HL - Barffour_Inflammation_CRP_serum_retinol_Zambian_children

Am. J. Trop. Med. Hyg., 98(1), 2018, pp. 334–343
doi:10.4269/ajtmh.17-0130
Copyright © 2018 by The American Society of Tropical Medicine and Hygiene

Comparability of Inflammation-Adjusted Vitamin A Deficiency Estimates and Variance in Retinol

Explained by C-Reactive Protein and α1-Acid Glycoprotein during Low and High Malaria
Transmission Seasons in Rural Zambian Children

Maxwell A. Barffour,1 Kerry J. Schulze,1 Christian L. Coles,1 Justin Chileshe,2 Ng’andwe Kalungwana,2 Margia Arguello,1
Ward Siamusantu,3 William J. Moss,1 Keith P. West, Jr.,1 and Amanda C. Palmer1*

1Department of International Health, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Baltimore, Maryland;
2Tropical Disease Research Centre, Ndola, Zambia; 3National Food and Nutrition Commission, Lusaka, Zambia

Abstract. Inflammation-induced hyporetinolemia (IIH), a reduction in serum retinol (SR) during inflammation, may bias
population estimates of vitamin A deficiency (VAD). The optimal adjustment for IIH depends on the type and extent of
inflammation. In rural Zambian children (4–8 years, N = 886), we compared three models for defining inflammation: α-1-
acid glycoprotein (AGP) only (inflammation present if > 1 g/L or normal if otherwise), C-reactive protein (CRP) only
(moderate inflammation, 5–15 mg/L; high inflammation, > 15 mg/L; or normal if otherwise) and a combined model using
both AGP and CRP to delineate stages of infectious episode. Models were compared with respect to 1) the variance in SR
explained and 2) comparability of inflammation-adjusted VAD estimated in low and high malaria seasons. Linear re-
gression was used to estimate the variance in SR explained by each model and in estimating the adjustment factors used
in generating adjusted VAD (retinol < 0.7 μmol/L). The variance in SR explained were 2% (AGP-only), 11% (CRP-only), and
11% (AGP–CRP) in the low malaria season; and 2% (AGP-only), 15% (CRP-only), and 12% (AGP–CRP) in the high malaria
season. Adjusted VAD estimates in the low and high malaria seasons differed significantly for the AGP (8.2 versus 13.1%)
and combined (5.5 versus 9.1%) models but not the CRP-only model (6.1 versus 6.3%). In the multivariate regression, a
decline in SR was observed with rising CRP (but not AGP), in both malaria seasons (slope = −0.06; P < 0.001). In this
malaria endemic setting, CRP alone, as opposed to CRP and AGP, emerged as the most appropriate model for quan-
tifying IIH.

BACKGROUND becoming conventional.15–17 This approach, originally pro-
posed by Thurnham et al.,17 is based on two important as-
Childhood vitamin A deficiency (VAD) is a serious public sumptions, namely 1) the inflammation-associated reduction
health problem, with adverse consequences for morbidity and in retinol coincides with the sequential rise in both CRP (during
mortality.1 It is estimated that up to 202 million preschool the early phase) and AGP (during the late phase), and 2) the
children globally may be deficient, defined as having a serum early and late phases of inflammation likely affect retinol
(or plasma) retinol concentration below 0.7 μmol/L.2 The concentrations differently. Thus, adjustment for both APPs
prevalence of VAD is highest in the World Health Organization may be necessary to fully account for IIH. A limitation of this
(WHO) African region, where children are unable to meet the approach is the potential for over-adjustment. This method
requirements for rapid growth owing to inadequate dietary inherently assumes that any observed difference in retinol,
intake and increased losses from prevalent infections.3–5 In between children with and without inflammation, is entirely
these settings, there remains a need for continued monitoring attributable to the inflammation and is transient, and can
of VAD to inform the design of appropriate interventions. The therefore be mathematically corrected to reflect the distribu-
assessment of vitamin A status is, however, problematic tion of values that would have occurred in the absence of
particularly in regions with a high prevalence of infections. It inflammation. However, the necessity or appropriateness of
has been shown that concentrations of serum retinol (SR), universally adjusting retinol concentrations among individuals
the commonly used biomarker of vitamin A status,6 are re- in the late stages of inflammation, where AGP (but not CRP) is
duced during the acute phase response to infections.7–10 This elevated, remains debatable.17,18 In a systematic review by
phenomenon, referred to as inflammation-induced hypo- Thurnham et al.,17 the ratio of plasma retinol concentration
retinolemia (IIH),11–14 is not completely understood and may among children in the late convalescent stage (elevated AGP
lead to misclassification of vitamin A status at the child and but normal CRP) to that of apparently healthy individuals was
population levels. estimated to be 0.94–1.34, suggesting that elevated AGP is
not always associated with a reduction in retinol. A recent
It is the recommendation of the WHO that to guide the in- study by Wessells et al.,18 which reported that the concen-
terpretation of serum or plasma retinol concentrations, the tration of retinol binding protein (RBP), a proxy indicator for
potential influence of inflammation be characterized by con- SR, was about 18% higher in children with elevated AGP (but
currently assessing one or more acute phase proteins normal CRP), compared with children with normal levels of
(APPs).15 Although methods for the adjustment of IIH are both APPs, further raises questions about the consistency of
evolving, α1-acid glycoprotein (AGP) and C-reactive protein the association between AGP and retinol, and hence, its utility
(CRP) remain the most commonly used APPs. Furthermore, in quantifying IIH. Application of adjustment factors in this
the practice of using both AGP and CRP to control for IIH is context, where AGP is positively (as opposed to negatively)
correlated with retinol or its binding protein would be in-
* Address correspondence to Amanda C. Palmer, Department of consistent with our current understanding of the changes in
International Health, Johns Hopkins University Bloomberg School of retinol over the course of the acute phase reaction.17,19,20
Public Health, 615 North Wolfe Street, W2041, Baltimore, MD 21205.
E-mail: [email protected]

334

OPTIMAL MODELING OF INFLAMMATION-INDUCED HYPORETINOLEMIA 335

Because of this potential for over-adjustment, and also cryovials and transported in liquid nitrogen to TDRC for stor-
considering the cost of running laboratory assays for both age at −80°C until analyzed. Data collection procedures were
CRP and AGP, it is critical that the APPs be assessed only if repeated in the same children at the endline assessment,
there is strong evidence of their relevance in characterizing which coincided with the high malaria transmission season
population vitamin A status. Ideally, any model for adjusting (March 2013).
IIH should meet the following minimum criteria: 1) the variance
in retinol explained by the model should increase when the Laboratory analyses. A reversed-phase HPLC procedure
intensity of the infection or inflammation is increased; and was used in the determination of baseline and endline SR
2) the model should produce a consistent and reliable estimate concentration. A commercial enzyme-linked immunosorbent
of VAD when applied at different time points in the same, non- assays kit was used to determine serum concentrations of AGP
intervened population. In this paper, we explored the extent to (AGP; Abcam, Cambridge, MA; catalog # ab108854). CRP was
which AGP and CRP meet these minimum criteria across two measured on an Immulite analyzer (Immulite 1000; Siemens
malaria seasons (low and high transmission) among children in Medical Solutions Diagnostics, Malven, PA; LKCRP1). Malaria
rural Zambia. In light of new evidence suggesting that malaria slides, stained with 3% Giemsa, were washed, dried, and read
may account for additional variance in retinol, beyond that independently by two technicians using light microscopy.
explained by the conventional APPs,18,21,22 a secondary aim Whenever necessary, a third independent reading was done
was to determine whether the inclusion of malaria in the opti- to resolve discordant pairs. For each positive slide, the corre-
mal correction model further changes the magnitude of the sponding thin film was read to determine the Plasmodium
VAD estimates. species. Laboratory procedures were carried out by Craft
Technologies (SR), the Johns Hopkins Bloomberg School of
MATERIALS AND METHODS Public Health (CRP) and TDRC (AGP and malaria slides).

Ethical clearance. Ethical approval was obtained from the Definitions. The definition of inflammation was model-
Institutional Review Board of the Johns Hopkins Bloomberg dependent. In the univariate AGP model, two inflammation
School of Public Health, Baltimore, MD, and the Ethics Review categories were defined, namely reference (AGP £ 1 g/L) or
Committee of the Tropical Diseases Research Center (TDRC), inflammation (AGP > 1 g/L). In the univariate CRP model, we
Ndola, Zambia. defined three categories of inflammation, namely reference
(CRP < 5 mg/mL), moderate (CRP 5–15 mg/mL), and high
Subjects and sample collection. This study included (CRP > 15 mg/L). In the reference model, four categories of
children 4–8 years of age from rural Mkushi District, a malaria inflammation were defined, namely reference (AGP £ 1 g/L and
endemic setting in Zambia. Data for this analysis were col- CRP £ 5 mg/L), incubation (AGP £ 1 g/L and CRP > 5 mg/L),
lected as part of baseline and endline assessments during the early convalescence (AGP > 1 g/L and CRP > 5 mg/L), and late
implementation of a cluster-randomized controlled trial convalescence (AGP > 1 g/L and CRP £ 5 mg/L). VAD was
designed to evaluate the impact of provitamin A carotenoid defined as SR concentration < 0.7 μmol/L. Children were
biofortified maize meal consumption on vitamin A status in considered to have malaria if they had Plasmodium falciparum
this population of preschool-aged children (registered as parasitemia of any density as defined by either microscopy or
NCT01696148 at www.clinicaltrials.gov). Detailed methods RDT or both, and malaria negative if both RDT and micros-
for the parent study have been published elsewhere.23 We copy were negative. This combined malaria definition was
obtained consent from a parent or legal guardian at the time of implemented because we observed that both RDT- and
the baseline assessment (September 2012), which coincided microscopy-defined malaria were associated with significant
with the low malaria transmission season and collected data elevations in both ferritin and soluble transferrin receptor
on the history of morbidity, dietary intake, and socioeconomic (sTfR). To characterize the baseline status of the study pop-
status. At a central site, we measured height with a Shorr ulation, we defined VAD as retinol < 0.7 μmol/L and iron
board to the nearest 0.1 cm, weight with a SECA 874 digital deficiency as ferritin < 12 μg/L in children < 5 years and ferri-
scale to the nearest 0.1 kg, mid-upper arm circumference with tin < 15 μg/L in older children, as previously described.25 Anemia
insertion tapes to the nearest 0.1 cm and tricipital skinfold with was defined as hemoglobin < 110 g/L for children < 60 months
a Holtain caliper to the nearest 0.1 mm. Axillary temperature and < 115 g/L in older children.25 We defined literacy as the
was taken with a digital thermometer, and children with high ability to read or write in English. We defined stunting and
fever (axillary temperature > 39°C) were referred to the nearest underweight as height-for-age and weight-for-age z-scores,
health center. From each child, trained laboratory technicians respectively, below −2 standard deviation of the WHO Growth
collected approximately 7 mL of venous blood into blood Reference. Fever was defined as axillary temperature > 37.5°C.
collection tubes (Covidien Monoject sterile tubes with no ad-
ditives). Malaria diagnosis was done in the field using a rapid Statistical analyses. Data from children who had complete
diagnostic test (RDT; SD Bioline Malaria Ag P.f, Standard baseline and endline data for SR, AGP, and CRP were in-
Diagnostics, Yongin, South Korea; 05FK50). Children test- cluded in this analysis. Because the trial intervention did not
ing positive were treated with Coartem in accordance with have an impact on SR concentrations, data were pooled to-
Zambian national guidelines.24 Hemoglobin was assessed using a gether from both the treatment and control arms of the trial.
Hemocue Hb 201 + hemoglobinometer (Angelholm, Sweden). We conducted sensitivity analyses to further test whether
In addition, thick and thin venous blood films were prepared outcomes of interest differed by treatment allocation.
using ∼2 and 10 μL of blood, respectively. Whole blood was
transported in cooler boxes containing ice packs to the field Exploratory analytic techniques including scatter plots, box
laboratory for processing. Samples were centrifuged for plots, and kernel density plots were used to examine the in-
10 minutes under dim light. Serum was aliquoted into pre-labeled dividual associations between retinol and CRP or AGP. We
explored evidence of linear or nonlinear associations using
locally weighted scatter plot smoothing techniques. Skewed
distributions, such as CRP, were log-normalized. When the

336 BARFFOUR AND OTHERS

visual displays showed evidence of a potential change in slope groups. In this CRP-malaria model, the following six groups
of the association between retinol and either of these two were defined: CRP < 5 mg/L without malaria, CRP < 5 mg/L
APPs, spline models were constructed to test the statistical with malaria, CRP < 5–15 mg/L without malaria, CRP < 5–
significance of the slope change. Subsequently, added vari- 15 mg/L with malaria, CRP > 15 mg/L without malaria, and
able plots were constructed to examine the association CRP < 15 mg/L with malaria. We also reestimated the VAD
between retinol and CRP or AGP adjusted for the other prevalence adjusted for both CRP and malaria using the same
biomarker. The added variable plot is a visual display of an procedures as described previously. Statistical significance
adjusted regression model, showing the association between was set at P < 0.05 except for interaction coefficient, in which
an outcome variable (Y ) and an explanatory variable (X1), case significance was defined as P < 0.1. All analyses were
controlling for interference from one or multiple covariates conducted with STATA 13 software (StataCorp, College Sta-
(X2, . . .). For instance, to examine the association between SR tion, TX).
and AGP and controlling for CRP, the slope of the added
variable plot, which is a plot of the Y-residual (Retinol − β0 − RESULTS
β2CRP) against the X-residuals (AGP − β0 − β2CRP) represents
an approximation of the slope of the regression line Retinol − Data from 886 children who had complete baseline and
β0 − β2CRP = β1AGP. Separate exploratory analyses were endline data for SR, AGP, and CRP were included in this
performed for the data collected in the low and high malaria analysis. The baseline characteristics of study participants are
seasons. presented in Table 1. Our study population, which included
children aged 68 months on average, is characteristic of a
Based on the exploratory analyses, two regression models, rural, malnourished population with a significant public health
one with AGP (dichotomized) as the only explanatory variable problem of both stunting (28%) and anemia (34%). At base-
(AGP-only model) and another with CRP (three-groups) as the line, nearly one-third of children reported fever in the past
only explanatory variable (CRP-only model) were constructed. 2 weeks, indicating a high prevalence of infection. The sub-
The CRP-only model was informed by our exploratory analy- group of children included in this analysis were statistically
ses, which showed a change in slope of the retinol–CRP curve similar to the general population of study participants with
at CRP concentration > 15 mg/L (P of interaction < 0.1). To respect to age, sex, and nutritional status.
compare these two models with the Thurnham et al.17 def-
inition of inflammation, we constructed a third model, the The distributions of SR, inflammation and malaria, in the low
reference model, using both AGP and CRP as described and high malaria seasons, are presented in Table 2. A change
previously. To estimate adjusted VAD prevalence, we first in malaria prevalence, from 22% in the low transmission
computed model-specific, adjusted retinol concentrations for season to 51% in the high transmission, was associated with a
each child. We estimated group-specific adjustment factors, corresponding increase in inflammation ( Table 2), whether
defined as the difference in mean SR concentrations com-
paring each inflammatory group within the model to the TABLE 1
respective noninflammatory group. For instance, in the AGP-
only model, adjustment factors were obtained by subtracting Baseline socio-demographic characteristics, nutritional status, and
the mean SR concentration of the inflammation group (AGP > morbidity history of study participants
1 g/L) from the mean of the reference group (AGP £ 1 g/L). The
same approach was applied in estimating the adjustment Description N Value
factors for the moderate and high CRP groups (using a refer-
ence of CRP, 5 mg/mL) and for the incubation, early- and late- Child characteristics 886 68.2 ± 14.9
convalescence groups (using children with normal CRP and Age, months 886 310 (35.0)
normal AGP as the reference). The subtraction of means (in- Age less than 60 months 886 440 (49.7)
stead of ratios) approach was used because the retinol data Female (%)
were normally distributed. To generate adjusted retinol con- 884 299 (33.8)
centrations for the inflammation group, the adjustment factor Nutritional status 877 69 (7.9)
was added to the measured concentration for each individual Anemia (%) 863 17.6 ± 3.2
within the group. We subsequently estimated adjusted or Iron deficiency (%) 862
unadjusted VAD using the adjusted or unadjusted retinol Weight, kg* 862 107.4 ± 9.3
concentrations, respectively. This same approach was ap- Height, cm* 862 248 (28.0)
plied to the three-group CRP model, and the four-group ref- Stunted† 111 (12.5)
erence model in estimating the respective model-specific Underweight† 873
adjusted VAD. For each model, we compared difference in the 883 251 (28.8)
VAD estimated between the low and high malaria transmission Morbidity history 875 7 (0.8)
season. Differences in the VAD estimates, whether unadjusted Fever in past 2 weeks (%) 874
or adjusted for inflammation using the three models were Axillary temperature > 37.5°C 499 (57.0)
tested using the using McNemar’s χ2 test. In addition, we also Cough in past 2 weeks (%) 852 51 (5.8)
estimated the variance in SR concentrations explained by Diarrhea in past 2 weeks (%) 874
each of the models using regression techniques. Finally, to 874 717 (82.2)
explore the potential impact of malaria on retinol concentra- Household characteristics 42 (4.8)
tion or VAD estimates, we included a malaria status indicator in Literate household head (%) – –
the CRP-only (the optimal) model, and subsequently reesti- Household with electricity (%) –
mated the concentration of SR in the different inflammation Occupation of household head (%) – 235 (26.9)
Farming/Farm labor 241 (27.6)
Self-used 159 (18.2)
Salaried worker

* Arithmetic mean ± standard deviation.

† Stunting and underweight defined as height-for-age and weight-for-age, respec-
tively, < −2 standard deviations of the WHO Growth Reference (WHO, 2006).39,40 Analyses
restricted to children who had complete data for serum retinol, α1-acid glycoprotein (AGP) and
C-reactive protein (CRP) at both baseline and endline (N = 886). We defined iron deficiency as
ferritin < 12 μg/L in children < 5 years and ferritin < 15 μg/L in older children.25 Anemia was
defined as hemoglobin < 110 g/L for children < 60 months and < 115 g/L in older children.25
We defined literacy as the ability to read or write in English. Fever was defined as axillary
temperature > 37.5°C.

OPTIMAL MODELING OF INFLAMMATION-INDUCED HYPORETINOLEMIA 337

TABLE 2 explained the highest percentage of the variance in retinol,
accounting for about 11% and 15% of the variation in retinol in
Distributions of retinol, malaria, and inflammatory indicators in the low and high malaria seasons, respectively. The AGP-only
Zambian children aged 4–8 years in low and high malaria trans- model explained the lowest variance in retinol, accounting
mission seasons (N = 744) for < 2% of the variation in retinol concentrations in both
seasons.
Low malaria season High malaria season
In both seasons, the adjusted VAD estimates were signifi-
Indicator (September 2012) (March 2013) P value cantly lower than the unadjusted estimate, regardless of the
model used ( Table 4). Adjustment for inflammation produced
Retinol, μmol/L 1.01 ± 0.28 1.00 ± 0.32 0.61 different effects on the VAD estimates, depending on the
Retinol < 0.7 μmol/L (%) 95 (10.7) 146 (16.5) < 0.001 malaria season and the adjustment model used. The adjusted
451 (51.0) < 0.001 VAD estimated by the CRP-only model was similar (6%) in
Malaria (%) 177 (21.2) both the low and high malaria seasons. Adjusted VAD esti-
Inflammation 1.71 ± 1.01 < 0.001 mated with the AGP and reference models differed signifi-
1.14 ± 0.87 655 (73.9) < 0.001 cantly between the low and high malaria seasons ( Table 4). Of
AGP (g/L) 388 (43.8) 2.00 ± 8.67 < 0.001 the roughly 11% of children who were classified as VAD using
AGP > 1.0 g/L (%) 0.79 ± 5.66 292 (33.0) < 0.001 unadjusted retinol concentration in the low malaria season, a
CRP, mg/L* 146 (16.5) significant proportion were subsequently determined to have
CRP > 5.0 mg/L (%) adequate SR concentrations following adjustment for in-
flammation with the AGP-only (25%), CRP-only (42%), and
AGP = α1-acid glycoprotein; CRP = C-reactive protein. reference model (47%). In the high malaria season, where the
* Mean (geometric for CRP and arithmetic for AGP and retinol) ± standard deviation, unless unadjusted VAD was 17% ( Table 4), the proportion of children
otherwise specified. Statistical test of difference between malaria seasons done with paired later determined to have adequate SR concentrations after
t test for continuous variables and McNemar’s Test for binary variables. adjustment were 62% for the CRP-only model, 22% for the
AGP-only model, and 45% for the reference model. The re-
based on AGP (44–74%) or CRP (17–33%). Although mean sults of the sensitivity analyses indicated that the associations
retinol concentration in the low and high transmission seasons of interest were not affected by the interventions delivered by
was similar (∼1.0 μmol/L), the unadjusted VAD prevalence the parent study ( Table 4).
increased from 11% in the low malaria season to 17% in the
high malaria season. The inclusion of malaria to the CRP-only model produced
inconsistent effects on the retinol concentrations and ad-
Figure 1 shows the scatter plots of the associations be- justment factors for the inflammatory groups ( Table 5). In the
tween SR concentrations and either AGP or the CRP (log low malaria season, the inclusion of malaria increased the
transformed), in the low and high malaria seasons. Overall, adjustment factor for the moderate CRP from 0.18 to
retinol concentrations declined with rising CRP concentra- 0.29 μmol/L, but decreased the adjustment factor for the high
tions, with an average reduction of about 0.06 μmol/L in retinol CRP group from 0.34 to 0.27 μmol/L. The reversed trend was
for every 10 mg/L increase in CRP in both seasons (P < 0.001). observed in the high malaria seasons, where an increase in
In both seasons, the CRP-associated decline began at con- adjustment was observed in the high but not the moderate
centrations lower than the 5 mg/L threshold commonly used CRP group. Among malaria-positive children who had normal
to define inflammation. In addition, an apparent change in levels of CRP (representation 16% of all children in the low
slope of the retinol–CRP curve was observed in the high malaria season, and 25% in the high malaria seasons) retinol
malaria season, increasing by about 5-fold from a rate of about levels were reduced by ∼0.14 μmol/L in the low malaria sea-
0.04 μmol/L for CRP values £ 15 mg/L to an average reduction son, but increased by 0.05 μmol/L in the high malaria season
of about 0.2 μmol/L for CRP values beyond 15 mg/L for every ( Table 5). Although over 50% of malaria cases were associ-
10 mg/L increase in CRP (P for interaction < 0.1). Similarly, ated with elevated CRP (moderate or high) in the high malaria
higher AGP concentrations were associated with lower retinol season, only 25% of malaria positives had elevated AGP in the
values, although unlike CRP, the AGP-associated decline was low malaria season. The additional adjustment for malaria
only apparent for concentration beyond the conventional reduced VAD prevalence from 6.3% (adjusted for CRP only) to
threshold for defining inflammation (1.0 g/L). After adjustment about 3.6% (P < 0.01) in the low malaria season, with no sig-
for CRP, however, we found no such negative correlation nificant effect in the high malaria season (Figure 3).
between AGP and retinol (Figure 2). Rather, we observed in the
high malaria season that retinol concentrations appear to in- DISCUSSION
crease with rising AGP values. With CRP, however, the neg-
ative correlation with retinol persisted even after adjusting for The last decade has seen a sustained global interest and
AGP (Figure 2). increased research efforts toward characterizing the nature
and effects of IIH.26 Yet, there is currently no consensus on the
Table 3 shows the mean retinol concentrations in appar- optimal model for quantifying the potential bias in VAD prev-
ently normal or inflammation groups as defined by the three alence estimates imposed by inflammation. We asserted that
models namely AGP-only, CRP-only, and the reference an optimal model should 1) explain a greater variance in retinol
model. In the AGP-only or CRP-only models, the mean retinol when the intensity of the inflammation is increased and
concentration in any of the inflammatory categories was lower 2) produce comparable estimates of VAD regardless of the
than the concentration in the respective normal group. Retinol intensity of the inflammation. Here, we evaluated the com-
adjustment factors for the various inflammatory groups parability of three models: an AGP-only model, a CRP-only
ranged from a low of 0.05 μmol/L for the AGP-only model to a
high of 0.31 μmol/L for the CRP-only model. In the reference
model, retinol concentration in the incubation and early con-
valescence groups (but not the late convalescence group)
were consistently lower than the normal group, with adjust-
ment factors ranging from 0.10 to 0.26 μmol/L. In the late
convalescence group, however, retinol concentrations were
slightly higher than the normal population during the high
malaria season. In both seasons, the CRP-only model

338 BARFFOUR AND OTHERS

FIGURE 1. Scatter plots of serum retinol (SR) against α1-acid glycoprotein (AGP) or C-reactive protein (CRP) during low and high malaria
transmission seasons among Zambian children (N = 886). The left panels show the scatter plots of the associations between SR concentrations and
AGP among rural Zambian children in the low (top panel) and high (bottom panel) malaria seasons, respectively. The right panels show the scatter
plots of the associations between SR concentration and log-normalized CRP in the low (top panel) and high (bottom panel) malaria seasons,
respectively change in slope at CRP concentration of ∼15 mg/L (P < 0.1).

model, and a reference model combining both AGP and CRP. than having to measure both AGP and CRP. In addition, this
We considered the variance in retinol explained and the ad-
justed VAD prevalence estimated in the low and high malaria model, compared with the recently proposed regression ap-
seasons. Based on the aforementioned criteria, we report that
in this population of rural Zambia children, the CRP-only proach, offers a relatively less complex approach to capture
model is the most appropriate for addressing inflammation
induced hyporetinolemia. Furthermore, we observed that the dose-dependent association between SR and the in-
AGP, whether used alone or in combination with CRP, is not tensity of inflammation. The CRP-only model also has bi-
ideal, and may in fact, be inappropriate for correcting IIH in ological plausibility. Evidence from animal models suggests
regions where malaria is endemic.
that levels of mRNA for both RBP and transthyretin drop
Of the three models evaluated, only the CRP model met the rapidly within the first 36 hours after induction of in-
two pre-specified criteria for the optimal modeling of IIH. flammation,19 at about the same time that CRP is expected
These findings suggest that IIH is largely explained by ele- to rise.20,27 The proposed three-group classification system
vated CRP. In practice, this novel model would be less costly used here, although unconventional, was informed by the

observed association between retinol and CRP in this pop-

ulation. In fact, had we adopted the conventional dichotomous
definition for inflammation (i.e., cut-off of 5 or 10 mg/L), the

OPTIMAL MODELING OF INFLAMMATION-INDUCED HYPORETINOLEMIA 339

FIGURE 2. Added variable plots showing the adjusted associations between serum retinol (SR) and α1-acid glycoprotein (AGP) or C-reactive
protein (CRP) in the low and high malaria season among rural Zambian children (N = 886). The added variable plot (also known as partial regression

plot) depicts the association between an outcomes variable (Y) and an explanatory variable (X1), controlling for interference from another ex-
planatory variable (X2). The slope of an added variable plot, which is a plot of Y-residual (Y − β0 − β2X2) against the X-residuals (X − β0 − β2X2),
approximates the slope of the regression line Y − β0 − β2X2 = β1X1. The left panels show the retinol residuals from the regression of SR concentration
on AGP and adjusted for CRP, in the low (top panel) and high (bottom panel) malaria seasons, respectively. In the left panel, the y axis is the retinol

residuals for the regression of retinol against CRP where the x axis is the AGP residuals for the regression of AGP against CRP. The right panels show

the residuals of the regression of SR concentration on CRP and adjusted for AGP, in the low (top panel) and high (bottom panel) malaria seasons,

respectively. In the right panel, the y axis is the retinol residuals for the regression of retinol against AGP where the x axis is the CRP residuals for the

regression of CRP against AGP. The low and high malaria seasons represent the periods of low malaria prevalence (September 2012) and high
malaria prevalence (March 2013), respectively. β = adjusted regression coefficient; P = statistical significance of the regression coefficient.

CRP-only model would still have explained a proportion of the observed that after adjusting for CRP, there was no associa-
variance in retinol greater than or at least comparable to that tion between SR and AGP assessed in the low malaria season.
explained by the combined AGP-CRP reference model in both Counterintuitively, SR concentrations assessed in the high
the low malaria season (9–10%) and the high malaria season malaria season appeared to increase with rising AGP con-
(11–14%). Additional research is needed to understand centrations. Similarly, the reference model showed that retinol
whether the observed differences in the strength of associa- concentrations increased (rather than decreased) in children
tion between CRP or AGP and retinol is unique to a malaria who had elevated AGP but normal CRP. This unexpected
endemic setting. observation is consistent with findings reported by Wessells
et al.18 among Burkinabe children, where concentrations of
Consistent with findings elsewhere,28,29 the univariate RBP was about 18% higher in children in the late convalescent
AGP-only model showed a decline in SR concentrations with stage relative to an apparently healthy group. The application
rising AGP values, in both malaria seasons. However, this of adjustment factors to this stage of inflammation, although
model explained only 2% of the variance in retinol in both statistically valid, would require the biologically implausible
malaria seasons despite a substantial increase in the pro- assumptions that the inflammation triggered an increase in
portion of children with inflammation (from 44% to 74%) SR. Considering that the mean retinol concentrations in the
across the two seasons. Perhaps of greater relevance, we also

340 BARFFOUR AND OTHERS

TABLE 3

Serum retinol concentrations and adjustment factors for model-specific inflammation groups and variance in retinol explained by three models for
correcting inflammation-induced hyporetinolemia in the low and high malaria seasons

Inflammation group N (%) Mean (μmol/L) AF (μmol/L)a Variance (%)

Low malaria season – – – 1.9
498 (56.2) 1.04 (1.02, 1.06) – –
AGP-only model 388 (43.8) 0.96 (0.94, 0.99) −0.08 (−0.04, −0.11)*** –
Normal (AGP £ 1 g/L) – 10.6
Inflammation (AGP > 1 g/L) – – – –
746 (83.6) 1.04 (1.03, 1.06) −0.20 (−0.14, −0.26)*** –
CRP-only model 0.84 (0.80, 0.89) −0.29 (−0.22, −0.36)*** –
Normal (CRP £ 5 mg/L) 87 (9.8) 0.75 (0.69, 0.82) – 10.5
Moderate (CRP = 5.1–15 mg/L) 59 (6.6) – –
High (CRP > 15 mg/L) – −0.24 (−0.16, −0.32)*** –
– 1.06 (1.04, 1.08) −0.26 (−0.20, −0.31)*** –
Reference model 458 (51.3) 0.82 (0.74, 0.90)
Normal (AGP £ 1 g/L; CRP £ 5 mg/L) 0.81 (0.76, 0.85) −0.04 (0.00, −0.08) –
Incubation (AGP £ 1 g/L; CRP > 5 mg/L) 42 (4.7)
Early convalescence (AGP > 1 g/L; 104 (11.7) 1.02 (0.99, 1.06) – 0.52
– –
CRP > 5 mg/L) 288 (32.3) – −0.05 (−0.01, −0.10)* –
1.04 (1.01, 1.07) – 15.0
Late convalescence (AGP > 1 g/L; – 0.99 (0.96, 1.01) – –
CRP £ 5 mg/L) 231 (26.1) −0.10 (−0.04, −0.16)** –
655 (73.9) – −0.31 (−0.26, −0.36)*** –
High malaria season 1.08 (1.06, 1.10) – 11.9
– 0.98 (0.92, 1.04) – –
AGP-only model 599 (67.1) 0.76 (0.71, 0.81) −0.10 (−0.27, 0.07) –
Normal (AGP £ 1 g/L) 113 (12.7) −0.20 (−0.15, −0.26)*** –
Inflammation (AGP > 1 g/L) 180 (20.2) –
1.05 (1.01, 1.08) 0.05 (0.00, 0.10) –
CRP-only model – 0.94 (0.75, 1.14)
Normal (CRP £ 5 mg/L) 218 (24.6) 0.84 (0.80, 0.88)
Moderate (CRP = 5.1–15 mg/L)
High (CRP > 15 mg/L) 13 (1.5) 1.10 (1.07, 1.13)
279 (31.5)
Reference model
Normal (AGP £ 1 g/L; CRP £ 5 mg/L) 376 (42.4)
Incubation (AGP £ 1 g/L; CRP > 5 mg/L)
Early convalescence (AGP > 1 g/L;

CRP > 5 mg/L)

Late convalescence (AGP > 1 g/L;
CRP £ 5 mg/L)

AGP = α1-acid glycoprotein; CRP = C-reactive protein.
a Adjustment factor for correcting inflammation induced-hyporetinolemia. *P < 0.05; **P < 0.01 ***P < 0.001; variance in serum retinol concentrations explained by the model. For the AGP-only
model, inflammation was defined as AGP > 1 g/L; in CRP-only model, inflammation was defined as moderate (5–15 mg/L) or high (> 15 mg/L). In the reference model (proposed by Thurnham et al),17 normal is
defined as normal AGP (£ 1 g/L) and normal CRP (< 5 mg/L); Incubation defined as elevated CRP with normal AGP; Early convalescence defined as elevated CRP with elevated AGP; late convalescence = normal
CRP with elevated AGP. The low and high malaria seasons represents the periods of low malaria prevalence (September 2012) and high malaria prevalence (March 2013) respectively.

late convalescence stage—1.02 μmol/L in the low malaria seasons. In the low malaria season however, only 25% of
season and 1.10 μmol/L in the high malaria season—falls within malaria cases were associated with elevated CRP, and in-
the normal physiologic range of SR concentrations,30 our terestingly, the mean retinol concentration was significantly
findings likely suggests that the rise in AGP may have co- reduced in all malaria-positive cases, with or without a con-
incided with the homeostatic mobilization of retinol from he- current elevation in CRP. The immune response to malaria
patic stores in the aftermath of an inflammatory episode.22
Adjusting for retinol in such circumstances would be un- TABLE 4
necessary and inappropriate.
Changes in prevalence of vitamin A deficiency after corrections for C-
There have been suggestions in the literature that the ad- reactive protein and/or α-1-acid glycoprotein in the low and high
ditional assessment of malaria may improve the interpretation malaria season
of SR and other nutritional biomarkers. Our findings suggest a
need for additional adjustment for malaria in the low trans- Vitamin A deficiency (%)
mission seasons, but not the high transmission season. The
additional adjustment for malaria in the high malaria season Adjustment model/season* All (N = 886) Pro-vitamin A† Controls†
did not produce a significant change in the VAD estimate (N = 389) (N = 497)
beyond the effect produced by adjusting for CRP alone likely
because most malaria cases (52%) were associated with Low malaria season, n (%) 95 (10.7) 42 (10.8) 53 (10.7)
elevated CRP in the high malaria season. Additionally, and Unadjusted 54 (6.1) 26 (6.7) 28 (5.6)
perhaps more importantly, malaria was associated with a re- CRP-only 73 (8.2) 30 (7.1) 43 (8.7)
duction in retinol only when CRP was concurrently elevated. AGP-only 49 (5.5) 21 (5.4) 28 (5.6)
The acute phase response in malaria is largely an innate Reference
mechanism, involving the activation of both CRP-dependent 146 (16.6)‡ 63 (16.2)‡ 83 (16.7)‡
and non CRP-dependent pathways.31 It is plausible that when High malaria season, n (%) 55 (6.3) 23 (5.9) 32 (6.4)
transmission is intense, the CRP-dependent mechanism as- Unadjusted 52 (13.4)‡ 64 (12.9)‡
sumes a more dominant role.31 The observed disparity in the CRP-only 116 (13.1)‡ 33 (8.5)‡ 48 (9.7)‡
association between malaria and CRP across the two malaria AGP-only 81 (9.1)‡
seasons may also be explained by the difference in the dura- Reference
tion of the acute phase response in the two transmission
AGP = α1-acid glycoprotein; CRP = C-reactive protein. The numbers represent frequency
and proportion of children with vitamin A deficiency. AGP-only model adjusted for AGP; CRP-

only model adjusted for CRP; Reference model adjusted for both AGP and CRP as proposed
by Thurnham et al.17 Children in the Pro-vitamin A group received a daily β-carotene
biofortified maize meal for 6 months. Control groups did not receive a β-carotene intervention.

* All adjusted estimates significantly lower than corresponding unadjusted estimates in

both seasons.

† Estimates were statistically similar between Provitamin A and control groups.
‡ Significantly higher than corresponding estimate in low malaria seasons (P < 0.01).

OPTIMAL MODELING OF INFLAMMATION-INDUCED HYPORETINOLEMIA 341

TABLE 5

Serum retinol concentrations among inflammation groups defined by C-reactive protein and malaria and estimated adjustment factors in low and
high malaria transmission seasons

Inflammation groupa N (%) Mean retinol (μmol/L) AF (μmol/L)b

Low malaria season (N = 836) 563 (67.3) 1.07 (1.05, 1.09) 0.00
Normal CRP-no malaria 58 (6.9) 0.89 (0.83, 0.95)*** −0.18 (−0.11, −0.25)***
Moderate CRP-no malaria 38 (4.6) 0.74 (0.67, 0.80)*** −0.34 (−0.25, −0.42)***
High CRP-No malaria 0.94 (0.89, 0.98)*** −0.14 (−0.09, −0.19)***
Normal CRP-positive malaria 132 (15.8) 0.79 (0.71, 0.86)*** −0.29 (−0.18, −0.39)***
Moderate CRP-positive malaria 25 (3.0) 0.80 (0.65, 0.94)*** −0.27 (−0.16, −0.39)***
High CRP-positive malaria 20 (2.4)
1.06 (1.02, 1.08) 0.00
High malaria season (N = 885) 374 (42.3) 0.97 (0.86, 1.07) −0.09 (−0.19, 0.00)
Normal CRP-no malaria 40 (4.5) 0.88 (0.76, 1.00) −0.18 (−0.04, −0.31)*
Moderate CRP-no Malaria 20 (2.3) 1.11 (1.07, 1.14)
High CRP-no malaria 0.99 (0.92, 1.07) 0.05 (0.01, 0.10)*
Normal CRP-positive malaria 220 (24.9) 0.75 (0.70, 0.80) −0.07 (−0.14, 0.01)
Moderate CRP-positive malaria 71 (8.0) −0.31 (−0.25, −0.36)***
High CRP-positive malaria
160 (18.1)

AGP = α1-acid glycoprotein; CRP = C-reactive protein; RDT = rapid diagnostic test.
a CRP defined as normal (< 5 mg/L), moderate (5–15 mg/L) or high (> 15 mg/L). Positive malaria defined as positive microscopy and/or RDT. No malaria defined as negative microscopy and

negative RDT.
b AF = Adjustment factor for correcting inflammation induced-hyporetinolemia; N (%) represent the total number (proportion) of children in each malaria-CRP category *P < 0.05; **P < 0.01

***P < 0.001; The low and high malaria seasons represents the periods of low malaria prevalence (September 2012) and high malaria prevalence (March 2013) respectively.

involves the initial activations of Th1-type pro-inflammatory shown in this population, the associations between retinol and
pathways for the clearance of parasites, and then later, the the APPs are not always linear. A related problem is the po-
activation of Th2-type anti-inflammatory mechanisms for tential to over-adjust by applying the same regression
continuing parasite clearance via the adaptive immune coefficient to every individual. To mitigate the problem of over-
mechanisms and downregulation of inflammation.32–34 It is adjustment, Larson et al suggested that adjustment factors be
plausible that that initial Th1 phase, which involves CRP, is applied to individuals at or above the first decile of the log-
relatively short-lived in the low malaria season, such that CRP normalized AGP and CRP concentrations, as opposed to
levels return to normal, before parasite clearance is complete. every individual.16 Had we adopted this approach, we would
In such conditions, the assessment of CRP alone, as opposed have adjusted the retinol concentrations in over 90% of our
CRP and malaria, may not suffice in characterizing the study population. Such a procedure, which would imply that
malaria-related inflammatory changes in retinol. nearly all children have some degree of IIH, may be unjustifi-
able, even in regions with a high burden of infection. Our data
A multivariate, linear regression approach for adjusting IIH suggest that the issue of over-adjustment may best be
has recently been proposed.16,35 The advantage of this ap- addressed by using context-specific thresholds, as opposed
proach is that it estimates context-specific adjustment to universally adopting the AGP and CRP deciles proposed
factors, based on observed associations between specific elsewhere.16,35
nutritional biomarkers and the acute phase reactants, as op-
posed to using predefined cut-offs for inflammation. However, A potential limitation of this study is that data were pooled
this approach has several limitations. First, the universal as- from across clusters of children, some of whom received a
sumption of linearity is inconsistent with the goal of generat- provitamin A carotenoid intervention for 6 months. Although
ing population-specific, data-driven adjustment factors. As the evidence from animal models suggests that a retinol in-
tervention may regulate the inflammatory response,35,36 there
FIGURE 3. Changes in vitamin A deficiency (VAD) estimates after was no intervention effect on either retinol or inflammation in
correction for C-reactive protein (CRP) alone, or with malaria, in the this population. Our sensitivity analyses confirmed that the
low and high malaria seasons. The bars show the prevalence of VAD intervention did not affect the outcomes of interest in this
either unadjusted (black), adjusted for CRP alone (gray) or adjusted for analysis. Another limitation, which is pertinent to all current
both CRP and malaria (white) in the low and high malaria seasons. A = adjustment procedures, is the assumption that the observed
significantly from unadjusted VAD (P < 0.01); B = significantly different reductions in retinol are completely transient and that retinol
from VAD adjusted for CRP-alone (P < 0.01). concentrations would return back to normal levels when the
inflammation resolves. Unlike other nutritional biomarkers,
such as ferritin for instance, which are mainly sequestered, as
opposed to excreted, some retinol is lost via increased urinary
output during inflammation. Hence in the case of retinol, the
resolution of inflammation may not completely correct the IIH.
Consequently, the tendency to overadjust is especially great
in the case of retinol. Until models are developed to appro-
priately account for the inflammation-induced urinary losses,
retinol adjustment factors must be interpreted cautiously. A
major strength of this study is the availability of multiple survey
data, which enabled the internal validation of these competing
models between two time points with varying burden of
inflammation.

342 BARFFOUR AND OTHERS

In conclusion, the appropriate assessment of population Review of Nutrition and Dietetics, Vol. 103. Basel, Switzerland:
vitamin A status is of global interest, not only because VAD Karger Publishing, 52–64.
remains a public health problem in several low-income 7. Abraham K, Muller C, Gruters A, Wahn U, Schweigert FJ, 2003.
countries, but also because of the need to estimate the im- Minimal inflammation, acute phase response and avoidance of
pact of available VAD control programs. Our findings highlight misclassification of vitamin A and iron status in infants—importance
an important criteria which can be incorporated into current of a high-sensitivity C-reactive protein (CRP) assay. Int J Vitam
models for characterizing IIH. It is important that the APPs for Nutr Res 73: 423–430.
characterizing inflammation in similar settings be judged on 8. Baeten JM et al., 2002. Vitamin A deficiency and the acute phase
the basis of their biological associations with retinol. We response among HIV-1-infected and -uninfected women in
propose that models for characterizing inflammation take Kenya. J Acquir Immune Defic Syndr 31: 243–249.
into consideration the variance in retinol explained by spe- 9. Filteau SM, Tomkins AM, 1994. Micronutrients and tropical in-
cific APPs and the consistency in adjusted VAD estimates fections. Trans R Soc Trop Med Hyg 88: 1–3, 26.
over varying intensities of infection or inflammation. Our 10. Kongsbak K, Wahed MA, Friis H, Thilsted SH, 2006. Acute-phase
findings also highlight areas for additional research. There is
a need to demonstrate the appropriateness of using CRP and protein levels, diarrhoea, Trichuris trichiura and maternal edu-
AGP, either alone or in combination, for quantifying IIH in
other settings. Furthermore, considering that vitamin A cation are predictors of serum retinol: a cross-sectional study of
modulates the inflammatory response to infections,37 in- children in a Dhaka slum, Bangladesh. Br J Nutr 96: 725–734.
cluding malaria,38 there is a need for research into models 11. Gieng SH, Green MH, Green JB, Rosales FJ, 2007. Model-based
that take into consideration the potential influence of base-
line vitamin A status. compartmental analysis indicates a reduced mobilization of
hepatic vitamin A during inflammation in rats. J Lipid Res 48:
Received February 18, 2017. Accepted for publication September 19, 904–913.
2017. 12. Gieng SH, Rosales FJ, 2006. Plasma alpha1-acid glycoprotein can
be used to adjust inflammation-induced hyporetinolemia in vi-
Published online November 20, 2017. tamin A-sufficient, but not vitamin A-deficient or -supplemented
rats. J Nutr 136: 1904–1909.
Acknowledgments: We thank the participating children, their families, 13. Semba RD, Muhilal , West KP Jr, Natadisastra G, Eisinger W, Lan
and Mkushi District officials for supporting the study’s implementa-
tion. We are also grateful to Dr. Mwanza and the Mkushi District Y, Sommer A, 2000. Hyporetinolemia and acute phase proteins
Medical Office for providing bed nets and Coartem to support malaria
prevention and control activities. We thank Bess Lewis and Lauren in children with and without xerophthalmia. Am J Clin Nutr 72:
Tanz for supporting data collection, Brian Dyer, Mitra Maithilee, and 146–153.
Lee Wu for their support in data management and Dr. Douglas Norris 14. Stephensen CB, 2000. When does hyporetinolemia mean vitamin
for providing inputs to this paper. A deficiency? Am J Clin Nutr 72: 1–2.
15. Thurnham DI, McCabe GP, 2012. Influence of Infections and In-
Financial support: This work was funded by HarvestPlus Challenge flammation on Biomarkers of Nutritional Status with Emphasis
Grant #8251, with support from the UK Department for International on Vitamin A and Iron. Report: Priorities in the assessment of
Development. The views expressed do not necessarily reflect those of
HarvestPlus. M. A. B. received partial support from the DSM Scholars vitaminn A and iron status in populations. Panama, September
Program through the Sight & Life Global Nutrition Research Institute at 15–17, 2010. Geneva, Switzerland: World Health Organization.
Johns Hopkins University and from Foreign Affairs, Trade and De- 16. Larson LM, Addo OY, Sandalinas F, Faigao K, Kupka R, Flores-
velopment Canada Grant #112305. Ayala R, Suchdev PS, 2017. Accounting for the influence of
inflammation on retinol-binding protein in a population survey
Authors’ addresses: Maxwell A. Barffour, Kerry J. Schulze, Christian L. of Liberian preschool-age children. Matern Child Nutr 13:
Coles, Margia Arguello, William J. Moss, Keith P. West, Jr., and
Amanda C. Palmer, Johns Hopkins Bloomberg School of Public e12298.
Health, Baltimore, MD, E-mails: [email protected], [email protected], 17. Thurnham DI, McCabe GP, Northrop-Clewes CA, Nestel P, 2003.
[email protected], [email protected], [email protected], kwest1@
jhu.edu, and [email protected]. Justin Chileshe and Ng’andwe Effects of subclinical infection on plasma retinol concentrations
Kalungwana, Tropical Disease Research Centre, Ndola, Zambia, and assessment of prevalence of vitamin A deficiency: meta-
E-mails: [email protected] and [email protected]. Ward analysis. Lancet 362: 2052–2058.
Siamusantu, National Food and Nutrition Commission, Lusaka, 18. Wessells KR, Hess SY, Ouedraogo ZP, Rouamba N, Ouedraogo
Zambia, E-mail: [email protected].
JB, Brown KH, 2014. Asymptomatic malaria infection affects
REFERENCES
the interpretation of biomarkers of iron and vitamin A status,
1. Sommer A, 1994. Vitamin A: its effect on childhood sight and life. even after adjusting for systemic inflammation, but does not
Nutr Rev 52: S60–S66. affect plasma zinc concentrations among young children in
Burkina Faso. J Nutr 144: 2050–2058.
2. WHO, 2008. Global Prevalence of Vitamin A Deficiency in Pop- 19. Rosales FJ, Ritter SJ, Zolfaghari R, Smith JE, Ross AC, 1996.
ulations at Risk, 1995–2005. WHO global database on vitamin A Effects of acute inflammation on plasma retinol, retinol-binding
deficiency. Geneva, Switzerland, World Health Organization. protein, and its mRNA in the liver and kidneys of vitamin
A-sufficient rats. J Lipid Res 37: 962–971.
3. Mahalanabis D, 1991. Breast feeding and vitamin A deficiency 20. Schweigert FJ, 2001. Inflammation-induced changes in the nu-
among children attending a diarrhoea treatment centre in tritional biomarkers serum retinol and carotenoids. Curr Opin
Bangladesh: a case-control study. BMJ 303: 493–496. Clin Nutr Metab Care 4: 477–481.
21. Hautvast JL, Tolboom JJ, West CE, Kafwembe EM, Sauerwein
4. Rahmathullah L, Raj MS, Chandravathi TS, 1997. Aetiology of se-
vere vitamin A deficiency in children. Natl Med J India 10: 62–65. RW, van Staveren WA, 1998. Malaria is associated with re-

5. Miller M, Humphrey J, Johnson E, Marinda E, Brookmeyer R, Katz duced serum retinol levels in rural Zambian children. Int J Vitam
J, 2002. Why do children become vitamin A deficient? J Nutr Nutr Res 68: 384–388.
132 (Suppl): 2867S–2880S. 22. Thurnham DI, Singkamani R, 1991. The acute phase response

6. McLaren DS, Kraemer K, 2012. Assessment of vitamin A status. and vitamin A status in malaria. Trans R Soc Trop Med Hyg 85:
Manual on Vitamin A Deficiency Disorders (VADD). World 194–199.
23. Palmer AC et al., 2016. Provitamin A-biofortified maize increases
serum β-carotene, but not retinol, in marginally nourished
children: a cluster-randomized trial in rural Zambia. Am J Clin
Nutr 104: 181–190.
24. Hamer DH, Ndhlovu M, Zurovac D, Fox M, Yeboah-Antwi K,

Chanda P, Sipilinyambe N, Simon JL, Snow RW, 2007. Im-

proved diagnostic testing and malaria treatment practices in
Zambia. JAMA 297: 2227–2231.
25. WHO/CDC, 2004. Assessing the Iron Status of Populations. Re-

port of a Joint WHO/CDC Technical Consultation on the

OPTIMAL MODELING OF INFLAMMATION-INDUCED HYPORETINOLEMIA 343

Assessment of Iron Status at the Population Level. Geneva, 33. Boeuf PS et al., 2012. Insights into deregulated TNF and IL-10
Switzerland: World Health Organization. production in malaria: implications for understanding severe
26. Tanumihardjo SA, Russell RM, Stephensen CB, Gannon BM, malarial anaemia. Malar J 11: 253.
Craft NE, Haskell MJ, Lietz G, Schulze K, Raiten DJ, 2016.
Biomarkers of nutrition for development (BOND)-vitamin A re- 34. Goncalves RM, Scopel KK, Bastos MS, Ferreira MU, 2012. Cy-
view. J Nutr 146: 1816s–1848s. tokine balance in human malaria: does Plasmodium vivax elicit
27. Gabay C, Kushner I, 1999. Acute-phase proteins and other sys- more inflammatory responses than Plasmodium falciparum?
temic responses to inflammation. N Engl J Med 340: 448–454. PLoS One 7: e44394.
28. Paracha PI, Jamil A, Northrop-Clewes CA, Thurnham DI, 2000.
Interpretation of vitamin A status in apparently healthy Pakistani 35. Larson LM, Namaste SM, Williams AM, Engle-Stone R, Addo OY,
children by using markers of subclinical infection. Am J Clin Suchdev PS, Wirth JP, Temple V, Serdula M, Northrop-Clewes
Nutr 72: 1164–1169. CA, 2017. Adjusting retinol-binding protein concentrations for
29. Wieringa FT, Dijkhuizen MA, West CE, Northrop-Clewes CA, inflammation: biomarkers reflecting inflammation and nutri-
Muhilal, 2002. Estimation of the effect of the acute phase re- tional determinants of anemia (BRINDA) project. Am J Clin Nutr
sponse on indicators of micronutrient status in Indonesian in- 106 (Suppl 1): 390s–401s.
fants. J Nutr 132: 3061–3066.
30. WHO, 2011. Serum Retinol Concentrations for Determining the 36. Serghides L, Kain KC, 2001. Peroxisome proliferator-activated
Prevalence of Vitamin A Deficiency in Populations. Vitamin and receptor gamma-retinoid X receptor agonists increase CD36-
Mineral Nutrition Information System. Geneva, Switzerland: dependent phagocytosis of Plasmodium falciparum-parasitized
World Health Organization (WHO/NMH/NHD/MNM/11.3). erythrocytes and decrease malaria-induced TNF-alpha secre-
31. Lima-Junior JC, Rodrigues-da-Silva RN, Pereira VA, Storer FL, tion by monocytes/macrophages. J Immunol 166: 6742–6748.
Perce-da-Silva DS, Fabrino DL, Santos F, Banic DM, Oliveira-
Ferreira J, 2012. Cells and mediators of inflammation (C- 37. Stephensen CB, 2001. Vitamin A, infection, and immune function.
reactive protein, nitric oxide, platelets and neutrophils) in the Annu Rev Nutr 21: 167–192.
acute and convalescent phases of uncomplicated Plasmodium
vivax and Plasmodium falciparum infection. Mem Inst Oswaldo 38. Serghides L, Kain KC, 2002. Mechanism of protection induced by
Cruz 107: 1035–1041. vitamin A in falciparum malaria. Lancet 359: 1404–1406.
32. Andrade BB, Reis-Filho A, Souza-Neto SM, Clareˆ ncio J, Camargo
LM, Barral A, Barral-Netto M, 2010. Severe Plasmodium vivax 39. de Onis M, Garza C, Victora CG, Onyango AW, Frongillo EA,
malaria exhibits marked inflammatory imbalance. Malar J 9: 13. Martines J, 2004. The WHO multicentre growth reference
study: planning, study design, and methodology. Food Nutr
Bull 25 (Suppl): S15–S26.

40. Rodd C, Metzger DL, Sharma A, 2014. Extending World Health
Organization weight-for-age reference curves to older children.
BMC Pediatr 14: 32.

High provitamin A carotenoid serum concentrations, elevated retinyl Downloaded from https://academic.oup.com/ajcn/article-abstract/102/2/497/4564660 by Texas Womans University user on 15 January 2019
esters, and saturated retinol-binding protein in Zambian preschool
children are consistent with the presence of high liver vitamin A stores1

Stephanie Mondloch,2,6 Bryan M Gannon,2,6 Christopher R Davis,2 Justin Chileshe,3 Chisela Kaliwile,4 Cassim Masi,4
Luisa Rios-Avila,5 Jesse F Gregory III,5 and Sherry A Tanumihardjo2*

2Interdepartmental Graduate Program in Nutritional Sciences, University of Wisconsin-Madison, Madison, WI; 3Tropical Diseases Research Centre, Ndola,
Zambia; 4National Food and Nutrition Commission of Zambia, Lusaka, Zambia; and 5University of Florida, Gainesville, FL

ABSTRACT Keywords: a-carotene, b-carotene, plant carotenoids, pyridoxal-5#-
Background: Biomarkers of micronutrient status are needed to best phosphate, retinol, retinol isotope dilution, school-age children, Zambia
define deficiencies and excesses of essential nutrients.
Objective: We evaluated several supporting biomarkers of vitamin INTRODUCTION
A status in Zambian children to determine whether any of the bio- Vitamin A (VA)7 status assessment of humans is challenging
markers were consistent with high liver retinol stores determined by
using retinol isotope dilution (RID). because liver VA concentrations are considered the gold standard
Design: A randomized, placebo-controlled, biofortified maize effi- of VA status (1) but are difficult to assess. Serum retinol (SR)
cacy trial was conducted in 140 rural Zambian children from 4 concentrations are homeostatically controlled over a wide range
villages. A series of biomarkers were investigated to better define of liver reserves (2) and are decreased during the acute-phase
the vitamin A status of these children. In addition to the assessment response (3, 4). Retinol isotope dilution (RID) is currently con-
of total-body retinol stores (TBSs) by using RID, tests included sidered the most-sensitive indirect biomarker of VA status (1, 2)
analyses of serum carotenoids, retinyl esters, and pyridoxal-5#- and can be used to determine the response to interventions (5);
phosphate (PLP) by using high-pressure liquid chromatography, however, a quantitative estimation of total-body retinol stores
retinol-binding protein by using ELISA, and alanine aminotransfer- (TBSs) or total liver reserves requires the numerical estimation
ase (ALT) activity by using a colorimetric assay. of dose absorption, partitioning in organs, and catabolism and
Results: Children (n = 133) were analyzed quantitatively for TBSs excretion (6). Researchers have used different methodologies
by using RID. TBSs, retinyl esters, some carotenoids, and PLP and assumptions when calculating VA status by using RID (6).
differed by village site. Serum carotenoids were elevated above most Because of surprising findings that indicated adequate through
nonintervened reference values for children. a-Carotene, b-carotene, hypervitaminotic VA status of Zambian preschoolers (7), these
and lutein values were .95th percentile from children in the US assumptions have been challenged (8). Therefore, other bio-
NHANES III, and 13% of children had hypercarotenemia (defined as markers of VA status were investigated to qualify the findings of
total carotenoid concentration .3.7 mmol/L). Although only 2% of high liver reserves in this group of children. Additional evidence
children had serum retinyl esters .10% of total retinol plus retinyl of actual VA status of a group could be evaluated with valid di-
esters, 16% of children had .5% as esters, which was consistent etary and health biomarkers that can be applied at the population
with high liver retinol stores. Ratios of serum retinol to retinol-binding level to support quantitative measurements.
protein did not deviate from 1.0, which indicated full saturation.
ALT activity was low, which was likely due to underlying vitamin 1 Supported by HarvestPlus (contract number 8256) and by an endowment
B-6 deficiency, which was confirmed by very low serum PLP titled “Friday Chair for Vegetable Processing Research” (to SAT) and Global
concentrations. Health Funds at the University of Wisconsin-Madison.
Conclusions: The finding of hypervitaminosis A in Zambian chil-
dren was supported by high circulating concentrations of caroten- 6 These authors are joint first authors.
oids and mildly elevated serum retinyl esters. ALT-activity assays *To whom correspondence should be addressed. E-mail: sherry@nutrisci.
may be compromised with co-existing vitamin B-6 deficiency. Nu- wisc.edu.
trition education to improve intakes of whole grains and animal- 7 Abbreviations used: AGP, a1-acid glycoprotein; ALT, alanine aminotransfer-
source foods may enhance vitamin B-6 status in Zambians. This ase; BCO1, b-carotene 15,15’-oxygenase; CRP, C-reactive protein; RBP, retinol
trial was registered at clinicaltrials.gov as NCT01814891. Am J binding protein; RID, retinol isotope dilution; SR, serum retinol; TBS, total-body
Clin Nutr 2015;102:497–504. retinol store; TDRC, Tropical Diseases Research Centre; VA, vitamin A.
Received April 3, 2015. Accepted for publication June 10, 2015.
First published online July 15, 2015; doi: 10.3945/ajcn.115.112383.

Am J Clin Nutr 2015;102:497–504. Printed in USA. Ó 2015 American Society for Nutrition 497

498 MONDLOCH ET AL.

The distribution and concentration of carotenoids varies Inclusion criteria were as follows: 5–7-y-old children living in Downloaded from https://academic.oup.com/ajcn/article-abstract/102/2/497/4564660 by Texas Womans University user on 15 January 2019
widely in fruit and vegetables (9). Some carotenoids are pre- the study area who were considered relatively healthy (no
cursors of VA that can be bioconverted by humans and other clinical infection or fever, weight-for-age and weight-for-height
animals; the most common of these carotenoids in the human z scores greater than 23, and hemoglobin concentration .70 g/L
diet are a-carotene, b-carotene, and b-cryptoxanthin. Other at recruitment), who had received antihelminthic treatment the
carotenoids in human circulation cannot be converted into week before recruitment, and had not received a high-dose VA
VA but serve other physiologic purposes (e.g., lutein in eye supplement in the past 6 mo. Blood collection (7 mL) was
health) (10), and serum concentrations can be used to verify performed by the TDRC, followed by centrifugation at the local
specific vegetables or fruit in the diet. Plant sources of pro- clinic. Malaria parasites were counted on thick blood smears
vitamin A carotenoids are a major source of VA and have been prepared in the field as described (24). Serum was transferred
estimated at providing a substantial percentage (w68%) of into 2 tubes, transported in nitrogen gas to the TDRC, shipped
total worldwide VA (11). Serum carotenoid concentrations are on dry ice, and stored at 2808C until analysis at the University
a function of a number of factors, the most pertinent being of Wisconsin-Madison for all VA biomarkers or the University
dietary intake (5, 12–15). Other factors associated with serum of Florida for concentrations of pyridoxal-5#-phosphate (PLP),
carotenoids are age, BMI, and the genetic variation in en- which is the physiologically active form of vitamin B-6.
zymes related to carotenoid absorption, transport, cleavage,
and degradation (16). Provitamin A carotenoid bioefficacy is TBSs and liver concentrations of VA
inversely related to VA status (5) and the VA content of the
diet (17), likely mediated by the VA-induced negative feedback TBSs and liver concentrations of VA were determined by using
of carotenoid transporter scavenger receptor-B1 and cleavage 13C-RID and applying the mass balance equation with the fol-
enzyme b-carotene 15,15’-oxygenase (BCO1) by transcription lowing assumptions: 90% dose absorption, fractional catabolic
factor intestine-specific homeobox (18, 19). Plasma carotenoids rate of 0.5%/d during the mixing period, equal serum and liver
and TBSs are increased in response to consumption of high- 13C-enrichment, and 80% of TBSs in the liver (7). A decrease to
carotenoid diets (5, 12, 14). Skin and serum carotenoids reflected 80% absorption was made for children with elevated CRP at the
both a low-carotenoid regimen followed by a high-carotenoid time of dosing. After a baseline blood draw, 1 mmol 13C2-retinyl
regimen (20). acetate dissolved in soybean oil was delivered directly to each
child by using a positive-displacement pipette and immediately
Serum retinyl esters are used as a biomarker of high VA stores followed by a high-fat–containing snack to facilitate absorption.
although the limitations of this assessment in children are not After a 14-d mixing period during which subjects consumed
clear because the cutoff of 10% of the total as retinyl esters was a controlled diet with limited VA (25), a second blood sample
chosen on the basis of adults with unknown TBSs. The retinyl was taken and the 13C:total C of SR at both blood draws was
ester concentration was first suggested as a biomarker on the basis determined by using gas chromatography–combustion isotope
of 3 patients with chronic intakes of pharmaceutical doses that ratio mass spectrometry to estimate TBSs and liver retinol
caused hypervitaminosis A (21). The presentation of retinyl concentrations (7).
esters in lipoproteins to cells instead of retinol on retinol-binding
protein (RBP) is hypothesized to cause VA toxicity (21, 22). Carotenoid and retinyl ester extraction and analysis
Hypervitaminosis A leads to liver fibrosis and elevated liver
enzymes in plasma. Additional assessments in Zambian children Samples were extracted for carotenoids and retinyl esters
were performed to gain more insight into the degree of hyper- by using a modified published procedure (26). To 1 mL (or all
vitaminosis A present (7). available) serum, ethanol (1.5 3 volume) with 0.1% butylated
hydroxytoluene as an antioxidant and 100 mL C23 b-apo-
METHODS carotenol as an internal standard were added. Samples were
extracted 3 times with 1.5-mL hexanes. Pooled hexane layers
Subjects were dried under nitrogen and reconstituted in 100 mL 50:50
(volume:volume) methanol:dichloroethane. To have high sensi-
All field procedures involving children were approved by the tivity to detect some of the minor retinyl esters and to ensure
Ethics Review Committee of the Tropical Diseases Research good separation of carotenoids, aliquots of the same extract were
Centre (TDRC) in Zambia and the Health Sciences Human run on 2 separate HPLC systems.
Subjects Institutional Review Board of the University of
Wisconsin-Madison. This trial was registered at clinicaltrials.gov For carotenoid analysis, 25 mL extract was injected onto
as NCT01814891; outcomes related to the intervention have been a Waters HPLC system (Waters) comprised of a C18 Resolve
reported (7), and biomarkers reported herein are before the in- (5-mm, 3.9 3 300-mm) analytic column (Waters) equipped with
tervention except for deworming, which was performed 1 wk a guard column, 2707 autosampler, 1525 binary pump, and 2998
before the first blood sample. Written informed consent was photodiode array detector. Samples were eluted at 2 mL/min by
obtained from parents or caregivers. The trial was conducted in using 95:5 (volume:volume) acetonitrile:water (solvent A) and
2012 in the Nyimba District of the Eastern Province of Zambia in 85:10:5 (volume:volume:volume) acetonitrile:methanol:dichloroethane
preschool children (n = 143 at initial enrollment) because of (solvent B) both with 10 mmol ammonium acetate/L as a mod-
a high prevalence of low SR concentrations in a previous survey ifier by using the following gradient method: 3 min at 100% A,
(23). The following 4 sites were chosen: 2 sites adjacent to the followed by a 7-min linear gradient to 100% B, a 15-min hold at
main paved roadway (coded as sites A and B) and 2 sites w8 km 100% B, 1-min linear gradient back to 100% A, and a 5-min hold
off the paved road (coded as sites C and D).

HIGH SERUM CAROTENOID CONCENTRATIONS IN ZAMBIA 499

at 100% A for re-equilibration. Chromatograms were evaluated at was assessed by using the Shapiro-Wilk test, and the homoge- Downloaded from https://academic.oup.com/ajcn/article-abstract/102/2/497/4564660 by Texas Womans University user on 15 January 2019
450 nm by using authentic HPLC-purified standards. neity of variance was assessed by using Levene’s test. For data
that satisfied assumptions, outcomes of interest were evaluated
For the analysis of retinyl esters, 50 mL serum extract was by using 1-factor ANOVA, and differences in study sites were
injected onto a Waters C18 Resolve (5-mm, 3.9 3 300-mm) determined by using least-significant difference tests. For data
column equipped with guard column. A Waters Delta 600 binary that failed assumptions, a nonparametric analysis was carried
pump and controller (Waters), 2487 Dual-Wavelength Absorbance out on ranked data. Proportions were compared by using x2
Detector (Waters), and a CR7A Chromatopac data processor analysis. Significance was defined as P # 0.05.
(Shimadzu) comprised the HPLC system. Chromatograms were
generated at 325 nm to quantify retinol and retinyl esters, which RESULTS
were confirmed by retinyl ester standards isolated and purified
from pig liver. The mobile phase was 1.5 mL 85:15 (volume: Subject characteristics
volume) acetonitrile:water/min with 10 mmol/L ammonium
acetate (solvent A) as an initial condition followed by a 10-min Enrollment occurred at the end of May 2012. Children (n =
linear gradient to 100% 80:20 (volume:volume) acetonitrile: 143) were recruited and consented, and 140 children met
dichloroethane (solvent B). Solvent B was held for 12 min baseline inclusion criteria and had blood samples taken (7).
followed by a 2-min linear reverse gradient to 100% A and an Because of statistically significant effects of site on nutritional
8-min hold at 100% A. outcome measures, baseline anthropometric data are presented
by site (Table 1). Of particular interest was that village B did not
Other assays have any cases of asymptomatic malaria, which was related to
higher hemoglobin concentrations than in other villages (P ,
Serum C-reactive protein (CRP) (Cayman Chemical Co.), 0.0001). Although enrollment occurred during the dry season
a1-acid glycoprotein (AGP) (Abcam), and RBP (Arbor Assays), when malaria transmission is low, children with active malaria
all of which are acute-phase proteins, were assayed by using would not have been enrolled because of exclusion criteria.
enzyme immunoassay kits. Alanine aminotransferase (ALT) ac-
tivity was assayed by using a colorimetric assay kit as recom- Markers of infection status were evaluated (7). Although
mended by the manufacturer (Sigma-Aldrich) as part of a strategy elevated AGP was universal in villages, CRP differed by site and
to gauge if any hepatocellular damage had occurred from ex- was lower in villages closer to the paved road (P , 0.0001)
cessive storage of retinyl esters in the liver. Once it was shown (Table 2).
that ALT activity was actually lower than normal, an inquiry into
vitamin B-6 status, which is a cofactor for ALT function, was Total liver retinol reserves
added to the protocol. With collaboration at the University of
Florida, PLP was determined by using HPLC followed by fluo- Calculated mean liver reserves for all subjects were 1.13 6 0.41
rescence detection (27). mmol retinol/g liver, with 59% .1 mmol/g (7), which is the cur-
rent cutoff for defining hypervitaminosis A (2). No reserves were
Statistical analysis ,0.1 mmol/g, which is the deficiency cutoff (2). Village B had
statistically significantly higher TBSs than those of villages lo-
Data are reported as medians (first and third quartile values, cated farther from the main road (i.e., villages C and D) (Table 2).
which are equivalent to 25th and 75th percentiles, respectively) to
control nonnormality in some outcome measures or means 6 Serum carotenoid concentrations
SDs for in-text summaries. Data were analyzed by using the
General Linear Model procedure in the Statistical Analysis After the primary outcome analysis (7), 123 samples had
System (version 9.4; SAS Institute). The normality of residuals sufficient serum to quantify the carotenoid profile and retinyl
esters. Carotenoids were not statistically significantly affected

TABLE 1
Baseline anthropometric and screening data for Zambian children (n = 140) by village1

A (n = 29) B (n = 36) C (n = 35) D (n = 40) P

Age,2 mo 72 (65, 80)3 68.5 (66.5,76) 69 (64, 75) 73 (66.5, 78.5) 0.34
0.57
Height, cm 107 (104, 113) 108 (105, 112) 107 (103, 110) 108 (103, 112) 0.76
Weight,2 kg 0.82
16.4 (15.4, 18.7) 17.5 (15.6, 18.7) 16.8 (15.8, 18.2) 16.9 (16.0, 18.6) 0.48
Height-for-age z score 21.2 (22.1, 20.9) 21.3 (22.0, 20.7) 21.6 (22.2, 20.8) 21.5 (22.0, 21.1) 0.14
Weight-for-age z score 21.4 (21.9, 21.0) 21.0 (21.7, 20.5) 21.3 (21.8, 20.8) 21.3 (21.8, 20.8) ,0.0001
20.6 (20.9, 20.3) 20.4 (20.6, 0.1) 20.4 (20.8, 0.2) 20.4 (20.7, 0.0) ,0.0001
BMI-for-age z score
Hemoglobin,2 g/L 118 (109, 123)b 125 (116, 128)a 117 (108, 123)b 112 (102, 120)b

Positive malaria blood smear, % 17.2 0 17.2 12.5

1Villages A and B were closest to the paved road, and villages C and D were 8 km from the road. Groups with uncommon
superscript letters were different: a . b. P values are for testing the null hypothesis that each variable was equal in groups by using an

ANOVA or chi-square test.
2Nonnormally distributed residuals; P value reflects a nonparametric analysis.
3Median; first and third quartile values in parentheses (all such values). First and third quartile values are also known as 25th and

75th percentiles, respectively.

500 MONDLOCH ET AL.

TABLE 2
The following nutritional biomarkers were analyzed in Zambian preschool children1

ABC DP

Total-body stores,2 mmol 740 (483, 973) [27]a,b,3 796 (671, 922) [34]a 668 (512, 803) [34]b 640 (510, 781) [38]b 0.031 Downloaded from https://academic.oup.com/ajcn/article-abstract/102/2/497/4564660 by Texas Womans University user on 15 January 2019
Estimated liver concentration,2 1.05 (0.82, 1.54) [27] 1.17 (0.98, 1.49) [34] 1.06 (0.78, 1.28) [34] 1.01 (0.83, 1.11) [38] 0.068

mmol/g 0.95 (0.79, 1.16) [27] 1.00 (0.88, 1.19) [36] 0.90 (0.65, 1.05) [32] 0.96 (0.75, 1.12) [36] 0.24
Serum retinol,2 mmol/L 0.96 (0.79, 1.14) [19] 1.18 (0.85, 1.47) [18] 0.86 (0.61, 1.32) [18] 0.90 (0.67, 1.28) [23] 0.70
0.96 (0.86, 1.10) [19] 1.02 (0.89, 1.20) [18] 0.95 (0.87, 1.09) [18] 1.04 (0.86, 1.21) [22] 0.45
Serum RBP, mmol//L 3.0 (1.4, 3.7) [29]b 4.0 (2.2, 5.2) [31]a 1.5 (1.2, 2.7) [29]b 2.6 (1.4, 3.6) [32]b 0.0009

Retinol:RBP, molar ratio 0.65 (0.50, 1.00) [29] 0.74 (0.51, 1.14) [31] 0.61 (0.40, 0.91) [27] 0.57 (0.29, 1.00) [32] 0.42
Retinyl esters,2 molar 0.49 (0.35, 0.62) [29]b 0.81 (0.47, 1.06) [31]a 0.46 (0.29, 0.73) [30]b 0.45 (0.26, 0.67) [32]b 0.0016
0.07 (0.05, 0.10) [27] 0.07 (0.03, 0.19) [22] 0.07 (0.05, 0.12) [24] 0.10 (0.07, 0.13) [28] 0.60
percentage of serum VA 0.95 (0.67, 1.23) [29]a,b 0.50 (0.39, 0.67) [31]c 0.77 (0.41, 1.15) [30]b,c 0.98 (0.81, 1.23) [33]a 0.0003
b-Carotene,2 mmol/L 0.04 (0.03, 0.06) [23] 0.02 (0.02, 0.03) [14] 0.04 (0.02, 0.06) [23] 0.04 (0.03, 0.06) [23] 0.07
a-Carotene,2 mmol/L 0.13 (0.09, 0.16) [17]c 0.34 (0.26, 0.57) [22]a 0.11 (0.10, 0.37) [9]b,c 0.20 (0.12, 0.54) [14]a,b ,0.0001
b-Cryptoxanthin,2 mmol/L 3.3 (2.8, 4.1) [19] 3.1 (2.6, 3.5) [24] 2.8 (2.4, 3.6) [25] 2.7 (2.2, 4.1) [20] 0.48
Lutein,2 mmol/L 16.7 (11.9, 21.1) [26]a 15.6 (11.2, 19.1) [22]a 9.1 (7.1, 14.1) [21]b 13.2 (8.5, 18.2) [21]a,b 0.0044
Zeaxanthin,4 mmol/L ,0.0001
Lycopene,2,4 mmol/L 7.1 [28] 12.9 [31] 31.0 [29] 19.4 [36] 1.0
ALT activity,2 U/L 93.1 [29] 93.5 [31] 92.9 [28] 97.4 [38]
PLP,2 nmol/L

Elevated CRP (.10 mg/L), % [n]

Elevated AGP (.1.2 g/L), % [n]

1Some biomarkers differed by site, which may have reflected local consumption of some foods. Villages A and B were closest to the paved road, and

villages C and D were 8 km from the road. Carotenoid values were reported only for separately identifiable peaks. Groups with uncommon superscript letters

were different: a . b . c. P values are for testing the null hypothesis that each variable was equal in treatment groups by using an ANOVA or chi-square test.
AGP, a1acid glycoprotein; ALT, alanine aminotransferase; CRP, C-reactive protein; RBP, retinol binding protein; VA, vitamin A.

2Nonnormally distributed residuals; P value reflects a nonparametric analysis.
3Median; first and third quartile values in parentheses; n in brackets (all such values). First and third quartile values are also known as 25th and 75th

percentiles, respectively.
4Values were not always quantifiable in samples that were ,100 mL serum.

by age, sex, BMI, CRP, or AGP; however, 3 individual carotenoids correlated with CRP (P = 0.035, r2 = 0.06) and AGP (P = 0.010,
(i.e., a-carotene, lutein, and lycopene; P # 0.0016) were affected r2 = 0.09).
by site, and therefore, results are presented by site (Table 2).
Serum ALT activity and pyridoxal-5#-phosphate
Serum total carotenoids had an overall mean concentration of
2.48 6 1.2 mmol/L (median 2.41 mmol/L) and did not differ by site. The liver enzyme ALT was evaluated in serum to determine
A common reference range for serum carotene is 0.9–3.7 mmol/L whether liver damage was present because of the hypervitaminotic
(28), which would encompass the overall mean. Sixteen samples state of some children. ALT activity was below normal (range:
(13%) had total carotenoid concentrations . 3.7 mmol/L, which 0.83–11.4 U/L), and only one child tested had normal activity
were considered hypercarotenemic. (10-40 U/L). ALT activity was not related to any other factors
evaluated. Because ALT activity was low, PLP concentrations
Fasting serum retinyl esters and ratio of retinol to were determined and were also below normal; 79% of children
retinol-binding protein had serum concentrations ,20 nmol/L, which is the suggested
deficiency cutoff (29), and 29% of values were in the extremely
Retinyl oleate, palmitate, and stearate were identified in serum low range of ,10 nmol/L. Although the r2 was low (r2 = 0.06),
extracts (n = 123). Other esters were present but sometimes PLP was statistically significantly negatively associated with
overlapped with other unidentified compounds, which could CRP (P = 0.020).
have been carotenoids that might have been detectable at 325
nm at the high concentrations in many of the samples. Retinyl DISCUSSION
esters were not statistically significantly affected by age, sex,
BMI, CRP, or AGP; however, the retinyl ester percentage of This study reports supporting biomarkers of VA status in a group
serum total VA (P = 0.0009) was affected by site, and there- of children with a large percentage diagnosed with hypervita-
fore, results are presented by site (Table 2). In all participants, minosis A by using an RID methodology (7). The children had
16% of subjects had retinyl esters .5% of serum total VA, relatively low weight- and height-for-age, and some children had
whereas 2% of subjects had retinyl esters .10% of serum total subclinical malaria. Provitamin A carotenoid concentrations be-
VA. In line with higher TBSs, retinyl esters in village B were fore the dietary intervention were higher than in most other global
also statistically significantly higher when based on total reti- populations, including in developed countries with minimal VA
nol equivalents in the serum, and the third quartile value (75th deficiency (Table 3). Fasting retinyl esters were slightly elevated
percentile) included the lower cutoff of .5% (Table 2). SR in some children, and the retinol-to-RBP ratio was 1.0. Although
and RBP concentrations did not differ by site. The ratio of we assayed ALT activity to assess potential liver involvement,
retinol to RBP was very close to 1.0 and did not differ by
site. As expected, but with low r2 values, RBP was negatively

HIGH SERUM CAROTENOID CONCENTRATIONS IN ZAMBIA 501

TABLE 3
Serum carotenoid concentrations in apparently healthy children from various regions of the world1

Retinol, b-Carotene, a-Carotene, b-Cryptoxanthin, Lutein, Lycopene,

Country Age, y n Notes mmol/L mmol/L mmol/L mmol/L mmol/L mmol/L Reference

Zambia 5–7 123 — 0.98 0.76 0.62 0.10 0.86 0.30 Current
Belize 3–9 493 — — 0.21 0.093 0.15 0.23+z 0.11 Apgar and

China 0.5–2 254 — 0.96 0.056 0.003 0.027 0.22+z — Gunter (30) Downloaded from https://academic.oup.com/ajcn/article-abstract/102/2/497/4564660 by Texas Womans University user on 15 January 2019
0.992 0.1252 0.0262 0.0902 0.093+z2 0.0172 Fan et al. (31)
Germany — 49 Native German 0.902 0.0932 0.0152 0.0612 0.11+z2 0.0172 Ru¨hl et al. (32)

— 32 Turkish immigrants, 0.942 0.0712 0.0112 0.0562 0.10+z2

well-adapted 1.17 0.16 0.049 0.16
1.102 0.312 0.0352 0.122
— 41 Turkish immigrants, 0.41 0.11 0.19 0.0192
— 0.0113 0.0033 0.0233
weakly adapted — 0.0233 0.0063 0.0343
— 1.29
Hungary 3.1–17.5 29 Noninfectious 0.78 0.13 — — 0.12+z ,0.007 Cser et al. (33)
0.68 0.02 0.06 0.422 0.0682 Das et al. (34)
India 2–11 50 Nonmalarial 0.66 0.45+z 0.18 Okuda et al. (35)
0.23 0.15 0.19 0.044+z3 0.007 Gamble et al. (36)
Japan 10–11 216 — 0.03 0.07 0.052+z3 0.012
1.17 — Adelekan et al. (37)
Marshall Islands 1–6 189 SR ,0.7 mmol/L 0.16 0.57 0.12 — 0.05 Ribaya-Mercado
0.61 0.030 0.020 0.50+z
89 SR .0.7 mmol/L 0.63 0.26 0.25 et al. (5)
0.34 0.47+z —
Nigeria 0.7–8 19 Nonmalarial 0.192 0.075 0.21 0.22 Ribaya-Mercado
0.052 — — et al. (12, 14)
Philippines 6.8–13.2 27 Baseline 1.29 0.070
0.46+z — Rankins et al. (38)
Philippines After 12 wk F/V 1.06 Wageesha et al. (39)
9–12 116 Baseline 0.87 —

Senegal 2–4 After 9 wk F/V 0.90 0.34+z 0.46 Ford et al. (13)
Sri Lanka 0.5–4.8 281 — — — 0.152 Spannaus-Martin
35 Individuals with
United States 6–7 1.35 et al. (40)
United States 0.4–6 carotenodermia
839 NHANES III —
77 — 1.09

1All values are means unless otherwise indicated. After Zambia (current study), countries are listed in alphabetical order. F/V, fruit and vegetable

intervention; SR serum retinol; +z, zeaxanthin included in the published lutein value.
2Median.
3Geometric mean.

a high prevalence of vitamin B-6 deficiency was discovered, Polymorphisms in BCO1 are associated with elevated fasting
which likely interfered with the ALT activity. b-carotene concentrations (42), and a lower conversion effi-
ciency of b-carotene, albeit at pharmacologic doses (43). Al-
Inferences from serum carotenoids consistent with high though statistically significant, a polymorphism associated with
liver VA stores a higher b-carotene concentration only explained 1.9% of the
variance at the population level (42), and other variations in
Aside from dietary intake, serum carotenoid concentrations genes related to the absorption, transport, and cleavage of ca-
can be altered by a number of factors including VA status, the rotenoids are likely involved in the individual variation in re-
transcriptional regulation of or genetic variations in genes related sponse to dietary carotenoids (16) as well as their transcription
to carotenoid metabolism, and BMI. VA status is likely the most factors (18).
important factor that influenced BCO1 activity (41). A negative-
feedback system induced by retinoic acid reduced b-carotene BMI was negatively correlated with all serum carotenoids
intestinal absorption and cleavage proteins (19) and led to less- except lycopene in the most-comprehensive survey of serum
efficient b-carotene conversion in rats that consumed a high-VA carotenoid concentrations in US children as part of the NHANES
diet (17), which explains the inverse relation between VA status III (13). Responses of plasma carotenoids to a fruit and vegetable
and provitamin A conversion in humans (5). Although serum intervention were also inversely related to BMI (14). Zambian
carotenoid increases were not analyzed by VA status in the study children in this study had low BMI-for-age z scores, which may
of Ribaya-Mercado et al. (5), TBSs and serum carotenoids in- be associated with higher serum carotenoid concentrations.
creased during the intervention, indicating that, while under However, even when NHANES III data were stratified by lowest
negative feedback of VA status, carotenoids were still being BMI category (#15th percentile), Zambian children mean values
absorbed intact in most subjects despite a likely downregulation were still above the 95th percentile for serum a-carotene,
of carotenoid transporter scavenger receptor-B1. The observa- b-carotene, and lutein (13).
tions of adequate VA intakes including provitamin A sources
(23), elevated TBSs (7), and high serum carotenoids in Zambian A comprehensive comparison of published serum carotenoid
children are consistent with these hypotheses. and retinol concentrations in apparently healthy children was
conducted (Table 3); children with an infection or active malaria
were excluded, and control groups are presented. For provitamin

502 MONDLOCH ET AL.

A carotenoids, Zambian children had much-higher concentrations Paradoxical response of ALT and its explanation by Downloaded from https://academic.oup.com/ajcn/article-abstract/102/2/497/4564660 by Texas Womans University user on 15 January 2019
of a-carotene and b-carotene than those of all other nonintervened impaired vitamin B-6 status
groups except for Nigerian children who were reported to have
been consuming red palm oil (37), which contains large amounts Serum ALT activity was assayed to determine whether el-
of a-carotene and b-carotene (44). a-Carotene is not well dis- evated VA stores resulted in hepatotoxicity (51); however,
tributed in vegetables but is present in pumpkin (9), which is values were below normal. PLP is a cofactor for ALT; there-
consumed in Zambia (23). b-Cryptoxanthin concentrations were fore, vitamin B-6 deficiency interferes with ALT-activity as-
comparable to those of other groups but lower than for US children sessment. Although inflammation was reportedly associated
(Table 3). For non–provitamin A, lutein concentrations were with lower PLP values (52), as observed here, the observation
higher than those of other nonintervened groups, and lycopene of extremely low PLP concentrations (29.3% of values ,10 nmol/L)
concentrations were higher than those of all groups except of and a very-high incidence (79%) of serum PLP concentrations
US children (13). These findings reflect the availability of or- ,20 nmol/L strongly indicated widespread vitamin B-6 de-
anges and tomato-based products in the United States. Oranges ficiency in these children. Functional biomarkers of vitamin
are not common in this part of Zambia, and tomatoes are eaten B-6 status (53, 54) are needed to confirm and extend these
freshly cooked and not concentrated. The only groups of children findings.
with carotenoid concentrations close to those of this Zambian
cohort are those in individuals who consumed high fruit and Maize is the Zambian staple food, and the common practice of
vegetable regimens that dramatically increased serum carot- refining maize and removing the nutritious germ and hull to
enoids (5, 12, 14) and in subjects with carotenodermia that improve consistency impacts intakes of vitamin B-6 and other
was due to excessive ingestion of carrot, pumpkin, and papaya nutrients (55). Although fish is consumed in this area, con-
(39). sumption is likely w50 g one or 2 times/wk (23). Other animal
source foods are expensive or scarce in rural Zambia. A modi-
Differences existed by site in serum carotenoid concentrations. fication to use whole-grain maize instead of refined could impact
Village B stood out from the others as having the highest a-carotene human nutrient intake and merits additional investigation. Site
and lycopene concentrations but lowest lutein concentrations, differences were noted, and children from village C had the
which likely reflected greater consumption of pumpkin and lowest PLP concentrations of all sites, and villages closest to the
tomatoes and less intake of leafy green vegetables relative to main road (i.e., villages A and B) had the highest concentrations,
other sites. All of these plant-source foods would have been which possibly reflected greater access to foods containing vi-
available during this harvest and early postharvest season (23). tamin B-6.

Inferences from retinyl ester distribution and retinol:RBP In conclusion, although this study used sophisticated RID
methods to estimate TBSs, which requires the use of mass
Evidence from animal and human studies suggested that spectrometry, less-technical methods were used to support the
serum retinyl esters are a potential biomarker for hypervita- original diagnosis of a high degree of hypervitaminosis A in these
minosis A. In hypervitaminotic rhesus monkeys, serum retinyl children. Carotenoid and retinyl ester profiles can be performed
esters as a percentage of the total ranged from 5.5% to 23% for with gradient HPLC. To have high sensitivity to detect minor
animals experiencing hypertrophy and hyperplasia of liver retinyl esters in addition to palmitate, we analyzed the carotenoid
stellate cells (45, 46) but with normal SR concentrations (i.e., and retinyl ester profiles on different systems. A method could be
1.21 6 0.28 mmol/L) (46). In 2 groups of postmenopausal developed to run both profiles simultaneously especially if retinyl
women, serum retinyl esters were not considered elevated palmitate is targeted. The findings of hypercarotenemia, saturated
(2.26 6 1.39% and 2.45 6 1.30%) despite dietary intakes that RBP, and elevated retinyl esters in some of the children support
were 2 times the current RDA of 700 mg retinol activity excessive stores of VA in this community.
equivalents/d (47). Therefore, we suggest that 5% retinyl
esters of total VA, which is more than twice the mean of these We thank Sara Arscott, Samantha Schmaelzle, and Fabiana Moura for as-
healthy women with high intakes, may be a more-useful sistance in coordinating the study in Zambia. We also thank Michael Grahn
cutoff in fasting blood samples to infer potential hypervita- for doing analyses, and Peter Crump for statistical consultation. We also thank
minosis A in children than is the 10% cutoff when more- Daniel Raiten for encouraging us to investigate vitamin B-6 status in this co-
quantitative methods, such as RID and liver biopsy, are not hort to explain low ALT activity.
available.
The authors’ responsibilities were as follows—SM, BMG, and SAT:
RBP is used as a surrogate for SR concentrations, which produced the first draft of the manuscript; SM: analyzed samples for ca-
were not elevated. Seventeen percent of subjects were mis- rotenoids and retinyl esters; BMG: was involved in the orchestration of the
diagnosed with VA deficiency on the basis of an SR cutoff field work and analyzed all samples on the gas chromatography–combustion
,0.7 mmol/L (7). In this study, the retinol-to-RBP molar ratio isotope ratio mass spectrometer; CRD: managed the laboratory and devel-
was essentially 1.0, suggesting that RBP was highly satu- oped HPLC analyses with SM; JC: was the point of contact for the Zambian
rated; free retinol was not circulating unbound to RBP, which ethics committee and supervised the team from TDRC in the collection of
is a concern in hypervitaminotic states (22), and apo-RBP blood samples, analysis of hemoglobin and malaria slides, and paper work
was not released into the serum from the liver, which can related to releasing serum and food samples from Zambia; CK: was instru-
occur in VA deficiency (48, 49). Furthermore, considering mental in the coordination of procurement of supplies and communication
that our children were not overweight, apo-RBP was not among partners; CM: coordinated with the local government for the conduct
shown circulating without its ligand, which is characteristic of the trial, oversaw all National Food and Nutrition Commission staff, and
of excess adipose tissue (50). met with SAT to discuss all aspects of the trial; LR-A and JFG: were re-
sponsible for PLP analyses; SAT: designed the study as the principal in-
vestigator, was involved in all aspects of the study, including manuscript
revision, and is the guarantor of the study; and all authors: reviewed and
approved the manuscript. HarvestPlus (www.harvestplus.org) is a global

HIGH SERUM CAROTENOID CONCENTRATIONS IN ZAMBIA 503

alliance of agriculture and nutrition research institutions working to in- 20. Jahns L, Johnson LK, Mayne ST, Cartmel B, Picklo MJ Sr., Ermakov Downloaded from https://academic.oup.com/ajcn/article-abstract/102/2/497/4564660 by Texas Womans University user on 15 January 2019
IV, Gellermann W, Whigham LD. Skin and plasma carotenoid response
crease the micronutrient density of staple food crops through biofortifica- to a provided intervention diet high in vegetables and fruit: uptake and
depletion kinetics. Am J Clin Nutr 2014;100:930–7.
tion. The views expressed in this article do not necessarily reflect those of
21. Smith FR, Goodman DS. Vitamin A transport in human vitamin A
HarvestPlus. None of the authors reported a conflict of interest related to the toxicity. N Engl J Med 1976;294:805–8.

study. 22. Mallia AK, Smith JE, Goodman DS. Metabolism of retinol-binding
protein and vitamin A during hypervitaminosis A in the rat. J Lipid Res
REFERENCES 1975;16:180–8.

1. Haskell M, Ribaya-Mercado JD, Vitamin A Tracer Task Force. 23. Hotz C, Chileshe J, Siamusantu W, Palaniappan U, Kafwembe E.
Handbook on vitamin A tracer dilution methods to assess status and Vitamin A intake and infection are associated with plasma retinol
evaluate intervention programs. Washington (DC): HarvestPlus; 2005. among pre-school children in rural Zambia. Public Health Nutr
2012;15:1688–96.
2. Tanumihardjo SA. Vitamin A: biomarkers of nutrition for de-
velopment. Am J Clin Nutr 2011;94:658S–65S. 24. Bresnahan KA, Chileshe J, Tanumihardjo SA. Quantification of food
and nutrient intakes in Zambian children with and without malaria
3. Stephensen CB, Gildengorin G. Serum retinol, the acute phase re- under controlled feeding conditions. Exp Biol Med (Maywood) 2014;
sponse, and the apparent misclassification of vitamin A status in the 239:45–51.
third National Health and Nutrition Examination Survey. Am J Clin
Nutr 2000;72:1170–8. 25. Schmaelzle S, Kaliwile C, Arscott SA, Gannon BM, Masi C, Tanumihardjo
SA. Nutrient and nontraditional food intakes by Zambian children in
4. Thurnham DI, Mburu AS, Mwaniki DL, De Wagt A. Micronutrients in a controlled feeding trial. Food Nutr Bull 2014;35:60–7.
childhood and the influence of subclinical inflammation. Proc Nutr Soc
2005;64:502–9. 26. Molldrem KL, Li J, Simon PW, Tanumihardjo SA. Lutein and b-carotene
from lutein-containing yellow carrots are bioavailable in humans. Am J
5. Ribaya-Mercado JD, Solon FS, Solon MA, Cabal-Barza MA, Perfecto Clin Nutr 2004;80:131–6.
CS, Tang G, Solon JAA, Fjeld CR, Russell RM. Bioconversion of plant
carotenoids to vitamin A in Filipino school-aged children varies in- 27. Ubbink JB, Serfontein WJ, de Villiers LS. Stability of pyridoxal-5-
versely with vitamin A status. Am J Clin Nutr 2000;72:455–65. phosphate semicarbazone: applications in plasma vitamin B6 analysis
and population surveys of vitamin B6 nutritional status. J Chromatogr
6. Gannon BM, Tanumihardjo SA. Comparisons among equations used 1985;342:277–84.
for retinol isotope dilution in the assessment of total body stores and
total liver reserves. J Nutr 2015;145:847–54. 28. Tung EE, Drage LA, Ghosh AK. Carotenoderma and hyper-
carotenemia: markers for disordered eating habits. J Eur Acad Der-
7. Gannon B, Kaliwile C, Arscott SA, Schmaelzle S, Chileshe J, Kalaungwana matol Venereol 2006;20:1147–8.
N, Mosonda M, Pixley K, Masi C, Tanumihardjo SA. Biofortified orange
maize is as efficacious as a vitamin A supplement in Zambian children even 29. National Research Council. Dietary Reference Intakes for thiamin, ribo-
in the presence of high liver reserves of vitamin. A Am J Clin Nutr 2014; flavin, niacin, vitamin B6, folate, vitamin B12, pantothenic acid, biotin,
100:1541–50. and choline. Washington (DC): The National Academies Press; 1998.

8. Lietz G, Haskell M, Melse-Boonstra A. Uncertainties of assessing total 30. Apgar J, Gunter W. Serum carotenoid concentrations in children in
body vitamin A stores in community settings in low-income countries using Belize and their reproducibility. Am J Clin Nutr 1996;64:726–30.
the stable-isotope dilution methodology. Am J Clin Nutr 2015;102:520–1.
31. Fan P, Shen Y, Li Y, Aiqin MA, Sun Z, Wang Z. [Serum retinol and
9. Arscott SA. Food sources of carotenoids. In: Tanumihardjo SA, editor. carotenoid of rural infants and young children in Linyi of Shangdong
Carotenoids and human health. New York: Springer Science and Province and analysis on their related influencing dietary factors.] Wei
Business Media; 2013. p. 3–19. Sheng Yan Jiu 2012;41:424–8 (in Chinese).

10. Johnson EJ, Hammond BR, Yeum KJ, Qin J, Wang XD, Castaneda C, 32. Ru¨hl R, Taner C, Schweigert FJ, Wahn U, Gru¨ber C. Serum carotenoids
Snodderly DM, Russell RM. Relation among serum and tissue con- and atopy among children of different ethnic origin living in Germany.
centrations of lutein and zeaxanthin and macular pigment density. Am J Pediatr Allergy Immunol 2010;21:1072–5.
Clin Nutr 2000;71:1555–62.
33. Cser MA, Majchrzak D, Rust P, Sziklai-László I, Kovács I, Bocskai E,
11. Micozzi MS, Brown ED, Edwards BK, Bieri JG, Taylor PR, Khachik F, Elmadfa I. Serum carotenoid and retinol levels during childhood in-
Beecher GR, Smith JC Jr. Plasma carotenoid response to chronic intake fections. Ann Nutr Metab 2004;48:156–62.
of selected foods and b-carotene supplements in men. Am J Clin Nutr
1992;55:1120–5. 34. Das BS, Thurnham DI, Das DB. Plasma a-tocopherol, retinol, and
carotenoids in children with falciparum malaria. Am J Clin Nutr 1996;
12. Ribaya-Mercado JD, Maramag CC, Tengco LW, Dolnikowski GG, 64:94–100.
Blumberg JB, Solon FS. Carotene-rich plant foods ingested with
minimal dietary fat enhance the total-body vitamin A pool size in 35. Okuda M, Sasaki S, Bando N, Hashimoto M, Kunitsugu I, Sugiyama S,
Filipino schoolchildren as assessed by stable-isotope-dilution meth- Terao J, Hobara T. Carotenoid, tocopherol, and fatty acid biomarkers
odology. Am J Clin Nutr 2007;85:1041–9. and dietary intake estimated by using a brief self-administered diet
history questionnaire for older Japanese children and adolescents. J
13. Ford ES, Gillespie C, Ballew C, Sowell A, Mannino DM. Serum ca- Nutr Sci Vitaminol (Tokyo) 2009;55:231–41.
rotenoid concentrations in US children and adolescents. Am J Clin
Nutr 2002;76:818–27. 36. Gamble MV, Palafox NA, Dancheck B, Ricks MO, Briand K, Semba
RD. Carotenoid status among preschool children with vitamin A de-
14. Ribaya-Mercado JD, Maramag CC, Tengco LW, Blumberg JB, Solon ficiency in the Republic of the Marshall Islands. Asia Pac J Clin Nutr
FS. Relationships of body mass index with serum carotenoids, to- 2004;13:336–40.
copherols and retinol at steady-state and in response to a carotenoid-
rich vegetable diet intervention in Filipino schoolchildren. Biosci Rep 37. Adelekan DA, Adeodu OO, Thurnham DI. Comparative effects of
2008;28:97–106. malaria and malnutrition on plasma concentrations of antioxidant mi-
cronutrients in children. Ann Trop Paediatr 1997;17:223–7.
15. Zhang J, Wang CR, Xue AN, Ge KY. Effects of red palm oil on serum
lipids and plasma carotenoids level in Chinese male adults. Biomed 38. Rankins J, Green NR, Tremper W, Stacewitcz-Sapuntzakis M, Bowen
Environ Sci 2003;16:348–54. P, Ndiaye M. Undernutrition and vitamin A deficiency in the De-
partment of Lingue`re, Louga Region of Sénégal. Am J Clin Nutr
16. Borel P. Genetic variations involved in interindividual variability in 1993;58:91–7.
carotenoid status. Mol Nutr Food Res 2012;56:228–40.
39. Wageesha ND, Ekanayake S, Jansz ER, Lamabadusuriya S. Studies on
17. van Vliet T, van Vlissingen MF, van Schaik F, van den Berg H. hypercarotenemia due to excessive ingestion of carrot, pumpkin and
b-Carotene absorption and cleavage in rats is affected by the vitamin A papaw. Int J Food Sci Nutr 2011;62:20–5.
concentration of the diet. J Nutr 1996;126:499–508.
40. Spannaus-Martin DJ, Cook LR, Tanumihardjo SA, Duitsman PK,
18. Lobo GP, Amengual J, Baus D, Shivdasani RA, Taylor D, von Lintig J. Olson JA. Vitamin A and vitamin E statuses of preschool children of
Genetics and diet regulate vitamin A production via the homeobox socioeconomically disadvantaged families living in the midwestern
transcription factor ISX. J Biol Chem 2013;288:9017–27. United States. Eur J Clin Nutr 1997;51:864–9.

19. Lobo GP, Hessel S, Eichinger A, Noy N, Moise AR, Wyss A, Palczewski 41. Lietz G, Lange J, Rimbach G. Molecular and dietary regulation of
K, von Lintig J. ISX is a retinoic acid-sensitive gatekeeper that controls b,b-carotene 15,15’-monooxygenase 1 (BCMO1). Arch Biochem Bio-
intestinal b,b-carotene absorption and vitamin A production. FASEB J phys 2010;502:8–16.
2010;24:1656–66.

504 MONDLOCH ET AL.

42. Ferrucci L, Perry JR, Matteini A, Perola M, Tanaka T, Silander K, Rice 49. Wahed MA, Alvarez JO, Khaled MA, Mahalanabis D, Rahman MM, Downloaded from https://academic.oup.com/ajcn/article-abstract/102/2/497/4564660 by Texas Womans University user on 15 January 2019
N, Melzer D, Murray A, Cluett C, et al. Common variation in the beta- Habte D. Comparison of the modified relative dose response (MRDR)
carotene 15,15’-monooxygenase 1 gene affects circulating levels of and the relative dose response (RDR) in the assessment of vitamin A
carotenoids: a genome-wide association study. Am J Hum Genet 2009; status in malnourished children. Am J Clin Nutr 1995;61:1253–6.
84:123–33.
50. Mills JP, Furr HC, Tanumihardjo SA. Retinol to retinol-binding protein
43. Lietz G, Oxley A, Leung W, Hesketh J. Single nucleotide poly- (RBP) is low in obese adults due to elevated apo-RBP. Exp Biol Med
morphisms upstream from the b-carotene 15,15’-monoxygenase gene (Maywood) 2008;233:1255–61.
influence provitamin A conversion efficiency in female volunteers. J
Nutr 2012;142:161S–5S. 51. Hathcock JN, Hattan DG, Jenkins MY, McDonald JT, Sundaresan PR,
Wilkening VL. Evaluation of vitamin A toxicity. Am J Clin Nutr 1990;
44. Sundram K, Sambanthamurthi R, Tan YA. Palm fruit chemistry and 52:183–202.
nutrition. Asia Pac J Clin Nutr 2003;12:355–62.
52. Paul L, Ueland PM, Selhub J. Mechanistic perspective on the re-
45. Penniston KL, Tanumihardjo SA. Subtoxic hepatic vitamin A con- lationship between pyridoxal 5#-phosphate and inflammation. Nutr Rev
centrations in captive rhesus monkeys. J Nutr 2001;131:2904–9. 2013;71:239–44.

46. Penniston KL, Thayer JC, Tanumihardjo SA. Serum vitamin A esters 53. Leklem JE. Vitamin B-6: a status report. J Nutr 1990;120:1503–7.
are high in captive rhesus (Macaca mulatta) and marmoset (Callithrix 54. da Silva VR, Rios-Avila L, Lamers Y, Ralat MA, Midttun Ø, Quinlivan
jacchus) monkeys. J Nutr 2003;133:4202–6.
EP, Garrett TJ, Coats B, Shankar MN, Percival SS, et al. Metabolite
47. Penniston KL, Weng N, Binkley N, Tanumihardjo SA. Serum retinyl profile analysis reveals functional effects of 28-day vitamin B-6 re-
esters are not elevated in postmenopausal women with and without striction on one-carbon metabolism and tryptophan catabolic pathways
osteoporosis whose preformed vitamin A intakes are high. Am J Clin in healthy men and women. J Nutr 2013;143:1719–27.
Nutr 2006;84:1350–6. 55. Hegedu¨s M, Pedersen B, Eggum BO. The influence of milling on the
nutritive value of flour from cereal grains. 7. Vitamins and tryptophan.
48. Muto Y, Smith JE, Milch PO, Goodman DS. Regulation of retinol- Plant Foods Hum Nutr 1985;35:175–80.
binding protein metabolism by vitamin A status in the rat. J Biol Chem
1972;247:2542–50.

nutrients

Review

Vitamin A Supplementation Programs and
Country-Level Evidence of Vitamin A Deficiency

James P. Wirth 1,*, Nicolai Petry 1, Sherry A. Tanumihardjo 2, Lisa M. Rogers 3, Erin McLean 4,
Alison Greig 5, Greg S. Garrett 6, Rolf D. W. Klemm 7,8 and Fabian Rohner 1

1 GroundWork, 7306 Fläsch, Switzerland; [email protected] (N.P.);
[email protected] (F.R.)

2 Department of Nutritional Sciences, University of Wisconsin-Madison, Madison, WI 53706, USA;
[email protected]

3 Department of Nutrition for Health and Development, World Health Organization, 1207 Geneva,
Switzerland; [email protected]

4 UNICEF Headquarters, New York, NY 10017, USA; [email protected]
5 Infant and Young Child Nutrition, Micronutrient Initiative, Ottawa, ON K2P 2K3, Canada;

[email protected]
6 Global Alliance for Improved Nutrition, 1202 Geneva, Switzerland; [email protected]
7 Helen Keller International, New York, NY 10010, USA; [email protected]
8 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD 21205, USA
* Correspondence: [email protected]; Tel.: +41-79-855-9038

Received: 30 October 2016; Accepted: 15 February 2017; Published: 24 February 2017

Abstract: Vitamin A supplementation (VAS) programs targeted at children aged 6–59 months
are implemented in many countries. By improving immune function, vitamin A (VA) reduces
mortality associated with measles, diarrhea, and other illnesses. There is currently a debate
regarding the relevance of VAS, but amidst the debate, researchers acknowledge that the majority of
nationally-representative data on VA status is outdated. To address this data gap and contribute to
the debate, we examined data from 82 countries implementing VAS programs, identified other VA
programs, and assessed the recentness of national VA deficiency (VAD) data. We found that two-thirds
of the countries explored either have no VAD data or data that were >10 years old (i.e., measured
before 2006), which included twenty countries with VAS coverage ≥70%. Fifty-one VAS programs
were implemented in parallel with at least one other VA intervention, and of these, 27 countries
either had no VAD data or data collected in 2005 or earlier. To fill these gaps in VAD data, countries
implementing VAS and other VA interventions should measure VA status in children at least every
10 years. At the same time, the coverage of VA interventions can also be measured. We identified
three countries that have scaled down VAS, but given the lack of VA deficiency data, this would be
a premature undertaking in most countries without appropriate status assessment. While the global
debate about VAS is important, more attention should be directed towards individual countries
where programmatic decisions are made.

Keywords: vitamin A; deficiency; supplementation; fortification; biofortification; MNPs; programs

1. Introduction

Vitamin A deficiency (VAD) is considered one of the most prevalent micronutrient deficiencies
worldwide, mainly affecting children in developing countries [1]. It is estimated that globally about
30% of children <5 years of age are vitamin A (VA) deficient, and about 2% of all deaths are attributable
to VAD in this age group [2]. VAD is also a major cause of preventable childhood blindness [1].
The transfer of VA in breast milk from the mother to the child depends on the status of the mother,

Nutrients 2017, 9, 190; doi:10.3390/nu9030190 www.mdpi.com/journal/nutrients

Nutrients 2017, 9, 190 2 of 18

and thus VAD often develops early in life [3,4], particularly in populations that consume diets low in
provitamin A carotenoids and/or populations prone to infections, which often lead to reduced intake
or depletion of VA stores [5–7].

Supplementation with high doses of preformed VA is currently one of the most widely-used
interventions delivering VA. At present, more than 80 countries worldwide are implementing universal
VA supplementation (VAS) programs targeted to children 6–59 months of age through semi-annual
national campaigns. Due to VA’s influence on immune function, supplementation with a high-dose of
VA is designed to reduce mortality associated with measles, diarrhea, and other illnesses [8] and not to
sustainably improve the VA status of populations. A high-dose of VA improves VA status for only
up to three months in children who have low dietary intake [9]. For this reason, while VAS provides
a protective dose in the presence of VAD, complementary interventions are needed for VAD control
such as VA (bio-) fortification, micronutrient powders, dietary diversity, nutrition education, and
prevention and control of infectious disease.

VAS programs began in the 1990s in response to evidence demonstrating the association between
VAD and increased childhood mortality [10,11]. Between 1990 and 2013, more than 40 efficacy
studies of VAS in children 6–59 months of age were conducted, and two systematic reviews and
meta-analyses have concluded that VA supplements can considerably reduce mortality and morbidity
during childhood [12,13].

In 2013, following the publication of these reviews, Awasti et al. [14] published results of a large
cluster-randomized effectiveness trial in India (the Deworming and Enhanced Vitamin A (DEVTA)
trial) that showed, conversely, that semi-annual VAS did not reduce mortality. In 2014, Fisker et al. [15]
reported the results of a randomized controlled trial in Guinea-Bissau, and concluded that “VAS had no
overall effect on mortality but was associated with reduced mortality in girls and increased mortality
in boys”. In 2015, Mason et al. [16] published a policy paper, referring to the results of the Awasti
study and suggesting that VAS programs are “less relevant” because of a decreasing prevalence of
measles and diarrhea and the fact that VA supplements do not impact the underlying VAD. Moreover,
they suggested a “policy shift” away from VAS towards other interventions (e.g., fortification) that
would sustainably reduce VAD. To date, the World Health Organization (WHO) continues to support
biannual supplementation to children 6–59 months of age in settings where VAD is a public health
problem, and recommends that “VAS should be used along with other strategies to improve vitamin A
intakes, such as dietary diversity and food fortification” [17].

In response, the article by Mason et al., several researchers [18–20] directly disputed their proposed
policy shift. In brief, researchers identified multiple design flaws of the DEVTA trial and cited repeated
criticisms of the DEVTA trial’s methods and conclusions [18], attributed near elimination of severe
VAD, xerophthalmia, and childhood blindness to VAS [19], and stated that more studies are needed
before any phase out of VAS can be conducted [20]. In line with the Global Alliance for Vitamin A
(GAVA) recommendations [21], these researchers [18] also suggested that countries should only
consider scaling back VAS programs in populations where there is evidence that VAD is no longer
a public health problem (i.e., <5% prevalence of VAD). Researchers on both sides of this debate
in the literature have acknowledged that more data on VA status are needed so that countries can
make informed programmatic decisions. Mason et al. [16] suggest “...increasing regular VA intakes,
while monitoring VAD changes”, and Bhutta and Baker [19] advocate for an “...increased quality,
frequency and disaggregation of measurement of vitamin A deficiency” and of effective coverage of
VAS programs. Moreover, Stevens et al., who examined global trends in VAD from 1993 to 2013 [2] and
reported a 10% reduction in the global VAD prevalence (from 39% in 1993 to 29% in 2013), noted that
the majority of data on VA status are outdated, and the available data make it difficult to accurately
estimate the global prevalence of VAD.

To address this data gap and contribute to the debate about VAS, we examined countries with a
routine VAS program and within these countries identified nationally-representative VA status surveys.
We identified the most recent survey to determine if data are current or outdated. Amongst these

Nutrients 2017, 9, 190 3 of 18

countries, we also examined the presence of VA fortification and biofortification, and micronutrient
powder programs. By identifying concretely what data exist in each country, and what programs exist
alongside VAS, we aim to assist national programmers and other stakeholders to identify their place in
the wider debate. What countries need data, or more-current data, to better implement and target their
VAS program? Where do data show that VAS should be scaled up? Where should it be scaled back
based on recent data and other programs in place?

2. Materials and Methods

2.1. VAS Programs

We used the United Nation’s Children’s Fund’s (UNICEF) State of the World’s Children (SOWC)
VAS coverage database [22] updated in November 2015, which reports supplementation coverage
between 2000 and 2014 to identify countries with possible VAS programs. The set of countries included
in the VAS coverage database was updated in 2014, and includes countries that had: (i) mortality rates
among children under-5 years of age (U5MR) >70 deaths per 100,000 population in the year 2000 as per
estimates released by the Inter-agency Group for Child Mortality Estimation (UN IGME) in September
2013; and/or (ii) a history of national level VAS programming; and/or (iii) a severe public health
problem of VAD in the year 2000 or earlier, defined according to WHO recommended cut offs for
declaring VAD a public health problem; or (iv) a history of national level VAS programming combined
with a moderate or mild public health problem of VAD around the year 2000 defined according to
WHO cut offs. The data related to VAD were based on nationally representative data from the grey
literature as well as data in the WHO Micronutrients Database [23].

From UNICEF’s database, we used the most recent estimates of the proportion of children
receiving two consecutive doses of VA supplements per annum. The proportion of children
6–59 months of age that were fully protected with two annual doses in one calendar year is estimated
by taking the lower of the two annual semester-wise coverage estimates. This employs an assumption
that children who received a dose in the semester with the lower coverage were also reached in the
semester with higher coverage. In cases where the VAS in both semesters was only delivered through
events (i.e., no VAS was delivered through routine contacts with the health system), if the timing
between doses was less than four or more than eight months apart, annual two dose coverage is
estimated to be zero.

2.2. Vitamin A Fortification, Bio-Fortification, and Micronutrient Powder Programs

Countries implementing VA programs were identified from multiple sources. The Global Alliance
for Improved Nutrition’s (GAIN) 2015 Fortifying our Future report [24] and website maps [25] were
used to identify countries implementing mandatory and voluntary fortification programs where VA is
added to vegetable oil. VA-fortified sugar programs were identified in two separate publications: the
2006 WHO food fortification guidelines [26] and a review article listing all food fortification programs
in Africa [27]. VA-fortified wheat flour and maize flour programs were identified in a publication by
Klemm et al. [28] and country profiles on the Food Fortification Initiative website [29]. A 2011 global
review of home fortification interventions [30] and preliminary results from UNICEF’s Nutridash
platform for 2015 [31] were used to identify national and sub-national micronutrient powder (MNP)
programs. MNP pilot and emergency programs were not included in our analysis. HarvestPlus’ 2015
global biofortified crop map [32] was used to identify all countries that have released crops (e.g., cassava,
maize, and/or sweet potato) biofortified with provitamin A carotenoids. Countries currently testing
biofortified crops were not included as part of our review.

2.3. Search Strategy for Vitamin A Deficiency Data and Inclusion and Exclusion Criteria

For countries identified in the UNICEF SOWC database, we first explored the WHO Micronutrients
Database for nationally-representative surveys measuring VAD amongst children 6–59 months of age.

Nutrients 2017, 9, 190 4 of 18

We then conducted a literature search in June 2016 using Web of Science, Google Scholar and PubMed
to identify other VA status surveys. Keywords used were: (vitamin A) OR (serum retinol) OR (plasma
retinol) OR (retinol binding protein) AND (country name). We also conducted searches of the websites
of WHO, UNICEF, GAIN, and the Micronutrient Initiative to identify unpublished surveys. Lastly,
the co-authors used their respective organizational networks to identify any surveys not identified
using the previous search approaches. We included nationally-representative surveys of children
6–59 months of age as the primary target age range, but were flexible by including surveys with
slightly older children and children with a smaller age range (e.g., children 6–35 months in Liberia).
We excluded surveys and studies that: (a) were representative of smaller-administrative areas (e.g.,
provinces, cities, urban areas); and (b) used convenience or non-random sampling to select subjects
(e.g., volunteers, clinic or hospital patients, and studies restricted to ethnic minorities, immigrant
populations, or non-representative subgroups).

We included studies or surveys that reported the national prevalence of sub-clinical VAD defined
as low serum/plasma retinol (ROH) or low retinol binding protein (RBP). Survey results were included
regardless of whether ROH or RBP were adjusted for inflammation, but inflammation-adjusted results
were prioritized, and were presented if available. Multiple approaches can be used to adjust ROH/RBP
results for inflammation, but we did not catalogue the approach used in each survey.

2.4. Deficiency and Coverage Definitions

For countries reporting VAD prevalence, we used the WHO guidelines [33] to classify the severity
of the public health problem in children 6–59 months, which considers a VA prevalence of 2%–9%
a mild public health problem, 10%–19% a moderate public health problem, and ≥20% a severe public
health problem. To interpret our findings, we followed the GAVA framework [21], which recommends
that countries with VAD prevalence <5% can consider scaling back VAS. Following the approach used
in UNICEF’s 2007 progress report [34], we used a coverage of ≥70% as a threshold to identify countries
with relatively high VAS coverage.

3. Results

3.1. VAS Coverage and Overlap with Other VA Programs

As shown in Table 1, the UNICEF’s SOWC VAS database is comprised of 82 countries, of which
77 had VAS coverage results. No coverage results were available for Equatorial Guinea, Kazakhstan,
Mexico, Morocco, and Turkmenistan. The coverage of VA supplements ranges from 0% to 99%, and
the median coverage is 70%. The coverage of VAS was ≥70% in 38 countries.

Of all 82 countries explored, 54 were implementing at least one other VA program. Forty-one
implemented programs mass fortifying vegetable oil, sugar, margarine, or wheat flour with VA, and 33
of these countries have mandatory fortification of at least one food. Vegetable oil fortification programs
were the most commonly implemented VA-fortification program, and were conducted in 35 countries.
Of the countries explored, provitamin A-biofortified crops have been released in 21 countries, 17 of
which were also fortifying a staple food (e.g., vegetable oil, sugar, and wheat flour). Twenty-one
countries were implementing MNP programs, twelve of which also implemented fortification and/or
biofortification programs.

3.2. Recentness of Vitamin A Deficiency Data

When examining countries by the recentness of the VAD prevalence data, we found 14 countries
with data collected after 2010, 13 countries with data from 2006 to 2010, seven countries with data
from 2001 to 2005, and 16 countries with data collected in or prior to 2000. For 32 countries, no
nationally-representative data on VAD prevalence could be found. As shown in the map (Figure 1),
countries that have not yet collected data on VAD prevalence are predominantly situated in the West

Nutrients 2017, 9, 190 5 of 18

African Sahel and Central Asia. In Central Africa, some data have been collected, but are largely
outdated. In total, only one-third of the countries explored collected VAD in the past 10 years.

Of the 50 countries that measured VAD prevalence, 35 and 15 countries measured serum/plasma
ROH or RBP concentrations, respectively. Twenty surveys accounted for inflammation in some manner,
either by excluding all children with inflammation or adjusting ROH or RBP concentrations.

3.3. VAS Program and Deficiency Data Comparison

A comparison of VAS coverage and availability of survey data identified 13 countries that
have high coverage (i.e., ≥70%) of VA supplements but have never collected national data on VAD
prevalence. In Africa, these countries include Burkina Faso, Chad, Congo, Guinea, Guinea Bissau, Mali,
Mauritania, Niger, and Sudan. In East and Central Asia, these countries include Myanmar, Democratic
People’s Republic of Korea, Tajikistan, and Uzbekistan. Seven countries (Benin, Botswana, Democratic
Republic of Congo (DRC), Nepal, Nigeria, Rwanda, and Zambia) had high VAS coverage (i.e., ≥70%)
and national VAD data collected more than 10 years ago (i.e., prior to 2006).

Ten countries (i.e., Bolivia, Bhutan, Djibouti, Eritrea, India, Kiribati, Micronesia, Sao Tome and
Principe, Swaziland, and Togo) had VAS coverage between 40%–69% but no national VAD data, and
four countries had the same VAS coverage but data collected prior to 2006 (i.e., Egypt, Honduras,
Lesotho, and Namibia).

In the 34 countries where VAD is a severe public health problem (measured by a national survey
in any time period), 19 had VAS coverage ≥70%, whereas 15 countries had moderate to low VAS
coverage (<70%). VAD is a moderate public health problem in eight countries, only two of which
(Rwanda and Vietnam) have VAS coverage ≥70%. Among the eight countries where VAD is considered
mild or not a public health problem (i.e., deficiency <10%), four have VAS coverage ≥70%, including
Cambodia, Indonesia, Kyrgyzstan, and Maldives. The above comparisons must be interpreted with
caution because for most countries, the year of VAD measurement is different from the year of
coverage assessment.

Nutrients 2017, 9, 190

Table 1. The most recent estimates of VAS coverage, presence of other VA program

Countries and Territories Coverage (%) of Year of Coverage VA Fortification, Biofortification, N
VAS Program Estimate and MNP Programs 2
Afghanistan
Angola 95 2014 fVO (v), fW (v)
6 2014 bSP (v)
Azerbaijan 58 2014
Bangladesh 0 2014 fVO (m), fW (v), MNP (v)
99 2014 fVO (m)
Benin 45 2013
Bhutan 40 2013 fVO (m), MNP (v)
Bolivia 70 2014
Botswana 98 2014 fVO (m), bSP (v)
Burkina Faso 68 2014 fVO (v), bP (v)
Burundi 71 2014 fVO (v), MNP (v)
Cambodia 96 2014 fVO (m), bP (v)
Cameroon 34 2014
Central African Rep 96 2014 fVO (m)
Chad 14 2014 bC (v), bP (v)
Comoros 99 2014
Congo 99 2014 MNP (v)
Côte d’Ivoire 99 2014 fVO (v)
66 2013
DRC 68 2008 bSP (v)
Djibouti
Egypt 49 2014 fVO (m), fW (m), bSP (v), bM (v)
Equatorial Guinea 71 2014 fMG (m), fS (m), MNP (v)
Eritrea 2 2012 fVO (m)
Ethiopia 27 2014 fVO (m)
Gabon 23 2014 MNP (v)
Gambia 19 2014 fMG (m), fS (m)
Ghana 99 2012 fVO (v), bSP (v)
Guatemala 98 2014 fVO (m), fW (m)
Guinea 30 2014
Guinea-Bissau 40 2005 fVO (m), fS (v), bSP (v)
61 2014
Haiti 84 2014 MNP (v)
Honduras
28 2014 fW (v)
India 54 2006 fVO (m), fS (m)
Indonesia 97 2010 bSP (v), MNP (v)
Kazakhstan 89 2014 fVO (m), fS (m), bSP (v), bC (v), MNP (v)
67 2014
Kenya 0 2014 fVO (v), bM (v)
Kiribati 99 2014
Kyrgyzstan 41 2014 fVO (m)
76 2013 fMG (m), MNP (v)
Laos 98 2013
Lesotho 39 2007
Liberia 89 2014
Madagascar
Malawi
Maldives

Mali
Marshall Islands

Mauritania
Mexico

6 of 18

ms, and VAD prevalence in countries in UNICEF’s SOWC VAS coverage database 1.

Year of Most Recent Biomarker 3 VAD Prevalence Severity of Source
Nationally-Representative VAD Survey (%) 4 VAD
ROH [35]
2013 ROH 50.4 Severe [36] *
1999 RBP 64.3 Severe [37]
2013 ROH 8.0 Mild [38]
2011 ROH 20.5 Severe [39] *
1999 82.0 Severe

1994 ROH 35.4 Severe [40] *

2005 ROH 27.9 Severe [41]
2014 RBP 9.2 Mild [42]
2009 RBP 35.0 †† Severe [43]
1999 ROH 68.2 Severe [44] *

2007 RBP 24.1 Severe [45]
1998/99 -ROH 61.1 Severe [46]

1995 ROH 11.9 Moderate [47] *

2006 ‡ ROH 37.7 Severe [48]

1999 ROH 64.0 Severe [49] *

2009/10 ROH 0.3 None [50]

2005 ROH 32.0 Severe [51] *
1996 ROH 13.8 Moderate [52]

2011 ROH <1 None [53]

2012 RBP 9.2 Mild [54]

2013 RBP 7.8 Mild [55]
2000 ROH 44.7 Severe [56]
1993 ROH 78.0 Severe [57] *
2011 RBP 13.2 Moderate [58]
2000 ROH 42.1 Severe [59]
2009 RBP 22.0 †† Severe [60]
2007 ROH 5.1 Mild [61]

1995 ROH 59.9 Severe [62]

2011/12 ROH 15.7 Moderate [63]

Nutrients 2017, 9, 190

Table 1

Countries and Territories Coverage (%) of Year of Coverage VA Fortification, Biofortification, N
VAS Program Estimate and MNP Programs 2

Micronesia 68 2006 MNP (v)
Mongolia 79 2014 fVO (m)
Morocco fVO (m), bSP (v), MNP (v)
Mozambique 99 2014 MNP (v)
Myanmar 94 2014
Namibia 62 2013 MNP (v)
85 2014 fS (m), MNP (v)
Nepal 4 2014 fVO (m), fS (m), bSP (v)
Nicaragua 95 2014 fVO (m), fW (m), bSP (v), bM (v), bC (v),

Niger 80 2014 bP (v)
MNP (v)
Nigeria 99 2014 fVO (m)
96 2014
DPR Korea 15 2012 fVO (m), fW (m), MNP (v)
Pakistan 83 2014 fVO (m), fS (m), bSP (v), MNP (v)
95 2014
Papua New Guinea 56 2013 fVO (m), bSP (v), MNP (v)
Philippines 89 2014 fVO (m), bC (v)
Rwanda 8 2014 MNP (v)
30 2014 fMG (v), fW (m)
Sao Tome and Principe 42 2013
Senegal 18 2014 MNP (v)
72 2014
Sierra Leone 99 2014 fVO (m), fS (m), bSP (v), MNP (v)
Somalia 43 2014
99 2014 fVO (m)
South Africa 88 2014 fVO (m), fW (m), bSP (v)
South Sudan 40 2013
fVO (m)
Sri Lanka 61 2013
Sudan 66 2014 fVO (v), fS (m), bSP (v), bM (v)
99 2014 fVO (v)
Swaziland 94 2014
Tajikistan 7 2014
Tanzania 93 2013
Timor-Leste 32 2014
Turkmenistan

Togo
Uganda
Uzbekistan
Vietnam
Yemen
Zambia
Zimbabwe

1 VA, vitamin A; VAD, vitamin A deficiency; VAS, vitamin A supplementation; UNICEF, Unit

oil, fMG = fortified margarine, fS = fortified sugar, fW = fortified wheat flour; bSP = biofo
plantain/banana; MNP = micronutrient powders. (m) = mandatory program, (v) = voluntar

measured as proportion of children with ROH or RBP concentrations <0.7 µmol/L, unless

inflammation in some manner (e.g., adjusting ROH or RBP concentrations, excluding children
Database on Vitamin A Deficiency; †† VAD prevalence measured as proportion of children
<0.78 µmol/L in Malawi; ‡ A more recent survey was conducted, but the results were not pu

7 of 18

1. Cont.

Year of Most Recent Biomarker 3 VAD Prevalence Severity of Source
Nationally-Representative VAD Survey (%) 4 VAD
[64]
2010 RBP 32.4 Severe [65]
1996 ROH 40.4 Severe [66]
2002 ROH 68.8 Severe
[67] *
1992 ROH 23.5 Severe [68]
1998 ‡ ROH 32.3 Severe [69]
2004 ROH 2.1 Mild
[70]
2001 ROH 29.5 Severe
[71]
2011 ROH 54.0 Severe [72]
2005 RBP 15.7 Moderate [73]
2013 ROH 20.4 [74]
1996 ROH 6.4 Severe
Moderate [75]
2010 ROH 17.7 [76]
2013 RBP 17.4 Severe [77]
2009 RBP 33.3 †† Moderate [78]
2012 ROH 43.6
Severe [79]
2006 ROH 29.3 Severe

Severe

2010 RBP 33.0 †† Severe [80]
2013
RBP 9.7 Mild [81]

2011 RBP 32.6 †† Severe [82]

2010 ROH 10.1 Moderate [83]

2003 ROH 54.1 Severe [84]
1999 ROH
35.8 Severe [85] *

ted Nations Children’s Fund; SOWC, State of the World’s Children.2 fVO = fortified vegetable
ortified sweet potato, bM = biofortified maize, bC = biofortified cassava, bP = biofortified
ry program; 3 ROH, serum/plasma retinol; RBP, retinol-binding protein. 4 VAD prevalence
s noted otherwise. Prevalences in italics indicate that prevalence calculation accounted for
n with any inflammation, etc); * Data source taken from the World Health Organization Global
n with RBP <0.825 µmol/L in Uganda, Somalia, and Tanzania; <0.83 µmol/L in Cameroon,
ublically available at the time of writing this manuscript.

NutrieNntustr2i0en1t7s, 290,1179,0 9, 190 
FigFuigruer1e. 1R. eRceecnetnntensessso of fn naatitoionnaallllyy-‐rreepprreesseenn

8 of 188  of 18

 

nnttaattiivvee ddaattaa oonn vvitiatamminin AA dedfeicfiiecniecnyc (yV(AVDA)D.  ).

Nutrients 2017, 9, 190 9 of 18

3.4. All Programs and Deficiency Data Comparison

Of the 38 countries with VAS coverage ≥70%, 30 implemented at least one other VA program.
Of these 30 countries, 16 had two or more (bio-)fortified foods or VA programs (e.g., vegetable oil
fortification and MNPs). Most notably, Nigeria has a VAS coverage of 80% and has mandatory
vegetable oil fortification and has released biofortified sweet potato, maize, cassava, and plantain
varieties. Similarly, Zambia has VAS coverage of 93%; has fortified vegetable oil and sugar; and has
released biofortified sweet potato and maize varieties.

Among all countries examined, we found 12 countries (i.e., Bolivia, Burkina Faso, Djibouti,
DPR Korea, Ghana, Guinea, Guinea Bissau, India, Mali, Mauritania, Niger, and Togo) that were
implementing VAS at the same time as another VA program but had no nationally-representative data
on VAD. Fifteen countries (i.e., Angola, Benin, Burundi, DRC, Egypt, Honduras, Lesotho, Madagascar,
Mozambique, Nepal, Nicaragua, Nigeria, Rwanda, Zambia, and Zimbabwe) currently implemented
VAS with another VA program but only collected national VAD data in or prior to 2005. More than
half of the 51 countries that implemented VAS at the same time as another VA program had no VAD
data or data collected prior to 2006.

4. Discussion

4.1. Implications of Outdated/Missing VAD Data

Our study finds that many countries implement VAS and other programs to improve VA status
without evidence of the national prevalence and severity of VAD. This has two critical implications.
First, it may prevent program planners from: (a) focusing resources on the most vulnerable areas; and
(b) from scaling down their programming in areas (e.g., high-income urban areas) where VA-related
mortality incidence and prevalence of VAD may be lower. Second, outdated data and the lack of
temporal comparisons prevent program planners from understanding trends in VAD. Understanding
deficiency trends is particularly important in countries that have scaled up VAS in parallel with other
VA interventions (e.g., (bio-)fortification, MNPs).

Monitoring the potential for excessive intakes is also a growing concern where more than one
staple food is (bio-)fortified or multiple interventions are occurring [86]. For example, VAD data in
West Africa is largely missing or outdated, but VAS and fortification programs have been implemented
concurrently since 2006 when a regional vegetable oil fortification project (Tâche d’Huile) was first
launched at the country level [87]. By 2013, an estimated 75% of the populations in eight West African
countries were consuming fortified vegetable oil [87]. Recent reports from Abidjan, Côte d’Ivoire,
found that 97% of vegetable oil was adequately fortified (i.e., 8 µg retinol equivalents/g) and increased
the nutrient intake of VA by 27% in children 6–23 months of age [88].

It is important to note, however, that overlapping VA interventions at the country level does
not necessarily imply excessive intakes will occur. First, the distribution and consumption of VA
(bio-)fortified foods and MNPs may vary at the sub-national level. Second, sub-optimal program
implementation may limit the intake of VA. To illustrate, Luthringer et al. [89] examined quality
assurance data from 20 national fortification programs and found that more than half of the
food samples tested were inadequately fortified and did not meet national fortification standards.
Despite this review by Luthringer, performance data on VA interventions is often missing or
outdated. This hampers the ability of national programmers to determine if vulnerable populations are
consuming sufficient levels of VA, and hampers their ability to identify potential excessive VA intake.

4.2. Filling the Data Gap

To address the gap in VAD data, national micronutrient surveys (measuring ROH or RBP in
children) should be implemented in countries where data are lacking or outdated. In some cases, using
more sensitive methodology (e.g., modified relative dose response (MRDR), isotope dilution tests [90])

Nutrients 2017, 9, 190 10 of 18

on randomly selected subsamples may be important, especially if widespread fortification programs
are in place.

In 2012, GAVA recommended that program managers prioritize survey results published in the
last ten years [21]. This was based on an assumption that the epidemiological landscape of developing
countries and degree of introduction of other VA programs had changed since the 1990s, and so the
VAD prevalence measured prior to 2000 would not sufficiently reflect the current context. While there
is no set interval between surveys that is recommended, data on VA status should be updated regularly,
particularly in settings where the consumption of VA-rich foods has changed by the introduction of
bio(fortification) programs or improved dietary diversity. Thus, we recommend that VA assessments
be conducted at least every ten years to account for changing consumption patterns and the increasing
coverage of (bio-)fortified foods.

When filling gaps in VAD data, surveys should also measure the coverage of VA interventions.
This ideally includes estimations of additional VA intake originating from (bio-)fortified foods and
supplements with VA (e.g., MNPs). As part of a recent survey in Abidjan, Côte d’Ivoire [88], additional
intake of VA from fortified vegetable oil was calculated using information of quantities of oil consumed
(mL/day) and actual VA levels in oil (µg/L), expressed as percentage of recommended nutrient
intake (RNI).

There are, however, challenges to implementing national micronutrient surveys, and these
challenges may, in part, explain some of the data gaps observed. Micronutrient surveys are more
complex than other commonly-implemented representative surveys (e.g., Demographic and Health
Surveys, Multiple Indicator Cluster Surveys, and Standardized Monitoring and Assessment of Relief
and Transitions (SMART) surveys) because they require the collection of questionnaire data and blood
samples for biomarker analysis. Although some DHS surveys (e.g., Cambodia) have incorporated
a micronutrient component, this practice is rare. Micronutrient surveys need a cold chain for blood
samples, which requires additional effort and resources and can pose challenges when surveying
remote areas. Another challenge to VA status surveys is the fact ROH and RBP are depressed in the
presence of inflammation. Thus, inflammation biomarkers, such as C-reactive protein and alpha 1-acid
glycoprotein, positive acute phase proteins, should be measured in VAD surveys to adjust ROH or
RBP concentrations, and be used to examine inflammation-adjusted and unadjusted VAD prevalence
rates. These inflammation markers will assist in interpreting low ROH values [91]. Various adjustment
approaches exist [92,93], and there is a growing consensus that VAD should only be estimated once
ROH or RBP are adjusted using both C-reactive protein and alpha 1-acid glycoprotein. Despite these
obstacles, micronutrient surveys have been successfully undertaken in countries with poor road
infrastructure, limited electricity networks, and only basic laboratory capacity. While it is difficult to
speculate as to why VAD data exist in some countries and not others, a notable pre-requisite for any
national VA assessment is demand from national nutrition stakeholders for representative data and
available resources.

4.3. Alignment of Vitamin A Deficiency Results and Supplementation Programs

The main rationale for implementing VAS programs is to prevent mortality. It is widely agreed that
VAD data are needed to better target VAS programs and to justify scaling up/down VAS programs [94].

Our analysis identified only three countries, Nicaragua, Guatemala, and Indonesia, that have
VAD prevalence <5%, the threshold recommended by GAVA. Nicaragua and Guatemala mandate that
sugar be fortified with VA, and Guatemala also mandates fortification of margarine. The coverage of
VAS was reported as 4% in Nicaragua and 19% in Guatemala in 2014. Due to evidence of low VAD
prevalence, the Guatemalan government recommended that VAS be scaled back to younger children
and no other foods be fortified. Furthermore, due to concern of excessive VA intakes in Guatemala
from fortified sugar, the government is considering lowering the amount of VA added to 7 mg/kg
sugar [95]. In Indonesia, the very low VAD prevalence observed by the 2011 national survey [53] is
contrasted by a 2011 evaluation of fortified vegetable oil in West Java [96] that found VAD prevalence

Nutrients 2017, 9, 190 11 of 18

between 10% and 18% in preschool children. Though the evaluation was only conducted in 24 villages,
it illustrates that nationally-aggregated VAD results may mask sub-national variations. In the case
of this evaluation, consumption of VA-fortified vegetable oil for one year markedly reduced VAD in
preschool children, particularly in children 24–59 months of age where VAD was reduced from 10%
at baseline to <1% at endline. As the fortification of vegetable oil with VA was made mandatory in
Indonesia in 2013 and was in full effect by 2015 [97] more recent studies in Indonesia examining VAD,
and potentially excessive intakes of VA, should be considered.

Kyrgyzstan offers an example of where national VAD results influenced VAS program policy.
VAD was estimated at 4.2% in 2009, and since 2010, the VAS program has been replaced by national
and universal distribution of an MNP that contains VA and other micronutrients [98,99]. However,
a national follow-up survey in 2013 found that VAD prevalence was 7.8% [55].

In addition, a national VA survey should also be conducted in Zambia, as extensive and recent
sub-national data suggest that children may consume excess amounts of VA [100]. Two recent studies
from Central and Eastern Zambia found that a much lower percentage of children had inadequate liver
reserves assessed with the MRDR test [100,101]. Another study in Eastern Zambia found that a large
proportion of children 5–7 years of age were experiencing hypervitaminosis A, assessed with ROH
isotope dilution [102], and documented hypercarotenodermia during mango season [103], likely due in
part to wide-scale sugar fortification on top of a traditional diet high in provitamin A carotenoids [100].
Another study in Central Zambia found that serum ROH concentrations in children 4 to 8 years old did
not respond to an intervention with provitamin A biofortified maize [104]. The author’s conclusion
was that the children were relatively VA adequate at baseline [104].

4.4. Global Debate and Engagement

In the international peer-reviewed literature, the debate on the appropriateness of VAS programs
focuses primarily, but not exclusively, on results from clinical trials. This debate is important, and
as many researchers rightly point out, there is consistent evidence that VAS programs can reduce
mortality and morbidity. We believe that the evidence is underpinning VAS and clearly demonstrates
its utility. However, our review suggests that country-level data on VAD are sorely needed, and that
several countries implement multiple VA interventions without a current understanding of national
VAD prevalence.

Properly done surveys yielding new VAD data will likely have programmatic implications.
These surveys may identify specific areas where VAD is rare and the VAS program should be scaled
back. In other cases, survey data could suggest that VAS programs should be implemented in parallel
with other VA programs (e.g., (bio-)fortification, and MNPs) to reduce mortality while sustainably
improving VA status. All that said, this type of speculation is futile without good representative data.

More country level data are needed to make evidence-based decisions, and international agencies
that fund VAS programs must support national stakeholders to fill the data gap. While VAD prevalence
estimates at a national level is the starting point for a national dialogue, program planners should
also examine the sub-national pattern of VAD to decide where and how to scale up/back their VAS
program [21]. While debates about efficacy and effectiveness are important, they do little to help
national programmers make decisions and, as such, we see this review as a contribution to focus the
discussion away from global extrapolations to country-specific discussions.

Lastly, in addition to regularly updated documentation of the VAD prevalence of various
population groups, the WHO database can be enhanced by a complementary repository of
micronutrient survey reports/publications. Unfortunately, many findings from national micronutrient
reports are not subsequently published online or in indexed journals that can be searched in perpetuity.

4.5. Limitations

Our study examined only national-level data, and thus data from sub-national but representative
studies was not included. Furthermore, our study is limited to publicly available data. Thus, it is

Nutrients 2017, 9, 190 12 of 18

possible that national policy makers in some countries included in our analysis have sufficient data on
VAD that was missed by our search criteria. Moreover, during our search, the authors noted that some
surveys were currently being implemented or results were not yet publicly released, such as national
surveys in India, Myanmar, Nepal, and Ethiopia. In addition, this inventory only presented the
prevalence of VAD as reported by survey reports and publications, and did not catalogue the various
blood collection methods used, analysis technique used, or if and how ROH or RBP concentrations
were adjusted for inflammation.

We reported the most recent VAS coverage as presented in UNICEF’s SOWC VAS coverage
database, but in some cases, this coverage may lead to misinterpretations of program coverage and
performance. For example, Sierra Leone and Bangladesh had a two-dose coverage of 8% and 0% in
2014, but a coverage of >90% in previous years [105]. This illustrates that implementation irregularities
in VAS programs continue to exist, despite being a well-established, scaled-up intervention in many
countries. In addition, as we only examined the 82 countries included as part UNICEF’s VAS coverage
database, our analysis likely excluded some countries where small scale VAS programs are conducted.

Lastly, although we present the fortification, biofortification, and MNP programs implemented in
each country, no comprehensive source of coverage results for these foods could be identified. The gap
in information about VA program coverage has been noted elsewhere, and has been identified as
a key challenge to identifying if children are being exposed to any or multiple VA interventions [106].
Biofortified staple crops are relatively new in the scheme of nutritional interventions. In Zambia by
2015, seeds of one biofortified maize variety had gained only 1% of the market [107]. Lastly, our
analysis did not include fortified blended flours (e.g., corn–soy blend plus), lipid-based nutrient
supplements, or fortified complementary food products (e.g., porridges).

5. Conclusions

VAS programs are implemented in ~80 low- and middle-income countries. Despite the widespread
implementation of this type of program, many countries have either no data or outdated data on VAD
prevalence. Changing consumption patterns and the expansion of VA (bio-)fortified foods warrants
that countries, particularly those implementing VAS programs, measure VA status in children at least
every 10 years, or when coverage and consumption data indicate that a shift in VAD prevalence
may have occurred. At the international level, UN agencies and non-governmental organizations
should provide technical support to undertake national micronutrient surveys. At the national level,
program planners should make evidence-based decisions, and if biochemical data of VA status are not
sufficient, it should be collected before changes to programs are made. Regarding the global debate
about VAS programs, we found that a few countries have already begun scaling down VAS programs,
and agree that it is an option in countries where VAD prevalence has been repeatedly shown to be
low. Where VAD is a public health problem, however, VAS programs can and should be implemented
in parallel with other programs to improve VA intake and reduce under-five mortality. However,
given the lack of VAD data at the country level, scaling down VAS programs may be a premature
undertaking. While the global debate about VAS is important, more attention should be directed
towards the situation in individual countries where programmatic decisions are made.

Acknowledgments: No funding was received for this research. The authors acknowledge Roland Kupka from
UNICEF and Katie Tripp from US Centers of Disease Control and Prevention for their critical review of earlier
drafts of the manuscript.

Author Contributions: J.P.W., N.P. and F.R. conceived the study. N.P. and J.P.W. compiled and analyzed the data,
and conducted the data analysis. J.P.W., N.P. and F.R. prepared the first draft of the manuscript. S.A.T., L.M.R.,
E.M., A.G. and R.D.W.K. provided survey reports, program data, and contributed to subsequent versions of the
manuscript. All authors reviewed and approved the final version for submission.

Conflicts of Interest: J.P.W., N.P. and F.R. are employees of GroundWork, a company providing technical support
for the implementation of nutrition surveys. L.M.R. is an employee of the World Health Organization, a global UN
agency that recommends that children 6–59 months of age in countries at risk of VAD should receive VAS every
4–6 months. E.M. is an employee of UNICEF, a UN agency that supports governments in the implementation

Nutrients 2017, 9, 190 13 of 18

of programs, including VAS programs. A.G. and R.D.W.K. are employees of Micronutrient Initiative and Helen
Keller International, respectively, non-for-profit agencies that provide technical and/or financial support for
the procurement of vitamin A capsules and implementation vitamin A supplementation programs. GSG is an
employee of GAIN, a not-for-profit organization that supports national food fortification programs in countries
where there is a demonstrated need and a food vehicle which can be fortified. S.A.T. is an employee of the
University of Wisconsin-Madison and declares no conflict of interest. The authors alone are responsible for
the views expressed in this publication and they do not represent the decisions, policies or views of the World
Health Organization, UNICEF, Helen Keller International, the Micronutrient Initiative, GAIN, the University of
Wisconsin-Madison, or GroundWork.

References

1. World Health Organization. Global Prevalence of Vitamin A Deficiency in Populations at Risk 1995–2005;
World Health Organization: Geneva, Switzerland, 2009. Available online: http://apps.who.int/iris/
bitstream/10665/44110/1/9789241598019_eng.pdf (accessed on 17 February 2017).

2. Stevens, G.A.; Bennett, J.E.; Hennocq, Q.; Lu, Y.; De-Regil, L.M.; Rogers, L.; Danaei, G.; Li, G.; White, R.A.;
Flaxman, S.R.; et al. Trends and mortality effects of vitamin A deficiency in children in 138 low-income
and middle-income countries between 1991 and 2013: A pooled analysis of population-based surveys.
Lancet Glob. Health 2015, 3, e528–e536. [CrossRef]

3. Engle-Stone, R.; Haskell, M.J.; Nankap, M.; Ndjebayi, A.O.; Brown, K.H. Breast milk retinol and plasma
retinol-binding protein concentrations provide similar estimates of vitamin A deficiency prevalence and
identify similar risk groups among women in Cameroon but breast milk retinol underestimates the
prevalence of deficiency among young children. J. Nutr. 2014, 144, 209–217. [PubMed]

4. Stoltzfus, R.J.; Underwood, B.A. Breast-Milk vitamin A as an indicator of the vitamin A status of women
and infants. Bull. World Health Organ. 1995, 73, 703–711. [PubMed]

5. Alvarez, J.O.; Salazarlindo, E.; Kohatsu, J.; Miranda, P.; Stephensen, C.B. Urinary-excretion of retinol in
children with acute diarrhea. Am. J. Clin. Nutr. 1995, 61, 1273–1276.

6. Mitra, A.K.; Alvarez, J.O.; Guay-Woodford, L.; Fuchs, G.J.; Wahed, M.A.; Stephensen, C.B. Urinary retinol
excretion and kidney function in children with shigellosis. Am. J. Clin. Nutr. 1998, 68, 1095–1103. [PubMed]

7. Mitra, A.K.; Alvarez, J.O.; Stephensen, C.B. Increased urinary retinol loss in children with severe infections.
Lancet 1998, 351, 1033–1034. [CrossRef]

8. Beaton, G.; Martorell, R.; Aronson, K.; Edmonston, B.; McCabe, G.; Ross, A.; Harvey, B. Effectiveness of Vitamin
A Supplementation in the Control of Young Child Morbidity and Mortality in Developing Countries; Administrative
Committee on Coordination–Subcommittee on Nutrition (ACC/SCN): Geneva, Switzerland, 1993.

9. Palmer, A.C.; West, K.P.; Dalmiya, N.; Schultink, W. The use and interpretation of serum retinol distributions
in evaluating the public health impact of vitamin A programmes. Public Health Nutr. 2012, 15, 1201–1215.
[CrossRef] [PubMed]

10. Sommer, A.; Tarwotjo, I.; Djunaedi, E.; West, K.P., Jr.; Loeden, A.A.; Tilden, R.; Mele, L. Impact of vitamin
A supplementation on childhood mortality. A randomised controlled community trial. Lancet 1986, 1,
1169–1173. [CrossRef]

11. Beaton, G.; Martorell, R.; Aronson, K.; Edmonston, B.; McCabe, G.; Ross, A.; Harvey, B. Vitamin A
supplementation and child morbidity and mortality in developing countries. Bol. Oficina Santitaria Panam.
1994, 117, 506–517.

12. Imdad, A.; Herzer, K.; Mayo-Wilson, E.; Yakoob, M.Y.; Bhutta, Z.A. Vitamin A supplementation for
preventing morbidity and mortality in children from 6 months to 5 years of age. Cochrane Database Syst. Rev.
2010. [CrossRef]

13. Mayo-Wilson, E.; Imdad, A.; Herzer, K.; Yakoob, M.Y.; Bhutta, Z.A. Vitamin A supplements for preventing
mortality, illness, and blindness in children aged under 5: Systematic review and meta-analysis. BMJ 2011,
343, d5094. [CrossRef] [PubMed]

14. Awasthi, S.; Peto, R.; Read, S.; Clark, S.; Pande, V.; Bundy, D.; the DEVTA (Deworming and Enhanced
Vitamin A) team. Vitamin A supplementation every 6 months with retinol in 1 million pre-school children in
north India: DEVTA, a cluster-randomised trial. Lancet 2013, 381, 1469–1477. [CrossRef]

15. Fisker, A.B.; Bale, C.; Rodrigues, A.; Balde, I.; Fernandes, M.; Jorgensen, M.J.; Danneskiold-Samsoe, N.;
Hornshoj, L.; Rasmussen, J.; Christensen, E.D.; et al. High-Dose vitamin A with vaccination after 6 months
of age: A randomized trial. Pediatrics 2014, 134, e739–e748. [CrossRef] [PubMed]

Nutrients 2017, 9, 190 14 of 18

16. Mason, J.; Greiner, T.; Shrimpton, R.; Sanders, D.; Yukich, J. Vitamin A policies need rethinking. Int. J. Epidemiol.
2015, 44, 283–292. [CrossRef] [PubMed]

17. World Health Organization. Guideline: Vitamin A Supplementation in Infants and Children 6–59 Months of Age;
WHO: Geneva, Switzerland, 2011.

18. West, K.P.; Sommer, A.; Palmer, A.; Schultink, W.; Habicht, J.P. Commentary: Vitamin A policies need rethinking.
Int. J. Epidemiol. 2015, 44, 292–294. [CrossRef] [PubMed]

19. Bhutta, Z.A.; Baker, S.K. Premature abandonment of global vitamin A supplementation programmes is
not prudent! Int. J. Epidemiol. 2015, 44, 297–299. [CrossRef] [PubMed]

20. Benn, C.S.; Fisker, A.B.; Aaby, P. Response to: J Mason et al. Vitamin A policies need 447 rethinking.
Int. J. Epidemiol. 2015, 44, 366–367. [CrossRef] [PubMed]

21. Global Alliance for Vitamin A (GAVA). Technical Consultation on Guidance to Vitamin A Supplementation
Programs for Children 6–59 Months of Age; GAVA: Ottawa, ON, Canada, 2012.

22. United Nations International Children’s Fund. Data: Monitoring the Situation of Children and Women. 2016.
Available online: http://data.unicef.org/nutrition/vitamin-a.html (accessed on 17 February 2017).

23. Vitamin and Mineral Nutrition Information System. Summary Tables and Maps on Global Prevalence of
Vitamin A Deficiency 1995–2005. Available online: http://www.who.int/vmnis/database/vitamina/status/
en/ (accessed on 17 February 2017).

24. Global Alliance for Improved Nutrition. Fortifying Our Future—A Snapshot on Food Fortification. 2015.
Available online: http://www.gainhealth.org/wp-content/uploads/2015/05/Fortifying-our-Future-A-
SnapShot-Report-on-Food-Fortification1.pdf (accessed on 17 February 2017).

25. Global Alliance for Improved Nutrition. Programs—Global Tracking. 2016. Available online: http://www.
gainhealth.org/programs/initiatives/#global-tracking (accessed on 17 February 2017).

26. World Health Organization. Guidelines for Food Fortification. 2006. Available online: http://www.who.int/
nutrition/publications/guide_food_fortification_micronutrients.pdf (accessed on 17 February 2017).

27. Sight and Life: Food Fortification in Africa. Available online: http://www.sightandlife.org/fileadmin/data/
Magazine/2013/27_3_2013/food_fortification_in_africa.pdf2013 (accessed on 17 February 2017).

28. Klemm, R.D.W.; West, K.P.; Palmer, A.C.; Johnson, Q.; Randall, P.; Ranum, P.; Northrop-Clewes, C. Vitamin A
fortification of wheat flour: Considerations and current recommendations. Food Nutr. Bull. 2010, 31,
S47–SS61. [CrossRef] [PubMed]

29. Food Fortification Initiative. Global Progress. 2016. Available online: http://www.ffinetwork.org/global_
progress/index.php (accessed on 17 February 2017).

30. United Nations International Children’s Fund; Centers for Disease Control and Prevention. UNICEF-CDC
Global Assessment of Home Fortification Interventions, 2011 Micronutrient Powders Data; Home Fortification
Technical Advisory Group: Geneva, Switzerland, 2012.

31. United Nations Children’s Fund. NutriDash Platform-2016; United Nations Children’s Fund: New York, NY,
USA, in press.

32. HarvestPlus. Nutritious Staple Food Crops: Who Is Growing What? International Food Policy Research Institute:
Washington, DC, USA, 2014.

33. World Health Organization. Serum Retinol Concentrations for Determining the Prevalence of Vitamin
A Deficiency in Populations. Vitamin and Mineral Nutrition Information System. Available online:
http://www.who.int/vmnis/indicators/retinol.pdf (accessed on 17 February 2017).

34. United Nations Children’s Fund. Vitamin A Supplementation: A Decade of Progress. 2007. Available online:
http://www.unicef.org/publications/files/Vitamin_A_Supplementation.pdf (accessed on 17 February 2017).

35. United Nations Children’s Fund. National Nutrition Survey Afghanistan (2013); Ministry of Public Health
(MoPH) and UNICEF: Karachi, Pakistan, 2013.

36. United Nations Children’s Fund. Assessing Vitamin A and Iron Deficiency Anaemia, Nutritional Anaemia among
Children Aged 0–60 Months in the Republic of Angola (1998–1999); UNICEF: Luanda, Angola, 2000.

37. Ministry of Health, United Nations Children’s Fund. Azerbaijan Nutrition Survey (AzNS), 2013; The U.S.
Agency for International Development; World Bank Group; UNICEF: Baku, Azerbaijan, 2014.

38. The International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR’B). National Micronutrients
Status Survey 2011–12; Final Report; Centre for Nutrition and Food Security; UNICEF; GAIN; Institute of
Public Health and Nutrition: Dhaka, Bangladesh, 2013.

Nutrients 2017, 9, 190 15 of 18

39. United Nations Children’s Fund. Benin: Enquete Nationale sur la Carence en Vitamine a et la Disponibilite en
sel Iode Dans les Menages 2000; Ministere de la Sante Publique; Direction de la Sante Familiale; The United
Nations Children’s Fund; USAID: New York, NY, USA, 1999.

40. United Nations Children’s Fund. United Nations Children’s Fund. Micronutrient Malnutrition in Botswana.
In A National Survey to Assess the Status of Iodine, Iron, and Vitamin A, 1996; Ministry of Health Wellness:
Gaborone, Botswana, 1996.

41. United Nations Children’s Fund. United Nations Children’s Fund. Burundi: Rapport de l’enquete Nationale
de Nutrition de la Population, 2005. In Ministere de la Sante Publique, Programme de Lutte Contre les Maladies
Transmissibles es Carentielles, 2006; The United Nations Children’s Fund: New York, NY, USA, 2005.

42. National Institute of Statistics; Directorate General for Health; ICF International. Cambodia Demographic
and Health Survey, 2014; National Institute of Statistics; Directorate General for Health; The DHS Program:
Phnom Penh, Cambodia, 2015.

43. Engle-Stone, R.; Ndjebayi, A.O.; Nankap, M.; Brown, K.H. Consumption of potentially fortifiable foods by
women and young children varies by ecological zone and socio-economic status in Cameroon. J. Nutr. 2012,
142, 555–565. [CrossRef] [PubMed]

44. United Nations Children’s Fund. Republique Centrafricaine, Enquête Nationale sur L’avitaminose A, la Carence
en fer et la Consommation du sel iode, 2000; Ministere Delegue a l’Economie au Plan et a la Cooperation
Internationale; Ministere de la Sante et de la Population; The United Nations Children’s Fund: New York,
NY, USA, 2000.

45. Rohner, F.; Northrop-Clewes, C.; Tschannen, A.B.; Bosso, P.E.; Kouassi-Gohou, V.; Erhardt, J.G.; Bui, M.;
Zimmermann, M.B.; Mascie-Taylor, C.G. Prevalence and public health relevance of micronutrient deficiencies
and undernutrition in pre-school children and women of reproductive age in Cote d’Ivoire, West Africa.
Public Health Nutr. 2014, 17, 2016–2028. [CrossRef] [PubMed]

46. Ministère de la Santé; Bureau National TDCI; UNICEF; CEPLANUT. Importance de la Carence en Vitamine A en
Republique Democratique du Congo; Ministère de la Santé: Kinshasa, Democratic Republic of the Congo, 2000.

47. The United Nations International Children’s Fund (UNICEF). National Survey for Assessment of Vitamin A
Status in Egypt, 1995; Nutrition Institute, The United Nations International Children’s Fund: New York, NY,
USA, 1995.

48. Demissie, T.; Ali, A.; Mekonen, Y.; Haider, J.; Umeta, M. Magnitude and distribution of vitamin A deficiency
in Ethiopia. Food Nutr. Bull. 2010, 31, 234–241. [CrossRef] [PubMed]

49. Bah, A.; Semega-Janneh, I.; Prentice, A.; Bates, C. Nationwide Survey on the Prevalence of Vitamin A and Iron
Deficiency in Women and Children in the Gambia; National Agency: Lyon, France, 2001.

50. Ministerio de Salud Publica y Asistencia Social (MSPAS). Guatemala: Encuesta Nacional de Micronutrientes
2009–2010, ENMICRON; Ministerio de Salud Publica y Asistencia Social: Guatemala City, Guatemala, 2012.

51. Ministere de la Sante Publique et de la Population. Enquete sur la Prevlence de la Carence en Vitamine A et de la
Deficience end iode end Haiti; Ministere de la Sante Publique et de la Population; United Nations Children’s
Fund: Port-au-Prince, Haiti, 2005.

52. Nestel, P.; Melara, A.; Rosado, J.; Mora, J.O. Vitamin A deficiency and anemia among children 12–71 months
old in Honduras. Rev. Panam. Salud Publica 1999, 6, 34–43. [CrossRef] [PubMed]

53. Sandjaja, S.; Budiman, B.; Harahap, H.; Ernawati, F.; Soekatri, M.; Widodo, Y.; Sumedi, E.; Rustan, E.;
Sofia, G.; Syarief, S.N.; Khouw, I. Food consumption and nutritional and biochemical status of 0.5–12-year-old
Indonesian children: the SEANUTS study. Br. J. Nutr. 2013, 110, S11–S20. [CrossRef] [PubMed]

54. Kenya Medical Research Institute and Ministry of Health. The Kenya National Micronutrient Survey 2011;
Kenya Medical Research Institute and Ministry of Health: Nairobi, Kenya, 2016.

55. Ministry of Health, National Statistics Committee, CDC, UNICEF. Follow-Up Survey of Nutritional Status
in Children 6–29 Months of Age, Kyrgyz Republic 2013; United Nations Children’s Fund: Bishkek, Kyrgyz
Republic, 2015.

56. Ministry of Health, National Institute of Public Health. Report on National Health Survey: Health Status of the
People of LAO PDR; Ministry of Health: Vientiane, Lao PDR, 2001.

57. Wolde-Gebriel, Z. National Survey on Iodine, Vitamin A and Iron Status of Women and Children in Lesotho;
Ministry of Health, UNICEF: Maseru, Lesotho, 1994.

58. United Nations Children’s Fund. Liberia National Micronutrient Survey 2011- Selected Preliminary Findings;
The United Nations Children’s Fund; Liberia Institute of Statistics: Monrovia, Liberia, 2011.

Nutrients 2017, 9, 190 16 of 18

59. U.S. Agency for International Development Micronutrient Progam (MOST). Enquête sur la Carence en
Vitamine A chez les Femmes et les Enfants et Enquête sur l’Anémie chez les Ecoliers de 6 à 14 Ans-Madagascar
2000. Available online: http://pdf.usaid.gov/pdf_docs/Pnadc603.pdf (accessed on 17 February 2017).

60. Ministry of Health. Malawi- National Micronutrient Survey 2009; Department of Nutrition, HIV and AIDS
in the Office of President and Cabinet (DHNA-OPC); Ministry of Health (MOH); National Statistics Office
(NSO); UNICEF; CDC: Lilongwe, Malawi, 2011.

61. United Nations Children’s Fund. National Micronutrient Survey 2007—Maldives. Available online:
http://www.unicef.org/maldives/micro_final_report-15_Aug_2010.pdf (accessed on 17 February 2017).

62. Palafox, N.A.; Gamble, M.V.; Dancheck, B.; Ricks, M.O.; Briand, K.; Semba, R.D. Vitamin A deficiency, iron
deficiency, and anemia among preschool children in the Republic of the Marshall Islands. Nutrition 2003, 19,
405–408. [CrossRef]

63. Villalpando, S.; Cruz Vde, L.; Shamah-Levy, T.; Rebollar, R.; Contreras-Manzano, A. Nutritional status of
iron, vitamin B12, folate, retinol and anemia in children 1 to 11 years old: Results of the Ensanut 2012.
Salud Publica Mex. 2015, 57, 372–384. [CrossRef] [PubMed]

64. Ministry of Health. Nutrition Status of Mongolian Population; Fourth National Nutrition Survey Report, 2010;
Minstry of Health; Public Health Institute; Nutrition Research Centre; UNICEF; WHO: Ulaanbaatar,
Mongolia, 2011.

65. Nasri, I.; El Bouhali, B.; Aguenaou, H.; Mokhtar, N. Vitamin A deficiency among Moroccan women and
children. Afr. Health Sci. 2004, 4, 3–8. [PubMed]

66. United Nations Children’s Fund. Mozambique: National Survey on Vitamin A Deficiency, Anemia, and Malaria;
The United Nations Children’s Fund; Ministry of Health; World Health Organization: Geneva, Switzerland, 2002.

67. United Nations Children’s Fund. Iodine Deficiency Disorders in Namibia and Data in the Status of Vitamin A and
Iron; International Council for the Control of Iodine Deficiency Disorders (ICCIDD); Ministry of Health and
Social Services (Namibia); United Nations Children’s Fund: New York, NY, USA, 1992.

68. Gorstein, J.; Shreshtra, R.K.; Pandey, S.; Adhikari, R.K.; Pradhan, A. Current status of vitamin A deficiency
and the National Vitamin A Control Program in Nepal: results of the 1998 National Micronutrient
Status Survey. Asia Pac. J. Clin. Nutr. 2003, 12, 96–103. [PubMed]

69. Gurdián, M.; Kontorovsky, I.; Alvarado, E.; Ramírez, S.; Hernández, R. Sistema Integrado de Vigilancia
de Intervenciones Nutricionales (SIVIN). In Informe de Progreso. Nicaragua, Ministerio de Salud, 2005;
Comision Tecnica de Seguridad Alimentaria y Nutricional: Managua, Nicaragua, 2005.

70. Maziya-Dixon, B.B.; Akinyele, I.O.; Sanusi, R.A.; Oguntona, T.E.; Nokoe, S.K.; Harris, E.W. Vitamin A
deficiency is prevalent in children less than 5 years of age in Nigeria. J. Nutr. 2006, 136, 2255–2261. [PubMed]

71. Bhutta, Z.A. National Nutrition Survey, Pakistan, 2011. Available online: http://gilanifoundation.
com/homepage/Free_Pub/HEALTH/National_Nutrition_Survey_of_Pakistan_2011.pdf (accessed on
17 February 2017).

72. Department of Public Health. National Micronutrient Survey Papua New Guinea, 2005—Summary Report;
UNICEF; Department of Public Health of Papua New Guinea; University of Papua New Guinea; Center for
Disease Control and Prevention: Port Moresby, Papua New Guinea, 2006.

73. Food and Nutrition Research Institute (FNRI); Metro Manila, Philippines. National Nutrition Survey
Philippines 2013. Personal communication, 2013.

74. United Nations Children’s Fund. National Nutrition Survey of Women and Children in Rwanda in 1996;
Ministry of Health (Rwanda); United Nations Children’s Fund; World Health Organization (WHO): Kigali,
Rwanda, 1997.

75. Laboratoire de Nutrition-Université Cheikh Anta DIOP (UCAD), Comité Sénégalais Pour la Fortification des
Aliments en Micronutriments (COSFAM), Micronutrient Initiative. Situation de Base du Statut en Vitamine A et
en fer chez les Enfants de 12–59 Mois et Chez les Femmes en age de Procreer (15–49 ans) Dans le Cadre du Programme
de Fortification des Aliments en Micronutriments au Senegal; Republique du Senegal: Dakar, Senegal, 2011.

76. Ministry of Health and Sanitation. 2013 Sierra Leone Micronutrient Survey. Available online:
http://groundworkhealth.org/wp-content/uploads/2016/03/SLMS-Report_FINAL_151203.pdf (accessed
on 17 February 2017).

77. Food Security and Nutrition Analysis Unit. National Micronutrient and Anthropometric Nutrition Survey,
Somalia 2009; Food Security and Nutrition Analysis Unit; SIDA; WHO; UNICEF; Ministry of Health:
Wellington, Australia, 2009.

Nutrients 2017, 9, 190 17 of 18

78. Shisana, O.; Labadarios, D.; Rehle, T.; Simbayi, L.; Zuma, K.; Dhansay, A.; Reddy, P.; Parker, W.; Hoosain, E.;
Naidoo, P.; et al. The South African National Helath and Nutrition Examination Survey, 2012 SANHANES-1: The
Health and Nutritional Status of the Nation; HSRC Press: Cape Town, South Africa, 2014.

79. United Nations Children’s Fund. Vitamin A Nutrition Status in Sri Lanka 2006; Ministry of Healthcare and
Nutrition; The United Nations Children’s Fund: New York, NY, USA, 2006.

80. National Bureau of Statistics (NBS). Micronutrients: Results of the 2010 Tanzania Demographic and Health
Survey. Available online: http://pdf.usaid.gov/pdf_docs/PNADY494.pdf (accessed on 17 February 2017).

81. United Nations Children’s Fund. Timor-Leste Food and Nutrition Survey 2013, Ministry of Health Final Report;
United Nations Chilren’s Fund; Ministry of Health: New York, NY, USA, 2015.

82. Uganda Bureau of Statistics; ICF International. Uganda Demographic and Health Survey 2011—Addendum to
Chapter 11; Uganda Bureau of Statistics; ICF International: Kampala, Uganda, 2012.

83. Laillou, A.; Pham, T.V.; Tran, N.T.; Le, H.T.; Wieringa, F.; Rohner, F.; Fortin, S.; Le, M.B.; Tran do, T.;
Moench-Pfanner, R.; et al. Micronutrient deficits are still public health issues among women and young
children in Vietnam. PLoS ONE 2012, 7, e34906. [CrossRef] [PubMed]

84. USAID. Report of the National Survery to Evaluate the Impact of Vitamin A Interventions in Zambia, July and
November 2003; Micronutrient Operational Strategies and Technologies (MOST); Centers for Disease Control
and Prevention (CDC); Food and Nutrition Commision of Zambia; University of Zambia; The United
Nations Children’s Fund: New York, NY, USA, 2003.

85. United Nations Children’s Fund. Zimbabwe National Micronutrient Survey: 1999; Ministry of Health and Child
Welfare; Nutrition Unit; The United Nations International Children’s Fund: New York, NY, USA, 2001.

86. Tanumihardjo, S.A. Vitamin A fortification efforts require accurate monitoring of population vitamin A
status to prevent excessive intakes. Procedia Chem. 2015, 14, 398–407. [CrossRef]

87. Sablah, M.; Klopp, J.; Steinberg, D.; Touaoro, Z.; Laillou, A.; Baker, S. Thriving public-private partnership
to fortify cooking oil in the West African Economic and Monetary Union (UEMOA) to control vitamin A
deficiency: Faire Tache d’Huile en Afrique de l’Ouest. Food Nutr. Bull. 2012, 33, S310–S320. [CrossRef]
[PubMed]

88. Rohner, F.; Leyvraz, M.; Konan, A.G.; Esso, L.J.C.E.; Wirth, J.P.; Norte, A.; Adiko, A.F.; Bonfoh, B.; Aaron, G.J.
The Potential of Food Fortification to Add Micronutrients in Young Children and Women of Reproductive
Age—Findings from a Cross-Sectional Survey in Abidjan, Cote d’Ivoire. PLoS ONE 2016, 11, e0158552.
[CrossRef] [PubMed]

89. Luthringer, C.L.; Rowe, L.A.; Vossenaar, M.; Garrett, G.S. Regulatory Monitoring of Fortified Foods:
Identifying Barriers and Good Practices. Glob. Health Sci. Pract. 2015, 3, 446–461. [CrossRef] [PubMed]

90. Tanumihardjo, S.A.; Russe, R.M.; Stephensen, C.B.; Gannon, B.M.; Craft, N.E.; Haskell, M.J.; Lietz, G.;
Schulze, K.; Raiten, D.J. Biomarkers of Nutrition for Development (BOND)-Vitamin A Review. J. Nutr. 2016,
146, 1816–1848. [CrossRef] [PubMed]

91. Raiten, D.J.; Ashour, F.A.S.; Ross, A.C.; Meydani, S.N.; Dawson, H.D.; Stephensen, C.B.; Brabin, B.J.;
Suchdev, P.S.; van Ommen, B.; Grp, I.C. Inflammation and Nutritional Science for Programs/Policies and
Interpretation of Research Evidence (INSPIRE). J. Nutr. 2015, 145, 1039s–1108s. [CrossRef] [PubMed]

92. Thurnham, D.I.; Northrop-Clewes, C.A.; Knowles, J. The Use of Adjustment Factors to Address the Impact of
Inflammation on Vitamin A and Iron Status in Humans. J. Nutr. 2015, 145, 1137s–1143s. [CrossRef] [PubMed]

93. Larson, L.; Addo, O.Y.; Sandalinas, F.; Baawo, K.; Faigao, K.; Kupka, R.; Flores-Ayala, R.; Suchdev, P.
Accounting for the Influence of Inflammation on Retinol Binding Protein in a Population Survey of Liberian
Preschool-Aged Children. FASEB J. 2016. [CrossRef]

94. Greig, A.; Neufeld, L. A Framework for Shifting from Universal Vitamin A Supplementation: Technical
Consultation on Guidance for VAS Programs for Children 6–59 Months of Age. Sight Life 2013, 27, 76–77.

95. Tanumihardjo, S.A.; Mokhtar, N.; Haskell, M.J.; Brown, K.H. Assessing the Safety of Vitamin A
Delivered Through Large-Scale Intervention Programs: Workshop Report on Setting the Research Agenda.
Food Nutr. Bull. 2016, 37, S63–S74. [CrossRef] [PubMed]

96. Sandjaja, J.I.; Jahari, A.B.; Ifrad, H.M.K.; Tilden, R.L.; Soekarjo, D.; Utomo, B.; Moench-Pfanner, R.; Soekirman;
Korenromp, E.L. Vitamin A-fortified cooking oil reduces vitamin A deficiency in infants, young children
and women: results from a programme evaluation in Indonesia. Public Health Nutr. 2015, 18, 2511–2522.
[CrossRef] [PubMed]

Nutrients 2017, 9, 190 18 of 18

97. Global Alliance for Improved Nutrition. Fortification of Vegetable Oil in Indonesia. Available online:
http://www.gainhealth.org/knowledge--centre/project/fortification--vegetable--oil--indonesia/ (accessed
on 17 February 2017).

98. Home Fortification Technical Advisory Group. Home Fortification with Micronutrient Powders (MNP).
Available online: http://hftag.org/2157_Projects.asp?tax_product_type=mnp&wpas=1&product--label=
Micronutrient%20powders%20(MNP) (accessed on 17 February 2017).

99. Centers for Disease Control and Prevention: International Micronutrient Malnutrition Prevention and
Control (IMMPaCt) —Kyrgyzstan. 2015. Available online: http://www.cdc.gov/immpact/projects/
kyrgyzstan.html (accessed on 17 February 2017).

100. Hotz, C.; Chileshe, J.; Siamusantu, W.; Palaniappan, U.; Kafwembe, E. Vitamin A intake and infection are
associated with plasma retinol among pre-school children in rural Zambia. Public Health Nutr. 2012, 15,
1688–1696. [CrossRef] [PubMed]

101. Bresnahan, K.A.; Chileshe, J.; Arscott, S.; Nuss, E.; Surles, R.; Masi, C.; Kafwembe, E.; Tanumihardjo, S.A.
The Acute Phase Response Affected Traditional Measures of Micronutrient Status in Rural Zambian Children
during a Randomized, Controlled Feeding Trial. J. Nutr. 2014, 144, 972–978. [CrossRef] [PubMed]

102. Gannon, B.; Kaliwile, C.; Arscott, S.A.; Schmaelzle, S.; Chileshe, J.; Kalungwana, N.; Mosonda, M.; Pixley, K.;
Masi, C.; Tanumihardjo, S.A. Biofortified orange maize is as efficacious as a vitamin A supplement in
Zambian children even in the presence of high liver reserves of vitamin A: A community-based, randomized
placebo-controlled trial. Am. J. Clin. Nutr. 2014, 100, 1541–1550. [CrossRef] [PubMed]

103. Tanumihardjo, S.A.; Gannon, B.M.; Kaliwile, C.; Chileshe, J. Hypercarotenodermia in Zambia: Which
children turned orange during mango season? Eur. J. Clin. Nutr. 2015, 69, 1346–1349. [CrossRef] [PubMed]

104. Palmer, A.C.; Siamusantu, W.; Chileshe, J.; Schulze, K.J.; Barffour, M.; Craft, N.E.; Molobeka, N.;
Kalungwana, N.; Arguello, M.A.; Mitra, M.; et al. Provitamin A-biofortified maize increases serum
beta-carotene, but not retinol, in marginally nourished children: A cluster-randomized trial in rural Zambia.
Am. J. Clin. Nutr. 2016, 104, 181–190. [CrossRef] [PubMed]

105. Sesay, F.F.; Hodges, M.H.; Kamara, H.I.; Turay, M.; Wolfe, A.; Samba, T.T.; Koroma, A.S.; Kamara, W.; Fall, A.;
Mitula, P.; et al. High coverage of vitamin A supplementation and measles vaccination during an integrated
Maternal and Child Health Week in Sierra Leone. Int. Health 2015, 7, 26–31. [CrossRef] [PubMed]

106. Klemm, R.D.W.; Palmer, A.C.; Greig, A.; Engle-Stone, R.; Dalmiya, N. A Changing Landscape for Vitamin A
Programs: Implications for Optimal Intervention Packages, Program Monitoring, and Safety. Food Nutr. Bull.
2016, 37, S75–S86. [CrossRef] [PubMed]

107. Tanumihardjo, S.; Ball, A.; Kaliwile, C.; Pixley, K. The research and implementation continuum of biofortified
sweetpotato and maize in Africa. Ann. N. Y. Acad. Sci. 2017. [CrossRef] [PubMed]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).