If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The objective of this study was to assess the effects of feeding negative dietary cation-anion difference (DCAD) dry cow diets on postpartum health. Cows from 4 commercial dairy farms in Ontario, Canada, were enrolled in a pen-level controlled trial from November 2017 to April 2019. Close-up pens (1 per farm), with cows 3 wk before expected calving, were randomly assigned to a negative DCAD [TRT; −108 mEq/kg of dry matter (DM); target urine pH 6.0–6.5] or a control diet (CON; +105 mEq/kg of DM with a placebo supplement). Each pen was fed TRT or CON for 3 mo (1 period) then switched to the other treatment for the next period, with 4 periods per farm. Urine pH was measured weekly until calving, and body condition score (BCS) was measured at enrollment and at 5 wk postpartum. Data from 15 experimental units [8 TRT and 7 CON, with 1,086 (TRT: n = 681; CON: n = 405) observational units (cows)] that received the assigned diet for >1 wk were included. The incidence of milk fever (MF), retained placenta (RP), metritis, hyperketonemia (blood β-hydroxybutyrate >1.2 mmol/L, measured weekly in wk 1 and 2), clinical mastitis within 30 DIM (MAST), displaced abomasum (DA) within 30 d in milk (DIM), purulent vaginal discharge (PVD, assessed once at wk 5), and number of disease events (≥1 or ≥2) were analyzed with logistic regression models with treatment, parity, BCS, and their interactions, accounting for pen-level randomization and clustering of animals within farm with random effects, giving 10 degrees of freedom to test treatment effects. Multiparous cows fed TRT had greater blood calcium between 1 and 4 DIM than multiparous cows fed CON, and the prevalence of subclinical hypocalcemia (total Ca ≤2.14 mmol/L) was lesser when fed TRT compared with CON (d 1: 73 ± 6% vs. 93 ± 4%; d 2: 65 ± 7% vs. 90 ± 5%), with no differences between treatments detected in primiparous cows. We detected interactions of treatment and BCS at enrollment for MF in multiparous cows and of treatment and parity for ≥2 disease events. Overconditioned (BCS ≥3.75) multiparous cows had reduced incidence of MF when fed TRT (TRT: 2 ± 1%, vs. CON: 13 ± 8%). We detected no treatment effects on RP, metritis, hyperketonemia, or PVD incidence. Cows fed TRT had lesser incidence of DA (1.7 ± 0.7% vs. 3.6 ± 1.6%) and tended to have lesser incidence of MAST compared with CON (1.8% ± 0.6% vs. 4.4 ± 1.4%). No treatment effect was detected on ≥1 disease events (TRT: 38 ± 7%, vs. CON: 42 ± 8%); however, multiparous cows on TRT were less likely to have ≥2 disease events than cows on CON (14 ± 4% vs. 23 ± 6%). Under commercial herd conditions, feeding prepartum diets with negative DCAD improved several measures of postpartum health.
The transition period, defined as 3 wk before and after parturition, is the subject of much research in the dairy industry because 30 to 50% of dairy cows experience at least 1 disease event (
). Many factors, including nutrition, environment, management, and metabolic challenges, contribute to the high incidence of postpartum disease. At the beginning of lactation, Ca metabolism faces a substantial, rapid challenge, leading to a variable degree and duration of low blood Ca concentration postpartum. Subclinical hypocalcemia (SCH) affects at least 50% of multiparous cows (
Because Ca plays an important role in muscle contraction, immune function, and nerve impulse transmission, it is expected that negative DCAD before calving would decrease postpartum disease incidence. Negative DCAD diets prepartum are among the nutritional strategies for prevention of hypocalcemia. Although its mechanism is not completely understood, it is believed that the mild metabolic acidosis caused by negative DCAD increases the sensitivity of tissues to the effects of parathyroid hormone. Active vitamin D availability is parathyroid hormone-dependent, and both hormones enhance Ca homeostatic mechanisms in the kidney, intestine, or bone (
reported that reducing the DCAD from +200 to −100 mEq/kg of DM in prepartum diets increased serum total calcium (tCa) concentration at parturition from 1.86 to 2.04 mmol/L.
fed 80 cows positive or negative DCAD diets prepartum (+130 vs. −130 mEq/kg of DM) and reported a milk fever (MF) incidence of 23.1 and 0%, respectively, and at least 30% reduction in SCH (using thresholds of plasma tCa ≤2.0 or ≤2.15, or ionized Ca ≤1.00 mmol/L within 3 DIM) in cows fed the negative DCAD diet. The reductions in the incidence of some diseases were supported by a recent meta-analysis on DCAD diets (
). With DCAD differences between control (CON) and treatment (TRT) groups of >205 mEq/kg of DM, reductions in retained placenta (RP; risk ratio = 0.6), metritis (odds ratio = 0.5), and MF (risk ratio = 0.6) were observed in the low DCAD groups (
). Studies reported in the meta-analyses included 96 to 84 cows per trial, so generally were not designed to detect effects on the risk of clinical disease. A need exists to perform controlled trials including multiple farms under field conditions, to provide results that can be generalized to commercial farms. Considering that the measurable effects of preventive practices are influenced by management, genetics, and environment, this work makes a novel contribution to a better understanding of the effects of negative DCAD diets in commercial settings.
The objective of this study was to assess the effects of a negative DCAD dry cow diet on early postpartum health on commercial dairy farms. With the expectation that negative DCAD diets enhance postpartum blood Ca concentration, we hypothesized that an acidogenic diet would decrease incidence of postpartum diseases in early lactation.
The effect of prepartum negative dietary cation-anion difference and serum calcium concentration on blood neutrophil function in the transition period of healthy dairy cows.
Controlled trial of the effect of negative dietary cation-anion difference prepartum diets on milk production, reproductive performance, and culling of dairy cows.
). The experimental protocol was reviewed and approved by the University of Guelph Animal Care Committee (Animal Utilization Protocol 3951). “Primiparous” refers to cows that were enrolled as nulliparous (3 wk before calving) and, at calving, initiated their first lactation; “multiparous” are cows that started their second or greater lactation.
Experimental Design and Treatment Groups
This pen-level randomized controlled trial was conducted on 4 commercial freestall dairy farms in Ontario, Canada. A convenience herd selection was made based on geographic location, herd size, and willingness to comply with the experimental protocol, which included not administering any prophylactic calcium supplements around calving. The study herds are described in Table 1. We collected the present data between November 2017 and April 2019. Data from each farm's computerized records (farms A, B, and C: Dairy Comp 305, Valley Ag Software; farm D: DairyPlan C21, GEA) were collected weekly.
Table 1Description of 4 commercial farms enrolled in a pen-level controlled trial of a negative DCAD diet for close-up dry cows
The single close-up pen on each farm was randomly assigned to a starting treatment, which was alternated approximately every 3 mo for a total of 4 periods (2 treatment and 2 placebo). The 2 treatments corresponded to a negative DCAD TRT [Soychlor, Landus Cooperative; −108 mEq/kg of DM (weighted average of the 4 farms); target urine pH 6.0–6.5] or a CON diet (+105 mEq/kg of DM, weighted average, with a placebo supplement) nutritionally similar to TRT but with a positive DCAD (Table 2). Treatments were added to the TMR mixer with other ingredients. The supplement was delivered in plain moisture-barrier bags labeled A or B (the only noticeable package difference) to blind farm staff to the treatment. The feeding rate of TRT was adjusted if necessary, as will be described, to achieve the targeted urine pH. The DCAD level of the CON group was set similar to the DCAD in each herd's diet before the study. Expected calving date was defined as 280 d after the last breeding date, and cows were followed until 60 DIM. When a pen's treatment was changed, there was a “washout” period of at least 8 d (that is, to be included in the analyses, cows had to be in the pen receiving the same assigned diet for at least 2 consecutive weekly farm visits). This exclusion criterion was based on the expectation that compensated metabolic acidosis would be achieved within 5 d (
), and allowing an additional margin of error. Therefore, cows were excluded if they were exposed to their assigned experimental diet for less than 8 d before parturition, if any Ca supplement was given as prevention for MF (a deviation from the trial protocol), if the cow was intractable for sampling, or if the cow was culled before calving (Table 3).
Table 2Ingredient composition (% of DM) and nutrient profile of prepartum diets for commercial farms in a pen-level controlled trial of a negative DCAD diet in the close-up dry period (CON = control diet; TRT = treatment diet)
Table 3Description of animals enrolled, reasons for exclusion, and animals included in the analyses in a pen-level controlled trial of a negative DCAD diet for close-up dry cows
Item
Farm A
Farm B
Farm C
Farm D
Total
Enrolled
427
335
423
183
1,368
Excluded
78
68
115
21
282
Calcium supplement given
2
6
0
1
9
Less than 8 d on assigned treatment
45
38
89
17
189
Culled before calving
16
4
3
1
24
Enrolled but had not calved by the end date of the trial
Farms planned to, on a weekly basis, move cows to the close-up pen 3 wk before parturition. However, producers did not record the exact date that each cow was moved. At enrollment (3 wk before expected parturition) and 5 wk postpartum, BCS of each cow was assessed on a 1-to-5 scale, measured in 0.25-point increments (
). To maintain blinding of the farmers, urine pH was measured weekly in all close-up pens, irrespective of treatment. Perineal stimulation was performed, and mid-stream urine samples were collected into a 20-mL plastic collection vial to measure pH with a portable meter (Laquatwin, Horiba Ltd.). If the average pH in a TRT pen was out of the target range of 6 to 6.5 for more than 2 wk, adjustments of the feeding rate were made. On 2 occasions an adjustment was required, with 2 farms making 1 increase each in the amount of TRT supplement. Urine pH and BCS assessments were performed by the same research assistant, who was not blind to the treatment, given the results of monitoring urine pH, but who did not participate in disease diagnosis. Urine pH results were not made known to the farmers.
On each farm, the close-up group diet was delivered daily (between 0800 and 0900 h). One dry cow TMR sample was collected from each farm for each TRT and CON diet (8 samples in total). The nutrient contents of the close-up TMR were analyzed by a commercial laboratory (Agri-Food Laboratories, Guelph, Ontario, Canada). Apart from the close-up DCAD level between treatments, the pre- and postpartum TMR were as similar as possible for cows within a farm and among the farms. Descriptions of the pre- and postpartum diets, ingredients, and chemical analyses are presented in Table 2, Table 4, respectively. Diets were formulated using feeds available or normally purchased on each farm. The target DM intake for close-up cows was 14, 13, 13, and 14 kg/cow/d for farms A, B, C, and D respectively.
Table 4Ingredient composition (% of DM) and nutrient profile of postpartum diets for commercial farms in a pen-level controlled trial of a negative DCAD diet in the close-up dry period
The sample size was dictated by logistical and financial constraints. Our primary outcomes were the incidence of clinical disease, which for MF in multiparous cows, displaced abomasum (DA), and clinical mastitis (MAST) were expected to have lactational incidence risks in CON between 5 and 10%. For context, detection of a decrease in the incidence of DA (a binary outcome) from 6% to 3% under cluster randomization (i.e., a pen-level study, as here), with a significance level of 5%, 4 periods with crossovers of treatment, an average cluster size of 50 (i.e., the number of cows per pen), and assuming an intraclass correlation coefficient of 0.05 (minimum of 0.01 and maximum of 0.10), 4 clusters per treatment (per period) are need to achieve >80% power. With 2 clusters per treatment, as here, power under these assumptions was ~55% (
A tutorial on sample size calculation for multiple-period cluster randomized parallel, cross-over and stepped-wedge trials using the Shiny CRT Calculator.
To assess blood Ca concentration, whole blood was collected from the coccygeal vessels during 2 different periods, corresponding to 12 experimental units (pen-treatments) in 2 subsets: (1) January to May 2018: n = 6 pen-treatments (9 ± 4 cows per pen-treatment; 57 cows), blood collected at d 1, 2, and 3 after parturition; and (2) November to December 2018, when sampling was extended based on emerging published data: n = 6 pen-treatments (14 ± 11 cows per pen-treatment; 83 cows), blood collected at d 1, 2, 3, 4, 5, 6, and 7 after parturition. These sampling periods were based on the availability of labor for daily sampling. During those 2 periods, all clinically healthy cows were sampled. Blood was collected into a 10-mL tube without anticoagulant (Vacutainer Precision Glide, Becton Dickinson). Within 3 h of sampling, tubes were centrifuged at 1,500 × g for 15 min, and serum was harvested and stored at −20°C in 2 aliquots until analysis. Total calcium measurement was performed at the Animal Health Laboratory, University of Guelph, using the Cobas Calcium Gen 2 kit (Roche Diagnostics) with a lower limit of quantification of 0.20 mmol/L, and intra- and interassay coefficients of variation of 1.7 and 2.1%, respectively.
Outcomes Assessment
Disease diagnoses were performed by the producers and farm staff for MF, RP, metritis, DA, and MAST and by the research team for hyperketonemia and purulent vaginal discharge (PVD). Purulent vaginal discharge and BHB data were provided to the producers at the end of each visit. All herds used the same disease definitions (provided and explained at the beginning of the trial). Milk fever was defined as weakness, dullness, low body temperature, trembling, or being unsteady on the feet (stage I) or not able to rise (stage II and III) at or in the 2 d after calving. Cows diagnosed with MF were treated according to each farm's protocols. Clinical MAST was defined as visibly abnormal milk (clots, flakes, or watery or bloody) with or without swelling of the udder (confirmed by human observation in the herds with automatic milking systems and by forestripping at each milking in the other herds). Metritis was foul-smelling, brown-red, watery vaginal discharge within the first 20 DIM with or without fever (≥39.5°C). Retained placenta was defined as failure to expel the fetal membranes by 24 h after calving. Displaced abomasum was diagnosed by the farm staff or by a veterinarian. For the diagnosis of PVD, at wk 5 postpartum, a researcher evaluated vaginal discharge with a Metricheck device (Simcro). A positive result for PVD was considered when cows scored 2 or above on a 4-point (scale 0–3) vaginal discharge scale (i.e., ≥50% pus;
). Blood BHB was assessed once per week in the first 14 DIM, and hyperketonemia was diagnosed if BHB ≥1.2 mmol/L (Precision Xtra Meter and β-ketone test strips, Abbott;
Parity (categorical variable: first vs. second or greater) and BCS at enrollment (categorical variable: overconditioned ≥3.75, vs. not overconditioned ≤3.5) were used to describe baseline data of the treatment groups. A chi-squared test was used to confirm similarity between treatment groups. For the TRT group, descriptive data on urine pH were categorized at thresholds of 7 and 6.5. Acid-base status changes Ca metabolism when urine pH ≤7 (
Statistical analyses were performed using SAS, version 9.4 (SAS Institute Inc.), with cow as the unit of observation but with pen as the experimental unit. The binary outcomes analyzed were MF, RP, metritis, ketosis, MAST ≤30 DIM, DA ≤30 DIM, PVD (assessed once at wk 5), or number of disease events (≥1 or ≥2). Cows were included in each analysis if they had the outcome of interest or were not culled before the end of the risk period for each disease. These outcomes were analyzed with logistic regression (GLIMMIX procedure of SAS) models with treatment, parity category, BCS category at enrollment, and their 2-way and 3-way interactions with treatment. Parity and BCS were offered to all models as pre-treatment covariables and removed if their P-value was >0.05. All models specified the error term for treatment with a random effect (farm × treatment × period term) to give the correct denominator degrees of freedom for experimental units in the study (
) and included a second random effect term for farm to account for clustering of animals within farm and unmeasured sources of variance at the farm level. Therefore, the main effect of treatment had 10 degrees of freedom for all models of disease outcomes.
The analysis of MF incidence was performed considering all cows or only multiparous cows, because as expected, no cases occurred in primiparous cows (
). A variable considering the difference in BCS between enrollment and 5 wk postpartum was analyzed with a linear regression model (MIXED procedure of SAS) including treatment, parity, and interaction of parity with treatment (treatment × parity), specifying the experimental unit (15 pen-treatments) and accounting for clustering of animals within farm with random effects (farm × treatment × period and farm, respectively). The logistic regression models are represented by the following equation:
where πijkl represents the probability of a cow at j parity with k BCS on i treatment in having the disease; i = treatment (CON or TRT), j = parity group, k = BCS at enrollment; βi = fixed effect of treatment; γj = fixed effect of parity; θk = fixed effect BCS at enrollment; (βγ)ij = effect of treatment by parity interaction; (βθ)ik = effect of treatment by BCS at enrollment interaction; (βγθ)ijk = effect of treatment by parity and BCS at enrollment interaction; Fl = random effect of farm l; Flip = random effect of farm l × treatment i × period p.
For tCa (continuous variable) and SCH serum tCa (≤2.14 mmol/L, categorical variable), ANOVA and logistic regression with linear and generalized linear mixed models (MIXED and GLIMMIX procedures of SAS), respectively, were built. The Shapiro-Wilk test was used to assess normality of the residuals fit in the mixed linear regression model. No transformations were necessary. Models included as covariables treatment, parity, sampling day, and interactions of parity and sampling day with treatment. Random effects specified the experimental unit (12 pen-treatments) and clustering of animals within farm (farm × treatment × period and farm, respectively). Based on the lowest Akaike information criterion, repeated measures were accounted for with a Toeplitz (2) and autoregressive covariance structure for tCa and SCH, respectively. In the final calcium models, the main effect of treatment had 6 degrees of freedom. A Tukey test was used to adjust for multiple comparisons. The model of total serum calcium is represented by the following equation:
where Yijkm = response to treatment i (CON or TRT) at the kth time (k = sampling time 1, 2, or 3 DIM) for m cow; μ = overall mean; βi = fixed effect of treatment; γj = fixed effect of parity; θk = fixed effect sampling time; (βγ)ij = effect of treatment by parity interaction; (βθ)ik = effect of treatment by sampling time interaction; (βγθ)ijk = effect of treatment by parity by sampling time interaction; Fl = random effect of farm; Flip = random effect of farm × treatment × period; εijklm = residual error within m cow, on treatment i, with j parity at k day; αm(lip) = repeated measure, m cow nested within the Flip effect. Repeated measures (α) were accounted for with a Toeplitz (2) structure based on superior fit to other covariance structures.
Based on the results observed for MF in multiparous cows, we built an additional model to assess the interaction of BCS on total serum Ca concentration. In the model equation previously shown, we included only multiparous cows, and, instead of parity, we included BCS at enrollment as a covariable.
The SCH logistic regression model is represented by the following equation:
where πijkml represents the probability of m cow at j parity in i treatment (CON or TRT) at the kth test day (k = test d 1, 2, or 3) having SCH; βi = fixed effect of treatment; γj = fixed effect of parity; θk = fixed effect of test day; (βγ)ij = effect of treatment by parity interaction; (βθ)ik = effect of treatment by sample day interaction; (βγθ)ijk = effect of treatment by parity by sample day interaction; Fl = random effect of l farm; Flip = random effect of l farm × i treatment × p period; αm(lip) = repeated measures, m cow nested within the Flip effect.
In all models, covariables and interaction terms were removed when P > 0.05 and P > 0.1, respectively. Results are expressed as least squares means with their standard error balanced for the proportions of covariables in the data, where applicable (using the OM option in the LSMEANS statement in SAS). Odds ratios with 95% confidence intervals are presented for logistic regression models.
Deviation from Study Design
The study design included 4 periods of 3 mo each per farm (e.g., TRT-CON-TRT-CON). Farm C started the trial on TRT and after 3 mo was switched to CON. This diet was offered for only 25 d (October 1 to 25, 2018) instead of the 3 mo originally planned. When cows fed CON started calving, early-lactation disease incidence and mortality dramatically increased. Of the 39 cows that calved in this period, 14 suffered from MF, 1 of which died. Because of this unexpected and unacceptable incidence of disease, the second study period (feeding CON) was stopped, and the farm was switched to the third period, with TRT. Because such an occurrence was not anticipated, we had not set out formal trial-stopping rules in advance, but an early-stopping rule was applied. This is a common and required practice in human medicine trials (
), where early evidence of an inferior treatment outcome with harm to the study subjects enrolled justifies study protocol changes. Therefore, on farm C, only 3 periods were applied, resulting in unbalanced experimental units (8 TRT and 7 CON). The disease data were analyzed with and without farm C (data not shown), and because the incidences of disease and the inferences about treatment were similar, data from farm C were included in the final analyses. The reduced number of experimental periods on this farm (3) is accounted for in the statistical models with specification of the experimental units.
RESULTS
A total of 15 experimental units (7 CON and 8 TRT; a total of 1,368 cows from 4 farms) were enrolled in this experiment. After applying the exclusion criteria, 1,086 observational units (360 primiparous and 726 multiparous cows) were included in the statistical analyses (Table 3). The numbers of cows per experimental unit are described in Supplemental Table S1 (https://data.mendeley.com/datasets/k3rwyr3wy2/1,
). At enrollment, 32% (118/360) of primiparous and 30% (215/726) of multiparous cows had BCS ≥3.75, with no difference between treatment groups (P > 0.6). The unadjusted average (± SD) urine pH were 8.1 ± 0.4 and 6.3 ± 0.6 in CON and TRT groups, respectively (Figure 1a). In the TRT group, during the 3 wk before parturition, 85% of urine samples had pH ≤7, and 72% were ≤6.5 (Table 5). Figure 1b represents urine pH distribution in TRT group by farm. The coefficient of variation of urine pH was 5.3, 6.5, 10.9, and 5.3% on farms A, B, C, and D, respectively. Supplemental Table S2 (https://data.mendeley.com/datasets/n7nmt7yyft/1,
) provides the crude risks of postpartum diseases (description only).
Figure 1Descriptive statistics of prepartum urine pH: (a) mean urine pH for control (CON) and treatment (TRT) groups during the 3 wk before partition. Error bars represent SD; (b) box plot graphs of urine pH from cows fed TRT by farm [bars represent maximum and minimum values and, in orange, interquartile range (Q1–Q3) and median].
Table 5Urine pH of samples collected during the study period in cows that received an acidogenic (−108 mEq/kg DCAD) dietary supplement for at least 2 wk before calving
Urine samples were collected weekly from dry cows in the experimental pen and categorized considering 2 pH cut-points. The experimental units were the close-up pen (n = 15 pen-treatments), with 1,086 observational units (cows).
1 Urine samples were collected weekly from dry cows in the experimental pen and categorized considering 2 pH cut-points. The experimental units were the close-up pen (n = 15 pen-treatments), with 1,086 observational units (cows).
An interaction of treatment and parity (P = 0.03) was observed when assessing BCS loss from enrollment to wk 5 postpartum. In multiparous cows, treatment did not have an effect on BCS loss (TRT −0.46 ± 0.02 vs. CON −0.45 ± 0.02-point difference; P = 0.88). In primiparous cows, TRT lost less condition than CON (−0.37 ± 0.02 vs. −0.45 ± 0.03-point difference, respectively; P = 0.03).
Effect of Treatment on Serum Calcium Concentration
Serum tCa at d 1, 2, and 3 was assessed in 12 experimental units (91 multiparous and 49 primiparous cows). When analyzing tCa as a continuous outcome, an interaction between treatment, parity, and sampling time was detected (P < 0.01). Considering serum samples at d 1, 2, and 3 postpartum, multiparous cows fed TRT had greater tCa at d 2 than CON (2.05 ± 0.03 vs. 1.89 ± 0.04 mmol/L; P < 0.01), but no differences were detected at d 1 or 3 (Figure 2a). Primiparous cows had 2.25 ± 0.04, 2.20 ± 0.04, and 2.24 ± 0.03 mmol/L tCa at d 1, 2, and 3, with no difference detected between groups (P > 0.68; Figure 2b). In multiparous cows, we did not detect an interaction of treatment with BCS (P = 0.23).
Figure 2Least squares means (±SE) concentrations of total Ca in serum from d 1 to 3 postpartum of (a) multiparous [n = 40 control (CON), 51 treatment (TRT)] and (b) primiparous (n = 18 CON, 31 TRT) cows fed CON or TRT diets (+105 and −108 mEq/kg DCAD, respectively). The experimental units were the close-up pen (n = 12 pen-treatments; 6 TRT and 6 CON), with 140 observational units (cows). Within a day, * denotes effect of DCAD (P < 0.05).
In the subset of cows with blood sampled for 7 d (n = 6 experimental units; 83 cows), in multiparous cows tCa was greater in TRT than CON from d 1 until d 4 (P < 0.05), with the greatest difference at d 2 (Figure 3a). In primiparous cows, tCa concentration tended to be greater at d 4 (P = 0.08) for cows fed TRT than cows fed CON, with no differences on the other days (Figure 3b).
Figure 3Least squares means (±SE) concentrations of total Ca in serum from d 1 to 7 postpartum of (a) multiparous [n = 22 control (CON), 28 treatment (TRT)] and (b) primiparous (n = 10 CON, 23 TRT) cows fed CON or TRT (+105 and −108 mEq/kg DCAD, respectively). The experimental units were the close-up pen (n = 6 pen-treatments; 3 TRT and 3 CON), with 83 observational units (cows). Within a day, * denotes effect of DCAD (P < 0.05); # denotes a tendency (P < 0.1).
Multiparous cows fed TRT had lesser prevalence of SCH (tCa ≤2.14 mmol/L) than multiparous cows fed CON at d 1 [93 ± 4% vs. 73 ± 6%; odds ratio 4.7 (95% CI 1.2–18); P = 0.03] and d 2 postpartum [90 ± 5% vs. 65 ± 7%; odds ratio 4.9 (1.5–16.3); P < 0.01; Figure 4a]. No differences between treatments were detected in primiparous cows (Figure 4b).
Figure 4Effect of DCAD fed prepartum [control (CON): +105 mEq/kg; treatment (TRT): −108 mEq/kg] on subclinical hypocalcemia (serum total Ca ≤2.14 mmol/L) at d 1, 2, and 3 postpartum for (a) multiparous (CON n = 40, TRT n = 51) and (b) primiparous (CON n = 18, TRT n = 31) cows. Values are LSM ± SE. The experimental units were the close-up pen (n = 12 pen-treatments; 6 TRT and 6 CON), with 140 observational units (cows). Within a day, * denotes effect of DCAD (P < 0.05).
The effect of treatment on health outcomes is summarized in Table 6. In model building, interactions of treatment with BCS were observed for MF in multiparous cows, and with parity for ≥2 disease events. There were 43 cases of MF, all in multiparous cows. An interaction was observed between BCS at enrollment and treatment, such that a lower incidence of MF was detected in overconditioned cows that received TRT, but no difference in MF incidence among cows with BCS ≤3.5 was detected (Table 7). No treatment effects occurred on the incidence of RP, metritis, ketosis, PVD, or disease events (≥1; Table 6). Cows fed TRT had a lower incidence of DA and tended to have a lower incidence of MAST than cows fed CON. Multiparous cows fed TRT were less likely to have ≥2 diseases in the early postpartum period than multiparous cows fed CON, but in primiparous cows no treatment effect on this outcome was detected (Table 7).
Table 6Effect of DCAD fed prepartum on postpartum diseases in a controlled trial of a negative DCAD diet for close-up dry cows
From 3 wk before expected parturition cows were fed control diet (CON, +105 mEq/kg) or treatment diet (TRT, −108 mEq/kg) DCAD. The experimental units were the close-up pen (n = 15 pen-treatments; 8 TRT and 7 CON), with 1,086 observational units (cows). LSM and SE are expressed as percentage (%).
DCAD = effect of level of DCAD (CON vs. TRT); parity = primiparous vs. multiparous; DCAD × parity = interaction between level of DCAD and parity; BCS = not overconditioned (≤3.5) vs. overconditioned (≥3.75) at enrollment (−3 wk); DCAD × BCS = interaction between level of DCAD and BCS.
Stratified LSM and SE are presented in Table 7 where interactions were detected.
—
—
—
—
—
0.75
<0.01
<0.01
—
—
1 From 3 wk before expected parturition cows were fed control diet (CON, +105 mEq/kg) or treatment diet (TRT, −108 mEq/kg) DCAD. The experimental units were the close-up pen (n = 15 pen-treatments; 8 TRT and 7 CON), with 1,086 observational units (cows). LSM and SE are expressed as percentage (%).
2 DCAD = effect of level of DCAD (CON vs. TRT); parity = primiparous vs. multiparous; DCAD × parity = interaction between level of DCAD and parity; BCS = not overconditioned (≤3.5) vs. overconditioned (≥3.75) at enrollment (−3 wk); DCAD × BCS = interaction between level of DCAD and BCS.
3 Stratified LSM and SE are presented in Table 7 where interactions were detected.
4 NA = not applicable; model was built considering multiparous cows only (n = 726).
From 3 wk before expected parturition cows were fed control diet (CON, +105 mEq/kg) or treatment diet (TRT, −108 mEq/kg) DCAD. LSM and SE are expressed as percentage of cows affected (%). The experimental units were the close-up pen (n = 15 pen-treatments; 8 TRT and 7 CON), with 1,086 observational units (cows).
1 From 3 wk before expected parturition cows were fed control diet (CON, +105 mEq/kg) or treatment diet (TRT, −108 mEq/kg) DCAD. LSM and SE are expressed as percentage of cows affected (%). The experimental units were the close-up pen (n = 15 pen-treatments; 8 TRT and 7 CON), with 1,086 observational units (cows).
2 Differences of DCAD × parity or DCAD × BCS LSM (multiple comparisons).
Cows fed diets with negative DCAD had improvements in some measures of postpartum health compared with cows fed a positive DCAD prepartum. In multiparous cows, tCa concentration was greater in the group fed negative DCAD diets in the days after calving, with a reduction in the prevalence of SCH (serum tCa ≤2.14 mmol/L) at d 1 and 2. Overconditioned multiparous cows fed negative DCAD diets had lower incidence of MF than overconditioned multiparous cows fed positive DCAD diets. Cows fed negative DCAD diets prepartum had lower incidence of DA and tended to have lower incidence of MAST in the first 30 DIM than cows fed positive DCAD diets. Multiparous cows fed negative DCAD diets were less likely to have multiple diseases postpartum than multiparous cows fed positive DCAD diets.
Because hypocalcemia is considered a gateway disease, we hypothesized that improving Ca metabolism through an acidogenic diet would decrease postpartum disease incidence. However, we cannot pinpoint the mechanism by which a negative DCAD before calving reduced the occurrence of certain diseases in the present study. Increased serum Ca and greater DMI postpartum in response to prepartum negative DCAD have been documented in a recent meta-analysis (
). These effects might enhance gastrointestinal motility and increase rumen fill, reducing risk factors associated with DA. The tendency for a lesser incidence of MAST in cows fed negative DCAD diets might be explained by improved teat sphincter closure after milking or by enhanced neutrophil function. Greater availability of active vitamin D, secondary to improved parathyroid hormone response, which might occur in cows fed negative DCAD, could also contribute to enhancement of the natural defenses of the mammary gland (
reported increased incidence of MAST with increasing days of exposure to negative DCAD diets prepartum. Our study was not designed to assess the duration of exposure to negative DCAD, and we do not have the data to comment on that variable in this trial.
Although hypocalcemia has been repeatedly associated with impaired immune function (
reported some improvement of neutrophil oxidative burst in cows fed negative DCAD, but in a subset of cows from the present study published elsewhere, no effect on neutrophil oxidative burst or phagocytosis was detected (
The effect of prepartum negative dietary cation-anion difference and serum calcium concentration on blood neutrophil function in the transition period of healthy dairy cows.
). That may explain the lack of effect of TRT on the incidence of RP, metritis, or PVD, for which neutrophil capacity is understood to be an important determinant. Overall, negative DCAD did not reduce the proportion of animals experiencing disease, but it reduced the proportion of animals affected by multiple diseases.
Serum Ca concentration was assessed in 2 data sets, with cows sampled for 3 or 7 d; the former included all of the cows in the latter. Serum Ca concentration was assessed only in clinically healthy cows. Both analyses showed greater serum tCa at d 2 postpartum in multiparous cows fed negative DCAD diets than in multiparous cows fed positive DCAD diets. In the 7-d subset, we observed greater tCa in multiparous cows fed negative DCAD diets each day from 1 to 4 DIM than in multiparous cows fed positive DCAD diets. These samples were collected approximately 8 mo later than those collected only for 3 d, so we speculate that unmeasured changes in diet or cow factors may have contributed to the difference in results. As expected, serum tCa concentration was greater in multiparous cows that received negative DCAD diets. Lowering the mean DCAD by 213 mEq/kg resulted in a tCa differences of +0.19, 0.24, +0.12, and +0.12 mmol/L at d 1, 2, 3, and 4 postpartum, respectively (data from the 7-d subset). This increment of tCa is in agreement with published meta-analyses.
described an increase of + 0.18 at calving and +0.12 mmol/L 24 h after calving in multiparous cows by lowering DCAD from approximately +200 to −100 mEq/kg. We did not measure ionized Ca in this study, for logistical and financial reasons. In primiparous cows, tCa concentration in the negative DCAD group tended to be greater than in the positive DCAD group only at d 4 (+0.12 mmol/L), whereas the meta-analysis by
demonstrated an increase in tCa in primiparous cows of +0.10 mmol/L at calving and +0.08 mmol/L 24 h after calving.
We expected that negative DCAD diets fed prepartum would reduce the incidence of MF in all multiparous cows compared with positive DCAD diets. However, in multiparous cows, we observed an interaction between BCS and treatment, indicating that overconditioned cows (BCS ≥3.75 before parturition) had much greater risk of MF, but that negative DCAD reduced the incidence in this group. Although it would be practically difficult to target acidogenic supplements to high-BCS cows, this difference in effect of DCAD on MF based on BCS is biologically plausible. Excessive body condition is associated with complications around parturition, including MF (
Postpartum body condition score and results from the first test day milk as predictors of disease, fertility, yield, and culling in commercial dairy herds.
Postpartum body condition score and results from the first test day milk as predictors of disease, fertility, yield, and culling in commercial dairy herds.
concluded that pastured cows with BCS ≥3.5 at calving were more likely to have MF. In overconditioned cows, in addition to enhancing Ca metabolism around parturition, a negative DCAD diet may mitigate the reduction of DMI around calving.
Relationship between overfeeding and overconditioning in the dry period and the problems of high producing dairy cows during the postparturient period.
suggested that BCS is a risk factor for MF due to decreased Ca intake and absorption around parturition. Considering the observed interaction of treatment with BCS at enrollment for MF in multiparous cows, we assessed serum calcium concentration in multiparous cows to investigate possible analogous differences in overconditioned cows. We did not detect a significant interaction of treatment with BCS, but our subgroup analysis was not designed to test this hypothesis. We encourage future studies designed to examine the relationship of adiposity with blood calcium concentrations.
Milk fever has been associated with increased risk of MAST (
Association of immediate postpartum plasma calcium concentration with early-lactation clinical diseases, culling, reproduction, and milk production in Holstein cows.
). Therefore, our results are among the first to indicate that negative DCAD diets tended to reduce the risk of MAST within 30 DIM in a population with low overall incidence of MAST, compared with positive DCAD diets. This outcome should be investigated in future large-scale clinical trials.
did not find differences in DA incidence or postpartum BHB when evaluating the effects of acidogenic diets. In our study, the cumulative incidence of hyperketonemia was lower than in other large studies (
A field trial on the effect of propylene glycol on displaced abomasum, removal from herd, and reproduction in fresh cows diagnosed with subclinical ketosis.
) but was not different between treatment groups. Nevertheless, cows fed negative DCAD had a reduced incidence of DA within 30 DIM than cows fed positive DCAD.
In previous studies, blood tCa within 24 h of parturition was not associated with RP or metritis risk (
Association of immediate postpartum plasma calcium concentration with early-lactation clinical diseases, culling, reproduction, and milk production in Holstein cows.
Association of immediate postpartum plasma calcium concentration with early-lactation clinical diseases, culling, reproduction, and milk production in Holstein cows.
, who found that lowering DCAD eliminated MF but had no effect on RP or metritis.
When analyzing DCAD, it is important to consider not only the targeted acidification level but also the magnitude of difference in DCAD between the studied diets, the number and characteristics of the animals included, and the study design. Although a physiologic inflection point exists when DCAD <0, both the absolute DCAD in the acidified group and the magnitude of difference between the negative DCAD and the control diets are important to contextualize an experiment. The fact that this study was performed on 4 commercial dairy farms with weekly visits represented some challenges in controlling variation in diets within and between farms, which would not be experienced in a research farm with greater control of the study elements. Nevertheless, we achieved accurate delivery of targeted DCAD in the negative DCAD group, with urine pH in the target zone.
Consistency of prepartum dietary calcium among and within 2 herds was one of the challenges encountered. The optimal amount of calcium fed to prepartum cows and its interaction with negative DCAD diets remain unclear. In a meta-analysis, based on a quadratic relationship between dietary Ca and predicted MF incidence, the predicted optimal dietary calcium should be <0.7 or >1.7% of DM (accounting for DCAD;
). Three of the farms in our study had mean prepartum dietary Ca >0.7 and <1.7% of DM. Results from a recent meta-analysis concluded that independent of DCAD, the risk of MF in multiparous cows seems to increase with prepartum dietary Ca content (Ca in the included studies ranged from 0.16 to 1.98% of DM;
). More research is needed to identify the optimal dietary supply of Ca with negative DCAD diets, but our study was weakened by not having greater consistency of dietary Ca within and between farms. However, this reflects the variability expected under field conditions of use of acidogenic supplements.
Producers were allowed to provide Ca treatment as soon as a cow demonstrated signs of clinical hypocalcemia. If producers provided Ca supplementation prophylactically, those cows were excluded from the trial; fortunately, only 9 instances of this occurred during the study. If not blinded to treatment, producers might have been inclined to prophylactically or differentially treat cows in CON. This is one reason we went to considerable trouble to have a placebo and to maintain blinding. We acknowledge that in the 2 instances where treatment inclusion levels in the diet were adjusted, producers may have guessed at the treatment assignments, but we did not explain these adjustments, and we did not observe any differential treatment of cows by the producers.
Out of necessity, treatment was applied at pen level on each farm in a cluster randomized design. A positive or negative DCAD diet was offered for 3 mo, and the pen was then switched to the opposite treatment, with a total of 4 periods per farm. Ideally both prepartum diets would have been fed simultaneously to 2 close-up pens on each farm. In that scenario, cows allocated to different pens in the same period would be exposed to the same conditions of feed, weather, and management, better isolating the effect of treatment. We acknowledge that within a farm, period and treatment effects are confounded. However, we think it unlikely that unmeasured changes in feed composition or pen social variables aligned perfectly with treatment periods. We accounted for unmeasured sources of variance and correlation of cows within a herd by including a random effect of farm in all statistical models. Despite the limitations, and considering the realities of the sizes of herd in Canada and many other dairy regions, switching diets every 3 mo, as performed in this study, was the best approach to minimize seasonal and nutritional confounding while providing replication of treatments within each farm. We correctly specified the experimental unit in all our analyses, which substantially but necessarily reduced the statistical power of the study. For this type of study design, it would have been desirable to have not only simultaneous treatments on each farm but also more experimental units with replication on more farms. However, that was not logistically or financially feasible.
Although we predicted more MF cases in the control group, we did not expect the dramatic increase in MF and related disease that occurred when farm C first switched to the control diet. We cannot infer a causal relationship. Before participating in the trial, this particular farm administered calcium bolus at parturition to all cows, which could have masked nutritional issues related to Ca metabolism. A stopping rule was applied, but this criterion was not defined before the trial. The trial continued on the farm, but treatment switched back to the negative DCAD and was not switched again, resulting in 3 experimental periods rather than 4. Although we did not explicitly disclose treatment allocation, the producer was likely not blinded to the treatment group afterward, which might have biased subsequent disease diagnosis. However, as previously described, including or entirely excluding the data from this farm did not change our inferences about the effects of treatment for the study.
We collected feed samples from positive and negative DCAD diets once on each farm. If were to repeat the trial, we would sample the TMR at least once per pen-treatment, to describe the diets and their variation more precisely. However, this would require not only more frequent sampling but many more farms, to be able to assess possible interactions of negative DCAD with other nutritional variables. Weekly measurement of urine pH prepartum demonstrated successful and acceptably consistent acidification when fed negative DCAD diets.
Although conducting the study under field conditions imposed limitations, it also lends strength to the external validity of the findings. Pen-level treatment assignment reflects the way negative DCAD is implemented in the field, but poses substantial challenges to execute a controlled trial with sufficient power for binary health outcomes. Nevertheless, we showed effects of feeding a negative DCAD on reducing the incidence of MF, DA, and MAST. The observed incidence risks of uterine diseases do not suggest that achieving the calculated number of pen-treatments would have produced different inferences for uterine health.
CONCLUSIONS
On commercial Holstein farms, prepartum negative DCAD diets reduced the incidence of displaced abomasum and tended to decrease clinical mastitis. The risk of milk fever was reduced in multiparous cows with BCS ≥3.75 that received negative DCAD diets. Negative DCAD did not reduce the risk of uterine diseases. This study helps to specify the expected effects of feeding a negative DCAD at the level studied on postpartum health outcomes.
ACKNOWLEDGMENTS
We thank the dairy producers involved in the project for their collaboration. Financial support and donation of Soychlor was provided by Landus Cooperative, parent company of Dairy Nutrition Plus. RCS was partially supported by a doctoral stipend from the Ontario Ministry of Agriculture, Food, and Rural Affairs. The authors have not stated any conflicts of interest.
REFERENCES
Bassler D.
Briel M.
Montori V.M.
Lane M.
Stopping randomized trials early for benefit and estimation of treatment effects.
Controlled trial of the effect of negative dietary cation-anion difference prepartum diets on milk production, reproductive performance, and culling of dairy cows.
The effect of prepartum negative dietary cation-anion difference and serum calcium concentration on blood neutrophil function in the transition period of healthy dairy cows.
A tutorial on sample size calculation for multiple-period cluster randomized parallel, cross-over and stepped-wedge trials using the Shiny CRT Calculator.
Postpartum body condition score and results from the first test day milk as predictors of disease, fertility, yield, and culling in commercial dairy herds.
A field trial on the effect of propylene glycol on displaced abomasum, removal from herd, and reproduction in fresh cows diagnosed with subclinical ketosis.
Association of immediate postpartum plasma calcium concentration with early-lactation clinical diseases, culling, reproduction, and milk production in Holstein cows.
Relationship between overfeeding and overconditioning in the dry period and the problems of high producing dairy cows during the postparturient period.
Our objective was to assess the effects of feeding negative dietary cation-anion difference (DCAD) prepartum diets on milk production, reproductive performance, and culling. Cows from 4 commercial farms in Ontario, Canada were enrolled in a pen-level controlled trial from November 2017 to April 2019. Close-up pens (1 per farm) with cows 3 wk before calving were randomly assigned to a negative DCAD (TRT; −108 mEq/kg of dry matter; target urine pH 6.0–6.5) or a control diet (CON; +105 mEq/kg of dry matter with a placebo supplement).
00355-6&id=gr1.jpg){kind=link}
00355-6&id=gr2.jpg){kind=link}
00355-6&id=gr3.jpg){kind=link}
00355-6&id=gr4.jpg){kind=link}