Differences In Milk Characteristics
Differences In Milk Characteristics Between A Cow Herd Transitioning To Organic Versus Milk From A Conventional Dairy Herd
1 Dairy & Functional Foods Research Unit, Eastern Regional Research Center, Agricultural Research Service, US Department of Agriculture, 600 E. Mermaid Lane, Wyndmoor, PA 19038, USA, 2 Soil Foodweb, Inc., 1750 SW 3rd St., #C, Corvallis, OR 97333, USA, and 3 Rodale Institute, 611 Sigfriedale Road, Kutztown, PA 19530, USA
Characteristics of conventional milk and milk from a herd transitioning from nongrazing to organic were studied by comparing adjacent farms over a 12-month period. Levels of short- and mediumchain fatty acids partially responsible for aroma and flavour were initially lower in the milk from the transitioning herd, but not after the cows had settled into an organic diet. Once that point was reached, the amount of a-linolenic acid in the transitioning herd milk exceeded that of the conventional herd. This case study demonstrates that subtle differences occur in the milk as cows transition to organic. Keywords Milk, Organic, Grass-fed, Fatty acids.
The market for organic and grass-fed dairy products continues to grow as consumers are willing to pay premium prices for food that they consider to be more healthful, flavourful or environmentally conscious than the majority of food purchases. Many Americans are willing to pay more for milk that does not contain antibiotics, pesticides, or hormones that are not naturally present (Bernard and Bernard 2009). The question is whether organic milk actually contains a higher level of beneficial compounds than conventional milk from cows with no access to pasture. Benbrook et al. (2013) examined organic and conventional milk from 14 commercial processors across the USA over an 18-month period, finding higher omega-3 fatty acid concentrations in the organic milk. This work did not deal with milk at the farm level, although it indicates that a transition to organic management could be beneficial to the consumers who purchase the resulting milk. Within the past 10 years, most of the research on the differences between milk from organic and conventional dairies has been conducted in Argentina (Schroeder et al. 2003, 2005), Italy (Bergamo et al. 2003; La Terra et al. 2010), United Kingdom (Ellis et al. 2007), Switzerland (Collomb et al. 2008) and Sweden (Fall and Emanuelson 2011). The only such studies comparing organic and conventional bovine milk in the USA published since 2003 were by Croissant et al. (2007) and Khanal et al. (2008). Croissant et al. investigated two herds 80 km apart, with each herd containing Holsteins and Jerseys. The geography and weather may have affected the characteristics of the milk in that study. Khanal et al. looked at only five cows for 45 days. Current regulations require a 12-month transitioning period for a conventional dairy to be certified as organic (US Department of Agriculture, 2015). Information on the changes that occur in the harvested milk during this period, especially the rate at which any changes may occur, is lacking. A more thorough comparison of organic and conventional milk, including sizable herds and excluding the variables of terrain and climate, is necessary to determine the differences, if any, between these two management systems. This study investigates the characteristics of the milk over a 12-month period from a farm with cows transitioning from nongrazing to organic pasture.
MATERIALS AND METHODS:
Milk Two farms located <1 km apart in Berks County, Pennsylvania, designated CONV (conventional) and TTO (transitioning to organic), supplied the milk for the study. The CONV farm had 64–74 cows, with a Holstein-to-Jersey ratio of 9:1 and a rolling herd average ranging from 9580 to 9933 kg milk over the course of the study. Cows received total mixed rations consisting of high-moisture maize silage, alfalfa haylage, ryelage (all produced on-farm) and a supplement of soya beans, wet brewer’s grain and minerals (Table 1). Cows did not have access to pasture. The TTO farm had 51–63 cows with a ratio of Holsteins to Jerseys and Jersey crosses of 3:1. The rolling herd average began at 10 230 kg milk and decreased during the transition phase to 8164 kg milk at month 7 and 6767 kg milk at month 12. Cows received dairy mineral supplements and averaged 53% of their dietary energy from fresh pasture during the grazing season, which ran from mid-April to mid-October. The pastures were well-established fields certified as organic and consisting of alfalfa, orchard grass, perennial rye, red clover, timothy and white clover. Weeds were also present and consumed by the cows, including dandelion, lamb’s-quarters, narrow leaf plantain and smooth pigweed. Somatic cell counts in the milk for both herds were <500 000 throughout the study. Fresh raw milk (1.2 L) was obtained from the farms’ bulk tanks in the morning after 5 min agitation, poured into 3.7- L resealable plastic storage bags (Ziploc, S.C. Johnson & Son, Racine, WI, USA) and immediately frozen. Milk was collected weekly over the course of 50 weeks from 6 May 2011 through 12 April 2012. No samples were collected at week 30, 34 and 35 (Thanksgiving, Christmas and New Year holidays). Milk was transported to the US Department of Agriculture laboratories, placed in a secondary vacuum bag (in case the original bag opened or ruptured), thawed, redistributed in smaller aliquots and refrozen for individual assays. Portions of the thawed milk were centrifuged to obtain lipid fractions for fatty acid assays.
Feed used in this study Conventional Organica Composition of ingredient Concentration in feed (g/100 g) Composition of ingredient Concentration in feed (g/100 g) Haylage 31.2–43 g/100 g dry matter, 3.3–7 g/100 g protein, 1 g/100 g fat 5.9–9.7 38 g/100 g dry matter, 7.4 g/100 g protein, 1.7 g/100 g fat, 3.8 g/100 g ash 37.5 Maize silage 34–38 g/100 g dry matter, 2.3–2.9 g/100 g protein, 1–1.3 g/100 g fat 21.1–31.4 35 g/100 g dry matter, 3.3 g/100 g protein, 1.5 g/100 g fat, 1.1 g/100 g ash 35.3 Ryelage and hay Ryelage: 31 g/100 g dry matter, 4 g/100 g protein, 1 g/100 g fat 19.2–29.7 Hay 8.9 Maize High moisture: 70 g/100 g dry matter, 6 g/100 g protein, 2.7 g/100 g fat 8.7–9.0 Dry and ground 7.7 Grain Wet brewer’s 23.6–24.4 Spelt 2.8 Soya beans Roasted and ground 3.1–3.7 Roasted whole 5.0 Mineral supplement Custom mixb 6.5 Custom mixc 0.44 Other ingredients CaCO3 0–0.21 Kelp meal 0.07 a Comprised 47% of diet. Remaining 53% was pasture; dried pasture samples contained 90.3–91.5 g/100 g dry matter, 14.2–17.3 g/100 g crude protein, 2.3–2.9 g/100 g crude fat and 8.1–9.6 g/100 g ash. Total diet averaged 50.1 g/100 g dry matter, 7.4 g/100 g crude protein, 2.5 g/100 g crude fat and 3.1 g/100 g ash. b Consisted of ML 100512 (Cargill Animal Nutrition, Minneapolis, MN): 31.4 g/100 g crude protein, 5.23 g/100 g crude fat, 4.5 g/100 g Ca, 0.4 g/100 g P, 2.7 g/100 g salt, 1 g/100 g Na, 2 g/100 g Se, trace amounts of MgO, Cu, Mn, Co, Zn, I, and vitamins A, D and E. c Consisted of 0.28 g/100 g salt, 0.07 g/100 g MgSO4, and 0.07 g/100 g RC Gold (yeast, lactic acid bacteria, and vitamins A, D and E; Fertrell, Bainbridge, PA)
Compositional data (solids, total fat, total protein and lactose) were obtained in accordance with AOAC Method 972.16 (AOAC International 2012) by MilkoScan Minor (FOSS, Eden Prairie, MN, USA). Ash was determined by AOAC Method 945.46 (AOAC International 2012) and dissolved in nitric acid (2 g/100 mL) to determine the mineral contents (Ca, Cu, Fe, K, Mg, Mn, Na and Zn) using an ICP-OES spectrometer (iCAP 6300 Duo, ThermoFischer Scientific, Madison, WI, USA). The pH was determined by PHM82 pH meter (Radiometer, Copenhagen, Denmark).
Fatty acid profiles:
The fat in each sample was obtained by centrifugation of the milk at 50009 g for 30 min at 10 °C. The fatty acids were converted to methyl esters using a procedure based on Christie (2003). Each lipid sample, weighing 100–125 mg, was dissolved in 2.5 mL hexane (Fisher Scientific, Fair Lawn, NJ, USA) prior to the addition of 100 lL sodium methoxide (25 g/100 g methanol); Sigma-Aldrich, St. Louis, MO, USA). After 5 min of shaking by inversion, 5 lL glacial acetic acid (J.T. Baker, Phillipsburg, NJ, USA) was added to lower the pH, and 1.0 g anhydrous CaCl2 (Mallinckrodt Specialty Chemicals, Paris, KY, USA) was added to capture any water. After an hour, the liquid was centrifuged at 700 g for 2–3 min. The supernate was pipetted into a 2-mL vial with a screw cap containing a Teflon-faced silicone septum (Supelco, Bellefonte, PA, USA), the hexane was evaporated under nitrogen, and 1.0 mL ethyl acetate (Burdick & Jackson, Muskegon, MI, USA) was added. An autosampler injected 1.0 lL into a HP 6980 gas chromatograph equipped with flame ionisation detection (HewlettPackard, Santa Clara, CA, USA) and an SP-2380 fused silica capillary column (60 m 9 0.25 mm; Supelco). The resulting chromatographic peaks were integrated with the instrument’s software to produce percentages of fatty acids in the fat. Multiplication by the concentration of fat in the milk yielded concentrations of fatty acids in the entire sample. The chromatographic reference standards were C4:0- C24:0 methyl esters and conjugated methyl linoleate (GLC448 and UC-59M, respectively; Nu-Chek-Prep, Elysian, MN, USA). Samples were analysed in duplicate and averaged over each month.
Water-soluble milk protein extracts were obtained using a protocol based on a procedure by Quiros et al. (2007). A 30-mL milk sample was centrifuged at 20 0009 g for 30 min at 4 °C, and the resulting supernate was subsequently filtered through Whatman no. 1 filter paper (GE Healthcare, Piscataway, NJ, USA) and lyophilised. Protein mixtures were dissolved in water containing 0.1 g/100 g TCA (Acros Organic, Fair Lawn, NJ, USA) to generate 10 mg/mL solutions. Injections of 100 lL were analysed by an Agilent 1200 series reverse-phase HPLC (Agilent Technologies, Santa Clara, CA, USA) using a Vydac C18 column (5 lm, 4.6 mm i.d. 9250 mm; Grace, Deerfield, IL, USA) and monitoring absorbance at 215 and 280 nm. The stationary phase was 0.1 g/100 g TCA in water. The liquid phase was 0.1 g/100 g TFA in acetonitrile, shifting from 0 to 20 g/100 g over 30 min, followed by 20 to 35 g/100 g over 5 min and finally 35 to 80 g/100 g over 45 min.
Milk volatiles were extracted using a SPME static headspace method. Samples for the SPME study were held at 80 °C until analysis. To avoid contamination by volatiles in the laboratory, samples were defrosted overnight at 24 °C in a dedicated refrigerator. For each run, 4 g NaCl (SigmaAldrich) was added to 10 mL milk to produce a saturated NaCl solution to enhance volatile collection. An internal standard of 0.10 mg/L 2-methyl-3-heptanone (SigmaAldrich) was added for quantitation purposes. Amber glass vials (20 mL) with polytetrafluoroethylene/silicone septum caps (Supelco) were used to minimise light exposure. Samples were vortexed for 10 s and then placed onto a CombiPAL autosampler (CTC Analytics, Swingen, Switzerland). Volatiles were adsorbed using a 2-cm, 50/30-lm-film-thick DVB/CAR/PDMS Stableflex SPME fibre (Supelco), while the sample was exposed to continuous agitation at 500 rpm for 5 min at 60 °C. Analytes were desorbed using a splitless injector at 250 °C for 5 min. An Agilent model 7890A gas chromatograph (Agilent) coupled with an Agilent model 5975C mass spectrometer was used to analyse the sample headspace components. The GC oven temperature was held at 33 °C for 4 min, then increased by 15 °C/min to 250 °C and held for 2 min. Separation was accomplished with a 0.6 mL/min flow of helium through a 30-m, 0.25-mm i.d., 0.25-lm-film-thick DB-5 column (Restek Corp., Bellefonte, PA, USA). Volatile compounds were identified by the NIST internal library in the ChemStation software (NIST/EPA/NIH Mass Spectra Library, version 2.0 days, December 2005) and by comparison with runs of known standards (Sigma-Aldrich) using the identical methodology. Relative abundance was calculated using the integrated ChemStation software by comparing peak areas with that of the internal standard. Samples were analysed in duplicate and averaged over each month.
Colour was measured using a HunterLab Color Quest XE spectrophotometer (Hunter Associates laboratory, Reston, VA, USA) fitted with a reflectance port of 2.54 cm in diameter according to the HunterLab Application for measuring translucent liquids. Frozen milk was thawed, warmed to 25 °C for 15 min and gently mixed. Approximately 100 mL of milk was poured into a 50-mm glass cell, placed at the port, covered for the measurement and then discarded. Three measurements of the CIE L*, a*, b*, WI and YI values were taken from each sample and averaged.
Statistics: Analysis of variance was performed using the GLM and mixed models in the Statistical Analysis System (SAS Institute Inc 2011) with farm and month as the dependent variables. Differences between means determined using the Bonferroni test are described as significant when P < 0.05.
RESULTS AND DISCUSSION
Composition Table 2 shows the compositional data for the milk over the 12 months of the study. No significant differences between the CONV and TTO milks were observed in any month. None of the eight minerals commonly found in milk that were measured in this study were significantly different between farms or over time. Values were similar to reported values (US Department of Agriculture 2013).
Overall composition of milk from conventional and transitioning-to-organic milking herds Conventional Transitioning RMSEa Proximate composition (g/100 g) Fat 3.63 3.63 0.270 Protein 3.29 3.26 0.118 Lactose 4.78 4.80 0.098 Total solids 12.5 12.4 0.348 Mineral composition (lg/g) Ca 1400 1400 125 K 1400 1400 254 Na 680 640 33 Mg 150 150 12 Zn 4.5 3.8 0.83 Cu 0.63 0.61 0.32 Fe 0.54 0.55 0.25 Mn 0.18 0.29 0.01
Monthly averages of fatty acids in the fat portion of raw milk from conventional (CONV) and transitioning-to-organic (TTO) milking herds, in months where differences are significant (P < 0.02) Month Farm 6:0a 8:0 10:0 12:0 14:0 14:1 15:0 16:0 16:1 17:0 18:0 18:1 t-18:1 18:2 18:3 CLA (mg/g fat) May CONV 17.2 11.7 26.9 30.7 111.9 11.7 11.6 14.4 204.8 TTO 8.4 6.7 13.8 16.8 77.4 6.1 8.1 26.5 307.2 June CONV 10.7 24.0 27.5 104.8 17.1 144.7 239.1 TTO 8.0 17.0 19.8 83.4 26.8 121.4 293.6 July CONV 18.1 142.1 5.6 TTO 22.5 120.1 8.2 August CONV 24.9 28.8 103.9 29.3 5.4 8.3 TTO 19.4 22.6 90.5 47.1 7.6 12.4 September CONV 6.8 33.3 TTO 13.7 25.1 October CONV 306.2 142.4 4.8 TTO 349.9 115.7 8.3 November CONV TTO December CONV 4.8 TTO 6.8 January CONV 298.4 33.7 33.9 5.0 TTO 346.1 21.6 25.5 7.0 February CONV 30.5 5.0 TTO 21.2 6.6 March CONV 13.2 33.0 4.5 TTO 11.3 22.4 6.9 April CONV 4.1 TTO 6.9 a Abbreviations for fatty acids: 6:0, caproic (hexanoic) acid; 8:0, caprylic (octanoic) acid; 10:0, capric (decanoic) acid; 12:0, lauric (dodecanoic) acid; 14:0, myristic (tetradecanoic) acid; 15:0, pentadecanoic acid; 16:0, palmitic (hexadecanoic) acid; 16:1 palmitoleic (hexadec-9-enoic) acid; 17:0, margaric (heptadecanoic) acid; 18:0, stearic (octadecanoic) acid; 18:1, oleic ([9Z]-octadec-9-enoic) acid; t-18:1, vaccenic ([E]-11-octadecenoic) acid; 18:2, linoleic ([9Z,12Z]-9,12-octadecadienoic) acid; 18:3, a-linolenic ([9Z,12Z,15Z]-9,12,15-octadecatrienoic) acid; CLA, conjugated linoleic acid.
Months in which the fatty acid profiles were significantly (P < 0.02) different between farms are shown in Table 3. The levels of caproic acid (hexanoic acid, abbreviated 6:0), caprylic acid (octanoic acid, 8:0), capric acid (decanoic acid, 10:0), lauric acid (dodecanoic acid, 12:0) and myristic acid (tetradecanoic acid, 14:0) were significantly lower in the TTO milk in month 1, 2 and 4 (May, June and August) of the study. A significant shift in fatty acids occurred in the rumens of the TTO farm cows as they transitioned to a grass diet. At the end of the summer, when the transitioning cows switched to stored grasses, differences began to centre on longer-chain fatty acids (16 carbon atoms and higher). Longer-chain fatty acids have high perception thresholds (elevated concentrations are required for humans to be able to detect their aroma or flavour) and play a small role in flavour, but smaller fatty acids have low thresholds and provide characteristic flavours to dairy products. All of the milk samples contained polyunsaturated fatty acids (PUFA) tied to positive health effects: rumenic acid (cis-9 trans-11 18:2), vaccenic acid (trans-11 18:1) and alinolenic acid (18:3). Rumenic acid is the predominant isomer of conjugated linoleic acid (CLA), which appears to act against atherosclerosis, cancer, diabetes and deposits of adipose tissue (Belury 2002). Vaccenic acid is the only known dietary precursor of CLA, and research suggests that its consumption may impart health benefits beyond those associated with CLA (Field et al. 2009). a-Linolenic acid appears to decrease the risk of cardiovascular disease (Hu et al. 1999; Burdge and Calder 2006). After a 2-month lag at the start of the transition, the TTO milk fat contained significantly more linolenic acid than the CONV milk fat in nearly every month. The rumenic and vaccenic acid contents of the CONV and TTO milks were similar, although the rumenic acid content in the TTO milk was significantly higher than that of the CONV milk in August. This disparity was evidently due to the dietary differences between the herds. At month 12, when the transitioning cows had settled into an organic diet, only the a-linolenic acid content was different between the two herds. Benbrook et al. (2013) also found higher levels of a-linolenic acid in organic milk. The TTO milk contained higher amounts of PUFA than the CONV milk. The recommended dietary requirement for PUFA is 6–11% of total energy (WHO 2008), so milk should be considered as adding to the total PUFA in the diet while not providing all the PUFA recommended.
The elution pattern of protein generally showed very little in terms of hydrophilic peptides, meaning that slight or no degradation of the proteins occurred. This result was expected as the samples were raw milk and therefore did not undergo any processing that might alter the structure or chemical integrity of the proteins. Also as expected, the major proteins were large intact hydrophobic proteins, consisting of the caseins and whey proteins. These proteins were particularly monitored in each sample to determine whether any changes were detected from sample to sample. Representative chromatographs from different months throughout the study show the overall distribution of proteins (Figure 1). The large peaks were composed of a number of different proteins (as1-, as2-, b- and j-caseins, and whey proteins) that coelute. While the relative heights of the peaks may have changed, showing that the overall concentrations of these proteins may be different with respect to each other, the composition of the proteins remained the same throughout the time period studied. The farming system employed in this study did not affect the overall protein composition of the milk obtained.
A total of 11 volatile compounds were identified in the raw whole milk. Acetone, 2-butanone, hexanal and octanoic acid (Figure 2) were the predominant volatile compounds in the milk, and other aldehydes, fatty acids and ketones were also present (Figures 3–4). All of these volatiles have been previously found in milk (Marsili 1999). The levels of some of the volatiles varied greatly throughout the year (Figures 2– 4) and were probably due to the composition of the feed or the pasture plants. The levels of most of the compounds in the TTO milks were in fairly small ranges throughout the year. In the CONV milk samples, hexanal, octanoic acid (Figure 2), pentanol (Figure 3), 3-methylbutanal and heptanone (Figure 4) reached their highest levels in the colder months. The TTO milk usually contained less hexanal, pentanol, butanoic acid and heptanone than the CONV milk, whereas nonanal was the only volatile in the TTO milk that was usually at a higher level than the CONV milk. The milks would therefore be expected to have a different aromas and flavours.
Milk colour was fairly stable between farms and over time with only a few differences noted. The whiteness values reported as L, which ranged from 64.3 to 72.1, were not significantly different, while the WI values ranged between 9.46 and 33.7 with only three means different from each other (means for TTO in March and April, were greater than TTO in May). No differences were noted for the a* values, which ranged from –1.90 to –2.29; negative a* values indicate the degree of greenness. Overall, the milk from TTO had significantly higher yellowness values than CONV milk, with the milk from TTO in July–November > CONV in March–April. The b* values ranged from 3.36 to 4.54 for TTO and 1.6 to 3.51 for CONV. The YI values ranged from 6.18 to 9.02 for TTO and 1.98 to 6.90 for CONV. Other studies have reported that organic milk is typically more yellow than milk from conventional herds, which could be attributed to the level of pasture in the diet, the breed of cows and the fat content of the milk (Hulshof et al. 2006; Croissant et al. 2007). b-Carotene imparts a yellow colour to milk fat (Panfili et al. 1994), is usually elevated in cattle on fresh forage and in cattle breeds such as Jerseys and is presumed to be the reason the TTO milk was yellower. Although both farms in this study had Jersey cows in the milking herd, TTO had one-fourth Jerseys in the herd compared to only one-tenth at CONV.
The fatty acid profiles of the CONV and TTO milk differed at the start of the transition period because of changes in the rumens of the transitioning animals. The profiles were similar after 6 months, except the TTO milk usually contained significantly more a-linolenic acid. The TTO milk was yellower than the CONV milk through most of the study. The levels of volatile compounds were variable throughout the 12 months period, but the composition and protein profiles were not different. A transition to organic pasture management appears to result in changes in fatty acid composition and volatile compounds in milk. A better understanding of factors affecting differences in milk from organic and conventional farms, including soil type, pasture quality and weather, is needed. The impact of seasonality on TTO and CONV milk also needs to be investigated.
The authors thank the following Agricultural Research Service employees for their contributions: Kerby Jones for his assistance in the fatty acid procedures, Susan Iandola for performing the GC-MS analyses, James Shieh for performing the compositional analyses and John Phillips for the statistical analyses. The authors also thank Rita Seidel of Rodale Institute for her assistance in obtaining the milk samples. Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the US Department of Agriculture. USDA is an equal opportunity provider and employer.
URL http://www.who.int/nutrition/topics/FFA_summary_rec_conclusion.pdf. Accessed 12/4/2014.
Modifying soil to enhance biological control of below ground dwelling insects in citrus groves under organic agriculture in Florida.