of the
admitted acute stroke patients were transferred to
the stroke unit the following day, provided a bed was
available. Among these non-consecutive acute
stroke patients, another 83 aphasics were identi®ed,
yielding a total of 119 aphasic patients.
The 119 patients with aphasia were seen at
follow-up visits 3, 6 and 18 months after the index
stroke. The aphasic tests were repeated at each
follow-up visit. The patients who recovered com-
pletely from aphasia were not assessed further.
Mortality, however, was registered for all 189
patients (106 + 83) during the entire 18-month
follow-up period. The study was approved by the
Ethics Committee of Karolinska Hospital. The pa-
tients or their relatives gave their informed consent
as appropriate.
Methods
Demographic data, vital signs, history of stroke,
previous aphasia, cardiovascular morbidity (ischae-
Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422
mic heart disease, heart failure, atrial ®brillation,
myocardial infarction and hypertension) and desti-
nation of discharge from the stroke unit were
registered. The aetiology of stroke was de®ned as
thrombosis, cardiac emboli and intracerebral hae-
morrhage, in accordance with the International
Classi®cation of Diseases, 9th edition [16]. Informa-
tion on general health, living conditions, cardiovas-
cular morbidity and mortality was collected at each
follow-up visit. Cardiovascular events were de®ned
as recurrent stroke, myocardial infarction, heart
failure, symptomatic arrhythmia or cardiovascular
death during follow-up.
Cerebral computed tomography (CT) was per-
formed in 98% of the aphasic patients and in 96% of
all stroke patients within the ®rst few days following
admission. All cerebral CT scans were re-evaluated
openly by a neuroradiologist with knowledge of the
type of aphasia. Of the patients with cerebral
infarcts, 39% had a second CT scan and this was
also evaluated. The most informative CT scan was
used. The volume of the lesion was estimated and its
location was noted and assessed as typical or
atypical of the different types of aphasia, according
to Naeser and Hayward [17].
Neurological assessment was performed according
to the Scandinavian Supervision Stroke Scale (SSSS),
ranging from 7 to 29 [18], where a score of 7
indicates no de®cits. Activities of Daily Living (ADL)
were assessed according to the Katz Index [19],
ranging from A (fully independent) to G (totally
dependent).
Assessment of aphasia
All aphasia tests were performed by the same speech
pathologist within a median time of 5 days (range
0±30) after onset of stroke. Ninety per cent of all
subjects were assessed within 11 days. In a few
cases, the tests had to be postponed due to uncon-
sciousness, stroke progression and confusion. The
assessments were based on tests available in Swedish
and generally used by speech pathologists. The tests
were:
1 Swedish translation of Reinvang's `Grunntest
for afasi' [20]. This test, which is based on the
Boston terminology and is similar to the Western
Aphasia Battery [21], measures ¯uency, naming,
comprehension and repetition, as well as writing
and reading. The sum of the total scores for the main
1 APHASIA IN ACUTE STROKE
variables yields the aphasia coef®cient, which meas-
ures the degree of aphasia. Percentile values give
each patient's raw score for the variables naming,
repetition and comprehension in relation to the
score of the whole group of these aphasics. The
relation between the percentile value of three
parameters and ¯uency estimated from spontaneous
speech gives the type of aphasia. Since all our
patients were tested in the acute stage of stroke, an
adjusted version, representative of the entire Reinv-
ang test, was used [22]. The test took about 15 min
to perform. The reliability and validity of `Grunntest
for afasi' have been reported elsewhere [22]. The
Reinvang test recognizes that, in fact, some aphasic
patients have a mixture of two or more aphasia
syndromes. These patients are therefore classi®ed as
mixed non-¯uent and mixed ¯uent, respectively.
Fluent aphasias are Wernicke's, conduction, trans-
cortical sensory, anomic and mixed ¯uent. Non-
¯uent aphasias are global, Broca, transcortical
motor and mixed non-¯uent.
2 Verbal communicative behaviour was assessed
by the Amsterdam-Nijmegen-Everyday-Language-
Test (ANELT) [23]. In this functional test the
understandability of the patient's message and the
intelligibility of the utterance are each rated on a
®ve-point scale, where 5 is normal. A score of 0 is
given when the patient, due to severe aphasia, is
incapable of taking instructions and/or producing
an answer. Each parallel test consists of 10 items
and takes about 15 min to perform. The items are
constructed as scenarios of familiar daily life situa-
tions, e.g. calling a doctor or talking to a sales clerk.
The ANELT understandability score indicates the
severity of the communicative disability, and was
chosen as the measure of the degree of aphasia.
ANELT is considered to be a reliable and valid test
[23].
3 In cases with dysarthria, the Token test was
used during follow-up to discriminate aphasia [24].
The version used has a range of 0±36 points and the
cut-off was set at 30 points to de®ne aphasia.
4 The speech pathologist also ranked the degree
of aphasia subjectively, on a scale of 1±4, where 4
denoted an inability to carry out the test and 3
represented severe aphasia.
Evolution of aphasia was de®ned as a change in
the type of aphasia during the course of recovery
and was assessed among patients who were tested
on at least three occasions and who improved.
Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422
415
The question of whether the aphasic patients
would receive speech therapy or not was decided on
clinical grounds without reference to this investiga-
tion.
Statistical methods
Data are presented as mean values ‹ SEM, unless
stated otherwise. Contingency tables were evaluated
by the chi-square test or, in case of small expected
frequencies, by Fisher's exact test. Statistical com-
parisons between groups were made using the
Mann±Whitney test or, in the case of a normal
distribution, Student's t-test. Within-group analyses
were made using the pairwise Student's t-test for
correlated means. For multivariate regression ana-
lyses, multivariate analysis of variance was used. All
analyses were carried out with the JMP, Version 3.1
(SAS Institute Inc., Cary, NC, USA) or the SAS
system (SAS Institute). A probability value (P) of
<0.05 was considered to be statistically signi®cant.
Results
Patients with aphasia vs. no aphasia
Among the 106 unselected consecutive acute stroke
patients, aphasia was found in 36 subjects (33%).
Six of them had recurrent stroke with aphasia; thus,
30 patients (28%) suffered aphasia for the ®rst time.
The mean age of the consecutive stroke patients was
76 years (range 34±99). The aetiology of stroke in
these patients was thrombosis in 58%, cardiac
emboli in 29%, intracerebral haemorrhage in 9%,
and cerebrovascular disease of unknown cause
(cerebral CT scan not done) in 4%. There were no
signi®cant differences in baseline characteristics
between acute stroke patients with and without
aphasia (Table 1).
Of the consecutive stroke patients, 24 died within
18 months. Three-quarters of the deaths were
attributed to cardiovascular disease. Short-term
in-hospital mortality tended to be higher among
the aphasic patients than among the non-aphasic
patients (11 vs. 3%; Table 1). Accordingly, long-
term prognosis (18 months) showed a twofold
higher mortality among aphasic patients compared
with non-aphasic patients (36 vs 16%; Table 1). In a
multivariate analysis including age, aetiology of
stroke, presence of atrial ®brillation and degree of
416 A . C . L A S K A et al.

Table 1 Baseline characteristics and mortality in consecutive neurological de®cit, increased mortality among
acute stroke patients with and without aphasia aphasic patients was related to the presence of atrial
Aphasics Non-aphasics P ®brillation (P < 0.001).

No. of patients 36 70
Mean age (years) 78.6 75.9 NS All patients with aphasia
Female (%) 52 54 NS
Cardiac emboli (%) 39 24 NS The total group of 119 aphasic patients consisted of
Atrial ®brillation (%) 41 24 0.06 the 36 subjects from the consecutive acute stroke
IHD (%) 11 26 0.08
group and 83 patients identi®ed among the non-
SSSS (mean) 15 13 NS
Mortality in the acute 11 3 0.10 consecutive acute stroke patients (Fig. 1). At 18
phase (%) months, 28 aphasics (24%) had recovered com-
Mortality during 36 16 0.02 pletely, 51 patients (43%) still had signi®cant
18 months follow-up (%)
aphasia and 25 (21%) of the initial 119 aphasic
SSSS, Scandinavian Supervision Stroke Scale; IHD, ischaemic patients had died.
heart disease; NS, not signi®cant.

Fig. 1 Flow chart of acute aphasic

patients and follow-up visits.
Aphasia was assessed by the
Reinvang and the ANELT tests. Ten
of the patients missing at 3 months
were seen at 6 months, and one
patient missing at 6 months was
seen at 18 months.

Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422

1 APHASIA IN ACUTE STROKE 417

The mean age of the 36 consecutive patients with
aphasia was higher than that of the non-consecutive
aphasic patients: 78 (range 53±95) vs. 74 (47±91)
years (P < 0.05). There were no differences in
gender, stroke aetiology, neurological de®cit, lesion
volume, type of aphasia or degree of aphasia
between the consecutive and non-consecutive apha-
sic patients. Case fatality during follow-up for 18
months was 36% among the consecutive aphasic
patients, as compared with 18% among non-con-
secutive patients (P < 0.05, by multivariate analysis
including age and degree of aphasia).
The distributions of the different types of aphasia
in the entire study population and among those with
remaining aphasia, as tested at each follow-up visit,
are shown in Fig. 2. Two patients could not be fully
assessed in the acute stage but were evaluated
during follow-up. In the acute stage there was no
age difference between patients with the different
types of aphasia. At 3 months, patients with global
aphasia tended to be younger [mean age 69 years
(range 56±79)], as compared with patients with
Wernicke's aphasia, whose mean age was 77 years
(47±90). Fluent aphasia was more common than
non-¯uent aphasia (77 vs. 40 patients; P < 0.01)
among our acute aphasic patients. The level of
speech function was higher among ¯uent aphasics
compared with non-¯uent aphasics (ANELT mean
2.9 vs. 0.6; P < 0.001).
Recovery
Recovery, expressed as change in the degree of
aphasia (ANELT understandability), is shown in
Fig. 3 for all aphasic patients who survived and were
present during follow-up. Improvements were seen
in all types of aphasia. A vast majority of the
patients had either global, Wernicke's or conduction
aphasia. Recovery in these three major groups is
also shown in Fig. 3.
Functional outcome, as measured by ANELT,
showed that patients with ¯uent aphasia reached a

Fig. 2 Distribution of different types of aphasia according to the Reinvang test, acute and at each follow-up. Patients who recovered
completely, were missed or died are also included. No aphasia ˆ complete recovery; not tested ˆ dead, missed or demented (see Fig. 1);
trans motor ˆ transcortical motor aphasia; trans sens ˆ transcortical sensory aphasia.

Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422

418 A . C . L A S K A et al.

Fig. 3 Recovery over time for all aphasic patients and for global, Wernicke's and conduction aphasia. Patients who died or were missed
during follow-up are excluded. The degree of aphasia assessed by ANELT is shown as median values with 25th and 75th quartiles for all
patients tested.

higher level of speech function than non-¯uent
aphasics (mean 4.3 vs. 2.5; P < 0.001).
The degree of recovery, measured as the difference
between ANELT in the acute stage and at
18 months, was not in¯uenced by age, gender,
severity of stroke, aetiology, previous stroke, con-
secutively or non-consecutively included patients,
type of aphasia, lesion side, lesion volume or ¯uency
in univariate analyses. However, in a multivariate
regression analysis including the above variables,
only age showed a signi®cant relation (P < 0.01),
i.e. the younger the patient, the higher the degree of
recovery. The initial severity of aphasia (ANELT
acute) correlated in a univariate analysis with the
degree of recovery (P < 0.001), i.e. the lower the
initial degree of aphasia, the higher the degree of
recovery.
The mean age of the 28 patients who recovered
completely was 77 years (range 52±95), as com-
pared with 72 years (47±95) among those who
did not (P ˆ 0.05), and their median initial
degree of neurological de®cit (range 7±29) was
12 points (25th-75th quartiles, 10±15), as com-
Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422
pared with 16 points (25th±75th quartiles, 10±
24) (P < 0.001). The patients with the mildest
degree of aphasia acutely, i.e. ANELT > 3.9
(n ˆ 28), all had cerebral infarcts and ¯uent
aphasia, and 19 of them (68%) recovered com-
pletely. Of those who did not recover completely,
one died, two had chronic aphasia and three were
missing at follow-up.
Of the total of 119 patients with aphasia, 26
suffered one or more cardiovascular events, inclu-
ding cardiovascular death, during the follow-up
period. The group with a cardiovascular event was
older (mean age 79 vs. 75 years, P < 0.05), had a
higher proportion of women (65 vs. 41%,
P < 0.05) and had more severe neurological de®cits
(SSSS mean 18 vs. 15 points, P ˆ 0.05). The group
also presented with more severe aphasia (ANELT
acute 0.0 vs. 1.8; P < 0.01). There were no
differences in aetiology, lesion volume or systolic
blood pressure.
There was no evident difference in recovery
among the left-handed aphasic patients (n ˆ 5)
compared with the right-handed patients.
1 APHASIA IN ACUTE STROKE 419

Cerebral CT
The lesion was clearly outlined on the cerebral CT
scan in 71 of the 119 patients with aphasia: 65 had
their lesion in the left hemisphere and six in the
right. The estimated median lesion volume was 15.5
mL (range 1±250). Cerebral infarcts were seen in 61
patients, and intracerebral haemorrhage in 10. Of
the six patients with right-hemisphere lesions, ®ve
were female, all of them right-handed, while the
single male was left-handed. Lesion volume and
lesion location for the different types of aphasia are
shown in Table 2. The estimated lesion volumes
related to the degree of aphasia both initially and at
18 months (P < 0.01). Among patients with visible
lesions, lesion volume did not differ signi®cantly
between those who recovered completely and those
who did not. There was no signi®cant difference in
long-term mortality (18 months) among those with
a visible lesion as compared with those without.
Evolution
Evolution of aphasia was observed in 34 of the 63
patients (54%) who could be tested at least three
times during 18 months and who improved
(Table 3). Evolution was not affected by age or
gender. Wernicke's aphasia was more frequent
among those whose aphasia evolved than among
those where it did not (47 vs. 15%, P ˆ 0.01).
Evolution was not found to be related to the initial
degree of aphasia measured with ANELT.
Discussion
The incidence of ®rst-ever aphasia in this study of
acute stroke patients was 28%, which con®rms
previous observations [1,2]. A 38% incidence of
aphasia in acute stroke patients was found by
Pedersen et al. [3] using the Scandinavian Stroke
Scale. However, that test is a blunter instrument and

Table 2 Initial type of aphasia and ®ndings on cerebral CT

No. of visible Lesion mean Involvement

Initial type of aphasia lesions volume (mL) of typical area Exceptions

Global (n = 25) 21 60 15 3 central lesions 1 occipital 2 frontal lobe

Broca (n = 1)
Mixed non-¯uent (n = 13)
Wernicke's (n = 29)
Conduction (n = 16)
Transcortical sensory (n = 12)
Mixed ¯uent (n = 17)
1 1
9 19
19 15
7 7
5 4
7 6.5
0 1 subcortical
4 1 pericallosa 4 subcortical
14 3 subcortical 1 medulla 1 frontal lobe
7
5
2 1 pericallosa 4 central lesions

Table 3 Evolution for 34 aphasic

patients during 18 months' follow- Initial type of aphasia Interim stage Evolved to
up
Global (n = 6) Mixed non-¯uent (4) Conduction (1)
Mixed ¯uent (1)
Broca (n = 1) Transcortical motor
Mixed non-¯uent (n = 3) Mixed ¯uent Mixed non-¯uent (2) Normal language (1)
Wernicke's (n = 16) Trans sens ± mixed ¯uent Conduction (2)
Mixed ¯uent Conduction (3)
Anomic (1)
Conduction Mixed ¯uent (3)
Normal language (1)
Wernicke's (1)
Trans sens (2)
Mixed non-¯uent (2)
Conduction (1)
Trans sens (n = 1) Trans motor Trans sens
Conduction (n = 5) Mixed ¯uent Mixed ¯uent (1) Normal language (4)
Mixed ¯uent (n = 2) Conduction Normal language (1) Trans sens (1)

Figures in parentheses indicate number of patients. Trans sens, transcortical sensory; trans moor,
transcortical motor.

Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422

420 A . C . L A S K A et al.
may include patients with no true aphasia. The
strength of this study lies in the unselected, con-
secutive nature of the sample. All subjects were
carefully diagnosed and we used previously well-
validated test instruments for aphasia. Indeed, 119
acute aphasic patients (in whom the type of aphasia
could be determined in 117) can be considered a
large study group that has not been described
earlier. In Sweden, more than 95% of patients with
an acute stroke are treated in hospital [25].
Accordingly, the present results may be considered
a good estimate of the incidence of aphasia in the
general population with acute stroke.
During the acute phase of stroke, patients are
often medically unstable and dif®cult to test. Many
of them do not have the mental or physical strength
or ability to participate in time-consuming aphasia
tests. For instance, according to Wade et al. [2], 28%
of acute stroke patients could not be tested, and
Pedersen et al. [3] excluded around 10% of the
patients in their study. In selected cases, we
performed the testing some days later. This enabled
us to test 90% of the subjects within 11 days. Our
results show that a combination of different and
adjusted aphasia tests performed by a speech
pathologist provided an opportunity to include all
aphasic patients and to assess the type and degree of
aphasia in detail for almost all of them.
Our hospital mortality (11%) was considerably
lower than reported previously [3]. This might be
due in part to the short mean duration of hospital
stays in Sweden, but also to lower mortality rates for
all stroke patients in recent years. Multivariate
analysis showed that the only factor with a signi-
®cant relationship to the higher mortality among
aphasic patients was atrial ®brillation. Stroke
patients with atrial ®brillation are older, have larger
lesions and higher short-term as well as long-term
mortality [26], more disabling strokes with a longer
duration of hospital stay and a lower rate of
discharge to their own homes [7]. Our ®ndings are
in accordance with previous observations of a
higher frequency of cardiac emboli in stroke patients
with aphasia [5, 6]. Thus, aphasia in acute stroke
patients suggests an increased mortality. Cardiac
embolism was, however, present in only half of these
deaths.
In the acute phase, almost 25% of the aphasic
patients had the global type of aphasia. At 18
months, the proportion had dropped to only a few
Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422
per cent, mainly due to a higher mortality among
patients with aphasia of this type. In contrast to our
®ndings, Ferro and Madureira [4] and De Renzi and
Ferrari [27] found patients with non-¯uent aphasia
to be younger than those with ¯uent aphasia.
However, the patients in those studies were not
assessed in the acute stage and, as Ferro and
Madureira [4] pointed out, they may have missed
older patients with severe strokes. It is therefore
important to know the time from stroke onset when
describing patient characteristics in relation to
different types of aphasia. The proportion of patients
with evolving aphasia was in agreement with
earlier studies [9,28]. Since the type of aphasia
changes dramatically over time (as clearly shown
in Fig. 2 and Table 3), it is also highly important to
state the time from onset of the stroke when
describing the distribution between different types
of aphasia.
Wernicke's aphasia characterized a quarter of the
aphasic patients in the acute stage but less than
10% at 18 months of follow-up, mainly due to a
high recovery rate and evolution in patients with
Wernicke's aphasia. The high proportion of Wer-
nicke's aphasia in this study might be due to the
high age of the study population. Earlier studies
[5,14] also indicate that older patients are more
prone to posterior infarcts, which elicit aphasic
syndromes with comprehension dif®culties.
Our results show that younger patients had a
higher recovery rate, which is in agreement with
others [9,15]. We also found, for all the different
types of aphasia, that the greatest improvement in
speech function occurred during the ®rst 3 months,
with only a little improvement later on [8,13,29].
Patients with severe speech impairment (global and
Wernicke's) improved more, although they did not
reach as high a level of speech function as those
with milder aphasia. Indeed, most patients (70%)
with an initially high degree of functional commu-
nication recovered completely. Accordingly, others
have shown that the milder the aphasia, the greater
the chance to recover to normal speech [2,3,15].
The apparently lower recovery rate among subjects
with mild aphasia is likely to be due to a ceiling effect
of the test instruments, i.e. an inability to record
improvement in these patients with mild aphasia.
All in all, the patient with the most impaired speech
function may have the greatest potential recovery
during rehabilitation.
1 APHASIA IN ACUTE STROKE
In this study, a cerebral CT scan was performed
on clinical grounds. The results suggest that lesion
location is of no help in predicting outcome. We
therefore agree with Knopman et al. [11], who
concluded that analysis of ¯uency is a better
predictor of outcome than analysis of the CT
®ndings. Moreover, the time factor must be taken
into account, as both the lesion as visualized on CT
and the type of aphasia evolve over time. The
optimal time for morphological description of a
cerebral CT scan may be about 3 weeks after stroke
onset [30].
In conclusion, acute stroke patients with aphasia
have high short- and long-term mortality, in most
cases due to concomitant cardiovascular disease.
The main factors in¯uencing speech recovery are
initial severity of aphasia and age. Identi®cation and
characterization of aphasia in acute stroke provide
valuable prognostic information. Knowledge about
spontaneous recovery is also an aid in planning
adequate rehabilitation. Further studies to evaluate
the possible in¯uence of training programmes and/
or pharmacological interventions on recovery and
quality of life for the aphasic stroke patient are
warranted.
Acknowledgements
We are most grateful to Arne Grepe MD PhD for
reassessment of the cerebral CT scans, and to
Peter Borenstein MD for valuable suggestions and
advice. The study was supported by the Swedish
Stroke Association Foundation and Funds, Karo-
linska Institutet, Danderyd Hospital Development
Funds (176), the Sera®mer Hospital Foundation,
the Lundbeck Foundation and the Foundations of
Eirs, SoÈderstroÈm-KoÈnigska Sjukhemmet and AÊ ke
Wiberg.
References
1 Brust JC, Shafer SQ, Richter RW, Bruun B. Aphasia in acute
stroke. Stroke 1976; 7: 167±74.
2 Wade DLHR, David R, Enderby P. Aphasia after stroke: natural
history and associated de®cits. J Neurol Neurosurg Psychiatry
1986; 49: 11±16.
3 Pedersen PM, Jorgensen HS, Nakayama H, Raaschou HO,
Olsen TS. Aphasia in acute stroke: incidence, determinants,
and recovery. Ann Neurol 1995; 38: 659±66.
4 Ferro JM, Madureira S. Aphasia type, age and cerebral infarct
localisation. J Neurol 1997; 244: 505±9.
Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422
421
5 Knepper LE, Biller J, Tranel D, Adams HP Jr, Marsh E. Etiology
of stroke in patients with Wernicke's aphasia. Stroke 1989; 20:
1730±2.
6 Bogousslavsky JCC, Regli F, Despland P-A, Van Melle G,
Kappenberger L. Cardiac source of embolism and cerebral
infarction ± clinical consequences and vascular concomitants.
Neurology 1991; 41: 855±9.
7 Lin HJWP, Kelly-Hayes M, BeiSeries AS, Kase CS, Bernjamin
EJ, D'Agostino RB. Stroke severity in atrial ®brillation. The
Framingham Study. Stroke 1996; 27: 1760±4.
8 Demeurisse G, Demol O, Derouck M, de Beuckelaer R,
Coekaerts MJ, Capon A. Quantitative study of the rate of
recovery from aphasia due to ischemic stroke. Stroke 1980;
11: 455±8.
9 Kertesz A, McCabe P. Recovery patterns and prognosis in
aphasia. Brain 1977; 100: 1±18.
10 Basso A, Capitani E, Moraschini S. Sex differences in recovery
from aphasia. Cortex 1982; 18: 469±75.
11 Knopman DS, Selnes OA, Niccum N, Rubens AB, Yock D,
Larson D. A longitudinal study of speech ¯uency in aphasia:
CT correlates of recovery and persistent non¯uency. Neurology
1983; 33: 1170±8.
12 Holland A, Greenhouse JDF, CS. Predictors of language
restitution following stroke: a multivariate analysis. J Speech
Hear Res 1989; 32: 232±8.
13 Sarno MT, Levita E. Natural course of recovery in severe
aphasia. Arch Phys Med Rehabil 1971; 52: 175±8.
14 Kertesz A, Sheppard A. The epidemiology of aphasic and
cognitive impairment in stroke: age, sex, aphasia type and
laterality differences. Brain 1981; 104: 117±28.
15 Marshall RC, Phillips DS. Prognosis for improved verbal
communication in aphasic stroke patients. Arch Phys Med
Rehabil 1983; 64: 597±600.
16 WHO. International Classi®cation of Diseases, 9th edn. Geneva:
WHO, 1967.
17 Naeser M, Hayward R. Lesion localization in aphasia with
cranial computed tomography and the Boston Diagnostic
Aphasia Exam. Neurology 1978; 28: 545±51.
18 RoÈden-JuÈllig A Ê , Britton M, Gustafsson C, Fugl-Meyer A.
Validation of four scales for the acute stage of stroke. J Intern
Med 1994; 236: 125±36.
19 Katz S, Ford A, Moskowitz R, Jackson B, Jaffe M. Studies of
illness in the aged. The index of ADL: a standardized measure
of biological and psychosocial function. J Am Med Assoc 1963;
185: 914±19.
20 Reinvang I. Aphasia and Brain Organisation. New York: Plenum
Press, 1985.
21 Kertesz A. Aphasia and Associated Disorders. New York: Grune
and Stratton, 1979.
22 Reinvang I, Sundet K. The validity of functional assessment
with neuropsychological tests in aphasic stroke patients. Scand
J Psychol 1985; 26: 208±18.
23 Blomert L, Kean M-L, Koster C, Schokker J. Amsterdam-
Nijmegen-Everyday-Language-Test. Construction, reliability
and validity. Aphasiology 1994; 8: 381±407.
24 De Renzi E, Vignolo LA. The Token test: a sensitive test to
detect receptive disturbances in aphasics. Brain 1962; 85:
665±78.
25 TereÂnt A. Increasing incidence of stroke among Swedish
women. Stroke 1988; 19: 598±603.
26 Britton M, Gustafsson C. Non-rheumatic atrial ®brillation as a
risk factor for stroke. Stroke 1985; 16: 182±8.
422 A . C . L A S K A et al.

27 De Renzi EFP, Ferrari P. The in¯uence of sex and age on the
incidence and type of aphasia. Cortex 1980; 16: 627±30.
28 Pashek GV, Holland AL. Evolution of aphasia in the ®rst year
post-onset. Cortex 1988; 24: 411±23.
29 Pickersgill MJ, Lincoln NB. Prognostic indicators and the
pattern of recovery of communication in aphasic stroke
patients. J Neurol Neurosurg Psychiatry 1983; 46: 130±9.
30 Mazzocchi FVL. Localisation of lesions in aphasia: clinical-CT
scan correlations in stroke patients. Cortex 1979; 15: 627±54.
Received 14 June 2000; revision received 20 December 2000;
accepted 20 December 2000
Correspondence: Ann Charlotte Laska MD, Division of Internal
Medicine, Karolinska Institutet Danderyd Hospital, S-182 88
Danderyd, Sweden (fax: +46 8-6226810; e-mail: ann-char-
lotte.laska@med.ds.sll.se).

Ó 2001 Blackwell Science Ltd Journal of Internal Medicine 249: 413±422