This is the old United Nations University website. Visit the new site at http://unu.edu
Nazmul Hassan and Kamaluddin Ahmad
Institute of Nutrition and Food Science, University of Dhaka,
Dhaka, Bangladesh
INTRODUCTION
Nutritional status, usually measured by anthropometry, is influenced by a large number of factors, one of which, of course, is food consumption (1). Studies showing effects of dietary supplementation on growth in energydeficient populations further support this hypothesis (2, 3). When a child's intake of food falls below the standard allowance, growth slows, and if low levels of intake persist, adult stature will be reduced. Similarly, if adults fail to meet their food requirements they lose weight. According to the malnutrition causality model this may lead to diminished ability to resist infection, to work, and to enjoy the normal satisfactions of life (4,5). This therefore implies that if one has an adequate food intake nutritional status will be satisfactory. The present study is an attempt to establish such a relationship empirically.
An increment in food intake is always associated with a concomitant increment in energy intake. Food intake therefore is quite indicative of energy intake. The present study expresses food intake in terms of energy. The anthropometric indices used include weight, height, skinfold thickness, and arm circumference.
METHODS
Data necessary for the study were collected from 50 per cent of the population in the 19811982 nutrition survey. This survey was conducted in 12 statistically selected rural locations in Bangladesh. The individual food intakes of a population of 2,000 were obtained through an intrafamily food distribution survey. The heights, weights, skinfold thicknesses, and arm circumferences of the sample population were recorded in an anthropometric survey.
RESULTS AND DISCUSSION
Tables 1 to 5 show the relationship between per capita calorie intake and anthropometric measurement for different age and sex groups in the study. The tables reveal that an increase in calorie intake leads to an increase in average height, weight, and arm circumference of children and adolescents in different age groups. However, the rate of increase in the anthropometric indices with increased calorie intake was observed to be greater in the lower age groups. This may be because of the higher growth response to small changes in food intake in younger children.
TABLE 1. Relationship between Per Capita Energy Intake and Anthropometric Measurements in 13YearOld Children
Calorie Intake  Sample Size 
Height (cm) 
Weight (kg) 
Skinfold (mm) 
Arm Circumference (cm) 
Below 400  48  70.3  7.7  7.4  12.6 
400600  15  75.4  8.6  8.6  13.0 
600800  25  75.9  9.0  7.8  13.0 
800 1,000  18  80.9  10.1  8.1  13.8 
1,0001,200  14  83.1  11.5  9.3  13.7 
1,2001,400  6  84.6  10.2  9.8  13.9 
1,400 and above  4  89.1  11.9  8.2  13.8 
Correlation coefficient  0.63  0.57  0.30  0.41 
TABLE 2. Relationship between Per Capita Energy Intake and Anthropometric Measurements in 46YearOld Children
Calorie Intake  Sample Size 
Height (cm) 
Weight (kg) 
Skinfold (mm) 
Arm
Circumference (cm) 
Below 800  18  88.0  11.5  8.0  13.7 
8001,000  23  92.7  12.2  7.9  14.1 
1,0001,200  37  94.0  12.2  7.6  13.6 
1,2001,400  41  97.8  13.5  7.4  14.1 
1,4001,600  15  98.8  14.8  7.5  14.6 
1,6001,800  6  101.2  13.9  7.0  13.8 
1,800 and above  11  102.0  13.4  6.6  14.2 
Correlation coefficient  0.37  0.25  0.21  0.09 
TABLE 3. Relationship between Per Capita Calorie Intake and Anthropometric Measurements in 79YearOld Children
Calorie Intake  Sample Size 
Height (cm) 
Weight (kg) 
Skinfold (mm) 
Arm
Circumference (cm) 
Below 1,000  7  104.0  14.2  6.5  14.1 
1,0001,200  12  111.9  16.6  5.3  14.7 
1,200 1,400  32  110.8  16.9  6.0  14.8 
1,400 1,600  21  113.8  17.1  5.4  14.8 
1,6001,800  16  113.0  18.0  5.7  15.1 
1,8002,000  13  114.4  18.0  6.0  15.2 
2,0002,200  9  119.2  19.8  7.0  15.6 
2,200 and above  11  116.3  18.4  4.8  14.7 
Correlation coefficient  0.32  0.30  0.08  0.16 
Positive correlation was calculated between calorie intake and anthropometric measurements according to different age groups. For the 13year age group the correlation coefficient between calorie intake and height was 0.63, 0.57 between calorie intake and weight, and 0.41 between calorie intake and arm circumference. For the 46year age group the corresponding correlation coefficients were 0.37, 0.25, and 0.09, respectively. The coefficients for 79yearolds were 0.32, 0.30, and 0.16, respectively, for each set of variables. Again, for males 1019 years old the coefficients between calorie intake and height were 0.45, 0.47 between calorie intake and weight, and 0.48 between calorie intake and arm circumference, while for females of the same ages the corresponding coefficients were 0.31, 0.35, and 0.37, respectively.
In order to establish the relationship more rigorously, a regression was run with nutritional status (height, weight, and arm circumference) as dependent variables and calorie intake as an independent variable. Three types of functions  simple linear, semilog, and double log were tried in an effort to explore which type of functional relationship would best explain the changes in nutritional status. Among the fitted functions, the double log model showed a good fit in all anthropometric measurements. R2 in this model was found to be the highest, and the standard error of b coefficients to be the lowest compared to all the other models used. Since the variation in nutritional status was more accurately explained by this model, the subsequent analysis was based on it.
TABLE 4. Relationship between Per Capita Calorie Intake and Anthropometric Measurements in 1019YearOld Male Adolescents
Calorie Intake  Sample Size 
Height (cm) 
Weight (kg) 
Skinfold (mm) 
Arm
Circumference (cm) 
Below 1,600  22  129.0  24.7  5.4  16.8 
1,6001,800  23  132.8  26.9  5.4  17.0 
1,8002,000  23  134.2  26.0  5.0  17.3 
2,0002,200  12  142.9  32.7  4.7  18.4 
2,2002,400  21  137.9  29.0  4.9  17.9 
2,4002,600  11  145.2  32.1  4.9  18.8 
2,6002,800  11  149.1  35.6  5.6  19.6 
2,800 and above  38  148.5  36.6  5.0  20.3 
Correlation coefficient  0.45  0.47  0.05  0.48 
TABLE 5. Relationship between Per Capita Calorie Intake and Anthropometric Measurements in 1019YearOld Female Adolescents
Calorie Intake  Sample Size 
Height (cm) 
Weight (kg) 
Skinfold (mm) 
Arm Circumference (cm) 

Below 1,400  31  132.7  27.4  7.2  17.8  
1,4001,600  27  136.1  29.3  6.0  18.3  
1,6001,800  20  134.9  29.0  7.8  18.6  
1,8002,000  15  1 37.8  30.3  7.2  18.7  
2,0002,200  20  141.1  34.0  7.9  20.1  
2,2002,400  19  142.4  34.2  7.0  19.6  
2,4002,600  7  146.1  38.1  8.8  21.3  
2,600 and above  22  145.0  37.5  8.8  21.2  
Correlation coefficient  0.31  0.35  0.17  0.37 
A double log function may be expressed:
Y = aX^{b} or
logY= log a+b logX
where Y is the various indices of nutritional status (height, weight, and arm circumference) and X is calorie intake.
The regression coefficients, standard error, and the coefficient of determination of the fitted model for the various indices of nutritional status are shown in table 6. It appears that about 58 per cent of the variation in weight,
63 per cent of the variation in height, and 46 per cent of the change in arm circumference were explained by calorie intake. The regression coefficient for all anthropometric indicators studied was found to be positive and highly significant at the 5 per cent level. The b coefficient in weight suggests that an increase of 1 per cent in calorie intake resulted in an increase of about 0.72 per cent in weight. About 0.31 per cent change in height was associated with a unit change in calorie intake, while the corresponding change was 0.25 per cent for arm circumference. The negative value of the constant (a) in the case of weight and a very low positive value for arm circumference could perhaps be attributed to the very low calorie intake of the study population compared to their requirement.
TABLE 6. Regression Coefficient (b) and Coefficient of Determination (R2) for Different Indices of Nutritional Status with Energy Intake
Index  Constant (a) 
b Coefficient  R^{2} 
Weight  0.9184  0.7235*  0.58 
(0.1635)  
Height  1.0959  0.3127*  0.63 
(0.0639)  
Arm circumference  0.4351  0.2562*  0.46 
(0.7303) 
Figures in parentheses indicate standard error.
*Significant at 5 per cent level.
In order to examine the respective effect of protein vs. calories on growth, a multiple regression was run with different indices of nutritional status as dependent variables and energy and protein intake as independent variables. The fitted model then became:
Y = aX_{1} ^{b1} X_{2}^{b2}
or
log Y = log a + b_{1} log X_{1} + b_{2}
log X_{2}
where X_{1} and X_{2} represent energy and protein intakes, respectively. The results are shown in table 7.
The results suggest positive and a much higher contribution of energy to weight, health, and arm circumference compared to protein. A unit change in energy intake was associated with 0.71 per cent change in weight, 0.31 per cent in height, and 0.27 per cent in arm circumference compared to an insignificant contribution of protein to growth. Such a low effect of protein might be because dietary protein is converted to energy, indicating that with the energydeficient diets in Bangladesh it is energy and not protein that determines growth.
Except for the 13year age group, skinfold thickness showed an irregular trend with calorie intake. The correlation coefficient was found to be negative, ranging between  0.05 and  0.21 for different age groups. In the 13yearold children, however, the relationship was the reverse, and the corresponding correlation coefficient was positive (0.30), indicating that skinfold thickness is sensitive to calorie intake in this age group.
However, the same exercise with other age groups in the population, i.e., those above 19 years, showed no regular trend. Height, weight, skinfold thickness, and arm circumference became more or less stable as adolescents moved to adulthood.
Energy is a prime requisite for body function and growth. When the diet is deficient in energy, the body must meet its energy need for obligatory functions by slowing down its growth process. The above study demonstrates that, under the dietary circumstances in Bangladesh, it is the level of energy intake that limits growth.
TABLE 7. Regression Coefficients (b) and Coefficient of Determination (R2) for Different Indices of Nutritional Status with Energy and Protein Intake
Index  Constant (a)  b Coefficients 
R^{2}  
b_{1}(calories)  b_{2}(protein)  
Weight  0.9064  0.7161 *  0.0073 **  0.58 
(0.0655)  (0.0627)  
Height  1.1062  0.3064*  0.0063**  0.63 
(0.0256)  (0.0245)  
Arm circumference  0.4047  0.2749*  0.0186*  0.46 
(0.0292)  (0.0280) 
Figures in parentheses indicate standard error.
* Significant at 5 per cent level.
** Insignificant at 5 per cent level.
ACKNOWLEDGEMENTS
The authors are grateful to the Ford Foundation for its financial assistance in performing this study. They also thank all the members of this Institute for their cooperation.
REFERENCES
1. L. Taylor, "Research Directions in Income Distribution, Nutrition and Economics of Food," Food Res Institute Stud., 16: 29 (1973).
2. R. Martorell, A. Lechtig, C. Yarbrough, H. Delgado, and R. E, Klein, "Energy Intake and Growth in an Energy Deficient Population," Ecol. Food Nutr., 7: 147 (1978).
3. C. Gopalan, M. C. Swaminathan, V. K. K. Kumari, and K. Vijayaraghavan, "Effect of Calorie Supplementation on Growth of Undernourished Children," Amer. J. Clin. Nutr., 26: 563 (1973).
4. R. Karim and F. J. Levinson, "Socioeconomic Constraints in Improving Nutritional Status in Bangladesh," Proceedings of the Third Bangladesh Nutrition Seminar, 1979.
5. N, S. Scrimshaw, C. E, Taylor, and J. E. Gordon, Interactions of Nutrition and Infection, World Health Organization Monograph Series, No. 57 (WHO, Geneva, 1968).