41 / 68
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 31, No 2, March/April 2020
AFRICA
93
proportions were compared by the large-sample
z
‑test and
the c
2
statistic, respectively. As HOMA-IR and circulating
concentrations of inflammatory markers were non-normally
distributed, they were transformed to improve on the
distribution (Table 1). Bivariate correlations were assessed from
Pearson’s correlation coefficients. Independent relations were
assessed from linear, logistic or stepwise regression analysis with
appropriate adjustors. Probability values were further adjusted
for non-independence of family members using the method of
maximum likelihood as implemented by the mixed procedure as
defined in the SAS package.
To determine the relative contribution of risk factors to
variations in eGFR, standardised
β
-coefficients were determined.
As DM is so strongly associated with renal dysfunction, in the
primary analysis we assessed relationships in the sample that
excluded those with DM. However, to compare the impact
of metabolic changes with that of DM on renal function at a
community level and to ensure that the exclusion of DM did not
result in a selection bias, we also performed the same analysis in
the whole sample, including those with DM.
Results
Table 2 shows the characteristics of the non-diabetic participants
in the study sample and Table 3 shows the characteristics
of all participants in the study sample. More women than
men participated and a high prevalence of hypertension and
obesity was noted, with many hypertensives being untreated or
uncontrolled; 33.9% of the non-diabetic sample had an eGFR
=
60–90 ml/min/1.73 m
2
and 4.4% an eGFR
=
20–60 ml/min/1.73
m
2
; 34.9% of the full study sample had an eGFR
=
60–90 ml/
min/1.73 m
2
and 6.4% an eGFR
=
20–60 ml/min/1.73 m
2
.
In the non-diabetic cohort, only blood glucose, low-density
lipoprotein (LDL) cholesterol concentrations and metabolic
syndrome features (MDRD eGFR only) showed weak
independent relationships with eGFR. No adiposity index or
other obesity-associated metabolic abnormality (lipids) was
independently associated with eGFRor creatinine concentrations
(Table 4).
However, in multivariate regression models, independent
of conventional risk factors, including metabolic syndrome
Table 1. Distribution of circulating inflammatory markers or the
homeostasis model of insulin resistance (HOMA-IR) before and after
transformation of data (
n
=
1 010)
Parameter
Skewness
Kurtosis
Shapiro
–
Wilk
Untransformed
C-reactive protein
4.22
31.2
0.64
Resistin
2.81
15.8
0.80
HOMA-IR
4.33
26.3
0.57
Logarithm of:
C-reactive protein
–
0.37
0.04
0.99
Resistin
0.12
0.52
1.00
HOMA-IR
0.36
–
0.24
0.98
Table 2. Participant characteristics without diabetes mellitus
Characteristics
All
With 24-hour BP
With PWV
Number (% female)
872 (63.1)
600 (63.0)
779 (61.4)
Age (years)
41.7
±
17.8
41.5
±
17.5
41.2
±
17.9
Body mass index (kg/m
2
)
28.6
±
7.7
28.1
±
7.4
28.0
±
7.2
% overweight/obese
23.4/38.2
24.8/35.2
23.9/35.6
Waist circumference (cm)
88.8
±
16.2 (
n
=
850)
87.9
±
15.8 (
n
=
584)
87.7
±
15.6 (
n
=
762)
% abdominal obesity
40.6
37.3
37.6
Waist:hip ratio
0.82
±
0.10 (
n
=
850)
0.82
±
0.10 (
n
=
584)
0.82
±
0.10 (
n
=
762)
Regular tobacco (% subjects)
15.6
15.5
16.9
Regular alcohol (% subjects)
21.3
21.8
22.1
% females postmenopausal
37.4
35.2
36.4
% hypertension
39.1
37.7
37.9
Current antihypertensive meds (%)
17.1
15.7
16.2
Glucose (mmol/l)
4.72
±
0.86
4.74
±
0.91
4.70
±
0.82
LDL-C (mmol/l)
2.58
±
0.93
2.58
±
0.96
2.57
±
0.94
HDL-C (mmol/l)
1.43
±
0.42
1.43
±
0.41
1.44
±
0.42
Triglycerides (mmol/l)
1.13
±
0.86
1.13
±
0.94
1.12
±
0.87
Metabolic syndrome (%)*
31.9/25.9/11.9/6.6
32.2/24.7/12.3/6.5
32.6/25.4/11.0/5.4
Insulin (
µ
U/ml)
8.05 (3.86 to 15.00)
8.46 (3.94–16.45)
7.87 (3.77–14.70)
HOMA-IR
1.60 (0.75–3.07)
1.74 (0.80–3.59)
1.58 (0.72–2.99)
Resistin (ng/ml)
10.6 (7.6–15.2)
10.5 (7.5–14.6)
10.6 (7.7–15.1)
C-reactive protein (ng/ml)
3.35 (1.27–7.72)
3.06 (1.17–7.55)
3.14 (1.22–7.44)
Office SBP/DBP (mm Hg)
127
±
22/83
±
13
126
±
22/83
±
12
126
±
22/83
±
12
24-hour SBP/DBP (mm Hg)
117
±
14/72
±
10 (
n
=
600)
117
±
14/72
±
10
117
±
14/72
±
10
Aortic PWV (m/s)
5.95
±
2.36 (
n
=
779)
6.06
±
2.36 (
n
=
529)
5.95
±
2.36
Aortic SBP (mm Hg)
119
±
22
118
±
22
118
±
22
Creatinine (
µ
mol/l)
74.0
±
16.8
74.4
±
16.4
73.9
±
16.7
eGFR (MDRD) (ml/min/1.73 m
2
)
96.1
±
25.8
95.6
±
25.4
97.0
±
25.6
eGFR (CKD-EPI) (ml/min/1.73 m
2
)
97.1
±
21.4
96.8
±
21.3
97.9
±
21.4
Data shown are mean
±
SD, median and interquartile range and proportions. PWV, pulse-wave velocity; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-
density lipoprotein cholesterol; HOMA-IR, homeostasis model of insulin resistance; SBP, systolic blood pressure; DBP, diastolic BP; eGFR, estimated glomerular
filtration rate; MDRD, Modification of Diet in Renal Disease equation; CKD-EPI, Chronic Kidney Disease Epidemiology equation. *Metabolic syndrome (%) is
percentage of individuals with one/two/three/four components of the metabolic syndrome.