- Email: [email protected]
Effect of Dietary Vitamin C on Concentrations of Ascorbic Acid in Plasma and Milk1
J. Dairy Sci. 84:2302–2307 American Dairy Science Association, 2001.
Effect of Dietary Vitamin C on Concentrations of Ascorbic Acid in Plasma and Milk1 W. P. Weiss Department of Animal Sciences Ohio Agricultural Research and Development Center The Ohio State University, Wooster 44691
ABSTRACT The addition of exogenous ascorbic acid to milk reduces the development of oxidized flavor. This experiment was conducted to determine whether feeding ascorbic acid to cows influenced vitamin C concentrations in milk. Thirty-two midlactation Holstein cows were fed a basal diet of 56% forage, 36.6% concentrate, and 7.4% roasted whole soybeans (dry basis) that was top-dressed with a premix that provided 0, 3, 16.5, or 30 g/d of L-ascorbic acid (provided by ascorbyl-2-polyphosphate) for 28 d. Supplementation had no effect on milk yield or composition or dry matter intake. Treatment linearly increased plasma concentrations of ascorbic acid (19.8, 22.3, 21.9, and 25.7 µmol/L, respectively) but had no effect on plasma dehydro-L-ascorbic acid (DHAA). Concentrations of ascorbic acid (103.7 µmol/L) and DHAA (9.5 µmol/L) in milk were not affected by treatment. Secretion of ascorbic acid into milk appeared to follow Michaelis-Menton kinetics, with a Vmax of 3.92 mmol/d and a Km of 3.59 µmol/L. Milk flavor as evaluated by a panel was normal for all samples after 1 d of storage. After 7 d of storage, the average flavor score was 2.5 (moderate oxidized flavor), but no differences among treatments were observed. Supplemental dietary vitamin C did not increase vitamin C concentration in milk, probably because the maximum potential secretion of the vitamin was occurring in unsupplemented cows. (Key words: vitamin C, ascorbic acid, oxidized flavor) Abbreviation key: AA = L-ascorbic acid; DHAA = dehydro-L-ascorbic acid. INTRODUCTION Vitamin C is not an essential dietary nutrient for adult dairy cattle; however, it is an important water-
Received January 26, 2001. Accepted May 14, 2001. E-mail: [email protected]. 1 Salaries and research support provided by state and federal funds appropriated to the Ohio Agricultural Research and Development Center, The Ohio State University. Additional support provided by Roche Vitamins, Parsippany, NJ.
soluble antioxidant found in many tissues of the cow. The concentrations of L-ascorbic acid (AA) plus dehydro-L-ascorbic acid (DHAA), because it is readily converted to AA, equal the vitamin C concentration in biological samples. Although not an essential nutrient, AA administered subcutaneously to steers improved various measures of neutrophil function (Roth and Kaeberle, 1985). Ascorbic acid is quickly degraded in in vitro rumen systems; the half-life of AA was 3.5 h compared with 6.9 h for AA from ascorbyl-2-polyphosphate (Macleod et al., 1999b). Even though AA appears to be degraded in the rumen, dietary supplementation of high doses (20 g/d from ascorbyl-2-polyphosphate or 40 g/d from L-ascorbic acid) of AA has increased plasma concentrations of AA in cattle (Macleod et al., 1999a; Hidiroglou, 1999, respectively). The effect of dietary supplementation of AA on milk AA concentrations has not been investigated. Changing the concentration of AA in milk could affect the development of oxidized flavor in milk. Barrefors et al. (1995) reported that for one of two test herds, oxidized milk had lower than average concentrations of DHAA and total vitamin C (no difference was found for AA) than normal flavored milk (based on principal component analysis). In the other test herd, concentrations of AA, DHAA, and total vitamin C was not significantly related to milk flavor. Exogenous addition of 200 mg/L of AA (lower concentrations were not tested) to skim milk reduced oxidized flavor development (Jung et al., 1998). In low concentrations, AA can reduce transition metals (e.g., copper) and promote lipid oxidation (Halliwell, 1996). The potential for small concentrations of AA in milk to initiate oxidation has not been tested. At higher concentrations, the antioxidant properties of AA should inhibit propagation of oxidation. Because AA supplementation could potentially affect milk flavor, a study was conducted to determine whether concentrations of AA in milk were changed when AA was supplemented via the diet and to determine the relationship between plasma and milk concentrations of AA. MATERIALS AND METHODS Thirty-two Holstein cows (DIM = 185; SE = 7) were fed a basal diet (Table 1) balanced to meet or exceed
2302
VITAMIN C IN MILK Table 1. Ingredient and nutrient composition of the basal diet. Component
% of DM
Alfalfa silage Corn silage Roasted whole soybeans Corn grain, ground Soybean meal, 44% CP Soybean hulls Wheat middlings Urea RUP supplement1 Molasses Macromineral supplement2 Trace mineral-vitamin supplement3 Nutrient CP, % NDF, % Cu, mg/kg Zn, mg/kg
15.20 41.10 7.40 16.04 9.82 3.41 2.55 0.24 1.08 0.69 2.26 0.21 17.6 30.2 15 71
1 Contained 33.3% fish meal, 33.3% blood meal, and 33.4% corn gluten meal. 2 Contained 52.65% limestone, 18.14% sodium bicarbonate, 11.06% dicalcium phosphate, 10.62% sodium chloride, and 7.53% magnesium oxide. 3 Contained 48.31% selenium premix (200 mg of Se/kg), 3.38% Zinpro 100 (Zinpro Corp., Edin Prairie, MN), 1.45% zinc oxide, 1.45% copper sulfate, and 45.41% vitamin A, D, and E premix (provided 5700 IU of vitamin A, 960 IU of vitamin D, and 20 IU of vitamin E/ kg of diet DM.
NRC (1989) recommendations. Roasted soybeans were included because they are related to increased oxidized flavor in milk probably via increased concentrations of linolenic and linoleic acids in milk fat (Timmons et al., 2001). The diet was formulated to provide 400 IU/d of supplemental vitamin E. All cows were fed that basal diet for at least 1 wk before the start of the experiment. Cows were placed into blocks (based on DIM and previous milk production) and randomly assigned to one of four treatments within each block. The basal diet was top dressed with 0, 3, 16.5, and 30 g of AA (provided as sodium and calcium ascorbyl-2-phosphate; Rovimix Stay C-35, Roche Vitamins, Parsippany, NJ). The 3 g/ d treatment was considered an economically reasonable inclusion rate (approximate cost $0.12/d); the 30 g/d treatment was chosen because that rate has been shown to elevate plasma AA concentrations in cattle. The 16.5 g/d treatment is the midpoint between the 3 and 30 g/ d treatments. Each morning, the AA supplement was mixed with a small portion of concentrate and given to the cows. After that mixture was consumed, the remainder of the ration was offered. Cows were housed in tie stalls, fed once daily, and milked twice daily. One day before supplementation began, samples of blood from the tail vein (approximately 3 h after feeding) and milk (p.m. and a.m.) were collected. On d 14 and 28, blood and milk (p.m. and a.m.) were sampled. Urine was spot
2303
sampled (one sample) from each cow on d 25 and 27 approximately 3 h after feeding. The a.m. and p.m. milk samples were composited (within cow) for each day and analyzed within 8 h after the a.m. milking for AA using HPLC (Timmons et al., 2001). Plasma and urine were analyzed immediately after sampling for AA using HPLC. Subsamples of plasma, urine, and milk were immediately acidified and reduced (dithiothreitol) to convert DHAA to AA (Timmons et al., 2001) and analyzed within 3 d for AA (this represents total vitamin C). The concentration of DHAA in plasma, urine, and milk was calculated by subtracting AA from total vitamin C. Plasma samples were also analyzed for α-tocopherol (Weiss et al., 1990). A subsample of milk was pasteurized and sent to another laboratory for flavor panel evaluation. Milk for flavor evaluation was stored in a refrigerator (4°C) for 1 and 7 d before evaluation by four panel members (blind to treatments) and scored (0 to 1, no oxidized flavor; 1 to 2, slight off flavor; 2 to 3, moderate off-flavor; 3 to 5, strong flavor). Separate samples of milk (a.m. and p.m.) were collected every 2 wk and analyzed for fat and CP (AOAC, 1990) with a B2000 Infrared Analyzer (Bentley Instruments, Chaska, MN) by Ohio DHI Cooperative (Powell, OH). Milk production and DMI data were analyzed using Proc GLM (SAS, 1999) with block and treatment in the model. The treatment effect was partitioned into linear, quadratic, and cubic orthogonal contrasts. Milk (milk flavor scores were averaged for the four panel members) and blood data were analyzed using the MIXED procedure (SAS, 1999). The model included block (random), treatment, day (repeated measure), and the treatment × day interaction. Based on Akaike’s criterion, the autoregressive covariance structure was used. Treatment effects were partitioned into linear, quadratic, and cubic (treatment only) contrasts. Treatment contrasts were adjusted for unequal spacing of treatments. Day effect was partitioned into linear and quadratic contrasts. In a preliminary statistical analysis, d 0 concentrations of AA and DHAA in blood and milk variables were used as covariates. Day 0 blood values were not related (P > 0.10) to values on d 14 and 28. Day 0 concentrations of AA (but not DHAA) was a significant source of variation (P < 0.05) for d 14 and 28 milk AA concentrations. Inclusion of the covariate did not change the statistical interpretation of treatment, day, and interaction effects. Therefore, means presented in this paper are not covariate adjusted. Urine data for d 25 and 27 were averaged and analyzed using the same model as was used for milk and DMI data. Correlations among plasma and milk variables were determined using Proc CORR (SAS, 1999). Journal of Dairy Science Vol. 84, No. 10, 2001
2304
WEISS
0.05) over time (0.10, 0.11, and 0.15 for d 0, 14, and 28, respectively; data not shown). Plasma concentrations of α-tocopherol were not affected by treatment or time (Table 2) and suggest that the cows were in good vitamin E status (Weiss et al., 1990). Concentrations of αtocopherol in plasma were not correlated (P > 0.20) with plasma AA or DHAA. Concentrations of AA and DHAA in milk were not affected by treatment or day (Table 3) and were similar to previously reported concentrations (Hartman and Dryden, 1978; Hidiroglou et al., 1995) for fresh raw milk. Ascorbic acid, but not DHAA, is secreted into milk (Hartman and Dryden, 1978). Concentrations of AA in milk can decrease rapidly by first being oxidized reversibly to DHAA, which can then be oxidized irreversibly to other compounds. The AA concentration in bovine (Timmons et al., 2001) and caprine (Jandal, 1996) milk decreased 70% after 3 d of storage at 4°C. In the present study, DHAA comprised about 9% of total AA in milk. Barrefors et al. (1995) reported concentrations of vitamin C (AA plus DHAA) similar to values in this study, but approximately 70% of the total vitamin C was DHAA; milk in that study was stored frozen before analysis, which can reduce AA concentrations (Jandal, 1996). Because DHAA in milk was formed from AA postsecretion, the concentrations of AA and DHAA were summed to represent total vitamin C in milk. That concentration times the volume of milk produced on the day milk was sampled equals the quantity of vitamin
RESULTS AND DISCUSSION Milk yield (mean = 30.2 kg/d; SE = 1.0), milk fat (3.60%; SE = 0.16), milk CP (3.22%; SE = 0.06), and DMI (mean = 20.8 kg/d; SE = 0.5) were not affected by treatment (data not shown). No (P > 0.15) treatment × day interactions were found for any blood or milk variable. The concentration of AA in plasma increased linearly (P < 0.05) as dietary vitamin C increased (Table 2), and concentrations changed quadratically (P < 0.05) with time (maximum at 14 d). The concentrations of plasma AA from control cows were similar to other reports (Hidiroglou et al., 1995; Hidiroglou, 1999; Macleod et al., 1999a). At d 28, mean plasma concentration of AA for cows fed 30 g/d of AA was about 25% higher than that of control cows, which is similar to the increase when 20 (Macleod et al. (1999a) or 40 (Hidiroglou, 1999) g/d of AA were supplemented to cows. Plasma DHAA was not affected by treatment but increased linearly (P < 0.05) over days. Ascorbic acid can be reversibly oxidized to DHAA, which can then be oxidized irreversibly to diketo-L-gulonic acid. Therefore, the sum of AA and DHAA represents the total concentration of biologically active vitamin C in the plasma. Similar to AA, the concentration of total vitamin C in plasma increased linearly (P < 0.01) with increasing supplementation and showed a quadratic effect (P < 0.05) over days (data not shown). The proportion of total vitamin C in plasma that was DHAA was not affected by treatment but increased linearly (P <
Table 2. Effect of ascorbic acid (AA) supplementation and day of supplementation on plasma and urine concentrations of AA, α-tocopherol, and dehydroascorbic acid (DHAA).1 Supplemental vitamin C 0 g/d
3 g/d
16.5 g/d
30 g/d
SE
µmol/L Plasma AAa Day 0 Day 14 Day 28 Mean (excluding d 0)b Plasma DHAAc Day 0 Day 14 Day 28 Plasma α-tocopherolc Day 0 Day 14 Day 28 Urine AAd Urine DHAAd
17.75 20.98 18.61 19.80
21.07 23.80 20.79 22.30
19.53 23.42 21.22 22.32
22.34 28.35 23.35 25.85
1.93 2.18 2.36 1.56
1.76 2.70 3.26
1.66 1.99 3.04
2.51 3.17 3.83
2.03 3.23 3.58
0.60 0.77 1.15
15.0 15.4 13.3
17.4 16.9 15.6
16.9 16.1 14.4
16.6 15.3 15.0
249.8 262.7
205.8 218.9
168.1 183.3
211.6 221.1
Quadratic day effect for AA (P < 0.05). Linear treatment effect (P < 0.05). c Linear day effect (P < 0.05); no treatment effect. d Quadratic treatment effect (P < 0.05). 1 No treatment × time interaction for any variable (P > 0.15). a b
Journal of Dairy Science Vol. 84, No. 10, 2001
1.29 1.20 1.18 20.0 21.1
2305
VITAMIN C IN MILK Table 3. Effect of ascorbic acid (AA) supplementation and day of supplementation on milk concentrations of AA and dehydroascorbic acid (DHAA) and milk flavor.1 Supplemental vitamin C
Milk AA, µmol/L Day 0 Day 14 Day 28
0 g/d
3 g/d
16.5 g/d
30 g/d
SE
92.07 103.01 95.53
112.17 116.79 107.80
103.12 104.50 100.65
98.50 105.24 104.07
5.27 6.67 6.4
Milk DHAA, µmol/L Day 0 Day 14 Day 28 Milk vitamin C2, mmol/d Day 0 Day 14 Day 28
8.58 8.03 9.48
7.07 6.43 9.48
10.90 7.16 13.15
10.44 11.38 12.30
2.88 2.30 3.30
3.02 3.44 3.14
3.78 3.89 3.53
3.43 3.35 3.28
3.52 3.56 3.66
0.21 0.27 0.20
Milk flavor3 Day 14 Day 28
2.73 2.28
2.10 2.64
2.71 2.37
2.47 2.28
0.51 0.45
No effects of treatment, day of supplementation, or time × day interaction (P > 0.10). Milk vitamin C = Concentrations of AA + DHAA times the volume of milk produced on the day milk was sampled for AA. 3 Flavor score (0 to 1 = normal; 5 = intensively oxidized) of milk stored 7 d in the refrigerator (4°C). Flavor scores of milk stored 1 d were all ≤1. 1 2
C secreted in milk per day. The quantity of vitamin C secreted in milk was not affected by treatment or day (Table 3). All milk samples stored for 1 d had normal flavor (scores 0 or 1), but after 7 d of storage, 40% of the samples had flavor scores ≥3 (detectable by consumers). Treatment had no effect on flavor scores for 1- or 7-dold milk at d 14 or 28 (Table 3). The mean flavor scores for 7-d-old milk were 2.5, 2.4, 2.5, and 2.4 (SE = 0.4) for the 0, 3, 16.5, and 30 g/d treatments, respectively (scores averaged for d 14 and 28). Flavor scores were not correlated (P > 0.20) with concentrations of any measured variable in milk or plasma. With the scoring system used in this experiment, an oxidized flavor would be detectable by most consumers when milk had a flavor score of 3 or greater, and milk with a score less than 2 would taste normal to most consumers (J.W. Harper, personal communication). Therefore, a composite of the 7-d-old milk from this experiment would have a detectable oxidized flavor by many consumers. Mean flavor score (same scale as used in the present experiment) of bulk tank milk from 20 herds in Ohio sampled four times was 2.1 after 8 d of storage (Timmons et al., 2001). Several factors influence the development of oxidized flavor in milk including fatty acid composition (especially polyunsaturated fatty acids), concentrations of transition metals such as copper, and the concentrations of antioxidants such as α-tocopherol (Timmons et al., 2001). Based on a large field study of herds feeding different concentrations of roasted soybeans, concentra-
tions of linoleic and linolenic acid in milk fat were the major factors related to oxidized flavor (Timmons et al., 2001). The development of oxidized flavor increased as the dietary concentration of roasted soybeans (mean, 4.2; range 0 to 15.3% of diet DM) increased. In that study, the concentration of DHAA in milk was positively correlated with the development of oxidized flavor, but the effect was quantitatively small (increasing DHAA concentration tenfold resulted in an average increase of one flavor score unit). Dietary vitamin E concentrations (mean, 50; range, 32 to 71 IU/kg of DM) were not related to the development of oxidized flavor in that study (Timmons et al., 2001). Those results suggest that the basal diet fed in the current experiment should have provided a good test for the effect of supplemental vitamin C on the development of oxidized flavor. Only about 40% of the 7-d-old samples had moderate to strong oxidized flavor; therefore, the number of observations in this study (8/treatment) might be limited for flavor scores. However, the lack of any numerical trend and the lack of changes in milk vitamin C concentrations suggests that the lack of a treatment effect on milk flavor score was not caused by insufficient statistical power. Mean concentrations of AA and DHAA in milk were about 4.5 and 3.5 times higher, respectively, than the plasma concentrations. Concentrations of AA in plasma were not correlated with concentrations of AA or total vitamin C (AA plus DHAA) in milk; however, a correlation (r = 0.42; P < 0.001) was found between concentrations of DHAA in plasma and milk. Because DHAA has not been shown to be secreted into milk, this correlation Journal of Dairy Science Vol. 84, No. 10, 2001
2306
WEISS
suggests that compounds that promote conversion of AA to DHAA (or conversely compounds that prevent conversion) are found in both plasma and milk and their concentrations are correlated. The lack of a linear relationship between plasma and milk concentrations of AA suggests that AA does not passively diffuse from plasma into milk. Michaelis-Menton kinetics were used to determine the relationship between plasma concentrations of AA and secretion of AA in milk, as described by Jensen et al. (1999) to quantify secretion of α-tocopherol in milk. The quantity of AA secreted in milk was determined by multiplying the concentration of total vitamin C (AA plus DHAA) at each time point by the volume of milk produced at each sample day. Total vitamin C, rather than AA, was used because any DHAA found in milk was probably produced after milk secretion. These values were then plotted against plasma concentrations of AA (Figure 1A). In Michaelis-
Menton terminology, the concentration of AA in plasma represents substrate concentration (S) and the amount of total vitamin C secreted in milk/d represents velocity of product secretion (V). In a Hanes plot, S is plotted against S/V to linearize the relationship between S and V. Plotting the concentrations of AA in plasma divided by the quantity of total vitamin C secreted in milk (i.e., S/V) versus the concentrations of AA in plasma (i.e., S; Figure 1B) resulted in the linear regression equation: S/V = 0.2548 (SE = 0.0233) × S + 0.00091 (SE = 0.00052), where S = concentration of AA in plasma (mmol/L) and V = quantity of AA secreted in milk (mmol/d). The inverse of the slope (1/0.2548) = mean maximum quantity produced per day (i.e., Vmax = 3.92 mmol) and the intercept times Vmax = Km = 0.0036 mmol/L. The linear Hanes plot (Figure 1B) suggests that AA secreted into milk follows Michaelis-Menton kinetics and that the mean maximum secretion is 3.92 mmol/d (SE = 0.33). The mean concentration of AA in plasma of control cows was approximately five times greater than the Km, suggesting that the maximum possible daily quantity of vitamin C is secreted into milk by midlactation cows with normal concentrations of AA in plasma. High concentrations of both AA and DHAA were found in urine (Table 2). A quadratic effect (P < 0.03) of treatment was observed with the highest concentration for cows fed 0 g/d, and the lowest concentrations were observed for cows fed 3 and 16.5 g/d. The reason for the quadratic effect is unclear. Urinary excretion of AA is essentially nil in humans when fed diets that are deficient or just at requirement in vitamin C; urinary excretion only occurs when excess vitamin C is consumed (Levine et al., 1996). If this is true for cows, then all cows were above requirements for vitamin C. Urine volume was not measured, so these data are equivocal; however, they suggest that intake and tissue synthesis of AA by cows was in excess of requirements. In conclusion, supplemental dietary vitamin C increased plasma concentrations of AA, but this did not correspond to increased concentration of AA in milk. Because concentrations of AA in milk were not altered by dietary vitamin C, no effects of treatment were observed on milk flavor. In this experiment, the maximum amount of AA that could be secreted into milk each day was 3.92 mmol. REFERENCES
Figure 1. A) Quantity of total vitamin C secreted daily in milk (ascorbic acid plus dehydroascorbic acid) versus concentration of ascorbic acid (AA) in plasma. The solid line represents the MichaelisMention equation; Milk vitamin C (mmol/d) = Vmax × [plasma AA]/ Km + [plasma AA]). B) Hanes plot of data in panel A. The solid line represents: Y = 0.2589X + 0.000912. The inverse of the slope = Vmax = 3.92 mmol/d; the intercept times Vmax = Km = 0.0036 mmol/L. Journal of Dairy Science Vol. 84, No. 10, 2001
Barrefors, P., K. Granelli, L.-A. Appelqvist, and L. Bjoerck. 1995. Chemical characterization of raw milk samples with and without oxidative off-flavor. J. Dairy Sci. 78:2691–2699. Halliwell, B. 1996. Vitamin C: Antioxidant or pro-oxidant in vivo? Free Rad. Res. 25:439–454. Hartman, A. M., and L. P. Dryden. 1978. The vitamins in milk and milk products. Pages 325–401 in Fundamentals of Dairy Chemis-
VITAMIN C IN MILK try, 2nd ed. B. H. Webb, A. H. Johnson, and J. A. Alford, ed. AVI Publishing Co., Inc., Westport, CT. Hidiroglou, M., M. Ivan, and T. R. Batra. 1995. Concentrations of vitamin C in plasma and milk of dairy cattle. Ann. Zootech. 44:399–402. Hidiroglou, M. 1999. Technical note: Forms and route of vitamin C supplementation for cows. J. Dairy Sci. 82:1831–1833. Jandal, J. M. 1996. Factors affecting ascorbic acid content and keeping quality of Shammi goat milk. Small Rumin. Res. 21:121–125. Jensen, S. K., A. K. B. Johannsen, and J. E. Hermansen. 1999. Quantitative secretion and maximal secretion capacity of retinol, betacarotene and alpha-tocopherol into cows’ milk. J. Dairy Res. 66:511–522. Jung, M. Y., S. H. Yoon, H. O. Lee, and D. B. Min. 1998. Singlet oxygen and ascorbic acid effects on dimethyl disulfide and offflavor in skim milk exposed to light. J. Food Sci. 63:408–412. Levine, M., C. Conry-Cantilena, Y. Wang, R. W. Welch, P. W. Wasko, K. R. Dhariwal, J. B. Park, A. Lazarev, J. F. Graumlich, J. King, and L. R. Cantilena. 1996. Vitamin C pharmacokinetics in healthy
2307
volunteers: evidence for a recommended dietary allowance. Proc. Natl. Acad. Sci. USA 93:3704–3709. Macleod, D. D., X. Zhang, J. J. Kennelly, and L. Ozimek. 1999a. Ascorbyl-2-polyphosphate as a source of ascorbic acid for dairy cattle. Milchwissenschaft 54:123–126. Macleod, D. D., X. Zhang, J. J. Kennelly, and L. Ozimek. 1999b. Pharmokinetics of ascorbic acid and ascorbyl-2-polyphosphate in the rumen fluid of dairy cows. Milchwissenschaft 54:63–65. National Research Council. 1989. Nutrient requirements for dairy cattle. 6th rev. ed. Natl. Acad. Sci., Washington, DC. Roth, J. A., and M. L. Kaeberle. 1985. In vivo effect of ascorbic acid on neutrophil function in healthy and dexamethasone-treated cattle. Amer. J. Vet. Res. 46:2434–2436. SAS. 1999. SAS/STAT User’s Guide, Version 8. SAS Inst., Cary, NC. Timmons, J. S., W. P. Weiss, D. L. Palmquist, and W. J. Harper. 2001. Relationships among dietary roasted soybeans, milk components, and spontaneously oxidized flavor of milk. J. Dairy Sci. (accepted). Weiss, W. P., D. A. Todhunter, J. S. Hogan, and K. L. Smith. 1990. Effect of duration of supplementation of selenium and vitamin E on periparturient dairy cows. J. Dairy Sci. 73:3187–3194.
Journal of Dairy Science Vol. 84, No. 10, 2001
Effect of Dietary Vitamin C on Concentrations of Ascorbic Acid in Plasma and Milk1 W. P. Weiss Department of Animal Sciences Ohio Agricultural Research and Development Center The Ohio State University, Wooster 44691
ABSTRACT The addition of exogenous ascorbic acid to milk reduces the development of oxidized flavor. This experiment was conducted to determine whether feeding ascorbic acid to cows influenced vitamin C concentrations in milk. Thirty-two midlactation Holstein cows were fed a basal diet of 56% forage, 36.6% concentrate, and 7.4% roasted whole soybeans (dry basis) that was top-dressed with a premix that provided 0, 3, 16.5, or 30 g/d of L-ascorbic acid (provided by ascorbyl-2-polyphosphate) for 28 d. Supplementation had no effect on milk yield or composition or dry matter intake. Treatment linearly increased plasma concentrations of ascorbic acid (19.8, 22.3, 21.9, and 25.7 µmol/L, respectively) but had no effect on plasma dehydro-L-ascorbic acid (DHAA). Concentrations of ascorbic acid (103.7 µmol/L) and DHAA (9.5 µmol/L) in milk were not affected by treatment. Secretion of ascorbic acid into milk appeared to follow Michaelis-Menton kinetics, with a Vmax of 3.92 mmol/d and a Km of 3.59 µmol/L. Milk flavor as evaluated by a panel was normal for all samples after 1 d of storage. After 7 d of storage, the average flavor score was 2.5 (moderate oxidized flavor), but no differences among treatments were observed. Supplemental dietary vitamin C did not increase vitamin C concentration in milk, probably because the maximum potential secretion of the vitamin was occurring in unsupplemented cows. (Key words: vitamin C, ascorbic acid, oxidized flavor) Abbreviation key: AA = L-ascorbic acid; DHAA = dehydro-L-ascorbic acid. INTRODUCTION Vitamin C is not an essential dietary nutrient for adult dairy cattle; however, it is an important water-
Received January 26, 2001. Accepted May 14, 2001. E-mail: [email protected]. 1 Salaries and research support provided by state and federal funds appropriated to the Ohio Agricultural Research and Development Center, The Ohio State University. Additional support provided by Roche Vitamins, Parsippany, NJ.
soluble antioxidant found in many tissues of the cow. The concentrations of L-ascorbic acid (AA) plus dehydro-L-ascorbic acid (DHAA), because it is readily converted to AA, equal the vitamin C concentration in biological samples. Although not an essential nutrient, AA administered subcutaneously to steers improved various measures of neutrophil function (Roth and Kaeberle, 1985). Ascorbic acid is quickly degraded in in vitro rumen systems; the half-life of AA was 3.5 h compared with 6.9 h for AA from ascorbyl-2-polyphosphate (Macleod et al., 1999b). Even though AA appears to be degraded in the rumen, dietary supplementation of high doses (20 g/d from ascorbyl-2-polyphosphate or 40 g/d from L-ascorbic acid) of AA has increased plasma concentrations of AA in cattle (Macleod et al., 1999a; Hidiroglou, 1999, respectively). The effect of dietary supplementation of AA on milk AA concentrations has not been investigated. Changing the concentration of AA in milk could affect the development of oxidized flavor in milk. Barrefors et al. (1995) reported that for one of two test herds, oxidized milk had lower than average concentrations of DHAA and total vitamin C (no difference was found for AA) than normal flavored milk (based on principal component analysis). In the other test herd, concentrations of AA, DHAA, and total vitamin C was not significantly related to milk flavor. Exogenous addition of 200 mg/L of AA (lower concentrations were not tested) to skim milk reduced oxidized flavor development (Jung et al., 1998). In low concentrations, AA can reduce transition metals (e.g., copper) and promote lipid oxidation (Halliwell, 1996). The potential for small concentrations of AA in milk to initiate oxidation has not been tested. At higher concentrations, the antioxidant properties of AA should inhibit propagation of oxidation. Because AA supplementation could potentially affect milk flavor, a study was conducted to determine whether concentrations of AA in milk were changed when AA was supplemented via the diet and to determine the relationship between plasma and milk concentrations of AA. MATERIALS AND METHODS Thirty-two Holstein cows (DIM = 185; SE = 7) were fed a basal diet (Table 1) balanced to meet or exceed
2302
VITAMIN C IN MILK Table 1. Ingredient and nutrient composition of the basal diet. Component
% of DM
Alfalfa silage Corn silage Roasted whole soybeans Corn grain, ground Soybean meal, 44% CP Soybean hulls Wheat middlings Urea RUP supplement1 Molasses Macromineral supplement2 Trace mineral-vitamin supplement3 Nutrient CP, % NDF, % Cu, mg/kg Zn, mg/kg
15.20 41.10 7.40 16.04 9.82 3.41 2.55 0.24 1.08 0.69 2.26 0.21 17.6 30.2 15 71
1 Contained 33.3% fish meal, 33.3% blood meal, and 33.4% corn gluten meal. 2 Contained 52.65% limestone, 18.14% sodium bicarbonate, 11.06% dicalcium phosphate, 10.62% sodium chloride, and 7.53% magnesium oxide. 3 Contained 48.31% selenium premix (200 mg of Se/kg), 3.38% Zinpro 100 (Zinpro Corp., Edin Prairie, MN), 1.45% zinc oxide, 1.45% copper sulfate, and 45.41% vitamin A, D, and E premix (provided 5700 IU of vitamin A, 960 IU of vitamin D, and 20 IU of vitamin E/ kg of diet DM.
NRC (1989) recommendations. Roasted soybeans were included because they are related to increased oxidized flavor in milk probably via increased concentrations of linolenic and linoleic acids in milk fat (Timmons et al., 2001). The diet was formulated to provide 400 IU/d of supplemental vitamin E. All cows were fed that basal diet for at least 1 wk before the start of the experiment. Cows were placed into blocks (based on DIM and previous milk production) and randomly assigned to one of four treatments within each block. The basal diet was top dressed with 0, 3, 16.5, and 30 g of AA (provided as sodium and calcium ascorbyl-2-phosphate; Rovimix Stay C-35, Roche Vitamins, Parsippany, NJ). The 3 g/ d treatment was considered an economically reasonable inclusion rate (approximate cost $0.12/d); the 30 g/d treatment was chosen because that rate has been shown to elevate plasma AA concentrations in cattle. The 16.5 g/d treatment is the midpoint between the 3 and 30 g/ d treatments. Each morning, the AA supplement was mixed with a small portion of concentrate and given to the cows. After that mixture was consumed, the remainder of the ration was offered. Cows were housed in tie stalls, fed once daily, and milked twice daily. One day before supplementation began, samples of blood from the tail vein (approximately 3 h after feeding) and milk (p.m. and a.m.) were collected. On d 14 and 28, blood and milk (p.m. and a.m.) were sampled. Urine was spot
2303
sampled (one sample) from each cow on d 25 and 27 approximately 3 h after feeding. The a.m. and p.m. milk samples were composited (within cow) for each day and analyzed within 8 h after the a.m. milking for AA using HPLC (Timmons et al., 2001). Plasma and urine were analyzed immediately after sampling for AA using HPLC. Subsamples of plasma, urine, and milk were immediately acidified and reduced (dithiothreitol) to convert DHAA to AA (Timmons et al., 2001) and analyzed within 3 d for AA (this represents total vitamin C). The concentration of DHAA in plasma, urine, and milk was calculated by subtracting AA from total vitamin C. Plasma samples were also analyzed for α-tocopherol (Weiss et al., 1990). A subsample of milk was pasteurized and sent to another laboratory for flavor panel evaluation. Milk for flavor evaluation was stored in a refrigerator (4°C) for 1 and 7 d before evaluation by four panel members (blind to treatments) and scored (0 to 1, no oxidized flavor; 1 to 2, slight off flavor; 2 to 3, moderate off-flavor; 3 to 5, strong flavor). Separate samples of milk (a.m. and p.m.) were collected every 2 wk and analyzed for fat and CP (AOAC, 1990) with a B2000 Infrared Analyzer (Bentley Instruments, Chaska, MN) by Ohio DHI Cooperative (Powell, OH). Milk production and DMI data were analyzed using Proc GLM (SAS, 1999) with block and treatment in the model. The treatment effect was partitioned into linear, quadratic, and cubic orthogonal contrasts. Milk (milk flavor scores were averaged for the four panel members) and blood data were analyzed using the MIXED procedure (SAS, 1999). The model included block (random), treatment, day (repeated measure), and the treatment × day interaction. Based on Akaike’s criterion, the autoregressive covariance structure was used. Treatment effects were partitioned into linear, quadratic, and cubic (treatment only) contrasts. Treatment contrasts were adjusted for unequal spacing of treatments. Day effect was partitioned into linear and quadratic contrasts. In a preliminary statistical analysis, d 0 concentrations of AA and DHAA in blood and milk variables were used as covariates. Day 0 blood values were not related (P > 0.10) to values on d 14 and 28. Day 0 concentrations of AA (but not DHAA) was a significant source of variation (P < 0.05) for d 14 and 28 milk AA concentrations. Inclusion of the covariate did not change the statistical interpretation of treatment, day, and interaction effects. Therefore, means presented in this paper are not covariate adjusted. Urine data for d 25 and 27 were averaged and analyzed using the same model as was used for milk and DMI data. Correlations among plasma and milk variables were determined using Proc CORR (SAS, 1999). Journal of Dairy Science Vol. 84, No. 10, 2001
2304
WEISS
0.05) over time (0.10, 0.11, and 0.15 for d 0, 14, and 28, respectively; data not shown). Plasma concentrations of α-tocopherol were not affected by treatment or time (Table 2) and suggest that the cows were in good vitamin E status (Weiss et al., 1990). Concentrations of αtocopherol in plasma were not correlated (P > 0.20) with plasma AA or DHAA. Concentrations of AA and DHAA in milk were not affected by treatment or day (Table 3) and were similar to previously reported concentrations (Hartman and Dryden, 1978; Hidiroglou et al., 1995) for fresh raw milk. Ascorbic acid, but not DHAA, is secreted into milk (Hartman and Dryden, 1978). Concentrations of AA in milk can decrease rapidly by first being oxidized reversibly to DHAA, which can then be oxidized irreversibly to other compounds. The AA concentration in bovine (Timmons et al., 2001) and caprine (Jandal, 1996) milk decreased 70% after 3 d of storage at 4°C. In the present study, DHAA comprised about 9% of total AA in milk. Barrefors et al. (1995) reported concentrations of vitamin C (AA plus DHAA) similar to values in this study, but approximately 70% of the total vitamin C was DHAA; milk in that study was stored frozen before analysis, which can reduce AA concentrations (Jandal, 1996). Because DHAA in milk was formed from AA postsecretion, the concentrations of AA and DHAA were summed to represent total vitamin C in milk. That concentration times the volume of milk produced on the day milk was sampled equals the quantity of vitamin
RESULTS AND DISCUSSION Milk yield (mean = 30.2 kg/d; SE = 1.0), milk fat (3.60%; SE = 0.16), milk CP (3.22%; SE = 0.06), and DMI (mean = 20.8 kg/d; SE = 0.5) were not affected by treatment (data not shown). No (P > 0.15) treatment × day interactions were found for any blood or milk variable. The concentration of AA in plasma increased linearly (P < 0.05) as dietary vitamin C increased (Table 2), and concentrations changed quadratically (P < 0.05) with time (maximum at 14 d). The concentrations of plasma AA from control cows were similar to other reports (Hidiroglou et al., 1995; Hidiroglou, 1999; Macleod et al., 1999a). At d 28, mean plasma concentration of AA for cows fed 30 g/d of AA was about 25% higher than that of control cows, which is similar to the increase when 20 (Macleod et al. (1999a) or 40 (Hidiroglou, 1999) g/d of AA were supplemented to cows. Plasma DHAA was not affected by treatment but increased linearly (P < 0.05) over days. Ascorbic acid can be reversibly oxidized to DHAA, which can then be oxidized irreversibly to diketo-L-gulonic acid. Therefore, the sum of AA and DHAA represents the total concentration of biologically active vitamin C in the plasma. Similar to AA, the concentration of total vitamin C in plasma increased linearly (P < 0.01) with increasing supplementation and showed a quadratic effect (P < 0.05) over days (data not shown). The proportion of total vitamin C in plasma that was DHAA was not affected by treatment but increased linearly (P <
Table 2. Effect of ascorbic acid (AA) supplementation and day of supplementation on plasma and urine concentrations of AA, α-tocopherol, and dehydroascorbic acid (DHAA).1 Supplemental vitamin C 0 g/d
3 g/d
16.5 g/d
30 g/d
SE
µmol/L Plasma AAa Day 0 Day 14 Day 28 Mean (excluding d 0)b Plasma DHAAc Day 0 Day 14 Day 28 Plasma α-tocopherolc Day 0 Day 14 Day 28 Urine AAd Urine DHAAd
17.75 20.98 18.61 19.80
21.07 23.80 20.79 22.30
19.53 23.42 21.22 22.32
22.34 28.35 23.35 25.85
1.93 2.18 2.36 1.56
1.76 2.70 3.26
1.66 1.99 3.04
2.51 3.17 3.83
2.03 3.23 3.58
0.60 0.77 1.15
15.0 15.4 13.3
17.4 16.9 15.6
16.9 16.1 14.4
16.6 15.3 15.0
249.8 262.7
205.8 218.9
168.1 183.3
211.6 221.1
Quadratic day effect for AA (P < 0.05). Linear treatment effect (P < 0.05). c Linear day effect (P < 0.05); no treatment effect. d Quadratic treatment effect (P < 0.05). 1 No treatment × time interaction for any variable (P > 0.15). a b
Journal of Dairy Science Vol. 84, No. 10, 2001
1.29 1.20 1.18 20.0 21.1
2305
VITAMIN C IN MILK Table 3. Effect of ascorbic acid (AA) supplementation and day of supplementation on milk concentrations of AA and dehydroascorbic acid (DHAA) and milk flavor.1 Supplemental vitamin C
Milk AA, µmol/L Day 0 Day 14 Day 28
0 g/d
3 g/d
16.5 g/d
30 g/d
SE
92.07 103.01 95.53
112.17 116.79 107.80
103.12 104.50 100.65
98.50 105.24 104.07
5.27 6.67 6.4
Milk DHAA, µmol/L Day 0 Day 14 Day 28 Milk vitamin C2, mmol/d Day 0 Day 14 Day 28
8.58 8.03 9.48
7.07 6.43 9.48
10.90 7.16 13.15
10.44 11.38 12.30
2.88 2.30 3.30
3.02 3.44 3.14
3.78 3.89 3.53
3.43 3.35 3.28
3.52 3.56 3.66
0.21 0.27 0.20
Milk flavor3 Day 14 Day 28
2.73 2.28
2.10 2.64
2.71 2.37
2.47 2.28
0.51 0.45
No effects of treatment, day of supplementation, or time × day interaction (P > 0.10). Milk vitamin C = Concentrations of AA + DHAA times the volume of milk produced on the day milk was sampled for AA. 3 Flavor score (0 to 1 = normal; 5 = intensively oxidized) of milk stored 7 d in the refrigerator (4°C). Flavor scores of milk stored 1 d were all ≤1. 1 2
C secreted in milk per day. The quantity of vitamin C secreted in milk was not affected by treatment or day (Table 3). All milk samples stored for 1 d had normal flavor (scores 0 or 1), but after 7 d of storage, 40% of the samples had flavor scores ≥3 (detectable by consumers). Treatment had no effect on flavor scores for 1- or 7-dold milk at d 14 or 28 (Table 3). The mean flavor scores for 7-d-old milk were 2.5, 2.4, 2.5, and 2.4 (SE = 0.4) for the 0, 3, 16.5, and 30 g/d treatments, respectively (scores averaged for d 14 and 28). Flavor scores were not correlated (P > 0.20) with concentrations of any measured variable in milk or plasma. With the scoring system used in this experiment, an oxidized flavor would be detectable by most consumers when milk had a flavor score of 3 or greater, and milk with a score less than 2 would taste normal to most consumers (J.W. Harper, personal communication). Therefore, a composite of the 7-d-old milk from this experiment would have a detectable oxidized flavor by many consumers. Mean flavor score (same scale as used in the present experiment) of bulk tank milk from 20 herds in Ohio sampled four times was 2.1 after 8 d of storage (Timmons et al., 2001). Several factors influence the development of oxidized flavor in milk including fatty acid composition (especially polyunsaturated fatty acids), concentrations of transition metals such as copper, and the concentrations of antioxidants such as α-tocopherol (Timmons et al., 2001). Based on a large field study of herds feeding different concentrations of roasted soybeans, concentra-
tions of linoleic and linolenic acid in milk fat were the major factors related to oxidized flavor (Timmons et al., 2001). The development of oxidized flavor increased as the dietary concentration of roasted soybeans (mean, 4.2; range 0 to 15.3% of diet DM) increased. In that study, the concentration of DHAA in milk was positively correlated with the development of oxidized flavor, but the effect was quantitatively small (increasing DHAA concentration tenfold resulted in an average increase of one flavor score unit). Dietary vitamin E concentrations (mean, 50; range, 32 to 71 IU/kg of DM) were not related to the development of oxidized flavor in that study (Timmons et al., 2001). Those results suggest that the basal diet fed in the current experiment should have provided a good test for the effect of supplemental vitamin C on the development of oxidized flavor. Only about 40% of the 7-d-old samples had moderate to strong oxidized flavor; therefore, the number of observations in this study (8/treatment) might be limited for flavor scores. However, the lack of any numerical trend and the lack of changes in milk vitamin C concentrations suggests that the lack of a treatment effect on milk flavor score was not caused by insufficient statistical power. Mean concentrations of AA and DHAA in milk were about 4.5 and 3.5 times higher, respectively, than the plasma concentrations. Concentrations of AA in plasma were not correlated with concentrations of AA or total vitamin C (AA plus DHAA) in milk; however, a correlation (r = 0.42; P < 0.001) was found between concentrations of DHAA in plasma and milk. Because DHAA has not been shown to be secreted into milk, this correlation Journal of Dairy Science Vol. 84, No. 10, 2001
2306
WEISS
suggests that compounds that promote conversion of AA to DHAA (or conversely compounds that prevent conversion) are found in both plasma and milk and their concentrations are correlated. The lack of a linear relationship between plasma and milk concentrations of AA suggests that AA does not passively diffuse from plasma into milk. Michaelis-Menton kinetics were used to determine the relationship between plasma concentrations of AA and secretion of AA in milk, as described by Jensen et al. (1999) to quantify secretion of α-tocopherol in milk. The quantity of AA secreted in milk was determined by multiplying the concentration of total vitamin C (AA plus DHAA) at each time point by the volume of milk produced at each sample day. Total vitamin C, rather than AA, was used because any DHAA found in milk was probably produced after milk secretion. These values were then plotted against plasma concentrations of AA (Figure 1A). In Michaelis-
Menton terminology, the concentration of AA in plasma represents substrate concentration (S) and the amount of total vitamin C secreted in milk/d represents velocity of product secretion (V). In a Hanes plot, S is plotted against S/V to linearize the relationship between S and V. Plotting the concentrations of AA in plasma divided by the quantity of total vitamin C secreted in milk (i.e., S/V) versus the concentrations of AA in plasma (i.e., S; Figure 1B) resulted in the linear regression equation: S/V = 0.2548 (SE = 0.0233) × S + 0.00091 (SE = 0.00052), where S = concentration of AA in plasma (mmol/L) and V = quantity of AA secreted in milk (mmol/d). The inverse of the slope (1/0.2548) = mean maximum quantity produced per day (i.e., Vmax = 3.92 mmol) and the intercept times Vmax = Km = 0.0036 mmol/L. The linear Hanes plot (Figure 1B) suggests that AA secreted into milk follows Michaelis-Menton kinetics and that the mean maximum secretion is 3.92 mmol/d (SE = 0.33). The mean concentration of AA in plasma of control cows was approximately five times greater than the Km, suggesting that the maximum possible daily quantity of vitamin C is secreted into milk by midlactation cows with normal concentrations of AA in plasma. High concentrations of both AA and DHAA were found in urine (Table 2). A quadratic effect (P < 0.03) of treatment was observed with the highest concentration for cows fed 0 g/d, and the lowest concentrations were observed for cows fed 3 and 16.5 g/d. The reason for the quadratic effect is unclear. Urinary excretion of AA is essentially nil in humans when fed diets that are deficient or just at requirement in vitamin C; urinary excretion only occurs when excess vitamin C is consumed (Levine et al., 1996). If this is true for cows, then all cows were above requirements for vitamin C. Urine volume was not measured, so these data are equivocal; however, they suggest that intake and tissue synthesis of AA by cows was in excess of requirements. In conclusion, supplemental dietary vitamin C increased plasma concentrations of AA, but this did not correspond to increased concentration of AA in milk. Because concentrations of AA in milk were not altered by dietary vitamin C, no effects of treatment were observed on milk flavor. In this experiment, the maximum amount of AA that could be secreted into milk each day was 3.92 mmol. REFERENCES
Figure 1. A) Quantity of total vitamin C secreted daily in milk (ascorbic acid plus dehydroascorbic acid) versus concentration of ascorbic acid (AA) in plasma. The solid line represents the MichaelisMention equation; Milk vitamin C (mmol/d) = Vmax × [plasma AA]/ Km + [plasma AA]). B) Hanes plot of data in panel A. The solid line represents: Y = 0.2589X + 0.000912. The inverse of the slope = Vmax = 3.92 mmol/d; the intercept times Vmax = Km = 0.0036 mmol/L. Journal of Dairy Science Vol. 84, No. 10, 2001
Barrefors, P., K. Granelli, L.-A. Appelqvist, and L. Bjoerck. 1995. Chemical characterization of raw milk samples with and without oxidative off-flavor. J. Dairy Sci. 78:2691–2699. Halliwell, B. 1996. Vitamin C: Antioxidant or pro-oxidant in vivo? Free Rad. Res. 25:439–454. Hartman, A. M., and L. P. Dryden. 1978. The vitamins in milk and milk products. Pages 325–401 in Fundamentals of Dairy Chemis-
VITAMIN C IN MILK try, 2nd ed. B. H. Webb, A. H. Johnson, and J. A. Alford, ed. AVI Publishing Co., Inc., Westport, CT. Hidiroglou, M., M. Ivan, and T. R. Batra. 1995. Concentrations of vitamin C in plasma and milk of dairy cattle. Ann. Zootech. 44:399–402. Hidiroglou, M. 1999. Technical note: Forms and route of vitamin C supplementation for cows. J. Dairy Sci. 82:1831–1833. Jandal, J. M. 1996. Factors affecting ascorbic acid content and keeping quality of Shammi goat milk. Small Rumin. Res. 21:121–125. Jensen, S. K., A. K. B. Johannsen, and J. E. Hermansen. 1999. Quantitative secretion and maximal secretion capacity of retinol, betacarotene and alpha-tocopherol into cows’ milk. J. Dairy Res. 66:511–522. Jung, M. Y., S. H. Yoon, H. O. Lee, and D. B. Min. 1998. Singlet oxygen and ascorbic acid effects on dimethyl disulfide and offflavor in skim milk exposed to light. J. Food Sci. 63:408–412. Levine, M., C. Conry-Cantilena, Y. Wang, R. W. Welch, P. W. Wasko, K. R. Dhariwal, J. B. Park, A. Lazarev, J. F. Graumlich, J. King, and L. R. Cantilena. 1996. Vitamin C pharmacokinetics in healthy
2307
volunteers: evidence for a recommended dietary allowance. Proc. Natl. Acad. Sci. USA 93:3704–3709. Macleod, D. D., X. Zhang, J. J. Kennelly, and L. Ozimek. 1999a. Ascorbyl-2-polyphosphate as a source of ascorbic acid for dairy cattle. Milchwissenschaft 54:123–126. Macleod, D. D., X. Zhang, J. J. Kennelly, and L. Ozimek. 1999b. Pharmokinetics of ascorbic acid and ascorbyl-2-polyphosphate in the rumen fluid of dairy cows. Milchwissenschaft 54:63–65. National Research Council. 1989. Nutrient requirements for dairy cattle. 6th rev. ed. Natl. Acad. Sci., Washington, DC. Roth, J. A., and M. L. Kaeberle. 1985. In vivo effect of ascorbic acid on neutrophil function in healthy and dexamethasone-treated cattle. Amer. J. Vet. Res. 46:2434–2436. SAS. 1999. SAS/STAT User’s Guide, Version 8. SAS Inst., Cary, NC. Timmons, J. S., W. P. Weiss, D. L. Palmquist, and W. J. Harper. 2001. Relationships among dietary roasted soybeans, milk components, and spontaneously oxidized flavor of milk. J. Dairy Sci. (accepted). Weiss, W. P., D. A. Todhunter, J. S. Hogan, and K. L. Smith. 1990. Effect of duration of supplementation of selenium and vitamin E on periparturient dairy cows. J. Dairy Sci. 73:3187–3194.
Journal of Dairy Science Vol. 84, No. 10, 2001