30 / 64
S28
AFRICA
CVJAFRICA • Volume 26, No 2, H3Africa Supplement, March/April 2015
urgently needed for the design of appropriate interventions and
the formulation of policy objectives within the framework of T4
(system and policy level) and T5 (global) translational science.
10
The objectives of this systematic review were to examine the
current burden and recent epidemiological trends of stroke in
Africa using available resources (existing epidemiological data
and models) while identifying knowledge gaps; and estimating
the future burden and proposing a responsive and holistic action
plan to control the epidemic. This comprehensive analysis will
include data on incidence, mortality, case fatality, prevalence,
DALYs, quality of life, vascular cognitive impairment, and cost
of care.
Methods
A systematic review of the literature was conducted according to
the Preferred Reporting Items for Systematic Reviews and Meta-
Analyses (PRISMA) guidelines.
11
PubMed database was searched
for ‘Africa’ combined with each of the following keywords:
‘stroke’, ‘cerebrovascular accident’, ‘intracerebral hemorrhage’
and ‘subarachnoid hemorrhage’. Further search was conducted
using combinations of the keywords and sub-Saharan African
countries such as ‘stroke Nigeria’. Other words were also
used in association with the keywords, country names and
Africa. These were ‘epidemiology’, ‘prevalence’, ‘incidence’ and
‘mortality’. Background references and citations were identified
and screened to obtain more articles. Articles were included in
the quantitative synthesis if they had an abstract in English,
were published between January 1960 and October 2014, and
described the epidemiological burden or determinants of stroke
in Africa whether it was original or not.
The search yielded a total of 1 274 articles (Fig. 1). All the
articles were initially screened by one reviewer. We excluded
404 articles that were indexed in both PubMed and AJOL, did
not have abstracts or full text in English, or were not based on
human studies. Two reviewers read the remaining 870 articles in
full to assess their eligibility for the quantitative synthesis. Fig.
1 shows the details of the review selection process. In addition,
data were extracted from Global Burden of Diseases (GBD)
model-derived figures.
Statistical analysis was performed to calculate percentage
change in age-adjusted stroke incidence, mortality and DALYs
for African countries between 1990 and 2010.
Results and Discussion
Incidence
Studies of stroke in Africa are mostly hospital-based case series.
Hospital-based data cannot provide prevalence or incidence
estimates (Tables 1, 2) because the population at risk (i.e. the
denominator) is not known. Moreover, they are also affected
by referral bias. Patients who die quickly from stroke or those
with mild stroke may not be captured.
12
Nevertheless, case series
provide information about the relative frequency of stroke in
comparison to other diseases requiring hospitalisation.
Stroke is the leading cause of medical coma in Nigeria.
13
It is also the leading cause of admissions from hypertension-
related complications, accounting for 40% of hypertensive
complications in the University of Port Harcourt Teaching
PubMed search
1112 publications
Articles were excluded
from the full text review if
they were duplicates
(
n
=
68), abstracts were
not available (
n
=
109),
only abstracts were
available (
n
=
78), they
were not based on human
studies (
n
=
62), or the full
text was not available in
English (
n
=
87)
Articles were excluded
from the quantitative
synthesis if they were not
relevant to populations
living in Africa (
n
=
342) or
did not have any measure
of effect estimate (
n
=
374)
Initial screening
(
n
=
1274)
AJOL search
162 publications
Full-text articles
assessed for
eligibility (
n
=
870)
Studies included
in qualitative
synthesis (
n
=
154)
Fig. 1.
Review selection process. AJOL: African Journals
online.
Table 1. Studies reporting crude incidence of stroke in Africa
Year Country/location/setting Author
Crude incidence per
100 000 per year
Age
Overall Male Female
Hospital-based
1984 Libya, Benghazi, urban Ashok
28
63 69 58 15+
1985 South Africa: Atteridgeville
and Mamelodi, suburban
areas of Pretoria, urban
Rosman
29
101 108 93 20+
1991 Zimbabwe, Harare, urban Matenga
18
31 30 32 All
1993 Libya, Benghazi, urban El Zunni
30
48 52 42 15+
2006 Mozambique, Maputo,
urban
Damasceno
19
149 174 128 15+
Population/community-based
1975 Nigeria, Ibadan, urban Osuntokun
20
26 25 13 All
1993 Egypt, Sohag, mixed*
Kandil
31
180 100 85 All
1993 Egypt, Sohag, urban
Kandil
31
150 90 53 All
1993 Egypt, Sohag, rural
Kandil
31
210 97 119 All
2006 Tanzania, Hai, rural
Walker
32
95 107 77 All
2006 Tanzania, Dares Salaam,
urban
Walker
32
108 115 100 All
2007 Nigeria, Lagos, urban
Danesi
24
25 28 21 All
2007 Egypt, Al-Kharga, mixed* Farghaly
22
250 270 230 All
2007 Egypt, Al-Kharga, rural
Farghaly
22
230 250 220 All
2007 Egypt, Al-Kharga, urban Farghaly
22
260 280 240 All
2012 Egypt,Al Quseir, urban El Tallawy
21
181 212 150 20+
*Combined rates including both rural and urban communities.