What a 31-yr multibreed herd taught us about the influence of B. indicus genetics on reproductive performance of cows

Abstract

Bos taurus × Bos indicus crosses are widespread in tropical and subtropical regions, nonetheless, quantitative information about the influence of B. indicus genetics on the reproductive performance of beef cattle is lacking. Herein, we determined the association between level of B. indicus genetics and reproduction from a 31-yr dataset comprising sequential breeding seasons of the University of Florida multibreed herd (n = 6,503 Angus × Brahman cows). The proportion of B. indicus genetics in this herd is evenly distributed by each 1/32nd or approximately 3-percentage points. From 1989 to 2020, the estrous cycle of cows was synchronized for artificial insemination (AI) based on detected estrus or timed-AI (TAI) using programs based on gonadotropin-releasing hormone and prostaglandin, and progestin/progesterone. All cows were exposed to natural service after AI and approximately 90-d breeding seasons, considering the day of AI as day 0. The proportion of B. indicus genetics of cows was associated negatively with pregnancy per AI, ranging from 51.6% for cows with 0%–19% of B. indicus genetics to 37.4% for cows with 81%–100% of B. indicus genetics. Similar association was found for estrous response at the end of the synchronization protocol, ranging from 66.3% to 38.4%, respectively. This reduced estrous response helped to explain the pregnancy results, once the pregnancy to AI of cows showing estrus was 2.3-fold greater than for those not showing estrus and submitted to TAI. Despite reduced pregnancy per AI, the increase in the proportion of B. indicus genetics of cows was not associated with a reduction in the proportion of pregnant cows at the end of the breeding season. Nevertheless, the interval from entering the breeding season to pregnancy was lengthened as the proportion of B. indicus genetics of cows increased. The median days to pregnancy was extended by 25 when the proportion of B. indicus genetics surpassed 78% compared with less than 20%. Thus, the increase in the proportion of B. indicus genetics of cows was related to a reduction in pregnancy per AI and lengthening the interval to attain pregnancy during the breeding season, but not with the final proportion of pregnant cows. As a result, reproductive management strategies directed specifically to cows with a greater proportion of B. indicus genetics are needed to improve the rate of pregnancy in beef herds.

Bos taurus vs. Bos indicus cows are popular choices in tropical and subtropical beef production systems, as they combine desirable carcass traits and adaptation to hot and humid environments. Here, we characterized the extent of the B. indicus influence on the reproductive performance of B. taurus vs. B. indicus crossbred cows in contemporaneous groups, under controlled breeding seasons, and on the long term. We found that the increase in the proportion of B. indicus genetics of cows resulted in a reduction in pregnancy per artificial insemination and lengthened the interval to attain pregnancy during the breeding season, but not the final proportion of pregnant cows at the end of the breeding season.

Introduction

Bos indicus breeds are widespread in tropical and subtropical regions and frequently used in crossbreeding programs with Bos taurus breeds. Bos indicus breeds are more tolerant to heat stress and more resistant to parasites than Bos taurus breeds (Ibelli et al., 2012; Mateescu et al., 2020), traits desirable in the tropics. Despite this adaptability to the tropics, the reproductive performance of B. indicus cattle, such as the proportion of cows calving, calf survival, and the proportion of cows weaning a calf are generally inferior to those for B. taurus (Chenoweth, 1994). This inferiority is in part attributed to the inadequate nutritional conditions that B. indicus cows are subjected in the tropics, which contribute to prolonging postpartum anestrus and puberty attainment (Baruselli et al., 2004; Nogueira, 2004; Ribeiro et al., 2015). Nevertheless, the prolonged anestrus postpartum and later attainment of reproductive maturity seem to be an intrinsic particularity of B. indicus cattle (Chenoweth, 1994). Shorter and less overt estrus and longer gestation duration are other examples of particularities contributing to impair reproductive performance of B. indicus (Chenoweth, 1994; Bó et al., 2003; Sartori and Barros, 2011). Moreover, when B. indicus-influenced cows are submitted to estrous synchronization programs developed for B. taurus, pregnancy per AI (P/AI) are generally suboptimal, that is, <50% (Yelich and Bridges, 2012). In addition to the aforementioned factors, when these cows are submitted to estrous synchronization programs, circulating progesterone concentrations are elevated, which dampens the success of these programs (Carvalho et al., 2008; Batista et al., 2017, 2020) and subsequent P/AI (Peres et al., 2009; Martins et al., 2014). Modifications incorporated to the synchronization programs targeting one or more of B. indicus’s particularities have increased the P/AI of B. indicus-influenced cows in South America. Some of these modifications include adoption of temporary calf removal or administration of equine chorionic gonadotropin at the end of synchronization protocol, use of intravaginal progesterone devices previously used, administration of a dose of prostaglandin analogue at the beginning or before the end of a synchronization program and use of exogenous estradiol (Meneghetti et al., 2009; Peres et al., 2009; Sá Filho et al., 2009, 2011; Williams and Stanko, 2020). Moreover, the use of estrous synchronization programs that include exogenous progesterone have proved to be beneficial to the subsequent pregnancy rates of B. indicus cows managed under controlled breeding seasons, because it hastens resumption of cyclicity in postpartum cows and attainment of reproductive maturity in heifers (Baruselli et al., 2018). The literature lacks studies that evaluate the association between level of B. indicus genetics and reproduction of beef cattle. This paucity of data is comprehensible because of the difficulty in finding a herd containing a representative and balanced number of cows within a known B. indicus spectrum, ranging from 0% to 100%, reared contemporaneously.

A population composed of straightbred and crossbred cows and sires that interbreed constitutes a multibreed population (Elzo and Wakeman, 1998). An Angus-Brahman multibreed population was established at the University of Florida in 1988 for purpose of genetic studies. Since its inception, the multibreed herd has been managed in yearly controlled breeding seasons, in which the genetic structure of the herd is maintained. The herd contains a representative and balanced number of cows within a spectrum of each 1/32nd, or approximately 3-percentage points of B. indicus genetics. We assessed and compiled the reproductive records of this herd from 1989 to 2020 aiming to interrogate the B. indicus influence on the reproductive performance of beef cows. Our overarching hypothesis was that the increase in the proportion of B. indicus genetics of cows is negatively associated with the reproductive success of B. indicus × B. taurus cows. Specifically, the aims were to evaluate the association between the proportion of B. indicus genetics of cows and (i) P/AI at first insemination following a synchronized estrous cycle, (ii) the proportion of pregnant cows at the end of the breeding season, (iii) interval between entering the breeding season and attainment of pregnancy.

Materials and Methods

Cows and mating strategy

Institutional Animal Care and Use Committee (IACUC) approval was not requested as data were obtained from a preexisting database. The study was a retrospective cohort study in which the factor studied was the level of B. indicus genetics of cows, and the responses of interest were those related to reproductive performance during a breeding season.

Data from individual cows were available during breeding seasons occurring between 1989 and 2020, for a total of 32 breeding seasons. However, the 2001 breeding season was disregarded due to incompleteness of the database. Thus, reproductive data relative to 31 breeding seasons were recovered from a herd belonging to the University of Florida. This herd is composed by animals with known admixtures of various proportions of Angus × Brahman genetics (Elzo and Wakeman, 1998). The crossbred cows (n = 6,503) are identified according to the 1/32nd (approximately 3-percentage points) fraction of Brahman and six different breed groups: 0%–19% (n = 1,216), 21%–34% (n = 1,085), 38% (Brangus) (n = 900), 41%–59% (n = 1,435); 63%–78% (n = 868) and 81%–100% (n = 1,000). The cows were managed contemporaneously, under grazing conditions (Paspalum notatum, Cynodon dactylon, or Lolium) with ad libitum access to water and minerals to meet or exceed maintenance requirements. Between late fall and early winter (November or December), when the availability and quality of perennial pastures were insufficient to attend the nutritional needs of cows, they were fed hay or haylage and protein supplements, such as soy hull and whole cottonseed, or sometimes energetic supplements, such as gluten feed and citrus pulp, according to the availability, cost, and needs of the cows. In the middle of winter (late January or February), the cows also started to graze winter cultivated rye cereal or ryegrass pastures. The cows remained under the winter nutritional strategies until late spring (May), when the conditions of perennial pastures improved. Breeding events occurred between the end of February and the beginning of August. All cows were exposed to an estrus synchronization program, followed by artificial insemination (AI) based on estrus. Cows not detected in estrus were inseminated by timed AI (TAI). Estrus was synchronized using different hormonal protocols that use gonadotropin-releasing hormone and prostaglandin F2α with or without an intravaginal progesterone insert, or an injection of estradiol valerate combined with an ear implant containing norgestomet, or melengestrol acetate feed supplementation. Day 0 (D0) of each breeding season was the day when the cows received the first AI. In some breeding seasons, there was a second or third AI based on return to estrus, before introducing bulls for natural service (NS). However, information about the occurrence of these services within the groups of cows submitted to AI and breeding seasons was incomplete or missing, thus precluding its use for analysis.

The AI were performed using conventionally frozen-thawed semen from commercial bulls or semen collected from bulls from the multibreed herd. All mating sires (i.e., AI and NS) were from the six breed groups aforementioned and a skilled technician inseminated all cows within a given year. The herd was constructed and was perpetuated through a diallel crossbreeding scheme, that is, the six sire groups are reciprocally mated with the six dam groups (Komender, 1987; Elzo and Wakeman, 1998). Two to five bulls per breed group were used in the mating program per year (Elzo and Wakeman, 1998), depending on the number of allotments of cows in the herd within a given year. For the NS, each allotment cows submitted to AI were equitably divided in six sub-allotments that contained a representative sample of the six breed groups of cows, and these sub-allotments were reciprocally exposed to bulls of one of the six breed groups throughout the breeding season. The bull:cow ratio for the total number of cows exposed ranged from 1:15 to 1:30. Bulls were introduced into the sub-allotments of cows 21.0 d ± SD: 8.7 (range 0 to 66 d) after AI (D0) and maintained in pastures with cows up to the end of the breeding season. The breeding seasons lasted approximately 90-d (88.4 d ± SD: 16.7, range 51 to 142), considering the interval between AI and removal of bulls from the sub-allotments of cows. All the bulls used on NS passed the breeding soundness exams performed prior to the initiation of every breeding season. We defined the end of the breeding season as the day when bulls were removed from the sub-allotments of cows. The variability on breeding seasons occurred because cows that calved later in the calving season entered the breeding season later and had fewer days of breeding. Also, breeding seasons were shortened progressively over the years, from the initial to the most recent years.

Pregnancy diagnosis was performed via transrectal palpation approximately 45 d after the end of breeding season during the first 22 yr of the project. Starting in 2011, pregnancy was diagnosed by transrectal ultrasonography between 30 and 45 d after AI and again from 30 to 45 d after the end of each breeding season. When only one pregnancy diagnosis was performed, then the interval between AI and the calving date was used to determine if the pregnancy was the result of the AI. When this distinction was not possible, either because doubtful or inconsistent intervals, we disregarded the data for this cow for calculation of P/AI. The association between proportion of B. indicus genetics and reproduction was evaluated using P/AI and the proportion of pregnant cows at the end of the breeding season (P/AI+NS). The P/AI was calculated as the proportion of cows that became pregnant to AI either following detected estrus or TAI. The P/AI+NS was calculated as the proportion of cows pregnant at the end of breeding season either by AI or NS.

Data curation and records

Data from each breeding season were organized in Excel spreadsheets by the ranch management. Information collected included cow identification, age, parity category (lactation 0, nulliparous; lactation 1, primiparous; lactation 2, secundiparous, lactation > 2, pluriparous), Angus and Brahman genetic fraction of cows (0% to 100% B. indicus), breed group of cows according to the proportion of Brahman (0%–19%, 21%–34%, 38% (Brangus), 41%–59%, 63%–78% and 81%–100%), date of last calving, date of AI, days postpartum (DPP) at AI, breed group of bulls, duration of the breeding season, and the subsequent calving date. Whenever available, we recovered body condition score (BCS; 1 = emaciated to 9 = obese, using 1.0 increments) and body weight (BW) at the beginning of synchronization protocol, pregnancy diagnosis after AI and removal of bulls. Also, we extracted data on estrous responses to the synchronization protocols between 2011 and 2020, when this information was recorded. During this period, cows were strictly under the Select Synch + CIDR estrous synchronization program for AI (Lamb and Mercadante, 2016).

For cows that became pregnant to AI, the day of pregnancy in the breeding season was day 0. For those not pregnant to AI, then the day of pregnancy was estimated using the mean gestation length for each genetic group. First, we calculated the mean duration of gestation in cows that became pregnant to AI, according to genetic group, by subtracting from the calving date the date of AI (Supplemental Table S1). Those values were used to estimate the date of pregnancy for cows that became pregnant to NS by subtracting the duration of gestation, according to each genetic group, from the day of calving. Once the day of pregnancy was established, then we calculated the day in the breeding season when cows became pregnant. The 21-d cycle pregnancy rate was calculated as the ratio between the pregnancy at the end of the breeding season (0 or 1) and the number of 21-d estrous cycles during the breeding season until the cow became pregnant. The number of 21-d estrous cycles was calculated as the days in the breeding season when pregnancy was detected divided by 21. It was assumed that each estrous cycle lasts 21 d.

The experimental unit for analysis was the cow, resulting in a database containing 6,404 records of P/AI and 6,503 records for proportion pregnant at the end of the breeding season and interval to pregnancy. The disparity was because we were unable to determine if the resulting pregnancy was by AI for 99 records. Moreover, incomplete records resulted in disparate numbers of results for different independent and dependent variables analyzed. For example, estrous response was available only for 1,685 records. For the 6,503 records, each breeding season contributed with 103–345 records, averaging 210 records per season. Between 1989 and 2017, heifers received the first AI with approximately 2 yr of age. Thus, until 2017, the database was only composed of 2-yr-old nulliparous (2.1 ± 0.08 yr, mean ± SD; 1.8 to 2.3), 3-yr-old primiparous (3.2 ± 0.09 yr; 2.8 to 3.5), 4-ye-old secundiparous (4.2 ± 0.09; 3.9 to 4.5), and pluriparous cows (7.2 ± 2.2; 4.0 to 16.5). Parity category of cows was defined according to the age at the AI service and the calving history. In 2018, the heifers started to receive the first AI service with approximately 1 yr of age resulting in alterations in primiparous and secundiparous categories. These three new categories (n = 523) are not part of database analyzed, but the P/AI and P/AI+NS results are depicted in a Supplemental Table S2 as a reference.

Ultrasonography evaluations

In the 2019 and 2020 breeding seasons, at the beginning and end of the estrus synchronization protocol, the ovaries of 250 pluriparous cows were scanned by transrectal ultrasonography (MicroMaxx, SonoSite, Inc., Bothell, WA; equipped with a 6.1 MHz linear transducer) to determine the diameter of largest follicle and presence of a corpus luteum. The exams were conducted by a single operator. The results of these evaluations are depicted in Supplemental Table S4.

Statistical analysis

All statistical analyses were conducted using SAS (version 9.4, SAS Institute Inc., Cary, NC, USA), considering cow as the experimental unit. The binary dependent variables (P/AI, P/AI+NS, estrous response, and detection of corpus luteum at ultrasonography) were analyzed by logistic regression using generalized linear mixed-effects models with the GLIMMIX procedure of SAS. In all models, the Kenward-Roger option was used to adjust denominator degrees of freedom to calculate the F-values.

Initial models to analyze binary data were built including the fixed effects of breed group (0%–19% vs. 21%–34% vs. 38% (Brangus) vs. 41%–59% vs. 63%–78% vs. 81%–100%), parity order (nulliparous vs. primiparous vs. secundiparous vs. pluriparous), and the interaction between breed group and parity order, and the random effect of breeding season (1–31).

Three additional statistical models were built to analyze P/AI. These models included the same fixed and random effects of the initial model, but also included the fixed effect of BCS category on D0 (≤ 4.50 vs. ≥ 5.0, defined according to the median [BCS = 5]), or the fixed effect of DPP category (30 to 60 vs. 61 to 90 vs. 91 to 120, defined according to each 30 d interval), or the fixed effect of estrous response (detected in estrus vs. not detected in estrus). For those three additional statistical models, the interaction between breed group and BCS category, or breed group and DPP category, or breed group and estrous response was included in the respective model if it resulted in P < 0.20. Finally, responses were analyzed using the initial model, but replacing breed group with the linear covariate of the proportion of B. indicus genetics (0% to 100%) to access its association with pregnancy rates or estrus response according to each 3% increment on B. indicus genetics of cows. The association was analyzed fitting models with proportion of Bos indicus genetics as linear and quadratic covariates and their interactions, and only factors with P < 0.20 remained in the final model.

The interval to pregnancy was analyzed according to the Cox’s proportional hazard regression using the PHREG procedure of SAS. The variable “time” was the interval in days from starting the breeding season to pregnancy. The statistical model included the fixed effects of breed group, parity order, the interaction between breed group and parity order, and breeding season; and only factors with P < 0.20 remained in the final model. Cows that did not become pregnant in the breeding season were censored on the last day of the respective season or if sold or died during the breeding season, whichever happened first. Adjusted hazard ratios (AHR) were generated using pluriparous or 81%–100% breed group as the reference. The LIFETEST procedure of SAS was used to calculate the median and mean interval to pregnancy. The survival plots were generated with GraphPad Prism, version 9.2.

The 21-d cycle pregnancy rate was analyzed by logistic regression with the GLIMMIX procedure of SAS fitting a binomial distribution. The model included the fixed effects of breed group, parity order, and the interaction between breed group and parity order, and the random effect of breeding season; and only factors with P < 0.20 remained in the final model.

The continuous-dependent variables calving interval, diameter of largest follicle at the beginning and end of synchronization program, and changes in BW and changes in BCS during the breeding season were analyzed by linear mixed-effects models using the MIXED procedure of SAS. The model included the fixed effects of breed group, parity order, and the interaction between breed group and parity order, and the random effect of breeding season. Whenever data were obtained from a single parity order, then parity and the interaction between breed group and parity order were removed from the statistical model.

Significance was set at P ≤ 0.05, and tendency for significance at 0.05 < P < 0.10. Results are reported as LSMEANS ± SEM unless otherwise indicated. Pairwise comparisons were performed after adjustment by the method of Tukey. In cases of interaction, the SLICE command was incorporated to the LSMEANS statement to generate the comparisons and respective P-values using Fisher’s protected least significant difference within categories of fixed effects in the model.

Results

Animals

The change of BW (n = 2,098) and BCS (n = 1,761) during the first-, D0 to D39, and last-half, D39 to D91, period of the breeding season did not differ (P > 0.10) among breed groups, as depicted in the supplemental material (Figure S1, Panels A–D). There was an association between parity order and change of BW and BCS (P < 0.001). For the BCS changes (Figure S2, Panels A and B), the association was significant only at the last-half (D39 to D91) of breeding season (Figure S2, Panel B). Nulliparous cows gained BCS when compared with any other of the three parity orders that maintained BCS. For BW changes (Figure S2, Panels C and D), nulliparous cows gained weight at the first-half (D0 to D39, Panel C) and last-half (D39 to D91, Panel D) periods of the breeding season when compared either with primiparous, secundiparous, or pluriparous cows. In the first-half of the breeding season, the three parity orders of suckled beef cows lost weight, whereas in the last-half of the breeding season, these parity orders maintained or gained weight. Moreover, the BCS, measured on D0 of the breeding season, was not associated (P > 0.10) with P/AI (BCS ≤ 4.5: 48.5% [383/789] vs. BCS ≥5.0: 46.4% [1,362/2,935]). Only 1.7% [63/3,724] of the cows presented BCS < 4.0.

Association between proportion of B. indicus genetics of cows and reproduction

There was an association (P < 0.0001) between the proportion of B. indicus genetics and P/AI, regardless of parity order (Table 1). The P/AI of the breed group with 0%–19% of B. indicus genetics was greater (P < 0.05) than every other breed group, with exception of that with 41%–59% of B. indicus genetics. On the other hand, the breed group with 81%–100% of B. indicus genetics had the smallest P/AI, which did not differ only from the breed group with 63%–78% of B. indicus genetics. Similarly, when P/AI was analyzed with the proportion of B. indicus genetics as continuous covariate in the model, there was a linear and negative association between the proportion of B. indicus genetics and P/AI (Figure 1, P < 0.0001). The equations for the P/AI predicted probabilities are described in the supplemental material (equations, item 1) according to the parity orders. In spite of this negative relationship, P/AI+NS did not differ among breed groups. There was, however, an association between parity order and P/AI+NS (P = 0.006). The Tukey adjusted method of multiple comparisons revealed the P/AI+NS of primiparous cows (86.5% [1,227/1,476]) was smaller than that of pluriparous (88.7% [2,311/2,664]). Whereas P/AI+NS of nulliparous (90.7% [1,232/1,395]) and secundiparous (88.4% [827/969]) cows did not differ from any parity order. The detailed P/AI and P/AI+NS data according to breed group and parity order are depicted in Supplemental Table S3 as a reference material.

Table 1.

Association between the proportion of Bos indicus genetics of cows and pregnancy per AI (P/AI) at first AI and pregnancy at the end of the breeding season

Item	Proportion of B. indicus						P-value
	0%–19%	21%–34%	38% (Brangus)	41%–59%	63%–78%	81%–100%	Breed	Parity	Breed × parity
P/AI1, % (n/n)	51.6a	44.3b,c	44.7b,c	46.3a,b	39.5c,d	37.4d	<0.0001	0.24	0.82
	(631/1,201)	(489/1,071)	(405/893)	(655/1,415)	(349/855)	(362/969)
P/AI + NS2, % (n/n)	89.2	88.6	88.7	88.6	88.5	88.2	0.99	0.006	0.84
	(1,056/1,216)	(935/1,085)	(773/900)	(1229/1,435)	(749/868)	(855/1,000)

Parity = 2-yr-old nulliparous, 3-yr-old primiparous, 4-yr-old secundiparous, or pluriparous.

^a,b,c,d Values without a common superscript differed between breed groups (P < 0.05) by applying Tukey test for multiple comparisons.

¹P/AI: Proportion of cows receiving AI, after estrous synchronization, that were pregnant to this insemination. Every cow submitted to the estrous synchronization was inseminated based on estrus or following timed TAI.

²P/AI + NS: Proportion of cows pregnant at the end of breeding season as result of AI or natural service.

Figure 1.

Predicted probabilities of pregnancy per AI according to the proportion of Bos indicus genetics of cows (P < 0.0001). Crossbred cows (Brahman × Angus) were submitted to estrus synchronization protocols for artificial insemination based on detected estrus or TAI. There was a negative relationship between the proportion of Bos indicus genetics of cows and P/AI, regardless of parity order of the cows (nulliparous, primiparous, secundiparous, and pluriparous). The open gray dots are the predicted probability according to a final model containing the fixed effect of parity order, breed group and year of the breeding season as random effect. The solid black line is the average of these predicted probabilities. The equations for the predicted probabilities according to the parity order are described in the supplemental material (equation, item 1).

Breed group and parity order were associated (P < 0.001) with the interval between the beginning of the breeding season and attainment of pregnancy independently. The interval to attain pregnancy was lengthened as the proportion of B. indicus genetics increased (Figure 2). Cows in the breed group with 0%–19% of B. indicus genetics had 1.20 times the hazard to become pregnant during the breeding season than the cows in the breed group with 81%–100% of B. indicus genetics (Table 2). This difference in rate of pregnancy resulted in 25 extra median days to pregnancy for cows with 81%–100% compared with cows with 0%–19% of B. indicus genetics (Table 2). Nevertheless, the 21-d cycle pregnancy rate was not associated (P = 0.11) with breed group. Regarding the parity order, compared to pluriparous, nulliparous, primiparous, and secundiparous cows took the longer time to attain pregnancy during the breeding season (Figure 3 and Table 3).

Figure 2.

Differences on Kaplan–Meier survival curves for time to pregnancy up to 140 d of the breeding season according to the proportion of Bos indicus genetics of cows (Brahman × Angus), P < 0.0001. Pregnancy results of 30 yearly breeding season containing the six contemporaneous breeding groups were compiled and analyzed (n = 6,385). All cows were submitted to an estrus synchronization protocol for AI service at the beginning of breeding season and, subsequently, natural service during seasons lasting 88.4 d (51–142 d). The Sidak test for multiple comparisons revealed differences (P < 0.05) between the survival curve of 81%–100%BR group and every other group, except for the 63%–78%BR group. The survival curve of 0%–19%BR differed (P < 0.01) from the ones of 38%BR and 63%–78%BR, only. Thus, in general, cows with 81%–100% of B. indicus genetics took the longest interval of time to attain pregnancy during the breeding season.

Table 2.

Association between proportion of Bos indicus genetics of cows and rate of pregnancy during the breeding season

Item	Proportion of B. indicus						P-value
	0%–19%	21%–34%	38% (Brangus)	41%–59%	63%–78%	81%–100%	Breed	Parity	Breed × parity
Cows, n	1,197	1,055	885	1,414	857	977
Interval to pregnancy 1 , d
Mean ± SEM	28.6 ± 1.2	30.7 ± 1.3	29.3 ± 1.2	31.3 ± 1.1	30.5 ± 1.2	34.0 ± 1.2
Median	0.0	18.0	20.0	18.0	22.0	25.0
(CI)	(0–1)	(11–20)	(14–24)	(12–22)	(18–25)	(23–26)
AHR	1.20*	1.11*	1.09	1.13*	1.07	1.00	0.007	<0.001	NS
(CI)2	(1.09–1.31)	(1.01–1.22)	(0.98–1.20)	(1.04–1.24)	(0.97–1.19)	(ref.)
21-d cycle pregnancy rate3, %	46.4	43.8	43.4	44.4	42.9	41.6	0.11	<0.0001	0.15
Calving interval, d	375.8 ± 1.3	375.0 ± 1.4	377.6 ± 1.5	376.9 ± 1.3	378.4 ± 1.5	379.0 ± 1.5	0.06	<0.0001	NS

Parity = 2-yr-old nulliparous; Primiparous = 3-yr-old cow; Secundiparous = 4-yr-old cows, or Pluriparous = ≥5-yr-old cows.

¹The day of AI was considered D0 of the breeding season. The interval to pregnancy due to NS was estimated by subtracting the mean gestation length, according to the breed group (Table S1), from the day of calving interval of subsequent calving date minus the AI date.

²AHR = adjusted hazard ratio; CI = confidence interval.

³Proportion of cows that became pregnant every 21-d in the breeding season.

^*Within a row, means with an asterisk differ from the reference (ref.) group (P < 0.05).

Figure 3.

Differences on Kaplan–Meier survival curves for time to pregnancy up to 140 d of the breeding season according to parity order, P < 0.0001. Pregnancy results of 30 yearly breeding seasons containing the six contemporaneous breed groups of crossbred cows (Brahman × Angus) were compiled and analyzed (n = 6,385). All cows were submitted to an estrus synchronization protocol, AI service, and, subsequently, natural service, during seasons lasting 88.4 d (51–142 d). The Sidak test for multiple comparisons revealed differences (P < 0.05) between the survival curve of pluriparous vs. nulliparous, pluriparous vs. primiparous, and pluriparous vs. secundiparous cows. Thus, pluriparous cows took the shortest interval of time to attain pregnancy during the breeding season, regardless of breed group.

Table 3.

Association between parity group of cows and rate of pregnancy during the breeding season

Item	Parity category				P-value
	Nulliparous	Primiparous	Secundiparous	Pluriparous	Parity effect
Cows, n	1,381	1,438	959	2,607
Interval to pregnancy 1 , d
Mean ± SEM	33.2 ± 1.1	32.5 ± 1.0	31.0 ± 1.3	27.6 ± 0.7
Median	25.0	21.0	18.0	15.0
(CI)	(25–26)	(17–24)	(11–21)	(12–18)
AHR	0.88*	0.89*	0.96	1.00	<0.001
(CI)2	(0.82–0.95)	(0.83–0.95)	(0.89–1.04)	(ref.)
21-d cycle pregnancy rate 3 , %	41.0c	41.6b,c,§	45.2a,b,§	47.3a	<0.0001
Calving interval, d	–	388.0a ± 1.2	372.7b ± 1.3	370.7b ± 1.1	<0.0001

¹The day of AI was considered D0 of the breeding season. The interval to pregnancy due to NS was estimated by subtracting the mean gestation length, according to the breed group (Table S1), from the day of calving interval of subsequent calving date minus the AI date.

²AHR = adjusted hazard ratio; CI = confidence interval.

³Proportion of cows that became pregnant every 21-d in the breeding season.

^*Within a row, means with an asterisk differ from the reference (ref.) group (P < 0.05).

^a,b,c Values without a common superscript differed between breed groups (P < 0.05) or with the same symbol tended to differ (P = 0.08) by applying Tukey test for multiple comparisons.

Influence of days postpartum and estrus expression on P/AI

The P/AI was altered according to the interaction between DPP and parity order (P = 0.03; Figure 4), regardless of breed group. There was a positive association between the increase in DPP and P/AI for primiparous and secundiparous cows, but not for pluriparous cows. In a second model, the P/AI, was associated with the occurrence of estrus before TAI (P < 0.0001), regardless of breed group or parity order. Cows showing estrus and inseminated based on estrus had greater P/AI (64.1% [579/915]) than those not showing estrus and submitted to TAI (28.1% [213/757]). Estrous response, nevertheless, was associated with breed group and parity category (P < 0.001). The estrous response was decreased as the proportion of B. indicus of cows increased (Figure 5A), and it was the least for primiparous cows (Figure 5B). Similarly, when replacing the breed group for the proportion of B. indicus genetics of cows as continuous covariate in the model, there was a linear and negative association between proportion of B. indicus genetics and estrous response (Figure 6, P < 0.0001). The equations for the estrous response predicted probabilities are described in the supplemental material (equations, item 1) according to the parity orders.

Figure 4.

Association of category of days postpartum (DPP) with pregnancy per AI (P/AI) according to the parity order of cows. The parity by DPP interaction effect (P = 0.03) revealed that the positive effect of DPP on P/AI occurred among primiparous and secundiparous cows, but not among pluriparous cows. The interaction effect was interpreted by using the SLICE option of the LSMEANS statement of SAS. In the same statement, we applied the DIFF option (Fisher’s protected least significant difference) to determine the difference across DPP within parity order. ^a,b,cWithin parity order, values without a common superscript differed between DPP categories (P < 0.05), and values with the same symbol tended to differ (P < 0.10).

Figure 5.

Estrous response to estrus synchronization protocol according to the proportion of Bos indicus genetics (A) or parity order (B) of cows. There were breed group and parity order effects (P < 0.0001) on estrus response. Means without a common superscript were different (P < 0.05) or with the same symbol tended to differ (P = 0.07) by applying Tukey test for multiple comparisons. Estrous response was reduced according to the increase in proportion of Bos indicus genetics of cows; and it was the lowest among primiparous cows.

Figure 6.

Predicted probabilities of estrous response according to the proportion of Bos indicus genetics of cows (P < 0.0001). Cows (Brahman × Angus) were submitted to estrus synchronization protocols for artificial insemination based on detected estrus or TAI. There was a negative association between the proportion of Bos indicus genetics of cows and estrous response, regardless of parity order of the cows (nulliparous, primiparous, secundiparous, and pluriparous). The open gray dots are the predicted probability according to a final model containing the fixed effect of parity order, breed group, and their interaction, and year of the breeding season as a random effect. The solid black line is the average of these predicted probabilities. The equations for the predicted probabilities according to the parity order are described in the supplemental material (equation, item 2).

The proportion of B. indicus genetics of cows and ovarian characteristics

The data of this section are depicted in the Supplemental Table S4. On D-10, the diameter of the largest ovarian follicle and proportion of cows bearing a corpus luteum were similar across breed groups. On D-3, the proportion of corpus luteum was also similar across the breed groups, but the diameter of the largest follicle was decreased as the proportion of B. indicus genetics of cows increased.

Discussion

This study assessed 31 yr of accumulated data from a herd containing an admixture of known proportions of Angus × Brahman genetics raised contemporaneously under the same subtropical environmental and nutritional conditions. The proportion of B. indicus (Brahman) genetics was negatively associated with P/AI of crossbred cows submitted to progesterone/progestin-based estrous synchronization protocols without exogenous estradiol as an inducer of ovulation. Likewise, estrous response to the protocol, a main fertility marker, was also associated negatively with the proportion of B. indicus genetics. Although supplemental estradiol, when used as inducer of ovulation, stimulates estrus expression and fertility in B. indicus cattle (Sá Filho et al., 2011; Martins et al., 2017), its use is not allowed in many countries, including the USA. Despite the suboptimal P/AI, the increase in B. indicus genetics of cows did not compromise the proportion of pregnant cows at the end of the breeding season. However, it lengthened the interval between the beginning of the breeding season and pregnancy. As much as an extra 25 d was necessary to attain 50% pregnancy rate during the breeding season when the proportion of B. indicus increased from 0%–19% to 81%–100%. In other words, this lengthened interval to pregnancy likely impacted negatively the kilograms of calves weaned, as late calving cows usually wean younger and lighter calves. Thus, under contemporaneous environmental and nutritional conditions, the B. indicus genetics dampened the reproductive performance of crossbred cows by reducing P/AI and lengthening time to pregnancy during the breeding season.

There was a negative association between the proportion of B. indicus genetics of cows and P/AI. The elevated proportion of cows in anestrus postpartum or prepubertal heifers at the beginning of the breeding season have been referred as two of the main causes of poor reproductive performance among B. indicus cattle reared on the tropics (Baruselli et al., 2004; Nogueira, 2004). These physiological conditions are typically associated to sub-optimal nutrition associated with low-quality tropical roughage (Baruselli et al., 2004; Nogueira, 2004; Ribeiro et al., 2015). Low nutrition planes have a negative influence on LH release in cows, prolonging the postpartum anestrus (Rasby et al., 1992; Rekwot et al., 2004). Similarly, in nulliparous cows, undernutrition negatively affects the LH pulse generation system, at the hypothalamic level, delaying the first ovulation (Gasser et al., 2006). From neuro-endocrine standpoint, differences between B. indicus and B. taurus cattle may also contribute to the prolonged anestrus postpartum and later attainment of reproductive maturity among B. indicus cattle. For example, ovariectomized Brahman cows had a reduced LH-response when treated with exogenous estradiol compared to the counterparts Brahman × Hereford and Hereford cows (Rhodes et al., 1978). However, at least at the nutritional standpoint, herein, there was no evidence that cows with greater proportion of B. indicus genetics were in nutritional disadvantage in relation to the cow with smaller proportion of B. indicus genetics. Cows were mostly managed to maintain their BW and BCS under a grazing management system, as evidenced by the variations on BCS and BW change along the breeding season. Importantly, during the winter periods, when the availability and quality of perennial pastures are lower, cows were supplemented and grazed winter pastures. As a consequence, at the beginning of the breeding season that took place during late winter and early spring, only 1.7% of the animals presented a very low BCS (i.e., <4.0). The ovarian ultrasonography exam conducted in a small subset of pluriparous cows indicated that the proportion of cows bearing a corpus luteum in the beginning of breeding season was similar between the breed groups, hence, suggesting the cyclicity was similar across cows with different proportion of B. indicus genetics. Thus, in this study, a prevalence of anestrus due to nutritional restriction did not seem the cause of suboptimal P/AI among cows with a greater proportion of B. indicus genetics.

The negative relationship between proportion of B. indicus genetics of cows and P/AI was not altered by DPP nor by parity order, but there was an interaction between parity order and DPP on P/AI. For primiparous and secundiparous cows, the increase in DPP was associated with increments on P/AI. Parity order and DPP are two important factors related with P/AI in estrous synchronization protocols (Stevenson et al., 2003, 2015; Echternkamp and Thallman, 2011). Generally, for suckled beef cows, the primiparous present the lowest P/AI as the result of an elevated proportion of cows still in anestrus in the beginning of synchronization program (Stevenson et al., 2003, 2015). In primiparous cows, the body growth requirements postpone the resumption of cyclicity (Short et al., 1990). In the same sense, because DPP are inversely related to the proportion of cows in anestrus, cows with less DPP tend to present reduced P/AI (Stevenson et al., 2000, 2003; Echternkamp and Thallman, 2011). This explains the beneficial effect of DPP on the P/AI of primiparous and secundiparous cows. Another challenging parity order is the Nulliparous cows, especially for straightbred B. indicus heifers, because of the excess of reproductively immature heifers at the beginning of breeding season as a 2-yr-old (Lima et al., 2020) and 1-yr-old or yearling (Martins et al., 2021). As yearlings, we verified that only 5.2% of Brahman heifers were cycling at the beginning of breeding season and P/AI (14.9%) and P/AI+NS after an approximately 90-d breeding season (53.1%) were overly low (Martins et al., 2021). The data presented in the supplemental tables from 2018 to 2020 referring to younger parity orders, add up to such reproductive inefficiency and help to explain why the nulliparous cows were inseminated historically only at 2 yr of age. Taken all together, it is possible to consider that parity of cows and DPP are related to the occurrence of anestrus. Primiparous cows are expected to be one of the parity orders with the least proportion of cyclic cows and DPP is inversely related with proportion of cyclicity. Based on that, the absence of interaction between breed group and DPP or breed group and parity order, indicated to us that the cyclicity status was not an explanatory factor for the suboptimal P/AI between cows with greater proportion of B. indicus genetics. In other words, independently of the cyclicity status of cows, P/AI was always less among cows with greater proportion of B. indicus genetics submitted to synchronization protocols.

Greater proportion of B. indicus genetics of cows reduced estrous response to the synchronization protocol as much as 27.9%; and primiparous cows presented the lowest proportion of estrus. This relationship held true, regardless of parity category. Thus, contributing to our idea that cyclicity was not a determining explanatory factor for the reduction in P/AI as the proportion of B. indicus genetics of cows increased. This is based on the fact estrus is an important fertility marker in synchronization protocols (Perry et al., 2007; Sá Filho et al., 2011; Jinks et al., 2013; Richardson et al., 2016). Indeed, we observed P/AI of cows inseminated in estrus was 2.3-folds greater than those non-estrus cows submitted to TAI. Strategies that stimulate follicular development during synchronization protocols improve estrous response and P/AI of Bos indicus-influenced cattle. Among others, there are three effective strategies. The first, consists of lowering the circulating levels of progesterone. When Bos indicus cows are treated with exogenous progesterone, such as CIDR, the circulating levels of progesterone are greater than in Bos taurus (Carvalho et al., 2008; Batista et al., 2020), and this leads to reduced follicular growth and estrous response (Dias et al., 2009; Peres et al., 2009; Martins et al., 2014). Common practices to decrease progesterone concentrations during the protocol consist of advancing the PGF injection (Carvalho et al., 2008; Dias et al., 2009; Peres et al., 2009; Williams and Stanko, 2020) and/or inserting an intravaginal progesterone device that was used previously (Dias et al., 2009; Peres et al., 2009; Martins et al., 2014). Such practices improved P/AI of B. indicus cyclic cows (Dias et al., 2009; Peres et al., 2009; Williams and Stanko, 2020). The second strategy consists in injecting exogenous estradiol, such as estradiol cypionate, in the end of synchronization protocol. The addition of estradiol stimulates estrous response and favors P/AI in beef cattle (Sá Filho et al., 2011; Jinks et al., 2013; Martins et al., 2017). The third strategy consists of providing LH-stimulus support for the final dominant follicular growth, that is especially beneficial for suckled B. indicus cows. The LH-secreting capacity in B. indicus is lower than that of B. taurus cows at comparable reproductive states during the postpartum period (D’Occhio et al., 1990). Common practices to provide LH-stimulus support consist of administering equine chorionic gonadotropin (eCG) or adopting temporary calf removal, such as for 48 h, in the end of synchronization protocol (Bó et al., 2003; Baruselli et al., 2004; Sá Filho et al., 2009). These practices increase P/AI when suckled Bos indicus cows are submitted to TAI. Overall, the use of estradiol-progesterone based protocols incorporating one or more of the modifications mentioned have warranted satisfactory P/AI (at least 50%) of straightbred B. indicus cows submitted to synchronization of ovulation for TAI in South America (Meneghetti et al., 2009; Sá Filho et al., 2009; Baruselli et al., 2012). Thus, the suboptimal P/AI between cows with greater proportion of B. indicus genetics in the present study is likely related to the lack of adoption of any of those modifications, designed to attend the reproductive physiological particularities of B. indicus. It is important to highlight that, independently of breed group, cows inseminated based on estrus presented greater P/AI, which indicates subfertility seemed not to be a condition of cows with a greater proportion of B. indicus genetics.

The proportion of B. indicus genetics of cows was not associated with the final proportion of pregnant cows at the end of the breeding season. However, cows with greater proportion of B. indicus genetics took longer to become pregnant during the breeding season. An initial delay during the first 30 d of the breeding season was expected as a consequence of the suboptimal P/AI, but cows with greater proportion of B. indicus genetics still became pregnant at a slower rate. Around D60 of the breeding season, cows with a lower proportion of B. indicus genetics reached a plateau point, in which not many cows became pregnant. Such plateau for the cows with 81%–100% was only reached at approximately D80 of the breeding season. In the same sense, according to the median results, this group of cows needed an additional 25 d period to attain a 50% of pregnancy during the breeding season. In contrast, more than 50% of cows with 0%–19% of B. indicus genetics were already pregnant on D0 of the breeding season. Explanation for such persistence of reduced pregnancy rates during the initial two-thirds of the breeding season is unclear. Our interpretation is that this delay to attain pregnancy is not due to an inherent subfertility condition of cows with greater B. indicus genetics, because P/AI of cows inseminated based on estrus was similar, regardless of the breed group. An obvious explanation for such results would be a larger proportion of cows in anestrus or reproductively immature at the beginning of the breeding season. However, the lack of breed group by parity order interaction on time to attain pregnancy indicated that the reproductive performance of cows with greater proportion of B. indicus genetics was always less than those with a smaller proportion of B. indicus, irrespective of differences of cyclicity related to parity order. Thus, other intrinsic characteristics of B. indicus cattle may have contributed to reduce the number of cows coming in estrus every 21 d, delaying time to pregnancy along the breeding season. The 25 d-period delay on time to pregnancy associated with longer pregnancy duration (291 d vs. 283 d) implies that fewer cows with greater proportion of B. indicus genetics calved at the first 30 d of the calving season; and cows calving earlier wean heavier calves (Rodgers et al., 2012; Lamb and Mercadante, 2016). Moreover, cows conceiving later in the breeding season were more subjected to winter nutritional restriction during late pregnancy. This is based on the fact that most breeding events occurred during the spring season, hence, resulting in a winter calving season. It is during mid-to-late pregnancy that the fetus undergoes rapid growth and therefore places a significant toll on the cows (Ferrell et al. 1976; Hoffman et al., 2017). Indeed, feed-restricted cows at the last trimester of pregnancy calved lighter calves with worse performance at weaning than those from adequately fed cows (Larson et al., 2009; Bohnert et al., 2013). Altogether, cows with greater proportion of B. indicus genetics took longer to attain pregnancy during the breeding season and also presented a longer pregnancy duration. Expected consequences are that these cows calved later, and also weaned younger and lighter calves; a negative impact on fetal programming is also anticipated.

In conclusion, in contemporaneous conditions, the proportion of B. indicus genetics did not influence the final proportion of cows pregnant at the end of the breeding season. However, greater proportions of B. indicus genetics of cows was associated with a lengthened interval between entering the breeding season and attaining pregnancy. This negative relationship was explained partially by the poorer pregnancy rates obtained in response to estrus synchronization protocols developed for B. taurus cattle, which do not consider the reproductive physiological particularities of B. indicus cattle. Estrous response at the end of the synchronization protocol was reduced as the proportion of B. indicus genetics of cows increased. Cows showing estrus before AI had greater P/AI than those not showing estrus, irrespective of proportion of B. indicus genetics of cows. Thus, subfertility did not seem to be an issue between cows with different proportion of B. indicus genetics. Overall, the increase in the proportion of B. indicus genetics of cows dampened reproductive performance by lengthening the time to pregnancy during the breeding season, but not the final proportion of cows pregnant at the end of the breeding season. Thus, similar calf-crops should be granted across different proportions of B. indicus genetics of cows, but cows with a greater proportion of B. indicus genetics are expected to calve later during the calving season.

Supplementary Data

Supplementary data are available at Journal of Animal Science online.

Figure S1. Scatter plots depicting the changes in body condition score (BCS; panels A and B) and body weight (BW; panels C and D) according to the proportion of B. indicus genetics of cows at the first- (D0 to D39) and last- (D39 to D91) half of 90-d breeding seasons (88.4 d, 51–142 d). The open gray dots represent the individual change of each cow (n = 1,761 for BCS and n = 2,098 for BW), while the solid black bar and value represent, respectively, the mean and LSMEANS within each breed group of cows. There was not an association (P > 0.10) between the proportion of B. indicus genetics of cows and changes in BW or BCS.

Figure S2. Scatter plots depicting the changes in BCS (panels A and B) and BW (panels C and D) according to parity order of cows at the first-half (D0 to D39) and last-half (D39 to D91) of the 90-d breeding seasons (88.4 d, 51–142 d). The open gray dots represent the individual change of each cow (n = 1,761 for BCS and n = 2,098 for BW), whereas the solid black bar and value represent, respectively, the mean and LSM within each breed group of cows. There was an association between parity order of cow and change of BCS at the last-half of breeding season and between parity order and change of BW at the first- and last-half of the breeding season (P < 0.001). ^a,b,c Values without a common superscript differed between parity orders (P < 0.05).

Glossary

Abbreviations

AI

artificial insemination

AHR

adjusted hazard ratio

BCS

body condition score

DPP

days postpartum

eCG

equine chorionic gonadotropin

GnRH

gonadotropin-releasing hormone

LH

luteinizing hormone

NS

natural service

P/AI

pregnancy per AI

P/AI+NS

pregnancy per artificial insemination and natural service

PGF

prostaglandin F2α

TAI

timed artificial insemination

Conflict of Interest

The authors report no declarations of interest.