![[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}
|
|
|
|
|
||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Universidad del Cauca, Popayan, Colombia;
* Département de Médecine Sociale et Préventive, Université de Montréal, Montréal, Canada;
Département de Nutrition, Université de Montréal, Montréal, Canada; and
** Universidad de Manizales, Manizales, Colombia
2To whom correspondence should be addressed. E-mail: be.alvarado.llano{at}umontreal.ca.
 |
ABSTRACT |
|---|
 TOP ABSTRACT SUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
KEY WORDS: • growth trajectories • multilevel models • infant feeding • morbidity • Colombia
Continued breast-feeding after 6 mo of life and its effects on infant growth in developing countries are matters of great controversy. Some studies have shown inverse associations (1), generally attributed to their cross-sectional nature; i.e., children with growth retardation would be more likely to continue to breast-feed (2). Longitudinal designs were used thereafter to detect reverse causality (3) and control for selection bias and confounding factors that could flaw study results (4–6). However, longitudinal results are still inconclusive. In Kenya and Guinea-Bissau, breast-feeding after 12 mo of age was associated with faster weight and length gains (7,8); in Senegal, children breast-fed during y 2 of life tended to growth faster in length (9); in Peru, a positive effect of breast-feeding on linear growth was observed in only a subset of children, aged 15 to 18 mo, whose animal-based food intake was low (4). In Sudan, continued breast-feeding beyond mo 6 of life was associated with slower length and weight gain; complementary food intake and mother’s socioeconomic status were important confounders (5). Breast-feeding after 6 mo, however, is advocated on the basis of its widely accepted protective effects on infectious morbidity and child mortality (10,11).
Given the lack of conclusive evidence and the specific social and cultural conditions affecting breast-feeding and weaning practices in Afro-Colombian communities, we decided to explore breast-feeding and complementary feeding practices, morbidity, and growth among Afro-Colombian children aged 6 to 18 mo using a prospective design. Previous results from the cross-sectional study conducted before cohort inception indicated that breast-feeding positively affected children > 12 mo old; i.e., breast-feeding after 12 mo resulted in higher length-for-age Z-scores, even when adjusting for diversity of food consumption and mother’s socioeconomic status (12). Qualitative field results (13) and quantitative data (12) from the same community support the hypothesis that mothers who perceived their infant’s health as "good" were more likely to breast-feed longer compared with mothers with a poor perception of their children’s health. This observation runs contrary to the data described in studies on Peruvian (3), Sudanese (5), Senegalese (14), and Guinean children (15). Complementary food intake began at a mean age of 3 mo; the diets of 70% of the 12-mo-old children consisted of animal-based foods (fish), tubers, rice, and legumes, rendering this population’s food intake higher in diversity than previously reported in another South American community (4). However, the cross-sectional nature of these findings limits causal inference.
In this paper, we present the results of the longitudinal study modeling trajectories of growth (length and weight) using Hierarchical Linear Models (HLM) (16,17). We focused on 3 specific questions: 1) What are the weight and linear growth trajectories of these infants (individual growth change)? 2) Does breast-feeding after 6 mo explain these trajectories once controlled by complementary food consumption and reported morbidity? 3) Do these trajectories vary as a function of maternal socioeconomic conditions, infant gender, and preterm delivery?
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Data collection.
Infant weight and recumbent length were measured during home visits, which took place at 
3-mo intervals (time between visits: 2.78 ± 0.59 mo). Values represent means ± SD unless otherwise noted. Anthropometric measurements were taken according to Lohman’s recommendations (19). Recumbent length was measured in all children using a graduated board with a fixed headboard and movable footboard (1 m/0.1 cm), and recorded to the nearest 0.1 cm; 4 measurements of the child’s length were taken and the mean recorded (interobserver technical error: 0.3 cm). Weight was measured using a suspended balance (Detecto Scale 25 kg/50 g) and recorded to the nearest 50 g. All children were barefoot and naked while being measured, and measurements were taken and recorded by the field investigator (B.E.A.). Children’s age estimates were based on date of birth and computed by Epi-NUT (20). The frequency of breast-feeding over the previous 24 h (number of feedings/d) and intake frequency of foods and beverages in the previous week were recorded monthly. The FFQ was developed on the basis of a previous dietary study conducted on the Pacific Coast of Colombia (21). We arranged the various types of food into 21 different food groups in a manner similar to that described by Marquis et al. (4). Mothers were asked to report the frequency of consumption of each food group and responses were scored from never in the last week (+1) to every day (+5). A total food consumption score was obtained by summing the single food group scores (Cronbach’s 
: 0.80). This total score is intended to represent a combined measure of food frequency (number of times per week the child consumed the food groups) and food diversity (number of food groups consumed in the last week). The higher the score, the higher the food frequency and/or diversity3 (22–25). Mothers were asked to record the occurrence of symptoms in their children (diarrhea, cough, fever, or other nonspecific symptoms such as head cold, rash, or skin lesions) on a daily basis. All symptom reports were based on mother’s perception. A primary health worker visited mothers weekly to monitor children’s health and data quality. A thick smear was performed when fever was reported within the last 48 h before the visit. There were 590 height and weight measurements, 1125 monthly data on breast-feeding and food frequency, and 36,151 observed child-days of morbidity.
Mothers’ socioeconomic conditions were assessed, and weight and height were measured at baseline using a portable wood anthropometer and a medical weight scale (Health-O-Meter, 180 kg/454 g).
Variables
Time-fixed factors. The contribution of a set of fixed factors (not time-varying) to length and weight change was considered. These factors included mother’s weight and height (to partially control for genetic factors); mother’s socioeconomic condition, including education (years of schooling); and a wealth index (possession of radio, stove, refrigerator, electricity, and telephone). Child’s sex and a dichotomous variable (preterm vs. term) for mother-reported gestational age at birth were also included as infant-related fixed factors. The latter was determined by asking in what month of pregnancy the infant was born.
Time-varying factors. These represent exposure in the interval before each measurement of length and weight. The use of lagged variables attempts to account for temporality assumptions. For each lagged period, we calculated the mean number of breast-feedings and the mean food consumption score. In addition, breast-feeding was recorded as a dichotomous variable: (+0) if never breast-fed in the lagged period and (+1) otherwise. We defined a lagged occurrence of 4 symptoms (diarrhea, cough, fever, and nonspecific) as the number of days on which the symptom was recorded divided by the total number of child-days observed during the lagged period (26) and the lagged occurrence of healthy days (days in which mothers did not record any symptoms).
Analyses
Statistical model.
Length and weight trajectories were modeled via HLM using HLM 5.0 software (27). HLM is appropriate for data with a nested structure, such as repeated-measures data in which several individual measurements (follow-ups) are nested or clustered within individuals (children). HLM separates within-child (Level 1) and between-child models (Level 2), estimating within-child and between-children variability:
 | (1) |
 | (2) |
 | (3) |
 | (4) |
In the Level 1 model, Yit is the length/weight for children i on follow-up t, T is the age for child i at time t, and X is a time-varying variable, i.e., breast-feeding, food score, and the lagged occurrence of symptoms (diarrhea, cough, fever, or other nonspecific). ß0i is the intercept, indicating child i’s fitted value of length or weight when both T and X equal 0. For the purpose of the study, T was centered on the youngest age (5 mo) in order to interpret ß0i as the length/weight at the onset of data collection; ß1i (T)it is a slope, indicating child i’s length/weight changes per mo; ßxi (X)it is a slope, indicating the association between time-varying variables and length/weight trajectories. The random effect for the Level 1 model is given by rit, and it is assumed to be normally distributed with mean 0 and variance 
2. The intercepts and slopes of the within-child model (Level 1) become the outcomes for the between-child model (Level 2). The 
parameters represent the mean level of the corresponding within-child parameters; 00 corresponds to the mean initial status for length/weight; 10 the average change per mo; and x0 the average effect of time-varying variables on length/weight. Any between-child variation in the regression coefficients is modeled in the Level 2 model as a function of a fixed factor Zi and random effects U0i and U1i. These random effects are assumed to be normally distributed with means 0 and variances 
002 and 112.
Hierarchical linear modeling allows us to test whether within-child and between-child regressions for a time-varying variable are different. The values of X in Eq. (1) are now transformed into deviations from each child mean calculated across the entire period of observations, i.e., (Xit – Xi) or child-centering. The resulting equations are as follows:
 | (5) |
 | (6) |
|
| (7) |
|
| (8) |
As proven by Snijders and Bosker (28), under this formulation, the between-child regression is now 0x + x0 and reflects the effects of time-varying variables on between-child differences in length/weight, whereas x0 is an estimator that reflects the effects of time-varying variables on within-child differences. For breast-feeding and food frequency, Xi was the child overall mean (
score obtained at each monthly observation/number of observations); and for morbidity variables, Xi represented the total occurrence for each reported symptom calculated for each child ( number of days for which a symptom was reported on each observation/total number of child days followed). In other words, if a child’s growth depends on the total time the child was breast-fed, then the between-child regression coefficient would capture the effect of breast-feeding on mean growth; but if the child’s growth also depends on breast-feeding at each time, then the within-child regression coefficient would capture this breast-feeding effect on growth rate.
In these analyses, the model construction follows Snijders and Bosker (28), beginning with Level 1 and continuing with Level 2. First, a model without covariates was fitted to test random variability in the intercept (Model 0) and to estimate the intraclass correlation coefficient (percentage of the total variability accounted for by the variability between children). Second, age centered at 5 mo was entered and random variability explored (Model 1). Polynomial functions were evaluated (entering age squared and a cubic age term), and a random slope tested. Then, time-varying variables, within-child centered (Xit – Xi), and their fixed effects at Level 2, Xi, were entered one at a time (Model 2). Combined and interaction effects among breast-feeding practices with food scores and symptom occurrence were also tested. When variables at Level 1 were all tested, fixed variables were entered (Model 3). Maternal characteristics were entered first, followed by child characteristics. Significant addition of Level-1 variables was modeled with a t test for fixed effects (Tx = x/SE x); and a deviance test for random effects. P-values < 0.05 were considered significant. We followed model assumptions and diagnostic procedures as suggested by Snijders and Bosker (28).
|
RESULTS |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
|
2 and 002 yielded an intraclass correlation coefficient of 31% for length and 75% for weight. Thus, the length variability was due mainly to within-child differences, whereas for weight variability, between-child differences were more significant. The overall mean lengths and weights (00) were 74.29 ± 2.42 cm and 9.15 ± 1.20 kg, respectively. For linear growth, a polynomial term was significant (20). Furthermore, random variation was significant for the intercept and slope in both length and weight (Table 2, Model 1), meaning that variability existed in length and weight at baseline and in the growth trajectories. The 95% CI at baseline was as follows: for length (59.89, 69.79 cm), and for weight (5.37, 9.99 kg). The 95% CI for growth rate was (1.13, 1.70 cm/mo) for length and (66.5, 319.4 g/mo) for weight. Initial status and growth rate, for either length (r = 0.042; P > 0.05) or weight (r = –0.13; P > 0.05), were not correlated.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Methodological limitations and strengths. Breast-feeding was modeled as a dichotomous variable. This modeling approach may underestimate the benefits of breast-feeding on the child’s growth. Our measurements did not capture the effects of the intensity of breast-feeding on child growth after 6 mo nor did they quantify the nutritional contribution from breast milk (29).
In this study, there is no reference against which complementary food consumption scores can be compared. Conventional dietary studies on nutrient intake are time consuming and costly; therefore, studies have validated food diversity scores on the basis of food groups (23,25). We developed a food group–based frequency questionnaire to identify the diversity and frequency in the child’s complementary diet as others have done (4,22–25,30,31). Our approach was also used for analysis of infant feeding practices in the national health surveys (31) and in other cross-sectional studies in Kenya (22) and Mali (23). Food diversity based on number of food groups consumed is associated with nutrient adequacy (23,32) and better nutritional scores in cross-sectional studies (23,24); it also modifies the effect of breast-feeding in cohort studies (4,6).
We modeled infant growth, feeding practices, and morbidity using HLM (28). Modeling growth with HLM offers flexibility in design and analysis, compared with more traditional approaches (16,17). Predictors can be discrete or continuous, time dependent or independent; the number of measurement follow-ups can vary among subjects; and all subjects for whom data were obtained during at least one follow-up are included in the analysis. In populations in which follow-up is difficult due to high mobility, growth modeling with HLM is highly effective. Moreover, modeling growth with HLM allows for the study of individual change (within-child variability) separate from change between children (between-child variability). Most previous studies estimated the effects of time-varying variables such as breast-feeding on the rate of growth by modeling only between-child variability and likely underestimating the effects of these time-varying variables on individual growth rate. Finally, measures taken on the same children are highly correlated in longitudinal studies, introducing errors in statistical parameters. HLM was specifically developed to account for correlated response variables (28).
Time-varying variable influences on growth. Our data suggest that breast milk in this population is an important source of nutrition, required for growth beyond 6 mo of age. We found that among children whose food consumption did not increase in frequency (number of times per week) and/or diversity (number of food groups consumed), the rates of weight gain (2.2 kg) and linear growth (4.6 cm) were higher among breast-fed children than in those who were not breast-fed. Our results are consistent with other observational studies in Kenya and Guinea-Bissau. Children aged 9–18 mo who breast-fed throughout their follow-up were 3 cm longer and 370 g heavier than children breast-fed <50% of the time (7). In Kenya, children weaned at 12 mo weighed 137 g less than breast-fed children (8).
Our longitudinal approach allows us to consider the possibility of reverse causality. Could a positive relation between breast-feeding and growth be related to a mother’s preference to continue breast-feeding her child when the child is in good health? Reverse causality is partially controlled by the structure of the Level 1 model in hierarchical linear modeling. It takes into account the correlation between repeated measures (28). For that reason, we pursued an additional analysis as proposed by Marquis et al. (3) to address the argument of reverse causality. A logistic regression of the probability of being breast-fed on follow-up (t) and of being stunted and wasted on follow-up (t – 1) was conducted. There were no significant associations, except that nonwasted children at age 14–16 mo were more likely to have been breast-fed at the time they were last measured (49%) than wasted children (0%; P = 0.04).
The fact that the food consumption score does not appear to generate an independent effect on growth within our population could be related to the lack of appropriate nutritional quality of the diet (25) or lack of adjustment of frequency or diversity of food consumption to the growing child. Food scores increased mainly with respect to 2 animal-based products, fish and cow’s milk, whereas consumption of vegetables and fruits never reached daily frequency (Fig. 1). Food diversity reaches a maximum at 15 mo of age, when the child eats the same food as the rest of the family (Table 1). Given the widespread food insecurity, low availability of foods, and poor socioeconomic conditions, increased diversity in the diet is severely constrained (24). Furthermore, unhygienic conditions could mitigate the value of added foods (33).
Between- and within-child models help us understand the effect of morbidity on infant growth. For linear growth, the total proportion of healthy time was more strongly associated with mean length (between-child coefficient) than with length gain (within-child coefficient). Conversely, weight gain was significantly associated with healthy days (within-child coefficients). These results are consistent with the observation of linear retardation after periods of ill health, whereas weight faltering was observed in days or weeks (34). In particular, cough and fever episodes were related to lower mean length and weight and slower weight gain, respectively, as others have shown (35–38). Malaria could be disregarded for most fever episodes because a thick smear was tested in children reporting fever on the 2 d before the home visit (n = 117), and only 3 malaria episodes were diagnosed.
No significant relation could be demonstrated between diarrhea and growth, even though this association was reported previously (39–41). In our data, 3 factors could partially explain the lack of associations: 1) low frequency and severity of diarrhea episodes; 2) the high frequency of breast-feeding in our population, and the cultural norm that leads mothers to increase frequency of breast-feeding when the child has diarrhea; and 3) a classification bias; we based our definition on mother’s perception because the usual criterion of >3 stools each day may be normal in breast-fed children (42). We did not test the validity of mothers’ reports in our study; others (43) have supported their validity. However, this is not the first study indicating a lack of association between diarrhea and growth after 6 mo of life (44–46).
Fixed-variable influence on growth. This longitudinal study adds evidence to support previous results concerning the effect of the mother’s socioeconomic conditions on child growth (38,47). Low levels of education and material deprivation in mothers are both associated with slower gains in length in our study. Child growth faltering among noneducated women could be explained by food insecurity, low social support from the child’s father, and infrequent utilization of preventive practices (48,49). Mother’s height was associated with greater length and weight and predicted length at 6 mo, but did not influence length or weight growth rates. Mothers’ heights may reflect the deprived social conditions during their own childhoods (50,51). Although preterm birth is associated with lower baseline weight and length, the preterm and term birth groups did not differ in growth rates. Studies among preterm infants showed a possible catch-up effect after 6 mo of age (52), although in our study, preterm children never reached the height and weight of children born at term. Effects found herein may be underestimated because mothers are more likely to report that the infant was born later (53). Finally, feeding patterns in the first 6 mo of a child’s life were explored in the actual study, but results were not presented. Children weaned before 6 mo did not differ from those that were breast-fed (P = 0.50). The timing and quality of the complementary foods introduced were not associated with child growth in our data (P > 0.20).
Modeling growth changes with HLM demonstrated theoretical and methodological advantages, allowing us to achieve the following: 1) to differentiate the effects of time-varying variables on growth rate (within-child differences) from effects on average growth (between-child differences), 2) to differentiate the effects of fixed variables on the initial growth measure from effects on growth rate, and 3) to include children with incomplete follow-up.
Our study supports the positive effect of breast-feeding for >6 mo on infant growth, and any intervention in Afro-Colombian populations should foster this practice. Interventions should focus on mothers with low levels of education to improve nutritional education, hygiene practices, and care skills that could counterbalance the effects of poverty. Similarly, housing and environmental living conditions should be improved to reduce the number of respiratory symptoms related to child growth faltering.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
3 A food consumption score of 40 means that the child either consumed 10 different food groups 3–5 times/wk (low diversity, high frequency), or 20 different food groups at a frequency < 2 times/wk (high diversity, low frequency).
Manuscript received 15 November 2004. Initial review completed 30 December 2004. Revision accepted 25 May 2005.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
1. Caulfield, L. E., Bentley, M. E. & Ahmed, S. (1996) Is prolonged breastfeeding associated with malnutrition? Evidence from nineteen demographic and health surveys. Int. J. Epidemiol. 25:693-703.
2. Martin, R. M. (2001) Commentary: does breastfeeding for longer cause children to be shorter?. Int. J. Epidemiol. 30:481-484.
3. Marquis, G. S., Habicht, J. P., Lanata, C. F., Black, R. E. & Rasmussen, K. M. (1997) Association of breastfeeding and stunting in Peruvian toddlers: an example of reverse causality. Int. J. Epidemiol. 26:349-356.
4. Marquis, G. S., Habicht, J. P., Lanata, C. F., Black, R. E. & Rasmussen, K. M. (1997) Breast milk or animal-product foods improve linear growth of Peruvian toddlers consuming marginal diets. Am. J. Clin. Nutr. 66:1102-1109.
5. Fawzi, W. W., Herrera, M. G., Nestel, P., el Amin, A. & Mohamed, K. A. (1998) A longitudinal study of prolonged breastfeeding in relation to child undernutrition. Int. J. Epidemiol. 27:255-260.
6. Adair, L. S. & Guilkey, D. K. (1997) Age-specific determinants of stunting in Filipino children. J. Nutr. 127:314-320.
7. Onyango, A. W., Esrey, S. A. & Kramer, M. S. (1999) Continued breastfeeding and child growth in the second year of life: a prospective cohort study in western Kenya. Lancet 354:2041-2045.[Medline]
8. Molbak, K., Jakobsen, M., Sodemann, M. & Aaby, P. (1997) Is malnutrition associated with prolonged breastfeeding?. Int. J. Epidemiol. 26:458-459.
9. Simondon, K. B., Simondon, F., Costes, R., Delaunay, V. & Diallo, A. (2001) Breast-feeding is associated with improved growth in length, but not weight, in rural Senegalese toddlers. Am. J. Clin. Nutr. 73:959-967.
10. Molbak, K., Gottschau, A., Aaby, P., Hojlyng, N., Ingholt, L. & da Silva, A. P. (1994) Prolonged breast feeding, diarrhoeal disease, and survival of children in Guinea-Bissau. Br. Med. J. 308:1403-1406.
11. Leon-Cavas, N., Lutter, C., Ross, J. & Martin, L. (2002) Prolonged breast feeding, diarrhoeal disease, and survival of children in Guinea-Bissau. Quantifying the Benefits of Breastfeeding: A Summary of the Evidence Panamerican Health Organization Washington, DC.
12. Alvarado, B. E. (2005) Prolonged breast feeding, diarrhoeal disease, and survival of children in Guinea-Bissau. Épidémiologie de la croissance infantile: étude de déterminants sociaux et biologiques auprès d’enfants 
gés de 6 à 18 mois en Colombie. Universite de Montreal Montreal, QC.
13. Tabares, E. & Alvarado, B. E. (2003) Estado nutricional y practicas alimentarias en los primeros 18 meses de vida en poblaciones Amerindias y AfroColombianas de la Costa Pacifica. AntroPacifico 1:17-26.
14. Simondon, K. B., Costes, R., Delaunay, V., Diallo, A. & Simondon, F. (2001) Children’s height, health and appetite influence mothers’ weaning decisions in rural Senegal. Int. J. Epidemiol. 30:476-481.
15. Jakobsen, M. S., Sodemann, M., Molbak, K. & Aaby, P. (1996) Reason for termination of breastfeeding and the length of breastfeeding. Int. J. Epidemiol. 25:115-121.
16. Raudenbush, S. W. (2001) Comparing personal trajectories and drawing causal inferences from longitudinal data. Annu. Rev. Psychol. 52:501-525.[Medline]
17. Boyle, M. H. & Willms, J. D. (2001) Multilevel modelling of hierarchical data in developmental studies. J. Child Psychol. Psychiatry 42:141-162.
18. Profamilia (2000) Encuesta Nacional de Demografia y Salud. Profamilia Profamilia, ed. Santa Fe de Bogota Colombia.
19. Lohman, T., Roche, A. & Fajardo, L. (1990) Encuesta Nacional de Demografia y Salud. Anthropometric Standardization Reference Manual Human Kinetics Champaign, IL.
20. Centers for Disease Control and Prevention (2000) Encuesta Nacional de Demografia y Salud. EPI-INFO: Software for Public Health, 6.04 CDC Atlanta, GA.
21. Minota, Y. (2000) Encuesta Nacional de Demografia y Salud. Validación de un cuestionario semi-cuantitativo de frecuencia de alimentos en Buenaventura, Costa Pacifica Colombiana Escuela de Salud Publica Universidad del Valle, Cali, Colombia.
22. Onyango, A., Koski, K. G. & Tucker, K. L. (1998) Food diversity versus breastfeeding choice in determining anthropometric status in rural Kenyan toddlers. Int. J. Epidemiol. 27:484-489.
23. Hatloy, A., Torheim, L. E. & Oshaug, A. (1998) Food variety—a good indicator of nutritional adequacy of the diet? A case study from an urban area in Mali West Africa. Eur. J. Clin. Nutr. 52:891-898.[Medline]
24. Hatloy, A., Hallund, J., Diarra, M. M. & Oshaug, A. (2000) Food variety, socioeconomic status and nutritional status in urban and rural areas in Koutiala (Mali). Public Health Nutr 3:57-65.[Medline]
25. Ruel, M. T. (2003) Is dietary diversity an indicator of food security or dietary quality? A review of measurement issues and research needs. Food Nutr. Bull. 24:231-232.[Medline]
26. Morris, S. S., Santos, C. A., Barreto, M. L., Cousens, S. N., Strina, A., Santos, L. M. & Assis, A. M. (1998) Measuring the burden of common morbidities: sampling disease experience versus continuous surveillance. Am. J. Epidemiol. 147:1087-1092.
27. Raudenbush, S. W., Brik, A. S., Cheong, Y. F. & Congdon, R. T., Jr (2000) Measuring the burden of common morbidities: sampling disease experience versus continuous surveillance. HLM 5: Hierarchical Linear and Nonlinear Modeling Scientific Software International Lincolnwood, IL.
28. Snijders, T.A.B. & Bosker, R. (1999) Measuring the burden of common morbidities: sampling disease experience versus continuous surveillance. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling SAGE Publications, Ltd London, UK.
29. Dewey, K. G. & Brown, K. H. (2003) Update on technical issues concerning complementary feeding of young children in developing countries and implications for intervention programs. Food Nutr. Bull. 24:5-28.[Medline]
30. Ruel, M. & Arimond, M. (2003) Update on technical issues concerning complementary feeding of young children in developing countries and implications for intervention programs. Measuring Childcare Practices International Food Policy Research Institute Washington, DC.
31. Arimond, M. & Ruel, M. T. (2004) Dietary diversity is associated with child nutritional status: evidence from 11 demographic and health surveys. J. Nutr. 134:2579-2585.
32. Ruel, M. T. (2003) Operationalizing dietary diversity: a review of measurement issues and research priorities. J. Nutr. 133:3911S-3926S.
33. Mortajemi, Y., Kaferstein, F., Moy, G. & Quevedo, F. (1993) Contaminated complementary food: a major risk for diarrhoea and associated malnutrition. Bull. WHO 71:79-92.[Medline]
34. Stephensen, C. B. (1999) Burden of infection on growth failure. J. Nutr. 129:534S-538S.
35. Beisel, W. R. (1977) Magnitude of the host nutritional responses to infection. Am. J. Clin. Nutr. 30:1236-1247.
36. Frongillo, E. A., Jr (1999) Symposium: Causes and Etiology of Stunting. Introduction. J. Nutr. 129:529S-530S.
37. Kolsteren, P. W., Kusin, J. A. & Kardjati, S. (1997) Pattern of linear growth velocities of infants from birth to 12 months in Madura, Indonesia. Trop. Med. Int. Health. 2:291-301.[Medline]
38. Ricci, J. A. & Becker, S. (1996) Risk factors for wasting and stunting among children in Metro Cebu, Philippines. Am. J. Clin. Nutr. 63:966-975.
39. Rowland, M. G., Rowland, S. G. & Cole, T. J. (1988) Impact of infection on the growth of children from 0 to 2 years in an urban West African community. Am. J. Clin. Nutr. 47:134-138.
40. Lopez de Romana, G., Brown, K. H., Black, R. E. & Kanashiro, H. C. (1989) Longitudinal studies of infectious diseases and physical growth of infants in Huascar, an underprivileged peri-urban community in Lima, Peru. Am. J. Epidemiol. 129:769-784.
41. Brown, K. H. (2003) Diarrhea and malnutrition. J. Nutr. 133:328S-332S.
42. Hyams, J. S., Treem, W. R., Etienne, N. L., Weinerman, H., MacGilpin, D., Hine, P., Choy, K. & Burke, G. (1995) Effect of infant formula on stool characteristics of young infants. Pediatrics 95:50-54.
43. Lartey, A., Manu, A., Brown, K. H., Peerson, J. M. & Dewey, K. G. (2000) Predictors of growth from 1 to 18 months among breast-fed Ghanaian infants. Eur. J. Clin. Nutr. 54:41-49.[Medline]
44. Briend, A., Hasan, K. Z., Aziz, K. M. & Hoque, B. A. (1989) Are diarrhoea control programmes likely to reduce childhood malnutrition?. Observations from rural Bangladesh. Lancet 2:319-322.
45. Checkley, W., Epstein, L. D., Gilman, R. H., Cabrera, L. & Black, R. E. (2003) Effects of acute diarrhea on linear growth in Peruvian children. Am. J. Epidemiol. 157:166-175.
46. Villamor, E., Fataki, M. R., Bosch, R. J., Mbise, R. L. & Fawzi, W. W. (2004) Human immunodeficiency virus infection, diarrheal disease and sociodemographic predictors of child growth. Acta Paediatr 93:372-379.[Medline]
47. Cleland, J. G. & Van Ginneken, J. K. (1988) Maternal education and child survival in developing countries: the search for pathways of influence. Soc. Sci. Med. 27:1357-1368.
48. Wagstaff, A., Bustreo, F., Bryce, J. & Claeson, M., WHO–World Bank Child Health and Poverty Working Group (2004) Child health: reaching the poor. Am. J. Public Health 94:726-736.
49. Schellenberg, J. A., Victora, C. G., Mushi, A., de Savigny, D., Schellenberg, D., Mshinda, H., Bryce, J., Tanzania Integrated & Management, of, Childhood Illness (2003) Inequities among the very poor: health care for children in rural southern Tanzania. Lancet 361:561-566.[Medline]
50. Spencer, N. J. & Logan, S. (2002) The treatment of parental height as a biological factor in studies of birth weight and childhood growth. Arch. Dis. Child. 87:184-187.
51. Ramakrishnan, U., Martorell, R., Schroeder, D. G. & Flores, R. (1999) Role of intergenerational effects on linear growth. J. Nutr. 129:544S-549S.
52. Arifeen, S. E., Black, R. E., Caulfield, L. E., Antelman, G. & Baqui, A. H. (2001) Determinants of infant growth in the slums of Dhaka: size and maturity at birth, breastfeeding and morbidity. Eur. J. Clin. Nutr. 55:167-178.[Medline]
53. Ywan, B. P., Suman, V.J. & Jacobsen, S.J. (1998) Maternal recall of distant pregnancy events. J. Clin. Epidemiol. 51:399-405.[Medline]
This article has been cited by other articles:
|
|
 |
|
  N. Mpontshane, J. Van den Broeck, M. Chhagan, K. K. A. Luabeya, A. Johnson, and M. L. Bennish HIV Infection Is Associated with Decreased Dietary Diversity in South African Children J. Nutr., September 1, 2008; 138(9): 1705 - 1711. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
