点击显示 收起
1 From the Program in International Nutrition and Community Nutrition, University of California, Davis, CA (SA-A, KHB, and KGD); the Department of Nutrition and Food Science, University of Ghana, Legon, Ghana (AL); the Department of Paediatrics, Nutritional Sciences and Public Health Sciences, University of Toronto, Toronto, Canada (SZ); and IRD, Département Sociétés et Santé, Paris, France (AB)
2 The opinions expressed herein are those of the authors and do not necessarily represent the views of the Institute of the International Life Sciences Institute (ILSI) or the US Agency for International Development (USAID). 3 Supported by the Nestlé Foundation with additional support from USAID's MGL Research Program through ILSI. 4 Reprints not available. Address correspondence to KG Dewey, Department of Nutrition, University of California, One Shields Avenue, Davis, CA 95616-8669. E-mail: kgdewey{at}ucdavis.edu.
ABSTRACT
Background: The low micronutrient content of complementary foods is associated with growth faltering in many populations. A potential low-cost solution is the home fortification of complementary foods with Sprinkles (SP) powder, crushable Nutritabs (NT) tablets, or energy-dense (108 kcal/d), fat-based Nutributter (NB).
Objective: The objective was to test the hypothesis that multiple micronutrients added to home-prepared complementary foods would increase growth and that the effect would be greatest in the presence of added energy from fat.
Design: We randomly assigned 313 Ghanaian infants to receive SP, NT, or NB containing 6, 16, and 19 vitamins and minerals, respectively, daily from 6 to 12 mo of age. We assessed anthropometric status at 6, 9, and 12 mo; micronutrient status at 6 and 12 mo; motor development at 12 mo; and morbidity weekly. Infants (n = 96) not randomly selected for the intervention (nonintervention; NI) were assessed at 12 mo.
Results: The groups did not differ significantly at baseline, except that the NB group had a higher proportion of boys and weighed slightly more. The dropout rate (15/313) was low. At 12 mo, after control for initial size, the NB group had a significantly greater weight-for-age z score (WAZ) (–0.49 ± 0.54) and length-for-age z score (LAZ) (–0.20 ± 0.54) than did the NT group (WAZ: –0.67 ± 0.54; LAZ: –0.39 ± 0.54) and the NT and SP groups combined (WAZ: –0.65 ± 0.54; LAZ: –0.38 ± 0.54); the difference with the NI group (WAZ: –0.74 ± 1.1; LAZ: –0.40 ± 1.0) was not significant. A lower percentage of the NI infants (25%) than of the intervention groups (SP: 39%; NT: 36%; NB: 49%) could walk independently by 12 mo.
Conclusion: All 3 supplements had positive effects on motor milestone acquisition by 12 mo compared with no intervention, but only NB affected growth.
Key Words: Multiple micronutrient supplements • home fortification • infant growth • motor development • Ghana
INTRODUCTION
The low micronutrient content of complementary foods in many disadvantaged populations has been associated with growth faltering (1), increased morbidity (2), and delayed motor milestone acquisition (3). A low-cost strategy to deal with this problem is home fortification of complementary foods with multiple micronutrient supplements. This practice could avoid duplication of effort with single nutrient supplementation programs and increase the usually low compliance associated with tablets and syrups taken between meals. We identified 3 types of multiple micronutrient supplements for this purpose: 1) Sprinkles (SP), a powder developed by Ped Med Inc, Canada (4); 2) foodlet, a crushable tablet developed by UNICEF (5); and 3) a peanut-based fortified spread developed by Nutriset SA, Malaunay, France (6). SP contained iron and zinc, nutrients that are deficient in many developing countries (7), folic acid, and vitamins A and C. We modified the foodlet (referred to herein as Nutritabs; NT) to include a larger set of micronutrients than in SP. We modified the fortified spread (referred to herein as Nutributter; NB) to include a still larger set of micronutrients plus added energy (108 kcal/d), mainly from fat (1.29 g linoleic/d and 0.29 g -linolenic/d).
When we began this study, SP had been shown to be efficacious in the treatment (8) of anemia, but there were no published data on the foodlet, and the fortified spread (in larger quantities) had been used only to treat severely malnourished children and not as a complementary food supplement. Recent studies on SP in Cambodia (9) and the foodlet in South Africa (10) have provided evidence of the positive effect of both supplements on iron status when mixed with home-prepared complementary foods. In Malawi (11), a version of fortified spread was found to increase weight, length, and hemoglobin concentrations in moderately malnourished infants. To date, no research has directly compared all 3 products. Our primary aim was to compare the acceptability and the effects of these 3 supplements on the growth and micronutrient status of infants from 6 to 12 mo of age. We also assessed morbidity and food intake and observed motor milestone acquisition at 12 mo. Our purpose was to determine whether providing a larger set of micronutrients (NT) would have any advantages over SP and whether the NB with added macronutrients would have any advantages over either of the other 2 supplements. We hypothesized that home fortification of complementary foods with micronutrient supplements would increase growth and that the effect would be greatest in the group given the supplement that included added energy from fat. Ghana was an appropriate site for the study because micronutrient deficiencies are common in infants there (12), and indications (13) are that the intake of essential fatty acids at this age may be low. In this study, we present data on growth, food intake, morbidity, and motor milestone acquisition.
SUBJECTS AND METHODS
Study area and participants
We carried out the study in Koforidua, in the Eastern Region of Ghana, between February 2004 and June 2005. The town has a population of 87 000; annual rainfall averages 2030 mm. The principal complementary food in most households is a fermented maize-based porridge. All infants attending weight-monitoring sessions in Koforidua between February and September 2004 were potentially eligible. Infants were eligible for inclusion if they were 1) 5 mo of age, 2) receiving any breast milk, 3) not known to be asthmatic or allergic to peanuts, and 4) planning to stay at the study site during the next 7 mo.
The study protocol was approved by the Human Subjects Review Committee of the University of California, Davis, and the Institutional Review Board of the Noguchi Memorial Institute for Medical Research, University of Ghana. Written permission to carry out the study in Koforidua was obtained from the Eastern Regional Health Administration, and written informed consent was obtained from the parents of each infant.
Study design
The study was designed as a community-based randomized trial involving 3 intervention groups and one nonintervention (NI) group. Intervention infants were recruited weekly at 5 mo of age. For logistical reasons, we could not include all infants who were eligible each week in the intervention groups. Therefore, during each week, we randomly selected 75% of eligible infants by entering identification numbers for all eligible infants in a dataset and using a SAS data step [ranuni (1) le 0.75] to select infants who would enter the intervention trial. After consent was obtained, we began collecting weekly morbidity data.
At 6 mo, the infants were randomly assigned (with the use of opaque envelopes with group designations) to receive SP, NT, or NB until 12 mo of age. There was no placebo group because of ethical concerns about providing no treatment to infants who may have tested low on certain assessments (eg, hemoglobin) at baseline. As an alternative, we included the NI group for assessment only at 12 mo of age. The NI infants were enrolled from among the 25% of those who were originally eligible but not randomly selected for the intervention groups (n = 170).
The primary outcome was growth, but we also assessed morbidity and observed motor milestone acquisition at 12 mo. In addition, we assessed energy intake from complementary foods, given that the energy content of NB could significantly affect the energy intake of infants in that group. Sample size was based on the detection of differences among 4 groups equivalent to a "medium" effect size [Cohen's d = (difference/pooled SD) = 0.5] (14). With a type I error of 0.05 and a 0.8 probability of detecting a true difference (1 – ß), the required sample size per group was 87. Allowing for 15% attrition in the 3 intervention groups, the target sample size for each of those groups was 102. At 12 mo, to achieve the target sample size of 87 for the NI group, we randomly selected 130 of the 170 eligible children, assuming that some of the parents would have moved away or declined to participate in the study. Because the NI group was recruited at 12 mo of age, it was not possible to collect baseline data at 6 mo. However, because the NI infants were randomly selected from the pool of initially eligible infants, they should have been similar to the other 3 groups initially.
Micronutrient supplements
SP was manufactured by Ped-Med Ltd, Toronto, Canada in single sachets (dose = 1 sachet/d). NT (1 tablet/d) was manufactured by Laboratoires Pharmaceutiques Rodael SA (Bierne, France) and supplied by Nutriset SA, which manufactured NB. NT was provided to the mothers in plastic bags, and the NB (20 g/d) was provided in foil packs with screw caps (net weight = 200 g). The nutrient composition of the 3 supplements is reported in Table 1. NT and NB were designed so that the daily dose would generally provide the needed amounts from complementary foods for all of the key nutrients (7, 15). Because of technical difficulties associated with adding the desired amounts of the 4 macrominerals (calcium, potassium, magnesium, and phosphorous), we included as much calcium, potassium, and phosphorous as possible in both supplements and maintained the amount of magnesium and manganese already contributed by the ingredients in NB. Because of logistic problems at the start of the study, we had to use SP from a batch that was available at the time, which had slightly higher contents of iron and zinc (and different chemical composition of these 2 minerals) than did the NT and NB. Similar types of SP have been used in Cambodia (9), and the doses were consistent with recommended daily intakes.
View this table:
TABLE 1. Nutrient composition of the micronutrient supplements used in the study
Procedures
Mothers (parents) of selected infants were visited in the home to verify eligibility, explain the study protocol in detail, and obtain written informed consent. We delivered all supplements to the intervention children on a weekly basis and instructed the mothers to administer the daily dose in a single meal, 7 d/wk. To ensure that children consumed the entire dose, we told mothers to mix the supplement with 1–2 tablespoons (15–30 mL) of the child's food. We supplied mothers with cups and spoons, which helped those with infants in the NB group to measure the recommended dose.
We collected background data at the time of recruitment. For the intervention children, data on morbidity (diarrhea, symptoms of respiratory infections, and fever) and daily supplement consumption, including leftovers, were collected weekly. Each month, we performed a 24-h dietary recall and calculated the energy intake from complementary foods by multiplying the grams of each food consumed by the energy content (kcal/100 g) of the food, which was obtained mainly from a food-composition database for Ghana (16) and, to a lesser extent, the WorldFood Dietary Assessment System (17). Energy intake from NB was calculated by multiplying the amount consumed by the energy density (108 kcal per 20 g). A fieldworker who had previously been trained for the World Health Organization (WHO) Multicenter Growth Reference Study (MGRS) completed head circumference, weight, and length measurements at 6, 9, and 12 mo and assessed 4 gross motor milestones (standing with assistance, walking with assistance, standing independently, and walking independently) at 12 mo using protocols described for the MGRS (18, 19). Measuring devices for weight (model 1583, Tanita, QuickMedical, Snoqualmie, WA) and length (model 447, Infantronic Digital Infantometer; QuickMedical) were regularly calibrated. We recorded infants’ birth dates and weights from Immunization and Weighing Cards and calculated weight-for-age (WAZ), length-for-age (LAZ), and weight-for-length (WLZ) z scores by using the WHO 2006 Child Growth Standards (20).
Statistical analysis
We performed data analysis using SAS version 8.1 (Cary, NC). Characteristics of groups at baseline were evaluated by using descriptive statistics, chi-square tests, and analysis of variance. We used a factor analysis with varimax rotation to create "household amenities" (such as type and location of toilet facility) and "ownership of electronic items" (such as television and radiocassette) factors from 14 socioeconomic status variables. Adherence to treatment was determined as the percentage of scheduled days the supplement was reportedly added to the child's food, and the median adherence with 95% CIs was calculated by bootstrapping (21). Because we based our sample size on the detection of differences between 4 groups, our aim was to compare all 4 groups whenever possible at 12 mo. When the lack of baseline data for the NI group precluded its inclusion in any comparison, we compared the 3 intervention groups only. Thus, differences in mean (±SD) WAZ, LAZ, WLZ, and head circumference at 12 mo (with Tukey adjustment for pairwise comparisons) were assessed twice, first between the 3 intervention groups and then between all 4 groups by using analysis of covariance. Longitudinal prevalence of illness (natural log-transformed percentage of surveillance days with illness) was compared between the intervention groups only, because there were no data for the NI group. The odds of walking independently by 12 mo were examined between all 4 groups by using logistic regression. In all these analyses at 12 mo, adjustments were made for child sex (due to group differences) and baseline values (whenever possible). Additional covariates that were significantly related to growth (maternal height), longitudinal prevalence of illness (maternal height and household amenity factor), or the odds of walking independently by 12 mo (household amenity factor) were identified through stepwise regression and were included in the above analyses because they explained a portion of the variance and, therefore, improved statistical power. In a few analyses, we compared the NB group with the SP and NT groups combined because our a priori hypothesis was that effects would be greater in the NB group than in the other 2 intervention groups. We used a path analysis to assess whether daily energy intake from complementary foods was an intermediary variable explaining any differences between intervention groups in growth or the ability to walk independently by 12 mo of age. The analysis was by intention to treat, ie, children were included whether or not they consumed the entire dose of supplement on each intervention day.
RESULTS
Of 612 eligible infants, we randomly selected 442 for the intervention, but 129 did not enroll because of the parents’ refusal or time constraints or the inability to locate their homes (Figure 1). The remaining 313 infants underwent blood sampling at 6 mo and were randomly assigned to 1 of 3 intervention groups, of whom 298 completed the intervention. A comparison of the 313 enrolled infants with the 129 who did not enroll was not possible because only the contact information of the latter was available. The 15 dropouts did not differ significantly in socioeconomic or baseline characteristics from the 298 who completed the study. Of the 130 infants randomly selected for the NI group, 96 were recruited.
FIGURE 1.. Study profile. SP, the multiple micronutrient Sprinkles, which generally provides the Recommended Nutrient Intake (RNI) of 6 vitamins and minerals; NT, the multiple micronutrient Nutritabs, which generally provides the RNI of 14 vitamins and minerals plus some calcium and potassium; NB, the multiple micronutrient Nutributter, which generally provides the RNI of 14 vitamins and minerals plus some calcium, potassium, phosphorous, magnesium, and manganese as well as energy (108 kcal/d), mainly from fat (including 1.29 g linoleic acid/d and 0.29 g -linolenic acid/d); NI, nonintervention children not randomly selected for intervention at 6 mo but assessed at 12 mo only.
At baseline, the groups did not differ significantly in demographic or socioeconomic characteristics (Table 2) or in the prevalence of morbidity (data not shown), except for a slightly higher proportion of males in the NB group. At 12 mo, median (95% CI) adherence to treatment did not differ significantly between groups (SP, 85.8 [82.3, 90.0]; NT, 87.5 [84.8, 90.3] and NB, 88.2 [86.1, 92.3]). The main reasons for nonadherence (percentage of scheduled days mother reported not feeding child with supplement) were child illness, mother forgetting the supplement, and child refusing the supplement (in decreasing order); the groups did not differ significantly in the frequency of any of these factors (data not shown).
View this table:
TABLE 2. Selected demographic and socioeconomic characteristics of the subjects who completed the study, by group1
After the exclusion of energy from the NB supplement, energy intake from complementary foods did not differ significantly between groups (Figure 2). When the contribution of NB was included, energy intake from complementary foods was significantly greater in the NB group than in the SP or NT group, by 85 kcal/d during the 7–12-mo period. Nearly all children were still breastfed at 12 mo of age (SP: 100%; NT: 100%; NB: 97%; and NI: 97%), and the mean daily breastfeeding frequency did not differ significantly between groups (including the NI group at 12 mo), except during the first month of supplementation, when it was slightly higher in the SP (13 ± 3/d) than in the NT (12 ± 3/d) or NB (12 ± 4/d) group (P = 0.004).
FIGURE 2.. Energy intake from complementary foods during 7–12 mo of age assessed with a monthly 24-h dietary recall. SP, the multiple micronutrient Sprinkles, which generally provides the Recommended Nutrient Intake (RNI) of 6 vitamins and minerals; NT, the multiple micronutrient Nutritabs, which generally provides the RNI of 14 vitamins and minerals plus some calcium and potassium; NB1 and NB2, the multiple micronutrient Nutributter, which generally provides the RNI of 14 vitamins and minerals plus some calcium, potassium, phosphorous, magnesium, and manganese as well as energy (108 kcal/d), mainly from fat (including 1.29 g linoleic acid/d and 0.29 g -linolenic acid/d); the values for NB1 exclude the energy provided by NB.
At baseline, the intervention groups did not differ significantly in head circumference or length (Table 3), but the NT group had a significantly lower weight and, consequently, lower WAZ and WLZ than did the SP and NB groups. There were no significant differences between groups in the percentage of infants with WAZ, LAZ, or WLZ < –2. During the intervention, infants in the NB group gained more weight (1.57 kg) and length (8.3 cm) than did those in the NT group (1.35 kg and 7.8 cm, respectively); the difference with the SP group (1.39 kg and 7.9 cm, respectively) was not significant (Figure 3 and Figure 4). After adjustment for baseline anthropometric status, child sex, and maternal height, WLZ and head circumference at 12 mo did not differ significantly between the 3 intervention groups, but the NB group had significantly greater WAZ and LAZ than did the NT group (Table 4) and than did the NT and SP groups combined (data not shown). When all 4 groups were compared, WAZ and WLZ were significantly greater in the NB group than in the NT group after control for child sex and maternal height.
View this table:
TABLE 3. Baseline (6 mo) anthropometric characteristics of subjects who completed the study, by treatment group1
FIGURE 3.. Mean (±SE) absolute weight gain from 6 to 9 mo () and from 9 to 12 mo (), by group. SP, the multiple micronutrient Sprinkles, which generally provides the Recommended Nutrient Intake (RNI) of 6 vitamins and minerals; NT, the multiple micronutrient Nutritabs, which generally provides the RNI of 14 vitamins and minerals plus some calcium and potassium; NB, the multiple micronutrient Nutributter, which generally provides the RNI of 14 vitamins and minerals plus some calcium, potassium, phosphorous, magnesium, and manganese as well as energy (108 kcal/d), mainly from fat (including 1.29 g linoleic acid/d and 0.29 g -linolenic acid/d). Bars with different lowercase letters (weight gain from 6 to 12 mo) are significantly different, P < 0.05 (ANOVA).
FIGURE 4.. Mean (±SE) absolute length gain from 6 to 9 mo () and from 9 to 12 mo (), by group. SP, the multiple micronutrient Sprinkles, which generally provides the Recommended Nutrient Intake (RNI) of 6 vitamins and minerals; NT, the multiple micronutrient Nutritabs, which generally provides the RNI of 14 vitamins and minerals plus some calcium and potassium; NB, the multiple micronutrient Nutributter, which generally provides the RNI of 14 vitamins and minerals plus some calcium, potassium, phosphorous, magnesium, and manganese as well as energy (108 kcal/d), mainly from fat (including 1.29 g linoleic acid/d and 0.29 g -linolenic acid/d). Bars with different lowercase letters (length gain from 6 to 12 mo) are significantly different, P < 0.05 (ANOVA).
View this table:
TABLE 4. Adjusted anthropometric values at 12 mo for the 3 intervention groups and for all 4 groups1
The path analysis showed that the greater energy intake from complementary foods in the NB group accounted for 43% (75 g/175 g) of the greater mean weight gain of the NB group compared with the SP group and for 46% (92 g/202 g) of the greater weight gain compared with the NT group (data not shown), but did not explain the difference in length gain (data not shown).
At 12 mo there were no significant group differences in the percentage of children who were able to stand independently (Figure 5). The percentage of children able to walk independently was significantly lower in the NI group (25%) than in the SP (39%), NT (36%), and NB (49%) groups. The differences between the 3 intervention groups were not significant, although the percentage for the NB group tended to be greater than that for the NT group (P = 0.07) and for the SP and NT groups combined (P = 0.06). Compared with the NI children, the odds (95% CI) of walking independently by 12 mo of age, adjusted for child sex and the household amenities factor, were greater in the SP (2.20; 1.11, 4.34; P = 0.024), NT (2.10;1.05, 4.07; P = 0.036), and NB (3.4; 1.67, 6.43; P = 0.001) groups. In path analysis (data not shown), energy intake from complementary foods at 12 mo (a proxy for mean intake from 7 to 12 mo) did not explain the effect of supplementation on the ability to walk independently by 12 mo.
FIGURE 5.. Percentages of children achieving the standing independently and walking independently milestones by 12 mo, by group. SP, the multiple micronutrient Sprinkles, which generally provides the Recommended Nutrient Intake (RNI) of 6 vitamins and minerals; NT, the multiple micronutrient Nutritabs, which generally provides the RNI of 14 vitamins and minerals plus some calcium and potassium; NB, the multiple micronutrient Nutributter, which generally provides the RNI of 14 vitamins and minerals plus some calcium, potassium, phosphorous, magnesium, and manganese as well as energy was added (108 kcal/d), mainly from fat (including 1.29 g linoleic acid/d and 0.29 g -linolenic acid/d); NI, nonintervention children not randomly selected for intervention at 6 mo but assessed at 12 mo only. Percentages with different lowercase letters are significantly different, P < 0.05 (logistic regression with adjustment for sex and the household amenities score).
During the intervention period, morbidity did not differ significantly between the 3 intervention groups, except for a slightly higher prevalence of cough in the NT group (Table 5). There were no significant between-group differences in the percentages of children with illness during the 1-wk period before the 12-mo interview (data not shown).
View this table:
TABLE 5. Longitudinal prevalence of morbidity between 6 and 12 mo of age for infants in the 3 micronutrient groups1
DISCUSSION
We found that supplementation with SP and NT did not increase growth but the that the NB group had greater weight and length gains than did the other 2 supplementation groups. Given the difference in mean WAZ between the NB and NT groups at 12 mo, we also expected a significant difference between the NB and NI groups, because the mean WAZ for the latter was similar to that of the NT group. The lack of statistical significance in the 4-group comparison was probably due to the fact that we had no baseline data for the NI group; thus, it was not possible to control for initial anthropometric values in the 4-group analyses. As a result, the adjusted means had large SDs, which made it difficult to show statistically significant differences between the 4 groups.
The lack of a significant effect of SP and NT on growth was contrary to our original hypothesis, which was formulated on the basis of previous findings that zinc supplementation (alone) can increase growth in underweight and stunted children (22). However, the results of subsequent multiple micronutrient supplement trials, published after our study was begun (9-11), have been mixed. Our results are consistent with those from Indonesia, Peru, South Africa, and Vietnam, which show that multiple micronutrient supplementation did not increase growth compared with a placebo (10, 23-26). In Cambodia, 6-mo-old infants who received 2 types of SP daily for 12 mo also did not differ significantly in growth from a placebo group (9). These results suggest that multiple micronutrient interventions alone may not improve the growth of infants in some populations. It is possible that the lack of effect of SP and NT in our sample was due to the relatively low prevalence of underweight or stunting at baseline, interference of other nutrients (particularly iron) with the utilization of zinc, poor absorption of zinc due to predominantly vegetarian diets, or a combination thereof. Interactions between iron and zinc have been observed in other studies. For example, in Indonesian infants receiving supplements from 6–12 mo of age (27), zinc alone (10 mg) increased WAZ and knee-heel length, and iron alone (10 mg) increased knee-heel length and psychomotor development, compared with a placebo, but iron + zinc had no significant effect on growth or development. In that study, the supplement was taken as a syrup between meals, and the iron:zinc molar ratio was 1.17:1, whereas in our study the supplements were mixed with food, and the iron:zinc molar ratio was 2.7:1. It is also possible that, if our intervention had been longer, as was the case in the 12-mo micronutrient supplementation trial in Mexico (28), we would have observed significant growth differences between the micronutrient-supplemented and NI groups.
Previous studies with fortified spreads reported positive effects on the weight gain (29) of severely malnourished children, but this is the first study to show that a highly fortified spread used as a complementary food supplement increased the weight and length of infants whose average anthropometric status was not low initially. Consumption of NB resulted in an average increase of 85 kcal in the daily energy intake from complementary foods. Average energy needs from complementary foods are 200 kcal/d at 6–8 mo and 300 kcal/d at 9–11 mo, assuming an average intake of breast milk at these ages (7). Thus, it is likely that infants in the NB group were more apt to meet daily energy requirements, and, in fact, the increased energy intake from complementary foods partially explained the greater weight gain in the NB group in a path analysis. Increased energy intake from complementary foods as a result of an educational intervention was also shown to increase infant weight in China (30) and length in India (31).
However, the increased energy intake from complementary foods in the NB group did not explain the difference in length gain, which implies that some other factor was involved. This factor was unlikely to be zinc (22), because all of the intervention groups received zinc. A possible explanation is the essential fatty acid content of the NB, which provided 1.29 g linoleic acid/d (65% of the recommended intake) (32) and 0.29 g -linolenic acid/d, (more than the recommended intake of 0.2 g/d) (32). In fact, other results from the study (not presented here) showed that the NB group had a significantly greater plasma -linolenic acid concentration and a lower percentage of saturated fatty acids, and that this difference in fatty acid profile explained much of the difference in linear growth among the intervention groups. This finding is consistent with evidence from a meta-analysis suggesting that -linolenic acid intake is associated with length gain in term infants (33). Studies in Burkina Faso and Congo Brazzaville also suggest that a low intake of essential fatty acids is associated with growth retardation (34).
Compared with the NI group, the odds of being able to walk independently by 12 mo were 2-fold greater in the SP and NT groups and 3.4 times those in the NB group. These results are in keeping with those from Bangladesh, which show that infants who received iron and zinc together and with other micronutrients from 6 to 12 mo of age had significantly higher scores on a motor development scale (Bayley II) than did those who received only riboflavin (35). In Indonesia, a cohort of 12-mo-old infants who received 280 kcal plus 12 mg Fe/d walked at an earlier age than did their counterparts who received only 25 kcal or 50 kcal plus 12 mg Fe (36). These results indicate the importance of providing supplements to children at risk of micronutrient deficiencies, even if growth is not improved. Infants who have well-developed motor skills and are explorative should be better prepared to take advantage of environmental resources and, in disadvantaged populations, may ultimately demonstrate better cognitive skills than children with delayed motor development (37).
As we found in a previous study in Ghana (38), the 3 intervention groups did not differ significantly in the prevalence of illness, except for cough, which was slightly higher in the NT group. It is unlikely that this small difference accounted for the difference in growth between the NT and NB groups because we did not find a significant relation between cough and growth.
We are unaware of any randomized trial that has compared the different approaches (ie, a micronutrient powder, a crushable tablet, and a fat-based spread) for delivering micronutrients to infants. A limitation of our design was that the home visits for the 3 intervention groups may have influenced the amount of attention given to the children and thus the children's health status, even if no supplements had been provided (the Hawthorne effect). Because the NI group did not receive frequent home visits, some of the difference in motor development may have been due to the Hawthorne effect rather than to the supplements. However, in a feeding trial in India among infants 4–12 mo of age, Bhandari et al (39) found no evidence for the Hawthorne effect. In that study there were 2 control groups (a visitation-only group and a nonintervention group with no extra visits) in addition to the 2 intervention groups (food supplements plus counseling and counseling alone). There were no significant differences in infant growth, breastfeeding practices, or energy intake from complementary foods between the 2 control groups. Although it is unknown whether this finding can be generalized to other populations, it suggests that the Hawthorne effect does not completely explain our results. A second limitation of our study is that it was not possible to mask the mothers or the fieldworkers who delivered the supplements to the study design. However, the anthropometrists, who also assessed motor milestones, were masked to group assignment. A strength of the study was the low dropout rate and the lack of any significant difference between the dropouts and participants in baseline characteristics. These results suggest that our findings are generalizable to the study population.
We conclude that the NB supplement, which provided a larger set of micronutrients plus some energy and fat, conferred greater benefits than did the SP and the NT supplements. Other results from this study indicate that all 3 supplements were well accepted and improved iron status (Adu-Afarwvah S, Lartey A, Brown KH et al, unpublished observations, 2006). Further research is needed on the efficacy and safety of multiple micronutrient supplements for home fortification of complementary foods, particularly with regard to nutrient interactions and effects on morbidity in malaria-endemic communities.
ACKNOWLEDGMENTS
We thank the families of the 409 infants, the study team members, including Godlove Addo (lead anthropometrist), Ebenezer Appiah-Dankyirah (Regional Director of Health Services), Francisca Djata (Head of Laboratory, Regional Hospital), the nurses (Reproductive Health Facilities), Veronica Yakubu (Municipal Nutrition Officer) and Bismark Sarkodie (Regional Nutrition Officer) for cooperation in Koforidua, Janet M Peerson (University of California, Davis) for statistical advice, Diane B Vandepeute (University of California, Davis) for administrative support, Paul Equipart for advice on the manufacture of NT, and Nutriset for the development and manufacture of NB.
The authors’ responsibilities were as follows—SAA and KGD: implementation of the study, analysis and interpretation of the data, and writing of the manuscript; AL and KHB: implementation of the study, interpretation of the data, and editing of the manuscript; AB: interpretation of the data; all authors: study design. SZ owns the IP rights to SP. The HJ Heinz Company has supported the technical development of SP on a cost-recovery basis. Any profit from royalty fees on the technology transfer of SP is currently donated to the Hospital for Sick Children Foundation. Until December 2003, AB was a paid consultant of Nutriset, the company that manufactured NB. None of the other authors had any potential conflicts of interest.
REFERENCES