A

K

I

i

0d

Effect of micronutrient supplement on health and nutritional status ofschoolchildren: study design

Bhattiprolu Sivakumar, Ph.D.a,*, Kamasundaram Vijayaraghavan, M.B.B.S., M.Sc.b,Shahnaz Vazir, Ph.D.b, Nagalla Balakrishna, Ph.D.b, Veena Shatrugna, M.D.c,

Kramadhati Venkata Rameshwar Sarma, M.D., M.Sc.d, Krishnapillai Madhavan Nair, Ph.D.a,Namala Raghuramulu, Ph.D.e, and Kamala Krishnaswamy, M.D.f

a Division of Biophysics, National Institute of Nutrition (Indian Council of Medical Research), Hyderabad, Indiab Division of Field Studies, National Institute of Nutrition (Indian Council of Medical Research), Hyderabad, India

c Division of Maternal and Child Health, National Institute of Nutrition (Indian Council of Medical Research), Hyderabad, Indiad Division of Education and Extension, National Institute of Nutrition (Indian Council of Medical Research), Hyderabad, India

e Division of Endocrinology and Metabolism, National Institute of Nutrition (Indian Council of Medical Research), Hyderabad, Indiaf National Institute of Nutrition (Indian Council of Medical Research), Hyderabad, India

Manuscript received April 6, 2005; accepted July 23, 2005.

bstract Objective: We tested the hypothesis that supplementation with a micronutrient-fortified beverageimproves micronutrient status and physical and mental development in apparently healthy school-children.Methods: The study was carried out in middle-income students in two residential schools thatcatered to children from a semi-urban population near Hyderabad, India. Included were 869 childrenwho were 6 to 16 y of age in grades 1 to 10. Because children at each grade were distributed acrosstwo classrooms (clusters) and were homogeneous, each grade was considered to consist of amatched pair. There were thus 10 pairs available for the study. Classes in each grade wererandomized to receive a micronutrient-enriched beverage or a placebo without added micronutrients.The study was double blinded and the duration was 14 mo, with supervised feeding of themicronutrient-enriched beverage. The effect of the micronutrients on the outcome variables growth,biochemical status, mental function, and bone health were assessed.Results: The number of matched pairs varied between seven and eight, and the required number ofchildren per treatment group ranged from 32 in the case of bone heath to 177 for body weight. Thepower of the outcome variables ranged from 74% to 100% and was adequate for successfulpairing. The effect assessed at the end of supplementation showed that the intervention wasbeneficial.Conclusions: Designing an intervention, choosing outcome variables, and implementing the pro-tocol in a typical Indian school setting were achieved. © 2006 Elsevier Inc. All rights reserved.

eywords: Matched-pair; Cluster; Randomized; Study design; Multiple micronutrients; Schoolchildren; Growth; Biochem-

Nutrition 22 (2006) S1–S7www.elsevier.com/locate/nut

ical status; Bone health; Mental function

[B[ocad

ntroduction

Micronutrient malnutrition (particularly of vitamin A,ron, and iodine) is a major public health problem in India

This study was supported by M/S GlaxoSmithKline, New Delhi, India.* Corresponding author. Tel.: 91-40-2701-8083.

cE-mail address: dirnin_hyd@yahoo.co.in (B. Sivakumar).

899-9007/06/$ – see front matter © 2006 Elsevier Inc. All rights reserved.oi:10.1016/j.nut.2005.07.009

1,2]. Deficiencies of other micronutrients such as zinc,-complex vitamins, and calcium have also been reported

3,4]. The deleterious effect of micronutrient malnutritionn growth and mental development is more severe in younghildren than in older children [4,5]. It is recognized that thedverse effects of micronutrient deficiencies on growth andevelopment, in particular mental function, physical work, and

ompromised immune status, warrant immediate attention

[m

g[[mmtesacdosv(bmqmt2itio

sisscmfemTg

S

B

dowaibtsd

woaitiosc

M

rwdmfiocg

S

iowvbufifr

S

dcsckcdscdhcpas

S2 B. Sivakumar et al. / Nutrition 22 (2006) S1–S7

3–7]. Food fortification is one strategy that has been recom-ended and adopted to control micronutrient malnutrition.Many studies have shown that micronutrients given sin-

ly or in combination have a positive influence on growth4,8,9], psychomotor coordination [10], mental function8,11], and bone health [10,12,13] in children. Studies withultiple micronutrient deficiencies have shown improve-ent in linear growth, psychomotor functions, and correc-

ion of biochemical abnormality when more than one nutri-nt was supplemented [8,10,13,14]. Supplementation ofchool-age Indian children with micronutrients such as ironnd B-complex vitamins has been reported to improve thehildren’s visual acuity, psychomotor function, and aca-emic achievement [15–17]. Schoenthaler et al. [18] dem-nstrated significant improvement in the mental function ofchoolchildren who were supplemented with minerals anditamins at 100% of the recommended dietary allowanceRDA) but not at 50% (or 200% level of codex RDA). Theeneficial response in physical and mental development toultiple micronutrients was noted only when energy re-

uirements were fulfilled [10,12,19]. Conditional improve-ent in function has been shown in some studies in which

he populations had a poor nutritional status initially [5,20–2]. The effect of micronutrient supplementation, singly orn combination, has thus been shown to be a variable, withhe observed effect being more pronounced when there is annitial deficit in growth (stunting) or function or deficiencyf some nutrient [10,20–22].

We therefore tested the hypothesis that “micronutrientupplementation can bring about an improvement in phys-cal and mental development in middle-income Indianchool children in a situation where there was apparently nohortage of food, nor was there a clinical evidence of defi-iency.” We designed a double-blind, placebo-controlled,atched-pair, cluster-randomized study to evaluate the ef-

ect of a beverage supplemented with multiple micronutri-nts on 1) micronutrient status, 2) growth, 3) morbidity, 4)ental function, and 5) bone health of school-age children.his report presents the details of the common methodolo-ies and the statistical analyses used.

tudy design

asis of study design

Studies aimed at assessing the effect of an interventionepend on appropriateness of the study design. The methodf choice for such studies is individual randomization,hich avoids systematic confounding variables of known

nd unknown nature. Further, to avoid any bias and toncrease the precision of the outcome variables, doublelinding is essential (i.e., neither the subject nor the inves-igator assessing the response is aware of the treatment theubject is receiving). Such a design, however, is extremely

ifficult to conduct in a large school situation because it t

ould be difficult to organize and monitor the distributionf beverage to each child every day. Therefore we adoptedcluster randomization, rather than an individual random-

zation, study design wherein clusters of individuals, ratherhan independent individuals, are randomly allocated to thentervention groups. This design has the added advantagesf avoiding treatment group contamination and enhancingubject compliance and is administration-friendly due to aluster-level intervention.

atched pair placebo control

The study population was a captive group consisting ofesidential schoolchildren who were 6 to 16 y of age andere stratified into grades 1 to 10. Each grade had childrenistributed across two classrooms (clusters) who wereatched with each other for their age. Each classroom thus

ormed a matched-pair cluster. The matched-paired designs a special case of the stratified design in which there arenly two clusters per stratum, as in this study, and oneluster is allocated to the treatment or the placebo controlroup in each grade.

ample size

The number of pairs and the required number of childrenn each treatment group to assess the effect of variousutcome measurements, such as increments in height andeight for growth, intelligence quotient for mental function,itamin A for effect on biochemical status, and whole-bodyone mineral content for bone health, were calculated bysing the formula of Donner and Klar [23]. The criteriaxed were a confidence interval of 95%, power of 80% andollow-up losses of 20% and 10% for the treatment and pair,espectively, for calculating the sample size (Table 1).

election of schools

The children of the residential schools that cater to mid-le-income, semi-urban populations near Hyderabad, India,omprised the subjects of this study. They were chosen asubjects because their dietary intakes were uniform andomparable. Food in these schools is prepared in a commonitchen and, from this point of view, would ensure betterompliance when supplements are made available, evenuring intermittent short holidays. Six schools were initiallyelected, and only two residential schools met the inclusionriteria of having a large number of male and female chil-ren. Four did not meet the inclusion criteria because theyad small numbers of children, taught only male or femalehildren, or were homes for orphans who are generallyresumed to consist of subjects who are likely to be far fromnormal state of mental happiness. Although most of the

ample was eventually derived from one residential school,

he second one was chosen mainly to increase the number of

gp

E

aoecpbt(op

R

agheea

C

tcpa

C

mca2dtrvn(

P

rtmwtwrtsmv

A

bT

TCm

V

HWIHVLT

o

do

�

S3B. Sivakumar et al. / Nutrition 22 (2006) S1–S7

irls. The subjects from the two selected schools had com-arable mean values for age, height, and weight.

nrollment of children

Written informed consent from the parent or guardiannd the school principal was obtained, and the willingnessf children to participate until completion of the study wasnsured by explaining to the children in detail all the pro-edures and methods involved in the study at the outset. Allrocedures followed were in accordance with and approvedy the institutional ethics committee of the National Insti-ute of Nutrition (Hyderabad, India). Although all childrenn � 930) in the school received the beverage supplements,nly those earmarked for the study (702 boys and 167 girls)articipated in growth and mental function evaluation.

andomization

In both schools all grades (1 to 10) had two classroomsnd the children were matched by age. Each class within arade was allocated randomly to a micronutrient-enrichedealth drink (supplement) or a placebo with a comparablenergy/macronutrient profile but without added micronutri-nts (Table 2). Each grade was considered a matched pairnd helped to avoid contamination (n � 9–10 pairs).

oding and decoding of the supplement

Beverages coded A and B, which were identical in color,exture, and taste, were served to children in the designatedlasses. The beverages were decoded (A, supplement; B,lacebo) after completion of data collection and preliminary

able 1alculation of sample size per treatment group and number of matched picronutrient supplementation on various outcome measurements

ariable (increment) Variance components* �

Between Within o2 w

eight (cm) 0.38 3.045 0.11eight (kg) 0.31 3.883 0.07

ntelligence quotient 3.43 79.328 0.04emoglobin (g/L) 68.72 311.144 0.22itamin A (�mol/L) 0.10 0.109 0.48umbar spine BMD 0.00018 0.001 2.61otal body BMC 823.69 6648.427 2.18

BMC, bone mineral content; BMD, bone mineral density* In the present study, the calculation of sample size was based on the for

f 80%, and follow-up losses of 20% (treatment) and 10% (pair).† Sample size per treatment {[(Zl��/2 � Z1��)2 o2/d2][1 � (m � 1)�]},

ifference, � � intracluster correlation coefficient, m � number of subjectf 38 for anthropometry).

‡ Sample size for pair � [(Z1�� � Z1��)2 Var(d)]/d2, where Var(d) � 2[component of variability between two clusters in a matched pair.

nalysis. o

omposition of health drink

The composition of the beverage (placebo and supple-ent) is presented in Table 3, and the base formulation

ontained wheat flour, malted barley, powdered skim milk,nd sugar. The supplement (54 g) provided 7.3 g of protein,g of fat, 208 kcal of energy, and 176 mg of calcium given

aily in two equal servings. The supplement, in addition tohe base formulation, provided 100% of codex RDA of iron,iboflavin (B2), pyridoxine (B6), vitamin B12, folate, anditamin C; 50% of the RDA for vitamins A, D, and B1,iacin, calcium, and iodine; and 15% of the RDA for zincTable 3).

reparation of health drink

The consistency and volume of the beverage that couldeasonably be consumed in one sitting were determined inrials before the start of study, and 27 g in 150 mL of tonedilk (milk containing 3.5% fat obtained from a local dairy)as found to be the most acceptable. Serving the beverage

o each child through sachets containing the portion sizeas found to be unsuitable in the schools. Therefore the

equired quantity of the beverage was prepared in bulk inhe kitchen of the residential schools just before serving. Amooth paste of the formulation in hot milk was prepared byanually mixing and then adding milk to the required

olume of the beverage.

dministration of health drink

Each day, 150 mL of the beverage was served withreakfast in the morning and before playtime in the evening.he two drinks were served separately, one after the other,

uired for a matched-pair cluster design study to assess the effect of

Design effect[1�(m �1) �]

Expecteddifference (d)

Sample size

Treatment† Pair‡

5.14 1.0 165 73.63 0.9 177 82.23 3.5 142 83.57 13.5 70 86.62 0.5 52 72.61 0.03 32 72.18 51.0 59 8

f Donner and Klar [23], and assuming a confidence interval of 95%, power

Z1��/2 � 1.96, Z1�� � 0.84, o2 � variance of increment, d � expecteduster (depends on variable and a minimum of 11 for BMD and maximum

� o2 AM], where o2w � within-cluster component of variability and o2AM

airs req

mula o

wheres in a cl

o2 w/m

r served in two specific places that had been allocated for

traaztg

C

opmpadc

b

tfttbstloGdest

D

1m

TRb

1234567891TP

TC

B(

N

PFEC

m

S4 B. Sivakumar et al. / Nutrition 22 (2006) S1–S7

he drinks and identified for the children. The teachers of theespective classes and the warden participated in the bever-ge distribution to ensure that the children received thessigned beverage. During the course of the study, albenda-ole (GlaxoSmithKline, Gurgaon, India) was administeredwice, at 6-mo intervals, to deworm the children in bothroups.

ompliance of beverage intake

The drinks were served to children under the supervisionf the field investigators who were responsible for the sup-lement distribution to ensure compliance with the require-ents of the study. If any child failed to turn up for sup-

lementation, that child was contacted and the reason forbsence was recorded. If the child was sick, the nature anduration of the sickness and treatment received were re-orded.

During the intermittent short holidays, there was noreak in the consumption of the beverages because most of

able 2andomization and classroom-wise number of children studied in each griochemical evaluation, and bone health

Grade Randomization Anthropometry Mental fu

Class room Baseline Final Baseline

1 2 S P S P S P

S P 18 27 16 25 10P S 41 60 37 51 35S P 41 39 35 31 40S P 52 47 38 37 49P S 64 54 54 47 62S P 56 43 48 39 55P S 56 56 50 46 56P S 55 42 48 36 54S P 31 28 29 28 31

0 P S 32 27 NA NA 29otal 446 423 355 340 421 3air 10 10 9 9 10

NA, not available; P, placebo group; S, supplement group

able 3omposition of nutrients in supplementary beverages given to children

asal formulation*placebo)

Basal formulation (placebo) with micr

utrient Amount Nutrient CodexRDA (100%)

rotein 7.3 g Iron 14 mgat 2.0 g Vitamin B2 1.6 mgnergy 208 kcal Vitamin B6 2 mgalcium 176 mg Vitamin B12 1 �g

Folic acid 200 �gVitamin C 80 mg

* The beverage was served twice daily as a 27-g serving with 150 mL of

alted barley, skimmed milk powder, and sugar.

he children remained in the school. When children pre-erred to go home for a long-term break (May to June 2000),hey were supplied with an appropriate number of sachets ofhe fortified product or placebo for consumption at homeefore they left. Children were given instructions to con-ume the drink twice daily, as was done during the schoolerm. Children were also given printed cards (in English, theocal language of Telugu, and Hindi) to mark the regularityf consumption of the drink on all days they spent at home.uidelines to fill the record were also provided. The chil-ren were also advised to take the sachets when they trav-led away from home. The children were instructed toubmit the completed records to the principal on their returno the school.

uration of supplementation

The study took place over a 14-mo period, with at least2 mo of direct supervised supplementation. During the 2o of summer vacation, care was taken to ensure a supply

aseline and end of study for anthropometry, mental function,

Biochemical evaluation Bone health

inal Baseline Final Baseline Final

P S P S P S P S P

9 15 4 6 4 5 NA NA NA NA32 39 12 37 12 31 5 20 4 1933 23 22 20 18 17 21 2 17 235 34 21 14 15 10 20 10 14 653 45 15 17 11 15 15 9 10 645 38 11 11 8 9 10 10 7 946 43 11 15 9 13 11 16 10 1443 32 20 16 19 14 21 17 20 1525 18 16 20 14 19 17 20 12 19A NA 23 17 NA NA 26 18 NA NA

21 287 155 173 110 133 146 122 94 909 9 10 10 9 9 9 9 8 8

nt (supplement)

Nutrient CodexRDA (50%)

Nutrient CodexRDA (15%)

Vitamin A 400 �g Zinc 2.3 mgVitamin D 2.5 �gVitamin B1 0.7 mgNiacin 0.9 mgCalcium 400 mgIodine 75 �g

ilk for 14 mo. The base formulation in both drinks contained wheat flour,

ade at b

nction

F

S

18503546524355422824 N93 310

onutrie

toned m

owt[

D

amc(2dmclga

A

h

S

Idtmu

Pi

cvioorww[swm(

pt

A

tfvdSie

P

Vcaap

R

mbaacct

E

sertA0n

P

trfoq

a

S5B. Sivakumar et al. / Nutrition 22 (2006) S1–S7

f supplements for consumption at homes, and complianceas assessed by measurement of urinary riboflavin concen-

ration, which indicates recent intake of the supplement24].

ata collection

Before starting the supplements in August 1999, data onnthropometry (height and weight), clinical deficiency,orbidity, mental function, and biochemical status were

ollected at baseline (March and July 1999) and after 4 moNovember to December 1999), 8 mo (February to March000), and 14 mo (September to October 2000). Children’sietary intake was assessed by the institutional diet surveyethod. Individually weighed dietary intake was recorded

onsecutively for 3 d in a 10% subsample [25]. The fol-ow-up of children studied was 80% in each treatmentroup. A cohort of all students who participated at baselinend final follow-up was considered for data analysis.

nalytical methods

Detailed methodologies of various outcome variablesave been described in separate reports [25–28].

tatistical analysis

Data were analyzed with SPSS 11.5 for Windows (SPSSnc., Chicago, IL, USA). The means of all pairs and stan-ard deviations for all variables at baseline and at the end ofhe study were calculated at the classroom level. In addition,ean increments (final value minus initial value at individ-

al level) of cohort data of nine pairs were calculated.

aired t test for comparison of means and meanncrements

The cluster means of all outcome measurements wereonsidered as the raw data for comparison of the meanalues of the supplement and placebo groups for assessingnitial differences by paired t test. Similarly, for assessmentf the effect of the micronutrient supplement, increments inutcome variables were tested. Even though paired t test isobust enough to test the significance between group means,ith substantial variation in cluster size from pair to pair,eighted-paired t test is the recommended method

23,29,30]. In the present study, because there were sub-tantial variations in cluster size within and between pairs,eighted-paired t test was used to assess the effect oficronutrient supplementation on all outcome variables

Table 2). o

Wilcoxon’s signed rank test was used for comparison ofroportions between groups. P � 0.05 was considered sta-istically significant.

djusting for covariates.Whenever baseline variables were different between

reatment groups, a regression model was applied to adjustor initial differences. The baseline value of the outcomeariables was considered a covariate and the increments asependent variables for adjusting for initial differences.imilar analysis was carried out for those outcome variables

n which the mean differences between treatment groupsxisted but were not statistically significant [23].

ower of outcome variables.Pepi package (Programs for Epidemiological Analysis,

ersion 3, Abramson JH and Gahlinger PM) was used toompute the power of a test to compare two means bypplying the observed difference between the supplementnd the placebo groups, their correlation, and the number ofairs.

esults and discussion

The study was a double-blind, placebo-controlled,atched-pair, cluster-randomized trial. The calculated num-

er of pairs and sample size in each treatment group and thatvailable at the end of the study were adequate (Tables 1nd 2). There were two classrooms per grade, which wereonsidered as paired clusters. Hence, there were 10 pairs oflusters at baseline and eight to nine clusters (n � 8–9) athe end of the study (Table 2).

ffectiveness of matching

Matching was assessed by verifying the standard Pear-on’s correlation between paired clusters for a given param-ter at baseline. For pairs of 10 or fewer, the correlationequired is at least 0.20. In this study, the correlation ob-ained for age, weight, height, intelligence quotient, vitamin, and whole-body bone mineral content were 0.99, 0.99,.99, 0.30, �0.56, and 0.97 respectively, suggesting that theumber of pairs required was 10 or fewer.

ower of outcome variables

The power of significant outcome variables obtained inhe study is presented in Table 4. The calculated poweranged from 0.74 for height-for-age z score to a high of 1.00or folate. This suggests that the differences observed in theutcome variables listed in Table 4 were not due to inade-uate sample size.

The sample size within a cluster was also found to bedequate for bone health measurements, except in the case

f children in grade 3 where there were only two children

(hmgstga

C

amrd2

E

ahwmvihsKptWmxp

C

mspoist

A

BSSUtC

R

[

[

[

TP

O

W����������D�

s

S6 B. Sivakumar et al. / Nutrition 22 (2006) S1–S7

Table 2). All pairs in the supplement and placebo groupsad comparable age (within a cluster) and school environ-ent at baseline [26]. Baseline functional parameters of

rowth, mental function, bone status, and micronutrienttatus were comparable between treatment groups, except inhe case of inadequacy of vitamin D (30.3% in supplementroup versus 21.6% in placebo group) and whole-body bonerea [25–28].

ompliance of beverage intake

The daily intake of supplement and placebo beverage forperiod of 14 mo was ensured during the study periodainly by supervised feeding. There was no instance of

efusal of the drink on any occasion during the study. Theropout rate was similar in both groups and varied from0% to 35% in various functional measurements.

ffect of micronutrient beverage on outcome variables

The detailed results on various outcome measurementsre reported elsewhere [25–28] and are briefly presentedere. Widespread prevalence of micronutrient inadequacyas evident in these schoolchildren. Micronutrient supple-entation significantly decreased inadequacies in folate,

itamins A, D, B2, and B12 [26]. There was a significantncrease in weight velocity and increments of z score ofeight and weight and shorter duration of illness in theupplement group compared with the placebo group [25].nox’s cube test score was significantly higher in the sup-lement group, thus suggesting better attention concentra-ion in this group compared with the placebo group [27].

hole-body bone area and bone density at the neck, fat-freeass, and percentage of fat, as assessed by dual-energy

-ray absorptiometry, was significantly higher in the sup-

able 4ower and P value of outcome variables based on nine pairs of sample

utcome variable P Power (1 � �)

eight velocity 0.017 0.944Height-for-age z score 0.016 0.739Weight-for-age z score 0.002 0.931RBC folate 0.000 1.000Vitamin A 0.002 0.957Vitamin D 0.000 0.991TSH 0.001 0.974Vitamin B2 0.017 0.861Whole-body area* 0.013 0.802Fat-free mass* 0.003 0.948Percentage of fat* 0.035 0.886uration of illness 0.036 0.875Attention concentration 0.016 0.834

�, increment (final minus baseline); RBC, red blood cell; TSH, thyroid-timulating hormone

* Based on eight pairs.

lement group than in the placebo group [28].

onclusions

Thus, by using a double-blind, placebo-controlled,atched-pair, cluster-randomized study, the hypothesis that

upplementation with a micronutrient-fortified beverage im-roves micronutrient status and physical and mental devel-pment in schoolchildren was proved. The study has policymplications in terms of providing micronutrients in theupplementary feeding programs currently in operation inhe country.

cknowledgments

The authors are grateful for expert advice from P.haskaram, M.D., A. Nadamuni Naidu, M.Sc., and B.esikeran, M.D., of the National Institute of Nutrition;udhakara Rao, M.D., of Henry Ford Hospital (Detroit, MI,SA); and Saroj Arya, Ph.D., of the National Institute for

he Mentally Handicapped (Hyderabad, India); and K. Ravihandra, Ph.D., of Osmania University (Hyderabad, India).

eferences

[1] A national nutrition policy. New Delhi: Government of India, De-partment of Women and Child Welfare, Ministry of Human ResourceDevelopment; 1993.

[2] Report of an ICMR Task Force study—District Nutrition Project,micronutrient deficiency disorders in 16 districts of India. Part 1. NewDelhi: Indian Council of Medical Research; 2001.

[3] Bhan MK, Sommerfelt H, Strand T. Micronutrient deficiency inchildren. Br J Nutr 2001;85:S199–203.

[4] Bhandari N. Bahl R, Taneja S. Effect of micronutrient supplementa-tion on linear growth of children. Br J Nutr 2001;85:S131–7.

[5] Kannani SJ, Poojara RH. Supplementation with iron and folic acidenhances growth in adolescent Indian girls. J Nutr 2000;130:S452–5.

[6] Grantham-McGregor SM, Fernald LC. Nutritional deficiencies andsubsequent effects on mental and behavioral development in children.Southeast Asian J Trop Med Public Health 1997;28:50–68.

[7] Wasantwisut E. Nutrition and development: other micronutrients’effect on growth and cognition. Southeast Asian J Trop Med PublicHealth 1997;28:78–82.

[8] Sandstead HH, Penland JG, Alcock NW, Dayal HH, Chen XC, Li JS,et al. Effects of repletion with zinc and other micronutrients onneuropsychologic performance and growth of Chinese children. Am JClin Nutr 1998;68:S470–5.

[9] Harahap H, Jahari AB, Husaini MA, Saco-Pollitt C, Pollitt E. Effectsof an energy and micronutrient supplement on iron deficiency anemia,physical activity and motor and mental development in undernour-ished children in Indonesia. Eu J Clin Nutr 2000;54:S114–9.

10] Bates CJ, Evans PH, Allison G, Sonko BJ, Hoare S, Goodrich S, et al.Biochemical indices and neuromuscular function tests in rural Gam-bian school children given a riboflavin, or multivitamin plus ironsupplement. Br J Nutr 1994;72:601–10.

11] Van Stuijverberg ME, Kvalsvig JD, Faber M, Kruger M, KenoyerDG, Benade AJ. Effect of iron-, iodine-, and beta-carotene–fortifiedbiscuits on the micronutrient status of primary school children: arandomized controlled trial. Am J Clin Nutr 1999;69:497–503.

12] Jahari AB, Haas J, Husaini MA, Pollitt E. Effects of an energy andmicronutrient supplement on skeletal maturation in undernourished

children in Indonesia. Eu J Clin Nutr 2000;54:S74–9.

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

S7B. Sivakumar et al. / Nutrition 22 (2006) S1–S7

13] Rosado JL. Separate and joint effects of micronutrient deficiencies onlinear growth. J Nutr 1999;129:S531–3.

14] Mwanri L, Worsley A, Ryan P, Masika J. Supplemental vitamin Aimproves anemia and growth in anemic school children in Tanzania.J Nutr 2000;130:2691–6.

15] Agarwal, DK, Upadhyay SK, Tripathi AM, Agarwal KN. Nutritionalstatus, physical work capacity and mental function in school children.Nutr Found India 1987;6:40.

16] Damodaran M, Sarma KVR, Tiar A, Naidu AN. Vitamin B-complexdeficiency and visual acuity. Br J Nutr 1979;41:27–30.

17] Bamji MS, Saroj A, Sarma KVR, Radhaiah G. Impact of long termlow dose B-complex vitamin supplements on vitamin status andpsychomotor performance of rural school boys. Nutr Res 1982;2:147–53.

18] Schoenthaler SJ, Amos SP, Doraz WE, Kelly MA, Wakefield J.Controlled trial of vitamin-mineral supplementation: Effect on intel-ligence and performance. Person Ind Differ 1991;12:351.

19] Aitchison TC, Durnin JV, Beckett C, Pollitt E. Effects of an energyand micronutrient supplement on growth and activity, correcting fornon-supplemental sources of energy input in undernourished childrenin Indonesia. Eu J Clin Nutr 2000;54:S69–73.

20] Perrone L, Salerno M, Gialanella G, Feng SL, Moro R, Di Lascio R,et al. Long-term zinc and iron supplementation in children of shortstature: Effect of growth and on trace element content in tissues. JTrace Elem Med Biol 1999;13:51–6.

21] Thu BD, Schultink W, Dillon D, Gross R, Leswara ND, Khoi HH.Effect of daily and weekly micronutrient supplementation on micro-nutrient deficiencies and growth in young Vietnamese children. Am J

Clin Nutr 1999;69:80–6.

22] Powers HJ, Bates CJ, Lamb WH, Singh J, Gelman W, Webb E, et al.Effects of a multivitamin and iron supplement on running perfor-mance in Gambian children. Hum Nutr Clin Nutr 1985;39:427–37.

23] Donner A, Klar N. Design and analysis of cluster randomization trialsin health research. New York: Oxford University Press; 2000.

24] Morell DB, Slater EC. The fluorometric determination of riboflavin inurine. Biochem J 1946;40:652–7.

25] Sarma KVR, Udaykumar P, Balakrishna N, Vijayaraghavan K, Siva-kumar B. Effect of micronutrient supplement on health and nutritionstatus of school children: growth and morbidity. Nutrition 2006;22:S8–14.

26] Sivakumar B, Madhavan Nair K, Sreeramulu D, Suryanarayana P,Ravinder P, Shatrugna V, et al. Effect of micronutrient supplement onhealth and nutritional status of schoolchildren: biochemical status.Nutrition 2006:22:S15–25.

27] Shahnaz V, Balakrishna N, Vijayapushpam T, Vijayaraghavan K,Sivakumar B. Effects of micronutrient supplement on health andnutritional status of schoolchildren: mental function. Nutrition 2006;22:S26–32.

28] Shatrugna V, Balakrishna N, Kamala K. Effect of micronutrientsupplement on health and nutritional status of schoolchildren: bonehealth and body composition. Nutrition 2006;22:S33–39.

29] Wears RL. Advanced statistics: statistical methods for analyzingcluster and cluster-randomized data. Acad Emerg Med 2002;9:330–41.

30] Ukoumunne OC, Guilliford M, Chinn S, Sterne JAC, Burney PGJ.Methods for evaluating area-wide and organization based interven-tions in health and health care: a systematic review. Health Technol

Assess 1999;3:1–110.