CARDIOVASCULAR JOURNAL OF AFRICA • Vol 21, No 1, January/February 2010
AFRICA
15
level (Table 3). Interestingly, the next best predictor in most of
the groups was
Δ
BMI and smoking status (Table 3). The total
amount of variability in the follow-up lipid levels explained by
the baseline study variables ranged from 31.1 to 44.9%.
Discussion
In this study, we presented tracking coefficients and predictors
of total cholesterol, HDL cholesterol, LDL cholesterol and tri-
glyceride levels in a population-based cohort study over a period
of four years. The percentage of boys initially in the extreme
quartile for TC, LDL cholesterol and triglicerides was 42.5, 54.8
and 40.4%, respectively. Similarly, for girls in the extreme quar-
tile it was 62.7, 53.8 and 38.2%, respectively. The correlation
coefficients for serum lipids ranged from 0.60 to 0.37. The best
predictor of follow-up level for each of the serum lipoprotein
cholesterol fractions was the corresponding baseline level. The
next best predictor in most of the groups was
Δ
BMI and smok-
ing status.
The highest level of stability in the extreme quartile was
54.8% for LDL cholesterol levels in the boys and 62.7% for LDL
cholesterol levels in the girls. Several previous studies among
adolescents have shown the constancy of high lipid and lipo-
protein levels.
23,24
In the CATCH study,
25
Kelder
et al
. revealed
that 54 to 55% of students remained in quintile five for serum
lipid levels four years later. In the Cardiovascular Risk in Young
Finns study,
26
approximately 50% of subjects who initially were
in the extreme quintiles for total cholesterol, LDL cholesterol
and HDL cholesterol levels were in the same quintiles after 12
years. In fact, in the Bogalusa heart study, approximately 50% of
children who had total cholesterol levels above the 75th percen-
tile at baseline still had this 12 years later.
27
The Young Finns
study
28
demonstrated that 60 and 55% of subjects initially in the
extreme quintiles for total cholesterol, LDL and HDL cholesterol
remained there after three and six years, respectively.
The correlations we observed over time are in agreement
with various other studies.
12,28
In the Busselton study,
29
a 27-year
longitudinal study, the age- and survey year-adjusted Pearson’s
correlation coefficients ranged from 0.35 to 0.55, and all were
statistically significant. In the Bogalusa heart study,
27
Larry
et al
.
observed correlation coefficients for total cholesterol levels meas-
ured 12 years apart ranging from 0.38 to 0.66. In the Amsterdam
Growth and Health study,
30
the correlation was significant, with a
coefficient of 0.71 for total cholesterol and 0.51 to 0.65 for HDL
cholesterol levels. Similar results were shown by Porkka
et al
.
26
in 12 years of tracking of serum lipid levels. In fact, correlation
coefficients were higher for TC, LDL-C and HDL-C (0.48–0.59)
than for triglyceride levels (0.33–0.37).
The Cardiovascular Risk in Young Finns study
27
observed
that the initial childhood or adolescent serum lipid value was
the most significant of each adult serum lipid variable. The total
amount of variability in the follow-up lipid levels explained by
the baseline study variables ranged from 4 to 41%.
26
Webber
et
al
. noticed that percentage variability was lower for triglyceride
than for cholesterol levels and ranged from 6 to 41%.
Conclusion
Serum lipids show constancy from childhood and adolescence to
adulthood, and therefore can be used as markers for the poten-
tial risk of developing premature atherosclerosis. Prevention of
coronary heart diseases must begin early on in childhood, and
should be driven by health education towards making healthy
lifestyle choices.
References
World Health Organisation. Global strategy on diet, physical activity and
1.
health: cardiovascular disease: prevention and control. WHO 2003.
EngelandA, Bjorge T, Tverdal A,
2.
et al
. Obesity in adolescence and adult-
hood and the risk of adult mortality.
Epidemiology
2004;
15
: 79–80.
Ghannem H, Fredj A. Transition épidémiologique et facteurs de risque
3.
cardiovasculaires en Tunisie.
Rev Epidemiol Sante Publique
1997;
45
(4):286–292.
Foubert L, Dejagers S, Bruckert E,
4.
et al
. Facteurs lipidiques de risque de
l’athérosclérose: qui traiter, quand et comment?
Ann Endocrinol
1997;
58
: 275–282.
Scandinavian Simvastatin Survival study group. Randomized trial
5.
of cholesterol lowering in 4444 patients with coronary heart disease:
The Scandinavian Simvastatin Survival study (4S).
Lancet
1994;
344
:
1384–1389.
Oren A, Vos LE, Cuno SP,
6.
et al
. Birth weight and carotid intima-media
thickness: new perspectives from the atherosclerosis risk in young adults
(ARYA) study.
J Intern Med
2004;
255
: 236–246.
Sung RYT, Yu CW, Chang SKY. Effect of dietary intervention and
7.
strength training on blood lipid levels in obese children.
Arch Dis Child
2002;
86
: 407–410.
Stavroula K, Meir J, Spiegelman D,
8.
et al
. Distribution of and factors
associated with serum homocysteine levels in children.
J Am Med Assoc
1999;
13
: 1189–1196.
Gidding SS. Cholesterol guidelines debate.
9.
Pediatrics
2001;
107
: 1229.
Reinehr T, Andler W. Change in the atherogenic risk factor profile
10.
according to degree of weight loss.
Arch Dis Child
2004;
89
: 419–422.
Wabitsh M, Hauner H, Heineze E. Body fat distribution and changes in
11.
TABLE 3. PREDICTORS OF SERUM LIPID LEVELS BY GENDERAFTER FOURYEARS OF FOLLOW UP FROM BASELINE
Total cholesterol
HDL-C
Triglycerides
LDL-C
Boys
Baseline total cholesterol
SBP
Smoking
∆
BMI
r
2
=
0.402
Basline HDL-C
Glycaemia
>
6.2 mmol/l
Sedentary lifestyle
r
2
=
0.349
Baseline triglycerides
Smoking
∆
BMI
r
2
=
0.311
Baseline LDL-C
Smoking
∆
BMI
Glycaemia
>
6.2 mmol/l
SBP
DBP
r
2
=
0.391
Girls
Baseline total cholesterol
SBP
Smoking
∆
BMI
r
2
=
0.435
Basline HDL-C
Glycaemia
>
6.2 mmol/l
Sedentary lifestyle
r
2
=
0.416
Baseline triglycerides
Smoking
∆
BMI
r
2
=
0.365
Baseline LDL-C
Smoking
∆
BMI
Glycaemia
>
6.2 mmol/l
SBP
DBP
r
2
=
0.449
