Dietary intervention strategies to enhance zinc nutrition: Promotion and support of breastfeeding for infants and young children

Abstract

Breastmilk is the only dietary source of zinc for exclusively breastfed young infants, and it remains a potentially important source of zinc for older infants and young children who continue breastfeeding beyond early infancy. Therefore, we examined available information on breastmilk zinc concentration and total milk consumption to develop estimates of the amount of zinc transferred in breastmilk to children of different ages. Breastmilk zinc concentration declines rapidly during the first few months postpartum and more slowly thereafter. Breastmilk supplies all of the theoretical zinc needs for at least the first several months of life, although the period during which breastmilk alone remains sufficient is uncertain. Breastmilk continues to provide more than half of children’s estimated zinc requirements after the introduction of complementary foods, even into the second year of life. Public health programs to promote and support breastfeeding should be included among the strategies to ensure adequate zinc status of young children.

Background

Promotion and support of appropriate breastfeeding practices should be considered among the recommended dietary strategies to enhance the zinc status of infants and young children, for two main reasons: breastmilk is an important potential source of bioavailable zinc, and breastfeeding protects against diarrhea [1–3], which causes excessive zinc losses [4], This paper will provide a summary of available information on the amount of zinc that can be provided to infants and young children through breastfeeding. The paper is divided into three sections, which address the following questions with regard to breastfeeding as a component of dietary intervention strategies to prevent zinc deficiency:

• Section 1: How much zinc is transferred in breastmilk to exclusively breastfed infants less than 6 months of age in relation to their physiological requirements for absorbed zinc?

• Section 2: How much zinc is transferred in breastmilk to infants and young children less than 24 months of age who are consuming breastmilk in addition to complementary foods?

• Section 3: What are the programmatic implications of the answers to the foregoing questions, and what are the remaining research needs?

Section 1

How much zinc is transferred in breastmilk to exclusively breastfed infants less than 6 months of age in relation to their physiological requirements for absorbed zinc?

Conclusions

The mean amount of zinc transferred in breastmilk to exclusively breastfed infants declines rapidly from ~4 mg/day during the first few days of life to ~1.75 mg/day by 1 month. Zinc transfer declines more slowly thereafter to ~0.7 mg/day at 6 months. The adequacy of breastmilk as the sole source of zinc during this period is difficult to determine because of uncertainty in the fractional absorption of zinc (FAZ) from breastmilk and doubts regarding the true physiological requirements for absorbed zinc during this period of life Nevertheless, using different assumptions for likely FAZ, based on available studies that suggest a mean FAZ of ~0.50 and current estimates of zinc requirements, breastmilk alone is probably adequate for exclusively breastfed term infants for at least 3 months and possibly for as long as 6 months, depending on which sets of estimates are applied and whether zinc accrued by the fetus during gestation is available for infant metabolism.

There is little information from zinc supplementation trials conducted among exclusively breastfed term infants less than 6 months of age, and the results of supplementation trials carried out among a broader age range of breastfed infants are inconsistent. In general three such trials carried out in industrialized countries failed to demonstrate a consistent functional advantage of providing additional zinc to infants less than 6 months of age, so the amount transferred in breastmilk is presumably adequate during this period. Breastmilk is clearly an important source of highly bioavailable zinc during this period of life and may be adequate as the sole source of zinc for exclusively breastfed term infants until ~6 months.

Detailed review of evidence

Overview of methodologie issues

Milk sampling

Accurate measurement of the amount of zinc transferred through breastmilk from mothers to their nursing children requires first, collection of representative samples of milk for analysis of zinc concentration and second, assessment of the total amount of milk consumed. The total amount of zinc transferred through breastmilk can then be determined by multiplying the milk zinc concentration by the total amount of milk consumed. To develop an appropriate milk sampling protocol for assessing milk zinc concentration, several investigators have examined whether milk zinc concentration varies by the time of day, portion of the feeding, left versus right breast, the gestational age at birth and postnatal age of the child, and the age, parity, and zinc status of the mother. The results of these studies are summarized in the following paragraphs. The single factor that appears to have the greatest impact on milk zinc concentration is the time postpartum. Thus, assessments of other factors affecting milk zinc concentration are interpretable only if the age of the child is held constant. The specific relation between time postpartum and milk zinc concentration is discussed in more detail in the next section.

Studies of milk zinc concentration in relation to time of day have yielded inconsistent results. Four studies found no relation between the time of day and milk zinc concentration [5–8], but three studies found that specimens obtained early in the morning had ~15% greater milk zinc concentration than specimens obtained in the afternoon or evening [9–11]. Thus, it might be advisable to obtain more than one sample during the day to gather more representative information on the total amount of zinc transferred, even though many studies suggest that just one sample might he sufficient.

Two studies examined milk zinc concentrations in different portions of the feeding, and neither found any significant differences in the zinc concentrations of foremilk and hindmilk [5, 12]. Only one study compared zinc concentrations in milk obtained from the left versus the right breast, and no consistent differences were detected [12].

Ten studies are available that compared the zinc concentration in the milk of mothers who delivered at term with that of mothers who delivered prematurely [13–22]. There were no consistent differences in the zinc concentrations of milk obtained from the respective sets of mothers. Other studies of the relations between maternal age and parity and milk zinc concentrations are difficult to interpret because of the failure to control adequately for infant age. Among the subset of studies that did compare results at specific, narrowly defined time periods postpartum, there are no consistent effects of maternal age or parity on milk zinc concentrations [7,23–26].

As reviewed by Hess and King in this issue [27], maternal zinc status and maternal zinc supplementation do not appear to affect milk zinc concentration. On the other hand, several longitudinal studies have found greater variability in milk zinc concentration between women than within women [10, 28], indicating that individual women have characteristic levels of zinc in their milk, possibly because of genetic factors influencing zinc transport to or within the mammary gland. The public health importance of these interindividual differences in breastmilk zinc concentration is unknown, because no relevant studies are available relating milk zinc concentration to infant growth.

In summary, there are few extraneous factors, other than time postpartum, that consistently affect breastmilk zinc concentration. Thus, it is possible to collect milk samples for analysis of zinc concentration from either breast, from any portion of the feeding, and at just one time or a few times per day, regardless of the age, parity, and nutritional status of the mother.

Pattern of distribution of data regarding milk zinc concentration and milk volume

To calculate the range of total zinc consumption from breastmilk by children of different ages, information is needed on the distribution of data concerning milk zinc concentration and total milk intakes during each age interval of interest. Few studies have reported information on the distribution of milk zinc concentration at different ages, so we reanalyzed data from a previously published study that collected milk samples from 71 infants at 2 weeks of age and again monthly from 1 through 7 9 months of age [28]. As shown in figure 1, these data indicate that the distribution of milk zinc concentration is skewed somewhat to higher values at each age, although the data can he normalized by logarithmic transformations.

FIG. 1.

Distribution of breastmilk zinc concentration according to time postpartum, and effects of different mathematical transformations of data. Data provided by Krebs et al, [28]

We also examined the results of seven previously completed studies of breastmilk intake to assess the age-specific distribution of milk volumes [28–34] and concluded that milk intake was generally normally distributed in all studies (Shapiro-Wilk statistic > 0.98 in all cases).

Studies of breastmilk zinc transfer to exclusively breastfed infants under 6 months of age

Identification and selection of studies

Data sets containing information on breastmilk zinc concentration were identified through a computerized database search (PubMed, accessed September 12, 2007) using the keywords “breast,” “milk,” “zinc,” and “infant” and no additional limits, which yielded 320 bibliographic citations. The titles were screened for relevance, and the full texts of the articles were obtained when possible. Additional data sets were located from the lists of citations included in the articles found during this initial search, through papers cited in a previously published review [35], and by contacting other experts in the field. Using this search strategy, we identified a total of 69 publications from 64 different studies that reported data on breastmilk zinc concentration.

From the set of 69 publications that were retrieved, we selected only those studies that presented data from mothers of healthy, term infants and provided information on both milk zinc concentration and time postpartum. Despite the conclusions of the methodologic review presented above, for the sake of caution we excluded studies that enrolled preterm infants or children with specific illnesses, unless data were reported separately for unaffected children. We also excluded studies of mothers who were receiving zinc supplements, unless the authors specifically stated that there was no effect of supplementation on milk zinc concentration or data from nonsupplemented women were presented separately. We also excluded studies of women who were receiving other mineral-containing supplements or other medicines. These exclusions resulted in a final set of 33 studies that were included in the combined analysis, as shown in table 1 [5, 7, 13–18, 21, 23, 26, 28, 36–61]. The studies that were excluded from the analysis are shown in table 2 [6, 8–11, 19, 22, 24, 25, 62–86], along with the reason(s) for the exclusion.

TABLE 1.

Publications included in the data analysis

Country, year [reference] author	Study group	Health and zinc status of mothers	Zinc concentration in milk according to stage of lactation (mg/L)	Description of milk sample	Growth or zinc status of infant
Argentina, 2001 [14] Ronayne de Ferrer	n = 20 mothers with term milk (mean gestational age of infants, 39.5 wk; range, 38–41 wk; mean birthweight, 3,303 g) and n = 24 middle- class mothers with preterm milk (mean gestational age of infants, 30.9 wk; range, 27–35 wk; mean birthweight, 1,440 g)	Apparently healthy	Mean ± SEM, term infants only; Days 2–5: 6.96 ± 0.69 Days 6–10: 4.27 ± 0.51 Days 11–15: 3.36 ±0.31 Days 16–30:2.44 ±0.11	Full breast sample collected by manual or pump expression between 10 and 12 am	Infant growth and zinc status not reported
Brazil, 1989a [41] Donangelo (excluded data from 31 to 280 days because of wide range of infant ages)	n =93 mothers of low SES with normal and uncomplicated gestations, both partially breastfeeding and exclusively breastfeeding. Mean maternal age, 25.9 ± 5.6 (SD) yr; mean birthweight, 3,277 ± 476 (SD) g	Mothers received iron, vitamin B12, and/or folate during pregnancy (but not lactation). 11% of mothers had serum zinc below cutoff for low zinc for adult women	Mean ± SEM, exclusively breastfeeding mothers: Days 1–5:5.94 ± 0.55 (n = 17) Days 6–30: 2.84 ± 0.37 (n = 13) Days 31–280: 1.65 ±0.38 (n = 10) Mean ± SEM, partially breastfeeding mothers: Days 31–280: 1.26 ± 0.19 (n = 21)	5–10 mL, collected by manual expression between 9 and 10 am before the infant was fed	Infant growth and zinc status not reported
Brazil, 1998 [15] Trugo	n = 28 mothers with term milk (mean birthweight, 3,380 ± 53 [SD] g) and n = 7 mothers with preterm milk (mean birthweight, 1,810 ± 47 [SD] g) at 1–5 days postpartum; n = 9 mothers with term milk (mean birthweight, 3,160 ± 29 [SD] g) and n = 15 mothers with preterm milk (mean birthweight, 1,700 ± 39 [SD] g at 6–36 days postpartum). All of low SES	All low SES; none received supplements. No information on maternal health and zinc status	Mean ± SD (range), term infants only: Days 1–5:7.26 ± 3.03 (1.3–13.7) (n = 22) Days 6–36:2.79 ± 1.14(1.2–4.2) (n = 8)	5–10 mL, of milk collected by manual expression between 9 and 10 am before the infant was due to be fed	Infant weights were adequate for gestational age. Infant zinc status not reported
Brazil, 2003 [21] da Costa	n = 50 mothers of term infants (37–41 wk gestation) > 20 yr of age, with at least 2 yr interparturition period; no supplements	Maternal health and zinc status not reported	Mean ± SD: Days 1–7: 6.97 ± 2.82 (n = 48)	Manual expression of 1.5 mL of colostrum from both breasts between the 1st and 7th days of lactation	Infant growth and zinc status not reported
Burundi, 1995 [53] Robberecht	n = 5 exclusively breastfeeding middle-class women (mean age, 32 yr) with infants aged 0–10 mo. Uncomplicated pregnancies	Healthy; zinc status not measured specifically	Mean ± SD: Days 2–4 (colostrum): 3.76 ± 0.51 (n = 1,8 samples) Days 5–14; 3.08 ± 0.62 (n = 1, 10 samples) 1 mo: 2.47 ± 1.17 (n = 3,12 samples) 2 mo: 1.64 ± 0.06 (n = 2,6 samples) 4 mo: 1.30 ± 0.06 (n = 2,5 samples) 10 mo: 0.75 ±0.02 (n = 2,6 samples)	Breastmilk samples from 5 lactating women [54]	Infant growth and zinc status not reported
Canada, 1982 [18] Mendelson	n = 10 mothers of term infants (mean gestational age, 39.0 wk; range, 38–40 wk; mean ± SD birthweight, 1,143 ± 305 g) and n = 14 mothers of preterm infants (mean gestational age, 28.7 wk; range, 26–33 wk; mean ± SD birthweight, 3,361 ± 550 g) All appropriate for gestational age	Normal pregnancy; maternal health and zinc status not reported	Mean ± SD, term infants only: Days 3–5: 5,35 ± 1.2 (n = 8) Days 8–10:4.1 ± 0.65 (n = 8) Days 15–17:3.37 ± 0.6 (n = 7) Days 28–30:2.6 ± 0.65 (n = 6)	Complete 24-h expression with breast pump	Infant growth and zinc status not reported
Canada, 1999 [20] Friel	n = 19 mothers 20–35 yr of age with full-term infants	Healthy, nonvegetarian; zinc not specifically measured	Mean ± SD: Days 2–3: 4.58 ± 0.52 (n = 17) Day 8: 3.99 ±0.88 (n = 16) Day 15: 3.54 ±0.74 (n = 16) Day 22: 3.37 ± 2.06 (n=15) Day 28: 2.5 ± 0.94 (n = 15) Day 36: 2.32 ± 1.41 (n = 15) Day 43:1.89 ± 0.65 (n = 15) Day 50:1.56 ± 0.96 (n = 15) Day 61:1.14 ± 0.55 (n = 12)	Full breast expression by electric pump during a regular feeding of the child between 10 am and 2 pm; a 15-mL aliquot was taken	Infant growth and zinc status not measured
China, 2007 [60] Xiang	n = 41 term, exclusively breastfeeding mother-infant pairs in rural China. Group I: n = 18 at 1 mo (mean maternal age, 26.6 yr). Group II; n = 23 at 3 mo (mean maternal age, 27.4 yr)	Mean maternal zinc intake, 7.5 mg/day; mothers apparently healthy	Mean ± SD: Group I (1 mo): 3.24 ± 0.23 Group II (3 mo): 1.94 ±0.17	5-mL milk samples were expressed manually at 5 time intervals (8–10 am. 10 am–12 pm, 12–2 pm, 2–4 pm, and 4–6 pm) after the breast had been suckled by the infant for 2 min	Weights and lengths were in normal ranges at birth, 1 mo, and 3 mo. Infant zinc status not reported
Finland, 1979 [57,58] Vuori	n = 27 mothers aged 20–35 yr (mean, 28 yr)	Healthy, well-nourished, all were primiparous mothers of term infants	Mean; Wk 2:4.0 (n = 20) Wk 3: 3.0 (n = 23) Wk4: 2.5 (n = 22) Wk 5: 2.4 (n = 25) Wk 6:2.1 (n = 17) Wk 7:1.9 (n =18) Wk 8–10:1.3 (n = 22) Wk 11–13: 1.1 (n = 18) Wk 14–16: 0.95 (n = 11) Wk 17–19: 0.78 (n = 12) Wk 20–22:0.75 (n = 13) Wk 23–25: 0.49 (n = 10) Wk 26–28: 0.52 (n = 9) Wk 29–31:0.44 (n = 6) Wk 32–34: 0.42 (n = 1) Wk 35–37: 0.48 (n = 2)	Fore- and hindmilk sample from each feeding (5–7/day in first 3 mo; 3–4/day thereafter) at 1–2 wk intervals up to 2 mo and 3–4-wk intervals thereafter	Infant growth (measured by height and weight) was comparable to reference norms for Finnish infants. Infant zinc status not reported
Germany, 1992 [55] Sievers	n = 10 healthy mothers aged 26–35 yr with term infants	Healthy; zinc not measured specifically	Median (10th, 90th percentile): Day 17: 3.6(2.4, 49) Day 35: 2.6 (1.6, 3.6) Day 56: 1.7 (1.1, 2.8) Day 85:1.3 (0.8, 1.4) Day 117:1.2 (0.6, 1.9)	At each feed 2 mL of fore- or hindmilk from nursed breast (n = 2,207 specimens)	Negative zinc balance observed in 11 of 33 breastfed infants; median daily zinc retention was positive in all cases. Infant growth not reported
Greece, 2005 [49] Leotsinidis	n = 180 mothers aged 25 ± 4 (SD) yr with full term infants, mean birthweight 3,396 ± 480 g	Healthy; zinc status not specifically measured. Anemia prevalence was 41.9%	Mean ± SD: Day 3:4.91 ± 1.73 Day 14:2.99 ± 0.92	10–20 mL collected in the morning by manual expression, 2 h after the previous breastfeeding	Infant growth and zinc status not reported
Honduras and Sweden, 2004 [40] Dömellof	n = 263 mother-infant pairs (total for Honduras + Sweden); mothers > 16 yr of age; exclusively breastfeeding at 4 mo. Infants were full term, with birthweight > 2,500 g, no illness	Maternal zinc status not reported. Prevalence of low ferritin (< 12 μg/h) was 32% in Honduras and 12% in Sweden	Mean ± SD at 9 mo: Honduras: 0.7 ± 0.18 (n = 108) Sweden: 0.46 ± 0.26 (n = 86)	10–40 mL collected by manual expression at 9 mo postpartum in the morning > 1 h after feeding	Infant growth and zinc status not reported
India, 1997 [16] Hemalatha (colostrum and transitional milk data not included because day of collection not reported)	Mothers of 68 preterm infants (gestational age at birth not given), 92 term breastfed infants, and 26 term formula-fed infants. All weaned with cereal at 6 mo	Maternal health and zinc status not reported	Mean ± SD, term infants only: Colostrum: 2.27 ±0.191 Transitional: 2.19 ± 0.297 Mo 1–3:1.90 ±0.20 Mo 4–6: 1.50 ± 0.225 Mo 7–12: 1.15 ±0.15	Full breast expression from the left breast between 10 and 11 am	Mean ± SEM plasma zinc (μg/dL): Cord: 118.3 ± 5.07 Mo 3:76.9 ± 8.85 Mo 6:74.2 ±5.12 Mo 9: 83.8 ± 7.84 Yr 1:114.6 ± 13.02 Infant growth not reported
Italy, 1999 [26] Chierici	n = 22 mothers of appropriate weight for gestational age term infants. Half were randomized to receive vitamin and mineral supplements	Healthy, nonsmokers, nonvegetarians; uncomplicated pregnancy, labor, and delivery	Mean ± SD, placebo group only (n = 11): Day 3: 8.16 ±2.96 Day 30: 3.9 ± 1.01 Day 90:2.87 ± 1.23	10 mL of milk was obtained from both breasts with a breast pump before the baby nursed	No difference in gains in infant weight, length, or head circumference between the 2 maternal study groups
Japan, 1982 [23] Higashi (colostrum data excluded because day of collection not specified)	n = 65 mothers 21–37 yr of age (mean, 27.3 yr) with healthy, full-term infants with birthweight > 2,500 g	Mothers judged to have good nutritional status based on clinical examination. Mean serum zinc ranged from 55 to 115 μg/dL	Mean ± SD: Colostrum: 10.39 ± 4.43 (n = 65) Wk 1:4.56 ±3.01 (n = 65) Mo 1: 2.66 ± 1.03 (n = 65) Mo 3: 1.14 ± 0.67 (n = 45) Mo 5: 1.05 ± 0.46 (n = 35)	10 mL of milk obtained by manual expression in the morning before the baby was due to be fed	Infant growth and zinc status not reported
Japan, 2005 [61] Yamawaki	n = 1,197 mothers under 40 yr of age (mean, 29.12 ± 3.99 [SD] yr) who did not smoke or use vitamin supplements, with infants with no symptoms of atopy and birthweight > 2,500 g (mean, 3,135 ± 346 [SD] g)	Zinc status not specifically measured	Mean ± SD: Days 1–5:4.75 ± 2.48 (n = 20) Days 6–10: 3.84 ± 1.39 (n = 38) Days 11–20: 3.37 ±0.89 (n = 40) Days 21–89: 1.77 ± 1.08 (n = 551) Days 90–180:0.67 ±0.8 (n = 476) Days 181–365:0.65 ± 0.43 (n = 39)	~50 mL of milk obtained at an intermediate time during suckling	Mean birthweight, 3,135 ±346 (SD) g. Infant zinc status not specifically reported
Nigeria, 1982 [36] Atinmo	n = 20 low-SES mothers with term infants	Maternal health and zinc status not reported	Mean ± SD: Days 2–4:5.98 ± 1.12 Days 8–14: 5.49 ±0.73 Mo 1–2:3.93 ±0.78	10 mL of milk collected by manual expression at noon just before infant feeding	Infant growth and zinc status not reported
Nigeria, 1984 [50] Mbofung	n = 240 low-SES mothers 17–40 yr of age who had given birth to a normal, full-term infant	Mothers judged healthy by clinical examination and medical history; zinc not specifically measured	Mean ± SEM; Days 1–7: 5.83 ± 0.13 (n = 96) Wk 4:3.2 ± 0.10 (n = 54) Wk 8: 3.0 ±0.10 (n = 34) Wk 12: 2.3 ±0.11 (n = 84) Wk 16: 1.88 + 0.13 (n = 40) Wk 20:1.08 ± 0.10 (n = 31) Wk 24:0.98 ± 0.08 (n = 21) Wk 32:0.82 ± 0.09 (n= 12) Wk 36: 0.68 ± 0.10 (n = 8)	8–12 mL of milk collected between 9 am and noon from the breast due for the next feeding by the mother using manual expression	Infant growth and zinc status not reported
Spain, 1997 [52] Ortega	n = 57 lactating women aged 18–35 yr with healthy infants with normal birthweight and a mean gestational age of 39.4 wk	Comparison of healthy mothers with low zinc intake (< 50% RI; mean serum zinc, 78.5 ± 9.8 μg/dL) and high zinc intake (≥50% RI; mean serum zinc, 87.0 ± 17 μg/dL) (p< .05)	Mean ± SD, mothers with low zinc intake: Days 13–14:3.05 ± 0.48 Day 40:1.88 ±0.41 Mean ± SD, mothers with high zinc intake: Days 13–14: 3.34 ±0.60 Day 40: 2.16 ±0.52 8% of mothers with low zinc intake and none with high zinc intake had < 0.75 mg/L breastmilk zinc	5-mL samples collected from each breast pre- and post-feeding between 10 and 11 am	Normal growth No difference in infant growth between groups
Spain, 2001 [56] Silvestre	n = 22 women, well-nourished and healthy with term infants	Well nourished; zinc not measured specifically	Mean ± SD: Days 2–4: 7.99 ± 3.23 Day 14:3.31 ± 1.06 Day 30: 2.41 ±0.90 Day 60: 1.40 ±0.65 Day 90: 1.05 ±0.71	Foremilk collected from both breasts prior to feeding; collection between 11 am and 4 pm	Healthy; infant growth and zinc status not reported
Turkey, 2005 [13] Ustundag	n = 20 mothers (mean age, 21 ± 1 [SD] yr) of term infants and n = 20 mothers (mean age, 21 ± 2 [SD] yr) of preterm infants (< 37 wk gestational age); all mothers were nonsmokers	Healthy mothers; weight ranged from 50 to 65 kg. Zinc not measured specifically	Mean ± SD, mothers of term infants only: Days 0–7:2.41 ± 0.28 Days 7–14: 2.28 ±0.19 Day 21:2.39 ±0.20 Mo 2:2.01 ±0.18	Milk collected by manual expression within 2 h of the first feeding in the morning between 8 and 11 am	Infant growth and zinc status not reported
UK, 1982 [44] Hibberd	n = 10 mothers of term infants	Healthy; zinc not specifically measured	Mean ± SD: Day 2:7.75 ± 1.33 Day 4:4.51 ± 1.32 Day 6:3.90 ± 1.23 Day 9:3.32 ± 1.27 Day 16: 2.81 ±1.16 Day 23: 2.33 ± 1.31 Day 30: 1.71 ± 1.15 Day 37: 1.71 ± 0.92	Complete expression from both breasts at each feeding with the use of an electric breast pump; an aliquot was taken for analysis and samples from each 24-h period were pooled	Infant growth and zinc status not reported
USA, 1979a [59] Vaughan	n = 38 Caucasian women aged 19–42 yr, mothers of term infants; 1–31 mo postpartum (average participation was 4 mo); 22 primiparous and 16 multiparous	“Good or excellent” nutritional status Maternal zinc intake (mg/day): Mo 4–6:12.7 Mo 7–9:17.3 Mo 10–12: 16.5 Mo 13–18: 16.1 Mo 19–31: 13.1 Serum zinc concentration (μg/dL): Mo 4–6: 230 Mo 7–9: 220 Mo 10–12: 140 Mo 13–18: 150 Mo 19–31:220	Mean ± SD: Mo 1–3:1.60 ± 0.23 (n = 28) Mo 4–6:1.05 ±0.15 (n = 39) Mo 7–9:0.75 ± 0.11 (n = 23) Mo 10–12: 0.63 ± 0.09 (n = 13) Mo 13–18: 0.69 ± 0.18 (n = 28) Mo 19–31: 0.60 ± 0.19 (n = 30) Total: 15.0 ± 4.2	150–200 mL of milk collected once per month at 4-wk intervals. Mothers instructed to collect milk over a 3–5-day period in the morning, afternoon, and evening at random intervals within feeding	Infant growth and zinc status not reported
USA, 1983a [7] Feeley	n = 102 mothers aged 16 to 38 yr with healthy full-term infants; no difficulties during delivery. Middle SES; 97 Caucasian, 3 African-American, 2 Asian-American	n = 39 (38%) took self-prescribed zinc supplements at a dose of 15 mg/day; n = 2 (2%) took self-prescribed zinc supplements at a dose of 25 mg/day No association between use of zinc supplements and milk zinc concentration	Mean ± SD: Days 4–7: 5.36 ± 0.02 (n = 91) Days 10–14:4.22 ± 0.01 (n = 163) Days 30–45: 2.99 ± 0.01 (n = 158) Means for each stage of lactation are significantly different (p< .05)	30 mL (1/3 foremilk, 1/3 halfway through feeding, 1/3 hindmilk) milk sample obtained early morning and late evening by the mother with the use of manual expression or pump	Infant birthweight ranged from 2.5 to 4.8 kg with a mean of 3.5 ± 0.5 (SD) kg. Infant zinc status not reported
USA, 1983b [43] Garza	Longitudinal study of n = 6 mothers 26–35 yr of age with healthy term infants with appropriate weight for gestational age; and exclusively breastfeeding until beginning of study (5–7 mo postpartum)	Healthy, nonsmoking, limited coffee, tea, and alcohol, no medications. Zinc not specifically measured	Mean ± SEM: Mo 6:1.2 ±0.2 (n = 5) Mo 6.5: 1.2 ±0.4 (n=6) Mo 7:1.2 ±0.3 (n = 6) Mo 7.5:0.8 ± 0.3 (n = 6) Mo 8:0.7 ± 0.3 (n = 5) Mo 8.5:0.8 ± 0.3 (n = 6) Mo 9:0.7 ±0.1 (n = 4)	Entire contents of a single breast obtained with an Egnell pump 3 to 4 h after the previous feeding, between 8 am and noon	Infant growth and zinc status not reported
USA, 1983c [51] Moser	n = 23 mothers of term infants recruited from private obstetrics clinic. Mean age was 30 ± 2 (SEM) yr; mean education was 15 ± 2 (SEM) yr	Mean ± SEM maternal zinc intake (mg/day): Wk 37 gestation: 9.8 ± 0.6 (n = 23) Mo 1 postpartum: 9.4 ± 0.5 (n = 23) Mo 3 postpartum: 12.8 ±1.8 (n = 20) Mo 6 postpartum: 9.6 ±0.7 (n=19) Mean ± SEM maternal plasma zinc (μg/dL): Wk 37 gestation: 63.8 ± 2.0 (n = 23) Mo 1 postpartum: 79.1 ±1.7 (n = 23) Mo 3 postpartum: 87.6 ± 2.2 (n = 21) Mo 6 postpartum: 84.4 ± 2.4 (n = 19)	Mean ± SEM: Mo 1:2.6 ±0.2 (n = 21) Mo 3:1.3 ±0.1 (n = 20) Mo 6: 1,1 ±0,1 (n = 18)	30 mL collected at the first morning feeding after 6 am containing ½ foremilk and ½ hind milk	Infant growth and zinc status not reported
USA, 1984a [17] Butte	Longitudinal study of n = 8 preterm infants (gestational age, 30–36 wk; mean, 33.9 wk; birthweight, 1,920 ± 70 [SD] g) and n = 13 term infants (gestational age, 37–42 wk; mean, 39.2 wk; birthweight, 2,990 ± 46 [SD] g)	Mothers of term infants 20–35 yr of age who did not consume > 2 servings/day of coffee, tea, alcohol, or carbonated beverages; parity ≤ 2, no routine medications. Maternal health was judged to be good by health history	Mean ± SD, term infants only: Wk 2:3.4 ± 0.8 Wk 4: 2.9 ± 0.9 Wk 6:2.1 ±0.9 Wk 8: 1.9 ±0.6 Wk 10:1.8 ± 1.0 Wk 12: 1.4 ±0.7	Full breast expression with breast pump between 8 and 12 am, at least 2 hours after feeding	Mean birthweight, 2.99 ± 0.46 (SD) kg. Infant zinc status not reported
USA, 1985a [38] Casey	n = 11 mothers aged 26–39 yr. White, middle-class, well nourished. Mean parity was 2.5 (range,2–4)	Apparently well-nourished, with uncomplicated pregnancy; zinc not measured specifically	Mean ± SD: Day 1:7.94 ± 5.65 (n = 7) Day 2:11.5 ±4.7 (n = 8) Day 3; 9.34 ± 1.97 (n = 9) Day 4:6.40 ± 0.66 (n = 9) Day 5:5.42± 1.11 (n = 8) Day 8 ± 2:4.74 ± 1.02 (n =10) Day 14 ± 3:3.88 ± 0.91 (n =10) Day 21 ± 3:3.71 ± 1.09 (n =9) Day 28 ± 3:2.98 ± 0.78 (n = 8)	5 mL collected from both breasts in the middle of the midmorning feeding (2 min after letdown)	Mean birthweight, 3.33 ± 0.46 (SD) kg (range, 2.70–4.05); mean weight at 1 mo, 3.95 ± 0.33 (SD) kg. Infant zinc status not reported
USA, 1985b [5, 45] Krebs (only data from placebo group included)	Middle-income, healthy lactating women, longitudinal trial, zinc supplement (ZS): n = 14, placebo (NZS): n = 39. Mean infant age was 3.4 mo	Healthy; some took zinc supplements during pregnancy [47]. NZS: mean dietary zinc intake 10.7 ±4.1 mg/day	Mean ± SD: Placebo group: Mo 1:2.65 ±0.81 Mo 9: 0.67 ± 0.40 Zinc supplement group: Mo 1: 2.83 ± 1.05 Mo 9: 0.82 ± 0.54 (at mo 9,8 subjects remained in placebo and 4 in zinc supplement group)	5 mL collected by manual expression. Pooled samples from 6 collections/day	Infant growth and zinc status not reported
USA, 1987 [37] Butte	n = 45 mothers of healthy, full-term, exclusively breastfed infants aged 0–4 mo	Healthy; zinc not measured specifically	Mean ± SD: Mo 1: 2.37 ± 0.82 Mo 2: 1.55 ± 0.62 Mo 3: 1.13 ±0.52 Mo 4: 1.03 ± 0.52 Interindividual variability (CV),0.33 Intraindividual variability (CV), 0.42	24-h alternate breast pooled sample: full breastmilk samples (entire content of 1 breast), at each feeding over 24-h period	Infant growth progressed satisfactorily. Infant length positively associated with zinc intake at mo 1 and 2
USA, 1989 [39] Casey	n = 13 well-educated mothers (mean age, 31.9 ± 4.4 [SD] yr), median income > US$35,000, with normal prepregnancy weight and weight gain during pregnancy. Uneventful pregnancy	Healthy mothers, nonsmokers, all multiparous. Dietary assessment: mean zinc intake, 10.9 ± 0.5 (SD) mg/day; 5/13 took supplemental zinc (additional 10–45 mg/day); no effect of supplement reported	Mean breastmilk zinc peaked at 2 days, then declined throughout lactation; decline was nonlinear and more rapid in 1st 2 wk. Higher zinc concentrations at 3, 6, 9, and 12 mo were not associated with supplemental zinc intake Mean ± SD (n = 13): Day 7: 4.70 ± 1.20 Mo 1:2.90 ±0.70 Mo 12:0.46 ± 0.30	2–5 mL obtained by manual expression from both breasts at midfeeding in the morning	Infant growth and zinc status not reported
USA, 1995 [28] Krebs (pooled data from maternal supplementation trial)	Healthy, lactating women. Prospective, randomized, double-blind, controlled trial; maternal zinc supplementation group (15 mg/day) (n = 40), placebo group (n = 31)	Healthy (plasma zinc was 79.8 μg/dL in both the zinc-supplementation and the non-zinc supplementation groups, compared with 88.3 μg/dL in the nonlactating control group) Randomized, controlled trial. Maternal zinc supplementation, 15 mg/day. No difference in milk zinc concentration between groups; thus, data were pooled	Mean ± SD (n = 71): Mo 0.5: 3.88 ± 0.99 Mo 1: 3.06 ± 1.12 Mo 2: 2.04 ± 0.83 Mo 3: 1.48 ± 0.65 Mo 4: 1.44 ±0.66 Mo 5: 1.16 ±0.62 Mo 6: 1.09 ± 0.67 Mo 7: 0.85 + 0.51 Mo 8: 0.87 ± 0.58 Mo 9: 0.78 ± 0.51	Milk collection: 1st 29 subjects, 5–10 mL hand- expressed midfeeding from both breasts during 3-day period. 2nd 42 subjects, 3 midfeeding samples/day for 3 days, with > 4 h between samples	Healthy infants. Mean weight-for-age percentile was 62 at 2 mo, 33 at 7 mo, and 25 at 9 mo of age. Mean length-for-age percentile was 43 at 2 wk, 28 at 7 mo, and 26 at 9 mo of age [48]
USA, 1996 [46] Krebs	n = 9 mothers (individually identified as; A–I) of normal, term, exclusively breastfed infants aged 2–5 mo	Healthy; zinc not measured specifically	Mean ± SD for each infant: Mo 2.0: A (1.37 ± 0.2), B (1.96 ± 0.26), Mo 2.3: C (1.01 ±0.28), D (1.96 ±0.25) Mo 4.0: E (2.76 ± 0.3),F (1.69 ± 0.32), Mo 4.5: G (1.40 ±0.21) Mo 4.8: H (1.33 ±0.15) Mo 5.0:1 (1.08 ± 0.14)	5–10 mL expressed at each feeding over 3 days	Mean weight change during 1-wk study period was 20 ± 7.9 g/day. Mean total zinc absorbed was 0.61 ± 0.23 mg/day

RI, recommended intake; SES, socioeconomic status

TABLE 2.

Publications excluded from the data analysis

Country, year [reference] author	Study group	Health and zinc status of mothers	Description of milk sample	Zinc concentration in milk according to stage of lactation (mg/L)	Growth or zinc status of infant	Rationale for exclusion
Bangladesh, 1990 [82] Simmer	n = 34 lactating mothers in rural area	No information on maternal health or zinc status	All milk extracted from both breasts over 24-h period using mechanical pump; 24-h infant test-weighings performed on 2 days	Mean ± SD: Mo: 1.96 + 0.99 Mo 2–3:1.25 ±0.54 Mo 6:0.93 ±0.41 Mo 9:0.73 ±0.61 Mo 12:0.54 + 0.27	No information on infant growth or zinc status	No information on preterm births
Bangladesh, 1996 [9] Hussain	n = 34 low-SES mothers in Dhaka in wk 6–36 of lactation. Mean age, 22 yr	2/3 of mothers had BMI <20	5 mL collected from alternate breasts by manual expression	Mean ± SD: 5–7 am: 2.10 ±0.83 2–3 pm: 1.74 ±0.53 10 pm: 1.84 ±0.69	No information on infant growth or zinc status	Data not presented by stage of lactation
Brazil, 1985 [66,67] Dorea	n = 8 mothers Living in a low-income community. Mean age, 22.5 yr	General decrease in WHZ over 0–6 mo lactation but remained > 90% WHZ reference population	~5 ml, collected at beginning and end of 1 nursing period at the same time of day	Breastmilk zinc concentration reported as a figure [67]	Infants grew normally during the study period. Breastmilk zinc was a significant predictor of weight gain (p = .0006) and linear growth (p = .02)	Numerical results not reported
Brazil, 1989b [71,72] Lehti	n = 25 Amazonian mothers	Generally inadequate nutritional status	Foremilk of the fullest breast collected by manual expression between 8;30 and 11:30 am	Mean ± SD: Mo 0–1: 2.2 ± 0.9 (n = 37) Mo 1–2:1.6 ± 0.9 (n = 59) Mo 2–3:1.4 ± 0.9 (n = 41)	No information on infant growth or zinc status	No information on preterm births
Brazil, 2007 [73] Melnikov	n = 117 mothers (mean age, 22.1 yr) and their infants. None of the infants showed malformations or clinically detectable impairment	Mothers were apparently healthy. Breastmilk zinc was not related to maternal age, parity, or history of miscarriage	Manual expression on day 2 postpartum	1.22 ± 0.78 mg% [sic]	No information on infant growth or zinc status	No information on preterm births
China, 2002 [81] Sian	n = 18 mothers (mean age, 26 ± 2 [SD] yr) of infants ~2 mo of age (mean age, 48±11 [SD] days)	Mean maternal zinc intake estimated by 3-day diet records, 7.8 ± 1.7 mg/day (n = 18)	5 mL collected by manual expression of midfeeding sample	At ~2 mo postpartum, mean breastmilk zinc concentration was 2.34 ± 0.83 mg/L. Milk zinc output was 2.01 ± 0.97 mg/day	No information on infant growth or zinc status	No information on preterm births
Côte d’Ivoire, [70] Lauber	n = 33 mothers in a rural area; 8 were primiparous. 19 infants were aged 1–3 mo and 10 infants were aged 4–15 mo	General health was good and nutritional status was fairly satisfactory	Aliquot taken from full breast expression by electric pump in early morning, from the breast not used to feed the infant the previous night	Mean ± SD: Mo 1:3.5 ±0.9 (n=4) Mo 6:2.3 ± 0,9 (n = 7) Mo 12:1.6 ± 0.9 (n = 8) Mo 18: 1.5 ± 1.3 (n = 6)	Weight -for-age declined from mo 5 onward and leveled off at mo 10 Height-for-age remained at 94%–96% of Harvard standards; practically no change in weight-for-height	No information on preterm births
Croatia, 1996 [24] Frković	n = 29 mothers of wide SES range; mean age, 28.9 yr (range, 17–45 yr); mean parity, 1.8	No information on maternal health or zinc status	80 mL collected	Mean (range): Days 2–12: 4.98 (1.69–11.60)	Mean cord blood zinc was 118 ± 21 μg/dL	No information on preterm births
Ethiopia, 2003 [83] Umeta	Rural Ethiopian mothers with infants aged 5–11 mo; n = 253 (83%) agreed to give breastmilk sample	Mothers of stunted and nonstunted children	~15-mL sample collected from right breast, ~60 min after last feeding	Age-adjusted breastmilk zinc (5–11 mo postpartum) of mothers of stunted infants (mean ± SEM): 0.60 ± 0.03 mg/L (n = 92) vs. 0.68 ± 0.02 mg/L (n = 161) in mothers of nonstunted infants (p = .02)	Infants of mothers with low breastmilk zinc were more stunted	No information on preterm births and wide infant age ranges
Finland, 1980 [84] Vuori	n = 28 mothers aged 24–35 yr (mean age, 28 yr). All mothers belonged to 2 highest (out of 3) SES groups. Mean height, 164 cm; mean weight, 54.4 kg prepregnancy, 59.1 kg after delivery, 57.6 kg at 1st survey week (6–8 wk postpartum), 55.5 kg at 2nd survey week (17–22 wk postpartum)	Lowest individual zinc intake was 9.3 mg/day. No correlation between zinc levels in diet and milk. Mean ± SD zinc intake (range): 13.7 ± 2.7 (10.0–19.4) mg/day at 1st survey week, 12.8 ± 2.8 (9.3–18.6) mg/day at 2nd survey week	8-mL aliquots at beginning and end of each feeding during a 24-h period and pooled to one sample	Mean ± SD: 1st survey week: 1.89 ± 0.74 2nd survey week: 0.72 ± 0.44 (p< .001)	No information on infant growth or zinc status	No information on preterm births
Finland, 1994 [25] Satmenperä	n – 200 exclusively breastfeeding mother-infant pairs (n = 116 exclusively breastfed to 6 mo; n = 36 exclusively breastfed to 7–9 mo). Mothers supplemented with 20 mg/day or 40 mg/day zinc or placebo	Healthy, nonsmoking mothers with uncomplicated pregnancy and delivery; full-term healthy infants of appropriate weight for gestational age	Milk was manually expressed; 10 mL of milk was pooled from the beginning and end of each feeding	Median (range) for placebo group (received iron): Days 4–5:4.75 (3.27–6.9) (n = 75) Mo 2:1.41 (1.1–2.19) (n = 77) Mo 4:0.9 (0.58–1.38) (n = 67) Mo 6: 0.67 (0.4–1.13) (n = 56) Mo 7.5: 0.61 (0.39–0.97) (n = 31) Mo 9: 0.6(0.38–0.95) (n = 14) Mo 10: 0.61 (0.42–0.87) (n = 8) Mo 11:0.43(0.33–0.57) (n =6) Mo 12:0.43(0.33–0.56) (n = 5)	No significant difference in serum zinc between exclusively breastfed infants and infants fed complementary foods at 7.5–9 mo. Milk zinc concentration greater in mothers given 40 mg/day zinc than in those given 20 mg/day zinc at 6 and 7.5 mo (p = 02)	All women received supplemental micronutrients
Guatemala, 2005 [65] Dhonukshe-Rutten	n = 56 women with healthy, exclusively breastfed infants aged 1–6 mo in periurban Guatemala City and low-income rural area	Zinc not specifically measured; 17/42 had intestinal parasites	Full expression with hand pump plus manual expression > 1 h after last feeding from breast not used in last feeding	Mean ± SD: Mo 1: 3.85 ± 0.76 (n = 4) Mo 2: 2.38 ± 0.53 (n = 14) Mo 3–6:1.53 ±0.64 (n = 29)	No information on infant growth or zinc status	Women with parasites had higher breastmilk zinc than non-infected counterparts: 2.30 ±1.10 (SD) vs. 1.80 ±0.82 mg/L
India, 1980 [6] Rajaiakshmi	Cross-sectional study: n = 412 low- income urban mothers, n = 100 high-income urban mothers, n = 208 low-income rural mothers Semilongitudinal study: n = 24 low-income mothers	High-income women had greater height, weight, MUAC, and triceps skinfold thickness than low-income women Mean hemoglobin, 13.0 ±0.15 (SEM) g/dL in low-income group and 12.9 ±0.26 (SEM) g/dL in high-income group	10 mL collected by manual expression just prior to infant feeding. Day-to-day variation assessed by daily collection for 5–7 days. Diurnal variation assessed by collections at 10 am and 6 pm in 6 women	Breastmilk zinc values according to stage of lactation were available from cross-sectional and semilongitudinal studies, as well as data on diurnal and day to-day variation in breastmilk zinc	No information on infant growth or zinc status	No information on preterm births
India, 1999 [19] Sharda	Mothers of 155 infants of gestational age 26–41 wk and birthweight 550–3,800 g Appropriate for gestational age and small for gestational age not defined; time of collection of milk samples not reported	Maternal serum zinc concentration ranged from 57.5 to 76.5 μg/dL Serum zinc concentration in nonpregnant women was 99.4 ± 14.4 (SD) μg/dL	Milk samples (colostrum) obtained by mechanical expression from both breasts	Mean ± SD: Gestational age 37–41 wk: 5.03 ± 0.078 (n = 5) Term, appropriate weight for gestational age: 5.04 ± 0.072 (n = 3) Term, small for gestational age: 4.98 ± 0.098 (n = 2) Gestational age 26–30 wk:4.75 ± 0.078 (n = 3) Gestational age 31–33 wk: 4.71 ±0.098 (n = 11) Gestational age 34–36 wk: 4.75 ± 0.10 (n= 11) Preterm, appropriate weight for gestational age: 4.72 ±0.12 (n = 17) Preterm, small for gestational age: 4.75 ±0.072 (n = 8)	Breastmilk zinc lower in preterm than in term infants and in low-birthweight than in normal-birthweight infants No difference in milk zinc concentration between mothers of term infants with different birthweights and mothers of preterm infants with different birthweights Mean ± SD milk zinc: Birthweight 0.5–1.5 kg: 4.75 ±0.085 mg/L (n = 12) 1.5–2.5 kg: 4.77 ± 1.64 mg/L (n = 14) 2.5–4.0 kg: 5,00 ±0.14 mg/L (n = 4)	Time of milk collection postpartum not reported
Indonesia, 1998 [68] Gross	n = 92 middle-income mothers in a poor urban area (East Jakarta), 25.4 ± 5.2 (SD) yr of age Mean infant age, 2.4 ± 1.4 (SD) mo (range, 0.1–5.2 mo)	Mean BMI, 22.0; 40% anemic (hemoglobin < 120 g/L); mean plasma zinc, 85.5 ± 24.2 (SD) μg/dL (increased with time postpartum); 29% had low plasma zinc concentrations	Entire milk content from the breast last suckled was collected with a manual pump during 5 consecutive days between 9 and 11 am	Median: 2.7 10th, 90th percentiles: 1.3,4.7 Range: 0.5–12.8 (n = 91)	Mean HAZ.WHZ, WAZ near 0 and distribution similar to NCHS reference population	Data from wide infant age range presented; values for milk zinc concentration according to infant age not presented numerically
Italy, 1993 [22,78] Perrone	n = 26 mothers aged 17–37 yr with term infants and 6 mothers with preterm infants. No supplementation during pregnancy or lactation	Apparent good health	Full breast manual expression at the 2nd nursing, 9–11 am	Median ± SD mg zinc/kg dry weight of breastmilk, term infants: Wk 1:36.4 ±2.8 (n = 46) Wk2:24.2 ± 1.6 (n = 15) Wk 3: 28.6 ± 6.8 (n = 19) Wk4:21.7± 1.4 (n = 59)	Healthy; zinc status not specifically measured	Data presented as pooled wide range of infant ages; zinc measured as mg per dry weight
Japan, 2003 [69] Honda	n = 68 Japanese mothers aged 19–38 yr	No information on maternal health or zinc status	Milk samples collected at 5–8 days postpartum Details of milk collection not specified	Mean ± SD, 5–8 days postpartum: Mother > 35 yr: 5.41 ± 1.44 Mother < 35 yr: 5.90 ± 1.83 Nullipara: 6.27 ± 1.92 Multipara: 5.35 ± 1.5 Cesarean delivery: 5.8 ± 1.58 No Cesarean delivery:5.83 ± 1.84	No information on infant growth or zinc status	No information on preterm births
Kenya, 2002 [77] Onyango	Mothers of 250 toddlers (mean age, 13.9 mo; range, 12–26 mo) from rural, low-income families	Not supplemented; zinc not specifically measured	Details of milk collection not specified	Breastmilk zinc content not directly measured	No information on infant growth or zinc status	Breastmilk composition taken from another study in The Gambia; no information on preterm births
Kuwait, 2000 [62] Al-Awadi	n = 34 mid-upper-class, lactating mothers of term infants living in Kuwait. Maternal age range was 25–40 yr	Healthy; none had plasma zinc below reference values	Manual expression between 6 and 8 am before the infant’s I st feeding	Mean ± SD: Mo 0–6: Kuwaiti: 3.2 ±0.12 Non-Kuwaiti: 2.4 ± 0.06 Mo 6–12: Kuwaiti: 2.4 ±0.14 Non-Kuwaiti: 1.9 ± 0.05 Mo 12–18: Kuwaiti: 2.0 ±015 Non-Kuwaiti: 1.7 ± 0.09	No information on infant growth or zinc status	Pooled data presented from wide range of infant ages
Nepal and USA, 1988 [74] Moser	n = 26 Nepalese mothers 2–6 mo postpartum, and n = 23 US mothers 3–6 mo postpartum	Mean ± SD plasma zinc (μg/dL): Nepal: 66.7 ± 2.0 USA: 88.3 ± 2.0 (p< .05) No significant difference in dietary zinc intake	Early morning, 5–10-mL prefeeding sample	Mean ± SD: Nepal: 1.10 ± 0.098 USA: 1.20 ±0.098	Mean ± SD infant plasma zinc (μg/dL): Nepal: 68.0 ± 5.9 USA: not determined	Wide age ranges and results not disaggregated by age
Nigeria, 2000 [76] Okolo	n = 15 mothers 26.2 ± 3.3 (SD) yr of age and infants 6.12 ± 0.26 (SD) mo of age	Healthy mothers; mean serum zinc, 117 μg/dL	Breastmilk collected at 6 mo of lactation; no information on collection method	Mean at ~mo 6:1.52	Mean serum zinc concentration was 164 μg/dL. Infant growth not reported	No information on preterm births
The Gambia and UK, 1990 [63] Bates	n = 56 mothers (mean age, 29.1 yr) in Gambia and n = 57 mothers (mean age, 31.8 yr) in UK; stage of lactation varied from 0 to 99 wk postpartum in Gambia and from 0 to 112 wk postpartum in UK	Maternal health and zinc status not reported	1–5 mL expressed between feedings	Mean: The Gambia Mo 1:4.22 Mo 2: 3.50 Mo 3: 2.78 Mo 4: 1.71 Mo 5: 1.69 Mo 6:2.15 UK Mo 2: 1.34 Mo 3: 2.06 Mo 4:0.87 Mo 5: 0.73 Mo 6; 0.73	No information on infant growth or zinc status	Mothers received micronutrient supplements; no information on preterm births
Turkey, 2006 [79] Rakicioğlu	n = 21 mothers, 27.3 ± 5.4 (SD) yr, nonsmokers, not taking supplements; infants aged 2–5 mo	Healthy; no chronic diseases	Manual full breast expression after the first nursing period of the day between 9 and 11 am	Mean ± SD: 1.5 ± 0.4 before Ramadan, 1.8 ± 0.3 after Ramadan (p = .001)	Increases in the body length and weight of infants were within the normal NCHS range	No information on preterm births
UK, 1993 [80] Richmond	Mothers of 39 normal, healthy children 17–61 wk of age	No information on maternal zinc status	No information on milk sample collection	Infant zinc intake from milk (geometric mean ± SD): 1.06 ± 2.10 mg/day	No information on infant growth	Zinc concentration of milk not presented. No information on preterm births
USA, 1976 [10] Picciano	Mothers in 6th to 12th wk of lactation (n = 50) with healthy, full-term infants. 42 mothers were 20–30 yr old and 8 were > 30 yr. 16 were primiparous and 34 were multiparous	37 of the 50 women were taking supplemental vitamins. “A few supplements contained less than 10% RDA for zinc”	1 40-mL sample/day at early morning feeding for 5 consecutive days (daily period). 2 additional samples at weekly intervals (weekly period) or within 1 day (within- day period)	Mean ± SD: Daily period: 1.68 ± 0.78 Weekly period: 1.59 ± 0.84 Within-day period: 1.58 ±0.81 Mean: Time of day: morning 1.71, midday 1.54, evening 1.49 (morning value was significantly and evening values) Age: 20–30 yr, 1.55; > 30 yr, 2.00 (p < .001) Parity: 0,0.39; ≥ 1.1.75 (p< .001) Lactation history: 0,1.46; ≥1,1.75 (p<.001)	Healthy; zinc status not specifically measured	Data not presented by stage of lactation; wide range of infant ages
USA, 1979b [8] Kirksey	n – 52 Caucasian, middle-SES mothers aged 18–31 yr. 33 were patients at an obstetrical clinic at enrollment; 19 were members of La Leche League (stages of lactation in these women ranged from 1 to 30 mo)	Healthy; zinc intake exceeded 2/3 of RDA (includes supplements); zinc intake from diet was less than 2/3 RDA for 30 women	33 obstetrical patients: 5–10 mL collected after milk letdown at 1st morning feeding on days 3 and 14 postpartum and prior to multivitamin supplementation 19 La Leche League patients: similar to obstetrical patients, except milk collected on 3 consecutive days	Mean breast milk zinc concentration decreased from 4.72 mg/L on day 3 to 3.20 mg/L on day 14. Correlation between zinc and iron in colostrum: r = 0.730. Levels of zinc in milk were similar for supplemented subjects (mean zinc intake, 28 mg) and unsupptemented subjects (mean zinc intake, 11 mg)	No information on infant growth or zinc status	Labels missing from data table
USA, 1982 [86] Zimmerman	n = 34 mothers recruited at hospital (7 preterm infants, 27 full-term infants)	Healthy mothers	5–15-mL samples collected after letdown and before feeding, at 1,15, and 30 wk (single samples)	Mean: Wk 1:3.50 Wk 15: 0.850 Wk 30; 0.550 No difference between breastmilk zinc of mothers of preterm and full-term infants	No information on infant growth or zinc status	Sample included 7 preterm infants
USA, 1984b [64] Dewey	n = 46 mothers during mo 7–20 of lactation aged 21–37 yr	Apparently healthy; zinc status not specifically measured	Full breast expression in the morning at the 2nd feeding of the day, either manual or with an Egnell pump	Mean ± SD: Mo 4–6:0.79 ± 0.47 (n = 38) Mo 7–11:0.42 ±0.22 (n = 27) Mo 12–20:0.33 ±0.26 (n = 4)	Weight-for-age was at or above the 25th percentile of NCHS standards for 6 of 8 infants at 7 mo and 5 of 6 infants at 11 mo. 4 of 6 infants had decreased gradually in weight for age since mo 7	No information on preterm births
USA, 1985c [85] Finley	n = 62 mothers with a mean age of 29 yr	Half of the mothers were vegetarian and half were non-vegetarian. Zinc status not specifically measured.	Manual expression of 1 breast in the morning at the 2nd feeding of the day	Data for milk zinc concentration according to month postpartum are presented as a figure	No information on infant growth or zinc status	No information on preterm births
USA, 1988 [11] Karra	n = 49 middle-income mothers; mean age, 28.8 yr; mean parity, 1.7	Apparently healthy and adequately nourished; zinc not specifically measured	Milk samples collected at every infant feeding over a 24-h period at 1,2,3,4,5, and 6 mo postpartum. 10 mL of milk was expressed manually or by breast pump	Data for milk zinc concentration according to month postpartum are presented as a figure	No information on infant growth or zinc status	All women received micronutrient supplements; numerical results not reported
USA, 1990 [75] Moser-Veillon	n = 20 lactating women aged 20–36 yr. Doubleblind, randomized, controlled trial. Multivitamins (n = 10) or multivitamins + 25 mg zinc (n = 10). Supplementation from 1 day to 36 wk postpartum	Healthy; mean plasma zinc values ranged from 77.2 to 96.1 μg/dL	Manual expression of half- foremilk, half-hindmilk, in the morning	Mean ± SEM, placebo group: Wk 1; 4.62 ±0.48 Wk2: 3.63 ±0.33 Wk 4; 2.73 ±0.18 Wk 12: 1.72 ±0.12 Wk 24:1.37 ±0.22 Wk 36:0.60 ± 0.22	No information on infant growth or zinc status	All mothers received a multivitamin supplement

BMI, body mass index; HAZ, height-for-age z-score; MUAC, mid-upper-arm circumference; NCHS, US National Center for Health Statistics; RDA, recommended daily allowance; SES, socioeconomic status; WAZ, weight-for-age z-score; WHZ, weight-for-height z-score

Milk zinc concentration according to infant age

To express the data on milk zinc concentration from different studies in consistent units, we applied conversion factors of 65.4 g of zinc per mole and 1.03 g of breastmilk per milliliter [35], and all results are presented as milligrams of zinc per liter of milk. The data available from all 33 acceptable studies are summarized in figure 2 and table 3. Each point in the figure represents the mean milk zinc concentration reported for the midpoint of a particular age interval in a given study. The data across studies are remarkably consistent, showing a relatively high concentration of zinc in colostrum and early transitional milk, a rapid fall in milk zinc concentration during the first 1 to 2 months postpartum, and a slower decline thereafter. Tests of various transformations of breastmilk zinc concentration and age indicated that a log-log relationship including a quadratic term for log(age) provided the best fit for the relationship between milk zinc concentration and time postpartum (r² = 0.88.p < .0001).

FIG. 2.

Breastmilk zinc concentration according to time postpartum, all studies. Ln(Zn concentration) = 0.975 −0.501*ln(age) − 0.063* ln (age)²

TABLE 3.

Breastmilk zinc concentration according to child’s age^a

Age group (mo)	No. of observation	Mean ± SD milk zinc concentration (mg/L)
< 1	74	4.11 ± 1.50
1 to 2	42	1.91 ±0.53
3 to 5	24	0.98 ± 0.35
6 to 11	24	0.77 ± 0.22
12 to 21	2	0.64 ± 0.05

^aData were extracted from studies presented in table 1. Figures for milk zinc concentration represent weighted means of results from all studies that provided information for respective age ranges.

Amount of milk consumed according to infant age

Information on the amount of milk consumed by children of different ages was extracted from the publication “Complementary feeding of young children in developing countries” by the World Health Organization (WHO) and the United Nations Children’s Fund (UNICEF) [87]. This document provided separate information for exclusively breastfed infants less than 6 months of age and for non-exclusively breastfed children less than 24 months of age, as shown in table 4.

TABLE 4.

Amount of breastmilk transferred from mother to child (g/day) according to child’s age and breastfeeding pattern^a

Breastfeeding pattern	Age group (mo)
	0–2	3–5	6–8	9–11	12–13
Exclusive breastfeeding	714 ± 131	784 ± 128	–	–	–
Partial breastfeeding	644 ± 159	706 ± 146	674 ± 151	616 ± 172	549± 187

^aData from Brown el al. [87], Values represent mean ± SD for children from developing countries.

Total zinc transfer to exclusively breastfed infants less than 6 months of age and relation to zinc requirements

To estimate the total amount of zinc transferred in breastmilk, we completed the following simulation exercise. Breastmilk zinc concentration was assumed to follow a log normal distribution, based on the information described above [28]. Because only sample means and standard deviations were available from the individual studies used to estimate milk zinc concentrations, the means and standard deviations of log(breastmilk zinc concentration) were estimated from the untransformed summary data by using the method of moments [88]. The relationships between infant age and the means and standard deviations of log(breastmilk zinc concentration) and of breastmilk intake were estimated with regression analysis. Finally, we generated 1.000,000 simulated data points, assuming that age was uniformly distributed between 0 and 6 months for exclusively breastfed children and between 0 and 18 months for partially breastfed children, and that log(breastmilk zinc concentration) + log(breastmilk zinc concentration)- and breastmilk volume were normally and independently distributed at each age. We then multiplied individual simulated values for breastmilk intake and milk zinc concentration to estimate total zinc transfer.

The results of the simulations of the total daily amount of zinc transferred in breastmilk according to child age are shown in figure 3 for exclusively breastfed infants less than 6 months of age. The mean amount of zinc transferred in breastmilk declines from ~4 mg/day during the first few days of life to ~1.75 mg/day by 1 month. Breastmilk zinc transfer declines further, albeit more slowly, to ~0.7 mg/day by 6 months. The prediction interval displayed in the figure indicates the ranges of values for daily milk zinc intakes that are likely to occur for children at a particular age. Because the available data do not permit adjustment for day-to-day (intraindividual) variability in milk zinc transfer, the portrayed range overestimates the distribution of children’s usual zinc intakes. Thus, the curves can be used to describe the ranges of daily zinc intakes, but they cannot be used reliably to assess the adequacy of usual zinc intake from breastmilk.

FIG. 3.

Simulated mean and 95% prediction interval of daily zinc transfer in breastmilk to exclusively breastfed infants according to infant age

Adequacy of breastmilk zinc transfer in relation to the zinc requirements of exclusively breastfed term infants less than 6 months of age

There are two possible theoretical approaches to assess the adequacy of breastmilk as a source of zinc for exclusively breastfed term infants: to compare usual zinc intake and assumed absorption from breastmilk with the physiological requirements for absorbed zinc, or to test for any functional response after providing supplemental zinc to such infants. The former approach is particularly challenging, for several reasons. First, not only is information needed on the usual total amount of breastmilk zinc consumed at each age, but knowledge is also required regarding the absorption of zinc from this food source. Furthermore, the true physiological requirements for absorbed zinc at each age must be known with some degree of certainty. Although information can be compiled on zinc consumption from breastmilk at different ages (as described above), there is less information on individuals’ usual zinc intakes or on zinc absorption from breastmilk. Moreover, because the FAZ varies according to the amount of zinc consumed in a particular meal [89, 90], and the amount of milk (and hence zinc) consumed at a single feeding may be extremely variable, the available information on mean FAZ from breastmilk should be considered a rough approximation, at best, of the true value for an individual child.

Two studies have been published on infant FAZ from breastmilk. One study, which reported results from nine exclusively breastfed infants 2 to 5 months of age who received zinc stable isotope tracers along with six feedings per day, found a mean FAZ of 0.54 (range, 0.40 to 0.68) [46]. Another study of 10 non-exclusively breastfed infants 5 to 7 months of age reported a mean FAZ of 0.50 (range, 0.27 to 0.78) when they were fed a total of 250 mL of expressed breastmilk, along with other foods, over the course of three feedings during an 8- to 12-hour study period [91]. Breastmilk zinc was highly bioavailable under the circumstances of these studies, although the actual study conditions were not necessarily consistent with usual breastfeeding practices, so the ultimate implications of these results remain uncertain. In developing countries, the mean frequency of breastfeeding is often greater than 12 times per day [1,30,32], so the amount consumed per feeding may be less than that provided in the foregoing studies. In that case, the zinc intakes per feeding would be less, and the FAZ might be even greater than was observed. Because of the broad range of FAZ reported from these studies and the uncertainty regarding the true FAZ from different amounts of breastmilk fed during real-life feeding conditions, we applied assumptions for FAZ values ranging from 0.40 to 0.60 to explore the range of zinc that might be absorbed from breastmilk by infants of different ages, according to these assumptions. 1 he results of these exploratory calculations are shown in table 5.

TABLE 5.

Estimated daily zinc intake among exclusively breastfed infants less than 6 months of age and estimated total absorbed zinc at three assumed levels of fractional zinc absorption (EAZ), and relation to estimated physiological requirements for zinc, accordine to infant’s ase

Age group (mo)	Milk zinc transfer (mg/day)	Amount of absorbed zinc (mg/day), according to assumed FAZ			Estimated average physiological requirements for zinc (mg/day)a
		FAZ = 0.40	FAZ = 0.50	FAZ = 0.60
< 1	2.52	1.01	1.26	1.51	1.3
1–2	1.37	0.55	0.68	0.82	1.3
3–5	0.86	0.34	0.43	0.52	0.7

^aRequirements from WHO and IZiNCG [92,93]

Several expert groups have published estimates of the physiological requirements for absorbed zinc. A WHO committee calculated the amount of zinc that must be absorbed each day to replace endogenous losses, based on data extrapolated from adults, and to allow for infant growth, based on estimates of zinc accretion in newly synthesized tissues of growing infants [92]. The WHO committee estimated that the requirements for absorbed zinc derived by this method are 1.3 mg/day during the first 3 months of life and 0.7 mg/day for the period from 3 to 5 months. The Steering Committee of the International Zinc Nutrition Consultative Group (IZiNCG) accepted the WHO recommendations hut further noted that these requirements may he partially offset by hepatic zinc reserves accumulated during gestation [93].

According to the comparisons in table 5, breastmilk alone would provide the WHO/IZiNGG estimated mean requirements [92,93] during the first month of life (without considering the possibility of drawing on hepatic zinc reserves) if the FAZ is greater than 0.50 but would provide less than the mean requirement thereafter. Nevertheless, IZiNCG concluded that breastmilk alone is an adequate source of zinc for approximately 6 months, based on the assumption that some reserves present at birth could be used to meet physiological demands and the fact that supplementation trials have not demonstrated consistent benefits of additional zinc supplementation for exclusively breastfed term infants, as discussed below.

Zinc supplementation trials among breastfed infants less than 6 months of age

Three intervention trials in which breastfed infants less than 6 months of age received either zinc supplements or placebo have been reported from industrialized countries. However, in two of these trials, the period of supplementation lasted until the infants attained 10 to 11 months of age, so the results do not specifically reflect the effect of providing additional zinc to infants less than 6 months of age. In the one trial completed among term infants less than 6 months of age, zinc-supplemented US girls gained 3 g/day more weight than their nonsupplemented counterparts from 4 to 7 months of age, hut there were no significant effects on the linear growth of girls or on the weight or length gains of boys [94]. In another study of US term infants 4 to 10 months of age, there was no effect of zinc supplementation on growth or other functional outcomes among either girls or hoys [95]. Finally, in a study conducted in France among infants 3 to 11 months old, most of whom were African immigrants, the zinc-supplemented hoys had significantly greater weight and length gains after 3 months of supplementation than hoys in the placebo group, but there were no significant effects of supplementation among girls [96]. However, these latter findings are difficult to interpret, because differential dropout rates by treatment group may have confounded the results. Moreover, the infants in this trial were receiving nonqualified amounts of other foods, so it is uncertain whether the groups differed only with regard to zinc supplementation. In summary, the results of these three studies carried out in industrialized countries are inconsistent, and none clearly demonstrates a uniformly positive impact of additional zinc on the growth of infants under 6 months of age. Thus, it seems that breastmilk alone is probably an adequate source of zinc for approximately the first 6 months of life for normal-birthweight, term infants in industrialized settings. We could not locate any relevant information from developing countries that was obtained specifically from exclusively breastfed infants.

Section 2

How much zinc is transferred in breastmilk to infants and young children less than 24 months of age who are consuming breastmilk in addition to complementary foods?

Conclusions

The amount of zinc transferred In breastmilk to partially breastfed infants less than 6 months of age is approximately 15% less than that described above for exclusively breastfed infants, because of the smaller volumes of milk consumed by partially breastfed infants. The age-related pattern of change in milk zinc intakes is similar for both groups of infants. After 6 months of age, breastmilk continues to be an important source of zinc, providing ~0,5 mg/day from 6 to 8 months and ~0.4 to 0.3 mg/day thereafter, a time when the estimated physiological requirements for absorbed zinc range from 0.84 to 0.5 mg/day, depending on age and the source of these estimates.

Detailed review of evidence

Studies of breastmilk zinc transfer to partially breastfed infants less than 24 months of age

Identification of studies and milk zinc concentration by child age

The same process that was described above was used to identify and analyze studies of breastmilk zinc transfer to partially breastfed infants and young children, and the same summary values for milk zinc concentration were used for the specific age intervals, as shown in figure 2.

Amount of milk consumed according to child age

As with the preceding analysis, we extracted information on breastmilk volumes from the WHO/UNICEF publication on complementary feeding [87]. For the present analysis, we selected the age-specific information for partially breastfed infants and young children from developing countries, as summarized in table 4.

Total zinc transfer in breastmilk to partially breastfed children less than 24 months, and relation to requirements

The amount of zinc transferred in breastmilk to partially breastfed children was estimated using the same simulation procedure described above. The analysis had to be limited to children less than 18 months of age because there was only one data point available for milk zinc concentration from children older than 18 months, and the available information on breastmilk volumes provided a single figure for the entire age range from 12 to 23 months. Because the volume of milk consumed by partially breastfed infants less than 6 months is approximately 15% less than the volume consumed by exclusively breastfed infants, the total milk zinc transfer to the former infants is correspondingly less, although the age-related pattern of change is similar, as summarized in figure 4 and table 6. Even after 9 months of age, breastmilk continues to provide an average of ~0.4 mg/day to partially breastfed infants.

FIG. 4.

Simulated mean and 95% prediction interval of daily zinc transfer in breastmilk to partially breastfed children according to age

TABLE 6.

Estimated daily zinc intake among partially breastfed children less than 18 months of age and relation to estimated average physiological requirements for zinc, according to child’s age

Age group (mo)	Milk zinc transfer (mg/day)	Estimated average physiological requirement for zinc (mg/day)a
		WHO and IZiNCG	IOM, 2000
< 1	2.18 ± 1.06	1.3	–
1–2	1.18 ±0.56	1.3	–
3–5	0.73 ± 0.39	0.7	–
6–8	0.51 ±0.33	0.8	0.84
9–11	0.39 ±0.31	0.8	0.84
12–17	0.29 ±0.31	0.8/0.5	0.74

^aWHO and IZiNCG provided similar estimates for children < 12 months of age, although IZiNCG assumed that breastmilk was adequate for exclusively breastfed infants < 6 months of age, so the estimates in the table refer to non-exclusively breastfed infants [92,93]. At 12 to 17 months of age, the two estimates provided by WHO and IZiNCG differ. The US Institute of Medicine (IOM) does not provide estimates of physiological requirements for zinc for children < 7 months of age [97].

Assessment of the adequacy of breastmilk zinc transfer in relation to theoretical requirements is even more problematic for partially breastfed children than for those who are exclusively breastfed. Because partially breastfed children are consuming other foods in addition to breastmilk, these other foods may contribute additional zinc and they may affect zinc absorption from breastmilk [98]. Thus, we have not attempted to assess the adequacy of breastmilk zinc transfer at different ages, and we simply present the mean amount of milk zinc that is transferred to demonstrate the potential contribution that breastmilk might provide in relation to these requirements. Clearly, breastmilk provides a potentially important proportion of the daily zinc requirement, even among partially breastfed children in the second year of life.

Section 3

What are the programmatic implications of the answers to the foregoing questions, and what are the remaining research needs?

The foregoing analyses indicate that breastmilk seems to provide an adequate amount of zinc to exclusively breastfed term infants less than 6 months of age. Breastmilk continues to be an important source of zinc for partially breastfed infants well into the second year of life. Thus, breastfeeding should be promoted and supported as an intervention that contributes to the prevention of zinc deficiency. WHO currently recommends that “infants should be exclusively breastfed during the first 6 months of life. Thereafter they should receive nutritionally adequate and safe complementary foods while breastfeeding continues up to 2 years of age and beyond” [99]. Thus, these infant and young child feeding recommendations are consistent with the current conclusion that breastfeeding should be promoted as an important dietary source of highly bioavailable zinc.

Information is available on how best to promote and support appropriate breastfeeding behaviors [100], and a detailed discussion of these issues is beyond the scope of the present review. Briefly, attention must be devoted to advocacy, policy formulation, health curriculum reform and in-service training of health service personnel, and the development of suitable community-based organizations to provide information, counseling, personal support, and referral when necessary [101].

Additional research is needed on the zinc requirements of infants and young children and the bioavailability of zinc from breastmilk when fed in usual amounts, either alone or with other foods. This information is needed to help determine both the period for which breastmilk alone is an adequate source of zinc and how much additional zinc is needed from complementary foods. Information is also needed on the period of adequacy of breastmilk for providing zinc to low-birthweight infants. Finally, different approaches for promoting optimal breastfeeding behaviors require further evaluation in the context of large-scale programmatic interventions.