![[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}
|
|
|
|
|
||
3
*
Center for Human Nutrition, Department of International Health,
Department of Biostatistics and
**
Department of International Health, The Johns Hopkins Bloomberg School of Public Health, Baltimore, MD 21205 and
ZVITAMBO Project, Harare, Zimbabwe
3To whom correspondence should be addressed. E-mail: humphrey{at}zvitambo.icon.co.zw.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONWHY CHILDREN BECOME VITAMIN...METHODSRESULTSLITERATURE CITED |
|---|
KEY WORDS: • vitamin A deficiency • children • liver stores • vitamin A supplementation • dietary intake.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONWHY CHILDREN BECOME VITAMIN...METHODSRESULTSLITERATURE CITED |
|---|
,2
). Improving the vitamin A status of young children in developing countries reduces child death rates by 20–50% (3
), which suggests that a substantial portion of their mortality is attributable to vitamin A deficiency. Considering that a large number of foods contain provitamin A carotenoids, many of which are accessible and inexpensive even for the very poor, why is vitamin A deficiency a widespread worldwide problem, especially among young children? Children become vitamin A deficient for two main reasons: 1) their mothers are deficient and produce breast milk low in vitamin A and 2) they are weaned onto diets that provide too little vitamin A. A third contributing factor is that they spend a substantial part of childhood being sick, when anorexia, malabsorption and increased catabolism further deteriorate their vitamin A status.
In this article, we examine these causes in detail. We then quantify their combined effect on the vitamin A stores of a typical child living in a developing country over the first 2 y of life. Each of these three factors (breast-milk vitamin A intake; dietary vitamin A intake; and the burden of disease and its effect on vitamin A consumption, utilization and loss) has been measured and reported in the literature for populations that live under a variety of geographic and cultural conditions. Thus, sensitivity analyses are presented to reflect these ranges of reported values. Next, we estimate the amounts of commonly available foods that contain vitamin A such a child needs to eat to overcome his or her deficiency. Finally, we present estimates of vitamin A stores, assuming the child receives the newly revised vitamin A supplementation regimen recommended by an informal technical consultation of the World Health Organization (WHO)4
(4
).
|
WHY CHILDREN BECOME VITAMIN A DEFICIENT |
|---|
|
TOPABSTRACTINTRODUCTIONWHY CHILDREN BECOME VITAMIN...METHODSRESULTSLITERATURE CITED |
|---|
The first cause of childhood vitamin A deficiency is maternal vitamin A deficiency. Until recently, the magnitude and implications of vitamin A deficiency among reproductive-aged women were relatively unappreciated and unaddressed compared with the problem among young children, primarily because the dramatic clinical signs of xerophthalmia are very rare in women. However, data reported over the past decade indicate that xerophthalmia, as an indicator of vitamin A deficiency, may be even more common among women than in children (1
,5
,6
). This causes substantial maternal and infant mortality (7
–9
) and maternal morbidity (10
). Moreover, deficient mothers produce breast milk with very low concentrations of vitamin A (11
), which puts the next generation at risk.
Mothers in developing countries are commonly vitamin A deficient for two main reasons: they consume diets low in vitamin A and they experience high fertility with prolonged breast-feeding. The new U.S. recommended dietary allowance (RDA) established by the Institute of Medicine (IOM) (12
) for nonpregnant, nonlactating women, aged 19–50 y is 700 µg of vitamin A per day. This increases to only 770 µg/d during pregnancy but nearly doubles to 1300 µg/d during lactation. Median intakes by American women in the Third National Health and Nutrition Examination Survey (NHANES III) (13
) were 583, 691 and 1054 µg/d among nonpregnant, nonlactating; pregnant; and lactating women, respectively, meeting 75–90% of their RDA. In addition, an estimated 44% of American women aged 20–39 y take a daily supplement (14
), which contributes another 1400 µg of vitamin A per day (13
). During pregnancy, most American women take a daily prenatal supplement containing 1000 µg of vitamin A (13
). These data can be compared with a median dietary intake of 403 µg/d by rural Bangladeshi women (15
), which provides 57% of their RDA if they are not pregnant or lactating and only 31% of their RDA during lactation.
In addition to consuming diets low in vitamin A, women in developing countries spend a substantial proportion of their lives breast-feeding, when vitamin A requirements are very high. In industrialized countries, women have, on average, 1.6 babies (16
) and breast-feed them for 5 mo (17
). Thus, these women spend 8 mo, or 2.2% of their 30 reproductive years (ages 15–45) breast-feeding. However, in the world’s least developed countries, women have an average of 5 children and breast-feed each one for 2 y (16
,17
). Hence, rural Bangladeshi women spend one-third of their reproductive years breast-feeding, when their dietary intake of vitamin A provides less than one third of their RDA.
Maternal vitamin A deficiency seems to have little impact on fetal status, when even well-nourished women transfer very little vitamin A to the baby (18
). However, during lactation, well-nourished women transfer about 250 µmol (71,500 µg) of vitamin A [130 L of breast milk consumed (19
,20
) containing 1.92 µmol of vitamin A per liter (55 µg/dL) (18
)], whereas women in developing countries transfer only about half that amount, because average milk vitamin A concentrations are about 1.05 µmol/L (30 µg/dL) (18
). Thus, all babies are physiologically vitamin A "depleted" at birth, having little in the way of vitamin A stores in their livers. Young infants in developing countries have even less vitamin A stores. But during lactation, breast-fed babies of well-nourished women accrue adequate stores, whereas breast-fed babies of vitamin A-deficient women remain depleted. Furthermore, if weaning foods are lower in vitamin A than the breast milk they partially replace (as they usually are), the child’s risk of deficiency increases further when breast-feeding stops.
Children’s diets provide too little vitamin A.
The second reason children become vitamin A deficient is that their diets during and after weaning contain insufficient vitamin A. Dietary vitamin A reference intakes for infants and young children established by IOM (12
) in the United States and by the Food and Agricultural Organization (FAO) (21
) are shown in Table 1
. Taken together, recommended or "safe" intakes range from 350 to 500 µg/d for infants and from 300 to 400 µg/d for 1- to 6-y-old children.
|
). For example, the median intake of retinol (preformed vitamin A in animal sources) by 1- to 3-y-old children in NHANES III (13
) was 404 µg/d, which exceeded their RDA of 300 µg/d by 35%. In studies of preschool children in Egypt, Mexico, Kenya (23
), and India (24
) median intakes of animal sources of vitamin A were 174, 119, 50 and 33 µg/d, respectively, providing only 11–58% of the RDA and leaving these children largely dependent on plant sources. In a study of Bangladeshi children, virtually the only source of preformed vitamin A consumed was breast milk; weaned children consumed only negligible amounts of vitamin A from animal sources (15
).
Previously, it was assumed that 6 µg of ß-carotene in all plant foods provided the same vitamin A activity as 1 µg of preformed retinol. However, recent data (Table 2
) reveal that the amount of ß-carotene from plant foods, which provides the equivalent vitamin A activity of 1 µg of preformed retinol, is far more than 6 µg and is also highly variable among foods, ranging from 8 µg to as much as 40–45 µg (22
,25
–30
). Based on the new data, IOM now recommends using a conversion factor of 12 µg of ß-carotene to 1 µg of retinol for a mixed fruit and vegetable diet (12
). They acknowledge, however, that colored fruits and cooked yellow tubers are much more efficiently converted to vitamin A than equal amounts of dark-green leafy vegetables. To clarify dietary calculation, IOM established a unit, the retinol activity equivalent (RAE), which sets 1 µg of all-trans-retinol = 12 µg of ß-carotene = 24 µg of 
-carotene or ß-cryptoxanthin. This new assumption makes very little difference in assessing the adequacy of American children’s diets, whose intake of animal sources of vitamin A already exceeds their RDA. In addition, an estimated 44% of young American children (1–5 y old) also take a supplement (14
), which provides an additional 750 µg of retinol (13
). Clearly, American children do not depend on plant sources to meet their vitamin A requirements, so the new conversion factor has no effect on their dietary adequacy. However, for children in Kenya, Egypt, Mexico, India and Bangladesh, whose animal food intake is modest or even absent, this new assumption of ß-carotene activity makes an important difference in interpreting their dietary vitamin A adequacy. Table 3
makes clear that, with the newly accepted conversion factors for dietary ß-carotene, mean "vitamin A" intake of children in these developing countries is substantially below their requirements. The percentage of the RDA met by these diets, using the latest corrected assumptions (1 µg of retinol requires 12 µg of dietary ß-carotene), is substantially lower than under the older assumptions (1 µg of retinol requires 6 µg of dietary ß-carotene). This helps explain why vitamin A deficiency is common among children who consume diets previously thought to meet or nearly meet their vitamin A requirements.
|
|
As we show quantitatively later in this article, inadequate vitamin A intake explains most childhood vitamin A deficiency. However, a third important contributing factor is that children in developing countries are so often ill. The prevalence of diarrhea among young children in developing countries ranges from 10 to 19% (31
–34
). Thus, a typical child in a developing country spends 36–70 d each year with diarrhea. The prevalence of acute respiratory illness among young children in developing countries is even higher, 25–60% (34
). Illness worsens vitamin A status primarily by reducing intake due to anorexia and malabsorption and increasing utilization through greater catabolism and urinary loss. Anorexia is a major determinant of reduced dietary intake during episodes of childhood diarrhea, the condition during which dietary intake has best been studied (35
). Diarrhea particularly seems to result in reduced intake of non–breast-milk foods; intake of breast milk is reduced to a lesser degree or not at all. Among infants and young children in Bangladesh who consumed about half their energy intake from breast milk, total caloric intake decreased only 3% during days of diarrhea (36
); among breast-fed Peruvian children, intake was reduced by 12% (37
). In another study from Peru, children did not reduce breast milk intake at all during days of diarrhea but took 23% fewer calories from non–breast-milk foods (38
). Among Guatemalan children who were completely weaned from the breast, energy intakes were reduced by 20% during diarrhea (39
). It is probable that reductions in vitamin A intake during illness are similar to reductions in energy; however, primarily for weaned children, vitamin A intake may be reduced proportionately more if milk, fruits and vegetables are particularly avoided during illness.
Malabsorption of vitamin A can occur during diarrheal illness and lower respiratory infection. Sivakumar and Reddy (40
) showed that absorption of labeled oily retinol by well children is nearly complete, but it is reduced by about 30% in children with gastroenteritis and respiratory infection. In another study among children with diarrhea, absorption of a 100,000-IU dose of vitamin A was reduced by 17 and 22% when administered in water and oral rehydration salt solution, respectively (41
).
Increased catabolism is frequently listed in textbooks as a consequence of illness, but there are limited data quantifying the vitamin A catabolic rate during health or disease. Increased catabolic losses are a result of the acute-phase response to infection, including fever and metabolic breakdown of muscle and adipose tissue (42
). Among healthy adults consuming a vitamin A-free but otherwise adequate diet, Sauberlich et al. (43
) estimated that 0.5% of total body vitamin A stores were lost daily. Haskell et al. (44
) reported that the catabolic rate in rural Peruvian 12- to 24-mo-old children was 2.2%/d. They speculated that the catabolic rate in healthier children might be between the catabolic rate of nongrowing adults (0.5%/d) and that of sick children. Because catabolic rates are applied to existing liver vitamin A stores, absolute losses of the vitamin decline as liver stores diminish (43
) and as deficiency progresses (45
). Vitamin A is not excreted intact in the urine of normal healthy people, but it can be excreted in substantial quantities during illness. Sixty years ago, Lawrie et al. (46
) reported that healthy subjects excreted no vitamin A but that sick patients, especially those with pneumonia, could excrete up to 3.5 µmol (1003 µg) during each day of acute illness. More recently, a research group at the University of Alabama and their collaborators measured urinary vitamin A excretion during infection among adults with sepsis and pneumonia hospitalized in intensive care units (ICUs) in the United States (47
) and among sick children in Peru (48
) and Bangladesh (49
). Although a small number of acutely ill adult patients in ICUs excreted up to 10 µmol (2869 µg) of retinol per day, most excreted much smaller amounts: the geometric mean loss among 29 ICU patients was 0.78 µmol/d (223 µg/d). Similarly, among young children with common childhood infections in Bangladesh, maximum losses were as high as 0.63 µmol/d (180 µg/d), but median losses were <0.007, <0.007, 0.038 and 0.126 µmol/d (<2.0, <2.0, 10.9 and 36.0 µg/d) among children with watery diarrhea, dysentery, pneumonia and sepsis, respectively. More than half the children with watery diarrhea and dysentery had no detectable retinol in their urine at all. Thus, urinary vitamin A loss seems to vary considerably among infections and among children with the same infections. Although some persons can certainly lose enough vitamin A through this route to precipitate xerophthalmia (justifying close monitoring and vitamin A supplementation during acute illness), urinary loss probably has only a minor impact on vitamin A status for the average child during an average day of illness.
Two childhood diseases, chicken pox and measles, can severely compromise vitamin A status. In a study by Campos et al. (50
), children were assessed for vitamin A status and then supplemented with 200,000 IU of vitamin A. Three months later, more than one third of the children came down with chicken pox. Before supplementation, about 40% of the children had deficient liver stores. Six months after supplementation, 90% of uninfected children were still vitamin A sufficient, but only 26% of the infected children had adequate stores. Measles results in markedly depressed circulating vitamin A concentrations (51
) and can precipitate xerophthalmia (52
,53
). Vitamin A supplementation during acute measles consistently and dramatically reduces measles case fatality rates (54
,55
). Thus, since 1987, both the Expanded Programme on Immunization and the Nutrition Unit of WHO, along with the United Nations Children’s Fund (UNICEF) and the International Vitamin A Consultative Group (IVACG), recommend high-dose vitamin A treatment of all cases of severe measles in places where the measles case fatality rate exceeds 1% (56
). Because children get chicken pox and measles only once in a lifetime, we do not include these diseases in calculations later in this article; given the substantial impact of these infections on vitamin A status, however, we would be remiss to exclude them from this discussion.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONWHY CHILDREN BECOME VITAMIN...METHODSRESULTSLITERATURE CITED |
|---|
The liver contains about 90% of total body vitamin A (57
). Thus, liver vitamin A concentration is the best estimate of total body stores and, therefore, the best indicator of vitamin A status. Because little vitamin A is transferred in utero (18
), all infants are born with very limited liver vitamin A stores, 
6 µmol total stores, or 0.04 µmol/g,5
which is less than a 2-wk supply (59
). Healthy, well-nourished infants rapidly accumulate liver vitamin A stores, achieving an adult concentration of 0.07 µmol/g (or about 20 µmol total stores) by 6 mo of age. This liver vitamin A concentration has been suggested as the minimum level that prevents deficiency, provides a suitable reserve for periods of stress and/or low intake and is fully consistent with good health (60
).
Liver stores of vitamin A accumulate as a function of the absorbed intake, basal requirements and catabolic losses of the vitamin. Dietary intake depends on the volume and vitamin A content of breast milk, foods and supplements consumed and the percentage absorbed from each source. Basal requirements are primarily influenced by age and body size. Catabolic loss of vitamin A is proportional to existing vitamin A liver stores and the catabolic rate of the vitamin. Illness may affect vitamin A stores by reducing intake, reducing absorption and increasing catabolic rate.
In this section, we develop a set of recursive equations to estimate the amount of vitamin A in the liver as a function of age. The amount of vitamin A in the liver at the end of day t is equal to the amount of vitamin A in the liver on the previous day (t - 1) plus the net change during day t[
(t)]. Let S(t) and S(t - 1) represent the micrograms of vitamin A stored in the liver at the end of day t and day (t - 1) after birth, respectively. Then,
 | (1) |
(t) is equal to total intake less basal needs and catabolic losses. To allow for the complexity consequent to illness, we develop separate equations for the expected net daily change in vitamin A depending on whether the child is healthy or sick during day t.
If the child is healthy during the day, the net daily change is
 | (2) |
where I0(t) is the absorbable daily intake of vitamin A for a healthy child [µg of vitamin A or µg of retinol equivalents (RE)]; U is the amount of vitamin A required for basal needs (µg), which is assumed not to vary with age; and r0 is the fractional catabolic rate (percentage of total body stores lost per day) for a healthy child. The absorbable intake of vitamin A for a healthy child is
 | (3) |
where b0(t), f0(t), and v(t) are the amounts of vitamin A absorbed from breast milk (µg of RE), food (µg of RAE) and supplements (µg), respectively, during day t for a healthy child. The amount of vitamin A from breast milk is
 | (4) |
where bv0(t) is the volume of breast milk consumed (mL) and bc0(t) is the absorbable vitamin A concentration of breast milk during day t. bc0(t) is the product of the vitamin A concentration of the breast milk in µg of RE/mL and the percentage of vitamin A absorbed from breast milk.
If the child is sick during day t, the net daily change is
 | (5) |
 | (6) |
where the components of eqs. 5 and 6 are defined analogously to eqs. 2 and 3 .
We define p(t) to represent the probability that a child is sick during day t. Then, the expected net daily change in vitamin A levels is a weighted average of eqs. 2 and 5
, where the weights are determined by p(t). That is,
 | (7) |
Various models could be postulated for the probability of illness, p(t). The simplest model is to assume that the probability a child is sick on day t is constant, such as 0.10. More complex models can also be developed to allow the probability a child is sick on day t to depend on age or even the amount of vitamin A in the liver the preceding day. The reason for considering such a model is that, as vitamin A stores decrease, the probability of illness may increase, which further negatively affects vitamin A stores, creating a negative feedback loop.
Regardless of the model that is chosen for p(t), these equations can be solved recursively for a given set of input parameters as follows. The initial value of vitamin A in the liver at birth is specified—for example, S(0) = 6 µmol (1700 µg). The expected net daily change in vitamin A is calculated from eq. 7 by using eqs. 2 –6 . Then, eq. 1 is used to estimate expected levels for the following day. These equations are then used recursively to obtain the expected vitamin A levels at each day of age.
Input parameters
To apply the equations developed in the preceding section, values for the input parameters must be specified (Table 4
). For each variable we selected a best estimate value as the most representative of a typical child in a developing country based on the available literature. These values were used in the primary analysis. However, recognizing the great variability that exists for each of these variables reflecting true between-individual variability and imprecision in their measurement, we also carried out sensitivity analyses substituting more extreme values for the input parameters. These sensitivity analyses demonstrate both the likely range of liver vitamin A stores among children living under various conditions in developing countries around the world and the robustness of our model in answering the key question: why do children become vitamin A deficient?
|
,62
).
Basal need (U).
In our primary analysis, we used 180 µg/d as the amount of vitamin A required for basal needs, or U, and assumed U to be constant over the first 2 y of life. This is the amount defined by FAO/WHO (21
) as the minimum daily intake of vitamin A needed to prevent the appearance of clinical signs of vitamin A deficiency and to permit normal growth among children 0–1 y of age. FAO/WHO stated that infants fed by vitamin A-deficient mothers whose milk provided 120 ± 15 µg of RE/d grow normally and do not show clinical signs of vitamin A deficiency. They then increased this value to 180 µg/d to account for variability in growth rate. Therefore, in a sensitivity analysis, we used the 120 µg/d values for U. However, it should be noted that subclinical vitamin A deficiency may begin to significantly increase morbidity and mortality risk before growth is impaired or clinical eye signs or symptoms appear (63
).
Catabolic rate (r0).
The fractional catabolic rate is the percentage of total body stores lost per day. There are limited data in the literature estimating this quantity. In a small volunteer sample of eight adult males, the catabolic rate was estimated to be 0.5%/d (43
) and is probably too low for infants whose turnover of liver vitamin A stores is expected to be higher to support growth (64
). In rural Peru, among children aged 12–24 mo (n = 105), a rate of 2.2%/d was calculated with data on plasma retinol kinetics (44
,64
). Because of the uncertainty in this measurement, we selected an intermediate value (1.5%/d) for our primary analysis, but, in sensitivity analyses, we substitute the higher and lower values, 0.5%/d and 2.2%/d. In all analyses, we assumed the rate to be constant from birth through 2 y.
Absorbable intake [I0(t)]. The daily intake of vitamin A from breast milk, food and supplements varies substantially among children and with age. Thus, we used age-specific mean intake data reported in the literature to plot breast milk intake (dL), breast milk vitamin A concentration (µg/dL) and dietary vitamin A intake (µg of RAE) as a function of age and then constructed a continuous curve by connecting the data points with straight lines. We assumed 90% of the vitamin A from breast milk and animal sources is absorbed and used. The new IOM conversion factor for ß-carotene of 12 µg of ß-carotene to 1 µg of retinol in mixed vegetable diets is used.
Breast milk intake [bv0(t)].
Brown (17
) compiled quantitative information from the literature on breast milk consumption by children and estimated the daily mean volume (mL/d) of breast milk consumed by age group and exclusivity of breast-feeding. In our primary analysis, we used the values estimated by Brown for children in developing countries who are exclusively breast-fed from birth through 6 mo and partially breast-fed from 6 mo of age through 2 y; we plotted the mean breast milk intake, or bv0, by the median age of the specified age group. With these data, we plotted intake values at 45, 135, 225 and 540 d. For the amount of colostrum and early milk consumed, we estimated the intake during days 1 and 7 to be 50 and 600 mL, respectively (65
). We assumed that children are completely weaned from the breast by 24 mo of age [bv0 (730) = 0 mL]. For bv0(t), we constructed a continuous curve by connecting the age-specific data points for breast milk consumption with straight lines. In a sensitivity analysis, we used breast milk intake values for children partially breast-fed from birth through 2 y (17
), assuming intake during day 1 and day 7 to be 50 and 510 mL, respectively.
Breast-milk vitamin A concentration [bc0(t)].
In our primary analysis, we estimated mature breast milk vitamin A concentration to be 1.05 µmol/L (30 µg/dL), an average value for several developing country studies (18
). Because the vitamin A concentration of colostrum is greater than that of mature milk (11
), we estimated the vitamin A content of breast milk from birth through 7 d to be 1.75 µmol/L (50 µg/dL). In two sensitivity analyses, we substituted breast milk vitamin A concentrations representative of women who are frankly deficient (mature milk = 20 µg/dL, colostrum = 35 µg/dL) and of women whose milk is considered to be vitamin A adequate (mature milk = 50 µg/dL, colostrum = 90µg/dL). In all analyses, we assumed the concentration of vitamin A in colostrum (0–7 d) and mature milk (>30 d) to be constant with age. For the period between 7 and 30 d, we constructed a straight line between two concentrations, representing a continuous decrease in the vitamin A content of transitional milk. In all analyses and over all time periods, we assumed that absorption of vitamin A from breast milk was 90%.
Non–breast-milk vitamin A intake [f0(t)].
In the primary analysis, we assumed exclusive breast-feeding from birth to 6 mo and then we assumed a complementary food diet was started that was similar in vitamin A content to that of the Kenyan children in the Nutrition Collaborative Research Support Program (NCRSP) study (23
). We assumed the vitamin A intake at 24 mo (730 d) of age to be the same as that of the 18- to 30-mo-old children studied (205 µg of RAE/d) (23
), and we estimated the intake during day t by using f0(t) (see below). In a sensitivity analysis, we assumed the complementary diet begun at 6 mo was similar to that of the 12- to 48-mo-old Indian children studied by Ramakrishnan et al. (24
). Using their data, we estimated the intake of 15-mo (450-d)-old children to be the same as that of the 12- to 17-mo-old study children (54 µg of RAE/d) and the intake of 1-y (365-d)-old children to be the average of the 18- to 23-mo and the 24- to 29-mo-old children studied (76 µg of RAE/d). For f0(t), we plotted the Indian data and connected the points with straight lines. In the sensitivity analysis examining the effect of partial versus exclusive breast-feeding during the first 6 mo of age, we assumed that weaning foods replacing breast milk before 6 mo of age (such as water, porridge, juice and tea) contain negligible amounts of vitamin A.
Impact of illness.
In our primary analysis, we assumed the probability of illness on any given day t, p(t), is constant, and we conservatively estimated a value of 10% from the low end of the range for the prevalence of diarrhea among young children in developing countries. During a day of illness, we assumed that breast milk volume remains the same, but absorption of vitamin A from the breast milk is reduced by 30%. Thus,
 | (8) |
For intake of vitamin A from food, we assumed that the amount of intake is 20% less due to anorexia and that, of this reduced intake, there is 30% reduced absorption. Thus,
 | (9) |
There are no data, as far as we know, that compare the catabolic rate during illness with the catabolic rate when healthy. However, we do know that, in general, catabolism increases during illness (42
). We have approximated that the catabolic rate increases by 150% during illness compared with the rate for healthy persons. Thus,
 | (10) |
In a first sensitivity analysis, we estimated the prevalence of illness to be 20%, the high end of reported ranges. In a second sensitivity analysis, we estimated vitamin A stores for an atypical child living in a developing country who is not sick at all during the first 2 y of life. In a third sensitivity analysis, we considered that, in addition to being acutely ill 10% of the time, the child is also infected with helminths from 6 mo of age onward, which results in a 20% reduction in vitamin A absorbed from all sources.
Vitamin A supplementation [v(t)].
The vitamin A supplementation regimen for mothers and children newly proposed by a WHO informal technical consultation (4
) is as follows. Mothers: two doses of 200,000 IU within the first 6 wk postpartum. The two doses should be given at least 24 h apart. Infants: three doses of 50,000 IU between birth and 6 mo; 100,000 IU at 6 mo; and 200,000 IU every 4–6 mo from 12 mo onward.
Maternal vitamin A supplementation increases the vitamin A content of breast milk, bc0, with a greater increase in concentration closer to the time of supplementation (Table 4)
. We assumed that 12% of each maternal dose was transferred to the infant through breast milk (0.12 x 200,000 IU = 24,000 IU = 25 µmol = 7200 µg) (59
) over 9 mo (66
). Infant vitamin A supplementation, or v(t), increases I0 during the day the infant receives the dose. On days when no supplement is given, v(t) = 0. We assumed that 50% of each high-dose supplement was absorbed (43
,67
), and the three doses of 50,000 IU were given at 42, 70 and 98 d; the 100,000-IU dose was given at 180 d; and the 200,000-IU doses were given at 365, 548 and 730 d.
Model for estimating the quantity of dietary vitamin A from non–breast-milk food sources that must be consumed over a given period of time to replete vitamin A stores
An additional question of interest is: How much vitamin A does a child need to consume from non–breast-milk foods to achieve a desirable amount of liver stores? To answer this question, we can use the recursive equations developed to estimate the vitamin A stores in the liver and solve for the amount of vitamin A needed from food. Specifically,
 | (11) |
where A = 0.7b0(t) + v(t) - u - 1.5r0S(t - 1), B = b0(t) + v(t) - u - r0S(t - 1), and recall that b0(t) = bv0(t)bc0(t). Required inputs are S(t), bv0(t), bc0(t), v(t), u, r0 and p(t).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONWHY CHILDREN BECOME VITAMIN...METHODSRESULTSLITERATURE CITED |
|---|
Estimated liver vitamin A stores among children in developing countries during the first 2 y of life
We now estimate the combined effect of low breast-milk vitamin A concentration, inadequate dietary vitamin A intake and prevalent illness on resulting liver vitamin A stores for a typical child in a developing country with the values defined in Table 4
for "primary analysis." In brief, this child begins life with 6 µmol (1700 µg) of vitamin A in his or her liver, has a basal requirement of 180 µg/d, loses 1.5% of total vitamin A stores each day due to normal catabolism of the vitamin, is exclusively breast-fed until 6 mo of age from a marginally vitamin A-deficient mother producing mature breast milk that contains 1.05 µmol/L (30 µg/dL), is fed a diet similar to the Kenyan children studied in the NCRSP study and is ill about 10% of the time. The liver vitamin A stores for this typical child are plotted in Figure 1
a. For comparison, we drew a line representing "minimally adequate" vitamin A stores, 0.07 µmol/g (20 µg/g). We also calculated and included lines representing the liver stores for a child whose dietary vitamin A intake is equivalent to that of the 50th percentile American child surveyed by NHANES III (13
), assuming that child also has a catabolic rate of 1.5%/d, is sick 10% of the time and did not take any vitamin A supplements.
|
8 mo, increased weaning food consumption together with continued breast-feeding allows this child to accumulate vitamin A, and liver stores grow to 25 µmol by 18 mo of age. Thus, at 18 mo this child approaches, but never attains, a minimally adequate liver vitamin A concentration (20 µmol/g). However, over the next 6 mo, the child’s breast-milk intake falls, needs and losses exceed intake and total stores fall. By 2 y of age, liver stores have fallen to 11 µmol and, if the child continues in this pattern, stores will soon be depleted. Peak liver vitamin A concentration briefly achieved at 18 mo is still less than half that of the average American child. Sensitivity analysis
Catabolic rate.
Substituting the slower adult rate of catabolism (0.5%/d), a very conservative estimate for growing infants, the developing country child’s liver stores remain less than the minimally adequate concentration for the entire first y of life (Fig. 1b)
. Liver stores then increase between 12 and 20 mo, when they peak at 60 µmol. However, as the child is weaned from the breast and breast-milk intake declines, stores decrease quickly. It is interesting to note that, although stores are adequate during the entire 2nd y of life, they remain, on average, about one-fourth that of the NHANES III child, assuming the same catabolic rate.
Substituting the higher catabolic rate of 2.2%, measured in Peruvian preschool children (44
,64
), total stores remain at about the level the child was born with (6 µmol) until about 10 mo of age. The child is then in slightly positive vitamin A balance and stores peak at 18 µmol at 18 mo of age. As breast milk volume decreases after 18 mo, the child is in negative balance and completely depletes stores shortly after 2 y.
Basal requirement.
Substituting 120 µg/d for U, basal requirement, this child’s intake of vitamin A is sufficient to build stores that more closely approximate and, by 1 y of age, even surpass minimally adequate levels (Fig. 1c)
. But, as the child consumes less breast milk, the weaning diet does not contribute enough vitamin A to meet the reduced basal needs and maintain the current level of stores, and liver vitamin A falls below 0.07 µmol/g before 2 y of age.
Exclusivity of breast-feeding during early infancy.
If the child is only partially breast-fed (as most babies in the world are by 3 mo of age), volume intake is about 15% lower (17
), and the child completely depletes liver stores at 3 mo (Fig. 1d)
.
Vitamin A intake.
Still assuming the low vitamin A breast-milk concentrations considered in the primary analysis [colostrum = 1.75 µmol/L (50 µg/dL), mature milk = 1.05 µmol/L (30 µg/dL)], but substituting the diet lower in vitamin A consumed by the Indian children studied by Ramakrishnan et al. (24
), the combined vitamin A intake from breast milk and weaning foods is not enough to accumulate any vitamin A stores, and the child’s liver is depleted by 21 mo of age (Fig. 1e)
.
|
. If the child is weaned onto a diet similar to that consumed by the Indian children studied by Ramakrishnan et al. (24
), liver stores fall below minimally adequate at 18 mo and are depleted at 23 mo of age.
Substituting a very low breast milk vitamin A concentration [colostrum = 1.2 µmol/L (35 µg/dL), mature milk = 0.70 µmol/L (20 µg/dL)] and assuming the Kenyan diet, the child’s liver stores are completely depleted by 1 mo of age and remain so throughout the entire first y of life (Fig. 1e)
. When the quantity of non–breast-milk dietary sources consumed increases in the 2nd y of life, the child briefly accumulates liver stores during a period between 12 and 18 mo, but stores peak at only 10 µmol, from which point they decrease. Substitute both the very low breast-milk vitamin A concentration and the Indian diet, and the child’s liver vitamin A stores are not only depleted by 1 mo of age but remain depleted over the entire first 2 y of life. An important observation here is that, with the assumptions we have used, once breast-milk vitamin A concentration falls much below 1.05 µmol/L (30 µg/dL), babies are in negative vitamin A balance, even if they are exclusively breast-feeding.
Sickness.
If we assume this child never gets sick, liver vitamin A stores exceed minimally adequate levels at 17 mo but only for a brief time before stores again fall below adequacy (Fig. 1f)
. Substitute a 20% probability of illness while maintaining all other assumptions in the primary analysis, and liver stores peak at only 21 µmol and are depleted shortly after 2 y.
Finally, helminths infect 20% of children worldwide (68
) and, in some populations such as that of coastal eastern Africa (69
), nearly all children are infected with helminths. We estimated liver vitamin A stores assuming the child was infected with helminths at 6 mo of age and older (Fig. 1g)
and assuming that infection reduces the absorption of vitamin A from breast milk and food by 20% (70
). With helminth infection, liver stores are decreased by about half and become negligible by 2 y.
Thus, regardless of what we assume about the probability of illness, inadequate vitamin A intake (from breast milk and from food) is the predominant cause of vitamin A deficiency in children. Helminth infection has a larger impact on vitamin A stores than such acute illnesses as diarrhea and respiratory infection, because it is a chronic infection that reduces absorbable intake of both breast milk and food over long periods of time. Although illness plays a less important role in determining vitamin A status than intake on a population basis, the probability of illness increases with vitamin A deficiency. Thus, the very children already consuming the least vitamin A are the ones most vulnerable to the further negative effects of illness, even if those effects are modest.
Can a child in a developing country attain minimally adequate vitamin A stores by 2 y of age with diet alone?
During the second y of life (12–23 mo), complementary food diets can be more varied and substantial because children have teeth and the motor skills necessary to consume more fruits, vegetables and other solid foods. Therefore, we need to consider whether, with an aggressive dietary education intervention, children with marginal or absent vitamin A stores at age 12 mo would be able to consume a sufficient quantity of locally available vitamin A-rich foods over the 2nd y of life to succeed in building minimally adequate liver stores of vitamin A. Using eq. 11
, we calculate the amount of vitamin A the child would need to consume from nonbreast milk foods to build minimally adequate vitamin A liver stores (0.07 µmol/g x 400 g = 28 µmol, or 8000 µg) by age 24 mo (730 d). Thus, if S (365) = 0 and S (730) = 8000, Figure 2
shows how much vitamin A the child would need to consume each day from 1 to 2 y of age from nonbreast milk food. On average, he or she must consume 160 µg/d if breast milk containing 1.05 µmol/L (30 µg/dL) is also being consumed and 200 µg/d if breast milk containing 0.7 µmol/L (20 µg/dL) is also being consumed.
|
If we estimate that spinach has 4000 µg of ß-carotene per 100 g and that 1 µg of retinol is equivalent to 26 µg of ß-carotene, this child would need to eat about 105 and 130 g of spinach each day to get 160 and 200 µg of RAE, respectively (Table 5
). In a study of vulnerable preschool Indonesian children, the usual portion size of spinach and other dark-green leafy vegetables, such as cassava leaves and swamp cabbage, was 15 g/d (71
)—or 1/10th of what they need.
|
If we consider the equivalent in animal products, assuming 90% of preformed vitamin is absorbed (40
), the child would have to consume 2.5 eggs or 0.360 L of cow’s milk per day or 9 g of chicken liver per week. More realistically, the child’s daily vitamin A intake should come from a combination of different vitamin A-containing foods. Table 5
shows how much fruits and vegetables a child would have to consume if 25 and 75% of the vitamin A in the diet is preformed vitamin A. These amounts are still greater than the amounts likely to be consumed by children.
What are the liver vitamin A stores if the reference child also receives the newly revised WHO vitamin A supplementation regimen or a 6-mo regimen beginning at 6 mo of age?
The resulting liver vitamin A stores for the primary analysis child who also receives the full newly revised WHO supplementation regimen are shown in Figure 3
a. During the first 6 mo, liver vitamin A stores approximate those of the NHANES III 50th percentile child. As breast milk consumption decreases and all the maternal dose has been transferred through breast milk by 9 mo of age, liver stores fluctuate greatly with administration of each high-dose supplement. Before the first 200,000-IU dose is given at 12 mo, stores dip below minimally adequate levels (0.07 µmol/g). For about 1 mo after each dose at 12 and 18 mo, vitamin A stores exceed that of the NHANES III child. Because breast milk volume steadily decreases between 18 and 24 mo, liver vitamin A stores fall below minimally adequate levels before the dose is received at 2 y. However, this child, who receives every dose of the regimen, is able to maintain a minimally adequate liver vitamin A concentration most of the time over the first 2 y of life. It is interesting to note that, even though the child receives every recommended supplement of vitamin A, liver stores only briefly (at 12, 18 and 24 mo) exceed the liver stores achieved by the 50th percentile American child from diet alone (NHANES III) and are still well below the proposed cut-off of 1.05 µmol (300 µg) per gram of liver for vitamin A toxicity (61
).
|
. From 12 mo onward, the child will achieve stores equivalent to that of the child receiving the newly revised WHO regimen. Young children become vitamin A deficient primarily because their mothers are vitamin A deficient and produce breast milk that is low in vitamin A, and their diets contain too little available vitamin A. A third contributing factor, especially for the most vulnerable children, is frequent illness. Exclusive breast-feeding cannot maintain adequate liver vitamin A stores in babies once the vitamin A content of their mothers’ milk falls below 1.05 µmol/L (30 µg/dL). To overcome vitamin A deficiency with fruits and vegetables alone, children need to increase current portion sizes about 10-fold.
Vitamin A supplementation programs are the most important and effective intervention currently available. The supplementation regimen recently proposed by WHO will meet the needs of most children but for some, just barely, especially if their mothers are extremely vitamin A deficient, their diets contain very little vitamin A, and they are sick more than average (estimated as 10% of the time). Therefore, supporting realistic dietary changes—for example, encouraging mothers to add even modest amounts of animal sources of vitamin A to their children’s diets; fortifying weaning foods, condiments and staples; and solar-drying mangoes to make them available more often during the year—continues to have a role in meaningfully improving the vitamin A status of children who eat a traditional developing country diet. However, even these interventions rarely result in sufficient intake in the absence of supplementation.
Editor’s note. These estimates and conclusions are based on data currently available. Future data, and more refined analytic techniques, are desirable and welcomed but, in the opinion of the authors and outside reviewers, are unlikely to alter the conclusions reached.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Preparation of this paper was supported by Cooperative Agreement HRN-A-00-97-00015-00 between the Office of Health, Infectious Disease and Nutrition; U.S. Agency for International Development (Washington); and the Center for Human Nutrition, Department of International Health, The Bloomberg School of Public Health, Johns Hopkins University (Baltimore, Maryland), with additional support from The Bill and Melinda Gates Foundation (Seattle, Washington) and the Task Force Sight and Life Research Institute.
4 Abbreviations used: EAR, estimated average requirement; FAO, Food and Agricultural Organization; ICU, intensive care unit; IOM, Institute of Medicine; IU, international units; IVACG, International Vitamin A Consultative Group; NHANES III, Third National Health and Nutrition Examination Survey; NCRSP, Nutrition Collaborative Research Support Program; RAE, retinol activity equivalent; RDA, recommended dietary allowance; RE, retinol equivalent; UNICEF, United Nations Children’s Fund; WHO, World Health Organization.
5 Approximate liver weights are 125 and 275 g at birth and 6 mo, respectively (58
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONWHY CHILDREN BECOME VITAMIN...METHODSRESULTSLITERATURE CITED |
|---|
1. West, K. P. (2002) Extent of vitamin A deficiency among preschool children and women of reproductive age. J. Nutr. 132:2857S-2866S.
2. Humphrey, J. H., West, K. P., Jr. & Sommer, A. (1992) Vitamin A deficiency and attributable mortality among under-5-year-olds. Bull. W.H.O. 70:225-232.[Medline]
3. Beaton, G. H., Martorell, R., Aronson, K. J., Edmonston, B., McCabe, G., Ross, A. C. & Harvey, B. (1993) Effectiveness of vitamin A supplementation in the control of young child morbidity and mortality in developing countries. Nutrition Policy Discussion Paper 13 1993 Administrative Committee on Coordination—Subcommittee on Nutrition, WHO Geneva. .
, NHANES III; - - -, min adequate; —-, primary analysis. (b) Sensitivity analysis of catabolic rate: estimated liver vitamin A stores in a typical developing country child by age and catabolic rate during health (r0): 1.5%/d (primary analysis), 0.5%/d (low estimate) and 2.2%/d (high estimate). For comparison, curves representing minimally adequate vitamin A stores (0.07 µmol/g) and estimated liver stores for a child whose dietary vitamin A intake is equivalent to that of a 50th percentile American child surveyed by NHANES III (13
, "very low"/India. (f) Sensitivity analysis of illness prevalence: estimated liver vitamin A stores in a typical developing country child by age and probability of illness [p(t)]: 10% over entire 2 y (primary analysis), 20% over entire 2 y (sensitivity analysis), and never sick (sensitivity analysis). For comparison, curves representing minimally adequate vitamin A stores (0.07 µmol/g) and estimated liver stores for a child whose dietary vitamin A intake is equivalent to that of a 50th percentile American child surveyed by NHANES III (13
