点击显示 收起
1 From the Department of Nutrition and Food Science (CPG, MD, and PBM-V) and the Biometrics Program, Department of Animal and Avian Sciences (LWD), University of Maryland, College Park, and the School of Public Health and Hygiene, Johns Hopkins University, Baltimore (SY).
2 Supported in part by Maryland Agricultural Experiment Station grant NFSC-96-39 and the US Department of Agriculture, Agricultural Research Service. Hoffmann-La Roche (Nutley, NJ) provided the placebo and ß-carotene capsules. 3 Address reprint requests to PB Moser-Veillon, Department of Nutrition and Food Science, University of Maryland, College Park, MD 20742. E-mail: pv6{at}umail.umd.edu.
ABSTRACT
Background: Information is lacking regarding normal changes in milk carotenoid concentrations in healthy, well-nourished women during the first month of lactation.
Objectives: This study investigated milk carotenoid concentrations during days 4–32 postpartum and assessed the effects of maternal ß-carotene supplementation.
Design: Subjects (n = 21; aged 19–39 y) were randomly assigned to receive ß-carotene (30 mg/d) or placebo from days 4 to 32 postpartum. Each subject provided 8 diet records and 8 milk samples during the study. Diet records were analyzed for energy, macronutrients, vitamins A and E, and carotenoids. Milk samples were analyzed with HPLC for concentrations of carotenoids, retinol, and -tocopherol. Data were analyzed by using repeated-measures analysis and orthogonal contrasts.
Results: No significant differences in average dietary intakes, body mass index, age, or parity were found between groups at baseline or after supplementation. Milk carotenoid concentrations decreased over time (P < 0.01), as did retinol and -tocopherol concentrations (P < 0.003). Concentrations of most carotenoids decreased to those reported for mature milk by day 32 postpartum. Milk lutein concentrations remained elevated throughout the study compared with values reported for mature milk, whereas plasma lutein concentrations decreased significantly over time. ß-carotene supplementation did not significantly change the milk concentrations of ß-carotene, the other carotenoids, retinol, or -tocopherol.
Conclusions: The lack of increase in milk ß-carotene despite supplementation suggests that transitional milk may be already nearly saturated with ß-carotene. The elevated milk lutein concentration and simultaneous decrease in plasma lutein suggest that lutein metabolism may be altered during early lactation.
Key Words: ß-carotene • carotenoids • lutein • retinol • -tocopherol • lactation • breast-feeding • women • breast milk • human milk
INTRODUCTION
Breast milk from healthy, well-nourished women contains virtually all the nutrients necessary for newborn infants and also contains a variety of growth and immune factors. Investigation of its chemical composition can provide information relevant to infant nutritional needs and also maternal nutritional requirements during lactation. In addition, relations between maternal dietary intake and plasma and milk composition may reflect aspects of nutrient metabolism, because many of the constituents of breast milk are derived from maternal blood. Finally, it is crucial to understand the effects of maternal nutrient supplementation on the composition of breast milk as more pregnant and lactating women are receiving supplements through national and international public health campaigns.
Several cross-sectional studies have described the carotenoid composition of colostrum (1–3) and mature breast milk (4–9). Two longitudinal studies measured total carotenoid concentrations during the first 5 d (10) and the first few weeks (11) of lactation. Only a few studies have examined the response to ß-carotene supplementation during well-established lactation (12–15). Studies published before the development of HPLC separation techniques reported only total carotenoid concentration (1, 4, 5), whereas later studies used sophisticated HPLC separations to report detailed information regarding concentrations of specific carotenoids. In general, interpretation has been limited by small sample size (13), lack of maternal dietary data (14), or omission of milk fat concentration as a variable (6, 8, 14).
The present study was part of a larger study designed to investigate the effects of ß-carotene supplementation on immune factors in breast milk. This article reports on the carotenoid profile of early human milk and the effect of maternal ß-carotene supplementation.
SUBJECTS AND METHODS
Subjects
Twenty-one pregnant women were recruited during their last trimester, primarily through midwifery clinics in the Washington, DC, metropolitan area. Women were eligible to join the study if they had previously breast-fed at least one infant, did not smoke, had not taken prenatal supplements containing ß-carotene, and did not plan to use oral contraceptives during the study period. The subjects provided informed consent after the study protocol was approved by the University of Maryland Institutional Review Board.
Treatment
Subjects were randomly assigned to 4 wk of supplementation with either ß-carotene (30 mg/d; n = 11) or placebo (n = 10) beginning on day 4 postpartum (day 0 of the study). Subjects received gelatin capsules containing beadlets with or without 10% ß-carotene (by weight) in a water-dispersible preparation; the placebo and ß-carotene capsules looked alike. Subjects consumed the supplements with breakfasts that included 250 mL whole milk or yogurt to facilitate absorption. Plasma ß-carotene concentrations were determined to monitor compliance.
Dietary data and sample collections
Subjects completed 24-h estimated diet records on study days -1, 2, 5, 8, 11, 14, 19, and 26. They also provided blood samples at baseline (day 0) and on the final day of supplementation (day 27); subjects were not fasting when blood was obtained. Blood collection is described elsewhere (16). Subjects provided milk samples on study days 0, 3, 6, 9, 12, 15, 20, and 27. They used electric breast pumps, and the same breast was pumped for all 8 samples. Subjects emptied one breast, gently mixed the contents, and transferred 10 mL milk into a plastic tube. The tube was then capped and kept in the dark on ice for up to 4 h. In the laboratory, 4 mL whole milk was transferred to a plastic storage tube and stored at -70°C until carotenoid processing.
Analyses
Diet records were analyzed with NUTRITIONIST IV (N-Squared Computing, Salem, OR) to estimate intakes of energy, carbohydrate, fat, protein, and vitamins A and E. Intakes of lutein + zeaxanthin, ß-cryptoxanthin, lycopene, -carotene, and ß-carotene were calculated by using the US Department of Agriculture–National Cancer Institute carotenoid database (17). To verify supplement dosages, 2 technicians, working independently, extracted the capsule contents with hexane and analyzed the ß-carotene content and purity with HPLC methods (18). Plasma analysis was described previously (16).
We used the creamatocrit method to estimate milk fat concentration (19). Frozen samples of whole milk were gently agitated on a mechanized shaker while warming to room temperature. A small sample was then drawn into a glass capillary tube and centrifuged to separate the fat and aqueous layers. A digital caliper was used to determine the volume of fat as a percentage of the total milk volume. This percentage was then used to extrapolate the concentration of fat in the milk according to the equation developed by Lucas et al (19).
Milk samples were analyzed to determine the concentrations of lutein + zeaxanthin, ß-cryptoxanthin, lycopene, -carotene, ß-carotene, retinol, and -tocopherol. All analyses were conducted under subdued lighting to avoid degradation of the carotenoids. Samples were allowed to thaw and come to room temperature on a shaker to prevent separation of the fat and aqueous phases. Ethanol was added to precipitate the proteins. Potassium hydroxide (40% in methanol) was added and samples were sonicated at 45°C for 30 min to digest the triacylglycerol and release the carotenoids. An internal standard (B-apo-8'carotenal; Fluka Chemical Corp, Ronkonkoma, NY) was added and samples were extracted twice with hexane. Combined extracts were washed and dried under nitrogen and then reconstituted in the mobile phase. Samples were analyzed with a Beckman System Gold pump (Beckman Instruments, Fullerton, CA) coupled to 2 scanning detectors (Module 166; Hewlett-Packard, Palo Alto, CA) and an automated 717 Plus Autosampler (Waters Associated, Milford, MA). A guard column (40 x 4.6 mm; Altech Associates, Guelph, Canada) preceded a Spherisorb column (3-µm particle size, 150 x 4.6 mm; Altech Associates). Samples were eluted isocratically with a mobile phase of acetonitrile:dioxane:methanol (83:13:4) and 0.01% triethylamine. Ammonium acetate (0.15 mol/L) was added to the methanol. Carotenoids, retinol, and -tocopherol were detected at 452, 325, and 295 nm, respectively. Standard curves were generated. Pooled serum and milk samples (in lieu of standard reference milk) were analyzed each day to check for precision. Standard reference serum (SRM 968-B; National Institute of Standards and Technology, Gaithersburg, MD) was analyzed periodically to verify accuracy. Milk carotenoid, retinol, and -tocopherol concentrations were calculated both per volume and per gram of milk lipid. In both plasma and milk analysis, lutein and zeaxanthin could not be completely resolved and were summed. Hereafter, all references to plasma or milk lutein concentrations refer to lutein + zeaxanthin.
Statistical methods
Data were analyzed with SAS, version 6.12 (SAS Institute Inc, Cary, NC). The mixed-model procedure accommodated unequal replication, heterogeneous variances, and correlated repeated measures. Plasma analysis included only 2 repeated measures (days 4 and 32 postpartum) but there were 8 repeated measures for nutrient and carotenoid intakes (days 3, 6, 9, 12, 15, 23, and 31 postpartum) and milk composition (days 4, 7, 10, 13, 16, 19, 25, and 32 postpartum). The model included the fixed effects of supplementation (placebo or ß-carotene), duration of the study, and the time-by-treatment interaction. Subject variation within supplementation groups and the residual variation were defined as random. Goodness-of-fit criteria provided in the mixed procedure were used to choose a variance and correlation structure that adequately described the random variation. Treatment means within time were compared by using t distribution probabilities. Time means within treatment were compared by using contrasts adjusted for repeated measures and a Bonferroni correction to control for experiment-wise error rates. Unless otherwise stated, statistical significance was set at P < 0.05.
RESULTS
All 21 women who enrolled in the study completed it. Subjects included 2 African American, 2 Hispanic, and 17 European American women. Most were middle-class and well educated; 3 had earned graduate degrees, 1 was a nurse, and 1 was a physician. There were no significant differences in body mass index, age, or parity between the 2 groups of women (Table 1). All but one of the women who began the study with low hemoglobin concentrations had normal concentrations by the end of the study. There were no significant differences in hemoglobin concentrations, either at baseline or postsupplementation, between the groups.
View this table:
TABLE 1 . Baseline characteristics of lactating women on day 4 postpartum1
Diet records were incomplete for one subject in each group; those records were deleted from the analyses. After we determined that there were no significant changes in nutrient intakes over time, intakes were calculated as the mean of the eight 24-h diet records (Table 2). There were no significant differences in mean intakes of energy, carbohydrate, fat, protein, vitamins A or E, lutein, ß-cryptoxanthin, lycopene, -carotene, or ß-carotene between the groups.
View this table:
TABLE 2 . Estimated daily dietary intakes in lactating women on days 4–32 postpartum1
The results of HPLC analysis showed that the ß-carotene supplement capsules contained 31–35 mg ß-carotene and an average of 1.2 mg -carotene. As we reported previously, plasma - and ß-carotene increased in the supplemented group but did not change significantly in the placebo group (16). ß-Carotene supplementation did not alter plasma concentrations of lycopene or ß-cryptoxanthin significantly (16). We also reported and discussed previously the significant decrease in plasma lutein concentrations in both the ß-carotene and the placebo groups (P < 0.0001 and P < 0.02, respectively; Figure 1) (16). Lutein intake did not change significantly during the study period (Figure 2).
FIGURE 1. . Mean (±SEM) plasma lutein concentrations in lactating women receiving placebo (; n = 10) or ß-carotene supplements (; n = 11) for 28 d. The main effect of time was significant (P < 0.05) in both groups but the treatment effect and the time-by-treatment interaction were not significant.
FIGURE 2. . Mean (±SEM) lutein intakes in lactating women receiving placebo (___; n = 9) or ß-carotene supplements (- - -; n = 10) for 28 d. There were no significant effects of time or diet (estimated dietary lutein intake).
Estimates of milk fat concentration in individual samples varied widely (range: 2.8–91.8 g/L). Because mean values did not change significantly over time and were not affected by ß-carotene supplementation, all data points were combined. The overall mean (±SE) milk fat concentration was 34.2 ± 1.19 g/L, which agrees with other published reports (4, 19). There were no significant overall effects of ß-carotene supplementation on milk concentrations of lutein, ß-cryptoxanthin, lycopene, or -carotene; therefore, overall mean values averaged for each time point (both groups combined) are shown in Figure 3. Milk concentrations of all 4 carotenoids decreased over time (P < 0.01 for all). Milk concentrations of retinol and -tocopherol were also unaffected by ß-carotene supplementation and decreased significantly over time (P < 0.005). Mean (±SD) retinol concentrations were 4944 ± 539 µmol/L initially and 2079 ± 207 µmol/L at the end of the study. -Tocopherol concentrations were 31 ± 4.6 µmol/L intially and 9.4 ± 1.2 µmol/L at the end of the study.
FIGURE 3. . Mean (±SEM) breast milk carotenoid concentrations over days 4–32 postpartum (n = 21). The decreases in these concentrations over time were significant for all 4 carotenoids (P < 0.01).
Milk ß-carotene concentrations tended to decrease over time in the placebo group, but this effect was NS (Figure 4). There was a significant time-by-treatment interaction for milk lutein concentration (P < 0.04) (data not shown); however, the interaction was no longer significant after we controlled for milk fat concentration.
FIGURE 4. . Mean (±SEM) breast milk ß-carotene concentrations in women receiving placebo (___; n = 10) or ß-carotene supplements (- - -; n = 11). There were no significant effects of time or supplementation.
DISCUSSION
This is the first longitudinal study to report milk carotenoid concentrations and estimated carotenoid intake during the first month of lactation and to investigate the effect of ß-carotene supplementation during this physiologic period. Average dietary intakes of energy, vitamin A, vitamin E, and carotenoids were published and discussed elsewhere (16). Lutein intake in the placebo group was higher than typical US intakes, as was the variability (6, 12). These data were driven primarily by one subject who reported consuming kale on most of the days she kept diet records. Because her plasma and milk lutein concentrations were also high, her data were judged to be consistent and were included in the statistical analyses.
The data indicate that the carotenoid profile of human milk appears to have nearly stabilized at 1 mo postpartum. Concentrations of -carotene, lycopene, and ß-cryptoxanthin all decreased significantly by the end of the study, reaching concentrations reported previously for mature milk (6–8, 12, 13). In the placebo group, milk concentrations of ß-carotene appeared to decrease, but we lacked the statistical power to show significance. Furthermore, these concentrations were equivalent to 5–10% of plasma concentrations, in good agreement with an earlier report (13). In contrast, milk lutein appeared to remain elevated when compared with concentrations published for mature milk (6–8, 12, 13). In addition, the milk lutein concentration was equivalent to 30% of the plasma lutein concentration. Because of the elevated lutein concentrations, ß-carotene represented a smaller proportion (<25%) of total milk carotenoids at 1 mo postpartum than that which was reported for mature milk (35%) (13). Lutein represented 25% of total milk carotenoids on day 4 postpartum, but represented nearly 50% on day 32 postpartum. This elevated milk lutein concentration contrasts sharply with the decrease in plasma lutein that we reported previously (16). This suggests that the flow of lutein into milk may account for the postpartum decrease in plasma lutein concentrations.
As yet, the role or roles of carotenoids in breast milk are not clearly understood. At least one report suggests that the extremely high concentrations transferred to the nursing infant during the first few days of lactation correct abnormally low plasma ß-carotene concentrations in the neonate (10). Lutein was the only carotenoid that remained elevated in the milk of these subjects for >4 wk, leading to speculation about its potential role for the infant. It has been proposed that lutein protects the retina of the eye from light-induced oxidative damage. It is therefore conceivable that maternal milk preferentially provides a higher concentration of lutein during the first months of life to accommodate the newborn's transition from the total darkness experienced in utero.
Other researchers have reported that ß-carotene supplementation of lactating women increases milk ß-carotene concentrations (12, 13, 15). In 2 of the studies, however, subjects were 1 mo postpartum, representing well-established lactation status with predictably lower milk ß-carotene concentrations. In the third study, subjects began the study with marginal malnutrition. In the current study, ß-carotene supplementation had no effect on milk ß-carotene concentrations; rather, milk ß-carotene concentrations were maintained at their initial values (day 4 postpartum) values (Figure 4). These concentrations were similar to that reported for transitional milk (1). The lack of increase after supplementation suggests that the high ß-carotene concentrations reported for colostrum and transitional milk represent ß-carotene saturation. Although Canfield et al (13) reported higher total concentrations of milk ß-carotene after an identical supplementation protocol, they also found correspondingly higher milk fat concentrations, which makes direct comparisons difficult. Milk fat concentrations in our study agree well with those in other reports of early milk, ie, on day 21 postpartum (20). The lack of effect of ß-carotene supplementation on concentrations of retinol, -tocopherol, and other carotenoids in milk agrees with the conclusions of Canfield et al (13).
We have presented detailed information regarding carotenoid concentrations in early milk. By 1 mo postpartum, most of these concentrations approached those reported for mature milk. Supplementation with ß-carotene did not produce any significant increase in milk concentrations, suggesting that early milk may be saturated with ß-carotene. Maternal supplementation with ß-carotene had no effect on concentrations of retinol, -tocopherol, or most other carotenoids in plasma or milk. Milk fat variability should be considered in the analysis of milk carotenoids. Further investigation is needed to elucidate any potential increases in plasma lutein concentration during pregnancy and to determine whether the observed postpartum decrease in plasma lutein is simply a normalization. Further research is also needed to investigate the cause and potential role of elevated milk lutein.
REFERENCES