Determinants of vitamin a deficiency in children between 6 months and 2 years of age in Guinea-Bissau
© Danneskiold-Samsøe et al; licensee BioMed Central Ltd. 2013
Received: 29 May 2012
Accepted: 17 January 2013
Published: 25 February 2013
The World Health Organization (WHO) classifies Guinea-Bissau as having severe vitamin A deficiency (VAD). To date, no national survey has been conducted. We assessed vitamin A status among children in rural Guinea-Bissau to assess status and identify risk factors for VAD.
In a vitamin A supplementation trial in rural Guinea-Bissau, children aged 6 months to 2 years who were missing one or more vaccines were enrolled, vaccinated and randomized to vitamin A or placebo. Provided consent, a dried blood spot (DBS) sample was obtained from a subgroup of participants prior to supplementation. Vitamin A status and current infection was assessed by an ELISA measuring retinol-binding protein (RBP) and C-reactive protein (CRP). VAD was defined as RBP concentrations equivalent to plasma retinol <0.7 μmol/L; infection was defined as CRP >5 ml/L. In Poisson regression models providing prevalence ratios (PR), we investigated putative risk factors for VAD including sex, age, child factors, maternal factors, season (rainy: June-November; dry: December-May), geography, and use of health services.
Based on DBS from 1102 children, the VAD prevalence was 65.7% (95% confidence interval 62.9-68.5), 11% higher than the WHO estimate of 54.7% (9.9-93.0). If children with infection were excluded, the prevalence was 60.2% (56.7-63.7). In the age group 9–11 months, there was no difference in prevalence of VAD among children who had received previous vaccines in a timely fashion and those who had not. Controlled for infection and other determinants of VAD, the prevalence of VAD was 1.64 (1.49-1.81) times higher in the rainy season compared to the dry, and varied up to 2-fold between ethnic groups and regions. Compared with having an inactivated vaccine as the most recent vaccine, having a live vaccine as the most recent vaccination was associated with lower prevalence of VAD (PR=0.84 (0.74-0.96)).
The prevalence of VAD was high in rural Guinea-Bissau. VAD varied significantly with season, ethnicity, region, and vaccination status.
KeywordsVitamin A deficiency Children Guinea-Bissau Risk factors Retinol-binding protein
Vitamin A deficiency (VAD) is a public health problem in many low-income countries. Worldwide, the prevalence of VAD is estimated to be 190 million in preschool-age children  causing 1–2 million deaths annually . VAD is defined to be of public health importance if the national prevalence reaches 15%  using a serum or plasma retinol concentration < 0.7 μmol/L as cut-off for VAD [1, 4]. Many countries lack data on vitamin A status. In Guinea-Bissau, no national vitamin A status survey has been conducted. Lacking biochemical vitamin A data, VAD estimates for preschool-age children are based on a regression model including the proxy variables: gross domestic product, under-5-mortality, and population growth rate. Using this method, Guinea-Bissau has been classified by the World Health Organization (WHO) as having a severe public health problem with an estimated 54.7% (9.9-93.0) of preschool-age children suffering from VAD . Despite estimates of a high prevalence of VAD, little xerophthalmia has been observed in Guinea-Bissau . In the present study we assessed the vitamin A status and investigated risk factors for VAD among 6-23-month-old children due to be vaccinated in rural Guinea-Bissau.
At enrolment into the trial, after written informed consent was obtained by the present parent or guardian, the children were weighed and measured, and the mother was interviewed about whether she had given the child medicine and whether the child coughed, vomited or had diarrhea or fever on the day of enrolment.
We originally planned to collect blood samples from 600 children at enrolment, prior to randomization, to describe baseline vitamin A status among the children and study risk factors for VAD. The study was observational and no sample size calculations were made. In the beginning of the trial all eligible children in a village were included in the study. However, enrolment proceeded faster than expected and to describe the full seasonal and geographical variation we continued sampling five children from each village until samples from both the rainy (June-November) and the dry (December-May) season had been obtained in all regions. Hence, a total of 1254 samples were collected.
The study was in compliance with the Helsinki Declaration and was approved by the Ministry of Health in Guinea-Bissau. The Danish Central Ethical Committee provided its consultative approval.
Handling and processing of blood samples
Dried blood spots (DBS) were prepared by letting finger prick blood drip onto filter paper (903 Protein Saver Card, Whatman International, UK). The papers were left to dry in a horizontal position protected from sunlight. After drying for a few hours, samples were bagged individually together with desiccant (Silica gel, Merck, NJ, USA) and left at ambient temperature for 1–14 days. Using this procedure, retinol-binding protein (RBP) levels remain stable for at least two weeks . Upon arrival in the capital, samples were frozen at −20°C until analysis at the National Laboratory in Guinea-Bissau. Before analysis, samples were inspected for the presence of desiccant and dissociated plasma. Any plasma dissociation, seen as a yellow discoloration outside the DBS, or lack of desiccant was noted.
RBP and C-Reactive Protein (CRP) were measured using a previously developed sandwich enzyme-linked immunosorbent assay (ELISA) . In brief, wells in two 96-well plates (Maxisorb C-shape, Nunc) were coated with polyclonal anti-human RBP antibody (A0040, Dako Denmark), and polyclonal anti-human CRP antibodies (Q0329, Dako, Denmark) respectively. DBS were eluted in buffer overnight and added to the wells after washing. After incubation and another round of washing, a second anti-human RBP (ICL, Dako, Denmark) and CRP antibody (Abcam, Dako, Denmark) conjugated to horseradish peroxidase was added. After a third round of incubation and washing, tetramethylbenzidine was added. After development of sufficient color, the reaction was stopped by adding sulphuric acid. Finally, the concentration was measured by spectrophotometry. All concentrations were measured in duplicates. If the coefficient of variation between two RBP measurements differed by > 20%, a second duplicate measurement was carried out. If the CRP value was ≥ 3 and ≤ 7 mg/L and the coefficient of variation between two CRP measurements differed by > 20%, a second duplicate measurement was carried out. If the CRP value was ≥ 7 or ≤ 3 mg/L and the coefficient of variation > 20%, duplicates differing more than 10% in optical density compared to the highest measured point on the standard curve were re-measured. A HPLC measured plasma sample was added as a quality control to each plate. If the quality control differed by more than 20% from the concentration measured by HPLC, the entire plate was re-measured.
Determining vitamin a status
The blood concentration of retinol is under homeostatic control and remains stable over a large range of vitamin A storage concentrations . Retinol is bound to RBP in the blood in a ≈1:1 complex but saturation tends to be lower with low retinol concentrations [9–11].
We used the WHO definition of subclinical VAD, which is a plasma retinol concentration of < 0.7 μmol/L. Previous estimates found by paired measurements of plasma retinol by HPLC and RBP by ELISA showed that a cut-off of 0.83 μmol/L for RBP was equivalent to a cut-off of 0.7 μmol/L for plasma retinol [9, 12]. Infection lowers plasma retinol concentrations and acute phase proteins may be used to correct for overestimation of VAD due to infection . We defined infection as a plasma concentration of CRP > 5 mg/L.
To estimate the cut-offs for VAD and infection using DBS, a total of 85 paired plasma and DBS samples were obtained and measured for RBP and CRP by ELISA. The paired measurements were used to form receiver operating characteristic curves. The areas under the curves were 0.75 and 0.90 for RBP and CRP respectively. The Youden index  was used to find the DBS cut-off value of 0.749 which corresponded to the highest combined sensitivity and specificity with regard to identifying individuals with VAD defined as plasma RBP < 0.83. The same was done for CRP, identifying the DBS cut-off which corresponded best, in terms of sensitivity and specificity, to a plasma CRP > 5 mg/L.
Determining geography and distances
Global positioning system (GPS) coordinates were measured at hospitals, at vaccination facilities, and at the center of the village. Coordinates were loaded into ArcMap 9.3 (ESRI, Redlands, CA) and transformed into a Universal Transverse Mercator geographic coordinate system. An elevation of zero using the Shuttle Radar Topography Mission Digital Elevation Model (U.S. Geological Survey, Reston, VA) was used to define the sea. Euclidian distance in ArcMap was used to calculate the distances between villages and nearest vaccination facility, hospital, and the sea, respectively.
Z-scores for anthropometry
Z-scores for weight-for-age, length-for-age, weight-for-length, and arm circumference-for-age were calculated using the WHO health growth standards  and the WHO Anthro version 3.1, June 2010 macro for Stata .
First, we assessed the overall prevalence of VAD in all enrolled children. To correct for an overestimation due to infection, we subsequently conducted an analysis in which we omitted all children with infection. Next, in a sensitivity analysis, we assessed the prevalence of VAD using a 10% lower or 10% higher cut-off (0.749 ± 0.0749) in order to test the importance of the cut-off for VAD.
Since only children missing a vaccine according to the vaccination schedule were included in the study, the estimates of the prevalence of VAD may not entirely reflect vitamin A status in the population. The measles vaccine is supposed to be given at 9 months of age but children can be given this vaccine up to 12 months of age without this being considered as a late vaccination. We therefore defined a subgroup of children as being ”timely vaccinated” if they were 9–11 months of age, had received the third dose of DTP/Pentavalent vaccine before 7 months of age and were missing only MV (Figure 2). Next, we studied the prevalence of VAD within this subgroup and included it in our simple model for risk factors for VAD (see below).
We examined the risk factors for VAD in Poisson regression models, obtaining prevalence ratios (PR) for VAD . For all continuous variables, we investigated whether there was evidence of a linear relationship between the variable and the prevalence of VAD by inspecting the prevalence of VAD in quintiles of the variable. If the association looked linear we tested for departure from linearity by inclusion of a quadratic term. Where evidence against linearity was not found, the variable was included as a continuous variable; otherwise the variable was included as categorical variables. A linear relationship with VAD could not be shown for the variables: length-for-age, distance to sea, distance to hospital, and distance to vaccination place.
Risk factors for vitamin A deficiency in rural Guinea-Bissau
VAD Prevalence (%)
Simple modelaPR for VAD (95% CI)
Simple model pc<
Large modelbPR for VAD (95% CI)
Large model pc<
≤ 5 mg/L
> 5 mg/L
Weight for aged,e
Arm circumference for aged,e
Length for aged,e
≥ −2, < −1
≥ −1, < 0
Weight for lengthd,e
Took medicine on inclusion day
Cough at the day of inclusiond
Fever at the day of inclusiond
Diarrhea at the day of inclusiond
Maternal educational level (years of school)d
Maternal age at child inclusion dated
≥ 14, ≤ 20
> 20, ≤ 30
> 30, ≤ 40
> 40, ≤ 55
Village distance to vaccination facility (km)
≥ 0, < 0.5
≥ 0.5, < 2
≥ 2, < 5
Distance to sea (km)
≥ 0, < 5
25 9 (24)
≥ 5, < 15
≥ 15, < 50
≥ 50, < 149
Distance to hospital (km)
≥ 0, < 10
≥ 10, < 20
≥ 20, < 30
≥ 30, < 70
Use of health services
Most recent vaccination typed,h
VAS received within 6 monthsd
Child use of bed netd
All variables were tested one by one in a simple model, controlling only for infection. In a larger model we included the obligatory variables as well as all variables from the simple model, and used automated stepwise backwards selection to exclude variables to a descending value of p<0.20. Wald tests were used to test for overall equality of categories within variable. If a variable missed information for less than 2% of the participants, the missing category was excluded from both the simple and the large model. This was the case for the variables: weight for age, length for age, weight for length, breast feeding, cough at the day of inclusion, fever at the day of inclusion, diarrhea at the day of inclusion, maternal ethnicity, maternal educational level, maternal age, most recent vaccination type, VAS within 6 months, and child use of bed net (Additional file 1: Table S1).
To assess whether there was a temporal effect of previous VAS on vitamin A status we also tested the effect of including children who had received VAS within 2, 4, and 6 months prior to inclusion into the trial in the simple and the large model.
For all risk factors we evaluated whether the effect on VAD varied by sex and by infection status, respectively, by inclusion of an interaction term in both the simple and the large model. All interactions were evaluated by Wald statistics; a p-value below 0.05 was considered significant. Unless otherwise stated only the results of the large model are presented.
Finally we conducted a sensitivity analysis where we tested if the conclusions regarding risk factors for VAD changed if we a) excluded all children with infection, or b) used the 10% lower or higher cut-off for VAD.
All statistical analyses were conducted using Stata/SE 11.1.
A total of 1254 samples were collected between August 2007 and November 2009. Of the 1254 samples, eight samples were lost, two came from children who should not have been included, two samples did not have a corresponding inclusion questionnaire, 71 samples had been stored without desiccant, and 69 had dissociated plasma. Hence, 1102 samples were included in the study. Due to missing information on a variable (Additional file 1: Table S1) the final number of children included in the large model was 1050.
Overall prevalence of VAD and infection
The overall prevalence of VAD was 65.7% (95% CI=62.9-68.5). The prevalence of infection was 31.4% (28.7-34.4). If we excluded all children with infection, the prevalence of VAD was 60.2% (56.7-63.7). Using a 10% lower cut-off it was 57.5% (54.6-60.5) and using a 10% higher cut-off it was 73.5% (70.9-76.1).
Risk factors for VAD
Sex of the child had no significant effect on the prevalence of VAD but girls tended to have slightly lower prevalence of VAD, the PR for VAD being 0.93 (0.86-1.01) (Table 1). The risk of VAD tended to decrease with increasing age in the simple model. However, the marginal effect disappeared after adjustment (Table 1). As expected, having an infection was associated with an increased risk of VAD, the PR being 1.27 (1.17-1.37).
Arm circumference was associated with VAD in the simple model, but not in the large model after adjustment (Table 1). None of the other anthropometric variables were associated with VAD. Twins had a higher risk of VAD (the PR being 1.30 (1.11-1.52)), while breastfeeding had no effect on VAD. Cough on the inclusion day was positively associated with VAD (PR=1.10 (1.01-1.21)), whereas fever and diarrhea and use of medicine on the date of enrolment were not (Table 1).
Fula and Pepel ethnicity were associated with an increased prevalence of VAD compared with Balanta ethnicity (PR=1.21 (1.03-1.42) and 1.38 (1.05-1.81), respectively) (Table 1). Maternal educational level and age were not associated with VAD.
There were considerable differences in the prevalence of VAD between the nine health regions (Figure 1 and Table 1). The regions of Bijagós/Bolama, Gabu, and Tombali had a more than 2-fold higher prevalence of VAD compared to the region of Biombo which is closest to the capital (the PRs being 2.46 (1.79-3.39), 2.10 (1.53-2.87), 2.07 (1.51-2.83) respectively). A distance equal to or above 5 km to a vaccination facility did not increase the risk of VAD in the simple model but were associated with a higher risk in the large model (PR=1.17 (1.01-1.36)). Although distance to the sea was associated with a higher prevalence of VAD in the simple model (Table 1), no linear relationship with VAD was found, and due to confounding by region the variable was not included in the large model. Distance to hospital was associated with VAD, but in a non-linear manner (Table 1).
Use of health services
Having received a live vaccine as the most recent vaccine rather than an inactivated vaccine was associated with a lower risk of VAD (PR= 0.84 (0.74-0.96) (Table 1)). We did not find an effect on the risk of VAD of having received VAS within the previous 6 months (Table 1). VAS within 2 or 4 was not associated with significantly lower risk of VAD either (Additional file 1: Table S2). Hospital delivery was associated with an increased risk of VAD (PR=1.16 (1.03-1.31)) compared to home delivery. The eight children who did not use bed net at any time of the year had a higher risk of VAD than children who used bed nets all year (PR=1.48 (1.11-1.97)).
A total of 380 children were 9–11 months of age. Of these, 124 children were timely vaccinated. The prevalence of VAD was 61% in the group of timely vaccinated children and 69% in those who had not been timely vaccinated. The risk of VAD was not significantly different between the two groups (PR=0.92 (0.76-1.10) in the timely vaccinated versus the non-timely vaccinated (Additional file 1: Table S3).
Interactions between risk factors and sex and infection
We found no interactions between sex and any of the risk factors retained in the large model. There were significant interactions between infection and weight for age (p<0.0001), ethnicity (p<0.001), season (p<0.01), and region (p<0.001), respectively; in children with infection the associations between these risk factors and VAD were less pronounced.
If children with infection were excluded, retaining 756 children in the analysis, cough on inclusion became insignificant in both models, arm-circumference-for-age was no longer significantly associated with VAD in the simple model, and most recent vaccination type became statistically significant in the large model (Additional file 1: Table S4).
Using the low cut-off for VAD thereby retaining only the most deficient children by setting the cut-off for VAD lower than 0.83, identified the same main risk factors for VAD in the large model with a few exceptions. Twinning status was no longer a risk factor and birth facility, and distance to hospital was no longer risk factors in any of the models (Additional file 1: Table S5).
Using a high cut-off for VAD also identified the same main risk factors in the large model except for birth facility that was no longer significant in the large model, and the use of bed net that was no longer significant in the both models (Additional file 1: Table S6).
Based on the collected blood samples young children in rural Guinea-Bissau may have a major VAD problem. The overall prevalence of VAD was 66%, 11% higher than the WHO estimate. Infection, twinning, cough on inclusion day, Fula and Pepel ethnicity, rainy season, living in the regions of Bijagós/Bolama, Gabu, and Tombali, and hospital delivery were associated with an increased risk of VAD. Receiving a live vaccine as the most recent vaccine was associated with a lower prevalence of VAD compared with having received an inactivated vaccine as the most recent vaccine. These factors remained significant or became even more pronounced when children with infection were excluded (Additional file 1: Table S5).
Strengths and weaknesses
No national vitamin A status study has been conducted in Guinea-Bissau and the present study is the first to describe the determinants of VAD. Using a DBS assay approach allowed for the collection of samples from many areas within the country, which would not have been possible if we had used plasma or serum samples due to lack of electricity.
A limitation that may possibly contribute to an overestimation of VAD in the study is the incomplete correction for infection by not including α1-acid-glycoprotein, and/or another slow acute phase protein in the models .
In the present study we took advantage of the infrastructure of a trial and only children eligible for the trial were bled. Thus, children were between 6–23 months of age and missed at least one routine vaccination, and no children were included in the first month after VAS provided in a campaign. Children with incomplete vaccination history have been shown to have a higher risk of VAD in other settings . However, in our study population in rural Guinea-Bissau it is common to be missing one or more vaccines; in a recent study from the Bandim Health Project we found that half of the children at one year of age in Guinea-Bissau are missing at least one vaccine . Furthermore, 67% of children in the age group from 9–11 months of age had not yet received a measles vaccine at the time of the home visit, and we did not find any statistical difference in the risk of VAD between the children who had received all the prior vaccines timely and children who had not. Hence, though we cannot exclude a slight overestimation of VAD due to the focus on children missing one or more vaccines, we believe that the study reflects the situation in rural Guinea-Bissau.
Since study children did not receive VAS within the preceding month, a short-lived increase in serum RBP after supplementation in a campaign is not likely to have affected the results . On the other hand, we might have overestimated the prevalence of VAD by not including this group. However, we were not able to document any effect of receiving VAS from 1 to 2 months before inclusion.
All geographic areas were represented in the sample. However, the included populations did not completely correspond to the respective population sizes in the various regions. Due to the inter-regional variation in risk of VAD this could have affected the overall estimate of VAD prevalence. However, calculating a weighted average the combined prevalence from all regions was 67%. The four months with the lowest prevalence of VAD corresponded to 48.2% of the study population, which will tend to underestimate the prevalence of VAD.
In conclusion, we may not have obtained a precise estimate of the true prevalence of VAD in the children in Guinea-Bissau, but neither setting the cut-off for VAD lower, nor exclusion of infected children or estimating the prevalence of VAS in timely vaccinated children resulted in a prevalence below 57.5% as estimated by the WHO. The analysis of determinants of VAD was also robust in the sensitivity analyses.
Consistency with other studies
Several studies of vitamin A status in children have investigated determinants for VAD. Since the Bandim Health Project have previously observed sex differences in the effect of vitamin A supplementation [5, 21–23] sex was included as an obligatory variable in the analysis. However, most studies have found no sex differences of VAD [9, 24, 25] and our study is in line with this, though with a tendency towards lower VAD in girls.
In countries where VAD is a public health problem, the WHO recommends VAS every 4–6 months . Several trials have evaluated the temporal effect of supplementation on VAD, but none including children within the study age group. We found only two studies investigating the effect of VAS on vitamin A status in the present age group. An observational study from Ethiopia reported increased VAD in children who had missed one or more doses of VAS within the last year . The other observational study collected paired blood samples before and after supplementation and found a significant increase in plasma retinol 4 months after VAS . We could not document any effect of missing VAS within the preceding 2, 4 or 6 months. Possible explanations could be geographical or methodological differences between the studies. The study from Ethiopia also reported an effect on VAD of missing a vaccine and coming from a Muslim household. We did not find a consistent effect of coming from a Muslim household. The Fulas are Muslims and had a high prevalence of VAD, but so are the Mandingas, who did not have a high prevalence of VAD.
Seasonal differences in vitamin A status have been documented in West Africa before, both in children  and in lactating and pregnant mothers [29, 30]. Our results also show strong seasonal differences, and these were particularly pronounced in children without infection.
We have previously shown that the inactivated DTP vaccine given after neonatal VAS was associated with reduced RBP levels in 4 months old girls . In the present study the prevalence of VAD was higher if the most recent vaccine was DTP compared with a live vaccine (MV, BCG, OPV), but the effect did not differ by sex in these older children.
The important dietary sources of pre-formed vitamin A are animal products such as egg and liver. In addition, β-carotene, which can be found in orange-yellow vegetables and fruits such as mango as well as in red palm oil [31–33] can be converted to retinoids . All of the above foods can be found in Guinea-Bissau. We found strong seasonal variation in the prevalence of VAD and a possible explanation could be food availability. Mango supply might partly explain seasonal variation in vitamin A status, either directly or through breastfeeding [29, 35] although the mango season starts a month later than the months with a low prevalence of VAD. The decreased prevalence of VAD in the dry season could also be explained by the harvest in December-January. Underestimation of infection may also partly explain why season is a strong predictor but this seems unlikely since no significant difference in the prevalence of a high CRP between seasons was found (30% (27–34) and 33% (29–37) in the dry and rainy season, respectively). It seems likely that vitamin A status follows the intake of vitamin A rich foods and to a lesser degree VAS since season was a strong predictor for VAD, but VAS within the preceding months was not. If the WHO guidelines were to be followed, special attention should be given to improving vitamin A status in risk situations, e.g. in the rainy season. However, several studies from Guinea-Bissau have shown that supplementing with vitamin A in the rainy season may have a negative impact on overall mortality [20, 36]. This emphasizes that VAS may have other effects than merely prevention of VAD.
The reason for the strong regional differences is not clear. Distance to the sea was confounded by region and might explain the high risk of VAD in the region of Gabu. We have no explanation why hospital deliveries were associated with a higher risk of VAD; this might be a spurious finding.
The increased risk of VAD amongst twins could possibly be explained by a lower nutritional status since twins are more often born low-birth-weight and run a higher risk of being stunted . However, due to the low number of twins (n=36) the result should be interpreted with caution.
The lower prevalence of VAD among children who had a live vaccine as their most recent vaccine supports previous observations that live vaccines are associated with beneficial effect on child health [38, 39].
We found that around two-thirds of children between 6 months and 2 years of age have RBP levels indicating VAD. Though we only included children who were not fully vaccinated, this accounts for a large proportion of the children in rural Guinea-Bissau, and there was no indication that children vaccinated timely had significantly lower risk of VAD. Though we may not have been able to control fully for the effect of infections, the results still point to VAD as a major public health problem in rural Guinea-Bissau. In our analysis of determinants we found that major risk factors were season, region, ethnicity, and vaccination status. The prevalence of VAS varied up to 2-fold between different groups, hence it is possible to identify subgroups of children at particular risk. In light of the missing effects of VAS on vitamin A status it may be better to opt for other methods to improve vitamin A status, i.e. through fortification, promoting vitamin A rich foods or more frequent low-dose supplementation.
Dried Blood Spot
Enzyme-Linked ImmunoSorbent Assay
Global Positioning System
Health and Demographic Surveillance System
Diphtheria-tetanus-pertussis-hepatitis B-H. influenzae type B
Retinol Binding Protein
Vitamin A deficiency
Vitamin A supplementation
World Health Organization.
The authors would like to thank The Oak Foundation, Øllingesøefonden, and The Danish International Development Agency (DANIDA) for their support and Alison Keenan for her language revision.
- WHO: Global prevalence of vitamin A deficiency in populations at risk 1995–2005. 2009, Geneva, http://whqlibdoc.who.int/publications/2009/9789241598019_eng.pdf,Google Scholar
- Humphrey JH, West KP, Sommer A: Vitamin A deficiency and attributable mortality among under-5-year-olds. Bull World Health Organ. 1992, 70: 225-232. http://whqlibdoc.who.int/bulletin/1992/vol70-No2/bulletin_1992_70(2)_225-232.pdf,PubMedPubMed CentralGoogle Scholar
- Sommer A, Davidson FR: Assessment and control of vitamin A deficiency: the Annecy Accords. J Nutr. 2002, 132: 2845-2850. http://jn.nutrition.org/content/132/9/2845S.full,Google Scholar
- WHO: Indicators for assessing vitamin A deficiency and their application in monitoring and evaluating intervention programmes. 1996, Geneva, http://www.who.int/nutrition/publications/micronutrients/vitamin_a_deficiency/WHONUT96.10.pdf,Google Scholar
- Benn CS, Aaby P, Balé C, Olsen J, Michaelsen KF, George E, Whittle H: Randomised trial of effect of vitamin A supplementation on antibody response to measles vaccine in Guinea-Bissau, west Africa. Lancet. 1997, 350: 101-105. 10.1016/S0140-6736(96)12019-5. http://dx.doi.org/10.1016/S0140-6736(96)12019-5,View ArticlePubMedGoogle Scholar
- Fujita M, Brindle E, Shofer J, Ndemwa P, Kombe Y, Shell-Duncan B, O'Connor KA: Retinol-binding protein stability in dried blood spots. Clin Chem. 2007, 53: 1972-1975. 10.1373/clinchem.2007.093104. http://dx.doi.org/10.1373/clinchem.2007.093104,View ArticlePubMedGoogle Scholar
- Erhardt JG, Estes JE, Pfeiffer CM, Biesalski HK, Craft NE: Combined measurement of ferritin, soluble transferrin receptor, retinol binding protein, and C-reactive protein by an inexpensive, sensitive, and simple sandwich enzyme-linked immunosorbent assay technique. J Nutr. 2004, 134: 3127-3132. http://jn.nutrition.org/content/134/11/3127.long,PubMedGoogle Scholar
- Thurnham DI, Northrop-Clewes CA: Optimal nutrition: vitamin A and the carotenoids. P Nutr Soc. 1999, 58: 449-457. 10.1017/S0029665199000592. http://dx.doi.org/10.1017/S0029665199000592,View ArticleGoogle Scholar
- Fisker AB, Lisse IM, Aaby P, Erhardt JG, Rodrigues A, Bibby BM, Benn CS: Effect of vitamin A supplementation with BCG vaccine at birth on vitamin A status at 6 wk and 4 mo of age. Am J Clin Nutrit. 2007, 86: 1032-1039. http://www.ajcn.org/content/86/4/1032.long,PubMedGoogle Scholar
- Gamble MV, Ramakrishnan R, Palafox NA, Briand K, Berglund L, Blaner WS: Retinol binding protein as a surrogate measure for serum retinol: studies in vitamin A-deficient children from the Republic of the Marshall Islands. Am J Clin Nutrit. 2001, 73: 594-601. http://www.ajcn.org/content/73/3/594.long,PubMedGoogle Scholar
- Semba RD, Yuniar Y, Natadisastra G, Muhilal , Gamble MV: Assessment of vitamin A status of preschool children in Indonesia using plasma retinol-binding protein. J Trop Pediatr. 2002, 48: 84-87. 10.1093/tropej/48.2.84. http://dx.doi.org/10.1093/tropej/48.2.84,View ArticlePubMedGoogle Scholar
- Gorstein JL, Dary O, Pongtorn , Shell-Duncan B, Quick T, Wasanwisut E: Feasibility of using retinol-binding protein from capillary blood specimens to estimate serum retinol concentrations and the prevalence of vitamin A deficiency in low-resource settings. Public Health Nutr. 2008, 11: 513-520. http://dx.doi.org/10.1017/S1368980007000821,View ArticlePubMedGoogle Scholar
- Thurnham DI, McCabe GP, Northrop-Clewes CA, Nestel P: Effects of subclinical infection on plasma retinol concentrations and assessment of prevalence of vitamin A deficiency: meta-analysis. Lancet. 2003, 362: 2052-2058. 10.1016/S0140-6736(03)15099-4. http://dx.doi.org/10.1016/S0140-6736(03)15099-4,View ArticlePubMedGoogle Scholar
- Schisterman EF, Perkins NJ, Liu A, Bondell H: Optimal cut-point and its corresponding Youden Index to discriminate individuals using pooled blood samples. Epidemiology. 2005, 16: 73-81. 10.1097/01.ede.0000147512.81966.ba. http://dx.doi.org/10.1097/01.ede.0000147512.81966.ba,View ArticlePubMedGoogle Scholar
- WHO Multicentre Growth Reference Study Group: WHO Child Growth Standards: Length/height-for-age, weight-for-age, weight-for-length, weight-for-height and body mass index-for-age. 2006, Geneva: Methods and development, http://www.who.int/childgrowth/standards/technical_report/en/,Google Scholar
- WHO: WHO Anthro for personal computers version 3.2.2. 2010, Geneva, http://www.who.int/childgrowth/software/en/,Google Scholar
- Barros AJ, Hirakata VN: Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol. 2003, 3: 1-13. 10.1186/1471-2288-3-1. http://dx.doi.org/10.1186/1471-2288-3-21,View ArticleGoogle Scholar
- Demissie T, Ali A, Mekonnen Y, Haider J, Umeta M: Demographic and health-related risk factors of subclinical vitamin A deficiency in Ethiopia. J Health Popul Nutr. 2009, 27: 666-673. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2928083/?tool=pubmed,View ArticlePubMedPubMed CentralGoogle Scholar
- Hornshøj L, Benn CS, Fernandes M, Rodrigues A, Aaby P, Fisker AB: Vaccination coverage and out-of-sequence vaccinations in rural Guinea-Bissau: an observational cohort study. BMJ Open. 2012, 2: e001509-10.1136/bmjopen-2012-001509. http://dx.doi.org/10.1136/bmjopen-2012-001509,View ArticlePubMedPubMed CentralGoogle Scholar
- Flores H, Campos F, Araujo RC, Underwood BA: Assessment of marginal vitamin A deficiency in Brazilian children using the relative dose response procedure. Am J Clin Nutr. 1984, 40: 1281-1289. http://www.ajcn.org/content/40/6/1281.long,PubMedGoogle Scholar
- Benn CS, Diness BR, Roth A, Nante E, Fisker AB, Lisse IM, Yazdanbakhsh M, Whittle H, Rodrigues A, Aaby P: Effect of 50,000 IU vitamin A given with BCG vaccine on mortality in infants in Guinea-Bissau: randomised placebo controlled trial. BMJ. 2008, 336: 1416-1420. 10.1136/bmj.39542.509444.AE. http://dx.doi.org/10.1136/bmj.39542.509444.AE,View ArticlePubMedPubMed CentralGoogle Scholar
- Benn CS, Rodrigues A, Yazdanbakhsh M, Fisker AB, Ravn H, Whittle H, Aaby P: The effect of high-dose vitamin A supplementation administered with BCG vaccine at birth may be modified by subsequent DTP vaccination. Vaccine. 2009, 27: 2891-2898. 10.1016/j.vaccine.2009.02.080. http://dx.doi.org/10.1016/j.vaccine.2009.02.080,View ArticlePubMedGoogle Scholar
- Benn CS, Fisker AB, Napirna BM, Roth A, Diness BR, Lausch KR, Ravn H, Yazdanbakhsh M, Rodrigues A, Whittle H, Aaby P: Vitamin A supplementation and BCG vaccination at birth in low birthweight neonates: two by two factorial randomised controlled trial. BMJ. 2010, 340: c1101-10.1136/bmj.c1101. http://dx.doi.org/10.1136/bmj.c1101,View ArticlePubMedPubMed CentralGoogle Scholar
- Arlappa N, Balakrishna N, Laxmaiah A, Raghu P, Rao VV, Nair KM, Brahmam GN: Prevalence of vitamin A deficiency and its determinants among the rural pre-school children of Madhya Pradesh, India. Ann Hum Biol. 2011, 38: 131-136. 10.3109/03014460.2010.498794. http://dx.doi.org/10.3109/03014460.2010.498794,View ArticlePubMedGoogle Scholar
- Rankins J, Green NR, Tremper W, Stacewitcz-Sapuntzakis M, Bowen P, Ndiaye M: Undernutrition and vitamin A deficiency in the Department of Linguere, Louga Region of Senegal. Am J Clin Nutr. 1993, 58: 91-97. http://www.ajcn.org/content/58/1/91.long,PubMedGoogle Scholar
- WHO/UNICEF/IVACG Task Force: Vitamin A supplements: a guide to their use in the treatment and prevention of vitamin A deficiency and xerophthalmia. 1997, Geneva: , http://whqlibdoc.who.int/publications/1997/9241545062.pdf,Google Scholar
- Pangaribuan R, Erhardt JG, Scherbaum V, Biesalski HK: Vitamin A capsule distribution to control vitamin A deficiency in Indonesia: effect of supplementation in pre-school children and compliance with the programme. Public Health Nutr. 2003, 6: 209-216. http://dx.doi.org/10.1079/PHN2002418,View ArticlePubMedGoogle Scholar
- Nathanail L, Powers HJ: Vitamin A status of young Gambian children: biochemical evaluation and conjunctival impression cytology. Ann Trop Paediatr. 1992, 12: 67-73. http://www.ncbi.nlm.nih.gov/pubmed/1376589,PubMedGoogle Scholar
- Bates CJ, Villard L, Prentice AM, Paul AA, Whitehead RG: Seasonal variations in plasma retinol and carotenoid levels in rural Gambian women. Trans R Soc Trop Med Hyg. 1984, 78: 814-817. 10.1016/0035-9203(84)90028-2. http://www.ncbi.nlm.nih.gov/pubmed/6533855,View ArticlePubMedGoogle Scholar
- Kæstel P, Martinussen T, Aaby P, Michaelsen KF, Friis H: Serum retinol is associated with stage of pregnancy and the acute phase response in pregnant women in Guinea-bissau. J Nutr. 2012, 142: 942-947. 10.3945/jn.111.155937. http://dx.doi.org/10.3945/jn.111.155937,View ArticlePubMedGoogle Scholar
- O'Neill ME, Carroll Y, Corridan B, Olmedilla B, Granado F, Blanco I, Van den Berg H, Hininger I, Rousell AM, Chopra M, Southon S, Thurnham DI: A European carotenoid database to assess carotenoid intakes and its use in a five-country comparative study. Br J Nutr. 2001, 85: 499-507. 10.1079/BJN2000284. http://dx.doi.org/10.1079/BJN2000284,View ArticlePubMedGoogle Scholar
- Scrimshaw NS: Nutritional potential of red palm oil for combating vitamin A deficiency. Food Nutr Bull. 2000, 21: 195-201.View ArticleGoogle Scholar
- You CS, Parker RS, Swanson JE: Bioavailability and vitamin A value of carotenes from red palm oil assessed by an extrinsic isotope reference method. Asia Pac J Clin Nutr. 2002, 11 (7): 438-442. 10.1046/j.1440-6047.11.s.7.1.x. http://dx.doi.org/10.1046/j.1440-6047.11.s.7.1.x,View ArticleGoogle Scholar
- Blomhoff R, Blomhoff HK: Overview of retinoid metabolism and function. J Neurobiol. 2006, 66: 606-630. 10.1002/neu.20242. http://dx.doi.org/10.1002/neu.20242,View ArticlePubMedGoogle Scholar
- Bates CJ: Vitamin A in pregnancy and lactation. Proc Nutr Soc. 1983, 42: 65-79. 10.1079/PNS19830008. http://dx.doi.org/10.1079/PNS19830008,View ArticlePubMedGoogle Scholar
- Fisker AB, Aaby P, Bale C, Balde I, Biering-Sørensen S, Agergaard J, Martins C, Bibby BM, Benn CS: Does the effect of vitamin A supplements depend on vaccination status? An observational study from Guinea-Bissau. BMJ Open. 2012, 2: e000448-BMJ Open. 10.1136/bmjopen-2011-000448. http://dx.doi.org/10.1136/bmjopen-2011-000448,View ArticlePubMedPubMed CentralGoogle Scholar
- Christiaensen L, Alderman H: Child Malnutrition in Ethiopia: Can Maternal Knowledge Augment The Role of Income. 2001, : The World Bank, Africa Region Working Paper Series, 22-http://www.worldbank.org/afr/wps/wp22/wp22-1.pdf,Google Scholar
- Aaby P, Martins CL, Garly ML, Balé C, Andersen A, Rodrigues A, Ravn H, Lisse IM, Benn CS, Whittle HC: Non-specific effects of standard measles vaccine at 4.5 and 9 months of age on childhood mortality: randomised controlled trial. BMJ. 2010, 341: c6495-10.1136/bmj.c6495. http://dx.doi.org/10.1136/bmj.c6495,View ArticlePubMedPubMed CentralGoogle Scholar
- Aaby P, Roth A, Ravn H, Napirna BM, Rodrigues A, Lisse IM, Stensballe L, Diness BR, Lausch KR, Lund N, Biering-Sørensen S, Whittle H, Benn CS: Randomized trial of BCG vaccination at birth to low-birth-weight children: beneficial nonspecific effects in the neonatal period. J Infect Dis. 2011, 204: 245-252. 10.1093/infdis/jir240. http://dx.doi.org/10.1093/infdis/jir240,View ArticlePubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2458/13/172/prepub