Cardiovascular Journal of Africa: Vol 27 No 6 (November/December 2016)

58 / 76

CARDIOVASCULAR JOURNAL OF AFRICA • Volume 27, No 6, November/December 2016 388 AFRICA
that BDNF may play an important role in cardiometabolic
morbidity.10,13 Therefore, we aimed to investigate associations
between cardiometabolic risk markers (glycated haemoglobin,
blood pressure and silent ischaemic events), cortisol and
cortisol:BDNF ratio in a bi-ethnic cohort.
Methods
This sub-study is part of the SympatheticActivity and ambulatory
Blood Pressure in Africans (SABPA) study carried out in 2008–
2009 and described elsewehere.14 The population consisted of
409 teachers from the Dr Kenneth Kaunda Education District,
South Africa. Selection ensured a socio-economically similar
population despite differences in cultural characteristics.
Exclusion criteria included tympanum temperature above
37.5°C, the use of anti-depressants, α- and β-blockers, and blood
donors or individuals vaccinated within a period of three months
prior to data collection. Additionally we excluded cortisone users
(n = 3), and the final sample comprised 406 individuals.
Participants were fully informed with regard to the study
procedure and signed an informed consent form. The study
was approved by the Ethics Review Board of the North-West
University (NWU-00036-07-S6).
During the 48-hour clinical data-collection process,
ambulatory blood pressure (ABPM), electrocardiogram
(Cardiotens CE120®, Meditech, Budapest, Hungary) and
accelerometer measures were obtained (Actical®, Mini Mitter,
Montreal, Quebec). The BP apparatus was fitted before 09:00,
with an appropriately sized cuff on the non-dominant side of the
participant. The participants were asked to record abnormalities
such as nausea, feeling stressed or having a headache on a
24-hour diary card. The apparatus was pre-programmed to
measure blood pressure every 30 minutes (08:00–22:00) and
hourly (22:00–06:00).
The ABPM and ECG data were analysed using the
CardioVisions 1.19 Personal Edition software (Meditech). An
average 24-hour systolic blood pressure (SBP) of ≥ 130 mmHg
and/or diastolic blood pressure (DBP) of ≥ 80 mmHg were used
as the criteria to define hypertension.16
Silent ischaemia was assessed by two-channel ECG recordings
(Cardiotens CE120®) for 20 seconds at five-minute intervals. An
ischaemic event was defined according to the following criteria:
horizontal or descending ST-segment depression of 1 mm,
duration of ST-segment episode lasting for one minute, and a
one-minute interval from the preceding episode.15
At 16:30, participants were transported to the North-West
University’s Metabolic Unit Research Facility where they were
introduced to the experimental procedures. They completed
psychosocial questionnaires under supervision of a registered
clinical psychologist. They received a standardised dinner and
were advised to go to bed at 22:00 and to fast overnight. At 05:45
they were woken, and the devices were removed after the last
ambulatory recording at 06:00. Anthropometric measurements
and fasting blood samples followed.
The participant’s daily physical activity was monitored over
24 hours, considering resting metabolic rate, with the Actical®
activity monitor (Mini Mitter Co, Inc, Bend, OR; Montreal,
Quebec, Canada). Gamma-glutamyl transferase (γ-GT) and
cotinine levels were used to assess alcohol intake and smoking
habits.
Anthropometric measurements were done in triplicate by level
two-accredited anthropometrists using calibrated instruments
(Precision health scale, A & D Company, Tokyo, Japan; Invicta
Stadiometer IP 1465, Invicta, London UK). Body mass and
height of the participants were measured while remaining in
underwear, for accuracy. Body surface area (BSA) (in m2) was
calculated according to the Mosteller formula. The mean of
three measurements was used to ensure accuracy. Inter- and
intra-observer variability was found to be less than 10%.
Fasting blood samples were obtained from the ante-brachial
vein branches with a winged infusion set using standardised
protocol, and were stored at –80°C until batch assay. Sequential
multiple analysers analysed serum gamma-glutamyl transferase,
high-sensitivity C-reactive protein (hsCRP) (low-grade
inflammation was defined when hsCRP was > 3 mg/l), cotinine
and HbA1c levels (glycated haemoglobin) (Konelab 20i, Thermo
Scientific, Vantaa, Finland; Unicel DXC 800- Beckman and
Coulter®, Germany and the Integra 400, Roche, Switzerland
respectively).
Quantikine colorimetric-sandwich immunoassays from R &
D Systems (catalogue number: DBD00) were used to determine
serum BDNF levels with an intra-assay and inter-assay precision
of 3.8–6.2% and 7.6–11.3%, respectively. Serum cortisol samples
were obtained before 09:00, avoiding the cortisol awakening
responses (CAR), 17 and analysed with ECLIA on Elecsys 2010,
Roche. The cortisol:BDNF ratio was calculated by converting
cortisol to the same SI unit as BDNF (from nmol/ml to pg/ml)
to obtain cortisol:BDNF.
Statistical analysis
Data analysis was done using Statsoft (Statistica V.12). The
Shapiro–Wilks test ascertained normality of data, and skewed
data were log normalised (log physical activity, log cotinine
levels, log γ-GT). Multiple comparisons were not done and a
priori hypotheses for all tests were performed.
Baseline characteristics of the bi-ethnic population were
comparedvia independent t-tests. Chi-squared (χ2) tests computed
proportions and prevalence. The raw data are presented as mean
± standard deviation in the descriptive table (Table 1) to ensure
clarity of clinical observations.
General linear models determined interactions on the main
effects (ethnic × gender) for all cardiometabolic variables
independent of a priori confounders (age, body surface area,
log physical activity, log cotinine and log γ-GT). There after
ANCOVAs, using least-square means, compared bi-ethnic
gender groups while adjusting for a priori confounders.
Pearson and partial correlation analyses determined
unadjusted and adjusted associations between HbA1c level,
24-hour BP, silent ischaemia and cortisol level, as well as
cortsol:BDNF, independent of a priori covariates. Forward
stepwise linear regression analyses determined associations in
several models between dependent variables: HbA1c level, blood
pressure, ischaemia and the independent variables: cortisol,
cortisol:BDNF and a priori covariates in the separate ethnic
gender groups.
Sensitivity analyses: forward stepwise linear regression
analyses were repeated after excluding HIV-positive status
teachers, and hypertension and diabetes medication users.
Significant values were noted as p ≤ 0.05.