The effects of farm management practices on liver fluke prevalence and the current internal parasite control measures employed on Irish dairy farms

Veterinary Parasitology 207 (2015) 228–240

Contents lists available at ScienceDirect

Veterinary Parasitology journal homepage: www.elsevier.com/locate/vetpar

The effects of farm management practices on liver fluke prevalence and the current internal parasite control measures employed on Irish dairy farms Nikolaos Selemetas a , Paul Phelan b , Padraig O’Kiely b , Theo de Waal a,∗ a b

UCD School of Veterinary Medicine, University College Dublin, Ireland Animal & Grassland Research and Innovation Centre, Teagasc, Grange, Dunsany, Co. Meath, Ireland

a r t i c l e

i n f o

Article history: Received 8 October 2014 Received in revised form 12 December 2014 Accepted 12 December 2014 Keywords: Liver fluke Fasciola hepatica Helminths Grazing management Dairy farms Ireland

a b s t r a c t Fasciolosis caused by Fasciola hepatica is responsible for major production losses in cattle farms. The objectives of this study were to assess the effect of farm management practices on liver fluke prevalence on Irish dairy farms and to document the current control measures against parasitic diseases. In total, 369 dairy farms throughout Ireland were sampled from October to December 2013, each providing a single bulk tank milk (BTM) sample for liver fluke antibody-detection ELISA testing and completing a questionnaire on their farm management. The analysis of samples showed that cows on 78% (n = 288) of dairy farms had been exposed to liver fluke. There was a difference (P < 0.05) between farms where cows were positive or negative for liver fluke antibodies in (a) the total number of adult dairy cows in herds, (b) the number of adult dairy cows contributing to BTM samples, and (c) the size of the total area of grassland, with positive farms having larger numbers in each case. There was no difference (P > 0.05) between positive and negative farms in (a) the grazing of dry cows together with replacement cows, (b) whether or not grazed grassland was mowed for conservation, (c) the type of drinking water provision system, (d) spreading of cattle manure on grassland or (e) for grazing season length (GSL; mean = 262.5 days). Also, there were differences (P < 0.001) between drainage statuses for GSL with farms on good drainage having longer GSL than moderately drained farms. The GSL for dairy cows on farms with good drainage was 11 days longer than for those with moderate drainage (P < 0.001). The percentage of farmers that used an active ingredient during the non-lactating period against liver fluke, gastrointestinal nematodes, lungworm, and rumen fluke was 96%, 85%, 77% and 90%, respectively. Albendazole was the most frequently used active ingredient for treatment against gastrointestinal nematodes (57%), liver fluke (40%) and lungworm (47%), respectively. There was a difference (P < 0.05) in the use of triclabendazole and albendazole between positive and negative farms, with triclabendazole use being more common in positive farms. This study highlighted differences in dairy management practices between Irish farms with dairy herds exposed or not exposed to liver fluke and stressed the need

∗ Corresponding author at: School of Veterinary Medicine, University College Dublin, UCD Veterinary Sciences Centre, Belfield, Dublin 4, Ireland. Tel.: +353 1716 6178. E-mail address: [email protected] (T. de Waal). http://dx.doi.org/10.1016/j.vetpar.2014.12.010 0304-4017/© 2014 Elsevier B.V. All rights reserved.

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

229

of fine-scale mapping of the disease patterns even at farm level to increase the accuracy of risk models. Also, comprehensive advice and professional support services to farmers on appropriate farm management practices are very important for an effective anthelmintic control strategy. © 2014 Elsevier B.V. All rights reserved.

1. Introduction

2. Materials and methods

Fasciolosis, caused by Fasciola hepatica (liver fluke), may have a great impact on global livestock productivity (Fox et al., 2011). Liver fluke is a significant trematode parasite of cattle in regions with a temperate climate (Bennema et al., 2011). The completion of the liver fluke life cycle is related to specific climatic and environmental conditions and, therefore, the use of both climatic and environmental variables can help in interpreting the observed temporal disease pattern (McCann et al., 2010a,b). The systematic use of anthelmintics and also drainage, fencing and pasture management practices are the major control measures against fasciolosis (Torgerson and Claxton, 1999). In Ireland the climate is mild, with higher precipitation in the western half of the country and on highlands (Walsh, 2012). The temperate climate permits a dairy farming system that is lower cost and more dependent on grazed grass compared to other EU countries (Creighton et al., 2011; Läpple et al., 2012). Extended grazing of grass from early spring to late autumn when conditions allow, and spring-calving that synchronises the peaks of both grass growth and milk production, are key features of dairy farming management (Dillon et al., 1995; Läpple et al., 2012). However, the humid climate together with the seasonal grazing and the reliance on grazed grass can increase exposure to parasites such as liver fluke (O’Farrell et al., 1986). Previous studies in Ireland have showed a herd prevalence of fasciolosis in autumn 2012 of 82%, using a liver fluke enzyme-linked immunosorbent assay (ELISA) on bulk tank milk (BTM) samples from dairy herds (Selemetas et al., 2015), and a prevalence of 65% of liver fluke infected livers from culled cattle during autumn 2002 and summer 2003 (Murphy et al., 2006). Parasites comprise a continuous and often high infectious pressure on grazing animals (Waller, 1999) and the degree of infection depends on previous parasitic exposure, the physiological state of animals and seasonal conditions affecting infection and grazing (Waller, 2006). While farm management factors are a major driver of liver fluke infection risk (Bennema et al., 2011), studies evaluating the association of farm management practices with liver fluke exposure and parasite control programmes in Ireland are limited. The objectives of this study were (i) to identify the association between different farm management practices with liver fluke prevalence on Irish dairy farms and (ii) to document the current measures employed by Irish farmers for the treatment and chemoprophylaxis of different internal parasites, especially liver fluke.

2.1. Study design and sample population In total, 450 dairy farmers throughout Ireland were invited to participate in a survey between October and December 2013. The farmers were selected by Teagasc (Irish Agriculture and Food Development Authority) dairy advisors with the aim of achieving an even distribution of farms throughout the country. In order to assure an equal sampling probability of counties with dairy cattle density, the counties were divided in four groups based on the national percentage of dairy farms in each county. Therefore, the target number of sampled farms per county was 10 for the least intensive dairy farming counties, 20 for the moderately intensive counties, 40 for the more intensive, and 80 farms for the most intensive county of Cork. Most of the contacted farmers participated in Teagasc discussion groups, with the remainder selected by Teagasc dairy advisors to help achieve the even geographical distribution of farms. A questionnaire and a sampling kit with equipment and instructions for taking a BTM sample was posted to the farmers.

2.2. Questionnaire Between October and December 2013 participating farmers were asked to complete a questionnaire on farm management practices and anthelmintic use, referring only in their adult dairy cows. The questionnaire was sectioned into four parts, (i) farm structure, (ii) pasture management, (iii) grazing management, and (iv) antiparasitic treatment control, and consisted of 44 questions, including 18 multiple choice, 16 open-ended and 10 binary closed-ended questions. The first section of questions asked about farm structure: total number of adult dairy cows on the farm (dairy herd size), adult dairy cows contributing to BTM sample and if sheep, goats, horses or deer shared pasture with the dairy cows. A second section of questions on pasture management determined (i) the total area of grazed grassland (ha) and the number of paddocks, which enabled the calculation of the stocking density (livestock units (LU) per ha of grassland; one LU is the feed requirement equivalent of one standard dairy cow (European Commission, 2009)) and the mean paddock size (ha), (ii) the overall quality of land drainage (good, moderate or poor) and (iii) the percentage of grazed grassland (selection from four categories <25%, 25–50%, 50–75%, >75%) with snail habitats (streams, ponds, dykes, flooded ditches) based on the subjective evaluation of farmers. A third section of questions

230

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

asked farmers when were the first and final grazing dates on average and the length (none, part-time or full-time) of daily access of dairy cows to pasture for each month, which enabled the calculation of (i) the absolute grazing annual length (GAL; last grazing date-first grazing date) in days and (ii) the grazing season length (GSL) in days, with each full-time day counting as one and each part-time day counting as 0.5 days. Farmers were also asked to elucidate if dairy cows graze together with replacement cows during the non-lactating period, if grazed grassland is mowed for conservation, about the type of drinking water facilities used and the extent of spreading cattle manure on grazed grassland. A fourth section of questions asked about the annual frequency, treatment months and brand name of the most recently used anthelmintics and the understanding of farmers regarding their efficacy against gastrointestinal nematodes (“gutworms”), liver fluke, rumen fluke and lungworm (separate questions for each). Finally, the farmers were asked which of the above parasites had been confirmed as present on their farm by laboratory analysis in the previous five years and if they treat a proportion of or the entire milking herd during non-lactating period. Hard copies were returned between November and December 2013 and farm addresses were at townland level. 2.3. Classification of brand name of anthelmintic product Within the survey, the suitability of a brand name of anthelmintic product for the control of a specific parasitic group was determined based on the claims of the product as provided by the Health Products Regulatory Authority of Ireland (HPRA, 2014). If a farmer stated within the survey that two or more anthelmintic products were used in a herd, both products were documented and included in the analysis. The responses that contained products not suitable for the control of a specific parasitic group (e.g. levamisole for liver fluke) were classified as ‘unsuitable’. If the farmer did not provide the name of anthelmintic or the product name was illegible, then the response was omitted from the analysis. 2.4. Collection of milk samples The BTM samples were collected by farmers as described previously (Selemetas et al., 2015). Briefly, all BTM samples were treated with a tablet containing 8 mg Bronopol and 0.3 mg Natamycin (Broad Spectrum Microtabs II; Advanced Instruments, Inc., USA) for preservation, before posting to the Veterinary Parasitology Laboratory of University College Dublin for testing for the presence of liver fluke antibodies. The BTM samples after reception at the laboratory were stored at 4 ◦ C for at least 8 h and then were centrifuged at 850 × g for 10 min. Following the removal of the fat layer the remaining non-lipid fluid was aliquoted and stored at −20 ◦ C until tested. 2.5. Detection of F. hepatica antibodies The levels of liver fluke antibodies were determined, using an in-house antibody-detection ELISA applied to BTM

samples (Selemetas et al., 2014). The ELISA is based on the F. hepatica recombinant mutant Cathepsin L1 (rmutFheCL1) antigen (Collins et al., 2004) and uses a monoclonal anti-bovine IgG1 conjugated to horse radish peroxidase (Prionics, Zurich, Switzerland). The antibody levels of each BTM sample were expressed as per cent positivity (PP) based on the ELISA optical density (OD), using the following equation: PP value =

OD of test sample × 100 mean OD of positive controls

2.6. ELISA cut-off value The cut-off value of the ELISA test to define the farms where cows were positive for liver fluke antibodies (positive farms), was determined using the Mixtools package in R3.0.3 language (R Core Team, 2013) and the algorithm of expectation-maximization (EM) of the normal mixEM function (Benaglia et al., 2009). This principle of this method is the fitting of finite mixtures of multivariate and univariate normal distributions by providing computational techniques for finite mixture model analysis, where subgroups are distributions that are almost entirely unspecified. This allowed a bimodal model of two normal distributions for the two groups (positive and negative farms) to be fitted to the histogram of PP values of BTM samples. 2.7. Mapping and georeferencing The geographical information systems software that was used for the construction of spatial maps and the spatial join analysis to designate the farms into classes of risk regions was ArcGIS version 10.1 (ESRI, Redlands, CA, USA). Each dairy farm was georeferenced by assigning its location to the centroid of the townland on high precision satellite images. The sampled farms were divided into four classes of risk regions (low, moderate, high, very high) according to the predicted probability of exposure to liver fluke as described previously (Selemetas et al., 2015). Briefly, the predicted risk of fasciolosis in Ireland was calculated by applying the Random Forest modelling technique (Breiman, 2001), which is a combination of tree predictors that are randomly generated by bootstrapping from the full dataset with replacement but with the same distribution as the complete dataset. Using mainly climatic, especially precipitation, and also soil variables as major predictors of the exposure to liver fluke, a risk map of the predicted exposure to liver fluke was constructed, with higher probability shown in western and south-western regions of Ireland (Selemetas et al., 2015). 2.8. Statistical analysis Data from the questionnaire responses were extracted into Excel and then imported into SPSS Statistics version 20.0 (IBM Corp., USA) for descriptive analysis and calculation of group means and 95% confidence intervals (CI). The significant differences in survey variables between positive and negative farms were identified using

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

231

Fig. 1. Histogram of the frequency distribution of per cent positivity (PP) values of the ELISA test for the bulk tank milk samples and the fitted normal distributions of the two subgroups (negative and positive farms).

the independent sample t-test for numerical variables and the Pearson’s chi-square test for the nominal variables. Also, the significant differences between multiple-groups nominal survey-variables for numerical variables were identified using the Kruskal Wallis test. The significance was set at P < 0.05. In order to examine the associations of the dependent variable of exposure to liver fluke against a number of independent variables in the survey, univariable and multivariable logistic regression analysis of the variables was performed using SPSS Statistics version 20.0. All non-binary dependent variables were dichotomised for binary logistic regression, using the approximation of the median for numerical variables. Firstly, a univariable analysis of dependent and independent variables was performed with independent variables scoring P-values <0.10 in the univariable analysis being included in the multivariable model. At the next step, a forward stepwise multivariable logistic regression analysis was completed with variables scoring P-values ≤0.05 being retained in the final model. 3. Results 3.1. Response rate Of the 450 dairy farmers that were contacted, 374 (83%) submitted a BTM sample providing a valid address of their farm, 370 (82%) completed the questionnaire and 369 (82%) farmers did both and only the latter participants were included in the analysis of the survey. 3.2. ELISA testing results A cut-off value of ≥16 PP in the ELISA test was used to determine the positive farms, as seen in the normal distributions of the bimodal model (positive and negative farms) fitted in the histogram of the frequency distribution of PP values of BTM samples (Fig. 1). The geographical distribution of the 369 sampled farms in this study is illustrated in Fig. 2. The ELISA testing showed that 78% (n = 288) of dairy herds had been exposed to liver fluke. The distribution of sampled farms according to their classification

into risk regions is shown in Fig. 3, with the prevalence of liver fluke exposure within the four risk categories (low, moderate, high, very high) being 71%, 69%, 76% and 92%, respectively.

3.3. Farm structure and pasture management The average dairy herd size of farms in the study was 97 (95% CI 90–111) cows with a range of 20–850 cows. The average number of dairy cows contributing to BTM samples of the sampled dairy farms was 77 (95% CI 72–88) cows with a range of 10–550 cows. There was a significant difference between positive and negative farms in the dairy herd size and the number of dairy cows contributing to BTM samples, with positive farms being larger for both comparisons (Table 1). The percentage of farms where dairy cows shared pasture with small ruminants, horses and deer was 6% (n = 22), 2.4% (n = 10), and 2.3% (n = 9), respectively. There was no significant difference in the presence of small ruminants (df = 2; 2 = 0.482; P = 0.84), horses (df = 2; 2 = 0.28; P = 1.00) or deer (df = 2; 2 = 1.15; P = 0.59) that shared pasture with dairy cows between the positive and negative farms. The average size of the total area of grassland was 38 (95% CI 36–43) ha with 23 (95% CI 22–24) paddocks and the corresponding ranges were 4–243 ha and 6–50 paddocks. The mean paddock size was 1.6 (95% CI 1.4–1.8) ha with a range of 0.2–6.6 ha. The mean stocking density was 2.7 (95% CI 2.6–2.8) LU/ha with a range of 1–5 LU/ha. There was a significant difference between positive and negative farms in the total size of grazed grassland and in mean paddock size, with positive farms having larger area of grassland and mean paddock size (Table 1). There was no significant difference (P > 0.05) between positive and negative farms in the number of paddocks of grazed grassland or in the stocking density (Table 1). The low-risk regions had significantly higher dairy herd size (df = 3; 2 = 12.59; P = 0.006), number of dairy cows contributing to BTM samples (df = 3; 2 = 8.06; P = 0.045) and stocking density (df = 3; 2 = 8.19; P = 0.042) compared to other risk regions (Table 2).

232

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

Fig. 2. Map of Ireland showing the location of dairy farms sampled (n = 369) and the distribution of those farms that tested negative (green dots) and positive (red dots) for liver fluke. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)

3.4. Importance of drainage Farmers classified their grassland between three drainage regimes: (i) good drainage in soils that never saturate even on very wet winter days, (ii) moderate drainage in soils that may saturate on wet winter days but return to normal on first dry day, and (iii) poor drainage in soils that saturate on wet winter days and water surplus drains at slow rate. The proportion of farmers that identified their farm’s drainage status as good (56%) and moderate (40%) was significantly higher (P < 0.001) compared

to those that identified it as poor (4%). There were no differences (P > 0.05) between drainage statuses for the dairy herd size, the number of dairy cows contributing to BTM samples, the stocking density or the total area of grassland (Table 3). However, there were significant differences between drainage statuses for GAL (df = 2; 2 = 19.08; P < 0.001) and GSL (df = 2; 2 = 18.47; P < 0.001), with farms on good drainage having longer GAL and GSL than moderately drained farms (Table 3). There was also a significant difference between positive and negative farms in the drainage statuses (P < 0.001)

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

233

Fig. 3. Map of Ireland showing the distribution of farms in four classes of risk regions (low, moderate, high, very high) according to their predicted probability of exposure to liver fluke based on a predictive distribution model using climatic and environmental risk factors (Selemetas et al., 2015).

with all farms declared as having good drainage being exposed to liver fluke and a tendency for more negative farms on moderate drainage (Table 4). There were no significant differences between different drainage statuses for the percentage of farmers that reported grazing of dry cows together with replacement cows (df = 2; 2 = 0.76; P = 0.73) or mowing of grazed grassland for conservation (df = 2; 2 = 1.11; P = 0.58). There were no significant differences between different drainage statuses for the percentage of farmers that reported having conducted previous testing for liver

fluke (df = 2; 2 = 2.8869; P = 0.26) and rumen fluke (df = 2; 2 = 1.23; P = 0.54) within the last five years and having positive test results (farmer-reported prevalence). The farmer-reported liver fluke prevalence for farms with good drainage was more than half that of their prevalence based on the ELISA testing results of the current study (Table 4). 3.5. Grazing management The average GAL of the study farms was 264.9 (95% CI 260.2–269.6) days and the average GSL was 262.5

234

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

Table 1 Summary statistics of the independent t-test for the comparison between positive and negative farms for liver fluke with the independent variables of survey questions. Variable

Positive farms Mean (95% CI)

Negative farms Mean (95% CI)

t-Test value

P-value

Dairy farms (n) Total adult dairy cows (n) Dairy cows contributing to bulk milk sample (n) Grassland area (ha) Grassland paddocks (n) Mean paddock size (ha) Stocking density (LUa /ha) Absolute grazing annual length (days) Grazing season lengthb (days) Annual treatments for liver fluke (n)

288 100.3 (92.0–117.1) 82.7 (73.1–92.3) 39.7 (36.9–45.2) 23.7 (22.6–24.6) 1.7 (1.6–1.8) 2.7 (2.5–2.9) 263.8 (258.2–269.4) 261.3 (257.6–265.1) 1.23 (1.15–1.30)

81 83.7 (72.2–94.4) 67.1 (57.2–77.0) 32.7 (28.2–38.4) 21.9 (20.1–24.5) 1.4 (1.3–1.5) 2.8 (2.6–3.0) 269.4 (262.9–276.0) 264.26 (258.5–270.1) 1.08 (0.91–1.25)

– 2.487 2.502 2.072 1.729 1.996 0.391 1.568 1.668 1.401

– 0.013 0.013 0.039 0.085 0.023 0.858 0.118 0.096 0.162

a Livestock unit (LU) is the grazing equivalent of one adult dairy cow having an annual milk production of three tonnes without supplementary concentrated feed. For more details, see: http://epp.eurostat.ec.europa.eu/statistics explained/index.php/Glossary:LU. b Each full-time grazing day counts as one and each part-time grazing day counts as 0.5 days.

(95% CI 259.2–265.6) days. There were no differences (P > 0.05) between positive and negative farms for GAL and GSL (Table 1). The GAL and GSL where dairy cows grazed on farms with good drainage was 14 and 11 days longer than for those with moderate drainage respectively (P < 0.001; Table 3). There were significant differences between risk regions of the predicted probability of exposure to liver fluke for GAL (df = 3; 2 = 18.181; P < 0.001) and GSL (df = 3; 2 = 15.742; P < 0.001), with farms in very high- and high-risk regions having shorter GSL and DAP

than farms in moderate and low-risk regions. There was no significant difference between positive and negative farms in the percentage of farmers that reported grazing of dry cows together with replacement cows (df = 1; 2 = 4.15; P = 0.06), or mowing of grazed grassland for conservation (df = 3; 2 = 15.74; P < 0.001), or spreading cattle manure on grassland (df = 1; 2 = 0.04; P = 0.90), or used different types of drinking water facilities (bite or nipple drinkers vs troughs or tanks) (df = 1; 2 = 1.50; P = 0.25).

Table 2 Summary statistics of the farm structure variables of survey questions between the four classes of risk regions of exposure to liver fluke (low, moderate, high, very high) based on the predictive distribution model (Selemetas et al., 2015) with 95% confidence intervals in parentheses. Variable

Total adult dairy cows (n) Dairy cows contributing to BTM sample (n) Stocking density (LUa /ha) a

Predicted risk regions of liver fluke exposure

P

Low

Moderate

High

Very high

132.2 (95.1–169.3) 95.1 (73.2–117.2)

93.4 (83.2–103.1) 72.2 (64.2–80.3)

89.2 (76.2–101.4) 72.3 (61.3–83.1)

93.3 (80.2–105.1) 78.4 (67.2–89.1)

3.3 (2.2–4.4)

2.59 (2.4–2.8)

2.7 (2.5–3.0)

2.6 (2.5–2.8)

0.006 0.045 0.042

Livestock unit.

Table 3 Percentages of study farms and farms in risk regions of exposure to liver fluke in each drainage status (with the corresponding number of farms in parentheses) and mean number of total dairy cows, number of dairy cows contributing to bulk tank milk (BTM) samples, grazed grassland (ha & paddocks), stocking density (LUa /ha), absolute grazing annual length (GAL) (days) and grazing season length (GSLb ) (days) on dairy farms differing in land drainage status (95% confidence intervals in parentheses). Drainage status

Farms % (n) Low risk farms % (n) Moderate risk farms % (n) High risk farms % (n) Very high risk farms % (n) Total dairy cows (n) Dairy cows contributing to BTM sample (n) Stocking density (LUa /ha) Grazed grassland (ha) Grassland paddocks (n) GAL (days) GSLb (days) a b

P-Value

Poor

Moderate

Good

4 (16) 0 (0) 5 (5) 6 (6) 4 (5) 81.5 (75.0–88.0) 71.8 (64.9–78.7) 2.2 (2.1–2.4) 36.1 (33.0–39.2) 24.3 (21.7–26.9) 246.3 (241.9–250.8) 247.0 (243.7–250.3)

40 (144) 34 (16) 33 (30) 37 (39) 50 (59) 91.7 (78.1–105.3) 75.6 (61.2–90.0) 2.6 (2.3–2.8) 35.7 (31.9–39.6) 22.8 (20.5–25.2) 256.7 (252.0–261.3) 255.2 (251.7–258.8)

56 (200) 66 (31) 62 (56) 57 (60) 45 (53) 103.1(86.1–120.2) 79.2 (66.5–91.9) 2.5 (2.3–2.7) 40.7 (34.6–46.8) 23.6 (20.6–26.6) 270.9 (267.2–274.6) 266.7 (264.1–269.3)

Livestock unit. Each full-time grazing day counts as one and each part-time grazing day counts as 0.5 days.

<0.001 0.964 0.989 0.997 0.977 0.297 0.671 0.107 0.498 0.259 <0.001 <0.001

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

235

Table 4 Negative and positive study farms, liver fluke prevalence and farmer-reported results of previous parasite testing within the last five years in dairy farms differing in land drainage status. Results are percentages of farmer’s answers within each drainage status (with the corresponding number of farmers in parentheses). Drainage status

Negative farms % (n) Positive farms % (n) Liver fluke prevalancea % (n) Farmer-reported prevalanceb Positive for liver fluke % (n) Positive for rumen fluke % (n) Positive for gastrointestinal nematodes % (n) Positive for lungworm % (n) a b

P-Value

Poor

Moderate

Good

24 (16) 0 (0) 0 (0)

76 (52) 32 (92) 64 (92)

0 (0) 68 (200) 100 (200)

<0.001 <0.001 <0.001

44 (7) 38 (6) 0 (0) 6 (1)

58 (83) 40 (57) 17 (25) 17 (24)

44 (87) 31 (62) 17 (34) 24 (48)

0.261 0.541 <0.001 <0.001

Proportion of farms tested positive for liver fluke in the current study based on ELISA testing on milk samples. Proportion of farms having previous positive parasite tests that were reported within the last five years.

3.6. Anthelmintic treatment Out of the 369 study farmers, the number of farmers that responded to the questions about treatment against gastrointestinal nematodes, liver fluke, lungworm and rumen fluke were 258 (70%), 351 (95%), 252 (70%) and 235 (70%), respectively. The proportion of different anthelmintic active ingredients used is presented in Fig. 4.

3.6.1. Liver fluke In total, 334 (95%) out of the 351 responding farmers reported using a treatment against liver fluke, with 311 (89%) farmers using a flukicide. Also, 256 (77%) farmers treated once per year against liver fluke and 78 (23%) farmers treated more than once per year (twice or three times). Approximately 96% of the farmers (n = 336) that treated their dairy cows against liver fluke, did so during the non-lactating period, mainly in December (61%). Approximately 2% of farmers (n = 7) used a product ‘unsuitable’ for treatment against liver fluke and 7% (n = 23) did not specify a product name. Albendazole was the most commonly used active ingredient against liver fluke (40%) with both triclabendazole and oxyclosanide the second most commonly used products (29%). The proportion of farmers that treated once per year against liver fluke by using either albendazole or triclabendazole or oxyclosanide was 70%, 88% and 72%, respectively. The use of triclabendazole was significantly higher among the farmers that treated once per year against liver fluke compared to the use of albendazole or oxyclosanide (df = 2; 2 = 2.38; P = 0.30). In addition, 82% of farmers (n = 274) treated the entire milking herd during the non-lactating period which was significantly higher than the percentage of 18% of farmers (n = 60) that treated only a proportion of the herd during the same period (df = 1; 2 = 9.07; P = 0.01). There was no significant difference (P > 0.05) in the annual number of treatments against liver fluke between the positive and negative farms (Table 1). There was a significant difference in the use of triclabendazole and albendazole between the positive and negative farms (df = 1; 2 = 4.34; P = 0.01), with triclabendazole being more likely to be used on the positive farms.

3.6.2. Gastrointestinal nematodes In total, 254 out of the 258 (98%) responding farmers reported using a treatment against gastrointestinal nematodes, with 211 (83%) farmers using an active ingredient effective against gastrointestinal nematodes. Also, 188 (74%) farmers treated once per year against gastrointestinal nematodes and 66 (26%) farmers treated more than once per year (twice or three times) with the same active ingredient. Approximately 85% of the farmers (n = 215) that treated their dairy cows against gastrointestinal nematodes did so during the non-lactating period, mainly in December (50%). Approximately 6% of farmers (n = 15) used a product ‘unsuitable’ for treatment against gastrointestinal nematodes and 17% (n = 43) did not specify a product name. Albendazole was the most frequently used active ingredient for treatment against gastrointestinal nematodes (57%), with eprinomectin the second most commonly used product (27%). However, among the farmers that treated more than once per year against gastrointestinal nematodes (n = 66), eprinomectin was the most frequently used active ingredient (44%) followed by albendazole (39%). 3.6.3. Lungworm In total, 242 out of the 252 (96%) responding farmers reported using a treatment against lungworm, with 225 (89%) farmers using an active ingredient effective against lungworm. Also, 206 (85%) farmers treated once per year against lungworm and 36 (15%) farmers treated more than once per year (twice or three times) with the same active ingredient. Approximately 77% of farmers (n = 186) that treated their dairy cows against lungworm did so during the non-lactating period, mainly in December (46%). Approximately 5% of farmers (n = 12) used a product ‘unsuitable’ for treatment against lungworm and 7% (n = 17) did not specify a product name. Albendazole was the most frequently used active ingredient for treatment against lungworm (47%), with eprinomectin the second most commonly used product (31%). 3.6.4. Rumen fluke In total, all 235 (100%) responding farmers reported using a treatment against rumen fluke, with 221 (94%) farmers using an active ingredient effective against rumen

236

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

Fig. 4. Proportion of surveyed Irish dairy farmers using each active ingredient for the treatment against gastrointestinal nematodes (GI nematodes), lungworm and liver fluke in dairy cows. Other includes abamectin, closantel, fenbendazole and moxidectin against GI nematodes, abamectin, moxidectin and oxfenbendazole against lungworm, and closantel and nitroxynil against liver fluke.

fluke (oxyclosanide). Also, 225 (96%) farmers treated once per year against rumen fluke and 10 (0.4%) farmers treated more than once per year (twice or three times). Approximately 90% of farmers (n = 212) that treated their dairy cows against rumen fluke did so during the non-lactating period, mainly in December (52%). Approximately 15% of farmers (n = 36) used a product ‘unsuitable’ for treatment against rumen fluke and 6% (n = 14) did not specify a product name. The proportion of farmers that used an active ingredient effective against rumen fluke was significantly higher among the positive farms for liver fluke compared to the negative farms (df = 1; 2 = 9.94; P = 0.002).

3.6.5. Risk regions There were significant differences between the predicted risk regions of exposure to liver fluke in the number of treatments for liver fluke (df = 1; 2 = 4.02; P = 0.045), with farmers in high-risk regions being more likely to treat their stock against liver fluke more times than farmers in low-risk regions. There were significant differences between the predicted risk regions of exposure to liver fluke in the active ingredients used against liver fluke (df = 6; 2 = 12.59; P = 0.048) and lungworm (df = 3; 2 = 7.93; P = 0.047), with albendazole being more likely to be used against both parasites in low-risk regions and with triclabendazole and eprinomectin being more likely to be used against liver fluke and lungworm respectively in high-risk regions. Finally, there were no significant differences between the predicted risk regions of exposure to liver fluke in the number of annual treatments against gastrointestinal nematodes (df = 3; 2 = 0.01; P = 0.98) and lungworm (df = 3; 2 = 0.01; P = 0.98) and in active ingredients used against gastrointestinal nematodes (df = 6; 2 = 2.19; P = 0.90) and lungworm (df = 6; 2 = 2.53; P = 0.86).

3.7. Previous parasite testing In total, 16% of study farmers (n = 59) reported having previous positive test results for gastrointestinal nematodes, 49% (n = 179) for liver fluke, 35% (n = 128) for rumen fluke, 20% (n = 73) for lungworm and 40% (n = 147) as having tested for none of the above. Of those farmers (n = 179) that reported having conducted previous testing for liver fluke and having positive test results, 153 (85%) had a positive ELISA testing result in the current study (Table 4). The number of confirmed previous test results for liver fluke was significantly higher in positive farms compared to negative farms (df = 1; 2 = 12.18; P = 0.02). There was no significant difference between farms with positive and negative previous test results for liver fluke in the annual number of treatments against liver fluke (df = 3; 2 = 3.49; P = 0.58). In addition, farms that had confirmed previous test results for gastrointestinal nematodes or rumen fluke or lungworm were more likely to treat against the above parasites (df = 3; 2 = 37.00; P < 0.001). 3.8. Multivariable logistic regression model Based on the results of the univariable regression analysis (Table 5) and the statistically significant multivariate model (P < 0.001), the exposure to liver fluke was positively correlated with the risk regions due to the predicted probability of exposure to liver fluke, adjacency of grazed grassland to snail habitats, poor drainage, treatment against rumen fluke and treating the entire herd for liver fluke (Table 6). The farms in very high-risk regions were approximately three, five and three times more likely to be exposed to liver fluke compared to farms in low-, moderate-, and high-risk regions respectively. The farms with grazed grassland close to snail habitats were about twice more likely to be exposed to liver fluke compared

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

237

Table 5 Univariable logistic regression analysis of association between exposure to liver fluke and all survey variables, with non-binary dependent variables being dichotomised for further multivariable logistic regression analysis. Survey variables

Responses

Total adult dairy cows Cows contributing to bulk tank milk sample Grassland area (ha) Grassland paddocks (n) Mean paddock size (ha) Stocking density (LUa /ha) Absolute grazing annual length (days) Grazing season lengthb (days) Drainage

Poor, moderate, good

Grassland close to snail habitats

<25%, 25–50%, 50–75%, >75%

Co-grazing of non-lactating cows with replacement cows

Yes, no

Mowing grazed grassland for conservation

Yes, no

Spreading cattle manure on grassland

Yes, no

Type of drinking system

Bite and nipple drinkers, troughs, tanks

Treating entire herd against liver fluke

Yes, no

Annual number of liver fluke treatments

1, 2, 3

Treatment against rumen fluke

Yes, no

a b

Binary variables

Liver fluke prevalence

P value

<80 cows ≥80 cows <60 cows ≥60 cows <35 ha ≥35 ha <25 paddocks ≥25 paddocks <1.5 ha ≥1.5 ha <2.5 LU/ha ≥2.5 LU/ha <270 days ≥270 days <270 days ≥270 days Poor and moderate

79.1% 78.1% 78.8% 77.8% 75.1% 82.4% 75.0% 80.1% 75.6% 81.4% 81.1% 76.0% 80.8% 76.0% 80.6% 76.1% 89.4%

<0.001

Good No (<25%)

68.5% 73.9%

0.001

Yes (≥25%) Yes No Yes No Yes No Bite and nipple drinkers

90.3% 72.3% 79.9% 78.6% 77.8% 77.7% 80.8% 73.9%

Trough, tanks Yes No 1 >1 Yes No

79.7% 81.6% 66.1% 79.3% 82.2% 83.4% 62.3%

0.091 0.077 0.039 0.086 0.025 0.086 0.097 0.097

0.044 0.851 0.720 0.052

0.008 0.151 <0.001

Livestock unit. Each full-time grazing day counts as one and each part-time grazing day counts as 0.5 days.

with farms where the grazing fields were more distant from snail habitats. Good drainage categorisation was not regarded as a significant predictor of exposure to liver fluke, whereas the poor drainage categorisation, which included

the farmers responses of poor and moderate drainage, was associated with lower odds of exposure to the parasite. The farmers that treated against rumen fluke were more than twice as likely to also face problems with liver fluke

Table 6 Multivariate model of association between the exposure to liver fluke and the most significant independent variables (risk-region of predicted probability of exposure to liver fluke, drained grassland, grassland close to snail habitats, treatment against rumen fluke and treating entire herd for liver fluke), displaying odds ratio and probability of each dichotomised variable and its subcategories containing group(s) of survey responses. In the case of risk region variable, the very-high risk class was compared with the other classes. Variable (dichotomised class) Risk region Low risk region Moderate risk region High-risk region Drainage Poor drainage Good drainage Grassland close to snail habitats No Yes Treatment against rumen fluke Treating entire herd against liver fluke

Class/survey response

Odds ratio

95% confidence interval

Very high vs low risk Very high vs moderate risk Very high vs high risk

3.01 4.98 2.93

1.13–8.06 2.12–11.63 1.23–6.94

Poor and moderate Good

0.37 1.16

0.18–0.75 0.18–7.29

<25% ≥25% Yes vs no Yes vs no

4.29 8.04 2.24 2.80

1.79–10.28 2.78–23.26 1.23–4.09 1.36–5.78

P value 0.003 0.028 <0.001 0.015 0.020 0.006 0.878 <0.001 0.001 <0.001 0.009 0.005

238

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

exposure by their dairy cows than farmers that did not treat for rumen fluke. Finally, the overall percentage of cases that were correctly predicted by the full logistic regression model was 82.2%. 4. Discussion The purposes of the current study were to examine the effect of several farm management practices on liver fluke prevalence on Irish dairy farms and to document the current anthelmintic control measures for all major internal parasites of grazing dairy cows in Ireland. The main sources of errors within questionnaire surveys such as the present study are non-response errors (reduced response rate), frame or coverage errors (misrepresentative sampling) and measurement errors (inaccurate response) (de Leeuw, 2005). The response rate of this survey was quite high (82%) and, therefore, the non-response error was relatively low. The survey population was widespread enough to cover the whole national area of dairy farming in Ireland and was proportional to the distribution of dairy farms in each county. Therefore, geographical coverage error was also likely to be low. However, the fact that the farmers participating in this survey used the Teagasc advisory service may have increased coverage error and decreased measurement error, as these farmers may have more technical expertise than farmers that do not use this advisory service. Another source of coverage error may also be the fact that the farms in the current study had larger herd size compared with the Irish dairy farm average (n = 67) based on the results for dairy farms in the Irish National Farm Survey for 2012 (Hennessy et al., 2013). The ELISA testing of BTM samples showed that 78% of the dairy herds had been exposed to liver fluke, which is similar to the 82% prevalence in dairy herds during autumn demonstrated by Selemetas et al. (2015), but higher than the prevalence found in culled cows (65%) during autumn and summer (Murphy et al., 2006). The higher prevalence in the current study compared to that found by Murphy et al. (2006) has occurred either due to different climatic and environmental conditions or due to different sampling technique, or because the BTM samples were collected between October and December, a time period where the peak of the annual cycle of fasciolosis is expected due to pasture contamination at the end of summer (SalimiBejestani et al., 2005a). The detection of exposure to liver fluke using an ELISA test on BTM samples is a popular, easy and rapid diagnostic method, but the results on a single BTM sample rely on seropositivity of the lactating cows contributing to the BTM sample and the in-herd prevalence (Duscher et al., 2011). Another limitation of this antibodydetection test is that the positive result reveals a previous exposure to the parasite rather than an existing infection (Salimi-Bejestani et al., 2005b). Although small ruminants, horses or deer are suitable hosts of liver fluke, there was no significant difference in the presence of these definite hosts that shared pasture with dairy cows between positive and negative farms, probably due to the small number of farmers confirming the presence of these potential hosts. In the current study it is surprising the high liver fluke prevalence among the dairy

farms with moderate and good drainage. This is in contrast with the findings by Bennema et al. (2011), who showed that low drained soil types were related with higher risk of fasciolosis, and by Selemetas et al. (2014) demonstrating that poorly drained soil was the main soil type in herds exposed to liver fluke. One explanation is that the drainage status of farms was a product of subjective evaluation as it was reported by the farmers themselves, with only 4% of farmers identifying their farm’s drainage status as poor. Another explanation may be that liver fluke infection levels may be higher on farms with longer GSLs (Bennema et al., 2011). Indeed, farms with good drainage had longer GAL and GSL than poorly and moderately drained farms. However, in the current study there were no significant differences between positive and negative farms for GAL and GSL, but farms in very high- and high-risk regions had lower GAL and GSL and higher liver fluke prevalence than farms in moderate and low-risk regions. Therefore, it is possible that under Irish conditions good drainage allows longer GAL and GSL, but the liver fluke exposure is more related to higher precipitation than poor land drainage classification and, therefore, farms in high-risk regions, which are characterised by high levels of precipitation (Selemetas et al., 2015), present lower GAL and GSL than farms in low-risk regions. The fact that farms with good drainage presented higher prevalence is extremely important, because many farmers on farms with good drainage might assume they are at a lower risk and may not treat their cows with anthelmintics active against liver fluke as frequently as recommended. These findings highlight the importance of communicating appropriate advice effectively to farmers on efficient parasite control. In the current study, 29% of the farmers that treated their dairy cows against liver fluke used triclabendazole, which is the only active ingredient effective against both adult and early immature liver flukes (Power et al., 2013). The remaining farmers used mainly albendazole and oxyclosanide that are active only against mature liver flukes and may need a follow-up treatment (Fairweather and Boray, 1999; Power et al., 2013). However, the majority of the latter farmers (70%) treated once per year against liver fluke and, therefore, it is likely that the control of liver fluke was not effective enough and improvements need to be made in the liver fluke antiparasitic programme on many Irish dairy farms. The popularity of albendazole and oxyclosanide, which is in agreement with the findings of Bloemhoff et al. (2014), can be attributed to the flukicide licensing restrictions (HPRA, 2013) that allow only the use of these two active ingredients for the control of liver fluke in lactating dairy cows. Albendazole was also the most popular active ingredient against gastrointestinal nematodes and lungworm, which is also in agreement with the findings of Bloemhoff et al. (2014), due to the ease of using a single active ingredient against several groups of parasites. However, this method of antiparasitic treatment by simultaneously applying the same class of drugs having the similar mode of action risks accelerating the appearance of anthelmintic resistance (Gasbarre, 2014). This highlights the need of an

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

improved policy on chemoprophylactic use and of communicating this information to Irish dairy farmers. A logistic regression analysis was conducted to predict the exposure of dairy herds to liver fluke using various variables of the survey questions as predictors. Based on the findings of the multivariate model, the exposure to liver fluke was associated with the increased risk of exposure to the parasite as predicted according to the spatial distribution model for Ireland (Selemetas et al., 2015). Indeed, the farms in very high-risk regions had constantly higher odds of exposure than the farms in other risk regions (Fig. 3), highlighting the need of fine-scale mapping of the disease patterns even at farm level in order to increase the accuracy of risk maps for fasciolosis. In addition, the exposure to liver fluke was highly associated with the proximity of grazed grassland to snail habitats like streams, ponds, dykes and flooded ditches and with poor drainage, but to a lesser extent, stressing the importance of impeded drainage to the transmission of fasciolosis (Charlier et al., 2011). The treatment against rumen fluke appears also to be related with the exposure to liver fluke, suggesting the existence of a possible epidemiological link between the controls of both parasites. Oxyclosanide is the only active ingredient effective against both rumen fluke and mature liver flukes (Paraud et al., 2009), with both parasites using the same intermediate snail host (Augot et al., 1996). Therefore, it is possible that either farmers use the same active ingredient for both parasites or they were confused regarding the question on parasites control. Finally, the exposure to liver fluke was positively correlated to the farmers decision to treat the entire herd against liver fluke during the nonlactating period, instead of treating only a proportion of the herd (selected cows, known infected cows or cows with lower milk production). Indeed, farmers whose herds were infected with liver fluke in the past or are currently exposed to the parasite are more likely to treat the entire herd during non-lactating period for practical reasons, mainly to reduce cost of labour and the risk of contaminating milk supply with drug residues. Significant differences between different categories of risk regions of exposure to liver fluke (Selemetas et al., 2015) in the number of treatments and the active ingredients used for liver fluke were noted in the current study. Dairy farmers in high-risk regions were more likely to treat their cows against liver fluke more times annually, and use triclabendazole which is effective against both adult and immature liver flukes (Fairweather and Boray, 1999), compared to farmers in low-risk regions that were more likely to treat their cows fewer times and use albendazole, which is not effective against all stages of liver fluke as triclabendazole (Richards et al., 1990). Regional differences in the parasite control and active ingredients used for the treatment against liver fluke were also found in the study by Bloemhoff et al. (2014), with dairy farmers in western part of Ireland being more likely to treat against liver fluke and also use nitroxynil compared to farmers from other areas. However, the farms of the latter study were divided into three regional categories at county level based on climatic conditions and soil type. The comparisons between administrative counties is arbitrary and not of epidemiological significance, because there is a big diversity of climatic and

239

environmental conditions, and co-existence of high- and low-risk regions of exposure to liver fluke, within counties (Selemetas et al., 2015). The presence of significant regional differences in dairy farm management, and also parasite control practices, highlight the complexity of risk factors affecting the epidemiology of parasitic diseases and the need for fine-scale data at farm level on farming management to support optimal parasite control. The proportion of farmers that used a product classified as ‘unsuitable’ for the control of gastrointestinal nematodes, liver fluke and lungworm was relatively low, at 6%, 2% and 5%, respectively. This is in agreement with the proportion of farmers (3%) that used a product ‘unsuitable’ for liver fluke, and in contrast with the proportion of farmers (30%) that used a product ‘unsuitable’ for nematodes, based on the findings of the study of Bloemhoff et al. (2014). The low proportion seen in the current study may be due to the technical expertise of the dairy farmers as a result of their access to the Teagasc advisory service. In addition, 77% of farmers treated their dairy cows against lungworm during the non-lactating period, mainly in December, although mass treatment before the grazing season is a more effective control measure against lungworm (Ploegera and Holzhauerb, 2012). This highlights the need for technical advice and advisory programmes to ensure an effective parasite control strategy. On the other hand, the proportion of farmers that used a product ‘unsuitable’ for rumen fluke was relatively high (15%), although there is only one active ingredient effective for treatment (oxyclosanide), which may reflect confusion among farmers on different parasites and anthelmintic products, and lack of accurate parasite control information. The current study identified the major farm management practices that can favour the exposure to liver fluke on Irish dairy farms and documented the current measures employed by Irish farmers on dairy management practices and chemoprophylactic use. Although most of the farmers presented a generally good level of understanding of farm and parasite control management practices, there appeared to be a shortfall in aspects of technical advice being applied on many Irish dairy farms. Farmers require appropriate and comprehensive advice in order to efficiently and profitably manage internal parasite challenges in their dairy farming enterprises. Surveys such as this can help identify areas of knowledge requiring advancement or areas of practice needing modification on dairy farms. Acknowledgements The authors gratefully acknowledge the assistance and cooperation of Teagasc advisors, and farmers in providing samples for this study. Map of Ireland was sourced from Urban Institute of Ireland. The research leading to these results has received funding from the European Community’s Seventh Framework Programme (FP7/2007–2013) under Grant Agreement No. 288975. References Augot, D., Abrous, M., Rondelaud, D., Dreyfuss, G., 1996. Paramphistomum daubneyi and Fasciola hepatica: the redial burden and cercarial

240

N. Selemetas et al. / Veterinary Parasitology 207 (2015) 228–240

shedding in Lymnaea truncatula subjected to successive unimiracidial cross-exposures. Parasitol. Res. 82, 623–627. Benaglia, T., Chauveau, D., Hunter, D.R., Young, D., 2009. mixtools: an R package for analyzing finite mixture models. J. Stat. Softw. 32 (6), 1–29. Bennema, S.C., Ducheyne, E., Vercruysse, J., Hendrickx, G., Claerebout, E., Charlier, J., 2011. Relative importance of management, meteorological and environmental factors in the spatial distribution of Fasciola hepatica in dairy cattle in a temperate climate zone. Int. J. Parasitol. 41, 225–233. Bloemhoff, Y., Danaher, M., Forbes, A., Morgan, E., Mulcahy, G., Power, C., Sayers, R., 2014. Parasite control practices on pasture-based dairy farms in the Republic of Ireland. Vet. Parasitol., http://dx.doi.org/ 10.1016/j.vetpar.2014.05.029. Breiman, L., 2001. Random forests. Mach. Learn. 45, 5–32. Charlier, J., Bennema, S.C., Caron, Y., Counotte, M., Ducheyne, E., Hendrickx, G., Vercruysse, J., 2011. Towards assessing fine-scale indicators for the spatial transmission risk of Fasciola hepatica in cattle. Geospatial Health 5 (2), 239–245. Collins, P.R., Stack, C.N., O’Neill, S.M., Doyle, S., Ryan, T., Brennan, G.P., Mousley, A., Stewart, M., Maule, A.G., Dalton, J.P., Donnelly, S., 2004. Cathepsin L1, the major protease involved in liver fluke (Fasciola hepatica) virulence: propeptide cleavage sites and autoactivation of the zymogen secreted from gastrodermal cells. J. Biol. Chem. 279, 17038–17046. Creighton, P., Kennedy, E., Boland, T., Shalloo, L., O’Donovan, M., 2011. A survey analysis of grassland dairy farming in Ireland, investigating grassland management, technology adoption and sward renewal. Grass Forage Sci. 66, 251–264. de Leeuw, E.D., 2005. To mix or not to mix data collection modes in surveys. J. Off. Stat. 21, 233–255. Dillon, P., Crosse, S., Stakelum, G., Flynn, F., 1995. The effect of calving date and stocking rate on the performance of spring-calving dairy cows. Grass Forage Sci. 50, 286–299. Duscher, R., Duscher, G., Hofer, J., Tichy, A., Prosl, H., Joachim, A., 2011. Fasciola hepatica – monitoring the milky way? The use of tank milk for liver fluke monitoring in dairy herds as base for treatment strategies. Vet. Parasitol. 178, 273–278. European Commission, 2009. Commission regulation 1200/2009 of 30 November 2009 implementing regulation 1166/2008 of the European Parliament and of the Council on farm structure surveys (FSS) and the survey on agricultural production methods (SAPM), as regards livestock unit coefficients and definitions of the characteristics. Off. J. Eur. Union L 329, 28. Fairweather, I., Boray, J.C., 1999. Fasciolicides: efficacy, actions, resistance and its management. Vet. J. 158, 81–112. Fox, N.J., White, P.C.L., McClean, C.J., Marion, G., Evans, A., Hutchings, M.R., 2011. Predicting impacts of climate change on Fasciola hepatica risk. PLoS ONE 6 (1), e16126. Gasbarre, L.C., 2014. Anthelmintic resistance in cattle nematodes in the US. Vet. Parasitol. 204, 3–11. Health Products Regulatory Authority (HPRA), 2013. Use of Flukicides in Dairy Cows, http://www.hpra.ie/homepage/medicines/ safety-notices/item?t=/use-of-flukicides-in-dairy-cows&id=a5d4f825 -9782-6eee-9b55-ff00008c97d0 (accessed 01.09.14). Health Products Regulatory Authority (HPRA), 2014. Veterinary Medicines Information, http://www.hpra.ie/homepage/veterinary/veterinarymedicines-information/find-a-medicine (accessed 01.09.14). Hennessy, T., Moran, B., Kinsella, A., Quinlan, G., 2013. Teagasc National Farm Survey Results 2012. Agricultural Economics & Farm Surveys

Department, Rural Economy and Development Programme, Teagasc, Athenry, Co., Galway, pp. 79. Läpple, D., Hennessy, T., O’Donovan, M., 2012. Extended grazing: a detailed analysis of Irish dairy farms. J. Dairy Sci. 95, 188–195. McCann, C.M., Baylis, M., Williams, D.J.L., 2010a. The development of linear regression models using environmental variables to explain the spatial distribution of Fasciola hepatica infection in dairy herds in England and Wales. Int. J. Parasitol. 40, 1021–1028. McCann, C.M., Baylis, M., Williams, D.J.L., 2010b. The seroprevalence and spatial distribution of Fasciola hepatica infected dairy herds in England and Wales. Vet. Rec. 166, 612–617. Murphy, T.M., Fahy, K.N., McAuliffe, A., Forbes, A.B., Clegg, T.A., O’Brien, D.J., 2006. A study of helminth parasites in culled cows from Ireland. Prev. Vet. Med. 76, 1–10. O’Farrell, K.J., Downey, N.E., Sherington, J., 1986. The effect of anthelmintic treatment at calving on the subsequent milk production characteristics of dairy cows. Irish Vet. J. 40, 116–123. Paraud, C., Gaudin, C., Pors, I., Chartier, C., 2009. Efficacy of oxyclozanide against the rumen fluke Calicophoron daubneyi in experimentally infected goats. Vet. J. 180, 265–267. Ploegera, H.W., Holzhauerb, M., 2012. Failure to eradicate the lungworm Dictyocaulus viviparus on dairy farms by a single mass-treatment before turnout. Vet. Parasitol. 185, 335–338. Power, C., Sayers, R., O’Brien, B., Furey, A., Danaher, M., Jordan, K., 2013. Review of studies on flukicide residues in cows’ milk and their transfer to dairy products. Irish J. Agric. Food Res. 52, 197–207. R Core Team, 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, http://www.R-project.org/ Richards, R.J., Bowen, F.L., Essenwein, F., Steiger, R.F., Buscher, G., 1990. The efficacy of triclabendazole and other anthelmintics against Fasciola hepatica in controlled studies in cattle. Vet. Rec. 126, 213–216. Salimi-Bejestani, M.R., Daniel, R.G., Felstead, S.M., Cripps, P.J., Mahmoody, H., Williams, D.J., 2005a. Prevalence of Fasciola hepatica in dairy herds in England and Wales with an ELISA applied to bulk-tank milk. Vet. Rec. 156, 729–731. Salimi-Bejestani, M.R., McGarry, J.V., Felstead, S.M., Ortiz, P., Akca, A., Williams, D.J., 2005b. Development of an antibody-detection ELISA for Fasciola hepatica and its evaluation against a commercially available test. Res. Vet. Sci. 78, 177–181. Selemetas, N., Phelan, P., O’ Kiely, P., de Waal, T., 2014. Weather and soil type affect incidence of Fasciolosis in dairy cow herds. Vet. Rec., http://dx.doi.org/10.1136/vr.102437. Selemetas, N., Ducheyne, E., Phelan, P., O’ Kiely, P., Hendrickx, G., de Waal, T., 2015. Spatial analysis and risk mapping of Fasciola hepatica infection in dairy herds in Ireland, using random regression Forest modelling and geographical information systems. Geospatial Health (in press). Torgerson, P., Claxton, J.R., 1999. Epidemiology and control. In: Dalton, J.P. (Ed.), Fasciolosis. CABI Publishing, Oxon and New York, pp. 113–149. Waller, P.J., 1999. International approaches to the concept of integrated control of nematode parasites of livestock. Int. J. Parasitol. 29, 155–164. Waller, P.J., 2006. Sustainable nematode parasite control strategies for ruminant livestock by grazing management and biological control. Anim. Feed Sci. Technol. 126, 277–289. Walsh, S., 2012. A Summary of Climate Averages 1981–2010 for Ireland, Climatological Note No. 14. Met Éireann, Dublin, 4pp.