点击显示 收起
1 From the Department of Animal and Veterinary Science, University of Idaho, Moscow, ID (EEM and MAM); the Arizona Respiratory Center, University of Arizona, Tucson, AZ (ALW); and the Department of Food Science and Human Nutrition, Washington State University, Pullman, WA (MKM)
2 Supported by the United Dairymen of Idaho, the Idaho Agricultural Experiment Station, Washington State University Agricultural Research Center, and grant no. AI42268 from the National Institutes of Health. 3 Reprints not available. Address correspondence to MA McGuire, University of Idaho, PO Box 442330, Moscow, ID 83844-2330. E-mail: mmcguire{at}uidaho.edu.
ABSTRACT
Background:trans Fatty acids (FAs) have been identified as negatively affecting human health. The trans FA composition of human milk fat must be examined to establish its influence on the nutritional quality of milk consumed by infants.
Objective: We sought to ascertain the individual and total trans FA isomers (sum of FAs containing at least one trans double bond) in human milk and to identify relations between individual FAs and milk fat concentration (% by wt).
Design: The FA composition of milk samples (n = 81) from women living in the southwestern United States was ascertained. The individual 18:1t isomers were separated. Correlations between each FA, total trans FAs, groups of similar FAs, and milk fat concentrations were examined.
Results: The mean total trans FA concentration was 7.0 ± 2.3% (range: 2.5–13.8%). The concentration of total 18:1t was 5.1 ± 2.0% (range: 1.5–11.6%), and 10t (range: 9–12t) was the most abundant isomer.
Conclusions: Milk fat from women living in the United States contains concentrations of trans FAs similar to those in milk from Canadian women but greater than those reported in milk from women in other countries. In decreasing order of concentration, the 10t, 11t, 9t, and 12t isomers represented 78.9% of the total 18:1t. These FAs generally originate from partially hydrogenated vegetable oils and ruminant fat in the diet. No relation was found between the concentration of total trans FAs and milk fat concentration.
Key Words: Milk fat • trans fatty acids • human milk • lipids • lactation • US women
INTRODUCTION
Human milk is considered to be the gold standard for nourishing infants. Mature milk generally contains from 3.5% to 4.5% lipid, mostly triacylglycerols (1). Milk fatty acids (FAs) are either synthesized in the mammary gland or are derived from plasma FAs that originate from the diet, synthesis in nonmammary tissues, or lipid stores (2). It is well established that the FA composition of the diet can alter milk FA composition (3).
trans FAs are defined as those that contain at least one double bond in the trans configuration. They are formed during the chemical hydrogenation of vegetable oils to manufacture commercial fats or during the microbial biohydrogenation of FAs in the digestive tract of ruminant animals. The consumption of trans FA–containing foods by lactating women results in the incorporation of trans FAs in their milk fat (3-5).
Some of the trans FAs found in human milk fat have been identified as having potentially negative effects on human health, primarily via altering lipid metabolism. For example, the consumption of some trans FAs is associated with higher total and LDL-cholesterol concentrations and lower HDL-cholesterol concentrations in humans (6). This increase in the ratio of LDL to HDL cholesterol has been associated with a higher risk of coronary artery disease (7). Furthermore, trans FAs have been implicated in the impairment of essential FA metabolism through their inhibition of the synthesis of long-chain polyunsaturated FAs (PUFAs) in humans and animals (8). This inhibition might result in a negative effect on infant growth (9, 10). Finally, trans FAs have also been implicated in a decrease in milk fat in humans (11), mice (12), and cattle (13). Specifically, the trans FA 18:210t12c, an isomer of conjugated linoleic acid (CLA), has been identified as causing a decrease in fat in the milk of cattle (14). Similarly, women consuming a 1.5-g CLA supplement containing 37% of the 18:210t12c isomer experienced a decrease in milk fat (11). The decrease in milk fat content would result in lower energy density of human milk.
Because of the potential negative effects of trans FAs, the Institute of Medicine has indicated that the intake of trans FA should be "as low as possible" (15). In addition, the Food and Drug Administration ruled that the trans FA content of foods must be included on nutrition labels (16). Even before these labeling laws, concern about the safety of trans FAs resulted in the documentation of their concentration in human milk fat. There have been several studies of the trans FA content of milk from women (17-26), but few analyzed the isomers of 18:1t. The main objective of the current study was to ascertain the isomeric and total trans FA concentrations (sum of all FAs that contain 1 trans double bond) of milk produced by US women. We hypothesize that total trans FAs will be inversely related to total milk fat content. To identify possible interactions between milk fat synthesis and specific FAs, we also examined correlations between each FA, total trans FA concentration, and milk fat concentration (% by wt).
SUBJECTS AND METHODS
Subjects
Healthy women between 32–35 wk gestation who planned to obtain care for their newborns through pediatric practices (which participated in identifying subjects and assisting in the original study) were eligible to participate in the study. The study population was 73.3% non-Hispanic white, and 78.5% of the participants had some college education; more details were provided in Oddy et al (27). All women were consuming self-selected diets. Because the parent study did not initially aim to examine the effect of maternal diet on milk composition, documentation of dietary intake was not part of the experimental design of the current study.
Written informed consent was obtained from all participants. The parent protocol and the protocol for the current study were approved by the University of Arizona Institutional Review Board.
Methods
The samples utilized in this study were collected as a part of the larger Infant Immune Study, which was designed to investigate immune system markers of asthma risk in infancy and early childhood (27). Milk samples were obtained during the period from November 1996 to March 2002 by complete expression of breast milk of women (n = 81) during the first month (
Milk lipid samples were thawed at room temperature, and a 2-g mass was extracted by using chloroform and methanol in a 2-to-1 ratio (29). The extracted lipid was converted to FA methyl esters (FAMEs) by using a base-catalyzed transesterification method (30). The FAMEs in hexane were analyzed on a gas chromatograph (Hewlett-Packard 6890 Series with auto injector; Agilent, Palo Alto, CA) that was fitted with a flame ionization detector and a 100 m x 0.25 mm (0.2-µm film) capillary column coated with CP-Sil 88 (Chrompack; Middelburg, Netherlands). The initial oven temperature was 70 °C, which was held for 3 min, and then it was increased to 175 °C at a rate of 3 °C/min and held for 3 min. Next, the oven temperature was increased to 185 °C at a rate of 1 °C/min and held for 20 min, increased to 215 °C at a rate of 3 °C/min (with no hold), and, finally, increased to 230 °C at a rate of 10 °C/min and held for 5 min. Peaks were identified by comparison with known FA standards. Response correction factors determined by the analysis of a butter-oil standard with certified values (CRM 164; European Community Bureau of Reference, Brussels, Belgium) were used to convert FA peak area percentage to percentage by weight.
The FAME samples (0.1 mg) were separated into saturated, monoene, and diene FA fractions (31). Dimethyl disulfide adducts were prepared from the monoene FAME fraction (32). To ascertain the occurrence of specific 18:1 t-isomers, samples were analyzed by gas chromatography (GC)–mass spectrometry (MS) [Hewlett-Packard (HP) 6890 equipped with a DB-225ms capillary column; Agilent]. An HP series II 5973 quadrupole MS controlled by HP MS CHEMSTATION software (PC version D.01.00; Agilent) was used for MS analysis. Column temperature was programmed to increase from 195 °C to 230 °C at a rate of 0.5 °C/min during a total run time of 70 min. The peak areas of 18:1t isomers that were eluted with cis isomer peaks were calculated by standardizing the peak areas identified on the GC-MS of the dimethyl disulfide derivatives with the 18:111t peak. These methods are newly applied to human milk FA analysis; however, we previously used the same methods to examine 18:1t isomers in rumen bacterial cultures (32).
Statistical analysis
Correlations between the percentage by weight of each FA, total trans FAs, groups of similar FAs, and milk fat concentration were examined by using Pearson correlation coefficients generated by the PROC CORR procedure of SAS (version 8.02; SAS Institute, Cary, NC). Analysis of variance (PROC GLM) was used to ascertain the relations between total trans FAs and FA percentages. Milk samples were categorized as having high, medium, or low concentrations of trans FAs on the basis of naturally occurring breaks in the data when plotted by total concentrations of trans FAs. Group least-squares means for each FA were analyzed for differences by using the Tukey procedure. Relations were declared significant at P 0.05.
RESULTS
Fatty acid concentrations
Overall, 83 FA peaks were identified, and the concentrations of 12 FAs were found to be > 1% by weight of the total FAs. The 3 major FAs were palmitic (16:0; 20.8%), oleic (18:19c; 29.4%), and linoleic (18:2n–6; 12.7%) acid (Table 1). The major FA groups (Table 2) were the saturated FAs (40.6%), of which 51.1% was palmitic acid, and the monounsaturated FAs (42.4%) of which 69.3% was oleic acid. Linoleic acid made up 74.6% of the PUFAs. The 18:29c11t isomer made up 83.3% of the total CLA isomers (Table 1). The mean total trans FA concentration was 7.0 ± 2.3% [range: 2.5–13.8% (Table 2)]. The 18:1t isomers appeared to be normally distributed, and 9–12t were the most abundant isomers (Figure 1). Milk fat content (Table 2) as measured by using the creamatocrit method was similar to other published data (1).
View this table:
TABLE 1. Concentration of major fatty acids (FAs) in human milk samples1
View this table:
TABLE 2. Summary of selected fatty acids (FAs) and total fat of human milk1
FIGURE 1.. Mean (±SD) distribution of individual trans 18:1 isomers as a percentage of total trans 18:1 isomers (18:1t) in human milk (n = 81).
Correlations among total fat, fatty acids, and fatty acid classes
For simplicity, only correlation comparisons for selected major FAs and FA groups are presented (Table 3). All correlations discussed here were significant (P < 0.05), unless otherwise stated. No correlations were observed between individual or total trans FAs and milk fat concentrations. Total 18:1t and total trans FAs were positively correlated with total CLA. No correlations were detected between linoleic acid, linolenic acid (18:3n–3), total n–3, or total n–6 FAs and total 18:1t or total trans FAs. However, total 18:1t was positively correlated with stearic acid (18:0) but negatively correlated with oleic acid. Total 18:1t, total trans FAs, and 18:210t12c (r = –0.384) were also negatively correlated with total 9-desaturase enzyme products (ie, the sum of 14:19c, 16:19c, 18:19c, and 18:29c11t).
View this table:
TABLE 3. Pearson correlation coefficients for relations between selected major fatty acids (FAs) and FA groups in human milk1
On the basis of the wide variation in the total trans FA concentrations of these milk samples, the samples were grouped into 3 categories to evaluate the differences between the groups: high (samples containing 8% total trans FAs; n = 21), medium (samples containing > 5% but < 8% total trans FAs; n = 48), and low (samples containing 5% total trans FAs; n = 12) trans FAs (Table 4). The grouping of the data was based on naturally occurring breaks in the data when plotted by total content of trans FAs. For simplicity, only data for major individual FAs and FA groups are shown. For stearic and linoleic acid concentrations, there were no significant differences between groups. Total saturated FA and PUFA concentrations did not differ significantly between the groups; however, monounsaturated FA concentrations were higher in the high trans FA group than in the low trans FA group (P < 0.05).
View this table:
TABLE 4. Comparison of composition of major fatty acids (FAs) and FA groups in human milk with different contents of trans FAs1
DISCUSSION
The current study used modern analytic methods to comprehensively establish the FA profiles of milk samples from US women. Wolff et al (33) detailed concerns about the accuracy of the data from earlier studies. The current study used a combination of GC and GC-MS analyses to identify specific 18:1t isomers. These techniques allowed for the identification of a wide range of total trans FAs (Table 1) and individual 18:1t isomers (Figure 1) in human milk fat. Previous studies (17-26) evaluated the total trans FA concentrations in milk from women. Several of these studies are summarized in Table 5. The current survey of milk samples provides information concerning the FA composition of milk during early lactation by US women living in Arizona. The total concentration reported is not as variable as that in milk from Canadian women (17). However, the mean concentrations in milk from women from Arizona and Canada were not significantly different. In a comparison with the data presented here, data from the Czech Republic (21) and Germany (18, 26) showed that milk from women in those countries contained intermediate amounts of total trans FAs, whereas milk from women in Nigeria (19) and China (20) contained lower amounts. The consumption of foods that contain greater concentrations of trans FAs, such as processed foods made with partially hydrogenated vegetable oils and ruminant fat (35), leads to greater intakes of total trans FAs, which are incorporated into milk fat (3, 5, 22). These differences in the means and ranges of total trans FAs between milk from women in different countries may be partially attributed to differences in the diets consumed by the women in these various areas of the world (5, 33). It will be interesting to see whether current efforts to decrease trans FA consumption in the United States will lead to a lowering of milk trans FA concentrations in US women.
View this table:
TABLE 5. Comparison of total trans fatty acids (FA) concentrations in human milk lipids of subjects in various countries1
Studies using similar techniques identified 10t and 11t as major 18:1t isomers (17, 18, 33, 36). Unfortunately, there has been little research on the dietary source of isomeric 18:1t, and patterns of 18:1t differ among foods (33, 35). All of the identified 18:1t isomers are typically found in both partially hydrogenated vegetable oils and ruminant fats. The 11t is the predominant 18:1t in ruminant fat (milk and beef), whereas 9t, 10t, and 11t predominate, and 12t is greater in partially hydrogenated vegetable sources (33, 35). The strong correlations among 9t, 10t, and 11t suggest that partially hydrogenated vegetable oils were the main dietary source of 18:1t for these lactating women. However, no reliable database of the isomeric 18:1t contents of foods exists from which this could be confirmed if dietary records were available. Thus, further investigation is needed.
As a result of the Institute of Medicine's call for the dietary intake of trans FAs to be as low as possible (15), new food-labeling laws have been adopted that require the labeling of the trans FA content of foods (16). If the food product contains > 0.5 g total trans FA per serving, not including CLA isomers, then the trans FA content of the food must be noted on the packaging. However, if the total is < 0.5 g per serving, then the product can be labeled as containing 0 g trans FAs. On the basis of these requirements and assuming that a single "serving" of human milk is 125 g, approximately one-half of the milk that we analyzed would require labeling. Nonetheless, whereas it would unquestionably be beneficial for breastfeeding mothers to decrease their consumption of high trans FA foods, human milk provides optimal nutrition for infants.
The concerns about trans FAs in the human diet have arisen because of the negative effect that trans FAs may have on heart health (37). Specifically, dietary supplementation with 7% total trans FAs and 5.5% 18:19t, a major trans FA formed during hydrogenation of plant oils (35), increased total and LDL cholesterol and decreased HDL cholesterol more than did diets rich in oleic, palmitic, or lauric and myristic acids (38). Furthermore, the treatment of linolelaidic acid (18:2tt) in human hepatoma HepG2 cells resulted in an LDL:HDL of 1.18, as compared with ratios of 0.70 and 0.96 for linoleic and palmitic acid treatments, respectively (39).
In some studies, trans FAs have been shown to alter the incorporation of essential long-chain PUFAs into milk fat (40). When rats were provided diets containing different amounts of trans FAs but the same proportions of linoleic and linolenic acids, trans FAs were incorporated into milk fat in a dose-dependent manner. However, rats receiving dietary trans FAs had higher concentrations of linoleic acid in milk fat, but total long-chain PUFAs did not differ significantly between groups (40). In contrast, Innis and King (41) found an inverse relation between essential all-cis n–6 and n–3 FAs and total trans FAs in human milk. A similar inverse relation between total 18:1t and linoleic or linolenic acids was noted in Canadian milk samples (17). However, in the current study, no correlation between total trans FAs or total 18:1t and linoleic, linolenic, total n–3, or total n–6 FA was detected.
Individual isomers of trans FAs may have quite different effects on lipid metabolism. For example, Mahfouz et al (42) showed that 18:110t inhibited 9-desaturase activity in rat liver microsomes, whereas 18:19t had little effect, and 18:111t had an intermediate effect. Dietary 18:210t12c reduced mRNA expression of the 9-desaturase enzyme in mammary explants from lactating dairy cows (43). Hence, trans FAs may have negative effects on the concentration of oleic acid and other 9-desaturase products found in human milk, which could lead to lower consumption of oleic acid, an antiatherogenic FA (44), by infants. In support of these observations, total trans FAs, 18:1t, and 18:210t12c (data not shown) were negatively correlated with oleic acid and total 9-desaturase products (Table 3).
Our current study found no correlation between total trans FAs, 18:210t12c, or any other individual trans FA and milk fat concentration, which is similar to results obtained in an the analysis of Canadian human milk samples (17). However, when women received an experimental treatment of 1.5 g CLA supplement containing 37% of the 18:210t12c isomer, a decrease in milk fat occurred (11). One difference noted between these studies was that the percentage by weight of 18:210t12c in the US milk samples was only 0.01% of total FAs, whereas that in the study by Masters et al (11) was >10-fold. Furthermore, the study of Masters et al was specifically designed to evaluate the decrease in milk fat induced by 18:210t12c interventions in a crossover experiment, whereas the current study was an evaluation of milk samples taken at the convenience of the subjects. These differences in experimental design and concentration of 18:210t12c may explain why no relation was identified between milk fat and 18:210t12c or total trans FAs.
This study shows that the concentrations of trans FAs in the human milk samples from Arizona are some of the highest reported in the United States and other countries. More regional information is needed to fully describe the variation in the concentrations of trans FAs in human milk samples throughout the United States. The effect of trans FAs on human health is an important area of research. Identifying the concentrations of trans FAs in human milk fat will aid in identifying possible relations between trans FAs and infant growth and health.
ACKNOWLEDGMENTS
ALW was the principal investigator of the parent study and provided the samples for this study. MAM and MKM obtained funding for the analysis of the samples. EEM analyzed the data and wrote the manuscript, with input from all authors. None of the authors had any financial or personal interest in any organization sponsoring the research, including advisory board affiliations, or any other personal or financial conflict of interest.
REFERENCES