MECHANISMS BY WHICH FEEDING SYNTHETIC ZEOLITE A AND DIETARY CATION ANION DIFFERENCE DIETS IMPACT FEED INTAKE, ENERGY METABOLISM, AND MILK PERFORMANCE: PART II.

The objectives of this study were to assess the effects of feeding 2 different diets, a low dietary cation-anion difference ( DCAD ) or a diet with synthetic zeolite A to multiparous Holstein cows during the close-up period on DMI and energy metabolism, as well as evaluate colostrum and milk production. A hundred and 21 multiparous Holstein cows, blocked by lactation number and expected parturition date were enrolled at 254 d of gestation and randomly assigned to 1 of 3 dietary treatments: control ( CON ; +190 mEq/kg; n = 40), negative DCAD ( DCAD , −65 mEq/kg; n = 41; Ultra Chlor; Vita Plus, Lake Mills, WI, USA), or a diet containing sodium aluminum silicate zeolite ( XZ ; +278 mEq/kg, fed at 3.3% DM, targeting 500 g/day; n = 40; X-Zelit, Protekta Inc., Lucknow, ON, Canada/Vilofoss, Graasten, Denmark). Prepartum DMI was measured daily using Insentec Roughage Intake Control ( RIC ) gates (RIC System, Holofarm Group, Netherlands). All cows received the same postpartum diet. Blood and urine samples were collected daily beginning 14 d before parturition (D-14) until parturition (D0), and on 1, 2, 3, 6, 9, 12, 15, 18, 21, 35, and 49 d postpartum. Colostrum collected within 6 h of parturition, weighed, and based on samples Brix value, IgG concentrations, and nutrient composition were analyzed. Prepartum, cows fed XZ diet had decreased DMI (11.70 ± 0.26, 13.88 ± 0.26, and 13.45 ± 0.25 kg/d for XZ, CON, and DCAD respectively) and lower rumination (487 ± 8.1, 531 ± 8.3, and 527 ± 8.5 min for XZ, CON, and DCAD respectively) compared with CON and DCAD fed cows. However, rumination was not different postpartum due to treatment. No prepartum or postpartum differences were observed for glucose or BHB concentrations in blood between dietary treatments. Colostrum collected from cows fed XZ had the highest IgG concentrations (91.10 ± 2.63, 78.00 ± 2.63, and 78.90 ± 2.63 mg/ mL for XZ, CON, and DCAD, respectively), but yield did not differ between dietary treatments. Additionally, cows in their third or greater lactation fed XZ had the highest milk production (51.0 ± 1.1 kg) during the first 49 d in milk. This study demonstrates that despite a decrease in DMI and rumination in cows fed XZ pre-partum, blood BHB concentrations were not altered. Additionally, cows fed XZ had higher colostral IgG concentrations and 3+ lactation cows fed XZ produced the most milk. These data suggest that feeding XZ prepartum may improve colostrum quality and improve milk yield in mature cows, and does not impact energy metabolism.


INTRODUCTION
Periparturient dairy cows undergo numerous metabolic and hormonal stresses caused by the events surrounding late pregnancy and early lactation.These challenges are due to the sudden dramatic changes in the cow's physiological demands at the onset of colostrum synthesis and milk production.Due to the alterations in metabolism to support lactation during this timeframe, dairy cows suffer from hypocalcemia (HC), negative energy balance (NEB), depressed feed intake, and immunosuppression (Gerloff, 2000;Ospina et al., 2010;Chapinal et al., 2011).Hypocalcemia is a well-established mineral disorder that occurs during the periparturient period and is considered to be a gateway disease that leads to other disorders that impact the immune system and metabolism.Further, incidences of HC and ketosis have been previously associated with increased incidences of displaced abomasum (DA), mastitis, retained placenta (RP), and metritis (LeBlanc  et al., 2005;Chapinal et al., 2011).Additionally, HC is well-established to have negative impacts on milk production (Santos et al., 2019;McArt and Neves, 2020).However, little is known about the effects of different prepartum diets on colostrum quality and quantity.
Calcium (Ca) homeostasis during the prepartum period has been shown to impact colostrum quality, with one potential explanation for this being that maintaining optimal Ca levels before calving can enhance the immune response during this transition period (Kimura et al., 2006;Martinez et al., 2012).Calcium has been shown to improve B cell function which can lead to increased IgG in the colostrum (Baba and Kurosaki, 2011).However, research has not continually supported such thoughts with studies observing no differences in colostrum quality when feeding a prepartum negative DCAD diet or supplementing the diet with synthetic zeolite A showing no alterations in colostrum quality, despite marked improvements in Ca homeostasis (Martinez et al., 2018;Kerwin et al., 2019;Rajaeerad et al., 2020).
Cows that suffer from subclinical hypocalcemia (SCH) have been shown to have increased blood nonesterified fatty acids (NEFA) and β-hydroxybutyrate (BHB) concentrations for the first 12 DIM compared with cows that do not experience SCH (Martinez et al., 2012).Researchers have concluded SCH increases mobilization of adipose tissue and impairs neutrophil function, increasing the risk of disease during early lactation (Martinez et al., 2012).Thus, to combat clinical hypocalcemia (CH) and subclinical hypocalcemia (SCH), research and industry has focused on feeding acidogenic diets to improve Ca homeostasis through this critical timeframe and avoid the aforementioned impairments.
Several studies have explored energy metabolism and milk production outcomes between cows receiving a negative DCAD diet relative to cows receiving a positive DCAD diet prepartum.Cows receiving a negative DCAD diet during prepartum had a decrease in DMI due to the induced metabolic acidosis, but had higher milk production and increased DMI postpartum (Joyce et al., 1997;Charbonneau et al., 2006;Lopera et al., 2018;Martinez et al., 2018).Feeding acidogenic diets prepartum has been shown to reduce the incidence of SCH and decrease blood NEFA, BHB, and liver triacylglycerol concentrations postpartum (DeGroot et al., 2010).However, less is known about the influence of supplementation of synthetic zeolite A prepartum, a phosphorus-binding supplement which helps mitigate hypocalcemia, on energy metabolism postpartum.Despite improvements in Ca homeostasis, previous studies have demonstrated decreases in DMI when feeding synthetic zeolite A during the prepartum period (Thilsing et al., 2007;Pallesen et al., 2008;Grabherr et al., 2009;Kerwin et al., 2019).Further, examination of inclusion rates of synthetic zeolite A in the prepartum diet has demonstrated increased circulating NEFA and BHB concentrations and authors attributed this to the reductions in DMI (Grabherr et al., 2009), while other research has shown no differences in hyperketonemia or energy balance postpartum when feeding synthetic zeolite A (Kerwin et al., 2019).Similarly, feeding synthetic zeolite A has been shown to decrease DMI, but despite DMI reductions, milk production was not impacted compared with cows fed a control diet (Thilsing et al., 2007;Grabherr et al., 2009;Kerwin et al., 2019).The objectives of this study were to evaluate the effects of feeding 2 different DCAD diets or a synthetic zeolite A diet during the prepartal period on energy metabolism, colostrum composition and production, and evaluate milk production in multiparous Holstein cows.Our study was guided by 2 main hypotheses.Our first hypothesis was that feeding synthetic zeolite A diet prepartum would lead to a reduction in DMI during the prepartum period.Our second hypothesis was that the supplementation of synthetic zeolite A diet would not induce an elevation in postpartum BHB concentrations, nor would it have any impact on the quality or quantity of colostrum produced or total milk production of cows.

MATERIALS AND METHODS
All experimental procedures were approved by the College of Agriculture and Life Sciences Animal Care at the University of Wisconsin -Madison under protocol number A006169-01 and strictly followed.Multiparous Holstein cows entering their second to sixth lactation (3.24 ± 0.10 average lactation parity) at the University of Wisconsin-Madison Emmons Blaine Dairy Research Center (UW; Arlington, WI) from November 2019 to April 2020 were utilized for this study.

Experimental Design, Cows, and Feeding
Experimental design, cows, and feeding are described in Frizzarini et al. (2023).Briefly, a hundred and 28 multiparous Holstein cows arranged in a randomized complete block design were enrolled in the experiment, of which 121 remained in the final data set.Cows that received Ca boluses, received their dietary treatment for less than 14 d, or longer than 35 d were not eligible for inclusion in the data set.Cows averaged (±SE) 829.20 ± 8.21 kg of BW when enrolled in the experiment (control: 828.29 ± 14.31 kg; negative DCAD: 834.07 ± 14.50 kg; zeolite A: 824.97 ± 14.14 kg).At 254 d of gestation, cows were moved to the close-up pen and started dietary treatment.The sample size determination was initially based on biologically important differences in serum total Ca concentrations between parturition and 2 DIM from previous work reported in Kerwin et al., 2019.Our study included more cows than the previously published study.The detectable difference in total Ca (tCa) concentration between the 3 treatments were at least 0.20 mmol/L, with α = 0.05 and β = 0.20.
Cows were blocked by expected date of parturition and lactation number and were randomly assigned to 1 of 3 prepartum experimental diets: control (CON; +190 mEq/kg; n = 40), negative DCAD (DCAD, −65 mEq/kg; n = 41; Ultra Chlor; Vita Plus), or a diet containing sodium aluminum silicate zeolite (XZ; +278 mEq/kg, targeting 500 g/day as fed; n = 40; X-Zelit, Protekta Inc.).The experimental prepartum diets were fed as a TMR once daily at approximately 1200h.After parturition all cows received the same lactating TMR once daily at approximately 0730h.Experimental diets were formulated based on the assumption of 13.5 kg of DMI, accounting for 5% refusals using NRC 2001 models (Supplementary Table 1).Feed intakes were recorded daily prepartum.During the close-up period, cows were housed in a single free-stall pen with 15 Insentec Roughage Intake Control (RIC) gates (RIC System, Holofarm group).The 15 RIC gates were randomly allocated to receive 1 of the 3 experimental prepartum diets, resulting in 5 RIC gates for each treatment allowing access to a group of 10 cows to 5 RIC gates that contained the same diet (Weld and Armentano, 2018).Each cow had an electronic ear tag (FDX EID TAG, Allflex) that allowed them access to any of the feeders containing the diet to which they were assigned.Twice a week, TMR samples were collected and composited and sent to Dairyland Laboratories Inc. for wet chemistry analysis (Supplementary Table 2).Postpartum cows were placed on a common diet balanced to NRC, 2001 requirements for a 650 kg Holstein cow milking 50 kg/d of ECM.Postpartum cows were housed in a freestall facility, bedded with sand and received a common TMR 1x per day and were milked 2x daily.Collection of individual cow feed intakes was not possible due to facility limitations for housing for postpartum cows.
At 254 d of gestation, all cows received an intramuscular shot for E. coli mastitis and multiple gramnegative bacterial diseases (Endovac -Dairy, Endovac Animal Health) for the prevention of mastitis, a subcutaneous shot of injectable trace mineral containing zinc, manganese, selenium, and copper (Multimin 90, Multimin North America), and a pour-on deworming treatment (Cydectin, Bayer HealthCare LLC, Animal Health Division).

DMI, Rumination time, BW, Back Fat Thickness, and BCS
Individual prepartum daily DMI was measured using the information provided by the RIC gates prepartum.Each cow also had a SmartBow ear-tag-based accelerometer system (Smartbow GmbH) used for the collection of rumination data throughout the experiment.All data was collected in real-time by a receiver device (Smartbow Wallpoint) and sent to a local server (Smartbow Station).Rumination data was summarized as total minutes of rumination per day.Cows were weighed once a week, every week (XR3000, Tru-Test ® ).Dry cows were weighed before feeding at approximately 1100h and lactating cows were weighed after feeding and before afternoon milking around 1200h.
Back fat thickness was measured at 3 time points: the day cows were enrolled in the experiment, the day of parturition (D0), and 49 DIM (D49).A portable B-mode ultrasound with a 5-MHz linear transducer was used to measure the backfat thickness (Ibex Pro, E.I Medical Imaging).Skin contact with the transducer was made using 70% ethanol.The transducer was positioned vertically at an imaginary line between the hooks and pins at the sacral examination site.Light pressure was applied to avoid fat compression (Schröder and Staufenbiel, 2006).After freezing the image, the distance between the skin and the profound fascia was calculated.Body condition score (BCS) with quarter points was determined when cows were enrolled in the experiment on D0, D14, D21, D35, and D49 and using the 5.0 scale (Edmonson et al., 1989).

Blood and Urine Samples
Blood and urine samples were collected daily approximately one hour before feeding beginning at enrollment until parturition.Blood and urine samples were collected postpartum at D1, D2, D3, D6, D9, D12, D15, D18, D21, D35, and D49.Blood samples were collected from the coccygeal vessel using 20-gauge Vacutainer needles (Greiner Bio-One GmbH; Exelint International, Co.).Whole blood was collected into 10-mL BD Vacutainer Serum (Becton, Dickinson, and Company) and 10-mL Lithium Heparin 158 USP Units (Becton, Dickinson, and Company) blood collection tubes and inverted gently.Samples from whole blood collected into lithium heparin tubes were analyzed for glucose concentrations using a hand-held biochemical analyzer (VetScan i-Stat, Abaxis).Blood samples collected into serum tubes were kept at room temperature and allowed to clot for 1 h before centrifugation.Serum and plasma were harvested after centrifugation at 3,000 x g for 20 min at 4°C.Samples were allocated into trip- licate aliquots and stored at −80°C until further analysis.Urine samples were collected into 5 mL Eppendorf by gently stimulating the area between the udder and the vulva.An electronic pH meter (Horiba LaquaTwin Compact PH Meter) was calibrated immediately before each use, and urine pH was measured daily to verify DCAD effectiveness and make dietary adjustments as needed when urine pH average increased above 6.5.

Colostrum, Milk Components, and Lactation Performance
Only cows that were milked within 6 h after parturition were included in the colostrum analyses, which resulted in a total of 89 cows (CON = 28, DCAD = 28, and XZ = 30).Calves were weighed by the staff at the calf unit before feeding colostrum.Colostrum quality is known to decrease with longer lag times from collection relative to calving (Morin et al., 2010;Sutter et al., 2019).Six hours was chosen as a cut-off based on the management system of the dairy farm and this is consistent with a decrease not higher than 25% on IgG concentrations in colostrum based on previous research (Morin et al., 2010).Before milking, 20 IU of oxytocin were given to all cows, teats were pre-dipped and then dry-wiped using a clean towel.After milking, colostrum weight was measured using a digital scale (CPWplus 75, Adam Equipment Inc.).Colostrum quality evaluation was performed using a digital Brix refractometer (Reichert, Brix/RI -Chek, Reichert Technologies), and 2 composite colostrum aliquots were collected.One sample was refrigerated at 4°C for colostrum composition analysis and sent to AgSource Laboratories (Full Service National Dairy Herd Improvement Association Certified Laboratory) that uses MilkScan FT+ and MilkoScan 7RM FTIR instruments.The other was frozen at −20°C for additional analyses (IgG, total Ca, and Al).Daily milk yield data for each cow was collected and recorded in Dairy Comp 305 (Valley Agricultural Software) by the farm staff until D49, and milk samples were obtained by the farm staff during both morning and evening milking on a weekly basis until D21.These samples were then averaged for component analysis.Energy-corrected milk (ECM) yield was calculated as [(0.3246X milk yield) + (12.86 X fat yield) + (7.04 X protein yield)] (NRC, 2001).Daily values were averaged into weekly means for statistical analysis.

Laboratory Analysis
Colostrum samples were used to measure IgG concentrations using a commercial Bovine IgG ELISA kit (Bethyl Laboratories Inc.) as previously described (Gelsinger et al., 2015).Briefly, colostrum samples were diluted 1:500,000 in a dilution buffer.The bovine IgG standards were prepared and used according to the manufacturer's instructions, ranging from 0.69 to 500 ng/mL.A reference sample was analyzed to assure the integrity of the assay, as well as all samples were analyzed in duplicates.The inter-assay and intra-assay coefficient of variation (CV) for the IgG assay was 11.9% and 7.06%, respectively.
To determine total calcium concentrations (tCa) in the colostrum, 1.25 g of colostrum was placed in a glass tube to be digested with a combination of nitric and perchloric acids.Samples were first digested at 90°C for 90 min and then at 200°C for 150 min.Next, samples were diluted with deionized water, the volume was brought up to 25 mL and then diluted with a lanthanum chloride solution to prevent other ion interference (AOAC International., 2012;methods 2.110b, 2.112, 2.113;Weaver et al., 2016).All samples were analyzed in duplicates using atomic absorption and the average percentage for the intra-assay CV was 5.43% using (Laporta et al., 2015).Aluminum (Al) concentration in colostrum was analyzed in a subset of cows (CON = 10, DCAD = 10, and XZ = 10) by Michigan State University Veterinary Diagnostic Laboratory.

Statistical Analyses
To evaluate the effect of the 3 different prepartum dietary treatments on colostrum quantity (weight) and colostrum quality (IgG, BRIX, Ca, and nutrient composition) the MIXED procedure of SAS (version 9.4, SAS Institute Inc.) was used.The cows were blocked by lactation number (lactation 2 (Lact2) vs lactation >3 (Lact3+)) and expected parturition date.Treatment and lactation were considered a fixed effect.Data were analyzed for normality, and when the assumption failed, data was transformed.Data is presented as LSMeans ± SEM or median and interquartile for data that were not normally distributed.Transformations were based on diagnostics plots and overall model fit.If transformations were necessary, analysis was performed and the P-value is shown.Statistical significance was declared if P ≤ 0.05, with tendencies for differences declared at 0.05 < P ≤ 0.10.Correlation comparison was performed in R (R Core Team, 2019).Spearman's correlation on residuals was conducted on Brix values, IgG content, time to collect colostrum after parturition, and colostrum weight.Values between ± 0.5 and ± 1.0 were considered strongly correlated, values between ± 0.30 and ± 0.49 moderately correlated, and values between 0 and ± 0.29 lowly correlated.
Prepartum and postpartum data for rumination, BW, BHB, and glucose were analyzed separately.Prepartum data was restricted from D-14 until D0 (parturition).A baseline measurement (26.45 ± 4.37 d before calving; mean ± SD) was determined using the first sample collected from each cow before the start of treatment and was used as a covariate for the respective analyses.A subset of cows was selected to analyze glucose concentrations, as measurements were collected from iStat analyzer (Abbott Laboratories) at time points when iCa was measured, cows were only included in the analysis if they had their first measurement before D-4 and were on the dietary treatment for a minimum of 21 d and a maximum of 28 d, resulting in 89 cows (CON: n = 30; DCAD: n = 31; XZ: n = 28).To measure the change in back fat thickness, BCS, and BW the difference from the start of dietary treatment to the day of parturition and the difference from the day of parturition to D49 were evaluated.This resulted in 2 new variables created: delta 1 and delta 2. Delta 1 represents the change in these parameters between the start of dietary treatment to calving (D0 -Prepartum Enrollment).Delta 2 represents the change in these parameters between weight after parturition (D0) to D49 (D49 -D0).All statistical analyses were conducted using the MIXED procedure of SAS (version 9.4, SAS Institute Inc.).Fixed terms in the model for rumination, BW, and BHB concentration were treatment, day, lactation, covariate, the interaction of treatment x day, and interaction of treatment x lactation.Fixed terms in the model for glucose concentration were treatment, day, lactation, the interaction of treatment x day, and interaction of treatment x lactation.Fixed terms in the model for delta 1 and 2 were treatment, lactation, and the interaction of treatment x lactation.Two separate analyses were used to fit the proper covariance structure per sampling day.Day was considered the repeated measure in both analyses.The first analysis included the prepartum period with daily sampling and, to account for autocorrelated errors, the ar(1) structure was utilized.Due to different sampling timeframes the analysis in the postpartum period the spatial power structure was utilized as a covariance structure.The random statement in all models included cow.Mean comparisons were adjusted by the Student's t-test.Data were analyzed for normality, and when the assumption failed, data was transformed.Transformations were based on diagnostics plots and overall model fit.Statistical significance was declared if P ≤ 0.05, with tendencies for differences declared at 0.05 < P ≤ 0.10.

RESULTS
Gestation length, dry period length, days on treatment diets, number of cows according to parity, 305 d milk produced in the previous lactation, days in milk in the previous lactation, time of colostrum milking, an average of calf weight, and number of calves per sex are presented in Table 1.There was no difference due to treatment in dry period length, parity, previous milk production, previous days in milk, time to the collection of colostrum after parturition, and calf weight between dietary treatments (P > 0.05; Table 1).Cows on the XZ diet had longer gestation length than cows on CON and DCAD diets (277.1 ± 0.5, 277.5 ± 0.5, and 281.0 ± 0.5 d, respectively; P < 0.01; Table 1).

Frizzarini et al.: MECHANISMS BY WHICH…
Whole blood glucose concentrations were affected by day and lactation during the prepartum period (P < 0.05; Table 2).A day effect was observed with decreasing glucose concentrations as cows approached parturition and a spike in glucose concentration at calving (Table 2).Cows in Lact2 had higher glucose concentrations relative to cows in Lact3+ during the prepartum period (79.1 ± 0.7 mg/dL and 77.9 ± 0.6 mg/dL, respectively).During the postpartum period, glucose concentrations were affected by day and an interaction between treatment and lactation was observed (P < 0.05; Table 2).Glucose concentrations peaked at D1 and steadily declined to hit the lowest concentration at D3 (Table 2).Cows fed the DCAD diet and in Lact3+ had the highest glucose concentration relative to all other groups during the postpartum period (66.4 ± 0.6 mg/dL).Cows in Lact3+ and fed XZ had the lowest glucose concentration during this period (63.0 ± 0.6 mg/dL).

Back Fat Thickness, Body Condition Score, and Body Weight
Table 3 presents the BW of cows that were fed the 3 prepartum diets.An effect of treatment was observed prepartum (P = 0.03), with cows fed the CON diet weighing the most, DCAD fed cows being of intermediate weight, and XZ fed cows weighing the least (830 [771,916], 837 [764,898], and 804 [748, 866] kg, respectively).No effect (P > 0.05) of treatment was observed during the postpartum period on BW.A parity effect was observed pre and postpartum, with cows in Lact3+ weighing more than cows in Lact2, both prepartum and postpartum (P < 0.01).

Colostrum
Average colostrum weight across all cows was 5. 90 [3.31, 8.16] kg.No effect (P > 0.05) of treatment, lactation, or time to collect colostrum were observed on colostrum weight (Table 4).The average time to collect colostrum after parturition was 95 [49,189] minutes.There were no differences between treatments (P > 0.05) due to colostrum collection time, and there was no association of colostrum collection time with either Brix percentage or colostrum IgG concentrations (P > 0.05).

Milk Yield and ECM
Milk yield for the first 7 weeks of lactation is presented in Figure 2. Average milk yield during the first 7 weeks of lactation was not different (P > 0.05) due to prepartum dietary treatment (48.27 ± 0.96, 46.73 ± 0.92, and 48.20 ± 0.94 kg for CON, DCAD, and XZ respectively).However, an overall treatment by lactation interaction (P = 0.03; Figure 3A) was observed for milk production.In Lact3+, cows fed XZ produced 51.00 ± 1.07 kg/d, cows fed CON produced 48.17 ± 1.05 kg/d, and DCAD fed cows produced 46.28 ± 1.08 kg/d.In Lact2, cows fed CON produced 48.38 ± 1.61 kg/d of milk, while cows fed DCAD produced 47.18 ± 1.49 kg/d of milk and cows fed XZ produced the least amount of milk at 45.39 ± 1.55 kg/d.Cows that were Lact 3+ fed XZ had increased milk production compared with Lact3+ DCAD fed cows (P = 0.002), but neither were different from the Lact 3+ CON cows (P > 0.05).Cows fed XZ that were Lact 3+ produced more milk than Lact 2+ XZ fed cows (P = 0.003), but there were no other differences between treatment groups.
The ECM for the first 3 weeks of lactation is presented in Figure 3B.Prepartum dietary treatment did not impact (P > 0.05) ECM postpartum (48.48 ± 1.21, 45.74 ± 1.16, and 46.54 ± 1.18 kg/d for CON, DCAD, and XZ respectively).However, there was a significant week effect on ECM (P < 0.01), with the highest ECM occurring in the third week of lactation, the second week of lactation had an intermediate ECM value, and the first week of lactation had the lowest ECM value (53.98 ± 0.81, 50.49± 0.81, and 36.30± 0.83 kg/d, respectively).

DISCUSSION
Currently, feeding negative DCAD diets or supplementing synthetic zeolite A are nutritional strategies implemented during the close-up period to mitigate hypocalcemia.Previous work has established that HC is linked to increased incidences of metabolic disorders such as ketosis and displaced abomasum (Shaver, 1997;Venjakob et al., 2021).Further, research has established that depressed DMI associated with the close-up  period can also contribute to NEB and result in ketosis (Shaver, 1997).Previous research has demonstrated that cows that fed a negative DCAD diet during the close-up period have depressed DMI (Charbonneau et al., 2006;Lopera et al., 2018;Martinez et al., 2018;Zimpel et al., 2018;Lean et al., 2019), which is similarly observed in cows supplemented with zeolite A (Thilsing et al., 2006;Pallesen et al., 2008;Grabherr et al., 2009;Kerwin et al., 2019).Therefore, we investigated the effects of feeding a positive DCAD, negative DCAD diet, or synthetic zeolite A supplement during the close-up period on energy metabolism, as well as colostrum and milk production in multiparous Holstein cows.We have demonstrated that synthetic zeolite A acts to improve calcium metabolism largely through binding and excretion of P into the feces (Frizzarini et al., in review).
The reduction in DMI when supplementing the close-up diet with synthetic zeolite A has been wellestablished in previous studies (Thilsing et al., 2006;Pallesen et al., 2008;Grabherr et al., 2009).In the present study, we also observed a decrease in DMI during the prepartum period among cows fed the XZ diet.The reduction in DMI was found to be 15.71% and 13.01% when compared with the CON and DCAD diets, respectively.Additionally, we found a decrease of 14.12% and 10.98% in DMI as a percentage of BW compared with the CON and DCAD diets, respectively.This coincided with an observed difference of 8.30% and 7.63% lower rate of rumination in XZ fed cows compared with cows on CON and DCAD diets.
In cows supplemented with synthetic zeolite A, there was a reduction in serum phosphorus (P) levels during the prepartum feeding period, as presented in the companion paper (Frizzarini et al., 2023).This decrease was observed as synthetic zeolite A binds to dietary P at low pH, rendering the dietary P unavailable for absorption (Thilsing et al., 2006;Kerwin et al., 2019), inducing a P restriction.A well-known symptom of P deficiency in various species, including cattle, is a decrease in feed intake or anorexia (Fuller et al., 1976;Knochel, 1977;Valk and Šebek, 1999).Studies have shown that P-deprived cows exhibited a decline in DMI during lactation when   10 [83.76, 322.29] 226.80 [95.46, 414.50] 181.36 [77.78, 351.54] 0.87 Protein (%) 14.65 [11.14, 16.85] 15.87 [12.55, 17.24] 15.48 [12.70, 16.95] 0.33 Protein (g) 726.24 [356.36, 986.54] 903.27 [503.75, 1235.49] 709.54 [321.97, 1248.62Colostrum Al concentration was analyzed in 30 samples (CON: n = 10; DCAD: n = 10; XZ: n = 10).a, b, c Letters differ when treatment means differ significantly (P ≤ 0.05). 3Colostrum was analyzed from cows that gave birth to 42 females (47%) and 47 males (53%).their dietary P was manipulated.For example, cows deprived of P (2.3 g P/kg of DM) from 3 weeks prepartum through the ensuing lactation began reducing their DMI in the second week of lactation (Puggaard et al., 2014).Similarly, cows that experienced moderate P deprivation (2.4 g P/kg of DM) over almost 2 lactations showed a decrease in DMI during the first dry period and second lactation (Valk and Šebek, 1999).In the experiment conducted by Puggaard and colleagues (2014), the average P concentration in plasma during prepartum was around 1.70 mmol/L, while in the pres-ent study, it was around 1.00 mmol/L, as presented in the companion paper (Frizzarini et al., 2023).This difference in blood P concentration could be the reason that Puggard et al. (2014) did not find an effect on DMI during the prepartum period, while in the present study, DMI and rumination decreased.Interestingly, no treatment effect was found on rumination during the postpartum period when the diets were the same for the 3 groups.Therefore, it remains unclear whether the decrease in DMI resulting from P deprivation is influenced by the duration of P deprivation or the proximity to parturition.The literature suggests that during P-deprivation, adding P to the diet alleviates the effects of the P removal from the diet (Grünberg et al., 2019).The decrease in feed intake may be associated with P deficiency because this mineral is necessary for ruminal microbial growth (Feng et al., 2015).However, anorexia has also been observed in non-ruminant species under dietary P restriction (Breves and Schröder, 1991) which suggests the reduced feed intake could be the metabolic disruption of the central nervous system during P deprivation, which impairs the synthesis of neurotransmitters (Grünberg et al., 2019).
In contrast to previous literature (Charbonneau et al., 2006;Zimpel et al., 2018;Lean et al., 2019) that reported a decrease in DMI among cows fed a negative DCAD compared with those fed positive DCAD diets during the prepartum period, our study did not observe a significant treatment effect on DMI, DMI as a percentage of BW, or rumination between cows fed DCAD versus CON diets.A meta-analysis conducted to analyze the response of prepartum dairy cows to lowering DCAD did show that reducing the DCAD in  the diet resulted in a reduction in DMI (Charbonneau et al., 2006).However, the exact reasons for this decrease in DMI are not consistent across the studies.Some proposed explanations include the low palatability of the anion source (Oetzel and Barmore, 1993) or the discomfort caused by metabolic acidosis (Vagnoni and Oetzel, 1998).Zimpel et al. (2018) showed that a negative DCAD diet induced a mild metabolic acidosis and led to a decreased DMI, but when the same diet was supplemented with alkalogenic salts, resulting in a positive DCAD diet, they did not observe a decrease in DMI.The authors then concluded that the reduction of DCAD is due to the metabolic acidosis that occurs when feeding DCAD diets leading to reduced DMI (Zimpel et al., 2018).
In the present study, we did not observe any significant effect of the dietary treatment during the prepartum period on BHB concentrations.Interestingly, regardless of treatment, BHB concentrations in all treatments remained below the threshold for subclinical ketosis (1.2 to 1.4 mmol/L of serum BHB; Ospina et al., 2010), which is the threshold associated with increased health risk and reduced milk production (Figure 1).This is supported by a lack of differences in BCS and backfat thickness across treatments.Interestingly, the least amount of BW was lost in the postpartum period in XZ, despite having the largest loss prepartum.Further, this was observed despite the decrease in DMI among cows fed the XZ diet compared with those fed the CON and DCAD diets during the prepartum period.These findings indicate that all the cows enrolled in this experiment had a low risk of developing diseases such as DA, metritis, clinical ketosis, and subclinical metritis, as indicated by subclinical ketosis metrics (LeBlanc et al., 2005;Duffield et al., 2009).Furthermore, despite the decrease in prepartum DMI for cows receiving the XZ diet, the cows that received the XZ diet and in Lact3+ produced more milk than the CON or the DCAD fed cows.This suggests that improving blood Ca concentrations, as observed in our companion study, and reducing the incidence of SCH and CH at the onset of lactation may be a pivotal piece for successful transitions and production in older cows (W.Frizzarini, unpublished data).Additionally, the supplementation of synthetic zeolite A compared with the negative DCAD diet during the prepartum period may provide advantages for populations of older cows.These findings highlight the potential benefits of the supplementation of synthetic zeolite A compared with CON or DCAD fed cows, and the improvement in calcium homeostasis in older cows does not compromise milk production or increase the risk of metabolic disorders, even in cases of decreased prepartum DMI.
In addition to focusing on postpartum energy metabolism in response to the prepartum diet, we were interested in the effects of synthetic zeolite A and negative DCAD on colostrum quantity and quality.While we found no effects of prepartum dietary treatment on colostrum quantity, IgG concentrations were increased by synthetic zeolite A supplementation compared with DCAD and CON fed cows.Colostrum quality and composition has been previously demonstrated to be affected by factors including prepartum diet, time of the year, sex of the calf, calving ease, and milk yield in the previous lactation (Borchardt et al., 2022).In this study, the length of the dry period was similar between prepartum dietary treatments and cows only calved between late December and early May, suggesting that the aforementioned factors were not responsible for the observed increase in IgG concentrations in the synthetic zeolite A supplemented cows.Additionally, previously published work has demonstrated that colostral IgG concentrations are relatively stable for the first 6 h postpartum and all of our analyzed samples were collected in this timeframe (Morin et al., 2010;Conneely et al., 2013).Only a small study of 13 animals per group revealed a difference in colostral IgG concentrations between 2 and 6 h postpartum (Moore et al., 2005).Further, colostrum weights were within previously described ranges and were only affected by the time it took to collect colostrum and the sex of the calf (Mann et al., 2016;Sutter et al., 2019).It is well-established that there is a negative correlation between colostrum quantity and quality.This is because the mammary gland is already transitioning to mature milk production post-parturition, as the time taken to harvest the colostrum still impacted colostrum weight, but not colostrum quality.This observation was still observed despite analysis only including cows that had colostrum harvested within 6 h of parturition.
In conclusion, feeding XZ resulted in decreased DMI and rumination time prepartum, and this was not observed in the negative DCAD or CON diets.However, feeding XZ prepartum did not result in increased body fat mobilization postpartum or changes in energy balance and metabolic analysis.In addition, cows supplemented with synthetic zeolite A had increased IgG concentrations in their colostrum and older cows (Lact3+) had increased milk production during the first 49 DIM when fed XZ during the close-up period compared with the CON and negative DCAD diets.These findings suggest that there may be specific positive implications for feeding XZ to mature cows in terms of colostrum and milk production and it may be derived through the improvements of blood Ca concentrations observed in our companion study during the transition period and thus collective cow health.While this study was initially designed to assess maternal metabolism, it revealed additional intriguing results in relation to older cows and colostrum composition.Future studies should be specifically designed to investigate the possible impacts of feeding XZ prepartum on colostrum quality and composition, as well as impacts on older cows, compared with DCAD and CON.

Frizzarini
et al.: MECHANISMS BY WHICH… Frizzarini et al.: MECHANISMS BY WHICH…

Frizzarini
et al.: MECHANISMS BY WHICH…

FrizzariniFigure 1 .
Figure1.Median and interquartile for prepartum and postpartum BHB concentrations (mM) for multiparous Holstein cows fed 1 of 3 prepartum diets: control (CON), negative DCAD (DCAD), or supplemented with synthetic zeolite A (XZ) during the close-up period.Prepartum and postpartum data were analyzed separately.There was a day (P < 0.01) and lactation (P < 0.01) effect during the prepartum period.There was a day (P < 0.01) and lactation (P = 0.05) effect during the postpartum period.

Frizzarini
et al.: MECHANISMS BY WHICH…

Frizzarini
et al.: MECHANISMS BY WHICH…

Frizzarini
Figure 2. Least squares means ± SE for weekly milk yield of multiparous Holstein cows fed 1 of 3 prepartum dietary treatments: control (CON), negative DCAD (DCAD), or with supplementation of synthetic zeolite A (XZ) diet during the close-up period.

Figure 3 .
Figure 3. Least squares means ± SE for (A) milk production (kg) according to lactation number and (B) ECM (kg) according to the week of lactation for multiparous cows fed 1 of 3 prepartum dietary treatments: control (CON), negative DCAD (DCAD), or supplementation of synthetic zeolite A (XZ) during the close-up period.An effect of the week (P < 0.01) was observed for ECM.There was a week effect (P < 0.01) and treatment by lactation effect (P = 0.03) for milk yield.a, b Letters differ when treatment by lactation means differ significantly (P ≤ 0.05).

Table 1 .
Least squares means ± SE or median [interquartile] for production variables of multiparous Holstein cows fed experimental prepartum diets beginning 254 d of gestation

Table 3 .
Average and least squares means ± SE or median [interquartile] for prepartum and postpartum BW, BCS, and backfat thickness of multiparous Holstein cows fed experimental prepartum diets beginning 254 d of gestation

Table 4 .
Least squares means ± SE or median[interquartile]for colostrum yield, brix, colostrum composition, and colostrum fatty acids in 89 multiparous Holstein cows fed experimental prepartum diets beginning at 254 d of gestation