- Email: [email protected]
Growth and intakes of energy and zinc in infants fed human milk
Growth and intakes of energy and zinc in infants fed human milk N a n c y F. Krebs, MD, MS, Carol J. Reidinger, MS, Alastair D. Robertson, PhD, a n d K. M i c h a e l H a m b i d g e , MD, ScD From the Department of Pediatrics, Pediatric Nutrition Center, and the Department of Medicine, Universityof Colorado Health Sciences Center, Denver The objectives of this study were to e x a m i n e the growth pattern of healthy infants who were fed human milk exclusively for >_5 months and its relationship to intakes of milk, energy, and zinc. Monthly anthropometric measurements were obtained on 71 infants through 7 months of a g e and on 43 through 9 months. Milk zinc concentrations were determined and milk intake was measured by 3-day test weighing. Mean (_+SD) c a l c u l a t e d energy intake from human milk was 106 _+ 20, 79 _+ 11, 70 _ 10, and 57 __. 12 k c a l / k g per d a y at 2 weeks and 3, 5, and 7 months of age, respectively; intake from milk plus solids was 70 _+ 13 k c a l / k g at 7 months. Zinc intakes from human milkwere 2.3 ___0.68, 1.0 _+ 0.43, 0.81 _+ 0.42, and 0.52 +_ 0.31 m g / d a y at these points. The maximum mean weight-for-age percentile for both sexes was 62 at 2 months; the mean percentile d e c l i n e d to 33 by 7 months and to 25 by 9 months of age. Mean length-for-age percentile d e c l i n e d from 43 at 2 weeks of a g e to 28 and 26 by 7 and 9 months of age, respectively. Energy intake at 2 weeks of a g e was positively associated with the weight increment from 2 weeks to 7 months of a g e (p = 0.003) and with a change in weight-for-age z scores from 2 weeks to 3 months of a g e (p = 0.028). Mean energy intakes of healthy breast-fed infants are lower through the first 7 months of life than current recommendations. Zinc intakes of breast-fed infants are likely to be adequate, on average, through 5 months of breast-feeding but subsequently are marginal without the introduction of weaning foods. Revision of standards for growth based on observations in breast-fed infants m a y be appropriate, but the possibility of growth-limiting nutrient deficiencies should also be investigated. (J PEDIATR1994;124:32-9)
Several studies have suggested a pattern of growth in infants fed human milk that differs from standards based primarily on the growth of formula-fed infants. 15 Infants who have Supported by grant 5-R01-DK1243 from the National Institutes of Health, grant RR00069 from the Clinical Research Center, grant T86-00279-023 from the PEW Nutrition Fellowship, and grant 5-T32-DK07658 (Training grant) from the National Institutes of Health. Submitted for publication March 18, 1993; accepted Aug. 11, 1993. Reprint requests: Nancy F. Krebs, MD, MS, University of Colorado Health Sciences Center, 4200 East Ninth Ave., Box C225, Denver, CO 80262. Copyright © 1994 by Mosby-Year Book, Inc. 0022-3476/94/$1.00 + .10 9/20/50758
32
been fed human milk for several months and whose growth is faltering relative to the National Center for Health Statistics (NCHS) standards do not necessarily increase their energy intake when given ready access to weaning foods. 6 This may be "normal" for the breast-fed infant, but an alternative hypothesis is that an inadequate supply of a single nutrient in mature human milk, such as zinc, may limit both appetite and growth. Mild zinc deficiency in infants and young children can depress growth, 7 and limited data indicate that food intake can also be depressed. 8 The results of an earlier pilot study suggested that subtle differences in the normal rate of decline in the zinc concentration of human milk as lactation progresses may result from suboptimal maternal dietary zinc intake.9 It appeared that those differences, though small, could result in a zinc intake by the
The Journal o f Pediatrics Volume 124, Number 1
fully breast-fed infant that was inadequate to sustain normal growth velocity, l° One of the objectives of this study was to determine the growth pattern of normal infants who were human milk exclusively for at least 5 months and to investigate the relationship of growth to milk intake. An additional objective was to determine whether infant growth is related to zinc intake from human milk.
METHODS Study design. We conducted a longitudinal study of infants who were to be fed human milk exclusively for at least 5 months. Subjects were the offspring of mothers who were enrolled before 2 weeks of age and who were randomly assigned to receive a daily 15 mg zinc supplement or placebo throughout lactation. Supplementation began at 2 weeks of age; data were collected then, at 1 month of age, and at monthly (_+ 1 week) intervals thereafter through at least 7 months of age. No significant effect of maternal zinc supplementation on milk zinc concentrations or milk output was found. 11 Infants will thus be considered as one group in this article. For the aspects of the study reported herein, data and sample collection consisted of infant anthropometry, milk samples for analysis of zinc concentrations, 3-day test weighing of infants for estimation of milk intake at 4 time points, and 3-day diet records concurrent with the test weighing once weaning foods were introduced. Subjects. We studied 40 female and 31 male infants through 7 months of age; 43 infants (23 girls) continued in the study through 9 months of age. All were healthy term infants who were of appropriate weight for gestational age. The mothers' mean age was 30.4 + 3.7 years; mean parity was 1.8 + 1.0, with a range of 1 to 5; all but one were white; socioeconomic status was middle to upper middle income; and education level was _> 12 years for 93% of the mothers (75% were college graduates). None of the mothers smoked, drank alcohol more than occasionally, or were known to abuse drugs. At the time of recruitment, mothers were enrolled only if they expressed the intention to breast-feed exclusively through 5 months and continue to nurse through 7 months. Informed written consent was obtained before data collection. The protocol was approved by the human subjects committee at the University of Colorado Health Sciences Center. Anthropometry. Infants' weights and lengths were obtained at 2 weeks and 1 month of age, and monthly thereafter. Naked weights were obtained on an electronic digital balance, accurate to 1 gm. Lengths were measured with the infant held in full extension on a portable measuring board with increments in millimeters. Head circumferences were measured at each visit with a laminated tape. All measurements were performed in duplicate by the same investiga-
Krebs et aL
33
tor (C.J.R.), who was assisted by the mother for the length measurements. Birth length data are not presented because of the variable techniques used in newborn nurseries to obtain these measurements. Milk and dietary intake. Human milk intake was measured by the test-weighing technique for 3 consecutive days at 2 weeks and at 3, 5, and 7 months of age. Electronic digital balances (Sartorius Corp., Bohemia, N.Y.), which integrated 100 rapid serial measurements to provide a mean weight to the nearest 1 gin, were set up in the subject's home t~: the test weighing. The mothers were given detailed instructions for the test-weighing procedure, including directions for proper use of the balances, and for use of consistent clothing and diapers for weights before and after feedings. Detailed log sheets were kept by the mothers for the 3 days to record weights. These were reviewed by a member of the research team at the end of each test-weighing period to identify any incomplete or erroneous entries. Research personnel were available by telephone to the mothers during the test-weighing periods for consultation if mechanical or logistical problems arose. We calculated 24hour intakes by finding the total intake for the 3-day period, dividing by the exact number of hours, and multiplying by 24. Intakes were not adjusted for insensible water losses. Participating mothers were lent electric breast pumps (Medela, Inc., Crystal Lake, Ill., and Ameda/Egnell Inc., Cary, Ill.) to allow provision of expressed human milk by bottle. Infant formulas were used only in small amounts in mixtures with dry cereals or other solid foods. Once solids had been introduced, 3-day diet records of measured food intake were kept for the infants concurrently with the test weighing. Nutrient analyses were performed with a computer program (Practorcare, Inc., San Diego, Calif.) that includes commonly used baby foods and has been modified in our laboratory to provide complete data for zinc content of foods by laboratory analysis of individual foods, manufacturers' data, or imputation from similar foods. Energy content of human milk was assumed to be 0.67 kcal/gm. Milk samples were obtained from each breast midway through the feeding on each side for every nursing session during the test-weighing periods for the first 29 subjects completing the study. Comparison of milk zinc concentrations by analysis of variance revealed no significant diurnal variation and no significant differences between sides (unpublished data). On the basis of these and previously published data, 12 the remaining 42 subjects collected three samples per day during the test-weighing periods. A single sample was obtained at the monthly visits without testweighing data. Samples of 5 to 10 ml were expressed by hand into zinc-free plastic vials and were frozen at - 2 0 ° C until analysis. Zinc concentration was determined on each sample by laboratory procedures described previously.9
34
Krebs et al.
The Journal of Pediatrics January 1994
Table I. Daily intakes of human milk and zinc from human milk, and calculated intakes of energy and zinc from weaning foods Age
Human milk In gm/day In gm/kg per day Zinc from human milk In mg/day Intake from weaning foods Emergy In kcal In kcal/kg Zinc In mg/day
2 Wk
3 Mo
5 Mo
7 Mo
600 +- 120 160 _+ 30
690 -+ 110 120 + 16
720 _+ 130 105 _+ 15
640 +_ 150 85 _+ 17
2.30 +_ 0.68
1.00 -+ 0.43
0.81 _+ 0.42 (16)*
0.52 _+ 0.31 (66)*
44_+ 21 6_ 3
109 + 74 15 + 10
0.17 + 0.14
0.47 _+ 0.43
Values (exceptthose in parentheses) are expressed as mean _+SD. *Number of subjects (male and female infants combined).
Blood tests. Samples were obtained by finger stick for spun hematocrits and plasma zinc concentrations at 6 months of age. Care was taken to clean the skin thoroughly before lancing and to avoid zinc contamination of the samples during collection. Plasma zinc concentrations were determined by a previously described method. 13 Data management and statistical analysis. The z scores and the percentiles for anthropometric data were calculated relative to the N C H S growth curves 14 by the CDC Anthropometric Software Package (Division of Nutrition, Center for Health Promotion and Education, Centers for Disease Control and Prevention, Atlanta, Ga.) Statistical analyses, including t test and stepwise linear regression, were performed with the Statistical Analysis Systems software program (1985; SAS Institute Inc., Cary, N.C.). Plotting of the longitudinal data was used extensively as a guide to statistical analysis. Percentiles and z scores for length, weight, and head circumference relative to age at the time of measurement were calculated and means were determined for each monthly cycle. The monthly weight and length increments were determined by dividing increments by the number of days between measurements and multiplying by 30. We also fitted anthropometric data, as functions of age from 2 weeks to 9 months, to a second-degree curve, smoothing the data and allowing the use of interpolated values to replace missing values. The z scores were then recalculated at the exact monthly time points. Separately for each interval (2 weeks to 3 months, 3 to 5 months, 5 to 7 months, 7 to 9 months), changes in fitted z score indexes over the interval were regressed on (1) subject attributes (sex and birth weight) and (2) dietary data at the start of the interval (energy [per kilogram] and zinc intakes, from human milk and supplemental food combined, and propor-
tion of energy and zinc intakes from human milk). In these regressions the method of forward selection and backward elimination was used. The significance level for entry and exit was p = 0.15. A specific effect was assessed after control for all other significant effects. RESULTS
Milk, energy, and zinc intakes. Results of test-weighing measurements are given in Table I. At each cycle, mean intakes of the male infants were 5% to 6% greater than those of the female infants. When intakes were compared relative to weight, however, they were virtually the same. Sixteen infants (23%) began intake of solids at 41/2 months. These solids contributed an average of 9% _+ 5% of energy for these infants at the 5-month cycle. At the sixthmonth visit, 55% of subjects had begun to consume weaning foods. By age 7 months, 66 infants (93%) were taking solids, providing an average of 15 + 11 kcals/kg per day, or 19% of total mean daily intake. The total energy intake from human milk and weaning foods at 7 months of age was 70 _+ 13 kcal/kg per day. Intakes of zinc from human milk declined sharply during the course of lactation (Table I). There were no significant differences between male and female infants. The addition of supplemental foods at 5 and 7 months of age contributed 5% _+ 11% and 40% _+ 26%, respectively, to the total mean zinc intakes, which were 0.85 _+ 0.42 mg/day and 1.0 + 0.5 rag/day at the two ages, respectively. Anthropometry. The mean birth weight was 3.38 +_ 0.45 kg. For the first 3 months the mean weight was greater than the median, with a peak at the 62nd percentile at 2 months of age, followed by a gradual decline to the 33rd percentile at 7 months of age (Fig. 1). The mean length percentile started at the 43rd percentile, increased slightly by 1 month
The Journal of Pediatrics Volume 124, Number 1
Krebs et al.
35
9o ~, 8o Z (D
70
~_ 6o
r~
o 40 l.a_ !
5o Ld
2O
•
Female
A
Male
"
------A-
1 10 I
0
1
J
[
2
,
I
~
3
r
4
,
I
5
,
I
6
~
I
7
,
I
8
,
I
9
AGE (MO)
Fig. t. Mean (_+SEM) weight-for-age percentiles versus age, by sex; solid lines indicate subjects followed through 7 months of age (40 female infants, 31 male infants); dashed lines indicate subjects followed through 9 months of age (23 female infants, 20 male infants). of age, and then gradually declined to the 28th percentile by 7 months (Fig. 2). At 7 months of age the boys were somewhat leaner than the girls, with mean weight-forlength percentile at the 37th and 60th percentiles, respectively. Regression analysis of the longitudinal data of actual z scores indicated significant differences between the sexes in weight-for-age percentiles from 2 weeks to 7 months and from 2 weeks to 9 months of age (p = 0.028 and 0.019, respectively). Significant interactions between age and sex were found in length-for-age z scores from 2 weeks to 9 months of age (p = 0.015) and for weight for length from 2 weeks to 7 months of age and from 2 weeks to 9 months of age (p = 0.004 and 0.003). The monthly increments were generally not different between male and female infants (Table II). The length increments from 2 weeks to 7 months of age for male and female infants were 15.6 cm and 14.1 cm, respectively (p = 0.001). The mean weight increment during this interVal was 3.71 _+ 0.88 kg and did not differ between the sexes. Growth and intake. Stepwise regression was performed foi the entire group for the dependent variables of changes in length-for-age, weight-for-age, and weight-for-length z scores during the intervals between the intake measurements. For the interval from 2 weeks to 3 months of age only, a significant negative association of birth weight with changes in Z scores for all three anthropometric measurements was observed (p _< 0.0002 for each change). Gender was also a significant variable at the 2-week to 3-month interval; male gender was associated with more negative
change in the weight-for-length z score (p = 0.003). Of the dietary variables examined, only the total energy per kilogram at 2 weeks of age was positively associated with the 2-week to 3-month change in the weight-for-age z score (p = 0.028) and in the weight-for-length z score (p = 0.0003); this association represents milk intake per kilogram at the 2-week test weighing. The weight and length increments during the interval from 2 weeks to 7 months of age were also examined as dependent variables in stepwise regression. For the length increment, male gender was positively associated (p = 0.0001) and birth weight negatively associated (p = 0.004). For the weight increment, energy intake per kilogram at 2 weeks of age was again positively associated (p = 0.003). Stepwise regression was also performed on subjects in the lowest quartile of the distributions of length-for-age, weightfor-age, weight-for21ength, and head-circumference z scores, with the sameindependent and dependent variables used for the same intervals noted above. There were no significant associations with the changes in length-for-age z scores at any of the intervals. Changes in the weight-for-age z scores at the 3- to 5-month and the 5- to 7-month intervals were positively associated with energy intake per kilogram at 3 and 5 months of age, respectively (p = 0.004 and 0.024). Changes in the weight-for-age z scores for the 5- to 7-month interval were negatively associated with the percentage of energy (p = 0.006) and positively associated with the percentage of zinc (p = 0.003) obtained from human milk at 5 months of age. The change in weight-forlength z score at 7 to 9 months of age was positively asso-
36
Krebs et al.
The Journal of Pediatrics January 1994
90
_~ 8o F-
~_ 60
.~
5o
~ 5o w 20
•
Male
10 I 1
,
I
,
2
,
~
3
4
, 5
I 6
,
I
~
I
7
8
t
I 9
AGE (MO)
Fig. 2. Mean ( ± SEM) length-for-age percentiles versus age, by sex; solid lines indicate subjects followed through 7 months of age (40 female infants, 31 male infants); dashed lines indicate subjects followed through 9 months of age (23 female infants, 20 male infants). Table II. M o n t h l y increments in weight a n d length for infants fed h u m a n milk, from b i r t h to 9 months of age Female Interval
(mo)
Weight (kg/mo) 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 Length (cm/mo) 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9
infants
n
Mean
40 40 39 39 39 39 23 23
0.80 0.66 0.53 0.42 0.36 0.31 0.27 0.18
40 40 39 39 40 39 23 23
2.66 2.81 2.19 1.55 1.68 1.23 1.33 1.41
Male
infants
± SD
n
Mean
± SD
± ± ± ± ± ± + ±
0.25 0.25 0.19 0.22 0.17 0.19 0121 0.21
31 31 31 31 29 29 20 19
0.89 0.63 0.58 0.43 0.36 0.30 0.33 0.28
± 0.33 ± 0.16" ± 0.20 + 0.17 ± 0.22 _+ 0.16 ± 0.18 ± 0.19
_ ± ± ± ± ± ± ±
0.82 0.86 0.83 0.89 0.95 0.82 0.58 0.75
30 31 31 31 29 29 19 19
3.22 2.77 2.49 1.93 1.73 1.50 1.28 1.32
± 0.99 ___0.99 _+ 0.85 ± 0.95 ± 0.87 _+ 0.83 ± 0.70 ± 0.63
*Significantly less than female increment (p = 0.008).
ciated with the percentage of zinc intake from h u m a n milk at 7 m o n t h s (p = 0.020). T h e changes in head-circumference z score for the 7- to 9-month interval were also positively associated with the percentage of zinc obtained from h u m a n milk at 7 m o n t h s of age (p -- 0.020). Biochemical data. P l a s m a zinc concentrations were det e r m i n e d at 6 months of age for 54 of the subjects; the m e a n was 11.9 +_ 2.6 # m o l / L (78 +_ 1 7 /~g/dl). A significant
negative association was found by stepwise regression between plasma zinc concentration and c h a n g e in weightfor-length z score between 5 and 7 m o n t h s of age for all subjects (p = 0.033) and for the subjects in the lowest quartile for this index (p = 0.011). Trends for the same relationship were found for the c h a n g e in length-for-age z score for the entire group and those in the lowest quartile (p = 0.07). Stepwise regression on the absolute i n c r e m e n t s
The Journal of Pediatrics Volume 124, Number 1
indicated a trend for the weight increment to be negatively associated with the plasma zinc concentration (p = 0.10). The mean hematocrit, obtained for 58 infants, was 0.38 ___0.3. There was no significant correlation between calculated iron intake from weaning foods and hematocrit. DISCUSSION The test-weighing method, although potentially disruptive to the mother-infant pair, has been shown to compare well with other techniques, including isotope dilution techniques, and is considered the "gold standard." The 3-day test-weighing period chosen for this study was thought to be a reasonable period to compensate for intraindividual daily variation while minimizing changes in the feeding pattern.15, 16 The mean milk intakes for this group of infants through the first 7 months of life fell at the lower end of the range reported in other studies. 5, 17, 18 A correction for insensible water losses was not attempted; the actual milk intake of the group could be expected to be higher, possibly as much as 5%. 15 Because of the difficulty of achieving representative sampling for bomb calorimetry, we used a standard conversion factor to calculate the energy intake of the infants. 19 At all points in the study the estimated energy intake relative to infant weight was less than current recommendations 2°, 21 and less than intakes reported for formula-fed infants. 22-24 The relatively rapid weight gain in the early months, despite lower-than-recommended energy intakes, suggests that energy requirements for the young infant fed human milk are lower than current recommendations. Investigating possible differences in energy requirements of human milk-fed and formula-fed infants, Butte et al. 23 found that the differences in energy intake were not totally accounted for by differences in energy expenditure or rate of growth, and may be related to energy digestibility and the composition of newly synthesized tissues. Weaning foods were consumed by nearly all the infants by the 7-month visit. At this time, the weight-adjusted energy intake was virtually the same as the average intake at 5 months of age. The weaning foods thus were apparently substituting calories and nutrients for those previously supplied by human milk, as also noted by Stuff and Nichols. 6 This observation suggests that a limitation of maternal milk supply was not directly governing energy intake. The stage of lactation is the most critical determinant for the zinc concentration of human milk, which declined by a factor of 5 from 0.5 to 9 months. 11 There was a corresponding decline in the calculated infant intakes of zinc from human milk during the course of the study. With the addition of weaning foods, the zinc intake at 7 months of age was essentially doubled to a mean of 1 mg/day. The larger percentage of total zinc intake supplied at 7 months by supple-
Krebs et al.
37
mental foods compared with that for energy, 40% versus 19%, reflects the relatively low concentration of zinc in human milk at this stage of lactation. The fractional absorption of the zinc contributed by solids would likely be lower than that from the milk, and the endogenous zinc losses in the feces may be higher. The bioavailability of the zinc in the milk could also be altered by the intake of other foods, presumably depending on the choice of solids and on such factors as fiber, phytate, protein, and iron, which may influence zinc bioavailability. In infants recovering from malnutrition, low plasma zinc concentrations were associated with accelerated catch-up growth. 25 Our findings suggest a similar relationship over a range of normal growth rates. To the extent that plasma zinc concentrations may reflect the size of relatively rapidly exchangeable pools of zinc available for metabolic processes, 26 the lower levels suggest a partial depletion of those pools to meet the demands of new tissue synthesis. After the third month, the gradual decline in mean weight-for-age percentiles during the next several months to approximately the 30th percentile (combined sexes) at 7 months of age is consistent with other reports. 1-4 The similar trend for the group followed through 9 months of age suggests that the progressive deviation from the N C H S median continues through at least 9 months. The monthly growth increments observed in this study are similar to those for breast-fed infants reported on by Heinig et al., 24 though the actual weights and lengths were greater in the latter cohort. Comparison of the data from the present study with reference data published by Roche et al. 27 for weight and length increments does not reflect the pattern of relatively rapid growth in the first 3 months in comparison with the N C H S percentiles. For both sexes, the maximum percentile for weight increments is the 25th percentile, occurring at the 3- to 4-month interval. In comparison with the incremental reference data from Guo et al., 2s which were based on a larger population, and with a higher percentage of infants fed human milk, the mean increments from the current study have higher percentiles. Whether the differences between the two sets of incrementalreference data are due to differences in sample size, mode of feeding, or other factors is unclear. A key issue is whether the gradual downward deviation from the median between 4 and 9 months of age creates any problems for infants fed human milk or whether this group has a track that should simply be viewed as different rather than faltering. One recent study found no association between energy intake and morbidity or achievement of major developmental milestones in infants fed human milk through the first year of life. 18 Teleologically, a slower infant growth rate after 4 months of life may be advantageous to the mother or the infant or both. Alternatively, it is ap-
38
Krebs et al.
propriate to question whether the more rapid growth typical of formula-fed infants is detrimental. The difficulty in measuring potentially subtle differences in motor and cognitive development and in immune function presents a significant challenge. The stepwise regressions for the overall weight increment, as well as the changes in weight-for-age z scores during shorter intervals, indicated that a greater energy intake, esPecially early in life, was associated with better weight gain and less deviation from the median. These findings suggest that low energy intake was causally related to less weight gain. Although the change in percentiles from 2 weeks to 7 months of age was greater for weight than for length, in both the present study and in some others 3, 4 the linear growth also slowed relative to the NCHS standards, particularly after 4 months and in the female infants. The possibility that the deviation from the median in the mean length for age may be due in part to a limiting intake of one or more nutrients cannot be dismissed. Stepwise regression indicated a significant relationship between the percentage of zinc intake from human milk at 5 months and the change in weight-for-age z score between 5 and 7 months of age for the subjects in the lowest quartile. There was also a trend (p = 0.10) for higher zinc intake from human milk in those infants whose length-for-age z scores were in the highest quartile at 7 months of age compared with those in the lowest quartile. Favorable absorption of zinc from human milk, together with efficient conservation of endogenous zinc by the gastrointestinal tract, 29 may be sufficient to ensure that the average intake of zinc from human milk by the fully breast-fed infant is still adequate at 5 months of age. The balance between the higher bioavailability but lower concentration of zinc in human milk, compared with the zinc in most solids, is impossible to predict from these data. A lower-than-average zinc intake from human milk at 5 months may be inadequate, as may be an average zinc intake from milk alone at 7 months. Significantly greater linear growth and weight gain have been reported in a group of breast-fed infants given zinc supplements between 4 and 9 months of age, but no data regarding the extent of breast-feeding or the introduction of other sources of nutrients were presented) ° The combination Of these findings and those from our study raises a note of caution in concluding that the average growth pattern of infants fed human milk is optimal. Assessing the adequacy of intake on the basis of energy alone may be an oversimplification because intake of one or more specific nutrients, including zinc, may also affect growth and possibly appetite. 8 Carefully designed supplementation trials, their inherent demands notwithstanding, may be the best way to address the possibility of specific nutrient deficiencies. Despite the controversy over the appropriate standards
The Journal of Pediatrics January 1994
for infant growth, some conclusions may be drawn. Given the well-documented relatively rapid weight gain in the early months of life in infants fed human milk, an individual breast-fed infant who "falters" during this time should not be regarded as normal. In assessing a 5- to 7-month-old breast-fed infant whose weight or length is crossing NCHS percentiles, one should consider appropriate introduction of solid foods, the possibility of a deficiency of one or more specific nutrients, and the differences in growth between the average human milk-fed versus the average formula-fed infant. Deviation from the growth standards for any cause should not necessarily be construed as an indication for early weaning. Appropriate investigation of potential problems should facilitate active support of the nursing motherinfant pair whenever possible. The full scope of the consequences of the differences in the growth of human milk-fed versus formula-fed infants will not be known without systematic long-term follow-up. Such long-term data will be critical in developing new standards and guiding clinical practice. We gratefully acknowledgethe enthusiastic dedication and hard work of the mothers and infants participating in this study. We also appreciate the generous support from the Medela and Ameda/Egnell companies in making available electric breast pumps for loan to the participating mothers and from Mericon Industries, which supplied the zinc and placebo tablets for the mothers, Finally, we would like to thank Susan Hartley, MS, RD, and Jamie Westcott for their substantial contributions to the dietary and laboratory data, respectively. REFERENCES
1. Whitehead RG, Paul AA. Growth charts and the assessment of infant feeding practices in the western world and in developing countries. Early Hum Dev 1984;9:187-207. 2. Dewey KG, Heinig JM, Nommsen LA, Peerson JM, Lonnerdal B. Growth of breast-fed and formula-fed infants from 0 to 18 months: the DARLING study. Pediatrics 1992;89:1035-41. 3. Duncan B, Schaefer C, Sibley B, Fonseca NM. Reduced growth velocity in exclusively breast-fed infants. Am J Dis Child 1984;138:309-13. 4. Salmenpera L, Perheentupa J, Siimes MA. Exclusivelybreast fed healthy infants grow slower than reference infants. Pediatr Res 1985;19:307-12. 5. Butte NF, Garza C, O'Brian-Smith, E, Nichols BL. Human milk intake and growth in exclusivelybreast-fed infants. J PEDIATR1984;104:187-95. 6. Stuff JE, Nichols BL. Nutrient intake and growth performance of older infants fed human milk. J PEDIAT~1989;115:959-68. 7. Hambidge KM, Casey CE, Krebs NF. Zinc. In: Mertz W, ed. Trace elements in human and animal nutrition. 5th ed. Orlando, Florida: Academic Press, 1986:1-137. 8. KrebsNF, Hambidge KM, Walravens PA. Increased food intake of young children receivinga zinc supplement. Am J Dis Child 1984;138:270-3. 9. Krebs NF, Hambidge KM, Jacobs MA, Rasbach JO. The effects of a dietary zinc supplement during lactation on longitu-
The Journal o f Pediatrics Volume 124, Number 1
10. 11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
dinal changes in maternal zinc status and milk zinc concentrations. Am J Clin Nutr 1985;41:560-70. Krebs NF, Hambidge KM. Zinc requirements and zinc intakes of breast-fed infants. Am J Clin Nutr 1986;43:288-92. Krebs NF, Reidinger C, Westcott JL, Hartley S, Hambidge KM. The effect of zinc supplementation during lactation on maternal zinc status and milk zinc concentrations [Abstract]. FASEB J 1991;5:A1289. Krebs NF, Hambidge KM, Jacobs MA, Mylet S. Zinc in human milk: diurnal and within feed patterns. J Pediatr Gastrenterol Nutr 1985;4:227-9. Hambidge KM, King JC, Kern JL, Westcott JL, Stall C. Prebreakfast plasma zinc concentrations: the effect of previous meals. J Trace Elem Electrolytes Health Dis 1990;4:229-31. Hamill PVV. NCHS growth curves for children, birth to 18 years. Washington, D.C.: (Publication No. [PHS] 78-1650:) National Center for Health Statistics, U.S. Department of Health, Education and Welfare, 1977. Woolridge MW, Butte NF, Dewey KG, Ferris AM, Garza C, Keller RP. Methods for the measurement of milk volume of the breast-fed infants. In: Jensen RG, Neville MC, eds. Human lactation: milk components and methodologies. New York: Plenum Press, 1985:5-20. Stuff JE, Garza C, Butte C, et al. Sources of variance in milk and caloric intakes in breast-fed infants: implications for lactation study design and interpretation. Am J Clin Nutr 1986; 43:361-6. Neville MC, Keller R, Seacat J, et al, Studies in human lactation: milk volumes in lactating women during the onset of lactation and full lactation. Am J Clin Nutr 1988;48:1375-86. Dewey KG, Heinig MJ, Nommsen LA, Lonnerdal B. Adequacy of energy intake among breast-fed infants in the DARLING study: relationships to growth velocity, morbidity and activity levels. J PEDIATR 1991;119:538-47. Lucas A, Ewing G, Roberts SB, Coward WA. How much energy does the breastfed infant consume and expend? Br J Nutr 1987;295:75-7. National Research Council, Food and Nutrition Board. Rec-
Krebs et al.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
39
ommended dietary allowances. 10th ed. Washington, D.C.: National Academy Press, 1989. Energy and protein requirements: report of a Joint FAO/ W H O / U N U Expert Consultation. (Technical Report Series 724.) Geneva: World Health Organization, 1985. Prentice AM, Lucas A, Vasquez-Velasquez L, Davies PSW, Whitehead RG. Are current dietary guidelines for young children a prescription for overfeeding? Lancet 1988;2:1066-9. Butte NF, Wong WW, Ferlic L, Smith OE, Klein PD, Garza C. Energy expenditure and deposition of breast-fed and formula-fed infants during early infancy. Pediatr Res 1990; 28:631-40. Heinig M J, Nommsen LA, Peerson JM, Lonnerdal B, Dewey KG. Energy and protein intakes of breast-fed and formula-fed infants during the first year of life and their association with growth velocity: the DARLING study. Am J Clin Nutr 1993; 58:152-61. Golden BE and Golden MHN. Plasma zinc, rate of weight gain, and the energy cost of tissue deposition in children recovering from severe malnutrition on a cow's milk or soya protein based diet. Am J Clin Nntr 1981;34:892-9. Miller LV, Hambidge KM, Naake VL, Hong Z, Westcott JL, Fennessey PV. Size of the zinc pools that exchange rapidly with plasma zinc in humans: alternative techniques for measuring and relation to dietary zinc intake. J Nutr (in press). Roche AF, Guo S, Moore WM. Weight and recumbent length from 1 to 12 mo of age: reference data for l-too increments. Am J Clin Nutr 1989;49:599-607. Guo S, Roche AF, Fomon S J, et al. Reference data on gains in weight and length during the first two years of life. J PE~IATR 1991;119:355-62. Krebs NF, Reidinger C, Westcott J, Miller LV, Fennessey PV, Hambidge KM. Whole body zinc metabolism in full-term, breastfed and formula fed infants. Proceedings of the Conference on Nutrient Regulation during Pregnancy, Lactation and Infant Growth. New York: Academic Press (in press). Walravens PA, Chakar A, Mokni R, Denise J, Lemonnier D. Zinc supplements in breastfed infants. Lancet 1992;340:683-5.
Several studies have suggested a pattern of growth in infants fed human milk that differs from standards based primarily on the growth of formula-fed infants. 15 Infants who have Supported by grant 5-R01-DK1243 from the National Institutes of Health, grant RR00069 from the Clinical Research Center, grant T86-00279-023 from the PEW Nutrition Fellowship, and grant 5-T32-DK07658 (Training grant) from the National Institutes of Health. Submitted for publication March 18, 1993; accepted Aug. 11, 1993. Reprint requests: Nancy F. Krebs, MD, MS, University of Colorado Health Sciences Center, 4200 East Ninth Ave., Box C225, Denver, CO 80262. Copyright © 1994 by Mosby-Year Book, Inc. 0022-3476/94/$1.00 + .10 9/20/50758
32
been fed human milk for several months and whose growth is faltering relative to the National Center for Health Statistics (NCHS) standards do not necessarily increase their energy intake when given ready access to weaning foods. 6 This may be "normal" for the breast-fed infant, but an alternative hypothesis is that an inadequate supply of a single nutrient in mature human milk, such as zinc, may limit both appetite and growth. Mild zinc deficiency in infants and young children can depress growth, 7 and limited data indicate that food intake can also be depressed. 8 The results of an earlier pilot study suggested that subtle differences in the normal rate of decline in the zinc concentration of human milk as lactation progresses may result from suboptimal maternal dietary zinc intake.9 It appeared that those differences, though small, could result in a zinc intake by the
The Journal o f Pediatrics Volume 124, Number 1
fully breast-fed infant that was inadequate to sustain normal growth velocity, l° One of the objectives of this study was to determine the growth pattern of normal infants who were human milk exclusively for at least 5 months and to investigate the relationship of growth to milk intake. An additional objective was to determine whether infant growth is related to zinc intake from human milk.
METHODS Study design. We conducted a longitudinal study of infants who were to be fed human milk exclusively for at least 5 months. Subjects were the offspring of mothers who were enrolled before 2 weeks of age and who were randomly assigned to receive a daily 15 mg zinc supplement or placebo throughout lactation. Supplementation began at 2 weeks of age; data were collected then, at 1 month of age, and at monthly (_+ 1 week) intervals thereafter through at least 7 months of age. No significant effect of maternal zinc supplementation on milk zinc concentrations or milk output was found. 11 Infants will thus be considered as one group in this article. For the aspects of the study reported herein, data and sample collection consisted of infant anthropometry, milk samples for analysis of zinc concentrations, 3-day test weighing of infants for estimation of milk intake at 4 time points, and 3-day diet records concurrent with the test weighing once weaning foods were introduced. Subjects. We studied 40 female and 31 male infants through 7 months of age; 43 infants (23 girls) continued in the study through 9 months of age. All were healthy term infants who were of appropriate weight for gestational age. The mothers' mean age was 30.4 + 3.7 years; mean parity was 1.8 + 1.0, with a range of 1 to 5; all but one were white; socioeconomic status was middle to upper middle income; and education level was _> 12 years for 93% of the mothers (75% were college graduates). None of the mothers smoked, drank alcohol more than occasionally, or were known to abuse drugs. At the time of recruitment, mothers were enrolled only if they expressed the intention to breast-feed exclusively through 5 months and continue to nurse through 7 months. Informed written consent was obtained before data collection. The protocol was approved by the human subjects committee at the University of Colorado Health Sciences Center. Anthropometry. Infants' weights and lengths were obtained at 2 weeks and 1 month of age, and monthly thereafter. Naked weights were obtained on an electronic digital balance, accurate to 1 gm. Lengths were measured with the infant held in full extension on a portable measuring board with increments in millimeters. Head circumferences were measured at each visit with a laminated tape. All measurements were performed in duplicate by the same investiga-
Krebs et aL
33
tor (C.J.R.), who was assisted by the mother for the length measurements. Birth length data are not presented because of the variable techniques used in newborn nurseries to obtain these measurements. Milk and dietary intake. Human milk intake was measured by the test-weighing technique for 3 consecutive days at 2 weeks and at 3, 5, and 7 months of age. Electronic digital balances (Sartorius Corp., Bohemia, N.Y.), which integrated 100 rapid serial measurements to provide a mean weight to the nearest 1 gin, were set up in the subject's home t~: the test weighing. The mothers were given detailed instructions for the test-weighing procedure, including directions for proper use of the balances, and for use of consistent clothing and diapers for weights before and after feedings. Detailed log sheets were kept by the mothers for the 3 days to record weights. These were reviewed by a member of the research team at the end of each test-weighing period to identify any incomplete or erroneous entries. Research personnel were available by telephone to the mothers during the test-weighing periods for consultation if mechanical or logistical problems arose. We calculated 24hour intakes by finding the total intake for the 3-day period, dividing by the exact number of hours, and multiplying by 24. Intakes were not adjusted for insensible water losses. Participating mothers were lent electric breast pumps (Medela, Inc., Crystal Lake, Ill., and Ameda/Egnell Inc., Cary, Ill.) to allow provision of expressed human milk by bottle. Infant formulas were used only in small amounts in mixtures with dry cereals or other solid foods. Once solids had been introduced, 3-day diet records of measured food intake were kept for the infants concurrently with the test weighing. Nutrient analyses were performed with a computer program (Practorcare, Inc., San Diego, Calif.) that includes commonly used baby foods and has been modified in our laboratory to provide complete data for zinc content of foods by laboratory analysis of individual foods, manufacturers' data, or imputation from similar foods. Energy content of human milk was assumed to be 0.67 kcal/gm. Milk samples were obtained from each breast midway through the feeding on each side for every nursing session during the test-weighing periods for the first 29 subjects completing the study. Comparison of milk zinc concentrations by analysis of variance revealed no significant diurnal variation and no significant differences between sides (unpublished data). On the basis of these and previously published data, 12 the remaining 42 subjects collected three samples per day during the test-weighing periods. A single sample was obtained at the monthly visits without testweighing data. Samples of 5 to 10 ml were expressed by hand into zinc-free plastic vials and were frozen at - 2 0 ° C until analysis. Zinc concentration was determined on each sample by laboratory procedures described previously.9
34
Krebs et al.
The Journal of Pediatrics January 1994
Table I. Daily intakes of human milk and zinc from human milk, and calculated intakes of energy and zinc from weaning foods Age
Human milk In gm/day In gm/kg per day Zinc from human milk In mg/day Intake from weaning foods Emergy In kcal In kcal/kg Zinc In mg/day
2 Wk
3 Mo
5 Mo
7 Mo
600 +- 120 160 _+ 30
690 -+ 110 120 + 16
720 _+ 130 105 _+ 15
640 +_ 150 85 _+ 17
2.30 +_ 0.68
1.00 -+ 0.43
0.81 _+ 0.42 (16)*
0.52 _+ 0.31 (66)*
44_+ 21 6_ 3
109 + 74 15 + 10
0.17 + 0.14
0.47 _+ 0.43
Values (exceptthose in parentheses) are expressed as mean _+SD. *Number of subjects (male and female infants combined).
Blood tests. Samples were obtained by finger stick for spun hematocrits and plasma zinc concentrations at 6 months of age. Care was taken to clean the skin thoroughly before lancing and to avoid zinc contamination of the samples during collection. Plasma zinc concentrations were determined by a previously described method. 13 Data management and statistical analysis. The z scores and the percentiles for anthropometric data were calculated relative to the N C H S growth curves 14 by the CDC Anthropometric Software Package (Division of Nutrition, Center for Health Promotion and Education, Centers for Disease Control and Prevention, Atlanta, Ga.) Statistical analyses, including t test and stepwise linear regression, were performed with the Statistical Analysis Systems software program (1985; SAS Institute Inc., Cary, N.C.). Plotting of the longitudinal data was used extensively as a guide to statistical analysis. Percentiles and z scores for length, weight, and head circumference relative to age at the time of measurement were calculated and means were determined for each monthly cycle. The monthly weight and length increments were determined by dividing increments by the number of days between measurements and multiplying by 30. We also fitted anthropometric data, as functions of age from 2 weeks to 9 months, to a second-degree curve, smoothing the data and allowing the use of interpolated values to replace missing values. The z scores were then recalculated at the exact monthly time points. Separately for each interval (2 weeks to 3 months, 3 to 5 months, 5 to 7 months, 7 to 9 months), changes in fitted z score indexes over the interval were regressed on (1) subject attributes (sex and birth weight) and (2) dietary data at the start of the interval (energy [per kilogram] and zinc intakes, from human milk and supplemental food combined, and propor-
tion of energy and zinc intakes from human milk). In these regressions the method of forward selection and backward elimination was used. The significance level for entry and exit was p = 0.15. A specific effect was assessed after control for all other significant effects. RESULTS
Milk, energy, and zinc intakes. Results of test-weighing measurements are given in Table I. At each cycle, mean intakes of the male infants were 5% to 6% greater than those of the female infants. When intakes were compared relative to weight, however, they were virtually the same. Sixteen infants (23%) began intake of solids at 41/2 months. These solids contributed an average of 9% _+ 5% of energy for these infants at the 5-month cycle. At the sixthmonth visit, 55% of subjects had begun to consume weaning foods. By age 7 months, 66 infants (93%) were taking solids, providing an average of 15 + 11 kcals/kg per day, or 19% of total mean daily intake. The total energy intake from human milk and weaning foods at 7 months of age was 70 _+ 13 kcal/kg per day. Intakes of zinc from human milk declined sharply during the course of lactation (Table I). There were no significant differences between male and female infants. The addition of supplemental foods at 5 and 7 months of age contributed 5% _+ 11% and 40% _+ 26%, respectively, to the total mean zinc intakes, which were 0.85 _+ 0.42 mg/day and 1.0 + 0.5 rag/day at the two ages, respectively. Anthropometry. The mean birth weight was 3.38 +_ 0.45 kg. For the first 3 months the mean weight was greater than the median, with a peak at the 62nd percentile at 2 months of age, followed by a gradual decline to the 33rd percentile at 7 months of age (Fig. 1). The mean length percentile started at the 43rd percentile, increased slightly by 1 month
The Journal of Pediatrics Volume 124, Number 1
Krebs et al.
35
9o ~, 8o Z (D
70
~_ 6o
r~
o 40 l.a_ !
5o Ld
2O
•
Female
A
Male
"
------A-
1 10 I
0
1
J
[
2
,
I
~
3
r
4
,
I
5
,
I
6
~
I
7
,
I
8
,
I
9
AGE (MO)
Fig. t. Mean (_+SEM) weight-for-age percentiles versus age, by sex; solid lines indicate subjects followed through 7 months of age (40 female infants, 31 male infants); dashed lines indicate subjects followed through 9 months of age (23 female infants, 20 male infants). of age, and then gradually declined to the 28th percentile by 7 months (Fig. 2). At 7 months of age the boys were somewhat leaner than the girls, with mean weight-forlength percentile at the 37th and 60th percentiles, respectively. Regression analysis of the longitudinal data of actual z scores indicated significant differences between the sexes in weight-for-age percentiles from 2 weeks to 7 months and from 2 weeks to 9 months of age (p = 0.028 and 0.019, respectively). Significant interactions between age and sex were found in length-for-age z scores from 2 weeks to 9 months of age (p = 0.015) and for weight for length from 2 weeks to 7 months of age and from 2 weeks to 9 months of age (p = 0.004 and 0.003). The monthly increments were generally not different between male and female infants (Table II). The length increments from 2 weeks to 7 months of age for male and female infants were 15.6 cm and 14.1 cm, respectively (p = 0.001). The mean weight increment during this interVal was 3.71 _+ 0.88 kg and did not differ between the sexes. Growth and intake. Stepwise regression was performed foi the entire group for the dependent variables of changes in length-for-age, weight-for-age, and weight-for-length z scores during the intervals between the intake measurements. For the interval from 2 weeks to 3 months of age only, a significant negative association of birth weight with changes in Z scores for all three anthropometric measurements was observed (p _< 0.0002 for each change). Gender was also a significant variable at the 2-week to 3-month interval; male gender was associated with more negative
change in the weight-for-length z score (p = 0.003). Of the dietary variables examined, only the total energy per kilogram at 2 weeks of age was positively associated with the 2-week to 3-month change in the weight-for-age z score (p = 0.028) and in the weight-for-length z score (p = 0.0003); this association represents milk intake per kilogram at the 2-week test weighing. The weight and length increments during the interval from 2 weeks to 7 months of age were also examined as dependent variables in stepwise regression. For the length increment, male gender was positively associated (p = 0.0001) and birth weight negatively associated (p = 0.004). For the weight increment, energy intake per kilogram at 2 weeks of age was again positively associated (p = 0.003). Stepwise regression was also performed on subjects in the lowest quartile of the distributions of length-for-age, weightfor-age, weight-for21ength, and head-circumference z scores, with the sameindependent and dependent variables used for the same intervals noted above. There were no significant associations with the changes in length-for-age z scores at any of the intervals. Changes in the weight-for-age z scores at the 3- to 5-month and the 5- to 7-month intervals were positively associated with energy intake per kilogram at 3 and 5 months of age, respectively (p = 0.004 and 0.024). Changes in the weight-for-age z scores for the 5- to 7-month interval were negatively associated with the percentage of energy (p = 0.006) and positively associated with the percentage of zinc (p = 0.003) obtained from human milk at 5 months of age. The change in weight-forlength z score at 7 to 9 months of age was positively asso-
36
Krebs et al.
The Journal of Pediatrics January 1994
90
_~ 8o F-
~_ 60
.~
5o
~ 5o w 20
•
Male
10 I 1
,
I
,
2
,
~
3
4
, 5
I 6
,
I
~
I
7
8
t
I 9
AGE (MO)
Fig. 2. Mean ( ± SEM) length-for-age percentiles versus age, by sex; solid lines indicate subjects followed through 7 months of age (40 female infants, 31 male infants); dashed lines indicate subjects followed through 9 months of age (23 female infants, 20 male infants). Table II. M o n t h l y increments in weight a n d length for infants fed h u m a n milk, from b i r t h to 9 months of age Female Interval
(mo)
Weight (kg/mo) 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 Length (cm/mo) 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9
infants
n
Mean
40 40 39 39 39 39 23 23
0.80 0.66 0.53 0.42 0.36 0.31 0.27 0.18
40 40 39 39 40 39 23 23
2.66 2.81 2.19 1.55 1.68 1.23 1.33 1.41
Male
infants
± SD
n
Mean
± SD
± ± ± ± ± ± + ±
0.25 0.25 0.19 0.22 0.17 0.19 0121 0.21
31 31 31 31 29 29 20 19
0.89 0.63 0.58 0.43 0.36 0.30 0.33 0.28
± 0.33 ± 0.16" ± 0.20 + 0.17 ± 0.22 _+ 0.16 ± 0.18 ± 0.19
_ ± ± ± ± ± ± ±
0.82 0.86 0.83 0.89 0.95 0.82 0.58 0.75
30 31 31 31 29 29 19 19
3.22 2.77 2.49 1.93 1.73 1.50 1.28 1.32
± 0.99 ___0.99 _+ 0.85 ± 0.95 ± 0.87 _+ 0.83 ± 0.70 ± 0.63
*Significantly less than female increment (p = 0.008).
ciated with the percentage of zinc intake from h u m a n milk at 7 m o n t h s (p = 0.020). T h e changes in head-circumference z score for the 7- to 9-month interval were also positively associated with the percentage of zinc obtained from h u m a n milk at 7 m o n t h s of age (p -- 0.020). Biochemical data. P l a s m a zinc concentrations were det e r m i n e d at 6 months of age for 54 of the subjects; the m e a n was 11.9 +_ 2.6 # m o l / L (78 +_ 1 7 /~g/dl). A significant
negative association was found by stepwise regression between plasma zinc concentration and c h a n g e in weightfor-length z score between 5 and 7 m o n t h s of age for all subjects (p = 0.033) and for the subjects in the lowest quartile for this index (p = 0.011). Trends for the same relationship were found for the c h a n g e in length-for-age z score for the entire group and those in the lowest quartile (p = 0.07). Stepwise regression on the absolute i n c r e m e n t s
The Journal of Pediatrics Volume 124, Number 1
indicated a trend for the weight increment to be negatively associated with the plasma zinc concentration (p = 0.10). The mean hematocrit, obtained for 58 infants, was 0.38 ___0.3. There was no significant correlation between calculated iron intake from weaning foods and hematocrit. DISCUSSION The test-weighing method, although potentially disruptive to the mother-infant pair, has been shown to compare well with other techniques, including isotope dilution techniques, and is considered the "gold standard." The 3-day test-weighing period chosen for this study was thought to be a reasonable period to compensate for intraindividual daily variation while minimizing changes in the feeding pattern.15, 16 The mean milk intakes for this group of infants through the first 7 months of life fell at the lower end of the range reported in other studies. 5, 17, 18 A correction for insensible water losses was not attempted; the actual milk intake of the group could be expected to be higher, possibly as much as 5%. 15 Because of the difficulty of achieving representative sampling for bomb calorimetry, we used a standard conversion factor to calculate the energy intake of the infants. 19 At all points in the study the estimated energy intake relative to infant weight was less than current recommendations 2°, 21 and less than intakes reported for formula-fed infants. 22-24 The relatively rapid weight gain in the early months, despite lower-than-recommended energy intakes, suggests that energy requirements for the young infant fed human milk are lower than current recommendations. Investigating possible differences in energy requirements of human milk-fed and formula-fed infants, Butte et al. 23 found that the differences in energy intake were not totally accounted for by differences in energy expenditure or rate of growth, and may be related to energy digestibility and the composition of newly synthesized tissues. Weaning foods were consumed by nearly all the infants by the 7-month visit. At this time, the weight-adjusted energy intake was virtually the same as the average intake at 5 months of age. The weaning foods thus were apparently substituting calories and nutrients for those previously supplied by human milk, as also noted by Stuff and Nichols. 6 This observation suggests that a limitation of maternal milk supply was not directly governing energy intake. The stage of lactation is the most critical determinant for the zinc concentration of human milk, which declined by a factor of 5 from 0.5 to 9 months. 11 There was a corresponding decline in the calculated infant intakes of zinc from human milk during the course of the study. With the addition of weaning foods, the zinc intake at 7 months of age was essentially doubled to a mean of 1 mg/day. The larger percentage of total zinc intake supplied at 7 months by supple-
Krebs et al.
37
mental foods compared with that for energy, 40% versus 19%, reflects the relatively low concentration of zinc in human milk at this stage of lactation. The fractional absorption of the zinc contributed by solids would likely be lower than that from the milk, and the endogenous zinc losses in the feces may be higher. The bioavailability of the zinc in the milk could also be altered by the intake of other foods, presumably depending on the choice of solids and on such factors as fiber, phytate, protein, and iron, which may influence zinc bioavailability. In infants recovering from malnutrition, low plasma zinc concentrations were associated with accelerated catch-up growth. 25 Our findings suggest a similar relationship over a range of normal growth rates. To the extent that plasma zinc concentrations may reflect the size of relatively rapidly exchangeable pools of zinc available for metabolic processes, 26 the lower levels suggest a partial depletion of those pools to meet the demands of new tissue synthesis. After the third month, the gradual decline in mean weight-for-age percentiles during the next several months to approximately the 30th percentile (combined sexes) at 7 months of age is consistent with other reports. 1-4 The similar trend for the group followed through 9 months of age suggests that the progressive deviation from the N C H S median continues through at least 9 months. The monthly growth increments observed in this study are similar to those for breast-fed infants reported on by Heinig et al., 24 though the actual weights and lengths were greater in the latter cohort. Comparison of the data from the present study with reference data published by Roche et al. 27 for weight and length increments does not reflect the pattern of relatively rapid growth in the first 3 months in comparison with the N C H S percentiles. For both sexes, the maximum percentile for weight increments is the 25th percentile, occurring at the 3- to 4-month interval. In comparison with the incremental reference data from Guo et al., 2s which were based on a larger population, and with a higher percentage of infants fed human milk, the mean increments from the current study have higher percentiles. Whether the differences between the two sets of incrementalreference data are due to differences in sample size, mode of feeding, or other factors is unclear. A key issue is whether the gradual downward deviation from the median between 4 and 9 months of age creates any problems for infants fed human milk or whether this group has a track that should simply be viewed as different rather than faltering. One recent study found no association between energy intake and morbidity or achievement of major developmental milestones in infants fed human milk through the first year of life. 18 Teleologically, a slower infant growth rate after 4 months of life may be advantageous to the mother or the infant or both. Alternatively, it is ap-
38
Krebs et al.
propriate to question whether the more rapid growth typical of formula-fed infants is detrimental. The difficulty in measuring potentially subtle differences in motor and cognitive development and in immune function presents a significant challenge. The stepwise regressions for the overall weight increment, as well as the changes in weight-for-age z scores during shorter intervals, indicated that a greater energy intake, esPecially early in life, was associated with better weight gain and less deviation from the median. These findings suggest that low energy intake was causally related to less weight gain. Although the change in percentiles from 2 weeks to 7 months of age was greater for weight than for length, in both the present study and in some others 3, 4 the linear growth also slowed relative to the NCHS standards, particularly after 4 months and in the female infants. The possibility that the deviation from the median in the mean length for age may be due in part to a limiting intake of one or more nutrients cannot be dismissed. Stepwise regression indicated a significant relationship between the percentage of zinc intake from human milk at 5 months and the change in weight-for-age z score between 5 and 7 months of age for the subjects in the lowest quartile. There was also a trend (p = 0.10) for higher zinc intake from human milk in those infants whose length-for-age z scores were in the highest quartile at 7 months of age compared with those in the lowest quartile. Favorable absorption of zinc from human milk, together with efficient conservation of endogenous zinc by the gastrointestinal tract, 29 may be sufficient to ensure that the average intake of zinc from human milk by the fully breast-fed infant is still adequate at 5 months of age. The balance between the higher bioavailability but lower concentration of zinc in human milk, compared with the zinc in most solids, is impossible to predict from these data. A lower-than-average zinc intake from human milk at 5 months may be inadequate, as may be an average zinc intake from milk alone at 7 months. Significantly greater linear growth and weight gain have been reported in a group of breast-fed infants given zinc supplements between 4 and 9 months of age, but no data regarding the extent of breast-feeding or the introduction of other sources of nutrients were presented) ° The combination Of these findings and those from our study raises a note of caution in concluding that the average growth pattern of infants fed human milk is optimal. Assessing the adequacy of intake on the basis of energy alone may be an oversimplification because intake of one or more specific nutrients, including zinc, may also affect growth and possibly appetite. 8 Carefully designed supplementation trials, their inherent demands notwithstanding, may be the best way to address the possibility of specific nutrient deficiencies. Despite the controversy over the appropriate standards
The Journal of Pediatrics January 1994
for infant growth, some conclusions may be drawn. Given the well-documented relatively rapid weight gain in the early months of life in infants fed human milk, an individual breast-fed infant who "falters" during this time should not be regarded as normal. In assessing a 5- to 7-month-old breast-fed infant whose weight or length is crossing NCHS percentiles, one should consider appropriate introduction of solid foods, the possibility of a deficiency of one or more specific nutrients, and the differences in growth between the average human milk-fed versus the average formula-fed infant. Deviation from the growth standards for any cause should not necessarily be construed as an indication for early weaning. Appropriate investigation of potential problems should facilitate active support of the nursing motherinfant pair whenever possible. The full scope of the consequences of the differences in the growth of human milk-fed versus formula-fed infants will not be known without systematic long-term follow-up. Such long-term data will be critical in developing new standards and guiding clinical practice. We gratefully acknowledgethe enthusiastic dedication and hard work of the mothers and infants participating in this study. We also appreciate the generous support from the Medela and Ameda/Egnell companies in making available electric breast pumps for loan to the participating mothers and from Mericon Industries, which supplied the zinc and placebo tablets for the mothers, Finally, we would like to thank Susan Hartley, MS, RD, and Jamie Westcott for their substantial contributions to the dietary and laboratory data, respectively. REFERENCES
1. Whitehead RG, Paul AA. Growth charts and the assessment of infant feeding practices in the western world and in developing countries. Early Hum Dev 1984;9:187-207. 2. Dewey KG, Heinig JM, Nommsen LA, Peerson JM, Lonnerdal B. Growth of breast-fed and formula-fed infants from 0 to 18 months: the DARLING study. Pediatrics 1992;89:1035-41. 3. Duncan B, Schaefer C, Sibley B, Fonseca NM. Reduced growth velocity in exclusively breast-fed infants. Am J Dis Child 1984;138:309-13. 4. Salmenpera L, Perheentupa J, Siimes MA. Exclusivelybreast fed healthy infants grow slower than reference infants. Pediatr Res 1985;19:307-12. 5. Butte NF, Garza C, O'Brian-Smith, E, Nichols BL. Human milk intake and growth in exclusivelybreast-fed infants. J PEDIATR1984;104:187-95. 6. Stuff JE, Nichols BL. Nutrient intake and growth performance of older infants fed human milk. J PEDIAT~1989;115:959-68. 7. Hambidge KM, Casey CE, Krebs NF. Zinc. In: Mertz W, ed. Trace elements in human and animal nutrition. 5th ed. Orlando, Florida: Academic Press, 1986:1-137. 8. KrebsNF, Hambidge KM, Walravens PA. Increased food intake of young children receivinga zinc supplement. Am J Dis Child 1984;138:270-3. 9. Krebs NF, Hambidge KM, Jacobs MA, Rasbach JO. The effects of a dietary zinc supplement during lactation on longitu-
The Journal o f Pediatrics Volume 124, Number 1
10. 11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
dinal changes in maternal zinc status and milk zinc concentrations. Am J Clin Nutr 1985;41:560-70. Krebs NF, Hambidge KM. Zinc requirements and zinc intakes of breast-fed infants. Am J Clin Nutr 1986;43:288-92. Krebs NF, Reidinger C, Westcott JL, Hartley S, Hambidge KM. The effect of zinc supplementation during lactation on maternal zinc status and milk zinc concentrations [Abstract]. FASEB J 1991;5:A1289. Krebs NF, Hambidge KM, Jacobs MA, Mylet S. Zinc in human milk: diurnal and within feed patterns. J Pediatr Gastrenterol Nutr 1985;4:227-9. Hambidge KM, King JC, Kern JL, Westcott JL, Stall C. Prebreakfast plasma zinc concentrations: the effect of previous meals. J Trace Elem Electrolytes Health Dis 1990;4:229-31. Hamill PVV. NCHS growth curves for children, birth to 18 years. Washington, D.C.: (Publication No. [PHS] 78-1650:) National Center for Health Statistics, U.S. Department of Health, Education and Welfare, 1977. Woolridge MW, Butte NF, Dewey KG, Ferris AM, Garza C, Keller RP. Methods for the measurement of milk volume of the breast-fed infants. In: Jensen RG, Neville MC, eds. Human lactation: milk components and methodologies. New York: Plenum Press, 1985:5-20. Stuff JE, Garza C, Butte C, et al. Sources of variance in milk and caloric intakes in breast-fed infants: implications for lactation study design and interpretation. Am J Clin Nutr 1986; 43:361-6. Neville MC, Keller R, Seacat J, et al, Studies in human lactation: milk volumes in lactating women during the onset of lactation and full lactation. Am J Clin Nutr 1988;48:1375-86. Dewey KG, Heinig MJ, Nommsen LA, Lonnerdal B. Adequacy of energy intake among breast-fed infants in the DARLING study: relationships to growth velocity, morbidity and activity levels. J PEDIATR 1991;119:538-47. Lucas A, Ewing G, Roberts SB, Coward WA. How much energy does the breastfed infant consume and expend? Br J Nutr 1987;295:75-7. National Research Council, Food and Nutrition Board. Rec-
Krebs et al.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
39
ommended dietary allowances. 10th ed. Washington, D.C.: National Academy Press, 1989. Energy and protein requirements: report of a Joint FAO/ W H O / U N U Expert Consultation. (Technical Report Series 724.) Geneva: World Health Organization, 1985. Prentice AM, Lucas A, Vasquez-Velasquez L, Davies PSW, Whitehead RG. Are current dietary guidelines for young children a prescription for overfeeding? Lancet 1988;2:1066-9. Butte NF, Wong WW, Ferlic L, Smith OE, Klein PD, Garza C. Energy expenditure and deposition of breast-fed and formula-fed infants during early infancy. Pediatr Res 1990; 28:631-40. Heinig M J, Nommsen LA, Peerson JM, Lonnerdal B, Dewey KG. Energy and protein intakes of breast-fed and formula-fed infants during the first year of life and their association with growth velocity: the DARLING study. Am J Clin Nutr 1993; 58:152-61. Golden BE and Golden MHN. Plasma zinc, rate of weight gain, and the energy cost of tissue deposition in children recovering from severe malnutrition on a cow's milk or soya protein based diet. Am J Clin Nntr 1981;34:892-9. Miller LV, Hambidge KM, Naake VL, Hong Z, Westcott JL, Fennessey PV. Size of the zinc pools that exchange rapidly with plasma zinc in humans: alternative techniques for measuring and relation to dietary zinc intake. J Nutr (in press). Roche AF, Guo S, Moore WM. Weight and recumbent length from 1 to 12 mo of age: reference data for l-too increments. Am J Clin Nutr 1989;49:599-607. Guo S, Roche AF, Fomon S J, et al. Reference data on gains in weight and length during the first two years of life. J PE~IATR 1991;119:355-62. Krebs NF, Reidinger C, Westcott J, Miller LV, Fennessey PV, Hambidge KM. Whole body zinc metabolism in full-term, breastfed and formula fed infants. Proceedings of the Conference on Nutrient Regulation during Pregnancy, Lactation and Infant Growth. New York: Academic Press (in press). Walravens PA, Chakar A, Mokni R, Denise J, Lemonnier D. Zinc supplements in breastfed infants. Lancet 1992;340:683-5.