J. Nutr. 2004 Ramakrishnan 898 903

Community and International Nutrition

Multiple Micronutrient Supplements during Pregnancy Do Not ReduceAnemia or Improve Iron Status Compared to Iron-Only Supplements inSemirural Mexico1,2

Usha Ramakrishnan,*,3 Lynnette M. Neufeld,† Teresa Gonzalez-Cossıo,†

Salvador Villalpando,† Armando Garcıa-Guerra,† Juan Rivera,† and Reynaldo Martorell*

*Department of International Health, Rollins School of Public Health, Emory University, Atlanta, GA; and†Centro de Investigacion en Nutricion y Salud, Instituto Nacional de Salud Publica (INSP), Cuernavaca,Morelos, Mexico

ABSTRACT The impact of iron-only supplements (FE) versus multiple micronutrient supplements containing iron(MM) during pregnancy on iron status was assessed in a subsample (n � 453) of women who participated in arandomized double-blind trial in Mexico. Compliance, monitored by observation, was high (�85%). The two groupswere similar at recruitment (�13 wk gestation) for various sociodemographic characteristics and for meanhemoglobin (Hb) concentrations and prevalence of anemia (Hb � 110 g/L; 11%). However, mean serum ferritin washigher (P � 0.05) in the MM group (n � 142) compared to the FE group (n � 148) and the prevalence of irondeficiency (serum ferritin � 12 �g/L) was lower in the MM group (44.4%) compared to the FE group (57.4%). Bythe third trimester, almost half the women were anemic in both groups, and mean Hb (g/L) was lower for the MMgroup (104.2; 95% CI: 102.5, 106.0) compared to the FE group (108.1; 95% CI: 106.4, 109.8) after adjusting forbaseline serum ferritin. In contrast, there were no differences in Hb concentrations at 1 mo postpartum or in meanferritin and prevalence of iron deficiency at 32 wk gestation and 1 mo postpartum (90.9 and 45.1% for the MMgroup; 92.6 and 47.3% for the FE group, respectively). In conclusion, rather than improve Hb or iron status relativeto FE-only supplements as hypothesized, MM supplements may have slightly reduced Hb concentrations duringpregnancy. Neither supplement was able to meet iron needs as evidenced by dramatic increases in anemia andiron deficiency by the end of pregnancy. J. Nutr. 134: 898–903, 2004.

KEY WORDS: ● iron ● multivitamin-mineral ● supplements ● pregnancy ● anemia

Anemia is a substantial public health problem in manydeveloping countries and has been associated with a range ofadverse consequences including poor mental development,reduced productivity, maternal mortality, and low birth weight(1–3). Iron deficiency is considered the main cause of anemia,especially among young children and pregnant women, whoare at increased risk due to their increased requirements (4).Anemia during pregnancy, however, remains a problem inmany settings despite the fact that routine provision of ironsupplements has been recommended for pregnant women(5,6). The failure of iron supplementation programs to reduceanemia in pregnant women has been attributed to various

factors that influence program delivery. These include lack ofavailability of supplements, poor coverage, inadequate pro-vider knowledge, and poor compliance due to lack of motiva-tion and/or side effects (6). However, more recently the effi-cacy of iron supplements has been questioned given thecomplex etiology of anemia.

Anemia may result from both nutritional and nonnutri-tional factors. Specifically, besides iron, deficiencies of micro-nutrients such as vitamins A, C, and B-12 and folate maycontribute to the development of anemia (4). These nutrientsmay affect hemoglobin (Hb)4 synthesis either directly or in-directly by affecting absorption and/or mobilization. For ex-ample, vitamin A has been shown to play a role in themobilization of iron for hematopoiesis and studies have shownthat vitamin A supplementation along with routine iron sup-plements during pregnancy substantially reduced the preva-lence of anemia when compared to only iron supplements(7,8). Similarly, vitamin C is known to enhance iron absorp-tion. Folate and vitamin B-12 deficiency can independently

1 Presented in part at the Experimental Biology Meeting, April 2001, Orlando,FL, and at the ILSI/Emory/CDC Conference on “Forging Effective Strategies forthe Control of Iron Deficiency,” Atlanta, GA, May 7–9, 2001 [Neufeld, L. M.,Ramakrishnan, U., Rivera, J., Villalpando, S., Gonzalez-Cossio, T. & Martorell, R.(2001). Prevalence of anemia and iron deficiency during pregnancy of womensupplemented with iron or iron and multiple micronutrients. FASEB J. 15: Abstract505.2].

2 Supported by the Thrasher Research Fund, NIH Grant HD-34531–05,UNICEF, New York, NY, Conacyt, and INSP, Mexico. None of the authors has anyfinancial or personal interest with the above organizations.

3 To whom correspondence should be addressed.E-mail: [email protected].

4 Abbreviations used: CRP, C-reactive protein; FE, iron only; Hb, hemoglobin;IDA, iron-deficiency anemia; INSP, Instituto Nacional de Salud Publica; MM,multiple micronutrients; RCT, randomized controlled trial.

0022-3166/04 $8.00 © 2004 American Society for Nutritional Sciences.Manuscript received 28 October 2003. Initial review completed 1 December 2003. Revision accepted 5 January 2004.

898

by guest on May 24, 2012

jn.nutrition.orgD

ownloaded from

http://jn.nutrition.org/

cause megaloblastic anemia that differs from microcytic iron-deficiency anemia by affecting DNA synthesis.

In many developing countries, diets that are poor in ironare also poor in several other nutrients due to low intakes ofanimal foods and high intakes of foods rich in absorption-inhibiting factors such as phytates. This has therefore raisedinterest in providing nutrients besides iron to reduce theprevalence of anemia (9,10), but few studies have evaluatedthe impact of multivitamin-mineral supplements on anemiaand iron status during pregnancy. Two randomized controlledtrials (RCT) from Tanzania (11) and Nepal (12,13) haveevaluated the benefits of multiple micronutrient supplementson pregnancy outcomes, but only one examined iron statusand found that multiple micronutrients did not improve he-matologic indicators when compared to patients who receivediron-folate supplements (13). We recently completed an RCTthat compared the effect of a daily prenatal multivitamin-mineral supplement to iron-only supplements on birth out-comes (14) and examine the effects on anemia and iron statusin this paper.

METHODS

Study setting and design. An RCT was carried out during 1997–2000 to compare the efficacy of a multiple micronutrient (MM)supplement compared to iron-only (FE) during pregnancy to improvebirth outcomes in a semirural community near the city of Cuernavacain Morelos, Mexico. This study was a collaborative project betweenthe Department of International Health at Rollins School of PublicHealth, Emory University (Atlanta, GA) and the Centro de Inves-tigacion en Nutricion y Salud, Instituto Nacional de Salud Publica(INSP) (Cuernavaca, Mexico). The MM supplement was designed toprovide 100–150% of the recommended dietary allowance (15) ofkey vitamins (700 �g retinol, 2 �g vitamin B-12, 66.5 mg vitamin C)and minerals (15 mg Zn) and was similar to supplements that arecommercially available. The control group received only iron, thestandard practice of the Ministry of Health in Mexico at the time thestudy was conducted. Both supplements contained 60 mg of iron inthe form of ferrous sulfate. Details of supplement content, studyeligibility, and recruitment are described elsewhere (14). Informedconsent was obtained from all women who agreed to participate andthey were then randomly allocated to either the MM or the FE group.

Data collection. At recruitment, the study physician and a teamof trained nurses conducted a prenatal examination that included adetailed obstetric history, physical examination, and anthropometricassessment at the study headquarters. The first supplement was con-sumed on site, following which women were visited at their homes 6days a week until delivery by trained workers who administered andrecorded the consumption of supplements. Socioeconomic status wasdetermined using a questionnaire regarding household building ma-terials, possessions, and occupation, from which an index of economicstatus was derived using factor analysis (14). Venous blood sampleswere collected at the field headquarters by trained nurses from willingsubjects at baseline, 32 wk of completed gestation, and 1 mo post-partum as part of routine prenatal and postpartum care. The bloodsamples were centrifuged at ambient temperature for 15 min at 2000� g. Serum was transferred to trace element–free microtubes, frozenimmediately at �20°C, and transferred within 1 week to �70°C untilanalysis.

Samples were analyzed at the INSP nutrition laboratories forserum ferritin and C-reactive protein (CRP) concentrations. Thequantitative measurement of ferritin in serum was determined bysandwich immunoassay (ELISA, Opus Behring Laboratories) usingcommercial kits (Dade Behring). CRP in serum was measured usingan immunonephelometry system (16,17), in which polystyrene par-ticles coated with monoclonal antibodies to CRP are agglutinatedwhen mixed with serum samples containing CRP.

Hemoglobin concentrations were measured in the field headquar-ters at the same time points by trained nurses using a portablephotometer (Hemocue) from a capillary blood sample obtained by

finger-prick. Appropriate referral and treatment for high-risk preg-nancies were provided by the study physician, who worked closelywith the local health authorities. Data entry and cleaning werecarried out on an ongoing basis with supervision by INSP staff at themain office in Cuernavaca. Additional data cleaning was carried outat Emory University.

Data analysis. The main outcome variables were measures ofiron status and anemia at 32 wk gestation and 1 mo postpartum.Because serum ferritin was not normally distributed, loge transformedvalues were used. A small value of 0.0001 was added to all values toensure appropriate transformation of “zero” values. Anemia was de-fined as hemoglobin concentrations below 110, 105, and 120 g/L atrecruitment, 32 wk gestation and 1 mo postpartum, respectively,whereas iron deficiency was defined as serum ferritin below 12 �g/L atall time points (18,19). Iron-deficiency anemia was indicated by thepresence of both anemia and iron deficiency. Gestational age wasbased on recalled date of last menstrual period and overweight wasdefined as BMI above 25 kg/m2 (20). Compliance was calculated byexpressing the total number of tablets consumed as a percentage ofthe total number that they could have consumed (6 d per wk fromrecruitment to delivery).

The study sample for this analysis was a subsample of all pregnan-cies assigned to treatment between July 1997 and December 31, 1999,that resulted in singleton live births and had data on measures of ironstatus at recruitment, 32 wk gestation, and 1 mo postpartum. Com-parisons between the final sample and those not included were donefor selected baseline characteristics and measures of compliance andthe comparability of the two treatment groups in the final sample wastested for selected sociodemographic, health, and nutrition charac-teristics of the women at recruitment. These comparisons were doneusing Student’s t tests for means for normally distributed variables andchi-square tests of proportions for categorical variables.

Mean hemoglobin and serum ferritin concentrations were com-pared between supplement groups using general linear models tocontrol for factors that differed between groups at recruitment. In thecase of binary outcomes, namely anemia, iron-deficiency anemia(IDA), and iron deficiency, multivariate logistic regression modelswere used to compare the treatment groups. In both cases, the role ofoutliers and suitability of the models were examined. In addition,effect modification by characteristics selected a priori to data analysis(iron status and maternal BMI at recruitment) was tested. All statis-tical analyses were conducted using SAS 8.2. A P value � 0.05 formain effects and P � 0.15 for interaction terms were consideredsignificant. Adjustments were also done for multiple comparisons asappropriate.

RESULTS

A total of 921 pregnancies were identified, of which 873were assigned to treatment after pregnancy was confirmed,eligibility was determined, and informed consent was obtainedand 645 of these pregnancies resulted in singleton live births.Loss to follow-up was about 25%, the reasons for which areexplained elsewhere (14). Hemoglobin measurements wereavailable at baseline, 32 wk gestation, and 1 mo postpartum fora subsample of 453 of 645 women (70.2%). Serum ferritinmeasurements were also available at the 3 time points for asmaller sample of 290 pregnancies (Fig. 1). Although theoriginal design was to obtain venous blood samples only for a30% random subsample of all pregnancies, blood samples wereactually obtained from a larger number of women at thevarious time points and baseline data were available for he-moglobin and serum ferritin for 802 and 683 pregnancies,respectively, that were assigned to treatment. Based on samplesize requirements and budgetary constraints, serum ferritinestimations were, however, done for the later time points (32wk gestation and 1 mo postpartum) only for the first 300samples that had baseline serum ferritin values.

The comparison of selected maternal characteristics at re-cruitment (Table 1) for the final sample of 453 pregnancies

PRENATAL MICRONUTRIENT SUPPLEMENTS AND ANEMIA 899


jn.nutrition.orgD

ownloaded from


with complete hemoglobin data at baseline, 32 wk gestation,and 1 mo postpartum indicates that the two treatment groupswere not different for the majority of characteristics includingage, parity, number of weeks pregnant at entry, hemoglobinconcentration, and maternal height. The groups were notdifferent for maternal schooling, ethnicity, and economic sta-tus. However, as reported earlier, almost a third of the womenwere overweight, with a higher proportion (P � 0.05) in theFE group (38.8%) compared to the MM group (32.3%). Theprevalence of anemia at recruitment did not differ betweengroups, but serum ferritin was lower in the MM group com-pared to the FE group in the final sample.

The comparison of the final sample (n � 453) to those withincomplete blood data and/or those who were lost to follow-up(n � 420) did not reveal any significant differences in baselinecharacteristics except that women in the final sample were lesslikely to be primiparous (Table 2). The prevalence of anemiaat baseline was also higher in the MM group (16.8%) com-pared to the FE group (7.3%) among those who were lost tofollow-up and/or had incomplete blood data, but no differ-ences in iron deficiency based on serum ferritin were found.Compliance and the total number of supplements consumedwhile in the study were lower for those lost to follow-up, asexpected. Within the final sample, the subsample of 290pregnancies with complete serum ferritin data did not revealany differences in baseline characteristics from the rest of thesample (n � 583; data not presented).

The prevalence of anemia was about 11% at recruitmentand increased to 38% by 32 wk gestation and 44% at 1 mopostpartum. Compared to anemia, the prevalence of iron de-ficiency was much higher at recruitment and almost all womenwere iron deficient at 32 wk gestation, and about half re-mained deficient at 1 mo postpartum. The prevalence of IDAduring pregnancy suggests that most of the anemia was due toiron deficiency. About two-thirds of anemia was due to irondeficiency at 1 mo postpartum. Comparisons of the prevalenceof anemia, iron deficiency, and IDA by treatment group (Ta-ble 3) reveal no differences at 32 wk gestation and 1 mopostpartum. Although the prevalence of anemia was similar atboth time points (P � 0.05), iron deficiency was higher (P� 0.026) at baseline in the iron group and controlling forthese differences indicated a higher risk of anemia (P � 0.014)and IDA (P � 0.026) in the MM group compared to the FEgroup at 32 wk gestation. The adjusted odds ratios (AOR)were 2.05 (95% CI: 1.16, 3.62) and 1.88 (95% CI: 1.08, 3.28)for anemia and IDA, respectively. Further adjustment forbaseline differences in BMI attenuated these findings (P� 0.025). There were no differences in any of the outcomes at1 mo postpartum after adjustment for baseline iron status andBMI, nor did either modify the effect of supplementation onany outcomes.

Comparison of mean hemoglobin concentrations at 32 wkgestation and 1 mo postpartum by treatment group (Table 4)revealed that although mean hemoglobin concentrations weresimilar in both groups before adjusting for baseline serumferritin, mean Hb concentrations were significantly higher inthe FE group compared to the MM group at 32 wk gestationfollowing adjustment for baseline serum ferritin. Comparisonof mean serum ferritin (log transformed) concentrations at 32wk gestation and 1 mo postpartum revealed no significantdifferences by treatment group before and after adjusting forbaseline serum ferritin.

TABLE 1

Comparison of maternal characteristics at baselineand supplement consumption in the MM

and FE treatment groups1

Multiplemicronutrients

(n � 227)Iron only(n � 226)

Maternal age, y 22.9 � 5.3 23.2 � 5.4Duration of pregnancy, wk 9.1 � 2.2 9.4 � 2.3Primiparous, % 33.0 32.7Schooling, y 7.1 � 3.5 7.1 � 3.2Economic status2 0.01 � 1.1 0.14 � 1.1Indigenous ethnicity, % 32.7 29.0Single mother, % 2.2 5.0*Height, cm 148.9 � 5.0 148.8 � 4.6Weight, kg 52.4 � 9.1 54.1 � 10.2BMI, kg/m2 23.6 � 3.7 24.4 � 4.4*Serum ferritin,3 �g/L 14.5 � 20.1 9.7 � 14.9*Hemoglobin, g/L 125.8 � 13.8 125.2 � 13.5Duration of supplementation, wk 29.4 � 2.9 29.4 � 2.6Supplements consumed, n 167.0 � 18.5 167.5 � 17.1Compliance, % 88.0 � 5.8 88.2 � 5.1

1 Values are means � SD unless otherwise specified. * Differentfrom multiple micronutrients, P � 0.05.

2 This index was derived using factor analysis and based on qualityof housing, occupation, and household possessions.

3 Median � interquartile range; n � 176 and 185 for the multiplemicronutrients and iron-only groups, respectively, and log-transformedvalues were used for comparisons.

FIGURE 1 Flow chart for study samples with hemoglobin andserum ferritin measurements at baseline, 32 wk gestation, and 1 mopostpartum. 1A total of 361 also have baseline serum ferritin. 2Finalsample with Hb and serum ferritin measures at baseline, 32 wk gesta-tion, and 1 mo postpartum.

RAMAKRISHNAN ET AL.900


jn.nutrition.orgD

ownloaded from


DISCUSSION

Our results clearly indicate that multiple micronutrientsupplements during pregnancy were not more efficacious inreducing anemia or iron deficiency compared to iron-onlysupplements. The prevalence of anemia and iron deficiencywas similar in both groups at 1 mo postpartum and thesenull findings cannot be attributed to low compliance or lackof statistical power. Poststudy calculations revealed that,using a two-tailed test with a significance level of 0.05, oursample sizes of 453 and 290 had �80% power to detect a

difference of at least 4 g/L (0.28 SD) and 0.35 log �g/L(0.35 SD) in mean hemoglobin and serum ferritin concen-trations, respectively, which represent small to mediumeffect sizes. In fact, although mean hemoglobin concentra-tions at 32 wk gestation were similar in both groups, thedifferences were significant in favor of the iron group fol-lowing adjustment for the baseline serum ferritin that dif-fered by intervention group. Similarly, although the unad-justed prevalence of anemia was not significantly differentbetween treatment groups at 32 wk gestation, the risk ofanemia nearly doubled for the MM group compared to theFE group following adjustment for baseline serum ferritin.Although the differences in hemoglobin cannot be ex-plained by differences in iron status, the significance ofthese findings is unclear. It should also be noted thatadjustment for baseline serum ferritin may not have beenrequired, because the difference in serum ferritin betweentreatment groups was not seen in the larger sample ofpregnancies that had baseline values of serum ferritin (n� 683). The potential for selection bias as a result of ahigher prevalence of anemia at baseline in the MM group(16.8%) compared to the FE group (7.3%) among those lostto follow-up and/or those who had incomplete blood data isalso a concern, although there were no differences in irondeficiency based on serum ferritin.

The lack of benefit of MM supplements to improve ironstatus during pregnancy may be due to several reasons. Vita-min A may not have benefited our study population becausethe prevalence of deficiency was much lower when comparedto the study in Indonesia that demonstrated reductions inanemia by vitamin A supplementation (7,21). Similarly, thecontribution of folate and vitamin B-12 deficiencies to anemiamay also have been less than expected as suggested by thesimilar prevalence of IDA and overall anemia at recruitment.Another important issue is the potential adverse interactionbetween iron and zinc (22,23). Zinc may have reduced thebioavailability of iron and therefore overridden any benefitsthat may have been attributable to other enhancers of iron

TABLE 2

Comparison of baseline characteristics, iron status, and supplement consumption of women includedin the analysis and those excluded due to incomplete blood data1

n Complete blood data n Incomplete blood data

Maternal age, y 453 23.1 � 5.3 417 23.0 � 5.6Duration of pregnancy, wk 453 9.1 � 2.2 410 9.4 � 3.2Primiparous, % 453 32.9 420 39.7*Schooling, y 453 7.1 � 3.4 351 6.8 � 3.3Economic status2 447 0.07 � 1.1 348 0.01 � 1.0Indigenous ethnicity, % 445 30.8 351 30.5Single mother, % 446 3.6 351 2.9Height, cm 453 148.8 � 4.8 418 148.4 � 4.9Weight, kg 453 53.3 � 9.7 419 53.7 � 10.0BMI, kg/m2 453 24.0 � 4.1 417 24.6 � 4.2Serum ferritin,3 �g/L 361 11.6 � 17.5 322 13.5 � 20.9Hemoglobin, g/L 453 125.5 � 13.6 350 125.0 � 15.1Duration of supplementation, wk 453 29.4 � 2.7 420 17.4 � 11.8*Supplements consumed, n 453 167.3 � 17.8 420 96.6 � 66.0*Compliance, % 453 88.1 � 5.4 420 72.5 � 21.0*Multiple micronutrient group, % 453 50.1 420 49.4

1 Values are means � SD unless otherwise specified. * Different from complete blood data, P � 0.05.2 This index was derived using factor analysis and based on quality of housing, occupation, and household possessions.3 Median � interquartile range; log-transformed values were used for comparisons.

TABLE 3

Prevalence of anemia, iron deficiency, and iron-deficiencyanemia during pregnancy and postpartum in the MM

and FE treatment groups

Multiplemicronutrients

(n � 142)Iron only(n � 148)

% %Anemia1

Baseline 13.4 14.932 wk gestation 47.9 42.61 mo postpartum 47.2 49.3

Iron deficiency2


Iron-deficiencyanemia3


1 Hb � 110, 105, and 120 g/L at baseline, 32 wk gestation, and 1 mopostpartum, respectively.

2 Serum ferritin � 12.0 �g/L.3 Hb � 110, 105, and 120 g/L at baseline, 32 wk gestation, and 1 mo

postpartum, respectively, and serum ferritin � 12.0 �g/L.



jn.nutrition.orgD

ownloaded from


absorption such as vitamin C. Although little is knownabout the bioavailability of nutrients in multiple micronu-trient supplements, results from an intervention trial inPeru showed that although iron interfered with zinc bio-availability (24), the inclusion of 15 mg zinc along with 60mg iron and folate did not adversely affect changes inhemoglobin and iron status during pregnancy when com-pared to just iron-folate (25,26). Our supplements con-tained similar amounts of these nutrients and it thereforeseems unlikely that competition between iron and zinc inthe supplement explains the lack of impact.

Although the prevalence of anemia at recruitment wasmuch lower than that reported in other developing countries,almost half the women had iron deficiency and these ratesincreased dramatically during pregnancy despite iron supple-mentation. The high prevalence of iron deficiency at 32 wkgestation and even at 1 mo postpartum was indeed surprisingand cannot be attributed to low rates of compliance. It is notclear whether our serum ferritin values were affected by otherfactors, but controlling for levels of serum CRP, a marker forinfections, did not alter our conclusions (results not shown).These findings suggest that iron supplementation during preg-nancy is not adequate to compensate the high demands, es-pecially in the context of poor prepregnant iron status, andemphasize the need for strategies to improve prepregnantnutritional status. Of particular interest are recent findingsfrom an RCT in Tanzania (27) indicating that compared to aplacebo, daily consumption of a multiple micronutrient-forti-fied beverage during pregnancy was effective in reducing ane-mia and preventing iron deficiency, especially among thosewho had IDA at baseline. Although it is not clear whether thisintervention was more effective than iron supplements, theprevalence of iron deficiency at follow-up (8 wk) was muchlower in the intervention group compared to both of our studygroups and this approach may be more sustainable and couldbe used to improve micronutrient status of all women ofreproductive age.

In conclusion, our study indicates that multivitamin-mineral supplements during pregnancy do not improve he-matologic and iron status compared to iron only. Thesefindings are consistent with recent data from Nepal whereno additional benefits were seen in hematologic status dur-ing pregnancy by multiple micronutrient supplements com-pared to iron-folate supplements (13). Furthermore, boththe above study and our work in Mexico (12,14) failed todetect any benefits of MM supplements for birth outcomesalthough an earlier trial conducted in HIV� women inTanzania demonstrated reductions in the prevalence of lowbirth weight and prematurity (11). Without doubt, we must

proceed with caution, especially since we do not have anycompelling evidence to date to provide multivitamin-min-eral supplements in lieu of iron supplements.

LITERATURE CITED

1. Lozoff, B. & Wachs, T. D. (2001) Functional correlates of nutritionalanemias in infancy and early childhood—child development and behavior. In:Nutritional Anemias (Ramakrishnan, U., ed.), pp. 69–88. CRC Press, Boca Raton,FL.

2. Haas, J. D. & Brownlie, T. (2001) Iron deficiency and reduced workcapacity: a critical review of the research to determine a causal relationship. J.Nutr. 131(2S-2): 676S–688S.

3. Ramakrishnan, U. (2001) Functional consequences of nutritional ane-mia during pregnancy and early childhood. In: Nutritional Anemias (Ramakrish-nan, U., ed.), pp. 43–68. CRC Press, Boca Raton, FL.

4. Allen, L. & Casterline-Sabel, J. (2001) Prevalence and causes of nu-tritional anemias. In: Nutritional Anemias (Ramakrishnan, U., ed.), pp. 7–22. CRCPress, Boca Raton, FL.

5. Mason, J., Lotfi, M., Dalmiya, N., Sethuraman, K., Gillenwater, K., et al.(2001) Progress in controlling micronutrient deficiencies. MI/Tulane University/UNICEF. The Micronutrient Initiative.

6. Ekstrom, E. C. (2001) Supplementation for nutritional anemias. In:Nutritional Anemias (Ramakrishnan, U.), pp. 129–152. CRC Press, Boca Raton,FL.

7. Suharno, D., Muhilal, D., Karyadi, C. E., West, J. G., Hautvast, A. J. &West, C. E. (1993) Supplementation with vitamin A and iron for nutritionalanaemia in pregnant women in west Java, Indonesia. Lancet 342: 1325–1328.

8. Shatrugna, V., Raman, L., Uma, K., & Sujatha, T. (1997) Interactionbetween vitamin A and iron: effects of supplements in pregnancy. Int. J. Vit. Nutr.67: 145–148.

9. Allen, L. H. (1993) The Nutrition CRSP: what is marginal malnutrition,and does it affect human function? Nutr. Rev. 51: 255–267.

10. Ramakrishnan, U. & Huffman, S. (2001) Multiple micronutrient malnutri-tion—what can be done? In: Nutrition and Health in Developing Countries(Semba, R. D. & Bloem, M., eds.), pp. 365–392. Humana Press, Totawa, NJ.

11. Fawzie, W. W., Msamanga, G. I., Spiegelman, D., Urassa, E.J.N.,McGrath, N., Mwakagile, D., Antelman, G., Mbise, R., Herrera, G., Kapiga, S.,Willett, W., Hunter, D. J. & the Tanzania Vitamin & HIV Infection Trial Team (1998)Randomized trial of effects of vitamin supplements on pregnancy outcomes andT cell counts in HIV-1-infected women in Tanzania. Lancet 351: 1477–1482.

12. Christian, P., Khatry, S. K., Katz, J., Pradhan, E. K., LeClerq, S. C.,Shrestha, S. R., Adhikari, R. K., Sommer, A. & West, K. P., Jr. (2003) Effects ofalternative maternal micronutrient supplements on low birth weight in rural Nepal:double blind randomised community trial. Br. Med. J. 326: 571–579.

13. Christian, P., Shrestha, J., LeClerq, S. C., Khatry, S. K., Jiang, T., Wagner,T., Katz, J. & West, K. P. (2003) Supplementation with micronutrients inaddition to iron and folic acid does not further improve the hematologic status ofpregnant women in rural Nepal. J. Nutr. 133: 3492–3498.

14. Ramakrishnan, U., Gonzalez-Cossıo, T., Neufeld, L. M., Rivera, J. &Martorell, R. (2003) Multiple micronutrient supplements during pregnancydo not increase birth size compared to iron-only supplements: a randomizedcontrolled trial in a semi-rural community in Mexico. Am. J. Clin. Nutr. 77:720 –725.

15. National Research Council (U.S.) (1989) Subcommittee on the TenthEdition of the RDAs. Recommended Dietary Allowances. National AcademyPress, Washington, DC.

16. Behring Nephelometer 100 Analyzer, Behring Laboratories, MesserGrisheim Gmbh, Frankfurt/M., W. Germany.

17. Whicher, J. T., Ritchie, R. F., Johnson, A. M., Baudner, S., Bienvenu, J.,Blirup-Jensen, S., Carlstrom, A., Dati, F., Ward, A. M. & Svendsen, P. J. (1994)

TABLE 4

Comparison of hemoglobin and serum ferritin during pregnancy and 1 mo postpartum in the MM and FE treatment groups1

Multiple micronutrients Iron only

Unadjusted Adjusted2 Unadjusted Adjusted2

Hemoglobin, g/L (n � 176) (n � 185)32 wk gestation 105.3 (103.1; 107.5) 104.2 (102.5; 106.0)* 107.1 (104.9; 109.2) 108.1 (106.4; 109.8)*1 mo postpartum 118.1 (115.8; 120.4) 117.5 (115.3; 119.7) 119.2 (116.9; 121.4) 119.7 (117.6; 121.9)

Loge serum ferritin, �g/L (n � 142) (n � 148)32 wk gestation 1.44 (1.29; 1.60) 1.40 (1.26; 1.54) 1.42 (1.28; 1.57) 1.47 (1.33; 1.60)1 mo postpartum 2.37 (2.19; 2.55) 2.34 (2.17; 2.50) 2.49 (2.33; 2.65) 2.52 (2.36; 2.69)

1 Values are means (95% upper and lower confidence level); * P � 0.01 for comparison by treatment group.2 Adjusted for baseline serum ferritin.

RAMAKRISHNAN ET AL.902


jn.nutrition.orgD

ownloaded from


New international reference preparation for proteins in human serum (RPPHS).Clin. Chem. 40: 934–938.

18. WHO/UNICEF/UNU (World Health Organization, United Nations Chil-dren’s Fund, United Nations University). (1998) Indicators for AssessingIron Deficiency and Strategies for Its Prevention. World Health Organization,Geneva.

19. Lynch, S. & Green, R. (2001) Assessment of nutritional anemias. In: Nutri-tional Anemias (Ramakrishnan, U., ed.), pp. 23–42. CRC Press, Boca Raton, FL.

20. WHO Expert Committee. (1995) Physical Status: The Use and Inter-pretation of Anthropometry, p. 47. WHO Technical Report Series 854, WorldHealth Organization, Geneva.

21. Hernandez-Cordero, S., Rivera, J., Villalpando, S., Gonzalez-Cossio, T.,Neufeld, L. M., Ramakrishnan, U. & Martorell, R. (2001) Multiple micronutrientsupplementation during pregnancy: effect on breast milk retinol concentration atone month postpartum. FASEB J. 15(4): Abstract 505.7.

22. Solomons, N. W. & Ruz, M. (1997) Zinc and iron interaction: conceptsand perspectives in the developing world. Nutr. Res. 17: 177–185.

23. Solomons, N. W. (1986) Competitive interaction of iron and zinc in thediet: consequences for human nutrition. J. Nutr. 116: 927–929.

24. O’Brien, K. O., Zavaleta, N., Caulfield, L. E., Wen, J. & Abrams, S. A(2000) Prenatal iron supplements impair zinc absorption in pregnant Peruvianwomen. J. Nutr. 130: 2251–2255.

25. O’Brien, K. O., Zavaleta, N., Caulfield L., Yang, L. E. & Abrams, S. (1999)Influence of prenatal iron and zinc supplements on supplemental iron absorption,red blood cell iron incorporation, and iron status in pregnant Peruvian women.Am. J. Clin. Nutr. 69: 509–515.

26. Zavaleta, N., Caulfield, L. E. & Garcia, T. (2000) Changes in iron statusduring pregnancy in Peruvian women receiving prenatal iron and folic acid sup-plements with or without zinc. Am. J. Clin. Nutr. 71: 956–961.

27. Makola, D., Ash, D. M., Tatala, S. R., Latham, M. C., Ndossi, G. &Mehansho, H. (2003) A micronutrient-fortified beverage prevents iron defi-ciency, reduces anemia and improves the hemoglobin concentration of pregnantTanzanian women. J. Nutr. 133: 1339–1346.



jn.nutrition.orgD

ownloaded from


Documents

J. Nutr. 2004 Ramakrishnan 898 903