 |
INTRODUCTION |
Obesity is becoming increasingly prevalent in developing countries around the world (Sichieri et al. 1995
). Although improvements in economic conditions potentially explain some of this trend (Popkin and Bisgrove 1988), it is possible that other factors are important also. In particular, we previously suggested (Sawaya et al. 1995) that protein-energy malnutrition, which remains prevalent throughout the developing world (Torun and Chew 1994), may be a significant contributing factor.
One of the first studies providing data on this paradoxical phenomenon was the study of men during the Dutch Famine of World War II, in which an increased incidence of obesity was observed in 19-y-old men whose mothers had suffered food restriction in the first 3 mo of pregnancy (Ravelli et al. 1976). This result was later confirmed in studies in animal models by us and others (Anguita et al. 1993, Jones et al. 1984). Extending the observation that food restriction during pregnancy increases the risk of obesity, we subsequently speculated that food restriction during childhood might have a similar effect (Sawaya et al. 1995). To verify this hypothesis, we conducted a study of the nutritional and socioeconomic conditions of families living in shantytowns in the city of São Paulo (Sawaya et al. 1995). We found, as hypothesized, that there was a significant association between stunting and high weight-for-height in children (Sawaya et al. 1995). This finding has recently been replicated in Russia, China and South Africa (Popkin et al. 1996).
Our finding of a link between malnutrition and obesity led us to conduct a further study to investigate the underlying mechanisms by which childhood malnutrition might promote obesity in later life. This report describes preliminary data from a 22-mo follow-up study on obesity risk factors in normal girls and girls with mild stunting living in shantytowns in Brazil.
| |
SUBJECTS |
A 22-mo prospective study was conducted in the shantytowns of the city of São Paulo, Brazil, in 1994-1995. Thirty school girls (7-11 y old), Tanner stage 1, were selected for the study. Half had mild stunting with height-for-age below 
1.4 Z-score (equivalent to <93% of the median) of the National Center for Health Statistics reference (Hamill 1977) as calculated using the program ANTHRO (Sawaya et al. 1995) but with normal weight-for-height (±1 Z-score or 90-110%); the other girls had normal height-for-age (above ±1 Z-score, or 90-110%) and normal weight-for-height (±1 Z-score, or 90-110%). These cutoffs were used to examine mild nutritional stunting (Torun and Chew 1994), because there were previous indications in our studies of metabolic alterations in malnourished children who are above the currently widespread cutoff for severe stunting of 2 Z-scores. The study protocol and informed consent forms were approved by the Ethics Committee of the Federal University of São Paulo, Brazil. All girls were clinically examined prior to the study, and blood, urine and fecal tests were performed. Those with positive results for anemia, any kind of acute infection, or presence of intestinal parasites were treated and studied after recovery.
View larger version (15K):
[in this window]
[in a new window]
| Fig 1.
Association between height-for-age (percentage of median) and waist/hip ratio in school girls living in shantytowns in the city of São Paulo, Brazil.
|
|
View larger version (15K):
[in this window]
[in a new window]
| Fig 2.
Association between baseline IGF-1 levels and change in height-for-age (percentage of median) during the 22-mo follow-up period, in school girls living in shantytowns in the city of São Paulo, Brazil.
|
|
Experimental design.
The study was conducted over a 8-d period and started with a morning measure of resting metabolic rate (RMR) and anthropometrics, followed by 7 d of weighed home food intake measurements. During this week, a questionnaire was used to collect socioeconomic family and living conditions. At the end of the week, another measurement of RMR was performed. The girls were brought up to the center a third time for blood collection while they were fasting. Anthropometric measurements were then repeated every 3 mo for 22 mo.
Resting metabolic rate.
On study d 1 and 8, the girls were awakened and brought to the center by car at 0630 h, after a 10-h overnight fast. At the center they rested supine for 30 min, and then RMR was measured by indirect calorimetry (Sensormetric, Vacumed Co.) for 30 min. The mean of the two measurements is used in the data analysis.
Anthropometric measurements and body composition.
Following measurements of RMR, a series of anthropometric measurements was performed in triplicate: weight, height, triceps and subscapular skinfolds (using a Lange Skinfold Caliper; Cambridge Scientific Industries, Cambridge, MD), and arm, waist and hip circumferences. The measurements were performed using standard techniques (Jelliffe 1966) while the subjects were fasting. Body fat and fat-free mass were calculated from skinfold thickness as described (Slaughter et al. 1988).
Food intake.
After the measurements of RMR and anthropometry on d 1, the girls returned home, and their food intake was weighed for 7 consecutive days. A trained research worker weighed breakfast, lunch and major day-time snacks, and a literate adult member of the family was trained to weigh dinner and night-time snacks. The energy content and composition of the diet were calculated using a computer program specially designed for Brazilian food (Anção et al. 1993).
Socioeconomic, family and living conditions.
The information about the socioeconomic, family and living conditions was collected during home visits by administration of a questionnaire.
Hormonal assays.
Blood samples were collected for measurement of insulin-like growth factor-1 (IGF-1), which was analyzed using a commercial RIA kit (Nichols Institute Diagnostics, San Juan Capistrano, CA) and other hormones (data not shown). Samples were collected between 0700 and 1000 h after a 10-h overnight fast.
Analysis.
Student's t test (two-tailed) was used for comparisons between groups of weight, height, age, body mass index (BMI), RMR and food intake. The tests of Wilcoxon, Spearman correlation and linear regression were used to compare other differences between the groups (Snedecor and Cochran 1967). Differences were considered significant at P < 0.05. Analysis of covariance was used to determine whether the slope of the relationship between dietary fat and change in weight-for-height differed between the stunted and non-stunted groups.
View larger version (20K):
[in this window]
[in a new window]
| Fig 3.
Association between baseline percentage of dietary energy from fat and change in weight-for-height (percentage of median) during the follow-up period, in stunted and nonstunted school girls living in shantytowns in the city of São Paulo, Brazil.
|
|
View larger version (17K):
[in this window]
[in a new window]
| Fig 4.
Possible model to explain a greater susceptibility of stunted children to weight gain when consuming a high fat diet.
|
|
| |
RESULTS |
Baseline measurements.
Table 1 shows the characteristics of the subjects. Differences in socioeconomic, family and living conditions were found between the two groups of subjects. Girls with mild stunting came from larger families with a lower family income per capita, had a later birth order in the family, higher maternal illiteracy, worse sanitary conditions, and more intestinal parasites prior to study. Girls with mild stunting also had lower height, weight, weight-for-age and height-for-age, but similar BMI, weight-for-height and percentage of fat and fat-free mass. There was an inverse association between height-for-age (expressed as a percentage of median) and waist-to-hip ratio in all subjects combined (Fig. 1). In addition, stunted girls had significantly higher waist-to-hip ratios than non-stunted girls (Table 1). There were no differences between the groups in blood levels of IGF-1. However, both non-stunted controls and stunted girls showed levels of IGF-1 below normal range (normal range for IGF-1: 9.3-24.0 nmol/L). Table 2 summarizes variables for baseline dietary energy intake and expenditure in the two groups. There was no significant difference between the groups for any of these variables.
Follow-up study.
Concerning the follow-up measurements at 22 mo, weight and height deficits in the stunted group were maintained over the period of study and, in addition, there was no difference between the groups in the trends in weight-for-age (data not shown). However, there was a significant association between the increase in height-for-age over the follow-up period and IGF-1 levels at the baseline measurement in all subjects combined (Fig. 2). In other words, girls with lower levels of IGF-1 at baseline had impaired linear growth over the 22-mo follow-up compared with girls with higher levels of IGF-1.
A further finding in the follow-up component of the study was that there was a significant association between dietary fat (percentage of energy intake) and the increase in weight-for-height over the study period in girls with mild stunting but not in non-stunted controls (Fig. 3). However, the difference between the two slopes was not significant, indicating that stunting did not alter the relationship between dietary fat and change in weight-for-height.
| |
DISCUSSION |
Our aim in this study was to examine whether there are identifiable risk factors for obesity associated with mild stunting, which previous research has suggested is a ubiquitous feature of chronic protein-energy malnutrition (Torun and Chew 1994). We found that several well-known socioeconomic risk factors for malnutrition were present in our group with mild stunting. These included a higher number of children per family, lower income per capita, later birth order in the family, higher maternal illiteracy, worse sanitary conditions, and more intestinal parasites. Thus, it is clear that, although both groups belonged to very poor families living in shantytowns, the stunted ones were poorer and that this poverty contributed to their stunting.
One important finding was that there was a significant and positive association between the IGF-1 level measured at baseline and the change in height-for-age during the 22-mo follow-up period. This observation confirms the strong association between IGF-1 levels and linear growth reported in previous studies (Froesch et al. 1985). Concerning the factors underlying the suspected susceptibility to obesity associated with stunting, we found no differences between girls with mild stunting and non-stunted controls in energy intake, energy expenditure or dietary macronutrient composition. The relationship between dietary fat and change in weight-for-height was significant in stunted girls but not control girls, although the slopes of the two relationships were not significantly different. Some evidence from developed countries has suggested that diets with a higher percentage of fat may promote excessive weight gain and adipose tissue accumulation (Klesges et al. 1992). Such findings raise the question of whether the high rate of obesity among recovered-malnourished children may in part be due to an increased susceptibility to high dietary fat intakes and in particular could, if proven, help to explain the recent dramatic increase in obesity in developing countries such as Brazil.
Figure 4 summarizes one possible model to explain why stunting may increase the susceptibility to weight gain during consumption of a high fat diet and at the same time is linked to an increased waist to hip ratio. Low energy intake during growth is known to promote decreased IGF-1 levels and a higher ratio of cortisol to insulin (Waterlow 1992). These hormonal changes lead to low muscle gain and the lower linear growth characteristics of malnutrition (Waterlow 1992). In addition, high cortisol levels are associated with a centralization of body fat such as in Cushing disease and truncal obesity (Pi-Sunyer and Xavier 1994) and thus could potentially explain the increased waist to hip ratio seen in our stunted group. Low IGF-1 levels could also potentially explain the increase in waist to hip ratio, because IGF-1 levels in obese females are reported to be inversely associated with the abdominal sagittal diameter and with visceral adipose tissue (Rasmussen et al. 1994). Finally, it has been suggested that low levels of IGF-1 may impair lipolysis (Boulware et al. 1992, Froesch et al. 1985). Consequently, it is possible that, in the presence of a high fat diet, the hormonal consequences of chronic protein-energy malnutrition may be to promote body fat gain because of a reduced ability to oxidize incoming fat from the diet.
In conclusion, this study raises the question of whether girls who are stunted may be more susceptible to becoming overweight when they consume a high fat diet than non-stunted girls. Hormonal changes resulting from energy restriction may provide one potential explanation for this phenomenon, and further research is needed on this important issue.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Abstract
Introduction
Abstract
