?

N-MH--Q1p{
AGRICULTURAL STRATEGIES FOR
MlCRONUTRIENTS • WORKING PAPER
1

HUMAN NUTRITION
FOOD AND MICRONUTRIENT RElATIONSHIPS

DORIS HOWES CALLOWAY

",

INTERNATIONAL FOOD POUCY RESEARCH INSTITUTE

Working papers of the International Food Policy Research Institute encompass
a wide range of subjects drawn from its research programs. The papers-primar-
ily data analyses, historical descriptions, or case studies-contain information
that tFPRI believes may be of interest to others. Working papers undergo informal
review but do not necessarily present final research results..
 HUMAN NUTRITION:
FOOD AND MICRONUTRIENT RELATIONSHIPS

Doris Howes Calloway

Working Papers on Agricultural Strategies for Micronutrients, No. 1

International Food Policy Research Institute

Washington, D.C.

March 1995
 CONTENTS

Foreword . v
1. Introduction. 1

2. Typology of Specific Nutrients 5

3. Food Patterns and Nutrient Adequacy in Three Populations .12

4. Strategies for Intervention . . . . . . . . . . . . . . . . . . . 19

iii
 TABLES

1. Malnutrition in developing countries, 1975-90 . . . . . . . . . . . . . . . . . . . . 3

2. Household and individual energy adequacy, intake, and percentage

from fat in Egypt, Kenya, and Mexico, 1984-86 . . . . . . . . . . . . 13

3. Average daily nutrient intake of toddlers in Egypt, Kenya, and Mexico, 1984-86 . 15

4. Predicted prevalence of inadequate nutrient intakes in toddler diets . . . 16

5. Frequency of toddler diets predicted to be inadequate in zero to six or

more nutrients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

ILLUSTRATION

1. Sources of energy for toddlers in Collaborative Research Support Program:

percentage of calories from eight food groups. . . . . . . . . . . . . . . . . . . . 14

iv
 FOREWORD

The International Food Policy Research Institute has been designated to take the lead in
coordinating activities related to human nutrition across the 16 centers that constitute the
Consultative Group on International Agricultural Research (CGIAR). A large part ofthat effort
is organized around a five-year project initiated in May 1993 with funding from the Office of
Health and Nutrition of the U.S. Agency for International Development. The objective of the
project is to identify and implement cost-effective alternatives within the CGIAR for increasing
micronutrient intakes. Taken together, micronutrient deficiencies affect a far greater number
of people in the world than protein-energy malnutrition, with serious consequences for health,
cognitive ability, and productivity.
Beginning with this paper, a series of Working Papers on Agricultural Strategies for
Micronutrients will report on activities undertaken on all aspects of the project, and so will
represent views and research findings from several disciplinary perspectives. The project will
undertake activities in two broad areas. One component involves a coordinated effort between
CGIAR centers and agricultural and nutrition research institutes in developing and developed
countries to breed for nutritionally improved staple food crops. The second component, a
collaborative effort between CGIAR centers and developing-country social science and
nutrition research institutions, will undertake collection and analysis of household survey
information to better understand the linkages between agricultural production, household
resource allocation, and nutrition outcomes for improved policy formulation.
The first three papers in the series provide perspectives from the human nutrition,
household economics, and plant nutrition/plant breeding disciplines. These papers were
commissioned for presentation at an organizational workshop convened in 1994 and attended
by an interdisciplinary group of persons, from within and outside the CGIAR, involved both in
research on micronutrients and in implementing programs to reduce micronutrient malnutri-
tion. Various research directions for the project were discussed. The authors of these three
papers were asked to provide (1) a summary of what is known about micronutrients in their
respective areas of expertise in a way that could be understood by those not trained in that
discipline, (2) their judgments as to what are the significant gaps in knowledge, and (3) their
opinions as to how project activities could best contribute to closing these gaps in knowledge
and to reducing the micronutrient malnutrition problem.
This first paper, which addresses micronutrients and human nutrition, was prepared by
Doris Howes Calloway, professor emerita, Department of Nutritional Sciences, the University
of California at Berkeley. Professor Calloway is not only well known for her distinguished
research on nutritional (including micronutrient) deficiencies in developing countries, but as
a former member of the Technical Advisory Committee of the CGIAR she has a keen
understanding of the CGIAR system and has worked tirelessly to incorporate nutritional
concerns into CGIAR activities. There could be no more appropriate person to initiate this
series of working papers.

Howarth E. Bouis
Series Editor

v
 ACKNOWLEDGMENTS

This paper was prepared for the International Food Policy Research Institute Workshop
on Food Policy and Agricultural Technology to Improve Diet Quality and Nutrition, held at
Annapolis, Maryland, 10-12 January 1994. lowe thanks to my colleague, Suzanne Murphy,
and to Howarth Bouis of IFPRI for reviewing earlier drafts ofthis paper, and to the Agricultural
Experiment Station, University of California at Berkeley, for general support.

vi
 1. INTRODUCTION

The ultimate goal of the Consultative Group on International Agricultural Research

(CGIAR) was restated in 1991 as being "to improve the nutrition and well-being of low-income
people [in the developing countries]." Achieving top billing for nutrition is gratifying, but it has
proved difficult to formulate a CGIAR research plan to address nutrition problems directly.
That is not surprising, given the original categorical crop focus of many CGIAR centers.
Because diets and food behaviors are complex, the nutrition problem also is complex. As the
centers now are moving toward consideration of crops and cropping patterns within ecological
and agricultural regions, it is timely to consider how specific nutrition topics can best be fitted
into the centers' evolving research priorities and strategies (TAC Secretariat 1992).
The purpose of this working paper is to provide background information on key relation-
ships between food and nutrient intakes, micronutrient status, and health outcomes and to
suggest research topics and activities that the centers should undertake related to micronu-
trient deficiencies. The background review is necessarily more indicative than comprehen-
sive. The three micronutrients given priority by the international nutrition community-iodine,
iron, and vitamin A-are considered along with a few others known to cause serious
malnutrition. Selected recent findings from consumption studies are cited to illustrate the role
offood sufficiency, composition, and preparation in achieving adequacy. A case is made for
the long-run utility of food rather than pharmaceutical interventions.

FOOD AND NUTRIENT RELATIONSHIPS

The drive to seek food is fundamental, a biological drive. Humans, like other animals,
seek energy sources and, given sufficient access to food, will usually meet or exceed their
energy demand. If a population group is reasonably active, energy intake is nicely regulated
to output, with only sufficient accumulation of adipose reserves as a physiologic hedge against
periods of unmet demand. Requirement determines the amount of energy consumed,
provided that the diet is not so bulky-low in fat, high in fiber and water-that the volume
needed to meet the energy requirement exceeds ingestive capacity. Diets with less than 15
percent of energy from fat and typically low in refined carbohydrates are not sufficiently
concentrated for children to meettheir energy needs (Wheeler 1980) and may not be sufficient
for adults who regularly engage in very heavy work.
There is no convincing evidence that humans, or other animals, have an innate ability to
select foods that supply specific nutrients other than sodium. If a population is successful in
an ecologic niche, by some process of trial and error in selecting for energy, it must have
found a dietary pattern that provides the minimum requirements of all the essential nutrients.
For humans, that means about 40-45 substances (in addition to the energy sources-carbo-
hydrates, fat, protein, and, to a limited extent, alcohol-the essential nutrients include 9 or
10 amino acids, 2 fatty acids, 13 vitamins, and 15-20 minerals). The amount and balance of
the nutrients may not be optimal (in which case the genetically potential size of an individual
may not be attained and the life span may be shortened), but the requirements for popula-
tion-reproduction must be met.
The balancing of nutrient intake with needs is a remarkable achievement. Success is in
large part due to the fact that natural foods contain mixtures of nutrients and that, in most
instances, there is no harm in eating several times as much of a nutrient as is required. (The
margin of safety is much smaller for most of the trace minerals and fat-soluble vitamins than
for other nutrients.) Foods also supply substances that are not classed as nutrients but have
 2

beneficial properties (for example, dietary fiber) or valued pharmacologic effects (for example,
caffeine). They also contain undesirable antinutritional and potentially toxic compounds (for
example, cyanide). Even with a sophisticated modern knowledge of dietary requirements,
intake of many less-studied nutrients remains a matter of chance rather than intent. Most of
the world's consumers manage with no specific knowledge of nutrition at all, selecting foods
deemed appropriate in their culture and preparing them in the way they have been taught.
The amount of essential nutrients required varies with age, size, sex, and reproductive
status. Within these categories, individual requirements for most nutrients are normally
distributed with a coefficient of variation of about 15 percent. There are no recognized
differences in qualitative or quantitative requirements across racial groups.
Because of their metabolic functions, the quantitative requirement for some nutrients
varies with the intake of others. Requirements for thiamin, riboflavin, and niacin are related
directly to total energy intake, and the requirement for vitamin B6 is related to protein intake.
Because vitamin E is an antioxidant, the need for it increases with the amount of polyunsatu-
rated fatty acids in the diet. To judge whether or not the intake of a single micronutrient is
adequate often requires knowledge of the intake of other nutrients, each of which may be
consumed above or below recommended levels.
Variability in the amount of nutrients required in the diet is much greater than that due to
inherent biological variability among people. Much of this added variability is due to the
presence of factors that affect the intestinal absorption of nutrients. The bioavailability of a
dietary nutrient-the amount available for absorption-sometimes varies with the form in
which the nutrient is present (for example, hemoglobin-iron is more readily absorbed than
other food forms) and to the presence of interfering substances in the diet (for example,
tannins). Bioavailability factors have been taken into account in developing internationally
recommended intakes of protein, iron, zinc, and vitamin A; knowledge is not sufficient to do
the same for other nutrients at this time.

POPULATIONS AT RISK
Priorities for research should relate to the prevalence and consequences of micronutrient
deficiencies. A first statistic to consider is the number of people who are generally malnour-
ished. Families who do not have enough to eat are usually poor; their diets are not likely to
include much nutrient-rich food and so are likely to be low in vitamins and minerals as well
as energy.
The United Nations Administrative Committee on Coordination, Sub-Committee on Nu-
trition (ACC/SCN) reported on the world nutrition situation in 1992 (UN ACC/SCN 1992). Its
estimates of general malnutrition in the developing countries are based on three indicators:
food-energy consumption, growth of children, and body weight of women (Table 1).
Malnutrition based on consumption is defined as the population with a yearly average
energy intake (kilocalories per capita per day) below the amount needed for a light activity
pattern (1.54 x the basal metabolic rate [BMR]). The factor of 1.54 x BMR is significantly less
than energy expenditures associated with agriculture and forestry occupations (moderate or
heavy activity requires 1.7-2.0 x BMR) (FAOIWHO/UNU 1985), and as the intake is the
average over a year, it will include some families whose intakes may be much lower in the
"hungry season." By this criterion, during 1988-90, 20 percent of the developing-country
population (including China) was malnourished-some 786 million persons.
Malnourished children are defined as those under five years of age whose weight-for-age
is more than 2 standard deviations below the reference population. Weights of children show
improvement over the past 15 years in that the proportion affected has fallen from 42 percent
 3

Table 1-Malnutrition in developing countries, 1975-90

Type of Malnutrition Percent Affected Number Affected
(millions)
1974-76 1988-90 1974-76 1988-90
General
Population (all ages) with energy intake (kilocalories/capita/day)
on average below 1.54 basal metabolic rate over one year 33.0 20.0 976 786
1975 1990 1975 1990
Children (under five years) with weight below -2 standard
deviations of reference 42.0 34.0 168 184
1980s
Women (15-49 years) with weight below 45 kilograms 45.0 400
Micronutrient malnutrition
Anemia: women (15-49 years old)
Hemoglobin: less than 12 grams per deciliter (nonpregnant)
or less than 11 grams per deciliter (pregnant) 42.0 370
Iodine deficiency disorders 5.6 211
Goiter (all ages)
Vrtamin A deficiency: children (under five years) with
xerophthalmia 2.8 13.8

Source: UN ACC/SCN 1992, chaps. 3 and 4.

Notes: Data on population with low energy intake (underfed) were calculated by the Food and AgriCUlture Organization of
the United Nations. The estimates are averages for 1974-76 and 1988-90. Underweight children results are estimated
by ACC/SCN for children aged 0 through 60 months, using a cutoff of -2 standard deviations of the median National
Center for Health Statistics reference. The estimate of underweight adult women is calculated from ACC/SCN's data
base on women's nutrition. The 45-kilogram cutoff is commonly used as a basis for comparison. The prevalence
estimates exclude pregnant and lactating women, but these are included in the calculation of numbers. Anemia
estimates are based on ACC/SCN's data base on women's nutrition. The cutoff points for anemia use the World
Health Organization (WHO) reference for pregnant and nonpregnant women. Iodine deficiency disorders estimates
are based on WHO and Intemational Council for Control of Iodine Deficiency Disorders data. Vitamin A deficiency
estimates are based on WHO data.

to 34 percent, but the number has increased from 168 to 184 million, owing to population
growth.
For women, the cutoff value for general malnutrition is 45 kilograms in body weight. By
this means, women are the most frequently malnourished group; 40 percent of women, or
400 million, fail to meet the criterion. 1
These statistics imply that the population of male adults is the group least likely to be
underfed. Women of childbearing age constitute more than one-half of the underfed world
population (400 million of the 786 million with low energy intakes). Almost another one-fourth
are children below age 5, and school-age children also are at risk in households with overall
low intake.

1The 45-kilogram standard will overestimate the population currently affected if stature is low, reflecting genetics or,
more probably, growth failure in childhood. A better indicator of the current nutritional situation is body mass index (8MI),
which takes stature into account, weightlheigh~ (kilograms/meter squared). The desirable range of 8MI for healthy adults is
accepted to be 20-25, based on actuarial statistics from developed countries. International experts (WHO 1990) now propose
a 8MI of 18.5 as defining grade 1 undernutrition. This cutoff would accept a 45-kilogram (99 pounds) weight as satisfactory
for women shorter than 155 centimeters (61 inches). Other data suggest another way in which women's 8MI can be used
as a marker of chronic energy deficit in a community; ifthe average 8MI of older women (women in their forties) is not greater
than that of younger women (those in their twenties), the quantity of food available is probably inadequate (Calloway 1989).
 4

There is little doubt that women and children are more vulnerable than men to micronu-
trient deficiency because of their added nutrient requirements for reproduction and growth.
The ACC/SCN reports that 42 percent of developing-country women are anemic, owing
largely to iron deficiency (Table 1). Anemia is even more prevalent in older infants and
preschoolers, reaching 70 percent or more in sample populations. Its prevalence generally
is lower in school-age children and quite low in men. Vitamin A deficiency is reported much
more commonly in children than adults; the characteristic eye sign, xerophthalmia, is present
in about 14 million young children (Table 1).
There is, however, little evidence-none available for poor, rural, developing-country
populations-that men's energy intakes relative to energy needs are greater than women's.
Nevertheless, because men's energy needs are higher than those of women and children but
their essential nutrient requirements per unit of energy generally are the same or lower, men's
nutrient needs can more easily be met from a nutritionally dilute diet (one with relatively low
nutrient content per calorie). If micronutrients seem not to be distributed within households
according to need, this may not indicate sex or age bias. It may simply be that the special
needs of children and of pregnant and lactating women are not sufficiently recognized; the
family eats from the same pot.
 2. TYPOLOGY OF SPECIFIC NUTRIENTS

The brief review in Chapter 1 provides a skeletal framework for addressing the question
of micronutrient deficiencies in relation to foods and diets. To cope with dietary-complexity
questions, it is helpful to think of the micronutrients as fitting a typology. There are three
general food-related reasons that specific vitamins and minerals are inadequate: uneven
distribution in food classes, environmental deficiency, and low bioavailability.

TYPE 1: UNEVENLY DISTRIBUTED IN FOODS

One reason a nutrient is likely to be low in a total diet is that the principal energy sources
do not provide much of it. Vitamins At 812, and C are examples of this category. As these
nutrients are dissociated from energy, they may be low in energy-adequate diets.

Vitamin A
Vitamin A in its essential form (retinol) is present only in animal-source foods. It is
fat-soluble and occurs mainly in fatty tissues (egg yolk, milk fat, fatty fish) and in liver, where
it is stored. Most vitamin A activity in diets is derived from conversion of precursor carotenoids
in plants to retinol. The carotenoids include 26 plant pigments ranging in color from creamy
yellow to red-orange; 16 carotenoids have some pro-vitamin A activity but only 3 W-carotene,
a-carotene, and cryptoxanthin) are significant as vitamin A sources for humans. The carote-
noids, in general, have antioxidant properties and may be beneficial, but only those that can
be converted to retinol can serve essential, specific nutrient functions.
The plethora of carotenoids has made for some confusion in quantifying the pro-vitamin
A activity of foods. Originally, vitamin A activity was measured biologically (usually a rat-
growth assay) and the efficacy of pure compounds and foods was established in this way.
This method was supplanted by less-costly chemical procedures, sometimes measuring only
total carotenoids, leading to overestimation of the vitamin A activity of foods. Reliable assay
methods are now available but many tables of food composition list outdated or erroneously
converted values.
Retinol is well absorbed but the carotenes in foods are not. Pure 13-carotene theoretically
yields 50 percent retinol, but because it is poorly absorbed (20-50 percent), 6 micrograms of
13-carotene is counted as equivalent to one microgram of retinol. Cryptoxanthin and a-caro-
tene are half as effective (12:1). The fractional absorption of carotenes decreases with
increasing dosage. Absorption is low if the diet is low in fat content and is generally impaired
by conditions that damage the intestinal tract.
The carotenes are quite safe and can be consumed in amounts much higher than is
needed for vitamin A adequacy. Retinol itself is toxic in large doses. Dosage of more than
100 times the recommended daily intake (RDI) at one time or in closely spaced administra-
tions is acutely toxic; chronic toxicity follows long-term dosage with retinol at a level of about
10 times the RDI.
Vitamin A is essential for visual function and for cell differentiation. Its deficiency results
in night blindness and ultimately in loss of vision, growth retardation, damage of epithelial
tissues (skin, lining of the respiratory tract and other mucous membranes, cornea), abnor-
malities in the enamel-forming cells of the teeth, and reproductive disorders. Vitamin A is
required by all animal species in which it has been tested.
 6

Deficiency of vitamin A has been identified in 37 countries as a public health problem,

particularly in countries where the supply of retinol is very low (UN ACC/SCN 1992). It is the
leading cause of preventable blindness in the world. Mortality due to such common childhood
diseases as measles and acute respiratory infections is higher among vitamin A-deficient
children than those whose diets are higher in vitamin A or who have been given intermittent,
large doses of pure vitamin A (8eaton, Martorell, and Aronson 1993; Fauzie, Herrera, and
Willett 1994).

Vitamin 812
Vitamin 812 is required by all higher animals, but plants do not supply any of this vitamin.
Vitamin 812 (cyanocobalamin) is synthesized only by certain bacteria (including the rumen
microflora) in the presence of cobalt. Intestinal microflora of many monogastric species also
manufacture 812, but as the site of production is below the site of absorption, the vitamin is
available only to coprophagous animals. Ruminants require cobalt in the diet but not
preformed 812; humans require vitamin 812 but not cobalt.
The quantitative requirement for 812 is very small, of the order of 1 microgram per day.
The vitamin is stored and conserved efficiently. The liver of a healthy, well-nourished
omnivorous adult contains a one-to-three-year supply.
Animal-source foods are the only consistent dietary source of vitamin 812. Strict vegetari-
ans have very low intakes (less than 0.5 microgram per day), derived from environmental
contamination. There are large gaps in the food composition data base for vitamin 812, both
as to its content in items commonly consumed in developing countries but not in developed
countries and as to the effects of food preparation and processing. Vitamin 812 can be
synthesized during bacterial (but not yeast or mold) fermentation of foods. Tempeh, a
mold-treated soybean product, is sometimes found to contain 812, but only when Klebsiella
bacteria are present and the soybean curd contains cobalt (Curtis et al. 1990). Recent
analyses found neither a commercially manufactured tempeh nor home-fermented maize
meal to contain measurable amounts of 8122 . Lactobacilli, the bacteria used in such prepa-
rations as yogurt, do not produce 812, and it is a dietary requirement for some species of
bacteria.
Vitamin 812 deficiency results in a characteristic macrocytic anemia and severe damage
to the central nervous system. The anemia was first recognized as "pernicious anemia"
related to diminished production in the stomach of acid and "intrinsic factor", which are
required for absorption of the vitamin. Malabsorption of 812 occurs with disorders that impair
absorption more generally, notably infection with the common intestinal pathogen Giardia or
with fish tapeworm.
Vitamin 812 deficiency has not been regarded as a pUblic health problem, but accumu-
lating evidence indicates that it may be. Recently, physicians have reported finding 812-
responsive neurological symptoms in older adults who do not have the typical anemia
(Lindenbaum, Healton, and Savage 1988). A newer method of assessment (the presence of
an abnormal product in the urine) was used to test the 812-status of breast-fed infants; the
findings were indicative of marginal status in the infants of vegetarian mothers whose breast
milk was low in 812 (Specker et a!. 1990).

2According to unpublished findings of Professor Susan Oace, University of California at Berkeley. The tempeh was
purchased in an Asian market in San Francisco. The fermented maize was obtained from several households in the Embu
District, Kenya (courtesy of S. Murphy and C. Neumann). The analytical method used is specific for vitamin B12, that is, it
does not measure similar but inactive corrinoids.
 7

Some 40 years ago, a group that attended malnourished children tested the effect of
vitamin 812 (10 micrograms per day) on the growth of 11 children, 3 whose rate of recovery
was slow and 8 others chosen at random from the population already progressing as
expected; none had recognized 812-deficiency symptoms. Five children responded "dramati-
cally." In an early attempt at cost-benefit analysis, the authors concluded that "growth
responses were equivalent to another 160-240 days of regular institutional care without the
help of 812" (Wetzel et al. 1949).

Vitamin C
Vitamin C occurs in many vegetables and fruits, often those that also provide carotene.
It is present at moderate but significant levels in two staple crops, potatoes and sweet
potatoes. Animal-source foods (except liver) are low in vitamin C, and because this nutrient
is destroyed by heat, cooked animal products provide almost none. Vitamin C is also damaged
by exposure to air and. being water-soluble, it is lost in discarded soaking and cooking liquids.
Old analytical methods tended to overestimate vitamin C. A new, specific methodology exists
but is not fully reflected in most food composition tables.
Vitamin C is involved in many metabolic pathways of the amino acids. It is required for
the synthesis of adrenal hormones and for the formation of collagen in skin, tendon, bone,
and connective tissue. It is important in wound-healing and in maintaining strength of blood
vessels. Vitamin C is related to iron nutrition in several ways: it is necessary for the formation
of hemoglobin and for deposition of iron in the liver, and it increases the absorption of
nonheme iron when the two are ingested together.
Ascorbic acid is an antioxidant, and as such it may play a role in protection against
pro-oxidant substances in the environment.
Clinical signs characteristic of scurvy are prevented by a small amount of vitamin C, 10-30
milligrams per day. Different views about the need to maintain high levels of ascorbic acid in
blood and tissues for nonspecific antioxidant protection have led to a wide range of recom-
mended intakes. About 100 milligrams per day is suggested as meeting a higher requirement
in persons who smoke. Once the tissue saturation level is reached, the excess vitamin is
excreted in the urine, mainly unchanged.
Major outbreaks of scurvy occurred in the past as a consequence of failure of potato crops
in Ireland and Norway. Outbreaks of scurvy have been reported in refugee camps but
deficiency of this severity is otherwise rarely seen today.

TYPE 2: ENVIRONMENTALLY DEFICIENT

Nutrients in this category are trace minerals that are required by humans but not by plants
and that commonly are deficient in soils and water. Iodine and selenium are examples.
Diseases due to deficiency of these minerals are more characteristic of geographic areas
than of specific dietary patterns.

Iodine
Iodine is a constituent of thyroid hormones. These potent hormones regulate cellular
energy metabolism and are required for protein synthesis and, in turn, for growth, formation
of nerves and bone, reproduction, and mental function. The most recognized physical sign
of iodine deficiency is goiter, an enlargement of the thyroid gland (Table 1).
 8

Goiter was recognized in China in the third millennium B.C. and was treated by dosage
with seaweed or burnt sponge. Iodine was identified as the curative substance in 1820, but
it was 100 years before the first formal test of iodine was undertaken for goiter prevention.
lodization of salt has proved to be an effective public health measure in countries where it is
mandated by law. Still, iodine deficiency disease (100) affects about 211 million persons (UN
ACC/SCN 1992).
Endemic 100 has different characteristics in different places, depending on the environ-
mental iodine level and on food consumption patterns. There is usually dwarfing, and some
degree of mental deficit. If severe iodine deficiency was experienced during intrauterine life
and early childhood, mental damage is severe and irreversible (called cretinism); deaf mutism
is common. Spontaneous abortion and neonatal death rates are high in both human and farm
animal populations. The picture that emerges with deficient but somewhat higher iodine
intakes is one of mental and physical slowness. The deficiency syndrome is worsened by the
presence of antithyroid substances (called goitrogens) in the diet or water supply. The most
commonly consumed goitrogen is cyanide in cassava, but several of the Brassica vegetables,
such as cabbage, also are goitrogenic.
The prevalence of 100 is lower where the food supply is commercialized, with foods being
brought in from iodine-adequate growing regions or the sea. Commercialization also intro-
duces other sources of iodine in the form of cleaning agents for bottling and food machinery
operations, conditioners for dough processing, and some artificial food colorants. Iodine is
increased in the milk supply by supplementation of the cows' feed, some drug treatments (for
example, for hoof rot), and, as noted, in sanitizing equipment. How much difference this
makes depends on the ability of a family to obtain or purchase these foods.
Toxicity becomes a concern when total dietary intakes reach about 1,500 micrograms,
10 times the ROI.

Selenium
In a large area of China, soils and water are low in both iodine and selenium. The
interaction of iodine and selenium in the expression of 100 is complex and still under
investigation (Arthur 1993).
Selenium is an antioxidant and substitutes, in part, for vitamin E. Selenium is needed for
formation of muscle and connective tissue, for sperm production, and for growth and
development. Both deficiency (white muscle disease) and toxicity (blind staggers) are
recognized in animals pastured on soils too low and too high, respectively, in selenium.
A particular type of heart muscle damage (Keshan disease), prevalent among young
children and women in some areas of China, is ascribed to deficiency of selenium (Arthur
1993). Human selenium poisoning from foods grown on seleniferous soil also has been
reported in China (Yang et al. 1983). Symptoms included loss of hair and nails and neurologic
damage, similar to toxic signs in farm animals. Habitual average selenium intake of those
affected was about 5,000 micrograms per day, 100 times the ROI, butthe lower limit of toxicity
is unknown.
Addition of selenium to fertilizer has proved effective in raising selenium levels in the food
supply in Finland; blood levels that had been low, reflecting soil deficiency, were raised to the
normal range (Makela et al. 1993).
The feasibility of correcting iodine deficiency through fertilizer has apparently not been
tested.
 9

TYPE 3: LOW IN BIOAVAILABILITY

Nutrients in this group are at risk due to low bioavailability. They are distributed widely in
foods and their total intakes generally correlate with energy intake. They differ from most
other nutrients in that their availability is improved by certain traditional home processing
methods. Iron, zinc, and niacin are in this category.

Iron
The most obvious symptom of iron deficiency is anemia. Anemia can be due to deficiency
of several other nutrients involved in production of red blood cells and hemoglobin (vitamins
812 and 86, folate, riboflaVin, copper, and so forth), but iron deficiency is most commonly
involved. Dietary deficiency is often complicated by blood loss due to infestation with
hookworm and schistosomes and, in women, heavy menstrual flow.
Severe anemia (defined as hemoglobin below 9 grams per deciliter) is associated with
adverse outcomes in pregnancy-increased maternal mortality, low birth weight, and prema-
turity. Impaired metabolism and temperature regulation and a number of immunologic
abnormalities are associated with iron-deficiency and anemia. Physical work capacity is
reduced with moderate anemia. Diminished physical and mental performance has been
reported in persons whose iron deficiency state is not severe enough to cause anemia, but
the evidence on this point is inconsistent.
Iron homeostasis is regulated at the step of intestinal absorption. Individuals who are
anemic or have depleted iron stores absorb iron at about twice the rate of those who are well
nourished. Lead shares an absorptive mechanism with iron, so lead absorption also is
increased in iron deficiency. As lead impairs cognitive function, this adds compleXity to
evaluation of the role of iron in mental performance.
Chronic iron toxicity, causing damage to the liver, results from repeated ingestion of iron
accumulated in acidic foods processed or stored in iron pots (such as beer or toddy) or
medicinal iron. Acute poisoning is most commonly reported in children who eat iron tablets;
a dose of 30 milligrams per kilogram of body weight is toxic, and 60 milligrams per kilogram
of body weight may be fatal.
Dietary characteristics that affect bioavailability of iron include the amount of iron from
hemoglobin and the presence of promoting factors-the consumption of meat, fish, and
poultry protein and the amount of ascorbic acid consumed along with iron from other
sources-and of inhibiting substances such as phytate and tannins. Summing these factors,
anemic persons would be expected to absorb about 20-30 percent of the iron in a typical
developed-country diet, but only 5-10 percent of that in typical low-cost, developing-country
diets. To achieve equal iron adequacy, the developing-country diet would have to provide two
or three times as much iron at the same caloric intake.

Zinc
Zinc is a component of 60 or more enzyme systems, many affecting the formation of
cellular proteins. As a consequence, symptoms of deficiency cover a wide spectrum. In both
humans and animals, zinc deficiency results in retarded growth and sexual development,
depressed immunity, loss of taste acuity and appetite, dermatitis, loss of hair, and poor
reproductive performance.
Like iron, zinc is closely regulated for homeostasis, mainly at the intestinal level. Zinc is
absorbed more efficiently by persons in low zinc status and proportionately more from small
than large doses. Zinc absorption is enhanced in the presence of animal-source protein and
 10

diminished by the presence of phytate, and, possibly, calcium. In Western-type diets, about
30-40 percent of zinc is available, but only about 15 percent is available from a diet low in
animal protein, low to moderate in calcium, and with a phytate-zinc molar ratio in the range
of 15-30.

Iron and Zinc. Few (if any) studies of iron deficiency have included evidence showing that only
iron is deficient and therefore solely responsible for the symptoms documented. A recent U.S.
study of pregnant women and girls from a poverty population found that iron and zinc intakes
were closely correlated and both were correlated with energy intake (Scholl et al. 1993). Both
iron and zinc deficiencies were shown to be risk factors for prematurity and their effects were
additive. After adjusting for caloric intake and other confounding factors, the odds of having a
very preterm delivery were increased to 1.86 for women with anemia but with higher zinc intake,
to 2.92 for those without anemia but with low zinc intake, and to 5.44 for those who were both
anemic and had low zinc intake.
Unsnarling the independent roles of iron and zinc presents an interesting challenge to
nutritionists. Resolution clearly will require laboratory study. In the real-life situation, diets low
in one are likely to be low in the other, and the bioavailability of both elements is affected
adversely by interfering substances commonly present in diets based heavily on cereals and
grain legumes.
Milling and refining of cereals is one way to reduce the content of the phytate and fiber
that interfere with mineral absorption, but in the process, the amounts of many micronutrients,
including iron and zinc, are reduced as well. There are, however, traditional food preparation
methods that significantly reduce the phytate content of cereals and legumes without
significantly lowering nutrient content. These include yeast-leavening, bacterial fermentation,
and germination.
Diets often contain much adventitious iron and zinc. As noted above, acidic foods that
are processed in iron pots take up iron, sometimes reaching quite high levels of readily
absorbable metallic iron. Zinc also is taken up from galvanized containers, and water collected
from galvanized roof covers will have added bioavailable zinc.
Minerals in the form of clay or plant ash may be added to foods to alter taste, color, or
texture. As harvested, foods often contain dust or bits of soil. The very high iron content of
teff, for example, is from this source, not the grain per se. Minerals in these forms are not well
absorbed.
The minerals that are not intrinsic to the food itself may be included in the home diet
(depending, for instance, on whether or not the product was washed thoroughly before
processing) and in food sampled for analysis. This irregular contamination plus natural
variability in the amounts of minerals taken up by plants from soils of different composition
argues for more attention to sampling and analysis of minerals in common foods.
The trace element content of animal-source foods tends to be more uniform than that of
vegetable matter but varies with the animal's diet to some extent. Veal, for instance, is lower
in iron than is beef as a result of calves being fed an iron-poor diet. Liver content is
exceptionally variable because it is a storage organ for iron, zinc, and many other substances.

Niacin
Pellagra is the classical deficiency disease associated with the B vitamin niacin. It is
uncommon now but once was a major cause of morbidity, including mental derangement
(fortunately reversible). The history of pellagra is illustrative of the complex relationships
between foods and nutrients.
 11

Pellagra appeared in Europe when maize was introduced from the Americas and became
rampant in the United States with the development of roller milling of cereals. The mystery of
why Europeans and Africans developed pellagra when native American people did not
remained unsolved for many years after niacin was identified as the principal deficiency in
pellagra-producing diets (Carpenter 1981). A Mexican student working at Cambridge discov-
ered that alkaline treatment of maize (soaking in lime water) releases niacin from the bound
form in which it is stored in mature grain, making it absorbable. A second factor proved to be
the low tryptophan content of low-cost diets, as that essential amino acid can be converted
to niacin (at a ratio of 60:1).3 Poverty, which resulted in small consumption of other sources
of niacin and tryptophan (milk, meat, and so forth) and roller milling, which removed nutrient-
rich cereal germ, account for the population distribution of pellagra. Pellagra is no longer
prevalent but still seen, usually associated with poverty and sometimes with alcohol abuse.
The niacin in mature cereal grains is about 30 percent absorbable without treatment.
While alkaline processing improves niacin availability, it is not altogether desirable; it is
destructive to riboflavin and thiamin, which also may be at risk of deficiency.

3The metabolic conversion of tryptophan to niacin requires iron and copper cofactors. The efficiency of this conversion
was shown to be reduced in iron-deficient chicks, another example of the interrelatedness of micronutrient deficiencies
(Oduho, Han, and Baker 1994).
 -/';-v .

3. FOOD PATTERNS AND NUTRIENT ADEQUACY

IN THREE POPULATIONS

Classification of nutrients into broad categories should make it possible to predict which
nutrients are likely to be at risk if the habitual pattern offood consumption is known. Nutrients
of Type 1 relate to diet diversity questions and those of Type 3 to the choice of basic staples
and methods of food preparation. To predict the risk of Type 2 nutrients (specific trace
elements) requires knowledge of local soil and water but, importantly, also of the extent to
which the population is reliant on locally produced foods. Good predictive ability will allow
improved targeting of research.
These issues are examined here, drawing on findings from research carried out in Egypt,
Kenya, and Mexico. 4 The Collaborative Research Support Program on Nutrition and Human
Function (NCRSP) involved field studies replicated in the three locations. The studies were
naturalistic, designed to test hypotheses about the impact of chronic, mild-to-moderate
energy deprivation on reproduction, cognition, and morbidity. The three archived data sets
have been used to look across the sites for comparability of associations between food and
nutrient intakes and between intakes and growth of children.
About 300 households were studied in each location for one year during 1984-86. Food
consumption was recorded two days each month for the household, the lead male and female
of the household, a toddler (for the period from 18 to 30 months of age), and/or a schooler
(between 7 and 9 years). As available tables of food composition were incomplete, in order
to convert food to nutrient intakes, the International Minilist (I ML) was created. This is a single
data base with about 40 nutrients and related factors in about 200 class-representative foods
(Murphy and Calloway 1991). Nutrient intakes were computed by cross-referencing actual
foods to IML listings.
For analysis of dietary patterns, foods were aggregated into eight conventional broad
categories (staples, fruits and vegetables, and so forth) comprising 26 subgroups according
to specific food type (maize, wheat, and so forth).

CHARACTERISTICS OF THE STUDY HOUSEHOLDS

The mean size of households was about eight persons in all sites (Table 2). The mean
household ages of 15-18 years reflect the selection of sample populations to include
reproductive function as an outcome, but the households are not atypical for developing
countries.
Households were examined for "general malnutrition" according to the ACC/SCN criterion
of average yearly intake below 1.54 x BMR. The total household energy intake was divided
by the summed BMR of all its members to generate this adequacy ratio. Household intake in
Mexico was 1.71 x BMR, but in Egypt and Kenya, values were lower, 1.57 and 1.49 x BMR,
respectively. The value for Mexico is consistent with energy needs for agricultural work and

4The Collaborative Research Support Program on Nutrition and Human Function (NCRSP) consisted of three field
research projects that followed a common protocol. Collaborating institutions were the National Institute of Nutrition, Mexico
(A Chavez) with the University of Connecticut (L. Allen and G. Pelto); the National Institute of Nutrition, Egypt (0 Galal) with
the University of Arizona (G. Harrison), the University of Kansas (N. Jerome), and Purdue University (A. Kirksey); and the
UniverSity of Nairobi (N. Bwibo) with the University of California, Los Angeles (C. Neumann). The University of California at
Berkeley was the managing institution (D. Calloway, G. Beaton, J. Balderston, S. Murphy, and H. Horan). The program was
supported by grants from USAID and the collaborating institutions Each project has reported its findings separately. Data
from the three projects were analyzed at Berkeley; much of the material included here is taken from the final report to USAID
(Calloway et al. 1992).
 13

Table 2-Household and individual energy adequacy, intake, and percentage

from fat in Egypt, Kenya, and Mexico, 1984-86
Egypt Kenya Mexicoa
Item (n=235) (n=286) (n=43)
(mean)b
Household size (persons) 7.9 (2.7) 7.7 (2.5) 8.5 (3.2)
Average age of household members (years) 18.4 (5.0) 15.4 (2.7) 17.3 (5.4)
C
Household energy adequacy ratio 1.57 (0.39) 1.49 (0.34) 1.71 (0.48)
Household energy intake per person (kilocalories
per capita per day) 1,838 (448) 1,606 (391) 1,874 (513)
Percentage offat in lead female dietd 21.7 (3.5) 10.9 (1.7) 20.3 (4.3)
Percentage of fat in toddler diet 24.4 (7.2) 12.2 (4.3) 23.4 (7.3)

aHouseholds from community number 1 except for percentage offat in diet, which is for all mothers and toddlers.
bStandard deviations are in parentheses.
cMean overall days of record of the total energy intake from household food, divided by the sum of household members'
basal metabolic rates (BMRs). BMRs were calculated for each person according to equations published in FAOIWHO/UNU
1985.
dValues shown are for the mothers of toddlers included in the nutrient analysis; n=96 for Egypt, 100 for Kenya, and 59 for
Mexico.

household labor in the absence of labor-saving appliances. That for Egypt also is consistent
with the easier living and working conditions in the periurban study village, Kalama. The value
for Kenya rates the population as generally malnourished and suggests that energy availabil-
ity may have limited household activity and children's growth.

DIET CHARACTERISTICS
The percentage of energy from each of eight food groups consumed by toddlers in the
three sample populations and comparable figures for U.S. preschoolers are shown in
Figure 1. Grains form a much greater portion of the diets of young children in the NCRSP
countries than in the United States-62 percent of calories in Mexico and 53 percent in Kenya
and Egypt as compared with 32 percent in the United States. The proportion of energy from
dairy products in the U.S. diets (19 percent) is two to three times the proportion in the NCRSP
countries (6-9 percent). The Egyptian children's diets have lower proportions of starchy
staples and higher proportions of total animal-source foods than those of Kenya and Mexico,
but they are still much more like the diets of these two countries than the U.S. diets. If dietary
levels of animal products and of unrefined grains are associated with the growth of children,
it should be possible to see these effects in the NCRSP countries, where intakes are often
at the extremes. Bivariate analysis shows the expected association between household
socioeconomic status (SES) and food-group intakes. Children from families of higher SES
categories generally consumed more foods that are not produced at home and are costly
(animal prOducts, separated fats, sugars and sweets, potatoes, wheat). Children in these
families were generally less likely to consume maize, legumes, and leafy green vegetables.
The predicted outcome of these differences in consumption pattern would be to increase
energy intake and the amount or bioavailability or both of most nutrients in the higher SES
diets (in part due to the reduced intake of interfering substances and in part to nutrients
supplied by animal products). Vitamins A and C might be at some risk, however, unless the
lesser intake of leafy greens is compensated for by selection of other fruits and vegetables
rich in vitamin C and carotene or animal products rich in retinol.
 14

Figure 1-5ources of energy for toddlers in Collaborative Research Support Program:

percentage of calories from eight food groups

Egypt CRSP Toddlers Kenya CRSP Toddlers

Sugar/sweets 10 Frultsivegetables 3

Fruits/vegetables 4 Root vegetables 1 a

Milk/cheese 9

Root vegetables 3 Sugar/sweets 6

Grains 54 Grains 52
MeaUeggs 9 Milk/cheese 7

Legumes/nuts 4 MeaUeggs 1
Fats/Oils 7
Legumes/nuts 10
Fats/Oils 3

Mexico CRSP Toddlers U.S. Preschoolers

Fats/Oils 9 Sugar/sweets 10

Sugar/sweets 7 Grains 31
Fruits/vegetables 11
MeaUeggs 6

Milk/cheese 6
Root vegetables 2
Grains 62 Milk/cheese 19
Legumes/nuts 7 Fats/oils 3
Frultsivegetables 1 Root vegetables 5 Legumes/nuts 4

MeaUeggs 17

Sources: Values for Egypt, Kenya, and Mexico are from Calloway et al. 1992; values for U.S. preschoolers are calculated
from USDA 1987.

NUTRIENT INTAKE
Toddler energy intake was much lower in Kenya than in Egypt and Mexico (Table 3).
Mean intake met the average toddler requirement (103.6 kilocalories per kilogram) (FAO/
WHO/UNU 1985) except in Kenya, where it was about 80 percent of the standard.
Total protein contributed 11-12 percent of energy in all groups, as is typical of diets
worldwide, but the amount derived from animal sources differed. Kenya toddlers' average
intake was only 4 grams per day (equal to about 125 milliliters (one-half cup) of milk or half
an egg), compared with higher but still ungenerous intakes of 10 grams per day in Mexico
and 14 grams per day in Egypt.
The Kenya toddler diet was lowest in fat in terms of both total amount and as a percentage
of dietary energy (percent fat). 5 In all groups, the percentage of fat intakes of toddlers is quite

51t seems anomalous that diets in both Kenya and the United States derived 3 percent of total calories from the food
group "fat and oils:' but the percentage of total dietary energy derived from fat was much higher in the United States than in
Kenya. The food group includes only separated fats and oils (such as butter, lard, and vegetable oils) added in home
preparation, whereas the total intake is the sum of these fats and oils plus the fat contributed by milk, meat, nuts, and so
forth, as well as that in processed commercial foods (such as chips and cookies). The U S. diet contained more of these
foods that contain fat.
 15

Table 3-Average daily nutrient intake of toddlers in Egypt, Kenya, and Mexico
Egypt Kenya Mexico
Nutrient (n=96) (n=100) (n=59)

Energy
Kilocalories 1,204 847 1,110
Kilocalories per kilogram 109 83 106
Protein (grams) 36 23 33
Animal protein (grams) 14 4 10
Fat (grams) 33 12 29
Sucrose (grams) 38 19 24
Dietary fiber (grams) 17 22 15
Phytate (milligrams) 796 1,066 1,666
Vitamin A (micrograms of retinol equivalents) 305 370 203
Vitamin D (micrograms) 0.31 0.06 0.44
Vitamin C (milligrams) 42 46 14
Riboflavin (milligrams per 1,000 kilocalories) 0.51 0.65 0.49
Vitamin B12 (micrograms) 1.4 0.61 1.4
Iron (milligrams) 6.8 7.0 6.8
Available iron (milligrams) 0.49 0.61 0.37
Zinc (milligrams) 5.2 3.7 5.4
Available zinc (milligrams) 1.2 0.49 0.68
Calcium (milligrams) 218 210 735
Note: Data are the average across one year between ages 18 and 30 months. Adapted from Beaton, Calloway, and Murphy
1992; Murphy, Beaton, and Calloway 1992; and Calloway et al. 1993.

close to that of their mothers (Table 2) and older siblings (data not shown), suggesting that
this statistic is a useful proxy for household diet quality. Fat energy in diets of the sample of
mothers of toddlers was about 11 percent in Kenya; it was much higher in Mexico (20 percent)
and Egypt (22 percent), but still lower than the 35-37 percent typical of developed countries
such as the United States.
Factors that interfere with absorption of nutrients were not uniform across the sites
(Table 3). All diets were high in fiber content but especially so in Kenya (26 grams per 1,000
kilocalories versus 14 grams per 1,000 kilocalories in Mexico and Egypt). Phytate intake was
high in all cases but lowest in Egypt, due to the use of somewhat more refined cereals and
yeast-leavening of bread. Phytate was twice as high in Mexico as in Egypt and intermediate
in Kenya. Tannin levels were increased in Egypt and Kenya diets, owing to consumption of
tea.
Because of the relatively low energy intake in Kenya, one might speculate that nutrients
in general might be low. Whether or not that is true, however, depends on the selection of
foods that are not principal contributors to energy, as well as those that are. Consumption of
carotene-rich fruits and vegetables was much higher in Kenya (for toddlers, 54 grams per
day) than in Mexico (5 grams per day) and Egypt (10 grams per day). Total vitamin A intake
was, however, similar in Kenya and Egypt, where the lower intake of carotene was offset by
higher consumption of retinol from animal foods (174 micrograms of retinol equivalents per
day in Egypt, 50 micrograms of retinol equivalents per day in Kenya). Total intake was lower
in Mexico, where intake of animal vitamin A did not reach compensatory levels (119
micrograms of retinol equivalents per day in Mexico). It is the micronutrients supplied
predominantly by animal products or affected by their presence that are likely to be deficient
in the diets of Kenya toddlers.
 16

PREDICTED PREVALENCE OF MICRONUTRIENT INADEQUACY

The prevalence of inadequate nutrient intakes can be estimated by a probability approach
(National Research Council 1986), with the assumption that intake and requirement are not
correlated and that requirements are distributed symmetrically. The probability that the
observed intake of an individual is below that individual's own requirement is computed; the
average of these values is the predicted prevalence of inadequacy in the group of toddlers.
The probability assessment shows the effect of the different patterns of food intake
(Table 4). The nutrients at highest risk across the countries are iron and zinc (which is due
to poor bioavailability and, in Kenya, to a lower total zinc intake from animal products). Calcium
is low, as expected, except in Mexico, where lime is added in tortillas. The probability of
vitamin inadequacy is high across a spectrum in Mexico: vitamins A and C, due to low fruit
and vegetable intakes; riboflavin, where a contributory factor is loss during alkaline processing
of maize; and vitamin 812, due to marginal intake of animal products. In Egypt, vitamin A,
riboflavin, and, for schoolers, vitamin 812 are at risk. In Kenya, vitamin 812 is at high risk,
whereas the risk of vitamin A inadequacy is low. No diet met the requirement for vitamin D,
but deficiency would be expected only in toddlers who were not sufficiently exposed to
sunlight.
The probability assessment does not identify individuals at risk of deficiency because the
individual requirement is unknown. It is of interest, however, to examine the likelihood that
diets are low in only one micronutrient. For this exercise, an intake was defined as low if there
was at least an 85 percent probability of inadequacy; seven micronutrients were examined:

Table 4-Predicted prevalence of inadequate nutrient intakes in toddler diets

Prevalence3
Mean Daily Egypt Kenya Mexico
Nutrient Requirement (n=96) (n=100) (n=59)

Iron, basal (milligrams) 0.52 65 36 88

Iron, to prevent anemia (milligrams) 0.52 36 13 43
Zinc, normative (milligrams)b 0.76 36 90 68
Zinc, basal (milligrams)b 0.54 10 57 25
Calcium (milligrams) 346 90 88 2
Phosphorus (milligrams) 346 2 6 0

Thiamin (milligrams) 0.39 1 0 0

Riboflavin (milligrams) 0.51 20 2 52
Vitamin B12 (micrograms) 0.33 3 44 8
Vitamin C (milligrams) 15 3 1 63
Vitamin A, normative (micrograms of retinol eqUivalents) 205 32 12 68
Vitamin A, basal (micrograms of retinol equivalents) 102 2 0 20
Vitamin D (micrograms) 7.70 100 100 100

Source: Adapted from published data from Beaton, Calloway, and Murphy 1992; Murphy, Beaton, and Calloway 1992; and
Calloway et al. 1993.
Note: Mean requirements calculated from 11-kilogram body weight; references and requirement distribution are given in
the sources cited.
apredicted prevalence of inadequacy was 0 for copper, magnesium, niacin, folate, and vitamin B6.
bNormative requirements allow for tissue reserves; basal requirements are sufficient to prevent clinical symptoms but not
storage. For iron, basal requirement, and requirement to prevent anemia are the same, but 50 percent greater absorption
efficiency is assumed in anemia.
 17

vitamins A, C, and 812, riboflavin, calcium, available iron, and available zinc (Table 5). 8y this
criterion, 32 percent of the toddler diets in Egypt, 10 percent in Kenya, and 17 percent in
Mexico were inadequate in only one nutrient; the low nutrient was in all but one case a
mineral-calcium or iron in Egypt, calcium or zinc in Kenya, and iron or zinc in Mexico. Diets
low in two nutrients were usually low in these same pairs of minerals. Vitamin A occurred as
the single inadequate nutrient in one diet in Mexico, and there were combinations of minerals
with vitamins A, C, and riboflavin where two or three nutrients were low in the Mexican diets.
Vitamin 812 was the second or third nutrient at risk in Kenyan diets. In each country, diets
likely to be low in more than one nutrient were lower in total energy and the proportion of
energy from animal-source foods than were diets with one or no nutrients rated low by the
criterion used.
In the study location in Kenya (Embu District), mild iodine deficiency is endemic (Neu-
mann, 8wibo, and Sigman 1992). Goiter was present in about 20 percent of the women in
the sample population and in 8 percent of schoolchildren. The project reported that 27 percent
of households used iodized salt about 50 percent of the time. In the absence of analytical
iodine values for locally produced and consumed foods (including iodized salt), no dietary
assessment of inadequacy is possible.

Table 5-Frequency of toddler diets predicted to be inadequate in zero to six

or more nutrients
Number of Inadequate Nutrients Egypt Kenya Mexico
(percent of sample)

0 12 10 7
1 32 10 17
2 33 35 24
3 9 36 24
4 9 6 15
5 3 3 7
6 or more 0 0 7

Note: Micronutrients included in this analysis are vitamins A, C, and B12, riboflavin, calcium, available iron, and available
zinc. Requirement standard is normative except for iron, which is basal (see Table 4). For this analysis, the cutoff
point for inadequacy is 85 percent probability. Columns may not sum to 100 because of rounding.

INTAKE AND OUTCOME

Many positive associations of nutrient intake variables with children's size and growth
were observed. These associations remain significant after adjustment for energy intake, so
the possibility is considered that larger children are simply eating more food and, thus, more
of all nutrients. Nutrients significantly associated with child size or growth in one or more
countries include vitamins A, C, 812, and riboflavin, and available (not total) zinc.
Many of the micronutrient-size associations remain significant in multivariate models that
include nondietary variables (socioeconomic status, sanitation score, sex, and maternal size).
In fact, a number of nondietary variables that showed significant bivariate correlations with
size and growth became nonsignificant when adjusted for nutrient variables. The nutrients
that were brought into these models contributed independent information about size and
growth.
 18

From the standpoint of practical application, it would be helpful if food or food pattern
intake or both could be substituted for individual nutrient intake for program-planning pur-
poses. The NCRSP findings suggest that it might be possible. In general, the pattern of food
intake was as strongly correlated with child size and growth as was nutrient intake.
Maize intake is uniformly associated with poorer growth and size in Mexico and Kenya,
where intakes were very large. This raises a question of the conditions under which greater
energy intake is bought at a price of diminished micronutrient availability.
It will surprise no one that milk and cheese intakes marked better growth of children. It
demonstrates once more that provision of naturally complementary foods is an efficient way
to supplement an array of nutrients that are low in most staple foods. The provision of more
food in some circumstances and different, nutritionally protective food in almost all may be
the route by which improved social and economic conditions promote the welfare of children
and their families.
 4. STRATEGIES FOR INTERVENTION

How the nutrition question is framed determines the choice among strategies for address-
ing it. Nutrition has been approached far too often from the perspective of a single nutrient
or a single discipline, leading to a narrow, if not plain wrong, definition of the problem, and to
inappropriate and hence ineffective remedies. 6 The fact of prevalent micronutrient deficiency
is not in dispute. But as seen here, where one nutrient is deficient, others probably are as
well; so, the nutritional status of those who are affected-and the associated health and
developmental outcomes-are unlikely to yield to strategies based on improvement of one
nutrient at a time.

HEALTH-SECTOR INTERVENTION
Yet there is a case to be made, at least in the short term, for specific intervention to prevent
the most serious manifestations of micronutrient deficiency, such as blindness and mental
retardation, that have lasting, crippling effects on societal development.
Periodic administration of large doses ofvitamin A and iodine to children clearly is effective
in preventing and correcting these serious deficiencies. Provision of iron tablets, particularly
to pregnant women, is an accepted public health practice. Ideally, and for cost-containment,
these pure supplements are given only to those whose need is unequivocal. The downside
is that these remedies require regular follow-up by health workers and consistent monitoring
of the population to assure program coverage. When claims on a national budget exceed
available funds, such programs often are not sustained. When the program stops, the problem
recurs unless the programs are incorporated as interim components of a more broadly based
strategy that involves both agricultural and health sectors.

FORTIFICATION OF FOODS
Fortification of an appropriate food vehicle with specific nutrients may be a better but still
imperfect solution. lodization of salt has been a clear success in many countries. The
experience in Kenya cited above is, however, cautionary: only a quarter of households used
fortified salt part of the time; use of iodized salt was positively correlated with economic status
and "modernity" of the adults. The reason cited for nonusage was cost. If a fortified product
is priced beyond their means, it cannot reach those most likely to be at risk of deficiency.
In theory, it should be possible to include in one or two products all the nutrients shown
to be at risk in a target population, in amounts that are needed by low-level consumers and
below the amount that may harm a high-level consumer of the product. In practice, this is
difficult to achieve at reasonable cost (because of nutrient-incompatibility, flavor and color
problems, stability, and so forth), even if the right vehicle and dosage can be assured.
Effective fortification requires a developed food processing industry and marketing
system. It adds to product cost. Effective fortification also requires a watchdog agency to
assure compliance. Even in the United States and Canada, which have long histories of food

6The "Protein Gap" is a textbook example. Assignment of causality for prevalent malnutrition to lack of protein per se
(when what was lacking was enough food and food of good quality) led to costly ventures into fish protein concentrate,
single-cell protein, a cupboardful of unexploitable weaning foods, and high-lysine maize.
 20

fortification and bureaucratic oversight, products do not always deliver what they claim and
are mandated by law to contain (Chen et al. 1993).

MORE AND BEITER FOOD

The logical solution to the nutrient problem is to assure intake of enough food of the right
kind. How to do this, where and how to intervene in the complex, multifactorial food system,
is at the crux of the CGIAR research agenda.
From their inception, the CGIAR centers have focused on two overarching nutrition-re-
lated goals. The first is to assure an adequate supply of food (defined as energy and protein)
sufficient to offset surging growth of population; the second is to reduce poverty in the rural
sector while lowering the price poor urban consumers must pay for their food.
The centers' research must be judged an unqualified success in meeting the first goal as
conceptualized. Hunger is less prevalent than in the past, despite increased numbers to be
fed, and where large numbers are affected, the root cause is not global insufficiency of food.
Mounting concern for sustainability of the natural resource base, coupled with continued
population growth, has added a new dimension to the centers' task. The centers' goal is
redefined: yields and yield-ceilings must be raised without damage to the environment, while
keeping poverty- and cost-reduction firmly in mind. Is it reasonable to add a requirement for
increased and perhaps different micronutrient content to the agenda? Perhaps.
Having mandates for staple crops, the commodity-oriented centers might direct their
attention to the Type 3 nutrients. Consider an example: cereals and grain legumes are the
principal source of iron in developing-country diets. Diets that meet calorie requirements
usually have enough total iron to meet human requirements, but bioavailability is so low that
the amount actually absorbed very often is insufficient. Crop genotypes have been identified
that have increased capacity for extraction of iron from soil and deposition in seed; these
genotypes are also more resistant to environmental stressors. Introduction of, for example,
iron-rich maize would improve developing-country diets, provided the iron were biologically
available. If the iron is linked to equally increased levels of interfering substances, as is likely,
the change might not only fail to improve iron nutrition, it could be damaging to zinc as well.
Full consideration of trade-offs requires more information about the new genotype (compo-
sition, production variables, and cost), the target population (food consumption and expen-
diture patterns, food preparation, and so forth), and the feasibility of alternative strategies for
improving iron nutrition.
If an agro-ecoregional center were established, it might approach nutrition differently.
Such a center logically would monitor the cropping patterns in a region as well as the
agricultural inputs (soil, water, fertilizer, labor). It would investigate consumption patterns of
both rural and urban populations in its purview. The center then would be well positioned to
predict which vitamins and minerals are likely to present problems and to test the feasibility
of modifying the crop and livestock mix to achieve the desired balance of nutrients. It would
not, I think, be able to do all this effectively without substantial effort in the social sciences
and access to modern nutrition information.
Some years ago, when the Technical Advisory Committee began a review of CGIAR
priorities and strategies, the author asked a colleague for her thoughts on the information
needed for assessment of interventions. Louise Fortmann (1990) responded as follows:

Socioeconomic understanding of how a particular crop is produced and used

in a given rural population system is critical for the production of effective and
usable innovations. Research is necessary to determine who grows a particu-
 21

lar crop, where, for what purposes, when, how, and why. This information is
necessary to determine how a given technological innovation can be linked
with a functioning agricultural enterprise. Such research is more complicated
than it might appear at first blush because, in some countries, the farm is not
unified at the household level. Thus, the question "who" must be answered
not only for different classes and ethnic groups but also for agricultural
enterprises managed by different members of the same household. The
question "where" must be answered not only in terms of geographic locations
and categories (such as soil type and rainfall) but also for social categories
such as tenurial niche (including common property). The question "for what"
is important both because of the need to take competing and sometimes
incompatible uses into account and because varying production objectives
dictate varying evaluation criteria. Yield per unit area is not always an
appropriate evaluative criterion. A farmer may be more interested in a crop,
variety, or practice that allows him/her to circumvent constraints such as
shortage of labor or shortage of cash or to meet a particular consumption
need. Social science research is also needed to enable technical researchers
to take into account what other crops and land uses may be displaced or how
the division of labor may be changed if the production parameters for this crop
are changed. Finally, social science research is frequently the only kind of
research that reveals the existence and function of strategic crops and
technologies that are rarely covered by traditional research institutes-for
example, the importance of camels in dryland Africa, the importance of fodder
trees for livestock production, or the importance of millet or amaranthus as
food crops. Similarly, given the more narrowly focused disciplinary boundaries
within which most biological scientific research is organized, social science
research, with its mandate to explain the context of agricultural production, is
more likely to reveal practices and technologies that cross disciplinary bound-
aries-the fodder benefits of sorghum, for example. Thus, in depth, social
science research allows the discovery of linkages to packages of practices
and knowledge missing from disciplinary views of agriculture. In sum, social
science research is essential to match crop research to the messy reality of
on-the-ground indigenous technical knowledge and practices.

A fundamental component of the research agenda is to define the nutrition problem in

program-relevant terms. The target population is not homogeneous, so one remedy is unlikely
to serve all. To intervene efficiently and effectively requires knowing fairly precisely what a
popUlation lacks and why. That knowledge is no less necessary for selecting crop-modification
strategies than for formulating policy. The more varied the talents that are brought to bear on
the task of problem-definition, the richer will be the knowledge base and the more secure the
design of intervention strategies.
At a minimum, what the centers and their partners must do is be aware of micronutrient
concerns, lest by inadvertence they do harm. This injunction applies not only to the mainte-
nance of the nutritional quality of the centers' crops but to the potential impact of new
technologies on the carefully balanced and often marginal intake patterns of the poor.
 22

REFERENCES

Arthur, J. R, ed. 1993. Interrelationships between selenium deficiency, iodine deficiency, and thyroid
hormones. American Journal of Clinical Nutrition 57 (2S): 235S-318S.

Beaton, G. H., D. H. Calloway, and S. P. Murphy. 1992. Estimated protein intakes of toddlers:
Predicted prevalence of inadequate intakes in village populations in Egypt, Kenya, and
Mexico. American Journal of Clinical Nutrition 55 (4): 902-911.

Beaton G. H., R Martorell, and K. J. Aronson. 1993. Effectiveness of vitamin A supplementation in

the control of young child morbidity and mortality in developing countries. ACC/SCN
State-of-the-Art Series, Nutrition Policy Discussion Paper 13. Geneva: World Health Or-
ganization.

Calloway, D. 1989. The functional consequences of malnutrition and implications for government
policy: Case studies in Egypt, Kenya, and Mexico. PEW/Cornell Lecture Series on Food
and Nutrition Policy. Ithaca, N.Y., U.S.A: Cornell University.

Calloway, D. H., S. P. Murphy, G. H. Beaton, and D. Lein. 1993. Estimated vitamin intakes of toddlers:
Predicted prevalence of inadequacy in village populations in Egypt, Kenya, and Mexico.
1993. American Journal of Clinical Nutrition 58 (3): 376-384.

Calloway, D., G. Beaton, J. Balderston, S. Murphy, and H. Horan. 1992. Village nutrition in Egypt,
Kenya, and Mexico: Looking across the CRSP projects. Final report to USAID. University
of California at Berkeley, Calif., U.S.A

Carpenter, K. J. 1981. Effects of different methods of processing maize on its pellagragenic activity.
Federation Proceedings 40: 1531-1535.

Chen, T. C., Q. Shao, H. Heath III, and M. F. Holick. 1993. An update on the vitamin D content of
fortified milk from the United States and Canada. New England Journal of Medicine 329:
1507.

Curtis, P. R, R E. Cullen, and K. H. Steinkraus. 1990. Symposium on indigenous fermented foods,

Thailand, 1977. Cited in B. K. Mital and S. K. Garg, Tempeh-Technology and food value.
Food Reviews International 6: 213-224.

FAOIWHO/UNU (Food and Agriculture Organization of the United NationslWorld Health Organiza-
tion/United Nations University). 1985. Energy and protein requirements: Report of a joint
FAOIWHO/UNU consultation. WHO Technical Report Series 724. Geneva: World Health
Organization.

Fauzie, W. W., M. G. Herrera, and W. C. Willett. 1994. Dietary vitamin A and the risk of mortality
among children. American Journal of Clinical Nutrition 59: 401-408.

Fortmann, L. 1990. Human linkages. Discussion paper prepared for the Technical Advisory Com-
mittee, Consultative Group on International Agricultural Research. University of California,
Berkeley, Calif., U.S.A

Lindenbaum, J., E. B. Healton, and D. G. Savage. 1988. Neuropsychiatric disorders caused by

cobalamin deficiency in the absence of anemia or macrocytosis. New England Journal of
Medicine 318: 1720-1728.
 23

Makela, A. L., V. Nanto, P. Makela, and W. Wang. 1993. The effect of nationwide selenium
enrichment of fertilizers on selenium status of healthy Finnish medical students living in
southwestern Finland. Biological Trace Element Research 36: 151-157.

Murphy, S. P., and D. H. Calloway. 1991. Developmentofa data base for the nutrition CRSP project.
In Proceedings of the Fifteenth National Nutrient Databank Conference, ed. M. R. Stewart,
125-130. Ithaca, N.Y., U.S.A.: The CBORD Group.

Murphy, S. P., G. H. Beaton, and D. H. Calloway. 1992. Estimated mineral intakes of toddlers:
Predicted prevalence of inadequacy in village populations in Egypt, Kenya, and Mexico.
American Journal of Clinical Nutrition 56 (3): 565-572.

National Research Council. 1986. Nutrient adequacy: Assessment using food consumption surveys.
Report of the Subcommittee on Criteria for Dietary Evaluation. Washington, D.C.: National
Academy Press.

Neumann, C., N. Bwibo, and M. Sigman. 1992. Diet quantity and quality: Functional effects on rural
Kenyan families. Kenya Project Final Report, Phase 11-1989-1992. University of Califomia,
Los Angeles, Calif., U.S.A.

Oduho, G. W., Y. Han, and D. H. Baker. 1994. Iron deficiency reduces the efficacy of tryptophan as
a niacin precursor. Journal of Nutrition 124: 444-450.

Scholl, T. 0., M. L. Hediger, J. I. Scholl, R. L. Fischer, and C.-S. Khoo. 1993. Low zinc intake during
pregnancy: Its association with preterm and very preterm delivery. American Journal of
Epidemiology 137: 1115-1124.

Specker, B. L., A. Black, L. Allen, and F. Morrow. 1990. Vitamin B12: Low milk concentrations are
related to low serum concentrations in vegetarian women and to methylmalonic aciduria in
their infants. American Journal of Clinical Nutrition 52: 1073-1076.

TAC (Technical Advisory Committee) Secretariat, Consultative Group on International Agricultural

Research (CGIAR). 1992. Review ofCGIAR priorities and strategies. AGRlTAC: IAR/92/18.
Rome: Food and Agriculture Organization of the United Nations.

UN ACC/SCN (United Nations Administrative Committee on Coordination/Sub-Committee on Nutri-

tion). 1992. Second report on the world nutrition situation, vol. 1. Geneva.

USDA (United States Department of AgriCUlture). 1987. Nationwide food consumption survey:
Continuing survey of food intakes by individuals, women 19-50 years and their children 1-5
years, 4 days, 1985. NFCS, CSFII Report No. 85-4. HyattSVille, Md.: USDA.

Wetzel, N. C., W. C. Fargo, I. H. Smith, and J. Helikson. 1949. Growth failure in school children as
associated with vitamin B12 deficiency-Response to oral therapy. Science 110: 651-653.

Wheeler, E. F. 1980. Nutritional status of savanna peoples. In Human ecology in savanna environ-
ments, ed. D. R. Harris, 225-238. New York: Academic Press.

WHO (World Health Organization). 1990. Diet, nutrition, and the prevention of chronic diseases: A
report of WHO study group. Technical Report Series 797. Geneva.

Yang, G. Q., S. Wang, R. Zhou, and S. Sun. 1983. Endemic selenium intoxication of humans in
China. American Journal of Clinical Nutrition 37: 872-881.
Doris Howes Calloway is professor emerita at the Department of Nutritional Sciences, University
of California at Berkeley.

-.2~ -