Artificial Insemination, Ovulation and the Role of the Boar

Introduction
Optimised insemination strategies on pig farms aim at reducing the number of inseminations, without reducing reproductive performance. At the moment, one insemination with good quality semen can result in good fertilisation results when insemination takes place in a 24 h period before ovulation. However, knowledge of this optimal period is not sufficient to limit the number of inseminations to only one in all sows, because the time of ovulation of sows can not be predicted within a 24 h range. Therefore, to optimise insemination strategies on farms, either the period during which one insemination leads to optimal fertilisation results must be increased and/or the prediction of ovulation time needs to be improved. Much work concerning the timing of insemination -relative to oestrus characteristics-and reproductive performance (litter size, farrowing rate) has been performed in the sixties and seventies. Since the early nineties, ultrasound has allowed assessment of the timing of ovulation in spontaneously ovulating sows. Since then, experiments have increased our knowledge on the timing of ovulation in sows relative to behavioural and physiological changes and on effects of insemination strategies on fertilisation.
Besides the timing of insemination, also the circumstances around insemination are important for fertilisation chances. In this overview, also recent data on the role of boar presence on uterine motility are presented (Langendijk, 2001).


Optimal period between insemination and ovulation
The life span of eggs after ovulation and the life span of a sufficient number of sperm cells capable of fertilisation within the oviduct, limits the period relative to ovulation in which inseminations can lead to successful fertilisation (for review, see Hunter, 1994). Studies from the sixties already showed that sows inseminated relatively early or late in oestrus showed a significant reduction in fertilisation rate (Hancock and Hovell, 1962) and in litter size and farrowing rate (Willemse and Boender, 1967). From these experiments, it is difficult to draw conclusions about the fertile life span of sperm cells and oocytes, because the time of ovulation varies considerably relative to the onset of oestrus (for review, see Soede and Kemp, 1997). Therefore, in many experiments, the process of fertilisation was studied using gilts with induced ovulation (e.g. Dziuk, 1970; Hunter, 1967; Hunter, 1981; Hunter et al., 1987). Those experiments showed the importance of processes such as sperm transport, capacitation of sperm cells and the filling and emptying of the sperm reservoir for the chances offertilisation.

In recent years, data have become available on the chances of fertilisation relative to spontaneous ovulation. It was found that in some sows, optimal fertilisation results can be achieved when insemination (with 3 billion sperm cells) takes place between as many as 40 h before ovulation and 16 h after ovulation (Soede et al., 1995a). However, to reach good fertilisation results in the majority of sows, it was found that insemination has to take place in the 12 ((Waberski et al., 1994a) or 24 h period before ovulation (Waberski et al, 1994b; Soede et al., 1995ab; Steverink et al,. 1997) This is illustrated in Figure 1. Figure 1 shows the fertilisation rates as assessed at Day 5 after ovulation in about 300 sows in which insemination took place one with a dose of 3 billion sperm cells. From those results, it is clear that fertilisation results are optimal when insemination takes place before ovulation. In the period of 0 to 24 hours before ovulation, the majority of sows have 100% fertilisation. Outside this interval, the percentage of sows with partial fertilisation and of sows in which none of the oocytes are fertilised increases considerably. Based on litter size and farrowing rate, the optimal interval between insemination and fertilisation was suggested to be even longer, between 28 h before ovulation and 4 h after ovulation (Nissen et al., 1997). The fact that an interval of more than approximately 24 h results in a reduction in fertilisation rate means that the fertile life span of sufficient numbers of sperm cells is limited to 24 h in the majority of sows. The fact that inseminations after ovulation result in a reduction in fertilisation rate means that the fertile life span of oocytes is about similar to the time it takes for sperm cells to capacitate and reach the site of fertilisation. Even in the optimal period, however, some sows have sub optimal or even no fertilisation (see Figure 1).

Figure 1:Percentage of sows having a fertilisation rate of 0%, 1-20%, 21-40%, 41-60%, 61-80%, 81-99% and 100% after one insemination at variable intervals from ovulation. Insemination was done with a dose of 3 billion sperm cells. Compiled data from Soede et al., 1995ab and Steverink et al., 1997



In recent years, experiments have been performed that studied the period in which inseminations lead to optimal fertilisation. Effects of the following factors were investigated:

In conclusion, from the above, it seems that factors reducing the number and/or the quality of sperm cells that reach the oviduct, shorten the period in which insemination results in optimal fertilisation. It can be concluded that good quality semen of at most 2 days of age ensures good fertilisation rates when inseminated between 0 and 24 h before ovulation. Strangely enough, a doubling of the dose hardly had a positive influence.

From recent research, several indications exist that a further improvement of sperm transport and/or sperm survival can be achieved. A few examples are given of the possibilities:

Prediction of the time of ovulation
On average, ovulation takes place at 35 to 45 h after onset of oestrus (standing response in presence of a boar). However, the variability between sows is large, between e.g. 10 and 85 (Weitze et al., 1994) or 10 and 58 h (Soede et al., 1995a) although sometimes less variation is found (35 to 43 h in 20 sows Mburu et al., 1995). Also parameters such as vaginal temperature and vaginal mucus conductivity have proven to be bad predictors for ovulation time (Soede et al., 1997; Stokhof et al., 1996).

In a number of German, Dutch, Swedish and Danish investigations in the nineties, the average timing of ovulation (percentage of oestrus) varied between 64% and 72% (see review by Soede and Kemp, 1997); the duration of oestrus (that is the period in which sows show a standing response in presence of a boar) explained 50-60% of the variation in the timing of ovulation. At the moment, therefore, the duration of oestrus seems the best estimator for the time of ovulation during oestrus.
Unfortunately, the duration of oestrus is highly variable and gives only a retrospective estimate of the time of ovulation.
Therefore, research should be aimed at finding other prospective estimators for the time of ovulation.

In the course of estrus, responsiveness of the sow to stimuli applied to evoke estrus behaviour (standing response) increases and decreases again (Willemse and Boender, 1967). This means that early and late in estrus the sow will be responsive to intense stimuli (e.g. a DMA (Detection-Mating-Area) or intensive boar contact). When she is in the middle of the estrus period she is expected also to be responsive to lower levels of stimuli (e.g BPT (Back-Pressure-Test) in absence of boar). By applying different levels of stimuli during the course of estrus different receptive phases may be distinguished (e.g. receptive to BPT in absence of a boar, receptive to BPT when fence line contact with boars is used or receptive to BPT when brought in a DMA). Langendijk et al. (2000b) studied whether applying different levels of boar stimuli during detection of estrus might distinguish more phases of responsiveness and whether this information could be used to yield a more accurate predictor of ovulation. The results are not very encouraging. A large percentage of sows will not show a standing response at lower levels of stimuli at all during the course of estrus. Therefore, in such sows different responsive phases during estrus can not be distinguished. Moreover, even in sows showing more responsive phases during the course of estrus, these phases were poorly related to the ovulation moment.

Possibly, a combination of parameters (maybe both physical and behavioural) can give a better estimate of ovulation time than only one parameter. Such an attempt was recently made by Langendijk et al. (1999b), who combined the redness of the inner vulva to oestrous behaviour characteristics. In sows that showed vulva reddening at the onset of boar oestrus (60% of the sows), the change from red to pink occurred on average at 18 h before ovulation with a range from 36 h before to 2 h after ovulation. In their study, no hormone profiles were measured, but it seems plausible that the change in vulva colour from red to pink is related to the changes in the oestrogen profile. If these results prove to be repeatable, it would mean that sows with a red inner vulva at onset of oestrus do not need to be inseminated before the end of vulva redness, which would result in a more efficient use of semen.

The ultimate prediction of ovulation time would be to be able to monitor the rise in LH concentrations or the decline in oestrogen concentrations in a simple way, but for pigs this may still be impractical. For the immediate future, possibilities to ‘predict’ ovulation time are limited to the prediction of the duration of oestrus for sows.


Towards more efficient insemination strategies
The variation in oestrous behaviour and the related low predictability of ovulation time makes it very difficult to inseminate sows in the optimal period before ovulation using one insemination only. Therefore, most sows are inseminated repeatedly during oestrus. One recurrent question is whether such repeated inseminations should be performed at 12 h or 24 h intervals.
Lamberson and Safranski (2000) made an economical comparison between different insemination schedules in combination with once or twice per day detection of estrus. In their model, the likelihood of conception was set at a maximum of 90% (when insemination took place at ovulation) and dropped to about 62% with insemination at 24 h before ovulation and also at 6 h after ovulation. Ovulation was programmed to occur at 40±12 h after onset of oestrus and the weaning to ovulation interval was set at 136±18 h. Using those parameters, economic returns were calculated to be highest with inseminations at 12 h intervals; at 12, 24 and 36 h after onset of oestrus with twice per day oestrus detection and at 0, 12, 24 and 36 h after onset of oestrus with once per day oestrus detection. These results are a logical result seeing the (relatively low) fertilisation rates in the model for inseminations in the period of more than 12 h before ovulation. Results from a survey on German pig farms showed that herds in which sows were inseminated every 12 h did not have a better reproductive performance than herds where sows were inseminated at 24 h intervals (VFV, 1999). Similarly, Castagna et al. (2001) and Thorup (2001) also conclude from their experiments that inseminations at 12 h and 24 h intervals can result in similar reproductive performance.
These latter results more reflect results from fertilisation experiments as mentioned before (Waberski et al., 1994b; Soede et al., 1995; Nissen et al., 1996; Steverink et al., 1997) that clearly suggest that inseminations at 24 h intervals should be sufficient. However, it is clear that factors such as inexperienced inseminators (Flowers, 1994) or the use of older semen (Waberski et al., 1997) may cause the need for shorter intervals between inseminations.

Inseminating sows as often as possible may not just be a waste of semen, labor and money, but may also have negative effects. Rozeboom et al. (1997) concluded that a second insemination during late- or met-oestrus decreased farrowing rate and litter size and De Winter et al. (1992) showed that uterine inoculation with bacteria resulted in more uterine infections during late- or met-oestrus than during early- or mid-oestrus. This phenomenon may be related to the oestrogen/progesterone ratio, affecting both the blood circulation and therewith the influx of leucocytes and the contraction activity of the uterus.
Therefore, very intensive insemination strategies can result in poor reproductive performance due to infections when many sows are inseminated during late oestrus.


Use of knowledge of oestrus duration
Because ovulation takes place at a relatively fixed two-thirds of oestrus, a prediction of the duration of oestrus is an indirect prediction of ovulation and therewith of the optimal time for insemination. Unfortunately, the duration of oestrus is very variable and seems to be influenced by many factors, such as housing conditions, stress conditions, season, parity, genetic background (see review by Soede and Kemp, 1997). Nevertheless, to some extent, prediction of the duration of oestrus is possible. First of all, the average duration of oestrus varies considerably between farms (from 31 to 65 h), but is highly consistent from month to month within a farm (Steverink et al., 1999a), which implies that the optimal timing of insemination during oestrus should vary between farms. Second, it has been well established that the duration of oestrus is related with the weaning-to-oestrus-interval (WOI) (e.g. Rojkittikhun et al., 1992; Weitze et al., 1994; Kemp and Soede, 1996); sows with a short (3-4 days) WOI on average have a long oestrus, associated with an advanced ovulation time (Weitze et al., 1994; Kemp and Soede, 1996). Steverink et al. (1999a) found that this relationship between WOI and duration of oestrus was significant in 43 (80%) of 54 farms. As a consequence, on most farms sows with a WOI of 6 days or more should be inseminated sooner after onset of oestrus to make sure that the first insemination is before ovulation. The effects of the relation between oestrus duration and optimal insemination time are illustrated in Table 1 for two farms that differ in average duration of oestrus (Steverink et al., 1999b). The data are generated with a simulation model for insemination strategies in pigs (PIGSIS; Steverink, 1999). The simulation confirms that the timing of the first insemination can easily be too late, especially on farms with a short duration of oestrus and sows with a WOI of 5 to 6 days. Too early first inseminations are not really a problem; sows can be inseminated again, but sows in which the first insemination takes place after ovulation should be avoided. Therefore, to get insight in the efficiency of the insemination strategy (and quality of oestrus detection) on farms, farmers should establish the duration of oestrus and their farms, and also taking into account the effect of Weaning-to-Oestrus Interval on the duration of oestrus. This also enables checking of the appropriateness of insemination timings of individual sows.


Table 1. Timing of first insemination relative to ovulation ¹ on a farm with an average oestrus duration of 36 h or 60 h, simulated with PIGSIS ² . On both farms, sows are inseminated at 24 h after observed onset of oestrus and every 24 h increase in Weaning-to-Oestrus-Interval is associated with a decrease in oestrus duration of 6 h.