Insemination timing influences reproductive performance rather than eCG (Equine Chorionic Gonadotropin) or synchronisation protocol in Murciano-Granadina does

Abstract

Counteracting goat reproduction seasonality and ensuring acceptable fertility rates are critical points of the success of selection of synchronisation protocols. The present study aims to investigate the factors that affect the performance of specific protocols to increase the probability of pregnancy results in the Murciano-Granadina breed within the official breeding program and way in which such factors interrelate. This study evaluated 11,008 artificial inseminations carried out between 2010 and 2013, using three different synchronisation protocols and eCG combinations. A categorical regression model (CATREG) was used to determine the impact of farm, doe age, birth number, kidding type, kid number, number of alive and dead born kids, synchronisation protocol, kidding month, type of semen, kidding season, eCG concentration, and their inter-actions. Results indicated that all variables, except for the number of alive and dead kids and semen type, significantly (p < 0.05) contributed to explaining the variability in pregnancy rates. Conclusively, rather than the role of eCG levels and artificial insemination (AI) protocol separately, Farm x eCG x AI Protocol interaction and the multifactorial influences on pregnancy rates should be considered when selecting does for insemination to maximise the outcomes of reproductive performance.

HIGHLIGHTS

• Farm × eCG × AI Protocol interaction and the multifactorial influences on pregnancy rates should be considered when selecting does for insemination

• Main factors to take into account would be those related to seasonality (kidding month, IA season), and also birth type

• It would be necessary adjust protocols (short term protocols has shown to have higher fertility) and eCG dosage for each farm

Keywords:

• Does
• fertility
• Murciano-Granadina breed
• synchronization protocol

Introduction

In recent years the demand for goat-derived products has considerably increased worldwide, which has led to the need to implement new selection technologies within breeding programs seeking the improvement of productivity and profitability of goat farms (Pizarro Inostroza et al. 2019, 2020a, 2020b). A fundamental tool for the dissemination of genetic improvement is artificial insemination (AI) which allows to increase the rate of genetic progress, farrowing control, planning based on market demands, interconnection of herds, reduction of reproductive diseases and also managing to reduce almost completely the influence of seasonality in reproduction and consequently the fluctuation of product (O’Hara et al. 2010; Cseh et al. 2012; Arrebola et al. 2014).

The seasonality of reproduction is present in most breeds of goats that inhabit in temperate latitudes which modulates sexual behaviour and biological functions through its stimulus on pituitary gland reaction to gonadotropin-releasing hormone (GnRH) and the pattern of luteinizing hormone (LH) secretion (Zarazaga et al. 2011; Zarazaga et al. 2019).

In Mediterranean areas, goats raised under the natural photoperiod, begin breeding season when day length starts to decrease (late summer or autumn), while it progresses to the end as day length increases (late winter or early spring) (Zarazaga et al. 2011). Such seasonality and time between births have been reported to affect the sustainability of goat farms and goat industry prices, which may indeed vary widely between the periods of maximum and minimum production with independence of the circumstances (Chemineau et al. 2007; Lopez‐Sebastián et al. 2014).

In does, the seasonality could be modified with different hormonal protocols which have been designed to control luteal function by the administration of exogenous intravaginal progesterone or progestogens devices associated with cloprostenol and gonadotropins (Texeira et al. 2016).

Although traditionally, such protocols were designed to last between 10 and 14 days (long-term protocols) (Luo et al. 2019) shorter protocols would be later designed as an attempt to improve fertility, follicular development, and ovulation (Camacho et al. 2019). In this context, a persistent dominant follicle, short follicular life, and asynchrony between oestrus and ovulation could be the cause of the reduction in fertility according to protocol, worsening in the long-term protocols (Viñoles et al. 2001). For instance, Martinez-Ros et al. (2018), concluded that administration of eCG (Equine Chorionic Gonadotropin) after either short- or long-term CIDR treatments assures the occurrence of fertile ovulations, whilst ovulatory success after protocols without eCG is dependent on the duration of the CIDR treatment; the best results being the yield after five days of device insertion in ewes, which supports our results.

The use of equine chorionic gonadotropin (eCG) is highly extended among these oestrus synchronisation protocols, regardless their duration, given it is directly related to the improvement of heat presentation rates (Santos et al. 2011), the distribution of oestrus (Luther et al. 2007), and the increase in ovulation rate and fertility (Shahneh et al. 2008).

In general terms, breeder association widespread the application of insemination programs, to simplify the synchronisation process of goats using a common synchronisation protocol (protocol length) in all their farms among which the only variable item is eCG concentration which depends on the artificial insemination season. However, as literature suggests, fertility rates are critically conditioned by the situation present in farms and/or their management, hence the application of potentially successful protocol can be truncated by an inappropriate or adverse combination of conditioning factors present in the environment where inseminations take place, which in turn hinders the maximisation of reproductive efficiency (Arrebola et al. 2014; Eşki et al. 2021).

Given the multifactorially affected nature of reproductive management success, the aim of the present study was to analyse conditioning factors of the reproductive performance and their interrelations to disentangle the background behind a higher reproductive efficiency in the Murciano-Granadina goat breed. The present study may help understand strategies to enhance the application and efficiency of large-scale reproductive management practices as consistent tool within breeding programs at an international scale through the example of an already internationally competitive autochthonous breed.

Material and methods

Sample selection

The original database consisted of a total of 11,008 AIs performed within the official breeding program of Murciano-Granadina goat breed corresponding to a period between 2010 and 2013. IAs had been performed during the period from 2010 to 2013 on 8126 Murciano-Grandina goats. Goats ranged from 9.07 to 142.27 months old at the moment of insemination and had been born from 1999 and 2012. Animals lacking a diagnosis of pregnancy or those for whom information in regards the IA protocol or the eCG concentration used was missing, were purged. Once the database was purged, records to be statistically processed consisted of 10,627 AI data from which 6420 data correspond to 8-day length protocol; 2860 data correspond to 13 days length protocol and 1160 data correspond to 14 days length protocol, respectively.

Murciano-Granadina goats and bucks considered in this analysis belonged to farms comprising the selection nucleus of the Murciano-Granadina goat breed. According to Pizarro Inostroza et al. (2020), farms of the selection nucleus are selected not only for their production and quality data but for the genetic connections among them. The decisión to chose farms within the oficial selection nuclues was made with the aim to have standardised farms participating in the study and reduce variability implied by such a relevant factor, all the farms considered in the study had received official National and International Sanitary Certificates. All farms were controlled and officially declared tuberculosis-free (C3), brucelosis-free (M4) (Order of 22 June 2018 and Directive 91/68/EEC), and scrapie RC free [Regulation (EC) No 999/2001 of the European Parliament and the Council]. Additionally, these farms followed voluntary control plans for caprine contagious agalactia (CCA) (National CCA Surveillance, Control, and Eradication Programme 2018–2020) and caprine arthritis encephalitis (CAEV) (Order AYG/287/2019 of 28 February 2019). Goats were clinically examined by an official veterinarian, and individuals presenting signs of illness or disease were officially declared, removed from the herds, and discarded from the analyses. Permanent stabling practices were followed by all farms considered, and ad libitum water, forage, and supplemental concentrate were provided.

Synchronisation protocols

Prior to the implementation of a synchronisation protocol, all the goats were subjected to transabdominal ultrasonography to discard does with possible gestation or pseudo-gestation. Similarly, the commercial brands and administered hormonal doses were the same for each synchronisation protocol. The protocols used were as follows:

8 days protocol (P8):

• DAY 1: Application of intravaginal fluorogestone acetate 45 mg sponge (Chrono-gest, MDS Animal Health, Madrid, Spain) and administration of 50 μg 0.2 mL/goat of PGF2α, cloprosterol (Estrumate, Animal Health, Madrid, Spain).

• DAY 6: Removal of sponges and administration of eCG (Foligon) 2 mL/goat at a variable concentration depending on season and farm history (222–400 IU).

• DAY 8: Artificial insemination by transcervical insemination technique is carried out approximately 46 hours after removal of the sponges.

• 13-days protocol (P13):

• DAY 1: Application of fluorogestone acetate sponges.

• DAY 9: Administration of PGF2α and eCG.

• DAY 11: Removal of sponges.

• DAY 13: Artificial insemination.

• 14-days protocol (P14):

• DAY 1: Ultrasonography (TA) and sponge insertion

• DAY 10: Administration of eCG and PGF2α.

• DAY 12: Removal of sponges.

• DAY 14: Artificial insemination.

Bucks and semen preparation

Semen was collected from a total of 55 bucks (ranging from 2 to 5 years old) during 2010 to 2013 period. All bucks were located at the Andalusian goat selection and improvement centre (Albolote, Granada, Spain). Bucks were fed with a commercial concentrate (0.5 kg) with ad libitum access to hay, water, and mineral supplementation blocks.

Ejaculates were collected with an artificial vagina and immediately placed in a water bath at 37 °C for individual evaluation. The selection criteria of ejaculates for AI were: Volume > 0.5 mL; Concentration > 3000 × 106 spz/mL and a mass motility > 4.

For insemination with chilled semen, the commercial extender Inra 96 (IMV Technologies, France) was used to prepare the straws at a concentration of 200 × 106 spermatozoa per straw, which were refrigerated at 5 °C and inseminated 4–6 h after preparation of the semen doses.

On the other hand, for the preparation of cryopreserved doses, the extender used was Triladyl (IMV Technologies, France) with 150 × 106 spermatozoa per straw. A Digitcool programmable cryo-freezer (IMV Technologies, France) was used for sperm cryopreservation. Straws were thawed in the field at a temperature of 37 °C for 30 s.

Artificial insemination procedure and pregnancy detection

The inseminations were carried out in 110 commercial farms of Murciano-Granadina goat breed. All the inseminations were developed approximately 46 h after the removal of the sponges. Does were cervically inseminated using a speculum with a built-in light source.

All the inseminated does were evaluated by transabdominal ultrasonography with a 5 MHz probe 42 days after insemination. Equally, the birth data was recorded to confirm that they came from insemination.

Data analysis

All statistical analysis were performed using the XLSTAT Version 2014.5.03 (Version 2014.5.03; Addinsoft Corp.: Paris, 2014). The categorical regression model (CATREG) was used to fit the statistical model, which describes how to change the probability of pregnancy results according to the influence of the independent variables and allows for identify of how variables were linearly dependent on the predictive factor considered. The resulting regression equations could be used to trace back, explain, or predict pregnancy ratio for any combination of studied variables.

Diagnosis of pregnancy acted as a dichotomous dependent variable in the categorical regression model, while the factors considered, had a categorical (farm, semen type, birth number, birth type, synchronisation protocol, insemination month and season, kidding month and season) or continuous numeric nature (covariates) (does age, kid number, live and dead born kids, eCG concentration). Additionally, the interactions of farm × protocol; farm × protocol × eCG and farm × protocol × eCG × insemination season were considered. Variable nature, description and the levels that they comprise can be found in Table 1.

Table 1. Nature, definition and levels of studied variables.

Variable	Variable Nature	Definition	Levels
Diagnosis of Pregnancy (Dependent)	Dichotomous	Outcome indicating pregnancy diagnosis	Positive, Negative
Factors considered:
Farm	Categorical	Farm where the study is conducted	Farm 1, Farm 2, Farm 3, …
Semen Type	Categorical	Type of semen used for insemination	Fresh, Frozen-Thawed
Birth Number	Categorical	Number of births by the doe	1–11
Birth Type	Categorical	Type of birth	Negative Pregnancy or Positive Pregnancy (Miscarriage with Lactation, Dead before kidding, Single, Double, Triple, Multiple (Quadruple or more) Single, Double, Triple, Multiple (Quadruple or more)
Synchronisation Protocol	Categorical	Method used to synchronise oestrus in does	Protocol 8 days, Protocol 13 days, Protocol 14 days, …
Kidding Month	Categorical	Month in which kidding occurred	January, February, March, …
Kidding Season	Categorical	Season in which kidding occurred	Spring, Summer, Fall, Winter
Insemination Month	Categorical	Month in which insemination occurred	January, February, March, …
Insemination Season	Categorical	Season in which insemination occurred	Spring, Summer, Fall, Winter
Doe’s Age at insemination	Continuous Numeric	Age of the doe at the time of insemination	9.07 − 142.27 months
Kid Number	Continuous Numeric	Number of kids born in the current kidding	0–6
Live Born Kids	Continuous Numeric	Count of live and dead kids born in the current kidding	0–5
Dead Born Kids	Continuous Numeric	Count of live and dead kids born in the current kidding	0–3
eCG Concentration	Continuous Numeric	Concentration of eCG hormone in the doe’s blood	222–400
Interactions considered:
Farm × Protocol	Continuous Numeric	Interaction between the farm and synchronisation protocol	170 levels
Farm × Protocol × eCG	Continuous Numeric	Interaction between farm, synchronisation protocol, and eCG concentration	330 levels
Farm × Protocol × eCG × Insemination Season	Continuous Numeric	Interaction between farm, synchronisation protocol, eCG concentration, and insemination season	330 levels

eCG: Equine chorionic gonadotropin.

Although additional factors may play a conditioning role such as individual effect of goat or buck, stress at the moment of insemination or heat stress tolerance, the difficulty to accurately control them, makes the standard error derived from incorrectly quantifying them, approaches the error derived from not considering them to be a conditioning factor, when indeed it is.

In these regards, including the effect of the goat or buck individually is not feasible given the distribution of data in our extense sample. First, not all the goats went to the same number of births, nor they had the same age, nor they were inseminated at the same age among others. These factors have already been considered in order to quantify their repercurssion on pregnancy diagnosis success. Thus, adding the goat as an additional factor may involve the occurrence of statistical redundancies, but still be unable to provide information on the particular factors within the goat factor responsible. Second, not all goats have been inseminated by all bucks thus, the comparison between does and buck possibilities is either not feasible or entrains a large error, which in turn distorts the results. This becomes even more challenging as bucks did not inseminate goats at the same moment, hence additional sources for variability may be added and become a source for redundancies.

In respect to stress, at the moment of insemination, several sources for stress maybe acting. This compels the implementation and develipment of a deep study aiming to disentangle and ascribe up to which levels of stress should be considered to derive from the mere insemination protocol and not other sources, for instance, the higher pronity of certain goats than others to be stressed when exposed to the same circumstances.

Furthermore, the goats in the study had differentially gone through insemination protocols from 1 to 4 inseminations, hence, stress adaptation in those goats going through several inseminations could be presumed. This compromises the reliability of stress measurements during the process, which translates into the aforementioned increase in the standard error derived from incorrect measuring, which approximates to the standard error generated from failing to consider stress as a conditioning factor, if it indeed plays a significant role.

As suggested by Chopra et al. (2023), with the aim to minimise the effects of stress during the moment of extraction and isnemination in goat reproductive husbandry, insemination protocols, coupled with goat sensitisation to insemination and effective buck trainingare implemented routinely. Insemination protocols ensure a systematic and well-timed approach to reproductive procedures, enhancing efficiency and reducing the duration of the process. Meanwhile, sensitising goats to insemination involves gradual exposure to the procedure, helping them become more accustomed to the handling and minimising anxiety. Furthermore, buck training focuses on acclimating male goats to the various aspects of reproductive interventions, such as handling and restraint. This comprehensive approach not only streamlines the extraction process but also creates a more comfortable and familiar environment for the animals, thereby reducing stress levels.

In regards the resistance/tolerance to heat stress, the improved resilience of Murciano-Granadina Goat Breeds has been reported in literature (Delgado et al. 2017). This strongly relies on the genetic pool of the breed which makes it outstand other goat breeds and small ruminat species such as sheep, as suggested by the studies on estimated breeding values (EBV) for thermal stress resistance found by Menéndez-Buxadera et al. (2015), showing a increased distribution of EBV observations towards resistant or tolerant in Murciano-Granadina goats. Furthermore, authors such as Arrébola et al. (2016) remarked the difficulty to control all the factors involved in the reproductive efficiency of goats. The authors conclude that numerous management and meteorological factors influence the success of the AI program in Murciano-Granadina goats. Indeed, an special attention is refered to the impossibility to control temperature and/or rainfall. The same authors, also explain that the possibility to schedule the dates of insemination based on forecasted temperaturas is yet to be explored.

In line with this, the results by Menéndez-Buxadera et al. (2015) explain that the genetic selection for tolerant/resistant animals may have been carried out indirectly as suggested but the despreciable differences between heritabilities for heat stress resistence when comfort and stress zones are compared (high correlations rangung between 0.65 and 0.85).

Hence, other husbandry and management factors/predictors of pregnancy diagnosis (hormonal treatments, age of the goat, time of AI) which can be standardised easily should be approached in order to maximise the outcomes expected from IA practices.

To ensure a strong linear relationship and independence between predictors multicollinearity tests were carried. This way, redundant variables were identified to avoid the overinflation of the explanatory potential of variance due to the presence of unnecessary variables. Variance inflation factor (VIF) was calculated as an indicator of multicollinearity following this formula: VIF = 1/(1–R²) where R2 is the coefficient of determination of the regression equation (González Ariza et al. 2021). As suggested Rogerson (2001) in the present study maximum VIF value of 5 was used. Tolerance (1−R²) is the amount of variability in a certain independent variable that is not explained by the rest. VIF was estimated using the discriminant analysis routine of the analysing data package of XLSTAT 2014 (Pearson Edition).

Results

Table 2 shows a summary of the results of VIF and tolerance of the studied variables. Those variables for which VIF values were smaller than 5 and were considered in the CATREG model that followed. Three evaluation rounds are needed to obtain all variables with a VIF value under 5. In the first round, a VIF of 3572.486 was reported for the interaction Farm × Protocol, while in the second round, the variable Farm presented a VIF value of 247.56. Hence both elements were discarded from future analyses.

Table 2. Multicollinearity analysis of studied parameters.

Statistic/traits	Tolerance (1−R^2)	VIF^a
Insemination age	0.6231	1.6049
Birth number	0.6254	1.5990
Kidding season	0.6491	1.5405
Kidding month	0.6800	1.4706
Birth type	0.8019	1.2470
ECG	0.8249	1.2122
Protocol	0.8613	1.1611
Farm × Protocol × ECG	0.9368	1.0675
Insemination season	0.9555	1.0466
Semen type	0.9846	1.0157
Alive	0.0000
Death	0.0000
Kid number	0.0000

^a Interpretation thumb rule: VIF ≥ 4 (highly correlated); 1 < VIF < 4 (moderately correlated); VIF = 1 (not correlated).

ECG: Equine chorionic gonadotropin.

Table 3 provides several indicators of the quality of the model (or goodness of fit). In this case, the model explains 99.99% of the variance of the dependent variable

Table 3. Model summary results.

Multiple R	R^2	Adjusted R^2	Apparent prediction error	Expected prediction error	Std. error	N
1	0.999	0.999	0.001	0.001	0.001	10.580

The model chosen is valid which as suggested by the statistical significance levels found and reported in Table 4 relative to ANOVA analysis. In these regards, the independent variables studied statistically significantly predict for the outcomes of diagnosis of pregnancy as the dependent variable, F (364, 10,621,414, p < .0005).

Table 4. Summary test of ANOVA to test the fit of the regression equation.

	Sum of Squares	Df	Mean square	F	Sig.
Regression	10,621.414	364	29.18	53604.729	0.001
Residual	5.586	10.262	0.001
Total	10.627	10.626

Table 5 showed the standardised coefficients (β) of studied variables, which pro-vide necessary information to predict fertility diagnosis. The variables birth number, insemination season, farm × protocol × eCG interaction, birth types, kidding months, protocols, kidding season, eCG doses, number of kids, insemination age, and alive kids contributed significantly to the model, contrary to variables semen type (0.991) and dead kids (0.087).

Table 5. Standardised coefficients and significance of categorical regression (CATREG) model.

	Standardised coefficients	Bootstrap (1000)
	(β)	Estimate of Std. Error	df	F	Sig.
Birth number	0.001	0	1	8.634	0.003
Insemination season	0.130	0.003	3	1645.807	0.001
Farm × protocol × eCG	0.130	0.003	329	1484.892	0.001
Birth type	0.449	0.007	6	3788.169	0.001
Kidding month	0.204	0.006	12	1254.608	0.001
Semen type	0.000	0	1	0	0.991
Protocol	0.007	0.003	3	5.734	0.001
Kidding season	0.213	0.003	4	4026.198	0.001
Death kids	−0.023	0.013	1	2.937	0.087
eCG	0.022	0.004	1	36.082	0.000
Kid number	−0.133	0.03	1	19.439	0.001
Insemination age	−0.001	0	1	6.398	0.011
Alive kids	−0.485	0.04	1	150.293	0.001

According to the aforereported results in Table 5, the model equation for fertility diagnosis (Y) was as follows:

Y = 0.449 × (Birth type) + 0.213 × (Kidding season) + 0.204 × (Kidding month) + 0.13 × (Insemination season) + 0.13 × (Farm x × protocol × eCG) + 0.022 × (eCG)+ 0.007 × (Protocol) + 0.001 × (Birth number) − 0.001 × (Insemination age) − 0.133 × (Kid number) − 0.485 × (Alived kids).

Table 6 shows correlations between the studied variables. There was a positive large correlation between the number of kids and alive kids (0.986); moderate between kidding season and kidding month (0.468); birth number and insemination age (0.588), and kidding month with alive kids (0.415), and a number of kids (0.415). On the other hand, large negative correlations were observed between insemination month and the interaction Farm x Protocol × eCG (−0.986); birth type with alive kids (−0.632), and kid number

Table 6. Correlations between the different studied variables.

	Insemination season	Protocol	Farm × Protol × eCG	Kidding month	Kidding season	Birth type	Semen type	eCG	Alive kids	Death Kids	Kid number	Birth number	Insemination age
Insemination season	1	0.013	−0.983	−0.017	−0.048	0.08	−0.055	0.033	−0.064	0.023	−0.064	−0.005	−0.019
Protocol	0.013	1	−0.084	0.027	−0.028	−0.006	−0.018	0.083	0.016	−0.014	0.018	−0.012	−0.022
Farm × Protocol × eCG	−0.983	−0.084	1	0.021	0.036	−0.072	0.043	−0.199	0.05	−0.019	0.05	0.005	0.015
Kidding month	−0.017	0.027	0.021	1	0.468	0.001	−0.004	−0.018	0.415	−0.037	0.415	0.014	0.035
Kidding season	−0.048	−0.028	0.036	0.468	1	0.326	−0.054	0.059	0.067	−0.012	0.066	−0.005	−0.019
Birth type	0.08	−0.006	−0.072	0.001	0.326	1	−0.067	−0.025	−0.632	0.221	−0.645	0.017	−0.071
Semen type	−0.055	−0.018	0.043	−0.004	−0.054	−0.067	1	0.078	0.057	−0.015	0.058	0.042	0.039
eCG	0.033	0.083	−0.199	−0.018	0.059	−0.025	0.078	1	0.065	−0.014	0.069	0.008	0.019
Alive kids	−0.064	0.016	0.05	0.415	0.067	−0.632	0.057	0.065	1	−0.005	0.986	0.019	0.098
Death kids	0.023	−0.014	−0.019	−0.037	−0.012	0.221	−0.015	−0.014	−0.005	1	−0.084	0.008	−0.008
Kid number	−0.064	0.018	0.05	0.415	0.066	−0.645	0.058	0.069	0.986	−0.084	1	0.019	0.096
Birth number	−0.005	−0.012	0.005	0.014	−0.005	0.017	0.042	0.008	0.019	0.008	0.019	1	0.588
Insemination age	−0.019	−0.022	0.015	0.035	−0.019	−0.071	0.039	0.019	0.098	−0.008	0.096	0.588	1
Dimension	1	2	3	4	5	6	7	8	9	10	11	12	13
Eigenvalue	2.756	1.994	1.597	1.569	1.09	1.014	0.997	0.876	0.428	0.408	0.257	0.011	0.002

Discussion

Our results (Tables 3 and 4) suggest that the designed model for the present study is efficient enough and permits the solid quantification of the variability of pregnancy rate results. Multicollinearity tests revealed the existence of redundance issues regarding the variables of Farm × Protocol interaction and farm, hence, these were removed from the analysis. The basis for these redundancies may esteem from the fact that considering the interaction between Farm × Protocol × eCG levels may already capture the variability data explained by the aforementioned variables, but also from the manner in which such variables combine and the fact that ecG levels may supply additional information which was not already accounted for and which in turn provides a new source of variability.

Although certain elements could have been presumed to be a source for multicollinearity problems, such as kidding or insemination month and season, the lack of redundancies between kidding month and season relies on the fact that all the months within preset seasons do not follow the same weather condition patterns among others. This make the relationship, distribution, and progression of data across seasons not to replicate such relationship, distribution, and progression across months, which then add for a potential source of additional variability which is not computed, were any of the variables excluded from the analysis. This becomes evident in the case of the redundancies found between insemination month and season, which suggests that insemination may rather be less influenced by the changes occurring across months and seasons than kidding. Hence, the mere consideration of one of those variables, in our case insemination season, serves to explain the variability in pregnancy diagnosis which could be ascribed to the effect of such factors.

The influence of farms on fertility has previously been reported in literature (Arrebola et al. 2014, 2016). Indeed, interfarm differences could be explained by several factors, such as climatic variations, production system (extensive, intensive, semi-extensive, or semi-intensive), altitude (plains, mountains, valleys), farm size (small, medium, large), the nutritional composition of food provided, herd composition, heat detection, exercise or herd health, among others (Nordstoga et al. 2010; Jaafar et al. 2018; Vacca et al. 2018).

According to Arrebola et al. (2012) intensive farms report better fertility results after AI, as a result of their greater degree of animal management when compared to other systems, which leads to less stress during the interaction between goat and human. Corteel and Leboeuf (1990) described a greater and more important susceptibility to stress from preovulatory events in goats and lower fertility after two cervical inseminations, caused and related by cortisol.

The effect of progesterone-based oestrus synchronisation protocols influenced fertility in the present study. As previously described by other authors (Evans et al. 2001; Menchaca and Rubianes 2002; Pietroski et al. 2013), long-term protocols were associated with lower fertility results as a consequence of sub-luteal serum concentration which is linked with undesired follicular development, ovulation, oocyte viability, and luteal functions. In this sense, short-term protocols were developed and normally associate with supraluteal concentrations of progesterone and in turn offer higher pregnancy rates since they positively influence follicular turnover (Karaca et al. 2009; Gore et al. 2020).

In the same way, in the present study, eCG administration influenced fertility rates as previously described by Hameed et al. (2020), who observed that the administration of eCG increased follicular growth, and oestrous response, ovulation rate, and pregnancy rate. However, repeated administration of eCG could derive in a decrease of does fertility since causes follicle luteinization, antibody production, ovarian hyperstimulation, or luteolysis during the early luteal phase (Sun et al. 2019).

A significant effect on herd fertility was reported for the season when AI takes place. Optimal results were reported for those AI taking place in summer especially when compared to winter. In this sense, previous studies Arrebola et al. (2012) suggested inseminations performed during the coldest months (autumn, and winter) offered the worst pregnancy results.

Likewise, Abecia et al. (2016) and Arrébola et al. (2016) observed how at low temperatures and high rainfall the percentages of pregnancy decrease significantly, which may explain a reduction in fertility in winter. In this regard, Arrebola et al. (2014) suggested that July is the month reporting the best fertility results. According to Mellado et al. (2006) in dairy goats, unlike in cattle, high temperatures do not have a negative impact on the percentage of pregnancies detected while the same authors observed that the decrease in fertility is rather influenced by the presence of rains.

In Spain, (>85%) does are routinely inseminated with chilled sperm (Arrebola et al. 2014), since obtained fertility results are higher (55–65%) than when frozen semen is used (35–38%) (Mocé et al. 2020). These differences may be due to the damages that sperm suffers during the freezing process, such as damage to the plasma membrane, and loss of integrity of the acrosome (Arando Arbulu et al. 2021), resulting in reduced fertility rates. However, studies carried out by Arrebola et al. (2012), stated that with both methods of sperm preservation (chilled and frozen) similar results to those are obtained in the Payoya breed as in the present study.

Does age also significantly influences fertility after artificial insemination as our results suggest. Ritar and Salamon (1983), associated this to the live weight of animals at the time of AI. These authors detected a linear increase in fertility as age progresses which would later be confirmed by Arrebola, et al. (2014), who observed higher fertility in goats that were older than 6 years of age.

The basis behind this may lay upon the fact that older goats, which are normally physically stronger and more dominant consequently eat better, which in turn turns into a challenge specially for younger does, if they are producing milk at the same time, given this could act as an stressor which may eventually cause body condition loss and the parallel hindering of fertility rates.

Interestingly, recent studies suggest the opposite, describing does age, and indirectly, parturition order, as a negative influencing factor of fertility as it progresses towards adulthood (Mocé et al. 2022). Indeed, such authors recommend insemination procedures should be performed after their first parturition, preventing the mating of does with more than one parturition whenever possible.

From the correlation of 32.6% between kidding season and birth type, one could infer that the closer the birth date is to winter, the greater the likelihood of a multiple births. This could be explained by the fact that animals that give birth in December have to be inseminated in July. Moreover, goats from the Mediterranean areas (Zarazaga et al. 2005; Gómez‐Brunet et al. 2010) kept under natural photoperiod, display the onset of breeding activity at the end of summer or autumn (day length is decreasing) and it ends in late winter or spring beginning when day’s length is increasing.

It is in that particular non-breeding season, when higher doses of eCG can generally be found to counteract the seasonality effect. Indeed, Martines-Ross et al. (Martinez-Ros et al. 2018) showed the ovulatory success and beneficial effect on fertility due to having found a higher incidence of ovulatory failures in animals not treated with eCG, known for the increase of the frequency in heat presentation, (Santos et al. 2011) ovulation rates, and fertility (Wildeus 2000).

All the aforementioned suggests the interaction between the kidding season and birth type factors would be connected to the insemination season, the reproductive seasonality, and the measures taken to counteract them (higher doses of eCG) which in turn would be the cause for the increase of prolificacy when the birth season approaches to winter.

Also, there is a 41.5% correlation between alive kids and kidding month, indicating that the closer to summer (June to September), the higher the live birth rate. When environmental conditions are unfavourable, this means episodes of extreme cold and humidity, juvenile survival decreases as derived from an indirect negative influence on offspring traits such as enhanced body growth, fat reserves, or mass (Théoret-Gosselin et al. 2015), which has also been confirmed by other authors who affirm that one of the main factors that can increase the mortality of kids is mainly low temperatures at the birth moment (Mellado et al. 2000).

There is a positive correlation of 22.1% between deaths and the birth type. In these regards, it is logical to pay attention to does and kid related factors as follows. As stated Landete-Castillejos et al. (2009) females in poor nutritional condition and with low body fat reserves generally give birth to lighter kids and are potentially unable to satisfy their progeny’s nutritional needs because they produce less or low-quality milk. Also, Mellado et al. (2000) asserted that another cause of death is the birth weight of the kids which is associated with the type of birth, and the size of the goat knowing the weakness of kids from multiple births (Mellado, 2008).

The strongest negative correlation was found between Farm × eCG × Protocol interaction and insemination season (−98.3%). What this means is that as we progress towards winter, the tendency will be to decrease eCG doses, with this choice being conditioned by the insemination season. The closer to winter, the lower the eCG dose and the more presence of 8-day protocol. Some authors (Mocé et al. 2020) have stated that the use of eCG was not necessary during the breeding season. In case of the AIs of our study, eCG was always used but the concentration of eCG administrated was lower during the breeding season (from Autumn to Winter).

Also, eCG and Protocol were shown to negatively condition the interaction Farm × Protocol × eCG. In turn, this interaction was shown to condition the diagnosis to a great extent, which indicates that the farm will be the most important factor in this interaction and the one that most condition the interaction of these three factors. Different combinations of protocol and IU of eCG have shown different fertility results depending on the farm.

This can be a consequence of multiple parameters like climatic variations, farm’s production system, altitude, farm’s size, the nutritional composition of the food given, herd composition, oestrus detection, health of the herd, or handling of the animals (Salvador et al. 2005; Nordstoga et al. 2010; Jaafar et al. 2018; Vacca et al. 2018).

A negative correlation was found between the number of births and type of birth (−64.5%). According to the coding performed to be able to statistically analyse the database, the highest values for the type of birth were attributed to negative pregnancy diagnosis, single birth and triple birth.

Our results suggest that only half of the single births (596/1276) were due to first birth animals and 294 to second birth animals. This is in line with Batista et al. (2009) who observed that the kidding rate and prolificacy were significantly higher in multiparous than in nulliparous goats. In terms of negative diagnoses, 1372 of the 3692 total were from first calving animals and 820 from second calving animals, which agrees that found by Arrebola et al. (2014) who asserted that goats over 6 years old of Payoya breed presented the highest fertility rate compared to younger goats.

Our results (Tables 3 and 4) suggest that the designed model for the present study is efficient enough and permits the solid quantification of the variability of pregnancy rate results. Multicollinearity tests revealed the existence of redundance issues regarding the variables of Farm × Protocol interaction and farm, hence, these were removed from the analysis. The basis for these redundancies may esteem from the fact that considering the interaction between Farm × Protocol × eCG levels may already capture the variability data explained by the aforementioned variables, but also from the manner in which such variables combine and the fact that eCG levels may supply additional information which was not already accounted for and which in turn provides a new source of variability.

The influence of farms on fertility has previously been reported in literature (Arrebola et al. 2014, 2016). Indeed, interfarm differences could be explained by several factors, such as climatic variations, production system (extensive, intensive, semi-extensive, or semi-intensive), altitude (plains, mountains, valleys), farm size (small, medium, large), the nutritional composition of food provided, herd composition, heat detection, exercise or herd health, among others (Nordstoga et al. 2010; Jaafar et al. 2018; Vacca et al. 2018).

According to Arrebola et al. (2012) intensive farms report better fertility results after AI, as a result of their greater degree of animal management when compared to other systems, which leads to less stress during the interaction between goat and human. Corteel and Leboeuf (1990) described a greater and more important susceptibility to stress from preovulatory events in goats and lower fertility after two cervical inseminations, caused and related by cortisol.

The effect of progesterone-based oestrus synchronisation protocols influenced fertility in the present study. As previously described by other authors (Evans et al. 2001; Menchaca and Rubianes, 2002; Pietroski et al. 2013), long-term protocols were associated with lower fertility results as a consequence of sub-luteal serum concentration which is linked with undesired follicular development, ovulation, oocyte viability, and luteal functions. In this sense, short-term protocols were developed and normally associate with supraluteal concentrations of progesterone and in turn offer higher pregnancy rates since they positively influence follicular turnover (Karaca et al. 2009; Gore et al. 2020).

In the same way, in the present study, eCG administration influenced fertility rates as previously described by Hameed et al. (2020), who observed that the administration of eCG increased follicular growth, and oestrous response, ovulation rate, and pregnancy rate. However, repeated administration of eCG could derive in a decrease of does fertility since causes follicle luteinization, antibody production, ovarian hyperstimulation, or luteolysis during the early luteal phase (Sun et al. 2019).

A significant effect on herd fertility was reported for the season when AI takes place. Optimal results were reported for those AI taking place in summer especially when compared to winter. In this sense, previous studies Arrebola et al. (2012) suggested inseminations performed during the coldest months (autumn, and winter) offered the worst pregnancy results.

Likewise, Abecia et al. (2016) and Arrébola et al. (2016) observed how at low temperatures and high rainfall the percentages of pregnancy decrease significantly, which may explain a reduction in fertility in winter. In this regard, Arrebola et al. (2014) suggested that July is the month reporting the best fertility results. According to Mellado et al. (2006) in dairy goats, unlike in cattle, high temperatures do not have a negative impact on the percentage of pregnancies detected while the same authors observed that the decrease in fertility is rather influenced by the presence of rains.

In Spain, (>85%) does are routinely inseminated with chilled sperm (Arrebola et al. 2014), since obtained fertility results are higher (55–65%) than when frozen semen is used (35–38%) (Mocé et al. 2020). These differences may be due to the damages that sperm suffers during the freezing process, such as damage to the plasma membrane, and loss of integrity of the acrosome (Arando Arbulu et al. 2021), resulting in reduced fertility rates. However, studies carried out by Arrebola et al. (2014), stated that with both methods of sperm preservation (chilled and frozen) similar results to those are obtained in the Payoya breed as in the present study.

Does age also significantly influences fertility after artificial insemination as our results suggest. Ritar and Salamon (1983), associated this to the live weight of animals at the time of AI. These authors detected a linear increase in fertility as age progresses which would later be confirmed by Arrebola et al. (2014), who observed higher fertility in goats that were older than 6 years of age.

The basis behind this may lay upon the fact that older goats, which are normally physically stronger and more dominant consequently eat better, which in turn turns into a challenge specially for younger does, if they are producing milk at the same time, given this could act as an stressor which may eventually caue body condition loss and the paralell hindering of fertility rates.

Interestingly, recent studies suggest the opposite, describing does age, and indirectly, parturition order, as a negative influencing factor of fertility as it progresses towards adulthood (Mocé et al. 2022). Indeed, such authors recommend insemination procedures should be performed after their first parturition, preventing the mating of does with more than one parturition whenever possible.

From the correlation of 32.6% between kidding season and birth type, one could infer that the closer the birth date is to winter, the greater the likelihood of a multiple births. This could be explained by the fact that animals that give birth in December have to be inseminated in July. Moreover, goats from the Mediterranean areas (Zarazaga et al. 2005; Gómez‐Brunet et al. 2010) kept under natural photoperiod, display the onset of breeding activity at the end of summer or autumn (day length is decreasing) and it ends in late winter or spring beginning when day’s length is increasing.

It is in that particular non-breeding season, when higher doses of eCG can generally be found to counteract the seasonality effect. Indeed, Martinez-Ros et al. (2018) showed the ovulatory success and beneficial effect on fertility due to having found a higher incidence of ovulatory failures in animals not treated with eCG, known for the increase of the frequency in heat presentation, (Santos et al. 2011) ovulation rates, and fertility (Wildeus, 2000).

All the aformentioned suggests the interaction between these two factors would be connected to the insemination season, the reproductive seasonality, and the measures taken to counteract them (higher doses of eCG) which in turn would be the cause for the increase of prolificacy when the birth season approaches to winter.

Also, there is a 41.5% correlation between alive kids and kidding month, indicating that the closer to summer (June–September), the higher the live birth rate. When environmental conditions are unfavourable, this means episodes of extreme cold and humidity, juvenile survival decreases as derived from an indirect negative influence on offspring traits such as enhanced body growth, fat reserves, or mass (Théoret-Gosselin et al. 2015), which has also been confirmed by other authors who affirm that one of the main factors that can increase the mortality of kids is mainly low temperatures at the birth moment (Mellado et al. 2000).

There is a positive correlation of 22.1% between deaths and the birth type. In these regards, it is logical to pay attention to does related (birth number, birth type, does age), kid related factors (live and dead born kids) and their interaction (kidding month, and kidding season, kid number) as follows. As stated Landete-Castillejos et al. (2009) females in poor nutritional condition and with low body fat reserves generally give birth to lighter kids and are potentially unable to satisfy their progeny’s nutritional needs because they produce less or low-quality milk. Also, Mellado et al. (2000) asserted that another cause of death is the birth weight of the kids which is associated with the type of birth, and the size of the goat knowing the weakness of kids from multiple births (Mellado, 2008).

The strongest negative correlation was found between Farm × eCG × Protocol interaction and insemination season (−98.3%). What this means is that as we progress towards winter, the tendency will be to decrease eCG doses, with this choice being conditioned by the insemination season. The closer to winter, the lower the eCG dose and the higher the use of 8-day protocols. Some authors (Salvador et al. 2005) have stated that the use of eCG was not necessary during the breeding season. In case of the AIs of our study, eCG was always used but the concentration of eCG administrated was lower during the breeding season (from Autumn to Winter).

Also, eCG and Protocol were shown to negatively condition the interaction Farm × Protocol × eCG. In turn, this interaction was shown to condition the diagnosis to a great extent, which indicates that the farm will be the most important factor in this interaction and the one that most condition the interaction of these three factors. Different combinations of protocol and IU of eCG have shown different fertility results depending on the farm.

This can be a consequence of multiple parameters like climatic variations, farm’s production system, altitude, farm’s size, the nutritional composition of the food given, herd composition, oestrus detection, health of the herd, or handling of the animals (Salvador et al. 2005; Nordstoga et al. 2010; Jaafar et al. 2018; Vacca et al. 2018).

A negative correlation was found between the number of births and type of birth (−64.5%). According to the coding performed to be able to statistically analyse the database, the highest values for the type of birth were attributed to negative pregnancy diagnosis, single birth and triple birth.

Our results suggest that only half of the single births (596/1276) were due to first birth animals and 294 to second birth animals. This is in line with Batista et al. (2009) who observed that the kidding rate and prolificacy were significantly higher in multiparous than in nulliparous goats. In terms of negative diagnoses, 1372 of the 3692 total were from first calving animals and 820 from second calving animals, which agrees that found by Arrebola et al. (2014) who asserted that goats over 6 years old of Payoya breed presented the highest fertility rate compared to younger goats.

Conclusions

The interaction between Farm, Protocol, and eCG levels can provide additional information not accounted for when considering these factors separately. Intensive farms with better animal management have less stress during insemination, which is crucial for high numbers of inseminations. Long-term protocols have lower fertility rates, while short-term protocols offer higher pregnancy rates. eCG increases follicular growth and ovulation rate, but repeated administration can decrease fertility. AI is more efficient in warmer months and closer to winter, with lower eCG doses and increaseduse of 8-day protocols. Sperm damage during freezing significantly influences fertility after artificial insemination, especially in older does with greater dominance and lower body condition. Poor nutrition leads to lighter kids and lower milk production, with a higher likelihood of multiple births closer to winter. Higher doses of eCG can counteract seasonality during non-breeding season, and unfavourable environmental conditions lead to decreased juvenile survival. Overall, factors like birth type, kidding season, kidding month, and insemination season are more important than eCG or AI protocol. It may be crucial to tailor specific protocols and eCG doses for each farm to maximise pregnancy rates.

Ethical approval

The study followed the premises described in the Declaration of Helsinki. The Spanish Ministry of Economy and Competitivity through the Royal Decree-Law 53/2013 and its credited entity the Ethics Committee of Animal Experimentation from the University of Córdoba permitted the application of the protocols present in this study as cited in the fifth section of its second article, as the animals assessed were used for credited zootechnical use. This national Decree follows the European Union Directive 2010/63/UE, from the 22nd of September of 2010. Furthermore, the present study works with records rather than live animals directly, and these records were obtained after minimal handling, hence no special permission was compulsory.

Acknowledgements

The authors would like to acknowledge to the National Association of Breeders of Murciano-Granadina Goat Breed and to the Andalusian goat selection and improvement center (Fuente Vaqueros, Granada, Spain).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data will be made available from the corresponding author F.J.N.G. upon reasonable request.

Additional information

Funding

Funding was not received for the development of the present study. The present research was carried out during the covering period of a Ramón y Cajal Post-Doctoral Contract with the reference MCIN/AEI/10.13039/501100011033 and the European Union “NextGenerationEU”/PRTR.