Re: Adult peripheral blood

Ierachmiel Daskal wrote:

>I  ( we the Blood banking Community) need your help:
>
>The American Association of Blood Banking and the FDA will try and
>implement by 03/01/04	a set of new regulations concerning the
>determination whether	bags of random human platelets are contaminated
>with bacteria.  Every hospital  in the US will be required to assay
>every unit of platelets prior to transfusion  to determine if there is a
>bacterial contamination present.
>
>This is an enormous task  since to date, there is no reliable
>methodology  ( pH dip sticks, glucose levels, Gram stain...)  to
>determine the bacterial contaminations in the sterile bags received from
>te Red Cross. The method must be a rapid one since the overall shelf
>life of platelets is 5 days only and hence the units cannot be
>quarantined for long time prior to release to patients.
>
>The question to the Flow Community is whether some flow technology  can
>be applied to determine if such contamination is present in a specimen
>obtained in a sterile manner form the unit to be transfused to
>patients..
>For example are there any "pan-bacterial wall" antibodies available that
>one use  , to check a sample from a bag of platelets to determine the
>extent of contamination if any. The major contaminants that we will be
>dealing will be skin flora ( ie Staph. epi. from the donor's skin and
>Gram negative rods).

Question number one: what level of contamination are you expected to be 
able to detect?

The maximum sample (i.e., core) flow rate in a flow cytometer is typically 
1.5 uL/second. Sampling 1 mL of platelet concentrate (an amount not likely 
to be missed from a 150 mL transfer bag), even if you ran it essentially 
undiluted, would take you about 11 minutes. Assuming you had some robust 
combination of parameters that could reliably detect bacteria and 
discriminate them from platelets and from any residual red cells and 
leukocytes, you'd be likely to miss one or two organisms; also, if the 
average number of organisms in 1 mL is one or two, Poisson statistics tells 
us that a substantial fraction of 1 mL samples will contain no organisms. 
You'd probably notice a few dozen organisms/mL and almost definitely catch 
100 or more.

Of course, it is unlikely that you will be able to run undiluted platelets 
in a flow cytometer; if you dilute them even 1:10, it will take you close 
to 2 hours to look at a single 1 mL sample. And maybe you want to detect 
bacteria at the same level a blood culture would; a typical blood culture 
inoculation volume is 10 mL, and the culture can show up positive after 
incubation if there is only a single viable organism in the original 10 mL 
inoculum. If you needed to detect that level of contamination without 
incubation, it would take most of a day to look at a single specimen.

So, flow cytometry on the raw specimen is impractical, even if you've got 
an ideal set of reagents. You will have to concentrate and/or incubate a 
few mL of sample to find what you're looking for. Some combination of 
fluorescent indicators of nucleic acid content (e.g., the various SYTO dyes 
from Molecular Probes) and of metabolism (fluorogenic enzyme substrates, 
membrane potential probes) will give you fairly strong signals and good 
discrimination; although some people have suggested using either antibodies 
or various lectins to detect bacteria, the general applicability of these 
procedures is far from proven, and the signals from fluorescent antibodies 
and lectins will not be as strong as those from nucleic acid dyes and 
metabolic indicators.

It will almost certainly be necessary to use detergent or other chemical 
treatment to get rid of most of the platelet mass in order to analyze the 
sample by flow cytometry; once you do that, it is probably just as easy to 
run the sample through a small-pore black membrane or aluminum oxide filter 
and look for bacteria using a fluorescence microscope or an image analyzer. 
Fluorescence miocroscopes are about to get a lot less expensive with the 
introduction of LED light sources that can effectively replace xenon and 
mercury arc and quartz halogen lamps for most applications, and inexpensive 
image analyzers with LED light sources and CCD detectors, more than up to 
the task, should be available as laboratory instruments (but probably not 
with FDA clearance as clinical devices) in a year or so. I hope to have a 
lot more to say about this in print in the next few months.

For what it's worth, I recall that the folks at Biometric Imaging were 
looking into detection of bacteria in platelet concentrates on their IMAGN 
2000 around the time that company was acquired by Becton-Dickinson. I keep 
hearing that the IMAGN 2000, which was unceremoniously yanked from the 
market some years back despite having acquired a following among blood 
bankers, is being resurrected, but I haven't been able to get details from 
anybody at BD. If they don't get it out soon, its job will be taken by a 
much cheaper, simpler, instrument.

I'd be happy to provide more details.

-Howard