Doi:10.1016/j.prevetmed.2004.01.006

Preventive Veterinary Medicine 63 (2004) 237–256 A meta-analysis of the milk-production response after anthelmintic treatment in naturally Javier Sanchez , Ian Dohoo , Jeromy Carrier , Luc DesCˆoteaux a Department of Health Management, Atlantic Veterinary College, University of Prince Edward Island, 550 University Avenue, Charlottetown, PEI, Canada C1A 4P3 b Faculté de Médécine Veterinaire, Université de Montreal, Sciences Cliniques, St-Hyacinthe, Que., Canada J2S 7C6 Received 29 April 2003; received in revised form 18 December 2003; accepted 16 January 2004 Abstract
Meta-analysis was used to estimate the effects of anthelmintic treatment on milk production in dairy cows. The literature search included peer-reviewed journals (both full articles and abstracts),conference proceedings and theses and included documents written in English, Spanish, French, Por-tuguese or Italian. The study outcome was defined as the difference in milk production (kg/cow perday) between treated and untreated cows. Random effect meta-analyses were performed on 75 tri-als published between 1972 and 2002. The combined estimate after controlling for publication biasand/or small-study effect was of 0.35 kg/cow per day. Significant variation among studies was de-tected and although several variables were associated with the study outcome, they did not significantlyreduce the unexplained variability among trials. Trials reporting the use of endectocides had highermilk-production response compared with trials using older anthelmintics. Similarly, whole-herd treat-ment trials or trials which applied the treatment in mid-lactation or strategically throughout the yearhad higher response compared with calving or dry-period treatment trials. Trials reporting the resultsas total 305-day milk production had lower response compared with trials which measured productionas daily milk weight. Primiparous cows trials had lower responses compared with multiparous cowstrials.
2004 Elsevier B.V. All rights reserved.
Keywords: Anthelmintic treatment; Dairy cattle; Milk production; Meta-analysis; Nematodes ∗ Corresponding author. Tel.: +1-902-566-0995; fax: +1-902-566-0823.
E-mail address: [email protected] (J. Sanchez).
0167-5877/$ – see front matter 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.prevetmed.2004.01.006 J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 1. Introduction
The milk-production responses obtained from field trials of anthelmintic treatments in adult dairy cows have been equivocal, and consequently, clear guidelines as to when an-thelmintic treatments should be applied have not been available. Although adult dairycows can harbor an important number of gastrointestinal parasites (mainly O. ostertagi)the lack of accurate diagnostic tests forthis group of animals (makes it difficult toestablish a threshold for anthelmintic treatment ( To obtain an overall estimate of the effect of deworming adult dairy cattle on milk production, a narrative review of more than 80 trials in dairycattle and concluded that a median increase in milk production of 0.63 kg/cow per day mightbe expected after anthelmintic treatment. Although traditional narrative reviews have beenused widely in veterinary literature, they are subjective in nature and therefore prone to areviewer bias They also do not easily take into account theprecision of the treatment response estimates; so, studies tend to be weighted equally.
On the other hand, a meta-analysis allows a reviewer to arrive at conclusions that might be more accurate than those from a non-quantitative, narrative review (A meta-analysis is a systematic review of the literature followed by a quantitativecompilation of all relevant results in which the precision of each individual trial is takeninto account. A meta-analysis can be biased by the exclusion or inclusion criteria used inthe study selection process or by the methods chosen to combine results from the selectedstudies However, these biases can be minimized when adetailed protocol specifying the selection of the studies and collection and analysis of thedata is followed.
Our objective was to use a meta-analysis to estimate the extent to which anthelmintic treatment (with a variety of drugs and treatment protocols) influenced milk production indairy cows.
2. Material and methods
The literature review was based on the following databases: CAB Abstracts (1972–2002) and Medline (1966–2001). The keyword combinations used were “anthelmintic dairy cattle”,“milk production nematodes”, “milk production anthelmintic”, “dairy cows dairy herdsanthelmintics”. A total of 416 references related to parasitism in cattle were identified.
References were removed if the article pertained to species other than dairy cattle, per-tained to the use of anthelmintics in ages other than lactating-age dairy cows, the trialanimals were artificially infected, the trial did not measure milk production or if the ar-ticle was not written in English, Spanish, French, Portuguese or Italian. The search wasnot restricted to peer-reviewed journals and it included abstracts, conference proceedingsand theses. In addition, all the references related to milk-production trials cited in a re-cent review paper (were identified. A total of 78 potential articles were J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 identified for the meta-analysis. Bibliographies of retrieved articles were examined for fur-ther references.
2.2. Outcome evaluated and data extraction The mean difference in milk production between treated and control group in kg/cow per day was used as the outcome. If the trial reported this outcome using any other time frame(e.g. actual 305-day milk yield, projected 305-day milk yield) or measurement (e.g. liters,pounds), the outcome was transformed to kg/cow per day.
The precision of the estimate reported was based on the standard errors (S.E.) or standard deviations (S.D.) of the treatment and control groups. If the paper reported separate estimatesfor each group, they were recorded as such. If the paper reported a common S.E. (or S.D.),that estimate was used for both groups. If the paper only reported a Z statistic or P, anestimate of a common S.E. was computed. For papers that only reported a P less than orequal to a given value (e.g. <0.05), then that given value was taken and the P and S.E.
computed as above. Finally, for studies that simply reported a non-significant effect, Pvalues of 0.15, 0.3 and 0.5 were assigned and compared. The P value that produced thesmallest (most conservative) estimate of the overall treatment effect was selected for thecalculation of the S.E.
Data only were extracted from clinical trials, although the studies need not have been The information described in was extracted. All this information was ex- tracted from the articles independently by two investigators using a structured data-collection Table 1Additional information extracted from the studies considered in the review of anthelmintic treatment andmilk-production response in adult dairy cows Journal indexed in Index Medicus, journal not indexed, abstracts/paperproceedings Whether a method of randomization was reported Whether blinded treatment measurement was reported Confounders controlled for in the analysis (i.e. age, farm, season, previousmilk production, etc.) Whether endectocides (e.g. ivermectin, moxidectin, eprinomectin,dectomectin) were used Dry off: calving, mid-lactation (between 0 and 200 day in milk) or strategic Whether the treatment was not applied to the whole herd at once Period of time (days) milk production was measured Milk production measure (daily weight, 305 actual, 305 projected, etc.) Whether the cows were on pasture-year-round, pasture-seasonal or partiallyconfined J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 form. The two datasets were then compared and all the disagreements were resolved by thesenior author (Sanchez) re-reviewing the source paper.
Fixed and random-effects meta-analyses were carried out to evaluate the effect of an- thelmintic treatment on milk production. The results from the random-effects meta-analysisare presented in this paper. A random-effects meta-analysis assumes that there is a normaldistribution of the study effects and the variance of the distribution is estimated from thedata. The method of DerSimonian and Laird (was used to estimate thevariance for the random-effects model. The heterogeneity statistic Q was used to determine if there was significant variability between studies. Under the nullhypothesis of a common treatment effect among trials, this Q statistic follows a chi-squareddistribution with K − 1 degrees of freedom, where K is the number of trials. Because a sig-nificant P (i.e. <0.05) for the Q statistic was observed, the results from the random-effectsmodel are presented in this manuscript.
Because several biases might influence the results of a meta-analysis, the following procedures were performed to detect (and if needed, to correct for) possible publica-tion bias or other small-study effects. First, the Begg’s andEgger’s tests were used in combination with a funnel plot If there was any evidence of publication bias from either of these tests or thefunnel plot, the “trim and fill” method suggested by was usedto estimate and correct for this publication bias. This method works by omitting studieswith large S.E. or low statistical significance (e.g. “small study”) until the funnel plot issymmetrical. Then, using the trimmed funnel, the center of the plot is estimated and theomitted studies are put back in along with their hypothetical “missing counterparts” aroundthe center ( 2.4.1. Trial quality characteristics To investigate factors which might have influenced trial results, weighted regression anal- yses (meta-regression) between the trial effect and trial quality characteristics (includingprecision of estimate) were performed. This was done in two steps. First, an unconditionalanalyses were carried out between the trial outcome and trial precision, randomization,blinding, control for confounders in the analysis and publication type. Subsequently, all un-conditionally significant variables (P ≤ 0.15) were retained and evaluated in a multivariableanalysis.
Meta-regression analyses also were used to evaluate the effects on the study outcome of: product formulation (endectocides or other drugs), parity of cows (primiparous, multiparousor all combined), time of treatment (dry off, calving, mid-lactation or strategic treatment),trial length (days after treatment for which milk production was measured), individual J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 treatment (versus whole-herd treatment), country where the trial was performed and pastureexposure.
A random-effects cumulative meta-analysis was performed using the 75 trials. This methodology computed an overall estimate of treatment effect at the time each trial waspublished. A cumulative meta-analysis may be used to identify (retrospectively) when atreatment effect reached conventional levels of statistical significance. However, we usedthe method to identify possible temporal patterns in the trial results.
The between-trial variance was obtained using the moment estimator of the variance (no adjustment for clustering of results within author wascarried out, because the number of trials per author was low. All analyses were carried outusing the statistical program Stata, Version 8 3. Results
Of the 78 articles identified by the literature review, 7 of them could not be retrieved (6 English and 1 Italian). Of the remaining 71, 7 articles were not used in the analyses for thefollowing reasons: 4 were review articles with no original data (2 were duplicates (and 1 only evaluated the effects of flukes on milk production The remaining 64 articles described 97 anthelmintic trials. Out of these, 8 articles (9 trials) did not contain data on the trial outcome (From the 9 trials that did not report the outcome ofinterest, 6 reported a non-significant effect of treatment on milk production while the other3 did not report any value. A further 11 articles (13 trials presented in listed inthe references) presented data in a manner that was not usable in the meta-analyses (usuallyno estimate of the precision of the results was available). Out of the 13 trials not usable inthe meta-analysis, 3 did not report the length of the milk production measurement so theoutcome could not be computed, 1 reported a negative effect and 9 reported a positive effectof treatment on milk production (2 were significant, 7 did not report the significance). These13 trials used, on average, 241 cows and ranged from 20 to 1643 animals. Of the 19 articlesidentified above as containing at least some trial results which were not usable, three alsoreported some trial results that were usable in the meta-analysis. Consequently, a total of48 articles with results from 75 trials were used for the meta-analysis (articles presented inlisted in the references). Forty-five of these articles were written in English, twowere in French and one was in Spanish. Summaries of the main trial characteristics of thetrials not usable and used in the meta-analysis are presented in vely.
Out of the 75 trials used in the meta-analysis, 16 reported a negative effect and the other Summary of the 13 trials not usable in the meta-analysis a Control for confounding (e.g. previous lactation, age, season, farm) in the analysis.
b DW = daily weight, 305 = 305-day total milk production (actual or projected), NR = not reported.
c Statistical significance reported: NR = not reported, NS = not significant, S = significant (P ≤ 0.05).
d Reason not being used: (A) no precision or P-value reported, (B) no measure of milk production reported; (C) no sample size reported.
Table 3Summary of the 75 trials used in the meta-analysis Control for confounding (i.e. previous lactation, age, season, farm) in the analysis.
b DW = daily weight, 305 = 305 total milk production (actual or projected).
c Statistical significance reported: NS = not significant (if actual P value not reported a value of 0.15 was assumed).
J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 Table 4Descriptions of the trial effect (mean difference kg/cow per day), sample size and number of trials used in themeta-analysis of anthelmintic treatment of naturally infected lactating dairy cows The sensitivity analysis using P values of 0.15, 0.3 and 0.5 produced overall treatment- effect estimates of 0.46, 0.47 and 0.49 kg/cow per day, respectively. Based on these results,the value of 0.15 was selected as the P value for studies which only reported results as“non-significant”. The DerSimonian and Laird pooled estimate of the mean difference inmilk production was 0.46 kg/cow per day (95% CI 0.36, 0.56). A forest plot presenting theresults from each trial and the combined effect is shown in statistical approachesfor the detection of publication bias or small-study effect showed differing results. Althoughthe Begg’s test reported a non-significant bias (P = 0.73), the Egger’s test reported a highlysignificant value (P < 0.001)—and a visual inspection of the funnel plot (suggestedthat publication bias might have been present. This plot is based on the fact that the precisionof the treatment effect will increase as the sample size (e.g. number of cows) increases.
Treatment effects from small studies will therefore scatter more widely at right end of thegraph. In the absence of bias, the plot will resemble a symmetrical funnel (If publication bias is present, few studies with small (or absent) treatment effects andlarge standard errors will be present, resulting in a gap in the lower right quadrant of thegraph. The random-effects “trim-and-fill” method reduced the combined pooled estimatefrom 0.46 to 0.35 (95% CI 0.25, 0.45). This method also indicated that an additional 12trials (small square boxes in would have been necessary to remove this apparentpublication bias (or other small-study effects).
Only 11 trials reported both a formal randomization procedure and blinding. The pooled estimate based only on these 11 trials (0.33) was similar to that reported by the “trim-and-fill”method suggesting a possible association between trial quality and effect estimate.
ws the results obtained from the meta-regression analyses of the associations between trial effect and trial quality characteristics. Both the unconditional and multivari-able analyses showed an association between trial effect and precision (as would have beenexpected based on the previous assessment of publication bias). Similarly, the trial out-come was associated with publication type and control for confounders. If control for other J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 Unitincreasein milk production (kg/cow/day) Fig. 1. Forest plot of the effects of anthelmintic treatment on milk-production response in lactating dairy cows(kg/cow per day). The overall estimate was derived from the random-effect meta-analysis. Lines with arrows aretruncated; length of the horizontal lines indicate 95% CIs of the effect estimate. The center of square represents thepoint estimate and the area of the square is proportional to the weight assigned to the trial. Dotted line representsthe overall estimate obtained from the random-effect meta-analysis. The < > at the bottom of the dashed lineshows the confidence interval for the overall effect. The solid vertical line marks the value where anthelmintictreatment would have no effect. Trial ID refers to values shown in Table 5Meta-regression analysis of the precision and two measures of methodological quality on the trial effect (n = 75) Abstracts/conference proceedings (n = 26) J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 Fig. 2. Funnel plot of the point estimates of the effect of anthelmintic treatment on milk-production response(kg/cow per day). Square points were hypothetical studies added by the “trim and fill” procedure to correct forpublication bias or small-study effect.
confounders was used in the statistical analysis, the mean difference in milk production wasapproximately 0.25 kg/cow per day lower than in trials that did not control for confoundersin the analysis. Correlations among trial quality characteristics were evaluated to deter-mine if collinearity was a problem in the multivariable analysis. The highest correlationbetween these predictor variables was 0.52 between control for confounders and publica-tion in non-indexed journals—suggesting that collinearity was not affecting the regressioncoefficients.
The results from the meta-regression analyses performed between the trial outcome and variables reflecting other trial characteristics are presented in the variablesevaluated in this analysis did not substantially reduce the variance between studies, four ofthem (time of treatment, individual treatment, trial length and milk measure) were signif-icantly associated with the trial effect. For example, studies that applied the anthelmintictreatment to mid-lactation cows or strategically throughout the year had an average re-sponse that was approximately +0.4 kg/cow per day higher than trials where the cows weretreated either during the dry period or at calving. On the other hand, trials in which individ-uals were assigned to treatment groups (versus whole-herd treatment) had a substantiallylower production response. These two trial characteristics were highly correlated: studiesin which individuals were treated generally applied the treatment at the time of calving butwhole-herd treatment encompassed all stages of lactation.
In relation to geographic location, trials were categorized as northern and southern: Northern trials were those carried out in Canada, Northern Unites States, north-west Europe.
Southern trials were those carried out in southern United States, New Zealand, Australia, Ar-gentina, India and Sri Lanka. Northern trials tended to have higher milk response compared J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 Table 6Unconditional meta-regression analyses based on 75 trials of anthelmintic treatment in lactating dairy cattle.
Table presents coefficients, standard errors, P and the moment estimator of the between-trail variance, τ (n = 75)) 305-Day actual or projected (n = 41) Other (e.g. 100 days total) (n = 11) a Model containing only the intercept.
with southern trials, but this difference was not significant. Pasture exposure was classifiedas pasture-seasonal and pasture-year-round. Only 59 trials reported information on pastureexposure (36 were pasture-seasonal and the remaining 23 were pasture-year-round). Nostatistically significant difference was found between level of pasture exposure and the trialoutcome (b = 0.11, P = 0.40).
The cumulative meta-analysis showed a significant effect after the first trial used in this analysis. However, a pronounced pattern was observed through time (During the1970s the trials had the highest treatment response. This estimate tended to decline during the1980s and start increasing again during the 1990s but without reaching the values observedinitially. Control for confounders (and especially controlling for farm effect) was related topublication year; studies carried out during the 1970s were less likely to control for farmeffect, so larger responses with significant effects were more likely to be reported (data notshown). Moreover, the type of drug used was related to the publication year. Older drugs(e.g. thiabendazole, morantel, levamisol) were more likely to be used during the 1970s,newer benzimidazole drugs during the 1980s and trials using endectocides (e.g. ivermectin)during the 1990s.
J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 Unitincreasein milk production (kg/cow/day) Fig. 3. Cumulative random effect meta-analysis of 75 trials of the effect on milk-production response after an-thelmintic treatment in naturally infected lactating dairy cows. The length of the line represent the 95% confidenceinterval for overall effect based on all preceding studies. The center (circle) on each line marks the point estimateat the time of trial’s publication. The dashed vertical line marks the current overall effect estimate (not adjustedfor publication bias).
4. Discussion
The combined unadjusted and adjusted estimates of 0.46 and 0.35 kg/cow per day, respectively, obtained from the 75 studies were smaller than the 0.63 reported byHowever, this estimate is an average effect of the milk-productionresponse from the 75 trials. One of the factors influencing the magnitude of the treat-ment response will be the level of GIN infection at the farm—which might explain thelarge variation in treatment response observed between trials (range of −2.17 to 3.16; Although not all the studies used in the review by used in the present meta-analysis, a similar median increase in milk production was obtained in thisstudy (which suggests some similarity between these two reviews. Only 57 trialsused in this meta-analysis matched with those evaluated by Gross et al. (n = 87). They hada median increase in milk production of 0.54 kg/cow per day (data not shown; the other 30studies did not have data suitable for the meta-analyses, did not meet the inclusion criteriaor were not retrieved). Using the same 57 studies, the combined estimate derived from the J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 random-effects model after correcting for a possible publication bias was 0.32 (95% CI0.21, 0.43)—similar to that from our full dataset.
The significant heterogeneity found in this analysis was expected because the differences in trial designs, GIN burdens, treatment protocols, drugs, geographic locations and agegroups would have influenced the treatment response.
The visual assessment of publication bias based on the funnel plot as well as the results from the Egger’s test indicated a publication bias in this analysis. Funnel-plotasymmetries also have been related to inclusion of trials of lower quality (e.g. studies whichare not double blind, studies with inadequate allocation of animals to the treatment group)(Lower quality trials have also been reported to overestimate the truetreatment effect ( When variables accounting for trial quality were evaluated, three of them were associ- ated with treatment effect and they showed similar trends to those reported by Two variables not significantly associated with treatment effect were the use of aformal randomization procedure and blinding of treatment measurement. Although studiesreporting blinding tended to have a lower effect, those reporting a formal randomized proce-dure had a higher response—which we had not expected. However, out that the results obtained from the meta-regression analysis should beinterpreted with some caution, especially when the trial characteristics have low variabilityacross studies (which was the case of the present meta-analysis), because the regressionanalysis can be biased by unmeasured confounders. On the other hand, trials publishedin indexed journals (e.g. Index Medicus) or trials that used better statistical methodolo-gies (which might reflect the quality of the published study) were associated with lowerproduction response. These analyses suggested that the overall estimate of 0.35 kg/cowper day would be less biased and so more appropriate to report as the overall combinedestimate.
Considering the meta-regression analyses of trial design characteristics, trials using macrocyclic lactone endectocides (e.g. ivermectin, moxidectin and eprinomectin) had ahigher milk response compared with those using either benzimidazoles or older anthelmintics(i.e. coumaphos, thiabendazole). In contrast, the same median in-crease between new and old anthelmintics. In theory, the new generation of anthelmintic ismore effective (especially against immature stages, including larval stages of O. ostertagi)(so a higher response might beexpected.
Trials that treated animals either in mid-lactation or strategically (several times during the year) had higher response compared with trials treating animals during the dry-offperiod or around calving. found a similar effect (animals treated inmid-lactation had twice as large a production response compared with those treated duringthe dry period or around calving). Production responses from whole-herd treatment schemeswere used in studies in which treatment was given mid-lactation or strategically throughoutthe year. The larger production response in whole-herd treatment trials might be relatedto the elimination of the parasites at one point with a more pronounced decrease in thepasture contamination (and consequently, less re-exposure to parasites) or to more frequenttreatments (strategic anthelmintic treatment). On the other hand, a larger treatment responsemight have been expected from cows treated at calving (shortly before their period of highest J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 milk production and physiologic demands). Clarification of this point will require furtherresearch.
Production response declined by 2 g for each additional day in the trial follow-up pe- riod. Similarly, when the outcome was reported as either 305-actual or 305-projected milkproduction, the response was lower compared with daily weight trials that reported effecton a per-day basis. Daily weight trials tended to measure milk production for shorter time(median = 95 days) compared with 305 days for whole-lactation trials.
Primiparous cow trials had a lower response than multiparous cow trials. This might reflect different susceptibilities to gastrointestinal parasites between these implicit age groups orthe higher production capacity of older cows. higher wormcounts in cows <3 years and those >10 years; that firstlactation animals had higher fecal egg counts compared with cows in other lactations. Onthe other hand, found that first lactation animals had lower opticaldensities from a crude indirect ELISA compared with older animals; this suggests that firstlactation animals might be more susceptible to gastrointestinal nematodes. It also has beenobserved that high-producing animals are more susceptible to GIN—which might suggestsa higher treatment response in the high producing group ( The distinct pattern observed in the cumulative meta-analysis (be related to the combined effect of improvement in study design (including statistical analysis) and/orchanges in efficacy of the anthelmintic used. The decline in the effect from 1972 to 1985might have been due to the use of better trial designs and analytic methods. Althoughcontrolling for a farm effect will have a bigger impact on the precision of the estimate,trials which controlled for a farm effect also tended to control for other variables in theanalysis. Controlling for confounders was associated with lower milk response (The increased response through the 1990s might reflect the greater efficacy of the endecto-cides. However, this effect might be confounded with the improvement in farm managementpractices and increasing productivity of cows over the years.
5. Conclusion
Our meta-analysis showed that, on average, an increase of milk production of ∼0.35 kg/ cow per day might be expected after anthelmintic treatment of naturally infected lactatingdairy cows. However, substantial between-trial variation was observed. While this was, inpart, due to differences in study design, it also emphasizes the need to have a reliable diagnos-tic test to identify cow/herds showing the detrimental effects of GIN parasitism to establisha more rational anthelmintic-control program. There was evidence of publication bias (i.e.
“small-study effect”) in the published literature. Variables such as formulation type, timeof treatment, period after treatment during which milk production was recorded, outcomemeasure recorded and parity were associated with the trial outcome (but only accounted forsmall amounts of the unexplained variance between studies). Although guidelines for an-thelmintic treatment in adult dairy cows could not be stated from the present meta-analysis,an important variation in treatment response was observed—suggesting that if a reliablediagnostic test for gastrointestinal parasitism in adult cows became available, a beneficialtreatment effect could be expected in herds classified as having parasite problems.
J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 References
Agneessens, J., Claerebout, E., Dorny, P., Borgsteede, F.H., Vercruysse, J., 2000. Nematode parasitism in adult dairy cows in Belgium. Vet. Parasitol. 90, 83–92.
Alvarez, E., Greco, J.A., Alberdi, J.G., Romano, A., 1977. Anthelmintic treatment of cattle by the ‘spot’ method, with levamisole. Gaceta Veterinaria 39, 18–22.
Barger, I.A., 1979. Milk production of grazing dairy cattle after a single anthelmintic treatment. Aust. Vet. J. 55, Barger, I.A., Lisle, K.A., 1982. Milk production of grazing dairy cows given monthly anthelmintic treatment.
Beatty, J.F., Arnold, B.L., Custer, E.W., McGee, W.H., Daniels, J.W., 1975. Effect of a coumaphos anthelmintic on production and quality. J. Dairy Sci. 58, 144.
Begg, C.B., Mazumdar, M., 1994. Operating characteristics of a rank correlation test for publication bias.
Bell, R.R., 1980. Effect of thiabendazole treatment on dairy cows in Texas. In: Proceedings of the 25th Annual Meeting of the American Association of Veterinary Parasitology, Washington, USA, p. 28.
Bhongade, H.G., Sarode, D.B., Rode, A.M., Sapre, V.A., 1993. Effect of anthelmintic therapy on haematobiochemical profiles of dairy cows. Indian J. Vet. Med. 13, 75–76.
Biondani, C.A., Steffan, P., 1988. Efecto de los parasitos gastrointestinales sobre la produccion lactea en rodeos lecheros. Vet. Argentina 42, 116–127.
Bisset, S.A., Marshall, E.D., Morrison, L., 1987. Economics of a dry-cow anthelmintic drenching programme for dairy cows in New Zealand. Part 1. Overall response in 47 herds. Vet. Parasitol. 26, 107–118.
Bliss, D.H., Jones, R.M., Conder, D.R., 1982. Epidemiology and control of gastrointestinal parasitism in lactating, grazing adult dairy cows using a morantel sustained release bolus. Vet. Rec. 110, 141–144.
Bliss, D.H., Todd, A.C., 1973. Milk production by Wisconsin dairy cattle after deworming with BAYMIX. Vet.
Med. Small Anim. Clin. 68, 1034–1038.
Bliss, D.H., Todd, A.C., 1974. Milk production by Wisconsin dairy cattle after deworming with thiabendazole.
Vet. Med. Small Anim. Clin. 69, 638–640.
Bliss, D.H., Todd, A.C., 1976. Milk production by Vermont dairy cattle after deworming. Vet. Med. Small Anim.
Block, E., Gadbois, P., 1986. Efficacy of morantel tartrate on milk production of dairy cows: a field study. J. Dairy Block, E., McDonald, W.A., Jackson, B.A., 1987. Efficacy of Levamisol on milk production of dairy cows: a field study. J. Dairy Sci. 70, 1080–1085.
Borgsteede, F.H., Tibben, J., Cornelissen, J.B., Agneessens, J., Gaasenbeek, C.P., 2000. Nematode parasites of adult dairy cattle in the Netherlands. Vet. Parasitol. 89, 287–296.
Brown, M.A., Maniscalco, V.J., 1974. Effects on milk production and internal parasites of dairy cattle from a ration supplemented with a parasiticide. The Southwestern Veterinarian Winter, pp. 51–53.
Carrier, J., DesCoteaux, L., Smith, R.A., 1999. Effect of an eprinomectin treatment at fall housing of milk production of dairy cows in Quebec. In: Proceedings of the 32nd Annual Conference American Associationof Bovine Practitioners, Nashville, TN, USA, September 23–26, 1999, pp. 234–235.
Corba, J., Reis, T., Gruebler, R., 1980. Effect of dehelmintization in dairy cattle with subclinical helminthoses on milk production. Helminthologia 17, 219–224.
Deeks, J.J., Altman, D.G., Bradbrun, M.J., 2001. Statistical methods for examining heterogeneity and combining results from several studies in meta-analysis. In: Systematic Reviews in Health Care: Meta-analysis in Context.
BMJ Books, London, pp. 285–312.
De Rond, J.C., de Jong, R., Boon, J.H., Brouwer, B., 1990. Influence of gastro-intestinal nematodes on the productivity of dairy cattle in the wet highlands of Sri Lanka. Trop. Anim. Health Prod. 22, 135–143.
DesCoteaux, L., Doucet, M., Caldwell, V., 2001. Evaluation of the impact of parasite control with the IVOMEC(R) SR Bolus given at breeding age on first lactation yield in Holstein heifers. Vet. Parasitol. 98, 309–314.
Duval, S., Tweedie, R., 2000. A nonparametric “Trim and Fill” method of accounting for publication bias in meta-analysis. J. Am. Stat. Assoc. 95, 89–98.
Eddi, C., Muniz, R.A., Caracostantogolo, J., Errecalde, J.O., Rew, R.S., Michener, S.L., McKenzie, M.E., 1997. Comparative persistent efficacy of doramectin, ivermectin and fenbendazole against natural nematodeinfections in cattle. Vet. Parasitol. 72, 33–41.
J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 Egger, M., Davey Smith, G., 2001. Principles of and procedures for systematic reviews. In: BMJ Books (Ed.), Systematic Reviews in Health Care: Meta-analysis in Context, London, pp. 23–43.
Egger, M., Smith, G.D., Schneider, M., Minder, C., 1997. Bias in meta-analysis detected by a simple, graphical Eysker, M., 1997. Some aspects of inhibited development of trichostrongylids in ruminants. Vet. Parasitol. 72, Eysker, M., Eilers, C., 1995. Persistence of the effect of a moxidectin pour-on against naturally acquired cattle nematodes. Vet. Rec. 137, 457–460.
Eysker, M., Ploeger, H.W., 2000. Value of present diagnostic methods for gastrointestinal nematode infections in ruminants. Parasitology 120 (Suppl.), S109–S119.
Fetrow, J., Johnstone, C., Bartholomew, R., 1985. Production responses of lactating dairy cows and heifers given thiabendazole at parturition. Am. J. Vet. Res. 46, 48–52.
Fisher, L.J., 1980. The economic response of lactating cows to treatment for parasites. Can. J. Anim. Sci. 60, Fisher, L.J., MacNeill, A.C., 1982. The response of lactating cows and growing heifers to treatment for parasites.
Fox, M.T., 1979. Deworming of cows—What are the benefits? Vet. Rec. 105, 315–316.
Fox, M.T., Jacobs, D.E., 1984. Observations on the effect of levamisole treatment on the production of dairy cows in England. Anim. Prod. 38, 15–22.
Frechette, J.L., Lamothe, P., 1981. Milk production effect of a morantel tartrate treatment at calving in dairy cows with subclinical parasitism. Can. Vet. J. 22, 252–254.
Froyd, G., 1973. Strategic treatment against subclinical parasitism. Helminthologia 14, 377–384.
Gasbarre, L.C., Stout, W.L., 1998. An ounce of prevention equals pounds of milk. Agric. Res. Washington 46, Gibbs, H.C., 1980. Effect of thiabendazole treatment of lactating dairy cows in Maine. In: Proceedings of the 25th Annual Meeting of the American Association of Veterinary Parasitology, Washington, USA, p. 27.
Glenn, J.S., Baker, N.F., Franti, C.E., Ver Steeg, J.D., 1982. Anthelmintic treatment of nonpastured dairy cows in California. J. Dairy Sci. 65, 2006–2010.
Gouffe, D., Raoult, J., Vallet, A., 1984. Strongyloses gastro-intestinales et production laitiere. Revue Med. Vet.
Gremillet, J.P., 1981. The use of thiabendazole to augment milk production. Bulletin Mensuel De La Societe Veterinaire Pratique De France 65, 603–610.
Gross, S.J., Ryan, W.G., Ploeger, H.W., 1999. Anthelmintic treatment of dairy cows and its effect on milk production. Vet. Rec. 144, 581–587.
Harris, B., Wilcox, C.J., 1976. Effect of anthelmintics on milk production. J. Dairy Sci. 59, 20–21.
Heider, L.E., Courtney, C.H., Herd, R. 1980. The effect of thiabendazole treatment on dairy cows in Ohio. In: Proceedings of the 25th Annual Meeting of the American Association of Veterinary Parasitology, Washington,USA, p. 28.
Hoste, H., Le Frileux, Y., Pommaret, A., 2002. Comparison of selective and systematic treatments to control nematode infection of the digestive tract in dairy goats. Vet. Parasitol. 106, 345–355.
Kloosterman, A., Albers, G.A.A., 1982. The effect of anthelmintic treatment or artificial trichostrongylid infection on milk production of cows. Parasitology 84, xxxiii–xxxxiv.
Kloosterman, A., Ploeger, H.W., Pieke, E.J., Lam, T.J.G.M., Verhoeff, J., 1996. The value of bulk milk ELISA Ostertagia antibody titres as indicators of milk production response to anthelmintic treatment in the dry period.
Vet. Parasitol. 64, 197–205.
Mathews, G.L., Gray, R.M., McGowan, A.A., 1983. Effect of anthelmintic treatment immediately after calving on milk production. Aust. Vet. J. 60, 116–119.
McBeath, D.G., Dean, S.P., Preston, N.K., 1979. The effect of a preparturient febendazole treatment on lactation yield in dairy cows. Vet. Rec. 105, 507–509.
McPherson, W.B., Gogolewski, R.P., Slacek, B., Familton, A.S., Gross, S.J., Maciel, A.E., Ryan, W.G., 2001.
Effect of a peri-parturient eprinomectin treatment of dairy cows on milk production. New Zeal. Vet. J. 49,106–110.
McPherson, W.B., Slacek, B., Familton, A., Gogolewski, R.P., Cramer, R.P., Gross, S.J., 1999. The benefit of topical eprinomectin on milk production in dairy cattle. In: Proceedings of the 17th International Conferenceof the World Association for the Advancement of Veterinary Parasitologists, Copehagen, Denmark.
J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 McQueen, I.P.M., Cottier, K., Hewitt, S.R., Wright, D.F., 1977. Effects of anthelmintics on dairy cow yields. New Michel, J.F., Richards, M., Altman, J.F.B., Mulholland, J.R., Gould, C.M., Armour, J., 1982. Effect of anthelmintic treatment on the milk yield of dairy cows in England, Scotland and Wales. Vet. Rec. 111, 546–550.
Miller, J.E., Baker, N.F., Farver, T.B., 1986. Anthelmintic treatment of pastured dairy cattle in California. Am. J.
Moher, D., Pham, B., Jones, A., Cook, D.J., Jadad, A.R., Moher, M., Tugwell, P., Klassen, T.P., 1998. Does quality of reports of randomised trials affect estimates of intervention efficacy reported in meta-analyses? Lancet 352,609–613.
Moore, E.D., Owen, J.R., Dowlen, H.H., Haws, L.D., Richardson, D.O., Smith, J.W., Bennett, S.E., 1974.
Comparison of anthelmintic treatments for replacement dairy heifers and dairy cows. Southern Div Memphis,Tennessee, 4–7 February, 1974. J. Dairy Sci. 57, 145.
Morhain, A., Legrand, J., 1981. Field tests with thiabendazole as a galactogen. Bulletin Mensuel De La Societe Veterinaire Pratique De France 65, 615–620.
Murphy, A., 1998. The effect of treatment with moxidectin, a long-acting endectocide, on milk production in lactating dairy cows. In: Proceedings of the XXth World Association for Buiattrics Congress, pp. 35–45.
Nødtvedt, A., Dohoo, I., Sanchez, J., Conboy, G., DesCoteaux, L., Keefe, G.P., Leslie, K., Campbell, J., 2002a.
The use of negative binomial modelling in a longitudinal study of gastrointestinal parasite burdens in Canadiandairy cows. Can. J. Vet. Res. 66, 249–257.
Nødtvedt, A., Dohoo, I., Sanchez, J., Conboy, G., DesCoteaux, L., Keefe, G.P., 2002b. Increase in milk yield following eprinomectin treatment at calving in pastured dairy cattle. Vet. Parasitol. 105, 191–206.
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.
Pfister, K., Reist, M., Meditna, D., 2001. The effect of eprinomectin or trichlorfon on milk yield of dairy cattle. In: Proceedings of the 18th International Conference of the World Association for the Advancement of VeterinaryParasitologists, Stresa, Italy.
Ploeger, H.W., Kloosterman, A., Bargeman, G., Wuijckhuise, L.V., Brink, R.R., Van den, B.R., 1990a. Milk yield increase after anthelmintic treatment of dairy cattle related to some parameters estimating helminth infection.
Vet. Parasitol. 35, 103–116.
Ploeger, H.W., Kloosterman, A., Borgsteede, F.H.M., 1990b. Effect of anthelmintic treatment of second-year cattle on growth performance during winter housing and first lactation yield. Vet. Parasitol. 36, 311–323.
Ploeger, H.W., Schoenmaker, G.J.W., Kloosterman, A., Borgsteede, F.H.M., 1989. Effect of anthelmintic treatment of dairy cattle on milk production related to some parameters estimating nematode infection. Vet. Parasitol.
34, 239–253.
Pluimers, E.J., 1979. Milk production increase following treatment of Dutch dairy cattle with thiabendazole. Vet.
Pouplard, L., 1978. Anthelmintic treatment of dairy cows. Vet. Rec. 103, 434.
Prosl, H., Supperer, R., Jones, R.M., Lockwood, P.W., Bliss, D.H., 1983. Morantel sustained release bolus: a new approach for the control of trichostrongylosis in Austrian cattle. Vet. Parasitol. 12, 239–250.
Rosenthal, R., DiMatteo, M.R., 2001. Meta-analysis: recent developments in quantitative methods for literature reviews. Annu. Rev. Psychol. 52, 59–82.
Sanchez, J., Dohoo, I., Nødtvedt, A., Markham, F., Leslie, K., DesCoteaux, L., Campbell, J., Keefe, G.P., 2002.
A longitudinal study of gastrointestinal parasites in Canadian dairy farms. The value of an indirect Ostertagiaostertagi ELISA as monitoring tool. Vet. Parasitol. 107, 209–226.
Sanyal, P.K., Singh, D.K., Knox, M.R., 1992. The effect of peri-parturient anthelmintic treatment on the productivity of dairy cattle in subtropical western India. Vet. Res. Commun. 16, 445–451.
Sanyal, P.K., Srivastava, S.M., Panchal, A.G., Singh, D.K., 1995. Effect of feeding fenbendazole incorporated urea molasses block on nematode infection and milk yield of dairy cattle. J. Vet. Parasitol. 9, 11–16.
Spence, S.A., Fraser, G.C., Chang, S., 1996. Responses in milk production to control of gastrointestinal nematode and paramphistome parasites in dairy cattle. Aust. Vet. J. 74, 456–459.
Sommefeldt, J.L., Clark, A.K., Hugghins, E.J., Shultz, J.B., Illg, D.J., 1986. Responses to controlling gastrointestinal parasitism in dairy cattle 69, 2723–2727.
Spence, S.A., Fraser, G.C., Dettmann, E.B., Battese, D.F., 1992. Production responses to internal parasite control in dairy cattle. Aust. Vet. J. 69, 217–220.
J. Sanchez et al. / Preventive Veterinary Medicine 63 (2004) 237–256 Stata Corporation, 2001. Stata Statistical Software, College Station, TX.
Sterne, J.A., Egger, M., Davey Smith, G., 2001. Investigating and dealing with publication and other biases. In: Systematic Reviews in Health Care: Meta-analysis in Context. BMJ Books, London, pp. 189–208.
Takagi, H., Block, E., 1986. Effects of feeding coumaphos to dairy cows at various stages of lactation on subclinical parasite infection and milk production. Can. J. Anim. Sci. 66, 141–150.
Tharaldsen, J., Helle, O., 1989. The effect of periparturient treatment with fenbendazole on the milk production of cows. Acta Vet. Scand. 30, 247–252.
Thomas, G.W., Barton, N.J., McGowan, A.A., Mickan, F.J., Miller, G., James, R.E., 1984. Effect of whole herd anthelmintic treatment on milk production of dairy cows. Res. Vet. Sci. 36, 266–269.
Thomas, R.J., Rowlinson, P., Nansen, P., Jorgensen, R.J., Soulsby, E.J.L., 1981. An evaluation of the anthelmintic treatment in a dairy herd. In: Nansen, P., Jorgensen, R.J., Soulsby, E.J.L. (Eds.), In epidemiology and controlof nematodiasis in cattle, EEC, Brussels ECSC, pp. 101–115.
Thompson, S.G., Higgins, J.P., 2002. How should meta-regression analyses be undertaken and interpreted? Stat.
Thompson, S.G., Sharp, S.J., 1999. Explaining heterogeneity in meta-analysis: a comparison of methods. Stat.
Todd, A.C., Bliss, D., Grisi, L., Crowley, J.W., 1978. Milk production by dairy cattle in Pennsylvania and North Carolina after deworiming (treatment at freshening and systematically over the first three months of lactation).
Vet. Med. Small Anim. Clin. 73, 614–619.
Todd, A.C., Bliss, D.H., Meyers, G.H., 1975. Milk production increases following treatment of subclinical parasitisms in Wisconsin dairy cattle. New Zeal. Vet. J. 23, 59–62.
Todd, A.C., Myers, G.H., Bliss, D., Cox, D.D., 1972. Milk production in Wisconsin dairy cattle after anthelmintic treatment. Vet. Med. Small Anim. Clin. 67, 1233–1236.
van Adrichem, P.W.M., Shaw, J.C., 1977. Effects of gastrointestinal nematodiasis on the productivity of monozygous twin cattle. 2. Growth performance and milk production. J. Anim. Sci. 46, 423–429.
Vercruysse, J., Claerebout, E., 2001. Treatment vs. non-treatment of helminth infections in cattle: defining the threshold. Vet. Parasitol. 98, 195–214.
Walsh, A., Younis, P.J., Morton, J.M., 1995. The effect of ivermectin treatment of late pregnant dairy cows in south-west Victoria on subsequent milk production and reproductive performance. Aust. Vet. J. 72, 201–207.
Wilk, J.C., Porterfield, I.D., McDaniel, B.T., 1980. Effects of the use of thiabendazole at calving in Holsteins and Wilkinson, A.C.R., 1992. Husk in heifers. Vet. Rec. 131, 567–568.
Yazwinski, T.A., Tucker, C., Copeland, S., Yazwinski, T., Guerino, F., 1999. Dose confirmation of moxidectin pour-on against natural nematode infections in lactating dairy cows. Vet. Parasitol. 86, 223–228.

Source: http://www.albanyrecreation.ac.nz/massey/fms/Colleges/College%20of%20Sciences/Epicenter/docs/ASVCS/Sanchez_et_al_2004.pdf?EA4D318619EC8AFC36FF3C7DA12BEE77