32 / 74
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 25, No 5, September/October 2014
226
AFRICA
from causes unrelated to thrombolysis therapy: cardiogenic heart
failure following an acute myocardial infarction, and recurrent
cerebral infarction and subsequent pneumonia. Eight patients
had evidence of asymptomatic intracranial haemorrhage,
attributed to haemorrhagic conversion of the ischaemic infarct.
In four of these patients, the CT scan was available to verify this
finding. Two patients had extracranial haemorrhage, including
one patient with a hip haematoma.
The median SITS-SICH score for our patient cohort was
4 (IQR 2–5). The patients were stratified into low, average,
moderate and high risk (see Table 1). In the GSH cohort, the
majority of patients had 3–5 points, or average risk, including
the two patients who developed SICH. There were no patients
who were scored as high risk (
>
9). The distribution of patient
risk in the GSH cohort differed from the SITS-SICH cohort,
with more patients classified as low or average risk, and no high-
risk patients (Table 4).
One GSH patient who had a SEDAN score of 3 suffered
a SICH (by both ECASS II and SITS-SICH criteria). There
was one death in this group, due to complications related to
pneumonia (Table 5).
Discussion
Urgent thrombolysis with IV tPA is a priority in the emergency
medical treatment of acute ischaemic stroke. Robust efficacy
data exists for this therapeutic modality, particularly when
administered within the first 90 min, and the window of benefit
has widened since it was first introduced from three to 4.5 hours
after onset of symptoms. Published data from the GSH Stroke
Unit reveal similar rates of SICH to those from developed and
several developing countries, which provide some reassurance to
clinicians concerned about the safety profile of this modality in
a South African setting. However, the risk of SICH remains, and
care should be taken to select patients who are likely to have the
most benefit at the lowest risk of SICH.
To be practical, the application of risk scores for SICH should
include information that is easily obtained in the emergency unit
(EU). They should contain independent risk factors for SICH
and take into account the interplay between these factors in an
individual patient.
The SITS-SICH score uses clinical variables that can be
attained relatively quickly and easily at the bedside in a resource-
constrained area. It has been validated in over 16 000 patients
from multiple centres, many of whom did not have prior
experience in thrombolysis. The SEDAN score uses clinical
information but it relies on the assessment of brain CT imaging
for subtle signs of stroke, which may be overlooked in a busy
EU setting by inexperienced reviewers. Many South African
centres use older (fewer slice) scanners that would decrease the
sensitivity in detecting such signs.
The overall rates of SICH seen in the SITS-SICH validation
cohort of 5.1% per ECASS II definition and 1.8% per SITS-
MOST definition compare with the GSH rates of 4.8 and 2.4%,
respectively. The SICH rate in the SEDAN score validation
cohort was 6.5%. There appeared to be a trend towards GSH
patients being slightly lower risk than either of the SITS-MOST
or SEDAN validation cohorts. This may reflect the more
cautious approach in patient selection being used at our centre.
The main limitation of this pilot study was that of small
sample size and low event rate in the GSH cohort. One is unable
to comment on the ability of either score to reliably predict the
risk of haemorrhage. However, the overall rate of SICH by the
ECASS II definitions was similar between the cohorts studied.
A further limitation was that CT brain scans taken prior to
2003 were not available for a review of the images, although
reports were present. Therefore, signs of early infarction and a
dense middle cerebral artery sign could not be evaluated as is
required for the SEDAN scoring system, nor could we confirm
the presence of a type II parenchymal haemorrhage, required for
the SITS-MOST definition of SICH.
Conclusion
The scores, in particular the SITS-SICH score, represent a
potentially useful clinical tool to aid in patient selection for
thrombolysis in ischaemic stroke. This study, piloting their use in
a South African cohort, suggests that they may be applicable in
our context but further research is required to validate their use.
Table 3. Baseline characteristics
Baseline characteristics (
n
= 42)
Median age, years (IQ range)
62 (50–66)
Weight on admission, kg (IQ range)
76 (67–80)
Preceding history of hypertension,
n
(%)
27 (66)
Median systolic BP on admission, mmHg (IQ range) 149 (134–175)
On anti-platelet therapy at admission,
n
(%)
11 (27)
Abnormal serum glucose on admission,
n
(%)
8 (20)
Mean time to thrombolysis (min)
169
Median NIHSS score on admission,
n
(IQ range)
14 (11–17)
CT brain scans (
n
=
23)
Early signs of infarction,
n
(%)
16 (70)
Hyperdense MCA,
n
(%)
7 (30)
Table 4. Comparison of SITS-SICH scores and risk of SICH by
ECASS II definition for GSH and SITS-MOST validation cohorts
Score
Total GSH
cohort
(
n
=
41)
% (
n
)
Total SITS
cohort
(
n
=
15 814)
% (
n
)
SICH
rate
(GSH)
%
SICH
rate
(SITS)
%
Low (0–2 points)
29 (12 /41)
22.7
0
1.6
Average (3–5 points)
53.7 (22/41)
55
9
4.7
Moderate (6–8 points) 17.1 (7/41)
21.4
0
8.9
High (
>
9 points)
0
1.1
0
23.2
Overall rate
4.6 (2/41)
5.1
Table 5. Comparison of SEDAN scores and risk of SICH by
ECASS II definition for GSH and SEDAN validation cohorts
Score
Total GSH
cohort
(
n
=
23)
% (
n
)
Total SEDAN
cohort
%
SICH rate
(GSH)
(
n
=
2)
% (
n
)
SICH rate
(SEDAN)*
%
0
4.4 (1)
12.4
0
0.9
1
30.4 (7)
27.5
0
3.5
2
34.8 (8)
28.3
0
5.1
3
26.1 (6)
20.9
16.7 (1)
9.2
4
0 (0)
8.6
0
16
5
4.4 (1)
2.2
0
27
6
0
0
0
0