42 / 76
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 27, No 6, November/December 2016
372
AFRICA
A younger age at baseline was also found to be associated
with increased weight gain and centralisation of fat mass over
the 5.5-year follow-up period, however this was not independent
of baseline BMI. Women under 25 years of age increased their
weight by an average of 1.2 kg/year, compared to 0.3 kg/year in
those who were older than 25 years. These marked age-related
differences in weight gain have been reported previously, with
increases in body weight being more pronounced in the younger
compared to the older age groups over the same time period.
4,6
These studies reported increases in body weight of 0.93, 0.73,
0.61 and 0.17 kg/year in participants from 21–35, 35–45, 39
and 59 years of age, respectively.
4-6
The CARDIA study, which
included black women of a similar age to our study, showed
that women between the ages of 18 and 20 years at baseline
increased their weight by an average of 1.2 kg/year, compared
to 0.9 kg/year over 10 years in those who were 27–30 years at
baseline.
28
Although several studies have shown an association between
a younger age and body weight gain, none have measured
changes in body composition or body fat distribution over time.
This study also showed a greater increase in trunk fat mass,
and abdominal VAT and SAT areas in the younger age group,
reflecting an increased centralisation of fat mass. However, when
entered into the multiple regression analysis, baseline age was no
longer significant in predicting weight gain and centralisation of
body fat over the follow-up period.
In the regression model, it was also shown that having a child
over the follow-up period was associated with less weight gain
over the 5.5-year follow-up period. This finding is in contrast to
previous research that has reported that child bearing is weight-
promoting.
29
Rosenberg
et al.
14
have shown in a group of black
women that the first child was associated with a 0.4 kg/m
2
larger
increase in BMI compared to those who had a second/additional
child. In other studies from the USA, a higher energy intake
30
and
lower SES
31
increased the risk of poor postpartum weight loss,
while a study from Brazil found that high prepregnancy weight
and higher gestational weight gain
32
both increased obesity risk.
Although the women in this study were of a very low SES
(Table 1), weight gain was found to be lower in women who
gave birth during the study period compared to those who
did not have children. It has been shown that with exclusive
breastfeeding, postpartum weight loss may be improved,
33
and
even though exclusive breastfeeding rates in SA are considered
low (1.4% of infants at six months), breastfeeding as part of
mixed feeding is still popular
33
and may contribute to lower
weight gain postpartum in these women.
Furthermore, recent data from SANHANES-1 reported that
one in three women (32.4%) experience hunger in the urban
informal (peri-urban) environment.
2
Unfortunately, since this
study did not assess breastfeeding rates or household food
insecurity, it is difficult to draw further conclusions as to the
effect of these factors on postpartum weight loss. Nonetheless,
given the poor SES of the study population, it is likely that more
children introduced into the house may promote greater food
insecurity, facilitating higher postpartum weight loss. Notably,
those who did not have children at baseline were significantly
younger than those who already had children, illustrating the
possible co-linearity between parity and age. The younger
women, who were also nulliparous, were also significantly less
active, further confounding the effect of parity on weight gain.
Longitudinal studies to determine risk of future weight gain
in high-income populations often use more static variables of
education, employment or income as proxies for SES. To our
knowledge, there are no longitudinal studies examining the
impact of changes in SES, on body weight or obesity risk, which
may be an important factor to consider in this highly mobile
population.
Although, in this study there was a significant improvement in
SES indicators over the follow-up period, specifically sanitation,
household assets, level of education and rate of employment,
it is still worth noting that less than 50% of the women had
completed secondary level education, were formally employed,
or came from households with running water and a toilet inside.
In spite of the improvements in SES within this population, the
traditional markers of SES, including level of education and
employment, and the changes in these markers over time, were
not independently associated with changes in body weight or
body fat distribution.
However, women who had access to sanitation at baseline,
representing a higher SES, had smaller weight gains over time.
Conversely, those who improved their sanitation (and hence SES)
over the study period had larger gains in body weight. This may
suggest that with improving SES, women may increase their body
weight, whereas if SES is stable, body weight might remain more
stable. In high-income populations, it has been shown that SES
is inversely associated with obesity, with a stronger relationship
in women than men.
12
Conversely, in LMICs, studies have shown
a positive association between SES and BMI,
34,35
while others
have shown an increasing prevalence of obesity in the lower
SES groups.
36
Therefore, sanitation rather than the traditional
measures of SES may better reflect the poor SES in this study.
Although this study did not find any associations with
change in body weight and baseline dietary intake, DQI-I or
physical activity, other large longitudinal studies have shown that
different dietary patterns or physical activity levels have been
associated with weight gain over time.
8,37,38
The most recent data
from SANHANES-1 and other studies also highlight poorer
dietary diversity in the urban informal (peri-urban) environment
compared to the urban formal areas,
2
which may illustrate
the interaction between SES and dietary quality. Therefore,
even though diet was not found to directly influence the body
composition changes in this study, dietary intake is likely to be
influenced by the SES of the women.
Furthermore, although the food frequency questionnaire
used in this study has been validated in black SA women, the
lack of association with changes in body composition may be
due to limitations with reporting of dietary intake and change
over time. Considering the body of evidence from other studies
showing the impact of change in dietary intake on increases in
body weight, this would be a priority to assess further.
Lastly, this study was unable to assess household food
security. Due to its interaction with both SES and dietary
quality, this would be vital to include in future research.
Although most women met the physical activity guidelines (
≥
30 min moderate- to vigorous-intensity physical activity per day,
American College of Sports Medicine), baseline physical activity
was not associated with weight gain. Most of the physical activity
was reported to be for transport, which is of low intensity and
may not confer any reasonable effect on energy balance.
39
As
with dietary intake, the use of subjective measures of physical