|
ANDROGEN ASSAYS
Wheeler, M.
Department of Chemical Pathology, St Thomas' Hospital, Guy's and
St Thomas' NHS Foundation Trust, London, SE1 7EH
In 2003, in an editorial in clinical chemistry, Herold and Fitzgerald
suggested that testosterone assays could be considered
random number generators. Admittedly they were considering
the use of the assays in the investigation of female
pathologies, but are the assays any better at the
higher concentration present in men? Questions that
one might ask include: 1. Do testosterone assays have
adequate precision and specificity to give the clinician
confidence in the results they receive for their patients.
2. How good is the agreement between different assays?
If a patient's sample is sent to a laboratory using
method A will there be a similar result from another
laboratory using method 8. 3. What factors influence
the result in terms of a) drugs b) concentration of
proteins and other constituents in the sample c) what
sample tube is used and d) when the sample is taken.
4. Should we measuring free testosterone as well as,
or instead of, testosterone? If so is an androgen
index (testosterone/SHBG) as good as free testosterone
determined in another way. This talk will address
these questions. Recent data from an evaluation of
automated direct testosterone methods, carried out
for the Medical and Health products Regulatory Agency
in the UK gives rise to much concern. In the female
testosterone range, not only was there a two-fold
difference between some methods but there could also
be a two-fold difference in the results by users of
the same method. Even at male testosterone concentrations
the precision of some methods was poor with a 3.5
mmol/L difference across the methods. Results are
affected by the concentration of sex hormone binding
globulin with the recovery being higher from male
serum. We are very much in the era of the black box,
the technologist being unable to modify the assay.
It is important that each method is evaluated before
it is introduced in the laboratory. Evaluation of
an oestradiol method from one manufacture showed that
the matrix of the calibrator was totally inappropriate
for human serum. Technologists have been concerned
about the possibility that the separating gel present
in SST blood tubes might affect results by either
absorbing small molecules or contributing interfering
substances into the serum. There are few well conducted
studies and some reports are conflicting. Recently
it has been shown that some Becton Dickinson tubes
affect some assays, including a testosterone assay.
The clinician can contribute to the confusion by not
appreciating the very significant circadian rhythm
of testosterone concentration present in young and
older men. So with all these limitations of total
testosterone assays, would free testosterone be a
more accurate indicator of androgen status. Our data
suggests it is not a good indicator in men with a
large overlap between normal men with primary hypogonadism.
If free testosterone concentration is determined,
Vermeulen suggests that a result that is mathematically
calculated from the testosterone and SHBG is better
than an androgen index. Overall data suggests testosterone
results should be interpreted with much caution. Perhaps
using the patient as a bioassay in as good as current
direct immunoassays. The introduction of methods using
tandem mass spectrometry may provide testosterone
assays that are specific and free from interference.
However both the instruments and the engineering support
has to undergo a lot of improvement before these become
suitable for the routine measurement of large numbers
of samples.
|
