AVPA/Conferences/February 2007/What's Happened to Marek’s Disease

From Poultry Hub

What's Happened to Marek's Disease?

A summary from the national MDV/CAV survey

Peter Groves¹, Steve Walkden-Brown² and Fakhrul Islam².

¹ Zootechny Pty Ltd, Bringelly, 2556
² School of Rural Science and Agriculture, University of New England, Armidale 2351

The national survey of Marek’s Disease and Chicken Anaemia Virus was conducted as part of the Australian Poultry CRC’s project #03-17. This was a collaborative project between UNE, RMIT University and Zootechny Pty Ltd. The project entailed three strands: virus pathotyping, infective dose titration and epidemiology. The survey constituted a part of the epidemiological strand of that project.
Only a summary of some outcomes of the MDV assessment will be reported here.

The survey involved 72 broiler farms (289 sheds) across Queensland, New South Wales, Victoria, South Australia, Tasmania & Western Australia. Data and sample collection extended across four seasons. Broiler farms were selected randomly, based on their date of placement and four sheds per farm were sampled. All major broiler companies cooperated and we achieved a 90% response rate.

The objectives of the survey were to identify factors which modify the risks of a farm being positive for MDV1 and/or CAV, to evaluate the effects of farms being positive for MDV1 in the face of HVT vaccination and to evaluate the effects of being CAV positive.

On each farm, a detailed management questionnaire was completed and dust samples were collected by settle plate at weekly intervals. Dust sampes were assayed for presence of MDV serotype1 by quantitative real time PCR (q-PCR). Blood and spleen samples were collected from 5 birds in each of 2 of sheds on each farm at 42 days of age. Blood samples were assayed for CAV antibody and spleens were assayed for MDV types 1 and 3 by q-PCR.

The detailed questionnaire collected information on proximity to other farms, biosecurity practices, hygiene, HVT vaccination, shed/ventilation design, bird density, breed, brooding factors and bird performance (mortality, growth, FCR).

State	Farms MDV+	Farms MDV-	P =
NSW	15	12	0.19
QLD	5	5
SA	6	2
TAS	1	3
Vic	9	10
WA	0	4

The distribution of MDV positive sheds by state shows no significant difference in positive rates. WA has a “zero cell” – inclusion of this would have serious statistical connotations for the analysis creating unstable estimates of coefficients. Tasmania, only having 1 positive out of only 4 samples could also cause an imbalance in the analysis. Further analysis had to proceed ignoring these two states.

In comparing farm ownership, the was no significant difference at all in distribution of positive/negatives due to broiler company in rate of being positive for MDV1 in dust at 42 days.

A univariate analysis was conducted over all recorded factors related to being MDV positive in dust. This was aimed at selecting variables that may contribute to a farm having at least one shed dust-positive for MDV1.
Selection of factors to be further studied was based on their univariate statistical probability of contribution being at least moderate (P<0.25, Hosmer & Lemeshow, 2000).

Many of these factors may be confounded, interactive or collinear. A multiple analysis is needed to control for the effects of combined factor effects.

Notably, HVT vaccination had no effect on the positive rate (not really expected that vaccination would prevent infection).

Forward stepwise multiple logistic regression was used to control for the combined presence of factors selected as significant contributors to the farm being MDV positive in dust at 42 days. Factors determined to contribute significantly to the model were:

There was a strong seasonal effect (>5x more risk of being positive in summer-autumn than in winterspring) Presence of another chicken farm within 2km Provision of clean clothing on farm had an apparent protective effect.

Larger farms (>5 sheds) and the use of surface water remained as risk factors and provision of farm clothing gave a significant protective association. These are far less important than the first 2 – and their 95% confidence intervals for their odds ratios include unity.

In considering the consequences of being MDV1 positive in dust for the farm, we analysed across performance factors with a summary shown in the following table.

Factor	MDV POS	MDV NEG	P =
Adjusted FCR	1.87	1.90	0.28
Mort % d0-14	1.68	1.49	0.41
Mort % Standardised to last 15 days	2.68	2.72	0.19
Mort Final	5.27	4.65	0.24
EOB age (days)	50.6	51.3	0.36
Bottom half of pool	55%	37%	0.17

Differences seen with flocks that were MDV dust-positive in this survey were not significantly different to negative sheds. If anything there was a tendency for MDV positive farms to be in the lower half of the broiler performance pool (farms were assessed for their position in their own operation’s pool and this does not contrast performance between farms of different operations). From our experience in the late 1990’s we would have expected late mortality to be elevated in farms with an established MDV1 infection. The data really needed a thorough examination for any confounding or interacting effects here.

We also needed to look at effect of vaccination when MDV1 was present, compared to when it was not detected in dust. Distribution of HVT vaccination across the breeds was not uniform. One breed had only 3 vaccinated sheds out of 64. Comparing mortality and growth performance without accounting for breed would be misleading.

There were no significant differences in late mortality (in one breed that received vaccination) between HVT vaccinated and non-vaccinated flocks, regardless of MDV1 status of the flock. We expected to see a difference here with MDV-positive flocks at least. In a survey like this we need to look for possible confounding and biases in this situation which may distort the raw data.

MDV1 status	Route of HVT vaccination	No. flocks	Standardised late mortality
Negative	subcutaneous	17	3.34 % a,b
Negative	in ovo	54	2.70 % a
Positive	subcutaneous	14	4.93 % b
Positive	in ovo	16	2.18 % a
		P =	0.0009

a,b – means without common postscripts differ significantly, P< 0.001

When we contrast the method of HVT vaccination application we get a different picture. There does appear to be a highly significant difference with flocks that were vaccinated at day old compared to those using in ovo vaccination with standardised late mortality, with the day old route associated with higher late losses where flocks were MDV positive in dust at 42 days. There are no immediately obvious reasons available from the data and this may need closer examination. This cannot differentiate between an actual effect of the day old vaccination or of differences for other reasons involving the hatcheries/operations using day old vaccination, so day old vaccination per se cannot be attributed with the sole effect here.

In ovo vaccinated flocks (one breed only) had significantly higher representation in the higher quartiles for average live weight than day old vaccinated flocks. In ovo vaccinated flocks also tended significantly to be processed at later ages and thus had higher average live weights. This could indicate that better health and growth rate in these flocks – a situation which usually contributes to that flock being chosen to be carried to an older age.

Conclusions so far:

Major risk factors for exposure to MDV1 identified were:

Flocks placed in Summer-Autumn (January to June), and

Proximity to another chicken farm (especially within 2 km)

Neither of these factors are under the farmer’s control.
Estimates of both the cost of being MDV dust-positive and any benefit from HVT vaccination are clouded by the route of vaccine application.

References

Hosmer, D.W, and Lemeshow, S. 2000. Applied Logistic Regression, 2nd Ed. John Wiley & Sons, New York.
Martin, S.W., Meek, A.H. and Willeberg, P. 1988. Veterinary Epidemiology, principles and methods. Iowa State University Press, Ames.