CARDIOVASCULAR JOURNAL OF AFRICA • Vol 21, No 6, November/December 2010
AFRICA
319
however, a significant negative correlation between serum CRP
and the number of laps completed by the boys, as well as calcu-
lated VO
2
max, indicating an inverse association between fitness
and serum CRP. In the girls, the association was in the same
direction, but not statistically significant.
Comparison of girls with a body fat percentage above 25%
(over-fat) with lean girls showed the lean group had a median
serum CRP of 0.30 mg/l (interquartile range 0.15–0.9) and
the over-fat group a median of 0.34 mg/l (interquartile range
0.2–1.0) (Table 3). There was therefore no significant difference
in serum CRP levels between the two groups.
Comparison of the data for serum CRP in the boys showed a
similar result, where the boys with higher body fat percentages
(
>
20%) had a median serum CRP value of 0.5 mg/l (interquar-
tile range 0.2–1.9) and the lean boys a median of 0.32 mg/l
(interquartile range 0.15–0.8). The Mann-Whitney
U
-test for
the difference in CRP levels in boys was
z
=
1.39 (
p
=
0.16),
indicating no significant difference, but a trend towards higher
serum CRP concentrations in the boys with higher percentages
of body fat.
Multiple regression analyses were performed with age, ciga-
rette smoking, body fat percentage, waist circumference and
physical activity level as predictor variables and serum CRP
(log transformed) as dependent variable. Children with a serum
CRP level
>
10 mg/l were excluded from these analyses. For the
boys, no variable was significantly associated with serum CRP.
In the girls’ group, age and waist circumference were statistically
significantly associated with serum CRP, although the adjusted
R
2
was small (adjusted
R
2
=
0.08,
p
=
0.007).
Discussion
The purpose of this study was to assess the association between
serum CRP levels, physical activity and body composition in
black adolescents in two township schools. The most active girls
had the lowest serum CRP values, similar to the findings of Isasi
et al
.,
19
with a significant difference between the CRP values
of active girls versus girls with low habitual physical activity.
The girls probably reported their physical activity more accu-
rately than the boys because their physical activity correlated
significantly with their bleep test scores, which measures actual
fitness.
27
The results for the boys, however, are difficult to explain
because the boys in the inactive group had lower serum CRP
values than those in the most active group. A possible reason
may be that most boys were relatively fit and active and that
other factors were more important determinants of their serum
CRP concentrations. Alternatively, the boys may not have report-
ed their physical activity as accurately as the girls, or some of the
boys may recently have become inactive. Another reason may
have been the absence of organised physical activity in the town-
ship, but that the boys participated in active play on an irregular
basis and could keep fit in this way.
The boys’ group on the whole, however, (inactive plus active)
showed trends as expected regarding physical activity and CRP
levels. The fitness level of the boys, measured by number of
laps completed in the bleep test, correlated negatively with body
fat percentage and abdominal skin fold, as well as with serum
CRP levels. If fitness is taken as a measure of habitual physical
activity, this was an indication of an inverse association between
physical activity and serum CRP levels in the boys.
As was expected, physical activity was higher in the boys than
the girls.
28
Studies have shown a negative correlation between
physical activity and body fatness.
1,2
It has been suggested that
decreased physical activity or increasing inactivity is probably
the main factor accounting for the reduction in total energy
expenditure, leading to a positive energy balance and increased
prevalence of obesity.
1,2
Egger and Swinburn
29
concluded that
even incidental activity can increase energy expenditure, and
intensity of activity also plays a role. Such incidental activity
TABLE 2. PEARSON PARTIAL CORRELATIONS FOR BOYS
AND GIRLSADJUSTED FORAGEAND SMOKING
Parameters
Boys
n p
Girls
n
p
BMI and log CRP
0.01 72* NS 0.21 112 0.04
Waist circumference and log CRP
–0.03 69 NS 0.22 112 0.03
Triceps skin fold and log CRP
0.10 69 NS 0.15 112 NS
Sub-scapular skin fold and log CRP 0.16 69 NS 0.20 112 0.05
Supra-spinal skin fold and log CRP 0.12 69 NS 0.28 112 0.004
Abdominal skin fold and log CRP
0.13 66 NS 0.21 112 0.03
Body fat percentage and log CRP
0.16 53 NS 0.14 106 NS
Bleep test laps completed and log CRP –0.30 71 0.012 –0.15 109 NS
Bleep test laps completed and PDPAR
score
0.06 112 NS 0.22 131 0.01
Bleep test laps completed and body fat
percentage
–0.26 81 0.02 –0.33 128
<
0.0001
*Number of subjects differs due to missing data.
TABLE 3. DESCRIPTIVE STATISTICS FOR GIRLSWITH
HIGH BODY FAT PERCENTAGEVERSUS GIRLSWITH
NORMAL BODY FAT PERCENTAGE
Body fat %
>
25%
(
n
=
35)
Body fat %
>
25%
(
n
=
101)
Variable
Mean/
number
Standard
deviation/%
Mean/
number
Standard
deviation/%
Age (years)
15.43
1.24
15.55
1.41
Weight (kg)
#
43.98
5.88
51.80
8.52
Height (cm)
155.6
0.04 155.0
6.67
BMI (kg/m
2
)
#
18.12
2.11
21.53
3.09
Waist circumference (cm)
60.55
4.24
66.28
5.78
WHR
0.75
0.04
0.76
0.05
Triceps skin fold (mm)
#
10.99
3.87
18.11
5.93
Sub-scapular skin fold (mm)
#
9.00
2.95
13.99
5.51
Abdominal skin fold (mm)
#
12.82
4.15
22.09
7.11
Fat %
#
20.60
4.64
32.06
4.74
Muscle mass (kg)
33.57
5.16
34.66
5.28
Tanner stage 1* (14 years)
2
5.7% 0
0
Tanner stage 2*
(14.4
±
0.6, 13–15 years)
5
14.3% 5
4.9%
Tanner stage 3*
(15.2
±
1.2, 14–17 years)
10
28.6% 46
45.5%
Tanner stage 4*
(15.7
±
1.2, 14–18 years)
16
45.7% 43
42.6%
Tanner stage 5*
(16.6
±
1.4, 15–18 years)
2
5.7% 7
6.9%
Median
Interquartile
range
Median
Interquartile
range
Serum CRP (mg/l)
0.30 0.15–0.91 0.34
0.20–1.23
Differentiated on account of fat percentage, girls with fat percentages
>
25%
were classified over-fat, according to Lohman.
28
*Tanner stage based on breast/genital development stage (1
=
no development,
5
=
mature)
21
with age, mean
±
SD (range),
#
significant differences between
groups. BMI: body mass index; CRP: serum C-reactive protein concentration;
WC: waist circumference; TSF: triceps skin fold.
