Jeffrey K. Griffiths, MD. The vitamin A paradox --- The Journal of Pediatrics. November 2000 . Volume 137 . Number 5


     Publishing and Reprint Information

      Human immunodeficiency virus
      Rate ratio
      Vitamin A supplementation

See related article, p 660.

Vitamin A supplementation has reduced child mortality worldwide,1 and it is
one of the most important public health advances of the
last century. In many countries, one half of all deaths in children under
the age of 5 years can be linked to malnutrition.2 Acute
respiratory tract infections (chiefly pneumonia) and diarrheal disease
remain the leading causes of pediatric death worldwide,
surpassing even malaria, human immunodeficiency virus, or tuberculosis. The
World Health Organization has determined that
pneumonia alone causes one fifth of all deaths in children under 5 years
and accounts for an astounding 8.2% of all worldwide
disability and premature death, when measured as disability-adjusted life
years or "DALYs."2 In a world in which food security for all
is still a dream, reducing the physiologic and immunologic consequences of
malnutrition is a worthy, cost-effective, and achievable
goal, if not an overwhelming moral mandate.

The immunosuppression of malnutrition appears to be most related to
micronutrient and vitamin deficiencies, rather than simple
caloric deficiency. Restitution of the immune response by the replacement
of these compounds is thus a strategy for combating
childhood mortality. Supplementation effectively, and independently,
complements the other 2 major strategies for improving
childhood health worldwide, vaccination and improved case management.3

Vaccine development is complex, costly, and time-consuming, and the
outcomes can be uncertain (eg, the rotavirus vaccine
experience). It is unlikely that inexpensive, effective vaccines will be
developed in the near future for all of the major respiratory and
diarrheal pathogens in the developing world. Vaccine delivery will also
continue to be a substantial problem. Improved case
management for respiratory and diarrheal diseases can lead to dramatic
decreases in mortality.4 This includes simple strategies
such as administration of antibiotics for pneumonia or dysentery and oral
rehydration therapy for watery diarrhea. However, antibiotic
resistance has already compromised this strategy and it will continue to
worsen. Thus micronutrient and vitamin supplementation
can potentially lead to important public health advances, even when the
other major strategies are not helpful or available. The
efficacy of these substances in reducing mortality is completely
independent of antibiotic resistance or vaccine availability and

VAS greatly reduces measles-specific mortality5 and has been shown to
reduce overall mortality in children 6 months and older in
the great majority of community studies.6 A consistent benefit against
diarrheal disease has usually been noted in these trials. As a
result, public health and child health authorities have recommended that
all children at risk for vitamin A deficiency receive VAS,
regardless of their nutritional or vitamin status. Commonly administered
doses of vitamin A are 100,000 IU for infants and 200,000 IU
for children 1 year of age and older, given every 4 to 6 months. VAS is
often given at the same time as immunizations during mass
vaccination campaigns.

Puzzlingly, though acute lower respiratory tract infection (pneumonia) is
the leading cause of death in children worldwide, large-scale
field trials of VAS have found "no evidence for a differential impact on
pneumonia by age" and that VAS "has no consistent overall
protective or detrimental effect on pneumonia-specific mortality in
children aged between 6 months and 5 years of age" (from the
authoritative 1995 World Health Organization review of pneumonia and VAS7).
How can this be possible? How can an agent that
reduces overall childhood deaths by ~30% in clinical trials have no effect
on the major cause of death, pneumonia?

The answer to this paradox may be an unpleasantly simple one that
eliminates any simple notion that vitamin A can be given
indiscriminately to all children. It appears increasingly possible that VAS
is protective against pneumonia in malnourished children
(who are likely to be vitamin A-deficient) and is paradoxically detrimental
for adequately nourished children. Essentially, there may
be a differential effect of vitamin A on the basis of the child's vitamin A
nutritional status. For example, a decreased rate of
pneumonia in supplemented malnourished children might be accompanied by an
increased rate in well-nourished supplemented
children. Overall deaths might still fall without a decrease in the overall
pneumonia rate, because deaths are concentrated in the
malnourished group. In this scenario the cases of pneumonia are shifted
from the group least likely to survive to the group most
likely to survive.

This simple explanation fits the authoritative global meta-analysis result
cited by the World Health Organization Vitamin A and
Pneumonia Working Group that worldwide, deaths, but not pneumonia rates,
are reduced by VAS.7 Recently, Sempertegui et al,8
my othercolleagues in Ecuador, and I found a large increase in the rate of
pneumonia in well-nourished children who received
low-dose (10,000 IU) weekly VAS when compared with control subjects (rate
ratio = 2.21, P = .005) and a clearly protective effect of
VAS for malnourished children (RR = 0.38, P = .01). We suggested that VAS
may only be appropriate for targeted populations if our
results are confirmed. It is squarely at the center of this evolving
literature that the very important results of Fawzi et al9 in this issue
of The Journal of Pediatrics are located.

Fawzi et al9 report that VAS, given at the time of hospitalization for
pneumonia and then 4 and 8 months after discharge and when
compared with placebo, was associated with an apparent increased risk of
respiratory tract infections (RR = 1.38, P = .005 for
cough and tachypnea) and clinic visits (RR = 1.34, P = .003) over the
following year in Tanzanian children aged 6 to 60 months. This
effect was most prominent in children who were seronegative for HIV (RR =
1.47, P = .001 for cough and tachypnea), whereas there
was an apparent reduction in the risk of cough with tachypnea in
HIV-positive children (RR = 0.54, P = .13). Further extending the
evidence of a differential response to vitamin A supplementation based on
nutritional status, they found that this particular VAS
regimen decreased the risk of acute diarrhea in children with wasting
disease and increased the risk in normally nourished children
and children with stunted growth during the follow-up period after their
initial hospitalization (P = .01 for interaction). Overall, children
who received vitamin A had a significantly lower risk of severe watery
diarrhea (RR = 0.57, P = .04) but not all diarrhea when
compared with placebo.

It is crucial to understand that Fawzi et al9 report important health
outcome differences between subgroups that may be presumed to
have varying degrees of vitamin A deficiency, based on their anthropometric
nutritional states, and HIV status. These significant
differences would not have been seen had they not analyzed their overall
data by nutritional state or HIV status. (They have also
reported that VAS decreases overall mortality in HIV-positive children10
and that VAS does not decrease but may worsen the
severity of pneumonia when given during acute episodes.11) The striking
results from our study in Quito, Ecuador, would have also
been masked by an overall analysis that did not include nutritional status
of the individual child. In the articles by Fawzi et al9 and
Sempertegui et al,8 there are citations and discussions of the other
meritorious studies that have suggested that VAS may be
detrimental in adequately nourished children. The importance of these 2
recent studies is their attention to the issue of respiratory
tract infections and the clarity and consistency of their results.

Nonetheless, these 2 studies have important differences. Although both had
the powerful advantages of being placebo-controlled,
double-blind studies, different treatment regimens were used in different
populations. The Tanzanian study enrolled children aged 6 to
60 months with acute pneumonia in a sea level region with a high (9%)
incidence of HIV and used episodic large doses of vitamin A.
The Ecuadorian study enrolled overtly healthy children aged 6 to 36 months
in an urban slum community at high altitude (~9500 feet)
and gave them weekly low doses of vitamin A. This regimen approximates the
recommended dietary allowance and was chosen
because of the purported lack of efficacy of large-dose, episodic VAS
against respiratory morbidity. The respiratory outcomes for the
2 studies followed similar symptoms and signs but are not identical. Cough,
cough and fever, and cough and tachypnea were
separately evaluated by Fawzi et al,9 whereas in Ecuador our primary
outcome variable was the incidence of acute lower respiratory
tract infection (pneumonia). This was defined as tachypnea and/or lower
respiratory tract secretions detected by auscultation with
one or more of the following: cough, fever, or chest retractions. Diarrhea
was defined as 3 or more liquid or semi-liquid stools in 24
hours or less in the Ecuadorian study and "according to a mother's
perception" in the Tanzanian one. Vitamin A was globally
protective against severe watery diarrheal disease in the Tanzanian study
but only in a subgroup of children aged 18 to 23 months in
Ecuador. Serum retinol levels, which are admittedly only poorly
representative of overall body stores, are available only for the
Ecuadorian children. HIV status was only assessed in Tanzania, because
pediatric HIV infection is still rare in Quito. Despite all of
this, the common finding of differential effects of VAS on subgroups of
children is striking. Both studies showed VAS to be protective
in children with wasting disease but not in healthy children or those with
stunted growth.

The potential explanations for this differential effect of VAS must include
a paradoxical (and as yet uncharacterized) adverse effect of
vitamin A on the immune system of healthy, vitamin A-replete children,
along with a beneficial effect in deficient children. Fortes et
al12 found that VAS reduced CD3+ and CD4+ cells in elderly nursing home
residents, but few if any relevant data from children exist.
This explanation would be a worrisome one for many reasons, including the
fact that VAS and vaccines are often co-administered.
Another potential explanation is that VAS in adequately nourished children
may augment the immune response to inappropriate
levels. As noted by Fawzi et al,9 increased cough or diarrhea may be an
indication of an improved inflammatory response secondary
to the pharmacologic effects of a large dose of vitamin A.

If this differential effect of VAS on pneumonia and diarrhea is real, then
the public health implications are enormous. Even in
countries with substantial malnutrition, there are many children who are
not vitamin A deficient and who might be harmed by VAS.
Targeted but not population-wide supplementation might then prove the
optimal choice, minimizing harm while maximizing benefit.
Mass administration campaigns, which are easiest to conduct when everyone
receives the intervention, would have to be altered to
include an assessment for vitamin A deficiency. This assessment would
require additional time and resources and the use of simple,
sound, and proven surrogate markers for vitamin A deficiency such as a low
weight for age (wasting). Of course, if this nutritional
assessment led to the targeting of VAS and other interventions to the
children most at need, this result would be a very positive one.

Global VAS has led to important decreases in childhood mortality. Its
abandonment or alteration should not be considered lightly.
Nonetheless, a number of studies now suggest that VAS is good overall and
bad for some. If VAS is harmful to some, then we
should not shirk our responsibility to appropriately target this necessary
vitamin to those who will benefit while minimizing harm to
others. The subgroups of children who benefit from VAS and the subgroups
who do not must be securely identified, and the exact
nature and circumstances of the benefits and risks must be delineated. As
it stands, children who are HIV-infected, who have
wasting disease, who are at risk for severe diarrhea or for measles, or who
have overt vitamin A deficiency all appear to benefit from
VAS, whereas HIV-seronegative, adequately nourished children may be harmed
by it. The extent of the harm is unknown. Alternative
dosing regimens may need to be studied if the currently used, large,
supraphysiologic doses of vitamin A are immunomodulating.
Fawzi et al9 have taken us a large step toward understanding the paradox of
vitamin A supplementation and are to be commended
for their work.



     1.  Beaton GH, Martoerll L, L'Abbé KA, Edmonston B, McCabe G, Ross AC,
et al. Effectiveness of vitamin A
     supplementation in control of young child morbidity and mortality in
developing countries. Final Report to CIDA (Canadian
     International Development Agency). Toronto, Ontario, Canada:
International Nutrition Program, University of Toronto; 1992.

     2.  Anonymous. Acute respiratory infections: the forgotten pandemic.
Bull WHO 1998;76:101-3.

     3.  Kirkwood BR, Gove S, Rogers S, Lob-Levyt J, Arthur P, Campbell H.
Potential interventions for the prevention of childhood
     pneumonia in developing countries: a systematic review. Bull WHO

     4.  Sazawal S, Black RE. Meta analysis of intervention trials on
case-management of pneumonia in community settings.
     Lancet 1992;340:528-33.

     5.  Hussey GD, Klein M. A randomized, controlled trial of vitamin A in
children with severe measles. N Engl J Med

     6.  Fawzi WW, Chalmers TC, Herrera MG, Mosteller F. Vitamin A
supplementation and child mortality. JAMA

     7.  Vitamin A and Pneumonia Working Group. Potential interventions of
the prevention of childhood pneumonia in developing
     countries: a meta-analysis of data from field trials to assess the
impact of vitamin A supplementation on pneumonia morbidity
     and mortality. Bull WHO 1995;73:609-19.

     8.  Sempertegui F, Estrella B, Camaniero V, Betancourt V, Izurieta R,
Ortiz W, et al. The beneficial effects of weekly
     low-dose vitamin A supplementation on acute lower respiratory
infections and diarrhea in Ecuadorian children. Pediatrics

     9.  Fawzi WW, Mbise R, Spiegelman D, Fataki M, Hertzmark E, Ndossi G.
Vitamin A supplements and diarrheal and
     respiratory tract infections among children in Dar es Salaam,
Tanzania. J Pediatr 2000;137:660-7.

     10.  Fawzi WW, Mbise RL, Hertzmark E, Fataki M, Herrera MG, Ndossi G,
et al. A randomized trial of vitamin A
     supplements in relation to mortality among HIV-infected and uninfected
children in Tanzania. Pediatr Infect Dis J

     11.  Fawzi WW, Mbise RL, Fataki MR, Herrer MG, Kawau F, Hertzmark E,
et al. Vitamin A supplements and severity of
     pneumonia among children admitted to hospital in Dar es Salaam,
Tanzania. Am J Clin Nutr 1998;68: 187-92.

     12.  Fortes C, Forastiere F, Agabiti N, Fano V, Pacifici R, Virgili F,
et al. The effect of zinc and vitamin A supplementation on
     immune response in an older population. J Am Geriatr Soc 1998;46:19-26.

    Publishing and Reprint Information
    Publishing and Reprint Information


     Director, Graduate Programs in Public Health, Department of Family
Medicine and Community Health, Tufts University
     School of Medicine, Boston, MA 02111
     J Pediatr 2000;137:604-7.
     Copyright © 2000 by Mosby, Inc.
     0022-3476/2000/$12.00 + 0  9/18/111162