CARDIOVASCULAR JOURNAL OF AFRICA • Vol 21, No 5, September/October 2010
266
AFRICA
could not confirm this observation in South African patients.
Indeed, hypertension and diabetes (the two bivariate predictors of
intermediate and long-term mortality identified in our patients)
are not consistently associated with an adverse cardiac outcome
in the peri-operative literature.
7,14
Diabetes was not significantly
associated with mortality in the only risk index of vascular surgi-
cal patients,
14
and it was not significantly associated with major
cardiovascular complications in the validation cohort of the
Revised Cardiac Risk Index.
7
A meta-analysis of hypertension in peri-operative patients
suggests that it is statistically associated with cardiac morbidity,
although the clinical importance of this finding is more difficult
to quantify.
15
Hypertension is not conventionally used to stratify
cardiac risk in non-cardiac surgery.
16
Similarly, in the Rotterdam
study, hypertension was not identified as a predictor of mortality
at one, five or 10 years following peripheral vascular surgery.
6
Although hypertension and diabetes are either not important
or inconsistently important risk predictors of mortality following
vascular surgery in European and American populations, they
were identified as important predictors of mortality in the South
African National Burden of Disease study.
10
This study identi-
fied three broad risk categories: mortality associated with sexu-
ally transmitted diseases, poverty and a western lifestyle. Risk
factors associated with a western lifestyle included hypertension
(second), tobacco smoking (third), high body mass index (fifth),
high cholesterol (seventh), diabetes (eighth), physical inactivity
(ninth) and low fruit and vegetable intake (tenth).
It is possible that South African vascular surgical patients
returning to a western lifestyle in the community (and the pres-
ence of associated risk factors) may be a more important predic-
tor of intermediate and long-term mortality than the established
clinical risk factors associated with peri-operative cardiac risk,
6,7
which also appear to be important predictors of intermediate and
long-term survival in developed-world patients.
If this is true, this study highlights an important public health
issue for a South African population in epidemiological transi-
tion, where the most important determinants of mortality are
continued exposure to a risk factor (such as hypertension and
diabetes) with little modification of these risk factors through
health surveillance and management. It is likely that to improve
survival in South African vascular surgical patients, a concerted
public health initiative is necessary. Community-based risk-
factor modification, surveillance and therapy should be consid-
ered of paramount importance.
The large number of patients ‘lost to follow up’ is also indica-
tive of a dysfunctional primary healthcare system. Vascular
surgical patients are patients who would benefit from continu-
ing risk-factor modification and surveillance. Indeed, chronic
medical therapy including statins, beta-blockers and angiotensin
converting inhibitors are associated with improved survival
following peripheral vascular surgery.
17
Our study could not confirm the efficacy of cardiac medi-
cal therapy in South African vascular surgical patients because
it was underpowered. Based on an HR of 0.68 for long-term
mortality following peripheral vascular surgery associated with
beta-blocker therapy,
17
a study of over 3 100 patients would be
required with a control event rate of 7.4%.
18
Although, there is increasing evidence that physiological data
may be predictive of intermediate and long-term survival follow-
ing major and intermediate-risk non-cardiac surgery,
19
this study
could not confirm this. The sample size in this study may have
been too small to show this association.
Our study had two limitations. Firstly, it was a retrospective
study, and therefore it is possible that not all the risk predictors
were recorded in the pre-operative medical charts.
Secondly, this study was limited by its sample size. It is likely
that physiological data may still be predictive of intermediate
and long-term survival and that chronic cardiac medication may
improve long-term survival. It is however, the largest study we
are aware of that has attempted to determine predictors of inter-
mediate and long-term mortality in South African patients.
Despite these study limitations, it appears that traditional
public health issues are more predictive of mortality for South
African vascular surgical patients than internationally accepted
peri-operative risk indices of cardiac morbidity and mortality,
which are predictive in developed-world patients. These findings
suggest that risk predictors for mortality are not necessarily the
same in South African patients, when compared with European
andAmerican patients. It is therefore imperative that we continue
to identify clinical predictors in South African patients, as it is
likely that they are different to those published in the interna-
tional literature.
Conclusions
In contrast to developed-world observations, peri-operative
clinical risk indices were not associated with intermediate and
long-term survival in South African vascular surgical patients.
Instead, hypertension was the only predictor of intermediate and
long-term survival retained in the multivariate model, which has
also previously been identified as the second most important
predictor of mortality in the South African National Burden of
Disease study.
10
TABLE 3. SURVIVAL CHARACTERISTICS OF PATIENTS
WITH HYPERTENSIONAND DIABETES
Group
Cases
(
n
)
Events
(
n
)
Censored
(
n
)
Mean survival time in
6-month blocks (95% CI)*
No hypertension 72
2
70
8.73 (8.4–9.1)
Hypertension
211 19
192
8.18 (7.8–8.5)
No diabetes
159
8
151
8.55 (8.2–8.9)
Diabetes
124 13
111
8.04 (7.6–8.5)
*Survival time in 6-month blocks; normal distribution with a skewness
statistic of 0.315.
TABLE 4. PREVALENCE OF CLINICAL RISK FACTORS IN
SOUTHAFRICANAND DUTCHVASCULAR PATIENTS
Clinical risk factor
South African
patients
(
n
=
283)
Rotterdam
study
6
(
n
=
1332)
p
-value
Male gender
182 (64%)
964 (72%)
0.007
Ischaemic heart disease
188 (66%)
601 (45%)
<
0.001
Congestive cardiac failure
6 (2%)
105 (8%)
<
0.001
Cerebrovascular accident
87 (31%)
101 (8%)
<
0.001
Diabetes
124 (44%)
229 (17%)
<
0.001
Hypertension
209 (74%)
609 (46%)
<
0.001
Serum creatinine
>
180
μ
mol.l
-1
15 (6%)
*
67 (5%)
†
0.88
Pre-operative beta-blockers
86 (30%)
335 (25%)
0.07
Pre-operative statins
66 (20%)
257 (19%)
0.17
*
Denominator 263 patients;
†
renal dysfunction.
