RE: serosurveys

<x-charset iso-8859-1>I would be particularly interested in avian serosurveys. I would like to
know if our bird populations are building some level of immunity. If so, we
may start to see some tapering off or cycling up and down of WNV. If not,
the possibility exists we could see larger outbreaks when conditions are
right (whatever that is) because there are plenty of susceptible birds to
fuel viral cycling. I am also beginning to suspect that the geographical
distribution of WNV activity may have less to do with the distribution of
mosquitoes than it does with the distribution of birds. How else to explain
the fact that we tend to see a lot of WNV activity (and 5 of our 7 human
cases in 2002) in a part of the County that is relatively the least
infested? Or have we just been fortunate not to have much spillover into
bridge vectors in the most infested areas yet? What I'm really interested in
knowing is are we likely to continue to see a few, sporadic human cases, or
are we in real danger of having a situation like Cook County, IL, with
hundreds of cases? Until we know otherwise, we have to guard against that
possibility, with all that that implies in the way of control strategies.

-----Original Message-----
From: Michael Gochfeld [mailto:gochfeld@eohsi.rutgers.edu]
Sent: Monday, October 28, 2002 2:06 PM
To: westnilevirus-L@cornell.edu
Cc: bpb01@health.state.ny.us
Subject: serosurveys

I am elated that some places are considering serosurveys according to Bryon
Backenson's email. He is right, they are indeed challenging to perform and
to
interpret. The easiest interpretations are from small villages where one
can often
see the history of a disease by the antibody distribution in people of
different age.

In Sprawlsville, it's not clear how much of a geographic area to cover, and
even
presuming that there is an
appropriately random sample, the statement of rate can only refer to the
area
represented by the sample. An assumption is that everyone in the population
has an
equal risk of exposure and infection. Other assumptions are that
immigration and
emigration don't occur or are irrelevant.

It's one thing if you're looking at an airborne exposure where a plume may
cover a
large areas (then all you have to do is
deal with atmospheric physics, discrete modeling, wind roses, etc). But for
a
mosquito borne illness, where the mosquitoes are presumably not
homogeneously
distributed, it's even more challenging. Individuals in the sample don't
have an
equal likelihood of being contacted by a mosquito or an infected mosquito.
Using GIS
and mapping the distribution of positive sera and negative sera against
known
mosquito habitats would be more useful and there is a rapidly emerging field
of
disease mapping, but I haven't heard that any of the serosurveys (still
precious few
by this time) have been examined that way.

When I did a student rotation at a virus lab in Trinidad, we did a lot of
serosurveys
(of people, birds, rodents, etc) for ARBO viruses, and were able to draw
conclusions
regarding past prevalence of infection and the degree of current protection.

I think when I mentioned the Romania data I added the caveat from one CDC
worker that
the serosurvey there had not been extremely accurate.

Mike Gochfeld

"Bryon P. Backenson" wrote:

> I may be wrong, but I believe that the "1 in 150" figure came not from the
> serosurvey done in Queens in 1999, but from the serosurvey done in 2000.
> This serosurvey was done in 3 locations (Staten Island NY, Suffolk County
> NY, Fairfield County CT), each of which covered a greater geographic area
> than the 1999 serosurvey did. The study consisted of taking blood and
> completion of a lengthy questionnaire. Approximately 750-850 individuals
> had blood drawn in each location, for a total of about 2400-2500
> individuals. Using the number of positives found, the populations of the
> area, and the number of hospitalized cases in each area, CDC crunched the
> numbers. The result were the figures we've seen replicated in numerous
> news reports:
>
> "Less than 1% of people in a given area will be exposed to West Nile
> virus."
> "Of those people that are exposed, about 1 in 150 will develop symptoms
> that are serious enough to require hospitalization."
>
> Sometimes the two statements above get misrepresented in the press.
>
> I'm not sure if the data from 2002 will alter these figures any. I know
> that some locations are considering doing serosurveys, but to be done
> right, these are extremely time- and labor-intensive, particularly when
you
> take into account the calculations that tell you the number of people
> required for statistical significance/adequate representation.
>
----------------------------------------------------------------------------
-----
>
> P. Bryon Backenson
> Research Scientist and Assistant Director
> Arthropod-Borne Disease Program
> New York State Department of Health
> ESP, Corning Tower Room 632
> Albany, NY 12237
> phone: 518-474-4568
> fax: 518-473-1708
> e-mail: bpb01@health.state.ny.us

</x-charset>
Received on Tue Oct 29 10:25:50 2002