Human Nutrition and Metabolism Dietary Fat Type Influences Total Milk Fat Content in Lean Women1,2 Nicole K. Anderson,* Kathy A. Beerman,* Mark A. McGuire,† Nairanjana Dasgupta,** J. Mikko Griinari,‡ Janet Williams,* and Michelle K. McGuire*3 *Department of Food Science and Human Nutrition, Washington State University, Pullman, WA 99164-6376; † Department of Animal and Veterinary Sciences, University of Idaho, Moscow, ID 83844-2330; **Program in Statistics, Washington State University, Pullman, WA USA 99162-3144; and ‡University of Helsinki, Department Animal Science, P.O. Box 28, Helsinki, Finland FIN-00014 ABSTRACT Trans fatty acids (TFA) are found naturally in some foods (e.g., dairy products) as well as many processed foods made with partially hydrogenated vegetable oils (PHVO). Data from a growing literature suggest that some TFA decrease milk fat in lactating animals. Because the physiologic effects of TFA in lactating women are unknown, this study was designed to investigate the effects of TFA consumption on human milk fat. A randomized, crossover design (n ⫽ 12) was used to study the effect of 3 dietary treatments: high PHVO (regular margarine), low PHVO (low TFA margarine), or low PHVO but high in naturally occurring TFA (butter) on milk fat. Treatments were administered for 5 d, with 7-d washout periods. Maternal adiposity was estimated by dual-energy X-ray absorptiometry. Milk and blood were collected on d 5 of each intervention period. In general, milk and serum fatty acid concentrations mirrored those of the dietary treatments. There were significant interactions between treatment and maternal adiposity on milk fat and infant milk consumption, as well as on serum glucose and nonesterified fatty acid (NEFA) concentrations. Consumption of regular margarine, compared with low TFA margarine, resulted in lower milk fat in leaner, but not in more obese women. Consumption of either regular or low TFA margarine, compared with butter, elevated serum NEFA concentrations in the more obese women. In summary, consumption of regular margarine, compared with low TFA margarine, decreased milk fat in lean women. Further studies are required to determine whether infant milk consumption might compensate for this potentially important change in milk composition. J. Nutr. 135: 416 – 421, 2005. KEY WORDS:

●

human milk

●

milk fat

●

lactation

●

breast-feeding

Human milk provides optimal nutrition for infants, with fatty acids (FAs) providing the majority of energy to the breast-feeding child. In general, it is thought that maternal diet does not influence total milk fat content, although body fat is positively related to this variable (1). However, we showed that when women consumed a diet low in dairy fat, they had lower milk fat than when they consumed more dairy fat (2). We theorized that consumption of partially hydrogenated vegetable oils (PHVO4; e.g., margarine) caused the milk

●

trans fatty acid

fat depression during the low-dairy period. Indeed, trans fatty acids (TFA) found in PHVO decrease milk fat in other lactating animals, such as cows (3–9). Estimates of TFA consumption in the United States range from 2.6 to 12.8 g/d, with foods containing PHVO contributing the majority of the TFA; some TFA isomers also occur naturally (10 –12). Interestingly, the Institute of Medicine recently reviewed the literature concerning the health implications of dietary TFA and recommended that “ . . .TFA consumption be as low as possible while consuming a nutritionally adequate diet” (13). Further, this report concluded that the tolerable upper limit of TFA intake should be zero. These recommendations were based on a positive relation between TFA intake and the risk of coronary heart disease. However, the effect of TFA consumption on human lactation is unknown. Clearly, additional research concerning TFA and maternal/child health is warranted. This study was designed to test the main hypothesis that consumption of regular margarine, compared with consumption of either butter or low TFA margarine, decreases milk fat in humans.

1 Presented in part at the 11th International Conference of the International Society for Research in Human Milk and Lactation, October, 2002, Mexico City, Mexico (McGuire, M. K., Anderson, N. K., McGuire, M. A., Mosley, E. E. & Beerman, K. Maternal adiposity influences the relationship between trans fatty acid isomers and lipid content in human milk); additional data were presented at Experimental Biology 02, April 2002, New Orleans, LA [Anderson, N., McGuire, M. K., McGuire, M. A., Beerman, K., Dasgupta, N., Koepp, A., Falen, R., Griinari, J. M. & Williams, J. (2002) Consumption of “no trans” margarine decreases human milk and serum conjugated linoleic acid (CLA) concentrations. FASEB J. 16: A662 (abs.)]. 2 Funded by the Western Dairy Center, Utah State University; Darigold, Seattle, WA; Ventura Foods, City of Industry, CA. 3 To whom correspondence should be addressed. E-mail: [email protected]. 4 Abbreviations used: B, butter; DEXA, dual energy X-ray absorptiometry; FA, fatty acid; HBF, high body fat; LBF, low body fat; LTM, low TFA margarine; NEFA, nonesterified fatty acids; PHVO, partially hydrogenated vegetable oil; RM, regular margarine; TAG, triacylglycerol; TFA, trans fatty acid.

SUBJECTS AND METHODS Lactating women (n ⫽ 12) from 2 to 12 mo postpartum (mean ⫾ SEM: 6.0 ⫾ 1.0 mo) were recruited from the Pullman, WA-

0022-3166/05 $8.00 © 2005 American Society for Nutritional Sciences. Manuscript received 19 December 2003. Initial review completed 11 March 2004. Revision accepted 2 December 2004. 416

TRANS FATTY ACIDS, BODY FAT, AND HUMAN MILK FAT

Moscow, ID area. All study procedures were approved by the Washington State University Institutional Review Board, and informed consent was obtained. This trial was designed as a randomized, crossover study with 5 periods: baseline (3 d), intervention I (5 d), washout I (7 d), intervention II (5 d), washout II (7 d), and intervention III (5 d). During the baseline period, maternal height and infant length were assessed (Seca® Measure-All™; ⫾ 0.1 cm), and maternal and infant weights were measured (Seca® Alpha, Model 770; ⫾ 0.1 kg and Seca® Model 727; ⫾ 1.0 g; respectively). During each intervention period, subjects consumed diets high in butter (B), regular margarine (RM), or low TFA margarine (LTM). During the B and RM periods, subjects minimized their intake of PHVO or dairy fats, respectively. During the LTM period, subjects limited their intake of both dairy fats and PHVO. To facilitate study compliance, subjects were provided with an ample supply of commercial butter (Darigold® Sweet Cream Butter, Grade AA, Darigold), RM (Gregg’s™ Gold-n-Soft® Real Margarine, Ventura Foods), or “low TFA” margarine (Saffola Soft Margarine®, Ventura Foods) during the corresponding period. In addition, subjects consumed 2 “study” muffins daily (1 before 1000 h and the other before 1600 h), each made with 37.8 ⫾ 1.0 g of the appropriate fat (i.e., B, RM, or LTM). Muffins weighed an average of 128 g each. Data concerning the energy, nutrient and FA composition of the study muffins are summarized in Table 1. Determination of adiposity. Maternal adiposity was estimated using dual-energy X-ray absorptiometry (DEXA; Hologic QDR® 4500 Acclaim™ Series) during the baseline period. Urine-based pregnancy tests (SureStep, Applied Biotech) were administered to

TABLE 1 Nutrient and fatty acid contents of muffins provided daily during B, LTM, and RM periods1 Dietary intervention Dietary component Energy and nutrients Energy, kJ Protein, g Carbohydrate, g Lipid, g Saturated fatty acids, g Cholesterol, mg Sodium, mg Fatty acids, mg 4:0 6:0 8:0 10:0 12:0 14:0 15:0 16:0 16:1(n-7) 18:1(n-9) t4:18:1 t5:18:1 t9:18:1 t10:18:1 t11:18:1 t16:18:1 18:2(n-6) c9,t11-18:2 t10,c12-18:2 18:3(n-3)

B

LTM

RM

4731 9 135 62 38 195 1492

4727 9 135 61 12 35 1465

4815 9 135 64 13 35 1599

2825 1270 620 1335 1475 4905 450 16,060 565 10,245 14 29 418 728 1431 555 3310 257 24 310

ND2 65 20 20 1475 235 10 4130 80 25,745 ND 55 1149 1211 1219 178 10,125 80 33 1585

ND 95 ND ND 5 55 ND 7351 55 7815 ND 76 2576 3464 4017 508 19,350 224 75 1675

1 Values for energy, protein, carbohydrate, lipid, SFA, cholesterol, and sodium were calculated from information provided on food labels; values for FA represent those derived from biochemical analyses on 4 representative muffins. 2 ND, not detectable.

417

subjects to confirm nonpregnant status before the DEXA scan. Because we became interested in the interaction between maternal adiposity and treatment on milk and serum composition, women were divided by median split into 1 of 2 body fat groups: low body fat (⬍30% body fat; LBF, n ⫽ 6) and high body fat (⬎30% body fat; HBF, n ⫽ 6). Blood and milk collection. Maternal blood (20 mL) was collected from nonfasting subjects between 0800 and 1000 h on d 5 of each intervention period, ⬃1 h after consumption of the study muffin. Serum samples were stored at ⫺20°C. On the final day of each intervention and washout period, milk samples were collected between 1300 and 1600 h by complete breast expression using an electric breast pump (Model SMR-B-R; Ameda-Egnell). Subjects pumped milk 2–3 h after a previous complete breast expression, while nursing the infant on the other breast until the milk flow had ceased or until the child was no longer nursing. Milk samples were collected from the same breast throughout the study and stored at ⫺20°C. Lipid extraction and FA determination. Milk, serum and food lipids were extracted utilizing a modified Folch methodology (14,15). Analyses of the majority of the FAs of the muffins and study fats were performed using GC (Hewlett Packard 6890 Series) fitted with a flame ionization detector. A programmed temperature gradient and detector temperatures of 255°C were used, and the FA profile was determined by split injection onto a CP-Sil 88 fused silica capillary column (100 m ⫻ 0.25 mm, Chrompack) (8). Concentrations of FA are presented as % identified FA. Individual FA were identified by evaluating retention times against those of pure standards (Matreya). Correction factors were determined by analyzing a butter standard (CRM 164; European Community Bureau of Reference). The t18:1 isomeric FA of the study muffins, experimental fats, milk, and serum were determined using the method of Mosley et al. (16). Briefly, methylated samples, using the methylation procedure described previously, were dried completely and dissolved in 0.6 mL of methylene chloride (1.6 mg of FAME:200 ␮L methylene chloride). After separation into saturated and trans monoene fractions, 200 ␮L of the sample were applied to the solid phase extraction column. Fractions were dried completely under nitrogen, dissolved in 200 ␮L of hexane and analyzed by GC (Hewlett Packard 6890 Series). A programmed isothermal temperature of 160°C and detector temperature of 255°C were used, and the FA profile was again determined using GC as described above. Milk protein and lactose and serum glucose. Milk protein and lactose were measured using the methods of Polberger and Lo¨nnerdal (17). Milk protein was determined spectrophotometrically (Bio-Rad Laboratories) with bovine serum albumin as a standard; intra- and interassay CV were 2.8 and 1.2%, respectively. Milk glucose was measured spectrophotometrically (Sigma Diagnostics®, Procedure No. 510) and used to calculate milk lactose after release of glucose from lactose by galactosidase; intra- and interassay CV were 1.3 and 1.1%, respectively. Serum glucose was determined using the identical kit; intra- and interassay CV were 2.7 and 1.1%, respectively. Serum insulin, NEFA, and triacylglycerol (TAG). Serum insulin was measured using RIA (Linco Research); intra- and interassay CV were 4.4 and 6.0%, respectively. Serum nonesterified fatty acids (NEFA) were determined by the method of McCutcheon and Bauman (NEFA-C Kit Wako Chemicals USA) (18,19); intra- and interassay CV were 2.3 and 1.0%, respectively. Serum TAG was analyzed using a series of coupled enzymatic reactions (20,21); the intra-assay CV was 0.9%. Maternal dietary intake and infant milk consumption. Maternal dietary intake was assessed from weighed written records on d 4 and 5 of each intervention period. Data from the dietary records, including selected micro- and macronutrients and FA, were analyzed using a computer-based dietary assessment program (Food Processor® Version 7.5; ESHA Research) (2,22). Infant milk consumption was estimated by weighing the infant before and after each feeding on d 4 of each intervention period using an electronic infant scale (Seca® Model 727; ⫾ 1.0 g). The initial weight was subtracted from the final weight at each feeding to estimate milk intake during the nursing session. Statistical analyses. Data were analyzed as a repeated-measures experiment with treatment (B, RM, and LTM), body composition

ANDERSON ET AL.

418

group (LBF or HBF), and the treatment ⫻ body composition group interaction in the model using standard ANOVA procedures (SAS Institute, Version 8.1). Initial analyses of milk fat data indicated that there were no sequence or carry-over effects from 1 intervention period to the next; thus, only data from the intervention periods were considered in further analyses. Comparisons between means were made using Tukey-Kramer adjustments. Main effects and simple comparisons were considered significant at P ⬍ 0.05; interactions were considered significant at P ⬍ 0.10. In this manuscript, data are typically presented by both dietary treatment and body composition group when the interaction between these factors was significant; data are collapsed by dietary treatment and/or body composition group when appropriate.

RESULTS All subjects completed the study; demographic variables and anthropometric measurements of the low (LBF) and high body fat (HBF) women and their infants are shown in Table 2. Women in the LBF group weighed less, were shorter, and had less body fat than did women in the HBF group (P ⬍ 0.05); they also tended to have fewer children (P ⫽ 0.06). Infant weight and length were not influenced by maternal adiposity. Maternal dietary intake. There was no interaction between treatment and body fat group and no independent effect of body fat group on any of the dietary intake variables examined (Table 3). Saturated fat and cholesterol intakes were higher during the B period compared with the LTM and RM periods. In general, daily intake of various FA (data not shown) reflected those found in LTM, RM, and B products (see Table 1). Higher levels of t18:1 isomers were consumed during the RM period compared with the B and LTM periods, except for t16 –18:1 which was greater during both the B and RM periods. Total TFA intakes (from study muffins and fats provided) were 5.7, 6.6, and 19.9 g/d during the B, LTM, and RM periods, respectively. Milk fat, lactose, and protein. The interaction between treatment and body fat group was significant for total milk fat content (Table 4). LBF, but not HBF women had lower (P ⬍ 0.05) milk fat during the RM period, compared with the LTM period. There was neither an interaction between treatment and body fat group nor independent effects of treatment or body fat

TABLE 3 Effect of B, LTM, and RM interventions on dietary intakes of lactating women1 Dietary intervention

Energy, kJ/d Fat, g/d % Energy from fat Protein, g/d Carbohydrate, g/d Fiber, g/d Saturated Fat, g/d Cholesterol, mg/d

B

LTM

RM

SEM

11,807 123 40 73 364 19 67b 414b

12,217 121 37 85 380 21 30a 190a

12,146 126 39 75 377 20 34a 200a

590 8 2 6 24 2 3 34

1 Values are means, n ⫽ 12; means in a row with superscripts without a common letter differ, P ⬍ 0.05.

group on milk lactose or protein (Table 4). However, HBF women tended to have lower overall milk lactose concentrations compared with LBF women (70.2 vs. 75.6 ⫾ 1.8 g/L; P ⫽ 0.07). Infant milk consumption. There was an interaction between dietary treatment and body composition on infant milk consumption (P ⫽ 0.05; Table 4). Infants nursed by LBF women tended to consume more milk during the LTM and RM periods, compared with the B period; infants nursed by HBF women tended to consume less milk during the LTM and RM periods. Serum glucose, TAG, NEFA, and insulin. There was an interaction between treatment and body fat group on serum glucose (P ⫽ 0.06), such that it tended to increase in the RM TABLE 4 Effect of B, LTM, and RM interventions on milk and serum components in lactating women and milk consumption in infants1 Dietary intervention B

LTM

RM

Pooled SEM

3.8ab 3.4ab 71.7 6.2

4.3b 2.9ab 73.0 6.3

2.6a 4.0ab 74.0 5.9

0.4

TABLE 2 Demographic variables and anthropometric measurements of LBF and HBF women and their infants at the time of enrollment1 Body fat group Variable Maternal Age, y Time postpartum, mo Parity, n Weight, kg Height, cm Body Mass Index, kg/m2 Body fat, % Infant Weight, g Length, cm

LBF

HBF

25.0 6.3 1.0 59.2 164.8 21.8 25.7

27.8 5.7 2.2 80.9* 170.8* 27.7* 36.2*

8002 67.3

7809 66.6

1 Values are means, n ⫽ 6; * different from LBF, P ⬍ 0.05.

SEM

1.2 1.0 0.3 4.0 1.6 1.2 1.9 453 1.6

Milk Fat,* % LBF HBF Lactose, g/L Protein, mg/mL Consumption,* g/d LBF HBF Serum Glucose,* mmol/L LBF HBF TAG, mmol/L Insulin, pmol/L NEFA,* mmol/L LBF HBF

740 850 4.31 4.94 1.43 81.3 3.01a 3.29a

868 631 4.29 5.11 1.34 69.5 2.88a 4.22b

812 633 4.93 4.87 1.32 72.9 2.96a 4.22b

1.6 0.4 102

0.21 0.16 11.81 0.39 0.39

1 Values are means; n ⫽ 12, unless data were subdivided by body composition group in which case n ⫽ 6; * significant treatment ⫻ body fat interaction, P ⬍ 0.10; means for a variable with superscripts without a common letter differ, P ⬍ 0.05.

TRANS FATTY ACIDS, BODY FAT, AND HUMAN MILK FAT

TABLE 5 Effect of B, LTM and RM consumption on serum fatty acid composition in lactating women1 Dietary intervention Fatty acid

B

LTM

RM

Pooled SEM

2.99 0.17 0.12 0.07a 0.19a 1.14a 0.49 15.21ab 1.12a 4.35 11.45a 0.05 0.07 1.29 0.63 0.52b 0.17 19.79b 0.34 0.09 1.12

0.33 0.02 0.02 0.01 0.04 0.11 0.07 0.94 0.12 0.26 0.94 0.01 0.08 0.24 0.19 0.10 0.02 1.46 0.06 0.02 0.20

g/100 g lipid 4:0 6:0 8:0 10:0† 12:0†§ 14:0†§ 15:0§ 16:0† 16:1(n-7)† 18:0 18:1(n-9)† t4-18:1 t5-18:1 t9-18:1 t10-18:1* t11-18:1† t16-18:1* 18:2(n-6)† c9,t11-18:2§ t10,c12-18:2 18:3(n-3)

2.76 0.17 0.12 0.14b 0.29ab 1.57b 0.50 17.82b 1.70b 4.53 12.51a 0.07 0.21 0.84 0.56 0.14a 0.15 15.61a 0.39 0.10 0.88

2.39 0.16 0.13 0.08a 0.35b 1.29ab 0.06 13.38a 1.20a 4.43 16.36b 0.05 0.03 0.81 1.16 0.22ab 0.16 16.00a 0.37 0.08 0.86

1 Values are means, n ⫽ 12; * significant treatment ⫻ body fat interaction, P ⬍ 0.10; † significant independent effect of treatment (P ⬍ 0.05); § significant independent effect of body fat group, P ⬍ 0.05;

means for a variable with superscripts without a common letter differ, P ⬍ 0.05.

period in the LBF women (Table 4). Further, women in the HBF group tended to have higher serum glucose concentrations compared with those in the LBF group, especially during the B and LTM periods. There was no treatment ⫻ body fat group interaction for serum TAG or insulin; there were also no main effects of dietary treatment or body fat group on these variables. However, HBF women tended to have higher overall serum TAG concentrations compared with LBF women (1.62 vs. 1.10 ⫾ 0.20 mmol/L; P ⬍ 0.1). There was an interaction between treatment and body composition group on serum NEFA concentration (P ⫽ 0.03; Table 4). During the B period, the HBF group had lower (P ⬍ 0.05) NEFA values than during the LTM and RM periods. This effect of treatment was not seen in the LBF women. Serum FAs. Only for t10 –18:1 and t16 –18:1 was there a significant interaction between body composition and body fat (Table 5). The LBF women had higher concentrations of t10 –18:1 during the LTM period compared with the RM period (1.60 vs. 0.47 g/100 g lipid, respectively; P ⬍ 0.05), whereas the HBF women tended to have higher concentrations during the RM and LTM compared with the B, periods (0.78, 0.72 and 0.59 g/100 g lipid, respectively; P ⬍ 0.1). There was an independent effect of dietary treatment for 8 FA. In general, differences in serum FA reflected similar differences in FA intake. Of the 4 FA for which there was an independent effect of body fat group, only c9,t11–18:2 was lower in the HBF women than in the LBF women (0.26 vs. 0.47 g/100 g lipid, respectively; P ⬍ 0.05). There was an independent effect of body fat group on 4 FA. Interestingly, 12:0, 14:0, and 15:0 were all significantly lower

419

in the LBF women compared with the HBF women. However, LBF women tended to have higher concentrations of c9,t11– 18:2 compared with HBF women (0.47 vs. 0.26 g/100 g lipid, respectively; P ⫽ 0.06). Milk FAs. Four of the FAs revealed a significant treatment ⫻ body fat group interactions (P ⬍ 0.10); these were 14:0, 15:0, 16:0, and 18:2(n-6) (Table 6). Nonetheless, for 14:0 and 15:0, both the LBF and HBF women tended to have higher concentrations of these FA during the B and LTM periods compared with the RM period. Both the LBF and HBF groups had higher concentrations of 16:0 during the B period compared with the LTM and RM periods. A significant independent effect of treatment (P ⬍ 0.05) was found for most of what are commonly considered the de novo FA (8:0 –16:0); in most cases, these FA were highest during the B period. There was no effect of body fat group on any milk FA. Because animal data suggest that t10 –18:1 causes milk fat depression, we also conducted correlation analyses to examine this possibility. However, milk fat was not correlated with either milk or serum t10 –18:1 concentration. Also, further analysis by intervention period did not reveal a correlation between milk or serum t10 –18:1 and total milk fat within a particular treatment. Additionally, there was no correlation between body fat and total milk fat after the separation into body fat groups. DISCUSSION The objective of this study was to investigate the effects of consumption of foods containing low amounts of TFA, high amounts of commercially prepared TFA (from PHVO), or high amounts of natural TFA (from butter) on human milk

TABLE 6 Effect of B, LTM, and RM intervention on milk fatty acid composition in lactating women1 Dietary intervention Fatty acid

B

LTM

RM

Pooled SEM

0.03b 4.72ab 0.13b 0.98a 3.74a 3.82a 0.15a 12.74a 1.10a 6.61 25.81a 0.01 0.56 0.81 2.47b 1.95b 18.50c 0.34ab 0.01 1.33b

0.01 1.46 0.01 0.06 0.26 0.31 0.02 0.57 0.13 0.49 0.68 0.01 0.19 0.28 0.37 0.32 0.72 0.05 0.01 0.07

g/100 g lipid 4:0† 6:0† 8:0† 10:0† 12:0† 14:0* 15:0* 16:0* 16:1† 18:0 18:1(n-9)† t4-18:1 t5-18:1 t9-18:1 t10-18:1† t11-18:1† 18:2(n-6)* c9,t11-18:2† t10,c12-18:2 18:3(n-3)†

0.04b 5.00b 0.17c 1.13b 3.89ab 6.18b 0.44b 18.64b 1.72b 7.11 26.72a 0.02 0.10 0.91 0.97a 0.64a 9.55a 0.47b 0.01 0.57a

0.01a 3.52a 0.09a 0.94a 4.32b 4.44a 0.17a 12.18a 1.26a 5.87 34.96b 0.01 0.20 0.57 1.38a 0.50a 13.43b 0.22a 0.01 1.14b

1 Values are means, n ⫽ 12; * significant treatment ⫻ body fat interaction, P ⬍ 0.10; † significant independent effect of treatment, P ⬍ 0.05; means for a variable with superscripts without a common letter differ, P ⬍ 0.05.

420

ANDERSON ET AL.

fat. Contrary to our initial hypotheses, milk fat was not altered by treatment when all data were combined. However, during the RM period, LBF women had significantly lower milk fat than during the LTM period. Clearly, lean women consuming commercially produced TFA (from PHVO) responded differently than did women with higher body fat. In this study, TFA consumption during treatments were within normal U.S. intakes (10), although the dietary interventions reflect the extremes of the typical range. Similarly, the milk fat, lactose, and protein values reported here concur with published data (17,23). Therefore, it is likely that our subjects were a representative sample of healthy, well-nourished, lactating U.S. women. As with any human intervention trial, there are several limitations to this study. Although women were asked to minimize their intakes of PHVO, dairy fats, and both dairy fats and PHVO during the B, RM, and LTM periods, respectively, we could not control for subjects who ate away from home and could not prepare foods with the margarines and butter provided to them. Second, our analysis of the margarine actually labeled as containing “no TFA” revealed that it did contain quite high amounts of some TFA isomers (albeit less than the RM) and, therefore, may not have been an ideal choice. However, the fact that we still found interactions between body fat group and dietary treatment on milk fat not using a TFA-free margarine is noteworthy. Despite these limitations, consumption of RM significantly decreased milk fat in the lean, but not the heavier women. These findings may help explain the consistent positive relation found between maternal adiposity and human milk fat (24 –26) in societies consuming large amounts of PHVO. For example, Lovelady et al. (25) found an association between body fat and milk lipid (r ⫽ 0.53; P ⫽ 0.002) and reported that dietary fat was positively correlated with milk lipid in lean (⬍27% body fat; r ⫽ 0.75; P ⫽ 0.01), but not obese women (ⱖ27% body fat). Together, our findings suggest that women with smaller fat stores may have less substrate to mobilize and utilize as milk fat, and that this might be especially important under the condition of TFA-induced milk fat depression. Hachey et al. (27) provided evidence that 10 –12% of milk fat is derived from mammary synthesis, 29% from dietary intake, and the remainder from synthesis or mobilization from other tissues (e.g., adipose tissue or liver). Because the most dramatic change in milk fat in lean women occurred between the LTM and RM periods, we examined what are considered the “de novo FA” (⬍16 carbons in length) to determine how this milk fat depression might be realized. However, overall changes in the enrichments of milk and serum 10:0, 12:0, and 14:0 were either small or nonexistent during the RM period compared with the LTM period (data not shown), and these concentrations seem to reflect dietary intake. As a result, the majority of the milk fat depression that accompanied the RM period in LBF women did not appear to be due to decreased synthesis in the mammary gland, but perhaps resulted from decreased uptake or utilization of “preformed FA” derived from the diet, liver, or the mobilization of adipose tissue. Further, our data showing that RM intake was associated with increased NEFA concentrations only in the HBF women suggest that the negative effect of some TFA isomers found in the RM might have been compensated for by increased adipose mobilization. It is possible that LBF women were unable to compensate due to lower body fat stores. If dietary TFA from PHVO can act as regulators of milk fat synthesis in humans, future studies should attempt to determine whether specific TFA isomers target enzymes involved in FA synthesis or lipoprotein lipase, which transfers lipid into mammary cells.

Further, regulation of adipose mobilization should be considered, perhaps with the use of stable isotope methodology. Data from the present study also suggest that maternal body fat and dietary treatment can interactively influence milk consumption by the infant. It is possible that increased demand by the infant was in response to consuming lower-fat milk. However, we did not find a consistent negative relation between milk fat content and milk consumption, as would be expected if this were true. Further work is warranted to determine the causal and longitudinal nature of these effects. Forthcoming studies designed to investigate the regulation of human milk fat should utilize sensitive measures of body fat (e.g., DEXA) to account for this important variable and measure alterations in the various precursors of milk lipid (e.g., NEFA, TAG, and glucose). Additional work will be required to determine whether the milk fat depression that occurred in lean women is of transient or chronic nature; also, it is important to determine whether changes in infant milk consumption might compensate for changes in milk fat concentration. Finally, in response to the recent report by the Institute of Medicine concerning the safety of TFA consumption by humans (13), subsequent studies should be considered to lend additional insight into the acceptable dietary intakes of TFA for breast-feeding women and their suckling infants. ACKNOWLEDGMENTS We thank Alfred Koepp, Roger Falen, and Erin Mosley for their technical support and proficiency in the laboratory.

LITERATURE CITED 1. Neville, M. C. & Picciano, M. F. (1997) Regulation of milk lipid secretion and composition. Annu. Rev. Nutr. 17: 159 –184. 2. Park, Y., McGuire, M. K., Behre, R., McGuire, M. A., Evans, A. & Shultz, T. D. (1999) High-fat dairy product consumption increases ⌬9c,11t-18:2 (rumenic acid) and total lipid concentrations of human milk. Lipids 34: 543–549. 3. Astrup, H. N., Vik-Mo, L., Ekern, A. & Bakke, F. (1976) Feeding protected and unprotected oils to dairy cows. J. Dairy Sci. 59: 426 – 430. 4. Selner, D. R. & Schultz, L. H. (1980) Effect of feeding oleic acid or hydrogenated vegetable oils to lactating cows. J. Dairy Sci. 63: 1235–1241. 5. Teter, B. B., Sampugna, J. & Keeney, M. (1990) Milk fat depression in C57b1/6J mice consuming partially hydrogenated fat. J. Nutr. 120: 818 – 824. 6. Wonsil, B. J., Herbein, J. H. & Watkins, B. A. (1994) Dietary and ruminally derived trans-18:1 fatty acids alter bovine milk lipids. J. Nutr. 124: 556 –565. 7. Teter, B. B., Sampugna, J. & Erdman, R. A. (1995) Specific dietary isomeric fatty acids fed to lactating mice are associated with milk fat depression. J. Dairy Sci. 78 (suppl. 1): 243. 8. Griinari, J. M., Dwyer, D. A., McGuire, M. A., Bauman, D. E., Palmquist, D. L. & Nurmela, K. V. (1998) Trans-octadecenoic acids and milk fat depression in lactating dairy cows. J. Dairy Sci. 81: 1251–1261. 9. Loor, J. J. & Herbein, J. H. (1998) Exogenous conjugated linoleic acid isomers reduce bovine milk fat concentration and yield by inhibiting de novo fatty acid synthesis. J. Nutr. 128: 2411–2419. 10. Carlson, S. E., Clandinin, M. T., Cook, H. W., Emken, E. A. & Filer, L. J. (1997) Trans fatty acids: infant and fetal development. Am. J. Clin. Nutr. 66: 715S–736S. 11. Allison, D. B., Egan, K., Barraj, L. M., Caughman, C., Infante, M. & Heimbach, J. T. (1999) Estimated intakes of trans fatty and other fatty acids in the US population. J. Am. Diet Assoc. 99: 166 –174. 12. Hunter, J. E. & Applewhite, T. H. (1991) Reassessment of trans fatty acid availability in the US diet. Am. J. Clin. Nutr. 54: 363–369. 13. Institute of Medicine (2002) Letter report on dietary reference intakes for trans fatty acids. In: Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. National Academy Press, Washington, DC. 14. Clark, R. M., Ferris, A. M., Brown, P. B., Hundrieser, K. E. & Jensen, R. G. (1982) Changes in the lipids of human milk from 2 to 16 weeks postpartum. J. Pediatr. Gastroenterol. Nutr. 1: 311–315. 15. Masters, N., McGuire, M. A., Beerman, K. A., Dasgupta, N. & McGuire, M. K. (2002) Maternal supplementation with CLA decreases milk fat in humans. Lipids 37: 133–138. 16. Mosley, E. E., Powell, G. L., Riley, M. B. & Jenkins, T. C. (2002) Microbial biohydrogenation of oleic acid to trans isomers in vitro. J. Lipid Res. 43: 290 –296.

TRANS FATTY ACIDS, BODY FAT, AND HUMAN MILK FAT 17. Polberger, S. & Lonnerdal, B. (1993) Simple and rapid macronutrient analysis of human milk for individualized fortification: basis for improved nutritional management of very-low-birth-weight infants? J. Pediatr. Gastroenterol. Nutr. 17: 283–290. 18. McCutcheon, S. N. & Bauman, D. E. (1986) Effect of chronic growth hormone treatment on responses to epinephrine and thyrotropin-releasing hormone in lactating cows. J. Dairy Sci. 69: 44 –51. 19. Sechen, S. J., Dunshea, F. R. & Bauman, D. E. (1990) Somatotropin in lactating cows: effect on response to epinephrine and insulin. Am. J. Physiol. 258: E582–E588. 20. Trinder, P. (1969) Determination of glucose in blood using glucose oxidase with an alternative oxygen acceptor. Ann. Clin. Biochem. 6: 24 –27. 21. Bucolo, G. & David, H. (1973) Quantitative determination of serum triglycerides by the use of enzymes. Clin. Chem. 19: 476 – 482. 22. Ritzenthaler, K. L., McGuire, M. K., Falen, R., Shultz, T. D., Dasgupta, N. & McGuire, M. A. (2001) Estimation of conjugated linoleic acid intake by

421

written dietary assessment methodologies underestimates actual intake evaluated by food duplicate methodology. J. Nutr. 131: 1548 –1554. 23. Jensen, R. G. (1995) The Handbook of Milk Composition. Academic Press, San Diego, CA. 24. Villalpando, S. & del Prado, M. (1999) Interrelation among dietary energy and fat intakes, maternal body fatness, and milk total lipid in humans. J. Mammary Gland Biol. Neoplasia 4: 285–295. 25. Lovelady, C. A., Nommsen, L. A., McCrory, M. A. & Dewey, K. G. (1993) Relationship of human milk lipid concentration to maternal body composition and dietary fat intake. FASEB J. 7: A200 (abs.). 26. Barbosa, L., Butte, N. F., Villalpando, S., Wong, W. W. & O’Brian, S. E. (1997) Maternal energy balance and lactation performance of Mesoamerindians as a function of body mass index. Am. J. Clin. Nutr. 66: 575–583. 27. Hachey, D. L., Thomas, M. R., Emken, E. A., Garza, C., Brown-Booth, L., Adlof, R. O. & Klein, P. D. (1987) Human lactation: maternal transfer of dietary triglycerides labeled with stable isotopes. J. Lipid Res. 28: 1185–1192.