Impacts of feeding lipid supplements high in palmitic acid or stearic acid on performance of lactating dairy cows

ABSTRACT Effects of feeding lipid supplements high in free fatty acids (FAs) of palmitate (C16:0) or stearate (C18:0) plus C18:1 cis 9&10 on milk yield and composition, apparent whole-tract apparent digestibility of FA, and the FA composition of milk lipids were studied. Four lactating Holstein cows with ruminal cannulae were used in a 4 × 4 Latin square assignment of four dietary treatments. Lipid supplements were enriched in free fatty concentrations of either palmitic acid (P) or stearic acid (S). The total mixed-ration contained 20 g/kg of lipid supplement that consisted of varying proportions of P to S. Treatments were: 100:0 P:S (P), 0:100 P:S (S), and two mixtures including 66:34 P:S (PS) and 34:66 P:S (SP). Milk yield and dry matter intake were not affected by lipid supplement, but the concentration and yield of fat in milk increased with increasing C16:0 in the lipid supplement. Increasing the C16:0 concentration in the lipid supplement increased its concentration in milk lipids while increasing C18:0 in the lipid supplement increased C18:0 concentration in milk fat. Whole-tract apparent digestibility of neutral detergent fiber (NDF) did not differ with lipid supplement, but organic matter digestibility tended to decrease with increasing C18:0 in the lipid supplement. Whole-tract digestibility of total FAs decreased with the increasing proportion of C18:0 to C16:0 in the lipid supplement. Apparent digestibility of C16:0 and C18:0 was not different within dietary treatment. The FA composition of the lipid supplement impacted both whole-tract digestibility of FAs and FA composition of milk lipids.


Introduction
Feeding lipids with different fatty acid (FA) composition impacts the FA composition of the milk fat (Grummer 1991;Palmquist et al. 1993) as well as yield of milk and milk fat (Palmquist & Jenkins 1980;Sutton & Morant 1989). Milk higher in monounsaturated FA composition impacted the physical properties of the butter, cheese, yogurt, and ice cream (Chen et al. 2004). Extruded linseed altered the FA composition of milk, which impacted butter firmness, fat loss in buttermilk, and butter moisture (Hurtaud et al. 2010). FA composition of the diet impacted the size of the milk fat globules (Briard et al. 2003;Wiking et al. 2004;Carroll et al. 2006) with large milk fat globules more susceptible to coalescence. Higher monounsaturated FA in milk produced softer dairy products (Chen et al. 2004). In addition to the FA composition of TAG, diet also impacts the FA composition of phospholipids (PL), which is a rich source of C18:2 and C20:4 (McCaughey et al. 2005). Spitsberg (2005) discussed the bovine milk fat globule membrane as a potential nutraceutical because the PL influenced cell function including growth and development but also might play a role in disease development.
Supplemental lipids in the diet of lactating cows of different breeds increased the yield of fat-corrected milk but did not affect dry matter intake (Carroll et al. 2006). Even though the yellow grease contained approximately 26 g linoleic acid/100 g FA, the changes in the FA composition of milk fat were small with little change in the polyunsaturated FA (Carroll et al. 2006). There are limits to the ability to change the unsaturated FA proportion in milk fat because rumen biohydrogenation (Harfoot 1978) yields an FA profile at the duodenum that does not reflect the unsaturated FA content of the diet (Avila et al. 2000). Thus, for the lipid supplement S the flow of C18:0 to the duodenum will be higher than what is present in the supplement as a consequence of biohydrogenation. However, there are concerns with dietary lipids that the amount of unsaturated FA and the proportion of unsaturated FA in the unesterified form could reduce Dry Matter (DM) intake and have detrimental effects of rumen fermentation (Jenkins & Jenny 1989;Eastridge & Firkins 1991). This concern has contributed to the development of rumen inert lipids.
Rumen inert FA supplements of palmitic acid, for example calcium salts and free FA, or stearic acid, have influenced milk fat content as well as changed its FA profile (Grummer 1991;Palmquist et al. 1993;Harvatine & Allen 2006;Jenkins & McGuire 2006) and increased milk yield (Mosley et al. 2007;Warntjes et al. 2008) of lactating dairy cows. However, there are concerns with respect to digestibility of the medium-and long-chain FA by lactating cows related to carbon chain length and saturation (Doreau & Ferlay 1994;Pantoja et al. 1996b;Dohme et al. 2004;Glasser et al. 2008;Schmidely et al. 2008) that are associated with rumen inert fats. Even with more common feed fats, for example tallow and yellow grease, there was decreased postruminal digestibility of FA in particular saturated FA (Avila et al. 2000).
Lipids are typical ingredients in dairy diets. Lipids have an impact on the FA composition of milk fat and the subsequent dairy products made for the milk. Lipids increase energy intake and enhance milk yield, but digestibility of FA can be a concern. Thus, the aims were to determine the impacts of feeding lipid supplements high in palmitic (C16:0) or stearic (C18:0) acid on milk yield, whole-tract apparent digestibility of FA, and the FA composition of milk lipids.

Materials and methods
Four multiparous, lactating Holstein dairy cows previously fitted with ruminal cannulae averaging 100 days-in-milk were used in a study with a 4 × 4 Latin square arrangement of treatments. Periods were 21 d in length with the last 4 d of each period used for data collection. Treatments were the type of dietary lipid supplement added to a total mixed-ration (TMR) to alter the dietary C16:0 to C18:0 composition ( Table 1). The source of C16:0 was Energizer-RP10® (P; IFFCO, Johor, Malaysia) and the source of C18:0 was a high stearic acid product (S; MS Specialty, Dundee, IL, USA). Both lipids were free FAs (Table 2), and added alone or in combinations to provide 20 g lipid/kg dry matter in the TMR. The ratio of P:S in the lipid supplement was 100:0, 66:34, 34:66, and 0:100. Thus concentrations of the lipid supplement (g/kg) in the four dietary treatments were 20 P+0 S (P), 13 P+7 S (PS), 7 P+13 S (SP), and 0 P+20 S (S). All other dietary ingredients were the same for the four TMR. Thus diets were similar in chemical composition (Table 3), except for the FA composition (Table 4). Chromic oxide was added to each TMR during mixing at 1 g/kg (as fed basis) of the TMR, as an external indigestibility marker to determine whole-tract apparent digestibility based on the method reported by Crocker et al. (1998).
Holstein cows were housed in a common pen and trained to use Calan electronic feeding gates (American Calan Inc., Northwood, NH, USA) to allow individual feed intake measurement. The pen contained a water trough for ad libitum water consumption and sand-bedded freestalls. All cows received an injection of exogenous recombinant somatotropin (Posilac®, Monsanto Co., St. Louis, MO, USA) on a 14-d schedule. The study was conducted from November 2005 to February 2006, and was approved by the Institutional Animal Care and Use Committee of the University of California at Davis (CA, USA).
Cows were fed twice daily after milking at approximately 08:00 and 20:00 h. Feed refusals were collected and weighed every second day. A sub-sample of each cow's feed refusal was collected and composited by cow at the end of each  , sliced  430  430  430  430  Beet pulp  140  140  140  140  Corn grain, steam-flaked  100  100  100  100  Barley grain, steam-rolled  100  100  100  100  Cottonseed, whole linted  100  100  100  100  Molasses, cane  50  50  50  50  Almond hulls  40  40  40  40  Mineral b  20  20  20  20  Energizer-RP10  20  13  7  0  Energy booster-100  0  7 13 20 a Treatments were: P = 20 g/kg of enriched palmitic acid product; PS = 13 g/kg of enriched palmitic acid product and 7 g/kg of enriched stearic acid product; SP = 7 g/kg of enriched palmitic acid product and 13 g/kg of enriched stearic acid product; S = 20 g/kg of enriched stearic acid product. b Mineral-mix contained not less than 8.03 g Ca and not more than 10.23 g Ca/100 g supplement, not more than 12.05 g Na, not less than 4.0 g P per 100 g supplement, not less than 16.16 mg/kg Se and not more than 17.00 mg/kg Se. Nutrient composition of diet samples each period (n = 4 for each treatment). b Treatments were: P = 20 g/kg of enriched palmitic acid product; PS = 13 g/kg of enriched palmitic acid product and 7 g/kg of enriched stearic acid product; SP = 7 g/kg of enriched palmitic acid product and 13 g/kg of enriched stearic acid product; S = 20 g/kg of enriched stearic acid product. c L = linear effect of dietary treatment; Q = quadratic effect of dietary treatment. Milk yield was recorded at each milking (07:30 and 19:30 h) using Westfalia milk metres with samplers (Westfalia, Naperville, IL, USA). Milk samples from each cow were collected from each milking during the last 4 d of each experimental period. Milk samples were preserved in 2-bromo-2-nitro-propane-1, 3-diol and maintained at 5°C until samples were taken to the laboratory, warmed to 40°C in a water bath, composited as evening and morning milkings, and analysed for fat, protein, lactose, and solids-not-fat (SNF). A 10 ml aliquot was obtained and stored at −20°C for later analysis of milk lipid FA where the lipid was extracted from the entire sample.
Fecal grab samples were collected from the rectum at 09:00 h, after milking but before feeding, and at 19:00 h before milking, during the last 4 d of each experimental period. Fecal samples from each cow were composited daily within period by pooling equal weights of fresh feces, and stored frozen at −20°C. Each fecal composite was dried in a forced-air oven at 50°C for 48 h, coarsely ground using a hammer mill, and subsequently ground through a 1-mm screen using a Wiley mill (Arthur A. Thomas, Philadelphia, PA, USA).
As the cows had ruminal cannulae, 400 ml rumen fluid samples were collected on day 21 at 06:00 and 10:00 h of each period for determination of pH and concentrations of volatile fatty acids (VFAs) and ammonia-N. Collected ruminal fluid was filtered through four layers of cheesecloth. Following measurement of pH, 2 ml of 25% metaphosphoric acid was added to 10 ml of rumen fluid, centrifuged, and stored frozen at −20°C until analysis.
Coccygeal blood samples were collected at 06:00 (i.e. prior to feeding) and 10:00 h (i.e. 2 h after feeding) from all cows on the last day of each experimental period. Samples were collected into sterile evacuated tubes containing 0.117 ml of 150 g/kg EDTA (K3; Becton Dickinson Vacutainer Systems, Rutherford, NJ, USA) and placed immediately on ice until samples were centrifuged at 3000 xg for 15 min. Plasma was stored at −20°C for later analyses of triglyceride (TG) and phospholipid (PL) fractions.
Diets and feces were analysed in duplicate for DM (#925.40), N (#984.13), ether extract (EE; #920.39C), and ash (#923.03) according to the procedures of AOAC (1995), whereas NDF was determined as described by Van Soest et al. (1991) with the use of alpha amylase and expressed inclusive of residual ash. The Acid Detergent Fiber (ADF), cellulose, and lignin(sa) were determined as described by Robertson and Van Soest (1981) with ADF expressed inclusive of residual ash. Ca, Cr, Cu, Fe, K, Mg, Mn, Na, and Zn in feeds were determined (#968.08) using an atomic absorption spectrometer (AOAC 1995). The P in feed was determined according to a Technicon autoanalyzer method N-4C (Kraml 1966). Total FA content and composition of lipid supplements, diets, and feces were determined according to the procedures of Joy et al. (1997), as modified by Crocker et al. (1998). Ammonia-N and FA in rumen fluid were determined by the methods of Joy et al. (1997).
Milk was analysed for fat, total protein, lactose, and SNF concentration (AOAC 1995, #972.16) with an infrared analyzer (Foss, Eden Prairie, MN, USA), and for urea N by a Technicon Autoanalyzer (Technicon Corporation, Emeryville, CA, USA) method N-10a (Marsh et al. 1957) in duplicate. The infrared milk analyzer was standardized using milk samples obtained from the California Department of Food and Agriculture (Food and Agriculture Code Section 32921) where nitrogen was determined by Kjeldahl (protein = % nitrogen * 6.38), milk fat was determined by Roese-Gottlieb, and lactose determined by the polarimetric method.
A 10 ml aliquot of milk was obtained from the composite milk sample and milk fat was isolated and analysed for total FA of the TG by gas liquid chromatography (Erickson & Dunkley 1964;Crocker et al. 1998;DePeters et al. 2001). The FA composition of milk lipids was determined from their methyl esters. The method used was as follows. Using glass volumetric pipets, 10 ml of milk was pipetted into glass 50 ml (150 mm × 25 mm) screw-top culture tubes and 1 ml of saturated sodium chloride and 18 ml of 0.0625 M HCl in ethanol were added. Tubes were sealed with Teflon lined caps and shaken by hand for 1 min. This was followed by addition of 10 ml of hexane and shaking on a Burrell wrist action shaker for 10 min. To separate the phases, tubes were centrifuged for 5 min at 500 xg and, using a Pasteur pipette, the upper hexane layer was removed and washed with 20 ml of double de-ionized water. After shaking and re-centrifuging at the same g force, the hexane layer was transferred to a vial containing 1.2 g of anhydrous sodium sulphate and allowed to dry for 15 min. Aliquots containing 10 mg of fat, which was determined previously using a FOSS (Eden Prairie, MN, USA) infrared milk analyzer, were transferred to glass 25 ml (125 mm × 18 mm) screw-top culture tubes for methylation. The hexane was evaporated under a stream of N at 60°C then washed with 1 ml of 2,2,4 trimethylpentane (iso-octane), evaporated again to dryness under N at 60°C and finally reconstituted with 2 ml of iso-octane. Methyl esters were then formed by addition of 100 µl of 2 M potassium hydroxide in methanol, vortexing briefly and allowed to stand at 25°C for 15 min. Esters were Treatments were: P = 20 g/kg of enriched palmitic acid product; PS = 13 g/kg of enriched palmitic acid product and 7 g/kg of enriched stearic acid product; SP = 7 g/kg of enriched palmitic acid product and 13 g/kg of enriched stearic acid product; S = 20 g/kg of enriched stearic acid product. b L = linear effect of treatment; Q = quadratic effect of treatment.
then washed by addition of 2 ml of saturated NaCl and 8.5 ml of deionized water and shaken with a Burell wrist action shaker for 10 min. In this approach some loss of C4:0 recovery may occur. The upper iso-octane phase was isolated by centrifuging at 500 xg for 10 min and an aliquot was placed into a 1.8 ml gas chromatography auto sampler vial and crimp sealed with an aluminium Teflon lined cap.
The methyl esters were separated and quantified using a Hewlett Packard 5890 gas chromatograph equipped with a flame ionization detector and a Supelco 2560, 100 m capillary column containing a 0.25 mm inside diameter, and a 0.20 micron film thickness. Hydrogen was used as the carrier gas with a linear flow rate of 27 cm/sec and a column head pressure of 250 kPa. One microliter of the ester mixture was injected using a Hewlett Packard 7673 auto injector and subjected to a split vent flow rate of 100 ml/min. The injector temperature was set at 210°C and the detector temperature set at 220°C with the column temperature initially set at 70°C for 10 min followed by a programmed increase to 175°C at 20°C per minute for 29 min and finally to 225°C at 5°C per minute for 12 min.
Ester peak areas were acquired from the chromatograph through Hewlett Packard's HPIB/RS-232 board connected to a personal computer using Windows 98 as the operating system. Unknown peak areas were quantified using Chrom Perfect version 3.52 (Justice Innovations, Inc., Palo Alto, CA) and compared to a known quantity of an external standard mixture containing all reported FA. This FA standard mixture was comprised of individual FA obtained from Nu Chek Prep (Nu Chek Prep, Elysian, MN).
A 5 ml aliquot of milk was obtained from the milk collected on Tuesday and Wednesday morning of all experimental periods and analysed for TG and PL FA using the methods described by Bitman et al. (1984). The milk was aliquoted into a 50 ml (25 × 150 mm) glass screw-top culture tubes and heated rapidly to 80°C and held for 1.5 min to deactivate lipase enzymes. After cooling, milk was extracted with 36 ml of chloroform: methanol (2:1) using a Burrell wrist action shaker for 10 min. The lower chloroform layer containing milk fat was transferred to another glass culture tube, dried under N, and reconstituted with 2 ml of hexane. Using only gravitational force and a Pasteur pipette transfers the entire 2 ml of hexane onto a Water's Sep-Pak silica cartridge (WAT051900), which had been previously prepared by rinsing with 20 ml of chloroform. The sample tube was washed with an additional 1 ml of hexane and added to the cartridge. The TG was eluted by forcing 4 × 10 ml aliquots of hexane:ethyl ether (1:1) through the cartridge with a glass syringe, and saving it in 50 ml glass screw-top culture tubes. The PL was subsequently collected into a 50 ml glass culture tube by forcing 2 × 10 ml aliquots of methanol, followed by 2 × 10 ml aliquots of chloroform:methanol:water (3:5:2), through the cartridge with a glass syringe. Since the TG fraction was much more concentrated than the PL fraction, only a 2 ml aliquot of the TG fraction eluent was used for the analysis, while all of the solvents from the PL fraction elution were used. For each fraction, solvents were evaporated under N at 60°C, followed by a 1 ml wash with iso-octane, which was again evaporated under N. Finally, 2 ml of iso-octane was added to each tube and the milk fat fractions were methylated, isolated, and quantified using the procedure described above for milk fat methylation.
Milk fat globule volume and number were determined at the California polytechnic University Dairy Product Center, San Luis Obispo, according to the methods described by Astaire et al. (2003). Data were analysed by the PROC MIXED procedures of SAS (2001) using the model: where Y mno = dependent variable, µ = overall mean, C m = random effect of cow (m = 1-4), P n = fixed effect of period (n = 1-4), T o = fixed effect of treatment (o = 1-4), and e mno = residual error. Treatment effect was considered significant if P ≤ .05. Contrasts for linear and quadratic effects were also declared if P ≤ .05. Ruminal pH, ammonia-N, VFA, and plasma PL fractions used cow as a random effect, diet and time as fixed effects, and diet by time as a repeated interaction.

Results
Intake of DM, both total and as a proportion of body weight, was similar among treatments, and averaged 25.9 kg/d and 3.97 kg/100 kg body weight, respectively, both of which are appropriate for the stage of lactation of the cows ( Table 5). Yields of milk and milk components were not affected by lipid supplement, but milk fat yield tended (P = .08) to decrease linearly with decreasing C16 in the lipid supplement. Concentrations of fat (P < .02), lactose (P < .04), and SNF (P < .03) in milk decreased linearly as the proportion of C16 in the supplemental fat decreased, whereas the concentration of protein in milk was not impacted. The production data were based on four cows fitted with rumen cannulae so power of test was limited. However, DMI among treatments differed by only 0.5 kg DM/d, a nonsignificant difference, which in our view is a small difference. The same was true for milk yield that differed 0.4 kg/d (nonsignificant difference) among treatment means. In addition, we limited feed intake to 97% of the previous day ad libitum intake during the collection phase. Most studies allow for a 5-10% feed refusal. Treatments were: P = 20 g/kg of enriched palmitic acid product; PS = 13 g/kg of enriched palmitic acid product and 7 g/kg of enriched stearic acid product; SP = 7 g/kg of enriched palmitic acid product and 13 g/kg of enriched stearic acid product; S = 20 g/kg of enriched stearic acid product. b L = linear effect of treatment; Q = quadratic effect of treatment. c Fat-corrected milk at 40 g/100 g milk.
However, we used the indicator method to determine digestibility and consequently we believe that there can be no refusal (orts). In addition, there is no way to account for the concentration of marker in the orts using the indicator method. Studies using pen-fed animals or animals fed ad libitum likely have issues with sorting and as a result cannot use the indicator method to accurately determine apparent digestibility.
The FA composition of the TG fraction of milk fat was influenced by dietary treatment (Table 6) as expected. The proportion of C16:0 in milk fat increased linearly (P < .01) with the increase in C16 in the lipid supplement, from 30.6 g/100 g lipid for S to 39.8 g/100 g lipid for P. The concentration of C18:0 in TG increased in response to increasing C18 in the lipid supplement (10.6 g/100 g lipid for P versus 13.5 g/100 g lipid for S; P < .01). Concentration of most trans C18:1 monoenes increased (P = .01) as the concentration of C18 in the lipid supplement increased. Finally, the concentration of total unsaturated FA in the cis form (UFA) increased (P = .02) in TG with decreasing C16 in the lipid supplement.
The concentration of C18:0 in the PL fraction increased (P < .02) while C16:0 tended (P < .07) to decrease linearly as the concentrations of C18 increased and C16 decreased in the lipid supplement fed (Table 7). There were few other notable changes, with the exception that the concentration of many C18:1 trans monoenes and C17:0 increased as the C16 content of the lipid supplement decreased.
Yield of FA was influenced by dietary treatment (Table 8). Yield of C16:0 in milk fat increased linearly with the increase of C16 in the dietary lipid supplement from 461.5 g/d for S to 658.5 g/d for P (P < .01). Finally, yield of total FA tended (P < .08) to decrease linearly as the proportion of C16 decreased in the diet (i.e. 1676.5 g/ d for P to 1509.4 g/d for S). Yields of total FA agree with yield of milk determined by infrared analysis (Table 5).
in OM digestibility primarily reflects the linear decrease in total tract digestibility of total FA with increasing proportion of C18:0 in the diet. Rumen fluid pH, molar proportions of individual VFA, and the total VFA concentration were unaffected by lipid supplement (Table 10). Ammonia-N concentration was affected by dietary treatment with a minimum concentration for treatment PS (Q; P < .05) compared with the other dietary treatments. As anticipated, there were changes due to time of rumen sampling for pH, ammonia-N, and VFA concentration, with all being higher 2 h post-feeding versus prior to feeding (data not presented).
Concentration, yield, and proportional distribution of milk N (Table 11) did not differ for dietary treatment. The diameter of particles as measured by D(3,2) differed (Table 12). Milk fat globules decreased in size as the proportion of P in the lipid supplement decreased. Specific surface area (SSA) increased as the proportion of P in the lipid supplement decreased reflecting the increased number of smaller particles where P was 0.98 and S was 1.2. The mean volume for P indicates that the larger milk fat globules contained are greater amount of the total milk fat. There was a tendency (P < .07) for a quadratic effect since the volume estimate increased with S in agreement with the slight increase in D(3,2) and SSA.

Discussion
Production performance (Table 5) of lactating cows was not impacted by the type of lipid supplement fed. Diets were similar in composition with the exception of the lipid supplement and in theory diets were similar in nutrient and Treatments were: P = 20 g/kg of enriched palmitic acid product; PS = 13 g/kg of enriched palmitic acid product and 7 g/kg of enriched stearic acid product ; SP = 7 g/kg of enriched palmitic acid product and 13 g/kg of enriched stearic acid product; S = 20 g/kg of enriched stearic acid product. b L = linear effect of treatments; Q = quadratic effect of treatments. c PUFA = Poly unsaturated fatty acids: C18:2, C18:3, C20:4, C20:5, C22:5, and C22:6. d UFAcis = Unsaturated fatty acids cis; C18:1 cis 9&10, C18:1 cis 11, C18:1 cis 12, C18:1 cis 13, and all PUFA FAs. Table 9. Whole-tract apparent digestibility (g/kg) of FAs, organic matter (OM), NDF, and nitrogen (N) from lactating cows. Treatments were: P = 20 g/kg of enriched palmitic acid product; PS = 13 g/kg of enriched palmitic acid product and 7 g/kg of enriched stearic acid product ; SP = 7 g/kg of enriched palmitic acid product and 13 g/kg of enriched stearic acid product; S = 20 g/kg of enriched stearic acid product . b L = linear effect of treatment; Q = quadratic effect of treatment. energy concentration at formulation so differences in DM intake and milk yield were not anticipated using 21-d experimental periods. Proportion of C16:0 to C18:0 in the TMR ranged from approximately 16:1 for P to 1:1 for S, although lipid supplements constituted a small portion of the diet by their inclusion at 20 g lipid supplement/kg diet DM. Indeed, the dietary ratio of C16:0 to C18:0 in the lipid supplement fed to lactating dairy cows had no effect on production performance measured as milk yield, dry matter intake (Table 5), whole-tract apparent digestibility of NDF (Table 9), as well as rumen pH and VFA (Table 10). In addition, the gross milk composition was not changed by the FA composition of the lipid supplement and milk fat yield only tended to increase as the ratio of C16:0 to C18:0 in the lipid supplement increased. Type of lipid supplement did not affect milk protein content or synthesis (Table 11). Dietary lipid was reported to potentially impact milk N (DePeters & Cant 1992), but in the present study all dietary treatments were similar in FA content and there was no unsupplemented lipid diet for comparison.
There was no change in the yield (Table 7) of short-chain (sum of C4:0 to C9:0) FA (112.0 P, 118.6 PS, 197.7 SP, and 107.4 S) or medium-chain (sum of C10:0 to C15:0) FA (253.1 P, 294.1 PS, 252.6 SP, and 267.0 S), which suggest that de novo synthesis, if reduced by lipid supplement, was not affected differently by the FA composition of the lipid supplements fed. Banks et al. (1976a) reported that yields of short-and medium-chain FA were lower in milk lipids of cows fed lipid supplements compared with a low fat control. However, using their data, yields of C6:0 to C14:0 FA were numerically similar for the palm oil/palmitic acid (84 g/d) and tallow (93 g/d) treatments. Likewise, yields of C4 to C14:0 were numerically similar for palmitic (60 g/d) compared with stearic (77 g/d) acids (Steele & Moore 1968). Because a portion of C16:0 is from de novo and a portion is derived preformed from the blood, it is difficult to use FA yields to predict de novo synthesis when C16:0 is a component of the dietary treatment.
The major change in FA concentration of milk TG was the increased C16:0 and decreased C18:0 and C18:1 cis 9&10 as Treatments were: P = 20 g/kg of enriched palmitic acid product; PS = 13 g/kg of enriched palmitic acid product and 7 g/kg of enriched stearic acid product; SP = 7 g/kg of enriched palmitic acid product and 13 g/kg of enriched stearic acid product; S = 20 g/kg of enriched stearic acid product. b Time = effect of time of the day, L = linear effect of treatment; Q = quadratic effect of treatment. Treatments were: P = 20 g/kg of enriched palmitic acid product; PS = 13 g/kg of enriched palmitic acid product and 7 g/kg of enriched stearic acid product; SP = 7 g/kg of enriched palmitic acid product and 13 g/kg of enriched stearic acid product; S = 20 g/kg of enriched stearic acid product. b L = linear effect of treatments; Q = quadratic effect of treatments. c NPN = Non-protein nitrogen.  the level of C16:0 in the diet increased ( Table 6). The small but significant changes in many C18:1 trans monenes reflect the unsaturated trans FA in the S lipid supplement.
Whole-tract apparent digestibility of OM tended (P < .08) to decrease linearly as the ratio of C16:0 to C18:0 in the lipid supplement decreased (Table 9). This decrease was small with diet P averaging 696 g kg -1 and diet S averaging 678 g kg -1 OM apparent digestibility. The decrease in apparent OM digestibility was not associated with a decrease of either apparent NDF or N digestibility but with a decrease in total FA digestibility. In agreement, previously Avila et al. (2000) found no effect of saturation ratio of the lipid supplement on rumen and whole-tract apparent digestibility of ash-free NDF. Whole-tract apparent digestibility of both total C16 and total C18 decreased as the ratio of C16:0 to C18:0 in the lipid supplement decreased. However, the digestibility of total C16 and total C18 was similar within diet. Previous research using lactating cows with duodenal and ileal cannulae (Børsting et al. 1992) found that FA digestibility of C16:0 was higher than C18:0 when dietary fat sources were encapsulated by a casein-formaldehyde process. In a review of fat digestibility studies, Lock et al. (2005) found that whole-tract digestibility of C16:0 was ∼750 g kg -1 and of C18:0 was ∼720 g kg -1 . In contrast, Elliott et al. (1996) found no difference in whole-tract apparent digestibility with decreasing levels of C16 intake due to feeding a prilled FA low in C16, or a flaked (or prilled) hydrogenated palm FA distillate. However, measured FA digestibility in ruminants is difficult to interpret due to rumen biohydrogenation of unsaturated FA. Doreau and Ferlay (1994) summarized data from 15 experiments and 64 diets and reported mean digestibility of 771 g kg -1 for C16:0 and 763 g kg -1 for C18:0. Dohme et al. (2004) reported that whole-tract digestibility of total FA was impacted by the concentration of C12:0, C14:0, and C18:0 in the diet. Total digestibility decreased in order of C12:0, C14:0, and C18:0 so that increasing dietary concentration of C18:0 resulted in the greatest reduction of whole-tract total FA digestibility. Decreases in lipid digestion when lipid intake increased were associated with a decrease in the digestibility of saturated FA (Wu et al. 1991;Pantoja et al. 1996aPantoja et al. , 1996b. Recently, Schmidely et al. (2008) did an exhaustive meta-analysis of FA digestion. In their appendix 1 table that describes the database input variables, apparent intestinal digestibility in dairy cows was 746 and 734 g kg -1 for C16:0 and C18:0, respectively, while for the overall data set that included sheep and growing cattle apparent intestinal digestibility was 767 and 754 g kg -1 for C16:0 and C18:0, respectively. Our results suggest that whole-tract digestibility of total C18 and total C16 was similar.
In non-ruminants the pig digestibility of FA decreases with increasing chain length and digestibility of unsaturated FA is greater than saturated FA digestibility (Gu & Li 2003). In ruminants there are similar relationships, but the differences are more modest. These similarities between ruminants and nonruminants allow generalizations. The decrease in the overall FA digestibility observed in the present study as the proportion of C18 in the lipid supplement increased may be a consequence of FA solubility in the small intestine. The flow of C18:0 to the duodenum should have increased with increasing C18 supplement as a consequence of rumen biohydrogenation (Leat & Harrison 1969;Avila et al. 2000). In ruminants the emulsification of FA is slightly different than for non-ruminants since triacylgylcerol hydrolysis followed by biohydrogenation yields unesterified FA that adheres to particulate matter and flow out of the rumen (Leat & Harrison 1969) and the possible formation of water-insoluble C16:0 and C18:0 soaps (Lough 1970). For absorption these FAs must be removed from the particulate matter in the process involving emulsification and micelle formation. The solubility of long-chain FA in bile salt solution was low for pH conditions of 5 and below with FA solubility increasing as pH increased (Friedman & Nylund 1980) possibly related to higher glycine conjugates. There is considerable species difference for bile acids (Jacobsen & Smith 1968). Bile of humans (Venturoni et al. 2012) and pigs (Watanabe & Tsuneyama 2012) were higher in glycine conjugates than taurine conjugates (Venturoni et al. 2012) in contrast to bile of sheep (Smith & Lough 1976) and cattle (Watanabe & Tsuneyama 2012), which were higher in taurine conjugates (Smith & Lough 1976). Taurine-conjugated bile acids were more effective at acidic pH than glycine-conjugated bile acids. Taurine conjugates were less likely to precipitate from solution by polyvalent cations than glycine conjugates (Hofmann & Small 1967). The acidic condition of the ruminant duodenum and jejunum limits the formation of insoluble calcium soaps of saturated FA. In general, lower pH was reported for the duodenum of ruminants than monogastrics (Freeman 1984). Taurocholic has a pKa of 2.0 while glycocholic has a pKa of 3.9 (Setchell et al. 2013) so taurine-conjugated bile acids should be more soluble at the lower duodenal pH of ruminants than glycine conjugates although Freeman (1984) states that conjugation of either glycine or taurine had little effect on solubilizing properties of bile salts. However, Madenci and Egelhaaf (2010) reported that critical micellar concentration was slightly larger for glycine conjugates than taurine conjugates. Smith and Lough (1976) reported that fully protonated long-chain FAs were poorly solubilized in sheep bile. The solubility of FA in aqueous solutions containing bile salts obtained from sheep gall bladders over a pH range of 2.0-7.4 was linoleic acid > oleic acid> elaidic acid > palmitic acid > stearic acid (Smith & Lough 1976).
Absorption of long-chain saturated FA was also reported to be low in the small intestine because of their high melting point relative to body temperature (Cheng et al. 1949) and because of their ability to form low-solubility calcium salts (Ramírez et al. 2001). Post-weaned pigs were fed diets supplemented with either medium-chain FA or long-chain FA and either without or with an added emulsifier (Price et al. 2013). Total lipid digestion was higher for the medium-chain FA diet (98.5%) compared with the long-chain FA diet (93.4%). For the long-chain FA diet, emulsification had a marked impact on ileal FA digestibility except for C18:1. At day 14 postweaning, although not significant, the digestibilities of the total fat, C16:0, and C18:0 with emulsification were 95.1%, 94.4%, 82.1%, respectively, compared with 92.1%, 92,1%, and 69.2% without emulsification. The pig was reported to have glycine conjugates (Freeman 1984). In broiler chickens, the digestibility of C16:0 and C18:0 was higher for the diet containing soybean oil compared with the diet containing tallow (Tancharoenrat et al. 2014). Ileal digestibility of oleic and linoleic acids (unsaturated FA) was higher than the digestibility of palmitic and stearic acids (saturated FA). These differences were attributed to the nonpolar nature of C16:0 and C18:0 that required emulsification prior to absorption, but that long-chain saturated FAs were also not readily incorporated into micelles. Chicks were reported to have taurine conjugates (Freeman 1984). In addition, the role of pH and melting temperature is also likely involved as well as other factors. For example, the conjugated-form cattle bile acids contain cholic acid and deoxycholic acid whereas pig bile acids contain ursodeoxycholic acid and hyodeoxycholic acid (Watanabe & Tsuneyama 2012). Cholic and deoxycholic acids are high hydrophobicity, which might explain why long-chain saturated FA digestibility can be higher in ruminants than nonruminants. Together, all of the previously discussed effects may have impacted the digestibility of long-chain FA in the present study. It is possible that the increasing amount of C18:0 reaching the duodenum as the proportion of C18 supplement increased impacted the solubility of FA into micelles with the result that total FA digestibility decreased.
Since the digestibility of both C16 and C18 was decreased with increasing C18 supplement, both would be provided in lower amounts for uptake by the mammary gland. Yield of milk fat decreased with increasing proportion of the C18 supplement. Yield of C16:0 decreased in a linear fashion (P < .01) with increasing C18 supplement. Palmitic acid in milk fat can be from de novo synthesis or preformed from the blood (Palmquist and Jenkins 1980). Long-chain FAs inhibit mammary de novo synthesis. Using the apparent FA digestibility estimates (Table 9), the calculated intake of total digestible C16 FA was 0.46 kg/cow/d for P compared with 0.23 kg/cow/d for S, a 50% decrease. Intake of all C18 was 0.56 kg/cow/d for P and only slightly increased with S, 0.61 kg/cow/d. Yield of all C16 FA in milk was 0.658 kg/cow/d for P compared with 0.461 kg/ cow/d for S, a 30% decrease. All C18 FA yield was 0.518 kg/ cow/d for P and 0.535 kg/cow/d for S, a 3% increase. Since the yields of C10:0, C12:0, and C14:0 did not decrease from P to S, it is likely that the decline in C16:0 in milk fat was a result of the lower digestibility of C16 with increasing C18 supplement.
Bile flow and the implications of bile to lipid ratio at the distal duodenum of steers fed increasing amounts of dietary fat were discussed (Plascencia et al. 2004). Bile secretion increased with increasing fat intake, and the bile to lipid ratio accounted for 69% of the variation observed for intestinal FA digestion. However, in the current study lipid intake was similar so differences in bile to lipid ratio at the small intestine should not significantly differ.
In conclusion, feeding a lipid supplement high in either C16 or C18 supported production performance of high-producing cows although digestibility of total FA declined with increasing proportion of C18 in the lipid supplement so consideration should be given to the type of lipid supplement fed. Type of lipid supplement impacted the FA composition of milk lipids as well as milk fat globule size distribution.