- Email: [email protected]
Longitudinal changes in milk composition of mothers delivering preterm and term infants
Ear!v Human Deoelopment. Elsevier
9 (1984) 153-162
153
EHD 00533
Longitudinal changes in milk composition of mothers delivering preterm and term infants Nancy F. Butte, Cutberto Garza, Carmen A. Johnson, E. O’Brian Smith and Buford L. Nichols USDA/ARS
Chrldren’s Nutrition Research Center, Section of Nutritlon and Gastroenterolog,: of Pedratrics, Baylor College of Medicrne, Houston, TX 77030. U.S.A. Accepted
for publication
Department
18 August 1983
Summary The concentrations of protein nitrogen (PN), non-protein nitrogen (NPN), energy, fat, sodium (Na), calcium (Ca), phosphorus (P), magnesium (Mg), and zinc (Zn) were determined in human milk from mothers giving birth to full-term (n = 13) and preterm infants (n = 8). Milk samples were collected under controlled conditions at two-week intervals for 12 weeks postpartum. Statistically significant differences in PN, Ca, and P concentrations were detected between the milk from mothers of preterm and term infants. The mean PN concentration in the preterm milk was statistically higher than that of term milk (198 vs. 164 mg N/dl), in contrast to the lower mean Ca (220 vs. 261 mg/l) and P (125 vs. 153 mg/l) concentrations detected in the preterm milk. NO other differences in mean nutrient concentration were observed between the two groups. Concentrations of PN. NPN. Na, P. and Zn decreased over time. The concentration of Mg increased slightly. The content of fat, energy, and Ca did not change. milk composition;
preterm
infants;
term infants
Introduction Optimal nutritional management of low birthweight infants remains controversial. Although human milk is used to feed these infants, its adequacy has been questioned. Davies [8] found that mature human milk fed to preterm infants did not Address all correspondence to: Nancy Pediatrics, Baylor College of Medicine.
0378-3782/84/$03.00
F. Butte. Ph.D.. Section of Nutrition & GI. 1200 Moursund, Houston. TX 77030. U.S.A.
0 1984 Elsevier Science Publishers
B.V.
Department
of
154
support the intrauterine growth rate. Fomon et al. [14] concluded from theoretical calculations that the concentrations of protein, calcium, sodium. and probably other minerals in mature human milk were insufficient to support tissue accretion at the intrauterine rate. Recent investigations [l-3,6,17,20,26], however, have demonstrated that the composition of milk from mothers giving birth prematurely differed from that of mothers delivering at term. It has been suggested that preterm milk may be more apt to meet the nutrient needs of premature infants than term milk. Atkinson and co-workers [2] reported that the total nitrogen concentration of human milk was higher in preterm than term milk during the first four weeks of lactation. Gross et al. [17] have found higher concentrations of N, Na, and Cl in preterm milk compared to term milk, and similar levels of K, Ca, P, and Mg throughout the first 28 days of life. Although the Ca, P, and Mg requirements may not be satisfied, these investigators suggested that preterm milk may be more appropriate than pooled, mature breast milk for the feeding of premature infants. Atkinson et al. [6] suggested that the Na, Cl, K, Mg, N, and energy, but not the Ca and P content of preterm milk would be sufficient to meet the nutrient needs of the premature infant during the.first four weeks of life. Macromineral balance studies by the same authors have confirmed the inadequacy of mother’s milk to sustain intrauterine retention rates of Ca, P, and possibly Mg in the very low birthweight infant (VLBW) [5]. The present study was designed to test whether the differences in nutrient concentrations reported between preterm and term milk persisted beyond the four weeks heretofore studied.
Materials and Methods Study design Twenty-one women who elected to breast-feed their infants were assigned to one of two groups based on the maturity of their infants at birth. Gestational age was assessed according to Dubowitz [ll]. Infants whose gestational age was greater than or equal to 37 weeks were classified as term (n = 13) and those whose gestational age was less than 37 wk as preterm (n = 8). Milk samples were provided by the mothers at 2-week intervals during the first 12 weeks of life. Subjects Prospective mothers were recruited by referral from cooperating hospitals in the Houston area. In order to minimize confounding factors, subject selection was restricted to women 20-35 years of age, of parity not greater than two, whose consumption of carbonated beverages, toffee, tea, and alcohol did not exceed approximately two servings per day, and who were not on routine medication or steroid contraceptives. Infants were free of any congenital or acquired diseases. The participating women gave their informed consent to the study protocol which was approved by the Institutional Human Experimentation Committees. The characteristics of the women enrolled in the study are summarized in Table 1.
IS5 TABLE Maternal
I and infant characteristics
Maternal age (yr) Gravidity Pari ty Height (cm) Prepregnancy weight (kg) Infant sex Gestational age (wk) Birthweight (kg) Birthweight (percentile)
Preterm (n = 8)
Term(n
27.5 (3.0) a 1.5 (l-3)h 0 159.7 (4.2) d 52.9 (5.2)’ 5M/3F 33.9 (2.3) (30-36) 1.92 (0.70) a 43.2 (37.4)’
26.6 (5.0) ’ 1.3 (l-2)h 0.4 (O-l)h 162.3 (5.5) ’ 54.9 (6.9)’ 8M/5F 39.2 (1.4) (37-42) 2.99 (0.46) * 27.5 (24.1) ’
’
=13)
‘
’ Mean (S.D.). h Mean (range). ’ Mean (S.D.) (range).
Throughout the study. maternal health was good by history. Any incidence of hypertensive disease, diabetes, anemia, and/or chronic medications was recorded when health screening was done during recruitment. Two preterm mothers were hypertensive during pregnancy. As ascertained by interview, al1 but two of the women were taking supplemental vitamin-mineral tablets at the time of the study. Smoking (approximately 1 pack per day) was reported by four of the term women. Infants are described by sex, gestational age, and birthweight in Table 1. There was a slight preponderance of males over females in both groups. Birthweight, corrected for gestational age, was assessed according to Usher and McLean intrauterine growth standards [27]. Raw weight was transformed into a percentile weight-for-gestational age. The majority of the infants, once at home, were breast-fed exclusively. In a few cases supplemental juice and cereal were introduced at approximately 6-8 weeks of life. The high-risk preterm infants who were retained in the nursery received formula and/or human milk depending on their physician’s orders; their mothers expressed breast milk at home for the purpose of this study and to maintain an adequate milk supply to feed their infants after discharge. The progress of the preterm infants thus reflected mixed feeding. The relative proportions of human milk and artificial formula were not recorded. Milk collection Samples of human milk were collected at the end of the 2nd, 4th, 6th, 8th. 10th. and 12th week post-partum under controlled conditions. Milk samples were collected between 8 a.m. and 12 noon with the use of an Egnell breast pump (Egnell, Inc., Cary, IL). The samples were collected at least 2 h after a feeding. The entire contents of one breast were expressed into sterile, acid-washed polypropylene bottles and kept at 4°C until delivered to the laboratory (within 4 h of collection). Immediately thereafter, volumes were noted and the samples stored at -20°C for analysis.
156
Biochemical analysis Nitrogen (N) was analyzed by the Kjeldahl method before and after trichloroacetic acid (24%; 1 : 1 volume) precipitation of protein; protein nitrogen was determined on the solubilized precipitant [21], and non-protein nitrogen (NPN) was estimated from the differente between total and protein nitrogen. The content of fat was determined by the Roese-Gottlieb extraction method [19. p. 2581. The content of Na, Ca, and Mg was estimated on dry-ashed samples by atomic absorption (AA) spectrophotometry [19, p. 221. Zinc levels were determined by an AA spectrophotometer equipped with a graphite furnace [19, p. 221. Phosphorus was determined colorimetrically on dry-ashed samples by a modification of Fiske’s method [13]. The energy content was determined by bomb calorimetry [23]. Al1 analyses were performed in duplicate. Statistical analysis Nutrient concentrations were subjected to analysis of variante and covariance with repeated measures (BMDPLV) [lol. Differences between preterm and term milk were examined over the 12 weeks of lactation with weight-for-gestational age and volume of milk specimens treated as covariates. Correlations between gestational age and nutrient concentration were examined at each week of observation.
Results The macronutrient content of the human milk samples is presented in Table 11. The protein nitrogen concentration decreased significantly over the twelve weeks of observation at a similar rate (6.3 mg/dl per wk) in both groups (P < 0.0001). Although the rate of change did not differ between groups, there was a significant differente in the overall mean nitrogen concentration between preterm milk (PTM) (198 + 32 mg/dl) (k S.D.) and term milk (TM) (164 -t 25 mg/dl) samples (P < 0.002). Differences between groups in PN content were statistically significant (P < 0.05) throughout the first 2 months of observation. The NPN content of human milk did not differ between groups and decreased significantly over time at a mean rate of 1.3 mg/dl per wk. The concentration of fat did not change significantly over time and no statistically significant differences in overall means were demonstrated between groups. Similarly, no statistically significant differences were seen in the energy concentration over time or between groups. The mineral content of the milk is displayed in Table 111. The calcium content of the milk did not change significantly over the 12 weeks of lactation. The overall concentration of sodium, phosphorus, and zinc decreased significantly over time and the magnesium concentration increased (P < 0.001). The mean calcium concentration of the TM (261 + 36 mg/l) was significantly higber (P = 0.05) than that of PTM (220 f 62 mg/l). The mean phosphorus leve1 of 153 + 30 mg/l for the TM was significantly higher than the 125 f 33 mg/l for the PTM (P < 0.03). The Ca/P ratio increased from 1.5 at 2 weeks to 2.1 at 12 weeks in
11
(33)
56.0
(29)
207.0
(1.00)
nitrogen.
L Overall
’ NPN = non-protein
means differ (P < 0.025).
means differ (P < 0.002).
h Overall
(S.D.).
(9.7)
68.3
a Mean
(9.3)
63.5
Preterm
Term
Energy (kcal/dl)
(0.94)
3.45
3.79
Term
(12)
55.0
Fat (g/dl) Preterm
(16)
57.0
Preterm
Term
NPN (mg/dl)’
(18)
241.0
Preterm
Term
nitrogen (mg/dl)
Protein
(33) *
53.0
Term
(ml)
2
(10.3) (6.6)
67.9 68.2
(1.24) (0.78)
4.04
(18)
54.0
3.95
(17)
50.0
(14) (26)
(26)
64.0
173.0
(16)
39.0
content
211.0
4
and macronutrient
Weeks post-partum
of milk expressed
Preterm
Volume
The volume
TABLE
69.0
54.0
69.8 71.4
(14.8)
( 11.7)
4.77
(10.2)
65.9
64.6
4.51
(12.2)
4.22
(1.06)
45.0
49.0
146.0
173.0
68.0
55.0
10
(1.28)
(15)
4.12
(15)
50.0
(24)
59.0
151.0
66.7
(1.18)
(29)
68.0
(39)
(67)
62.0
189.0
8
(8.6)
( 13.6)
(1.17)
(1.33)
(13)
(25)
(26)
(42)
(32)
(44)
66.0
63.0
4.24
3.82
37.0
45.0
145.0
169.0
66.0
61.0
12
by mothers of preterm and term infants
68.0
(1.18)
4.50
(11)
3.96
(17)
52.0
(16)
(38)
(24)
(32)
46.0
162.0
203.0
6
of human milk produced
(12.8)
(12.7)
(1.57)
(1.45)
(14)
(14)
(26)
(28)
(30)
(28)
67.7
66.1
4.31
3.92
48.0
52.0
164.0
198.0
65.0
54.0
mean
Overall
(10.1)
(12.3)
(1.15)
(1.25)
(14)
(18)
(25)’
(32) c
(28)’
(39) h
111
3.4
(0.8)
(1.0)
’ Overall means differ (P < 0.03).
h Overall means differ (P = 0.05).
a Mean (S.D.).
4.1
Preterm
Term
Zinc
266.0 (98)
220.0 (77)
(8)
(8)
Term
33.0
Sodium Preterm
36.0
179.0 (36)
Term
Term
149.0 (38)
Phosphorus Preterm
Magnesium Preterm
214.0 (48) ’
255.0 (53)
Preterm
2
Weeks post-partum
Term
Calcium
(mS/L)
Content
(7) (6)
2.9
3.2 (0.9)
(0.6)
184.0 (54)
165.0 (34)
31.0
36.0
164.0 (25)
132.0 (36)
254.0 (52)
(9)
2.1
2.7 (0.9)
(1.1)
173.0 (65)
149.0 (30)
35.0
40.0 (11)
150.0 (32)
128.0 (33)
267.0 (24)
218.0 (62)
6
(9)
1.9
2.9
(0.6)
(0.6)
153.0 (47)
131.0 (30)
36.0
42.0 (13)
148.0 (20)
135.0 (31)
258.0 (22)
236.0 (66)
8
by mothers of preterm and term infants
213.0 (71)
4
Mineral content of human milk produced
TABLE
(9)
1.8
2.0
(1.0)
(0.9)
150.0 (49)
135.0 (37)
38.0
41.0 (10)
142.0 (34)
102.0 (27)
270.0 (25)
215.0 (61)
10
(9)
1.4
1.6
(0.7)
(0.3)
130.0 (41)
150.0 (44)
39.0 (10)
42.0
136.0 (27)
104.0 (30)
260.0 (26)
223.0 (65)
12
(8)
2.2
2.8
(0.8)
(0.8)
168.0 (57)
166.0 (51)
35.0
40.0 (10)
153.0 (30) c
125.0 (33)’
261.0 (36) h
220.0 (62) ’
mean
Overall
both groups due to the constant calcium and decreasing phosphorus concentrations. Magnesium increased (P < 0.001) slightly from an initial 34 mg/l at 2 weeks to 40 mg/l at 12 weeks. NO differences in the mean concentrations or rates of rise were seen between groups. The sodium concentrations for PTM and TM were not statistically different. Sodium decreased at a mean rate of 9.4 mg/l per wk. Zinc concentration decreased steadily over time at a rate of 0.2 mg/l per wk. The groups did not differ statistically with respect to the overall mean content or rate of change in concentration of zinc. The covariate, weight-for-gestational age, was not significant in the analyses between groups. The protein nitrogen concentration was correlated negatively to gestational age (r range = - 0.47 to -0.79, P < 0.05). Statistically significant correlations were not demonstrated between gestational age and any of the other nutrients. Variability in nutrient concentrations was evident in the scatter of individual values at each time interval. Zinc was particularly variable (CU = 0.46), followed by sodium (CU = 0.39). NPN (CU= 0.33) and fat (CU= 0.28). The coefficient of variation of the concentrations of al1 other nutrients was approximately 20%. The overall mean volume of the milk specimens provided by the preterm mothers was less than that expressed by the term mothers (P < 0.025). Volume tested as a covariate was insignificant, however, in the analyses between groups.
Discussion Higher protein nitrogen concentrations in preterm milk were sustained throughout the first two months of lactation. The mechanism responsible for the higher protein levels in preterm milk is not known, although the hormonal balance and metabolic regulation associated with shorter gestational periods possibly may alter protein synthesis. The lipid concentration and calorie content of the milk neither changed over time nor varied between groups. This is in accordance with Gross [17], who failed to find differences in lipid levels between PTM and TM, but at variante with observations by Anderson [l], who reported a 30% higher concentration of fat and 10-20% higher energy density in PTM. Gross energy was determined by bomb calorimetry in om study, whereas in other studies [1,17,26] energy values were calculated from the macronutrient composition, although Anderson [l] did verify the calculated values by bomb calorimetry on a subset of samples. The energy content of TM cited by Anderson was lower than usually reported levels for mature milk. The overall gross energy values for PT and TM were 66.1 and 67.7 kcal/dl in our study, and 68.0 and 58.5 kcal/dl in Anderson’s study [l]; the metabolizable energy values for PT and TM were 65.4 and 62.3 kcal/dl, respectively, in Gross’ study [17]. Although the mode of milk expression was similar between studies, samples collected by Anderson represented 24-h expressions in contrast to single, morning expressions in our study. Diurnal patterns of lipid concentration in human milk [24] indicate lower levels in the morning. The gestational ages studied by Anderson [l]
160
(26-33 wk) were younger than those in our study (30-37 wk), which may have had some unknown influence. The tempora1 trends in mineral concentration in PTM were similar to those observed previously in mature milk [22]. Sodium and zinc have been shown to decrease and calcium to remain constant in mature milk as lactation progressed [22]. In this study, unlike other reports of TM [6,17], Mg demonstrated a slight increase over time and P a slight decrease. The discrepancy probably is attributable to the longer period of observation in the present study. In contrast to previous reports that describe PTM [6,17,26], the calcium and phosphorus content was significantly lower than that of TM. Significantly lower concentrations of phosphorus were reported in PT milk by Lemons [20]. These low mineral levels in conjunction with high variability suggest that human milk may be an unreliable nutrient source for preterm infants and may indeed be growth limiting [9]. The premature infant fed breast milk exclusively for prolonged periods of time is particularly at risk of developing calcium- and phosphate-depletion [15,25]. Preterm milk failed to supply adequate calcium and phosphorus to sustain the intrauterine accretion rates of these minerals [5]. Hypophosphatemia developed in the infants fed preterm milk, and there was radiographic evidente of inadequate bone mineralization. Differences in the volumes of milk produced by women delivering at term and prematurely have been offered as an explanation for the observed compositional differences of their milks. The differences in PN, Ca, and P observed between PTM and TM were not due to differences in the volume of milk specimens. If there had been a compensatory mechanism to volume, one would expect the nutrient concentrations to have been altered in the same direction. This was not observed. Investigators studying the nutritional management of the premature infant have strived to achieve intrauterine growth rates and minimize metabolic stress. Tentative nutrient requirements have been suggested which are based on the theoretical premise that attainment of the intrauterine growth rate is desirable for optimal clinical outcome [29]. There was concern, however, that nutrient levels required to support such accelerated growth rates would tax the immature metabolic and excretory systems of these infants. Nitrogen retention rates similar to intrauterine accretion rates were attained, with apparent metabolic tolerante, by premature infants fed solely on preterm human milk [4,5,7] or formula [4,5,17]. Mature human milk failed to support intrauterine growth rates or nitrogen retention rates [4,16]. The feeding of human milk to VLBW infants has not been evaluated in terms of nutrient accretion rates beyond the first 6 weeks of life. Because of the declining concentrations of protein, sodium, phosphorus, and zinc in preterm milk, metabolic studies are needed to assess the limitations of human milk as an exclusive nutrient source for VLBW infants beyond the first few weeks of life. In addition, future studies should monitor the composition of weight gain, since there is ample evidente from animal studies [12,18,28] of marked alterations in nutrient retentions in animals fed imbalanced diets. This study has documented significantly higher concentrations of PN, but lower concentrations of Ca and P in PTM compared with TM. The feeding of VLBW
161
infants with PTM for prolonged periods of time requires careful scrutinization, not only in terms of growth, but also in the quality of that growth as wel1 as other developmental indices.
Acknowledgments The publication of this work is supported by a contract from the NICHD, DHEW No. l-HD-8-2-28, and USDA/ARS, Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children’s Hospital. The authors gratefully acknowledge the field supervision of J. Hopkinson, S. Post, and C. Heinz; the technical assistance of J. Sachen; the secretarial help of G. Quinones and M. Boyd; and editorial review by E.R. Klein.
References 1 Anderson, G.H.. Atkinson. S.A. and Bryan, M.H. (1981): Energy and macronutrient content of human milk during early lactation from mothers giving birth prematurely and at term. Am. J. Clin. Nutr., 34, 258-265. 2 Atkinson, S.A., Anderson, G.H. and Bryan, M.H. (1980): Human milk: Comparison of the nitrogen composition in milk from mothers of premature and full-term infants. Am. J. Clin. Nutr.. 33. 811-815. 3 Atkinson, S.A., Bryan. M.H. and Anderson, G.H. (1978): Human milk: Differente in nitrogen concentration in milk from mothers of term and premature infants. J. Pediatr., 93, 67-69. 4 Atkinson, S.A., Bryan, M.H. and Anderson, G.H. (1981): Human milk feeding in premature infants: Protein. fat, and carbohydrate balances in the first two weeks of life. J. Pediatr.. 99. 617-624. 5 Atkinson, S.A., Radde, I.C. and Anderson, G.H. (1983): Macromineral balances in premature infants fed their own mother’s milk or formula. J. Pediatr., 102, 99-106. 6 Atkinson, S.A., Radde, I.C., Chance. G.W., Bryan. M.H. and Anderson, G.H. (1980): Macromineral content of milk obtained during early lactation from mothers of premature infants. Early Hum. Dev., 4. 5-14. 7 Chessex, P., Reichman, B., Verellen. G., Putet. G., Smith, J.M.. Heim. T. and Swyer. P.R. (1983): Quality of growth in premature infants fed their own mother’s milk. J. Pediatr.. 102. 107-112. 8 Davis, D.P. (1977): Adequacy of expressed breast milk for early growth of preterm infants. Arch. Dis. Child.. 52. 296-301. 9 Day, G.M., Chance, G.W.. Radde, I.C., Reilly, B.J.. Park, E. and Sheepers, J. (1975): Growth and mineral metabolism in very low birthweight infants. 11. Effects of calcium supplementation on growth and divalent cations. Pediatr. Res. 9, 568-575. 10 Dixon, W.J. and Brown, M.B. (1979): Biomedical Computer Programs P-Series BMDP-79. Berkeley, University of California Press. 11 Dubowitz, L.M.S., Dubowitz. V. and Goldberg, C. (1970): Clinical assessment of gestational age in the newborn infant. J. Pediatr., 77, 1-10. 12 Filer. L.J. and Churella, H. (1963): Relationship of body composition. chemical maturation. homeostasis, and diet in the newborn mammal. Ann. N.Y. Acad. Sci., 110, 380-394. 13 Fiske. C.H. and Subbarow, Y. (1925): The colorimetric determination of phosphorus. J. Bio]. Chem.. 66. 375. 14 Fomon, S.J.. Ziegler, E.E. and Vazques, H.D. (1977): Human milk and the smal1 premature infant. Am. J. Dis. Child., 131, 463-467. 15 Greer, F.R., Steichen, J.J. and Tsang, R.C. (1982): Calcium and phosphate supplements in breast milk related rickets. Am. J. Dis. Child., 136, 581-583.
162
16 Cross. S.J. (1983): Growth and biochemical response of preterm infants fed human milk or modified infant formula. N. Engl. J. Med., 308, 237-241. 17 Cross, S.J., David, R.J., Bauman. L. and Tomarelli, R.M. (1980): Nutritional composition of milk produced by mothers delivering preterm. J. Pediatr.. 96. 641-644. 18 Henry. Y.. Gueguen, L. and Rerat, A. (1979): Influence of the leve1 of dietary phosphorus on the voluntary intake of energy and metabolic utilization of nutrients in the growing rat. Br. J. Nutr.. 42, 127-137. 19 Horwitz. W. (Ed.) Official Methods of Analysis of the Association of Official Analytical Chemist, 12 edn., Washington, DC. Association of Official Analytical Chemists. p. 22, 258. 20 Lemons. J.A., Moye. L., Hall. D. and Simmons. M. (1982): Differences in the composition of preterm and term human milk during early lactation. Pediatr. Res., 16, 113-117. 21 Lonnerdal. B., Forsum. E. and Hambraeus. L. (1976): The protein content of human milk 1. A transversal study of Swedish normal material. Nutr. Rep., 13, 125-134. 22 Macy. LG. (1949): Composition of human colostrum and milk. Am. J. Dis. Child., 78. 589-603. 23 Miller. D.S. and Payne, P.R. (1959): A ballistic bomb calorimeter. Br. J. Nutr., 13. 501-508. 24 Picciano. M.F. and Guthrie, H.A. (1976): Copper. iron. and zinc contents of mature human milk. Am. J. Clin. Nutr., 29. 242-254. 25 Rowe, J.C., Wood. D.H., Rowe, D.W. and Raisz, L.G. (1979): Nutritional hypophosphatemic rickets in a premature infant fed breast milk. N. Engl. J. Med., 300, 293-296. 26 Schanler. R.J. and Oh, W. (1980): Composition of breast milk obtained from mothers of premature infants as compared to breast milk obtained from donors. J. Pediatr.. 96, 679-681. 27 Usher, R. and McLean, F. (1969): Intrauterine growth of live-bom Caucasian infants at sea level: Standards obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J. Pediatr., 74, 901-910. 28 Weil, W.B. and Wallace, W.M. (1963): The effect of variable food intakes on growth and body composition. Ann. N.Y. Acad. Sci.. 110, 358-373. 29 Ziegler, E.E., Biga, R.L. and Fomon, S.J. (1981): Nutrient requirements of the premature infant. In: Textbook of pediatrie nutrition, pp. 29-39. Editor: R.M. Suskind. Raven Press, New York, NY.
9 (1984) 153-162
153
EHD 00533
Longitudinal changes in milk composition of mothers delivering preterm and term infants Nancy F. Butte, Cutberto Garza, Carmen A. Johnson, E. O’Brian Smith and Buford L. Nichols USDA/ARS
Chrldren’s Nutrition Research Center, Section of Nutritlon and Gastroenterolog,: of Pedratrics, Baylor College of Medicrne, Houston, TX 77030. U.S.A. Accepted
for publication
Department
18 August 1983
Summary The concentrations of protein nitrogen (PN), non-protein nitrogen (NPN), energy, fat, sodium (Na), calcium (Ca), phosphorus (P), magnesium (Mg), and zinc (Zn) were determined in human milk from mothers giving birth to full-term (n = 13) and preterm infants (n = 8). Milk samples were collected under controlled conditions at two-week intervals for 12 weeks postpartum. Statistically significant differences in PN, Ca, and P concentrations were detected between the milk from mothers of preterm and term infants. The mean PN concentration in the preterm milk was statistically higher than that of term milk (198 vs. 164 mg N/dl), in contrast to the lower mean Ca (220 vs. 261 mg/l) and P (125 vs. 153 mg/l) concentrations detected in the preterm milk. NO other differences in mean nutrient concentration were observed between the two groups. Concentrations of PN. NPN. Na, P. and Zn decreased over time. The concentration of Mg increased slightly. The content of fat, energy, and Ca did not change. milk composition;
preterm
infants;
term infants
Introduction Optimal nutritional management of low birthweight infants remains controversial. Although human milk is used to feed these infants, its adequacy has been questioned. Davies [8] found that mature human milk fed to preterm infants did not Address all correspondence to: Nancy Pediatrics, Baylor College of Medicine.
0378-3782/84/$03.00
F. Butte. Ph.D.. Section of Nutrition & GI. 1200 Moursund, Houston. TX 77030. U.S.A.
0 1984 Elsevier Science Publishers
B.V.
Department
of
154
support the intrauterine growth rate. Fomon et al. [14] concluded from theoretical calculations that the concentrations of protein, calcium, sodium. and probably other minerals in mature human milk were insufficient to support tissue accretion at the intrauterine rate. Recent investigations [l-3,6,17,20,26], however, have demonstrated that the composition of milk from mothers giving birth prematurely differed from that of mothers delivering at term. It has been suggested that preterm milk may be more apt to meet the nutrient needs of premature infants than term milk. Atkinson and co-workers [2] reported that the total nitrogen concentration of human milk was higher in preterm than term milk during the first four weeks of lactation. Gross et al. [17] have found higher concentrations of N, Na, and Cl in preterm milk compared to term milk, and similar levels of K, Ca, P, and Mg throughout the first 28 days of life. Although the Ca, P, and Mg requirements may not be satisfied, these investigators suggested that preterm milk may be more appropriate than pooled, mature breast milk for the feeding of premature infants. Atkinson et al. [6] suggested that the Na, Cl, K, Mg, N, and energy, but not the Ca and P content of preterm milk would be sufficient to meet the nutrient needs of the premature infant during the.first four weeks of life. Macromineral balance studies by the same authors have confirmed the inadequacy of mother’s milk to sustain intrauterine retention rates of Ca, P, and possibly Mg in the very low birthweight infant (VLBW) [5]. The present study was designed to test whether the differences in nutrient concentrations reported between preterm and term milk persisted beyond the four weeks heretofore studied.
Materials and Methods Study design Twenty-one women who elected to breast-feed their infants were assigned to one of two groups based on the maturity of their infants at birth. Gestational age was assessed according to Dubowitz [ll]. Infants whose gestational age was greater than or equal to 37 weeks were classified as term (n = 13) and those whose gestational age was less than 37 wk as preterm (n = 8). Milk samples were provided by the mothers at 2-week intervals during the first 12 weeks of life. Subjects Prospective mothers were recruited by referral from cooperating hospitals in the Houston area. In order to minimize confounding factors, subject selection was restricted to women 20-35 years of age, of parity not greater than two, whose consumption of carbonated beverages, toffee, tea, and alcohol did not exceed approximately two servings per day, and who were not on routine medication or steroid contraceptives. Infants were free of any congenital or acquired diseases. The participating women gave their informed consent to the study protocol which was approved by the Institutional Human Experimentation Committees. The characteristics of the women enrolled in the study are summarized in Table 1.
IS5 TABLE Maternal
I and infant characteristics
Maternal age (yr) Gravidity Pari ty Height (cm) Prepregnancy weight (kg) Infant sex Gestational age (wk) Birthweight (kg) Birthweight (percentile)
Preterm (n = 8)
Term(n
27.5 (3.0) a 1.5 (l-3)h 0 159.7 (4.2) d 52.9 (5.2)’ 5M/3F 33.9 (2.3) (30-36) 1.92 (0.70) a 43.2 (37.4)’
26.6 (5.0) ’ 1.3 (l-2)h 0.4 (O-l)h 162.3 (5.5) ’ 54.9 (6.9)’ 8M/5F 39.2 (1.4) (37-42) 2.99 (0.46) * 27.5 (24.1) ’
’
=13)
‘
’ Mean (S.D.). h Mean (range). ’ Mean (S.D.) (range).
Throughout the study. maternal health was good by history. Any incidence of hypertensive disease, diabetes, anemia, and/or chronic medications was recorded when health screening was done during recruitment. Two preterm mothers were hypertensive during pregnancy. As ascertained by interview, al1 but two of the women were taking supplemental vitamin-mineral tablets at the time of the study. Smoking (approximately 1 pack per day) was reported by four of the term women. Infants are described by sex, gestational age, and birthweight in Table 1. There was a slight preponderance of males over females in both groups. Birthweight, corrected for gestational age, was assessed according to Usher and McLean intrauterine growth standards [27]. Raw weight was transformed into a percentile weight-for-gestational age. The majority of the infants, once at home, were breast-fed exclusively. In a few cases supplemental juice and cereal were introduced at approximately 6-8 weeks of life. The high-risk preterm infants who were retained in the nursery received formula and/or human milk depending on their physician’s orders; their mothers expressed breast milk at home for the purpose of this study and to maintain an adequate milk supply to feed their infants after discharge. The progress of the preterm infants thus reflected mixed feeding. The relative proportions of human milk and artificial formula were not recorded. Milk collection Samples of human milk were collected at the end of the 2nd, 4th, 6th, 8th. 10th. and 12th week post-partum under controlled conditions. Milk samples were collected between 8 a.m. and 12 noon with the use of an Egnell breast pump (Egnell, Inc., Cary, IL). The samples were collected at least 2 h after a feeding. The entire contents of one breast were expressed into sterile, acid-washed polypropylene bottles and kept at 4°C until delivered to the laboratory (within 4 h of collection). Immediately thereafter, volumes were noted and the samples stored at -20°C for analysis.
156
Biochemical analysis Nitrogen (N) was analyzed by the Kjeldahl method before and after trichloroacetic acid (24%; 1 : 1 volume) precipitation of protein; protein nitrogen was determined on the solubilized precipitant [21], and non-protein nitrogen (NPN) was estimated from the differente between total and protein nitrogen. The content of fat was determined by the Roese-Gottlieb extraction method [19. p. 2581. The content of Na, Ca, and Mg was estimated on dry-ashed samples by atomic absorption (AA) spectrophotometry [19, p. 221. Zinc levels were determined by an AA spectrophotometer equipped with a graphite furnace [19, p. 221. Phosphorus was determined colorimetrically on dry-ashed samples by a modification of Fiske’s method [13]. The energy content was determined by bomb calorimetry [23]. Al1 analyses were performed in duplicate. Statistical analysis Nutrient concentrations were subjected to analysis of variante and covariance with repeated measures (BMDPLV) [lol. Differences between preterm and term milk were examined over the 12 weeks of lactation with weight-for-gestational age and volume of milk specimens treated as covariates. Correlations between gestational age and nutrient concentration were examined at each week of observation.
Results The macronutrient content of the human milk samples is presented in Table 11. The protein nitrogen concentration decreased significantly over the twelve weeks of observation at a similar rate (6.3 mg/dl per wk) in both groups (P < 0.0001). Although the rate of change did not differ between groups, there was a significant differente in the overall mean nitrogen concentration between preterm milk (PTM) (198 + 32 mg/dl) (k S.D.) and term milk (TM) (164 -t 25 mg/dl) samples (P < 0.002). Differences between groups in PN content were statistically significant (P < 0.05) throughout the first 2 months of observation. The NPN content of human milk did not differ between groups and decreased significantly over time at a mean rate of 1.3 mg/dl per wk. The concentration of fat did not change significantly over time and no statistically significant differences in overall means were demonstrated between groups. Similarly, no statistically significant differences were seen in the energy concentration over time or between groups. The mineral content of the milk is displayed in Table 111. The calcium content of the milk did not change significantly over the 12 weeks of lactation. The overall concentration of sodium, phosphorus, and zinc decreased significantly over time and the magnesium concentration increased (P < 0.001). The mean calcium concentration of the TM (261 + 36 mg/l) was significantly higber (P = 0.05) than that of PTM (220 f 62 mg/l). The mean phosphorus leve1 of 153 + 30 mg/l for the TM was significantly higher than the 125 f 33 mg/l for the PTM (P < 0.03). The Ca/P ratio increased from 1.5 at 2 weeks to 2.1 at 12 weeks in
11
(33)
56.0
(29)
207.0
(1.00)
nitrogen.
L Overall
’ NPN = non-protein
means differ (P < 0.025).
means differ (P < 0.002).
h Overall
(S.D.).
(9.7)
68.3
a Mean
(9.3)
63.5
Preterm
Term
Energy (kcal/dl)
(0.94)
3.45
3.79
Term
(12)
55.0
Fat (g/dl) Preterm
(16)
57.0
Preterm
Term
NPN (mg/dl)’
(18)
241.0
Preterm
Term
nitrogen (mg/dl)
Protein
(33) *
53.0
Term
(ml)
2
(10.3) (6.6)
67.9 68.2
(1.24) (0.78)
4.04
(18)
54.0
3.95
(17)
50.0
(14) (26)
(26)
64.0
173.0
(16)
39.0
content
211.0
4
and macronutrient
Weeks post-partum
of milk expressed
Preterm
Volume
The volume
TABLE
69.0
54.0
69.8 71.4
(14.8)
( 11.7)
4.77
(10.2)
65.9
64.6
4.51
(12.2)
4.22
(1.06)
45.0
49.0
146.0
173.0
68.0
55.0
10
(1.28)
(15)
4.12
(15)
50.0
(24)
59.0
151.0
66.7
(1.18)
(29)
68.0
(39)
(67)
62.0
189.0
8
(8.6)
( 13.6)
(1.17)
(1.33)
(13)
(25)
(26)
(42)
(32)
(44)
66.0
63.0
4.24
3.82
37.0
45.0
145.0
169.0
66.0
61.0
12
by mothers of preterm and term infants
68.0
(1.18)
4.50
(11)
3.96
(17)
52.0
(16)
(38)
(24)
(32)
46.0
162.0
203.0
6
of human milk produced
(12.8)
(12.7)
(1.57)
(1.45)
(14)
(14)
(26)
(28)
(30)
(28)
67.7
66.1
4.31
3.92
48.0
52.0
164.0
198.0
65.0
54.0
mean
Overall
(10.1)
(12.3)
(1.15)
(1.25)
(14)
(18)
(25)’
(32) c
(28)’
(39) h
111
3.4
(0.8)
(1.0)
’ Overall means differ (P < 0.03).
h Overall means differ (P = 0.05).
a Mean (S.D.).
4.1
Preterm
Term
Zinc
266.0 (98)
220.0 (77)
(8)
(8)
Term
33.0
Sodium Preterm
36.0
179.0 (36)
Term
Term
149.0 (38)
Phosphorus Preterm
Magnesium Preterm
214.0 (48) ’
255.0 (53)
Preterm
2
Weeks post-partum
Term
Calcium
(mS/L)
Content
(7) (6)
2.9
3.2 (0.9)
(0.6)
184.0 (54)
165.0 (34)
31.0
36.0
164.0 (25)
132.0 (36)
254.0 (52)
(9)
2.1
2.7 (0.9)
(1.1)
173.0 (65)
149.0 (30)
35.0
40.0 (11)
150.0 (32)
128.0 (33)
267.0 (24)
218.0 (62)
6
(9)
1.9
2.9
(0.6)
(0.6)
153.0 (47)
131.0 (30)
36.0
42.0 (13)
148.0 (20)
135.0 (31)
258.0 (22)
236.0 (66)
8
by mothers of preterm and term infants
213.0 (71)
4
Mineral content of human milk produced
TABLE
(9)
1.8
2.0
(1.0)
(0.9)
150.0 (49)
135.0 (37)
38.0
41.0 (10)
142.0 (34)
102.0 (27)
270.0 (25)
215.0 (61)
10
(9)
1.4
1.6
(0.7)
(0.3)
130.0 (41)
150.0 (44)
39.0 (10)
42.0
136.0 (27)
104.0 (30)
260.0 (26)
223.0 (65)
12
(8)
2.2
2.8
(0.8)
(0.8)
168.0 (57)
166.0 (51)
35.0
40.0 (10)
153.0 (30) c
125.0 (33)’
261.0 (36) h
220.0 (62) ’
mean
Overall
both groups due to the constant calcium and decreasing phosphorus concentrations. Magnesium increased (P < 0.001) slightly from an initial 34 mg/l at 2 weeks to 40 mg/l at 12 weeks. NO differences in the mean concentrations or rates of rise were seen between groups. The sodium concentrations for PTM and TM were not statistically different. Sodium decreased at a mean rate of 9.4 mg/l per wk. Zinc concentration decreased steadily over time at a rate of 0.2 mg/l per wk. The groups did not differ statistically with respect to the overall mean content or rate of change in concentration of zinc. The covariate, weight-for-gestational age, was not significant in the analyses between groups. The protein nitrogen concentration was correlated negatively to gestational age (r range = - 0.47 to -0.79, P < 0.05). Statistically significant correlations were not demonstrated between gestational age and any of the other nutrients. Variability in nutrient concentrations was evident in the scatter of individual values at each time interval. Zinc was particularly variable (CU = 0.46), followed by sodium (CU = 0.39). NPN (CU= 0.33) and fat (CU= 0.28). The coefficient of variation of the concentrations of al1 other nutrients was approximately 20%. The overall mean volume of the milk specimens provided by the preterm mothers was less than that expressed by the term mothers (P < 0.025). Volume tested as a covariate was insignificant, however, in the analyses between groups.
Discussion Higher protein nitrogen concentrations in preterm milk were sustained throughout the first two months of lactation. The mechanism responsible for the higher protein levels in preterm milk is not known, although the hormonal balance and metabolic regulation associated with shorter gestational periods possibly may alter protein synthesis. The lipid concentration and calorie content of the milk neither changed over time nor varied between groups. This is in accordance with Gross [17], who failed to find differences in lipid levels between PTM and TM, but at variante with observations by Anderson [l], who reported a 30% higher concentration of fat and 10-20% higher energy density in PTM. Gross energy was determined by bomb calorimetry in om study, whereas in other studies [1,17,26] energy values were calculated from the macronutrient composition, although Anderson [l] did verify the calculated values by bomb calorimetry on a subset of samples. The energy content of TM cited by Anderson was lower than usually reported levels for mature milk. The overall gross energy values for PT and TM were 66.1 and 67.7 kcal/dl in our study, and 68.0 and 58.5 kcal/dl in Anderson’s study [l]; the metabolizable energy values for PT and TM were 65.4 and 62.3 kcal/dl, respectively, in Gross’ study [17]. Although the mode of milk expression was similar between studies, samples collected by Anderson represented 24-h expressions in contrast to single, morning expressions in our study. Diurnal patterns of lipid concentration in human milk [24] indicate lower levels in the morning. The gestational ages studied by Anderson [l]
160
(26-33 wk) were younger than those in our study (30-37 wk), which may have had some unknown influence. The tempora1 trends in mineral concentration in PTM were similar to those observed previously in mature milk [22]. Sodium and zinc have been shown to decrease and calcium to remain constant in mature milk as lactation progressed [22]. In this study, unlike other reports of TM [6,17], Mg demonstrated a slight increase over time and P a slight decrease. The discrepancy probably is attributable to the longer period of observation in the present study. In contrast to previous reports that describe PTM [6,17,26], the calcium and phosphorus content was significantly lower than that of TM. Significantly lower concentrations of phosphorus were reported in PT milk by Lemons [20]. These low mineral levels in conjunction with high variability suggest that human milk may be an unreliable nutrient source for preterm infants and may indeed be growth limiting [9]. The premature infant fed breast milk exclusively for prolonged periods of time is particularly at risk of developing calcium- and phosphate-depletion [15,25]. Preterm milk failed to supply adequate calcium and phosphorus to sustain the intrauterine accretion rates of these minerals [5]. Hypophosphatemia developed in the infants fed preterm milk, and there was radiographic evidente of inadequate bone mineralization. Differences in the volumes of milk produced by women delivering at term and prematurely have been offered as an explanation for the observed compositional differences of their milks. The differences in PN, Ca, and P observed between PTM and TM were not due to differences in the volume of milk specimens. If there had been a compensatory mechanism to volume, one would expect the nutrient concentrations to have been altered in the same direction. This was not observed. Investigators studying the nutritional management of the premature infant have strived to achieve intrauterine growth rates and minimize metabolic stress. Tentative nutrient requirements have been suggested which are based on the theoretical premise that attainment of the intrauterine growth rate is desirable for optimal clinical outcome [29]. There was concern, however, that nutrient levels required to support such accelerated growth rates would tax the immature metabolic and excretory systems of these infants. Nitrogen retention rates similar to intrauterine accretion rates were attained, with apparent metabolic tolerante, by premature infants fed solely on preterm human milk [4,5,7] or formula [4,5,17]. Mature human milk failed to support intrauterine growth rates or nitrogen retention rates [4,16]. The feeding of human milk to VLBW infants has not been evaluated in terms of nutrient accretion rates beyond the first 6 weeks of life. Because of the declining concentrations of protein, sodium, phosphorus, and zinc in preterm milk, metabolic studies are needed to assess the limitations of human milk as an exclusive nutrient source for VLBW infants beyond the first few weeks of life. In addition, future studies should monitor the composition of weight gain, since there is ample evidente from animal studies [12,18,28] of marked alterations in nutrient retentions in animals fed imbalanced diets. This study has documented significantly higher concentrations of PN, but lower concentrations of Ca and P in PTM compared with TM. The feeding of VLBW
161
infants with PTM for prolonged periods of time requires careful scrutinization, not only in terms of growth, but also in the quality of that growth as wel1 as other developmental indices.
Acknowledgments The publication of this work is supported by a contract from the NICHD, DHEW No. l-HD-8-2-28, and USDA/ARS, Children’s Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children’s Hospital. The authors gratefully acknowledge the field supervision of J. Hopkinson, S. Post, and C. Heinz; the technical assistance of J. Sachen; the secretarial help of G. Quinones and M. Boyd; and editorial review by E.R. Klein.
References 1 Anderson, G.H.. Atkinson. S.A. and Bryan, M.H. (1981): Energy and macronutrient content of human milk during early lactation from mothers giving birth prematurely and at term. Am. J. Clin. Nutr., 34, 258-265. 2 Atkinson, S.A., Anderson, G.H. and Bryan, M.H. (1980): Human milk: Comparison of the nitrogen composition in milk from mothers of premature and full-term infants. Am. J. Clin. Nutr.. 33. 811-815. 3 Atkinson, S.A., Bryan. M.H. and Anderson, G.H. (1978): Human milk: Differente in nitrogen concentration in milk from mothers of term and premature infants. J. Pediatr., 93, 67-69. 4 Atkinson, S.A., Bryan, M.H. and Anderson, G.H. (1981): Human milk feeding in premature infants: Protein. fat, and carbohydrate balances in the first two weeks of life. J. Pediatr.. 99. 617-624. 5 Atkinson, S.A., Radde, I.C. and Anderson, G.H. (1983): Macromineral balances in premature infants fed their own mother’s milk or formula. J. Pediatr., 102, 99-106. 6 Atkinson, S.A., Radde, I.C., Chance. G.W., Bryan. M.H. and Anderson, G.H. (1980): Macromineral content of milk obtained during early lactation from mothers of premature infants. Early Hum. Dev., 4. 5-14. 7 Chessex, P., Reichman, B., Verellen. G., Putet. G., Smith, J.M.. Heim. T. and Swyer. P.R. (1983): Quality of growth in premature infants fed their own mother’s milk. J. Pediatr.. 102. 107-112. 8 Davis, D.P. (1977): Adequacy of expressed breast milk for early growth of preterm infants. Arch. Dis. Child.. 52. 296-301. 9 Day, G.M., Chance, G.W.. Radde, I.C., Reilly, B.J.. Park, E. and Sheepers, J. (1975): Growth and mineral metabolism in very low birthweight infants. 11. Effects of calcium supplementation on growth and divalent cations. Pediatr. Res. 9, 568-575. 10 Dixon, W.J. and Brown, M.B. (1979): Biomedical Computer Programs P-Series BMDP-79. Berkeley, University of California Press. 11 Dubowitz, L.M.S., Dubowitz. V. and Goldberg, C. (1970): Clinical assessment of gestational age in the newborn infant. J. Pediatr., 77, 1-10. 12 Filer. L.J. and Churella, H. (1963): Relationship of body composition. chemical maturation. homeostasis, and diet in the newborn mammal. Ann. N.Y. Acad. Sci., 110, 380-394. 13 Fiske. C.H. and Subbarow, Y. (1925): The colorimetric determination of phosphorus. J. Bio]. Chem.. 66. 375. 14 Fomon, S.J.. Ziegler, E.E. and Vazques, H.D. (1977): Human milk and the smal1 premature infant. Am. J. Dis. Child., 131, 463-467. 15 Greer, F.R., Steichen, J.J. and Tsang, R.C. (1982): Calcium and phosphate supplements in breast milk related rickets. Am. J. Dis. Child., 136, 581-583.
162
16 Cross. S.J. (1983): Growth and biochemical response of preterm infants fed human milk or modified infant formula. N. Engl. J. Med., 308, 237-241. 17 Cross, S.J., David, R.J., Bauman. L. and Tomarelli, R.M. (1980): Nutritional composition of milk produced by mothers delivering preterm. J. Pediatr.. 96. 641-644. 18 Henry. Y.. Gueguen, L. and Rerat, A. (1979): Influence of the leve1 of dietary phosphorus on the voluntary intake of energy and metabolic utilization of nutrients in the growing rat. Br. J. Nutr.. 42, 127-137. 19 Horwitz. W. (Ed.) Official Methods of Analysis of the Association of Official Analytical Chemist, 12 edn., Washington, DC. Association of Official Analytical Chemists. p. 22, 258. 20 Lemons. J.A., Moye. L., Hall. D. and Simmons. M. (1982): Differences in the composition of preterm and term human milk during early lactation. Pediatr. Res., 16, 113-117. 21 Lonnerdal. B., Forsum. E. and Hambraeus. L. (1976): The protein content of human milk 1. A transversal study of Swedish normal material. Nutr. Rep., 13, 125-134. 22 Macy. LG. (1949): Composition of human colostrum and milk. Am. J. Dis. Child., 78. 589-603. 23 Miller. D.S. and Payne, P.R. (1959): A ballistic bomb calorimeter. Br. J. Nutr., 13. 501-508. 24 Picciano. M.F. and Guthrie, H.A. (1976): Copper. iron. and zinc contents of mature human milk. Am. J. Clin. Nutr., 29. 242-254. 25 Rowe, J.C., Wood. D.H., Rowe, D.W. and Raisz, L.G. (1979): Nutritional hypophosphatemic rickets in a premature infant fed breast milk. N. Engl. J. Med., 300, 293-296. 26 Schanler. R.J. and Oh, W. (1980): Composition of breast milk obtained from mothers of premature infants as compared to breast milk obtained from donors. J. Pediatr.. 96, 679-681. 27 Usher, R. and McLean, F. (1969): Intrauterine growth of live-bom Caucasian infants at sea level: Standards obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J. Pediatr., 74, 901-910. 28 Weil, W.B. and Wallace, W.M. (1963): The effect of variable food intakes on growth and body composition. Ann. N.Y. Acad. Sci.. 110, 358-373. 29 Ziegler, E.E., Biga, R.L. and Fomon, S.J. (1981): Nutrient requirements of the premature infant. In: Textbook of pediatrie nutrition, pp. 29-39. Editor: R.M. Suskind. Raven Press, New York, NY.