Epidemiologic Investigations of HPAI and Disease Prevention · • Plumes of virus–associated...

Epidemiologic Investigations

of HPAI and Disease Prevention

Brian McCluskey, DVM, MS, PhD, Dip. ACVPM

USDA, APHIS, Veterinary Services

Epidemiologic Investigations

• Field-based observational studies

• Geospatial analyses

• On-farm sampling efforts

• Phylogenetic investigations

“The study of the distribution and determinants of disease in populations with the goal of using those determinants

to prevent disease”

Outbreak Investigation:

Case Series

• 81 turkey farms in IA, MN, ND, SD and WI

– 63 meat production

– 11 breeder farms

– 6 raising breeder candidates

– 1 unknown type

• Developed clinical signs between March 30 and May 2, 2015.

Animal and Plant Health Inspection Service Veterinary Services

HPAI Investigation - Questionnaire

INSTRUCTIONS

The purposes of these investigations are to assess potential pathways of initial introduction of HPAI

viruses onto commercial poultry operations and potential lateral transmission routes of HPAI viruses

from infected premises to noninfected premises.

Following confirmation of an HPAI virus introduction into a commercial flock, an investigation should be

initiated as soon as possible, no later than 1 week following detection. The investigator(s) assigned

should be integrated into other response activities but their primary focus is on completion of the

introduction investigation.

Analysis conducted by Dr. Andrea Beam and Dr. Dave Dargatz, Veterinary Services

Biosecurity

Number Respondents Level or Response Percent Farms

House with family on property 81 Yes, common drive 38

Yes, no common drive 22 No 40

Signage (“no admittance” or “biosecure area”) 80 83

Gate to farm entrance 79 Yes, locked 9 Yes, not locked 18 No 73

Farm area fenced in 81 11 Freq veg. mowed (per month) 81 < 4 40

4 + 60 Facility free of debris/trash 81 89 Wash/spray area for vehicles 81 46

Designated parking workers/visitors 80 49

Changing area for workers 81 Yes, shower 27 Yes, no shower 46 No 27

Workers wear dedicated coveralls 81 73 Workers wear rubber boots or Boot covers 81 100

Barn doors lockable 81 Yes, routinely locked 40 Yes, not routinely lock 22 No 38

Foot pans at barn entrances 81 Yes, in use 99 Footbath type 81 Dry 12

81 Liquid 98 Ante area 81 98 Rodent bait station 81 Yes, checked q 6 weeks 95 Fly control 81 41 Raccoons, possums, foxes seen in or around barns 81 28

Wild turkeys, pheasants, quail seen around poultry 81 26

Biosecurity audits

81 43

Equipment Number Respondents

Level or Response Percent Farms

Farm specific (NOT shared with other farms) 75 Company vehicles/ trailers 65

77 Feed trucks 19

80 Gates/panels 91

80 Lawn mowers 63

78 Live haul loaders 8

68 Poultry trailers 31

72 Pre-loaders 15

79 Pressure sprayer/washer 57

77 Skid-steer loader 61

67 Tillers 87

70 Trucks 56

58 Other 66

Wild Bird Characteristics

Number Respondents

Level or Response Percent Farms

Wild birds around farm

78

Waterfowl

63

79

Gulls

33

78

Small perching birds

96

78

Other water birds

15

78

Other birds

28

Houses bird proof

79 62

Wild birds seen in house

81 35

Birds seen year round

77 90

Seasonality to presence of some birds

79 84

Bird location

76 Away from facilities 49

77 On farm, not in barns 66

76 On farm, in barns 26

Case-Control Study - Layers

• Study included all detected cases as of

May 15, 2015 in Iowa and Nebraska

• Total of 28 cases and 31 controls

• Survey administered to site managers

• At risk period was two weeks prior to

detection

• Controls were matched by at risk period

• Factors examined at farm and barn level

Analysis conducted by Lindsey Garber and Kathe Bjork, Veterinary Services

Case Control Study - Layers

FACTOR*

Farm Level

Percent

case farms

Percent

control

farms

Odds Ratio P-value Average

attributable

fraction

In an existing control zone 50 10 32.0 .002 31.7%

Rendering trucks near barns 29 3 22.3 <.001 14.0%

Garbage trucks near barns 61 23 14.7 <.001 28.1%

Visitors change clothes 77 93 0.08 /

12.6**

.01 7.6%**

Company service person visit in past

14 days 50 19 5.0 <.001 15.0%

*Reference level = absence of factor ** do not change clothes

Case Control Study - Layers

FACTOR*

Barn level

Percent

case farms

Percent

control

farms

Odds Ratio P-value Average

attributable

fraction

Barn entry with a hard surface entry pad

cleaned and disinfected

28.6 53.6 0.16/6.9** 0.01 33.7%**

Disposing of dead birds near a barn

(within 30 yards)

60.7 35.5 2.8 0.002 20.2%

Having ceiling or eaves inlet ventilation

type (compared with curtain, sidewall or

tunnel types)

48.2 67.7 0.33/3.0** <0.001 23.4%**

*Reference level = absence of factor **absence of factor

Biosecurity

“It is frustrating how difficult it is to prove

scientifically which biosecurity efforts work

or are the most important or effective”

Dr. Francisco Revriego

European Commission

“Biosecurity is inconvenient” Dr. John Glisson U.S. Poultry and Egg Association

Wind related spread of EA/AM H5N2 HPAI virus

• Evaluate the role of aerosol transmission between commercial turkey flocks in Minnesota (n=98)

• Used EPA’s AERMOD modeling system

• Wind speed, direction data collected from 105 National Weather Service stations across Minnesota

• Other meteorological parameters were also collected

Analysis conducted by Sasi Malladi (Univ of Minn), Todd Weaver, Amy Delgado, Phillip Riggs and Andrew Fox, Veterinary Services

Figure 1. Aerosol exposure regions representing a 0.01 (lighter region) and 0.05 (darker region)

daily likelihood of transmission from an infected premises at 4 mph (green) and 32 mph wind

speeds (brown).

Epidemiological Analyses

• Plumes of virus–associated particles

were modeled for all infected farms

based on average wind speed and

direction over various time frames.

• Case-control study examining effects of

cumulative daily exposure to modeled

plumes accounting for age and and

susceptibility of birds on each farm.

• Repeated measures examining daily and

cumulative risk of disease associated

with modeled plumes.

Results • Cumulative exposure of a farm to

multiple plumes over a 6-11 day period

was the most stable predictor of

increased odds of disease.

• Proximity to other infected farms had a

consistent association with increased

odds of becoming a case in all three

analyses.

• Given current data it is difficult to rule in

or out the contribution of aerosol

transmission of farms up to 3.5 km from

a shedding farm.

On-farm Air Sampling

Sampling and analysis conducted by Montse Torremorell et al., Univ of Minnesota

On-farm Air Sampling

Inside 5 m 70-150 m 500-1000 m

Turkeys

Positive 40 (36%) 7 (21%) 0% NT

Suspect 26 (23%) 17 (50%) 8 (38%) NT

Negative 45 (41%) 10 (29%) 13 (62%) NT

Layers

Positive 28 (78%) 22 (24%) 1 (4%) 0 (0%)

Suspect 8 (22%) 16 (18%) 9 (32%) 8 (13%)

Negative 0 (0%) 52 (58%) 18 (64%) 54 (87%)

Total

Positive 68 (46%) 29 (23%) 1 (2%) 0 (0%)

Suspect 34 (23%) 33 (27%) 17 (35%) 8 (13%)

Negative 45 (31%) 62 (50%) 31 (63%) 54 (87%)

Total 147 (100%) 124 (100%) 49 (100%) 62 (100%)

Ct <35: positive; Ct 35-40: suspect; Ct >40 negative.

Environmental risk factors

related spread of EA/AM

H5N2 HPAI virus

• State-wide dataset of poultry operations

in Minnesota was merged with records

from EMRS.

• Included information on production type

and the location of the farm in latitude

and longitude.

• Potential risk factors considered were

poultry density, distance to nearest body

of water, density of corn (within 0.5, 1 and

3 km) and distance to major highways.

• Logistic regression and boosted

regression tree analyses were conducted.

Analysis conducted by Amy Delgado, Phillip Riggs and Andrew Fox, Veterinary Services

Results

Risk Factor Levels Adjusted Odds Ratio

95% Conf. interval P-value

Species Chicken Turkey

--- 9.91

(3.00-32.37)

<0.001

Distance to nearest infected neighbor

>10 km <10 km

--- 10.67

--- (5.75, 19.8)

<0.001

Corn density (acres) within 3 km

<1191 acres 1192-1999 acres 2000-2739 acres >2740 acres

--- 5.15 8.42 12.12

--- (1.63-12.29) (2.80-25.34) (4.18-35.13)

<0.001

Results

Predictor Relative Influence

Distance to nearest infected neighbor 43.5%

Corn acreage within 3 km 14.8%

Distance to highways 71, 12, or 4 12.8%

Distance to nearest water 10.4%

Distance to any major highway 7.9%

Poultry density in county 6.8%

Production type of operation 2.4%

Species present on farm (chicken vs turkey) 0.4%

57

HPAI/LPAI Outbreak

Indiana – January 2016

• Objective

– Identify potential farm- and barn-level risk

factors for AI H7N8 on turkey farms in Indiana

during the 2016 outbreak

• Sample

– All case farms (n=9)

– Control farms from surrounding geographic

area in the disease control area (n=30)

58

Case-Control Study

• Differences identified between case farms

and barns compared to controls

• Proximity to dead bird disposal and litter

compost areas were risk factors in

previous studies

• Higher percentage of case farms had

visitors entering barns

Case-Control Study Highlights

FACTOR Percent case farms Percent control farms

Brooders raised on farm

33 62

Wild mammals around

poultry barns during the

14‐day risk period

88 41

Visitor type during 14-day risk

period:

Veterinarian

Service person

Catch crew

13

88

25

0

72

7

Pond within 350 yd 100 60

Results – Farm Level

Factor Percent case

barns

Percent control

barns

Type of bird:

Brooder

Grower

0

100

26

74

Age of birds (weeks):

<8

8 – 13

14+

0

25

75

44

15

41

People who entered barn during 14‐day risk period:

Veterinarian

Service person

Occasional worker

Any

29

88

38

100

4

61

21

68

Hard surface entry pad cleaned and disinfected 33 68

Median distance to dead bird disposal area (yds) 59 100

Results – Barn Level

• County-level analysis – Dubois and

surrounding counties

• Observed weather patterns 6 weeks prior

to outbreak

62

Geospatial Assessment for H7N8 HPAI/LPAI Indiana Outbreak

• Mean high

temperature

(50.88˚F)

• Warmer than

last 2 years

• Mean low

temperature

(30.04˚F) –

above

freezing

Temperature

• 2016 precipitation greater than 2015, less than 2014 – perhaps not single factor of importance

• Paired together – temperature & precipitation conducive to virus survival in the environment

Precipitation

• H5N2 2015 Minnesota outbreak – cropland may contribute to presence of infected waterfowl

• Cropland coverage in SW Indiana counties ranged from 3.5-78.0%

• Cropland did not appear to differ between the infected and uninfected counties

• Dubois County on lower range of wetland and open waters; not considered a factor

Percent

Cropland

Percent

Wetland

Cropland & Wetland/Open Water

• Virus circulating in waterfowl must find susceptible poultry population

• Dubois County has higher density of poultry compared to surrounding counties

Turkey & Chicken

inventory by county

Susceptible Population

Questions?