Regarding your review of procalcitonin applied in paediatrics, we share y=
our concern over early
detection of invasive bacterial infectious disease in young infants, as a=
common problem in our
every-day emergency room routine.
A year ago, drawn by the favorable reviews of procalcitonin use in adults=
as an accurate marker
of disease, we started our own line of investigation on the subject, in c=
hildren at the ED. The
main objective was to evaluate the usefulness of procalcitonin in the ear=
ly diagnosis of
bacterial invasive infections in children under 2 years of age. The resul=
ts have been very
satisfactory, and are pending publishment.
Sensitivity obtained by procalcitonin (cutoff 0=925 ng/ml) has been aroun=
d 90% (versus 83=923% for
C-reactive protein, CRP), being both specificity and predictive values hi=
gher than those por CRP
(cutoff at 20 mg/L). In infants with less than 12 hours of fever, sensiti=
vity was also higher for
procalcitonin (80% vs 60%), showing it=92s value as an earlier rising bio=
Specificity was 100%. Note that our cutoff values are rather lower than t=
hose used by other
authors (Gendrel for one, and Hatherill also, with cutoff values at 2 ng/=
ml), and application has
been used on a fairly wide sample of infants that consult on our emergenc=
For the time being we are not applying this determination on a routine ba=
sis at the paediatric
emergency ward, even though we do believe in it=92s potential as a marker=
for severe disease in the
infant, with clear advantages over CRP. We shall continue our study in in=
fants, and in our
country a multicentric investigation project is taking shape so as to rec=
ord a broader spectrum
of cases to further prove the application and importance of procalcitonin=
in the febrile infant
and throughout paediatric age.
Dr. Juan Jos=E9 Garcia
Hospital Sant Joan de D=E9u
Jay Pershad escribi=F3:
> Dr Fisher:
> Did read your commentary and the French paper (Gendrel et al). Looks
> promising indeed. BTW, you have a very educational site and I enjoy rea=
> your analysis.
> W.r.t to the paper, I had 3 concerns:
> 1. As you stated, to be really useful, any test has to be applied to th=
> relevant study population (ED), i.e. when they present with fever and a=
> relatively well looking. Then, one would like to have a screening test =
> bacteremia/potential SBI, with high sensitivity and good PPV (those who=
> the disease amongst those with a positive test) . This would help decid=
> which ones need antibiotics.
> There is significant "sampling bias" in this study population. It inclu=
> all patients who were febrile and were sick enough to warrant admission=
> deemed by the ED staff. Therefore, patients would be more likely to hav=
> abnormal test because they are sicker (shock, meningitis etc) This woul=
> tend to overestimate the sensitivity.
> 2. Their PPV was calculated based on a prevalence of bacterial infectio=
> 25% and 50%. In reality, this prevalence is lower, hence most screening
> tests for OB/SBI tend to have lower PPV's.
> 3. Their age range was broad.(1 month to 15 years). I would have prefer=
> if they had provided the ROC curves for the younger age group who are =
> risk for occult bacterial infections..i.e. where clinical exam is not
> discriminatory. Personally, I expect a 7 yr. old say to provide clinica=
> clues of invasive or non invasive bacterial infection versus viral
> illnesses. I really don't need a lab. test in that age group to make th=
> I liked many things about the paper:
> 1. Their attempt at confirming viral illnesses with PCR for enteroviral
> disease etc.
> 2. What looked really promising was that, in non infectious inflammator=
> conditions, like JRA, KD etc the level of < 2 was discriminatory. This =
> unlike many other markers that we have.
> 3. PCT is relatively easy and inexpensive to assay.
> 4. It has shown promise in sepsis states from the adult literature.
> Jay Pershad
> For more information, send mail to [log in to unmask] with the =
message: info PED-EM-L
> The URL for the PED-EM-L Web Page is:
For more information, send mail to [log in to unmask] with the message: info PED-EM-L
The URL for the PED-EM-L Web Page is: