Vitamin A and carotenoid toxicity

Rune Blomhoff

Abstract When the intake of vitamin A is inadequate to meet

the body’s needs, the liver stores are depleted to main-
Large differences in time and dose needed to induce tain serum retinol at a normal concentration. If intake
hypervitaminosis A have been observed. High doses of is low over a prolonged period of time, the serum
vitamin A in food and oily solutions are well tolerated, retinol concentration will decrease. A consequence of
whereas emulsified preparations have higher toxicity. vitamin A deficit in target cells is clinical vitamin A
Chronic hypervitaminosis seems to be induced follow- deficiency, which is characterized by ocular features
ing daily doses of 300,000 to 600,000 IU of vitamin A (Bitot’s spots, night-blindness, and xerophthalmia)
(90–180 mg of retinol) in oily preparations for many and keratomalacia leading to corneal drying (xerosis),
months or years, whereas teratogenicity may be induced corneal ulceration, and corneal necrosis. Generalized
by daily doses as low as 40,000 IU of vitamin A (12 impaired resistance to infection is also reduced.
mg of retinol) in oil during the first trimester. For the Vitamin A deficiency is a public health problem
provitamin A, β-carotene, serious adverse effects have in over 120 countries. It is estimated that about 100
been reported in large-scale prospective randomized million children are subclinically vitamin A deficient,
trials: four years of supplementation with 20 to 30 mg 3 million are clinically vitamin A deficient, and over 1
β-carotene per day was associated with increased risk of million childhood deaths are associated with vitamin
lung cancer and cardiovascular disease among smokers A deficiency annually. At least 5 million children
and workers exposed to asbestos. These results strongly develop xerophthalmia each year, of whom up to half a
suggest that high doses of β -carotene should not be million go blind. In vitamin A–deficient areas, women
recommended for any group until the safety of such doses of childbearing age are also at high risk for vitamin
can be established. A deficiency because of increased vitamin A require-
ments during pregnancy and lactation [1].
Trials indicate that wherever vitamin A deficiency
Introduction exists, routine vitamin A supplements given between
6 and 72 months of age can be expected to reduce
Humans need vitamin A for normal functioning of mortality by an average of 23%, based on eight field
the retina, maintenance of epithelial surfaces, immune studies [2]. In addition, vitamin A supplements as part
competence, growth, development, and reproduction. of measles management can reduce the mortality by
The fat-soluble vitamin A is obtained either from more than 50% [3]. Recent studies in areas where
preformed vitamin A present in milk, eggs, butter, and vitamin A is suboptimal also suggest that vitamin A
fish liver oils, or as provitamin A carotenoids in dark- supplements given to mothers immediately after birth
green leafy vegetables, red palm oil, and red- or orange- will improve their vitamin A status and the vitamin A
colored fruits and vegetables, such as carrots, mangoes, content of their breastmilk, which in turn seem to
papayas, and sweet potatoes. Most of the absorbed contribute to improved health outcomes of both moth-
vitamin A is transported from the small intestine to ers and infants. Due to these developments, research
the liver, where it is stored as retinyl palmitate. It is in this field is now focusing not only on the efficacy of
released into the bloodstream in combination with vitamin A interventions but also on the assessment of
a specific binding protein, from which target cells the sustainability of vitamin A interventions [2].
throughout the body take it up. Several strategies are needed to ensure adequate
vitamin A intakes for all people, including the use
The author is affiliated with the Institute for Nutrition of periodic large doses, weekly or daily, of vitamin A
Research in the University of Oslo in Oslo, Norway. supplements, fortification of commonly consumed

320 Food and Nutrition Bulletin, vol. 22, no. 3 © 2001, The United Nations University.
Vitamin A and carotenoid toxicity 321

foods, and educational programs to improve the diet. all-trans retinol

In areas of vitamin A deficiency, it is believed that OH
correcting vitamin A deficiency is likely to have an
impact on reducing childhood mortality that is at least
as great as that of any single immunization. Because retinyl palµitate O
vitamin A is stored in the liver, the impact of large- O C15H31
dose supplementation is relatively long lasting, and the
appropriate dosages could potentially be given weekly,
monthly, or more infrequently [1, 2]. 11-cis-retinal
The present international practice has been to use
100,000 IU of vitamin A (30 mg of retinol) for infants
6 to 12 months of age and 200,000 IU of vitamin A CHO
(60 mg of retinol) for children aged 12 months and all-trans-retinoic acid
over, given up to once every six months. For vitamin COOH
A supplementation to the mothers, a 200,000-IU (60
mg of retinol) dose has been given, preferably at the
time of delivery but no later than within six weeks 11-cis-retinoic acid
of delivery.
The last review of vitamin A safety was done by the
COOH
International Vitamin A Consultative Group (IVACG)
several years ago [4], and since then important new β-carotene
developments have taken place and have justified
an update. In addition, UNICEF, the World Health
Organization (WHO), and other nongovernmental
organizations have put a lot of effort into vitamin FIG. 1. Structural formulas of some natural retinoids and
A deficiency control since the Montreal Conference beta-carotene
on Hidden Hunger in 1989. The increasing use of
vitamin A supplements has raised questions regarding synthetic compounds, such as TTNN and Ch-55* [7],
the safety of vitamin A, particularly in young children which do not fit into this definition of retinoids, have
(less than six months old) and pregnant women and been shown to be much more active than retinol or
in connection with immunizations. We aim to review retinoic acid in assays for vitamin A or retinoid activ-
the safety of vitamin A and carotenoids related to ity. Sporn and Roberts [7] proposed in 1985 that “a
such intervention strategies. For evaluating the safety retinoid should be defined as a substance that can
of vitamin A and carotenoids, it is helpful to have elicit specific biologic responses by binding to and
an insight into their metabolism and functions in activating a specific receptor or set of receptors.” In
the body. This report will therefore first give a short practice, most researchers today use a combination
introduction to these aspects. of these two definitions, that is, the class of retinoids
consists not only of retinol analogues (with or without
biologic activity), but also of several compounds that
Metabolism of vitamin A and carotenoids are not closely related to retinol but elicit biologic
vitamin A or retinoid activity.
Vitamin A is a term reserved to designate any com- Retinol (molecular weight, 286 daltons) and its
pound possessing the biological activity of retinol derivatives are hydrophobic compounds that are highly
[5] (fig. 1). The term retinoids, which was introduced unstable in the presence of oxygen and yield a mix-
in 1976 by Sporn et al. [6], was designated by the ture of dehydrated and double-bonded rearrange-
International Union of Pure and Applied Chemis- ment products in acids. Light catalyzes double-bond
try–International Union of Biochemistry (IUPAC- isomerization of most retinoids.
IUB) [5] to include “compounds consisting of four Vitamin A exists in the plant world only in the form
isoprenoid units joined in a head-to-tail manner; all of precursor compounds, such as β-carotene. This
retinoids may be formally derived from a monocyclic is a member of a large class of naturally occurring
parent compound containing five carbon-carbon carotenoids, with about 50 compounds with vitamin
double bonds and a functional terminal group at the A activity. In all cases, it is a requirement for vitamin
terminus of the acyclic portion.” By this definition,
retinoids would include both the naturally occurring * TTNN: 6-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-
forms of vitamin A and the many synthetic analogues naphthalenyl)-2-naphthalene carboxylic acid.
of retinol, with or without biological activity. One Ch-55: (E)4-[3-(3,5-di-tert-butylphenyl)-3-oxo-1-
problem with this definition is the fact that several propenyl]-benzoic acid.
322 R. Blomhoff

A activity that after a split of the molecule at least ecules of triacylglycerol and phospholipids packed
one intact molecule of retinol or retinoic acid can be together in a characteristic manner with carotenoids,
obtained. retinyl esters, and other fat-soluble vitamins and
The dietary sources of vitamin A are provitamin cholesteryl esters and a few specific apolipoproteins.
A carotenoids from vegetables and preformed retinyl These huge lipoproteins (100–2,000 nm in diameter)
esters from animal tissues. Traditionally, one interna- are exocytosed into the intestinal lymph and then move
tional unit (IU) of vitamin A is defined as 0.3 µg of all- into the general circulation, where several processes,
trans-retinol. For nutritional purposes, a better term such as triacylglycerol hydrolysis and apolipoprotein
is retinol equivalents (RE), which is used to convert exchange, result in the formation of chylomicron rem-
all sources of vitamin A and carotenoids in the diet nants. Almost all retinyl esters and carotenoids present
into a single unit. Thus, 1 µg of all-trans-retinol equals in the chylomicrons remain with the particle during
1 RE. Generally, 1 µg of retinol has been assumed to conversion to chylomicron remnants [11–14](fig. 2).
be biologically equivalent to 6 µg of β-carotene or 12 Although chylomicron remnants are mainly cleared
µg of mixed dietary carotenoids [4]. However, the basis by the liver, extrahepatic uptake of remnants may be
for these conversion factors is weak. Recent findings important in the delivery of retinol and carotenoids to
from population intervention studies with vitamin A extrahepatic tissues, such as bone marrow, peripheral
status endpoints (serum retinol response and total body blood cells, spleen, adipose tissue, skeletal muscle,
vitamin A by isotope dilution) suggest that the current and kidney. In light of the importance of retinoids for
conversion estimate of 6:1 is a gross overestimate of the regulating gene expression and cellular differentiation,
bioavailability of β-carotene in most foods [8–10]. chylomicrons may be an important transport complex
During absorption, essentially all of the retinyl for delivering retinol and carotenoids to tissues with
esters are enzymatically converted to retinol in the intensive cell proliferation and differentiation, such as
intestinal lumen prior to absorption by intestinal bone marrow and spleen [12, 13]. Since this delivery
cells. Carotenoids, on the other hand, are internalized of vitamin A to tissues increases relatively linearly with
unchanged by the absorptive cells, where they are the dietary intake of preformed vitamin A (but not
partially converted to retinol (fig. 2). Retinol is then carotenoids), acute toxic effects of preformed vitamin
esterified to long-chain fatty acids, forming retinyl A might also be mediated by this uptake rather than
esters, before incorporation into chylomicrons. Chy- retinol bound to retinol-binding protein (RBP) or free
lomicrons consist of aggregates of thousands of mol- retinol or retinoic acid in plasma (fig. 2).

FIG. 2. Major pathways for retinoid transport in the body. ROH, Retinol; RE, retinyl ester; CM,
chylomicron; CMR, chylomicron remnant; RBP, retinol-binding protein; TTR, transthyretin; RA,
retinoic acid; RAR, retinoic acid receptor; RXR, retinoic X receptor
Vitamin A and carotenoid toxicity 323

Most of the absorbed dietary vitamin A is, however, epidermis in organ culture could be changed from
delivered to hepatic parenchymal cells (hepatocytes). keratinized to mucus-producing tissue by treatment
The retinyl esters are hydrolyzed, and retinol may be with retinol or retinyl acetate. During the last 15 years,
transferred to RBP, which is found in high concentra- the ability of retinoids to affect the gene expression
tion in endoplasmic reticulum. Binding of retinol to and differentiation of epithelial cells in vivo and in
RBP apparently initiates a translocation of retinol- vitro has been studied in great detail.
RBP from endoplasmic reticulum to the Golgi com- The importance of retinoids for the proper function
plex, followed by secretion of retinol-RBP from the of both male and female reproductive organs has
cells [11–14]. also been well documented. Classically, it has been
In vitamin A–sufficient states, most of the chylomi- assumed that retinol and retinoic acids perform sepa-
cron remnant retinyl esters taken up by hepatocytes rate functions and that both are needed for normal
are transferred as RBP-bound retinol to perisinusoidal reproduction. Recent data indicate, however, that
stellate cells in the liver for storage [15, 16]. In mam- retinoic acid alone may fulfil all functions of retinoids
mals, 50% to 80% of the body’s total retinol (retinol in this process [23]. Present evidence clearly indicates
plus retinyl esters) is normally present in the hepatic that retinoids can influence growth and differentiation
stellate cells. Ninety-eight percent of the stellate cell of various hematopoietic progenitor cells. It is also
vitamin A is present in the form of retinyl esters packed evident that retinoids play an important role in the
together in cytoplasmic lipid droplets. The normal complex interplay of cells and soluble factors that
reserve of vitamin A in stellate cells is adequate to constitute the immune system [24].
last for several months [17, 18]. The normal massive Vitamin A deficiency, as well as vitamin A excess,
storage of retinyl esters in stellate cells and the cells’ produces a number of malformations of the embryo
ability to control mobilization of retinol ensures that of all vertebrates. These abnormalities involve almost
the blood plasma retinol concentration is defended all organs in the body (e.g., the central nervous system,
close to 1 to 2 µM, in spite of normal fluctuations in eye, face, dentition, ear, limbs, urogenital system, skin,
daily vitamin A intake. lungs, heart, and hematopoietic system, as well as body
The transport of carotenoids has been described axis development). Thus, retinoids are regarded as
in much less detail. The fractions of carotenoids that essential molecules that orchestrate many aspects of
are absorbed from the diet in an unchanged form normal embryonic development [25].
enter the lymphatics in association with chylomicrons Thus, today we know that vitamin A is of central
and follow their transport mainly to the liver. Follow- importance for many biological processes, such as
ing processing of chylomicron remnant carotenoids cell differentiation, proliferation, and apoptosis, and
in parenchymal cells, some of the carotenoids are that the mechanism of action of vitamin A in these
probably converted to retinoids. The carotenoids processes involves regulation of gene expression. This
do not accumulate in liver cells but are mobilized is mediated by nuclear receptors that are specific for
as components of the very-low-density lipoprotein vitamin A metabolites that regulate gene expression
(VLDL) particles that are converted to VLDL remnants by binding to short DNA sequences in the vicinity of
and low-density lipoproteins (LDL) in the circulation target genes [11, 12].
by a process resembling chylomicron remnant forma- In addition to the role of carotenoids as vitamin A
tion. Thus, no specific transport protein exists for precursors, the carotenoids have a separate function as
carotenoids in plasma. The carotenoids may to a cer- antioxidants. Oxidative stress is believed to play a role
tain degree accumulate in adipose tissues following in cancer and respiratory tract infections. Therefore,
ingestion of large doses [19, 20]. the antioxidant function of carotenoids turns out to be
another interesting facet of the function of this family
of molecules [26].
Physiological function of vitamin A and
carotenoids
Requirements of vitamin A and carotenoids
We learned from early research that vitamin A func-
tions as a chromophore in the visual process, and that In most societies, vitamin A needs are met by the
vitamin A deficiency and vitamin A excess dramati- combined consumption of carotenoid precursors of
cally change the differentiation of epithelial cells. As vitamin A in plants and preformed vitamin A (retinol
early as 1925, Wolbach and Howe [21] showed that and retinyl esters) in animal foods. There has been no
vitamin A deficiency in rats led to the replacement consensus in different countries as to the amount of
of differentiated mature epithelium with squamous vitamin A and carotenoids that should be consumed in
keratinizing epithelial cells in tissues from various order to maintain optimal health. Whereas no recom-
parts of the body. Twenty-eight years later, Fell and mendations are generally given for carotenoids, most
Mellanby [22] reported that the phenotype of chick countries have based their recommendations between
324 R. Blomhoff

500 and 1,000 µg of retinol equivalents (RE) for adults rations [30–34]. It is useful to differentiate between
[27–29]. The recommendations are also based on a the acute vitamin A intoxication caused by short-
surprisingly small number of studies. term ingestion of excessive amounts of vitamin A
Much of the world relies on WHO and the Food [35–38] and the chronic hypervitaminosis A resulting
and Agriculture Organization (FAO) to publish and from long-term intake of more moderate vitamin A
disseminate the technical information that is used for doses [39–42]. Thus, acute hypervitaminosis A can
national dietary allowances or as a basis for produc- be defined as any toxicity manifested following the
ing national nutrient requirement standards. The ingestion of a single very high dose or several repeated
establishment of human nutrient requirements is very high doses over a few days. Furthermore, chronic
a prerequisite for countries to develop food-based hypervitaminosis A can be defined as any symptoms
dietary guidelines for their populations. During the resulting from continued ingestion of high doses
1980s, WHO and FAO reviewed the literature on for months or years. Toxicities associated with acute
vitamin A requirements. However, progress in vitamin and chronic excess of vitamin A intake have been
A research since then requires updating knowledge extensively reviewed [43, 44]. The symptoms of acute
of the subject. WHO and FAO are now in the process and chronic hypervitaminosis A are summarized
of revising their recommendations. Table 1 presents in table 2.
the recommended dietary intake of vitamin A from
WHO/FAO, the European Union, and Recommended
Dietary Allowances/National Academy of Sciences.
TABLE 2. Symptoms of hypervitaminosis A

Symptoms of hypervitaminosis A Acute hypervitaminosis A

Increased CSF pressure. Bulging of fontanelle in infants,
Symptoms of hypervitaminosis A may occur in the
headache and blurred vision in adolescents and adults
skin, nervous system, musculoskeletal system, circula- Loss of appetite, nausea, vomiting, abdominal pain
tion (e.g., plasma proteins), and internal organs. The Peeling of skin, cheilitis, loss of hair
toxicity varies with dose and body mass, age, sex, Fatigue, lassitude, vertigo, somnolence, disturbance of
disease conditions, concurrent drug administration, consciousness, occasionally meningeal irritation
and environmental chemical exposures. Hemorrhages, nose bleeding
Toxic reactions provoked by large doses of vitamin Edema
A are well known to occur following either intake of Occasionally tenderness of the long bones
liver rich in vitamin A (e.g., polar bear, seal, halibut, or Occasionally hepatomegaly and splenomegaly
whale) or excessive administration of vitamin A prepa- Chronic hypervitaminosis A
Desquamation and peeling of the skin, erythema, pruri-
tus, disturbed hair growth
TABLE 1. Comparison of recommended dietary intakes of Desquamation of mucous membranes: cheilitis, angular
vitamin A (RE/day) stomatitis, gingivitis, glossitis
Pain and tenderness of the bones: restricted movement
Food and Radiological detectable bone changes (hyperostosis
Nutrition corticalis)
FAO/ Board, Increased CSF pressure, headache predominantly in the
WHO EU NAS, RDA occipital region
Group Age (yr) 1988 [27] 1992 [28] 1989 [29] Hyperirritability, sleep disturbance: EEG changes only
exceptionally
Infants 0–1 350 350 375 Papillary edema, diplopia
Children 1–3 400 400 400 Anorexia and loss of weight
Children 4–6 400 400 500 Hepatomegaly, sometimes with splenomegaly
Children 7–10 400 500 700 Edema and swelling
Males 11–12 500 600 1,000
Males 12–15 600 600 1,000 Clinical-chemical findings
Males 15+ 600 700 1,000
Females 11–12 500 600 800 Retinyl ester concentration in plasma is always increased
Females 12–15 600 600 800 Increased serum retinol dependent of dose
Females 15+ 500 600 800 Calcium concentration in serum is often increased
Pregnant 600 700 800 Activity of serum alkaline phosphatase is often increased
Lactating 1st 6 mo 850 950 1,300 SGOT and SGPT activities are occasionally increased with
Lactating 2nd 6 mo 850 950 1,200 chronic intoxication
Sources: refs. 27–29. Source: refs. 30–44.
Vitamin A and carotenoid toxicity 325

Lowest dose needed to induce acute and It is known from a series of other experimental
chronic hypervitaminosis A studies that vitamin A is absorbed much better from
emulsified than from oily solutions. Early studies in
The large differences in time and dose needed to experimental animals, infants, children, and adults
induce hypervitaminosis A from supplements may demonstrated that aqueous emulsions of retinol and
at first glance look surprising and may lead errone- retinyl esters resulted in higher plasma peak values,
ously to the assumption that individual tolerance to higher liver values, and lower fecal losses than oily
vitamin A varies over a wide range. Much of the vari- vitamin A preparations [46–48]. These observations
ance may, however, be related to the physical formula- have recently been confirmed [49–52].
tions of the ingested preparations, that is, whether The safe dose of vitamin A is therefore related to
vitamin A exists in an oily solution or is emulsified the physical form of the vitamin A preparation. It is
in an aqueous and equivalent formulation. If vitamin conceivable that preparations of vitamin A blended
A is ingested as an oily solution, fairly high doses with a detergent or water, or lyophilized and packed
seem to be tolerated—independently of age—until in dry tablets, have a lower threshold for toxicity
hypervitaminosis A symptoms appear. As reviewed than vitamin A in oily emulsions or in foods such as
by Korner and Vollm [45] and presented in figure 3, liver. This knowledge seems to have been forgotten
a daily intake of 5,000 IU of vitamin A (1.5 mg of by most investigators in the field and has not been
retinol) per kilogram of body weight for 3.5 years is taken into account when considering the toxicity and
necessary before symptoms appear. In contrast, in the teratogenicity of vitamin A in almost all more recent
case of aqueous emulsions, the period before symp- studies.
toms appear after administration is only approximately In agreement with the threshold levels presented
seven months with identical doses. Therefore, one may in figure 3, controlled studies aimed at secondary
have to take into account an approximately sixfold cancer prevention also demonstrated that daily doses
higher toxicological activity for emulsified vitamin A of 300,000 to 600,000 IU of vitamin A (90–180 mg
preparations as compared with oily preparations. Thus, of retinol) as retinol in oily preparations for many
it is impossible to draw conclusions on the safe dose months or years are usually required to produce signs
of vitamin A when literature reports do not present of hypervitaminosis A [53] in adults. However, toxic
details of the vitamin A preparation administered. It symptoms developed earlier when similar doses of
would very important for further studies to validate emulsified water-soluble formula were used, as in an
the suggestions from Korner and Vollm [45]. Italian study [54] in which doses of 300,000 IU of reti-
Both graphs in figure 3 show a clear relation between nol (90 mg of retinol) were given daily for 12 months
the daily ingested amount of vitamin A and the dura- to patients with resected stage I non-small-cell lung
tion of intake with regard to the appearance of hyper- cancer. Serum levels of γ-glutamyltranspeptidase rose
vitaminosis A symptoms. Administration of very high during treatment but were significantly higher only
doses significantly reduces the time before onset of after two years. Serum triglyceride levels increased 63%
hypervitaminosis A, whereas “moderate” doses prolong over the first year of treatment and were significantly
the lag period [45]. higher than those of controls at 8 and 12 months
[55, 56]. The majority of adverse events in the Italian
study were dermatological (dryness, desquamation,
Daily intake (IU/kg body weight)

45000
and itching).
40000
Higher toxicity with emulsified preparations than
35000 with oily preparations was also observed in the Indian
30000 National Program in the 1970s [57]. Reversible toxic
25000 reactions were too frequently reported when an annual
20000 oral dose of 200,000 to 300,000 IU (60–90 mg of reti-
nol) was given as an emulsified preparation. However,
15000
these effects were nearly eliminated when the program
10000 changed to comparable amounts of vitamin A in an
5000 oily solution.
0
0 500 1000 1500

Time (days) May vitamin A intake marginally above the

recommended dietary intake increase the
FIG. 3. Time until the appearance of chronic hypervitami- risk of osteoporosis?
nosis A after daily doses of preformed vitamin A in oily
solution ( ) and water-soluble preparations ( ). Source: Recently, Melhus et al. [58] in an epidemiological
modified from ref. 45 study found that subject-reported dietary intakes of
326 R. Blomhoff

retinol greater than 5,000 IU (1.5 mg of retinol) per retinoids. Doses of 13-cis-retinoic acid (isotretinoin)
day were associated with reduced bone mineral density and all-trans-retinoic acid (tretinoin) of 0.4 to 2.0
and increased risk of hip fracture. The relation was mg/kg/day in the first trimester of pregnancy caused
dose-dependent, and no attenuation of the effect of spontaneous abortions and serious birth defects,
excess vitamin A on the risk of hip fracture was seen including malformations of the brain, heart and major
in multivariate analyses that adjusted for major risk arteries, craniofacies, and thymus [67].
factors for osteoporosis. These results are consistent
with experimental data from both in vitro and in
vivo studies. Animal studies [59–61] have shown the Lowest teratogenic dose in experimental
importance of vitamin A in the bone remodeling animals
process. Vitamin A deficiency results in retarded bone
growth [60], and hypervitaminosis A and pharmaco- High doses of preformed vitamin A (retinol and retinyl
logical doses of retinoids are associated with acceler- esters) undoubtedly cause birth defects in experimental
ated bone resorption, bone fragility, and spontaneous animals. The teratogenic dose of retinol and retinyl
fracture [59, 62–65]. ester varies substantially among species [67, 68]. It is
It was surprising that Melhus et al. [58] observed therefore not easy to predict the lowest teratogenic
these effects at a dose of over 5,000 IU of vitamin A dose in humans from these data.
(1.5 mg of retinol) per day. Results from animal experi- In animal experiments, hypervitaminosis A pro-
ments and clinical experience with retinoids would duced malformations in almost all organ systems. The
suggest that much higher doses are needed to induce type and incidence of malformation depends on the
these adverse effects. A limitation of the study by stage of gestation and dose and, to a lesser extent,
Melhus et al. [58] is the possibility of information bias on the species and strain. Reported structural defects
resulting from the retrospective questioning of vitamin in animals include defects of the brain, spinal cord,
A intake. The study probably had a high degree of face, eye, all parts of the ear, teeth, salivary glands,
random error in the assessment of retinol intake. Other aortic arch, heart, lungs, gastrointestinal tract, liver,
limitations were possible confounding factors, such gallbladder, urinary system, genitalia, pituitary, thyroid,
as lack of data on thyroid hormone therapy, physical thymus, skull, vertebrae, ribs, extremities, muscles, and
activity after hip fracture had occurred, and family situs inversus [43, 44, 67–69].
history of osteoporosis. Moreover, the possibility of It is not always easy to trace how the dose was given
other unidentified dietary confounders cannot be in each experiment. Because the physical form of
excluded. This epidemiological study must therefore the vitamin A preparation has a significant effect on
not be taken as a proof for a cause-and-effect relation- the dose-effect curve, these data may represent an
ship between retinol intake and the risk of hip fracture. underestimation of the doses needed for adverse effects
Further studies are required. when compared with vitamin A in oily solutions and
in foods such as liver.
The lowest teratogenic dose of vitamin A and retin-
Preformed vitamin A and birth defects oids reported in experimental animals is given in table 3.
The lowest dose in rabbits causing teratogenic effects
As early as 1953, Cohlan [66] observed that the admin- was 5.5 mg/kg/day or 18,300 IU/kg/day. A comparable
istration of large doses of retinoids to pregnant experi- dose for a 50-kg pregnant woman would represent an
mental animals resulted in a variety of embryonic intake of 915,000 IU/day (277 mg of retinol). Thus, if
malformations. Although well known to experimental humans are as sensitive to the teratogenicity of vitamin
scientists, it came as a shock to dermatologists when A as the most sensitive experimental animal, the data
they observed similar teratogenic effects in patients would tend to suggest that doses of several hundred
with keratinizing disorders treated with high doses of IU are needed to induce adverse birth defects.

TABLE 3. Lowest teratogenic dose of retinoids (mg/kg/day) in animals and humans

Cynomolgus
Retinoid Rat Mouse Rabbit Hamster monkey Human
Retinol 50–90 25–50 5.5 15 6 Not
determined
All-trans-retinoic acid 0.4–2.0 3 2 7 7.5 Not
determined
13-cis-retinoic acid 150 100 10 25 2.5–5.0 0.4
Source: refs. 43, 44, 67–69.
Vitamin A and carotenoid toxicity 327

When animal data are extrapolated to humans, safety and 13-cis-retinoic acid are generally believed to be
factors are usually applied to the animal doses. Typically, responsible for the teratogenic effect in humans [67].
a 10-fold factor is applied for species differences and Serum concentrations of retinyl palmitate metabo-
a further 10-fold factor for interhuman differences. If lites in humans following administration of vitamin
this commonly used, but somewhat arbitrary, 100-fold A in liver or various types of supplements have been
safety factor is applied, a daily intake of 9,150 IU (2.8 studied (table 4). It should be noted that vitamin A
mg of retinol) for a 50-kg woman is the upper safe intake from liver varies with diet and species. The
limit. This intake is, however, only 3.5 times higher normal concentrations of all-trans-, 13-cis-, 4-oxo-
than the recommended daily vitamin A intake for all-trans-, and 4-oxo-13-cis-retinoic acid in human
pregnant women in many countries (800 µg of retinol) serum are somewhat higher than in rats: 1, 1, 0.5, and
and is very close to the actual average daily intake for 2–4 ng/ml, respectively. When volunteers were given
women in many areas of the world. Thus, the ordinary 10,000 to 25,000 IU of vitamin A (3–7.5 mg of retinol)
safety evaluations used by toxicologists to set limits for in oil, plasma peak levels were in the normal range.
xenobiotic and pollutant exposures are doubtful for After dosing of 50,000 to 500,000 IU of vitamin A
establishing a safe level of vitamin A for humans. (15–150 mg of retinol) in oil or 250,000 to 346,000 IU
of vitamin A (75–104 mg of retinol) in liver, plasma
concentrations of the retinyl palmitate metabolites
Plasma threshold levels of preformed increased somewhat but were still much lower than
vitamin A causing teratogenic effects the expected teratogenic threshold. High plasma
concentrations comparable to a teratogenic dose were
An alternative approach would be to determine observed when retinyl esters were administered in
the lowest serum level of retinoid metabolites that a water-soluble physical form. These data suggest
corresponds to the teratogenic effect observed that the physical form that is used to administer the
following isotretinoin (13-cis-retinoic acid) treatment vitamin is an important determining factor and that
in humans. It should then be possible to give human the following order of potencies exists: liver < oily
volunteers various doses of vitamin A to determine solutions < water-soluble forms. Furthermore, it is
the intake causing teratogenic levels in human plasma. unlikely that the teratogenic threshold is reached
Following dosing of 0.4 mg of 13-cis-retinoic acid when humans are dosed with 10,000 to 500,000 IU
per kilogram in humans, peak plasma concentrations (3–150 mg) of retinol in oil or liver, in contrast to
of 13-cis-retinoic acid between 50 to 150 ng/ml have 500,000 IU (150 mg) of retinol in water-soluble or
been observed [70–72] (table 4). Although it has not emulsified form.
been demonstrated conclusively, all-trans-retinoic acid

TABLE 4. Plasma peak concentrations following dosing with retinoids (ng/ml)a

All-trans- 13-cis- 4-oxo-all-trans- 4-oxo-13-cis-

Dose retinoic acid retinoic acid retinoic acid retinoic acid
Lowest teratogenic dose of 13-cis-retinoic acid Not 50–210 Not 600–800
(0.4 mg/kg) [70–72] determined determined
Normal plasma concentration [73, 74] 1 1 0.5 2–4
Dosing of humans with 10,000 IU (3 mg) retinol 2 2 Not 3
in oil for 60 days [75] determined
Dosing of humans with 25,000 IU (7.5 mg)retinol 2 2 Not 5
in oil for 60 days [75] determined
Dosing of humans with 50,000 IU (15 mg) retinol 4 10 Not 19
in oil for 20 days [73] determined
Dosing of humans with 250,000 IU (75 mg) retinol 2 22 1 32
in fried liver [76]
Dosing of humans with 500,000 IU (150 mg) retinol 4 31 6 43
in fried liver [74]
Dosing of humans with 346,500 IU (104 mg) retinol 5 13 Not Not
in oil [77] determined determined
Dosing of humans with 500,000 IU (150 mg) 87 68 17 64
water-soluble retinol [74]
a. In these studies, retinol was administered as retinyl palmitate, which has similar teratogenicity to retinol.
328 R. Blomhoff

Duration of treatment needed for A supplements used. Although these case reports
teratogenic effect are inconclusive, they are useful for planning well-
controlled prospective studies.
The timing during embryonic development that the No further case reports have been published since
threshold levels of teratogenic retinoids are exceeded 1985 on the teratogenic effect of vitamin A in humans,
is another parameter to be considered in risk assess- in spite of widespread use of high-dose vitamin A
ments. To obtain abnormal second visceral arches supplements. Several million capsules, each with doses
in rat embryos, it is necessary to expose embryos in of 25,000 to 200,000 IU of vitamin A (7.5–60 mg
vitro to teratogenic concentrations of 13-cis-retinoic of retinol), are consumed each year in a number of
acid for three to six hours during the sensitive period countries [79].
[78]. Exposure for less than three hours did not cause It is important to note that there have been no case
abnormal arches, even if the concentration of 13-cis- reports of birth defects from vitamin A exposure due
retinoic acid was greatly increased. It is not known to dietary sources [79]. Intake from a dinner with 150 g
whether this finding applies to other retinoids. This of liver would represent about 100,000 to 200,000 IU
might suggest that in humans teratogenic retinoids of vitamin A (30–60 mg of retinol) [86], but liver is
would need to exceed threshold levels for a certain not likely to be consumed frequently.
period of time to cause teratogenicity. This is charac-
teristic of isotretinoin exposure in the human, where
plasma 13-cis-retinoic acid and 4-oxo-13-cis-retinoic Case-control studies of teratogenicity from
acid are continuously elevated due to the twice-daily preformed vitamin A
dosing regimen. This repeated exposure in the human
is associated with teratogenicity [78]. Several case-control studies have been published that
The necessary duration of exposure in the human focused on the relation between vitamin A consump-
is unknown, but it is most likely to be an important tion and birth defects in humans. Martinez-Frias
parameter. The area under the curve has been used and Salvador [87] evaluated 11,293 cases of congeni-
to determine teratogenicity, but this variable may be tal malformations in Spain with controls who were
inappropriate without consideration of the duration of matched for sex, hospital, and day of birth. Maternal
exposure above the threshold level. Malformations may intake of vitamin A from diet and from supplementa-
be induced only when the threshold level is exceeded tion during pregnancy was assessed by a question-
during a certain period. naire. The authors found no statistically significant
association between exposure to vitamin A and birth
defects. The odds ratio for birth defects from exposure
Case reports of teratogenicity from to vitamin A in multivitamin complexes (mean dose,
preformed vitamin A 20,263 IU of vitamin A = 6.1 mg of retinol) was not
statistically different from 1, whereas exposure to
Several anecdotal reports exist on malformations in vitamin A supplements alone (mean dose, 63,636
infants whose mothers consumed supplements with IU of vitamin A = 19.1 mg of retinol) had an odds
25,000 to 150,000 IU (7.5–45 mg) of retinol per day ratio of 9.9 (p = .006). When they compared birth
during pregnancy. In 1985, Rosa et al. [79] reviewed defects in children of mothers exposed to supplements
the literature and found 18 suspicious birth defect containing various doses of vitamin A with children of
outcomes from pregnancies with high vitamin A dose mothers not taking any supplement, they observed a
exposure. Some of these were reported in proceedings decrease in birth defects, although it was not significant
from meetings, some had been reported to the Food in the group exposed to up to 40,000 IU of vitamin
and Drug Administration (FDA), and six had been A (12 mg of retinol) daily. In the group consuming
reported earlier in scientific journals [80–85]. All more than 40,000 IU of vitamin A (12 mg of retinol), a
but one of the 18 birth defects reported with a high significant increase in birth defects (OR = 2.7, p = .06)
dosage of vitamin A occurred with long-term exposure was found. This study therefore suggests that daily
continuing past conception of 25,000 to 150,000 IU doses of up to 40,000 IU are safe, whereas doses of
(7.5–45 mg) of retinol per day. The single exception more than 40,000 IU may increase the risk of birth
was a massive accidental exposure (500,000 IU or defects. One should, however, be cautious when draw-
150 mg of retinol) occurring in the first month of ing a conclusion based only on the 16 of 11,293 cases
pregnancy [83]. Some of the exposures were in com- of birth defects that had been exposed to doses of
binations with high dosages of other vitamins, such as vitamin A of 10,000 IU (3 mg of retinol) or more in
vitamin D and E, and there is no information about this study.
dietary intake of vitamin A in combination with sup- Werler et al. [88] conducted a study in the United
plementary intake. In addition, very little information States of 2,658 infants with birth defects possibly due
is available as to the physical form of the vitamin to abnormal development of cranial neural crest cells
Vitamin A and carotenoid toxicity 329

and 2,609 controls. Vitamin A supplementation was multivitamin use had a 43% lower risk of having
defined as daily use for at least seven days of retinol infants with conotruncal defects than did mothers
alone or with vitamin D, or fish oils. Information who reported no use. Thus, periconceptional mul-
on the vitamin A dose and dietary intake was not tivitamin use is associated with a reduced risk of
available. Women who used vitamin A supplements in conotruncal defects.
early pregnancy had approximately a twofold increased
risk of giving birth to an infant with malformation of
cranial neural crest-derived structures. However, since Prospective cohort study of teratogenicity
these findings were based on few exposed controls, risk from preformed vitamin A
estimates were unstable and were also compatible with
no association. Since supplements containing 25,000 In a nonrandomized, prospective cohort study, Roth-
to 100,000 IU of vitamin A (7.5–30 mg of retinol) man et al. [92] obtained information on the diets of
were common during the years when the infants in the 22,748 pregnant women during their first trimester of
study by Werler et al. [88] were born, it is reasonable pregnancy. Information on the outcome of pregnancy
to suggest that the users of vitamin A supplements was obtained from the obstetricians who delivered the
included some who consumed doses up to 100,000 IU babies or from the women themselves. Of the 22,748
(30 mg of retinol) per day. women, 339 had babies with birth defects; 121 of these
In a geographically based case-control study in the babies had defects occurring in sites that originated in
United States, Mills et al. [89] interviewed women the cranial neural crest. The proportion of total birth
whose pregnancies produced offspring with neural defects and cranial neural crest-derived birth defects
tube defects (n = 548) or major malformations other appeared to be relatively constant for the women
than neural tube defects (n = 387) and normal control who had total intakes of 0 to 15,000 IU of vitamin A
subjects (n = 573) to determine periconceptional vita- (0–4.5 mg of retinol) per day (table 5).
min A supplement exposure levels. The proportion of For defects associated with cranial neural crest tissue,
women consuming doses of vitamin A between 8,000 the ratio of the prevalence among the babies born to
and 25,000 IU (2.4–7.5 mg of retinol) was no greater women who reported consuming more than 15,000
in any malformation group than in the normal control IU of preformed vitamin A (4.5 mg of retinol) per
group. Thus, Mills et al. [89] found no association day from food and supplements, as compared with the
between periconceptional vitamin A exposure at doses prevalence among babies whose mothers consumed
consumed by most women during organogenesis. 5,000 IU (1.5 mg of retinol) or less per day, was 3.5
Recently, Czeizel and Rockenbauer [90] performed a (95% CI, 1.7–7.3). For vitamin A from supplements
paired analysis of cases with congenital abnormalities alone, the ratio of the prevalence among the babies
and healthy controls in a large population-based data born to women who consumed more than 10,000
set of the Hungarian Case-Control Surveillance of IU of vitamin A (3 mg of retinol) per day to that
Congenital Abnormalities. Of 35,727 pregnant women among the babies whose mothers consumed 5,000 IU
who had control infants without defects, 9.5% received
vitamin A. Of 20,830 pregnant women who had off- TABLE 5. Birth defects in the prospective cohort study of
spring with congenital abnormalities, 7.9% consumed Rothman et al.[92]
vitamin A supplements, a rate that is significantly
lower than that of the control group. Thus, this study Cranial
indicated that low or moderate doses of vitamin A Total neural crest Total birth
pregnancies defects defects
(<10,000 IU ) during the first trimester of pregnancy Daily dose (IU) (no.) (%) (%)
are not teratogenic but have some protective effect
on the fetus. Total intake
Another case-control study [91] in the United States 0–5,000 6,410 0.5 1.3
of conotruncal heart defects, which arise in part from 5,001–10,000 12,688 0.5 1.5
10,001–15,000 3,150 0.6 1.3
cranial neural crest cells, found a lower risk among
≥15,001 500 1.8 3.0
women who used multivitamin preparations that
included vitamin A during the periconceptional period From food
0–5,000 21,755 0.5 1.5
than among those who did not. Botto et al. [91] identi-
5,001–10,000 805 0.6 1.7
fied 158 infants with conotruncal defects and 3,026 ≥10,001 188 1.1 2.7
unaffected, randomly chosen control infants born to
From supplements
mothers residing in metropolitan Atlanta, Georgia,
0–5,000 11,083 0.5 1.3
USA. Periconceptional multivitamin use was defined 5,000–8,000 10,585 0.5 1.6
as regular use from three months before conception 8,001–10,000 763 1.2 1.7
through the third month of pregnancy, as reported by ≥10,001 317 2.2 3.2
the subject. Mothers who reported periconceptional
330 R. Blomhoff

of vitamin A (1.5 mg of retinol) or less per day was extended periods. Hypercarotenemia has been seen
4.8 (95% CI, 2.2 to 10.5). Since the mean vitamin A in people who consume large quantities of food rich
intake in this group of women was 21,675 IU per day, in β-carotene, such as carrots. These individuals also
it is likely that some women took supplements with develop yellow palms and soles, a condition technically
much more than 25,000 IU per day. It would be helpful known as hypercarotenodermia. This condition can be
to know just how much of the apparent association clearly differentiated from jaundice, because the whites
between vitamin A consumption and birth defects in of the eyes are yellow only in patients with jaundice.
this study resulted from the consumption of vitamin These symptoms disappear with reduced intake [100].
A at these higher levels. A rare genetic inability to convert β-carotene to vita-
Although the risk estimate is statistically significant, min A [101, 102], hypothyroidism, diabetes mellitus,
it is important to note that it is based on just five and hepatic and renal disease can cause hypercaro-
excess defects in the high-supplement group. Other tenemia [98]. High lycopene intake also produces
factors that could have given rise to any or all of these hypercarotenodermia; however, a deeper orange is
five events include diagnostic misclassification (espe- usually observed than with β-carotene [100].
cially when a quarter of all reports were based solely Canthaxanthin is an approved food color additive,
on maternal reports and were not verified further); but it has been used without regulatory approval
maternal abuse of other vitamins, drugs, or noxious for attaining a skin color similar to a suntan. Exces-
agents (since most women who took a vitamin A sup- sive intake produces discolored plasma and feces,
plement also consumed other vitamins at unstated which probably has no physiological significance;
doses); and, of course, chance. however, crystalline deposits occurred in the retinas of
This study therefore indicated that intake of supple- all subjects who ingested 60 mg, and a change in retinal
ments at some level above 10,000 IU of vitamin A (3 function after long-term treatment was observed in a
mg of retinol) increases both total and cranial neural few persons with such deposits [103]. The dose of can-
crest-derived defects. Rothman et al. [92] suggested thaxanthin required for this ocular response appears
that an apparent threshold for birth defects occurs to be more than 30 mg/day [104]. The changes in
near 10,000 IU (3 mg of retinol) from supplements retinal function are corrected over a period of months
per day. However, this threshold is not consistent with or years [103, 105]. Similar canthaxanthin retinopathy
other data on teratogenicity [90], and the study has induced in monkeys was not associated with retinol
been criticized on several scientific grounds [93–96]. dysfunction or abnormal morphology [106].
Furthermore, no information is available as to the The retinas of 26 patients with protoporphyria
physical form of supplements consumed in the various who received treatment with β-carotene (200 30-mg
case-control studies or the prospective study. Water- capsules per month for several months) for periods
soluble forms, which might have been responsible of 1 to 10 years were not found to have crystalline
for a major fraction of the birth defects described, deposits. Asymptomatic retinopathy, consisting of
are about six times more toxic than vitamin A in oily multiple, bright-yellow, glistening crystalline deposits
solutions, but this has not been considered in any of in and around the maculae, was observed in 6 of 50
these studies. patients who had ingested more than 200 30-mg doses
of canthaxanthin as a photoprotectant and systemic
skin colorant [107].
Safe dose of carotenoids Experimental and epidemiological investigations
have suggested that food rich in β-carotene might
Experimental studies in animals have shown that reduce the risk of cancer, particularly lung cancer
β-carotene is not mutagenic or teratogenic. Doses and coronary heart disease. Therefore, several trials
up to 180 mg of β-carotene per day have been used have tested whether large doses of supplementary
for many years to treat patients with erythropoietic β-carotene might reduce the incidence of these dis-
protoporphyria, with no evidence of vitamin A toxicity eases. Although most of these trials have reported no
and without the development of abnormally elevated toxic effects, serious adverse effects were reported in
blood vitamin A concentrations [97, 98]. This is due four large-scale prospective randomized trials.
to the observation that conversion of β-carotene and In the ATBC (α-Tocopherol and β-Carotene Sup-
other provitamin A carotenoids seems to be regulated plements) Study, 29,133 Finnish men aged 50 to 69
by the vitamin A status of individuals. Therefore, high years who smoked five or more cigarettes daily were
intakes of carotenoids do not lead to abnormally high randomly assigned to receive α-tocopherol (50 mg),
vitamin A concentrations or symptoms of hypervita- β-carotene (20 mg), α-tocopherol and β-carotene, or a
minosis A [99]. placebo daily for five to eight years (median, 6.1 years)
Hypercarotenemia, or high serum concentrations [108]. Disappointingly, however, the results showed
of carotene, may occur when individuals take supple- that β-carotene supplementation was associated with
ments containing 20 mg or more of β-carotene for an increase in lung cancer risk of about 20%.
Vitamin A and carotenoid toxicity 331

Similar findings were observed in the CARET (Beta- β-carotene groups than in the placebo group.
carotene and Retinol Efficacy Trial) Study l [109], It is worth noting that both the ATBC and the
which tested the combination of 30 mg of β-carotene CARET studies included only smokers and workers
and 25,000 IU of retinyl palmitate taken daily against exposed to asbestos. The results of these four studies
a placebo in 18,314 men and women at high risk of strongly suggest, however, that high doses of β-carotene
developing lung cancer. The CARET intervention was (20–30 mg) should not be recommended for any group
stopped 21 months early because of clear evidence of until the safety of such doses can be established.
no benefit and substantial evidence of possible harm;
there were 28% more lung cancers and 17% more
deaths in the active intervention group receiving the Conclusions
daily combination of 30 mg of β-carotene and 25,000
IU of retinyl palmitate. When considering the toxicity of vitamin A, it is
The CARET study also observed that the active- important to take into account the observation that
treatment group had a 26% increase in the relative emulsified preparations seem to be approximately six
risk of death from cardiovascular disease [110]. Thus, times more toxic than oily preparations or foods such
after an average of four years of supplementation, as liver. Chronic hypervitaminosis seems to be induced
the combination of β-carotene and vitamin A had following daily doses of 300,000 to 600,000 IU of
no benefit and may have had an adverse effect on vitamin A (90–180 mg of retinol) in oily preparations
the incidence of risk of death from lung cancer, from for many months or years, whereas teratogenicity may
cardiovascular disease, and from any cause in smokers be induced by daily doses as low as 40,000 IU of
and in workers exposed to asbestos. vitamin A (12 mg of retinol) in oil during the first
Rapola et al. [111] studied the frequency of major trimester of pregnancy. The controlled, periodic distri-
coronary events in 1,862 men enrolled in the ATBC bution of a single 50,000-IU dose of vitamin A (15 mg
Study (smokers aged between 50 and 69 years) who of retinol) confers no apparent acute risk on young
had had a previous myocardial infarction. In agree- infants, whereas a 100,000-IU (30 mg of retinol) dose
ment with the CARET study, there were significantly is associated with a minimum risk of transient acute
more deaths from coronary heart disease in the side effects.

References
1. WHO. Global prevalence of vitamin A deficiency. MDIS literature. Eur J Clin Nutr 1996;50(suppl 3):S38–53.
Working Paper No. 2. WHO/NUT/95.3. Geneva: World 9. Reddy V. Vitamin A status and dark green leafy vegeta-
Health Organization, 1995. bles. Lancet 1995;346:1634–5.
2. Beaton GH, Martorell R, Aronson KJ, Edmonston B, 10. Parker RS. Methodological considerations in determin-
McCabe G, Ross AC, Harvey B. Effectiveness of vitamin A ing vitamin A and carotenoid activity in humans. Food
supplementation in the control of young child morbid- Nutr Bull 2000;21:124–9.
ity and mortality in developing countries. ACC/SCN 11. Sporn MB, Roberts AB, Goodman DS, eds. The retinoids.
State of the Art Series Nutrition Policy Discussion Paper Vols. 1 and 2. Orlando, Fla, USA: Academic Press, 1984.
No 13. Geneva: United Nations Administrative Com- 12. Blomhoff R, ed. Vitamin A in health and disease. New
mittee on Coordination/Sub-Committee on Nutrition York: Marcel Dekker, 1994.
(ACC/SCN), 1993. 13. Blomhoff R, Green MH, Berg T, Norum KR. Transport
3. Villamor E, Fawzi WW. Vitamin A supplementation: and storage of vitamin A. Science 1990;250:399–404.
implications for morbidity and mortality in children. J 14. Wolf G. A history of vitamin A and retinoids. FASEB
Infect Dis 2000;182(suppl):S122–33. J 1996;10:1102–7.
4. Bauernfeind JC. The safe use of vitamin A. A report of the 15. Blomhoff R, Helgerud P, Rasmussen M, Berg T, Norum
IVACG. Washington, DC: Nutrition Foundation, 1980. KR. In vivo uptake of chylomicron [3H]retinyl ester by
5. International Union of Pure and Applied Chemistry– rat liver: evidence for retinol transfer from parenchymal
International Union of Biochemistry (IUPAC-IUB) to nonparenchymal cells. Proc Natl Acad Sci USA
Joint Commission on Biochemical Nomenclature. 1982;79:7326–30.
Nomenclature of retinoids: recommendations. Arch 16. Blomhoff R, Berg T, Norum KR. Transfer of retinol
Biochem Biophys 1983;224:728–31. from parenchymal to stellate cells in liver is mediated by
6. Sporn MB, Dunlop NM, Newton DL, Smith JM. Preven- retinol-binding protein. Proc Natl Acad Sci USA 1988;
tion of chemical carcinogenesis by vitamin A and its 85:3455–8.
synthetic analogs (retinoids). Fed Proc 1976;35:1332–8. 17. Blomhoff R, Rasmussen M, Nilsson A, Norum KR,
7. Sporn MB, Roberts AB. What is a retinoid? Ciba Found Berg T, Blaner WS, Kato M, Mertz JR, Goodman DS,
Symp 1985;113:1–5. Eriksson U. Hepatic retinol metabolism. Distribution
8. de Pee S, West CE. Dietary carotenoids and their role of retinoids, enzymes, and binding proteins in isolated
in combating vitamin A deficiency: a review of the rat liver cells. J Biol Chem 1985;260:13560–5.
332 R. Blomhoff

18. Blomhoff R, Wake K. Perisinusoidal stellate cells of B. A rare cause of hydrocephalus in an infant: chronic
the liver: important roles in retinol metabolism and vitamin A poisoning. Arch Fr Pediatr 1986;43:501–2.
fibrosis. FASEB J 1991;5:271–7. 42. Teo ST, Newth J, Pascoe BJ. Chronic vitamin A intoxica-
19. Parker RS. Absorption, metabolism, and transport of tion. Med J Aust 1973;2(7):324–6.
carotenoids. FASEB J 1996;10:542–51. 43. Hathcock JN, Hattan DG, Jenkins MY, McDonald JT,
20. Castenmiller JJ, West CE. Bioavailability and bioconver- Sundaresan PR, Wilkening VL. Evaluation of vitamin A
sion of carotenoids. Annu Rev Nutr 1998;18:19–38. toxicity. Am J Clin Nutr 1990;52:183–202.
21. Wolbach SB, Howe PR. Tissue changes following 44. Biesalski HK. Comparative assessment of the toxicology
deprivation of fat soluble A vitamin. J Exp Med 1925;42: of vitamin A and retinoids in man. Toxicology 1989;57:
753–77. 117–61.
22. Fell HB, Mellanby E. Metaplasia produced in cultures of 45. Korner WF, Vollm J. New aspects of the tolerance of
chick ectoderm by high vitamin A. J Physiol 1953;119: retinol in humans. J Vitam Nutr Res 1975;45:363–72.
470–88. 46. Lewis JM, Bodansky O, Birmingham J, Cohlan SQ.
23. Eskild W, Hansson V. Vitamin A functions in the repro- Vitamin A—comparative absorption, excretion and
ductive organs. In: Blomhoff R, ed. Vitamin A in health storage of oily and aqueous preparations. J Pediatr
and disease. New York: Marcel Dekker, 1994:531–59. 1947;31:496–508.
24. Semba RD. The role of vitamin A and related retinoids in 47. Lewis JM, Cohlan SQ. Comparative absorption of
immune function. Nutr Rev 1998;56(1 Pt 2):S38–48. various types of vitamin A preparations. Med Clin
25. Morriss-Kay GM, Sokolova N. Embryonic development North Am 1950;34:413–24.
and pattern formation. FASEB J 1996;10:961–8. 48. Kalz F, Schafer A. Vitamin A serum levels after ingestion
26. Stahl W, Sundquist AR, Sies H. Role of carotenoids in of different vitamin A preparations. Can Med Assoc
antioxidant defense. In: Blomhoff R, ed. Vitamin A in J 1958;79:918–9.
health and disease. New York: Marcel Dekker, 1994: 49. Warwick WJ, Hansen LG, Sharp H. Absorption of
275–87. vitamin A in patients with cystic fibrosis. Absorption
27. FAO/WHO. Requirements of vitamin A, iron, folate and is best with emulsified vitamin A alcohol. Clin Pediatr
vitamin B12. FAO Food Nutr Ser 23. Rome: Food and (Philadelphia) 1976;15:807–10.
Agriculture Organization/World Health Organiza- 50. Eckhoff C, Bailey JR, Collins MD, Slikker W Jr, Nau H.
tion, 1988. Influence of dose and pharmaceutical formulation of
28. Reports of the Scientific Committee for Food. Food, vitamin A on plasma levels of retinyl esters and retinol
science and techniques. Series No. 31. Luxembourg: and metabolic generation of retinoic acid compounds
Commission of the European Communities, 1993. and beta-glucuronides in the cynomolgus monkey.
29. National Research Council. Recommended dietary Toxicol Appl Pharmacol 1991;111:116–27.
allowances. 10th ed. Washington, DC: National Acad- 51. Collins MD, Eckhoff C, Slikker W, Bailey JR, Nau H.
emy Press, 1989. Quantitative plasma disposition of retinol and retinyl
30. Cleland JB, Southcott RV. Illnesses following the eating of esters after high-dose oral vitamin A administration
seal liver in Australian waters. Med J Aust 1969;1:760–3. in the cynomolgus monkey. Fundam Appl Toxicol
31. Clark L. Hypervitaminosis A: a review. Aust Vet J 1971; 1992;19:109–16.
47:568–71. 52. Johnson EJ, Krasinski SD, Howard LJ, Alger SA, Dutta
32. Misbah SA, Peiris JB, Atukorala TM. Ingestion of SK, Russell RM. Evaluation of vitamin A absorption
shark liver associated with pseudotumor cerebri due to by using oil-soluble and water-miscible vitamin A
acute hypervitaminosis A. J Neurol Neurosurg Psychiat preparations in normal adults and in patients with gas-
1984;47:216. trointestinal disease. Am J Clin Nutr 1992;55:857–64.
33. Rodahl K, Moore T. Vitamin A content and toxicity of 53. Gossel TA, Bricker JD. Vitamins. In: Gossel TA, Bricker
bear and seal liver. Biochem J 1943;37:166–8. JD, eds. Principles of clinical toxicology. New York:
34. Rodahl K. Toxicity of polar bear liver. Nature 1949; Raven Press, 1994:403–17.
164:530–3. 54. Pastorino U, Infante M, Maioli M, Chiesa G, Buyse M,
35. Marino GG, Iacolucci JP. Acute adult vitamin A toxicity: Firket P, Rosmentz N, Clerici M, Soresi E, Valente M.
report of a case. J Am Osteopath Assoc 1987;87:563–5. Adjuvant treatment of stage I lung cancer with high-
36. Furman KI. Acute hypervitaminosis A in an adult. Am dose vitamin A. J Clin Oncol 1993;11:1216–22.
J Clin Nutr 1973;26:575–7. 55. Pastorino U, Chiesa G, Infante M, Soresi E, Clerici
37. LaMantia RS, Andrews CE. Acute vitamin A intoxica- M, Valente M, Belloni PA, Ravasi G. Safety of high-
tion. South Med J 1981;74:1012–4. dose vitamin A. Randomized trial on lung cancer
38. Gangemi M, Maiuri F, Di Martino L, Pettoello M. chemoprevention. Oncology 1991;48:131–7.
Intracranial hypertension due to acute vitamin A 56. Infante M, Pastorino U, Chiesa G, Bera E, Pisani P,
intoxication. Acta Neurol (Napoli) 1985;7:27–31. Valente M, Ravasi G. Laboratory evaluation during
39. Bartolozzi G, Bernini G, Marianelli L, Corvaglia E. high-dose vitamin A administration: a randomized
Chronic hypervitaminosis A in infants, and children. study on lung cancer patients after surgical resection. J
Description of 2 cases and critical review of the litera- Cancer Res Clin Oncol 1991;117:156–62.
ture. Riv Clin Pediatr 1967;80:231–90. 57. Underwood BA. Vitamin A in animal and human
40. Baglin A, Hagege C, Franc B, Richaud M, Prinseau J. nutrition. In: Sporn MB, Roberts AB, Goodman DS,
A systemic-like disease: chronic vitamin A poisoning. eds. The retinoids: biology, chemistry, and medicine.
Ann Med Interne (Paris) 1986;137:142–6. 1st ed. Vol. 1. Orlando, Fla, USA: Academic Press, 1984:
41. Gottrand F, Leclerc F, Chenaud M, Vallee L, Gaudier 281–392.
Vitamin A and carotenoid toxicity 333

58. Melhus H, Michaëlsson K, Kindmark A, Bergström R, acid, and 14-hydroxy-4,14-retro-retinol in human plasma
Holmberg L, Mallmin H, Wolk A, Ljunghall S. Excessive after liver consumption. Life Sci 1996;59:169–77.
dietary intake of vitamin A is associated with reduced 77. Tang G, Russell R. Formation of all-trans-retinoic acid
bone mineral density and increased risk for hip fracture. and 13-cis-retinoic acid from all-trans-retinyl palmitate
Ann Intern Med 1998;129:770–8. in humans. J Nutr Biochem 1991;2:210–3.
59. Moore T, Wang Y. Hypervitaminosis A. Biochem J 1945; 78. Ritchie H, Webster W. Parameters determining isotretin-
39:222–8. oin teratogenicity in rat embryo culture. Teratology
60. Mellanby E. Vitamin A and bone growth: the revers- 1991;43:71–81.
ibility of vitamin A deficiency changes. J Physiol 79. Rosa FW, Wilk AL, Kelsey FO. Teratogen update:
1947;105:382–99. vitamin A congeners. Teratology 1986;33:355–64.
61. Fell H, Mellanby E. The effect of hypervitaminosis A 80. Stänge L, Carlström K, Eriksson M. Hypervitaminosis
on embryonic limb buds cultivated in vitro. J Physiol in early human pregnancy and malformations of the
1952;116:320–49. central nervous system. Acta Obstet Gynecol Scand
62. Armstrong R, Ashenfelter K, Eckoff C, Levin A, Shapiro 1978;57:289–91.
S. General and reproductive toxicology of retinoids. 81. Von Lennep E, El Khazen N, De Pierreux G, Amy
In: Sporn MB, Roberts AB, Goodman DS, eds. The JJ, Rodesch F, Van Regemorter N. A case of partial
retinoids: biology, chemistry, and medicine. 2nd ed. sirenomelia and possible vitamin A teratogenesis.
Vol. 1. New York: Raven Press, 1994:545–72. Prenat Diagn 1985;5:35–40.
63. Kamm J, Ashenfelter K, Ehmann C. Preclinical and 82. Bernhardt IB, Dorsey DJ. Hypervitaminosis A and
clinical toxicology of selected retinoids. In: Sporn MB, congenital renal anomalies in a human infant. Obstet
Roberts AB, Goodman DS, eds. The retinoids: biology, Gynecol 1974;43:750–5.
chemistry, and medicine. 1st ed. Vol. 2. Orlando, Fla, 83. Mounoud RL, Klein D, Weber F. A case of Goldenhar
USA: Academic Press, 1984:287–326. syndrome: acute vitamin A intoxication in the mother
64. Nieman C, Obbink H. The biochemistry and pathology during pregnancy. J Genet Hum 1975;23:135–54.
of hypervitaminosis A. Vitam Horm 1954;12:69–99. 84. Pilotti G, Scorta A. Hypervitaminosis A in pregnancy
65. Bauernfeind JC. The safe use of vitamin A: a report and neonatal malformations of the urinary tract.
of the International Vitamin A Consultative Group Minerva Ginecol 1965;17:1103–8.
(IVACG). New York: Nutrition Foundation, 1980. 85. Morriss GM, Thomson AD. Letter: Vitamin A and rat
66. Cohlan SQ. Excessive intake of vitamin A as a cause of embryos. Lancet 1974;2:899–900.
congenital anomalies in the rat. Science 1953;117:535–6. 86. Scotter MJ, Thorpe SA, Reynolds SL, Wilson LA, Lewis
67. Nau H, Chahoud I, Dencker L, Lammer EJ, Scott WJ. DJ. Survey of animal livers for vitamin A content. Food
Teratogenicity of vitamin A and retinoids. In: Blomhoff Addit Contam 1992;9:237–42.
R, ed. Vitamin A in health and disease. New York: 87. Martinez-Frias ML, Salvador J. Epidemiological aspects
Marcel Dekker, 1994:615–63. of prenatal exposure to high doses of vitamin A in
68. Geelen JA. Hypervitaminosis A induced teratogenesis. Spain. Eur J Epidemiol 1990;6:118–23.
CRC Crit Rev Toxicol 1979;6:351–75. 88. Werler MM, Lammer EJ, Rosenberg L, Mitchell AA.
69. Shenefelt RE. Gross congenital malformations. Animal Maternal vitamin A supplementation in relation to
model: treatment of various species with a large dose of selected birth defects. Teratology 1990;42:497–503.
vitamin A at known stages in pregnancy. Am J Pathol 89. Mills JL, Simpson JL, Cunningham GC, Conley MR,
1972;66:589–92. Rhoads GG. Vitamin A and birth defects. Am J Obstet
70. Lucek RW, Colburn WA. Clinical pharmacokinetics of Gynecol 1997;177:31–6.
the retinoids. Clin Pharmacokinet 1985;10:38–62. 90. Czeizel AE, Rockenbauer M. Prevention of congenital
71. Brazzell RK, Vane FM, Ehmann CW, Colburn WA. Phar- abnormalities by vitamin A. Int J Vitam Nutr Res 1998;
macokinetics of isotretinoin during repetitive dosing to 68:219–31.
patients. Eur J Clin Pharmacol 1983;24:695–702. 91. Botto LD, Khoury MJ, Mulinare J, Erickson JD. Peri-
72. Khoo KC, Reik D, Colburn WA. Pharmacokinetics conceptional multivitamin use and the occurrence of
of isotretinoin following a single oral dose. J Clin conotruncal heart defects: results from a population-
Pharmacol 1982;22:395–402. based, case-control study. Pediatrics 1996;98:911–7.
73. Eckhoff C, Collins MD, Nau H. Human plasma all- 92. Rothman KJ, Moore LL, Singer MR, Nguyen U-SDT,
trans-, 13-cis- and 13-cis-4-oxo-retinoic acid profiles Mannino S, Milunsky A. Teratogenicity of high vitamin
during subchronic vitamin A supplementation: compar- A intake. N Engl J Med 1995;333:1369–73.
ison to retinol and retinyl ester plasma levels. J Nutr 1991; 93. Brent RL, Hendricks AG, Holmes LB, Miller RK.
121:1016–25. Teratogenicity of high vitamin A intake. N Engl J Med
74. Buss N, Tembe E, Prendergast B, Renwick A, George 1996;334:1196–7.
C. The teratogenic metabolites of vitamin A in women 94. Khoury MJ, Moore CA, Mulinare J. Vitamin A and
following supplements and liver. Hum Exp Toxicol birth defects. Lancet 1996;347:322.
1994;13:33–43. 95. Watkins M, Moore C, Mulinare J. Teratogenicity of high
75. Chen C, Mistry G, Jensen B, Heizmann P, Timm U, van vitamin A intake. N Engl J Med 1996;334:1196–7.
Brummelen P, Rakhit AK. Pharmacokinetics of retin- 96. Werler MM, Lammer EJ, Mitchell AA. Teratogenicity of
oids in women after meal consumption or vitamin A high vitamin A intake. N Engl J Med 1996;334:1195–6.
supplementation. J Clin Pharmacol 1996;36:799–808. 97. Mathews-Roth MM. Beta-carotene therapy for eryth-
76. Arnhold T, Tzimas G, Wittfoht W, Plonait S, Nau H. ropoietic protoporphyria and other photosensitivity
Identification of cis-retinoic acid, 9,13-di-cis-retinoic diseases. Biochimie 1986;68:875–84.
334 R. Blomhoff

98. Meyers DG, Maloley PA, Weeks D. Safety of antioxidant 108. Albanes D, Heinonen OP, Taylor PR, Virtamo J, Edwards
vitamins. Arch Intern Med 1996;156:925–35. BK, Rautalahti M, Hartman AM, Palmgren J, Freedman
99. Olson JA. Benefits and liabilities of vitamin A and LS, Haapakoski J, Barrett MJ, Pietinen P, Malila N, Tala
carotenoids. J Nutr 1996;126(4 suppl):1208S–12S. E, Liippo K, Salomaa ER, Tangrea JA, Teppo L, Askin
100. Lascari AD. Carotenemia. A review. Clin Pediatr (Phila- FB, Taskinen E, Erozan Y, Greenwald P, Huttunen JK.
delphia) 1981;20:25–9. Alpha-tocopherol and beta-carotene supplements and
101. McLaren DS, Zekian B. Failure of enzymic cleavage of lung cancer incidence in the Alpha-tocopherol, Beta-
beta-carotene. The cause of vitamin A deficiency in a carotene Cancer Prevention Study: effects of base-line
child. Am J Dis Child 1971;121:278–80. characteristics and study compliance. J Natl Cancer
102. Monk BE. Metabolic carotenaemia. Br J Dermatol 1982; Inst 1996;88:1560–70.
106:485–7. 109. Omenn GS, Goodman GE, Thornquist MD, Balmes J,
103. Weber U, Goerz G, Baseler H, Michaelis L. Canthax- Cullen MR, Glass A, Keogh JP, Meyskens FL Jr, Valanis B,
anthin retinopathy. Follow-up of over 6 years. Klin Williams JH Jr, Barnhart S, Cherniack MG, Brodkin CA,
Monatsbl Augenheilkd 1992;201:174–7. Hammar S. Risk factors for lung cancer and for interven-
104. Köpcke W, Barker FM, Sxhalch W. Canthaxanthin tion effects in CARET, the Beta-Carotene and Retinol
deposition in the retina: a biostatistical evaluation of Efficacy Trial. J Natl Cancer Inst 1996;88:1550–9.
411 patients. J Toxicol Cutan Ocular Toxicol 1995;11: 110. Omenn GS, Goodman GE, Thornquist MD, Balmes J,
89–104. Cullen MR, Glass A, Keogh JP, Meyskens FL, Valanis
105. Arden GB, Oluwole JO, Polkinghorne P, Bird AC, Barker B, Williams JH, Barnhart S, Hammar S. Effects of
FM, Norris PG, Hawk JL. Monitoring of patients taking a combination of beta carotene and vitamin A on
canthaxanthin and carotene: an electroretinographic and lung cancer and cardiovascular disease. N Engl J Med
ophthalmological survey. Hum Toxicol 1989;8:439–50. 1996;334:1150–5.
106. Goralczyk R, Buser S, Bausch J, Bee W, Zuhlke U, Barker 111. Rapola JM, Virtamo J, Ripatti S, Huttunen JK, Albanes
FM. Occurrence of birefringent retinal inclusions in D, Taylor PR, Heinonen OP. Randomised trial of
cynomolgus monkeys after high doses of canthaxanthin. alpha-tocopherol and beta-carotene supplements on
Invest Ophthalmol Vis Sci 1997;38:741–52. incidence of major coronary events in men with previ-
107. Poh-Fitzpatrick MB, Barbera LG. Absence of crystalline ous myocardial infraction. Lancet 1997;349:1715–20.
retinopathy after long-term therapy with beta-carotene.
J Am Acad Dermatol 1984;11:111–3.