Cardiovascular Journal of Africa: Vol 31 No 6 (November/December 2020)

13 / 64

CARDIOVASCULAR JOURNAL OF AFRICA • Volume 31, No 6, November/December 2020 AFRICA 293
with diabetes (n = 37) at baseline examination excluded] per
one SD increment of log-transformed values of NT-proBNP
adjusted for age, gender, waist circumference and follow-up
time to re-examination. NT-proBNP levels were divided into
quartiles to explore the relationship between NT-proBNP and
prevalent insulin resistance at baseline (adjusted for age and
gender) as well as incident diabetes (adjusted for age, gender,
waist circumference and follow-up time to re-examination). All
analyses were performed using SPSS Windows version 23.0 and a
two-tailed p-value < 0.05 was considered statistically significant.
Results
Baseline characteristics of the differences betweenCaucasians and
Africans are listed in Table 1. Africans had overall significantly
worse metabolic status at baseline compared to Caucasians, with
higher systolic and diastolic BP, BMI, and blood glucose, insulin,
HOMA-IR and TG levels, and more prevalent diabetes cases.
NT-proBNP levels were significantly higher in women compared
to men (56.7 and 35.7 ng/l, respectively), however, there were
no significant differences between Caucasians and Africans
with regard to age, gender or NT-proBNP levels at baseline
examination (Table 1). No significant differences between black
and white women were observed (p = 0.861).
NT-proBNP and cross-sectional association with continuous
metabolic parameters: baseline characteristics of the study
samples are listed in Table 1. In cross-sectional linear regression
analyses at baseline, each one SD increase in baseline values
of NT-proBNP was significantly and inversely associated with
body weight (β –2.23; p = 0.042), BMI (β –1.01; p = 0.007), waist
circumference ( β –1.82; p = 0.033), HbA 1c (β –0.14; p = 0.009),
insulin (β –1.66; p = 0.002), HOMA-IR (β –0.47; p = 0.006) and
TG (β –0.04; p = 0.002) (Table 2).
NT-proBNP and cross-sectional associations with
dichotomous metabolic parameters: in cross-sectional gender-
and age-adjusted analyses at baseline, each one SD increment
of NT-proBNP was associated with reduced odds of prevalent
overweight (Caucasians: OR: 0.79; 95% CI: 0.62–1.00; p = 0.045
and Africans: OR: 0.72; 95% CI: 0.57–0.92; p = 0.009), IGT
(glucose: OR: 0.77; 95% CI: 0.60–0.99; p = 0.040 and HbA1c: OR:
0.78; 95% CI: 0.62–0.99; p = 0.038), prevalent MetS (OR: 0.76;
95% CI: 0.60–0.96; p = 0.040), hypertriglyceridaemia (OR: 0.64;
95% CI: 0.47–0.87; p = 0.004) and insulin resistance (OR: 0.57;
95% CI: 0.43–0.76; p < 0.001) (Table 3).
The relative risk of insulin resistance at baseline decreased
significantly across quartiles of baseline values of NT-proBNP.
Compared with the lowest quartile of NT-proBNP, the OR
(95% CI) for prevalent IR in subjects belonging to quartiles two,
three and four was 0.83 (0.44–1.57), 0.30 (0.14–0.63) and 0.25
Table 1. Characteristics of the study sample at baseline
Study sample Total Caucasian African p-value
Number 397 203 194
Gender (% women) 50.1 50.7 49.5 0.803
Age (years) 44.7 ± 9.6 45.1 ± 10.9 44.4 ± 8.1 0.467
Systolic BP (mmHg) 132.9 ± 17.7 129.4 ± 15.1 136.6 ± 19.5 < 0.001
Diastolic BP (mmHg) 87.3 ± 12.8 84.1 ± 10.3 90.7 ± 14.2 < 0.001
Weight (kg) 82.9 ± 20.0 84.1 ± 21.2 81.7 ± 18.5 0.234
Body mass index (kg/m2 ) 28.9 ± 6.6 27.6 ± 5.9 30.2 ± 7.1 < 0.001
Waist circumference (cm) 93.5 ± 15.8 93.2 ± 16.1 93.7 ± 15.4 0.717
Triglycerides (mmol/l) 1.05 (0.72–1.53)
1.32 ± 1.06
0.97 (0.69–1.45)
1.21 ± 0.79
1.01 (0.77–1.57)
1.47 ± 1.29
0.023
Plasma glucose (mmol/l) 5.8 (5.4–6.3)
6.2 ±1.8
5.7 (5.4–6.2)
5.9 ± 0.67
5.9 (5.4–6.4)
6.5 ± 2.5
0.002
HbA1c (%) 5.8 ± 0.9 5.5 ± 0.4 6.1 ± 1.2 < 0.001
Insulin (μU/ml) 13.4 ± 9.4 12.1 ± 8.5 14.8 ± 10.2 0.004
HOMA-IR 3.5 ± 3.1 3.2 ± 2.8 3.8 ± 3.4 0.041
NT-proBNP (ng/l) 31.5 (18.1–53.8)
46.3 ± 48.0
34.1 (19.9–55.1)
46.0 ± 47.0
29.9 (16.9–53.6)
46.5 ± 49.2
0.335
Prevalent diabetes, n (%) 11 (2.8) 2 (1.0) 9 (4.6) 0.027
CVD, n (%) 43 (10.8) 24 (11.8) 19 (9.8) 0.517
Hypertension, n (%) 179 (45.1) 63 (31.0) 116 (59.8) < 0.001
Kidney disease, n (%) 9 (2.3%) 5 (2.5) 4 (2.1) 0.789
Values are means (± SD) or median (25–75th interquartile range). AHT = antihyper-
tensive treatment; CVD = cardiovascular disease (coronary events or stroke);
HOMA-IR = homeostatic model assessment of insulin resistance; NT-proBNP =
N-terminal pro-brain natriuretic peptide
Table 2. Associations of one SD increment of NT-proBNP and glucometabolic traits
Whole population Caucasian African
Beta (SE) p-value Beta (SE) p-value Beta (SE) p-value
Weight (kg) –2.23 (1.09) 0.042 –0.75 (1.55) 0.628 –3.29 (1.41) 0.021
BMI (kg/m2) –1.01 (0.37) 0.007 –0.61 (0.50) 0.223 –1.17 (0.50) 0.019
Waist (cm) –1.82 (0.85) 0.033 –1.68 (1.19) 0.157 –1.92 (1.17) 0.101
HbA1c (%) –0.14 (0.05) 0.009 –0.01 (0.01) 0.039 –0.03 (0.01) 0.012
Glucose (mmol/l) –0.01 (0.01) 0.062 –0.02 (0.01) 0.072 –0.03 (0.02) 0.081
Insulin (μU/ml) –1.66 (0.52) 0.002 –0.11 (0.05) 0.019 –0.15 (0.05) 0.001
HOMA-IR –0.47 (0.17) 0.006 –0.57 (0.23) 0.014 –0.45 (0.26) 0.082
TG (mmol/l) –0.04 (0.01) 0.002 –0.07 (0.04) 0.124 –0.13 (0.04) 0.002
Age- and gender-adjusted linear regressions with outcome as a continuous variable. BMI = body mass index; NT-proBNP = N-terminal pro-brain natriuretic peptide;
TG = triglycerides; Mets = metabolic syndrome; HOMA-IR = homeostatic model assessment of insulin resistance.
Table 3. Associations of one SD increment of NT-proBNP and
prevalence of glucometabolic states at baseline
SABPA baseline
Dichotomous variables OR 95% CI p-value
Waist
Caucasian cut-off 0.79 0.62–1.00 0.045
African cut-off 0.72 0.57–0.92 0.009
IGT (glucose > 5.6 mmol/l cut-off) 0.77 0.60–0.99 0.040
TG (> 1.7 mmol/l cut-off) 0.64 0.47–0.87 0.004
Prevalent MetS 0.76 0.60–0.96 0.023
IGT (HbA1c > 5.7%) 0.78 0.62–0.99 0.038
Risk of belonging to HOMA-IR Q4 0.57 0.43–0.76 < 0.001
Logistic regressions are adjusted for age and gender. SD = standard deviation;
NT-proBNP = N-terminal pro-brain natriuretic peptide; TG = triglycerides;
IGT = impaired glucose tolerance; Mets = metabolic syndrome; HOMA-IR
= homeostatic model assessment of insulin resistance; Q4 = upper quartile of
HOMA-IR. Ethnic waist cut-off points are defined according to Alberti et al.19
and Botha et al. 18