Measuring and Monitoring FMD Occurrence Melissa McLaws EuFMD
Acknowledgements
•Chris Bartels: EuFMD
•Naci Bulut: FMD Institute, Ankara, Turkey
•Theo Knight Jones: Institute for Animal Health, Pirbright, UK
•Shams Amin Abo Gabal, Rehab Abdel-Kader El Bassal, Animal Health Research Institute, Cairo, Egypt
Soheir Hassan Abdel-Kader, Iman Ali Farag, Amaal Ibrahim Attya Mansour, General Organization for Veterinary Services, Cairo, Egypt
•General Directorate for Food and Control, Ankara, Turkey
•Iranian Veterinary Organization, Tehran, Iran
Key points • Data from Turkey and Iran show:
– Yearly incidence rate 10-30 X higher when measured using NSP serosurveys compared to reports of suspected cases
– Provincial (or district) level FMD incidence measured by 1) reports and 2) serological data were poorly correlated
• FMD reporting should be encouraged and reports analysed continuously (monthly) to detect ‘events’
• Carefully designed serosurveys should be conducted regularly (annually if possible)
– Unbiased measure of FMD infection, best information about risk factors
Monitoring FMD is key principle of PCP
Measuring FMD occurrence to:
1. Assess FMD risk
– Within country:
oSpecies, sector, husbandry system, area
oTo target control and inform FMD impact assessment
2. Monitor changes over time
– Detect ‘events’ (epidemics)
– Efficacy of control strategy
But what is the best way to do it?
Incidence (%)6.00 - 12.004.00 - 6.002.00 - 4.001.00 - 2.000.50 - 1.000.00 - 0.50No data
no. villages reporting in 2010 / ttl villages in province
2010 Village Incidence FMD reports
0
100
200
300
num
ber
of re
po
rts
Jan
2009
Apr
200
9
Jul 2
009
Oct 2
009
Jan
2010
Apr
201
0
Jul 2
010
Oct 2
010
month
Turkey:2009-2010
All FMD reports
Monitoring FMD Occurrence: Case Reports
Turkey
2009-2010
2010 Village level incidence
Monitoring FMD Occurrence: Case Reports
• Most are laboratory confirmed & serotyped
0
100
200
300
400
num
ber
of o
utb
reaks
Jan
2001
Jan
2002
Jan
2003
Jan
2004
Jan
2005
Jan
2006
Jan
2007
Jan
2008
Jan
2009
Jan
2010
Jan
2011
Jan
2012
Month
All Outbreaks
05
01
00
150
200
250
num
ber
of o
utb
reaks
Jan
2001
Jan
2002
Jan
2003
Jan
2004
Jan
2005
Jan
2006
Jan
2007
Jan
2008
Jan
2009
Jan
2010
Jan
2011
Jan
2012
Month
Serotype A
05
01
00
150
200
250
num
ber
of o
utb
reaks
Jan
2001
Jan
2002
Jan
2003
Jan
2004
Jan
2005
Jan
2006
Jan
2007
Jan
2008
Jan
2009
Jan
2010
Jan
2011
Jan
2012
Month
Serotype O
05
01
00
150
200
250
num
ber
of o
utb
reaks
Jan
2001
Jan
2002
Jan
2003
Jan
2004
Jan
2005
Jan
2006
Jan
2007
Jan
2008
Jan
2009
Jan
2010
Jan
2011
Jan
2012
Month
Serotype Asia1
O Asia-1 A
Turkey: 2001-2012
Monitoring FMD Occurrence: Active surveillance
• Clinical cases
• NSP serology
– Sampling young stock gives picture of FMD infection in recent months
Incidence80.00 - 100.0060.00 - 80.0040.00 - 60.0020.00 - 40.000.00 - 20.00No data
2010 Village Seroprevalence
Turkey: 2010
• Compare FMD incidence measured using survey data and reports in Turkey and Iran
• Well developed passive reporting systems
• Extensive serosurveys
Turkey (spring 2009) Turkey (spring 2010) W. Azerbaijan, June 2011
-32,670 samples -64,765 samples -8349 samples
78 provinces, 334
districts, 554 villages
74 provinces, 460
districts, 946 villages
1 province, 14 districts
281 epi-units
-60 samples /village -60 samples /village - 30 samples/epi unit
-cattle only
-mostly 4-18 months
50% cattle, 50% SR
-mostly 4-18 months
-cattle only
-6-24 months
How to relate reporting and serological data?
Incidence6.00 - 12.004.00 - 6.002.00 - 4.001.00 - 2.000.50 - 1.000.00 - 0.50No data
no. villages reporting in 2010/ttl villages in province
2009 Village Incidence FMD reports
Incidence80.00 - 100.0060.00 - 80.0040.00 - 60.0020.00 - 40.000.00 - 20.00No data
2009 Village Seroprevalence
Incidence80.00 - 100.0060.00 - 80.0040.00 - 60.0020.00 - 40.000.00 - 20.00No data
2010 Village Seroprevalence
How to relate reporting and serological data?
2010 2009
Serology
Reports
Incidence (%)6.00 - 12.004.00 - 6.002.00 - 4.001.00 - 2.000.50 - 1.000.00 - 0.50No data
no. villages reporting in 2010 / ttl villages in province
2010 Village Incidence FMD reports
0
100
200
300
num
ber
of re
po
rts
Jan
2009
Apr
200
9
Jul 2
009
Oct 2
009
Jan
2010
Apr
201
0
Jul 2
010
Oct 2
010
month
Turkey:2009-2010
All FMD reports
Relative Incidence200.00 - 500.00100.00 - 200.0080.00 - 100.0060.00 - 80.0040.00 - 60.0020.00 - 40.000.00 - 20.00No data
Serological/Report Village Incidence 2009
Relative comparison (serological ÷ report incidence)
Relative Incidence200.00 - 500.00100.00 - 200.0080.00 - 100.0060.00 - 80.0040.00 - 60.0020.00 - 40.000.00 - 20.00No data
Serological/Report Village Incidence 2010
2009:
Median 30X
2010:
Median 11X
Absolute comparison (Serology minus Report incidence)
Incidence difference (%)80 - 10060 - 8040 - 6020 - 400 - 20-10 - 0No data
Serological and Report Village Incidence 2009: Absolute difference
Incidence difference (%)80 - 10060 - 8040 - 6020 - 400 - 20-10 - 0No data
Serological and Report Village Incidence 2010: Absolute difference
Median difference: 20% Median difference: 29%
W. Azerbaijan serosurvey
• 80.2% of epi-units had at least 5 calves with a high titre (>70% inhibition)
• 18% observed clinical signs in their stock in the previous 12 months (questionnaire)
Incidence85.71 - 100.0080.71 - 85.7171.43 - 80.7166.67 - 71.4350.00 - 66.67
2011 Incidence FMD serosurvey
Incidence28.57 - 40.0027.27 - 28.5719.81 - 27.2716.67 - 19.818.57 - 16.670.00 - 8.57
Incidence FMD reports (sample)
Incidence13.77 - 32.769.97 - 13.779.18 - 9.977.81 - 9.187.37 - 7.811.52 - 7.37
16 mo prior to survey
Incidence FMD reports
District-level FMD Incidence:
Serosurvey Clinical signs: Survey Clinical signs: Offical reports
No significant correlation (Spearman’s)
How to relate reporting and serological data?
Epi-unit level Turkey
2009
Turkey
2010
W. AZB
Median serological incidence 20% 33% 81%
GISVet survey
Median report incidence 0.22% 2.8% 9% 20%
Median relative incidence 30.8 11.8 8.2 3.7
Median incidence difference 20% 29% 69% 65%
50
60
70
80
90
100
se
rolo
gic
al in
cid
en
ce
0.00 10.00 20.00 30.00 40.00report incidence
W. Azerbaijan serosurvey
02
04
06
08
01
00
se
rolo
gic
al in
cid
en
ce
0 5 10 15report incidence
Turkey 2010
Differences: Serological and Report data
May be due to:
1. Under-reporting
2. Subclinical infection
• FMD signs observed by farmers on:
– 17% NSP+ epi-units in W. Azerbaijan
– 41% NSP+ calves (0-12 months) in Asia-1 outbreak investigation in Turkey
– 86% NSP+ calves (0-12 months) in Asia-1 outbreak investigation in Turkey
• Will vary by FMD strain, species infected, vaccination status
Differences: Serological and Report data
May be due to:
3. NSP antibody induced by vaccine rather than infection
• Especially if use unpurified vaccine
• Less likely if youngstock targeted in survey (?)
4. Reflection of previous year’s cases?
• Compared Turkish serosurvey results to previous year’s reports and correlation not improved
Both approaches contribute to assessing and
monitoring FMD risk
• Reports of suspect clinical cases
– Essential for early detection & response (PCP Stage 3 and higher)
– Real-time, linked to control
– Laboratory confirmation including serotype
– Cost-effective
– Effectiveness dependent on favourable ‘attitude’ to reporting
– Subject to bias, under-reporting
Egypt • Nationwide serosurvey in 2011
– 5299 ruminant samples tested in 310 villages
• NSP-Ab positive:
– 17.6% samples (95%CI: 16.6 – 18.6%)
– 78% villages (95% CI: 73-82%)
• 2011: 15 reports of suspect disease
0
50
100
150
200
250
300
350
0
10
20
30
40
50
60
70
80
Nile delta
Central Upper East West Total n
um
ber
villa
ges
sam
ple
d
% p
osit
ive
Egypt 2011: Large ruminant results in serosurvey
% villages with >=2 seropositive % cattle % buffalo Nr villages tested
Both approaches contribute to assessing and
monitoring FMD risk
• NSP serology – Detects clinical and subclinical FMD (ie measures FMD
infection)
– Less biased (with careful design!)
• Unit of analysis: what is a ‘case’?
• Target young stock
– Combine with questionnaire about risk factors
– Resource intensive
• Survey can be combined with post-vaccination monitoring (SP serosurvey)
Key points • Data from Turkey and Iran show:
– Yearly incidence rate 10-30 X higher when measured using NSP serosurveys compared to reports of suspected cases
– Provincial (or district) level FMD incidence measured by 1) reports and 2) serological data were poorly correlated
• FMD reporting should be encouraged and reports analysed continuously (monthly) to detect ‘events’
• Carefully designed serosurveys should be conducted regularly (annually if possible)
– Unbiased measure of FMD infection, best information about risk factors
Thank you!
•EuFMD
•FMD Institute, Ankara, Turkey
•Institute for Animal Health, Pirbright, UK
•Animal Health Research Institute, Cairo, Egypt
•General Organization for Veterinary Services, Cairo, Egypt
•General Directorate for Food and Control, Ankara, Turkey
•Iranian Veterinary Organization, Tehran, Iran
2010 epidemic in Turkey: Use of reports to detect surge in cases
ANIMAL LEVEL
SEROPREVALENCE
VILLAGE
LEVEL
2009 2010 2009 2010
<4 mo 16.67%
(0-49%)
- 20% 33%
4-12
mo
8.9%
(8.4-9.3%)
12.9
(12.5-13.2)
13-18
mo
9.1%
(8.4-9.7%)
15.4
(14.9-15.9)
19-24
mo
19.1% (18.0-
20.1%)
17.3
(16.3-18.2)
>24 mo 16%
(2.8-28.4%)
21.3
(18.0-24.5) 0
100
200
300
num
ber
of re
po
rts
Jan
2009
Apr
200
9
Jul 2
009
Oct 2
009
Jan
2010
Apr
201
0
Jul 2
010
Oct 2
010
month
Turkey:2009-2010
All FMD reports
Serosurveys
0
100
200
300
400
num
ber
of o
utb
reaks
Jan
2001
Jan
2002
Jan
2003
Jan
2004
Jan
2005
Jan
2006
Jan
2007
Jan
2008
Jan
2009
Jan
2010
Jan
2011
Jan
2012
Month
All Outbreaks