In article <3ti593$sg0 at lastactionhero.rs.itd.umich.edu>,
jperrin at umich.edu (John Perrin) wrote:
>I am posting this question for anyone that might do testing for Borrellia
>antibodies in spinal fluid. 1.) How do you process csf for western
>blot? 2.) Are your samples diluted for testing as serum specimens are or
>are they run without dilutions? 3.) What types of interpretations do
>you use? i.e. If you obtain an equivicol result on a non diluted specimen
>is that clinically significant?
>Please respond to either this posting or e-mail any answers you might
>have directly to me at jperrin at umich.edu> Thanks in advance.
>There has been a variety of methods used and none are well accepted. One
approach is to measure Ig concentrations in serum and CSF and adjust values
based on a ratio. Different ratios have been used. Others simply test
undiluted or moderately diluted CSF. Since the concentrations of the various
proteins in the western are not well standardized and there is some variation
of band locations seen across the blot, the CSF concentrations used are
probably not that important.
I work in a commercial diagnostic company that does not make westerns,
although we use the technique for development. If I were to design a
commercial method, we would need to show respectable specificity (>95%) first
and then test sensitivity. For borrelia, you also need to test CSF from
patients with symptoms similar to borreliosis and documented not to have
borreliosis. Using samples from asymptomatic subjects don't really predict
specificity. Typically the FDA requires 50-200 such samples be tested. You
would probably test at several CSF dilutions or you could vary borrelia
antigen concentrations in the blot. After you have satisfactory specificity,
then you can see how sensitive the test is. That's assuming you have enough
defined samples to estimate sensitivity. Obviously, you also must define the
interpretation criteria used in the western and keep it consistent.
You may also wish to include known false positive type sera. For borrelia, the
best recognized "problem sera" usually affect IgM blots. These are EBV,
parvovirus, and rheumatological serology positives. In addition, other
spirochetal infections (syphilis, relapsing fever, etc) cause significant
cross-reactivity. We have also observed that H. pylori serology positive
samples signficantly cross-react with flagellin.
One reference I might suggest are a series of articles by Paul Fawcett at the
DuPont Institute or a recent article in J. Clin. Micro by Russ Johnson from
U. of Minnesota. They has addressed many of the above issues for western
blots, but don't say much about CSF. Since neuroborreliosis is more
common in Europe, there are also several articles from Europe discussing the
CSF issue. However, none offer a clear-cut method that can be applied without
research.
Bryan Kiehl
GenBio
15222-A Avenue of Science
San Diego, CA 92128
(619) 592-9300, ext 309
(619) 592-9400 [FAX]
