Relation of serum retinol to acute phase proteins and malarial morbidity in Papua New Guinea children

¹ From the Nutrition and Veterinary Science Departments, The Pennsylvania State University, University Park; the Departments of International Health and Molecular Microbiology and Immunology, Johns Hopkins School of Hygiene and Public Health, Baltimore; and the Papua New Guinea Institute of Medical Research, Goroka.

² Supported by a grant from USAID's OMNI Research Program through the Human Nutrition Institute of the International Life Sciences Institute.

³ Address reprint requests to FJ Rosales, Nutrition Department, The Pennsylvania State University, 126 Henderson Building South, University Park, PA 16802. E-mail: fxr5{at}psu.edu.

ABSTRACT  
Background: Acute phase proteins (APPs) are associated with malaria-induced hyporetinemia (serum retinol <0.70 µmol/L); however, the degree of the association is not well documented.

Objective: The association between malaria-induced hyporetinemia and APPs was assessed.

Design: In a cross-sectional study, 90 children with serum retinol concentrations from <0.35 to >1.05 µmol/L were selected from children in a clinical trial of vitamin A supplementation. Serum was collected before treatment allocation. Retinol binding protein (RBP) concentrations were determined by radioimmunoassays, and transthyretin, ₁-acid glycoprotein (AGP), ₁-antichymotrypsin, C-reactive protein (CRP), haptoglobin, and albumin concentrations by radial immunodiffusion assays.

Results: Children in the subsample had high rates of splenomegaly and Plasmodium-positive blood-smear slides (P < 0.01); AGP (Pearson's r = -0.40, P < 0.001) and CRP (r = -0.21, P = 0.04) were inversely correlated with retinol. The negative APPs RBP, transthyretin, and albumin were positively and significantly associated with retinol. All APPs, except ₁-antichymotrypsin, were significantly correlated with splenomegaly. Of the positive APPs, AGP correlated with CRP (r = 0.37, P < 0.001), indicating chronic inflammation. In a stepwise regression analysis, 73% of retinol's variability was explained by RBP and transthyretin. The model predicted that a 1-SD increase in RBP or transthyretin increases retinol by 0.38 or 0.47 µmol/L, respectively, whereas an equivalent increase in AGP decreases retinol by 0.12 µmol/L.

Conclusions: The RBP-transthyretin transport complex of retinol is not altered by inflammation. Positive APPs are useful markers of type and severity of inflammation; however, except for AGP, it is unlikely that they can correct for malaria-induced hyporetinemia.

Key Words: ₁-Acid glycoprotein • albumin • ₁-antichymotrypsin • C-reactive protein • haptoglobin • hyporetinemia • multiple linear regression • Plasmodium falciparum • retinol binding protein • serum retinol • transthyretin • children • Papua New Guinea

INTRODUCTION  
Acute phase proteins (APPs) are serum proteins, the concentrations of which change during an acute phase response. Empirically, APPs are classified on the basis of how they change during an acute phase response (1): those proteins that increase are classified as positive APPs and those that decrease [eg, transferrin, albumin, transthyretin, and retinol binding protein (RBP)] are classified as negative APPs (2, 3). On the basis of in vitro studies that have shown the molecular events leading to the initiation and termination of an acute phase response, APPs are classified into groups in place of their response to different cytokines and cofactors that regulate their hepatic syntheses (4, 5).

The functions of APPs are many; some positive APPs participate in host-adaptive and host-defense mechanisms by binding to foreign substances or by having opsonizing activities and modulating phagocytic cell functions. Other APPs have more specific actions, such as inhibiting serine proteinases or serving as transport proteins with antioxidant activity, like ceruloplasmin (2, 4). On the other hand, negative APPs such as transthyretin, RBP, transferrin, and albumin have no apparent immune function (2, 3). Their main role is to transport nutrients and, therefore, their reduction during infection and inflammation may lower the concentration of specific nutrients (3).

The serum concentration of retinol, the alcohol form of vitamin A, decreases during malarial infections. This reduction has been characterized as a direct consequence of the inflammatory response to Plasmodium infections (6, 7) because several APPs were found to be associated with a reduction in retinol (8–10). However, it is not clear which of these APPs is a useful predictor of plasma retinol during malarial infection. This information would be useful because, as some researchers have suggested, the inclusion of a measure of the acute phase response would help interpret plasma retinol concentrations during malarial infection (6, 8).

The purpose of the present study was to assess the association between serum retinol and several APPs in children with malaria-related morbidity and to determine which of these APPs is a significant predictor of serum retinol. C-reactive protein (CRP) was chosen because it is a marker of severe malarial infection (11, 12), ₁-antichymotrypsin (ACT) because it has a longer half-life than CRP (13), and ₁-acid glycoprotein (AGP) because its concentration increases during malarial infection (14). Haptoglobin is a positive APP, but in malarial infections its concentration is reduced because haptoglobin helps clear free hemoglobin from recent hemolysis, and thus, it is a marker of recent Plasmodium attacks and fever (11, 12). Negative APPs such as albumin, RBP, and transthyretin were chosen because the first is used as a nutritional marker, and the others form the transport complex that specifically carries retinol in the circulation.

SUBJECTS AND METHODS  
Study population
This cross-sectional analysis included 90 children who participated in a community-based vitamin A supplementation trial. Briefly, preschool children in a Plasmodium falciparum–endemic area of Wosera, Papua New Guinea, were enrolled in a randomized, double-masked, placebo-controlled trial (15). At enrollment, children received a physical examination that included an ophthalmic evaluation for signs of vitamin A deficiency, determinations of axillary temperature with an electronic thermometer, and an assessment of spleen size in the standing position with use of Hackett's grading system (16). Body weight was also measured during the physical exam. A 5-mL sample of venous blood was drawn for serum retinol analysis by HPLC and for hemoglobin concentration measurements with Hemo-Cue (R & D Systems, Minneapolis). Children with a midupper arm circumference < 12.5 cm were excluded from the study (15). The community-based study was approved by the Institutional Review Board of the Johns Hopkins School of Medicine and the Medical Research Council of Papua New Guinea Ministry of Health. A total of 484 children were enrolled and the results of the efficacy study were published previously (15, 17).

The children in the present study were selected on the basis of their serum retinol concentrations at baseline and to allocate 15–20 children per stratum on the basis of cutoff concentrations of serum retinol recommended by Underwood (18). This stratification scheme provided a proportional and linear distribution of children with serum retinol concentrations ranging from <0.35 (ie, vitamin A deficiency) to 1.05 µmol/L, a concentration at which subclinical vitamin A deficiency is unlikely to occur. Approval to conduct this study was obtained from The Pennsylvania State University Biosafety Committee.

Serum samples
Venous blood collected at enrollment was kept in a dark box at ambient temperature for <6 h before centrifugation. Serum was obtained by centrifugation at 1500 x g for 5 min at room temperature, and aliquots were prepared and immediately stored at -70°C. The samples were transported to the Johns Hopkins School of Hygiene and Public Health (Baltimore) in liquid nitrogen (15), and 100-µL aliquots of serum were stored at -70°C at The Pennsylvania State University.

Radioimmunoassay
Human serum RBP was purified previously by one of us (19) and its concentration was determined by spectrophotometry (20). The purified serum RBP was iodinated with ¹²⁵I (Amersham Life Science Products, Piscataway, NJ) with the lactoperoxidase method of Miyachi et al (21) according to the procedure described previously by Smith et al (22) for rat RBP. An antibody to human RBP purchased from Accurate Chemical & Scientific Corp (Westbury, NY) was titrated to a final dilution of 1:8000, which bound 60% of the ¹²⁵I. The standards for this assay were prepared from pooled human serum obtained from healthy volunteers. The mean (±SD) concentrations of serum RBP (2.25 ± 0.3 µmol/L, or 47.2 ± 6.4 µg/mL) and of transthyretin (5.5 ± 0.3 µmol/L, or 302.3 ± 15.2 µg/mL) in this pool were determined by radial immunodiffusion (RID) with commercially available kits (The Binding Site Inc, San Diego). The antibody-binding capacity to human serum RBP was compared with that of purified human RBP in serial dilutions, which were run in parallel. The antibody bound equally to either serum or purified RBP and it had a linear response between 40 and 150 µg/L. Serum samples for this assay were diluted (1:150–1:300) in assay buffer, and triplicate aliquots were assayed with labeled RBP and antiserum to human serum RBP. In each assay, 2 aliquots from the pooled serum (ie, 2.25 ± 0.3 µmol RBP/L) were run, each in triplicate; on the basis of these determinations, we estimated the intraassay CV (<5%) and the interassay CV (<7.0%), and the accuracy of the assay was compared with an external standard (ie, the CV was ±3% of 2.25 ± 0.3 µmol serum RBP).

Single radial immunodiffusion assay
Single RID assays were prepared to measure serum concentrations of transthyretin, CRP, haptoglobin, AGP, ACT, and albumin. The RID assay used was a modification of the method developed by Mancini et al (23). The antiserum for each protein was purchased from DAKO Corporation (Carpinteria, CA). The samples were diluted in 7% bovine serum albumin (Sigma Chemical Co, St Louis) and 25 mmol tris buffer/L, pH 7.5. Calibrators, external standards, and samples were applied to wells in 5-µL volumes. For the calibrators, a plot of the diameter squared on the y axis and the concentration of the antigen on the x axis gave a linear function (23). On the basis of this linear function, sample concentrations were calculated.

Calibration and standardization of RID assays
Calibrators from DAKO were titrated for each serum protein, and 4 different concentrations were selected for each assay. During this procedures, it was determined that none of the human antisera cross-reacted with bovine serum albumin. The College of American Pathologist's (Northfield, IL) reference standard for serum proteins was used as an external standard for all assays, except for ACT, for which The Binding Site reference standard was used. The accuracy of each RID assay was checked against 2 different dilutions of external standards. Quality controls were conducted for each plate and the intra- and interassay CVs were determined. The limit of detection was 1.0 mg/L for CRP, 0.03 g/L for ACT, 0.02 g/L for haptoglobin, 0.01 g/L for AGP, 0.02 g/L for transthyretin, and 0.04 g/L for albumin.

Statistical analysis
Exploratory and confirmatory analyses were conducted to determine the distribution of the variables and their associations. Weight-for-age (WAZ) and weight-for-height (WHZ) z scores were calculated with the software ANTHRO [software version 1.01; Centers for Disease Control and Prevention (CDC), Atlanta, and the World Health Organization (WHO), Geneva]. Children were classified as wasted if they had WHZ scores > 2 SDs below the reference median (ie, WHZ < -2) and as chronically undernourished if they had WAZ scores > 2 SD below the corresponding CDC/WHO international reference median (ie, WAZ < -2) (24). Chi-square tests and Spearman correlation coefficients were used to determine the significance of an association between categorical variables (eg, serum retinol stratum versus inflammation status) and Pearson correlation coefficients were used for continuous variables (25). A correlation matrix was determined by using SPSS 4.0 (SPSS Inc, Chicago). A multiple regression model was developed to assess the significance of various covariates in explaining variations in serum retinol concentrations. Because of collinearity among the covariates, stepwise regression analysis was used to select the variables in the model (26). The effect of specific covariates, such as age group and splenomegaly, were assessed by determining whether or not the mean residual sum of squares (ie, the variance) and the estimates of the parameters (ie, the coefficient) remained constant in nested models (26). The standardized regression coefficient ß (ie, the change in serum retinol relative to a 1-SD change in the covariate) was used to evaluate the relative importance of covariates (27).

RESULTS  
Characteristics of children in the subsample
The nutritional status and demographic and malariometric indexes of the study sample are presented in Table 1. There was a significantly higher proportion of girls and older children in the subsample of children included in this analysis than in the remainder of the population. The subsample also had significantly higher rates of splenomegaly and Plasmodium-positive blood-smear slides than the rest of the population. The age distribution of children in the subsample was bimodal, with 2 crests at 30 and 54 mo and a recession at 36 mo. Children were classified into those younger and those older than 36 mo. The proportion of children older than 36 mo was inversely associated with serum retinol concentration (Spearman's r = -0.20, P = 0.05). Age differences were not apparent by sex group and serum retinol concentrations did not differ significantly between boys and girls.

View this table:
TABLE 1.. Distribution of demographic, anthropometric, biochemical and malariometric characteristics of children in the subsample compared with the rest of the children participating in a vitamin A supplementation trial in Papua New Guinea, 1996  
Among children in the subsample, the rate of chronic undernutrition based on WAZ scores was 42%; however, few (5%) of these children were wasted. Eighty-one percent of the children had palpable spleens; their average enlarged spleen score was 3.0 (ie, Hackett's classification: "spleen with lowest palpable point projected more than half-way to the umbilicus but not below a line drawn horizontally through it") (16). Neither undernutrition (ie, WAZ and WHZ scores < -2) nor the prevalence of anemia (ie, hemoglobin < 110 g/ L) was significantly associated with serum retinol (data not shown). In contrast, the proportion of children with enlarged spleens was inversely associated with serum retinol concentration (Spearman's r = -0.30, P < 0.001).

Relation of serum retinol to acute phase proteins during Plasmodium infection
Positive APPs were higher and negative APPs were lower as serum retinol concentrations decreased per stratum (Table 2). The proportion of children with inflammation on the basis of AGP concentrations was significantly greater for those with lower retinol concentrations, whereas the proportion of children with inflammation on the basis of serum CRP, ACT, or haptoglobin concentrations did not show a linear trend with strata of serum retinol concentration. Negative APPs (RBP, transthyretin, and albumin) were positively and linearly associated with serum retinol, whereas AGP was inversely associated with serum retinol (Figure 1). Although not shown in Figure 1, CRP (Pearson's r = -0.21, P = 0.04) was inversely correlated with the serum retinol concentration.

View this table:
TABLE 2.. Distribution of acute phase proteins: retinol binding protein (RBP), transthyretin, C-reactive protein (CRP), ₁-antichymotrypsin (ACT), ₁-acid glycoprotein (AGP), haptoglobin, and albumin by serum retinol concentration in children from a malarial endemic area, Wosera, Papua New Guinea, 1996¹  

View larger version (25K):
FIGURE 1. . Scatter plots of serum retinol concentration on the serum acute phase proteins retinol binding protein, transthyretin, ₁-acid glycoprotein, and albumin. Each point represents an individual child. The coefficient of determination (R²) was determined by simple linear regression (see Methods).

 
Interrelation among acute phase proteins and clinical variables in children with malaria-related morbidity
Among positive APPs, serum AGP was positively and linearly associated with serum CRP and ACT concentrations, but not with serum haptoglobin (Table 3). CRP was positively correlated with serum ACT, although the correlation was lower than for serum AGP. Serum AGP and CRP individually were inversely associated with serum RBP, transthyretin, and albumin concentrations. Splenomegaly (ie, Hackett's classification) was positively associated with serum AGP and CRP and inversely correlated with serum RBP, transthyretin, and albumin. None of these APPs was significantly associated with the abundance of P. falciparum or with age in months. Splenomegaly was positively associated with P. falciparum infection.

View this table:
TABLE 3.. Pearson correlation matrix among clinical characteristics and acute phase proteins: retinol binding protein (RBP), transthyretin, ₁-acid glycoprotein (AGP), albumin, ₁-antichymotrypsin (ACT), and haptoglobin in children from a malarial endemic area, Wosera, Papua New Guinea¹  
Predictors of serum retinol concentration in children with malaria-related morbidity
Covariates included in the regression models were selected on the basis of their significant association with serum retinol (Tables 2 and 3) by using a forward stepwise procedure (26). In model 1, which included all subjects, both RBP and transthyretin were strong predictors of serum retinol (Table 4). The effects of splenomegaly and the age of the children on serum retinol were assessed in nested models. Models 2A and 2B included children with splenomegaly (Hackett's classification > 0) stratified by age group. The association between serum retinol, RBP, and transthyretin did not change, as indicated by the estimate of the variance and the estimates of each predictor in the models. The saturation of serum RBP (ie, the calculated molar ratio of serum retinol to RBP) was within reference limits (32) and did not differ significantly between children younger and older than 36 mo (P > 0.05, t test).

View this table:
TABLE 4.. Relation of serum retinol to retinol binding protein (RBP), transthyretin, and ₁-acid glycoprotein (AGP) assessed by a multiple regression analysis and by the saturation of RBP in children from a malarial endemic area, Wosera, Papua New Guinea, 1996  

DISCUSSION  
The Wosera area of Papua New Guinea is known for a high presence of Plasmodium infections (33, 34); undernutrition is also prevalent. A study reported that 29% of children aged 2–10 y suffered from chronic growth deficit (ie, height-for-age z score < -2), but few of these children were wasted (35). Recent evidence indicates that stunting may be a predictor of survival of Plasmodium infections in this population (36). The vitamin A status of children from the Wosera area is not known. Although a survey conducted in East Sepik, Papua New Guinea, reported that 91% of preschool children had marginal vitamin A deficiency (ie, serum retinol < 0.70 µmol/L) (37), the prevalence of children with Bitot's spots was low, 0.19%, as determined by ophthalmic surveys conducted in other areas of Papua New Guinea (37). Children in the present study were selected to ensure adequate representation in the different strata of serum retinol, from <0.35 to >1.05 µmol/L. More than 75% of these children had splenomegaly or a Plasmodium-positive slide. Therefore, the serum of these children provided a unique opportunity to assess the relation between serum retinol and APPs.

Positive APPs were variably associated with malariometric indicators. These positive APPs are induced by a combination of cytokines, except for ACT, which can be induced by interleukin 6 (IL-6) alone (2). Proinflammatory cytokines and IL-6 concentrations increased in patients with malarial infections, especially in acute infections with P. falciparum (7). However, only 12% of the children in this study had serum ACT concentrations >0.6 g/L (ie, indicative of inflammation) and their serum ACT concentrations were not associated with splenomegaly or with P. falciparum abundance. On the other hand, 34% of these children had serum CRP concentrations 10 mg/L (ie, indicative of inflammation) and serum CRP was significantly associated with splenomegaly; however, CRP did not rise to the concentrations observed previously in acute bacterial or viral infections (38). The observed changes in serum CRP suggested a chronic inflammation (2). This was corroborated by the change in serum haptoglobin and AGP. Low haptoglobin concentrations represent recent hemolysis during malaria infections (11, 12).

In the present study, the rate of hypohaptoglobinemina was mild compared with studies in which acute malarial infections have been assessed (11, 12). The serum AGP concentration was positively and significantly associated with serum CRP. This result is important because it establishes the overall type of inflammation in this population (eg, acute compared with chronic inflammation). In acute inflammation, CRP concentrations increase 100-to 1000-fold, whereas in chronic inflammation, the increase is less. Serum AGP concentrations increase similarly in both acute and chronic infections, and thus, AGP is poorly correlated with CRP in acute inflammations (38). Thus, children in the present study had a chronic inflammatory response to continuous and prolonged exposure to Plasmodium infections.

Previous studies in populations with malaria-related morbidity indicated that serum retinol decreases as part of the inflammatory response to malarial infections (6–10, 39). Although few of these studies estimated the associations of serum retinol to APPs , their results showed that the reduction in serum retinol is a consequence of inflammation. The reduction in serum retinol occurred in populations in whom vitamin A deficiency would not be considered a public health problem, such as in children who have taken vitamin A supplements (8) and in healthy French (7) and Thai (6) adults. The present study had a cross-sectional design, therefore a cause-effect inference was limited. Nonetheless, the highly significant and negative associations of serum retinol with splenomegaly and AGP in children with normal RBP saturation ratios (Table 4) strongly suggest that hyporetinemia (ie, a serum retinol concentration <0.70 µmol/L) is a consequence of inflammation.

Some researchers have suggested the use of positive APPs to adjust measured values of serum retinol for the effect of inflammation (6, 8). As noted above, malaria-induced inflammation reduces serum retinol independently of vitamin A status, and thus, serum retinol may be a less sensitive indicator of vitamin A status. Although positive APPs were useful as indicators of inflammation in children with malaria, it is unlikely that they could be used to correct malaria-induced hyporetinemia. Changes in positive APPs in serum were variably associated with serum retinol and malariometric indicators. In part, this was because their serum concentrations represented their synthesis (eg, an inflammatory stimulus) and their degradation (eg, their biological function). Degradation may not be related to the metabolism of vitamin A (eg, haptoglobin clears free hemoglobin from recent hemolysis). In contrast, RBP and transthyretin were highly and consistently associated with serum retinol (Table 4) because their metabolism is intimately linked to that of vitamin A. Although AGP was significantly correlated with serum retinol in the univariate analysis, this association was not significant after changes in RBP and transthyretin were controlled for in the multivariate analysis. These results suggest that this association should be further assessed to understand whether it is specific to chronic inflammation, and thus, whether a change in AGP can be used as a correction factor for measured serum retinol. In the present study, a change in AGP concentrations equivalent to 1 SD resulted in a reduction in serum retinol of 0.12 µmol/L (Table 4).

In summary, this study determined the degree of association between serum retinol and concentrations of various APPs in children with malaria-related morbidity. The regression model showed that RBP and transthyretin were significantly associated with serum retinol, emphasizing their biological function, which is not altered by chronic inflammation. This evidence strongly suggests that hyporetinemia is a consequence of chronic inflammation during malarial infection.

ACKNOWLEDGMENTS  
We acknowledge the contributions of Moses Baisor and Jack Tarika and the participation of children from the Wosera area of Papua New Guinea.

REFERENCES  

1. Koj A. Definition and classification of acute-phase proteins. In: Gordon AH, Koj A, eds. The acute-phase response to injury and infection. Amsterdam: Elsevier Science Publishers, 1985:139–44.
2. Whicher JT, Westacott CI. The acute phase response. In: Whicher JT, Evans SW, eds. Biochemistry of inflammation. Boston: Kluwer Academic Publishers, 1992:243–69.
3. Schreiber G, Tsykin A, Aldred AR, et al. The acute phase response in the rodent. Ann N Y Acad Sci 1989;557:61–86
4. Steel DM, Whitehead AS. The major acute phase reactants: C-reactive protein, serum amyloid P component and serum amyloid A protein. Immunol Today 1994;15:81–8.
5. Wegenka UM, Buschmann J, Lutticken C, Heinrich PC, Horn F. Acute-phase response factor, a nuclear factor binding to acute-phase response elements, is rapidly activated by interleukin-6 at the posttranslational level. Mol Cell Biol 1993;13:276–88.
6. Thurnham DI, Singkamani R. The acute phase response and vitamin A status in malaria. Trans R Soc Trop Med Hyg 1991;85:194–9.
7. Tabone MD, Muanza K, Lyagoubi M, Jardel C, Pied S, Amedee-Manesme O. The role of interleukin-6 in vitamin A deficiency during Plasmodium falciparum malaria and possible consequences for vitamin A supplementation. Immunology 1992;75:553–4.
8. Filteau SM, Morris SS, Abbott RA, et al. Influence of morbidity on serum retinol of children in a community-based study in northern Ghana. Am J Clin Nutr 1993;58:192–7.
9. Das BS, Thurnham DI, Das BD. Plasma -tocopherol, retinol, and carotenoids in children with falciparum malaria. Am J Clin Nutr 1996;64:94–100.
10. Friis H, Mwaniki D, Omondi B, et al. Serum retinol concentrations and Schistosoma mansoni, intestinal helminths, and malarial parasitemia: a cross-sectional study in Kenyan preschool and primary school children. Am J Clin Nutr 1997;66:665–71.
11. Hurt N, Smith T, Tanner M, et al. Evaluation of C-reactive protein and haptoglobin as malaria episode markers in an area of high transmission in Africa. Trans R Soc Trop Med Hyg 1994;88:182–6.
12. McGuire W, D'Alessandro U, Olaleye BO, et al. C-reactive protein and haptoglobin in the evaluation of a community-based malaria control programme. Trans R Soc Trop Med Hyg 1996;90:10–4.
13. Thompson D, Milford-Ward A, Whicher JT. The value of acute phase protein measurements in clinical practice. Ann Clin Biochem 1992;29:123–31.
14. Van Hensbroek MB, Kwiatkowski D, Van Den Berg B, Hoek FJ, Van Boxtel CJ, Kager PA. Quinine pharmacokinetics in young children with severe malaria. Am J Trop Med Hyg 1996;54:237–42.
15. Shankar AH, Genton B, Semba RD, et al. Effect of vitamin A supplementation on morbidity due to Plasmodium falciparum in young children in Papua New Guinea: a randomized trial. Lancet 1999; 354:201–7.
16. World Health Organization. Malaria terminology. Report of a drafting committee appointed by the World Health Organization. World Health Organ Monogr Ser 1953;38.
17. The International Life Science Institute Research Foundation's Human Nutrition Institute, The International Vitamin A Consultative Group (IVACG). Report of the XVIII International Vitamin A Consultative Group Meeting. Sustainable control of vitamin A deficiency: defining progress through assessment, surveillance, evaluation. Washington, DC: US Agency for International Development, 1997.
18. Underwood BA. Vitamin A in human nutrition. Public health considerations. In: Sporn MB, Roberts AB, Goodman DS, eds. The retinoids. Biology, chemistry and medicine. 2nd ed. New York: Raven Press, 1994:211–27.
19. Lenich ML, Ross AC. Production and secretion of retinol-binding protein by a human hepatoma cell line, HepG2. J Lipid Res 1987; 28:941–8.
20. Muto Y, Goodman DS. Vitamin A transport in rat plasma. Isolation and characterization of retinol-binding protein. J Biol Chem 1972;247:2533–41.
21. Miyachi Y, Vaitukaitis JL, Nieschlag E, Lipsett MB. Enzymatic radioiodination of gonadotropins. J Clin Endocrinol Metab 1972; 34:23–8.
22. Smith JE, Borek C, Goodman DWS. Regulation of retinol-binding protein metabolism in cultured rat liver cell lines. Cell 1978;1 5:865–73.
23. Mancini G, Carbonara AO, Heremans JF. Immunochemical quantitation of antigens by single radial immunodiffusion. Immunochemistry 1965;2:235–54.
24. World Health Organization Woking Group. Use and interpretation of anthropometric indicators of nutritional status. Bull World Health Organ 1986;64:929–41.
25. Rosner B. Fundamentals of biostatistics. 2nd ed. Boston: Duxbury Press 1986:409.
26. Weisberg S. Applied linear regression. 2nd ed. New York: John Wiley & Sons, 1985:196–221.
27. Norusis MJ. SPSS/PC+ statistics 4.0 for the IBM PC/XT/AT and PS/2. Chicago: SPSS Inc, 1990:B–94.
28. Calvin J, Neale G, Fotherby KJ, Price CP. The relative merits of acute phase proteins in the recognition of inflammatory conditions. Ann Clin Biochem 1988;25:60–6.
29. Kanakoudi F, Drossou V, Tzimouli V, et al. Serum concentrations of 10 acute-phase proteins in healthy term and preterm infants from birth to age 6 months. Clin Chem 1995;41:605–8.
30. Benvenga S, De Luca F, Vermiglio F, Pandullo E, Bartalena L, Trimachi F. Changes in radioimmunoassayable prealbumin (TBPA) serum levels from birth to adulthood. Horm Metab Res 1986;18:73.
31. Spiekerman AM. Proteins used in nutritional assessment. Clin Lab Med 1993;13:353–69.
32. Mourey MS, Siegenthaler G, Amedee-Manesme O. Regulation of metabolism of retinol-binding protein by vitamin A status in children with biliary atresia. Am J Clin Nutr 1990;51:638–43.
33. Genton B, Al-Yaman F, Beck HP, et al. The epidemiology of malaria in the Wosera area, East Sepik Province, Papua New Guinea, in preparation for vaccine trials. I. Malariometric indices and immunity. Ann Trop Med Parasitol 1995;89:359–76.
34. Genton B, Al-Yaman F, Beck HP, et al. The epidemiology of malaria in the Wosera area, East Sepik Province, Papua New Guinea, in preparation for vaccine trials. II. Mortality and morbidity. Ann Trop Med Parasitol 1995;89:377–90.
35. Gibson RS, Heywood A, Yaman C, Sohlstrom A, Thompson LU, Heywood P. Growth in children from the Wosera subdistrict, Papua New Guinea, in relation to energy and protein intakes and zinc status. Am J Clin Nutr 1991;53:782–9.
36. Genton B, Al-Yaman F, Ginny M, Taraika J, Alpers MP. Relation of anthropometry to malaria morbidity and immunity in Papua New Guinea children. Am J Clin Nutr 1998;68:734–41.
37. World Health Organization, United Nations Children's Fund. Global prevalence of vitamin A deficiency. Geneva, Switzerland: World Health Organization, 1995. (WHO/NUT/95.3.)
38. Fassbender K, Zimmerli W, Kissling R, et al. Glycosylation of ₁-acid glycoprotein in relation to duration of disease in acute and chronic infection and inflammation. Clin Chim Acta 1991;203:315–28.
39. Samba C, Galan P, Luzeau R, Amedee-Manesme O. Vitamin A deficiency in pre-school age Congolese children during malarial attacks. Part 1: Utilization of the impression cytology with transfer in an equatorial country. Int J Vitam Nutr Res 1990;60:215–23.