10 / 67
CARDIOVASCULAR JOURNAL OF AFRICA • Vol 24, No 2, March 2013
8
AFRICA
were on aspirin, 29.1% on oral anticoagulation and 34.5% on no
stroke-prevention strategy (Table 3).
Follow-up data six months after the patients’ index
presentations were available for 124 of the 162 patients (76.5%)
recruited to the study. Of these, eight had died, and the cause
of death was sepsis related in three, cardiovascular in two and
unclear in the rest. Fifteen patients had made at least one hospital
revisit, 11 of whom had been re-admitted. Five of these were due
to decompensation of heart failure, four due to major bleeding,
and one each due to an acute coronary syndrome and an acute
ischaemic stroke, respectively. Seven patients had reported an
accidental fall at home and this accounted for two of the four
patients admitted with a major bleed. At the six-month follow
up, 100 of 124 (80.6%) patients were on a rate-control strategy,
while the remaining 24 were on a rhythm-control strategy (19
amiodarone, three flecainide, one propafenone, one sotalol).
Discussion
Epidemiologically, the results of this survey were relatively
similar to a study from Cameroon,
7
with the mean age of patients
being in the mid-sixties. There was a higher preponderance of
males in our study. A dominance of native Africans was noted,
followed by Asian and Caucasian races. However, the majority
of the Asians and Caucasians in our study had been long-term
residents in Kenya.
As in the developed countries, AF prevalence in Africa
increases with increasing age. Over 80% of our patients were
over 50 years of age, and more than 50% of the total study
population were in the over-70-year age bracket. This observation
was also reflected in the data from the Heart of Soweto study,
8
where a steep increase in case presentation was noted after 50
years of age.
Consistent with existing data, the relationship between
hypertension and the development of AF has been traditionally
documented, and is thought to be due to changes in myocyte
ultrastructure and physiology.
10
In our population, despite the
relatively high background prevalence of RHD,
11,12
hypertension
was present in over two-thirds of our AF patients. This supports
the notion of aggressive screening and treatment for hypertension,
as this is singled out as the most common modifiable risk factor
driving this arrhythmia epidemic, even in developing countries.
Uncontrolled hypertension then progresses to hypertensive
heart disease and heart failure, which again was the next most
common predisposition, present in over one-third of our patients.
There are no local guidelines and protocols available for the
management of AF, and treating physicians would decide on a
rate- or rhythm-control approach, based on either their preference
or on international guidelines. Due to the pro-arrhythmogenicity
of anti-arrhythmic agents, their inappropriate usage may account
for an increased mortality rate of AF in our setting.
It was surprising to note that 15% of patients presented
with a TE event as the index presentation of AF. It was even
more surprising that one in every five patients who needed to
be on anticoagulation was on aspirin, an aspirin–clopidogrel
combination or no treatment at all. With a five-fold increased
stroke risk with AF, our data not only serves to highlight the
gravity of the morbidity associated with undiagnosed AF, but
more importantly, inappropriate stroke risk stratification by the
treating physicians.
Use of the CHA
2
DS
2
VASC scoring system
13
may improve the
predictive index of a TE event, especially in the CHADS
2
score
0–1 cohort. However, its non-validation in the native African
population currently limits its utility in our study setting.
Prevalence of the disease in the elderly and the risk of major
bleeding with oral anticoagulation form a sinister recipe for
complications in AF. This was highlighted by accidental falls
reported in 5.6% of patients, and re-admissions for major
bleeding in 3.2% of patients. In an environment with poor
infrastructure and emergency medical services, physician choice
of antiplatelets over oral anticoagulants, even in patients needing
anticoagulation may be understandable.
The study site, being a private teaching hospital, may
have introduced a sampling bias, hence the significantly low
proportion of patients with rheumatic valvular disease in our
cohort. Additionally, performing echocardiography for only
patients who had clinical evidence of RHD may also have
accounted for the low numbers of valvular AF in our study,
despite a significant background prevalence of RHD.
Our study had several other limitations. (1) Due to the
retrospective design, classification of AF subtypes had to be
made on available medical records. For patients presenting in
the last six months of the study, an inadequate length of follow
up prevented appropriate categorisation of these patients. (2) An
echocardiographic evaluation was not available for all patients.
This would have provided supplementary data on the presence
and severity of structural heart disease, and assist in better
characterising both valvular and non-valvular AF. (3) A longer
follow-up period to assess for hard end-point complications of
AF was not possible due to patients returning for medical care
only when severely ill.
Conclusion
ClinicalcharacteristicsofAFinKenyaaresimilartodatafromother
parts of Africa. However, non-valvular AF is more predominant
in our setting, with hypertension, heart failure and diabetes being
the most common associated co-morbidities. A rate-control
modality using digoxin and beta-blockade is the predominant
treatment strategy. Stroke risk assessment and stratification is
sub-optimally performed in Kenya, with a significant proportion
of eligible patients not receiving anticoagulant therapy. Regional
data on risk factors and complications of AF should be pooled to
generate locally tailored guidelines in an attempt to achieve better
adherence to the provision of AF care.
We thank Dr Barbara Karau, Benedict Akoo and the Medical Records
Department of the Aga Khan University Hospital. This study complies
with the Declaration of Helsinki, and the research protocol was approved
by the ethics committee of the Aga Khan University Hospital, Nairobi, and
informed consent of the subjects was obtained.
TABLE 3. STROKE RISK STRATIFICATIONAND MANAGEMENT
CHADS
2
score
0
1
≥
2
Total number
29/156 (18.6) 26/156 (16.7) 101/156 (64.7)
No. on anticoagulation (%)
5 (4.4)
11 (40)
80 (79.8)
No. on ASA (%)
13 (47.8)
7 (40)
17 (17.8)
No. on DAT (%)
0
0
2 (1.6)
No. on no treatment (%)
11 (47.8)
8 (20)
2 (0.8)
ASA: aspirin, DAT: dual antiplatelet therapy.