| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research Communications |
1 From the Department of Nutrition, University of California, Davis; Instituto de Investigación Nutricional, Lima, Perú; and Instituto de Salud del Niño, Lima, Perú.
2 Supported by USAID grant HRN 5600-G-00-3027-00.
3 Reprints not available. Address correspondence to KH Brown, 3150 Meyer Hall, Department of Nutrition, University of California, Davis, CA 95616. E-mail: khbrown{at}ucdavis.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: Our objective was to determine the effects of dietary viscosity and energy density on total daily energy consumption by young, non-breast-fed children.
Design: We measured the amount of food consumed and the duration of meals during 3 substudies, in each of which 3 study diets were offered for 4 consecutive days each in random sequence: high energy density, high viscosity (HD-HV); high energy density, low viscosity (HD-LV); and low energy density, low viscosity (LD-LV). The viscosity and energy density of the unmodified starch-containing HD-HV diet were varied across substudies to determine whether the effect of amylase liquefaction was related to the initial characteristics of the HD-HV diet. The viscosity of the HV diets ranged from 79000 to 568000 mPa•s; energy density of the HD diets ranged from 
4.18 to 4.93 kJ (1.00–1.18 kcal)/g. Viscosity of the LV diets was 3000 mPa•s and the energy density of the LD diets was 2.47 kJ (0.6 kcal)/g.
Results: In each substudy, children consumed more of the LD-LV diet (g•kg body wt-1•d-1) than of the other diets and more of the HD-LV diet than of the HD-HV diet (P < 0.001). Energy consumption from the HD-LV diet was greater than from the other diets (P < 0.001), but the energy intakes from the latter diets were not significantly different.
Conclusion: Amylase liquefaction of HD-HV porridges resulted in increased energy consumption by young children.
Key Words: Viscosity • energy density • amylase • infant feeding • complementary feeding • Peru • weaning food • infants
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In a recent review, Ashworth and Draper (14) suggested that these disparate results may be explained by differences between studies in the viscosity of the starch-containing (untreated) diet and in the age ranges of the children studied. Specifically, viscosity may limit intake only when the consistency of the diet exceeds some particular threshold above which children can no longer readily chew and swallow the diet; this viscosity-limiting effect on intake may occur only in very young children with immature neuromuscular control of chewing and swallowing. Thus, to define the effect of reduction of viscosity in different situations, it is necessary to characterize the consistency of the starch-containing, untreated diet and to restrict the study to young children of a reasonably narrow age range. Moreover, because experimental studies have found that children's energy intakes increase when they consume diets of greater energy density (15), it is critical that comparisons of viscosity be made while keeping dietary energy density constant. In practice, caregivers often reduce the viscosity of porridges by adding water, thereby simultaneously decreasing their energy density and confounding the interpretation of any effect of viscosity reduction on total energy intake.
To assess the independent effects of dietary viscosity and energy density on total daily energy consumption by young children, we completed 3 substudies, in each of which 3 diets were compared: high density, high viscosity (HD-HV); high density, low viscosity (HD-LV); and low density, low viscosity (LD-LV). In the different substudies, the viscosity and energy density of the starch-containing (untreated) HD-HV diets were altered slightly to determine whether the effect of liquefaction depended on the initial characteristics of the HD diet. Liquefaction of the HD-LV diet was achieved by the addition of locally produced amylase-rich flour prepared from germinated maize (G López de Romaña, V Bennett, E Morales, J González, KH Brown, unpublished observations, 1995). When added to porridge, amylase can hydrolyze starch into low-molecular-weight dextrins and maltose, thus creating a less viscous product that retains the energy density of the original preparation. We hypothesized that energy intake from the HD porridges would be greater after liquefaction of the untreated HD-HV preparations. We further expected that the magnitude of this effect of liquefaction would decrease with starch-containing (untreated) HD-HV diets of lower initial viscosity. We also postulated that energy intakes would be greater from HD diets than from LD diets, regardless of viscosity. Finally, we predicted that the duration of meals would be less with the LV diets than with the HV ones.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Procedures
Three different substudies were completed, with 6 children in each. In all substudies the children received each of the 3 assigned study diets during consecutive diet periods that lasted 4 d each. Each of the 6 possible dietary sequences was randomly assigned to different children to avoid any possible confounding effect of the order of the diets. As noted above, diets within all 3 substudies consisted of high-density, high-viscosity (HD-HV); high-density, low-viscosity (HD-LV), and low-density, low-viscosity (LD-LV) porridges.
Study diets
Recipes were developed with wheat flour, pea flour, full-fat dry milk, sugar, and a 1:1 mixture of soybean and cottonseed oils (Table 1
). Amylase-rich flour (ARF), prepared at the Instituto de Investigación Nutricional from germinated Montaña maize, was used to liquefy the diets. The LD-LV diets were prepared by simply adding more water to the HD-HV diets. In substudy 1, energy densities were either 4.93 kJ (1.18 kcal)/g or 2.47 kJ (0.59 kcal)/g in the HD and LD diets, respectively. The viscosity was either 568000 or 2500–3500 mPa•s in the HV and LV diets. In substudy 2, the energy density of the HD-HV and HD-LV diets was decreased to 4.6 kJ (1.1 kcal)/g by adding more water. This resulted in a reduction in viscosity of the HD-HV diet to 175900 mPa•s. Viscosity of the LV diets and energy density of the LD diets remained unchanged. In substudy 3, the percentage of energy provided by fat was increased slightly, as was the ratio of pea flour to wheat flour. The energy density of the HD diets was decreased to 4.18 kJ (1.0 kcal)/g. This set of modifications resulted in a further reduction in the viscosity of the HD-HV diet to 72500–86000 mPa•s. The viscosity of the LV diets for substudy 3 was 2900–3000 mPa•s.
|
Diets were prepared once daily by trained personnel. All ingredients were added before cooking, stirred until blended, and then cooked on low heat for 30 min. The temperature was then increased to 95°C, and the mixture was cooked for an additional 5 min. The porridges were then weighed and water was added, as necessary, to account for any evaporative losses during cooking.
Feeding protocol
To ensure that adequate amounts of the diet could be offered, the amount prepared for each meal was calculated to meet the maximum amount consumed by children of this age group in previous studies (15, 17) plus an additional 50%. To account for any evaporative losses, diets were weighed in serving bowls immediately after cooking, covered with plastic film, and refrigerated. Serving bowls were color coded to ensure that each child received the proper diet in his or her diet sequence.
After reheating the food to 50°C by partially immersing the feeding bowl in a pan of hot water, nursing aides spoon-fed the porridges to each child according to a standardized protocol. The diet was offered until the child refused further intake, and was then offered again after 2 "rest periods" of 2 min each. When the food was refused a third time, the meal was considered complete. Meals were offered 5 times a day, except in substudy 1, in which 2 subjects received 4 meals a day. Adjustment for these latter subjects in the analysis did not alter results. On the basis of the results of previous studies (15, 17), 4 d were considered adequate to achieve a stable intake for each diet. Before beginning the study, the HD-LV diets were provided for 2 d so that the children could become accustomed to the spoon-feeding routine. A multiple vitamin-mineral supplement (Centrum Jr; Lederle Laboratories, Wayne, NJ) was added to the first meal of the day during all days of the study. Plain water was offered freely after the completion of the meal but no other foods were permitted.
Measurements
Preweighed feeding bowls containing the porridges were reweighed after completion of the meal to determine the amount of food consumed. Bibs were weighed before and after feeding, as was any spilled food, to account for possible losses. The duration of each meal was also recorded.
Each child was weighed daily without clothes before the first meal of the day with use of a balance sensitive to 10 g. The children's clinical status was assessed with regard to temperature, respiratory rate, cardiac frequency, presence of vomiting, and number and consistency of stools. Abdominal circumference was measured before the first meal of the day and again in the evening. Those children presenting with intercurrent diarrhea, vomiting, or fever were withdrawn from data collection until 2 d after their return to normal health.
Statistical analysis
Because there were no significant differences in energy intake on days 1–4 of each diet period, all 4 d of each treatment were included in the analysis. The major outcome variables were the total amount of the diet consumed (g/d and g•kg body wt-1•d-1), the total energy intake (kJ•kg body wt-1•d-1), and the mean duration of meals (min/meal). These values were compared within individuals for each dietary treatment by using repeated-measures analysis of variance (ANOVA) followed by Tukey's test for post hoc comparisons of diets. The results were analyzed separately for each of the 3 substudies. Study results were also combined in a repeated-measures ANOVA that included a substudy main effect and substudy-by-diet interaction to assess whether the results differed by substudy. Analyses were done by using SAS for WINDOWS, release 6.12 (SAS Institute, Cary, NC).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. There were no significant differences in the children's baseline characteristics by substudy, nor in the indicators of clinical status by dietary period within substudy.
|
, the children consumed significantly more of the study diet during the LD-LV than during the HD-LV diet period (175.6 ± 49 compared with 136 ± 34 g•kg body wt-1•d-1), and more of the latter diet than of the HD-HV diet (86 ± 23 g•kg body wt-1•d-1) (P < 0.001 for all comparisons). Despite the greater intake of the LD-LV diet, significantly more energy was consumed during the HD-LV diet period (Table 3). Energy intakes from the HD-HV and the LD-LV diets were not significantly different. Significantly less time was required per meal for consumption of the HD-LV (31 ± 12 min/meal) and LD-LV (29 ± 5 min/meal) diets than was required to consume the HD-HV (35 ± 9 min/meal) diet (P < 0.05; Table 4).
|
|
|
, the mean amount consumed in substudy 2 was greater during the LD-LV than during the HD-LV diet period (188 ± 50 compared with 128 ± 34 g•kg body wt-1•d-1); consumption of the latter diet was significantly greater than that of the HD-HV diet (94 ± 38 g•kg body wt-1•d-1) (P < 0.001 for all comparisons). Mean energy consumption (kJ•kg body wt-1•d-1) was significantly greater during the HD-LV diet period than during the HD-HV or LD-LV diet periods (P < 0.001; Table 3). Mean energy intakes during the HD-HV and LD-LV diet periods did not differ significantly. There were no significant differences in the amount of time needed to complete the meal, by diet period (Table 4).
Substudy 3 (viscosity of HD-HV diet = 79000 mPa •s)
In the final substudy, the mean amount consumed during the LD-LV diet period was significantly greater than that during the HD-LV diet period (205 ± 27 compared with 159 ± 27 g•kg body wt-1•d-1); the mean amount of the latter diet consumed was more than that of the HD-HV diet (125 ± 33 g • kg body wt-1 • d-1; P < 0.001; Figure 1). Once again, there were no significant differences in the mean energy consumption (kJ•kg body wt-1•d-1) when the HD-HV and the LD-LV diets were compared, but the HD-LV diet provided significantly more energy than each of the other diets (P < 0.001; Table 3). The average duration of each meal did not differ significantly by diet period (Table 4).
Comparisons among substudies
There were no significant differences among substudies in the mean amount consumed, the mean energy intake, or the mean duration of meals. To assess whether the effect of viscosity reduction on total daily energy consumption varied by substudy, we calculated the ratio of energy intakes during the HD-LV periods to that during the HD-HV periods. As shown in Table 3, the mean ratios tended to decline progressively from 1.72 to 1.29 from substudy 1 to substudy 3, but these differences were not significant, possibly because of the small number of children enrolled in each substudy. There was a negative linear relation between initial weight-for-length z score and total daily energy intake (P = 0.036; Figure 2).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Several factors may modify the effect of dietary characteristics on children's total energy intake. For example, the children's age, neurologic development, nutritional status, and presence of infection as well as feeding practices and the caregivers' feeding behaviors may all independently influence the effect of any particular dietary manipulation. Thus, it is important to consider several methodologic issues that may affect the generalizability of the results of the present study. The children enrolled in the research protocol had a mean height-age (the age at which the child's height would be at the 50th percentile for children of the same sex in the reference population) of 10 mo (range: 4–19 mo), which may be a better indication of their level of physical development than their mean chronologic age of 13 mo (range: 8–17 mo). The results of this study may not necessarily apply to older children. All of the study subjects had been hospitalized for treatment of moderate-to-severe malnutrition or infections before entry into the study. Thus, their energy needs were undoubtedly greater than those of healthy children of the same biological age. The inverse relation between the children's total daily energy intakes and weight-for-height z scores observed in the present study confirmed the results of an earlier report (17). None of the children in the present study was being breast-fed, so it is uncertain whether similar responses to the diets would occur in breast-fed children. The study diets and daily micronutrient supplements were provided in sufficient amounts to ensure that the children could satisfy their theoretical nutrient needs, and they were fed in standardized fashion by trained study aides to remove possible variability due to differences in feeding techniques.
The average energy intake from the HD-HV and LD-LV diets ranged from 430 to 522 kJ•kg body wt-1•d-1 during the different substudies. This range of intake is more than the average energy needs of healthy children of this age, although it is perhaps only marginally adequate for children recovering from malnutrition and infection, such as those enrolled in this study. In each substudy, the children consumed significantly more energy from the HD-LV diets than from the other 2 diets by a factor of 30–70%. Thus, amylase liquefaction of the HD-HV diets promoted increased total daily energy intake. It appeared that there was less of an effect of adding amylase to the HD-HV diets with lower initial viscosity in substudies 2 and 3, but these trends across substudies were not significant. At the observed magnitude of difference in energy intake between the HD-HV and HD-LV diets in the different substudies, a total sample size of 21 children per substudy would have been necessary to provide adequate statistical power to test this hypothesis. Unfortunately, insufficient resources were available to expand the study sample.
The HD-HV diets in the different substudies ranged from a very thick and dough-like preparation to one that was similar in consistency to creamy mashed potatoes. It was possible to ensure uniform preparation of the study diets by controlling the amount of each ingredient and the preparation techniques and by measuring the resulting viscosities. Differences in the viscosities of the baseline diets may account for some of the discrepancies in previously published studies, although this is difficult to ascertain because variability in the laboratory instruments and measurement techniques make it very difficult to compare the viscosity results reported in different studies. The inconsistent findings of prior studies may also have been due to differences in the ages, health conditions, nutritional status, or feeding practices of the study subjects. However, when we compared the results of previous studies in which the effect of viscosity was examined for diets of the same energy density, there were no obvious explanations for the inconsistent results (Table 5).
|
Note that there were no consistent trends in energy intake from the HD-HV diets in the different substudies despite the progressively lower viscosity of the HD-HV diets in substudies 2 and 3. There are several possible explanations for this apparent lack of response to viscosity reduction across substudies. First, unlike the comparisons within each substudy, the comparisons across substudies were done in different children. Thus, interindividual variability may have obscured any differences that occurred, particularly in view of the small sample sizes in each substudy. Second, as viscosity was reduced across substudies, energy density was also reduced. Thus, the 2 sets of changes may have counteracted each other, as discussed above. A third, particularly intriguing possibility is that the effects of amylase liquefaction observed within the individual substudies were not due only to viscosity reduction per se, but to other aspects of amylase treatment. For example, it is conceivable that prior hydrolysis of dietary starch resulted in less undigested carbohydrate reaching the colon, where it can produce discomfort and possibly suppression of appetite (19). Further studies will be necessary to clarify these issues.
A reduction in child feeding times would be an obvious advantage for caregivers with busy work schedules. As observed in previous studies (10, 15), liquefaction of the HD-HV diet in the present study decreased feeding duration, although this effect was restricted to the first substudy only. We cannot explain why this outcome was not observed in all substudies, although it may be related to the viscosity of the unmodified diets.
The potential benefit of viscosity reduction for increasing energy intake and decreasing feeding times of children in developing countries must be balanced against the cost and potential adverse consequences of these dietary manipulations. The material cost of amylase or ARF is small, but the time involved in production of ARF can be considerable if it is prepared in the household. This burden has been a limiting factor in the use of ARF in some settings (20). Additionally, the possible risks of microbial contamination of ARF and transfer of food-borne toxins must be considered (21). For these reasons, use of an industrial source of amylase may prove to be a safer and more acceptable alternative where suitable local production and distribution systems are available. It may also be possible to develop acceptable low-viscosity diets without the need for amylase through judicious use of fats and simple sugars, which provide additional energy but do not increase viscosity. However, these sources of energy also reduce the nutrient density (amount per unit of energy) of the final product, so adequate nutrient sources must be provided to mitigate the potentially adverse effects of fats and sugars on nutrient density. In the present study, adequate nutrient density was achieved by using a micronutrient supplement and increasing the amount of legumes in the study diets.
In summary, we found that amylase liquefaction of high-energy-density, high-viscosity diets increased total daily energy intakes by young children and, in some cases, shortened the duration of their feeding times. It appears that the benefit of liquefaction may diminish when the unmodified baseline diet is of low viscosity. However, additional studies aree needed to determine whether a particular threshold of viscosity exists at which no further benefit of liquefaction can be detected. Moreover, studies are needed in children of different ages to determine the range of ages for which viscosity reduction is most beneficial. Despite the fact that these latter issues remain unresolved, amylase liquefaction of energy-dense, starch-containing diets may provide a useful means of increasing energy intakes by young children in populations with marginally adequate energy consumption. By contrast with the potential advantage of low-viscosity, high-energy-density food preparations for children with low or marginal energy consumption, these products might be considered disadvantageous in populations beset with high rates of childhood obesity.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M M. Islam, M. Khatun, J. M Peerson, T. Ahmed, M A. H Mollah, K. G Dewey, and K. H Brown Effects of energy density and feeding frequency of complementary foods on total daily energy intakes and consumption of breast milk by healthy breastfed Bangladeshi children Am. J. Clinical Nutrition, July 1, 2008; 88(1): 84 - 94. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M M. Islam, J. M Peerson, T. Ahmed, K. G Dewey, and K. H Brown Effects of varied energy density of complementary foods on breast-milk intakes and total energy consumption by healthy, breastfed Bangladeshi children. Am. J. Clinical Nutrition, April 1, 2006; 83(4): 851 - 858. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. S. Mamiro, P. W. Kolsteren, J. H. van Camp, D. A. Roberfroid, S. Tatala, and A. S. Opsomer Processed Complementary Food Does Not Improve Growth or Hemoglobin Status of Rural Tanzanian Infants from 6-12 Months of Age in Kilosa District, Tanzania J. Nutr., May 1, 2004; 134(5): 1084 - 1090. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
