Vitamin A Deficiency

Introduction

Global progress is being made in combating ocular consequences of vitamin A deficiency (VAD), but efforts need to be intensified if the mortality-risk consequences associated with subclinical deficiency are to be fully controlled. Because VAD is caused by habitually low intake of vitamin A in relation to requirement, sustainable solutions lie in improving intake and controlling preventable factors that elevate requirement. The underlying cause of inadequate consumption and high need is related to poverty. Economic, social and environmental deprivation limits accessibility to, and consumption of, vitamin A-containing foods. This is particularly the case when physiological needs are greatest, namely during periods of rapid growth and development, i.e., infancy, early childhood and pregnancy, and during lactation when breastmilk carries the vitamin in a highly absorbable form from mother to infant. The major non-physiological factor that increases need is frequency of infections, including diarrhoea and other febrile illnesses such as measles. Strategies for sustained elimination of VAD, therefore, will have to consider both improving vitamin A consumption and limiting the risk of infection.

Indicators

Vitamin A status is best assessed by the total body content of vitamin A, which can be viewed as a continuum from deficiency to excess, with obvious health consequences at either extreme (Figure 5). The extremes are marked by specific indicators, for example in the case of severe depletion, by ocular signs (xerophthalmia, including night blindness) and very low serum retinol levels (<0.35 mmol/L). Physiological functions, however, are impaired before tissues are depleted. It is this stage of subclinical depletion of body stores that defines VAD and is thought to be the beginning of an increased occurrence in the severity of infectious illness and risk of death. Monitoring the prevalence of disease severity and deaths is too non-specific for attribution to a single nutrient, and, unfortunately, currently available biological indicators lack specificity and sensitivity for identifying subclinically depleted body stores. This has necessitated using prevalence values below arbitrary cut-offs specific for different vitamin A indicators. Blood retinol levels are the most commonly measured indicator of vitamin A status in surveys and, using a cut-off of 0.7 mmol/L, a prevalence > 10% has been set as defining a public health problem (WHO, 1996, p. 7).

Consequences

The most obvious health consequences of severe VAD involve the visual system, affecting vision in low light or darkness, and dryness (xerosis) and disruption in the integrity of the surface of the conjunctiva and cornea (Bitot's spot, corneal clouding, ulceration). Occurrence of these signs is associated with elevated risk of blindness and death. Hidden consequences that occur even before eye signs are detectable include changes in surface linings of the gastrointestinal, respiratory, excretory and reproductive systems. In addition, the integrity of the immune system is impaired. Risk for severe disease and death is increased by these hidden changes. Only in the last decade have the mortality-associated risks of sub-clinical VAD been appreciated. A recent study from Nepal found that mortality risk is not limited to children. During pregnancy, even relatively mild night blindness was associated with greatly increased mortality risk (West et al., 1997). VAD also contributes to impaired growth and development, and to inefficient utilization of iron for haemoglobin production (Underwood and Arthur, 1996).

Goals

In 1990, the end of decade goal set for vitamin A was the virtual elimination of VAD and all its consequences, including blindness. Global progress in achieving this goal is tracked in this report by trends both in reduction of prevalence of xerophthalmia and low serum levels of vitamin A. The expectation is that as deficient populations progressively ascend on the continuum toward adequate vitamin A status (Figure 5), clinical signs will virtually disappear and the tower portion of population-based blood vitamin A distribution curves will shift toward adequacy. The risk of health consequences from VAD will remain, however, until serum retinol levels reach adequacy, which for relatively healthy vitamin A-sufficient populations means above 1.05 mmol/L (Flores et al., 1991).

To achieve and maintain adequate serum retinol levels among deprived populations, attention will need to be given not only to an increased dietary intake of vitamin A but also to infectious disease control, including protecting, promoting and supporting breastfeeding, immunizations and parasite control programmes. Acute and chronic infections independent of vitamin A status suppress serum retinol levels (Filteau et al., 1993). Hence, tracking serum retinol levels of vitamin A may in part serve as a proxy for progress in achieving other health-related goals even after clinical eye signs have been eliminated. This is the rationale WHO used in establishing two cut-off prevalences marking ascension toward adequacy on the vitamin A status continuum, e.g., less than 10% prevalence of serum levels under 0.7 mmol/L as a public health problem specific to VAD, but less than 5% for elimination of risk of all consequences of VAD (WHO, 1996). This section of the paper reports current prevalence of VAD, trends in progress toward the goal and progress in implementing control strategies.

Table 12: Trends in Prevalences of Clinical Signs of VAD Calculated from Instances where Multiple Surveys have been Reported

Country	Year	Indicator	Prevalence (%)	Percent Change per 10 years	Trend (percentage points/10 yrs)
India	1976	X1B	1.40
	1979		0.90
	1988-90		0.70	-42.0%	-0.58
Nepal	1979-80	X1B	0.60
	1981		0.64
	1996		0.50*	-10.0%	-0.06
Sri Lanka	1975-76	X1B	1.10
	1987		0.33	-60.0%	-0.64
Indonesia	1977-78	X1B	1.01
	1992		0.30	-48.5%	-0.49
Aceh (Regional)	1977	X1B	2.42
	1982-83		1.23
	1989		1.28	-40.0%	-0.95
Philippines	1982	Total	3.20
	1993		0.50	-78.0%	-2.50
Ethiopia	1980-81	X1B	1.00
	1996		0.50*	-32.3%	-0.32
Niger	1988	XN	2.01
	1992		2.47	+57.2%	+1.15
Bhutan	1976	Total	1.30
	1989		0.70	-35.6%	-0.46

Source: WHO (1995) and UNICEF et al. (1997b, p.20)

* Estimated X1B, actual data was XN only. See Annex 5 for explanation of X1B and XN.

Note: In cases where there are three different surveys, the difference between the earliest and the latest years is taken.

Magnitude

Clinical VAD Early in the 1980s, xerophthalmia was estimated to afflict 4-8 million pre-school age children and to cause half million cases of childhood blindness, two-thirds of whom died (Sommer et al., 1981). A large amount of data has been compiled since these estimates were made (WHO, 1995), and additional information obtained from UNICEF field offices and country reports at recent meetings (IVACG, 1997). In the last decade, the degree of increased activity related to VAD is reflected by the fact that 72 countries have conducted nationally representative surveys, 32 of which included assessment of ocular signs and symptoms and 40 of which included serum retinol. An additional 16 countries have surveys planned² (UNICEF et al., 1997b). Xerophthalmia rates in most surveys were based on night blindness (XN) and Bitot's spot (X1B) as the two most commonly reported clinical signs, and prevalence of serum retinol levels under 0.7 mmol/L as the most common expression of subclinical VAD. To date, only a few of these surveys were re-assessment surveys that could be used to document trends, particularly following implementation of control programmes.

² The survey in Eritrea was reported at the IVACG XVIII in Cairo in September 1997.

Table 13: Prevalence Estimates of Clinical VAD for Children aged 0-60 months, 1985-1995

Region	Percent Clinical Prevalence		Numbers of Children Affected (millions)		Percent change per 10 years	Rate of Prevalence Change, 1985-95, percentage points/10 yrs
	1985	1995	1985	1995
South Asia	1.79	0.95	2.67	1.58	-47%	-0.84
East Asia/Pacific	0.43	0.25	0.66	0.40	-42%	-0.18
Latin America & the Caribbean	0.35	0.24	0.17	0.12	-31%	-0.11
East/South Africa	1.80	1.06	0.69	0.53	-41%	-0.74
West/Central Africa	1.40	0.87	0.53	0.45	-38%	-0.53
Middle East/North Africa	0.63	0.27	0.24	0.12	-57%	-0.36
Total	1.06	0.63	5.00	3.30	-41%	-0.43

Source: UNICEF et al., (1997b, p.21)

Multiple ocular survey data are available for the eight countries shown in Table 12. From these limited repeat surveys, country-specific trends are suggested. The limited comparability of the data make further inferences difficult. In India, Indonesia and the Philippines, vitamin A supplement distribution programmes have been in effect for several years, and Sri Lanka has used supplements in a programme targeted to high-risk groups. All of these countries have experienced reduction in clinical VAD, which may be attributable in part to the supplementation programme. In Nepal and Niger, minimal or no improvement occurred during a period when there were no broad based control programmes in effect. Both countries now have implemented programmes and future surveys will indicate if these programmes are associated with an improved situation. The limited data in Table 12 are too few for calculating global trends; for this a model approach was used.

Data from 35 surveys of clinical VAD prevalence, judged to be sufficiently comparable for estimating regional prevalence at different time points, were used to develop a model from which the situation in 1985 could be more precisely interpolated. These 35 surveys are listed in Annex 4. This established a baseline against which to evaluate progress. Annex 5 provides a description of the method used.

Table 13 presents the clinical VAD prevalence and trend estimates by region for 1985 and 1995. The model estimates that five million children were clinically afflicted in 1985 in comparison to 3.3 million in 1995. The model estimate for 1995 is somewhat higher than the 2.8 million estimated by WHO using a different methodology (WHO, 1995, p. 4).

Regional variations in prevalence reductions are evident. Their meaning in terms of achieving elimination goals is clear when calculated in terms of percentage points per 10 years (pp/10 yrs). Global reduction in prevalence was occurring at a rate of 0.43 pp/10 yrs, with region-specific variations from a low of 0.11 in Latin America and the Caribbean to a high of 0.84 in South Asia. The end-of-decade global goal of elimination of the risk of VAD-related blindness will not be achieved at these rates.

Table 14: Year when Elimination of Clinical VAD will Occur at Current Reduction Rates, by region

Region	Percentage points/10 yr	Percent Clinical Prevalence (%)		Expected Years of Elimination of Clinical VAD
		1990 (year of the World Summit for Children)	2000
South Asia	-0.84	1.37	0.53	2006
East Asia/Pacific	-0.18	0.34	0.16	2008
Latin America/Caribbean	-0.11	0.29	0.18	2014
East/South Africa	-0.74	1.43	0.69	2009
West/Central Africa	-0.53	1.34	0.81	2015
Middle East/North Africa	0.36	0.35	-	-
Total	-0.43	0.84	0.41	2009

Table 14 shows the year when the risk of blindness - part of the decade goal - would be achieved, predicted by the model and barring circumstances that could disrupt current progress, e.g., civil unrest and disasters. Only the Middle East and North Africa region would achieve the year 2000 goal. Latin America and the Caribbean, with the lowest prevalence (0.29%) in 1990 would not reach the goal until 2014 and a similar time is needed by Western and Central Africa, where the prevalence in 1990 was four times higher (1.34%). Hence, although trends are encouraging, global reduction rates would have to increase and regional rates will need to be considerably accelerated.

This discussion of prevalence and trends has been based primarily on data available for pre-school-age children. In areas of endemic VAD, there are increasing numbers of reports that prevalences of night blindness in pregnant and breastfeeding women are similar to or exceed those in children (IVACG, 1997a). Due to limited survey information, this vulnerable group is only recently beginning to enter the global estimates of those with VAD. In view of recent reports from Nepal showing a reduction in maternal mortality brought about by weekly supplements of beta-carotene or vitamin A equivalent to 23,000 IU vitamin A (West et al., 1997), priority should be given to assessing populations of VAD pregnant women and including them in programmes.

Subclinical VAD The results of 42 surveys of sub-clinical VAD, assessed biochemically on the basis of low serum retinol in children under five years old, were compiled from WHO (1995) and UNICEF field office reports (UNICEF et al., 1997b). These are listed in Annex 6. Some of these surveys are subnational and extrapolation to a national basis has limitations. The cut-off of serum retinol less than 0.7 mmol/L although not usually associated with ocular signs, carries mortality-related risk. The prevalence of children under five who are subclinically deficient is much higher than for clinical VAD. Reported ranges exceed 70% in Burkina Faso and Mali, and are more than 40% in CD'Ivoire, Honduras and Mauritania. The prevalence rate in Indonesia is about 58%, according to a 1991 survey.

Global trends in sub-clinical prevalence estimates are less easily tracked than clinical deficiency. Confounding disease-related factors that vary in intensity from country to country, making cross-country and regional comparisons difficult to evaluate, influence serum retinol levels. Few data are available from representative repeat serum retinol surveys, particularly at a national level; in fact repeat survey data are currently available from only four countries, all of them in Latin America. These limited data are insufficient to develop a reliable predictive model from which to interpolate global or regional prevalence at different time points. Presumably global progress in reducing clinical VAD is being reflected by progress in reducing subclinical VAD, but it is not possible to substantiate this until sufficient data from repeat surveys become available.

Source: Adapted from UNICEF et al., 1997b, p. 31

Nonetheless, some notion of how far we must progress to eliminate subclinical VAD is gained by noting the prevalence of inadequate serum levels from surveys done this decade. Figure 6 summarizes data since 1990 for 38 countries where surveys were nationally representative or thought to reflect the situation in at least half of the country. Although data for the United Kingdom used a cut-off prevalence below 0.87 mmol/L, these data as well as a value from Hispanic children living in the southwestern USA (Pilch, 1987) are included to provide relative comparisons³. The figure reveals that most Latin American and Caribbean countries, except Dominican Republic, Ecuador, Mexico and Nicaragua, have achieved or are close to achieving a prevalence under 10% which designates a public health problem, while most of the African region, Indonesia and parts of the Western Pacific lag behind.

³ There is virtually no recent data for minority children in Canada.

Status of Programmes to Reduce VAD

Of the 78 countries where VAD is known to be a public health problem, 61 (78%) have policies supporting regular vitamin A supplementation of children (UNICEF et al., 1997b, p. 8). The majority of countries implementing vitamin A supplementation programmes have adopted the policy recommended by WHO/UNICEF/IVACG (1997). This calls for children aged 6-12 months to receive a 100,000 IU dose, and for children aged 12 months and over to receive a 200,000 IU dose of vitamin A once every four to six months. Fifteen countries use alternative supplementation regimens in which a lower dose, usually 50,000 IU, is provided. Coverage has been increased in recent years by providing supplements during National Immunization Days (NIDs) or other mass campaigns such as periodic anti-helminth campaigns. Thirty-four of 78 countries which have used NIDs for delivery increased coverage rates of children under five (UNICEF et al., 1997b, p. 22 and p. 32). While NIDs typically take place on a yearly basis, several countries have organized additional periods, such as Vitamin A week' or Vitamin A month', to achieve extensive biannual distribution.

It should be noted that in many countries NIDs are organized to eradicate polio and may be discontinued after the year 2000. In the long term, it is critical to integrate vitamin A supplementation into ongoing routine child immunization programmes or to develop other more sustainable ways to ensure that children receive supplements regularly until adequate intake from food sources is achieved. In many countries, vitamin A supplementation of young children has been successfully integrated into routine maternal child health visits, or is often linked to community-based nutrition improvement programmes.

Where VAD is a public health problem, WHO/UNICEF/IVACG (1997) recommend that all breastfeeding mothers receive a high dose supplement within eight weeks of delivery to improve her own status and raise the level of vitamin A in her breastmilk. Of 78 countries with recognized VAD, 46 have adopted this policy although coverage is highly variable (UNICEF et al., 1997b, p. 11).

Food-based approaches to increase the quantity and quality of micronutrients in diets, particularly to improve vitamin A intake, include home-based horticulture promotion and fortification of processed foods. Fortification of commonly consumed foods, such as sugar, cereals, flours, margarine and oils, is being pursued by governments in over half of the countries with a recognized VAD problem. In South East Asia, instant noodle fortification is occurring and other vehicles are being explored elsewhere. Although supplements continue to be routinely provided to address the immediate situation found in some of these countries, it is anticipated that a phase-in of fortified foods will sustain the progress achieved and allow supplementation to be phased-out. In other countries, such as Guatemala where xerophthalmia is rare but subclinical deficiency is still prominent, mandatory fortification of sugar has moved the country toward adequate vitamin A status. This has occurred in the absence of supplementation, and it is expected that the progress will continue and be sustained as long as fortification continues.

Lessons learned from the Guatemala fortification experience are valuable guidance for other Central American countries that have fortified sugar or have sugar fortification under preparation - Colombia, Ecuador, El Salvador, and Nicaragua - and for several countries in Asia, Africa and South America where feasibility is being explored. Reliance on a single fortification approach to control micronutrient deficiencies is a fragile strategy, even though it may not require changes in food-habits or knowledge. This was illustrated in Guatemala when the vitamin A fortification programme temporarily stopped in the 1980s and evidence of VAD reappeared. Similarly, in Venezuela when national economic adversity temporarily stopped flour fortification with iron, anaemia prevalence increased (Institute of Medicine, 1997). In the past, identification of a single widely consumed food vehicle to fortify has dominated efforts, but the potential to fortify different products to take advantage of more localized food availability and consumption patterns is increasingly clear. In addition, social marketing of fortified products is advised so as to inform consumers of the benefit of making wise food choices, both from natural food sources and what is likely to become a variety of fortified foods available in the future.

Another lesson learned from fortification experiences is the need for early government-private sector alliances, as well as alliances with other stake-holders (Ending Hidden Hunger, 1991). This is important to create among the private sector a sense of social responsibility for control of micronutrient deficiencies, and, in turn, elicit government recognition of private-sector economic risk concerns. Where possible, accommodating some private sector needs has provided incentives for voluntary private sector involvement. For example, in the Philippines where margarine is fortified, a government endorsement visible on the product as an 'acceptance seal' and for cost-recovery of initial capital investments was found to be an appropriate incentive. Now in the Philippines, the private-sector on its own initiative is considering a variety of food products for fortification. Multiple nutrient fortification is also being explored. In Venezuela, for example, wheat and maize flour are fortified with both iron and vitamin A.

Horticulture approaches are increasingly recognized for their effectiveness and potential sustainability in improving not only vitamin A status, but also micronutrient status generally. Recent evaluation of the large-scale horticultural intervention in Bangladesh has shown effectiveness in improving vitamin A status of those households that have home gardens, including increased consumption by children of vitamin A-rich garden produce (UNICEF, 1997a). Moreover, homestead gardening lends itself to adapting traditional preservation and preparation practices to improve nutrient retention, for example the indirect rather than direct sun solar drying of mangoes in Haiti and West Africa. Since several food sources of provitamin A also are rich sources of vitamin C, and contain iron, increased consumption of these foods addresses multiple nutrient needs. The importance of combining increased vitamin A levels in the food supply with nutrition education and appropriate social marketing that promotes consumption by the vulnerable groups is increasingly recognized (IVACG, 1997).

Socio-economic progress itself in many countries, including direct poverty-reduction programmes, probably accounts for part of the notable improvement in reducing prevalence of VAD. Specific interventions, however, have undoubtedly made an important contribution to VAD control, both through provision of vitamin A itself and through other health promotion and social measures. For example, during the early years of life, promotion of breastfeeding and supplementation of breastfeeding mothers immediately post-partum (breast milk being the infant's primary source of vitamin A), broad immunization coverage and oral rehydration therapy (ORT) use in diarrhoeal disease control are likely to have contributed significantly. Measles immunization may account for another part of the improvement, as higher measles immunization coverage can be shown to be associated with lower clinical VAD prevalence, taking account of economic status. And programmes that enhance the social status of women and empower them have been linked to family nutrition improvements.

Attention is given to the importance for sustainability of embedding VAD control programmes into community programmes. Supplementation programmes for the 6- to 72-month population accelerated around 1994, and using the measles immunization contact for delivery around nine months of age expanded. Thus supplementation and measles immunization may account for part of the global improvement in the last few years, particularly in the reduced prevalence of xerophthalmia.

It is crucial to take advantage of the present opportunity to firm up the evaluation experience from various health promotion programmes and to determine the relative impact of large-scale supplementation programmes. It is noteworthy, for example, that Indonesia, where supplementation has been a national programme since 1974, declared the country xerophthalmia-free at a public health level in 1994. However, Figure 6 shows that subclinical VAD persists in Indonesia at over 50% prevalence. Recently the Indonesian government has expanded its efforts to control VAD through supplementation of mothers and promotion of consumption of foods rich in vitamin A (UNICEF, 1997a), while private sector involvement in providing fortified foods, e.g., instant noodles, is being pursued.

With the continuation of present trends, clinical VAD may be eliminated in many parts of the world on average in the next 15 to 20 years. But many, especially children, will remain affected by sub-clinical deficiency, unless intervention programmes that include sustainable solutions are actively implemented and/or given increased emphasis. This is necessary to underpin the trend toward decreased prevalence of VAD. Therefore, it will be necessary to sustain major efforts to deal with both clinical and sub-clinical VAD for at least the next two decades.