Fertility results after do-it-yourself and commercial company artificial insemination in dairy herds in Northern Ireland.

DAIRY herd fertility has been declining rapidly over the past 30 years (Royal and others 2000, Lucy 2001, Mee 2004), and has been associated with a rapid increase in milk production achieved by the widespread use of Holstein/Friesian genetics. In the UK this decline has been calculated as a 1 per cent reduction in pregnancy rates to first service per annum over a 20-year period (Royal and others 2000). Current pregnancy rates to first service in the UK and USA stand at approximately 40 per cent (Butler 1998, Royal and others 2000, Mayne and others 2002). Slightly higher pregnancy rates to first service of 48 and 49 per cent have been quoted in the Republic of Ireland (Buckley and others 2000) and Australia (Morton 2000a), respectively, although these rates were based on pregnancy diagnosis rather than calving rates. If this current trend continues, conception rates could well be below 30 per cent by 2010 (Royal and others 1999). Many factors have been associated with this decline in fertility, including increased milk production, nutritional stress, health problems and various management practices (Lucy 2001). Diskin and others (2000) accurately described artificial insemination (AI) of cattle as ‘the most important reproductive technology developed in the past 30 years’. Do-it-yourself artificial insemination (DIYAI) is now common practice on dairy farms, and, if used properly, is a cost-effective fertility management tool. However, there is much potential for DIYAI to contribute to poor fertility rates through inadequate storage and handling of semen and inaccurate deposition of semen into the female reproductive tract. Recent studies have suggested that DIYAI is contributing to lower fertility rates on dairy farms (Morton 2000b, O’Farrell and Crilly 2001), while Buckley and others (2003) recorded no difference in pregnancy rates to first service between DIY operators and commercial AI operators in well managed spring-calving dairy herds in the Republic of Ireland. In autumn 1998, a study was set up to investigate the key factors that influence fertility in dairy herds in Northern Ireland. The objectives of the study were to record reproductive performance, to identify key factors affecting performance, and, in turn, to establish management and/or treatment interventions that could be used to maintain or improve reproductive performance. Results from the first year of the study have already been published (Mayne and others 2002). Fertility performance was generally poor with a mean conception rate to first service of 37·1 per cent and a large variation between herds of 21 to 66 per cent (Mayne and others 2002). In total, 18 of the 19 herds surveyed practised DIYAI as their main method of service, so a study was conducted in winter 2001 and spring 2002 to determine if DIYAI was contributing to poor fertility performance in these herds. Six herds were selected for the study on the basis that they routinely carried out DIYAI. Mean herd size varied from 72 to 214 cows and mean conception rate to AI over the preceding three years varied from 20·4 to 47·7 per cent between herds (Table 1). The study was carried out during peak periods of AI (Table 1). Farmers were asked to carry out either DIYAI or AI using commercial company technicians (CCAI) on alternate days over the service period. Farmers were given a breeding season calendar to identify the days when DIYAI or CCAI were to be carried out. Farmers were instructed to carry out DIYAI according to their normal procedure. For CCAI, farmers were asked to contact their local commercial inseminator to arrange a visit if any cows required insemination. When the commercial inseminator arrived on the farm the farmer was instructed to select a straw of his choice from his liquid nitrogen flask, which was then given to the commercial inseminator to thaw and inseminate into the selected cow. Where possible, all farms on all days (either DIYAI or CCAI) employed the ‘am/pm’ rule for AI, that is, cows seen to be on heat in the morning were inseminated in the afternoon/evening of the same day, and cows seen to be on heat in the afternoon/ evening were inseminated the following morning. Detailed records of heats, services, AI operators, semen used and any abnormalities detected at service were recorded by the farmers for all services. CCAI inseminators were asked to keep duplicate service details for inseminations carried out by themselves, as well as assessing semen storage conditions and AI facilities. To assess the accuracy of heat detection, farmers were also asked to collect foremilk samples from each cow, for progesterone analysis, at the milking after which oestrus was detected. Milk progesterone concentrations were determined using an ELISA kit (Ridgeway Science) based on the method of Sauer and others (1986). Successful conception to a given insemination was assessed by non-return to heat within 100 days of service and by a subsequent confirmed calving. The 100-day non-return rate was used in addition to calving rates, as recovery of calving data proved difficult due to technical and personnel problems. Results were analysed using StatXact (Cytel Software) by a two-step process. Step 1 tested the assumption that the odds for the difference in the conception rate between DIYAI and CCAI were the same for each farm, using the method of Breslow and Day (1980). If this was not significantly different then a common odds ratio (OR) was estimated, which was then tested to see if it was significantly different from step 1 using the test of Mantel and Haenszel (1959). All probabilities quoted are based on the asymptotic version of the tests, as numbers were sufficient for this to be the case. Significance was determined by P<0·05. The concentration of progesterone in foremilk is lower than in composite milk due to a lower milk fat content (Heap and others 1976, Pope and others 1976). However, McCoy and others (2001) found a significant relationship (r2=0·83) between the milk progesterone concentrations in composite and foremilk samples. In the present study, cows recorded as Mean herd Conception Herd size* rate to AI* (%) Trial period