We need to be a bit careful that we don't combine largely benign
pneumococcal bacteraemia with entirely vicious pneumococcal meningitis
in this discussion. Since we have had HiB vaccination, and more
recently have broken the back of the Meningococcus type B epidemic in NZ
we are seeing much lower numbers of children with bacterial meningitis.
For the lucky country (ies) with routine pneumococcal vaccination this
will be even less common.
I think that there is a risk that we may stop considering the
possibility of this disease as carefully as we have done in the past. I
also think that we have more of an aversion to performing diagnostic LPs
than really necessary; as a procedure it is pretty simple and if done
with appropriate analgesia etc probably less noxious than placing an NG
tube or performing a urethral catheterisation.
Regardless of the above I would virtually never perform an LP on a well
looking child with no meningism and yes I have been caught out badly by
a child with pneumococcal meningitis which I was late to recognise. I
see there was another case in the Sydney paper this morning. I don't
think WCC or CRP is the solution. PEM is not an exact science and it is
pretty hard to really know what I would have done on the second
presentation since the outcome is now known and I wasn't there at the
time. Most likely I would not have done the LP.
Cheers and best wishes for the new year to all.
Starship Children's ED
From: Pediatric Emergency Medicine Discussion List
[mailto:[log in to unmask]] On Behalf Of David Herd
Sent: Friday, 21 December 2007 19:54
To: [log in to unmask]
Subject: Re: fever and antibiotics
Ed, (started jotting a short note but got out of control...)
This is a rare case and unfortunately the typical case that will lead
many people to over investigate. This discussion could trigger a rash
of LPs around the world. I have seen a few cases of pneumococcal
meningitis who went on to develop holes in their brain and always
think of these children as I send someone home with no investigations
(which I do far more often than not). I personally think pneumococcus
is worse than meningococcal meningitis and your case highlights a few
reasons. First they often do not look unwell and can have misleading
symptoms including diarrhoea. They are flushed and tachycardic (hot
shock) and often respond well to fluids which can be falsely
reassuring. Third, many (most?) will get better without treatment.
Fourth, when they decide to get worse they often do so dramatically
with serious outcomes. Finally I think pneumococcus is the one
disease in Paeds ED where CRP has some utility. Provided the illness
is greater than 24 hours duration the presence of systemic
pneumococcus invariably gives rise to a CRP > 100. After all CRP was
initially isolated as a response to the c-polysaccharide on
pneumococcus. A procalcitonin would have been elevated and has an
advantage over CRP in that it only takes 6 hours to be elevated. It
would not have been useful in this scenario. A white cell count would
have been elevated but a normal count would not exclude the illness.
These pneumococcal children still look "well" even to an experienced
clinician. If they look unwell the CRP will be > 300 - but then we
would be already treating. With the CRP at 250 I would have
investigated (mostly clinically) for deep abscess, spine
osteomyelitis, disciitis and then done an LP. Weird and wonderfuls
need to be considered (travel history, pets, environmental exposure,
infective contacts, metabolic, haematologic). We also have invasive
staphylococcus in NZ. Odds are that something bacterial was likely
I like the term number needed to investigate which you could define
as the number of positive or negative tests that will lead to
management change relative to the number of inconclusive tests. This
is similar to pre-test probability but sort of implies that the test
will either diagnose or exclude the disease. A test that does not
alter your management should not be ordered. The sensitivity,
specificity and subsequent likelihood ratios must be pretty good for
LP microscopy (gram stain positive) with a few more positives showing
on culture. Only PCR on CSF would be more specific. A number of false
positives will caused by viral and sterile meningitis. A negative CSF
microscopy is also very good making the test sensitive as well. False
'negative' microscopy (lack of CSF pleocytosis) can occur but
generally the child will either be well or overwhelmed. Another
related term would also be number needed to harm by investigation.
How much pain and anxiety do we induce by performing unnecessary tests?
How many LPs would need to be done on similar children to her at her
first visit to find a pathogen (that needed treatment). I would guess
a thousand. Probably higher if she had the 7-serovalent pneumococcus
vaccine (that covers 80% of the invasive disease). Then the NNI might
be 1 in 5000. Therefore I would not have LP'd on visit one. At day
10 when she returned to ED on ceclor you would probably have to LP
maybe 50 children to find the one with pneumococcus. The sensitivity
and specificity of the clinical exam to exclude meningitis has not
been rigourously tested but obviously experienced clinicians are
better at picking this up than training doctors - usually because
they have missed it when they were training. The more astute
clinician would also say something like 'bah-humbug' when questioned
about the evidence base of their assessment. Having said that I
would have to apply my own instant gestalt assessment (my 'blink')
also known as intuition or sixth-sense or 'the art of medicine'. I
would have discussed my impressions with the family and let them know
the options and risks. I may have LP'd on the second visit depending
on the clinical assessment and the families wishes. If we elected not
to LP then we could assume 98% don't have the disease. Of the 2% with
pneumococcus some will get better (in blood cultures taken in ED up
to 95% resolve). So even if she had pneumococcus the chance of her
developing further problems is 0.02x0.05 or 1 in a thousand. If I see
100 of these children a year I will miss one every 10 years, probably
still more than I would like to see. On her third visit to the ED she
would have earned an LP regardless of the clinical assessment as
outlined above. Number needed to investigate must have been less than
10 at this point.
I don't think this is easy at all. Anyone who claims it is straight
forward in my opinion lacks insight.
Merry Christmas. Thanks for the case.
David Herd BSc FRACP
Paediatric Emergency Physician
Auckland, New Zealand
On 20/12/2007, at 3:41 PM, Ed Oakley wrote:
> Thanks for all the replies to date
> Obviously there was a reason for posting this case. I described
> the first presentation of this child and on this occasion she had a
> UA and culture - both of which turned out to be normal.
> Her fever persisted and she returned to her family doctor 2 days
> later and was put on ceclor - apparently was still well with no focus.
> A further 4 days later she returned to the ED with ongoing fever -
> she was still very well and had no focus (now 10 days into the
> illness). At no time was there any sign for kawasaki's disease
> other than the fever. At this visit the antibiotic were ceased,
> she was observed for 12 hours and sent home.
> Obviously I would ask what you would do at this point. In the
> interest of not causing further chest pain to those reading I will
> finish the story.
> She returned a further 3 days later with ongoing fever (now off a/b
> for 4 days). she was occasionally complaining of not feeling well
> but looked well and was eating junk food in the ED
> FBE - WCC 15, CRP >250 other bloods ok, bl cult -ve
> Admitted for observation
> LP done 12 hours later (as she had deteriorated on the ward)
> confirmed pneumococcal meningitis. Child became very unwell - ICU
> stay needed - but eventually got home.
> This has raised for us the question of when to investigate for
> partially treated meningitis, when to investigate persistent fever
> in kids on antibiotics.
> Of the 24 replies - to the list and directly to me - only one would
> have tapped her at the first visit. I agree this is a common
> presentation but should we have tapped on the second visit??
> Thanks again in anticipation of your replies to help us determine
> what to do in future such cases
> Ed Oakley
> Paediatric Emergency Physician
> Department of Emergency Medicine
> Royal Children's Hospital,
> Flemington Rd
> Parkville Victoria 3052
> Tel: +61 3 9345-6592
> Fax: +61 3 9345-5938
> email: [log in to unmask]
> 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:
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: