Bill Bartlett wrote:

> Does anyone have opinion, or hard evidence, as to the utility of prolactin
> measurements in the fit versus pseudo-fit scenario?
>
> Dr WA Bartlett
> Consultant Clinical Scientist
> Clinical Biochemistry
> Birmingham Heartlands Hospital
> Birmingham B9 5SS
>
> Tel. No. 0121 766 6611 Ext 5461.
> Mobile 0374 103338

Our laboratory provides, amongst other things, a routine service to an
adjacent supra regional neurology centre (HPNC) and I have have been providing
this test for a number of years using finger prick blood spots. I am glad to
share experiences, and would be interested if anyone else has adopted an
similar, or different, approach .

The main problem with the test is collecting samples at the right time. The
peak rise occurs after 20-30 minutes but by 60 minutes the prolactin has just
about returned to normal, so it is easy to miss it and record a false
negative, for example in casualty. If the patient is admitted, sometimes
there is a long wait for a fit and often EEG is available to help establish a
diagnosis.These are probably the main reasons it isn't used much in the
hospital setting.

Nevertheless neurologists do find it difficult to be sure of the diagnosis of
epilepsy, a label which can have serious medical and social implications for
the patient. Rather like hypoglycaemic attacks, it is sometimes hard to be
sure about a fit which only has been reported when the patient is away from
clinical observation. For the last four years we have offered the test as a
kit method to enable the patient to collect samples at the time of a fit. They
(or their partners) collect on card a baseline blood spot, one 20 minutes
after the start of the attack and another after at least 60 minutes (which
should confirm the baseline result). The prolactin is eluted and measured. The
test seems reliable and the analyses are quite simple and reproducible, if a
little time consuming. They do depend on having an analyzer that can cope with
the dilute eluents, We currently use a TOSOH AIA21 (Eurogenetics).

We did audit the test with the neurologists after a couple of years, but since
then have not actively promoted the test. My opinion is that this is useful,
and perhaps as importantly so do our neurologists, (or presumably they would
not keep taking the trouble to contact me asking for kits to be sent to their
patients).

I currently send out a few kits per month. However only a minority are
returned by the patients with blood spots for analysis, sometimes after
considerable delay. How keen the patient is to do the test, of course, may
also be informative to the clinician in some cases.

As far as hard evidence goes, before the audit I pulled together a
bibliography of about 50 papers, which show a high degree of agreement, so
there is little doubt that the effect is genuine. It is distinct from
non-specific stress effect which also of course can raise prolactin, as well
as other hormones, but which doesn't cause the typical rapid rise and fall.
(I am sure Medline or another web search would produce a similar list of
references.) Unfortunately most are small studies involving a handful of
patients which would not meet the current definition of hard evidence.
Furthermore, they almost all use different cut-offs making generalisations
about sensitivity or specificity difficult. The cut-offs can be either hard,
such as a numerical cut-off or a percentage increase over baseline, or soft
(e.g. mean +/- x sds). The hard cut-offs are more practical for routine use.
There are remarkably few reported cases of normals or pseudo seizure exceeding
any reasonable cut-off level, while most false negatives have been just below
a high cut-off (e.g. 800 or 1000 mU/L). We adopted a cut-off of a rise of
2.5x the baseline (or 60 min+) and a rise at 20 min to at least 500 mU/L. This
allows for a grey borderline area when only one criterion is met. My feeling
is that there is a gradation of degree of rise depending on the type and
severity of the epilepsy. Unfortunately it would be difficult to get enough
cases to produce a definitive study.

Apart from patient compliance, the main worry with a negative result is that
the timing of the peak sample may be wrong. A 'classic' steeple type response
is pretty convincing, but a negative result needs to be treated cautiously
because of the timing difficulty.

Nevertheless a negative finding can be useful to the physician. To give one
anecdote, one of our neurologist felt a certains patient's epilepsy was not
genuine but the patient was threatening medical legal action if he withheld
antiepileptics. The finger prick test showed an absence of prolactin rise,
'objective' support of his clinical diagnosis for which he was no doubt
grateful.

So overall, my vote is - not a perfect test but it can help.

regards

Steve


Dr Stephen Frost
Principal Biochemist
The Princess Royal Hospital
Haywards Heath
West Sussex
UK

[log in to unmask]






%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%