点击显示 收起
1 From the Department of Nutrition, University of California, Davis, Davis, CA.
2 Supported by grant no 94-37200-2536 from the US Department of Agriculture. 3 Reprints not available. Address correspondence to MJ Heinig, Department of Nutrition, University of California, One Shields Avenue, Davis, CA 95616-8669. E-mail: mjheinig{at}ucdavis.edu.
ABSTRACT
Background: It has been documented that growth patterns differ between breastfed and formula-fed infants. Some investigators have suggested that these differences may be related to differences in zinc nutriture.
Objective: The objective of this study was to examine the effect of zinc supplementation on growth, morbidity, and motor development in healthy, term, breastfed infants.
Design: We conducted a randomized double-blind intervention comparing zinc supplementation (5 mg/d as zinc sulfate) with placebo in breastfed infants aged 4–10 mo. Growth and indexes of body composition and gross motor development were measured monthly from 3 to 10 mo. Morbidity data were collected weekly.
Results: Eighty-five infants were enrolled, and 70 completed the study. The baseline characteristics, attained weight or length at 10 mo, growth velocity, gross motor development, and morbidity did not differ significantly between groups, even after control for potentially confounding variables.
Conclusions: The dietary zinc intake of these breastfed infants appeared to be adequate, given that zinc supplementation did not affect growth, development, or risk of infection (although sample size for detection of differences in development or infection was limited). Previously described differences in growth between breastfed and formula-fed infants in such populations do not appear to be due to differences in zinc nutriture.
Key Words: Zinc • randomized supplementation trial • infant growth • motor development • breastfeeding
INTRODUCTION
One of the first signs of zinc deficiency in both humans and experimental animals is impaired growth (1-8). Current recommendations for zinc intake during early infancy are based on the assumption that, during the first 6 mo of life, breastfed infants meet their needs via human milk and initial stores (9). However, the zinc concentration of breast milk declines during lactation (10), and when the infant is 4–5 mo old, the amount of zinc consumed relative to requirements may become marginal (11). Although reports of overt zinc deficiency among breastfed infants are rare (12, 13), the slower weight gain of breastfed than of formula-fed infants in industrialized countries (14, 15) has prompted some to question whether zinc may be a limiting nutrient for the growth of breastfed infants (16, 17).
Studies in both breastfed and formula-fed infants have shown that infants are able to adapt to lower intakes of zinc by increasing fractional absorption and decreasing endogenous losses of zinc (18-20). However, the extent to which endogenous losses can be reduced appears to be limited (19). Using stable isotope methods to evaluate zinc homeostasis, Krebs et al (20) determined that breastfed infants aged 2–5 mo had adapted nearly maximally and that mean net absorption of zinc was marginally adequate to meet estimated needs. Although the bioavailability of zinc from human milk is high (11, 20-22), any impairment of zinc absorption from breast milk, such as that caused by phytate or other components in the diet (23), may put the infant at risk of zinc deficiency, even in relatively affluent populations.
The effect of zinc supplementation on child growth has been studied extensively in developing countries, but relatively little information is available from industrialized countries (24). In a meta-analysis of clinical trials, Brown et al (24) reported that a growth response to zinc supplementation was evident in populations in which the average initial height-for-age z score was < –1.5 but not in those in which stunting was less prevalent. Thus, it is unclear whether children in industrialized countries would benefit from increased zinc intakes. Most of the zinc supplementation trials have been conducted in children aged > 1 y or in formula-fed infants. Of the 6 published studies that have included term, breastfed infants, 4 were conducted in developing countries (25-28), and 1 was conducted in a predominantly African population living in France (15). One other zinc supplementation study of breastfed infants in the United States has been reported in preliminary form (16).
The objective of our study was to ascertain whether zinc supplementation increases weight or length gain among breastfed infants in a relatively affluent population. Although the primary outcome variable was growth, we also included the "functional" measures immune competence, morbidity, and motor development. In humans and in animal models, zinc deficiency has been reported to impair immune function (29-32).
SUBJECTS AND METHODS
Subjects and study design
The study was designed as a randomized, double-blind, placebo-controlled trial of zinc supplementation in infants from 4 to 10 mo of age. Mothers who planned to fully breastfeed (ie, no other milk or formula was given on a daily basis) for 10 mo were recruited during the first 3 mo after delivery, primarily through local pediatricians' offices. Recruitment and data collection were completed between November 1994 and August 1997.
Selection criteria included 1) healthy term infant weighing > 2500 g at birth; 2) healthy nonsmoking mother 19 y of age, with no chronic medical condition that could interfere with lactation; 3) mother planned to fully breastfeed for 10 mo (ie, would not give formula on a daily basis) and not to introduce complementary foods before 4 mo; and 4) mother planned to remain in the study area throughout the study period.
Written informed consent was obtained from all participants. The protocol was approved by the Committee on the Use of Human Subjects in Research at the University of California, Davis.
Infants were randomly assigned to receive 5 mg elemental zinc (as zinc sulfate) or placebo in drops given each day between the ages of 4 and 10 mo. Infants were stratified by sex, and random assignment to groups was done by using the Moses-Oakford algorithm, as described by Meinert and Tonascia (33). The zinc supplement was prepared in the pharmacy at the University of California, Davis, Medical Center and shipped to the study office, where an assistant, who was not in contact with the study subjects, labeled the bottles with 1 of 4 colors (2 colors were assigned to each group to reduce the chance that a group assignment would accidentally be revealed). Each mother-infant pair was assigned to a color group so that neither the primary investigator nor the mothers would know whether their infants received the zinc supplement. Mothers were provided each month with a supply of zinc or placebo and instructed to give the dose by dropper each morning 1 h before feeding any complementary foods. Mothers were given a calendar to mark each day that the zinc or placebo was given. These calendars were reviewed during each monthly visit, and the number of days that the drops were missed was recorded to calculate adherence to treatment. Mothers were also asked to return the bottles each month so that any unused portion could be measured.
Sample size was calculated on the basis of growth data from the Davis Area Research on Lactation, Infant Nutrition and Growth Study (14), in which mean weight gain from 4 to 6 mo and from 6 to 10 mo was 20% higher in the formula-fed group than in the breastfed group. Assuming that the effect of zinc supplementation would be of similar magnitude, we chose 20% as the expected difference between groups. A final sample size of 35 per group (70 total) was calculated to be sufficient to detect a 20% difference in weight or length gain ( = 0.05, ß = 0.20).
Anthropometry
Birth weight and length were recorded from physicians' records. Starting when their infants were 3 mo old, mothers were asked to bring their infants to a central facility each month for measurements of length, weight, head circumference, and skinfold thickness at the triceps, subscapular, flank, and quadriceps sites. All infants who remained available for measurements were followed to age 10 mo. Mothers who were unable to bring their infants to the office were visited in their homes. All measurements were performed by one of us (MJH), according to the procedures described by the World Health Organization (34). The weight-for-age, length-for-age, and weight-for-length z scores were calculated by using National Center for Health Statistics reference data (35). Maternal and paternal weight and height were also measured according to standard procedures (14). Estimated prepregnancy weight was obtained by maternal recall, and maternal weight was measured at 6 mo after delivery.
Blood samples
At infant ages 4 and 10 mo, blood samples were obtained by venipuncture from infants whose mothers were willing to permit this procedure. Blood samples were taken only when infants were free of symptoms of any illness for 10 d before the procedure. To minimize the influence of diurnal and postprandial variation on indexes such as plasma zinc, all of the samples were collected in the morning (between 0800 and 1000), 2 h after a feeding. Hemoglobin and hematocrit were analyzed according to standard procedures. After centrifugation, plasma and erythrocytes were separated and frozen for later analysis. Appropriate precautions were taken to avoid trace element contamination. Plasma zinc, iron, and copper were analyzed by using flame atomic absorption spectrophotometry (36). Plasma ferritin was analyzed by radioimmunoassay (Diagnostic Products Corp, Los Angeles, CA). Immunoglobulin G2 and G4 were analyzed by using radial immunodiffusion (The Binding Site Inc, San Diego, CA). After extraction (37), erythrocyte superoxide dismutase was analyzed according to the method outlined by Marklund and Marklund (38).
Infant dietary intake
Mothers were asked to record on the calendar provided the date when each new food was introduced. From 6 to 10 mo, mothers were asked to keep once a month a detailed 1-d diet record of the amounts (using standard household measures) of nonbreast-milk foods and fluids consumed. Nutrient intake from complementary foods was calculated by using FOOD PROCESSOR II software (version 14; ESHA Research, Salem, OR) and information from food manufacturers. At each visit, the mother was also asked to fill out a brief food-frequency questionnaire regarding her infant's intake during the previous week.
Morbidity
Throughout the study, mothers were asked to record on the calendar provided any symptoms of infant illness, using standardized guidelines for the description of symptoms. Mothers were contacted by phone once a week to obtain and clarify this information. Physician contacts and diagnoses were also recorded. Morbidity data were coded and grouped into 5 major categories: respiratory illness, diarrhea, otitis media, fever (without other symptoms), and other illnesses. Respiratory illness was defined as colored nasal discharge, congestion (difficulty breathing), or coughing or sneezing (or both) accompanied by nasal discharge. Diarrhea was defined as 3 watery stools in a 24-h period, with or without vomiting. Otitis media was identified by physician diagnosis. Fever was defined as an elevated temperature (38 °C) without accompanying symptoms. The most frequent conditions in the "other illnesses" category were viral rashes, vomiting without diarrhea, and eczema.
Motor development
Each month, infant motor development was assessed by using the Alberta Infant Motor Scale (AIMS) (39). The AIMS is a 58-item observational instrument designed to evaluate motor maturation from birth to the age of independent walking; it is based on the sequential development of postural control relative to 4 postural positions: supine (9 items), prone (21 items), sitting (12 items), and standing (16 items). The infant's score is based on the number of behaviors performed spontaneously during the observation. One of us (JMH) conducted all of the AIMS assessments in the study office, and the conditions for testing were kept the same for all infants (eg, quiet room, carpeted, and infant alert and recently fed). The validity and reliability of the method were previously tested in a sample of 506 infants, in whom the correlation coefficient with the Bayley Scale of Motor Development was 0.98 (39).
Statistical analysis
Initial analyses included descriptive characteristics of both groups of subjects from age 4 to 10 mo and of those who left the study before age 10 mo. For continuous variables, differences in characteristics between groups were assessed by using Student's t test and the Mann-Whitney U test. Differences in categorical variables were assessed by using chi-square tests. Outcome variables were analyzed on an intention-to-treat basis. Analysis of covariance was used to examine the effect of zinc supplementation on growth and developmental scores from ages 4 to 6, 6 to 10, and 4 to10 mo after control for initial size and infant sex. When variables were not normally distributed, transformation to normalize the distribution was attempted (ie, natural log was taken of ferritin values). Logistic regression was used for morbidity and blood indexes when no transformation could normalize the distribution of the variables. Statistical analyses were performed with SPSS software (version 14; SPSS Inc, Chicago, IL).
RESULTS
After screening, 92 mother-infant pairs were found to be eligible for participation in the study. Seven mothers declined to participate, 1 because of a lack of transportation and 6 because they believed the study procedures would be inconvenient. Eighty-five infants were enrolled in the study and were randomly assigned to treatment groups: 41 to the zinc group and 44 to the placebo group (Figure 1). Three of the 85 infants (1 in the zinc group and 2 in the placebo group) were lost to follow-up because the family moved out of the area or the mother decided not to continue for personal reasons. Twelve infants (7 in the zinc group and 5 in the placebo group) became ineligible after enrollment because they stopped breastfeeding or consumed supplementary formula on a daily basis. Of these 12 infants, 2 (1 in each group) were weaned because of "poor growth," as perceived by the infant's physician; 3 refused to continue breastfeeding; and 7 were given formula on a daily basis because their mothers returned to work. The 12 infants who became ineligible stopped taking the zinc or placebo drops, but measurements of growth and motor development were continued to 10 mo of age. There were no significant differences in demographic characteristics or growth from age 4 to 10 mo between the 70 infants who remained eligible and those who became ineligible after enrollment. The results shown below are for the 70 infants who remained eligible; inclusion of the 12 infants who became ineligible did not alter the findings.
FIGURE 1.. The trial profile.
The differences between the zinc and placebo groups in any of the baseline characteristics were not significant (Table 1). Average maternal age (31 y) and educational level (16 y) were relatively high, and most of the mothers were multiparous, white, and relatively affluent. Birth weight averaged 3700 g. Complementary foods were introduced at an average age of 22 wk in both groups.
View this table:
TABLE 1. Characteristics of subjects in the zinc and placebo groups1
The proportion of days (out of 182 d) on which the zinc or placebo drops were given was 93% in the zinc group and 88% in the placebo group. Two infants in the placebo group never accepted the drops, and 1 infant in each group missed > 60 d because of family crises. When these 4 infants were excluded, the proportion of days on which the drops were given was 94% in the zinc group and 93% in the placebo group. Because exclusion of these infants from the analyses did not affect any of the results, the results below are based on the entire cohort (ie, with the use of intention to treat).
The differences between groups in the average nutrient intake from complementary foods between ages 6 and 10 mo were not significant (Table 2). These values include non-breast-milk foods only and do not include the zinc in the drops given to the zinc group. The differences in the frequency of consumption of various categories of complementary foods offered to the infants also were not significant (Table 3). The foods most commonly offered to the infants were fruit, infant cereals, and vegetables. Formula was consumed occasionally by 12 infants (2 in the placebo group, 10 in the zinc group), but this consumption was either temporary (<1 wk) or infrequent (3 d/wk), so the average number of formula feedings per day was very low and did not differ between groups.
View this table:
TABLE 2. Daily nutrient intake from complementary foods of infants (6–10 mo old) in the zinc and placebo groups1
View this table:
TABLE 3. Frequency of consumption of foods from various food groups in the zinc and placebo groups1
The differences between groups in weight or length gain from age 4 to 10 mo and in any other of the anthropometric indexes were not significant (Table 4). Mean weight gain was 348 ± 75 and 359 ± 89 g/mo and mean length gain was 1.48 ± 0.15 and 1.48 ± 0.19 cm/mo in the zinc and placebo groups, respectively. The results were similar when analyzed separately for males and females or for those with lower or higher weight or length at birth or at 4 mo (data not shown). The weight-for-age z scores from ages 4 to 10 mo are shown in Figure 2. The differences between groups in attained weight and in weight gain from ages 4–6 or 6–10 mo were not significant. Infant weight-for-age z scores declined in both groups from age 4–10 mo, and they fell below the National Center for Health Statistics median by age 9 mo. The differences between groups in length-for-age or in length gain from ages 4–6 or 6–10 mo were not significant (Figure 3). Although the average initial weight-for-length z score was slightly lower in the zinc group than in the placebo group, the differences between groups in the change in weight-for-length z scores between ages 4 and 10 mo were not significant (Figure 4).
View this table:
TABLE 4. Anthropometric indexes for infants in the zinc and placebo groups at 4 and 10 mo1
FIGURE 2.. Mean (±SEM) weight-for-age z scores of zinc-supplemented infants (n = 33) and those receiving placebo (n = 37) from 4 to 10 mo of age, obtained by analysis of covariance after control for initial size and infant sex. The change in z scores between ages 4 and 10 mo did not differ significantly between groups (–0.89 ± 0.08 and –0.84 ± 0.08, respectively; t test).
FIGURE 3.. Mean (±SEM) length-for-age z scores of zinc-supplemented infants (n = 33) and those receiving placebo (n = 37) from 4 to 10 mo of age, obtained by analysis of covariance after control for initial size and infant sex. The change in z score between ages 4 and 10 mo did not differ significantly between groups (–0.40 ± 0.07 and –0.38 ± 0.07, respectively; t test).
FIGURE 4.. Mean(±SEM)weight-for-lengthzscoresofzinc-supplemented infants (n = 33) and those receiving placebo (n = 37) from 4 to 10 mo of age, obtained by analysis of covariance after control for initial size and infant sex. The change in z score between ages 4 and 10 mo did not differ significantly between groups (–0.49 ± 0.11 and –0.39 ± 0.09, respectively; t test).
The differences between groups in gross motor development based on AIMS scores at each age were not significant, as shown in Figure 5. All scores were at or above the 25th percentile for age. None of the infants was classified as being at risk (<–1 SD) or abnormal (<–2 SD). In addition, the incidence (Figure 6) or prevalence (data not shown) of diarrhea, otitis media, respiratory illness, fever only, and total illness over the study period did not differ significantly between groups. The values shown are adjusted for day care use and the presence of siblings in the home.
FIGURE 5.. Mean (±SEM) total attained score on the Alberta Infant Motor Scale for zinc-supplemented infants (n = 33) and those receiving placebo (n = 37) from 4 to 10 mo of age. The differences between groups by analysis of covariance after control for initial size and infant sex were not significant.
FIGURE 6.. Mean (±SEM) incidence of illness in zinc-supplemented infants (n = 33) and those receiving placebo (n = 37) throughout the study period. The differences between groups by Mann-Whitney U test were not significant.
Very few mothers were willing to allow us to draw blood from their infants. However, there were no differences in growth or any of the baseline characteristics between those who allowed us to take blood samples and those who did not (data not shown). There were no differences between groups in any of the hematologic indexes at 4 mo (n = 7/group; data not shown). There were no significant differences between groups in change in hematologic indexes from age 4 to 10 mo (data not shown). However, only 4 infants provided blood samples at both time points. The median values and ranges for the hematologic variables in the subset of infants from whom we obtained blood samples at ages 4 mo, 10 mo, or both are shown in Table 5. Plasma zinc concentrations were higher and copper concentrations were lower in the zinc group than in the placebo group at age 10 mo (P < 0.05, Mann-Whitney U test). The differences between groups in erythrocyte copper–zinc superoxide dismutase activity and in any of the other hematologic indexes at age 4 or 10 mo were not significant.
View this table:
TABLE 5. Hematologic and biochemical indicators in a subset of infants from the zinc and placebo groups at ages 4 and 10 mo1
DISCUSSION
Zinc supplementation had no significant effect on the growth of the infants in our population. These results conflict with those of 2 other zinc supplementation trials conducted in breastfed infants in industrialized countries (16, 17). In a study of 57 breastfed infants aged 4–9 mo, Walravens et al (16) reported significantly greater weight and length gains in infants (primarily in the males) receiving 5 mg Zn/d for 3 mo than in those receiving a placebo. The study was conducted in Paris with infants of mostly low-income immigrant families from Africa, and the authors concluded that zinc deficiency had limited growth velocity in this cohort. However, information was lacking on whether these infants were provided with complementary foods according to current guidelines. If complementary foods were given too early or were low in zinc or high in phytate (or both), it is possible that the response of the zinc-supplemented group may not have been typical of infants given higher-quality complementary foods at the appropriate age. Furthermore, infants were enrolled in the study at variable ages and the attrition rate in the zinc-supplemented group appeared to be greater. The differential attrition may have created a bias between groups, particularly if the dropouts in the zinc-supplemented group grew less rapidly than did the infants who remained. In a study of breastfed infants in the United States, Krebs et al (17) reported somewhat greater weight gain (but not length gain) from age 2 to 6.5 mo in females receiving 5 mg Zn/d than in those receiving placebo; no effect of zinc supplementation was seen in the male infants. Complementary foods were introduced after 5 mo, but no information is published on the types and quantities of foods offered to these infants. Data on functional outcomes such as morbidity and development were not available from either of these supplementation studies.
Although growth outcomes did not differ between groups in our study, weight-for-age declined over time in both groups, as has been reported among breastfed infants in other populations (40). This decline occurred despite the introduction of complementary foods at an average of 22 wk of age in both groups. It should be noted, however, that mean initial weight-for-age and length-for-age were well above the reference median, and mean length-for-age z scores remained above the median throughout the study. Zinc intake from complementary foods at 6–10 mo of age (1.4 mg/d) did not differ significantly between groups, and infants in both groups consumed meat, poultry, or fish an average of 2–3 d/wk. Animal source foods are an important source of bioavailable zinc (41). Nonetheless, if we assume that the infants received an average of 0.5 mg Zn/d from human milk (42), total dietary zinc intake in our cohort (not including supplements) averaged < 2 mg/d, which is less than the estimated average requirement of 2.5 mg/d for infants aged 7–12 mo (9). Similar intakes among US breastfed infants have been reported by Krebs (41) and in the Continuing Survey of Food Intakes by Individuals for 1994–1996 and 1998 (43; J Arsenault, personal communication, 2005). In a randomized trial in which infants at 5 mo of age were given either beef or iron-fortified rice cereal as the first complementary food (44), no effect on weight or length at age 9 mo was observed despite significant differences in zinc intakes at age 5–7 mo. These results, and the lack of response to zinc supplementation in our cohort, suggest that most breastfed infants in relatively affluent populations absorb enough zinc to meet their needs for growth. This contrasts with the generally positive effect of zinc supplementation of breastfed infants in disadvantaged populations (25-28). It is likely that infants in such populations are at greater risk of zinc deficiency not only because of dietary factors but also because of lower birth weights and higher morbidity associated with intrauterine malnutrition, poor sanitation, and greater risk of infectious disease. Hepatic metallothionein at birth apparently is a source of zinc during the early postnatal period (11), and hepatic zinc reserves may be correlated with birth weight, as is the case for iron. This variability in zinc reserves may account for some of the variability in the results of zinc supplementation trials.
In our cohort, there was no significant difference between groups in attained AIMS scores. No other studies of zinc supplementation and motor development in industrialized countries were found. Undernourished children have been shown to have deficient or delayed cognitive and motor development (41, 45). Zinc supplementation has been shown to affect activity patterns, but not motor development, of rural Guatemalan infants (46). Thus, it is not surprising that we did not find effects on motor development, especially given the fact that the subjects in our study were generally well nourished. The change in AIMS score between ages 4 and 10 mo was 35.8 ± 4.5 and 35.5 ± 5.7 in the zinc and placebo groups, respectively; to show statistical significance for such a small difference would have required a sample size of >1000/group.
There were also no significant differences between groups in either prevalence or incidence of illness in our cohort. However, the sample size in the current study was not calculated to examine differences in morbidity rates, and thus the study may have lacked sufficient power to detect significant differences in these outcomes. Rates of illness in our cohort were much lower than those found in other supplementation studies carried out in developing countries (47). Marginal zinc deficiency has been shown to impair immune function in animal models, including infant rhesus monkeys (48). Zinc supplementation has been reported to improve immune response and decrease morbidity in infants and children in developing countries (47, 49-52), but no information was found regarding morbidity and zinc supplementation in the populations of industrialized countries.
We were able to obtain blood samples from only a small number of infants (14 at age 4 mo, 19 at age 10 mo). Despite the small sample size, plasma zinc concentrations were significantly higher in the zinc group, as has been seen in almost all supplementation trials regardless of setting (24). Plasma copper concentrations were significantly lower in infants who received zinc supplements than in those who did not, although the difference in red blood cell copper–zinc superoxide dismutase activity, which is considered a marker of long-term copper status, was not significant (possibly because of the small sample size). Zinc has been shown to impair copper absorption (53), and infant rhesus monkeys receiving zinc-fortified formula had lower plasma copper than did those fed unsupplemented formula (48). However, Sazawal et al (54) found no effect of zinc supplementation for 4 mo on plasma concentrations of copper in infants in India. Oral zinc supplements have also been shown to inhibit iron absorption in humans (55). We did not observe significant differences between groups in hemoglobin, hematocrit, or serum iron or ferritin, but the small sample size severely limited the statistical power of these comparisons.
To summarize, we found no significant effect of zinc supplementation on attained weight or length at 10 mo, growth velocity, gross motor development, or morbidity. The sample size was sufficient to detect a 300-g difference between groups in weight gain from ages 4 to 10 mo (based on the actual variability in growth in these subjects), but more subtle effects may not have been detected. Furthermore, the study cohort was composed of healthy, term infants with relatively high birth weight and generally adequate complementary-food diets after 6 mo of age, who were thus at relatively low risk of zinc deficiency. Nonetheless, we conclude that the slower growth rate of breastfed infants than of formula-fed infants, is not related to zinc nutriture in the population in the current study.
ACKNOWLEDGMENTS
We thank Janet Peerson for her statistical expertise, Angie Lee-Ow for preparation of the supplements, and Shannon Kelleher for her technical assistance.
All authors were responsible for the study design. MJH and KGD were responsible for data acquisition and statistical analysis. All others were responsible for interpretation of the data. MJH and KGD were responsible for drafting the manuscript. All authors contributed to the revision of the manuscript. None of the authors had any personal or financial conflict of interest.
REFERENCES