SEARCH

SEARCH BY CITATION

Keywords:

  • Obesity;
  • infant feeding;
  • longitudinal;
  • cohort studies;
  • adiposity

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Objective. To evaluate the association between breastfeeding duration, introduction of solid or semi-solid foods before four months of age and overweight/obesity at 11 years. Methods. Prospective population-based birth cohort study, including 1 204 adolescents aged 11 years who were born in Pelotas (Brazil), in 1993, and were previously interviewed at birth, six and 12 months of age. Five explanatory variables were used: duration of any breastfeeding, duration of exclusive or predominant breastfeeding, ever breastfeeding, introduction of solid or semi-solid foods before four months of age, and feeding pattern at four months of age. Risk of overweight was defined as body mass index (BMI) ≥ 85th percentile, and obesity was defined as BMI ≥ 85th percentile plus triceps and subscapular skinfolds ≥ 90th percentile, using the National Center for Health Statistics curves. Results. The prevalence of risk of overweight and obesity at 11 years were 23.2% and 11.6%, respectively. The lowest prevalence of both outcomes was found among subjects who were breastfed for one to three months. However, tests for linear trend or heterogeneity did not result in a significant association between breastfeeding indicators and anthropometry at 11 years. Adjustment for sex, skin color, birth weight, maternal schooling, smoking during pregnancy and maternal pre-pregnancy BMI did not alter the findings. Conclusions. Our results do not support the hypothesis that breastfeeding has a long-term lasting effect on anthropometry in this population.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

As overweight children and adolescents are at an increased risk of becoming obese adults [1], [2], it is important to identify risk factors for overweight and obesity in early life. Most studies on the long-term effects of breastfeeding on health suggest that breastfeeding has a protective effect on the prevalence of obesity in childhood [3] or adolescence [4] although not all studies have found consistent results [5], [6]. Few of them explored the effect of the introduction of solid or semi-solid foods on the prevalence of overweight or obesity later on, and results of such studies are inconsistent [5], [7].

Relatively few studies on this issue were carried out in low or middle-income settings [8–10]. In the 1993 birth cohort from Pelotas, Brazil, there was no association between breastfeeding and overweight at four years [9], while an analysis of the 1982 cohort in the same city found no effect of breastfeeding on body composition at 18 years [8]. On the other hand, Siqueira and coworkers [10] found that children and adolescents aged 6–14 years from a private school in Sao Paulo, who were never breastfed, presented the highest prevalence of obesity.

The present study aimed at assessing the effect of: total breastfeeding duration; predominant breastfeeding duration; introduction of solid or semi-solid food before four months of age on overweight and obesity at the age of 11 in children from the 1993 Pelotas (Brazil) birth cohort study.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Pelotas is a southern Brazilian city with 320 000 inhabitants, located near the Uruguayan border. In 1993, all hospital deliveries were identified and children have been contacted on several occasions. Their mothers were interviewed soon after delivery, and children were weighed and measured. At 6 and 12 months of age, all low birth weight children and a systematic sample of 20% of the remaining were targeted for follow-up. Because of unequal sampling probabilities for low and normal birth weight children, weighted analyses were performed.

The following exposures were considered: total breastfeeding (BF) duration in months; predominant breastfeeding (PBF) duration in months (water, teas or juice were incorporated in addition to breast milk, but solid or semi-solid foods were not), ever breastfed (yes/no), introduction of solid or semi-solid foods (for example, fruit, vegetables, beans, rice, eggs, yogurts, etc) before four months of age (yes/no). Exclusive breastfeeding was very rare at the time of the study, so that predominant breastfeeding was used as an explanatory variable. If the child had been weaned or other foods introduced by the age of 6 months, information from this visit was used. If weaning happened between 6 and 12 months, information from the latter interview was used.

Feeding pattern at four months of age was also considered, with the following categories: (a) PBF; (b) breast milk plus solids or semi-solids; (c) breast milk plus non-human milk plus solids or semi-solids; and (d) non-human milk plus solids or semi-solids. No children received breast plus non-human milk without foods, or non-human milk without foods.

In 2004–5, all cohort participants were eligible for a follow-up visit, except the 141 who were known to have died. A citywide and a school census were carried out to help locate cohort members. Weight was measured using a SECA (Birmingham, UK) scale (100 g accuracy). Height was measured using a portable stadiometer (1 mm accuracy). Triceps and subscapular skin folds were measured using a CESCORF caliper (0.1 mm accuracy). Body mass index (BMI) was calculated. Nutritional status was categorized using the World Health Organization recommendation [11]; risk of overweight was defined as BMI ≥ 85th percentile, and obesity was defined as BMI ≥ 85th percentile, and triceps and subscapular skinfolds ≥ 90th percentile, using the National Center for Health Statistics (NCHS) curves.

The following variables were considered as possible confounders: maternal schooling (highest grade achieved), maternal smoking during pregnancy (yes/no), maternal pre-pregnancy BMI, birth weight (in grams), family income (in minimum wages), sex, skin color (white/black and mixed). At all stages of the cohort study, 10% of the interviews were repeated by field supervisors for quality control purposes.

With the sample size included, it was possible to detect statistically significant relative risks of 0.7 or smaller with a statistical power of 80%, and confidence level of 95%. Analyses were performed using Stata and SPSS, including chi-square tests for heterogeneity or linear trend in the unadjusted analyses, and Wald tests in the adjusted models. Poisson regression was used in the multivariable analysis, as suggested for binary outcomes with high prevalence [12].

All stages of the 1993 cohort were approved by the Federal University of Pelotas Medical School Ethics Committee. Mothers or guardians provided written informed consents in each visit. Further details of the methodology of the 1993 Pelotas (Brazil) birth cohort study are available elsewhere [13].

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Out of the 5 265 live born children in 1993, only 16 were not included in the birth cohort study, and 1 460 were sampled for follow-up at six and 12 months; 96.8% and 93.4% respectively, of these children sampled were traced. At 11 years, 4 452 adolescents were traced; added to the 141 known to have died, they represented 87.5% of the original cohort. In this analysis, we included the 1 204 subjects with data available at birth, six and 12 months, and 11 years.

Table I describes the sample in terms of sex (49.6% boys), skin color (3/4 white), birth weight (9.8% <2 500 g), family income (18% of the families earning less than one minimum wage per month), maternal schooling (27.2% <4 years), maternal smoking during pregnancy (1/3 of the mothers smoked during pregnancy), maternal pre-gestational BMI (22.6% >25 kg/m2). With respect to breastfeeding, only 3.9% of children had never been breastfed, whereas almost one quarter were breastfed for less than one month and only 18.3% were breastfed for 12 months or more. The proportion of subjects who were predominantly breastfed for less than one month was 27.4%, and 14.9% were predominantly breastfed for at least four months.

Table I. . Description of the sample in terms of maternal and child's characteristics.
Variable%
  • a

    The highest number of missing values was 37 (pre-pregnancy BMI).

  • PBF = Predominant breastfeeding; BF = Breastfeeding; BMI = Body mass index.

Adolescent's sex 
Male49.6
Female50.4
  
Adolescent's skin color
White70.9
Black + mixed29.1
Birth weight (g)
< 2 5009.8
≥ 2 50090.2
Family income (in minimum wages)
≤ 118.0
1.1–345.5
3.1–622.3
6.1–107.4
≥ 106.8
  
Maternal education (years)
02.5
1–424.7
5–848.2
≥ 924.6
Smoking during pregnancy
Yes33.1
No66.9
Maternal BMI (pregestational)
< 18.5 kg/m28.8
18.5–24.99 kg/m268.7
25.0–29.99 kg/m216.6
≥ 30 kg/m26.0
Duration of total breastfeeding (months)
< 122.7
1–2.926.9
3–5.914.7
6–8.97.5
9–11.99.8
≥ 1218.3
Duration of predominant breastfeeding (months)
< 127.4
1–1.919.2
2–2.921.4
3–3.917.1
≥ 414.9
Ever breastfed
Yes96.1
No3.9

With respect to solid or semi-solid foods, 85.1% of children had already received it before four months of age. In terms of the variable combining PBF with introduction of other foods, most children received both non-human milk and other foods at the age of four months. With regard to the use of non-human milk, alone or in combination with breast milk, 21.2% received cow's milk, 33.5% formula and 18.1% both. The frequencies of risk of overweight and obesity at 11 years of age were 23.2% and 11.6%, respectively (data not shown in Table I).

Table II presents the association between breastfeeding and the potential confounding variables. Total breastfeeding duration was significantly associated with sex, skin color and maternal schooling. Predominant breastfeeding was associated with skin color, birth weight, maternal schooling, and maternal pre-pregnancy BMI. Ever breastfeeding was significantly associated with birth weight; the association with skin color was of borderline significance (P = 0.06). Maternal schooling was associated with the introduction of solid or semi-solid foods before four months of age.

Table II. . Relationship between breastfeeding, infant feeding and child, and family characteristics (n = 1 204).
 % males% white skin colorBirth weight (g) (mean±s.d.)% poor (1st quintile of income)Maternal schooling in years (mean±s.d.)Maternal BMI in kg/m2 (mean±s.d.)% maternal smoking
  • *

    χ2 test (sex, skin color, poverty and maternal smoking) and one-way analysis of variance (birth weight, maternal schooling and maternal body mass index – BMI). PBF = Predominant breastfeeding; BF = Breastfeeding; BMI = Body mass index; s.d.=Standard deviation.

Duration of total breastfeeding (months)
< 156.676.83 114 (608)22.06.2 (3.3)23.0 (4.4)32.1
1–2.948.970.73 175 (524)16.36.7 (3.6)22.2 (3.8)39.5
3–5.955.671.23 179 (475)21.06.7 (3.4)22.3 (3.3)30.5
6–8.950.571.43 231 (507)13.28.9 (4.2)22.7 (3.7)28.9
9–11.939.076.33 216 (485)19.07.3 (3.7)22.7 (3.6)32.2
≥ 1242.160.63 210 (493)23.06.0 (3.0)23.0 (3.6)29.0
P*0.0030.030.270.19<0.0010.560.10
Duration of predominant breastfeeding (months)
< 153.577.53 103 (586)19.76.4 (3.6)23.3 (4.3)32.3
1–1.949.869.63 158 (493)21.06.2 (3.5)22.8 (3.7)32.5
2–2.947.965.03 242 (524)18.26.6 (3.2)22.6 (3.7)39.3
3–3.944.264.63 184 9471)19.47.2 (3.8)22.6 (3.6)32.5
≥ 450.676.53 231 (512)19.67.4 (3.5)22.3 (3.3)27.2
P*0.31<0.010.010.960.0010.0010.11
Ever breastfed
Yes49.470.43 186 (517)19.86.7 (3.6)22.8 (3.8)33.4
No53.283.02 945 (724)12.86.0 (3.2)22.7 (4.2)25.5
P*0.610.060.0020.230.180.840.26
Introduction of solid or semi-solid foods before 4 months
Yes49.570.13 167 (530)19.46.6 (3.5)22.9 (3.9)34.1
No50.678.53 166 (562)19.67.4 (3.5)22.3 (3.3)27.2
P*0.790.080.160.960.0060.080.07
Feeding pattern at four months of age
PBF ≥ 4 months50.676.53 231 (512)19.67.4 (3.5)22.3 (3.3)27.2
BF + solid or semi-solid food38.765.93 223 (447)18.16.9 (3.5)22.6 (3.6)36.4
BF + solid or semi-solid food + milk47.060.23 183 (516)20.46.4 (3.7)23.1 (3.8)26.9
Milk + solid or semi-solid food53.674.63 147 (556)19.36.6 (3.5)22.9 (4.0)36.1
P*0.005<0.010.160.960.020.020.02

Table III shows that the risk of overweight was not related to breastfeeding duration, although a protective effect was apparent for children who were breastfed for one to three months compared with those breastfed for 12 months or longer. Results were unchanged after adjustment for sex, skin color, birth weight, maternal schooling, smoking during pregnancy and pre-pregnancy BMI. As for the feeding pattern at age four months, children who received non-human milk were at approximately 30% lower risk of overweight than those who did not. Additional analyses (not presented) showed similar risk of overweight for children receiving cow's milk and formula.

Table III. . Prevalence of overweight, crude and adjusted prevalence ratios (PR) for overweight at 11 years.
  Crude analysisAdjusted analysisa
 Prevalence of overweightPR (CI 95%)P-valuePR (CI 95%)P-value
  • a

    Likelihood-ratio test for heterogeneity adjusted for infant's sex, infant's skin color, infant's birth weight, family income (first quintile), maternal schooling, smoking during pregnancy and maternal pregestational body mass index.

  • PBF = Predominant breastfeeding; BF = Breastfeeding.

Duration of total breastfeeding (months)
< 125.70.99 (0.71–1.38)0.090.92 (0.66–1.27)0.06a
1–2.917.30.66 (0.46–0.95) 0.61 (0.43–0.86) 
3–5.921.00.81 (0.54–1.21) 0.81 (0.55–1.18) 
6–8.925.30.97 (0.61–1.54) 0.77 (0.48–1.23) 
9–11.929.71.14 (0.77–1.70) 1.04 (0.70–1.56) 
≥ 1226.21.00 1.00 
Duration of predominant breastfeeding (months)
< 123.40.81 (0.58–1.14)0.440.80 (0.57–1.12)0.64a
1–1.922.90.79 (0.55–1.14) 0.82 (0.57–1.17) 
2–2.921.00.73 (0.50–1.06) 0.71 (0.49–1.01) 
3–3.921.00.73 (0.49–1.07) 0.74 (0.51–1.10) 
≥ 428.91.00 1.00 
Ever breastfed
Yes23.31.09 (0.61–1.93)0.770.95 (0.53–1.72)0.87a
No21.31.00 1.00 
Introduction of solid or semi-solid foods before 4 months
Yes22.30.77 (0.58–1.03)0.080.77 (0.58–1.02)0.07
No28.91.00 1.00 
Feeding pattern at four months of age
PBF ≥ 4 months28.91.000.051.000.05a
BF + solid or semi-solid food28.91.00 (0.70–1.45) 1.00 (0.70–1.43) 
BF + solid or semi-solid food + milk19.10.67 (0.44–0.99) 0.69 (0.47–1.02) 
Milk + solid or semi-solid food21.30.74 (0.55–1.00) 0.73 (0.54–0.99) 

Table IV shows the corresponding analyses for obesity. Although the risk for children breastfed for 1–2.9 months was still somewhat lower, the differences are not significant. As in the preceding analyses, there was no association with ever breastfeeding or age of introduction of solids and semi-solids. Unlike what was observed for overweight, there was no association with the feeding pattern at four months.

Table IV. . Prevalence of obesity, crude and adjusted prevalence ratios (PR) for obesity at 11 years.
  Crude analysisAdjusted analysisa
 Prevalence of obesityPR (CI 95%)P-valuePR (CI 95%)P-value
  • a

    Likelihood-ratio test for heterogeneity adjusted for infant's sex, infant's skin color, infant's birth weight, family income (first quintile), maternal schooling, smoking during pregnancy and maternal pregestational body mass index.

  • PBF = Predominant breastfeeding; BF = Breastfeeding.

Duration of total breastfeeding (months)
< 114.71.10 (0.68–1.80)0.280.99 (0.60–1.64)0.17a
1–2.98.00.62 (0.36–1.08) 0.58 (0.33–1.02) 
3–5.99.70.74 (0.39–1.39) 0.73 (0.40–1.32) 
6–8.911.00.81 (0.38–1.73) 0.64 (0.30–1.35) 
9–11.914.41.08 (0.58–2.01) 1.14 (0.62–2.10) 
≥ 1213.11.00 1.00 
Duration of predominant breastfeeding (months)
< 111.90.96 (0.56–1.65)0.830.85 (0.49–1.47)0.81a
1–1.913.01.08 (0.61–1.90) 1.05 (0.59–1.85) 
2–2.911.30.91 (0.51–1.64) 0.87 (0.50–1.53) 
3–3.99.30.75 (0.39–1.43) 0.75 (0.40–1.42) 
≥ 412.21.00 1.00 
Ever breastfed
Yes11.71.30 (0.47–3.59)0.611.34 (0.45–3.91)0.60a
No8.51.00 1.00 
Introduction of solid or semi-solid foods before 4 months
Yes11.50.93 (0.58–1.50)0.780.88 (0.55–1.41)0.60
No12.21.00 1.00 
Feeding pattern at four months of age
PBF ≥ 4 months12.21.000.621.000.61 a
BF + solid or semi-solid food12.71.05 (0.57–1.95) 1.09 (0.60–1.99) 
BF + solid or semi-solid food + milk12.00.97 (0.53–1.76) 0.93 (0.51–1.68) 
Milk + solid or semi-solid food10.80.88 (0.53–1.45) 0.81 (0.49–1.33) 

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Our results fail to show a consistent association between breastfeeding or introduction of complementary foods and the risk of overweight or obesity in young adolescents. The only significant findings were reduced risks of overweight among children breastfed for one to three months, and for those receiving non-human milk at the age of four months, but these were not significant for obesity. Although it is tempting to attribute this finding to a critical window, this result must be interpreted with caution, because such findings were not observed in other studies [4], [14], nor even in our population [8], [9].

Some limitations of our study need to be mentioned. First, duration of exclusive breastfeeding in Pelotas in 1993 was so short that we were not able to analyze this variable separately. Second, body composition data (lean and fat mass separately) are not available, although our definition of obesity at 11 years includes two skinfolds, as recommended by the World Health Organization [11]. Strengths of this study include the prospective design, with information on breastfeeding collected in infancy and early childhood, minimizing the likelihood of recall bias. Furthermore, we included in our analysis the main confounding factors reported in previous studies [15], thus reducing the likelihood of residual confounding. In particular, residual confounding due to socio-economic factors is unlikely to account for this association because longer breastfeeding duration was not associated with high socio-economic and educational level [8], [9], unlike that which is observed in high-income countries [16]. The high follow-up rate minimizes the likelihood of selection bias.

Horta et al. [17] concluded that breastfeeding was inversely associated with overweight in adolescence using meta-analytic techniques (pooled odds ratio: 0.69; 95% CI: 0.60; 0.80). Other meta-analyses have found similar results [15], [18], [19], although Owen et al. [15] concluded that publication bias was likely in the studies on this issue. A few studies from developed countries reported no association [6], [20]. Virtually all studies included in these meta-analyses derive from high-income countries.

We now have three studies that did not find a higher risk of overweight or obesity for children who were not breastfed, or who were breastfed for short periods. These findings came from two separate birth cohorts from Pelotas, using as outcomes BMI and body composition at 18 years [8], and weight-for-height Z-score at four years [9]. We propose three possible explanations for this discrepancy with most of the literature. One possibility is that unlike studies from developed countries, residual confounding by socio-economic status is not present in our analyses (Table II). Second, there may be local peculiarities regarding the type of local weaning foods, which are different from those used in developed countries. Third, a recent study in our population showed that sleeping metabolic rate was higher among 8-month-old babies who received cow's milk in addition to breast milk, compared with those who received only breast milk [21]. Such higher rates may offset the additional nutrient intake associated with non-human milk. Further research is needed to elucidate these possible pathways.

Even if breastfeeding has no long-term lasting protective effect on the risk of obesity, its other benefits, including short-term protection against infectious diseases and mortality [22], and long-term benefits [17], [19] support its continued promotion.

Acknowledgement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References
  • 1
    Whitaker RC, Wright JA, Pepe MS, Seidel KD, Dietz WH. Predicting obesity in young adulthood from childhood and parental obesity. N Engl Med. 1997; 337(13)86973.
  • 2
    Guo SS, Roche AF, Chumlea WC, Gardner JD, Siervogel RM. The predictive value of childhood body mass index values for overweight at age 35 y. Am J Clin Nutr. 1994; 59(4)8109.
  • 3
    Armstrong, J, Reilly, JJ, Child Health Information Team. Breastfeeding and lowering the risk of childhood obesity. Lancet. 2002;359(9322): 20034.
  • 4
    Gillman MW, Rifas-Shiman SL, Camargo CA, et al. Risk of overweight among adolescents who were breastfed as infants. JAMA. 2001; 285(19)24617.
  • 5
    Hediger ML, Overpeck MD, Kuczmarski RJ, Ruan WJ. Association between infant breastfeeding and overweight in young children. JAMA. 2001; 285(19)245360.
  • 6
    Parsons TJ, Power C, Manor O. Infant feeding and obesity through the lifecourse. Arch Dis Child. 2003; 88(9)7934.
  • 7
    Kramer MS. Do breast-feeding and delayed introduction of solid foods protect against subsequent obesity?. J Pediatr. 1981; 98(6)8837.
  • 8
    Victora CG, Barros F, Lima RC, Horta BL, Wells J. Anthropometry and body composition of 18 year old men according to duration of breast feeding: birth cohort study from Brazil. BMJ. 2003; 327(7420)901.
  • 9
    Araúújo CL, Victora CG, Hallal PC, Gigante DP. Breastfeeding and overweight in childhood: evidence from the Pelotas 1993 birth cohort study. Int J Obes (Lond). 2006; 30(3)5006.
  • 10
    Siqueira RS, Monteiro CA. Breastfeeding and obesity in school-age children from families of high socioeconomic status. Rev Saúúde Púública. 2007; 41(1)512.
  • 11
    WHO Expert Committee on Physical Status: The use and interpretation of anthropometry: report of a WHO expert committee. WHO Technical Series-854. Geneva; 1995.
  • 12
    Barros AJ, Hirakata VN. Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol. 2003; 3: 21.
  • 13
    Victora CG, Araujo CLP, Menezes AMB, et al. Methodological aspects of the 1993 Pelotas (Brazil) Birth Cohort Study. Rev. Saúúde Púública. 2006; 40(1)3946.
  • 14
    Shields L, O'Callaghan M, Williams GM, Najman JM, Bor W. Breastfeeding and obesity at 14 years: a cohort study. J Paediatr Child Health. 2006; 42(5)28996.
  • 15
    Owen CG, Martin RM, Whincup PH, Smith GD, Cook DG. Effect of infant feeding on the risk of obesity across the life course: a quantitative review of published evidence. Pediatrics. 2005; 115(5)136777.
  • 16
    Bauchner H, Leventhal JM, Shapiro ED. Studies of breastfeeding and infections. How good is the evidence?. JAMA. 1986; 256(7)88792.
  • 17
    Horta, BL, Bahl, R, Martines, JC, Victora, CG. Evidence on the long-term effects of breastfeeding: systematic reviews and meta-analyses. Study commissioned by the World Health Organization; 2006.
  • 18
    Arenz S, Rüückerl R, Koletzko B, von Kries R. Breast-feeding and childhood obesity-a systematic review. Int J Obes Relat Metab Disord. 2004; 28(10)124756.
  • 19
    Ip, S, Chung, M, Raman, G, Chew, P, Magula, N, De Vine, D, et al. Breastfeeding and Maternal and Infant Health Outcomes in Developed Countries. Evidence Report/ Technology Assessment No. 153 (Prepared by Tuffs-New England Medical Center Evidence- based Practice Center, under Contract No. 290-02-0022). AHRQ Publication No. 07-E. 007. Rockville, MD: Agency for Healthcare Research and Quality, April, 2007.
  • 20
    Poulton R, Williams S. Breastfeeding and risk of overweight. JAMA. 2001; 286: 144950.
  • 21
    Haisma H, Wells JC, Coward WA, et al. Complementary feeding with cow's milk alters sleeping metabolic rate in breast-fed infants. J Nutr. 2005; 135(8)188995.
  • 22
    WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality. How much does breastfeeding protect against infant and child mortality due to infectious diseases? A pooled analysis of six studies from less developed countries. Lancet. 2000;355: 4515.