The growth and nutritional status of the breast-fed infant Imogen S. Rogers*, Pauline M. Emmett, Jean Golding Unit qf Pediutric and Perinatal Epidemiology. Institute of Child Health, University of Rrisroi. 24 Tyndall Avenue, Brisfol L&S8lTQ, UK Abstract: The literature on the relationship between early infant feeding and growth shows that after the first 3 or 4 months, breast-fed infants in the developed world are lighter than formula-fed infants with markedly lower adiposity. There is some evidence of a slightly lower rate of lineal growth over the first year or so. These differences in weight and length do not apparently persist beyond the first few years of life. In the developing world the situation is very different. The growth curves of breast-fed infants of malnourished mothers may falter between the third and sixth month of life. However, the generally poor quality of the supplementary foods offered in the developing world and the increased risk of diarrhoeal infections mean that supplementary feeding before the age of 6 months is unlikely to lead to a growth advantage and may well lead to growth faltering. 0 1997 Elsevier Science Ireland Ltd. Keywords: Breast feeding; Growth; Bottle feeding; Developing mothers; Supplementary feeding countries; Malnourished 1. Introduction It has been generally accepted that for infants of well-nourished mothers unsupplemented breast feeding should provide adequate nutrition for the first four to six months of life and support satisfactory growth during this time. In the first part of this paper we consider the evidence for this and determine the pattern of growth of infants in reasonably affluent environments. In the second part we consider the possible factors which may affect the growth and nutritional status of the child, such as illness in the child or malnutrition of the mother. Finally we examine the effects of breast feeding beyond one year. *Corresponding author. 0378-3782/97/$17.00 P/I SO37R-3782(97)00061-3 Q 1997 Elsevier Science Ireland Ltd. All rights reserved. S158 IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S174 2. Energy requirements of breast-fed infants There are three main components to the energy requirements of an infant -the energy for basal metabolism, including the maintenance of body temperature, the energy for growth and the energy for physical activity. Wells and Davies [l] measured sleeping metabolic rate as a proxy for basal metabolic rate. They studied 50 infants at lZweeks-old and found that the breast-fed group had a lower mean sleeping metabolic rate than the formula-fed group. This difference was related to differences in fat-free mass between the two groups. Butte et al. [2] had previously shown similar differences between breast and formula-fed infants at 1 and 4 months. The energy requirements of a breast-fedinfant may therefore be less than a formula-fed one. In many cultures babies are wrapped tightly for several weeks after birth, as in rural Nigeria where the baby is strapped tightly to the mother’s back in such a way that its movement is greatly restricted [3]. This results in a considerable reduction of the baby’s energy requirements for temperature maintenance and physical activity, and these babies have indeed been observed to have low intakes of breast milk. It has been suggestedthat the wrapping of babies in this way might ensure good growth in babies who are receiving restricted quantities of breast milk. A study in Oxford produced evidence that the young infant regulatesthe amount of milk it takes according to its needsfor growth [4]. The authors first studied bottle-fed infants in four groups: small-for-dates infants (i.e. more than 2 standard deviations below the mean for gestational age and sex), large-for-dates infants (more than 2 standard deviations above the mean for gestational age and sex), infants whose mothers had hypertension in pregnancy and a control sample who fell into none of these three categories.They were able to show that at two months the small-for-dates group took significantly more milk per unit bodyweight than either the infants of the hypertensive mothers or the controls, and the large-for-dates group took significantly less milk per unit bodyweight than the other three groups. Consequently the mean weight gain per day was greatest for the small-for-dates babies and lowest for the large-for-dates babies. However, due to the large difference in bodyweight between the small-for-dates and large-for-dates babies, the mean absolute intakes of the small-for-dates babies was lower than for the large-for-datesbabies. A similar pattern of weight gain was shown for groups of breast-fed infants, although no attempt was made to try to measurethe amount of breast milk the child took. Many studies have shown that among breast-fed infants the smaller infants take smaller volumes of milk. Michaelsen [5] working in Copenhagen, found that the strongest determinant of the variation in milk intake between infants at 2 and 4 months of age was the weight of the infant, accounting for 42% and 35% of the variance, respectively. Paul et al. [6] working in Britain and Neville et al. [7] working in USA came to similar conclusions. (The determinants of breast milk volume have been discussedin more detail elsewhere [S].) The British Committee on Medical Aspects of Food Policy has estimated the requirementsof infants for energy and a number of nutrients. However, in their report they emphasizethat the dietary referencevalues (DRV) given apply only to formulafed infants. IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S/l4 “For most nutrients the DRV of infants who are not wholly breast-fedrepresentsat least the sameamount of each nutrient from formulas and other foods as the wholly breast-fedinfant of the sameage would receive . , . . . the efficiency of absorption of energy and some nutrients from formulas is less than from breast-milk and some adjustment may be needed. Thus in some cases the DRVs for infants aged O-3 months who are formula-fed are in excess of those which might be expected to be derived from breastmilk. This should not be taken to mean that provision of these nutrients from breastmilk is unsatisfactory [9].” A study in Cambridge, England [lo] of the growth of 48 infants who received breastmilk for at least six months, carried out test weighings of the infants before and after each feed to estimate their milk intake. From the results, Whitehead and Paul were able to show that the previously estimated dietary requirements in the first 8 months of life were grossly in excess of the measured intakes of this group of breast-fed infants. The infants’ growth performance was good relative to the American growth standardsfor weight and height produced by the National Centre for Health Statistics (NCHS), only slowing a little after 4 months. Similar conclusions were reached by Lucas et al. [ 111 who estimated the energy expenditure of breast-fed infants using doubly-labelled water. The energy intakes of these infants were about 20% less than the recommendedvalues, but the mean weight, length and head circumference of the infants lay above the NCHS 50th centile at 5 and 11 weeks. The results of the Lucas study were supported by a meta-analysis of all published doubly-labelled water studies of energy expenditure by healthy children [ 121,which had a combined sample size of 355. This yielded energy requirements of 110, 95, 85, 83, 83, 84 and 85 kcal/kg/day at 1, 3, 6, 9, 12, 24 and 36 months. considerably less than the FAO/WHO/UNU recommendeddietary allowances [I 31 which originated from a meeting held in 1981. A study of 161 well-nourished infants of normal length over the first 4 months of life took place in the Netherlands [14]. Mothers were interviewed about the diet of their young babies, and the amount of breast milk ingested was obtained by test-weighing the child before and after each feed. Growth of the child up to 4 months of age was, in the main, determined by the sex of the child, birth weight, smoking habits of the mother and energy intake; boys and infants of smoking mothers grew faster than girls and infants of non-smoking mothers. The energy intake was lower in breast-fed infants than in bottle-fed infants and the breast-fed infants showed lower weight gain than the bottle-fed infants. 3. Growth related to the feeding of colostrum An immediate difference between breast- and bottle-fed babies is seen in the first week of life. A prospective study in Scotland [15] in 1987 to 1988 of 150 babies that had been born at term and weighed 2500 g or more showed that the proportion of the birthweight that had been lost by 5 days post delivery was 6.6% among the 63 breast-fed babies and 3.1% among the 51 formula-fed babies (p <0.0005); this occurred in spite of the modem practice of feeding shortly after delivery. The S160 IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S174 difference may reflect the time taken to fully establish breast feeding, or the fact that colostrum is about 15 kcals/ 100 ml lower in energy than formula. In many cultures colostrum is not fed to the baby. This is thought to be disadvantageousbut very little evidence is available. In Guinea-Bissau, a study that followed 734 children showed that those with a delayed onset of breast feeding (and therefore less colostrum ingested) did not differ with regard to weight gain, infectious disease,or mortality up to 3 years of age [16]. 4. Growth of breast fed infants in the developed world In Western Australia, a study of 394 healthy infants [17] found that, between birth and three months, weight gains were similar regardlessof feeding method. From 3 to 6 months of age weight gains were greater in infants who were bottle-fed from birth (or breast fed for only a short time) when comparedwith those that were exclusively breast-fed for 6 months or longer. No information on linear growth was given. The difference between the breast- and bottle-fed infants continued up to 12 months of age but the study did not follow them any further. Other studies have been concernedwith comparisonsof growth over a much longer period of time. In the USA the DARLING Study, followed 80 infants to 18 months [18]. The mean weight of the formula-fed infants remained at or above the NCHS growth standardsmedian throughout the first 18 months, but the mean weight of the breast-fed infants dropped below the median from 6 to 8 months onwards; between 6 and 18 months it was significantly lower than the formula-fed group. In contrast however, the length and the head circumference of the two groups were similar throughout. This study of high socio-economic status families in California suggests that current growth charts are probably inappropriate for breast-fed children. The growth of the 48 Cambridge infants originally studied by Whitehead and Paul [lo] has now been followed over several years. By the second month their weights were above the NCHS 50th centile, with a peak relative to the growth standardsat 3 months. After this point there was a relative decline acrossthe centiles until the tenth month. Linear growth showed a similar pattern although the decline across the centiles was smaller. The adiposity of the infants (as measuredby triceps skinfold) differed more widely from the growth standardsthan did weight and length, being only equivalent to the 10th centile from 2 to 10 months [19]. The relatively slower growth in weight and length continued into the secondyear of life [6], however, by 3 years of age both weight and length were almost back to the 50th centile [20]. Using data from the British cohort of 1946, Douglas [21] looked at growth as measuredby the length and weight of the child up to the age of 2 years. He found that although the breast-fed child was the same length as the bottle-fed child at 2 years, the bottle-fed child weighed more. It is uncertain how far these results would apply to formula-fed children today, becausethe composition of the formula milks used in the 1940s was very different from those used today. It is a common problem when comparing the growth of breast- and formula-fed infants that papersdo not give the name of the formula used. Pre-1974 in the UK and IS. Rogers et al. I Early Human Development 49 Suppl (I.9971 SIJ7-S174 most Commonwealth countries ‘formula milk’ was just roller dried cows’ milk. Post-1974 major changes were made to formula milks to bring their composition closer to that of breast milk. This resulted in major changesin, for example, protein intakes, which may well have had an effect on the growth patterns of formula-fed infants. A study in Saudi Arabia suggests [22] that young children tend to be stunted relative to NCHS standards, despite a relatively affluent and presumably wellnourished population. The lengths and weights of 400 children aged from 6 to 24 months attending a well-baby clinic were measured.Seventeenpercent of the children were stunted i.e. more than 2 SD. below the NCHS standardsfor height-for-age. In multiple regression analysis bottle-fed children were less likely to be stunted (p <: 0.05), while children fed starchy solid foods and older children were more likely to be stunted (p < 0.01 and p < 0.001 respectively). The duration of breast feeding had no effect. Theseresults are difficult to interpret without more detail on the actual food intakes of the children. It may be that genetic factors play an important part in their short stature. However, the authors suggestedthat inappropriate use of weaning foods was a major factor. 5. Growth of infants related to the nutritional status of the mother When the mothers are well-nourished, breast-fed infants show satisfactory weight gains for the first four to six months, after which time the weight curve flattens unless supplementary foods are introduced. As suggestedearlier in this paper, appropriate growth curves for breast-fed infants may be less than current standards.Whitehead and Paul [6] have compared the growth curves of infants from various parts of the developing world with both American NCHS standardsand the growth curves of their group of well-nourished breast-fed infants from Cambridge. In each case the growth curves deviated from the NCHS standardsearlier than from the growth curves of the Cambridge infants. (Like the breast-fed infants in the DARLING Study the Cambridge infants were from affluent families. It seems likely that their apparent growth faltering relative to the NCHS standardsreflects the fact that these standards are inappropriate for breast-fed infants.) However, studies in various parts of the world where a large proportion of the population is malnourished suggest that the growth curves of breast-fedinfants may falter after 3 or 4 months indicating that the milk production of the mothers may be inadequate in some way after this point. Work carried out in Bolivia highlighted an association between growth and maternal nutritional status [23]. It showed infant weight gain between 3 and 6 months to be correlated with maternal anthropometry, both during pregnancy (34 weeks gestation) and at 3 months post-partum. However, this study did not control fog confounding variables such as socio-economic status or episodes of illness. These factors were considered in a study of 339 children aged between 3 and 36 months in Bangladesh [24]. It showed that the child’s nutritional status as indicated by its weight for age (as a percentageof the NCHS median) was associatedwith the body massindex of the mother (p < O.OOl),the socio-economic status of the family S162 IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S174 (p < 0.001) and whether or not the child was being breast fed. The featuresremained significant in a multiple regression analysis allowing for the mother’s body mass index, socio-economic status,the age of the child, mother’s education and the child’s birthrank (p = 0.0011). Butte et al. studied the milk and supplementary food intake of 30 rural Mesoamerindian infants at 4 and 6 months of age in relation to their growth [25]. The growth of these infants faltered at 6 months relative to NCHS standards and to a referencesampleof North American breast-fedinfants. However, energy intakes were similar to those of breast-fed infants from more privileged environments [26], and weight and length velocities were not correlated with energy intakes from milk or total food. Although recognising that the lack of correlation between growth and energy might reflect the small sample size, the authors suggestedthat the actual cause of the growth faltering might be micronutrient rather than energy deficiency. Among severely malnourished women there is evidence that the milk output is significantly lower than among well-nourished women. However, the infants of severely malnourished women tend to be of low birth weight. Size appearsto be a major determinant of breast-milk output, and the infants of severely malnourished women have a similar milk intake per kg body weight to those of well-nourished women [8]. The output of milk by moderately malnourished women and its macronutient composition is very similar to that of well-nourished women. Although supplementationof lactating women may result in an improvement in maternal health 1271studies have failed to demonstratea clear effect on milk output [28]. However, supplementation of the mother’s diet has been shown to improve the growth of the infant in several cases.The vitamin and mineral content of the mother’s milk appears to be more responsiveto supplementationthan the macronutrient content or output. It may be that this accounts for the observed improvements in infant growth on maternal supplementation. (The effects of supplementing the mother’s diet on breastmilk composition and volume are described in more detail elsewhere [8].) 6. Growth and nutritional mineral intake status of the infant in relation to vitamin and In certain situations the growth of exclusively breast-fed infants might be limited by the intake of a number of micronutrients. Lactating women with low intakes of vitamin B, may be at risk of supplying inadequate amounts of this vitamin to their infants in their breastmilk. During the first 28 days post-partum, supplementation either of exclusively breast-fed infants or their mother was found to improve the vitamin B, status of the infants. Vitamin B, intake and status was a significant predictor of growth during this period. This implies that the growth of the infants of unsupplementedmothers was restricted by their low vitamin B, intake [29]. A number of cases of rickets have been reported in breast-fed infants [30,31]. These were generally black infants who had received minimal exposure to sunlight, had been breast-fed without supplementationfor more than six months, and often had received only vegetarian foods when supplementationstarted.Although this combina- IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S1?4 Slhi tion of circumstances is extremely unusual in the developed world, it has been recommendedthat all breast-fed infants should be supplementedwith vitamin D, so as to eliminate the possibility of nutritional rickets [32]. The vitamin K content of human milk is very low, and does not seem to be responsiveto maternal supplementation,other than at pharmacological levels 133,341, Thus, breast-fed infants who do not receive supplementary vitamin K are at an increasedrisk of haemorrhagic diseaseof the newborn, a fatal disorder resulting from vitamin K deficiency. However, even in unsupplementedbreast-fedinfants the risk of this disorder remains extremely low especially once synthesis by the intestinal flora has become established. Several studies seem to show that a number of minerals may not be adequately supplied over 6 months by an exclusive diet of breast milk. The iron storesof a term infant becomedepleted by about 4 months of age unless replenished by an adequate supply of iron from the diet [35]. A study comparing exclusively breast- or formula-fed infants found that the breast-fedinfants had a negative iron balance from 3 to 6 months while the formula-fed infants increasedtheir total body iron during this time. The prevalence of anaemiaat 9 months in breast-fedinfants was higher than in formula-fed infants despite the introduction of iron-rich supplemental foods to the diet of breast-fedinfants from 6 months onwards. The authors concluded that the iron in human milk was insufficient to meet the demands for growth, and that supplemental iron should be given from the fourth month [36]. Similarly a longitudinal study of Japaneseinfants found that while the prevalence of anaemia at 4 months oi age and the mean haemoglobin levels were similar in breast- and formula-fed infants. by 9 months the breast-fed infants had a higher prevalence of anaemia and lower mean haemoglobin levels 1371.Several other studies have shown that while breast feeding protects against iron deficiency in comparison to feeding an unfortified cow’s milk formula, the iron status of breast-fedinfants is inferior to those fed iron-fortified formula by the age of about six months [38-401. That this may not be altogether a bad thing is suggestedby a study in Guam 1411 which showed that infants who had been fed a formula with high iron as opposed to low iron content were at substantially greater risk of diarrhoea. Normal full-term infants exclusively breast fed for 4 to 6 months, are unlikely to becomeseverely iron deficient despite the low iron content of human milk. The authors, therefore, postulate that the maintenance of a relatively low concentration of iron in the neonatal gastrointestinal tract may be one of nature’s strategies for protecting infants from bacterial infection during the first months of life. A study of 93 US infants ]42], in contrast, found no difference in the incidence of gastrointestinal symptoms between those fed formulas with and without iron fortification. However, the level of exposure to infectious organisms of the infants in this study is likely to have been very different to that in Guam. The zinc concentration of human milk falls markedly over the first few months of lactation. Zinc deficiency is known to result in growth faltering, and there has been concern about whether the zinc intake of breast-fed infants is adequatefor growth. The zinc requirements of premature infants are higher than those of full-term infants. and a few cases of acquired zinc deficiency have been reported in breast-fed S164 IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S174 premature babies [43-451. However, the evidence of the relationship between zinc intake and growth in breast-fed infants is inconsistent. In a longitudinal study of eight Brazilian mothers and their breast-fedinfants [46] the zinc concentration of the breast milk was found to be a significant predictor of both length and weight in the first six months (while fat concentration was not). Krebs et al. [47] followed 71 American infants for 9 months and found tenuous evidence of a relationship between growth and zinc intake in breast-fed infants. For those infants in the lowest quartile stepwisemultiple regression indicated there was a significant relationship between the changein weight-for-age z scorebetween 5 and 7 months and the percentageof zinc intake from human milk at 5 months. However, there was no significant associationbetween zinc intake and changesin length-for-age at any age. A study of Bangladeshi infants from birth to 12 months [48] estimated their zinc intake from breast milk to be only lo-30% of the American recommendeddietary allowance (RDA) [49]. The authors suggestedthat the apparently sub-optimal dietary intake of zinc by these infants might be partly responsible for their poor growth. However, although the zinc concentration of the mother’s milk was measuredin this study the actual milk intake by the infants was not. Walravens et al, [50] found that after 3 months of zinc supplementation the rate of growth of a group of breast-fed infants had increased relative to a placebo group. However, these were not exclusively breast-fed infants. The supplementaryfoods given to the test and placebo groups were stated to be no different, but their composition in terms of zinc, phytate and other factors influencing zinc absorption was not described. A study in Finland [51] found zinc supplementationof lactating mothers to have no effect on the growth of their breast-fed infants, and no evidence of a relationship between zinc intake and growth. In the breast-fed cohort in the DARLING study [52] there was no association between breast-milk zinc concentration or the intake of the infant and growth from 6 to 12 months; indeed from 3 to 6 months there had been a negative association with growth. This is contrary to expectation if zinc intake is a limiting factor in the growth of breast-fed infants. Krebs and Hambridge estimated the zinc requirements of infants allowing for growth and the replacement of zinc lost in the urine and sweat [53]. These estimated requirements were strongly affected by growth velocity, and declined from a high of 780 pg per day in male infants at 1 month to 480 pg per day in the fifth month. (The zinc concentration of human milk declines over time in a similar way to these estimated requirements.) They suggested that the American RDAs for zinc [49] (which are 66% higher from 6 to 12 than from 0 to 6 months) are inappropriate. Assuming a high bioavailability of zinc from human milk they calculated that the intake of zinc by breast-fed infants should be sufficient to meet growth requirements until at least 5 to 6 months. 7. The effect of supplementary solids on growth The ability of breast feeding to support optimal growth of the infant at about 6 months of age without significant caloric supplementationremains a matter of debate, IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-Sl74 particularly where the nutritional status of the mother is poor. However me introduction of supplementary feeding at or before six months does not necessarily increase the nutrients available to the infant. The feeding of foods other than breastmilk can introduce the risk of contamination and gastrointestinal infectious ]54]. As has been indicated such infections may have a negative effect on the nutritional status of the infant which outweighs any benefit from the food. in addition it has been suggestedthat the introduction of supplementary food may compromise growth by reducing the intake of breastmilk and effectively replacing a high quality food with a low quality one [55]. Furthermore, it appearsthat feeding breast milk in combination with solids reduces the bioavailability of nutrients such as iron in the breast milk [56]. Stuff and Nichols [57] found that the introduction of solid food to breast-Fed infants older than 16 weeks reduced their milk consumption and resulted in no overall increase in energy intake/kg bodyweight. There were similar results in the DARLING study [58]. Both these studies were conducted on relatively advantaged populations in the USA, however, a recent study on a group of low-income women in Honduras also suggested that infants self-regulate their energy intake on the introduction of solid foods [59]. Women who were exclusively breast feeding at 4 months were randomly assigned to either continue exclusive breast feeding until 6 months, to give solid foods and to nurse ad libitum until 6 months, or to give solid foods and maintain the baseline nursing frequency. There was no significant difference between the three groups in weight and length gain or energy intake of the infants. In contrast, a study in Indonesia [60] where it is common to force-feed infants, from as early as the third day after birth, found no relationship between energy intake from breastmilk and from additional foods at ages 1-24 weeks, In this population breast-milk intake was apparently unaffected by supplementary feeding. Studies relating infant feeding practices and growth in different populations do not have consistent results. The Indonesian study did not find a significant correlation between the feeding of additional foods and weight from 0 to 12 months, except at 2 weeks when there was a negative correlation [61]. In a study of infants in rural China the duration of exclusive breast feeding was found to be positively associatedwith growth, suggesting a negative effect on growth of early supplementary feeding 1621. The feeding of rice flour before the age of 7 months was negatively associatedwith growth (rice flour is very low in protein). However, the feeding of soybeanmilk, liver and pork blood products more than once a week from 7 to 9 months was positively associatedwith growth. In Yemeni children between 3 and 12 months of age, there was no association between the introduction of supplementary food and nutritional status (as measuredby weight-for-length, length-for-age or weight-for-age) 1631.A longitudinal study of 45 infants in Dundee, Scotland found an association of early feeding of solids with greater weight at 8, 13 and 26 weeks 1641.However, the association was not found after the age of 26 weeks, and it appearedthat the weight difference precededthe introduction of solid foods, rather than being causedby it. In California. Nommsen et al. 1651found no difference between the weight-for-age, weight-for-length and length-for-age z scoresof infants, at age 1 year, introduced to solids before 23 weeks and those introduced after 23 weeks. In rural Filipino infants younger than 6 months, breastfeedinghad a strong positive S166 IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S174 effect on growth [55]. This positive effect was diminished by supplementaryfeeding, even after adjusting for the incidence of diarrhoea. Thus it seems likely that the growth reduction associated with supplementary feeding reflected a trading of breast-milk intake for inferior weaning foods. However, in Ghanaian infants the growth from 0 to 6 months was faster in those given a traditional fermented weaning porridge as a supplementaryfood than in those who were exclusively breast fed [66]. (The fermenting of weaning porridge reduces its bulk and increases the bioavailability of nutrients. The drop in pH during fermentation may also make the porridge microbiologically cleaner, reducing the risk of diarrhoeal infection.) The growth advantageof feeding the weaning porridge seemedto be particularly large where the breast-milk intake of the infant was low. In rural Thai infants, from 0 to 6 months of age, breast-milk intake in grams and energy intake from breast-milk alone were significant predictors of weight while energy intake from supplementswas not [67]. However, in infants older than 6 months weaned infants were heavier and ate more protein than unweaned infants. This reflected the low protein concentration of the breast milk as compared to the weaning food. A longitudinal study of Bedouin Arab infants found that infants who were breast fed (exclusively or in combination with other foods) had a reduced rate of stunting at age 6 months compared with weaned infants [68]. However, considering only the exclusively breast-fed infants the relative risk of stunting at 6 months was not significantly different from one, reflecting either a beneficial effect of supplementary feeding or the small size of the exclusively breast-fedgroup. Infant feeding practices from 6 to 12 months were not associatedwith stunting, but infants who were stunted at 6 months were at a greater risk of being stunted at 12 months. A study by Lutter et al. [69] found the effect of supplementary feeding on the growth of breast-fed children to be highly dependenton its timing. They fed a high energy supplement to children at risk of malnutrition over the first year of life, and compared their growth to that of unsupplementedcontrols. Supplementationhad the greatest effect on growth between 3 and 6 months (the period of weaning) and between 9 and 12 months (the time of peak diarrhoea incidence). However, as the mothers themselves were also supplementedduring pregnancy and lactation it was not possible to ascribe the growth advantage purely to the supplementary food received by the child. Furthermore, the authors cautioned against the risk of disrupting breastfeedingpatterns when supplementaryfeeding is introduced before 6 months. 8. Growth of the child breast-fed beyond 12 months Beyond the age of 6 months the nutrient supply from the breast milk alone is no longer adequate to support growth and supplementary feeding becomes necessary. Very few children in the West are breast fed at all after their first birthday. However, breastfeedingwell beyond this age is common in much of the developing world. This might be thought to be advantageousto the child, providing a high quality source of food. For example, breast milk made an important contribution to the fat intake of I.S. Rogers et al. / Early Human Development 49 Suppl (1997) Sl.V-Sl74 Slh? 16month old Dutch macrobiotic infants [70]. In addition the duration of breastfeeding is positively correlated to the interval to the next conception 171J. Thus, in deprived communities, breast feeding might indirectly improve the food supply to the child by increasing birth spacing and thus reducing the pressureon family resources. The extremely high bioavailabity of the nutrients in breast milk means the actual contribution to the diet may be higher than a simple comparison of the intakes of breast milk and weaning foods would suggest[72]. During infections breast milk may play an especially important role in the diet as breast milk intakes seem to be maintained while the consumption of weaning foods can fall drastically. A study of children suffering from diarrhoea in Bangladesh found that those who were still breast fed consumed 77.0 kcal/kg/day as compared to 49.9 kcal/kg/day in those who were completely weaned [73]. Studies in Malawi [74]. Indonesia [75] and Bangladesh[76] have found the incidence of xerophthalmia to be reduced in children still receiving breast milk. However, an apparently contradictory association between prolonged breast feeding and malnutrition has been reported from several surveys. Victora et al. 1771 examined the association between duration of breast feeding and nutritional status in a sample of 802 Brazilian children aged from 12 to 35.9 months. The prevalence of malnutrition was lowest in children breast fed for 3 to 6 months, but breast feeding after 6 months was negatively associated with length-for-age. Furthermore. the prevalence of malnutrition (low weight-for-length) was considerably higher amongst those children still being breast fed, even after controlling for several confounding factors. These included age of the child, district of residence, family income, ethnic background and hospitalisation for an infectious disease. However, the authors did not investigate the use of preventive health care services or the effects of illnesses not requiring hospitalisation. Brakohiapa et al. found breastfeeding beyond 19 months to be associated with malnutrition in a group of children visiting a hospital in Ghana [78]. However. socio-economic status was not controlled for in the analysis. In a subsample of malnourished children total protein and energy intake was observedto rise sharply on weaning. It has been suggestedthat the increased food intake of the children may have reflected a change in feeding behaviour by the mother rather than an increase in their appetite [79,80]; the children were not observed to reject supplementary focxl prior to weaning. Mothers in Mali have also reported that their children eat much more and often show improved growth on weaning [81]. However, extra food was given to the child at the time of stopping breast feeding, partly to replace the breast milk and partly as a distraction from nursing; in some casesthis went on for as long as 6 months. A study was made of 510 children living in a rural area of Nepal, aged 3- 10 years at the time of follow up [82]. These children were breast fed for an average of 35 months (S.D. 11.3 months), and the duration of exclusive breast feeding was on average 18.2 months (S.D. 7.7 months); the children had severeevidence of stunting; 65% were less than the 10th centile of height-for-age in the NCHS growth standards. They were also between 1 and 1.5 kilograms lighter than U.S. children of the same height. The authors carried out multiple regression anaIysis and showed significant S168 IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S174 associationswith the age of the child, the area in which the child lived, the household income and a number of dietary factors including the duration of breast feeding. Prolonged breast feeding was associated with greater fat stores but with reduced stature and low haemoglobin levels. Although a number of other studies have also found an association between prolonged breast feeding and malnutrition [83] others have found breast feeding beyond one year to be positively associated with good nutritional status. A prospective study of growth in the Philippines found breast feeding to enhance growth from 6 to 24 months. This effect remained even when the reduced incidence of diarrhoea in breast-fedchildren was allowed for [55]. Briend et al. found a greater mean arm-circumference for age in breast-fed than in weaned Bangladeshi children 1841. (It is interesting that in a study of American children Agras et al. found breastfeedingbeyond 5 months to be associatedwith greater adiposity at 6 years of age [85].) Furthermore, some studies have found no association between prolonged breastfeedingand growth, even after allowing for socio-economic variables. Included among these are studies on groups of Bedouin Arabs [68] and Yemeni children [63]. 9. Growth and illness Just how much normal growth is restricted because of ill-health (whether in the developed or developing world) is still unclear, although some good studies have been published. For example, a longitudinal study of growth among 910 infants in a poor-urban population of Pakistan [86] revealed that although all mothers started breast feeding, 50% of mothers introduced a breast-milk substitute from one month of age (this was generally buffalo milk). 75% of children had had at least one episodeof diarrhoea and 60% had had at least two. The study found no association between the growth of the child and the age at weaning, but there was a strong relationship between growth and the total number of episodes of diarrhoea using multiple regression analysis. A study in the Gambia followed a cohort of 126 newborns for the first 2 years of life [87]. For the first 6 months, the weight-for-age exceededthe NCHS standardsbut after this there was a falling off. Two diseasescontributed to the weight faltering; diarrhoeal diseases were estimated to cause one half of the deficit and lower respiratory tract infections one quarter. 115 of these children had known dates of weaning, and the effect of diarrhoea on growth was shown to be strongest in those children who had been weaned (- 14.4 grams per day of illness) compared with the breast-fed ones (-3.6 grams per day). The difference between the two was statistically significant (p < 0.01). During the second year of life, growth was shown to be essentially normal, and unrelated to the frequency of infections. In Indonesia 33 breast-fed infants were followed at 3 weekly intervals [88]. The children were divided according to the amount of time the mother spent breast feeding. Only at the lowest level of breast feeding did illness have any negative impact on the growth of the child. It should be noted that in this study sample diarrhoeal diseasewas far less of a problem than respiratory illness. An interesting study of diarrhoea and its effect on nutrition was undertaken in I.S. Rogers et al. I Early Human Development 49 Suppl (1997) Sf57-S174 s I h,J Nigeria [89]. In a community based study 45 children who had been identified with diarrhoea had their diets examined during and after the diarrhoeal episodes. The authors showed that the energy intakes during diarrhoea (85.4 kcallkglday), were slightly lower than during health (95.9 kcal/kg/day, p < 0.05) but that there were no differences in the frequency or duration of breast feeding during this time. They point out that the magnitude of the difference in the children’s daily intake associatedwith illness was small relative to the deficit in their intake when comparedto the amounts recommended for this age group. Similarly a study of pre-school children in Guatemala [90] showed a reduced intake of both energy and protein during illness, particularly if diarrhoea was present.The authors of a separateGuatemalanstudy [91] describedthe anorexia exhibited by children with infection. They reported that weight loss and height arrest was associatedwith a variety of infections, and indicatd that although the finding was present in well nourished children it was more pronounced in those who were already growth retarded. The long-term effects of such growth faltering have, however, been questioned by a study from the International Centre for Diarrhoeal Researchin Bangladesh [92]. By studying 230 children at weekly intervals between 6 and 35 months, the authors were able to confirm growth faltering at the time of a diarrhoeal episode, but they found that this was rapidly followed by a growth spurt so that overall there was no deficit in the attained weight and height of these children a few weeks after the episode. 10. Reverse causality Those studies which have shown an association between breast feeding and malnutrition are generally cross-sectional [83]. Such studies do not allow proper investigation of the possibility of reverse causality. In many societies children are weanedon reaching certain developmental stagessuch as walking independently 1931, or when they are ‘sturdy and healthy’ [94]. The weaning of sickly children may be delayed, creating a non-causal association of breast feeding and malnutrition. Ghanaian children who were sickly have been reported to be weanedlater than those who could walk and talk [95]. Briend and Bari found that on average Bangladeshi children had a lower weight-for-age when breast-fed than when weaned [96]. However, the difference although statistically significant was very small suggestingit could have arisen from confounding factors. In addition, children, on average,had a higher weight-for-age when they were about to be weaned than when they were not, suggesting that the difference in nutritional status occurred prior to weaning. A longitudinal study of 849 children in West Africa found a higher prevalence of malnutrition in breast-fed than in weaned children [97]. However, in 162 children weaned during the survey there was no change in nutritional status on weaning. Furthermore children with low weight-for-age were breast fed for longer than the rest suggesting that the association between breast feeding and malnutrition in this group arose by selection. (Several studies have also found that larger infants tend to be weaned earlier as has already been described.) Furthermore, many of the studies reporting an association between prolonged breast feeding and malnutrition have not s170 IS. Rogers ei al. 1 Early Human Development 49 Suppl (1997) S157-Sl74 allowed for social class (lower social class being associated both with prolonged breast feeding and with malnutrition in developing countries). 11. Discussion The relationship between nutrient intake and growth is not an easy one to investigate. In the first 4 months of life the growth component of energy expenditure is large but by 6 months it is much less important and by 12 months it is minimal. Furthermore, many factors including heredity, altitude, disease and malabsorption may affect growth independent of energy or nutrient intake. The studies do not demonstratea consistent effect of supplementaryfeeding on infant growth. It seems likely that before 6 months the milk supply of a well-nourished mother is sufficient to support adequate growth and that any increase in growth observed on supplementation is unlikely to represent a functional advantage. Where the mother is malnourished the growth of her breast-fed infant may falter from about 3 months. However, most studies in the developing world have failed to show a growth advantage of supplementary feeding of infants younger than 6 months. After 6 months of age the growth advantageof supplementaryfeeding will be dependenton a number of factors, including the nutritional quality of the weaning food, the effect on the breast milk intake of the child, and the risk of introducing the child to infection. Although it has been reported anecdotally an anorectic effect of partial breast feeding has not been conclusively demonstrated.The increasedfood intake which has been observedon weaning could well arise from a changein feeding behaviour by the mother. Furthermore, breast feeding is known to protect against diarrhoea and other infections, a major cause of short-term growth faltering. It is probable that the observed association between prolonged breast feeding and malnutrition is not a causal one, but is likely to have arisen from confounding factors such as poverty and the delayed weaning of sickly children. The way in which growth varies within the individual is well recognised.From the evidence presentedhere there is someindication that infections in the first year of life do cause a faltering of growth, but whether this is caught up later is unclear. One major question concerns the advice offered to mothers on how long to breast feed and what the viable alternatives are. As discussedin a recent editorial [98], there is still considerabledebateas to the appropriatetime for starting supplementation,and at what age children should be weaned, with some strong opinions being expressed. One school of thought is that women should continue to breast feed and delay the introduction of supplementary food regardless of whether the child becomes undernourished or marasmic, in order to reduce the risk of diarrhoeal disease.Others think this advice is totally inappropriate. References [l] Wells JCK, Davies PSW.The effect of diet and sex on sleeping metabolic rate in 12-week-oldinfants. Eur J Clin Nutr 1995;49:329-35. IS. Rogers et al. I Early Human Development 49 Suppl 11997) Sl57-3174 s I7I [2] Butte NF, O’Brian SE, Garza A. Energy utilisation of breast-fed and formula-fed infants. Am J Clirt Nutr 1990;51:350-8. [3] Omolulu A. Breast-feeding practice and breast-milk intake in rural Nigeria. Hum Nutr: Appl Nun 1982;36A:445-51. 141Ounsted M, Sleigh G. The infant’s self-regulation of food intake and weight gain. Lancer 1975;i:1393-405. [S] Michaelsen KF, Larsen PS, Thomsen BL et al. The Copenhagencohort study on infant nutrition and growth: breast-milk intake, human milk macronutrient content, and influencing factors. Am J Clin Nutr 1994;59:600-11. 161WhiteheadRG, Paul AA. Growth charts and the assessmentof infant feeding practices in the western world and in developing countries. Early Hum Dev 1984;9:187-207. 171Neville MC, Keller R, Seacat J et al. Studies in human lactation: milk volume in lactating women during the onset of lactation and full lactation. Am J Clin Nun 1988;48:1375-86. [8] Emmett PM, RogersIS. Properties of human milk and their relationship with maternal nutrition. Early Hum Dev 1997;49 Suppl:S7-S28. [9] Department of Health, Dietary Reference Values for Food Energy and Nutrients for the United Kingdom. London: HMSO, 1991. [lo] Whitehead RG, Paul A. Infant growth and human milk requirements. Lancet 1981;i:161-3. [l 11 Lucas A, Ewing G, Roberts SB et al. How much energy does the breast fed infant consume and expend? Br Med J 1987;295:75-7. 1121Prentice AM, Lucas A,Vasquez-VelasquezL et al. Are current dietary guidelines for young children a prescription for overfeeding? Lancet 1988;31:1066-9. [13] FAO/WHO/ENU. Energy and protein requirements WHO Tech. Rep. Series 724. Geneva: WHO, 198.5. 1141Hoffmans MDAF, Obermann-DeBoer GL, Florack EIM et al. Determinantsof Growth During Early Infancy. Hum Biol 1988;60:237-49. [15] Salariya EM, RobertsonCM. Relationships between baby feeding types and patterns,gut transit time of meconium and the incidence of neonataljaundice. Midwifery 1993;9:235-42. [ 161 GunnlaugssonG, da Silva MC, SmedmanL. Age at breast feeding start and post-natal growth and survival. Arch Dis Child 1993;69:134-7. [17] Hitchcock NE, Gracey M, Gilmour AI. The growth of breast fed and artificially fed infants from birth to twelve months. Acta Paediatr Stand 1985;74:240-5. [ l8J Dewey KG, Heinig MJ, Nonnnsen LA et al. Growth of breast fed and formula fed infants from 0 to 18 months: The DARLING study. Pediatrics 1992;89:1035-41. [ 191 Paul AA, Black AE, Evans J et al. Breastmilk intake and growth in infants from two to ten months.J Hum Nutr Diet 1988;1:437-50. (201 Paul AA, Black AE, Whitehead RG. Follow up of energy intake and growth of children initiall) breastfed. Proc Nutr Sot 1988;47:92A. 1211Douglas JWB. Breast feeding in Great Britain. Part II. Infant feeding in relation to survival health and development. .J Ob Gynaecol Br Commonwealth 1950;57:349-61. [22] Al-Fawaz IM, Bamgboye EA, AI-Eissa YA. Factors influencing linear growth in Saudi Arabian children aged 6-24 months. J Trop Pediatr 1994;40:235-9. 1231Novotny R, Haas JD. Maternal anthropometry and infant growth with exclusive breast feeding in La Paz, Bolivia. J Trop Pediatr 1987;33:309-14. 1241Rahman M, Roy SK, Ali M et al. Maternal nutritional status as a determinant of child health. J Trap Pediatr 1993;39:86-8. (251 Butte NF, Villalpando S, Wong WW et al. Human milk intake and growth faltering of rural Mesoamerindian infants. Am J Clin Nutr 1992;55:1109-16. (261 Heinig MJ. NommsenLA, PeersonJM et al. Energy and protein intakes of breast-fedand formula-fed infants during the first year of life and their associationwith growth velocity: the DARLING Study. Am J Clin Nutr 1993;58:152-61. (271 Prentice AM, Lunn PG, WatkinsonM et al. Dietary supplementationof lactating Gambian women. 11. Effect on maternal health, nutritional statusand biochemistry. Hum Nutr: Clin Nutr 1983;37(3:65-74. [28] Prentice AM. Can maternal dietary supplementshelp in preventing infant malnutrition? Acta Paediatr Scaod 1991;374Suppl:S67-S77. S172 IS. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S174 [29] Kang-Yoon SA, Kirksey A, Giacoia G et al. Vitamin B-6 status of breast-fed neonates:influence of pyridoxine supplementationon mothers and neonates.Am .I Clin Nutr 1992;56:548-58. [30] Bachrach S, Fisher J, Parks JS. An outbreak of vitamin D deficiency rickets in a susceptible population. Pediatrics 1979;64:871-7. [31] Edidin DV, Levitsky LL, Schey W et al. Resurgenceof nutritional rickets associatedwith breastfeeding and special dietary practices. Pediatrics 1980;65:232-5. [32] Anon . Rickets in a breast-fed infant. Nutr Rev 1984;42:380-2. [33] Greer FR, Marshall S, Cheny J et al. Vitamin K Status of lactating mothers, human milk, and breast-feeding infants. Pediatrics 1991;88:751-6. [34] Pietschnig B, Haschke F, Vanura H et al. Vitamin K in breast milk: no influence of maternal dietary intake. Eur J Clin Nutr 1993;47:209-15. [35] International Nutritional Anaemia Consultative Group. Iron deficiency in infancy and childhood. New York, NY: The Nutrition Foundation, 1979. [36] Calvo EB, Galindo AC, Aspres NB. Iron status in exclusively breast-fed infants. Pediatrics 1992;90:375-9. 1371Hokama T. A study of the hemoglobin levels in breast-fed infants in one village of Okinawa prefecture. Acta Paediatr Jap 1993;35:138-40. [38] Pizarro F, Yip R, Dallman PR et al. Iron status with different infant feeding regimens: Relevance to screening and prevention of iron deficiency. J Pediatr 1991;118:687-92. [39] Saarinen UM, Siimes MA. Iron absorption from breast milk, cow’s milk, and iron-supplemented formula: An opportunistic use of changesin total body iron determined by hemoglobin, ferritin and body weight in 132 infants. Pediatr Res 1979;13:143-7. [40] Garry PJ, Owen GM, Hooper EM et al. Iron absorption from human milk and formula with and without iron supplementation.Pediatr Res 1981;15:822-8. [41] Haddock RL, Cousens SN, Guzman CC. Infant diet and salmonellosis. Am J Pub1 Health 1991;81:997-1000. [42] Oski FA. Iron-fortified formulas and gastrointestinal symptoms in infants: a controlled study. Pediatrics 1980;66:168-70. [43] Connors TJ, Czamecki DB, Haskett MI. Acquired zinc deficiency in a breast-fed premature infant. Arch Dermatol 1983;119:319-21. [44] Aggett PJ, Atherton DJ. Symptomatic zinc deficiency in breast-fedpremature infants. Arch Dis Child 1980;55:547-50. [45] Zimmerman AW, Hambridge MK. Low zinc in mother’s milk and deficiency syndromein breast-fed premature infants (abstracted).Pediatr Res 1980;14:518. [46] Doer JG. Is zinc a rate limiting nutrient in human milk? Nutr Res 1993;13:659-66. [47) Krebs NF, Reidinger CJ, Robertson AD et al. Growth and intakes of energy and zinc in infants fed human milk. J Pediatr 1994;124:32-9. 1481Simmer K, Ahmed S, CarlssonL et al. Breast milk zinc and copper concentrationsin Bangladesh.Br J Nun 1990;63:91-6. [49] Committee on Dietary Allowances. Food and Nutrition Board, Commissionon Life SciencesNational Research Council. Recommended Dietary Allowances, 9th. issue. Washington, DC: National Academy Press, 1980. [50] Walravens PA, Chakar A, Mokni R et al. Zinc supplements in breastfed infants. Lancet 1992;340:683-5. [51] SalmenperaL, PerheentupaJ, Nanto V et al. Low zinc intake during exclusive breast-feedingdoesnot impair growth. J Pediatr GastroenterolNutr 1994;18:361-70. [521 Dewey KG, Heinig MJ, Nommsen LA et al. Zinc statusin breastfed infants. Lancet 1992;340:1417. 1531Krebs NF, Hambridge KM. Zinc requirementsand zinc intakes of breast-fed infants. Am J Clin Nutr 1986;43:288-92. [54] Golding J, Emmett PM. Gastroenteritis, diarrhoea and breast feeding. Early Hum Dev 1997;49 suppl:s83-s1o3. [55] Adair L, Popkin BM, VanDerslice J et al. Growth dynamics during the first two years of life: a prospective study in the Philippines. Eur J Clin Nutr 1993;47:42-51. IS. Rogers et al. I Early Human Development 49 Suppl (1997) SI57-S174 [56] O&i FA, Landaw SA. Inhibition of iron absorption from human milk by baby food. Am J Dis Child 1980: 134:459-60. 1.571Stuff JE, Nichols BL. Nutrient intake and growth performance of older infants fed human milk .I Pediatr 1989;115:959-68. 1581 Heinig MJ, Nommsen LA, Peerson J et al. Intake and growth of breast-fed and formula-fed infants in relation to the timing of introduction of complementary foods: the DARLING Study. Acta Paediatr Stand 1993;82:999-1006. I.591 Cohen RJ, Brown KH, Canahuati J et al. Effects of age of introduction of complementary foods on infant breast milk intake, total energy intake, and growth: a randomised intervention study !.II Honduras. Lancet 1994;343:288-93. [60] van Steenbergen WM, Kusin JA, Kardjati S et al. Nutritional transition during infancy in East Java. Indonesia: 1. A longitudinal study of feeding pattern, breast milk intake and the consumption of additional foods. Eur J Clin Nutr 1991;45:67-75. [61] Kusin JA, Kardjati S, van Steenbergen WM et al. Nutritional transition during infancy in East Java. Indonesia: 2. A longitudinal study of growth in relation to the intake of breast milk and additional foods. Eur J Clin Nutr 1991;45:77-84. [62] Guldan GS, Zhang M, Zhang Y-P et al. Weaning practices and growth in rural Sichuan infants: 1~ positive deviance study. J Trop Pediatr 1993;39:168-75. [63] Jumaan AO, Serdula MK, Williamson DF et al. Feeding practices and growth in Yemeni children. .I Trop Pediatr 1989;35:82-6. [64] Forsyth JS, Ogston SA, Clark AC et al. Relation between early introduction of solid food to infants and their weight and illnesses during the first two years of life. Br Med J 1993;306:1572-6. 16.51Nommsen LA, Heinig MJ, Lonnerdal B. Appropriate timing of complementary feeding of breast-fed infants. FASEB 1989:3:A1054. 1661 Armar-Klemesu MA, Wheeler EF, Brakohiapa LA et al. Infant feeding and growth in rural Ghana: Is the use of the traditional fermented porridge a case for early supplementation? J Trop Pediatr 1991;37:111-5. [67] Jackson DA, Imong SM, Silprasert A. Infant weight in relation to nutritional intake and morbidity in northern Thailand. Eur J Clin Nutr 1988;42:72.5-39. [68] Forman MR, Guptill KS, Chang DN et al. Undernutrition among Bedouin Arab infants: the Bedouin Infant Feeding Study. Am J Clin Nutr 1990;51:343-9. [69] Lutter CK, Mora JO, Habicht J-P et al. Age-specific responsiveness of weight and length to nutritional supplementation. Am J Clin Nutr 1990;51:359-64. [70] Dagnelie PC, van Staveren WA, Verschuren SAJM et al. Nutritional status of infants aged 4 to 18 months on macrobiotic diets and matched omnivorous control infants: a population-based mixedlongitudinal study. Eur J Clin Nutr 1989;43:31 l-23. 1711 Rogers IS. Lactation and fertility. Early Hum Dev 1997;49 Suppl:S185-S190. (721 Prentice A. Breast feeding and the older infant. Acta Paediatr Stand 1991;374:78-88. 1731 Hoyle B, Yunus M, Chen LC. Breast-feeding and food intake among children with acute diarrhea1 disease. Am J Clin Nun 1980;33:2365-71. [74] West KP, Chirambo M, Katz J et al. Breast-feeding, weaning patterns, and the risk of xeropthalmia 111 Southern Malawi. Am J Clin Nutr 1986;44:690-7. 1751 Tarwotyo I, Sommer A, Soegihartot T. Dietary practices and xeropthalmia among Indonesian children. Am J Clin Nutr 1982;35:574-81. [76] Rahman H. Jalil MA, Rahman A et al. Nutritional status of children in rural Bangladesh. Dhaka: Institute of Public Health and Nutrition and Helen Keller International. Mohakhali Health Complex, 1984. (771 Victora CG, Vaughan JP, Martines JC et al. Is prolonged breastfeeding associated with malnutrition” Am J Clin Nutr 1984;39:307-14. 1781 Brakohiapa LA, Bille A, Quansah E et al. Does prolonged breastfeeding adversely affect a child’s nutritional status? Lancet 1988;ii:416-8. 1791 Tangermann RH, Serdula MK, Colebunders R et al. Breastfeeding beyond twelve months. Lance1 1988;ii:1016. 1801 Rohde JE. Breastfeeding beyond twelve months. Lancet 1988:ii:1016. s174 1.S. Rogers et al. I Early Human Development 49 Suppl (1997) S157-S174 [81] Dettwyler KA. Breastfeeding and weaning in Mali: Cultural context and hard data. Sot Sci Med 1987;24:633-44. [82] Martorell R, Leslie J, Moock PR. Characteristicsand determinantsof child nutritional statusin Nepal. Am J Clin Nutr 1994;39:74-86. [83] Grummer-StrawnL. Does prolonged breast-feedingimpair child growth? A critical review. Pediatrics 1993;91:766-71. [84] Briend A, Wojtyniak B, Rowland MGM. Breast feeding, nutritional state and child survival in rural Bangladesh. Br Med J 1988;296:879-82. [85] Agras WS, Kraemer HC, Berkowitz RI et al. Influence of early feeding style on adiposity at 6 years of age. J Pediatr 1990;116:805-9. [86] Jab1F, Karlberg J, Hanson LA et al. Growth disturbancein an urban area of Lahore, Pakistan related to feeding patterns, infections and age, sex, socioeconomicfactors and seasons.Acta Paediatr Stand 1989;350 Suppl:S44-s54. [87] Rowland MGM, Rowland SGJG, Cole TJ. Impact of infection on the growth of children from 0 to 2 years in an urban West African community. Am J Clin Nutr 1988;47:134-8. [88] Launer L, Habicbt JP, Kardjati S. Breast feeding protects infants in Indonesia against illness and weight loss due to illness. Am J Epidemiol 1990;131:323-31. [89] Dickin KL, Brown KH, Fagbule D et al. Effect of diarrhoea on dietary intake by infants and young children in rural villages of Kwara State, Nigeria. Eur J Clin Nutr 1990$4:307-17. [90] Martorell R, Yarbrough C, Yarbrough S et al. The impact of ordinary illnesses on the dietary intakes of malnourished children. Am J Clin Nutr 1980;33:345-50. [91] Mata JL, Kromal RA, Urrutia JJ et al. Effect of infection on food intake and the nutritional state: perspectives as viewed from the village. Am J Clin Nutr 1977;30:1215-27. [92] Briend A, Hasan KZ, Aziz KMA et al. Are diarrhoea control programmeslikely to reduce childhood malnutrition? Observations from rural Bangladesh. Lancet 1989;ii:319-22. [93] Lauber E, Reinhart M. Prolonged lactation performance in a rural community of the Ivory Coast. J Trop Pediatr 1981;27:74-7. [94] Morley D, Bicknell I, Woodland M. Factors influencing the growth and nutritional status of infants and young children in a Nigerian village. Tram Royal Sot Trop Med Hyg 1968;62:164-5. 1951Brakohiapa LA. Prolonged breastfeeding.Lancet 1988;ii:1360-1. [96] Briend A, Bari B. Breastfeeding improves survival but not nutritional status of 12-35 months old children in rural Bangladesh. Em J Clin Nutr 1989;43:603-8. [97] Molbak K, Gottschau A, Aaby P et al. Prolonged breast feeding, diarrhoeal diseaseand survival of children in Guinea-Bissau. Br Med J 1994;308:1403-6. 1981Unsigned editorial. Growth and development. Paediatr Perinat Epidemiol 1993;7:1-4.