Fat composition of milk replacer influences growth performance, feeding behavior, and plasma fatty acid profile in ad libitum-fed calves.

Fat composition in milk replacers (MR) for calves differs from bovine milk fat in multiple ways. The aim of the study was to investigate the impact of different approaches of formulating fat in MR on growth, ad libitum intakes of MR and solid feeds, as well as blood metabolites in dairy calves. Upon 24 to 96 h after birth, 63 calves were acquired from dairy farms and incorporated into the study. Calves were blocked based on arrival day and randomly assigned within each block to one of 3 treatments differing in MR fat composition (n = 21 per group): VG was based on vegetable fats including 80% rapeseed and 20% coconut fat; AN was formulated with animal fats including 65% lard and 35% dairy cream; and MX with a mixture of 80% lard and 20% coconut fats. All 3 MR contained 30% fat, 24% crude protein, and 36% lactose and were formulated to have a fatty acid profile resembling that of milk fat. From arrival onwards (3.1 ± 0.84 d of age; LSM ± SD), calves were group-housed and were offered an ad libitum supply of MR at 135 g/L (13.5% solids). Weaning was gradual and induced between wk 7 to 10, after which calves were only fed solid feeds. Starter feed, chopped straw, and water were offered ad libitum throughout the study. Calves were weighed, and blood was collected weekly until d 84 after arrival. Prewean-ing ADG was greater in calves fed AN (915 g/d) than other treatments (783 g/d), whereas no differences were detected in the weaning and postweaning phases. Preweaning MR intakes were greater in calves fed AN than MX from wk 2 to 6 and were also higher in calves fed AN than VG in wk 5 and 6. Consistently, the number of rewarded visits during the ad libitum phase was greater in calves fed AN than MX, whereas VG showed no differences. This led to a higher preweaning ME intake in calves fed AN than in VG and MX. Serum cholesterol


INTRODUCTION
Recent research in calf nutrition recommends increasing milk supply to levels closer to what calves would voluntarily consume, and to increase the fat content (≥23%, DM basis) in milk replacers (MR) to better align with bovine whole milk composition (Yohe et al., 2020;Wilms et al., 2022a).Higher dietary fat intakes from liquid feed reduced hunger-related behavior during the weaning transition in ad libitum fed calves (Echeverry-Munera et al., 2020), improved fecal scores (Amado et al., 2019), reduced the number of therapeutic interventions (Berends et al., 2020), and reduced calf mortality (Urie et al., 2018).However, when increasing the fat inclusion in MR, it is crucial to consider fat composition as MR have a different fatty acid (FA) profile and triglyceride (TG) structure than milk fat (Esselburn et al., 2013;Hageman et al., 2019;Mellors et al., 2023).
The differences in FA profile and TG structure between milk fat and the fat fraction of MR are related to the type of fat used.Fat structure of MR is also Fat composition of milk replacer influences growth performance, feeding behavior, and plasma fatty acid profile in ad libitum-fed calves.
affected by the production methods which affect fat globule size and structure, as well as the emulsification in solution.In Europe and other parts of the world, vegetable fats such as palm, coconut, and rapeseed are the primary fat sources used (Yohe et al., 2021;Welboren et al., 2021a,b), whereas, in North America, fat blends containing lard, tallow, and coconut are more common (Hill et al., 2007;Esselburn et al., 2013).The FA profile has proven to be important for neonatal health and development, as some FA influence gastrointestinal and immune functions (Guilloteau et al., 2009;Górka et al., 2018;Welboren et al., 2023).Furthermore, the balance between omega-6 (n-6) and omega-3 (n-3) polyunsaturated FA (PUFA) seems to influence metabolic inflammation and synthesis of long-chain PUFA necessary for cellular processes (Calder, 2008;Schmitz and Ecker, 2008).Essential FA and specialized pro-resolving mediators formed in the cells by the metabolism of PUFA are critical for growth, organogenesis, inflammation regulation, and neurodevelopment (Cheruku et al., 2002;Helland et al., 2003;Ramiro-Cortijo et al., 2020).In addition to the FA profile, the positioning of longchain saturated FA (SFA) on the glycerol backbone is an important factor in fat absorption and digestion in infants (Carnielli et al., 1996;Lucas et al., 1997;Karupaiah and Sundram, 2007).The physiochemical differences between milk fat and the fat in MR are also related to the structural organization of fat globules.In whole milk, milk fat globules are surrounded by a surface layer composed of proteins and polar lipids that envelop each individual fat droplet (Lindmark Månsson, 2008).
A previous study on MR fat sources showed that the inclusion of 17% lard in MR compared with 18% vegetable fats (75% palm; 20% coconut; 5% rapeseed) did not affect ADG of calves fed 2.0 L of MR 3 times daily (Huuskonen et al., 2005).In contrast, Hill et al. (2007) found an increase in ADG in calves fed twice daily 2.0 L of MR containing 20% fat with a combination of animal and vegetable fats (80% lard, 15% coconut, and 5% canola) compared with calves fed an alternative MR formulation with 20% lard.These results could be explained by the low medium-chain FA content of lard, such as caproic and lauric acid, which are metabolized more rapidly than long-chain FA (Odle, 1997).Thus, replacing some of the lard with fat sources containing a higher proportion of medium-chain FA, such as coconut oil, may modulate metabolic responses.However, a study conducted by Swank et al. (2012) reported that inclusion of coconut oil in MR did not affect growth performance, as well as serum TG and serum nonesterified FA (NEFA), in Jersey calves that were fed twice daily with 2.0 L of one of 3 MR.These MR formulations contained 28% fat and included 100% lard, a blend of 80% lard and 20% coconut oil, or a blend of 60% lard and 40% coconut oil.
Most studies examining the effects of MR fat sources on calf growth and metabolism are limited to MR containing low fat levels (18% DM) and low feeding rates (4.0 to 6.0 L/d).However, these conditions largely differ from the volume of milk that preweaned calves would voluntarily consume, which ranges from 9.0 to 10.0 L/d (Jasper and Weary, 2002;Wilms et al., 2021), as well as from the fat content of bovine whole milk.Feeding MR high in fat at a high feeding plane may exacerbate any growth and or metabolic responses in relation to fat composition in MR.In most published literature, the FA profile and TG structure in MR fat have seldom been balanced to be closer to bovine milk fat.In this way, evaluating the effects of fat composition in MR on calf metabolism and health is difficult.An interesting approach would be to characterize the total FA profile in the plasma, as these serve as markers (Ma et al., 1995;Nikkari et al., 1995) for dietary fat composition and can provide insights on how calves metabolize different dietary fats.
There is a need to understand how fat characteristics in MR, such as FA profile and TG structure affect calf metabolism and development.This study investigated how fat composition in MR with a high fat content affect growth performance, feeding behavior, and blood metabolites of calves before and after weaning.The hypothesis was that a MR formulated with animal fats from lard and dairy cream would present nutritional benefits compared with formulations with vegetable fats and a mixture of vegetable and animal fats.The benefits of the animal fat treatment are expected based on its FA profile and a TG structure close to that of milk fat.

MATERIALS AND METHODS
This study was conducted at the Calf and Beef Research Facility of Trouw Nutrition Research and Development (Sint Anthonis, the Netherlands) between November 2020 and March 2021.All procedures complied with the Dutch law on Experimental Animals, which follows the principles of ETS123 (Council of Europe, 1985 and the 86/609/EEC Directive), and were approved by the animal welfare authority (Centrale Commissie Dierproeven, CCD, the Netherlands).The project application code is AVD2040020173425.

Experimental Design and Animals
In total, 63 Holstein Friesian male calves were obtained from 8 collaborating commercial dairy farms, and the study was conducted in a randomized complete block design.At the farm of origin, a standardized protocol for colostrum management, including 3 feedings of colostrum in the first 24 h; 3.0 to 4.0 L within the first 3 h after birth, followed by 2 feedings of 2.0 L, was applied.Afterward, meals of 3.0 L of MR (Sprayfo Delta, Trouw Nutrition, Deventer, the Netherlands) were offered twice a day until the collection day.Between 1 and 4 d after birth, calves were transported in a transportation van to the research facility and immunoglobulins (IgG) levels in blood were measured within 48 to 96 h after birth, using a portable Multi-Test Analyzer (Mai Animal Health, Elmwood, WI).Body weight at arrival was 46.6 ± 4.5 kg (mean ± SD), and age at arrival was 3.1 ± 0.84 d.Upon arrival, calves were group housed in one of 4 pens, including 9 animals each.The experiment was performed twice (in 2 rounds), but in the second round, only 3 pens were used, thus resulting in 63 calves over the whole experiment.Each round of the experiment was of longitudinal nature as calves would arrive to the research facility as cows would calve.The recruitment of calves took place between the 8th of June 2020 to the 25th of September 2020 for the first round and between the 30th of November 2020 and the 16th of March 2021.Group pens measured 5.0 m × 5.9 m with flax straw-bedded resting area of 5.0 m × 2.8 m and rubber-slatted floors in front of the feeders.Within each group pen, calves were assigned to one of 3 blocks.Blocking was based on age at arrival, and day of birth, and within each block, calves were randomly assigned to one of 3 experimental MR diets fed right upon arrival.This means that the block effect also included the study round.The randomization was performed using the random function [RANDBETWEEN (0,100000)] in Microsoft® Excel® (Microsoft 365 MSO, Version 2212, Build 16.0.15928.20278).The experimental period was divided into 3 phases, including pre-weaning (P1), weaning transition (P2), and post-weaning (P3).In P1 (d 0 to 42 after arrival), calves were ad libitum fed MR, while in P2 (d 43 to 70 after arrival).Finally, in P3 (d 71 to 84 after arrival), calves were fed exclusively solid feeds.Calves had free access to starter pellets, chopped wheat straw, and water from arrival onwards.The temperature in the calf facility was at a minimum of 12°C and a maximum of 31°C (targeted between 15 and 28°C) and relative humidity between 60 to 80%.Calves were exposed to daylight and artificial light from 0600 to 2200 h, and a nightlight for the remainder of the day.

Dietary treatments and Feeding
The 3 experimental diets (n = 21 per treatment) consisted of: 1) a MR with only vegetable fats composed of 60% unhardened rapeseed oil mixed with 40% of Racomelt fat blend (Cargill, Minnesota, US) (VG); 2) a MR formulated with 100% animal fats, including 65% of packers lard and 35% of liquid dairy cream (AN); and 3) an MR with a mix of animal and vegetable fats, including 80% of packers lard and 20% of coconut oil (MX).The Racomelt product from Cargill (Minnesota, US) was a mixture of 50% coconut oil interesterified with 50% fully hydrogenated rapeseed oil.While using only 2 fat sources, the VG, and MX MR treatments represent 2 rather distinct approaches to the fat composition of MR.The AN treatment represented an attempt to align with the FA profile of milk fat without including 100% milk fat, as this would defeat one of the main purposes of MR, which is increasing the net dairy product yields of dairy farms.All MR were composed of 63.0% skim milk powder, 30.0%oils and fats, 5.0% sweet whey powder, and 2% premix (vitamins, minerals, and additives).During the process of MR production, the fats were sprayed dried.For refined vegetable oils, quality parameters included a maximum of 0.1% free fatty acids (FFA) and maximum of 1 mEq/kg peroxide.The 3 MR formulas were formulated to have the same macronutrient content (30% fat, 24% crude protein, and 36% lactose; dry matter (DM) basis; Trouw Nutrition, Deventer, the Netherlands).Although analytical values presented slight variations (Table 1), it can be considered that these diets were isonitrogenous and isoenergetic.All 3 diets were formulated to be close to the FA profile of bovine milk (Moate et al., 2007) fat while remaining within formulation constraints and using only 2 fat sources (Table 2).However, it should be noted that the entire FA profile of milk fat was not considered, as some FA only exist in milk fat.The MR concentration was 135.0 g/L (13.5% solids) to get closer to the solid content of WM.The farm staff responsible for scoring and sampling the calves were unaware of the treatment group to which each calf belonged as a letter (A, B, or C) was assigned to each treatment, thus ensuring a blinded approach.Milk replacers were fed ad libitum by automated milk feeders (C1000, Förster-Technik, Engen, Germany), whereas chopped wheat straw and starter pellet (solid feed; CBRF calf starter, ForFarmers, Lochem, The Netherlands) were provided via individual automated feeders (CRFI, Biocontrol, Rakkestad, Norway), and water via an automated system (Förster-Technik).
All 3 dietary treatments were available in each pen and supplied to the calves individually using electronic recognition.This means that each pen included 3 blocks of 3 calves each and one automated milk feeder able to dispense all 3 MR treatments through the same teat.The number of daily visits to the automatic milk dispensers and drinking speed was monitored automatically.Visits were recorded as rewarded when the animal received MR, whereas unrewarded visits were visits where the calf did not receive MR. Drinking speed was monitored automatically for each calf and was based on the total rate in minutes per liter consumed.The minimal time interval between 2 milk meals was set at 30 min, meal size was set at a minimum of 0.5 L and maximum 2.0 L, and the maximum drinking speed was set at 0.

Measurements
Milk replacers and starter pellets were sampled for analyses at the start of the study.For each feed 3 samples were taken from 3 different bags to gather a representative sample.Intakes of MR and water were automatically recorded throughout the study period.Growth was measured by weighing the calves using a custom scale (W2000; Welvaarts Weegstemen, Hertogenbosch, the Netherlands) on arrival and then weekly thereafter at fixed days from wk 1 to 12 at 1300 h.Calves arrived at the facility on either Monday or Wednesday and were then weighed on the same day each week.Blood samples were collected from the jugular vein at the same time as BW measurements (on arrival and in wk 1, 2, 3, 4, 6, 8, 10, and 12) in one 9 mL and one 6 mL serum/gel tubes (Vacutest, Italy), one 9 mL and one 4 mL lithium heparin (LH) tubes for plasma (BD Vacutainer ® , United States), and 2 4 mL NaF tubes (BD Vacutainer®, United States).Serum tubes were set for 30 min at room temperature and centrifuged at 1,500 g for 10 min at 20°C (Rotina 380 R, Hettich, Tuttlingen, Germany).Lithium heparin and NaF tubes were placed on ice before centrifugation and were centrifuged at 1,500 g for 15 min at 4°C within 10 min after the blood was collected.Serum and plasma samples were stored in 3-fold or 2-fold (NaF) 2.0 mL cryotubes (set A/B/C).Set A for serum and LH was stored at −80°C, whereas sets B and C, as well as the NaF samples, were stored at −18°C until the analyses were performed.

Chemical Analysis
Milk replacers and starter pellets were processed and analyzed at MasterLab (Boxmeer, the Netherlands).Milk replacer samples and starter pellets were analyzed for DM, crude ash, crude fat, crude protein, minerals (calcium, phosphorus, chloride, sodium, and potassium), and carbohydrates (lactose and glucose).Dry matter content was determined by drying to a constant weight in a 103°C stove for 4 h (EC 152/2009;EC, 2009).Crude ash was analyzed by incineration in a muffle furnace by combustion for 4 h at 550°C (EC 152/2009;EC, 2009).Crude fat was determined by treating the sample with hydrochloric acid and subsequent extraction with petroleum (EC 152/2009;EC, 2009).Crude protein content was analyzed by combustion, according to the Dumas method (Etheridge et al., 1998;ISO 16634-1:2008).Macro-minerals were analyzed using inductively coupled plasma mass spectrometry (Perki-nElmer ICP-MS 300D) according to NEN-EN 2017(2017).Chloride was analyzed as described by Wilms et al. (2019).Carbohydrates in MR were determined by titrimetric method according to 1971 71/250/EEG for lactose and EG 152/2009 for glucose.The FA profile of the fat fraction in MR was analyzed by QLIP (Zutphen, Netherlands) by gas chromatography according to NEN-ISO 15885 (2003).
Analyses of blood chemistry (bilirubin, albumin) and organ function [gamma-glutamyl transpeptidase (GGT), lactate dehydrogenase (LDH), and alkaline phosphatase (ALP)] were performed at SynLab-vet GmbH (Leverkusen, North Rhine-Westphalia, Germany).Plasma total cholesterol, HDL-cholesterol, and TG were measured in Utrecht University (the Netherlands) using a Clinical Chemistry Analyzer (AU680, Beckman Coulter).Low-density (LDL) cholesterol was subsequently calculation using the following formula: LDL-cholesterol = total cholesterol -HDL-cholesterol -(TG × 0.45) Plasma fibrinogen concentrations were measured in Bonn University according to the fibrinogen functional turbidimetric assay, imitating the conversion of physiological fibrinogen to fibrin in plasma (Stief, patent 2008/16).For the standard curve, 20.9 mg/mL fibrinogen from bovine plasma (341573-1 GM, Sigma) was diluted in a 1:3 series in 30 mL sterile phosphate-buffered saline (PBS) (pH 7.4).In addition, samples (300 µL) were diluted in BPS.Afterward, 50 µL of each sample, standard and control were pipetted in duplicates on a microtiter plate and 100 mL FIFTA reagent (0.6 g bovine serum albumin, 10 mL PBS and 3 µL thrombin) was added to each well.Finally, the extinction was measured photometrically at an absorbance of 405 nm at 37°C for 10 min.A subset of plasma samples comprised of 6 calves per treatment (complete blocks) and 4 time points (arrival, wk 4, 8, and 12) were analyzed for FA profile by gas chromatography at the University

Calculations and Statistical Analysis
Based on the outcome of a previous study from Wilms et al. (2022a) 2022) using the same starter feed as in the present study.Total ME intakes included MR and starter feed, while straw was not considered.The energy conversion ratio (ECR) was calculated by dividing the daily total ME intake by ADG.Continuous variables were analyzed using mixed-effects model with PROC MIXED in SAS (SAS 9.4M6, SAS Institute Inc., Cary, NC) with the calf as the experimental unit, and the statistical model was as follows: where: Y ijk is the dependent variable; µ is the overall mean; T i is the fixed effect of the ith treatment; V j is the random effect of the jth block; W k is the fixed effect of the kth week entering the model as a repeated measure; TW ik is the fixed effect interaction between the ith treatment and the kth week; and ε ijk is the random error associated with the jth block at the kth week with the ith treatment.In the case of BW and ADG, arrival BW was included as a baseline covariate (µ 0 ).For variables with equally spaced time points (e.g., BW, intakes, blood acid-base), the covariance structure (autoregressive covariance [AR(1)] or the heterogeneous autoregressive covariance [ARH(1)]) were selected by the lowest corrected Akaike Information Criterion.For variables with unequally spaced time points (e.g., blood parameters), the Toeplitz (TOEP) or heterogeneous Toeplitz (TOEPH) covariance structure was used.
Data that did not meet the assumptions of normal-    (2007) including the average fatty acid profile from milks described in 28 scientific publications.
dard deviation of each variable).Principal component analysis (PCA) was applied to provide an overview of plasma FA composition based on treatments (AN, VG, and MX) at 4 sampling time points (0, 4, 8, and 12 wk).Hierarchical clustering (represented as a heat map) was performed for plasma FA composition using Euclidean distance and Ward's minimum variance method (ward.D) using MetaboAnalyst 5.0 (Pang et al., 2021).Random forest analysis was performed using MetaboAnalyst 5.0 to identify key blood FA that discriminated between dietary treatments and were ranked by mean decrease accuracy.

General Health and Growth Performance
The concentration of IgG measured at arrival did not differ between treatments (21.3 ± 1.50 g/L; mean ± SD).Three calves (2 VG and 1 MX) were removed post-inclusion from the study because of failure to adapt to the automated drinking machines.From these same 3 calves, 2 were removed in the first week and the third one in the third week after arrival.In addition, one calf suffering from severe diarrhea in the AN group was removed in wk 2 after arrival.This resulted in having 19 calves in VG and 20 calves in AN and MX.Partial data collected from these animals were not used in the statistical analyses.The number of calves receiving therapeutic interventions for diarrhea and respiratory disease, as well as the total number of therapeutic interventions, did not differ across treatment groups (data not shown).Throughout the entire experimental period of 12 wk, BW tended to be greater for AN than VG (P = 0.10), whereas MX did not differ from the other groups (Table 3; Supplemental Figure 1.A).The average daily gain in P1 was greater for AN than VG (P = 0.02; Supplemental Figure 1.B), whereas MX did not differ from other groups.When considering the overall study period, ADG from AN calves tended to be greater than VG (P = 0.10), while MX did not differ from other groups.

Feed Intakes and Intake Behavior
Results for liquid and solid feed intakes are presented in Table 3. Preweaning MR intakes were greater for AN than for MX (P < 0.01; Figure 1.A), and there was a trend for greater MR intakes in AN than in VG (P = 0.09).In contrast, no differences were detected between VG and MX.There was also a treatment by time interaction for MR intake when considering the overall rearing period (P = 0.04), in which MR intake were higher in AN than MX in wk 2 to 6 but not in the weaning period.In addition, MR intake was higher in VG than MX in wk 3 and higher in AN than VG in wk 5 and 6.In contrast no differences were present in starter feed and straw intakes across treatments.The higher MR intake in AN led to a greater ME intake in that group as compared with VG and MX in the preweaning phase (P < 0.01), but not in P2 and P3 (Figure 1.B).Throughout the experimental period, the ECR was greater for VG than for AN and MIX (P = 0.02), and these differences were mostly marked during P1.
Throughout the ad libitum phase (P1), the number of rewarded visits to the automated feeders was higher for AN than for MX (P = 0.03), whereas VG did not differ from other groups (Table 4).The number of unrewarded visits was higher in AN than VG and MX (P < 0.01) throughout the study period (P1 and P2).In addition, there was a treatment by time interaction during P2 (P = 0.02), in which the number of unrewarded visits was higher in AN than other groups in wk 7 and was higher in VG than MX in wk 8.This led to a greater total number of visits (unrewarded and rewarded) to the automated MR feeders in P1 for AN compared with MX (P = 0.05) with VEG in between.Throughout P1 and P2, no differences were found between treatment groups for the number of breakoffs, the meal size per visit, and the absolute drinking speed.However, the time per meal was higher in VG than other groups in P1 but not in P2 (P < 0.01).

Blood Parameters
Table 5 provides average blood parameters measured weekly in the first 4 wk of life and at d 42, d 56, d 70, and d 84.Plasma cholesterol was greater in VG than in other treatment groups (P < 0.01; Figure 2.A).Consistently, the plasma HDL-and LDL-cholesterol were greater in calves fed VG as compared with other groups (P < 0.01; Figure 3.A and 3.B).When expressed as a percentage of total plasma cholesterol, HDL-cholesterol (Figure 3.C) was lower in VG than other groups (P < 0.01), whereas LDL-cholesterol (Figure 3.D) was higher in VG than other groups (P < 0.01).Additionally, serum albumin (Figure 2.B) and total protein were lower in VG compared with other groups (P < 0.01).For plasma cholesterol (P < 0.01), HDL-cholesterol (P < 0.01), serum albumin (P = 0.04), treatment-bytime interactions were also present.This reflected the marked treatment differences in the preweaning phase, in contrast to the postweaning phase, in which differences were no longer evident.In addition, serum TG was lower in VG than in AN (P = 0.05).Regarding organ function, serum concentrations for GGT, LDH, and alkaline phosphatase did not differ between the 3

Plasma Fatty Acid Profile
The concentration of FA groups and selected FA measured weekly in plasma are presented in Table 6.The concentrations of C12:0, C14:0, and C18:0 were greater in VG than in AN and MX (P < 0.01).The concentration of medium chain SFA were greater in VG (P < 0.01), while long chain SFA did not differ.Both LA (P = 0.04) and ALA (P < 0.01) were greater in VG than in other groups.The n-3 PUFA concentrations were greater in VG (P < 0.01).Total PUFA and total FA in plasma were greater in VG than in AN and MX (P = 0.02).The n-6 to n-3 ratio was the lowest in VG, followed by AN and then by MX (P < 0.01).
The FA profile, expressed as percentage of total FA is presented in Supplemental material S2.Calves fed VG had a higher proportion of medium chain SFA (P < 0.01), and a lower proportion of long chain SFA (P < 0.01) than other treatment groups.The proportion of MUFA was lower in MX than AN in wk 4 and 12 (P = 0.01), whereas VG was not different from other groups.The proportion of n-6 PUFA was higher in AN than MX in wk 4 (P < 0.01).The proportion of n-3 PUFA was higher in VG than other treatment groups (P < 0.01), leading to a greater total PUFA proportion in plasma of VG calves (P = 0.05).The ratio of n-6 to n-3 PUFA was the highest in MX and the lowest in VG (P < 0.01).Finally, significant interactions (P < 0.05) between treatment and time were present for FA in plasma both when expressed as concentration in plasma (C12:0, C14:0, total MUFA, ALA, n-3 PUFA) and as a percentage total FA in plasma (C12:0, C14:0, C16:0, total MUFA, LA, ALA, total PUFA).This reflects the increase of the concentration of these FA through the ad libitum milk phase in calves fed VG as compared with AN and MX followed by a decline during weaning leading to no differences at 12 wk of age.This was however not the case for C22: 4n -6 and C12: 1cis -11.

Multivariate Analyses
The results of the PCA are shown in Figure 4.These illustrate a clear distinction between the time of sampling at arrival (wk 0), when calves had ad libitum access to MR (wk 4), at weaning (wk 8), and in the post weaning phase (wk 12), but not between dietary treatments (VG, AN, and MX) based on plasma FA composition.The first 3 principal components explained 38%, 21%, and 8% of the total variance between treatments, respectively.Figure 5 shows the clustering result for pareto-scaled (mean centered and divided by the range of each variable) plasma FA concentration in plasma shown as heat map.In Figure 6A, the results of the random forest analysis were used to identify the major FA in plasma that were affected by the dietary treatments, and these FA are shown in order of their mean decrease in accuracy.Among the plasma FA, C18: 1cis -14, C22: 4n -6, C20: 3n -3, C18: 3n -3, and C20: 5n -3 were the most affected by dietary treatments.Blood concentrations of C18: 1cis -14, C20: 3n -3, C18: 3n -3, C20: 5n -3, C20: 1cis -11, C24: 1n -9 were higher in calves fed VG, and those of C22: 4n -6 were lower than in calves fed AN and MX MR at wk 4 and 8 (Figure 6B).

DISCUSSION
This study examined how fat composition in MR would affect growth performance, feeding behavior, and blood metabolites in calves fed ad libitum.Calves fed AN consumed higher volume of MR than other treatment groups, while starter feed intake did not differ across treatment groups.In this way, preweaning ME intake was higher in calves fed AN than calves fed VG and MX.This resulted in higher preweaning ADG in calves fed AN (915 g/d) than VG and MX (~783 g/d).The plasma FA profile of calves fed MR with different fat sources were strongly influenced by dietary treatments at wk 4 and 8 but not at wk 12 when dairy calves were completely weaned.

Performance, Feeding Behavior, and Health
The MR diet did not affect the percentage of calves treated for diarrhea, which was not expected based on previous literature (Hill et al., 2007;Liu et al., 2022;Dell'Anno et al., 2023) as the AN treatment contained 0.88% (% total FA) of butyric acid.A study by Hill et al. (2007) showed that supplementing sodium butyrate in MR (20% fat) fed 2.0 L twice daily resulted in a reduction in abnormal fecal scores and a reduced number of medical treatments compared with no supplementation.Since the MR diets were isoenergetic, the higher MR intake in AN resulted in a greater total ME intake in the preweaning phase compared with other treatment group.Despite the higher MR intake in calves fed AN, starter feed intake was not negatively affected which aligns with other studies in which no differences in starter feed intakes were reported in calves drinking different volumes of milk in the preweaning phase (Seibt et al., 2021;Echeverry Munera et al., 2021;Wilms et al., 2022a).Due to the similar starter feed intakes across treatment groups, there was only a trend for higher ME intake in AN than other groups when considering the entire 12-wk experimental period.

Wilms et al.: FAT COMPOSITION IN MILK REPLACER
In previous studies, MR formulations with 100% lard, as compared with 100% vegetable fats, did not improve the growth of calves (Huuskonen et al., 2005;Hill et al., 2007).However, when a source of mediumchain FA (ΣC6:0 to C12:0) was included in a lard-based fat mixture (80% lard, 15% coconut, and 5% canola), ADG was greater than with lard alone in calves fed 2.0 L of MR (20% fat, DM) twice daily (Hill et al., 2007).In the current study, VG contained a greater proportion of medium-chain FA (22% total FA) than MX (12%) and AN (7%).As a reference, milk fat contains 11% medium-chain FA (Wilms et al., 2022b).The slower growth observed in calves fed VG compared with AN was linked to a higher preweaning ECR in the VG group, indicating a lower feed efficiency.Since the ECR did not differ between AN and MX, it is likely that the greater preweaning growth of calves fed AN compared with MX is solely related to the greater intake of MR during the ad libitum MR phase.Previous studies examining the macronutrient profile in relation Means with a different superscript are significantly different (P ≤ 0.05). 1 In the preweaning phase (P1), calves were group housed and fed ad libitum (135 g/L) to wk 6 after arrival.Weaning (P2) was gradual and took place between wk 7 to 10 after arrival.In wk 11 and 12 (P3), calves were fed solids only.Data are also analyzed by considering the entire 12-wk rearing period. 2 Treatments included three milk replacers: VG was a MR with vegetable oils (80% rapeseed oils and 20% coconut, n = 19), in AN, the MR contained animal fats (65% lard, 35% dairy cream, n = 20), and in MX, a mixture of 80% lard and 20% coconut oil was used (n = 20). 3In the case of BW and ADG, arrival BW was included as a baseline covariate.
*SEM is expressed as log, whereas LSM is back transformed after the log-transformation.
to lactose, fat, and protein levels in MR assumed that ad libitum fed calves would regulate their MR intake based on the energy density of the meal (Berends et al., 2020;Echeverry-Munera et al., 2021;Wilms et al., 2022a).However, in the current study, the MR diets were isoenergetic, and therefore, it is unexpected that calves would consume different amounts of MR when offered ad libitum.
Several human and rodent references have shown that the physicochemical properties of dietary fats influence the feeling of satiety after ingestion and food intake (Feltrin et al., 2004;Lawton et al., 2007;Maljaars et al., 2019).This is mainly determined by the FA chain length (Feltrin et al., 2004;Meyer et al., 1998) and the degree of saturation of dietary fats (Lawton et al., 2007;Kozimor et al., 2013;Maljaars et al., 2019).Some FA including, C12:0 and C18:0, have been demonstrated to regulate appetite by influencing gastric emptying and hormone secretion, such as cholecystokinin (CCK;Fel-trin et al., 2004;Meyer et al., 1998).In an experiment including the same MR as in the present study, calves fed VG as a slower abomasal emptying than AN and MX (Wilms et al., 2023b, in review).Thus, one hypothesis could be that differences in voluntary MR intake might be attributed to the composition in mediumchain FA.The AN treatment contained a lower proportion of medium-chain FA (7% total FA) compared with VG (22% total FA) due to the inclusion of coconut oil (Odle, 1997).After ingestion, medium-chain FA are mainly transported to the liver as free FA, whereas long-chain FA are absorbed as chylomicrons (Schmitz and Ecker, 2008).Therefore, the medium-chain FA are a rapidly available energy source that can influence satiety signaling in calves.However, differences in MR intake were most pronounced between the AN and MX treatments, with no significant difference in the total proportion of medium-chain FA and, more specifically, in the proportion of C12:0.Accordingly, the concentration of medium-chain FA in plasma was higher in VG than in AN and MX. Kozimor et al. (2013) showed that liquid meals rich in MUFA might elicit a weaker satiety response than liquid meals rich in PUFA or SFA in normal-weight women.However, the proportions of MUFA and SFA in the 3 diets were relatively balanced, which does not support this hypothesis in the current study.Alternatively, the positioning of FA on the glycerol backbone of dietary fats could influence intake, but this is not supported by the current data because AN and MX had a similar TG structure expected from the high inclusion of lard.
The lower MR intake in calves fed VG and MX could also have been caused by the inclusion of coconut oil, which was not present in AN.However, this does not explain why the intake of MR was the lowest in MX since VG and MX contained the same inclusion level of coconut oil.Overall, the FA profile of MX was very similar to that of AN, except for butyric and caproic acids.Therefore, it can be hypothesized that the presence of some levels of butyric acid in AN resulted in improved gastrointestinal tract (GIT) development, as previously observed when supplementing sodium butyrate to calves in MR or starter feed (Guilloteau et al., 2009;Niwińska et al., 2017).Enhanced GIT development may also have increased the ingestion capacity of calves fed AN, resulting in heavier calves with higher ME requirements.In addition, metabolic effects induced by butyric acid may have modulated MR intake.In vitro, butyrate increases gene expression of peptide YY and proglucagon (Zhou et al., 2006), which are involved in the regulation of food intake.Alternatively, these differences in MR intake could be related to a lower fat digestibility of AN, resulting in increased MR intake to reach a target level of ME intake.However, this was not supported by Wilms et al. (2023b, in review), where no differences in digestibility of the same MR diets were observed at 3 and 5 wk of age in calves fed restricted volumes of MR.Finally, the palatability of MR could influence the voluntary intake of calves.However, calves fed AN did not consume larger MR meals but had an increased number of daily visits to the automated milk feeders.The effort associated with getting up, going to the automated milk feeder, and waiting in line suggests that the taste of AN was probably not the reason driving the increased MR intake in this group.In addition, absolute drinking speed was not affected by fat sources in MR, although VG calves had longer meal durations than other groups.

Blood Metabolites and Plasma Fatty Acids
Plasma cholesterol is the balance between the input of cholesterol into the blood from both endogenous synthesis and the intestines, as well as the cholesterol efflux from the blood via hepatic extraction and elimination (Stellaard, 2022).A study by Mcnamara et al. (1987) showed that dietary fat quality is a more important determinant of plasma cholesterol concentrations than dietary cholesterol intake.However, the dogma stating that blood cholesterol is affected by dietary cholesterol intake has been refuted (Enas, 1996).Indeed, the contribution of cholesterol synthesis to circulating cholesterol is substantially more important than cholesterol absorption from the intestines (MacKaye and Jones, 2011) Means with a different superscript are significantly different (P ≤ 0.05). 1 In the preweaning phase (P1), calves were group housed and fed ad libitum (135 g/L) to wk 6 after arrival.Weaning (P2) was gradual and took place between wk 7 to 10 after arrival.Data are also analyzed by taking the entire 12-wk rearing period in overall.
2 Relative drinking speed is the drinking speed relative to the average drinking speed on the day of arrival. 3 Treatments included three milk replacers: VG was a MR with vegetable oils (80% rapeseed oils and 20% coconut, n = 19), in AN, the MR contained animal fats (65% lard, 35% dairy cream, n = 20), and in MX, a mixture of 80% lard and 20% coconut oil was used (n = 20).*SEM is expressed as log, whereas LSM is back transformed after the log-transformation.
calves fed a MR high in fat (23%), including a blend of palm and coconut fats, compared with a MR with a lower inclusion (17% fat) of the same fat blend.The authors attributed the higher serum cholesterol to the higher dietary fat intake.Since MR intake was greater in AN, it could be expected that plasma cholesterol would be higher in this group due to the greater fat intake.However, serum cholesterol was noticeably higher in calves fed VG than in the other treatment groups.
According to Hegsted et al. (1965), dietary fat composition is more important than dietary fat amount regarding blood cholesterol determination.Interestingly, the PUFA content was higher in the VG MR (13% total FA) than in other MR treatments (9% total FA).High plasma cholesterol in calves fed VG contradicts human nutrition work describing a lowering of plasma cholesterol with diets high in PUFA, whereas the opposite is observed with diets high in SFA (Stamler, 1978).The current data also contrasts with higher serum total cholesterol concentrations in breastfed infants compared with infants fed with formulas including vegetable fats (Shamir et al., 2003).This is consistent with similar observations where piglets fed sow milk had higher serum cholesterol than piglets fed formula with vegetable fats (Rioux and Innis, 1993).Previous work performed decades ago described that feeding MR (14% fat) with a high proportion of PUFA from the vegetable origin (soybean oil, corn oil) in the fat fraction led to a marked accumulation of cholesterol in the blood plasma, liver, and adipose tissues compared with calves fed MR with tallow only (Wiggers et al., 1977;Barrows et al., 1980;Richard, 1980).In contrast, different inclusion levels of tallow in MR did not result in different blood and tissue cholesterol concentrations in these same experiments.It is worthwhile mentioning that soybean and corn oils have a much higher PUFA proportion (55% total FA) than the VG treatment of the current study (13% total FA), however, fat inclusion was also smaller than in the current study (14 vs. 30%).The metabolic pathways responsible for the contradictory effect of high dietary PUFA on serum cholesterol in humans and (non-ruminating) ruminants are not well-understood.In ruminants, the rumen biohydrogenation process plays a crucial role in converting unsaturated FA to SFA (Jenkins et al., 2008).As a result, the FA leaving the rumen and absorbed in the intestines are predominantly saturated.In humans, the cholesterol-lowering effect of dietary PUFA was attributed to a redistribution of cholesterol between blood plasma and body tissues while keeping the total body balance of cholesterol stable (Avigan and Steinberg, 1965;Bieberdorf and Wilson, 1965;Spritz et al., 1965).Hence, the observed cholesterol accumulation in the body related to high dietary PUFA and low SFA could be attributed to alterations in cholesterol synthesis or excretion rates in calves.The study of Richard et al. (1980) showed that in response to feeding MR with soybean and corn oils to calves, there was a shift in plasma cholesterol from the high-density lipoprotein (HDL; 2-thirds of total cholesterol) to the low-density lipo- Means with a different superscript are significantly different (P ≤ 0.05). 1 Abbreviation: TG = triglyceride, GGT = Gamma-glutamyltransferase, LDH = Lactate dehydrogenase, AST = Aspartate transaminase, BHB = β-hydroxybutyric acid. 2 Treatments included three milk replacers: VG was a MR with vegetable oils (80% rapeseed oils and 20% coconut, n = 19), in AN, the MR contained animal fats (65% lard, 35% dairy cream, n = 20), and in MX, a mixture of 80% lard and 20% coconut oil was used (n = 20).*SEM is expressed as log, whereas LSM is back transformed after the log-transformation.protein (LDL) fraction (2-third of total cholesterol).In contrast, calves fed a MR with tallow or vegetable shortening oils had an equal distribution between the HDL and LDL-cholesterol fractions.Consistently, the proportion of LDL-cholesterol increased through time in calves fed VG as compared with AN and MX and the consumption of dry feed decreased serum cholesterol concentrations leading to no differences at 12 wk of age.Finally, serum TG concentrations was lower in calves fed VG as compared with calves fed AN.These results align with Wilms et al. (2023b, in review), in which postprandial TG concentration were lower in calves fed VG as compared with AN.This is also consistent with Wilms et al. (2023a, in press) in which calves fed a whole milk powder had increased postprandial TG as compared with calves fed MR containing different fat levels from palm and coconut oils.
Serum bilirubin, GGT, LDH, and alkaline phosphatase were not affected by dietary fat composition in MR.In contrast, serum albumin concentration was lower in calves fed VG than in calves fed AN and MX.Serum albumin is synthesized and secreted by the liver (Moshage et al., 1987).It has many functions, including serving as a transport molecule for blood metabolites such as bilirubin and amino acids, providing protection against oxidative stress, and serving as an antiinflammatory molecule (Garcia-Martinez et al., 2013).In rats, supplementation with n-3 PUFA is reported to increase serum albumin concentration (Abdou and Hassan, 2014).However, ALA is an n-3 PUFA and was higher in VG than in AN and MX.Calves fed MX had lower serum aspartate aminotransferase concentrations than calves fed VG and AN.Aspartate aminotransferase is an enzyme found in various tissues in the body, such as the liver and muscle, and is used as an indicator of tissue damage and liver function (Whitehead et al., 1999).However, the calves in the present study were in good general health, and aspartate aminotransferase alone is not sufficient to assess liver health (Yu et al., 2019).
Plasma FA composition can be used as a marker of dietary fat composition (Glatz et al., 1989;Ma et al., 1995;Nikkari et al., 1995).Specific plasma FA profiles have been associated with the occurrence of metabolic diseases such as insulin resistance and diabetes, in humans (Pelikanova et al., 2001;Laaksonen et al., 2001).For example, in humans, an elevated risk of coronary heart disease has been observed in association with higher concentration of long-chain SFA and lower concentration of n-3 PUFA in plasma phospholipids (Liu et al., 2019).Different fat composition in MR resulted in a distinct plasma FA profile in plasma measured in the preweaning and weaning phases, whereas no differences were found at arrival.Also, in the postweaning phase, only a few FA were still different among dietary treatment groups.The total plasma concentration of FA was greater in calves fed VG, although calves fed AN had a greater preweaning fat intake.This is consistent with the higher plasma cholesterol in VG calves, as cholesterol is part of the lipid fraction of plasma.Overall, the composition of plasma FA was quite similar between AN and MX, which was expected since at least 65% of the lipid fraction was lard in both treatment groups.Interestingly, the concentration of SFA in plasma was greater in VG calves, although the VG treatment had a lower SFA content in the fat fraction.In the SFA fraction of plasma, the concentration of medium-chain FA was higher in the plasma of VG calves, consistently with the high proportion of medium-chain FA in the VG treatment.Medium-chain FA are readily absorbed because they do not rely on incorporation into chy- lomicrons for absorption (Hageman et al., 2019) and are oxidized more rapidly than long-chain FA, as their uptake into mitochondria is independent of carnitine transport (Marten et al., 2006).Once they are taken up into the hepatic portal vein and transported to the liver, their oxidation releases energy (Lapillonne et al., 2014).Furthermore, despite a significantly lower proportion of C16:0 in the VG treatment (8% total FA) than in other groups (~26% total FA), plasma C16:0 concentrations were similar in all treatment groups.This is because saturated FA can be synthesized endogenously from nonfat sources (e.g., carbohydrates, AA) or by β-oxidation of unsaturated FA (Brenna, 2002).
The PCA analysis showed that the rearing period was an important determinant of plasma FA profile as there was a clear distinction between arrival (wk 0), the ad libitum phase (wk 4), at weaning (wk 8), and after weaning (wk 12), but not between dietary treatments based on blood plasma FA composition.The FA that were the most affected by the MR diets, as clustered in the random forest analysis, including MUFA (C18: 1cis -14; C20: 1cis -11; C12: 1cis -11), n-3 PUFA (C20: 3n -3; C18: 3n -3, ALA; C20: 5n -3, EPA; C22: 5n -3), and n-6 PUFA (C22: 4n -6).For these selected FA, few differences seemed to be present between calves fed AN and MX, apart from C20: 5n -3 and C22: 5n -3, which were greater in calves fed AN than MX.Although the MUFA content was the same in all 3 treatment groups, plasma MUFA concentrations were elevated in calves fed VG compared with AN and MIX, apart from C12: 1cis -11.However, the nutritional significance of MUFA for neonates has not been sufficiently studied (Delplanque et al., 2015), and these results are difficult to interpret.Both plasma n-6 and n-3 PUFA concentrations were greater in calves fed VG, which is consistent with the higher proportions of these FA in the VG treatment.The greater LA and ALA plasma concentrations in the VG treatment likely served as precursors for longer-chain PUFA found in plasma (Jensen et al., 1997).Linoleic acid is primarily elongated to ARA (C20: 4n -6) and dihomogamma LA (20:3n-6; Noble et al., 1978), whereas ALA is primarily preceded by EPA (C20: 5n -3) and DHA (22:6n-3;Sardesai, 1992).All the above-mentioned FA were highly elevated in plasma from calves fed VG compared with the plasma from calves fed AN and MX.In neonates, essential FA and specialized pro-resolving mediators, produced through PUFA metabolism within cells, play a crucial role in growth, organogenesis, inflammation regulation, and neurodevelopment (Cheruku et al., 2002;Helland et al., 2003;Ramiro-Cortijo et al., 2020).Docosahexaenoic acid has been associated with cognitive and visual development, as well as immune functions (Smith and Rouse, 2017), while EPA plays a role in the production of eicosanoids, which serve diverse functions, including mitigation of inflammation in the body and brain (Meydani, 1996;Endres et al., 1989).Similarly, ARA is crucial for growth, brain development, and overall health of infants by acting as a precursor to eicosanoids important for the immune system (Hadley et al., 2016;Jasani et al., 2017).However, the markedly higher concentrations of total FA in plasma coupled with high serum cholesterol may be related to an accumulation of lipids in the plasma of VG-fed calves.This may indicate that certain FA are not metabolized by VG calves which might be concerning.Excessively high levels of LA in infant formulas can lead to adverse effects due to proinflammatory functions of certain oxygenated metabolites (Innis, 2007;Ramsden et al., 2012).

CONCLUSION
The study confirmed the expected greater nutritional value of blend of 65% lard and 35% dairy cream over 100% vegetable oils and a mixture of 80% animal fat and 20% vegetable oil.Although health parameters were not improved with AN, MR intakes, and preweaning growth were greater in calves fed AN compared with VG and MX.Calves fed VG presented high concentrations of plasma cholesterol with a higher proportion of LDL-cholesterol, which is a negative indicator of metabolic health.Plasma FA concentrations mirrored the FA composition of the diet, and consequently, large differences were observed between VG and other treatments.However, differences did not persist in the postweaning phase and further research is required to understand the biological relevance of fat composition in MR on long-term metabolism.
5 L/min until d 21 and 0.63 L/min thereafter.From d 43 to d 70, calves were gradually weaned from a maximum daily consumption of 8.0 L on d 43 to 2.0 L on d 70 by reducing the daily intake by 0.214 L/d.A minimum time interval of 1.5 h was set between visits to the MR feeder.At d 71, calves were weaned and monitored until d 84.Starter pellets and chopped straw were fed ad libitum during the pre-and post-weaning period.
Wilms et al.: FAT COMPOSITION IN MILK REPLACER of Guelph (Ontario, Canada) as described in Heileson et al. (2021).
ity of residuals had to be log-transformed (base 10).After the log-transformation, the distribution of the data was tested again, and the data were normally distributed.Comparisons of means across treatments were conducted with the PDIFF option of the LSMEANS statement of the MIXED procedure of SAS.Significant interactions between treatment and time were explored with the SLICE option of the LSMEANS statement using the PDIFF option of the MIXED procedure of SAS.Discrete variables (e.g., therapeutic interventions) were analyzed using mixed effects logistic regressions.These analyses were conducted with PROC GENMOD in SAS.Results in tables and figures are presented as least squares means (LSM) with the standard error of the means (SEM).In the case of log-transformed data, the SEM was not transformed back to the original scale(Bland and Altman, 1996).Significance was declared at P ≤ 0.05, and the trend threshold was set at 0.05 < P ≤ 0.10.Multivariate Analyses.Multivariate statistical analyzes of plasma FA composition were performed using the web-based metabolomics data processing tool MetaboAnalyst 5.0 (http: / / www .metaboanalyst.ca;Pang et al., 2021).Data were transformed using the generalized logarithm transformation, and then Pareto scaled (mean divided by the square root of the stan-Wilms et al.: FAT COMPOSITION IN MILK REPLACER Wilms et al.: FAT COMPOSITION IN MILK REPLACER treatments.Serum AST concentration was higher for VG and AN compared with MIX (P = 0.01).

Figure 1 .
Figure 1.Intake of milk replacer (A) and metabolizable energy (B), measured daily in calves fed ad libitum from the third week after arrival onward.Treatments included 3 milk replacers: VG was a MR with vegetable oils (80% rapeseed oils and 20% coconut, n = 19), in AN, the MR contained animal fats (65% lard and 35% dairy cream, n = 20), and in MX, a mixture of 80% lard and 20% coconut oil was used (n = 20).

Figure 2 .
Figure 2. Plasma total cholesterol (A) and serum albumin (B) concentrations measured at 1300 h at arrival and weekly thereafter on d 7, d 14, d 21, d 28, d 42, d 56, d70 and d 84 in calves fed ad libitum.Treatments included 3 milk replacers: VG was a MR with vegetable oils (20% coconut and 80% rapeseed oils, n = 19), in AN, the MR contained animal fats (65% lard and 35% dairy cream, n = 20), and in MX, a mixture of 80% lard and 20% coconut oil was used (n = 20).

Figure 3 .
Figure 3. Concentration of high-density lipoprotein (HDL) cholesterol (A) and low density (LDL) cholesterol (B) in plasma, as well as the proportion of HDL-cholesterol (C) and LDL-cholesterol (D) on total cholesterol.Blood samples were measured at 1300 h at arrival and weekly thereafter on d 7, d 14, d 21, d 28, d 42, d 56, d70 and d 84 in calves fed ad libitum.Treatments included 3 milk replacers: VG was a MR with vegetable oils (80% rapeseed oils and 20% coconut, n = 19), in AN, the MR contained animal fats (65% lard and 35% dairy cream, n = 20), and in MX, a mixture of 80% lard and 20% coconut oil was used (n = 20).

Figure 6 .
Figure 6.Random forest (RF) analysis (A) identified important blood fatty acids affected by dietary treatments based on the mean decrease accuracy.The blood concentrations of selected FA (B) identified by RF analysis.Treatments (n = 6 calves per treatment): VG was a MR with vegetable oils (80% rapeseed oils and 20% coconut), in AN, the MR contained animal fats (65% lard and 35% dairy cream), and in MX, a mixture of 80% lard and 20% coconut oil was used.
Wilms et al.: FAT COMPOSITION IN MILK REPLACER conducted in the same research facility as the present study investigating MR macronutrient profiles on growth performance and metabolism, it was estimated that 19 calves per treatment group would be sufficient.Metabolizable energy (ME) content of the MR was calculated as follows: ME (Mcal/kg) = [0.057× CP (%) + 0.092 × crude fat (%) + 0.0395 × lactose (%)] × 0.93 (NRC, 2001).The ME of the starter feed was calculated using equations from NRC (2001) and was as follows: ME (Mcal/kg) = (1.01 × DE -0.45) + 0.0046 × (fat in starter feed), in which DE was the digestible energy.The DE value was measured by Amado et al. (

Table 1 .
Wilms et al.: FAT COMPOSITION IN MILK REPLACER Nutrient composition of milk replacers and starter pellet fed ad libitum to group-housed calves (n = 59) 1 Expressed in g/kg DM unless specified otherwise.2Treatmentsincluded three milk replacers: VG was a MR with vegetable oils (80% rapeseed oils and 20% coconut, n = 19), in AN, the MR contained animal fats (65% lard, 35% dairy cream, n = 20), and in MX, a mixture of 80% lard and 20% coconut oil was used (n = 20).

Table 2 .
Fatty acid profiles of milk replacers differing in fat sources fed ad libitum to group-housed calves (n = 59) 2From Moate et al.

Table 3 .
Wilms et al.: FAT COMPOSITION IN MILK REPLACER Feed intakes measured and growth performance (least squares means) per rearing period in ad libitum fed calves receiving milk replacers differing in fat sources (n = 59).Feed intakes were measured daily, whereas BW was measured weekly

Table 4 .
. Previous work by Echeverry-Munera et al. (2021) reported greater serum total cholesterol in Wilms et al.: FAT COMPOSITION IN MILK REPLACER Feeding behavior (least squares means) per rearing period measured daily in ad libitum fed calves receiving milk replacers differing in fat sources (n = 59)

Table 6 .
Wilms et al.: FAT COMPOSITION IN MILK REPLACER Concentration of fatty acid groups and selected fatty acids in plasma measured weekly at 1300 h on arrival, and on d 28, 56, and 84 after arrival in ad libitum fed calves receiving milk replacers differing in fat sources (n = 18)