Associations of subclinical hypocalcemia with fertility in a herd of grazing dairy cows

The main objective was to assess the associations of subclinical hypocalcemia (SCH), diagnosed at parturition (SCH-0) and 7 d in milk (SCH-7), with fertility in a herd of grazing dairy cows. Additional objectives were to characterize Ca concentration on 0 and 7 d in milk (DIM), assessing the risk factors for SCH-0 and SCH-7 and also the relationship with health status (metritis, endometritis, subclinical ketosis, and culling). A prospective observational study was carried out in a dairy farm in Argentina. Holstein cows (n = 126) were body condition scored (BCS, 1–5) on ?21 ± 3, 0, 7 ± 3, and 28 ± 7 DIM and blood was collected on 0 and 7 ± 3 DIM to determine Ca and β-hydroxybutyrate concentrations. Calcium concentrations <2.0 and <2.14 mmol/L were used to define SCH-0 and SCH-7, respectively. The associations of SCH with (1) the odds for pregnancy to first service (P1AI) and pregnancy by 100 DIM (P100) were evaluated by logistic models, (2) the services per pregnancy was evaluated by a Poisson regression model, and (3) the hazards of insemination and pregnancy were evaluated with proportional hazards regression models whereas median days from calving to first insemination and to pregnancy were estimated by Kaplan-Meier method. Additionally, Ca concentration was assessed by linear regression models, and the associations of SCH-0 and SCH-7 with the odds for metritis, endometritis, subclinical ketosis, and culling were evaluated by logistic models. Calcium concentrations were similar at 0 and 7 DIM (2.40 vs. 2.41 mmol/L, respectively); they were higher in cows calving in fall than in summer (2.58 vs. 2.24 mmol/L), and they also were higher in primiparous than in multiparous cows (2.53 vs. 2.28 mmol/L, respectively). The proportion of cows having SCH-0 and SCH-7 was 27.3 and 39.3%, respectively. Fall-calving cows had lower odds for SCH-0 [odds ratio (OR) = 0.31, 95% confidence interval (CI) = 0.12–0.86] than summer-calving cows, multiparous cows had higher odds for SCH-0 (OR = 3.96, 95% CI = 1.09–14.39) than primiparous cows, and cows with prepartum BCS ≥3.00 had higher odds for SCH-0 (OR = 4.03, 95% CI = 1.17–13.89) than in cows with BCS <3.00. Conversely, parity and prepartum BCS were not important predictors for SCH-7. Surprisingly, SCH-0 was not a risk factor for SCH-7. Cows with SCH-0 had lower odds for P1AI (OR = 0.26, 95% CI = 0.07–0.99) than normocalcemic cows, given that P1AI was 14 versus 38%, respectively. The hazard of first service was not associated with SCH-0 (hazard ratio = 1.03, 95% CI = 0.63–1.70) but cows with SCH-0 had lower hazard of pregnancy (hazard ratio = 0.39, 95% CI = 0.16–0.98) and took 32 d longer to get pregnant (105 vs. 73) than normocalcemic cows. Conversely, SCH-7 was not associated with fertility. Finally, SCH-0 and SCH-7 were associated with the odds for subclinical ketosis and metritis, respectively. In conclusion, SCH-0 but not SCH-7 is associated with reduced fertility in a herd of grazing dairy cows, but both were associated with health status.     

Effect of gonadotropin-releasing hormone administered at the time of artificial insemination for cows detected in estrus by conventional estrus detection or an automated activity-monitoring system

The objectives of this study were to determine the effects of GnRH at the time of artificial insemination (AI) on ovulation, progesterone 7 d post-AI, and pregnancy in cows detected in estrus using traditional methods (tail chalk removal and mount acceptance visualization) or an automated activity-monitoring (AAM) system. We hypothesized that administration of GnRH at the time of AI would increase ovulation rate, plasma progesterone post-AI, and pregnancy per AI (P/AI) in cows detected in estrus. In experiment 1, Holstein cows (n = 398) were blocked by parity and randomly assigned to receive an injection of GnRH at the time of estrus detection/AI (GnRH, n = 197) or to remain untreated (control, n = 201) on 4 farms. The GnRH was administered as 100 µg of gonadorelin acetate. Ovarian structures and plasma progesterone were assessed in a subset of cows (GnRH, n = 52; control, n = 55) in experiment 1 at the time of AI and 7 d later. In experiment 2, a group of 409 cows in an AAM farm were enrolled as described for experiment 1 (GnRH, n = 207; control, n = 202). Data were categorized for parity (primiparous vs. multiparous), season (cool vs. warm), number of services (first vs. > first), DIM (>150 DIM vs. ≤150 DIM), and for AAM cows in experiment 2 for activity level (high: 90–100 index vs. low: 35–89 index). Pregnancy diagnosis was performed between 32 and 45 d post-AI (P1) and 60 to 115 d post-AI (P2). In experiment 1, there was no difference in plasma progesterone at day of estrus detection (control = 0.09 ng/mL vs. GnRH = 0.16 ng/mL), 7 d later (control = 2.03 ng/mL vs. GnRH = 2.18 ng/mL), and ovulation rate (GnRH = 83.2% vs. control = 77.9%) between treatments. There were no effects of GnRH in experiment 1 for P/AI at P1 (control = 43.3% vs. GnRH = 38.6%), P2 (control = 38.4% vs. GnRH = 34.5%), and for pregnancy loss (control = 9.8% vs. GnRH = 8.2%). In experiment 2, there were no effects of GnRH for P/AI at P1 (control = 39.6% vs. GnRH = 40.1%), P2 (control = 35.0% vs. GnRH = 37.4%), and for pregnancy loss (control = 9.5% vs. GnRH = 6.2%). There was a tendency for a parity effect on P/AI for P1, but not P2 or for pregnancy loss. High-activity cows had greater P/AI in P1 (low activity = 27.9% vs. high activity = 44.1%), P2 (low activity = 21.8% vs. high activity = 41.2%), and lower pregnancy loss (low activity = 20.7% vs. high activity = 5.1%), but there were no interactions between treatment and activity level. The current study did not support the use of GnRH at estrus detection to improve ovulatory response, progesterone 1 wk post-AI, and P/AI. More research is needed to investigate the relationship between GnRH at the time of AI and activity level in herds using AAM systems.     

Association between genomic daughter pregnancy rates and reproductive parameters in Holstein dairy cattle

The objective of this observational study was to evaluate the relationship between genomic daughter pregnancy rate (GDPR) with reproduction parameters such as pregnancy at first artificial insemination (AI), pregnancy per AI, and pregnancy losses (PL). A total of 12,949 events from 3,499 Holstein cows were included. Cows were enrolled as nulliparous (n = 1,220), primiparous (n = 1,314), or multiparous (n = 965). Cows were bred either after a timed AI protocol, timed embryo transfer (ET), or spontaneous estrus. Most lactating cows were bred following a timed AI protocol based on estradiol and progesterone, and most nulliparous were artificially inseminated following estrus detection. Hair samples were collected from the tail switch and cows were genotyped using a SNP platform (Clarifide, Zoetis). Cows that were bred by timed AI were evaluated for estrous behavior using tail chalk. Tail chalk was applied on the head of the tail 2 d before timed AI and the chalk was evaluated at AI (no estrus: 100% of chalk remaining or ≥50% of chalk remaining; Estrus: <50% of chalk remaining). Pregnancy diagnosis was performed at d 32 and 60 after AI using ultrasonography, and the presence of a heartbeat was considered a positive diagnosis. Pregnancy loss was defined as a pregnant cow on d 32 that was nonpregnant on d 60. As GDPR increased, the odds of pregnancy at first AI increased [odds ratio (OR) = 1.28, 95% CI = 1.20–1.35], the odds of pregnancy per AI increased (OR = 1.31, 95% CI = 1.25–1.36), and the odds of PL decreased (OR = 0.66, 95% CI = 0.60–0.72). Most cows that were bred on the day of the timed AI demonstrated estrus (n = 6,075; 92.9%). The odds of demonstrating estrus on the day of timed AI increased as GDPR increased (OR = 1.31, 95% CI = 1.17–1.48). There was no interaction between GDPR and parity or breeding management for pregnancy at first AI, pregnancy per AI, and PL. In conclusion, the odds of pregnancy at first AI and pregnancy per AI increased as GDPR increased. Moreover, the odds of PL increased as GDPR decreased. Greater GDPR was also associated with greater occurrence of estrus on the day of timed AI. These results suggest that selecting for higher GDPR could result in better reproductive performance, but this would need to be assessed with additional research.     

Effect of nerve growth factor-β administered at insemination for lactating Holstein dairy cows bred after timed-artificial insemination protocol

The objectives of this study were to determine the effects of nerve growth factor-β (NGF), purified from bulls' seminal plasma and administered at the time of artificial insemination (AI), on progesterone post-AI, interferon-stimulated genes (ISG), and pregnancy per AI (P/AI) for lactating Holstein dairy cows enrolled in a timed-AI protocol. We hypothesized that administration of NGF at the time of AI would increase plasma progesterone post-AI, upregulate relative abundance of ISG, and improve P/AI in lactating dairy cows. Holstein cows (n = 557) from a single commercial dairy farm were blocked by parity and randomly assigned to receive an intramuscular injection containing 296 µg of bovine purified NGF at the time of AI, diluted in 2 mL of phosphate-buffered saline (NGF: n = 275), or receive only the 2 mL of phosphate-buffered saline (control: n = 282). Plasma progesterone and corpus luteum size were assessed in a subset of cows (NGF: n = 32; control: n = 36) at d 7, 14, and 19 post-AI. Relative mRNA abundance of ISG (ISG15, MX1, MX2, and RTP4) was assessed in peripheral blood leukocytes on d 19 post-AI. Pregnancy diagnosis was performed at 37 and 65 d post-AI. There was an interaction effect between treatment and parity for plasma progesterone; however, plasma progesterone and ISG did not differ between treatments. There were no effects of NGF for P/AI at 37 d post-AI (NGF = 40.0% vs. control = 41.6%), 65 d post-AI (NGF = 36.0% vs. control = 38.1%), and for pregnancy loss (NGF = 8.4% vs. control = 7.7%). The current study revealed that effects to NGF in lactating Holstein cows were minor and contingent with parity for progesterone, and no improvement in ISG relative abundance and P/AI were observed.     

Hormonal treatment before and after artificial insemination differentially improves fertility in subpopulations of dairy cows during the summer and autumn

Reduced conception rate (CR) during the hot summer and subsequent autumn is a well-documented phenomenon. Intensive use of cooling systems can improve summer and autumn reproductive performance, but is unable to increase CR to winter and spring levels. We examined whether combined hormonal treatments—to increase follicular turnover before artificial insemination (AI) and progesterone supplementation post-AI—might improve fertility of cooled cows during the summer and autumn. The experiment was conducted from July to November in 3 commercial herds in Israel and included 707 Holstein cows at 50 to 60 d in milk (DIM). Cows were hormonally treated to induce 2 consecutive 9-d cycles, with GnRH administration followed by PGF_2α injection 7 d later, followed by an intravaginal insert containing progesterone on d 5 ± 1 post-AI for 14 d. Both untreated controls (n = 376) and treated cows (n = 331) were inseminated following estrus, and pregnancy was determined by palpation 42 to 50 d post-AI. First-AI CR data revealed a positive interaction between treatment and cows previously diagnosed with postpartum uterine disease [odds ratio (OR) 2.24]. Interaction between treatment and low body condition score tended to increase the probability of first-AI CR (OR 1.95) and increased pregnancy rate at 90 DIM (OR 2.50) and at 120 DIM (OR 1.77). Low milk production increased the probability of being detected in estrus at the end of synchronization within treated cows (OR 1.67), and interacted with treatment to increase probability of pregnancy at 90 DIM (OR 2.39) relative to control counterparts. It is suggested that when administered with efficient cooling, combined hormonal treatment in specific subgroups of cows, that is, those previously diagnosed with postpartum uterine disease or those with low body condition score or low milk yield might improve fertility during the summer and autumn. Integration of such an approach into reproductive management during the hot seasons might improve treatment efficiency and reduce expenses.     

Effect of body condition change and health status during early lactation on performance and survival of Holstein cows

Body condition score (BCS) and disease records are commonly available in dairy operations. However, the effect of BCS changes (ΔBCS) considering specific health profiles has not been investigated extensively. The objective of this study was to assess the effects of different levels of ΔBCS on fertility, milk yield, and survival of Holstein cows diagnosed with reproductive disorders (REP; dystocia, twins, retained fetal membranes, metritis, and clinical endometritis), other health disorders (OTH; subclinical ketosis, left displaced abomasum, lameness, clinical mastitis, and respiratory disease), or with no disease events (HLT) within 40 days in milk (DIM). Data included lactation information from 11,733 cows calving between November 2012 and October 2014 in 16 herds across 4 geographical regions in the United States (Northeast, Midwest, Southwest, Southeast). Cows were evaluated for BCS at 5 ± 3 DIM (BCS5) and at 40 ± 3 DIM (BCS40) and the difference between BCS40 and BCS5 was classified as excessive loss of BCS (EL; ΔBCS ≤−0.75), moderate loss (ML; ΔBCS = −0.5 to −0.25), no change (NC; ΔBCS = 0), or gain of BCS (GN; ΔBCS ≥0.25). Multivariable logistic regression was used for assessing potential associations between the outcomes of interest and ΔBCS and health. The effect of the interaction term ΔBCS by health group was not statistically significant for any of the study outcomes. The odds of resumption of ovarian cyclicity (ROC), in GN, NC, and ML cows were 1.94 (95% CI: 1.57–2.40), 1.59 (1.28–1.97), and 1.27 (1.10–1.47) times greater than the odds of ROC in EL cows, respectively. The odds of pregnancy at 150 DIM (P150) in GN cows were 1.61 (1.20–2.17) times greater than the odds of P150 in EL cows. Cows with REP or OTH disorders had smaller odds of ROC compared with HLT cows [REP: OR = 0.65 (0.56–0.76) and OTH: OR = 0.79 (0.68–0.92)]. For pregnancy outcomes, REP cows had smaller odds of pregnancy at the first artificial insemination compared with HLT cows [0.70 (0.58–0.84)]. Similarly, REP cows had smaller odds of being diagnosed pregnant by 150 and 305 DIM compared with HLT cows [P150: 0.73 (0.59–0.87), P305: 0.58 (0.49–0.69)]. Overall, average daily milk within the first 90 DIM was greater in EL (39.5 ± 1.13 kg/d) and ML (38.9 ± 1.11 kg/d) cows than in NC (37.8 ± 1.12 kg/d) and GN (36.2 ± 1.12 kg/d) cows. On the other hand, average daily milk within the first 90 DIM was lower in REP (37.0 ± 1.11 kg/d) cows compared with OTH (38.7 ± 1.12 kg/d) and HLT cows (38.6 ± 1.11 kg/d). The magnitude of ΔBCS and the health status of early lactation cows should be considered when assessing subsequent cow performance and survival.     

Effect of pegbovigrastim on fertility and culling in grazing dairy cows and its association with prepartum nonesterified fatty acids

This randomized controlled trial on 4 commercial grazing dairy farms investigated whether treatment with pegbovigrastim (PEG) affected fertility and culling as measured during the full lactation. We also explored the effect of potential interactions of PEG treatment with parity, prepartum body condition score, prepartum nonesterified fatty acid concentration (pre-NEFA), and early-lactation clinical disease on these outcomes. Holstein cows were randomly assigned to 1 of 2 trial arms: a first PEG dose approximately 7 d before the expected calving date and a second dose within 24 h after calving (PEG: primiparous = 342; multiparous = 697) compared with untreated controls (control: primiparous = 391; multiparous = 723). Cox's proportional hazards regression models were used to analyze rate of first insemination, rate of pregnancy [within 150 and 305 d in milk (DIM)], and hazard of culling. Additional analyses were performed on data that were stratified by parity group and pre-NEFA class (low ≤0.3; high >0.3 mM). In high pre-NEFA cows, PEG treatment increased the rate of first insemination [hazard ratio (HR) = 1.15]. Early-lactation clinical mastitis (CM) and uterine disease (UD: retained placenta, metritis, or both) were associated with a reduced rate of pregnancy within 150 DIM (HR = 0.49 and 0.78, respectively). Pegbovigrastim treatment in high pre-NEFA cows with CM and UD increased the rate of pregnancy within 150 DIM (HR = 1.75 and 1.46, respectively). In high pre-NEFA cows, PEG treatment resulted in a lower hazard of culling (HR = 0.79). No treatment effect was detected in low pre-NEFA cows. This study shows that the effect of PEG treatment on fertility and culling interacts with pre-NEFA. In high pre-NEFA cows, PEG treatment increased the rate of first insemination, counteracted the negative association of early-lactation CM and UD with the rate of pregnancy, and decreased the hazard of culling.     

Effects of pegbovigrastim administration on periparturient diseases,milk production,and reproductive performance of Holstein cows

The aim of this study was to evaluate the effects of treating Holstein cows with pegbovigrastim on periparturient diseases, milk production, and reproductive performance while exploring the mode of action of an immunomodulatory protein. Cows were randomly allocated to 1 of 2 treatments, untreated control (CTR, n = 423) and pegbovigrastim (PEG, n = 417). At 7 d from the anticipated calving date (d ?7), cows allocated to PEG received a subcutaneous injection of 15 mg of pegylated recombinant bovine granulocyte colony stimulating factor (pegbovigrastim injection, Imrestor, Elanco Animal Health, Greenfield, IN). A second injection was administered within 24 h after calving (d 0). Blood samples were obtained from a subset of cows (CTR, n = 103; PEG, n = 102) at ?7 and 0, 3, 7, and 14 d relative to parturition. Samples were used for hemogram and quantification of haptoglobin, nonesterified fatty acids, β-hydroxybutyrate, and trace and macro minerals. Vaginal cytobrush was performed on the same subset cows at d 0, 7, and 14 to assess the relative neutrophil count. Additionally, colostrum samples were collected to measure IgG, IgM, IgA, and lactoferrin concentrations. Postpartum disease occurrence was recorded from calving until 30 d in milk (DIM). Weekly milk yield was recorded for the first 12 wk after calving. Cows treated with PEG had a 3- to 4-fold increase in circulating polymorphonuclear leukocyte, lymphocyte, and monocyte numbers, with a peak at 3 d after treatment followed by a gradual decline, but the counts remained significantly greater compared with CTR at 14 DIM. The administration of PEG did not affect the incidence of clinical and subclinical mastitis, retained fetal membranes, metritis, puerperal metritis, and endometritis. Primiparous cows treated with PEG tended to have lower odds of developing hyperketonemia than CTR [odds ratio (OR) = 0.57, 95% confidence interval (CI) = 0.23 to 1.42]. Cows treated with PEG had higher odds of being diagnosed with lameness within 30 DIM compared with CTR (OR = 1.79, 95% CI = 1.16 to 2.76); however, we found no significant differences by 60 DIM. Treatment with PEG increased the odds of displaced abomasum (OR = 8.27, 95% CI = 1.02 to 66.6). Cows treated with PEG had higher odds of being diagnosed with 1 or more clinical diseases compared with CTR cows (OR = 1.39, 95% CI = 1.02 to 1.90). We observed no differences in linear scores or milk composition between treatments. Furthermore, primiparous cows treated with PEG produced more milk than CTR primiparous cows during the first 12 wk postpartum (PEG = 37.51 ± 0.66; CTR = 35.91 ± 0.65 kg), but no differences were observed on energy-corrected milk. Treatment did not alter reproductive performance; additionally, cows diagnosed with metritis or puerperal metritis and treated with PEG tended to have higher proportion of neutrophils in the vaginal mucosa when compared with CTR metritic cows. Although PEG treatment increased circulating polymorphonuclear leukocyte, monocyte, and lymphocyte numbers, as expected, it was detrimental to cow health because it increased morbidity.     

Association among gestation length and health,production, and reproduction in Holstein cows and implications for their offspring

Objectives were to evaluate associations among gestation length (GL) and performance in Holstein cows and their offspring. A total of 8,095 Holstein cows and 3,635 female offspring born alive on 2 farms using only artificial insemination (AI) were evaluated. Gestation length averaged 276 ± 6 d in the 8,095 dams, and it was categorized as short (SGL; at least 1 SD below the population mean; mean = 266 d, range 256 to 269 d), average (AGL; population mean ± 1 SD; mean = 276 d, range 270 to 282 d), or long (LGL; at least 1 SD above the population mean; mean = 285 d, range 283 to 296 d). Responses evaluated in dams included incidence of diseases within 90 d in milk (DIM), pregnancy at first AI and by 300 DIM, and time to pregnancy. Milk yield and removal from the herd by culling or death were recorded for the first 300 DIM. Responses evaluated in female offspring born alive included removal from the herd and reproductive performance. Within primiparous cows, those with SGL had greater incidence of stillbirth, retained placenta, and metritis than primiparous with AGL or LGL; however, within multiparous cows, those with SGL or LGL had greater incidence of dystocia, stillbirth, retained placenta, and metritis than cows with AGL. Morbidity and rate of morbidity were greater for SGL and LGL than AGL. Rate of removal of dams from the herd was 38% faster for SGL than AGL. Milk production was greatest in AGL cows, but responses depended on parity. For primiparous cows, milk production was less in SGL and LGL than in AGL (AGL = 35.4, SGL = 34.6, LGL = 33.0 ± 0.4 kg/d), whereas for multiparous cows, production was less in SGL but greater in LGL than in AGL (AGL = 41.6, SGL = 38.6, LGL = 42.4 ± 0.3 kg/d). A smaller proportion of cows with SGL received at least one AI, but pregnancy at first AI did not differ among groups. Rate of pregnancy was 11% slower for LGL multiparous than for AGL multiparous. By 300 DIM, pregnancy was greater in AGL than SGL. Pregnancy by 300 DIM in multiparous cows was also greater for AGL than LGL. Heifers from dams with GL that deviated from AGL had greater mortality postweaning (AGL = 3.2 vs. SGL = 6.5 vs. LGL = 5.4%). The rate of removal from the herd was greater for SGL (adjusted hazard ratio = 1.78; 95% CI: 1.26 to 2.52) and LGL (adjusted hazard ratio = 2.00; 95% CI: 1.45 to 2.76) than for AGL heifers. Pregnancy at first AI was lowest for LGL and by 500 d of age a larger proportion of AGL heifers were pregnant than LGL (AGL = 82.3 vs. SGL = 79.2 vs. LGL = 74.0%). Cows with GL within 1 SD of the population mean (270 to 282 d) had improved health, production, and reproduction. Heifers from cows with GL within 1 SD of the population mean had improved health and reproduction. Gestation length affects performance of both dams and their offspring.     

Metabolic parameters in transition cows as indicators for early-lactation culling risk

A retrospective cohort study was performed with the objective of determining whether the serum concentrations of nonesterified fatty acids (NEFA), β-hydroxybutyric acid (BHBA), or calcium were associated with the risk of culling within 60 d in milk (DIM) in Holstein cows, and to establish thresholds for each metabolite that were predictive of increased culling risk. Data from 5,979 cows in Ontario (Canada) and several US states were obtained from 4 previously reported studies. For each metabolite and each of 3 sampling weeks (-1, +1, and +2 relative to calving), an optimal threshold was calculated based on having the maximum combined sensitivity (Se) and specificity (Sp) and used to categorize the serum concentrations into high and low risk groups. Logistic regression models were built for each metabolite and each week of sampling, as well as considering together all metabolites in wk -1 and wk +1 relative to calving. Cow was considered the experimental unit and herd as a random effect. Considered separately, precalving NEFA ≥ 0.4 mmol/L [odds ratio (OR)=1.8; 95% confidence interval (CI)=1.4 to 2.2], NEFA ≥ 0.8 mmol/L in wk +1 relative to calving (OR=2.0; 95% CI=1.5 to 2.6) and NEFA ≥ 0.8 mmol/L in wk +2 (OR=4.2; 95% CI=1.9 to 9.4 for cows in lactation 2; OR=2.1; 95% CI=1.4 to 3.3 for cows in lactation ≥ 3) were each associated with an increased risk of culling within the first 60 DIM. Similarly, BHBA ≥ 0.7 mmol/L in wk -1 (OR=1.8; 95% CI=1.3 to 2.5), BHBA ≥ 1.2 mmol/L in wk +1 (OR=1.8; 95% CI=1.4 to 2.2), and BHBA ≥ 1.6 mmol/L in wk +2 (OR=3.2; 95% CI=1.6 to 6.4 for cows in lactation 2; OR=2.3; 95% CI=1.6 to 3.3 for cows in lactation ≥ 3) were each associated with an increased risk of culling within the first 60 DIM. Likewise, calcium ≤ 2.3 mmol/L in wk -1 (OR=1.6; 95% CI=1.2 to 2.2), calcium ≤ 2.2 mmol/L in wk +1 (OR=1.5; 95% CI=1.2 to 1.9), and calcium ≤ 2.3 mmol/L in wk +2 (OR=2.3; 95% CI=1.1 to 3.1) were each associated with an increased risk of culling within the first 60 DIM. When all metabolites were analyzed together, serum NEFA and calcium concentrations in wk -1 and serum NEFA concentration in wk +1 remained in the models. In conclusion, elevated serum NEFA and BHBA concentrations and lower serum calcium concentrations within 1 wk before calving through 2 wk after calving were associated with an increased risk of culling in early lactation. Measuring the concentration of selected metabolites around parturition may help to develop monitoring and intervention strategies to prevent early culling in transition 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

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). Each pen was fed TRT or CON for 3 mo (1 period), and then switched to the other treatment for the next period (4 periods per farm). Data from 15 experimental units (8 pen treatments in TRT and 7 in CON), with a total of 1,086 observational units (cows), were included. The effect of treatment on milk yield at the first 3 milk recording tests of lactation was assessed with linear regression models accounting for repeated measures. The risk of pregnancy at first artificial insemination and culling by 30, 60, and 305 d in milk (DIM) were analyzed with logistic regression models, and effects on time to first AI, pregnancy, and culling were assessed with Cox proportional hazards models. All models included treatment, parity, and their interactions, accounting for pen-level randomization and clustering of animals within farm with random effects, giving 10 degrees of freedom for treatment effects. Multiparous cows fed TRT produced more milk at the first (42.0 vs. 38.8 ± 1.2 kg/d) and second (44.2 vs. 41.7 ± 1.3 kg/d) milk tests. However, multiparous cows fed TRT tended to have 0.2 percentage units less milk fat content at these tests. Although multiparous cows fed TRT tended to have greater energy-corrected milk at the first test (least squares means ± standard error: TRT = 46.1 ± 0.9 vs. CON = 43.8 ± 1 kg/d), there were no differences observed in energy-corrected milk at the second or third tests. In primiparous cows, there was no effect of treatment on milk production. Multiparous cows fed TRT had greater pregnancy to first insemination (TRT = 42 ± 3 vs. CON = 32 ± 4%) and tended to have shorter time to pregnancy [hazard ratio (HR) = 1.20; 95% CI: 0.96–1.49]. In primiparous cows fed TRT, time to pregnancy was increased (HR = 0.76; 95% CI: 0.59–0.99). Culling by 30 DIM tended to be less in TRT (3.3 ± 1.1%) than CON (5.5 ± 1.8%). No effect of treatment on culling by 305 DIM was detected in primiparous cows, but in multiparous cows, the TRT diets decreased the odds of culling (21.3 ± 1.9 vs. 31.7 ± 2.8%) and daily risk of culling to 305 DIM (HR = 0.64; 95% CI: 0.46 to 0.89). Under commercial herd conditions, prepartum negative DCAD diets improved milk production and reproductive performance, and reduced culling risk in multiparous cows. In primiparous cows, TRT diets had no effect on milk yield or culling, but increased the time to pregnancy. Our results suggest that negative DCAD diets should be targeted to multiparous cows.     

The effect of presynchronization with prostaglandin F2α and gonadotropin-releasing hormone simultaneously, 7 d before Ovsynch,compared with Presynch-10/Ovsynch on luteal function and first-service pregnancies per artificial insemination

Pregnancy per artificial insemination (P/AI) following Ovsynch is optimized when cows ovulate to the first GnRH of Ovsynch. Fertility programs are designed to presynchronize cows to d 6 or 7 of the estrous cycle to increase the chances of ovulation of a first-wave dominant follicle to the first GnRH of Ovsynch. The hypothesis of this experiment was that simplification of a presynchronization program through the combination of PGF_2α and GnRH on the same day, 7 d before Ovsynch, would allow for similar P/AI compared with Presynch-10. Lactating dairy cows (n = 432) 41 to 47 d in milk (DIM) were randomly assigned to 2 treatments within parities for first service. Control cows received Presynch-10/Ovsynch consisting of the following: PGF_2α–14 d–PGF_2α–10 d–GnRH–7 d–PGF_2α–56 h–GnRH–16 h–AI. Treated cows received PGF_2α and GnRH–7 d–GnRH–7 d–PGF_2α–56 h–GnRH–16 h–AI. All cows received a supplemental injection of PGF_2α 24 h after the PGF_2α of Ovsynch to enhance complete luteolysis. All cows received timed AI between 75 and 81 DIM. Blood was collected to assess circulating concentrations of progesterone (P4), and the number and size of corpora lutea (CL) were recorded using ultrasonography on day of PGF_2α of Ovsynch. The administration of PGF_2α simultaneously with GnRH and 7 d before Ovsynch (PG+G) had similar P/AI at 28 (46 vs. 48%), 35 (43 vs. 43%), 49 (39 vs. 39%), and 77 d post-AI (38 vs. 39%) compared with Presynch-10. No differences were observed in P/AI in primiparous versus multiparous cows at 28 (52 vs. 45%), 35 (48 vs. 41%), 49 (45 vs. 37%), and 77 d post-AI (43 vs. 36%). No difference existed between treatments in percentage of cows with functional CL at PGF_2α of Ovsynch, total luteal area (mm²), or serum concentrations of P4 at time of PGF_2α of Ovsynch, regardless of parity. Number of CL had a tendency to be greater for multiparous PG+G vs. Presynch-10 cows (2.34 ± 0.09 vs. 2.15 ± 0.08) but not in primiparous cows (1.95 ± 0.10 vs. 1.98 ± 0.11). In summary, administering both PGF_2α and GnRH on the same day, 7 d before the start of Ovsynch, appears to be a simple and effective alternative to Presynch-10 Ovsynch.     

The effect of a double dose of cloprostenol sodium on luteal blood flow and pregnancy rates per artificial insemination in lactating dairy cows

Inadequate luteolysis in fertility programs is a problem for lactating dairy cows treated with a single dose of PGF_2α. The proportion of cows with complete luteolysis can be increased by administering 2 doses of PGF_2α 24 h apart. This study hypothesized that a double dose of cloprostenol sodium (1.0 mg) could take the place of 2 doses 24 h apart due to its enhanced half-life. Cows were allocated to receive 1 of 3 treatments: negative controls: 0.5 mg of cloprostenol sodium (single; n = 337); positive controls: two 0.5-mg doses of cloprostenol sodium 24 h apart (two/24; n = 313); and treated: 1.0 mg of cloprostenol sodium (double; n = 298) at the final PGF_2α of Double-Ovsynch. Cows received artificial insemination (AI) 16 h after final GnRH of Double-Ovsynch. Pregnancy diagnosis was determined at 24, 34, 62, and 184 d post-AI. Pregnancy loss was categorized in the following periods: between 24 to 34, 34 to 62, and 62 to 184 d post-AI. Ultrasonography (B-mode and color Doppler) was used to assess luteal function pre- and posttreatment with various doses of cloprostenol sodium. Luteal volume and luteal blood flow (LBF) from d 7 and 14 corpora lutea were determined before treatment (d −1), and 2 and 4 d after treatment. No evidence was observed of an effect of treatment on pregnancy rates per AI at 24, 34, or 62 d post-AI. No effect was observed of treatment on pregnancy losses occurring between d 24 and 34, 34, and 62, and between 62 and 184 d post-AI. However, third-plus parity cows treated with the single treatment had greater pregnancy loss compared with two/24 and double between d 24 and 34 post-AI. Third-plus parity cows that received the double treatment had lower LBF 2 and 4 d after treatment compared with cows treated with single. Amount of LBF present 4 d after treatment was not a predictor of pregnancy or pregnancy loss. A double dose (1.0 mg) of cloprostenol sodium may be a feasible alternative for fertility programs based on nondifferent outcomes to the two/24 treatment, lower pregnancy losses, and reduced LBF disappearance following treatment in third-plus parity cows in comparison with the single treatment.     

Association between early postpartum rumination time and peak milk yield in dairy cows

Limited information is available on the relationship between rumination time (RT) in the early postpartum period and milk production later in lactation. Therefore, the objectives of this study were to (1) investigate the association of change in RT and average RT during the immediate postpartum period with peak milk yield (PMY) in dairy cows, and (2) determine the best model based on days in milk (DIM) to evaluate this association. Cows from 33 free-flow automatic milking system farms were included in this study, where retrospective milk production and RT data were collected for 12 mo. Cows were categorized by parity number into parity 1 (P1, n = 1,538), parity 2 (P2, n = 1,354), or parity ≥3 (P3+, n = 1,770). For each cow, PMY was identified as the highest daily milk yield up to 180 DIM for P1 and 120 DIM for P2 and P3+ cows. Five change in RT variables and 5 average RT variables were created corresponding to the first 2 to 6 DIM. Change in RT variables were the slope coefficients for change in RT/d related to DIM = 1 extracted from simple linear regressions, and average RT variables were the arithmetic mean RT. Five models analyzing PMY and corresponding variables calculated over the first 2 to 6 DIM had fixed effects of average RT, change in RT, parity, average RT × parity interaction, change in RT × parity interaction, and a random intercept for farm. Peak milk yield occurred at (median) 75, 44, and 46 DIM for P1, P2, and P3+, respectively. Overall PMY was (mean ± standard deviation) 54 ± 11 kg and it increased as parity increased. A positive association was found between change in RT and PMY, and average RT and PMY for P2 and P3+ cows in all 5 models corresponding to the first 2 to 6 DIM, indicating that greater average RT and quicker increase in RT after calving are associated with greater PMY for multiparous cows. Although the model including all 6 DIM had the greatest accuracy, results indicated that rumination data collected over the first 2 DIM may also provide adequate information for the association of average RT and change in RT with PMY in P2 and P3+ cows. For each 100 min/d increase in change in RT over the first 6 DIM, PMY increased by 4.3 (95% confidence interval: 2.2–6.3) and 4.8 (95% confidence interval: 3.2–6.5) kg for P2 and P3+ cows, respectively. Peak milk yield increased by 2.3 (95% CI: 1.7–2.8) and 2.2 (95% confidence interval: 1.7–2.6) kg for each 100 min increase in average RT over the first 6 DIM for P2 and P3+ cows, respectively. No association was observed between rumination behaviors and PMY for P1 cows. Results from this study indicate that the length of time for multiparous cows to achieve a stable RT in the early postpartum period combined with average RT during the same period may be useful in predicting their overall lactation milk production.     

Responses to metabolic challenges in dairy cows with high or low milk yield during an extended lactation

Responses of dairy cows with high or low milk yield (MY) beyond 450 d in milk (DIM) to 3 metabolic challenges were investigated. Twelve multiparous Holstein-Friesian cows that calved in late winter in a pasture-based system were managed for a 670-d lactation by delaying re-breeding. Cows were selected for either high MY (18.9 ± 1.69 L/cow per d; n = 6) or low MY (12.3 ± 3.85 L/cow per d; n = 6) at 450 DIM. Cows were housed indoors for 2 periods of 12 d at approximately 460 and 580 DIM. Each cow was fed freshly cut pasture (460 DIM) or pasture silage (580 DIM) plus 6.0 kg of DM barley grain daily (approximately 200 MJ of total metabolizable energy/cow per day). At all other times, cows were managed as a single herd and grazed pasture supplemented with cereal grain to an estimated intake of 180 MJ of metabolizable energy/cow per d. Cows were fitted with a jugular catheter during the final week of each experimental period. Over a period of 3 d, each cow underwent an intravenous glucose tolerance test (0.3 g/kg of body weight), an insulin tolerance test (0.12 IU of insulin/kg of body weight), and a 2-dose epinephrine challenge (0.1 and 1.6 µg/kg of body weight). Cows selected for high MY had greater milk and milk solids yields between 450 and 580 DIM than low MY cows (17.3 vs. 10.8 ± 1.49 kg of milk/d and 2.4 vs. 1.5 ± 0.23 kg of milk solids/d). The results indicated that whole body and peripheral tissue responsiveness to insulin may vary between cows of high and low MY. Following the glucose tolerance test, high MY cows had a lower plasma insulin response with a greater glucose area under the curve than low MY cows. Further, high MY cows had slower plasma glucose clearance compared with low MY cows during an insulin tolerance test. The plasma nonesterified fatty acid (NEFA) responses to the IVGTT and the ITT were similar between cows of high and low MY, but the clearance of NEFA from the plasma following both the IVGTT and ITT were slower at 580 compared with 460 DIM. The sensitivity to epinephrine was greater in high MY cows compared with low MY cows as the glucose and NEFA area under the curve and the percentage change in NEFA were greater in high MY after the low dose epinephrine challenge. However, the lipolytic but not the glucose appearance in response to epinephrine was greater in high MY cows than low MY cows. Following the high dose of epinephrine, the glucose response was lower, but the NEFA response was greater in high MY compared with low MY cows. Cows able to sustain greater MY to 580 DIM had a greater propensity for lipid mobilization, possibly enhancing nutrient partitioning to the mammary gland during the late stages of an extended lactation.     

Effect of incomplete milking during the first 5 days in milk on udder and reproductive tract health: Results from a randomized controlled trial

The aim of this study was to investigate the effect of an incomplete milking on risk of mastitis and reproductive tract disease. Multiparous dairy cows (n = 878) from 13 commercial herds were enrolled in a randomized controlled trial. Cows were randomly assigned to either a control (milked conventionally) or a treatment group, which consisted of an incomplete milking (10–14 L of milk collected/d) from 1 to 5 d in milk (DIM). Quarter milk samples were collected at approximately 11 and 18 DIM to measure somatic cell count (SCC). Quarters were considered negative for intramammary infection if SCC was <100,000 cells/mL and positive if SCC was ≥200,000 cells/mL. To calculate intramammary infection incidence, negative quarters of the initial samples collected were tested again 1 wk later. This was done to deter incidence of positive quarters. To calculate elimination rate, positive quarters were tested again 1 wk later to detect mastitis elimination. Farmers recorded clinical mastitis events. Cows were also examined at approximately 35 DIM with a Metricheck device (Simcro, Hamilton, New Zealand) for detection of purulent vaginal discharge (PVD) and with an endometrial cytobrush for presence of leukocytes [endometrial cytology for smear (ENDO) and for leukocyte esterase test (LE)]. A threshold ≥3 was used to define a positive PVD or LE test, whereas a polymorphonuclear cell count ≥6% was used to define a positive ENDO. Five generalized mixed models with cow or herd as random intercepts were used to determine the effects of incomplete milking on odds of new intramammary infection, odds of intramammary infection elimination, and odds of a positive PVD, LE, or ENDO status. To investigate time until first clinical mastitis event, a Cox model with a herd frailty term was used. The odds of new intramammary infection and intramammary infection elimination for incompletely milked cows were 0.90 [95% confidence interval (CI): 0.49, 1.7] and 2.9 (95% CI: 1.4, 6.0) times those of conventionally milked cows, respectively. The hazard of clinical mastitis in incompletely milked cows was 0.96 (95% CI: 0.59, 1.6) times that of conventionally milked cows. The odds of PVD, LE, and ENDO for incompletely milked cows were 1.4 (95% CI: 0.89, 2.1), 1.3 (95% CI: 0.88, 1.8), and 1.2 (95% CI: 0.81, 1.7) times those of conventionally milked cows. These results suggest that incomplete milking during the first 5 DIM increases the odds of a decrease in SCC from 11 to 18 DIM but does not affect odds of increase in SCC in the same period. The incomplete milking had no effect on clinical mastitis incidence in the first 90 DIM or on reproductive tract health at 35 DIM.     

Characterization of anogenital distance and its relationship to fertility in lactating Holstein cows

Anogenital distance (AGD) serves as a marker for prenatal androgenization, reproductive development, and fertility in humans and rodents. The primary objectives of this observational study in lactating dairy cows were to (1) characterize the distribution and variability of AGD, (2) determine the relationship among AGD and potential postnatal AGD determinants of age and height, and (3) evaluate the associations between AGD and pregnancy to first artificial insemination (P/AI) and cumulative pregnancy by 250 d in milk (DIM) within parity groups (first, second, and third+ parities). The secondary objective was to evaluate the association between AGD and testosterone concentrations. The AGD (mm), age (yr), and height at hip (cm) at the time of AGD determination, and aforesaid reproductive outcomes were determined in 921 Holstein cows (first, second, and third+ parity; n = 360, 256, and 305, respectively). Plasma concentrations of testosterone were determined in a subset of 93 cows. Overall, AGD had a normal distribution and high variability [mean (±standard deviation); 131.0 ± 12.2 mm], was weakly associated with cow age and height (coefficient of determination = 0.09 and 0.04, respectively), and had an inverse relationship with P/AI in first- and second-parity cows, but not in third+ parity cows. For every 1 mm increase in AGD, the odds of P/AI decreased by 3.4 and 2.4% for first- and second-parity cows, respectively. The optimal AGD threshold to predict probability of P/AI was 127.1 mm for both first- (sensitivity: 66.4; specificity: 56.6%) and second-parity cows (sensitivity: 46.0; specificity: 70.4%). Accordingly, first- and second-parity cows were categorized into either short or long AGD (≤ or >127.1 mm), and associations with reproductive outcomes were evaluated. First-parity cows with long AGD had lower P/AI (30.9 vs. 53.6%) and decreased likelihood (hazard ratio: 0.68) of pregnancy by 250 DIM than those with short AGD. Similarly, second-parity cows with long AGD had reduced P/AI (28.3 vs. 44.4%) and a tendency for decreased likelihood (hazard ratio: 0.76) of pregnancy by 250 DIM than in cows with short AGD. The association between AGD and testosterone was weak and nonsignificant. In summary, AGD in Holstein cows was normally distributed, highly variable, and weakly associated with age and height. Besides, AGD had an inverse relationship with P/AI and cumulative pregnancy by 250 DIM in first- and second-parity cows; however, such a relationship was not evident in older (third+ parity) cows.     

Effect of prostaglandin F2α on subclinical endometritis and fertility in dairy cows

The objectives were to determine the effects of PGF_2α treatment on the prevalence of subclinical endometritis (SCE) and fertility of dairy cows. A total of 406 Holstein cows (167 primiparous and 239 multiparous) from 5 herds were used. Uterine lavage for diagnosis of SCE, PGF_2α treatment, evaluation of body condition scores (BCS), and collection of blood samples for estrous cyclicity determination were performed at 21, 35, and 49 d in milk (DIM). Polymorphonuclear cells (PMN) were quantified and thresholds for diagnosing SCE were selected by receiver operating characteristics analysis. Cows classified as having SCE at 35 DIM (≥6.5% PMN) and 49 DIM (≥4.0% PMN) had increased time to pregnancy; however, cows classified as having SCE at 21 DIM (≥8.5% PMN) did not. Median days to pregnancy were delayed by 30 (151 vs. 121 d) and 40 (169 vs. 129) d for cows classified as having SCE at 35 and 49 DIM, respectively. Treatment with PGF_2α did not affect the prevalence of SCE either at 35 (37.9 vs. 38.4%) or at 49 DIM (34.0 vs. 40.4%). Treatment with PGF_2α did not affect time to first insemination (AI; median 76 DIM for cows treated with PGF_2α; 79 DIM for control. Nonetheless, PGF_2α treatment increased pregnancy to first AI in all the cows (35.5 vs. 24.1%) and hazard ratio (HR) of pregnancy in cows with BCS ≤2.5 when all of the cows were evaluated (HR = 1.5; 95% confidence interval; CI = 1.1 to 2.0) and when only cows without SCE were evaluated (HR = 1.8; 95% CI = 1.2 to 2.7). Treatment with PGF_2α did not affect the hazard of pregnancy in cows with SCE at 49 DIM (HR = 0.9; 95% CI = 0.6 to 1.3). In these farms, treatment with PGF_2α did not affect SCE or time to first insemination, but did increase first-service pregnancy per AI and decreased time to pregnancy in cows with low BCS.     

Reproductive performance and herd exit dynamics of lactating dairy cows managed for first service with the Presynch-Ovsynch or Double-Ovsynch protocol and different duration of the voluntary waiting period

The objective of this experiment was to evaluate the reproductive performance and herd exit dynamics of dairy cows managed for first service with programs varying in method of submission for insemination and voluntary waiting period (VWP) duration. Holstein cows from a commercial farm in New York were randomly allocated to receive timed artificial insemination (TAI) after the Double-Ovsynch protocol (GnRH, 7 d later PGF_2α, 3 d later GnRH, 7 d later GnRH, 7 d later PGF_2α, 56 h later GnRH, and 16 to 18 h later TAI) at 60 ± 3 d in milk (DIM) (DO60 = 458), TAI after Double-Ovsynch at 88 ± 3 DIM (DO88 = 462), or a combination of AI at detected estrus (starting at 50 ± 3 d in milk) and TAI with the Presynch-Ovsynch protocol (PGF_2α, 14 d later PGF_2α, 12 d later GnRH, 7 d later PGF_2α, 56 h later GnRH, and 16 to 18 h later TAI; PSOv = 450). Subsequent artificial insemination (AI) services were conducted at detected estrus or the Ovsynch protocol (32 ± 3 d after AI GnRH, 7 d later PGF_2α, 56 h later GnRH, and 16 to 18 h later TAI) for cows not reinseminated at detected estrus. In a subgroup of cows, cyclicity (based on progesterone concentration), uterine health (vaginal discharge and uterine cytology), and BCS were evaluated at baseline (DO60 and DO88 = 33 ± 3 DIM; PSOv = 34 ± 3 DIM), beginning of the synchronization protocol (DO60 = 33 ± 3 DIM; DO88 = 61 ± 3 DIM; PSOv = 34 ± 3 DIM), and within ?5 (PSOv) or ?10 d (DO) of the VWP end (DO60 = 50 ± 3 DIM; DO88 = 78 ± 3 DIM; PSOv = 45 ± 3 DIM). Effects of treatments were assessed with multivariable statistical methods relevant for each outcome variable. Cows in the DO88 treatment had delayed time to pregnancy during lactation (DO60 vs. DO88 hazard ratio = 1.53, 95% confidence interval = 1.32 to 1.78; PSOv vs. DO88 hazard ratio = 1.37, 95% confidence interval = 1.19 to 1.61) and, within multiparous cows, the DO88 and PSOv treatments had greater risk of leaving the herd than cows in the DO60 treatment (DO88 vs. DO60 hazard ratio = 1.49, 95% confidence interval = 1.11 to 2.00; PSOv vs. DO60 hazard ratio = 1.39, 95% confidence interval = 1.03 to 1.85). Cows in the DO88 treatment had improved uterine health, greater BCS, and reduced incidence of anovulation than cows in DO60 and PSOv; however, overall pregnancy per AI 39 ± 3 d after AI was similar for the 3 treatment groups. In summary, reproductive management strategies that led to similar average DIM to the first service (～60 d) through a combination of AI at estrus with TAI (PSOv) or all TAI (DO60) resulted in reduced time to pregnancy after calving when compared with an all TAI program (DO88) with a VWP of 88 d. Within the multiparous cow group, those that received all TAI with a VWP duration of 60 d were less likely to leave the herd than cows in the other treatments.     

Mining data from milk infrared spectroscopy to improve feed intake predictions in lactating dairy cows

Feed intake is one of the most important components of feed efficiency in dairy systems. However, it is a difficult trait to measure in commercial operations for individual cows. Milk spectrum from mid-infrared spectroscopy has been previously used to predict milk traits, and could be an alternative to predict dry matter intake (DMI). The objectives of this study were (1) to evaluate if milk spectra can improve DMI predictions based only on cow variables; (2) to compare artificial neural network (ANN) and partial least squares (PLS) predictions; and (3) to evaluate if wavelength (WL) selection through Bayesian network (BN) improves prediction quality. Milk samples (n = 1,279) from 308 mid-lactation dairy cows [127 ± 27 d in milk (DIM)] were collected between 2014 and 2016. For each milk spectra time point, DMI (kg/d), body weight (BW, kg), milk yield (MY, kg/d), fat (%), protein (%), lactose (%), and actual DIM were recorded. The DMI was predicted with ANN and PLS using different combinations of explanatory variables. Such combinations, called covariate sets, were as follows: set 1 (MY, BW^0.75, DIM, and 361 WL); set 2 [MY, BW^0.75, DIM, and 33 WL (WL selected by BN)]; set 3 (MY, BW^0.75, DIM, and fat, protein, and lactose concentrations); set 4 (MY, BW^0.75, DIM, 33 WL, fat, protein, and lactose); set 5 (MY, BW^0.75, DIM, 33 WL, and visit duration in the feed bunk); set 6 (MY, DIM, and 33 WL); set 7 (MY, BW^0.75, and DIM); set-WL (included 361 WL); and set-BN (included just 33 selected WL). All models (i.e., each combination of covariate set and fitting approach, ANN or PLS) were validated with an external data set. The use of ANN improved the performance of models 2, 5, 6, and BN. The use of BN combined with ANN yielded the highest accuracy and precision. The addition of individual WL compared with milk components (set 2 vs. set 3) did not improve prediction quality when using PLS. However, when ANN was employed, the model prediction with the inclusion of 33 WL was improved over the model containing only milk components (set 2 vs. set 3; concordance correlation coefficient = 0.80 vs. 0.72; coefficient of determination = 0.67 vs. 0.53; root mean square error of prediction 2.36 vs. 2.81 kg/d). The use of ANN and the inclusion of a behavior parameter, set 5, resulted in the best predictions compared with all other models (coefficient of determination = 0.70, concordance correlation coefficient = 0.83, root mean square error of prediction = 2.15 kg/d). The addition of milk spectra information to models containing cow variables improved the accuracy and precision of DMI predictions in lactating dairy cows when ANN was used. The use of BN to select more informative WL improved the model prediction when combined with cow variables, with further improvement when combined with ANN.