![[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}
|
|
|
|
|
||
*
Department of Pediatrics, University of California, Davis;
Nutrition Research and Development Center, Bogor, Indonesia; and
**
Department of Education, California State University, Sacramento
2To whom correspondence should be addressed. E-mail: epollitt{at}ucdavis.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSDependent and moderating...RESULTSDISCUSSIONLITERATURE CITED |
|---|
KEY WORDS: • micronutrients • development • iron deficiency • supplementary feeding • energy • children
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSDependent and moderating...RESULTSDISCUSSIONLITERATURE CITED |
|---|
) and its early onset and long course (2
,3
). Malnutrition seldom occurs over a few days but builds over months or years; with time and changes in severity, it affects different interrelated psychobiological domains (e.g., cognition, emotion, locomotion) (4
). Briefly, experiential, maturational, and nutritional factors shape the developmental trajectories of the malnourished child.
The developmental nature of the cause and effect process is also observed in the variance of the timing of response to the same factor. For instance, early supplementary (energy) feeding among poorly nourished children will have an earlier effect on emotional regulation (crying) than on vocalizations (5
). Repeated measurement of the same attribute within a domain will allow the detection of the onset of the effect. Clinical trials testing for the effects of supplementary feeding on development are weakened by the inclusion of singular measures of the developmental outcome. The likelihood of failing to reject a null hypothesis when it should be rejected is strengthened by repeated measures of the outcome that allow for the construction of developmental trajectories.
This study compares the effects of a micronutrient intervention with and without a high energy supplement on the developmental trajectories of poorly nourished children from 12- and 18-mo-old cohorts. A specific intent was to determine whether the micronutrient intervention without energy was associated with a decline in mental test performance. The safety of the intervention was further explored by including the height of the children as a moderating variable. The developmental trajectories were calculated quantitatively on the basis of the scores from the Bayley Scales of Infant Development (6
) obtained through seven repeated measurements at fixed intervals over 1 y.
On the basis of the same study, we reported previously (5
) that within the 18-mo cohort, the group that received a energy + micronutrient supplement (E + M)3
had an advantage on the Bayley Scale of Mental Development over groups that received either skim milk + micronutrients (M) or solely skim milk (S). The functional advantage of E + M over S was interpreted as an indication of the salutary benefits of the energy + micronutrient supplement. However, we interpreted the differences between M and S as suggestive evidence of a modest decline in the subjects that received solely the micronutrient supplement. In particular, the higher mental test scores obtained at two test periods by the children that received S compared with those that received M, provided the basis for this tentative interpretation. Further, this interpretation was in agreement with previous information suggesting that micronutrient supplementation in severely malnourished children had adverse effects (7
,8
). The present study further assessed this issue of a developmental decline, and its analysis was restricted to E + M and M. The children that received S were excluded from this analysis because they were tested four rather than seven times as were the children in the EM and M groups (5
). An additional step to test the validity of the interpretation of an adverse effect from the micronutrient supplementation without energy was justified by the nutritional importance currently attributed in international public health to micronutrient interventions (9
).
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSDependent and moderating...RESULTSDISCUSSIONLITERATURE CITED |
|---|
This study results from second wave analyses of a randomized trial of early, daily supplementary feeding for 12 mo in two cohorts of children (12 and 18 mo old). Originally, all subjects were recruited from 24 day care centers (DCC) located on six tea plantations in Pangalengan, West Java. Each DCC included subjects in both the 12- and 18-mo cohorts. Criteria for case inclusion included age and anthropometrics that suggested poor nutrition. The present study focused selectively on the children that received a micronutrient supplement either with a high energy supplement (E + M) or with skim milk (M).
The success of randomization by DCC was tested in between-group comparisons of socioeconomic, community, family, anthropometric, biochemical, behavioral and mental variables. Of 
130 comparisons, only four variables discriminated the groups. This was considered a chance finding. Thus, the children in the 24 DCC enrolled in the study were considered homogeneous across groups.
For the purposes of this study, within cohort, we estimated the median length-for-age for each treatment group (i.e., E + M and M) and classified children as tall or short, relative to the cohort/treatment/length-for-age median. Four classes (2 treatment x 2 sizes) were then formed for each cohort. All subjects had been tested with the same developmental scales a total of 7 times. Children were tested at baseline (Time 1) and every 2 mo through the year-long treatment, with the final testing occurring 12 mo after baseline (Time 7). For a comprehensive description of the methodological and substantive aspects of the study, see (10
).
As noted, the original experimental design was modified for the purposes of this study by the exclusion of the S group and the post-hoc inclusion of length as a moderating variable. This adjustment and the consequent reduction of the number of subjects within cells reduced the sensitivity of the study design, which then allowed for the detection of associations between variables but not for causal judgments. Sample size, however, is not the only determinant of the sensitivity of a research design. Another critical factor is the validity and reliability of the measurement of the outcome variable because the statistics on this variable determine the significance of an effect or a statistical association. The slope of the curve that represents a developmental trajectory among infants and toddlers is a tight approximation to the construct that we intend to test. It shows whether an effect observed during a period of rapid change is continuous.
The research protocol was reviewed and approved by the Ethics Research Committee of the University of California, Davis, and authorized by the Director General of the National Ministry of Health of the Republic of Indonesia. Children were enrolled in the study after informed consent was obtained from their parents.
Dietary supplement.
The supplement was given two times/d, 6 d/wk, at the DCC. The origin of the supplement, a description of their containers and storage procedures, and the modes of distribution are described by Durnin et al. (11
). The diet of the children participating in the study came from four sources: food eaten at home; food provided 6 d/wk at the DCC; supplements provided at the DCC and breast milk. The nutritional intake of the subjects throughout the period of the study was described (11
). Table 1
shows the composition of the two treatments, and the percentage of the recommended daily allowance (12
) provided by each nutrient.
|
Originally, the number of repeated measures obtained and the desired power (P < 0.10) to detect differences between supplements of 0.5 SD of the mental development scores indicated the use of 
15 subjects per group. The inclusion of a height criterion within cohort to also classify the subjects into tall and short groups was a post-hoc decision; therefore, this variable was not part of the calculation of sample size.
Children were invited to enroll in the trial provided they met the following criteria: no signs of chronic disease; length-for-age <1 SD, and weight-for-length between -1 and -2 SD below the median of the reference standards of the World Health Organization/National Center for Health Statistics (WHO/NCHS). There were a total of 75 subjects, with 42 children in the 18-mo cohort and 33 children in the 12-mo cohort. The numbers of subjects in each cohort/treatment/length category group ranged from 8 to 12.
Nutritional indicators.
To compare subjects’ heights and weights with WHO/NCHS reference data, Z-scores were computed for each child using the age and sex appropriate means and SD from the recently revised and updated NCHS database derived from the National Health and Nutrition Examination Survey (NHANES) (13
) (Table 2
). Within the 12-mo cohort, the mean baseline length Z-score ranged from -2.53 (M short) to -1.29 (E + M tall); the baseline weight Z-scores ranged from -3.05 (M short) to -2.18 (E + M tall). Among the children in the older cohort, the pretreatment mean length Z-scores ranged from -3.22 (E + M short) to -1.59 (E + M tall) and the mean weight scores ranged from -3.42 (E + M short) to -2.26 (E + M tall). Consistently, the mean weight Z-scores were lower than the mean length Z-scores. In the 18-mo cohort, there was an improvement in weight in all four groups; this was not the case in three of the four groups in the 12-mo cohort. In most groups, there were no obvious gains in length relative to the reference curves. The effects of the supplements on anthropometrics in the original study were reported by Beckett et al. (14
).
|
). Baseline mean hemoglobin (Hb) for the 12-mo cohort ranged from 107.8 to 123.0 g/L. Post-treatment Hb values were >120 g/L in all groups. Mean Hb concentrations for the 18-mo cohort at baseline were all >110 g/L. Post-treatment Hb means ranged from 119.1 to 125.6 g/L. All post-treatment iron indicators were in the normal range for age, according to NHANES III reference data (15
). A three-way analysis of covariance (ANCOVA; Cohort x Treatment x Length Category) model with baseline to post-treatment change in Hb as the dependent variable and standardized baseline Hb as a covariate showed no main or interactive effects on Hb change. See also (16
).
|
|
Dependent and moderating variables |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSDependent and moderating...RESULTSDISCUSSIONLITERATURE CITED |
|---|
The mental and motor developmental assessments were done in Sundanese with the 1969 version of the Bayley Scales of Infant Development (6
) every 2 mo for 12 mo. Considerable efforts were made to maintain the original instructions and the degree of difficulty of the items that were modified [for more information, see (5
)]. Raw scores, rather than standardized scores, were used because we were interested primarily in testing for group differences and the use of the standardized scores did not help in the interpretation of the findings. In fact, such scores could be misleading for readers who are not familiar with the developmental scales and may attribute normative values based on U.S. populations to standardized scores.
Stability coefficients were calculated for the adjacent evaluations (2-mo interim) yielding six coefficients for each cohort by treatment combination. [Less specific analyses of stability coefficients, predictive validity and sensitivity of the tests have been published; see (17
).] On the mental scale, there was little test-retest stability in the two treatment groups of the 12-mo cohort (less than half of 12 coefficients were significant). Further, intergroup differences in the sizes of the coefficients followed different patterns; the highest coefficients were the first two (Time 1–2; Time 2–3) in the E + M group, and the last two (Time 5–6; Time 6–7) in the M group. In the 18-mo cohort, 10 of the 12 stability coefficients were moderately high, r > 0.30 (18
). There was greater evidence of test-retest stability on the motor scale for both the 12- and 18-mo cohorts, with 9 of the 12 coefficients > 0.30 (data not presented).
Length category.
Within each cohort/treatment group, the median length-for-age at baseline was used to classify children as tall or short. Children with lengths equal to or above the median length for the cohort/treatment group were classified as tall and those with lengths below the median length for the cohort/treatment group were classified as short.
Statistical analysis.
Individual growth curves were computed for each child using the slopes and intercepts, and population curves for each subgroup were computed. Linear and quadratic models were used to determine which curve best fit the data, as indicated by one sample t tests with each of the 3 beta coefficients (Beta 0, Beta 1 and Beta 2).
Pretreatment mental and motor scores were standardized to a mean of zero and variance of one. Three-way ANCOVA models including cohort, supplement and length category with standardized baseline scores as covariates were used to compare the slopes of the seven mental and motor scores. Main effects and interactions that were not significant at P 0.10 were removed so that the best model could be fit to the data. Group differences were considered significant at P < 0.05. All analyses were conducted with SPSS v.6.1.4 (19
).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSDependent and moderating...RESULTSDISCUSSIONLITERATURE CITED |
|---|
The unadjusted scores of the mental test data for both cohorts increased with each assessment (Table 4
). In the model fitting analyses, the second graduate quadratic term (Beta 2) of the slopes for the 12- and 18-mo cohorts were not different from zero, indicating that the quadratic model did not fit the data. Accordingly, the comparative analyses were restricted to the linear model.
|
). The main finding was a Cohort x Length Category x Supplement interaction (P = 0.008) (see adjusted model, Table 5
). When examined by cohort, there were no interactions between length category and supplement for the 12-mo cohort. However, within the 18-mo cohort, ANOVA showed that slopes differed (P = 0.03) (Fig. 1
). Pairwise comparisons indicated that slopes for the short E + M (b = 6.13) and short M (b = 4.68) children differed greatly (P < 0.05); slopes for tall E + M (b = 5.35) and tall M children (b = 5.39) in the 18-mo cohort did not differ. The individual growth curves for the short E + M subjects (Fig. 2,upper panel
) and short M subjects (Fig. 2
, lower panel) in the 18-mo cohort revealed the nature of the changes in unadjusted test scores over time and illustrated the comparatively large changes in interindividual variability within the M short group at the last two testing points.
|
|
|
, lower panel). Motor development.
As with the mental scores, the raw motor scores increased over time for each of the eight groups (Table 7
). The only Beta 2 value that was different from zero was found in the quadratic model of the 18-mo-old children that received the M supplement. Therefore, the linear model was used to test for intergroup differences.
|
). Treatment did not have a main effect or significant interaction with any other variable or covariate.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSDependent and moderating...RESULTSDISCUSSIONLITERATURE CITED |
|---|
) tentative interpretation based on the lower test scores of the children that received the micronutrient supplement without energy compared with those that received only skim milk. The micronutrient supplement without extra energy given over 1 y to 18-mo-old poorly nourished children was associated with a longitudinal decline in mental test performance. The new analysis improved the precision of the results by pointing out that the unfavorable response was restricted to the shortest children within this already disadvantaged group. However, the micronutrient supplement did not in itself affect mental test performance among the shortest of the short. The individual curves of the mental test scores did not all show a downward trend. Further, individual curves of some children changed from comparativelyvery low initial scores to comparatively high scores at the end; however, the incremental change for most children was negligible. The micronutrient supplement without energy did not adversely affect the physical growth measurements among the shortest of the short. In fact, both length and weight improved from the 1st to the 7th evaluation in most children within the 18-mo cohort, independent of treatment or length category. A post-hoc analysis of the difference between the growth changes of the short children that received the E + M and M supplements revealed no differences. These findings on physical growth suggest that among the shortest children, it is unlikely that an increase in the susceptibility to infection mediated the relationship between the micronutrient supplement and the comparatively poor performance in the mental scale.
No adverse effects on the mental scale were detected in the 12-mo cohort. However, no conclusive statements are warranted on this particular domain at this developmental period because of the poor sensitivity of the Bayley scale for this age group. We have reported related psychometric problems previously (17
), but it was pertinent to revisit the problem (Subjects and Methods) because of the additional evidence for concern. The test-retest stability coefficients were calculated to assess whether the construct behind the Bayley Mental Scale remained partly invariant over time. This was not the case in the two groups of the 12-mo cohort as it was with the 18-mo cohort. We found that the assumption of invariance must be rejected for the 12-mo cohort and that the patterns of stability differed between the E + M and the M groups. Our conclusion is the Bayley Mental Development scale is useful to quantify changes in mental development after 18-mo of age, but not at earlier ages.
To our knowledge, there are no other published studies on the developmental effects of micronutrient supplements with and without energy supplements on poorly nourished children. Data from the trial in Pangalengan, Indonesia reported elsewhere (16
) showed that the micronutrient supplement remedied or prevented iron deficiency anemia, independent of whether the supplement was given with or without energy. The change in iron was accompanied by a change in motor activity but did not improve mental scores or the behaviors observed under natural field conditions. An interpretation was that the micronutrient supplement did not reverse the developmental delays that resulted from other nutrient deficiencies (e.g., energy). At issue is that another study recently reported (20
) that despite the presence of malaria, stunting and wasting, iron supplementation improved motor and language development in a group of toddlers in Pemba, Zanzibar (Tanzania). Comorbidity did not seem to prevent the salutary effects of iron repletion in iron-deficient anemic children; there is contradictory evidence from other studies that oral supplementation increases clinical malaria attack rates in endemic populations (21
).
Here we found that the two treatments did not affect scores on the Bayley Scale of Motor Development. However, in a previous paper from the Pangalengan study (22
) that focused on motor milestones, we showed that within the 12-mo cohort, 6 mo after launching the study, 100% of the children that received E + M could walk by themselves and run, whereas only 73% of the children that received M were similarly skillful. The Bayley Motor Scale, however, is more inclusive than the scale of motor milestones. The latter is restricted to the development of gross motor skills that lead to bipedal locomotion, and the Bayley considers numerous fine and gross motor behaviors.
Interactions among the timing of the study, the personal history of the children and the nature of the supplement given changed the contour of the developmental trajectory of a relatively invariant mental attribute. Once it emerged, this change persisted throughout the period of intervention. Only through a follow-up would we be able to learn whether such effects lasted into late childhood and beyond. Quality of life will also determine such future development.
It is possible that the shortest children were the most chronically undernourished and that they lagged in their response to the micronutrient supplement. However, the fact remains that the shortest children that received the E + M supplement did show a favorable response. A second consideration is that the differences in response were manifested over a 12-mo developmental trajectory; it is not a finding restricted to a single post-treatment evaluation. In fact, the pattern of the developmental trajectories constitutes the strongest finding of the study. However, a conclusive statement on causality is not warranted because the body size variable (i.e., length category) was a post-hoc addition to the design and was not randomized.
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
3 Abbreviations used: ANCOVA, analysis of covariance; DCC, day care center; E + M, energy plus micronutrients; M, micronutrients only; NHANES, National Health and Nutrition Examination Survey; S, skim milk only; WHO/NCHS, World Health Organization/National Center for Health Statistics.
Manuscript received 22 March 2002. Initial review completed 23 April 2002. Revision accepted 11 June 2002.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSDependent and moderating...RESULTSDISCUSSIONLITERATURE CITED |
|---|
1. Pinstrup-Anderson, P., Pelletier, D. & Alderman, H. (1995) Child Growth and Nutrition in Developing Countries 1995 Cornell University Press Ithaca, NY. .
2. Cebu Study Team (1991) Underlying and proximate determinants of child health: the Cebu longitudinal health and nutrition study. Am. J. Epidemiol. 133:185-201.
3. Martorell, R., Yarbrough, C., Yarbrough, S. & Klein, R. E. (1980) The impact of ordinary illness on the dietary intakes of malnourished children. Am. J. Clin. Nutr. 33:345-350.
4. Pollitt, E. (2001) The developmental and probabilistic nature of the functional consequences of iron-deficiency anemia in children. J. Nutr. 131:669S-675S.
5. Pollitt, E., Saco-Pollitt, C., Husaini, M. A. & Huang, J. (2000) Effects of an energy and micronutrient supplement on mental development and behavior under natural conditions in undernourished children in Indonesia. Eur. J. Clin. Nutr. 54(Suppl.):S80-S90.
6. Bayley, N. (1969) Bayley Scales of Infant Development 1969 Psychological Corporation New York, NY. .
7. Oppenheimer, S. J., Macfarlane, S.B.J., Moody, J. B., Bunari, O. & Hendrickse, R. G. (1986) Effects of iron prophylaxis on morbidity due to infectious disease: report on clinical studies in Papua, New Guinea. Trans. R. Soc. Trop. Med. Hyg. 80:596-602.[Medline]
8. Doherty, C. P., Kashen Sarkar, M. A., Salim Shakur, M., Ling, S. C., Elton, R. A. & Cutting, W. A. (1998) Zinc and rehabilitation from severe protein-energy malnutrition: higher-dose regimens are associated with increased mortality. Am. J. Clin. Nutr. 68:742-748.[Abstract]
9. UNICEF/UNU/WHO/MI (1999) International Nutrition Foundation and Micronutrient Initiative 1999:36 Boston MA and Ottawa, Canada. .
10. Pollitt, E., Durnin, J.V.G.A., Husaini, M. & Jahari, A. (2000) Effects of an energy and micronutrient supplement on growth and development in undernourished children in Indonesia: methods. Eur. J. Clin. Nutr. 54(Suppl.):S16-S20.
11. Durnin, J.V.G.A., Aitchison, T. C., Beckett, C., Husaini, M. & Pollitt, E. (2000) Nutritional intake of an undernourished infant population receiving an energy and micronutrient supplement in Indonesia. Eur. J. Clin. Nutr. 54(Suppl.):S43-S51.
12. Food and Nutrition Board (1989) Recommended Dietary Allowances 10th ed. 1989:284-285 National Academy Press Washington, DC. .
13. National Center for Health Statistics, Centers for Disease Control (2000) Growth Charts. http://www.cdc.gov/growthcharts/(accessed 4 February 2002).
14. Beckett, C., Durnin, J.V.G.A., Aitchison, T. C. & Pollitt, E. (2000) Effects of an energy and micronutrient supplement on anthropometry in undernourished children in Indonesia. Eur. J. Clin. Nutr. 54(Suppl.):S52-S59.
15. Looker, A. C., Dallman, P. R., Carroll, M. D., Gunter, E. W. & Johnson, C. L. (1997) Prevalence of iron deficiency in the United States. J. Am. Med. Assoc. 277:973-976.[Abstract]
16. Harahap, H., Jahari, A. B., Husaini, M. A., Saco-Pollitt, C. & Pollitt, E. (2000) Effects of an energy and micronutrient supplement on iron deficiency anemia, physical activity and motor and mental development in undernourished children in Indonesia. Eur. J. Clin. Nutr. 54(Suppl.):S114-S119.
17. Pollitt, E. & Triana, N. (1999) Stability, predictive validity, and sensitivity of mental and motor development scales and pre-school cognitive tests among low-income children in developing countries. Food Nutr. Bull. 20:45-52.
18. Cohen, J. (1988) Statistical Power Analysis for the Behavioral Sciences 1988 Lawrence Erlbaum Associates Hillsdale, NJ. .
19. Statistical Package for Social Sciences (1996) Version 6.1.4 1996 SPSS Chicago, IL. .
20. Stoltzfus, R. J., Kvalsvig, J., Chwaya, H., Montresor, A., Albonico, M., Tielsch, J., Savioli, L. & Pollitt, E. (2001) Effects of iron supplementation and antihelminthic treatment on motor and language development of preschool children in Zanzibar: double blind, placebo controlled study. Br. Med. J. 323:1389-1393.
21. Oppenheimer, S. J. (2001) Iron and its relation to immunity and infectious disease. J. Nutr. 131:616S-635S.
22. Jahari, A. B., Saco-Pollitt, C., Husaini, M. & Pollitt, E. (2000) Effects of an energy and micronutrient supplement on motor development and motor activity in undernourished children in Indonesia. Eur. J. Clin. Nutr. 54(Suppl.):S60-S68.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
