24 / 62
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 31, No 4, July/August 2020
186
AFRICA
hospital in Gaborone, the largest city and capital of Botswana
with a population of about 232 000 inhabitants.
25
The warfarin
clinic runs once weekly and serves about 30 patients per week
(roughly 40% have MHVs).
INR testing and consultations occur on the same day. Warfarin
tablets are available in PMH and also in the peripheral clinics. The
study cohort consisted of patients aged ≥ 18 years who were on
warfarin for at least 30 days and with at least three INR readings.
Ethical approval was obtained from the University of
Botswana, Ministry of Health and wellness, and PMH ethical
review boards. All participants provided informed consent before
their inclusion in the study.
Data were collected between September 2017 and January
2018 from consecutive patients with MHVs at PMH. Through
personal interviews and a review of medical records, patients’
age, gender, residence, occupation, level of education, co-existing
medical conditions and drug history were documented. Other
information was the presence or absence of known risk factors
for thromboembolic events such as hypertension, diabetes
mellitus and human immunodeficiency virus (HIV) infection.
Information on operated valves and dates of replacement were
extracted from patients’ medical records. As most patients
underwent valve replacements outside Botswana, data on the type
(model) of MHVs were unavailable because of inaccessibility of
surgical notes.
Study outcomes were the occurrence of any bleeding and
thromboembolic events since the valve replacement. Due to
different times of follow up in different patients, the rates of
bleeding and thromboembolic complications are presented as
events per 100 patient-years. Major bleeding was defined as overt
bleeding leading to a decrease in the haemoglobin level of at least
2 g/dl or transfusion of at least two units of packed red blood
cells, occurring at a critical site such as intracranial, intraspinal,
intra-ocular, retroperitoneal, intra-articular or pericardial, or
intramuscular with compartment syndrome.
26
Thromboembolic
complications included ischaemic stroke, transient ischaemic
attack, myocardial infarction, pulmonary embolism, deep-vein
thrombosis and systemic embolism.
27
Using recent INR readings, each patient’s time in therapeutic
range (TTR) was calculated to assess the level of anticoagulation
control using the Rosendaal method.
22
TTR is the number of
person-days that each patient stayed within an INR of 2.5 to 3.5,
divided by the total number of person-days on warfarin.
22
We
used INR values from at least two valid intervals separated by 56
days (eight weeks) or less, without an intervening hospitalisation.
Individual patient’s TTRs were used to calculate the overall
TTR for the clinic. A TTR value below 65% is defined as poor
anticoagulation control.
28,30
Statistical analysis
Data were entered and analysed using SPSS for Macintosh,
version 24.0 (IBM Corporation). Continuous variables are
presented as mean with standard deviation (SD) for normally
distributed data, and median with interquartile range (IQR) for
asymmetrical distribution. Categorical and nominal variables are
presented as absolute and relative frequencies (%). Comparisons
of demographic and clinical characteristics between patients
with and without thromboembolic and bleeding complications
were analysed with independent
t
-tests, Mann–Whitney
U-
or
Pearson’s
χ
2
tests. A two-sided
p
-value of < 0.05 was considered
statistically significant.
To assess for independent predictors for bleeding and
thromboembolic complications, a multivariate logistic regression
model was used. All factors with a
p
-value < 0.25 on bivariate
analysis were added to the multivariable model. We report
adjusted odds ratios (ORs), 95% confidence intervals (CIs) and
p
-values.
Results
The study included 142 patients whose mean (SD) age was 42
(12) years (Table 1). The majority of participants were female,
and over two-thirds of the patients were less than 50 years old.
Many participants (56%) resided in Gaborone and the majority
had formal education. About 44.4% of the participants were
unemployed and with no regular source of income. The most
common co-morbidities were hypertension, atrial fibrillation
and HIV (all on antiretroviral therapy). Ninety per cent of
participants had either mitral or aortic valve replacement. The
median (IQR) duration since valve operation was four years
(1.8–10.0). A total of 568 blood INR tests were assessed, with
only 28.1% of them being in the therapeutic range. The median
(IQR) TTR was 29.8% (14.1–51.0) and about 14.8% of the
patients had a TTR ≥ 65%.
Twenty (14.1%) patients reported significant bleeding events,
and the rate of major bleeding was 1.5 per 100 patient-years
(Table 2). Gastrointestinal bleeding was the commonest major
bleeding event.
Thromboembolic events occurred in 32 (22.5%) patients. Overall,
the rate of occurrence of thromboembolic complications was 2.8
per 100 person-years. Of the 32 patients with thromboembolic
events, 25 (78.1%) had stroke/TIA and seven (21.9%) had valve
thrombosis. Hypertension (
p
= 0.451), atrial fibrillation (
p
= 0.879),
HIV (
p
= 0.568) and diabetes (
p
= 0.510) were not associated with
thromboembolic events. Also, there was no gender difference in
bleeding and thromboembolic complications.
Thromboembolic events were more common among people
in Gaborone than those from outside the city (
p
= 0.044).
Patients with a longer duration of warfarin use were more likely
to suffer bleeding and thromboembolic events than those with
a shorter duration. On multivariate analysis, the duration of
warfarin use (OR 1.06, 95% CI: 1.01–1.11) and an increased level
of education (OR 2.25, 95% CI: 1.17–4.33) were independent
predictors of bleeding complications (Table 3).
Discussion
In this study, 14.1 and 22.5% of patients with MHV prostheses
reported major bleeding and thromboembolic complications,
respectively. The rate of major bleeding was 1.5 events per 100
person-years while that of thromboembolic complications was
2.8 events per 100 person-years.
The rate of major bleeding in our cohort is lower than the
rates previously reported in other settings, with event rates as
high as 3.9 per 100 person-years.
12,31
It is, however, difficult to
compare bleeding complication rates across studies because of
the variation of factors such as patient characteristics, study
methods, duration of follow up and the level of anticoagulation
control.