It's  many years since my lab offered PTH assays (using our own
reagents and assay systems, not kits) as part of its SAS menu, but in
those days we recognized that "PTH" in blood represents a can of
worms containing a miscellaneous collection of different molecules -
intact PTH, fragments, etc. The introduction of 'two site' assays
increased assay specificity, since many  fragments would not react
simultaneously with both antibodies. But this in itself does not
guarantee that the assay system responds only the intact polypeptide,
and is unaffected by fragments (some of which may interact with one
or other of the antibodies used, possibly causing assay bias).

In short even though PTH is an analyte of relatively simple molecular
structure, and a 'right answer' to a measurement of its concentation
in blood is in principle attainable , the 'right answer' will be
obtained only if the assay system responds solely to the intact
molecule. Also two PTH assay systems using different antibodies are
likely to agree only if  both respond only to the intact molecule
even though both rely on the same "pure" PTH preparation as
calibrant.  If they don't, this is good evidence that one or other,
or both, are also responding to fragments.  In other words one or
other or both assays constitute invalid 'analytical assays', and are
therefore, in effect, 'comparative assays' as defined by Finney many
years ago.

Kit manufacturers faced with this sort of problem often cover up by
altering the numbers on the standards so as to bring  reference
ranges more or less into line. (A notorious example of this was the
misrepresentation of the values of FT4  standards by certain
manufacturers)    But this is in principle dangerous because it lulls
users into a false sense of security, concealing the possibility that
the various assays may give different results with certain
pathological samples.

Johnathan Middle is therefore right in asking  "which (of two
methods) is right?" if they yield different results. But I question
his argument that the expression of  results of  measurements in SI
units implies that they " must therefore be traceable back to a
primary standard which contains a certified molar content of a single
defined chemical entity ('PTH') and this entity must be the same as
the single defined chemical entity ('PTH') in all patient samples".

We can express the results of a gravimetric measurement in Kgs  (a
well known SI unit) , but this does not imply that that the
measurement of the weight of a mixture of different substances  (and
the expression of the result in Kgs) is inadmissible.  In other
words, the expression of a result in SI units does not imply that the
measurement is of the amount or concentration of a "single  defined
chemical entity".  JM is, I think, confusing  the issue of the units
used to express assay results with that of whether an assay
constitutes  valid "analytical assay", i.e an assay that responds
only to a single molecularly homogeneous substance present in
standards and unknowns.

In practice, many - probably most - of the proteins measured by
immunoassay do not fit this description, and are of heterogeneous
molecular composition.   HCG is a typical example.  An assay of HCG
in blood is thus a "comparative assay" . It has long been known that
such assays cannot be standardised.  Nor do they have  "right
answers".  But this has nothing to do with the nature of the units in
which results are expressed

Regrettably  organisations such as NIBSC,  and others have - by
distributing international refence preparations whose analyte
"content" is expressed in international units  (claimed as
constituting  a "common currency") - been the originators and
principal perpetuators of the myth that the results of different
comparative assays (on individual samples) should be identical when
expressed in such units, and that if they are not there is something
wrong with the assays or their standardisation.

Such organisations  have stubbornly resisted the implication that the
only way of dealing with this problem (other than by developing
assays for each component of a heterogeneous analyte, using, eg,
microarray chip technology) is by  abandoning the use of either  SI
or international units to express assay results , and to replace them
by  "method dependant" units. Hopefully this would at last destroy
the fiction that the results of different (comparative) immunoassays
should be numerically the same or necessarily possess the same
diagnostic value.




--
Regards

Roger

Molecular Endocrinology
University College London Medical School
London W1N 8AA

Fax +44 20 7580 2737
Phone +44 20 7679 9410

The information in this e-mail is confidential and may be legally
privileged. It is intended for the exclusive attention of the addressee
stated above and should not be copied or disclosed to any other
without the express agreement of the sender.