Effects of prepartum concentrate feeding on reticular pH, plasma energy metabolites, acute phase proteins, and milk performance in grass silage fed dairy cows

We investigated how concentrate feeding during the last 21 d of pregnancy affects reticular pH, inflammatory response, dry matter (DM) intake, and production performance of dairy cows. We hypothesized that adding concentrates to dairy cow’s diet before calving reduces the decrease in reticular pH postpartum and thus alleviates inflammatory response. We also hypothesized that prepartum concentrate feeding increases DM intake postpartum and consequently improves milk performance. Two feeding experiments were conducted using a randomized complete block design. In each ex-periment, 16 multiparous Finnish Ayrshire cows were paired based on parity, expected calving date, body weight, and milk yield of the previous lactation. Within the pairs, cows were randomly allocated on one of the 2 dietary treatments 21 d before expected calving. In exp. 1, diets were ad libitum feeding of grass silage as a sole feed or supplemented with increasing amounts of concentrate offered separately (increased to 4 kg/d by d −7). In exp. 2, diets were ad libitum feeding of a total mixed ration containing either grass silage, straw, and rapeseed meal (64%, 28%, and 8% on DM basis, respectively) or grass silage, straw, and cereal-based concentrate mixture (49%, 29%, and 30% on DM basis, respectively). Following calving, all the cows were fed similarly and observed until d 56 postpartum. Feed intake and milk yield were recorded daily, and reticular pH was monitored continuously by reticular pH bolus. Blood samples were collected at the beginning of the experiments, 7 d before the expected calving date, 1 d (in exp. 1) or 5 d (in exp. 2), 10 d, and 21 d postpartum. In exp. 1, concentrate feeding increased metabolizable energy intake and tended to increase DM and crude protein intake prepartum. Moreover, prepartum concentrate feeding increased the concentrations of plasma β-hydroxybutyrate and insulin, but differences in nonesterified fatty acids, glucose, or acute phase proteins were not observed. After calving, prepartum diet did not affect DM or nutrient intake, plasma energy metabolites, or milk production in exp. 1. Although prepartum concentrate feeding before calving increased reticular pH on the first day of lactation, it elevated plasma concentrations of serum amyloid-A and haptoglobin postpartum in the grass-silage based diets. In exp. 2, adding concentrates to the diet based on a mixture of grass silage and straw did not affect prepartum DM intake or plasma concentrations of nonesterified fatty acids, glucose, or insulin. Adding concentrates to prepartum diet increased plasma concentration of β-hydroxybutyrate before calving as in exp. 1. After calving, prepartum concentrate feeding increased DM and nutrient intake during the second week of lactation in exp. 2, but no effects were observed thereafter. In contrast to our hypothesis prepartum concentrate feeding decreased reticular pH after calving (6.47 vs. 6.59) in exp. 2, but no differences in inflammatory markers were observed. Based on this study, close-up concentrate feeding in diets based on grass silage with or without straw does not alleviate the decrease in reticular pH or mitigate inflammatory response postpartum.


INTRODUCTION
The transition period is a crucial time for the health and production performance of dairy cows.Unsuccessful transition from pregnancy to lactation impairs milk performance, health, and fertility of the cows (Drackley, 1999).In early lactation, a rapid increase in the energy requirements for milk performance exceeds Effects of prepartum concentrate feeding on reticular pH, plasma energy metabolites, acute phase proteins, and milk performance in grass silage fed dairy cows energy intake resulting in a negative energy balance.To maintain a high milk yield and to ensure sufficient energy intake, the energy content of the diet is typically raised by adding concentrates, such as cereal grains, to the lactation diet.Concentrates, however, are rich in rumen fermentable starch, and therefore increasing the concentrate proportion in a diet decreases ruminal pH by enhancing VFA production in the rumen.A decline in ruminal pH may result in SARA, which adversely affects the health, feed intake, and lactation performance of dairy cows (Kleen et al., 2003;Plaizier et al., 2012).A sudden change from a high-forage and low-starch dry period diet to a high-concentrate lactation diet may predispose cows to SARA (Nordlund et al., 1995;Kleen et al., 2003), and adding concentrates to a dairy cow's diet during the last weeks of gestation (i.e., a close-up period) is assumed to alleviate the risk of the disorder (Humer et al., 2018).Prepartum concentrate feeding has been reported to enhance ruminal papillae growth (Dirksen et al., 1984;Dieho et al., 2017) and consequently VFA absorption rate.
During SARA, increased concentrations of positive acute phase proteins, such as serum amyloid-A (SAA) and haptoglobin (Hp), have been observed (Gozho et al., 2005;Khafipour et al., 2009a).Serum amyloid-A and Hp are widely used indicators for inflammatory response, and elevated concentrations have been observed in the proximity of parturition (Murata et al., 2004;Trevisi et al., 2009).Some studies have also reported a decrease in negative acute phase protein concentrations (e.g., albumin, retinol-binding protein) (Bertoni et al., 2008;Trevisi et al., 2009).Inflammation is an energy-consuming process, which may impair DMI and consequently aggravate negative energy balance during the transition period (Bradford et al., 2015;Kuhla, 2020).Intensive body fat mobilization following negative energy balance may be associated with local inflammation in adipose tissue, thus contributing to systemic inflammation (Kuhla, 2020).
The studies of the close-up diet's effects on postpartum inflammation have been contradictory.In recent studies, the high starch content in the prepartum diet increased SAA and Hp levels postpartum (Shi et al., 2020;Haisan et al., 2021).In contrast, in the study of Little et al. (2017) supplementing grass silage with concentrates during the close-up period decreased inflammatory response based on serum albumin concentration.Further, Roche et al. (2015) observed that in lean cows (BCS 4 on a scale of 1 -10) increasing energy intake from 75% to 100% or 125% of the energy requirements raised postpartum plasma albumin concentration indicating reduced inflammation.
Although adding concentrates to the close-up diet may have some physiological advantages, e.g., increased blood glucose levels, decreased blood nonesterified fatty acid (NEFA) and BHB concentrations, benefits in terms of lactation performance are not often reported (Minor et al., 1998;Doepel et al., 2002;Yang et al., 2022b).In fact, several previous studies have shown that regardless of the close-up concentrate feeding, a single unchanged diet with controlled energy intake throughout the dry period has led to favorable responses, such as improved energy balance and decreased liver total lipid concentration, in the metabolism of dairy cows (Dann et al., 2006;Richards et al., 2020).In these studies, one of the major forage sources during the dry period has been corn silage, and the starch content of the prepartum treatments has varied from 120 to 270 g per kg DM.In Northern latitudes, however, corn is not a commonly cultivated crop, and dry cows are typically fed with grass silage.If the digestibility of grass silage is too high for dry cows, straw is often used to dilute energy content of the total mixed ration (TMR).As grass silage lacks starch, the shift from the dry period diet to the high-starch lactation diet may be too drastic and disturb rumen function during the transition period.Moreover, in Northern Europe, it is still rather common to offer forages and concentrates separately, which is reported to predispose cows to SARA (Maekawa et al., 2002b).However, studies investigating the effects of close-up concentrate feeding on rumen pH and inflammatory response in grass silage-based diets are scarce.
We hypothesized that adding concentrates to the close-up diet alleviates the decrease in reticular pH and decreases plasma positive acute phase protein levels in early lactation.We also hypothesized that prepartum concentrate feeding increases DMI postpartum and thus, improves milk performance.The overall objective of this study was to investigate how close-up concentrate feeding of dairy cows affects reticular pH, inflammatory response, DMI, and lactation performance in prepartum diets based on either grass silage or a mixture of grass silage and straw.

Cows, diets, housing, and experimental design
Two feeding experiments, approved by the Finnish National Animal Ethics Committee (Hämeenlinna, Finland), were conducted at University of Helsinki research farm in Helsinki, Finland.All experimental procedures were conducted following the guidelines established by the European Union Directive 2010/63/EU and the current Finnish legislation on animal experimentation (Act on the Protection of Animals Used for Scientific or Educational Purposes 497/2013).Animal welfare and health were monitored daily, and if symptoms of illness or distress occurred, the veterinarian was consulted.In both experiments, 16 multiparous Finnish Ayrshire cows were used in a complete randomized block design.Cows were paired according to parity, BW, milk yield of previous lactation, and expected calving date (Table 1).Within the pairs, the cows were randomly allocated to one of the 2 dietary treatments from 21-d before expected calving until parturition and observed until d 56 postpartum.Experimental procedures and allocation to the treatments were available to all experimental staff.
In both experiments, cows were housed on tie stalls bedded with wooden shavings during the close-up period.In experiment 1 (Exp1), cows were moved to individual calving pens approximately 5 d before expected calving and returned to the tie stalls approximately 4 d after parturition.In experiment 2 (Exp2), cows were moved to the calving pens approximately 3 d before expected calving and back to tie stalls approximately 5 d after parturition.During the stay in the tie stalls, cows were milked twice daily (at 0600 and 1630 h).The cows were moved to a loose housing system on an average 19 d postpartum in Exp1 and 14 d postpartum in Exp2.In the loose housing system, cows were milked with the automatic milking system (AMS; Lely Astronaut A3, Lely International, Maasluiss, Netherlands).

Feeding in Exp1
In this experiment, the treatments comprised feeding first cut grass silage as a sole feed (FOR1) offered ad libitum or supplemented with an increasing amount of a cereal-based concentrate mixture for 21 d before expected calving (Figure 1 panel A) (CONC1).The concentrate mixture contained (on DM basis) 39% oat, 32% barley, 15% faba bean, 10% molassed sugar beet pulp, 3% mineral supplement (Seleeni-E-Melli TMR, Raisioagro Ltd., Raisio, Finland), and 1% propylene glycol.The chemical compositions of the feed ingredients are presented in Supplemental Table S1.The grass silage was distributed to cows twice daily (0900h and 1900h) and concentrates were offered separately at 0600h and 1700h to cows in the CONC1 group.If cows did not eat all the concentrates, the orts were collected and weighed before morning feeding.All cows also received 100 g/d of a dry cow mineral mixture (Tunnu-Melli, Raisioagro Ltd.) dressed on top of the silage during the morning feeding.In addition, cows in the FOR1 group received an increasing amount (30 g/d from d −21 to d −15, 75 g/d from d −14 to d −7, and thereafter 125 g/d until parturition) of the same mineral supplement (Seleeni-E-Melli TMR, Raisioagro Ltd.) used in the concentrate mixture in the CONC1 diet.However, to avoid a risk of hypocalcemia, we removed this mineral supplement from both prepartum diets in the middle of the experiment and increased the amount of the dry cow mineral mixture (Tunnu-Melli) to 200 g/d.
After calving, both groups were fed similarly.From d 1 to d 10 cows were offered separately ad libitum the same first cut grass silage as fed prepartum, increasing amount of the same concentrate mixture given to CONC1 group prepartum supplemented with a commercial protein feed from d 2 (Amino-Maituri, Raisioagro) (Figure 1 panel A).From d 11 cows were offered ad libitum lactation partial mixed ratio (PMR) of grass silage (1st or 2nd cut, 65% on DM basis) and the concentrate mixture (35% on DM basis).The cows in the last 2 pairs received second cut grass silage from the lactation wk 5 since the first cut silage ran out before the study had been completed for all the cows.As the pairs were formatted according to the expected calving date, both cows in each pair received the same grass silage during lactation.In addition to lactation PMR, all cows received commercial protein and compound feeds (Amino-Maituri and Maituri 10 000, Raisioagro) as shown in Figure 1 panel A. The commercial feeds were fed separately during the stay in the tie stall and from the AMS after moving to loose housing.
After calving, grass silage and lactation PMR was distributed 3 times daily (0900h, 1300h, and 1900h) and concentrates were offered separately at 0600h, 1300h, and 1700h during the stay in tie stalls.After moving to loose housing, commercial concentrates were fed from the AMS during the milkings.

Feeding in Exp2
The prepartum treatments contained either a high fiber TMR of first cut grass silage (64% on DM basis),  After calving all cows received the same diet.From d 1 to d 10 they received ad libitum grass silage and increasing amounts of same concentrate mixture as prepartum with mineral supplement for lactating cows and commercial protein feed (Amino-Maituri, Raisioagro Ltd.; light blue).From d 11 cows received lactation partial mixed ratio (PMR) ad libitum, commercial protein feed (Amino-Maituri) and compound feed (Maituri 10 000, Raisioagro Ltd.; dark blue) offered separately.Panel B: Feeding and blood sampling in Exp2.From d −21 to calving cows were fed either with prepartum TMR containing grass silage, straw, and rapeseed meal (FOR2) or prepartum TMR containing grass-silage, straw, and cereal-based concentrate mixture (CONC2).After calving all cows were offered lactation PMR ad libitum, commercial protein feed (Amino-Maituri, Raisioagro Ltd.; light blue) and compound feed (Maituri 10 000, Raisioagro Ltd.; dark blue).Amounts in the figure are presented as fresh weight.
barley straw (28% on DM basis), and rapeseed meal (8% on DM basis) (FOR2) or TMR of the grass silage (49% on DM basis), barley straw (21% on DM basis), and a cereal-based concentrate mixture (30% on DM basis) (CONC2) offered ad libitum.The prepartum concentrate mixture for CONC2 contained 40% oat, 26% barley, 13% rapeseed meal, 10% rolled faba bean, 10% molassed sugar beet pulp, and 1% propylene glycol on DM basis.The chemical compositions of the TMR components are presented in the Supplemental Table S1.In addition to TMR, all cows were separately offered 200 g/d the dry cow mineral mixture (Tunnu-Melli, Raisioagro Ltd.) during the close-up period.The TMR was prepared in a feeding wagon (TMR-SUK M2, Pellon Group, Ylihärmä, Finland) twice a week.The TMR was distributed twice daily (0900h and 1800h), and the mineral mixture was dressed on top of the ration during the morning feeding.
After calving, both groups were fed similarly.Cows were fed with lactation PMR of grass silage (45.5% on DM basis), whole crop barley silage (19.5% on DM basis) and concentrate mixture (35% on DM basis).The concentrate mixture in the postpartum ration was similar to the prepartum, except the proportion of rapeseed meal was 10%, and a mineral supplement (Seleeni-E-TMR, Raisioagro Ltd.) was added with a proportion of 3%.Cows also received increasing amounts (Figure 1 panel B) commercial protein and compound feeds (Amino-Maituri and Maituri 10 000, Raisioagro Ltd.).During the stay in tie stalls the lactation PMR was distributed twice daily (0900h and 1800h), and the commercial concentrates were offered separately 3 times daily at 0600h, 1300h, and 1700h.In addition, all cows received 200 g/d mineral supplement (Seleeni-E-TMR, Raisioagro Ltd.) dressed on top of the ration during the morning feeding.After moving to the loose house system commercial concentrates were offered from AMS similarly as in Exp1.

BW and BCS
In both experiments, cows were weighed once a day on 2 consecutive days at the beginning of the study and one week before expected calving (CV 9600 Scale, Solotop Oy, Helsinki, Finland).When necessary, feeding of cows was delayed after blood sampling and weighing.In the case of delayed calving cows were weighed once a week on 2 consecutive days until calving.After calving, while staying in tie stalls, the cows were weighed once a d 3 h after morning milking 1 d, 2 d, 9 d and 10 d postpartum in Exp1, and 2 d, 3 d, 9 d and 10 d postpartum in Exp2.Body condition score with a scale of one to 5 (Edmonson et al., 1989) was determined by the same person at the same time points as BW.After moving to the loose housing system, the BW was measured by AMS during the milkings (3 times per day on average in both experiments) and body condition scoring was done 4, 6, and 8 weeks postpartum.

Blood Samples and Analysis
Blood samples were taken once a day before morning feeding at 0900h from the coccygeal blood vessel at the beginning of the experiment, 7 d before expected calving and 1 d (±1) (Exp1) or 5 d (±1) (Exp2), 10 d (±1) and 21 d (±1) after calving.In the case of delayed calving blood samples were collected once a week until parturition.The first 2 samples postpartum were taken 3 h after morning milking before forage or PMR distribution.At d 21 the sample was taken at 0800h right after milking in AMS.Blood samples were collected in 9 mL vacuum tubes (Vacuette, Geiner Bio-One GmbH, Kremsmünster, Austria) containing either EDTA (for glucose, insulin, NEFA and BHB) or heparin (for SAA and Hp) as anticoagulant and kept in ice after collection until centrifugation.The samples were centrifuged within 30 min after collection (2220 × g for 10 min, at 4°C) to separate plasma.The plasma samples were stored at −20°C until analysis.Plasma concentrations of glucose, insulin, and NEFA were analyzed as described by Salin et al. (2012), and the concentration of plasma BHB was analyzed as reported by Puhakka et al. (2016).Commercial enzymatic kits were used for the analysis of plasma SAA (TTP8020 phase TM Range Multispecies SAA ELISA kit, Tridelta Development Ltd. Maynoot, Co. Kildare, Ireland) and Hp (TP801 phase TM Range Haptoglobin kit, Tridelta Development Ltd.).

Reticular pH
Reticular pH was monitored continuously by wireless, intra-ruminal pH bolus (SmaXtec, Austria) validated by Klevenhusen et al. (2014).Before application, boluses were calibrated according to the manufacturer's instructions (1-point calibration with a buffer of pH 7.00).Bolus (123 mm × 35 mm, ~220 g) was inserted into the reticulorumen via esophagus with a balling gun after a 2-h feed restriction 22 d (Exp1) or 28 d (Exp2) before the expected calving date.Bolus measured reticular pH every 10 min with the measurement range of pH 3 -9 (according to manufacturer) and transmitted the data in a real time to a receiver installed in the barn.

Feed Samples, Analysis, and Measurements
Feed intake was recorded daily by Roughage Intake Control system (Insentec BV, Marknesse, Netherlands).In Exp1, silage samples were taken twice a week and samples of the concentrate mixture were taken once weekly.In Exp2, samples of prepartum TMR components were taken during the preparation of TMR twice a week.Samples of postpartum forages were taken twice weekly and of postpartum concentrates once weekly.For the commercial concentrates, samples were taken monthly in each experiment.The chemical composition of individual feed components and their proportions were used to reconstitute the composition of TMR.
The DM of the forages was determined by oven drying (103°C, 24 h) weekly after sample collection to adjust the composition of the TMR and PMR.Forage samples were stored at -20°C and concentrate samples at 4°C until further analysis.Samples of the forages and the concentrate mixtures were pooled for 4 consecutive weeks.For the commercial concentrates in Exp1, samples were pooled for 2 consecutive months.In Exp2, the commercial concentrate samples were analyzed monthly.Samples from the mineral supplements were taken once a month and pooled into one sample in each experiment for the analysis of DM and ash.
The pooled samples were dried in the oven (50°C, 48 h) and were ground through a 1 mm screen (Sakomylly KT-3100, Koneteollisuus Oy, Helsinki, Suomi) for subsequent feed analysis.The samples were analyzed for OM (ashing at 550°C for 12 h), and NDF using sodium sulfite (Van Soest et al., 1991) with FiberTherm FT12 analyzer (Gerhardt, Königswinter, Germany).To determine NDF content of the experimental concentrates and barley silage, α-amylase was used in addition to sodium sulfite.The NDF content is reported on ashfree basis.Crude protein was analyzed as described by Puhakka et al. (2016) and total fat with petroleum ether extraction and hydrolysis with HCl as described by Lamminen et al. (2017).The starch content of the concentrates and barley silage was measured by colorimetric method (Salo and Salmi, 1968) using a spectrophotometer (Shimadzu UV-VIS mini 1240, Shimadzu Europa GmbH, Duisburg, Germany).
Subsamples of each feed ingredient were dried at 103°C for 24 h to determine DM.The DM content of the silages was corrected for the loss of volatile compounds (ammonium-N, ethanol, lactic acid, VFA) according to Huida et al. (1986).Silage lactic acid, ethanol, ammonium-N, and water-soluble carbohydrate contents were measured using colorimetric methods reported by Puhakka et al. (2016).The concentrations of VFA were measured using the liquid chromatography method that Puhakka et al. ( 2016) have described.

Milk Samples and Analysis
Milk yield was recorded for every milking (WB Auto Sampler, Tru-Test, Aucland, New Zealand) during the stay in the tie stalls and with AMS after moving to the loose house.Milk samples were taken for 4 consecutive milkings at wk 1, 2, 4, 6, and 8 postpartum.The samples were preserved with Bronopol preservative (Valio Ltd., Seinäjoki, Finland) and sent to a commercial laboratory (Valio Ltd.) for mid-infrared analyses (MilkoScan FT+; Foss Electric A/S, Hillerød, Denmark) for fat, CP, lactose, urea, and for SCC measurements (Fossomatic FC, Foss Electric A/S).

Calculations and Statistical Analysis
In Exp1, one cow in the CONC group was excluded from the experiment at d 10 postpartum due to clinical ketosis.All postpartum results of this cow were excluded from the statistical analyses.Thus, the total number of cows in the final postpartum data was 7 for CONC1 and 8 for FOR1 in Exp1.In Exp2, one cow in CONC2 group was treated for anorexia during the first 2 weeks after calving and one in FOR2 had interdigital phlegmon at wk 5 postpartum.All results of these cows from those periods were excluded from the statistical analyses.The acute phase protein results of the first postpartum blood sampling were removed from the cows with the retained placenta (1 cow in FOR1 and 2 cows in CONC1 in Exp1; 2 cows in FOR2 and 1 cow in CONC2 in Exp2).
The chemical composition of each feed ingredient and daily DMI were used to calculate chemical composition of the experimental diets presented in Table 2.The metabolizable energy content of the experimental forages was calculated by multiplying in vitro digestible OM in DM with 0.016 and feed table values were used for concentrate mixtures (Luke 2021).In vitro digestible OM in DM was determined according to Nousiainen et al. (2003).Energy intake was corrected for the associated effects of feeds and feeding level (Luke 2021), and ECM yield was calculated according to Sjaunja et al. (1991).Metabolizable energy balance was calculated as the difference between the corrected energy intake and the ME requirements calculated according to Luke (2021).
Prepartum and postpartum data were analyzed separately.Daily means were used to calculate feed intake, mean and minimum reticular pH prepartum and from d 1 to d 10 postpartum.Weekly means were calculated for the analyses of feed intake, milk performance, and mean and minimum reticular pH from wk 1 to 8 postpartum.Weekly means were used also for the analyses of ME intake and BW.Two different thresholds for increased SARA risk, reticular pH 6.00 suggested by  2018) and pH 6.30 by Sato et al. (2012) were used in calculations.Daily time below the threshold was calculated for each day by summing the count of 10-min periods below the threshold and dividing the sum by the number of days in the period considered.
Repeated measures for variables over time, i.e., feed intake, milk performance, postpartum plasma energy metabolites, mean and minimum pH, postpartum energy balance, and postpartum BCS and BW were analyzed as repeated measures ANOVA using MIXED procedure of SAS version 9.4 (SAS Institute Inc., Cary, NC) with the statistical model as follows: where Y ijk is the dependent variable, μ is the overall mean, D is the effect of the diet, T is the effect of time, and TD is the interaction of T and D, P is the effect of pair and PT is the interaction of P and T, and E is residual error.D, T, and TD were considered as fixed effects, and P and PT as random effects.Three covariance structures (compound symmetry, unstructured, and the first order autoregressive) were tested, and the one with the lowest Bayesian information criterion was selected.Instead of the first order autoregressive structure, spatial power structure was used for the variables with unevenly spaced sampling intervals (e.g., milk components).
Plasma energy metabolites prepartum, prepartum BW and BCS, change in BW and BCS, and daily time below pH threshold were analyzed by ANOVA with a following model: where Y ii is the dependent variable, D is the fixed effect of the diet, P is the random effect of the pair, and E is residual error.In addition, we used pre-experimental concentrations of plasma energy metabolites as covariates to analyze their concentrations at d −7.If the effect of the covariate was significant, the result is presented in the tables in addition to the unadjusted result.
The normality of residuals was tested using the UNIVARIATE procedure and Shapiro-Wilk test.If the residuals were not normally distributed, the variables were transformed (log, square, square root, 1/square root) to achieve normal distribution of the residuals.If data were not normally distributed after the transformations, interquartile range (IQR) was used to detect outliers.Observations that were more than 1.5 IQR below the lower quartile or more than 1.5 IQR above the upper quartile were considered outliers, and those outliers were excluded from the analysis.Three outliers were removed from milk fat yield results in Exp2 (2 in FOR2, 1 in CONC2) and 1 (CONC1) from milk urea concentration in Exp1.The data are presented as least squares means and the largest standard error of the mean is reported in the tables.Data for plasma acute phase proteins were not normally distributed after the aforementioned procedures and were consequently analyzed using Friedman's non-parametric test.The concentrations of plasma acute phase proteins are presented as arithmetic means in the figures.Treatment effects were considered as significant at P ≤ 0.05 and a tendency at 0.05 < P ≤ 0.10.

Prepartum DMI, Plasma Energy Metabolites, BW, and BCS
In Exp1, prepartum NFC and ME intake were greater (P < 0.01 and P = 0.04, respectively; Table 3), and DM and CP intake tended to be greater for cows fed with CONC1 than FOR1 (P = 0.06 and P = 0.07, respectively).Body weight, change in BW, or BCS did not differ between the groups.The plasma concentrations of BHB at the beginning of the experiment and one week before calving were greater for cows in the CONC1 than in the FOR1 groups (P = 0.03 and P = 0.02, respectively), whereas the concentrations of NEFA were greater for cows in the FOR1 than in the CONC1 group (P = 0.05) (Table 4).The difference in NEFA, however, disappeared when pre-experimental concentration was used as a covariate.In Exp2, no differences in prepartum DM or nutrient intake were observed, except intake of NFC was greater (P < 0.01) and intake of NDF tended (P = 0.09) to be lower in the CONC2 group than in the FOR2 (Table 3).The concentration of BHB tended to be greater (P = 0.09) for the cows in the CONC2 group than in the FOR2 7 d before expected calving, but no differences were observed in plasma glucose, insulin, or NEFA (Table 4).Prepartum concentrate feeding did not affect BW, change in BW, or BCS prepartum.

Postpartum DMI, Plasma Energy Metabolites, BW, BCS, and Milk Performance
In Exp1, prepartum concentrate feeding had no effect on postpartum DM (Figure 2 panel A) or nutrient intake (Table 5), or plasma glucose, insulin, NEFA, or BHB levels (Table 6).Also, the prepartum diet did not affect milk performance (Table 7).Postpartum BW or BCS did not differ between the groups (Figure 2

panel C).
In Exp2, rate of DMI increase differed (interaction P = 0.02) between CONC2 and FOR2 groups at early lactation (Figure 2 panel B), as well as the intake of nutrients and ME (Table 5).At wk 3 DMI in the FOR2 group reached the same level as the CONC2 group, and DMI remained similar level for both groups for the rest of the trial.We did not observe any differences between the groups in the concentrations of plasma energy metabolites (Table 6) or postpartum BW or BCS (Figure 2 panel D).The prepartum diet did not affect milk or ECM yield, or milk composition, but milk fat yield was greater at the second week of lactation (interaction P = 0.008) in CONC2 compared with FOR2 (Table 7).

Reticular pH
In Exp1, prepartum concentrate feeding did not affect average reticular pH or time below pH 6.00 or 6.30 before or after calving (Table 8).However, the reticular pH was significantly higher for cows in the CONC1 group than in the FOR1 group on the first day of lactation (interaction P = 0.02; Figure 3 panel A), but no differences were observed thereafter.

Plasma Acute Phase Proteins
After calving, prepartum concentrate feeding tended to increase plasma concentrations of SAA and Hp on the first day of lactation in Exp1 (P = 0.07; Figure 4 panel A and C).In addition, the concentration of SAA was greater in the CONC1 group than in the FOR1 group at d 21.At the beginning of the trial or 7 d before expected calving, differences in plasma concentrations of acute phase proteins were not observed.In Exp2, prepartum concentrate feeding did not affect the concentrations of acute phase proteins either before or after calving (Figure 4 panel B and D).

DISCUSSION
We investigated if adding concentrates to a dry cow's diet based on grass silage with or without straw reduces the decrease in reticular pH, alleviates inflammatory response, and increases DMI and milk performance postpartum.Supplementing the dry period diet based on a mixture of grass silage and straw with concentrates decreased reticular pH postpartum but did not result in increased concentrations of plasma SAA and Hp.In contrast, prepartum concentrate addition did not affect reticular pH postpartum in grass silage-based diets, but it elevated the concentrations of plasma inflammatory markers.Regardless of the forage type, prepartum concentrate feeding had no effect on postpartum DMI, plasma concentrations of energy metabolites, or milk performance.

Prepartum DMI and Plasma Energy Metabolites
Adding concentrates to the grass-silage based closeup diet increased DMI during the last weeks of pregnancy which was also observed in the studies of Keady et al. (2001), McNamara et al. (2003), and Little et al. (2017).However, concentrate feeding with the mixture of grass-silage and straw did not affect DMI prepar- tum.The lack of effect on prepartum DMI in Exp2 is most likely explained by the lower digestibility of basal diet and offering feeds as TMR.The inclusion of straw to the grass silage-based diets during the dry period reduces OM digestibility and therefore decreases DMI (Dewhurst et al., 2000;McNamara et al., 2003;Salin et al., 2018).When feed ingredients are offered separately as in Exp1, concentrates partially replace the forage intake which increases the concentrate proportion in the diet and results in greater passage rate of the feeds from the rumen (Maekawa et al., 2002a).Higher plasma NEFA and lower BHB concentrations in FOR1 compared with CONC1 at the beginning of Exp1 may indicate that animal characteristics contributed to differences between treatments at 7 d before calving.However, covariate adjusted concentrations of NEFA suggest no significant treatment effect in lipid mobilization prepartum.Further, in the absence of a difference in ME balance before calving the treatment effect on lipid mobilization was unlikely.Consequently, the greater BHB concentration in the CONC1 group than in FOR1 was most likely originating from the increased rumen butyrate production (Rabelo et al., 2003;Shi et al., 2020).
Increasing the proportion of concentrates in the diet during the close-up period elevates plasma insulin level prepartum (Kokkonen et al., 2004) which was seen 7 d before calving in Exp1.Adding concentrates to the dairy cow's diet increases ruminal propionate production stimulating insulin secretion (Sutton et al., 1988) In addition to increased propionate production, plasma insulin level was likely increased by the greater ME intake in CONC1 than FOR1 (Andersen et al., 2004).In Exp2, however, prepartum concentrate feeding did not affect plasma insulin concentration, which was most likely due to the lack of difference in ME intake between the groups and more evenly distributed daily concentrate supply from TMR than from separately offered feed ingredients (Sutton et al., 1986).

Postpartum DMI, Plasma Energy Metabolites, and Milk Performance
Close-up concentrate had only marginal effects on the DMI increase rate after calving.Consistently with DMI, no treatment effects were observed on milk performance.The results agree with former studies in which adding concentrates or increasing the energy density of the diet has not affected either DMI or milk yield (Keady et al., 2001;Little et al., 2017;Richards et al., 2020).A temporary increase in milk fat yield in the CONC2 group can be attributed to increased mobilization of body reserves (Pullen et al., 1989;Keady et al., Rissanen et al.: Prepartum concentrate for transition cows Corrected for the associated effects of the feeds and feeding level according to Luke (2021).2001; Roche et al., 2009).However, we did not observe differences either in plasma NEFA concentrations on d 10 or d 21, or change in BW or BCS after calving in Exp2.The lack of differences in plasma energy metabolites, or ME balance supports the conclusion that prepartum concentrate feeding had no major effect on cow's postpartum energy status in the current study.

Reticular pH
In contrast to our hypothesis and results by Dieho et al. (2017) adding concentrates to the close-up diet decreased the mean reticular pH and tended to increase time below pH 6.30 during the first 7 weeks of lactation when the diet consisted of a mixture of grass-silage and straw.Including concentrates in dry cows' diet is expected to improve the VFA absorption capacity of the rumen (Dirksen et al., 1984;Dieho et al., 2017), which has been considered to decrease the susceptibility to SARA (Penner et al., 2009).However, in line with the previous study of Shi et al. (2020), our results do not confirm that adding concentrates to a close-up diet would alleviate the decrease of ruminal pH during lactation.It is noteworthy that in the study of Shi et al. (2020), cows fed with high starch prepartum diet (261 g/kg DM, mainly from barley silage) likely experienced SARA already before calving.Previous exposure to SARA during the close-up period may have caused a greater reduction in postpartum rumen pH in cows fed with high starch prepartum diet, as Dohme et al. (2008) showed that SARA is more severe when recurring.
Symptoms of SARA, such as decreased DMI, milk fat concentration, and fat: protein ratio, have been observed when reticular pH is below 6.00 for 120 min per day (Villot et al., 2018).In our study, however, we did not observe differences in DMI or milk fat content between the groups.In contrast, a total of 18 individual cows (7 in CONC groups and 11 in FOR) had reticular pH below 6.00 for at least 120 min daily after calving on one or more days, despite the moderate starch content of the postpartum diets (150 and 190 g/kg DM) in the 2 experiments.The observation indicates that many of the cows encountered periods of decreased reticular pH although the dietary starch content was moderate compared with SARA induction studies.Substantial inter-animal variation in rumen pH profile and susceptibility to SARA is well established (Gao and Oba, 2014;Humer et al., 2015;Dijkstra et al., 2020).For instance, Humer et al. (2015) reported a considerable difference in time below reticular pH 5.80 between SARA tolerant and SARA susceptible cows (15 min/d vs. 750 min/d, respectively).Cows' risk of SARA and rumen pH response to high grain diets are likely affected by individual characteristics of cows, such as feeding behavior (Plaizier et al., 2018;Stauder et al., 2020), rumen microbial composition (Khafipour et al., 2009b;Yang et al., 2022a) or parity (Maekawa et al., 2002b;Neubauer et al., 2020).According to the authors' knowledge, there are no studies investigating the effects of prepartum concentrate feeding on reticular pH in cows with different SARA susceptibility.

Plasma Acute Phase Proteins
Earlier studies have shown that low rumen pH and SARA may result in systemic inflammation (Plaizier et al., 2012;Zebeli et al., 2012;Aschenbach et al., 2019).Besides SARA, stress and/or transient feed reduction can increase plasma concentrations of positive plasma acute phase proteins (Zhang et al., 2013;Kvidera et al., 2017;Horst et al., 2021).Elevated concentrations may be related to increased rumen permeability, allowing the translocation of bacterial endotoxins from the rumen to blood circulation (Khafipour et al., 2009a;Steele et al., 2011).We hypothesized that prepartum concentrate feeding would alleviate the decline of rumen pH postpartum and therefore decrease plasma concentrations of SAA and Hp.In contrast to our hypothesis, adding concentrates to the close-up diet increased the plasma concentration of these acute phase proteins postpartum in Exp1.In line with the results in Exp1, Shi et al. (2020) observed that feeding a diet containing 26% of starch during the close-up period tended to increase plasma SAA concentration after a grain challenge on d 7 postpartum.They observed also that the concentration of plasma SAA before the grain challenge tended to be higher for cows fed the high-starch prepartum diet than the low-starch prepartum diet, possibly due to the previous exposure to SARA during the close-up period.In Exp1, however, increased SAA concentration in the CONC1 diet cannot be attributed to a reduction of reticular pH, because reticular pH was greater in CONC1 than FOR1 at d 1 postpartum, and no differences between treatments were observed thereafter.It has been suggested that overfeeding in optimally conditioned cows, or restricting energy intake in thin cows during the close-up period may induce local inflammation in adipose tissue stimulating lipolysis which may increase fatty acid flow to the liver and subsequently result in an inflammatory response (Ametaj et al., 2005;Vailati-Riboni et al., 2016).In Exp1, the BCS of the cows was above the mid-point of the scale (3.41 on scale 1-5) at the enrollment of the study, which may have contributed to the greater increase in SAA and Hp in the CONC1 group.However, in the absence of a difference in plasma NEFA in Exp1, the effect of adipose tissue lipolysis on plasma concentrations of SAA and Hp cannot be confirmed.

CONCLUSIONS
In contrast to our hypotheses adding concentrates to the close-up diet based on a mixture of grass silage and straw decreased the reticular pH postpartum.Moreover, prepartum concentrate supplementation increased postpartum concentrations of plasma SAA and Hp in the high-quality grass silage-based diet, but the increase cannot be attributed to differences in reticular pH.We conclude that prepartum concentrate feeding in grass-silage based diets with or without straw does not alleviate the reduction in reticular pH postpartum or mitigate inflammatory response potentially caused by decreased rumen pH.Further, offering concentrates during the last weeks of gestation had no sustained effect on feed intake, energy metabolism, and lactation performance postpartum.Since there is substantial inter-animal variation in reticular pH and SARA susceptibility, more research is required to investigate prepartum nutritional management options for cows with high SARA susceptibility.

Figure 1 .
Figure 1.Panel A: Feeding and blood sampling in Exp1.From d −21 to parturition cows were fed either with grass silage as a sole feed (FOR1) or supplemented with cereal-based concentrate mixture offered separately (CONC1).After calving all cows received the same diet.From d 1 to d 10 they received ad libitum grass silage and increasing amounts of same concentrate mixture as prepartum with mineral supplement for lactating cows and commercial protein feed (Amino-Maituri, Raisioagro Ltd.; light blue).From d 11 cows received lactation partial mixed ratio (PMR) ad libitum, commercial protein feed (Amino-Maituri) and compound feed (Maituri 10 000, Raisioagro Ltd.; dark blue) offered separately.Panel B: Feeding and blood sampling in Exp2.From d −21 to calving cows were fed either with prepartum TMR containing grass silage, straw, and rapeseed meal (FOR2) or prepartum TMR containing grass-silage, straw, and cereal-based concentrate mixture (CONC2).After calving all cows were offered lactation PMR ad libitum, commercial protein feed (Amino-Maituri, Raisioagro Ltd.; light blue) and compound feed (Maituri 10 000, Raisioagro Ltd.; dark blue).Amounts in the figure are presented as fresh weight.

Figure 3 .
Figure 3.Effect of prepartum concentrate feeding on reticular pH 1-10 d postpartum in Exp1 (panel A; P-value for diet × time interaction = 0.02 and pooled SEM = 0.04) and on reticular pH 1-8 wk postpartum in Exp2 (panel B; P-value for diet × time interaction = 0.02 and pooled SEM = 0.061).In panel A: FOR1 = prepartum diet containing grass silage as a sole feed; CONC1 = Prepartum diet containing grass silage supplemented with cereal-based concentrate mixture offered separately.In panel B: FOR2 = prepartum TMR containing grass silage, straw, and rapeseed meal; CONC2 = prepartum TMR containing grass-silage, straw, and cereal-based concentrate mixture.All data are presented as LSM, and tendency for differences (P ≤ 0.10) at the time point is illustrated with ○.

Figure 4 .
Figure 4. Effects of prepartum concentrate feeding on plasma concentration of serum amyloid-A and haptoglobin in Exp1 (panel A and C, respectively) and in Exp2 (panel B and D, respectively).In panel A and C: FOR1 = prepartum diet containing grass silage as a sole feed; CONC1 = Prepartum diet containing grass silage supplemented with cereal-based concentrate mixture offered separately.In panel B and D: FOR2 = prepartum TMR containing grass silage, straw, and rapeseed meal; CONC2 = prepartum TMR containing grass-silage, straw, and cereal-based concentrate mixture.The treatment effects were tested with Friedman's non-parametric test at each time point.Significant difference is illustrated with asterisk (* P < 0.05) and with ring for tendency (○ P ≤ 0.10).

Table 1 .
Rissanen et al.: Prepartum concentrate for transition cows Animal data at enrollment (d −21 relative to expected calving)

Table 2 .
Rissanen et al.:Prepartum concentrate for transition cows Average nutrient composition of the diets (% of DM, unless otherwise noted) 1

Table 3 .
Rissanen et al.:Prepartum concentrate for transition cows Effect of concentrate feeding during the close-up period on dry matter, nutrient, metabolizable energy (ME) intake, and energy balance in multiparous dairy cows prepartum

Table 4 .
Rissanen et al.:Prepartum concentrate for transition cows Effects of prepartum concentrate feeding on plasma energy metabolites in multiparous dairy cows prepartum 2CONC = In Exp1 prepartum diet containing grass silage supplemented with cereal-based concentrate mixture offered separately.In Exp2, prepartum TMR containing grass-silage, barley straw, and cereal-based concentrate.

Table 5 .
Effect of concentrate feeding during the close-up period on dry matter, nutrient and metabolizable energy (ME) intake in multiparous dairy cows 1-8 wk postpartum In Exp1 prepartum diet containing grass silage supplemented with cereal-based concentrate mixture offered separately.In Exp2, prepartum TMR containing grass-silage, straw, and cereal-based concentrate.

Table 6 .
Effects of concentrate feeding before calving on plasma energy metabolites in multiparous dairy cows after postpartum CONC = In Exp1 prepartum diet containing grass silage supplemented with cereal-based concentrate mixture offered separately.In Exp2, prepartum TMR containing grass-silage, straw, and cereal-based concentrate.

Table 7 .
Rissanen et al.:Prepartum concentrate for transition cows Effect of concentrate feeding during the close-up period on milk production and milk composition in multiparous dairy cows postpartum

Table 8 .
Rissanen et al.:Prepartum concentrate for transition cows Effect of concentrate feeding on close-up period on reticular pH in multiparous dairy cows