From: sbharris@ix.netcom.com(Steven B. Harris)
Newsgroups: sci.med
Subject: Re: Low MCV on Sysmex SE-9000
Date: 1 Apr 1999 08:03:12 GMT

In <01be7c02$dbd809e0$6e222fca@default> <dburns@gil.com.au> writes:

>A Sysmex SE-9000 has recently been installed into the laboratory
>(metropolitan public hospital laboratory).
>
>When the MCV of approx. 30 samples are compared with those of the Coulter
>STKS the averall average MCV of the two analysers is quite close i.e. 0.3
>fl difference between the two analysers. We are happy with this since the
>STKS will remain as a back-up analyser.
>
>When any of the individual samples are compared, however, there is often
>2 - 3 fl difference between the analysers. Not unexpected considering
>different analysers are involved.
>
>The problem lies with the fact that some patient samples have MCV which
>are up to 10 fl lower on the Sysmex than on the Coulter STKS. This is
>quite reproducible on each analyser. The problem is detected by high MCHC
>produced by the Sysmex.The Coulter gives normal MCHC on these samples.
>Blood films do not show spherocytes as a cause of the raised MCHC. These
>samples are not turbid from very high white cell counts or lipaemia, nor
>are cold agglutinins causing the high MCHC.
>
>Overall, there is a very good correlation between the two analysers. The
>red cell counts, haemoglobin and MCV show a very good correlation. This
>is not a calibration problem. The problem is confined to selected
>patients, several of whom have low serum sodium results, but this is an
>inconsistent finding. Neither sodium, potassium, chloride, urea or
>glucose (all of which have been shown to influence MCV) are consistently
>abnormal in these patients. No consistent patient history has been found.
>
>I encountered the same problem in another laboratory using a Sysmex
>NE-8000 and a Coulter T540.
>
>A significant finding is that spun packed cell volumes compare well with
>Coulter results. The Sysmex Hct and MCV are too low compared to spun
>PCV's.
>
>Since the electrolyte concentrations of Isoton III and Cellpack are quite
>different, I have tried diluting these samples in Cellpack for up to 10
>minutes prior to repeating the analysis on the Sysmex. This has been
>shown to influence MCV on some analysers. Unfortunately this has had no
>effect.
>
>I am seeking a practical approach to this problem. I would like to
>resolve the problem by using the Sysmex SE-9000, rather than running all
>these samples on the backup instrument.
>
>Thank you for any assistance you are able to offer.
>
>Dominic Burns.
>
>dburns@gil.com.au



    Interesting.  As I understand it, a counter type blood analyser
works by measuring the change in capacitance when cells pass one by one
through a capillary tube.  You count them as they go past, and their
size (volume = mcv) is measured directly by the capacitance change.
Another sensor looks at total blood hemoglobin content
spectrophotometrically, and all other values are calculated from these
numbers.

   What you're saying is some patients have RBCs which the SE-9000
counts properly but simply says are too small (they don't induce enough
capacitance change).  Multiplying these numbers gives the packed cell
volume (hematocrit), and if this, when measured by simply spinning a
sample, agrees with the Coulter but not the SE-9000 machine, and they
both give the same RBC count, that pretty much nails the problem as an
SE-9000 RBC capacitance measurement error (MCV reading error) for some
patient's RBCs.  If this shows up as a high mean cell hemoglobin
concentration (and it is proportionately high), that only means that
the two machines have hemoglobin sensors that are opperating correctly.
They obviously give the same gross blood hemoglobin content numbers
(Hb), and this, when divided by the RBC number and the falsely low MCV
of the SE-9000, gives falsely high MCHC's.  The machine, with the same
number of cells and the same amount of hemoglobin, naturally calculates
with a falsely low cell size that each cell's normal amount of
hemoglobin (MCH, which should also be normal if your SE-9000's counter
and hemoglobin sensors are okay) is packed too tightly into those cells
which it thinks are fasely small.

   If the problem really is reproducable between patients, then there
must be some difference in the way the two machines measure MCV, which
produces an error for some particular patients.  But they're both okay
on RBC count and hemoglobin spectroscopy.  You need to find out exactly
how they operate with regard to measuring MCV.  Perhaps my assumption
that they all use capacitance is wrong.  Whatever the problem, if it's
reproducable for certain patients, it's unlikely to be a particular
machine malfunction, but rather a generic problem with the SE-9000
design.  You can, of course, check this by having some samples from
some of these patients sent out to somebody who also has one of these
machines.   If they come back normal, your machine is bad in a very,
very odd way, and the company should replace it.  If not, the entire
design is bad, and the company is going to be very very very unhappy
with you for finding this out.

   Fascinating problem, tho.