CARDIOVASCULAR JOURNAL OF AFRICA • Vol 24, No 4, May 2013
AFRICA
111
and mobilise without complication. This is regardless of the
troponin level, which is for prognostic purposes only; exposing
a significant proportion of patients who would be classified
as high risk according to the ESC guidelines to potentially
sub-optimal care according to these guidelines.
1
Despite these clear local best-practice guidelines, very little
is known regarding the demographics, actual management and
referral patterns of patients suffering an NSTEMI in South
Africa and how this influences the outcome of those patients.
This study aimed to determine whether the management of an
NSTEMI differs depending on the hospital to which the patient
presents (patients presenting to secondary hospitals being less
likely to receive early invasive management), and if so, whether
this is a consequence of geographical remoteness or level of care,
and how this influences outcome.
Methods
After obtaining ethical approval, including a waiver of informed
consent from the University of Stellenbosch’s Health Research
Ethics Committee (reference no: N11/09/288), a retrospective
cohort study was conducted looking at adults presenting with
NSTEMIs to TBH, KBH andWH. This was done over a one-year
period from September 2010 to August 2011. Patients presenting
during the first six months of the study time were analysed in
terms of clinical risk profile and in-patient management, and
then subsequent management up to six months post admission.
These hospitals were chosen for their unique similarities and
differences. TBH is situated in Parow, Cape Town and is one
of two academic referral centres in the city. It has 1 310 beds
and provides a tertiary service to about 2.64 million people.
12
In addition it provides a secondary service to the immediate
surrounding areas, this latter group being the subject of this
study. The Division of Cardiology within the Department of
Medicine at TBH manages all ischaemic chest pain and has 28
beds with three full-time cardiologists.
KBH and WH are both secondary hospitals, similar except for
their physical proximity to their tertiary referral centre, namely
TBH. KBH and WH have 282 and 269 beds, respectively, with
84 and 55 of those beds being assigned to the Departments
of Medicine. Both hospitals have two full-time specialist
physicians. Like TBH, KBH is also situated in Parow, 4.6 km
from TBH, while WH is situated in the Boland/Overberg region
of the Western Cape, approximately 94 km (over an hour) away
from TBH.
Patients 18 years and older presenting to the Departments
of Medicine at KBH and WH, and to the secondary service
provided by the Division of Cardiology, Department of Medicine
at TBH with an NSTEMI from September 2010 until February
2011 were included in this retrospective study. NSTEMI was
defined as angina-type chest pain in an unstable pattern,
requiring hospitalisation and associated with elevated troponin
levels (troponin I
≥
1.0
µ
g/l; troponin T
≥
0.1 ng/ml) and no signs
of ST-segment elevation.
1
Patients with the following were excluded: renal failure
(creatinine
>
200
µ
mol/l), patients who developed an NSTEMI
during hospitalisation for a condition other than ACS, including
surgery within two weeks, cerebrovascular accident (CVA),
anaemia (haemoglobin
<
9 g/l), septicaemia (fever and evidence
of systemic infection), warfarin therapy, known high bleeding
risk, life expectancy less than six months, patients referred
from other secondary hospitals for tertiary care. Previously
documented left bundle branch block (LBBB) without new
changes were allowed.
Cases were identified from the records of the National Health
Laboratory Service at KBH, WH and TBH. All recorded positive
cardiac troponin levels from September 2010 to February 2011
were collected. After obtaining permission from the chief
medical superintendent, the original medical records of all these
cases were requested and screened and those identified as having
suffered an NSTEMI without exclusion criteria were included in
the study.
Data were collected anonymously from the medical records
of those cases identified for inclusion. Two data sets were
collected, data during the index admission and data from follow-
up visits over the following six months. Those cases for which
no information was available at six months were included in the
initial data set and documented as lost to follow up for the second
data set (Fig. 1).
Statistical analysis
The statistical analysis was done in conjunction with the
University of Stellenbosch’s Centre for Statistical Consultation.
Data from the two secondary-level hospitals (WH and KBH)
were combined into a single data set, referred to as the secondary
hospitals (SH). This data set was then analysed and compared
with the TBH data. Descriptive statistics and chi-squared
comparisons were done for categorical data. A
p
-value
<
0.05 in
a two-tailed test of proportions was considered significant.
Unless stated otherwise, continuous data is displayed as
mean
±
standard deviation (SD). Analysis of variance was done
on this data and a
p
-value
<
0.05 was considered significant.
Where statistically significant differences in management or
outcome were found between the SH cohort and the TBH cohort,
a sub-analysis was done comparing WH and KBH to ascertain
whether these differences were due to differences in management
between these hospitals.
Results
The baseline characteristics of the two groups were similar
except for more documented dyslipidaemia and prior aspirin use
in the TBH group (Table 1).
The groups were similar in terms of their clinical presentation
(Table 2). In 58% of patients it was their first presentation with
chest pain. There was a large variation in time to presentation
from the onset of pain (mean: 24.21
±
33.75 hours, median:
7 hours). On presentation, patients had a heart rate of 85.73
±
24.85 beats per minute. Cardiac failure was documented in 39%,
with a relatively equal distribution between Killip II, III and IV
failure.
13
Very few patients (5.26%) had a normal ECG, with the
most frequent abnormality being ST-segment depression, seen
in 46%.
Patients presenting to TBH had a significantly higher TIMI
score than those presenting to the SH (
p
=
0.0046). This could
not be accounted for by differences between WH and KBH,
where the TIMI score was 3.412
±
1.064 and 3.615
±
1.134,
respectively (
p
=
0.5587).
11
This difference in risk stratification
was not reflected in the Grace risk score.
14
Most cases were treated with aspirin (87%) and LMWH
