Study: Migratory birds not to blame for spread of avian flu

[Blue]

Re: Study: Migratory birds not to blame for spread of avian flu

I think it's the part about the Tamiflu that bothers me the most!

[HenryN]

Re: Study: Migratory birds not to blame for spread of avian flu

The owner of the small backyard farm was offered Tamiflu and all other poultry on the farm were culled?

Not THAT'S interesting... Hello! I don't remember hearing this part before.

Here is a media report on the quarantine of a second farm and the "inconclusive" necropsy (translation: looked like Qinghai H5N1):


Second P.E.I. farm quarantined in bird-flu probe

Last Updated: Sunday, June 18, 2006 | 9:24 PM ET

http://www.cbc.ca/canada/story/2006/...rdflu-sun.html


A second farm on Prince Edward Island has been placed under quarantine as health officials continue to investigate the discovery of avian flu virus in a domestic goose in that province.
Jim Clark, national manager of the avian influenza working group of the Canadian Food Inspection Agency, said they have not discovered any signs of bird flu on the second farm.
However, the CFIA decided to quarantine the farm as a precautionary measure when they discovered there had been movement of people and perhaps poultry between that farm and another where the virus was found last Friday.
Both farms have a small, mixed flock of free-range birds.
The alarm was raised last week when a goose tested positive for an H5 virus on a farm in western P.E.I. Four of 11 geese died on that farm, leading to a post-mortem that proved inconclusive, Clark said.
The goose came from a small flock of 40 birds, mainly geese, ducks and chickens.
<NOSCRIPT></NOSCRIPT><NOSCRIPT></NOSCRIPT>

Samples from the dead birds have been sent to the CFIA laboratory in Winnipeg to determine whether the virus found in the dead goose was the deadly H5N1 strain that has raised fears of a global pandemic since it was discovered in China in 2003. It has spread to 48 countries since then, but has not been discovered in the Americas.
"There's nothing that's obvious for us that would suggest … it is anything except a low pathogenicity virus," Clark told the Canadian Press, adding that he won’t know for sure until he sees the results of the tests on Tuesday or Wednesday.

[HenryN]

Re: Study: Migratory birds not to blame for spread of avian flu

Here is a media report on the quarantine of a second farm and the "inconclusive" necropsy (translation: looked like Qinghai H5N1):



"There's nothing that's obvious for us that would suggest ? it is anything except a low pathogenicity virus," Clark told the Canadian Press, adding that he won?t know for sure until he sees the results of the tests on Tuesday or Wednesday.

The geese were OBVIOUSLY DEAD.

The "for us" qualifier speaks volumes.

[Blue]

Re: Study: Migratory birds not to blame for spread of avian flu

The owner of the small backyard farm was offered Tamiflu...

This speaks volumes also. Since when do they offer Tamiflu with exposure to
LPAI?

[HenryN]

Re: Study: Migratory birds not to blame for spread of avian flu

The owner of the small backyard farm was offered Tamiflu...

This speaks volumes also. Since when do they offer Tamiflu with exposure to
LPAI?

The reading between the lines of the public and private info left little doubt that the four goslings on Prince Edward Island died of Qinghai H5N1.

On the public side:

The four geese had neurological symptoms. Qinghai H5N1 causes neurological symptoms in waterfowl, low path does not.

The four geese were found dead the next day. Qinghai H5N1 causes sudden death in waterfowl. Low path H5 does not.

The necropsy was "inconclusive". Qinghai H5n1 causes organ damage, low path H5 does not.

Only one of the four dead birds was tested. Protocol calls for testing of 3-4 birds.

A press conference was called to abnnounce H5 positive PCR. A press conference would be called after MULTIPLE positives and it was called something like 10 days after the birds died.

The press conference was so long after the deaths that media reports got teh date of death wrong because they thought "Monday" was in reference to the week of the press conference, instead of the Monday in the prior week.

All other birds on the farm were culled and birds on the adjacent farms were quarantined.

Farm owners were offered Tamiflu.

Each of the above strongly suggests Qinghai H5N1. In combination, the above leaves little doubt.

However, the private information was even more disturbing.

Local officials were NOT happy with distribuition of a press release.

The PCR positive was weak and just above the cut-off, again signaling Qinghai H5N1 and indoicating that the low level in DEAD birds would mean that virtually all live birds with H5N1 would test negative.

Information on the testing was forthcoming until I asked about the size of the insert. H5N1 HA would have 4 additional amino acids at the HA cleavage site, producing an insert that was 12 BP larger, which could be seen on the sizing gel. That information was considered "private" and could not be disclosed, even though the sample was from a baby GOOSE.

The sequences from low path H5 from 2005 and 2006 in Canada and 2006 in the US STILL has not been made public.

The story on H5N1 in Canada is quite clear.

[Mamabird]

Re: Study: Migratory birds not to blame for spread of avian flu

I'm curious as to why we have not discovered additional dead wildfowl in North America. After all, Qinghai is lethal, and is evident in Russia, Europe and Africa, it spreads rapidly.

[HenryN]

Re: Study: Migratory birds not to blame for spread of avian flu

I'm curious as to why we have not discovered additional dead wildfowl in North America. After all, Qinghai is lethal, and is evident in Russia, Europe and Africa, it spreads rapidly.

The amount of testing of dead waterfowl in North America is VERY limited. Most of the testing in the US and Canada was on live healthy birds. Testing of clusters of dead birds was minimal, and many dead birds tested were road kill or birds that flew into windows.

On PEI, the H5 PCR in a dead goose was a weak positive, so most H5 in live birds would be undetectable or reported as negative..

[canagica]

Re: Study: Migratory birds not to blame for spread of avian flu

The amount of testing of dead waterfowl in North America is VERY limited. Most of the testing in the US and Canada was on live healthy birds. Testing of clusters of dead birds was minimal, and many dead birds tested were road kill or birds that flew into windows.

On PEI, the H5 PCR in a dead goose was a weak positive, so most H5 in live birds would be undetectable or reported as negative..

To give an idea of how many dead birds were tested last year as part of the avian influenza surveillance effort, 7355 dead birds (morbidity/mortality, not hunter-killed) were tested, out of ~145000 total (live+m/m+hunter-killed) tested, or roughly 5% of samples.

What's really interesting is to look at the proportion of wild birds tested that year: USGS summarizes those data in quarterly reports. I did the math for 2006, and there were 66486 dead birds reported in 06; if we take the data from the HEDDS dead bird tested number (7355), then we learn that about 11% of the dead birds reported at various federal, state, and local levels were tested. These reports summarize the diagnoses for each die-off, and the diagnoses for about 3.8% (2533) of the dead birds were undetermined or the investigation is still open.

So-
Im not sure if it is accurate to say that testing of dead birds/clusters of dead birds was minimal, and from looking at the diagnoses, very few are characterized as window kills or road kill.

It would be interesting to know, in a cluster of say 5000 ducks that are diagnosed with Botulism Type C, how many birds are needed to be tested in order to come up with a conclusion for the outbreak...that would give us some idea of whether this level of testing is "very limited".

[HenryN]

Re: Study: Migratory birds not to blame for spread of avian flu

To give an idea of how many dead birds were tested last year as part of the avian influenza surveillance effort, 7355 dead birds (morbidity/mortality, not hunter-killed) were tested, out of ~145000 total (live+m/m+hunter-killed) tested, or roughly 5% of samples.

What's really interesting is to look at the proportion of wild birds tested that year: USGS summarizes those data in quarterly reports. I did the math for 2006, and there were 66486 dead birds reported in 06; if we take the data from the HEDDS dead bird tested number (7355), then we learn that about 11% of the dead birds reported at various federal, state, and local levels were tested. These reports summarize the diagnoses for each die-off, and the diagnoses for about 3.8% (2533) of the dead birds were undetermined or the investigation is still open.

So-
Im not sure if it is accurate to say that testing of dead birds/clusters of dead birds was minimal, and from looking at the diagnoses, very few are characterized as window kills or road kill.

It would be interesting to know, in a cluster of say 5000 ducks that are diagnosed with Botulism Type C, how many birds are needed to be tested in order to come up with a conclusion for the outbreak...that would give us some idea of whether this level of testing is "very limited".

Since 19/20 birds tested were alive, testing of dead birds was minimal. As I recall, many of the dead birds were from trauma (as in road kill or windows). Most of the effort was clearly directed at live healthy birds, with the expected negative results (which would be negative because the assays used were not sensitive enough to detect Qinghai H5N1 in live wild birds).

[canagica]

Re: Study: Migratory birds not to blame for spread of avian flu

Since 19/20 birds tested were alive, testing of dead birds was minimal. As I recall, many of the dead birds were from trauma (as in road kill or windows). Most of the effort was clearly directed at live healthy birds, with the expected negative results (which would be negative because the assays used were not sensitive enough to detect Qinghai H5N1 in live wild birds).

Sorry to split hairs, but 5% of the 2006 samples were from mortality events, 30% were hunter-killed, 30% were live birds, and 35% were environmental samples (USGS HEDDS)

If you compare the total number of birds reported in the 2006 USGS NWHC quarterly reports of morbidity/mortality events (~66500 birds), ~7400 of them were tested (hence about 11% of the birds involved in mortality events were tested). The diagnoses are listed for each record, and if you look you will see that <5% (6/133 events) are trauma related. I didn't total the number of birds in the trauma events, but they are a very small proportion (probably <2%) of the total number of birds reported.

[HenryN]

Re: Study: Migratory birds not to blame for spread of avian flu

Sorry to split hairs, but 5&#37; of the 2006 samples were from mortality events, 30% were hunter-killed, 30% were live birds, and 35% were environmental samples (USGS HEDDS)

If you compare the total number of birds reported in the 2006 USGS NWHC quarterly reports of morbidity/mortality events (~66500 birds), ~7400 of them were tested (hence about 11% of the birds involved in mortality events were tested). The diagnoses are listed for each record, and if you look you will see that <5% (6/133 events) are trauma related. I didn't total the number of birds in the trauma events, but they are a very small proportion (probably <2%) of the total number of birds reported.

Reality check. There were 7,355 sick, dead or dying birds tested. There were a total of 144,461 bird samples tested. More than 19/20 were tests of live healthy birds, either as those trapped, those killed by hunters, those used as sentinals, or those providing environmental samples collected with pooper scoopers.

As noted multiple times now, the testing of dead or dying birds was MINIMAL. 19/20 tests were of samples from healthy birds (which have levels of H5N1 that are below the detection limits of the tests).

[canagica]

Re: Study: Migratory birds not to blame for spread of avian flu

Reality check. There were 7,355 sick, dead or dying birds tested. There were a total of 144,461 bird samples tested. More than 19/20 were tests of live healthy birds, either as those trapped, those killed by hunters, those used as sentinals, or those providing environmental samples collected with pooper scoopers.

As noted multiple times now, the testing of dead or dying birds was MINIMAL. 19/20 tests were of samples from healthy birds (which have levels of H5N1 that are below the detection limits of the tests).

I think you are missing the point of my earlier question: the 7355 tested were out of a total of 66K reported in wildlife mortality events. Is that minimal? I don't know, hence the question of how many birds do you have to test in order to say (e.g.) that a mortality event that killed 5K ducks was due to botulism? Do we need to test all 66K, or does the current sampling scheme have some rationale behind it that is valid? The number of birds per mortality event ranges from a handful to thousands, so what would be a logical sampling regime?

Again, if anyone has seen anything published about the sensitivity of the tests (false negative rates, false positive rates, etc), it would enlighten this discussion immensely.

[HenryN]

Re: Study: Migratory birds not to blame for spread of avian flu

I think you are missing the point of my earlier question: the 7355 tested were out of a total of 66K reported in wildlife mortality events. Is that minimal? I don't know, hence the question of how many birds do you have to test in order to say (e.g.) that a mortality event that killed 5K ducks was due to botulism? Do we need to test all 66K, or does the current sampling scheme have some rationale behind it that is valid? The number of birds per mortality event ranges from a handful to thousands, so what would be a logical sampling regime?

Again, if anyone has seen anything published about the sensitivity of the tests (false negative rates, false positive rates, etc), it would enlighten this discussion immensely.

Here is the published data on Qinghai in live birds in Europe, Africa, the Middle East, and Africa. ZERO positives (other than the teal in Egypt which was positive after initially testing as negative).

The minimal is in reference to the effort directed at detecting H5N1 in dead birds in North America. The effort was directed at testing live birds. There are many more dead birds than those reported. Canada even began asking residents to report dead birds. The fact that 19/20 tests were in live birds, indicates efforts to find and test dead birds were not very great (as in minimal).

[HenryN]

Re: Study: Migratory birds not to blame for spread of avian flu

I think you are missing the point of my earlier question: the 7355 tested were out of a total of 66K reported in wildlife mortality events. Is that minimal? I don't know, hence the question of how many birds do you have to test in order to say (e.g.) that a mortality event that killed 5K ducks was due to botulism? Do we need to test all 66K, or does the current sampling scheme have some rationale behind it that is valid? The number of birds per mortality event ranges from a handful to thousands, so what would be a logical sampling regime?

Again, if anyone has seen anything published about the sensitivity of the tests (false negative rates, false positive rates, etc), it would enlighten this discussion immensely.

Here is data to show how abysmal the testing was in the United States

http://wildlifedisease.nbii.gov/ai/LPAITable.pdf

Detection of H5N1 was limited to 17 collections (see above link). The samples were said to be low path H5N1. However, the number of positives from dead or dying birds was ZERO, because testing of those birds was minimal.

However, even with samples that were H5N1 positive, H5N1 was only isolated six times. Nothing was isolated in 7 of the collections, and 4 collections yeilded serotypes other than H5N1 (H5N2, H6N2, H1N1, H4N1).

These data are well into the abysmal category, where low path H5N1 could not be isolates from samples that were H5N1 positive, and presumably gave sequences which were used to declare the samples low path.

The abysmal record on detecting or isolating H5N1 low path shows why no high path was found.

[canagica]

tests not sufficient?

Here is data to show how abysmal the testing was in the United States

http://wildlifedisease.nbii.gov/ai/LPAITable.pdf

Detection of H5N1 was limited to 17 collections (see above link). The samples were said to be low path H5N1. However, the number of positives from dead or dying birds was ZERO, because testing of those birds was minimal.

However, even with samples that were H5N1 positive, H5N1 was only isolated six times. Nothing was isolated in 7 of the collections, and 4 collections yeilded serotypes other than H5N1 (H5N2, H6N2, H1N1, H4N1).

These data are well into the abysmal category, where low path H5N1 could not be isolates from samples that were H5N1 positive, and presumably gave sequences which were used to declare the samples low path.

The abysmal record on detecting or isolating H5N1 low path shows why no high path was found.

I'm fairly certain that birds dying in wildlife mortality events are tested for multiple pathogens, AI being one of them. So-back to my original question: is testing 11% of the birds killed in reported wildlife mortality events enough? I'm not sure that testing one in ten dead/morbid birds is "minimal"; what percentage should qualify as sufficient?

As for the LPAI results, The quality of the initial test that garnered these results is described thusly:

"A series of AI rapid screening tests are performed that cannot differentiate between highly pathogenic AI (HPAI) and low pathogenic AI (LPAI) viruses. The initial rapid screening tests are highly sensitive and can detect active and inactive viruses in samples. Because these rapid screening tests are highly sensitive, it is not uncommon to have positive results for a specific subtype on the initial screen test and yet not be able to isolate a virus of that subtype."

Taken at face value, this statement indicates that the test has a high rate of false positives. So-we shouldn't be surprised that H5 wasn't even isolated from some of the samples that initially tested positive.

It would be useful to know how often LPAI H5N1 has been isolated in studies conducted over the last 50 years, with their sample sizes, to assess how frequently waterfowl have carried this subtype. Then we might objectively determine how frequently they ought to be found...Has anyone spent the time collating those data?