On Jan 20, 3:39am, GIBB at snuffy.fhhosp.ab.ca wrote:
> Subject: RE: RSV testing: PCR Overkill?
> Kevin O'Donnel asks:
>> > Are there many cases where PCR is not overkill? Or to put the
> > question another way, aside from screening for mutations, for which
> > applications in the clinical or plant pathology fields is PCR now the
> > method of choice?
>> This is a very good question. Here is my personal answer - and I'd be
> interested in any other views.
Let me add some views as a clinical microbiologist and someone who does PCR.
>> There does seem to be a feeling in some quarters that PCR will replace all
> current diagnostic microbiology. That could be true, but I do remember
> promises being made for antigen detection kits about 10 years ago. They
> going to convert all of microbiology into bedside tests - it was just a
> of making the right monoclonals.
One advantage of PCR over antigen detection is that once specimen preparation
is optimized then the procedure becomes organism independent. That is, any
primer pair will have some temperature optimum and it will work. With
antibodies every new antibody has its own quirks (some can immunoprecipitate,
some Western, some can be used in fluorescence, etc.). Primers pairs can also
be made by anyone once the sequence is known.
>> Another impresion I come across is that any lab which cannot do PCR at the
> of a hat must be living in the dark ages. In my opinion there are just
> cautious, and not jumping on the bandwagon.
There is certainly a degree of prestige with doing the latest hi-tech thing,
but, I would venture that more problems crop up with underexperienced labs
attempting procedures beyond their technical skill. Everyone doesn't need to
do it and in terms of allocation of health care dollars, everyone shouldn't.
>> PCR does have a place now for diagnosis and management of hepatitis C.
>> It probably has a place in diagnosis of TB, though it is not yet clear
> where. It is not sensitive enough to screen out negative specimen prior to
> culture, and does not (yet) give sensitivity results. It may be used as a
> quick test to confirm that a smear-positive specimen does indeed contain
> Mycobacterium tuberculosis, but that is an expensive approach and not
> much faster than "conventional" Bactec culture followed by probe-based
Take a look at some of the non-PCR nucleic based procedures. SDA is starting
to look good (granted, not much clinical data yet) and could provide good
clinical info. This issue is complicated because we have a tendency to
evaluate clinical tests by how they improve the lab's productivity (cheaper
test, faster test, simpler test, etc.). It will take longer (perhaps not much
with pressure from hospitals themselves) to evalaute the overall savings in
total health care dollars from faster lab diagnoses (I'm thinking about TB
verses non TB mycobacterial infections, for example).
> It is useful in pathology for dealing with fixed tissue.
This is still virgin area to be tapped.
>> There are some other specific organisms that are hard to recover by
> conventional methods, and may be detected better by PCR, but I don't think
> PCR is established as the best diagnostic method for any of these.
Until PCR kits are common, it is unlikely any will be established as "best."
The current "homebrew" nature of PCR means that average success rate is held
back by alot of inferior procedures.
> In many
> cases these organisms are not common, or not commonly looked for.
True, but in many instances those organisms can be the cause of alot of
health dollars in total. In addition, antigen detection kits may not be cost
effective because of a limited shelf life and infrequent use. In addition, I
think with multiple PCR procedures in place, the technical skill of the lab
will be better maintained and easier to maintain.
>> There are some "high-volume" examples where PCR (and other amplification
> methods) are being investigated annd promoted by large companies to replace
> current technologies. Chlamydia and gonococcus testing come to mind. I
> think that amplification technology offers a huge advantage in these cases,
> though there are some benefits. (The biggest benefit will perhaps be to
> company that comes out with a really good system - or is that being too
> cynical?) Benefits include the ability to test urine rather than cervical
> swabs for Chlamydia, for example.
Testing urine also means that less expense is generated in sample procurement
and opens the possibility of more frequent testing, particularly in non
>> There are some disadvantages to PCR over culture for any of these
> Lack of antibiotic sensitivity tests is one, and lack of a whole genome for
> PFGE etc is another.
Some sensitivity issues may be resolvable by PCR (+/- for resistance gene),
but certainly not all.
> These are potentially soluble problems. The basic
> problems of false-positive results has not yet been answered however. The
> automated systems may well get round it, but for any "home-made" PCR system
> there is a very real potential problem. There is a very good paper on
> point in J Clin Micro (Feb 1994, vol 32, p 277-84) showing that some
> well-organised labs had large numbers of false positives in doing PCR for
Standardized kits will go a long way to helping this. In addition, we are
still on the steep part of the learning curve.
> No diagnostic microbiology lab which hopes to survive can afford to ignore
> etc. We need to study it, and we need to get some expertise. But I don't
> think we need feel obliged to actually do much of it right now, with the
> exception of Hep C.
We have pressure for other tests from clinicians. CSF Toxo in AIDS patients
is one. CSF HSV testing looks pretty good as well. Here at UVa, we also do
Bordatella PCR instead of DFA for confirmation (I know, not common).
Beyond PCR for diagnostic purposes, there are other issues. One is that I
think there are applications for DNA based assessments for management in
addition to diagnosis (quantitation of Hep C, for example).
I also firmly believe (and I'm sure many will take issue with me), that at
some point in the future, taxomony will be based on nucleic acid sequence and
not arbitrary biochemical reactions. Perhaps one of the better outcomes of
the genome sequencing project, will be the development of rapid, self
contained, and most importantly, cheap, automated sequencers. I expect long
before I am retired, species identification will be based on ribosomal RNA