Golden Rice: What Role Could It Play in Alleviation of Vitamin A Deficiency?
By Richard Robertson(UIUC), Laurian Unnevehr (UIUC) and David Dawe (IRRI) 
Vitamin A Deficiency (VAD) is an important nutritional problem in the developing world (World Bank, 1993). Vitamin A’s primary physiologic role is in vision and maintenance of the general health of the eye, with a myriad of secondary roles, such as maintenance of the immune system. Supplementation or increased consumption of carotenoids in deficient populations has been found to substantially reduce morbidity and mortality for children (Sommer, 1997). VAD is prevalent among the poor in Asia, because their diets are dependent on rice, which does not contain Vitamin A precursors (FAO, 1993).
Golden Rice (GR) was developed to provide a new, alternative intervention to combat VAD by genetically engineering rice to contain beta-carotene (the vitamin A precursor found in plants) in the endosperm of the grain (Toenniessen, 2000). GR’s genetic engineering heritage has inspired conflicting opinions concerning its possible effectiveness and desirability. Critics have argued that the resources devoted to GR development could be better used to support traditional interventions. Proponents have argued that GR represents an important complementary intervention that could succeed in reaching more of those at risk for VAD.
This paper examines the potential benefits from golden rice in two ways. First, if GR can be developed into a viable field crop, can it deliver significant amounts of beta-carotene into the diets of high-VAD-risk children? We use children’s food intake data from the Cebu region of the Philippines to simulate the potential benefits of GR consumption. Accurate food intake data for at risk populations are relatively rare, and the publicly available Cebu Longitudinal Health and Nutrition Survey provides a unique data set to simulate the potential benefits of GR. Second, how would GR compare to existing interventions in terms of monetary cost, effectiveness of delivery, and coverage of subpopulations at risk? We examine evidence from the Philippines and Indonesia to see how GR would compare to other intervention programs.Vitamin A Deficiency in Asia
Of the various macro- and micronutrient deficiencies afflicting poor populations, VAD is unique with its well defined and debilitating effects. VAD is caused by insufficient consumption of foods containing vitamin A (vA) and beta-carotene (which is metabolized into vitamin A). Preformed vA is found in animal products while beta-carotene resides in dark green, yellow, and orange plants as well as in animal products.
Vitamin A (in the form of various retinol derivatives) is a fat soluble vitamin stored primarily in the liver with various roles, the primary being sight and general eye health. When a person does not eat enough vA and beta-carotene for long periods of time, the most obvious results are nightblindness, Bitot’s spot (foamy patches on the surface of the eye), xeropthalmia, and in extreme cases, blindness (NAS, 2001). VAD has also been associated with increased child mortality and vulnerability to infection, particularly measles and diarrhea (Sommer, 1997). This increased morbidity and mortality are apparent even before the appearance of xeropthalmia, and supplementation has had a dramatic effect in reducing mortality among children ages 6 months to 6 years in Nepal, Indonesia, India, and Ghana (Sommer, 1997).
VAD is assessed through looking for the clinical signs or by monitoring blood serum levels (<10µg retinol/dL considered deficient; <20µg/dL considered low). However, due to the liver’s storage capacity, bloodstream levels are depressed only when liver stores are virtually gone. Thus, a deficiency detectable in this way is indicative of a chronic problem. Lastly, VAD is detected by dietary analysis. This method is not as definitive as the other two, but is the primary method of identifying broad populations at risk.
Vitamin A requirements, as estimated by the National Academy of Sciences (2001) are summarized in Appendix Table 1. There are two different numbers reported for requirements: the Estimated Average Requirement (EAR) and the Recommended Daily Allowance (RDA). The EAR measures the average requirement over all persons in the category. The RDA is set so as to provide 97-98% of individuals with a sufficient intake and is thus 40% higher than the EAR in the case of vitamin A. For children under 13 years of age, the RDA is in the range of 300 to 600 mg per person per day, depending upon age and gender. These RDA’s are expressed in Retinol Activity Equivalents (RAE’s).
There are several ways to measure vitamin A and beta-carotene; all are expressed in micrograms (g). An International Unit (IU) of retinol is 0.3 µg and an IU of beta-carotene is 0.6 µg. Retinol Equivalents (REs) were devised to normalize by biological activity since carotenoids have less of an effect than retinol for the same mass. Thus, 1 RE is 1 µg Retinol, 6 µg beta-carotene or 12 µg other carotenoids. However, more recent evidence has led some nutritionists to believe that beta-carotene is even less effective than previously thought (NAS, 2001). Thus, the NAS report (2001) introduces the idea of Retinol Activity Equivalents (RAEs) to distinguish from the earlier REs. So, 1 RAE is 1 µg Retinol, 12 µg beta-carotene, or 24 µg of other carotenoids. These relationships are summarized in Appendix Table 2. One implication of the recent thinking regarding RAEs is that plant based diets may be even more deficient than previously thought, and hence more people who consume such diets may be at risk for VAD (UN ACC/SCN, 2000).
There are different estimates of VAD prevalence. Drawing on a variety of sources, the Fourth Report on the World Nutrition Situation reports varying estimates of 2.8 or 3.3 million preschool children with clinical VAD, and an additional 251 or 75-140 million are subclinical (sources are WHO/UNICEF and the Micronutrient Initiative/UNICEF/Tulane, respectively). It is noted that, "[d]espite the discrepancies in these estimates, which ultimately reflect the paucity of real data, it is clear that vitamin A deficiency remains a major public health problem of immense proportions." (UN ACC/SCN, 2000, p. 30).
A lack of diversity in diets of the poor is the determining factor in VAD. In particular, poor people whose diets are based on rice are at greater risk for VAD, since rice contains no beta-carotene. Table 1 shows data on the importance of rice in Asian diets and the estimated per capita intake of vA from the mid-1980s. Most of the less developed Asian countries have estimated intakes of less than 120 REs/day putting them well under the EAR or the RDA of 400 plus RAE’s, depending on age and gender. In these poor Asian countries rice accounts for a large percentage of calories and of food expenditures. More recent data confirm that in Bangladesh, Myanmar, and Vietnam, rice continues to account for at least two-thirds of caloric intake for the average citizen (FAOSTAT, 2001) with the share being even higher for many of the poor. Rice accounts for half of all expenditures for the poorest 10% of the urban population in Bangladesh. Even for the poorest 60% of the urban population, nearly 40% of all expenditures go to rice (Bangladesh Bureau of Statistics, 1998). Dark green leafy vegetables are the most affordable dietary source of vA, but the bioavailability of the beta-carotene lags that of animal products, which are much more expensive and thus less accessible for the poor (Helen Keller International, 1999). Asian countries with low income and rice-based diets are also those classified as having a significant public health problem with xerophthalmia in a study by West and Somer (1987), which includes India, Bangladesh, Nepal, Indonesia, the Philippines, Vietnam, Cambodia, and Laos.
In this paper, we use data from the island of Cebu in the Philippines, and thus a summary of the VAD situation in the Philippines is useful for context. In the Philippines, there are still large numbers of at-risk groups with deficient or low levels of plasma VA. The Fifth National Nutrition Survey in the Philippines (National Nutrition Council, 2001) conducted in 1998 found that VAD (based on blood assays) affected 38% of children ages 1-5 years, 22% of pregnant women, and 17% of lactating women. These fractions are a bit higher than those of the Fourth National Nutrition Survey of 1993 when they were 35%, 16%, and 16% (Kuizon et al, 1995). The 1993 survey also found night blindness in 0.4% of children under 6, 1.3% among 7-14 year olds, and 1.1% over the whole population. The reason for the increase in VAD among children and women is not clear. It is possible that the 1998 figures are abnormally high due to the effects of the currency crisis that began in the second half 1997 and the impact of the severe El Nińo that hit the region at around the same time.
In Indonesia, Helen Keller International (1999) found high rates of nightblindness (from 1 to 3.5%) among non-pregnant mothers in nearly all areas where their survey was conducted. These rates were much higher than what was found among children aged 12-23 months. The same study cites similar rates of nightblindness in Bangladesh of 1.7% among non-pregnant non-lactating women and 2.4% among lactating women. All of these data seem to indicate that VAD is still a problem in Asia, in spite of significant intervention efforts, which we discuss next.Intervention Programs
VAD is a result of both inadequate incomes and market failures in health. Households with low incomes eat less diverse diets and obtain less vitamin A. Rising income does not always result in increased Vitamin A intake, at least at low levels of development, as estimates of income elasticities for Vitamin A in several developing countries are quite low (Behrman, 1995). Households may not understand the importance of Vitamin A, either due to lack of knowledge or due to the lack of immediate connection between intake and improvements in health (Rogers, 1995). Even when fully informed, households may not choose Vitamin A rich foods due to dietary preferences, relative prices, or the costs of household production associated with such foods. Furthermore, households may not fully internalize all of the benefits of improved child and maternal health, justifying interventions to fully realize this public good. The health benefits of Vitamin A are clearly substantial, but are difficult to quantify in terms of statistical value of life or other economic measures. Thus, interventions to address VAD are usually assessed on a cost-effectiveness basis - i.e., which intervention achieves public health goals at least cost (World Bank, 1993).
Because VAD is recognized as an important public health problem, a number of interventions have been used to combat it. The major approaches include supplementation through administration of mega-doses of VA, fortification of widely consumed foods, nutrition education programs, and promotion of home gardens. We focus on the first two interventions as those with widest impact, and provide information about these programs in the Philippines as context for our case study.
Nutrition education and the promotion of home gardens are recognized as long-term solutions: diverse diets being the goal. Few would disagree that such programs are important, possibly necessary, regardless of what other alternatives are pursued. Unfortunately, such interventions are costly, labor intensive, work slowly, and do not necessarily have a direct effect, although there are often desirable side-effects, such as increasing household income (UN ACC/SCN, 2000). However, it seems unlikely these programs are adequate by themselves and that others interventions are necessary to address VAD in both the short and the long run.
Supplementation with VA capsules is now widely practiced in many countries. Once the retinol in the capsules is stored in the liver, it is excreted slowly and can sustain VA requirements for 4-6 months (dela Cuadra, 2001). Thus, semi-annual distribution is feasible and is the preferred method because it reduces the labor costs associated with distribution. USAID has supported supplementation programs in Nepal, Indonesia, and the Philippines (USAID).
From 1993 to 1996, the Philippines was able to deliver VA supplements to a large percentage (88-93%) of the target group, children between the ages of 12-59 months. However, coverage declined to 78% in 1997 and funding for the program is currently in danger. Furthermore, despite the high rates of coverage in the five years from 1993 to 1997, the National Nutrition Survey of 1998 mentioned above still found widespread problems with VAD. Thus, while numerous studies have found supplementation to be effective in reducing VAD, it does not appear to be a "magic bullet" that can eliminate all VAD.
Coverage of supplementation has been less in other countries such as Indonesia. A survey by Helen Keller International (1999) covering more than 30,000 households nationwide found that VA capsules were reaching just 40% of 12-59 month old children in Jakarta slum areas, and 56% in Ujung Pandang. Coverage was higher in Surabaya and Semarang, reaching 65-80%. In rural areas, coverage was 65-75% in various provinces of Java, but just 40% on Lombok in poorer Eastern Indonesia.
Furthermore, in both of these countries, supplementation does not reach sizable numbers of individuals at risk outside of the primary target group, children aged 12-59 months. For example, while pregnant and lactating mothers are widely recognized to be at risk of VAD, only about 20% of Indonesian women within 30 days of delivery were receiving supplementation from 1996 to 1999 (Helen Keller International, 1999).
In the Philippines, supplementation was found by Capistrano et al (1998) to be a higher cost alternative than fortification. They estimated the cost of achieving one person year of adequate VA intake through supplementation at US$10-15 (based on the exchange rate at that time of about 40 pesos to the US$). Most of these costs were accounted for by personnel costs (rural health workers who administer the supplements), which were about 80% of total costs. The cost of the capsules themselves was less important. However, the same authors concluded that supplementation reached more individuals than fortification of wheat flour.
Thus, although supplementation may be inexpensive in an absolute sense, these programs appear to have some problems sustaining government financial support. Even when the programs operate at full strength, they still do not reach all people at risk of VAD. The evidence is clear that supplementation programs are effective, but it is equally clear that more options are needed for eliminating VAD.
Fortification is another strategy for reducing VAD that is widely practiced. Some of the general criteria for a good fortification vehicle are widespread consumption, no downside risks to certain segments of the population, and a food regulatory system that can enforce adherence to requirements for processing and labeling. If production of the commodity is relatively centralized, this substantially reduces the burden on the regulatory authorities. In Asia, VA fortification is most common in margarine (India, Indonesia, Philippines, Malaysia) and in condensed and filled milk (Malaysia, Thailand, Philippines). Some wheat flour and cooking oil in the Philippines is also fortified. In other parts of the world, margarine is also an important vehicle for VA fortification, but other foods are used as well, the most common being maize meal in parts of Africa and sugar in Central America (Mora, et al, 2000).
In the Philippines, VA fortification began in 1992 with margarine. Studies showed that storage and thermal stability were good, and that daily consumption of fortified margarine reduced the prevalence of low serum retinol among a sample of three to six year rural children by 60% (Solon et al, 2000). Studies on the fortification of wheat flour in the Philippines began more recently, and have shown that fortification was effective at sharply reducing the percentage of children with inadequate stores of VA in the liver. As of yet, however, most millers do not participate. But recent legislation (Republic Act 8976) will make fortification of wheat flour, sugar, and cooking oil with VA mandatory in November 2004.
While many fortification programs are successful, others have failed for various reasons. One example is the fortification of mono-sodium glutamate (MSG), a common food additive in Asian cooking. In both Indonesia and the Philippines, production of MSG is highly concentrated, with one firm in the Philippines controlling 90% of the market. It is also widely consumed by the population. A two-year trial done in the Philippines showed that MSG fortification led to increases in mean serum retinol and a reduction in xerophthalmia in the sample children. But fortification was never pursued beyond this stage because fortification imparted a yellowish tinge to the MSG crystals, and the manufacturer was afraid this would damage the products’ image of purity (Solon et al, 2000). Fortification of sugar in the Philippines also failed, in this case because of financial problems experienced by the sugar mills in the wake of the Asian currency and financial crisis in the late 1990s (Solon et al, 2000). This experience is in contrast to the successful experience with the VA fortification of sugar in Guatemala (Mora et al 2000).
Fortification has wide reach, but, compared to supplementation, may result in less coverage of specified target groups. Based on analysis of food consumption data, Capistrano et al (1998) estimated that fortification of wheat flour with 490 retinol equivalents (REs) per 100 grams flour would still leave 29% of all children 12-59 months old with inadequate VA intake. Not surprisingly, coverage of the poor was low, especially in rural areas, where consumption of wheat products is lower. This coverage is less than the coverage achieved by the supplementation program in the Philippines between 1993 to 1997.
On the other hand, Capistrano et al (1998) also found that fortification of wheat flour was more cost-effective in terms of cost per person-year of adequate VA intake achieved. They estimated these costs at about US$3-5 (again based on the exchange rate at that time of about 40 pesos to the US$). The vast majority of these costs are the costs of the fortificant. Thus, fortification is also an effective vehicle for reducing VAD, and, at least in the Philippines, seems to be substantially lower cost than supplementation. This comes with a disadvantage of less coverage, however, because wheat flour is primarily consumed by the non-poor.
When thinking about fortification in Asia, rice seems like an obvious candidate. However, it is not technically feasible to fortify rice with Vitamin A (Dexter, 1998). Thus, GR was developed to meet this need with genetic modification to accomplish the fortification with Vitamin A. GR, once adapted to farmers fields, would be able to deliver VA (through beta-carotene) in a preferred staple in a way that does not necessarily require the institutional infrastructure of supplementation and fortification. We turn now to examination of its potential role in the diet of a population at high risk for VAD.Case Study of Potential GR Impacts in Cebu, the Philippines
Cebu is a relatively poor region of the Philippines and has historically been found to have high incidences of VAD (Solon et al 1978). During the National Nutrition Survey of 1998 Cebu had a VAD prevalence of 28.6% among pregnant women and 9.4% among lactating women, indicating that VAD is still a problem.
A study in the 1970s in Cebu documented the incidence of VAD and estimated some of the determinants of VA intake (Solon et al (1978). This study found a high degree of VAD (47%), especially among older children. Their findings regarding determinants of VA intake suggest some of the underlying economics of the household that influence VA intake. These include: increased food expenditures raised vA intake in urban areas but not in rural areas (the impact of animal products in urban areas), home gardens had a significant impact raising vA intake only in hinterland areas (presumably isolated from markets and more likely to consume production), and vA intake was decreased when the mother worked outside the home (less time to prepare traditional dishes using DGLV). These findings point to the potential importance of home garden production (lower cost of DGLV), access to markets (influences relative prices), animal product consumption (reflects income level), and opportunity costs of mother’s time in explaining VA intake. In our analysis, we examine the impact of home gardens and animal product consumption in data from 1994; in future work we plan to examine other economic determinants of VA intake.Data and Methods
The data used are from the 1994 Cebu Longitudinal Health and Nutrition Survey. This survey identified and is following about 2000 families who had a child between May 1983 and April 1984. The mothers were selected while still pregnant. In addition to being interviewed at that time, the families were surveyed immediately after the birth of the child and every two months for the first two years. Follow-up surveys were conducted in 1991-92, 1994-95, and 1998-99 expanding the range of information collected. We use the 1994 data for this paper as it includes detailed food intake data for 1,839 children aged 10 to 12 years. The consumption data record how much of which foods were eaten by the child throughout the course of one full day.
These data were selected for our analysis because they provide recent, accurate food intake information for a population known to be at risk for VAD. In addition, the survey includes various household data that relate to consumption decisions and the determinants of VA intake. These data are publicly available through the University of North Carolina (see web site at http://www.cpc.unc.edu/cebu/ ). An important drawback to use of these data is that Cebu is one of the places in Asia where rice is not the only staple. White flint corn (mainly in the form of grits) is of almost equal importance in Cebu (very little rice is grown and is thus imported from other Philippine regions), which will limit the potential impact of GR.
The consumption of rice, corn, vegetables, and animal products are determined (as averages) for the original children in the survey who range from 10 to 12 years of age in 1994. These measurements are further broken down by approximate asset quintile as measured by total reported value of houses and lands owned by the household. Total household assets provide a measure of permanent income and are more easily obtained from these data than annual income.
We use the consumption patterns of rice and white corn grits to estimate the effects of substituting GR and yellow corn grits. Beta-carotene provided by GR is that reported by Ye et al (2000); yellow corn grits VA content is from the USDA food composition tables. For GR, we use two possible values, one reflecting the actual beta-carotene in the variety that now exists and another higher value reflecting the potentially achievable level in the future (Ye et al, 2000). These increases in VA from hypothetical substitutions are considered as fraction of the EAR and RDA, as well as compared to an estimate of current VA intake and potential VA intake from wheat fortification.Results
Food consumption relates to asset level as expected: rice increases, corn decreases, vegetables decrease slightly, fish intake remains roughly constant, and animal products dramatically increase (Table 2). These preferences are typical of rural Southeast Asia where maize and rice are both staples (e.g. see Foster and Leathers, ch.8, 1999 for similar data from East Java).
A unique aspect of the Cebu survey is that it allows examination of the influence of home gardens. About two-thirds of the households surveyed had home gardens. Children in these households consume less rice, more corn, more vegetables, and fewer animal products at all asset levels (Appendix Table 3). This may be due to proximity to markets, which would reduce the relative prices of preferred foods. Isolation from markets would increase incentives for home garden production. However, the difference in vegetable consumption is very small for those with home gardens, and thus the presence of the home garden has less influence than expected on VA intake. Because the differences are small, we do not consider home garden households separately in the GR substitution analysis.
The amounts of cereals consumed determine the effects of substitution with beta-carotene containing alternatives. Estimates are made for two different substitution scenarios and the effects due to rice and corn are recorded along with a total possible effect in Table 3. The high substitution scenario assumes substitution of all rice and corn and an optimistic view of the beta-carotene content of GR (2.0 µg beta-carotene/g uncooked milled rice). In a low substitution scenario, only 1/3 of rice and ˝ of the corn grits are substituted for beta-carotene containing alternatives. In this scenario, GR is assumed to deliver the beta-carotene in the best existing strains (1.6 µg/g). For comparison, the beta-carotene content of the existing GR is 19.56 RE per cup of cooked rice, and yellow cooked grits have 14.5 RE per cup. Switching to GR delivers more beta-carotene per unit consumed, but the relative impact of the two hypothetical substitutions will depend on the relative importance of rice and maize in these children’s diets.
For the top and bottom quintiles, in the low substitution scenario the increases are 20 and 9 RE, respectively (Table 3). This corresponds to only 4.6 and 2.1% of the EAR, or 3.4 and 1.5% of the RDA. This amount is a very modest, and possibly insignificant, increase in intake. In the high substitution scenario, the VA intake increases from GR for the top and bottom quintiles are 76 and 34 RE, respectively, which are 17.2 and 7.8 percent of the EAR. This complete substitution results in more dramatic increases in intake. Because rice is less important in the diets of bottom quintile, GR has less impact on VA intake for this group. For the low substitution scenario, yellow corn actually would contribute more than GR for the bottom quintile. So the impact of GR is clearly dependent on the current role of rice in the diets of these Cebuano children.
In order to gauge the relative importance of GR, let us consider their current likely vA intake. The primary sources of vitamin A and beta-carotene are vegetables and animal products. Solon et al (1978) found that 81% of the vA intake was from vegetables and fruits, 9% was derived from animal products and another 9% from cereals, for a total VA intake ranging from 51 RE in the worst region to 156 RE in the best. Children in the 1994 CLHNS consumed about 0.2 cups of vegetables per day (about 15g) and 60-135g of animal products per day depending on asset level. Using a blanket estimate of 23 µg beta-carotene/g vegetables, this indicates that roughly 60 REs are derived from the vegetables. Similarly, using a blanket estimate of 0.5 µg retinol/g animal products, 30-70 REs are consumed from animal sources. Thus, animal product consumption and VA intake from animal products has increased since the 1970s, as would be expected.
Our crude estimate of REs from both vegetables and animal products gives a total of 90-160 REs/day on average which is much lower than the EAR of about 430 RAEs and the RDA of 600 RAEs for this age group. The bottom and top quintile estimated intakes are 103.6 and 113 RE, respectively, with the top quintile obtaining relatively more from animal products. The low level of intake relative to requirements for all rural households makes the potential additional intake from GR important.
As wheat flour fortification will become mandatory in the Philippines, we examined its likely impact in these Cebuano children’s diets. The standard for fortification will be 490 RE/ 100 g of flour. This will provide 30% of the RDA in two 30 g pieces of pan de sal (a wheat roll commonly consumed the Philippines). Children in the top and bottom quintiles consume 84 and 53 g of wheat products, respectively, and thus would gain 190 and 120 RE’s from fortification. Thus, the widespread consumption of wheat products across income classes will result in fairly good coverage and impact from fortification of wheat flour. Because wheat flour can be fortified with very high levels of VA, these expected increases in intake are larger than those estimated for GR, which has more modest levels of VA, and is not the sole staple in this region.Implications for the Hypothetical Benefits from Golden Rice
The estimates presented in this paper indicate that the potential impact of GR on VA intake will vary with dietary patterns. It would clearly be more important where rice is the main staple and if its beta-carotene content were enhanced. In this particular region, its impact is modest due to the fact that rice is not the main staple, and is less important in the diets of the poor. Only full substitution results in significant contributions to VA intake. When compared to the current intake of vitamin A and beta-carotene, it is clear that it could provide a significant relative boost to total intake. And, GR would probably have a larger effect than a change from white to yellow grits. Given the amount of wheat consumed by this population, wheat fortification would actually have a larger impact on intake. But current intake is only about a quarter of the EAR, and therefore both wheat fortification and GR would be needed to move these diets close to the EAR.
The Cebu region may provide the low end estimate of potential GR impact, due to the fact that rice is not the main staple. By cautious extension, GR would probably be more significant elsewhere in Asia where rice is much more important, and VA intakes are even lower. We do not have comparable intake data from another country, but some comparisons are possible. For example, per capita rice consumption for the poor in Bangladesh is double that of the poor in the Philippines, on average (Barker and Herdt, 1985, p.167). Thus, GR substitution could provide much more VA in Bangladesh than it would appear to in Cebu.
Another insight from this analysis is the importance of both wheat and animal products in assessing VA intake and potential benefits from fortification. Animal product consumption is rising rapidly throughout Asia (Delgado et al 1999), and this will influence VA intake and the incidence of VAD. This is particularly true in light of the revised understanding of how much VA is delivered from plant sources. Wheat consumption is also rising throughout Asia, and this will make fortification more effective. Past assessments of VAD may not have adequately accounted for rapid changes in consumption patterns, and the likely resulting changes in regional or subpopulation incidence of VAD.
As previously stated, if accepted, Golden Rice has several desirable features to recommend it as a vitamin A intervention strategy because of its unique position as the overwhelmingly dominant staple in Asia. It could be disseminated through the existing channels for modern varieties. It is conceivable that it could deliver beta-carotene without the institutional, industrial, and logistical infrastructure required for supplementation and fortification. If culturally accepted and agronomically sound, it has the potential to provide very widespread coverage. Probably the most attractive quality is that it will be quite sustainable with minimal effort if the previous conditions are met. In terms of costs, it can probably compete well with supplementation and fortification because after the sunk costs of development, no more would need to be spent than is spent on other rice research anyway. Even a very modest effect could be more cost effective than other interventions. In the Philippines, the national supplementation program in 1996 cost 1.4 billion pesos ($35 million with the 40 peso exchange rate) while the national fortification projections range from 150-270 million pesos ($3.75-6.75 million). By comparison, the International Rice Research Institute’s annual budget for Golden Rice is likely to be no more than $2-3 million per year. While the costs of the other interventions would increase dramatically if scaled up to include, say, all of Asia, Golden Rice would require only modest additional investments to achieve greater geographic coverage.
Although GR compares favorably with other interventions in terms of costs and coverage, there are still many other unresolved issues regarding Golden Rice. First, the amount of VA delivered in the grain needs to be larger in order for GR to have impact. Second, consumer acceptance is clearly an important issue. It is unknown whether GR will taste differently from conventional white rice. There are other specialty rices consumed in small amounts in different Asian countries that have colored grains, such as red, black, and purple, but the prospects for widespread acceptance of this very different looking commodity are uncertain. Third, it is clear that the release of GR would require extensive nutrition education efforts, which we have not included in the costs of the GR strategy as discussed above. However, perhaps its current and future publicity will serve to heighten awareness of VAD, which might have spillover benefits for support and participation in other interventions. Fourth, there are still questions of bioavailability, stability of beta-carotene during rice storage, and thermal stability after cooking. Fifth, there are potential food safety concerns to be resolved. The daffodil genes used to create beta-carotene in the rice grain might create allergenic materials in GR, and their safety for human consumption will need to be evaluated (Toenniessen, 2001). Finally, there is the issue of GR cost, which will depend in part on whether it can be produced in much the same way that current MVs are produced. Can the current GR japonica variety be adapted deliver similar or better yields than currently available indica MVs? In addition to production costs, it is likely that GR will require separate channels for production, processing, storage, and marketing, which may add to its cost and consumer price. Thus, there remain a large number of questions that to be answered before GR can be a viable intervention.
It appears that GR has the potential to be a low-cost, wide-coverage intervention. While it can deliver substantial amounts of VA under certain scenarios, it is unlikely to meet all requirements and thus would be an ineffective stand-alone strategy. GR is best viewed as a possible addition to the menu of options for combating this public health problem and a possible complement to existing interventions. In order to make it most complementary to existing interventions, it may be useful to examine where and for whom it is likely to provide greatest benefit. Such analysis would help to guide research to adapt GR to local growing conditions and tastes for particular regions and subpopulations, and would maximize the potential benefit from this biotechnology breakthrough.
Bangladesh Bureau of Statistics, 1998.
Behrman, J. R. 1995. Household Behavior and Micronutrients: What we know and what we don’t know. Agricultural Strategies for Micronutrients, Working Paper 2, IFPRI, Washington DC.
Capistrano M., D. Dado, H. Maglalang, M.V. Magpantay, N.R. Juban. Cost-Effectiveness Analysis of Vitamin A Intervention Programs in the Philippines. Mimeo. Manila, Philippines.
Datta, S., and H. E. Bouis. 2000. Application of biotechnology to improving nutritional quality of rice. Food and Nutrition Bulletin, 21(4): 451-456.
De la Cuadra, A.C. The Philippine micronutrient supplementation programme. Food and Nutrition Bulletin 21:4, 512-514.
Delgado et al 1999. The Livestock Revolution. IFPRI.
Dexter, P.B. 1998. Rice Fortification for Developing Countries, USAID Opportunities for Micronutrient Interventions Program, http://www.mostproject.org/rice4.PDF .
FAOSTAT database, at http://apps.fao.org/, accessed May 2001.
FAO and IRRI. 1993. Rice in Human Nutrition.
FAO and ILSI. 1997. Preventing Micronutrient Malnutrition - A Guide to Food Based Approaches. Available on line at http://www.fao.org/docrep/X5244E/X5244E00.htm.
Fiedler JL, Dado DR, Maglalang H, Juban N, Capistrano M, Magpantay MV. 2000. Cost analysis as a vitamin A program design and evaluation tool: a case study of the Philippines. Soc Sci Med 51(July):223-42.
Foster, P and H.D. Leathers. 1999. The World Food Problem. Lynne Rienner Publishers.
Helen Keller International. 1999. "Increasing coverage of high-dose vitamin A capsules to prevent crisis-induced re-emergence of vitamin A deficiency." Indonesia Crisis Bulletin, Year 1, Issue 8. December.
Kuizon, M.D., L.A. Perlas, J.R. Madriaga, R.L.Cheong, J.A. Desnacido, J.M. Marcos, R.T. Fuertes, D.H. Valdez. 1995. Biochemical Nutrition Survey, Fourth National Nutrition Survey of the Philippines, 1993. Food and Nutrition Research Institute, Department of Science and Technology, Metro Manila.
MASIPAG. 2001. Grains of Delusion: Golden Rice seen from the ground.
Mora, J.O, O. Dary, D. Chincilla, and G. Arroyave. 2000. Vitamin A Sugar Fortification in Central America, USAID, http://www.mostproject.org/newsugarcover.pdf
National Nutrition Council. 2001. Fifth National Nutrition Survey of the Philippines, 1998.
National Academy of Sciences. 2001. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc. <http://books.nap.edu/books/0309072794/html/86.html>
Solon, F.S., B.M. Popkin, T.L. Fernandez, M.C. Latham. 1978. Vitamin A deficiency in the Philippines: a study of xerophthalmia in Cebu. American Journal of Clinical Nutrition. 31(February): 360-368.
Solon, F. L.E. Sanchez-Fermin, L.S. Wambangco. 2000. Strengths and weaknesses of the food fortification programme for the elimination of vitamin A deficiency in the Philippines. Food and Nutrition Bulletin 21:2, 239-245.
Sommer, A. 1997. Vitamin A Prophylaxis. Archives of Disease in Childhood. 77(September): 191-194.
Toenniessen, G.H. 2001. Vitamin A Deficiency and Golden Rice: The Role of the Rockefeller Foundation. Available on the web at http://www.rockfound.org/
UIUC Department of Food Science & Human Nutrition. Nutrition Analysis Tool, Version 2.0. http://www.nat.uiuc.edu
United Nations ACC/SCN. 2000. Fourth Report on The World Nutrition Situation,
USAID, Asia/ Near East Sector Overviews: Population, Health, and Nutrition, http://www.usaid.gov/regions/ane/newpages/sectors/phnp2.htm
US Department of Agriculture http://www.nal.usda.gov/fnic/cgi‑bin/nut_search.pl
West, K.P. Jr. and A. Somer. 1987. Delivery of Oral Doses of Vitamin A to prevent Vitamin A deficiency and nutritional blindness: A State of the art review. Nutrition Policy Discussion Paper no. 2, United Nations Administrative Committee on Coordination-Subcommittee on Nutrition.
World Bank, 1993. Investing in Health: World Development Report 1993. Oxford University Press.
Ye, X., S. Al-Babili, A. Klöti, J. Zhang, P. Lucca, P. Beyer, and I. Potrykus. 2000. Engineering the Provitamin A (beta-Carotene) Biosynthetic Pathway into (Carotenoid-Free) Rice Endosperm. Science Jan 14: 303-305.
Table 1: Rice Consumption, Energy availability, and Vit A intake in Selected Asian Countries
Source: FAO and IRRI (1993). Unfortunately these data refer to the mid 1980s; we could not find later data for Vit A intake on a country basis.
Table 2: Index Children’s Consumption of Foods by Household Asset Level
Source: 1994 CLHNS survey; rice, maize and vegetables are in cups; wheat, fish, and animal products are in grams.
Table 3: Impact of Hypothetical Staple Substitutions on VA Intake for Top and Bottom Quintiles of Household Asset Distribution
Notes: High substitution replaces all rice with GR and all maize with yellow maize. GR assumed to have future potential level of beta-carotene. Low substitution replaces 1/3 of rice with GR and ˝ of maize with yellow maize. GR is assumed to have beta-carotene of current best strain. EAR and RDA assumed are midpoint of those for girls and boys 9-13 years in Appendix Table 1.
Appendix Table 1. Vitamin A Requirements
Source: National Academy of Sciences, Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc (2001).Appendix Table 2. Vitamin A Unit Conversions
Appendix Table 3: Index Children’s Consumption of Different Foods by Household Asset Level and Whether Household Has a Home Garden
Source: 1994 CLHNS survey; rice, maize and vegetables are in cups; wheat and animal products are in grams.
 Authors are graduate student and professor, Department of Agricultural and Consumer Economics, University of Illinois at Urbana-Champaign, USA; and economist, International Rice Research Institute, Los Banos, Philippines.
 The World Bank estimated that Vitamin A deficiency accounted for 11.8 million disability adjusted life years lost in the world in 1990; with an additional 39.1 million DALYs lost from associated diseases. This is roughly one-quarter of the total global burden of disease from malnutrition.
 See FAO and ILSI (2000) for a review of food based approaches, including home gardens, and a discussion of their complementarities and the need for multiple approaches to achieve the long run goal of eliminating micronutrient deficiencies.
 One technology that has been tested in Indonesia and Brazil fortifies rice flour with Vit A and then extrudes the flour into rice kernel shapes, which are blended with regular rice (Dexter, 1998). This would seem to be an expensive and cumbersome method to implement on a nation-wide basis.
 This series of surveys is carried out jointly by the Carolina Population Center of the University of North Carolina and the Nutrition Center of the Office of Population Studies at the University of San Carlos, the Philippines.
 From an initial census of 28,000 households in the region, a sample of 2600 families was selected and surveyed. Of these, 2300 households remained in the survey in 1994. Data are collected for the "index child" included in the first survey, and also for younger siblings born since 1984.
 The reported amounts are as consumed and after cooking, so do not compare to raw commodities in production data. Also, not everyone is eating this much each day. Consider vegetables, where only 1/3 of the children ate any on the interview day, averaging about 0.5 cups. However, over a longer period of time, it is assumed that the portion rotates through the population resulting in an effect not unlike averaging over all children regardless of whether they ate on that particular day, that is, about 0.2 cups.
 As economists, we are unsure how to compare the RE estimates from Ye et al (2000) with the new RAE measure proposed by NAS (2001). It is possible that GR delivers only half as much Vit A, if its beta-carotene content is converted to RAE’s. However, that would mean that the poor in Cebu are currently consuming much less VA since vegetables are their primary source. In this case, GR’s relative contribution could be more important.
 There does not seem to be much difference in VA intake across income levels. However, if the vegetable sources are reduced according the new RAE measure proposed by the NAS, then the poor would consuming much less due to their greater dependence on vegetable sources.
 To match the amount delivered by wheat in these diets, then GR would need to have around 7 µg/g , in contrast to the current 1.6 µg/g. That estimate is dependent on the amounts of wheat and rice consumed.
 Of course, this ignores the cost of the original Golden Rice development in Europe. However, this can be treated as sunk cost because much of what was accomplished has independent basic science merit and will have spillover effects in other aspects of biotechnology research.