AFRICA
S29
CVJAFRICA • Volume 26, No 2, H3Africa Supplement, March/April 2015
Hospital, Nigeria.
14
In several studies from the West African
sub-region, it emerged as the leading cause of adult neurological
admissions, constituting up to 65% of such admissions.
15
Furthermore, a steady increase in stroke admissions has been
observed in some institutions that have monitored their stroke
admissions over time. In Tanzania, stroke admissions increased
from 23 per 100 000 in 1935 to 86 per 100 000 in 1962.
16
In Ghana,
the number of stroke patients admitted per year increased from
about 50 in 1960 to 622 in 1993, and the percentage of total adult
medical admissions due to stroke increased from less than 2% in
1960 to about 12% in 1993.
16
Stroke admissions to hospital are clearly rising in Africa.
Although this could be due to increased patronage of orthodox
medicine, increasing stroke incidence in an ageing population
in the throes of epidemiological transition is a more plausible
explanation.
12
Using hospital data, five studies estimated stroke crude
incidence rates ranging from 31/100 000 per year in Harare,
Zimbabwe in 1991
8,17,18
to 149/100 000 per year in Maputo,
Mozambique in 2006
8,17,19
(Table 1). In a meta-analysis by
Adeloye, the pooled estimate of 77.39/100 000 per year (95%
CI
=
51.31–103.48) from hospital-based studies
8
was lower
than from community-based studies. This may suggest that the
available hospital-based African studies underestimated stroke
incidence as a result of exclusion of fatal or mild cases who do
not present in these hospitals.
Stroke incidence, estimated on the basis of representative
community samples with rigorous case ascertainment and
accurate diagnosis over a minimum period of three years,
provides far more information about stroke burden than hospital-
based studies. Nevertheless, such studies require considerable
resources and rigorous methods.
16
There are several community-based incidence studies from
sub-Saharan Africa (Table 1). From the (1973–75) stroke registry
in Ibadan, Nigeria, the crude annual incidence of first-ever stroke
was 26 per 100 000. However, this is likely an underestimate,
because of difficulties with case ascertainment resulting from the
very large population, small study staff, and non-inclusion of
those who patronised traditional healers.
20
InTanzania, stroke incidencewas recorded in twodemographic
surveillance sites: Hai (rural) and Dar-es-Salaam (urban) from
2003–2006. Patients with stroke were identified by the use of a
system of community-based investigators and liaison with local
hospital and medical centre staff. Patients who died from stroke
before recruitment were identified via verbal autopsy, which
might have included non-incident strokes.
16
Overall crude annual
stroke incidence rates were 94.5 per 100 000 in Hai and 107.9
per 100 000 in Dar-es-Salaam (Table 1). When age-standardised
to the WHO world population, annual stroke incidence rates
were 108.6 per 100 000 in Hai and 315.9 per 100 000 in Dar-es-
Salaam.
16
Age-standardised stroke incidence rates in Hai were
similar to those reported in developed countries. However,
age-standardised incidence rates in Dar-es-Salaam were higher
than those published from developed countries. This could be
because of differences in the prevalence of risk factors, which
emphasises the importance of health screening at a community
level.
16
A recent door-to-door survey of every household in Al Quseir
(urban), Egypt
8,17,21
from 2009 to 2012 reported a crude annual
incidence of 181 per 100 000 population but the age-standardised
incidence was not calculated (Table 1). Furthermore, Farghaly
et
al
. performed a door-to-door screening in Al Kharga district,
Egypt,
8,17,22
from 2005 to 2009 and reported a crude annual
incidence of 250 per 100 000 population (Table 1). Although the
age-standardised incidence was likely to be higher than that in
Tanzania (Dar-es-Salaam), which is the global highest,
23
it was
not reported.
Generally, population-based crude incidence rates were higher
than hospital-based rates, ranging from26.0/100 000 person years
in Ibadan, Nigeria in 1979,
8,17,20,24
to 250/100 000 person years in
Al-Kharga, Egypt in 2007
8,17,22
(Table 1). The random-effects
meta-analysis of crude population-based incidence rates was
112.94/100 000 person years (95% CI
=
90.7–135.0).
8
However,
this meta-analysis included incidence studies with incomplete
case ascertainment,
24
conducted over one year rather than the
recommended three-year period.
8,12,16,17
The studies reporting low
rates, therefore, could have been marked by underestimation of
the stroke burden in Africa, and the pooled estimate
8
reported
might therefore be much lower than the true rates.
Crude rates provide valuable information that reflects the
public health burden of stroke, given the age distribution for the
country (i.e. if a specific country has a large number of strokes
because it has a relatively large elderly population, they must
nevertheless care for this larger number of people), whereas
adjusted rates allow a more comparable basis between the risk
of stroke across the life course of residents of the country and
for comparison between countries.
23
Crude rates underestimate
the impact of stroke on a country, particularly when strokes are
occurring at younger ages, as occurs in Africa.
Nevertheless, the annual crude incidence rate in Egypt was
higher than reports by Béjot
et al
. in France (113.5 per 100 000),
Corso
et al.
in Italy (223 per 100 000), Vega
et al
. in Spain (113.5
per 100 000),
22
and Pandian
et al
. in India (119 to 145/100 000).
25,26
Table 2. Population/community-based studies reporting
prevalence of stroke survivors in Africa
Year Country/location/setting Author
Crude prevalence
per 100 000
Age
Overall Male Female
1982 Nigeria, Igbo-Ora, rural
Osuntokun
37
58
–
– All
1985 Tunisia Kelibia, mixed* Atia-
Romdhane
41
42
–
– All
1988 Ethiopia, central Ethiopia,
rural
Tekle
Haimanot
34
15
–
– 20–85
1993 Egypt, Sohag, mixed*
Kandil
31
508 520 490 All
1993 Egypt, Sohag, urban
Kandil
31
410 460 470 All
1993 Egypt, Sohag, rural
Kandil
31
540 510 570 All
1994 Tanzania, Hai, rural
Walker
42
127 155 103 15+
2002 South Africa: Agincourt
Health and Population Unit,
Limpopo province, rural
Connor
43
243 188 296 15+
2006 Nigeria, Lagos, urban
Danesi
38
114 151 69 All
2009 Benin, Cotonou, urban Cossi
44
460 610 360 15+
2009 Egypt, Al-Kharga, mixed* Farghaly
22
560 610 510 All
2009 Egypt, Al-Kharga, urban Farghaly
22
580 620 530 All
2009 Egypt, Al-Kharga, rural
Farghaly
22
520 580 458 All
2010 Tanzania, Hai district, rural Dewhurst
39
2300 2971 1752 70+
2010 Egypt, Assuit, urban
Khedr
35
963 1174 736 All
2013 Egypt, Qena, mixed*
Khedr
40
922 1103 726 All
*Combined rates including both rural and urban communities.