A Meta-analysis Comparing The Effect Of Vaccines Against Mycoplasma Hyopneumoniae On Daily Weight Gain In Pigs

Preventive Veterinary Medicine 54 (2002) 265±278

A meta-analysis comparing the effect of vaccines against Mycoplasma hyopneumoniae on daily weight gain in pigs C.S. Jensena,b,*, A.K. Ersbùllc, J.P. Nielsena a

Department of Clinical Studies, The Royal Veterinary and Agricultural University, Dyrlñgevej 88, DK-1870 Frederiksberg C, Denmark b Research Centre for the Management of Animal Production and Health (CEPROS), Foulum, DK-8830 Tjele, Denmark c Department of Animal Science and Animal Health, The Royal Veterinary and Agricultural University, GrùnnegaÊrdsvej 8, DK-1870 Frederiksberg C, Denmark Received 23 February 2001; accepted 28 January 2002

Abstract Our aims were to evaluate the published literature concerning the effect of swine vaccination against Mycoplasma hyopneumoniae on the average daily weight gain (ADWG). This was done by re-evaluating the in¯uence of selected factors on ADWG by a meta-analysis of published studies from 1991 to 1999, ful®lling certain inclusion criteria. With ADWG as the outcome, an analysis of variance was performed for such variables as treatment, vaccination schedule, age during study, housing system and publication quality. Each clinical trial was considered as a random effect and the numbers of pigs in each trial were weightings. Of 63 published studies, 16 describing three commercial vaccines ful®lled the criteria for the metaanalysis. Due to few studies with one of the vaccines …n ˆ 3†, only two vaccines were included. Vaccinated pigs gained an average of 592 g …S:E: ˆ 15† with Stellamune1 and 590 g …S:E: ˆ 15† with Suvaxyne1 compared to non-vaccinated pigs that gained an average of 569 g …S:E: ˆ 14† …P < 0:01† when adjusted for age during the study. Vaccine type, vaccination schedule, housing system and publication quality were not signi®cantly associated with ADWG. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Meta-analysis; Mycoplasma hyopneumoniae; Vaccine; Swine; Respiratory disease

* Corresponding author. Present address: Department of Clinical Studies, The Royal Veterinary and Agricultural University, Dyrlñgevej 88, DK-1870 Frederiksberg C, Denmark. Tel.: ‡45-35-28-28-43; fax: ‡45-35-28-28-38. E-mail address: [email protected] (C.S. Jensen).

0167-5877/02/$ ± see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 5 8 7 7 ( 0 2 ) 0 0 0 0 5 - 3

266

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

1. Introduction Swine enzootic pneumonia causes major economic losses in the pig industry worldwide (Ross, 1999). Mycoplasma hyopneumoniae is the primary agent responsible, but secondary infections from Actinobacillus pleuropneumoniae, Pasteurella multocida and Streptococcus suis also can increase the severity of the disease (Christensen et al., 1999). In Denmark, 6.9 million doses of vaccines against M. hyopneumoniae were sold in 1998, for an annual production of approximately 2 million pigs (Anon., 1998). Three different commercial vaccines against M. hyopneumoniae were available in Denmark in the study period: Stellamune11 (RespiSure1 outside Europe), Suvaxyn12 and Hyoresp1.3 The vaccines were used to reduce the severity of pneumonia, and thereby increase productivity by improving the average daily weight gain (ADWG) (Dayalu et al., 1992; Dayalu, 1994; Nash, 1996). Both the Stellamune1 and Suvaxyn1 vaccines contain inactivated whole cell cultures of M. hyopneumoniae combined with an oil adjuvant. Many studies have investigated the effect of vaccination on productivity (e.g. VraaAndersen and Christensen, 1993; Charlier et al., 1994; Lium et al., 1994; Scheidt et al., 1994; Trayer, 1994). However, several factors differed between studies, including the age of pigs at vaccination, ef®cacy measure (e.g. ADWG and lung lesion), study periods, housing system and group size. These variations between study designs make it dif®cult to draw generalizable conclusions regarding the ef®cacy of any particular vaccine. Metaanalysis is a method that takes this variability into account and increases the sample size by combining results from several studies, thereby increasing the power to detect differences in outcomes (Dickersin and Berlin, 1992). Our main aim was to evaluate the effect of commercial vaccines against M. hyopneumoniae on the ADWG of growing pigs using a meta-analysis of published studies. Further, the second aim was to examine the in¯uences of vaccine schedule, age of pigs, housing system and publication quality. 2. Materials and methods 2.1. Literature review We restricted the review to studies investigating the effect of vaccines against M. hyopneumoniae and published from 1991 to 1999. The studies were written in either English or Danish, and identi®ed using computerized literature databases (BA/RRM, AGRIS, CAB, CC, AGRICOLA, MEDLINE). All studies were reviewed to obtain information on vaccine type used, sample size and assignment of pigs to vaccinated and non-vaccinated groups, route of vaccination, vaccination schedule, housing systems, and age during the study, ef®cacy measure and variance of ef®cacy measure. 1

Stellamune1 (RespiSure1 outside Europe) produced by Pfizer. Suvaxyn1produced by Fort Dodge. 3 Hyoresp1 produced by Merial. 2

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

267

Table 1 Variables provided from the individual studies evaluating the effects of vaccination against M. hyopneumoniae, and included in the meta-analysis Variable

Indicate

Levels

ADWG Trial Treatment

ADWG A number given each trial Which treatment the pigs were given

Trial sample size

Number of pigs included in a treatment group within a trial Whether or not the pigs were vaccinated The period during which the daily weight gain was measured

Continuous 1±28 Non-vaccinated Vaccinated with Stellamune1 Vaccinated with Suvaxyn1 Continuous

Vaccination status Age during study Housing system Vaccine schedule

Housing system during the study period How old the pigs were at vaccination

Publication quality Whether the study was published in an refereed publication or not

Yes No From birth to slaughter From weaning to slaughter From around 30 kg to slaughter At the original farm Moved to another farm Pigs vaccinated twice, first time before weaning Pigs vaccinated twice, first time at Weaning Pigs vaccinated at other times Yes No

We used `study' to de®ne the published study (one publication), and a `trial' to de®ne the ®eld trials conducted within a study. Some studies included more than one trial and some studies combined results from several herds into one estimate. 2.2. Inclusion criteria for studies Only studies using one of three commercial intramuscular vaccines (e.g. Stellamune14 (RespiSure1 outside Europe), Suvaxyn15 and Hyoresp16) were included in the metaanalysis. From this, only studies where vaccines were randomly assigned to pigs within herds naturally infected with M. hyopneumoniae, and with vaccinated and non-vaccinated pigs housed in the same facilities were included for further analysis. Further information necessary for inclusion were the sample size of vaccinated and non-vaccinated pigs, and both the age of pigs at vaccination and whether the pigs were moved or stayed on the original farm. Finally, the studies also had to provide ADWG, and describe the period during which the ADWG was measured. 2.3. Statistical analysis The vaccination schedule was classi®ed into three groups: (1) pigs vaccinated twice, with the ®rst vaccination before weaning; (2) pigs vaccinated twice, with the ®rst 4

Stellamune1 (RespiSure1 outside Europe) produced by Pfizer. Suvaxyn1produced by Fort Dodge. 6 Hyoresp1produced by Merial. 5

268

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

vaccination at weaning; (3) pigs vaccinated at other times. Variables used in the analysis are listed in Table 1. Heterogeneity of the effect of vaccination among the ®eld trials on the ADWG was tested by performing an analysis of variance including trial, vaccine status and a 2-way interaction term between trial and vaccine status (Dickersin and Berlin, 1992). Table 2 Vaccination studies …n ˆ 39† not meeting the inclusion criteria for the meta-analysis of the effect of vaccination against M. hyopneumoniae on the ADWG Reference authors

Reason for exclusion based on first missing information according to the list of inclusion criteria

Bili'c et al. (1992) Bilic et al. (1996) Blagovi'c et al. (1992) Busse and BoÈhne (1996) Charlier et al. (1994) Chiu et al. (1992) Christensen and Vraa-Andersen (1993) De Jong et al. (1996) Diekman et al. (1994) Dohoo and Montgomery (1996) Guadagnini et al. (1998) Harrison and Bryant (1999) Jayappa et al. (1998) Karge et al. (1998) Kobisch et al. (1994) Maes et al. (1999) Martinon et al. (1998a) Martinon et al. (1998b) Martinon et al. (1998c) Miller et al. (1992) Moore et al. (1997) Morrow et al. (1994) Murphy et al. (1993) MunÄoz et al. (1996a) MunÄoz et al. (1996b) Okada et al. (1998) Petersen et al. (1991) Petersen and Weiss (1992) Pihl and Bñkbo (1996) Reynaud et al. (1992) Sheldrake et al. (1991) Sheldrake et al. (1993) Stripkovits and Miller (1993) Thacker et al. (1998a) Thacker et al. (1998b) Thacker et al. (1998c) Thacker et al. (1999) Weiss and Peterson (1992) Weng et al. (1992)

Vaccination of sows No randomization at pig level No randomization at pig level No randomization at pig level No randomization at pig level No commercial vaccine used Daily weight gain not reported No randomization at pig level No information concerning vaccine type used No randomization at pig level No randomization at pig level Combined vaccines Combined vaccines No randomization at pig level Vaccination of sows No randomization at pig level Daily weight gain not reported Daily weight gain not reported in figures Daily weight gain not reported in figures No commercial vaccine used Daily weight gain not reported Daily weight gain not reported in figures Aerosol vaccine administration Missing information of group size Pigs not housed in same facilities No commercial vaccine used Daily weight gain not reported Missing information on group size No randomization at pig level No commercial vaccine used Intraperitoneal vaccine administration Intraperitoneal vaccine administration Pigs not naturally infected Vaccination of sows Incomplete information on vaccine type used Incomplete information on vaccine type used Combined vaccines Missing information on group size Oral vaccine administration

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

269

A signi®cant interaction indicates heterogeneity. A funnel graph was made to check for publication bias (Everitt, 1995). With the ADWG as the outcome, an analysis of variance was performed with treatment, age during study, vaccination schedule, housing system and publication quality as ®xed effects. The variable describing `trial' was expressed as a random effect. The effect of vaccination (both vaccines) on ADWG was evaluated using an orthogonal contrast for comparison of vaccinated to non-vaccinated pigs. All variables were included in the initial model. Backward elimination of the ®xed effects was performed to establish the ®nal model. Table 3 Vaccination studies …n ˆ 13† included in the meta-analysis on the effect of vaccines against M. hyopneumoniae on the ADWG Reference authors

Vaccine type used

No. of pigsa

ADWG ADWG vaccinatedb non-vaccinatedb

Age during studyc

Housing during triald

Dalloli et al. (1998) Diekman et al. (1999)

Suvaxyn1 Stellamune1 Stellamune1 Suvaxyn1 Stellamune1

634 108 108 100 911

712 760 700 707 815

690 780 700 675 799

3 3 3 3 3

1 1 1 2 1

4 1 1 1 1

‡ ‡ ‡ ‡ ‡

7 4 4 3 4

Suvaxyn1 Stellamune1 Stellamune1 Stellamune1 Stellamune1 Stellamune1 Stellamune1 Stellamune1 Stellamune1 Stellamune1 Suvaxyn1 Suvaxyn1

120 211 468 240 306 327 210 100 100 988 132 200

709 716 608 662 418 422 412 826 826 811 680 554

719 719 583 599 390 404 387 766 766 790 653 544

3 3 3 2 1 1 1 3 3 3 3 2

1 1 1 1 1 1 1 1 1 2 1 1

1 1 1 1 1 4 1 6 1 4 1 1

‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡

4 4 4 3 3 7 3 8 3 6 3 3

Suvaxyn1 Suvaxyn1 Suvaxyn1 Suvaxyn1 Suvaxyn1 Suvaxyn1 Suvaxyn1 Suvaxyn1 Suvaxyn1 Suvaxyn1 Suvaxyn1

200 200 200 200 200 200 200 200 200 252 302

568 628 638 623 630 588 605 618 635 638 660

544 621 621 611 611 585 585 594 594 607 607

2 2 2 2 2 2 2 2 2 3 3

1 1 1 1 1 1 1 1 1 1 1

4 1 4 1 4 1 4 1 4 10 10

‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡

6 3 6 3 6 3 6 3 6 12 12

Hogg (1991) Le Grand and Kobisch (1996) Lium et al. (1994) Maes et al. (1998c)e Martinod (1996) Pejsak et al. (1992) Scheidt et al. (1994)f Trayer (1994) Uhlenhopp (1991) Vraa-Andersen and Christensen (1993)g

Wallgren et al. (1998) a

Total number of vaccinated and non-vaccinated pigs used in the trials. ADWG for pigs given by the reference authors. c From birth to slaughter (1), from weaning to slaughter (2), during finishing (3). d On the original farm (1), moved to another farm (2). e Identical to Maes et al. (1998b). f Identical to Scheidt et al. (1993). g Identical to Vraa-Andersen (1994a) and Vraa-Andersen et al. (1994b). b

Vaccine schedule (weeks after farrowing)

270

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

The unit of concern was the group of pigs assigned to a treatment within a trial. Trials conducted at the same herd were treated as independent. All analyses were performed as weighted analyses of variance using trial sample size as weights (Armitage and Berry, 1987). The assumptions for performing the analyses of ADWG were evaluated using residual plots and test for normality. All analyses were performed using the procedure PROC MIXED from the statistical package SAS1 (Statistical Analysis System, SAS Institute, 1996). A signi®cance level of 5% was used. 3. Results Sixty-three relevant publications evaluating the effect of swine respiratory vaccines were identi®ed. Of these, three were review papers (Dayalu et al., 1992; Dayalu, 1994; Nash, 1996) and six were of trials for which the results also were published elsewhere (VraaAndersen, 1994a; Vraa-Andersen et al., 1994b; Scheidt et al., 1993; Bili'c et al., 1994; Maes et al., 1998a,b). Of the remaining 54 publications, 39 studies were not used in the Table 4 Herd specific information …n ˆ 13† from studies included in the meta-analysis on the effect of vaccines against M. hyopneumoniae (MH) on the ADWG Reference

No. of herds

Dalloli et al. (1998) 1 Diekman et al. (1999) 1 1 Hogg (1991) 1 Le Grand and 1 Kobisch (1996) Lium et al. (1994) 4 Maes et al. (1998c) Martinod (1996) Pejsak et al. (1992) Scheidt et al. (1994) Trayer (1994) Uhlenhopp (1991) Vraa-Andersen and Christensen (1993)

5 3 2 1 1 1 1

1 1 1 1 Wallgren et al. (1998) 2 a

Production system

a

b

Herd size Pigs c

Sows

± All-in all-out Continuous ± ±

AP , PM PRRS, APa, PMb PRRS, APa, PMb ± PMb

± ± ± ± ±

Conventional herds, continuous Continuous ± ± ± All-in all-out in finishing ± SPFd herd infected with MH

±

±

30±75

Influenza, PRRS APa

850±3120 ± ± ± 2000 ± 3000

± ± 650±920 ± ± 200 ±

2500 6000 5000 120000 ±

± ± ± ± 85±95

SPF herd infected with MH SPF herd infected with MH Conventional herd Conventional herd ±

A. pleuropneumoniae. P. multocidae. c The symbol `±' indicates that the information is not available. d Specific pathogen free. e Atrophic rhinitis. b

Other respiratory infections

PMb PMb APa APa APa APa, PMb, ARe APa, ARe APa, ARe

900 ± ± 560 ±

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

271

Table 5 Description of the methods, and findings on lung lesions reported for field trials included in the meta-analysis Reference

Dalloli et al. (1998) Diekman et al. (1999) Hogg (1991) Le Grand and Kobisch (1996) Lium et al. (1994) Maes et al. (1998a,b) Martinod (1996) Pejsak et al. (1992) Scheidt et al. (1994) Trayer (1994) Uhlenhopp (1991) Vraa-Andersen and Christensen (1993) Wallgren et al. (1998) a b

Detection methods of lung lesions

Findings regarding lung lesions Non-vaccinateda

Vaccinateda

Each lobe separately Lung disease Mean lung score Mean lung score Lesion score Prevalence of pneumonia Mean area affected Mean area affected Mean area affected Lung score Lung score Lung score Lung score Lesion score Lesion score Lung damage Mean lesion score Reported in another paperb

See reference 0.70% 1.30% 10.90% 1.79 50.00% 3.27% 3.27% 17.08% 15.90% 7.90 9.70 12.70 12.00% 12.00% 10.41% 23.68 ±

See reference 0.70% 1.70% 5.10% 0.78 33.00% 0.85% 2.12% 5.97% 4.70% 4.40 5.80 6.30 6.00% 4.00% 5.71% 16.06 ±

Percentage of lung affected Percentage of lung affected

24.60% 24.60%

4.60% 8.30%

Findings of lung lesions in non-vaccinated and vaccinated pigs. Christensen and Vraa-Andersen (1993).

analysis for various reasons (Table 2). Thus, 14 studies (equaling 28 trials) met the inclusion criteria (Table 3); 12 trials tested Stellamune1, 16 Suvaxyn1 and ®ve Hyoresp1. Hyoresp1 was excluded from the analysis due to an insuf®cient number of trials (Kyriakis et al., 1999 and Reynaud et al., 1998). Variance of ADWG was reported only in three studies and therefore could not be used in our analysis (Lium et al., 1994; Wallgren et al., 1998; Maes et al., 1999). Information on the number of herds used in ®eld trials, production systems, herd size and information concerning respiratory diseases other than M. hyopneumoniae is given in Table 4. Due to missing information, these variables were not included as covariates in the same analysis as the evaluation of lung lesions (Table 5). Table 6 Descriptive results of the effect of Stellamune1 and Suvaxyn1 vaccines on the ADWG in the meta-analysis (n ˆ 28 trials) Vaccine

Vaccinations status of pigs

Sample size

Number of trials

Mean ADWG (g)

S.D.

Stellamune1

Vaccinated Non-vaccinated Vaccinated Non-vaccinated

2131 1946 1818 1871

12 12 16 16

665 640 637 616

164 164 48 48

Suvaxyn1

272

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

Fig. 1. Paired results on the ADWG for non-vaccinated and vaccinated pigs from the trials included in the metaanalysis (n ˆ 28 trials).

Larger group sizes were used in ®eld trials with Stellamune1 compared to Suvaxyn1. There was also a tendency towards a higher crude ADWG in both groups in the Stellamune1 ®eld trials compared to Suvaxyn1 ®eld trials (Table 6). The test of heterogeneity for ADWG among ®eld trials by vaccine status was not signi®cant …P ˆ 0:90†, indicating homogeneity among ®eld trials, and thus the results could be pooled (Fleiss, 1986). A funnel graph further demonstrated no publication bias (Fig. 2), and the variable `publication quality' was also non-signi®cant in the analysis. A similar linear relationship between ADWG for vaccinated pigs compared to nonvaccinated pigs was seen for the two vaccines (P ˆ 0:86 for identical lines) (Fig. 1). The ®nal model for ADWG included treatment …P < 0:001† and age during the study …P < 0:001† (Table 7). The remaining variables (Table 1) were non-signi®cant. Table 7 Results from the final model from the analysis of variance of ADWG in the meta-analysisa Variable

Level

Estimateb Degree of freedom F-test (g) Numerator Denumerator

P

Treatment

Vaccine 1 Vaccine 2 Control From birth to slaughter From weaning to slaughter From around 30 kg to slaughter

592b 590b 569c 409b 624c 718d

Age during study

a b

2

26

25.4

<0.001

2

26

31.5

<0.001

Levels for the same variable with the same letters are not significant different on a 5% significance level. Estimate given as least square means.

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

273

4. Discussion Many studies have shown that the vaccination of pigs against M. hyopneumoniae can increase the ADWG in infected herds (e.g. Vraa-Andersen and Christensen, 1993; Charlier et al., 1994; Lium et al., 1994; Scheidt et al., 1994; Wallgren et al., 1998; Maes et al., 1999). Unfortunately, many publications fail to report whether the effect was statistically signi®cant (e.g. Pejsak et al., 1992; Le Grand and Kobisch, 1996; Dalloli et al., 1998; Hogg, 1991). Meta-analysis increases the power of individual and relatively small studies by combining results from independent studies. The increased power results in a higher precision of the estimatesÐthereby decreasing the variance and more correctly pointing out signi®cant results (Dickersin and Berlin, 1992). A meta-analysis often uses the inverse variance of the outcome (effect measure) as a weight (Fleiss, 1986). The inverse variance of the measurements in each trial is used so that more precise trials (those with small variance) will be given a higher weight in the metaanalysis (Armitage and Berry, 1987). However, reported ADWG variations were only available in a few studies. As an alternative, the analysis was weighted by trial sample size. The rationale for this is that ®eld trials with larger sample sizes tend to have a lower variance, compared to ®eld trials with smaller sample sizes (Rosenthal, 1991). To reduce bias, only randomized studies were included in the meta-analysis. Randomization in the reported vaccine studies was mostly at the pig level (e.g. Hogg, 1991; Dalloli et al., 1998; Maes et al., 1998c; Diekman et al., 1999), one study at litter level (Maes et al., 1999) and one at pen level (De Jong et al., 1996). In two studies, the assignment of pigs was based on `weeks' (e.g. pigs born in week 1 were vaccinated, pigs born in week 2 were non-vaccinated, etc.) (Dohoo and Montgomery, 1996; Pihl and Bñkbo, 1996). Some studies did not state how the pigs were assigned, except that vaccinated and non-vaccinated pigs were housed in the same facilities (e.g. Blagovi'c et al., 1992; Charlier et al., 1994; Bilic et al., 1996; Karge et al., 1998). Only studies with randomization at the pig level were included in the meta-analysis to avoid bias. Different biases can occur when performing a meta-analysis. All available databases were searched to avoid sampling bias. The trend towards publishing only positive results (Rosenthal, 1979) can lead to publication bias. This is very dif®cult to assess because it is often not possible to identify unpublished studies. A funnel graph can be used to investigate publication bias, because publication bias will tend to skew the pyramid shape by selectively excluding studies with small or non-signi®cant effects (Everitt, 1995). No indication of publication bias was seen from the funnel graph (Fig. 2) and the variable `publication quality' was non-signi®cant in the analysis of variance. Seventy-eight percent of the studies were excluded because they did not ful®ll all of the inclusion criteria. This might lead to exclusion bias. Most of the studies were excluded for not reporting ®gures on the ADWG or group size, which were necessary for the analysis. There was a large variation between studies in the way that lung lesions were evaluated (Table 5), and no direct standardization of the diagnostic methods was possible. For this reason, the analyses were not based on the results of lung lesions, although this is a more direct measure of vaccine effect. Only ®ve ®eld trials were performed with Hyoresp1

274

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

Fig. 2. Funnel graph showing sample size against increase in ADWG between vaccinated and non-vaccinated pigs for the studies included in the meta-analysis and comparing publication quality yes/no (n ˆ 28 trials).

which was excluded from the meta-analysis. This might be due to more recent marketing (1998) of Hyoresp1 compared to the other two vaccines. The effect of vaccination might be underestimated because only a partial vaccination of the pig population was performed in the vaccine studies. However, in practice all pigs on a farm are vaccinatedÐleading to greater herd immunity than in ®eld trials. In a study estimating the cost of pneumonia (Straw et al., 1989), the ADWG decreased by 37 g for every 10% of a pig's lungs affected by pneumonia. The results of own investigation (Table 5) could not be directly related to those of Straw et al. (1989) due to the wide variation among methods to characterize lung lesion among different studies. In the present study, the average vaccination effect was a 21 g increase in ADWG (Table 5). Because the prevalence of pneumonia in vaccination trials either differed or was not reported, then the level of lung lesion reduction corresponding to a 21 g increase in ADWG is unknown. The economic value of increases in ADWG might differ among herds, but a 21 g increase might be too small to justify vaccination from an economic viewpoint. 5. Conclusion In conclusion, the present meta-analysis of 28 published ®eld trials investigating the effect of vaccinating pigs against M. hyopneumoniae showed that vaccinated pigs had a 21 g higher ADWG compared to non-vaccinated pigs when adjusted for age during the study. Vaccine type, vaccination schedule, housing system and publication quality had no signi®cant effect on ADWG, and the funnel graph suggested that there was no publication bias. Acknowledgements The authors thank Dr. Margit Andreasen for discussions and comments on the manuscript. The present study was ®nanced by Research Centre for the Management of Animal

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

275

Production and Health (CEPROS), Foulum, DK-8830 Tjele, as a part of the project; ``Animal health economics. Decision support systems for disease control in pig and cattle herds'', CEP97-5/411-231.

References Anon., 1998. Annual Report, Danish Veterinary Laboratory. Armitage, P., Berry, G., 1987. Statistical Methods in Medical Research, 2nd Edition. Blackwell Scientific Publications, London. Bili'c, V., Lipej, Z., Simpraga, B., Sudri'c, F., 1992. Usage of RespiSure vaccine in Croatia. In: Proceedings of the 12th IPVS Congress, Hague, The Netherlands, p. 326. Bili'c, V., Lipej, Z., Zuri'c, M., Stigli'c, N., 1994. Immunoprophylaxis of mycoplasmal pneumonia in fattened pigs. In: Poomvises, P., Ingkaninun, P. (Eds.), Proceedings of the 13th IPVS Congress, Bangkok, Thailand, p. 193. Bilic, V., Lipej, Z., Valpotic, I., Habrun, B., Humski, A., Njari, B., 1996. Mycoplasmal pneumonia in pigs in Croatia: first evaluation of a vaccine in fattening pigs. Acta Vet. Hung. 44 (3), 287±293. Blagovi'c, S., Fluksek, V., Lausin, M., Stigli'c, N., Cazin, P., 1992. Clinical evaluation of the protective capabilities of an adjuvanted M. hyopneumoniae vaccine (RespiSure) in swine. In: Proceedings of the 12th IPVS Congress, Hague, The Netherlands, p. 327. Busse, F.W., BoÈhne, I., 1996. Vaccination against enzootic pneumonia (EP) or mycoplasma induced respiratory disease (MIRD) in two breeding farms. In: Proceedings of the 14th IPVS Congress, Bologna, Italy, p. 234. Charlier, P., Jambers, B., Martinod, S., Legrand, A., 1994. Efficacy of Stellamune mycoplasma in European field trials. In: Poomvises, P., Ingkaninun, P. (Eds.), Proceedings of the 13th IPVS Congress, Bangkok, Thailand, p. 136. Chiu, Y.T., Weng, C.N., Liu, C.I., 1992. Protective efficacy of Mycoplasma hyopneumoniae vaccine against mycoplasmal pneumonia combined with Pasteurella multocidae infection. In: Proceedings of the 12th IPVS Congress, Hague, The Netherlands, p. 323. Christensen, G., Vraa-Andersen, L., 1993. Forebyggelse af almindelig lungesyge. Dan Dyrlaegeforen Medlemsb 76 (5), 177±183. Christensen, G., Sùrensen, V., Mousing, J., 1999. Diseases of the respiratory system. In: Straw, B.E., D'Allaire, S., Menegeling, W.L., Taylor, D.J. (Eds.), Proceedings of the Eighth Symposium on Diseases of Swine. Iowa State University Press, Ames, IA, pp. 913±940. Dalloli, C., Geroldi, S., Guadaginini, P.F., 1998. Efficacy of Mycoplasma hyopneumoniae vaccine (Suvaxyn M. Hyo) under field conditions. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 156. Dayalu, K.I., 1994. Beneficial effects of a Mycoplasma hyopneumoniae vaccination (Stellamune) as evaluated under both experimental and field conditions. In: Proceedings of the 13th IPVS Congress, Bangkok, Thailand, p. 137. Dayalu, K.I., Keich, R.L., Charlier, P., Martinod, S., 1992. Evaluation of the beneficial effects of a Mycoplasma hyopneumoniae vaccine (RespisureTM)Ðresults from controlled and field studies. In: Proceedings of the 12th IPVS Congress, Hague, The Netherlands, p. 302. De Jong, M.F., Jedema, E.J., Sampimom, O., 1996. Mycoplasma hyopneumoniae vaccination in 10-week-old piglets. Results of a field trial. In: Proceedings of the 14th IPVS Congress, Bologna, Italy, p. 220. Dickersin, K., Berlin, J.A., 1992. Meta-analysis: state-of-the-science. Epidemiol. Rev. 14, 154±176. Diekman, M.A., Grant, A.L., Kelly, D.T., Pusaateri, A.E., Wagner, J.R., Scheidt, A.B., 1994. Reproductive and lean growth performance of gilts reared in all-in/all-out facilities: effect of vaccination against Mycoplasma hyopneumoniae and exposure to ammonia. J. Anim. Sci. 72 (2), 30. Diekman, M.A., Scheidt, A.B., Grant, A.L., Kelly, D.T., Sutton, A.L., Martin, T.G., Cline, T.R., 1999. Effect of vaccination against Mycoplasma hyopneumoniae on health, growth, and pubertal status of gilts exposed to moderate ammonia concentrations in all-in-all-out versus continuous-flow system, swine health. Swine Health Prod. 7 (2), 55±61.

276

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

Dohoo, I.R., Montgomery, M.E., 1996. A field trial to evaluate a Mycoplasma hyopneumoniae vaccine: effects on lung lesions and growth rates in swine. Can. Vet. J. 37, 299±302. Everitt, B.S., 1995. The Cambridge Dictionary of Statistics in the Medical Sciences. Cambridge University Press, Cambridge. Fleiss, J.L., 1986. Analysis of data from multiclinical trials. Contr. Clin. Trials 7, 267±275. Guadagnini, P.F., Alborali, L.G., Meloni, S., Martelli, P., 1998. Herd health status in finishers; lung lesions and productive performance after Mycoplasma hyopneumoniae vaccination. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 156. Harrison, R., Bryant, D., 1999. Field performance using a Mycoplasma±Atrophic rhinitis±Pasteurella±Erysipelas combination Bacterin/toxoid: ProSystem 1B*P*M*E. AASP, pp. 207±208. Hogg, A., 1991. Response to Mycoplasma hyopneumoniae. AASP, pp. 23±24. Jayappa, H., Lehr, C., Tracewell, C., Wasmoen, T., 1998. Prevention of Diseases in Swine with A Mycoplasma hyopneumoniae and Haemophilus parasuis Combination Vaccine. AASP, pp. 251±254. Karge, I., SchroÈder, L., Elsen, R., 1998. Vaccination of young piglets against Mycoplasma hyopneumoniaeÐ evaluation of economic parameters. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 148. Kobisch, M., LabbeÂ, A., Morvan, P., Carolet, R., 1994. Evaluation of a Mycoplasma hyopneumonia vaccine in pigs experimentally infected with Mycoplasma hyopneumoniae and Pasteurella multocida. In: Poomvises, P., Ingkaninun, P. (Eds.), Proceedings of the 13th IPVS Congress, Bangkok, Thailand, p. 194. Kyriakis, S.C., Alexopoulos, C., Viemmas, J., Sarris, K., Lekkas, S., Koutsoviti-Papadopoulous, M., Saoulidis, K., 1999. Field experience on performance improvement after vaccination with HYORESP in Mycoplasma hyopneumoniae infected commercial pig unit. Eur. Focus, 21. Le Grand, A., Kobisch, M., 1996. Enzootic pneumonia: comparison of the effect of pulse medication and vaccination. In: Proceedings of the 14th IPVS Congress, Bologna, Italy, p. 223. Lium, B., Lund, A., Skomsùy, A., 1994. A field study on vaccination against Mycoplasma hyopneumoniae infection in pigs. In: Poomvises, P., Ingkaninun, P. (Eds.), Proceedings of the 13th IPVS Congress, Bangkok, Thailand, p. 191. Maes, D., Verdonck, M., Castryck, F., Vrijens, B., Verebek, W., Viaene, J., de Kruif, A., 1998a. The effect of vaccination against Mycoplasma hyopneumoniae in pig herds with an all-in/all-out production system. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 152. Maes, D., Verdonck, M., Castryck, F., Miry, C., Gielis, D., Vrijens, B., de Kruif, A., 1998b. The effect of vaccination against Mycoplasma hyopneumoniae in pig herds with a continuous production system. In: Done, S., Thomson, J., Varley, M. (Eds.), 15th IPVS Congress, Birmingham, England, p. 153. Maes, D., Deluyker, H., Verdonck, M., Castryck, F., Miry, C., Lein, A., Vrijens, B., De Kruif, A., 1998c. The effect of vaccination against Mycoplasma hyopneumoniae in pig herds with a continuous production system. J. Vet. Med. Ser. B 45, 495±505. Maes, D., Deluyker, D., Verdonck, M., Castryck, F., Miry, C., Vrijens, B., Verebek, W., Viaene, J., de Kruif, A., 1999. Affect of vaccination against Mycoplasma hyopneumoniae in pig herds with all-in/all-out production system. Vaccine 17, 1024±1034. Martinod, S., 1996. Protection against Mycoplasma hyopneumoniae and Actinobacillus pleuropneumoniae infections using a mycoplasma inactive vaccine RespiSure under field conditions. In: Proceedings of the 14th IPVS Congress, Bologna, Italy, p. 221. Martinon, O., Tiberghien, M.P., Reynaud, G., Blanchet, M., Brun, A., Milward, F., 1998a. Efficacy of Mycoplasma hyopneumoniae Bacterin (Hyoresp) in challenge studies. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 157. Martinon, O., Tiberghien, M.P., Reynaud, G., Blanchet, M., Brun, A., Milward, F., 1998b. Efficacy of a `on shot' schedule of Mycoplasma hyopneumoniae Bacterin (Hyoresp). In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 284. Martinon, O., Tiberghien, M.P., Reynaud, G., Blanchet, M., Brun, A., Milward, F., 1998c. Duration of immunity of a Mycoplasma hyopneumoniae Bacterin (Hyoresp). In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS, Birmingham, England, p. 285. Miller, S.K., Ross, R.F., Erickson, B.Z., Gerber, D., Schultz, R.H., Chladek, D.W., 1992. Response of pigs vaccinated with Mycoplasma hyopneumoniae vaccines to challenge with virulent Mycoplasma hyopneumoniae. In: Proceedings of the IPVS Congress, Hague, The Netherlands, p. 324.

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

277

Moore, C., Daigneault, J., Hoover, T.C., 1997. Efficacy of Different Strategies of Mycoplasma Vaccination As Measured by Serology. AASP, pp. 135±136. Morrow, W.E.M., Iglesias, G., Stanislaw, C., Stepphenson, A., Erickson, G., 1994. Effect of a mycoplasma vaccine on average daily gain in swine. Swine Health Prod. 2 (6), 13±18. MunÄoz, A., PallareÂs, F.J., Ramis, G., 1996a. Study of the effect of vaccination with Stellamune mycoplasma1 on enzootic pneumoniaÐrelated clinical and productive parameters under field conditions. In: Proceedings of the 14th IPVS Congress, Bologna, Italy, p. 229. MunÄoz, A., PallareÂs, F.J., Ramis, G., 1996b. Use of a general linear model to evaluate the effect of vaccination of swine with Stellamune mycoplasma1 on zootechnical performance and financial gain. In: Proceedings of the 14th IPVS Congress, Bologna, Italy, p. 222. Murphy, D.A., van Alstine, W.G., Clark, L.K., Albregts, S., Knox, K., 1993. Aerosol vaccination of pigs against Mycoplasma hyopneumoniae infection. Am. J. Vet. Res. 54 (11), 1874±1880. Nash, W.A., 1996. Mycoplasma hyopneumoniae vaccination and disease control. Pig. J. 37, 78±94. Okada, M., Okonogi, H., Ono, M., Sakano, T., Sato, S., 1998. Evaluation of Mycoplasma hyopneumoniae vaccine in pigs under experimental and field condition. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 290. Pejsak, Z., Tarasiuk, K., Cazin, P., 1992. Vaccination against mycoplasmal pneumoniaÐa field study on RespiSure. In: Proceedings of the 12th IPVS, Hague, The Netherlands, p. 325. Petersen, G.R., Weiss, D., 1992. Vaccination against mycoplasma hyopneumoniae in swine. Pig. Vet. J. 28, 35±39. Petersen, G., Weiss, D., Egan, J., Korshus, J., Peters, R., Miron, M., 1991. Response to Mycoplasma hyopneumoniaeÐvaccination in nursering piglets. AASP, pp. 17±21. Pihl, K., Bñkbo, P., 1996. Vaccination mod mycoplasma-lungesyge hos svin. Vet. Inform. 6, 3±5. Reynaud, G., Brun, A., Milward, F., Lacoste, F., Vandeputte, J., 1992. Clinical results obtained with an inactivated vaccine against porcine mycoplasmosis. In: Proceedings of the 12th IPVS Congress, Hague, The Netherlands, p. 303. Reynaud, G., Martinon, O., Chouvet, C., Brun, A., Milward, F., 1998. Clinical field trial with Mycoplasma hyopneumoniae bacterin (hyoresp). In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 150. Rosenthal, R., 1979. The ``fiel drawer problem'' and tolerance for null results. Psychol. Bull. 86 (3), 638±641. Rosenthal, R., 1991. Meta-analytic Procedures for Social Research. Saga Publication. Ross, R.F., 1999. Mycoplasmal diseases. In. Straw, B.E., D'Allaire, S., Menegeling, W.L., Taylor, D.J. (Eds.), Proceedings of the Eighth Diseases of Swine. Iowa State University Press, Ames, IA, pp. 495±510. SAS Institute, 1996. SAS/STAT Software, Changes and Enhancement, Version 6.11. SAS Institute, Inc., Cary, NC, 1996. ISBN 1555442749. Scheidt, A.B., Mayrose, V.B., Van Alstine, W.G., Clark, L.K., Cline, T.R., Einstein, M.E., 1993. The Benefit to Cost Ratio of Vaccinating Pigs for Mycoplasmal Pneumonia in a Swine herd Affected by Enzootic Pneumonia. AASP, pp. 29±30. Scheidt, A.B., Mayrose, V.B., Van Alstine, W.G., Clark, L.K., Cline, T.R., Einstein, M.E., 1994. The effects of vaccinating pigs for mycoplasmal pneumonia in a swine herd affected by enzootic pneumonia. Swine Health Prod. 2 (1), 7±11. Sheldrake, R.F., Gardener, I.A., Saunders, M.M., Romalis, L.F., 1991. Intraperitonal vaccination of pigs to control Mycoplasma hyopneumoniae. RVTSA 51, 285±291. Sheldrake, R.F., Romalis, L.F., Saunders, M.M., 1993. Serum and mucosal antibody responses against Mycoplasma hyopneumoniae following intraperitonal vaccination and challenge of pigs with M hyopneumoniae. RVTSA 55, 371±376. Straw, B.E., Tuovinen, V.K., Bigras-Poulin, M., 1989. Estimation of the cost of pneumonia in swine herds. JAVMA 12, 1702±1706. Stripkovits, L., Miller, D.J.S., 1993. Comparative Studies of the Efficacy of Mycoplasma Hyopneumonia Vaccine (RespiSure), Lincomycin and Tiamulin Against Experimental M. hyopneumoniae Infection of Piglets. AASP, pp. 17±26. Thacker, B., Boettcher, T., Anderson, T., Thacker, E., Young, T., 1998a. The influence of passive immunity on serological responses to Mycoplasma hyopneumoniae vaccine. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 154.

278

C.S. Jensen et al. / Preventive Veterinary Medicine 54 (2002) 265±278

Thacker, E.L., Thacker, B.J., Boettcher, T.B., Jayappa, H., 1998b. Comparison of antibody production, lymphocyte stimulation, and protection induced by four commercial Mycoplasma hyopneumoniae bacterins. Swine Health Prod. 6 (3), 107±112. Thacker, E.L., Thacker, B.J., Jayappa, H., 1998c. Evaluation of humeral and cellular immune response and protection induced in pigs by five commercial mycoplasma vaccines. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 151. Thacker, E.L., Halbur, P.G., Thacker, B.J., 1999. Effect of Vaccination on Dual Infection with Mycoplasma hyopneumoniae and PRRV. AASP, pp. 375±377. Trayer, T., 1994. Performance of Pigs Vaccinated with RespiSure in All-in/All-out Housing. AASP, pp. 23±25. Uhlenhopp, E.K., 1991. Suvaxyn Respired MH: Findings From the Field. AASP, pp. 123±131. Vraa-Andersen, L., 1994a. Respiratory diseases in Danish slaughter swine application of epidemiological methods. Ph.D. Thesis. KVL, Denmark. Vraa-Andersen, L., Christensen, G., 1993. Forebyggelse af almindelig lungesyge. Dan Dyrlaegeforen Medlemsb 76 (4), 129±132. Vraa-Andersen, L., Christensen, G., Kuiper, R., 1994b. Vaccine efficiency trial with Suvaxyn M. hyo in Denmark. In: Poomvises, P., Ingkaninun, P. (Eds.), Proceedings of the 13th IPVS, Bangkok, Thailand, p. 192. Wallgren, P., LoÈfstedt, M., Heldmer, E., 1998. Strategic vaccination against Mycoplasma hyopneumoniae to avoid marketing contagious animals from multiplying herds. In: Done, S., Thomson, J., Varley, M. (Eds.), Proceedings of the 15th IPVS Congress, Birmingham, England, p. 149. Weiss, D.L., Peterson, G.R., 1992. Mycoplasma hyopneumoniae Bacterin field efficacy study. In: Proceedings of the 12th IPVS Congress, Hague, The Netherlands, p. 305. Weng, C.N., Tzan, Y.L., Liu, S.D., Lin, S.Y., Lee, C.J., 1992. Protective effects of an oral microencapsulated Mycoplasma hyopneumoniae vaccine against experimental infected pigs. RVTSA 53, 42±46.