2003 The American Society for Nutritional Sciences J. Nutr.

133:1339-1346, May 2003

Community and International Nutrition

A Micronutrient-Fortified Beverage Prevents Iron Deficiency, Reduces Anemia and Improves the Hemoglobin Concentration of Pregnant Tanzanian Women
Diklar Makola2, Deborah M. Ash, Simon R. Tatala*, Michael C. Latham, Godwin Ndossi* and Haile Mehansho Division of Nutritional Sciences, Cornell University, Ithaca, NY,; * Tanzania Food and Nutrition Centre, Dar es Salaam, Tanzania and; Procter & Gamble Company, Cincinnati, OH
2

To whom correspondence should be addressed at Kettering Medical Center, 3535 Southern Blvd., Kettering, OH 45459. E-mail: dmakola@yahoo.com.

ABSTRACT
Maternal malnutrition continues to be a major contributor to adverse reproductive outcomes in developing countries, despite longstanding efforts to fortify foods or to distribute medicinal supplements to pregnant women. The objective of this study was to test the effect of a micronutrient-fortified beverage containing 11 micronutrients (iron, iodine, zinc, vitamin A, vitamin C, niacin, riboflavin, folate, vitamin B-12, vitamin B-6 and vitamin E) on the hemoglobin, iron and vitamin A status of pregnant women in Tanzania. A group of 259 pregnant women with gestational ages of 8 to 34 wk were enrolled in a randomized double-blind controlled trial in which study women received 8 wk of supplementation. Hemoglobin, ferritin and dried blood spot retinol were measured at baseline and at the end of the supplementation period. The supplement resulted in a 4.16 g/L increase in hemoglobin concentration and a 3 g/L increase in ferritin and reduced the risk of anemia and iron deficiency anemia by 51 and 56%, respectively. The risk of iron deficiency was reduced by 70% among those who had iron deficiency at baseline and by 92% among those who had adequate stores. The micronutrientfortified beverage may be a useful and convenient preventative measure, one that could help improve the nutritional status of women both before and during pregnancy and thereby help avoid some of the potential maternal and fetal consequences of micronutrient deficiencies. KEY WORDS: dietary supplement iron, anemia micronutrient pregnant women vitamin A

Anemia of pregnancy, defined as hemoglobin concentration < 110 g/L, is a major public health problem affecting 35 to 75% of pregnant women in developing countries and 18% in industrialized countries (1 ). Although the most common cause of anemia is iron deficiency (2 ), deficiencies of vitamin B-12, folate, vitamin A and even zinc contribute either singly or in combination to maternal anemia (3 ,4 ). Women in developing countries have a high prevalence of iron deficiency but also tend to be deficient in other micronutrients such as zinc (5 ), vitamin A (6 ,7 ), folate and vitamin B-12 (4 ). Chronic food shortage, frequent pregnancies and prolonged lactation lead to maternal nutritional depletion characterized by deficiencies of nutrients such as iron and vitamin A (8 ,9 ). The cultural practice of "eating down" (limiting intake while pregnant to produce smaller infants) may also contribute to anemia (10 ). Iron deficiency is also exacerbated by blood loss during menstruation and at delivery and by the iron requirements of the fetus and placenta (11 ). Pregnancy increases the need for energy and micronutrients (12 14 ). Even with an optimal diet, a woman must enter pregnancy with an estimated minimum of 300 mg of stored iron to avoid an iron deficit (11 ). Iron stores in fertile women in industrial countries average only 150 mg (15 ). In the developing world, the majority of women who enter pregnancy are anemic. Anemia in these populations is often exacerbated by a high prevalence of infections with intestinal parasites (16 ,17 ), malaria and HIV (18 ) and the occurrence of hemoglobinopathies (19 21 ). Numerous programs aimed at reducing the incidence of anemia and iron deficiency in affected populations have historically focused on either medicinal supplementation with iron, folate and vitamin A or fortification of foods with iron (19 ). However, because iron deficiency rarely occurs in isolation and is often accompanied by other micronutrient deficiencies (4 ), the use of multiple micronutrient supplements instead of iron-only supplements is being strongly considered (14 ,22 ). Very few studies evaluating the efficacy of micronutrient supplements on iron status and anemia have been published and thus the efficacy of this approach has not been established (23 ). However, combining iron and vitamin A supplementation in pregnant women has already been shown to improve both vitamin A and iron status (24 ,25 ). The present study evaluated the effect of a micronutrient-fortified dietary supplement (delivered in the form of a highly palatable orange-flavored beverage) on the iron (serum ferritin), vitamin A (serum retinol) and anemia (hemoglobin) status of pregnant women attending antenatal clinics in the Mpwapwa and Kongwa districts in Tanzania. An earlier efficacy trial (26 ) demonstrated that school children in the Mpwapwa District who received the supplement for 6 mo had significantly better growth, in addition to improved iron status and serum retinol levels.

MATERIALS AND METHODS

Study design. This study was a randomized, placebo-controlled double-blind effectiveness trial of a micronutrient-fortified dietary supplement conducted in pregnant women in Tanzania. Enrollment into the study took place during 2 wk in August of 1999 and the postintervention

evaluations were performed 8 wk after enrollment. The participants were recruited from women attending prenatal clinics in the hospitals and surrounding health centers of Mpwapwa and Kongwa districts. The area is arid and largely without agricultural irrigation; crops are primarily maize, sorghum, cassava, millet, cowpeas and groundnuts. Fruits and vegetables are seasonal and include guavas, papayas, mangos, oranges, bananas, baobab fruits, tomatoes, amaranth, cassava leaves and sweet potatoes. Malaria is endemic, with sharp increases in the disease incidence during the rainy seasons. Intestinal helminthic infections are common. Sample size requirements were calculated by the use of an expected hemoglobin concentration difference of 8 g/L (with an estimated variance of 10.5 g/L) between the experimental and control groups. To detect this difference with a power of 80% and a confidence level of 95%, and after correcting for a 25% dropout rate, each of the two groups required 211 women participants. Pregnant women who believed that they were between 12 and 34 wk pregnant were invited to participate in the study. Exclusion criteria included a pregnancy that was determined to be either <12 wk or >34 wk on uterine palpation; a hemoglobin concentration of <80 g/L or a serious medical condition; or complication of pregnancy such as cardiac disease, pneumonia and threatened abortion. Information on the parity and the gravidity of current pregnancy, a history of other illnesses before and during the pregnancy and the use of other iron/folic acid supplements was obtained. Height, weight, mid-upper arm circumference (MUAC) and skinfold thickness were measured both at entry into the study and during the postintervention assessment 8 wk later. A determination of gestational age was made on the basis of the womans recall of her last menstrual period (LMP) as well as palpation of uterine height of fundus (HOF) and measurement of symphysis fundal height (SFH). At each of the six study centers, a block randomization (10 subjects in each block) was used to assign women into either the micronutrient-fortified (experimental) group or the nonfortified (control) group. Micronutrient supplementation. The micronutrient supplement was developed and produced by Procter & Gamble Company as an orange-flavored micronutrient-fortified powdered beverage mix containing 11 micronutrients. The supplement is unique in that the iron was stabilized in an aqueous system and was found to be highly bioavailable (23.4% when consumed alone and 10.7% when consumed with rice), thus providing 2.57 and 1.16 mg of bioavailable iron, respectively (27 ). Each serving contained 368 kJ of energy, with two servings providing 736 kJ to each participating woman. The constituents and concentrations of the various micronutrients were agreed upon by an extensive collaboration between nutritionists at Cornell University, UNICEF, Tanzania Food and Nutrition Centre (TFNC) and the Procter & Gamble Company, and were meant to provide a nutritionally adequate, pleasant-tasting product whose taste, color and composition were retained during storage, distribution and use (Table 1). The beverage mix was packaged in 25-g packets. A nonfortified beverage mix of identical appearance, color and taste was packaged in similar, but different colored, 25-g packets and served as the placebo. The placebo packages also contained 368 kJ of energy.

View this table: In this window In a new window TABLE 1Composition of micronutrient-fortified dietary beverage12 At enrollment and baseline examination, study participants were supplied with a 2-wk supply of packets. Subjects were instructed to add the contents of each packet to 250 mL of clean boiled water. Each subject was asked to consume the beverage thus produced twice a day with meals. Participants were asked to visit the study centers every 2 wk to collect new supplies and to bring the empty packets for counting. Intermittent home visits were made by field members of the research team to evaluate compliance with supplementation and to encourage attendance at their follow-up visits. During the biweekly visits, the participants received routine antenatal examinations that included weight, blood pressure, height of fundus checks and a physical examination. Ordinarily, nurses in the clinic perform the physical examinations with only high-risk pregnancies being referred to the physician. However, during the conduct of this study, all women who took part in the study were examined by one of the physicians participating in the study during most of the clinic visits. A history of experience with the supplied beverage was obtained. Empty packets were counted and recorded as a measure of compliance. Participating women were questioned with regard to concurrent medicinal iron/folic acid supplementation. Just before the women left the antenatal clinic, staff provided the mothers with a 2-wk supply of an iron/folic acid supplement that contained 60 mg of elemental iron and 500 g of folic acid to be taken on a daily basis. Women who were found to have parasitic infections were treated with a single dose of Albendazole (400 mg). Follow-up examinations. Postintervention assessments were performed after 8 wk of supplementation. Based on a previous pilot study, it was postulated that supplementation for 8 wk would be adequate to demonstrate an effect if one existed and that most participants would not have delivered by the time of follow-up. Laboratory determinations. Qualified TFNC laboratory technicians performed phlebotomies to collect 5 mL of venous blood (2 mL serum) at entry into the study and at follow-up. The blood was used for immediate determination of hemoglobin concentration by use of the HemoCue portable photometer (HemoCue AB, Angelhom, Sweden). The accuracy of the HemoCues was checked daily with a control cuvette provided with the machines; precision was 0.03 g/L. The remaining blood was allowed to clot at room temperature, after which it was transported the same day to the Mpwapwa District Hospital laboratory in a cool box with cooling elements. On arrival at the laboratory, the samples were centrifuged at 3000 x g for 5 min and the serum was placed in three vials and frozen at -40C. After about 1 mo, the serum was transported in a nitrogen gas tank to the TFNC laboratory in Dar es Salaam, where it was kept frozen at -40C to await later analysis

of ferritin and C-reactive protein (CRP) levels. Serum ferritin was assayed by use of a quantitative ELISA (28 ). Samples were assayed in duplicate to measure within-run precision, and aliquots of a pooled serum sample were used to measure interassay precision. Lymphocheck control serums were used to assess accuracy and precision; the coefficient of variation was <5%. CRP were determined by an ELISA based on the method developed by Naik and Voller (29 ). The lower limit of detection was 7.8 mg/L. Capillary (finger-prick) blood was also collected at baseline and at follow-up and was used to prepare dried spots on filter paper for subsequent retinol and thyroid-stimulating hormone (TSH) assays. The quantitative determinations of TSH in blood dried on filter paper were made by use of the Dried Blood Enzyme Immunoassay (peroxidase/color) in vitro diagnostic kit (30 ). To monitor the precision of the assay, quality control serums were included in each run and the CV was about 12%. The retinol assays were performed at Craft Technologies (Wilson, NC) by use of the dried blood spot (DBS) retinol methodology developed by Craft et al. (31 ). The correlation between serum retinol and DBS retinol on a subsample of the study was r2 = 0.85 (N. E. Craft, Craft Technologies, personal communication). Stool samples were collected during enrollment into the study and used to determine the presence of helminth eggs and estimate parasite load. Statistical analysis. Categorical data are presented as frequencies, whereas continuous data are presented as means SD. To achieve normality, skewed variables such as ferritin, mothers age, MUAC and skinfold thickness were log-transformed, and CRP was square-root transformed in addition to removal of outliers. Statistical analysis involving the transformed variables used transformed data, although the results are presented in original units. Comparisons between groups when using continuous variables were made by use of independent t tests, if the variables were normally distributed, and Mann-Whitney U tests were used for non-normal data. The 2 test was used to compare categorical variables such as anemia status (hemoglobin < 110 g/L) and iron depletion status (ferritin < 12 g/L) between the two treatment groups. The dependent t test was used to determine changes in continuous variables such as hemoglobin and ferritin concentration that occurred within each treatment group. Multiple regression and binary logistic regression techniques were used to adjust for possible confounding by variables such as gestational age, CRP, history suggestive of malaria, vitamin A and baseline hemoglobin or ferritin concentrations. The SPSS 9 for Windows (SPSS Inc., Chicago, IL) and Epi Info version 6.04b were used for statistical analysis. Treatment codes were broken only after all the data analysis had been completed. Ethics and consent. The study proposal was approved by the Cornell University Committee on Human Subjects and the Research and Ethics Committee of the TFNC in Tanzania. One author of the study, who is also a local physician, explained the purpose and requirements of the study to the women and invited them to participate. Only those women who gave informed consent were enrolled in the study. The study participants benefited from the regular obstetric and medical care received from the physicians and other health workers involved in the conduct of the study.

RESULTS
There were 579 women who volunteered for the study and, of these, 439 met the study criteria and were enrolled (140 were excluded primarily because of advanced gestational age or because their hemoglobin concentration was <80 g/L). Of those enrolled, 121 women were lost to followup and 59 mothers delivered before their 8-wk postintervention visit (Fig. 1 ). Because delivery affects hemoglobin and other variables that were measured, the main statistical analyses were restricted to the 259 (59% of enrolled women; 127 experimental and 132 placebo) women who were still pregnant and were still in the study after 8 wk of supplementation, regardless of their gestational age at entry into the study. The mothers who were either lost to follow-up or were excluded from the study because they delivered before the end of the study had more advanced gestational age and consequently weighed more than mothers who remained in the study (Table 2). However, there were no differences between treatment groups with regard to delivering early or being lost to follow-up. There were also no differences in baseline iron and vitamin A status between mothers who completed the study and those who did not.

View larger version (38K): In this window In a new window FIGURE 1 Selection of study participants and reasons for loss to follow-up in a randomized double-blind trial to investigate the effect of a micronutrient-fortified beverage in pregnant Tanzanian women. View this table: In this window In a new window TABLE 2Comparison of anthropometric, obstetric, parasitic infestation and supplement intake measurements of pregnant Tanzanian women enrolled in a micronutrient-fortified beverage clinical trial1 There were no significant differences between treatment groups at baseline with regard to age, education level, parity or gravidity, previous stillbirths and abortions, weight, height, body mass index, skinfold thickness and mid-upper arm circumference (Table 2). When the mothers recall of LMP was used to determine gestational age (n = 188 is based on the number of women who could recall their LMP), women in the fortified-beverage group had a higher gestational age (25

compared with 23 wk, independent t test, P = 0.01) than those in the control group. The three different indicators of gestational age were significantly correlated to one another [gestational age LMP to gestational age HOF (Spearmans r = 0.63, P < 0.001, n = 187); gestational age LMP to gestational age SFH (Spearmans r = 0.62, P < 0.001, n = 184); and gestational age HOF and that of gestational age SFH (Spearmans r = 0.87, P < 0.001, n = 253)]. The authors chose to use the gestational age LMP variable as an indicator of gestational age in all further analysis because the variable was normally distributed and was correlated to the other measures of gestational age. Of the 259 women, 69% were believed to have consumed at least 90% of the prescribed supplement, whereas 93.4% consumed at least 70%. There was no observed difference in compliance between treatment groups. The prevalence of parasitic infestation was low and did not differ between treatment groups, nor did history of malarial infection or routine iron/folic acid supplementation. Anemia. At baseline, 61.4% of the study participants were anemic (hemoglobin level <110 g/L), with no significant difference in the proportion of anemic women between the experimental and the control groups. At the end of 8 wk of supplementation, both groups experienced a decrease in the proportion of women with anemia. The proportion of anemic women declined from 59.1 to 48.5% among women in the placebo group and from 63.8 to 37% among women in the fortified group. At the end of supplementation, the placebo group had a significantly higher proportion of anemic women compared with the fortified group (48.5 compared with 37%, P = 0.019). At the end of the study, women belonging to the fortified-beverage group were 51% less likely to be anemic than were women in the placebo group (logistic regression: OR 0.49, CI 0.280.85, P = 0.01), after adjusting for baseline anemia status and vitamin A concentration. Hemoglobin change. Both groups experienced a significant increase in hemoglobin concentrations, but mothers in the fortified-beverage group had a higher mean increase of 8.62 g/L, whereas mothers in the placebo group had a mean increase of only 4.46 g/L (Table 3). The increase experienced by women in the fortified group was 4.16 g/L higher than the increase experienced by women in the placebo group (P = 0.015). Furthermore, the mean postsupplementation hemoglobin concentration of women in the fortified-beverage group was significantly higher than the postsupplementation hemoglobin concentration of women in the placebo group (114 g/L compared with 109 g/L, P = 0.02) (Table 3). View this table: In this window In a new window TABLE 3Comparison of hemoglobin, ferritin, serum retinol and CRP concentrations before and after supplementation and between experimental and placebo women1

During stepwise multivariable regression, the following factors, in order of importance, were found to be important predictors of change (or increase) in hemoglobin: treatment group; treatment group and gestational age interaction term; baseline hemoglobin; gestational age; baseline serum retinol; baseline ferritin < 30 g/L; and CRP (Table 4). Belonging to the fortified group and baseline serum retinol were positively associated, whereas baseline hemoglobin, CRP, ferritin < 30 g/L and treatment group x gestational age interaction were negatively associated with change in hemoglobin. This model was able to predict 36% of the variation in the hemoglobin change variable. According to this model, the effect of the treatment on change (increase) in hemoglobin depended on the mothers gestational age at entry into the study with those mothers who had earlier gestational ages and experienced a greater increase in hemoglobin than those whose gestational age was more advanced at entry into the study. This model suggested that women with lower baseline hemoglobin and serum ferritin < 30 g/L experienced a higher increase in hemoglobin. View this table: In this window In a new window TABLE 4Effect of fortified micronutrient supplement on change (increase) in hemoglobin concentration (g/dL) between baseline and 8-wk follow-up in pregnant women in Tanzania (N = 174)1 Iron deficiency and iron deficiency anemia. Serum ferritin levels provide an indication of body stores of iron rather than circulating iron levels. At baseline, there was no significant difference in the mean ferritin concentration of the two groups (18.47 compared with 18.91 g/L) (Table 3). However, at the 8-wk follow-up, the mean ferritin level was significantly higher (4.5 g/L) in the fortified-beverage group compared with that of the placebo group (P = 0.009). The fortified beverage was shown to result in a net improvement in ferritin concentration of 5 g/L. Women in the fortified-beverage group had a significant mean increase of 3 g/L, whereas women in the placebo group experienced a nonsignificant mean decrease of 2 g/L (Table 3) (P = 0.17). The proportion of iron-depleted (ferritin < 12 g/L) women at entry into the study was equally distributed between the two treatment groups (54.1% in the experimental group compared with 45.9% in the placebo group; 2 = 0.90, df 1, P = 0.34). At follow-up, although there was still no significant difference in the proportion of iron-depleted women between the two groups, 42.1% of iron-depleted women were now in the experimental group and 57.9% in the placebo group ( 2 = 1.75, P = 0.23). Further analysis showed that 9.6% (8 of 83) of women in the fortified group and 20.4% (18 of 88) of women in the control group who were iron replete at entry into the study became iron deficient during the course of the study. Based on the above proportions, the risk of becoming iron deficient during the 8 wk of the study was reduced by 53% among women in the fortified group compared to women in the placebo group Relative Risk (RR) = 0.47, MH2 = 0.05). After adjusting for whether study women were iron deficient at entry into study, the baseline CRP, the follow-up CRP and the interaction between iron depletion status and treatment group, the supplement resulted in a 70% reduction in the risk of being iron deficient at follow-up

in those who were iron deficient at baseline and a 92% decrease in those who were not iron deficient at baseline. By use of a ferritin cutoff of 12 g/L and a hemoglobin cutoff of 110 g/L, 19% of the 259 mothers were classified as having iron deficiency anemia. There was no significant difference in the prevalence of iron deficiency anemia in the two groups, both at entry and at the end of 8 wk of supplementation. However, after adjusting for whether the subject had iron deficiency anemia at entry into the study, supplementation reduced the risk of having iron deficiency anemia at follow-up by 56% (logistic regression: RR = 0.44, CI 0.191.02, Wald 3.67, P = 0.05). Vitamin A status. DBS retinol levels were determined both at the baseline and follow-up examinations; 54% of women had levels below normal (1.05 mol/L). Women in both groups experienced a significant decline in mean serum retinol levels (Table 3). There were no differences in DBS retinol concentration in women in either the fortified-beverage or placebo groups, either at entry or at the end of the 8 wk of supplementation. Thyroid function. There was no effect on TSH (Table 3).

DISCUSSION
Data from this study confirmed that micronutrient deficiencies are prevalent in the female population of Tanzania and the prevalence of anemia (63%) and vitamin A deficiency (54%) among study participants is consistent with findings obtained by other Tanzanian researchers (32 35 ). Furthermore, this study showed that after 8 wk of supplementation, a micronutrientfortified dietary supplement containing various micronutrients in physiological doses reduced the risk of anemia during pregnancy by 51% and the risk of iron deficiency anemia by 56%. The supplement reduced the risk of becoming iron deficient or developing iron deficiency during the study by 53%. The supplement further increased hemoglobin concentration by 4.16 g/L. These findings should be evaluated in the light of a recent meta-analysis of controlled iron supplementation trials, in which Sloan et al. (36 ) found that a daily dose of 114 65 mg of iron is associated with a 10.0 0.13 g/L increase in the hemoglobin levels of women in an experimental group compared with those of control subjects and an increase of 12 0.23 g/L when a dose of 132 76 mg was given in combination with other nutrients or antimalarial prophylaxis. In our study, a multiple micronutrient supplement containing 10.8 mg of iron together with other micronutrients resulted in a 4.16 g/L increase in hemoglobin. A number of studies have shown that serum ferritin levels tend to decline as pregnancy proceeds (37 44 ). Some of these same studies (38 44 ) also demonstrated that iron supplementation using daily doses of elemental iron, ranging from 60 to 120 mg, reduces this decline in ferritin concentration. In the meta-analysis referred to above, Sloan et al. (36 ) found that in studies reporting ferritin results, iron supplementation alone (132 77 mg) increases serum ferritin by

9.48 g/L and hemoglobin by 8.7 g/L. In our study, women receiving 10.8 mg of iron daily and 1050 retinol equivalents (RE) of vitamin A daily had a significant increase in serum ferritin (mean 3 g/L). The increase in hemoglobin observed in our study could be explained partly by an increase in iron absorption attributed to vitamin C (45 ) and increased use of iron in hemoglobin synthesis attributed to vitamin A and riboflavin (45 ). Micronutrients such as folic acid and vitamin B-12 are also likely to have contributed to the increase in hemoglobin by correcting megaloblastic anemia (45 ). Because the study was not designed to test the effect of individual nutrients, nor was it designed to test the effect of the supplement on megaloblastic anemia, the current study was not able to test the above explanations. The relatively smaller increase in ferritin observed in this study in relation to hemoglobin increase may also be attributable to the utilization of iron in hemoglobin synthesis. In a study conducted in Indonesia by Muslimatun et al. (44 ), a combination of 6000 RE vitamin A and 120 mg elemental iron given weekly resulted in a significant decrease in serum ferritin, and the authors believe that the decline in serum ferritin in the vitamin A + iron group is attributed to increased utilization of iron in hemoglobin synthesis. In Indonesian women receiving 2400 RE daily for 8 wk, Suharno et al. (25 ) found that women in the supplemented group had a 0.2 mol/L increase in serum retinol compared with that of women in the control group. Given that the participants in this current study received 1050 RE daily for 8 wk during pregnancy, it is not surprising that this supplement did not have a significant effect on serum retinol. An increase in the concentration of vitamin A in the supplement and/or longer duration of supplementation should be considered. The supplement was popular with the participating women and compliance rates were high, with at least 69% of the women consuming at least 90% of the intended dose of the supplement. On the contrary, the intake of the iron/folic acid supplements was low, with only 3 of 259 women reporting use of supplements at entry into the study and 72 (28%) reporting use during some part of the study. It would seem that the increased attention the researchers paid to the womens intake of iron/folic acid supplements and the encouragement of the researchers influenced the ANC health workers to dispense more iron/folic acid supplements. The iron/folic acid supplementation did not influence the effect of the supplement on hemoglobin, ferritin or vitamin A change, nor was there any significant association between iron and folic acid supplement use and the outcomes of interest in this study. The overall increase in hemoglobin in this study was most likely the result of seasonal factors associated with lesser rates of malaria attack, increased food availability during the time of the study and the increased use of iron/folic acid supplements; the treatment of intestinal helminth infestations among women taking part in the study could also have contributed to the increase in hemoglobin. This increase occurred in both treatment groups. The study had some limitations that either were inherent in the design or were associated with difficulties in the conduct of the study. The exclusion of women with hemoglobin concentrations < 80 g/L (n = 36) resulted in the removal of women who would have had a greater potential to benefit from the supplementation. However, despite the exclusion of these women, the micronutrient supplement did have a beneficial effect on hemoglobin. The study was designed to test the efficacy of a multiple micronutrient supplement on the micronutrient status of pregnant

women who were also receiving the usual iron/folic acid supplements provided at ANC clinics in Tanzania. However, the low compliance with the usual iron/folic acid supplementation makes the study more of an effectiveness study, in which the effect of the micronutrient supplement was evaluated under usual ANC practices in rural Tanzania. The two treatment groups did not differ in the proportion of women who took iron/folic acid supplements during the study and iron/folic acid supplement intake had no effect on hemoglobin change or anemia status. During multivariable analysis, whether study participants took iron/folic acid supplements did not seem to confound the effect of the micronutrient on hemoglobin or anemia status. Although iron/folic acid intake did not seem to introduce bias into results of the study, the results would have been more valid if all participating women had been supplied with iron/folic acid supplements by the researchers. To be able to recruit an adequate number of study participants, women with a wide range of gestational ages and in different trimesters of pregnancy were included in the study. Although the researchers were able to control for gestational age in multivariable analysis, the authors were not able to stratify the subjects by trimester because some of the strata were too small. The multivariable analysis did suggest that women who were at earlier gestational ages benefited the most from the supplementation. The inclusion of women who were 34 wk pregnant also resulted in 13.4% of women delivering before the end of the 8 wk of supplementation, thus contributing to the high loss to follow-up rate. The 41% loss to follow-up was a major limitation of this study and had a potential to introduce bias into results of the study. However, because the study participants who completed the study and those who did not had similar baseline characteristics, except for gestational age at entry into the study and weight at follow-up (for those who had delivered at follow-up), the impact of the loss to follow-up was minimized. The modest improvements in iron status in this study need to be placed in context. The supplement provided physiological, not therapeutic, doses of micronutrients. The supplement, although not sufficient to cure serious deficiency, was able to modestly improve the iron status of iron-deficient pregnant women. This was consistent with the preventative, rather than curative, objective of this physiological-dose approach. This dietary supplement was not intended as a therapy for anemia, but rather as a preventative measure. Comparisons should be made, not with medicinal iron supplements given to cure anemia, but with interventions such as food fortification, food diversification and nutrition education that were designed to improve micronutrient intake and prevent deficiency. The approach of these interventions is also preventative rather than curative. Ideally, if women of childbearing age consumed the dietary supplement routinely, pregnancy could be undertaken in good nutritional status, thereby optimizing their chances for a safe delivery and a healthy child. On the basis of results of this study, we concluded that dietary supplements such as the one tested here offer a possible approach to be used in concert with, not in place of, other strategies to control the anemia prevalent in pregnant women. Furthermore, although this study failed to find a significant effect on vitamin A and thyroid status in pregnant women, the supplement, because it contains multiple micronutrients, has the potential for controlling several micronutrient deficiencies at the same time.

The micronutrient content of the supplement could be further tailored to meet the needs of micronutrient-deficient populations. Because the minerals and vitamins are in physiological doses, the use of the supplement would not require medical consultation, thus empowering families and providing them with a tool to use in their goal of improving maternal health and nutrition.

ACKNOWLEDGMENTS
We thank Wilbald Lorri, Managing Director, and the staff of the Tanzania Food and Nutrition Center, especially M. Maganga, J. Amri and R. Kitwenga, for important assistance with this research. We greatly appreciate the efforts of the Procter & Gamble Company of Cincinnati, who developed and supplied the supplement at no cost. Finally, we extend thanks to the Tanzanian women who enthusiastically participated in the study.

FOOTNOTES
1

Supported by a Micronutrient Initiative grant, Procter & Gamble Company, UNICEF Tanzania, Tanzania Food and Nutrition Centre and Cornell University.
3

Abbreviations used: CRP, C-reactive protein; DBS, dried blood spot; FAO, Food and Agriculture Organization of the United Nations; HOF, height of uterine fundus; LMP, last menstrual period; MUAC, mid upper arm circumference; NRC, National Research Council; SFH, symphysis fundal height; RE, retinol equivalents; RR, relative risk; TFNC, Tanzania Food and Nutrition Centre; TSH, thyroid-stimulating hormone. Manuscript received 22 December 2002. Initial review completed 16 January 2003. Revision accepted 14 February 2003.

LITERATURE CITED
1. 1. World Health Organization (WHO) (1992) The Prevalence of Anemia in Women: A Tabulation of Available Information 2nd ed. 1992 World Health Organization Geneva, Switzerland.
y

2. Latham, M. C. (1997) Human Nutrition in the Developing World 1997 Food and Agriculture Organization of the United Nations Rome, Italy. 3. Hercberg, S., Galan, P. & Dupin, H. (1987) Iron deficiency in Africa. World Rev. Nutr. Diet. 54:201-236.[Medline] 4. Van den Broek, N. R. & Letsky, E. A. (2000) Etiology of anemia in pregnancy in South Malawi. Am. J. Clin. Nutr. 72:247S-256S.[Abstract/Free Full Text] 5. Huddle, J. M., Gibson, R. S. & Cullinan, T. R. (1998) Is zinc a limiting nutrient in the diets of rural pregnant Malawian women?. Br. J. Nutr. 79:257-265.[Medline]

6. Azais-Braesco, V. & Pascal, G. (2000) Vitamin A in pregnancy: requirements and safety limits. Am. J. Clin. Nutr. 71:1325S-1333S.[Abstract/Free Full Text] 7. Underwood, B. A. & Arthur, P. (1996) The contribution of vitamin A to public health. FASEB J. 10:1040-1048.[Abstract] 8. Viteri, F. E. (1994) The consequences of iron deficiency and anemia in pregnancy. Adv. Exp. Med. Biol. 352:127-139.[Medline] 9. Merchant, K. M. (1994) Maternal nutritional depletion 1994SCN News 11. 10. Brems, S. & Berg, A. (1988) Eating Down during Pregnancy: Nutrition, Obstetric and Cultural Considerations in the Third World. Discussion paper prepared for ACC/SCN 1988 World Bank Washington, DC. 11. Bothwell, T. H. (2000) Iron requirements in pregnancy and strategies to meet them. Am. J. Clin. Nutr. 72:257S-264S.[Abstract/Free Full Text] 12. Hughes, A. (1991) Anemia in pregnancy. Maternal Health and Safe Motherhood 1991 Division of Family Health, WHO Geneva, Switzerland. 13. Viteri, F. E. (1992) Iron, a global perspective. Ending Hidden Hunger: A Policy Conference on Micronutrient Malnutrition 1992 Atlanta GA. 14. Ladipo, O. A. (2000) Nutrition in pregnancy: mineral and vitamin supplements. Am. J. Clin. Nutr. 72:280S-290S.[Abstract/Free Full Text] 15. Hallberg, L. (1981) Iron balance in pregnancy and lactation. Foman, S. J. Zlotkin, S. eds. Nutritional Anemias 1981:13-28 Raven Press New York, NY. . 16. Van Den Broek, N. (1998) Anaemia in pregnancy in developing countries. Br. J. Obstet. Gynaecol. 105:385-390.[Medline] 17. Fleming, A. F. (1969) Iron status of anaemic pregnant Nigerians. J. Obstet. Gynaecol. Br. Commonw. 76:1013-1017.[Medline] 18. Van Den Broek, N. R., White, S. A. & Neilson, J. P. (1998) The relationship between asymptomatic human immunodeficiency virus infection and the prevalence and severity of anemia in pregnant Malawian women. Am. J. Trop. Med. Hyg. 59:10041007.[Abstract] 19. Rush, D. (2000) Nutrition and maternal mortality in the developing world. Am. J. Clin. Nutr. 72:212S-240S.[Abstract/Free Full Text] 20. Ratten, G. J. & Beischer, N. A. (1972) The significance of anemia in an obstetric population in Australia. J. Obstet. Gynaecol. Br. Commonw. 79:228-237.[Medline]

y y

21. Harrison, K. A. & Ibeziako, P. A. (1973) Maternal anaemia and fetal birthweight. J. Obstet. Gynaecol. Br. Commonw. 80:798-804.[Medline] 22. Huffman, S. L., Baker, J., Shumann, J. & Zehner, E. R. (1998) The Case for Promoting Multiple Vitamin/Mineral Supplements for Women of Reproductive Age in Developing Countries 1998 Linkages Project, Academy for Educational Development Washington, DC. 23. Ramakrishnan, U., Manjrekar, R., Rivera, J., Gonzales-Cossio, T. & Martorell, R. (1999) Micronutrients and pregnancy outcome: a review of the literature. Nutr. Res. 19:103-159. 24. Panth, M., Shatrugna, V., Yasodhara, P. & Sivakumar, B. (1990) Effect of vitamin A supplementation on haemoglobin and vitamin A levels during pregnancy. Br. J. Nutr. 64:351-358.[Medline] 25. Suharno, D., West, C. E., Muhilal, , Karyadi, D. & Hautvast, J. (1993) Supplementation with vitamin A and iron for nutritional anemia in pregnant women in West Java, Indonesia. Lancet 342:1325-1328.[Medline] 26. Latham, M. C., Ash, D., Ndossi, G., Mehansho, H. & Tatala, S. (2001) Micronutrient dietary supplementsa new fourth approach. Arch. Latinoam. Nutr. 51:37-41. 27. Mehansho, H. (2002) Eradication of iron deficiency anemia through food fortification: the role of the private sector. J. Nutr. 132:831S833S.[Abstract/Free Full Text] 28. Li, P. K., Humbert, J. R. & Cheng, C. S. (1978) Evaluation of a commercially obtainable ferritin test kit in relation to the high-dose parabolic phenomenon. Clin. Chem. 24:1650-1651.[Free Full Text] 29. Naik, P. & Voller, A. (1984) Serum C-reactive protein levels and falciparum malaria. Trans. R. Soc. Trop. Med. Hyg. 78:812-813.[Medline] 30. Miyai, K., Ishibashi, K. & Kawashima, M. (1981) Two-site immunoenzymometric assay for thyrotropin in dried blood samples on filter paper. Clin. Chem. 27:14211423.[Abstract/Free Full Text] 31. Craft, N. E., Bulux, J., Valdez, C., Li, Y. & Solomons, N. W. (2000) Retinol concentrations in capillary dried blood spots from healthy volunteers: method validation. Am. J. Clin. Nutr. 72:450-454.[Abstract/Free Full Text] 32. Massawe, S. N., Urassa, E. N., Nystrom, L. & Lindmark, G. (1999) Effectiveness of primary level antenatal care in decreasing anemia at term in Tanzania. Acta Obstet. Gynecol. Scand. 78:573-579.[Medline]

33. Matteelli, A., Donato, F., Shein, A., Muchi, J. A., Leopardi, O., Astori, L. & Carosi, G. (1994) Malaria and anemia in pregnant women in urban Zanzibar, Tanzania. Ann. Trop. Med. Parasitol. 88:475-483.[Medline] 34. Hinderaker, S. G., Olsen, B. E., Bergsjo, P., Lie, R. T., Gasheka, P. & Kvale, G. (2001) Anemia in pregnancy in the highlands of Tanzania. Acta Obstet. Gynecol. Scand. 80:18-26.[Medline] 35. Bergsjo, P., Seha, A. M. & Ole-Kingori, N. (1996) Hemoglobin concentration in pregnant women. Experience from Moshi, Tanzania. Acta Obstet. Gynecol. Scand. 75:241-244.[Medline] 36. Sloan, N. L., Jordan, E. & Winikoff, B. (2002) Effects of iron supplementation on maternal hematologic status in pregnancy. Am. J. Public Health 92:288293.[Abstract/Free Full Text] 37. Kaneshige, E. (1981) Serum ferritin as an assessment of iron stores and other hematologic parameters during pregnancy. Obstet. Gynecol. 57:238-242.[Medline] 38. Taylor, D. J., Mallen, C., Mcdougall, N. & Lind, T. (1982) Effect of iron supplementation on serum ferritin levels during and after pregnancy. Br. J. Obstet. Gynaecol. 89:1011-1017.[Medline] 39. Guldholt, I. S., Trolle, B. G. & Hvidman, L. E. (1991) Iron supplementation during pregnancy. Acta Obstet. Gynecol. Scand. 70:9-12.[Medline] 40. Milman, N., Agger, A. O. & Nielsen, O. J. (1991) Iron supplementation during pregnancy. Effect on iron status markers, serum erythropoietin and human placental lactogen. A placebo controlled study in 207 Danish women. Dan. Med. Bull. 38:471476.[Medline] 41. Milman, N., Graudal, N. & Agger, A. O. (1995) Iron status markers during pregnancy. No relationship between levels at the beginning of the second trimester, prior to delivery and post partum. J. Intern. Med. 237:261-267.[Medline] 42. Milman, N., Graudal, N., Nielsen, O. J. & Agger, A. O. (1997) Serum erythropoietin during normal pregnancy: relationship to hemoglobin and iron status markers and impact of iron supplementation in a longitudinal, placebo-controlled study on 118 women. Int. J. Hematol. 66:159-168.[Medline] 43. Trugo, N.M.F., Donangelo, C. M., Seyfarth, B.S.P., Henriques, C. & Andrade, L. P. (1996) Folate and iron status of non-anemic women during pregnancy: effect of routine folate and iron supplementation and relation of erythrocyte folate with iron stores. Nutr. Res. 16:1267-1276.

44. Muslimatun, S., Schmidt, M. K., Schultink, W., West, C. E., Hautvast, J. A., Gross, R. & Muhilal, (2001) Weekly supplementation with iron and vitamin A during pregnancy increases hemoglobin concentration but decreases serum ferritin concentration in Indonesian pregnant women. J. Nutr. 131:85-90.[Abstract/Free Full Text] 45. Fishman, S. M., Christian, P. & West, K. P. (2000) The role of vitamins in the prevention and control of anaemia. Public Health Nutr. 3:125-150.[Medline] 46. National Research Council (NRC) (1989) Recommended Dietary Allowances 1989 National Academy Press Washington, DC. 47. FAO/WHO (1988) Requirements for Vitamin A, Iron Folate and Vitamin B-12. Report of a Joint FAO/WHO Expert Consultation 1988 Food and Agriculture Organization of the United Nations/World Health Organization Rome, Italy. 48. FAO/WHO (1985) Requirements of Vitamin A, Thiamin, Riboflavin, and Niacin 1985 Food and Agriculture Organization of the United Nations/World Health Organization Rome, Italy.