Contents - Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
The relationship between energy intake and diarrhoeal disease in their effects on child growth: biological model, evidence, and implications for public health policy
Chessa K. Lutter, Jean-Pierre Habicht, Juan A. Rivera, and Reynaldo Martorell
The effects of both inadequate energy intake and diarrhoeal disease in the aetiology of childhood malnutrition are well established. Their relative importance is still debated, however, and the resultant uncertainty affects public health policies designed to improve child nutrition. This paper summarizes several earlier publications, and presents the results graphically to show that the same synergistic relationship holds across both urban and rural populations as well as across populations with differing levels of diarrhoea and malnutrition. The results show that the effect of inadequate energy intake on nutrition status depends on the level of diarrhoea, just as the negative effect of diarrhoea on nutrition status depends on the level of energy intake.
The public health implications of these findings are threefold. First, an immediate strategy to improve child nutrition should focus on mechanisms to improve energy intake and dietary quality in general, and in and around the diarrhoeal episode in particular. Second, concurrently and as a long-term strategy, environmental conditions that put young children at risk for diarrhoea have to be improved. Finally, inasmuch as diarrhoea does not have a negative effect on growth among infants receiving the major part of their energy from breast milk, breast-feeding has a special and previously unrecognized role in the relationship between energy intake and diarrhoea as they affect nutrition status. Thus, efforts to promote increased energy intake and dietary quality among infants and young children should be coupled with efforts to promote breastfeeding exclusively through at least the first 4 months of life and partially through 24 months, when children are most at risk for diarrhoea.
Although the effects of both inadequate energy intake and diarrhoeal disease in the aetiology of early childhood malnutrition have been well established, debate about their relative importance affects public health policies designed to improve child nutrition. Understanding the nature of the relationship among energy intake, diarrhoea, and child growth is important for identifying public health interventions most likely to result in improved nutrition status as well as the conditions under which specific interventions are likely to have the greatest benefit.
The positive effect of increased energy intake resulting from nutritional supplementation at 36 months is approximately 2 cm growth in length for Guatemalan and Colombian children [1, 2]. Calculated estimates of the cumulative effect of diarrhoea on length by the age of three years have ranged from 2.5 to 10 cm . The overall effect of energy intake on malnutrition is not independent of that of diarrhoea. however, because diarrhoea also affects intake, energy requirements, and the amount of energy available at the cellular level. Thus, we hypothesized that diarrhoea and inadequate energy intake are biologically interrelated in a synergistic fashion .
In their classic monograph. Scrimshaw et al.  first described the synergistic relationship between malnutrition and infection on nutrition status. They stated: "The simultaneous presence of malnutrition and infection results in an interaction that is more serious for the host than would be expected from the combined effect of the two working independently." Statistically, synergism is described as an interaction that occurs when the effect of one independent variable on the dependent variable depends on the level of another independent variable such that their combined effects are greater than the sum of their individual effects .
A graphic representation of the model proposed by Lutter et al.  is presented here. This model predicts that the biological relationship between the effects of inadequate energy intake and diarrhoea on nutrition status is synergistic.
Effects of the variables
The effect of diarrhoea on energy balance
Diarrhoea, particularly when accompanied by fever, can compromise energy balance through three mechanisms: (1) reduced dietary intake [6-8]; (2) increased faecal loss because of malabsorption of micronutrients and macronutrients , decreased intestinal transit time, and loss of nutrients into the gut [10, 11]; and (3) increased catabolism because of an acceleration in basal metabolic rate.
Each mechanism has a different effect on energy balance. For example, reduced intake leads to a decrease in energy ingested, and increased faecal loss results in reduced availability of ingested nutrients. The third mechanism, increased catabolism, results in increased requirements. Thus, the first two mechanisms cause a reduction in energy available at the cellular level at the same time that the third causes an increase in requirements. The overall effect can lead to an energy deficit during illness and, depending on the aetiology of illness, can extend for some time thereafter.
The effect of energy intake and diarrhoea on nutrition status
Inasmuch as troth inadequate energy intake and diarrhoea affect overall energy balance, the effect of either one on nutrition status is likely to depend on the level of the other. Thus, the positive effect of additional energy intake on child growth increases as the level of diarrhoea increases, and the negative effect of additional diarrhoea increases as the level of energy intake decreases. When nutrition status is not compromised because of inadequate energy intake relative to requirements, diarrhoea should not have a negative effect on growth, as has been observed among well-nourished children in the United States. However, where nutrition status is compromised because of inadequate energy intake relative to requirements, the negative effect of diarrhoea on growth will differ depending on the degree to which nutrition status is compromised.
A graphic three-dimensional representation of these relationships is shown in figure 1 (see FIG. 1. Graphic representation of the relationship between the effect of energy intake and diarrhoea on nutrition status), where nutrition status increases along the Y (vertical) axis, the level of diarrhoea increases along the X (horizontal) axis, and energy intake increases along the Z axis (away from the observer). The incremental effect of energy intake on nutrition status for children without diarrhoea is measured by the slope of the line segment ah relative to the Z axis; this slope is positive. The incremental effect of energy intake on nutrition status for children with the highest level of diarrhoea is measured by the slope of the line segment dc relative to the Z axis; this slope is positive (and greater than ab). The incremental effect of diarrhoea on nutrition status for the lowest level of energy intake is measured by the slope of the line segment ad relative to the X axis; this slope is negative. Finally, the incremental effect of diarrhoea on nutrition status for the highest level of energy intake is measured by the slope of the line segment bc relative to the X axis; this slope is zero in the figure. The synergism is shown by the greater slope of ad compared to bc, which reflects the greater negative incremental effect of diarrhoea for children with low energy intakes relative to those with high energy intakes. The surface abed represents estimated nutrition status for different combinations of energy intake and diarrhoea.
This biological synergism can be modelled for statistical estimation with the following equation:
where Y is nutrition status, D is a measure of the level of diarrhoea, E is a measure of the level of energy intake, a, b, c and d are parameters to be estimated. and e is the error term. The relationship between this equation and figure 1 is seen by noting that the parameter b is the slope of the line ad relative to the X axis, c is the slope of the line cd relative to the Z axis. and d is the difference between ad and bc relative to the X axis. The parameter a denotes the intercept.
Support for the model
Results from three recent publications provide evidence for this underlying biological model.
In the context of evaluating the effect of nutritional supplementation on child growth, we first proposed that the relationship between energy intake and diarrhoea as it related to growth was synergistic: that the pernicious effect of diarrhoea would depend on the energy intake of the child, just as the effect of increased energy intake on growth would depend on the level of diarrhoea . We therefore tested the hypothesis that the positive effect of supplementation (resulting in a significant increase in energy intake) on growth would depend on the level of diarrhoea.
Length and diarrhoeal morbidity were compared at 36 months of age for two cohorts of urban Colombian children: supplemented from birth, and unsupplemented. Energy intakes were 1,329 ± 44 and 1,171 ± 45 kcal/day for unsupplemented children at 18 and 36 months of age respectively. Supplementation resulted in a significant increase of 220 and 253 kcal/ day respectively at these two ages. Although unsupplemented children were ill with diarrhoea a total of 83 days between birth and 36 months of age. compared to 73 days for supplemented children, this difference failed to reach statistical significance.
The effects of energy intake and diarrhoea on attained length at 36 months of age were assessed by linear regression, modelled mathematically as described above, except that energy intake was defined as a dichotomous variable and given a value of l for the children who were supplemented and 0 for those who were not. Diarrhoea was defined as the number of days of diarrhoea between birth and 36 months of age. Analysis of variance was also used to describe the results.
For the same prevalence of diarrhoea, significant differences were seen in the slopes of the regression equation but not in the intercepts for supplemented versus unsupplemented children (see FIG. 2. Effect of nutritional supplementation and number of days ill with diarrhoea on length at 36 months (Developed from data in ref. 3)). Lack of significant difference between the two intercepts indicates that in the absence of diarrhoea there was no difference in attained length between supplemented and unsupplemented children; both supplemented and unsupplemented children who experienced no diarrhoea had a mean length of 87.7 cm at 36 months. The slope for unsupplemented children, however, was significantly different from zero (-0.03 cm per day ill; p<.001), showing that each day with diarrhoea was associated with a reduction of 0.03 cm in attained length at the age of 36 months. The cumulative effect of this deficit for those unsupplemented children with the highest level of diarrhoea was 5 cm (87.7 - 78.7 cm). In contrast, the slope for supplemented children did not differ significantly from zero, showing that diarrhoea had no effect on attained length at 36 months. The mean length of these children was 87.7 cm at 36 months, regardless of the level of diarrhoea.
A dose response in the effectiveness of supplementation by quartile of diarrhoea was also shown by two-way analysis of variance (p < .001 for interaction term; table 1). The difference in length between supplemented and unsupplemented children in the lowest quartile of diarrhoeal disease was small and not statistically significant. The differences became larger and their significance increased in the next three quartiles of diarrhoea, so that the difference in the highest quartile was nearly 5 cm.
TABLE 1. The effect of quartile of diarrhoeal disease and supplementation category on mean length (cm) at 36 months
|Number of days of diarrhoea||Supplemented||Unsupplemented||Difference|
|>60- 105||30||88.6||30||86.0||2.6* *|
Source Ref. 3.
Interaction for two-way analysis of variance model, p <.001. *p<.005 **p<.001. ***p<.0001
Villages in Progreso, Guatemala
Two groups [12, 13] examined growth in length in rural Guatemalan children between 3 and 36 months of age in relation to the percentage of time ill with diarrhoea and supplemental energy from two traditional beverages: one with a high amount of energy (atole) and one with a low amount of energy (fresco). The interval of 3-36 months was used because Guatemalan infants were largely breast-fed until 3 months of age and experienced very little diarrhoea. Thus, the interval was that when the biological effect of increased energy intake (achieved through supplementation) and diarrhoea would be expected to occur.
Children supplemented with atole consumed significantly more supplemental energy than those supplemented with fresco: 124 + 85 versus 16 + 13 kcal/ day. Energy from the home diet did not differ significantly between the two groups: children supplemented with atole consumed 840 + 239 kcal/day at home compared to 889 + 291 kcal/day for those supplemented with fresco. The percentages of time ill with diarrhoea were 8.5 + 7.1 and 8.2 + 7.2 days respectively, which were not significantly different.
The effect of supplement type and diarrhoea, defined as the percentage of time ill, on growth in length between 3 and 36 months of age was examined using the same statistical model described above. Supplement type was described as a dichotomous variable and given the value of 1 for the children who were supplemented with atole and 0 for those supplemented with fresco.
The results of the regression equation showed significant differences in both the intercept and slopes (see FIG. 3. Effect of type of nutritional supplement and percentage of time ill with diarrhoea on incremental length between 3 and 36 months (Developed from data in ref. 13). For children receiving atole, the percentage of time with diarrhoea did not have a statistically significant effect on growth. They grew 29.5 cm, regardless of the level of diarrhoea. For children receiving fresco, however, the percentage of time with diarrhoea was negatively and significantly related to growth. For this group the difference in growth between those with the highest and lowest levels of diarrhoea was 3.7 cm (28.0 - 24.3 cm). However, the differences in intercepts indicate that even at very low levels of diarrhoea the additional energy provided through supplementation positively and significantly affected
The results from the study in Colombia indicate the direction of the interaction between energy intake and diarrhoeal disease on growth, and therefore provide an a priori direction for hypothesis testing in the study from Guatemala. Thus' results in table 2 show the probability values for a one-tailed test, which were calculated from the probability values based on a two-tailed test, which the authors reported.
TABLE 2. The effect of diarrhoeal disease and supplement type on incremental length. 3-36 months
|Variable||b coefficient||p value|
|Supplement type (T)||14.42||.001|
|Interaction (T x D)||0.57||.05|
Source: Ref. 13. growth (table 2). In the absence of diarrhoea, growth for children supplemented with atole was 1.5 cm greater (29.5 - 28.0 cm) than that of children supplemented with fresco.
The effects on weight gain of energy intake, proportion of energy from breast milk, diarrhoeal prevalence, and age were evaluated in poor urban Peruvian infants . Between I and 6 months of age, weight gain was positively associated with energy intake and the proportion of energy from breast milk, but was not associated with prevalence of diarrhoea. In contrast, between 6 and 12 months the interaction between energy intake and prevalence of diarrhoea on weight gain was found significant (p < .(X12). Among infants whose usual energy intake was greater than 75% of recommended amounts, no relationship between diarrhoeal prevalence and weight gain was found. However, among infants whose usual intake was less than 75% of the recommendation, a negative relationship was seen.
Many studies have been undertaken to establish the direction of causality in the relationship between diarrhoea and nutrition status. and these have recently been reviewed . However. the biological model presented in this paper and supported with evidence from Colombia, Guatemala, and Peru argues that this relationship cannot be viewed in isolation from energy intake. The dissimilarities that were found among them result from differences in the amounts of energy ingested as well as the source of energy rather than differences in the underlying biological model.
Although the percentage of time ill with diarrhoea was roughly comparable in both the Colombian and Guatemalan populations, energy intakes differed remarkably. The unsupplemented Colombian children consumed almost twice as much energy as the Guatemalan children (1,329 versus 778 kcal/day), and the net increase from supplementation was approximately 230 kcal/day in Colombia compared to only 124 kcal/ day in Guatemala.
Thus, the finding in Guatemala of a supplement effect even at low levels of diarrhoea among children consuming the high-energy supplement can be attributed to the larger deficit in their energy intakes, such that even in the absence of diarrhoea both supplemented and unsupplemented children had inadequate energy intakes . This was not the case in Colombia, where additional energy from supplementation did not contribute to improved growth in the absence of diarrhoea, because in the absence of the energy cost of diarrhoea, home diets were sufficient to meet energy needs.
The study in Peru provides an example of how the source of energy affects the relationship between energy intake and diarrhoea on nutrition status. Numerous studies have confirmed the fact that breast-milk consumption is unchanged (or may be increased) during diarrhoea [6, 15, 16]. In contrast, estimates of the reduction in energy intake from other food sources range from 15% to 20% [8, 13]. The overall energy cost resulting from reduced food intake during diarrhoea was hypothesized to depend on the proportion of energy coming from breast milk versus other sources . Also. although no data are available on the energy cost of faecal losses among breast-fed infants during diarrhoea, data do show that losses among non-breast-fed infants depend in part on the diet consumed [18, 19]. The overall energy cost of faecal loss during diarrhoea therefore also was hypothesized to depend on the proportion of energy coming from breast milk versus other foods .
Because breast milk represented the major source of energy for Peruvian infants between I and 6 months of age, diarrhoea did not affect their energy balance in the same way as for older infants receiving less of their total energy from that source [ 14]. Among older infants receiving a smaller proportion of their total energy from breast milk, diarrhoea and energy intake had the same interactive effect on growth as was found among Colombian and Guatemalan children.
In Colombia and Guatemala the supplements provided not only additional energy but also protein, vitamins, and minerals. Therefore, the relative importance of energy intake versus that of overall improved dietary quality cannot be determined: both are likely to be important.
In the Peruvian study, the relationship between diarrhoea and growth varied with age and usual energy intake . In populations where breast-feeding is still prevalent, age is important because it serves as a proxy for the percentage of energy coming from breast milk.
Together, the results from these studies show that the relationship between inadequate energy intake and diarrhoea is synergistic and affects nutrition status in a manner far greater than the simple additive effects either one alone would predict. For example, Guatemalan children who consumed high amounts of the high-energy supplement (>100 kcal/day) and had high levels of diarrhoea (ill >10% of the time) grew approximately 3 cm more between 3 and 36 months than those who consumed the low-energy supplement but had comparable levels of diarrhoea. In Colombia, unsupplemented children in the highest quartile of disease were 5 cm shorter than unsupplemented children in the highest quartile of disease at 36 months of age. These differences are far greater than the 2 cm reported as the difference between supplemented versus unsupplemented Colombian and Guatemalan children at 36 months of age in these same studies [ 1 , 2].
The public health implications of these findings arc threefold:
First. an immediate strategy to improve child nutrition should focus on mechanisms to improve energy intake in general, and during and immediately after the diarrhoeal episode in particular. Because of anorexia it may he difficult to increase dietary intake during or immediately after a diarrhoeal episode. Research is needed to determine foods that are more readily accepted during illness and that could be used to maintain an energy source. There may also be a period after an episode when appetite is increased. To the extent that appropriate foods are available to meet this increased appetite, the negative effect of diarrhoea on energy balance may be offset. Research is necessary to determine whether this increase in appetite occurs and, if so, when.
Second, concurrently and as a long-term strategy, environmental conditions that put young children at risk for diarrhoea must be improved. This will include improvements in the infrastructure of water and sanitation, coupled with specific interventions to improve household food preparation and hygiene so as to lessen exposure to pathogens [20, 21]
Third, breast-feeding has a special and previously unrecognized role in the relationship between energy intake and diarrhoea with respect to nutrition status. Its protective effect on risk of diarrhoea has long been recognized; however, it has only recently been demonstrated that when infants who are receiving a major proportion of their energy from breast milk do get diarrhoea, the illness does not have a negative effect on growth . Thus, efforts to promote increased energy intake among infants and young children should by coupled with efforts to promote exclusive breast-feeding through at least the first 4 months of infancy and partial breast-feeding through 24 months, when children arc most at risk for diarrhoea.
The authors gratefully acknowledge the help of Drs. Randall Lutter and Douglas Robson in designing and interpreting the three-dimensional figures.
Contents - Previous - Next