Advanced Cases in Pediatric Fever Without a Source

1
Advanced Cases in Pediatric Fever Without a Source

• Andi Marmor
• June, 2004

2
Key Questions

• What are the risk factors for SBI and UTI in
  febrile infants?
• How effective is the pneumococcal vaccine?
• Partial vaccination
• Technical difficulties when the best laid plans
  go awry
• How do you collect urine?
• Do viruses count as a fever source?

3
Fever Without a Source A Quick Review

• For nearly 20 of febrile children, no source of
  infection can be identified after thorough
  history and physical exam
• A small proportion of these children, although
  well-appearing, will have a serious bacterial
  infection (SBI) or occult urinary tract infection
  (UTI)
• Guidelines have been developed to help physicians
  identify and treat those children at high risk
  for these conditions

4
Age Groups for Estimating Risk of SBI in
Well-Appearing Infants

• Guidelines for management of infants with fever
  without a source are based on groupings of
  infants into 3 age groups based on both their
  risk of SBI/UTI and the most likely bacterial
  causes of SBI
• Neonate (0-28 days)
• Infant 1-3 mo
• Infant 3-36 mo

5
Neonates (lt28 days)

• Causes of SBI/UTI
• E. Coli, GBS, Listeria, Salmonella
• What counts as a fever source?
• Clinical exam is unreliable, and even infants
  with viral symptoms may be at risk for SBI
• Prevalence of SBI in well-appearing infants lt28
  days with Tgt38
• 4-12
• UTI
• Prevalance of UTI is high for boys and girls
• Associated with a 15-20 risk of bacteremia

6
RecommendationsNeonates, T gt38

• CBC, blood cultures
• Cath UA and urine culture
• LP
• Antibiotics
• Ampicillin and gentamicin IV, or ampicillin and
  cefotaxime IM
• Admission

7
Infants 1-3 months of age

• Causes of SBI/UTI
• E. Coli (UTI), GBS, S. pneumonia, N.
  meningitidis, Hib
• What counts as a fever source?
• Named viral syndrome
• Otitis media
• Other viruses?
• Prevalence of UTI in this age group is about 9
  overall, (highest in uncirc boys, but only 2 in
  circumcised boys)

8
Infants 1-3 months of age Predictors of SBI

• Studies in the early 90s established criteria
  for dividing well-appearing febrile infants this
  age into groups at high or low risk for SBI based
  on WBC count
• WBC 5-15 Risk of SBI (NOT including UTI) is
  1-3
• High risk 10-20

9
Recommendations 1-3 months, Tgt38

• Cath urinalysis and urine culture on all infants
• If UA is positive, begin treatment for
  pyelonephritis and consider admission
• CBC and blood culture
• If WBCgt15K, antibiotics (ceftriaxone IM/IV)
• Lumbar puncture
• If signs of CNS irritability, and strongly
  consider if giving antibiotics
• Follow up
• The next day (2nd dose if antibiotics were given)
  
• Admit if unable to follow up

10
Infants 3-36 months, Tgt38.5

• Causes of SBI
• S. pneumoniagtgtgtN. meningitidis, Hib
• Causes of UTI
• E. ColigtgtgtKlebsiella, Proteus, Strep spp
• Risk highest in girls and in uncirc boys up to
  6-12 mo
• Risk for SBIbefore pneumococcal vax
• Overall risk of SBI in these infants estimated
  2-6
• WBC count useful to stratify infants into high
  risk (10) and low risk (1)

11
Hooray for the pneumococcal vaccine!

• 7-valent polysaccharide conjugate vaccine
• Approximately 97 of pneumococcal isolates that
  cause IPD are represented in PCV-7
• Recommended since August, 2000
• 2,4 and 6 months with booster at 12-15 mo

12
Vaccine Efficacy

• PCV-7 tested in a large NC Kaiser-based
  randomized controlled trial of 37,868 children
• Efficacy against IPD from vaccine serotypes
• Fully vaccinated children (4 doses) 97.4
• Those receiving one or more dose of vaccine 94.
  
• Efficacy against IPD from any pneumococcal
  serotype,
• Those receiving one or more doses 89.1

13
Vaccine Efficacy Post-licensure

• Multiple post-licensure studies have supported
  the expected reduction in invasive pneumococcal
  disease (IPD)
• 78-85 drop in rates of IPD in children lt2 years
  of age.
• Rates of disease from non-vaccine serotypes have
  not increased
• However, IPD and SBI are still possible, even in
  vaccinated children.

14
How should vaccine change our management?

• Since IPD is responsible for the majority of SBI
  in infants gt3 months of age
• And the vaccine is at least 90 effective against
  IPD
• The risk of SBI in vaccinated children is lt1,
  regardless of WBC count.
• Therefore, a CBC is unlikely to significantly
  impact the assessment or management of vaccinated
  children.

15
Is this change in management cost-effective?

• Lee et al (2001) conducted a cost-effectiveness
  analysis of various management strategies for
  infants with FWS
• Conclusion empiric CBC/blood cx NOT cost
  effective if rates of SBI lt0.5
• Costs gt300,000 per life saved
• Rates of SBI lt0.5 in vaccinated infants, based
  on current data

16
Recommendations for vaccinated children 3-36 mo
of age

• Is the child effectively immunized?
• At least two doses (3 is better!)
• 2 weeks from 2nd dose
• Screen for UTI as for the unvaccinated child
• Well-appearing, vaccinated children are low risk,
  so blood tests not likely to change management!

17
Case 1

• Rutabaga is a 9 week old male infant with fever
  at home to 103, parents gave Tylenol.
• In clinic, T is 37.6, vitals otherwise normal for
  age, baby is well-appearing
• Exam/hx hint of a cough, mild papular rash
  onchest, feeding well, older sibs with colds
• Received 1st dose of Prevnar 3 weeks ago
• Uncircumcised

18
What are the key parts of Rutabagas Hx/PE in
estimating his risk of SBI/UTI?

• Age 1-3 mo
• Appearance Non-toxic
• Fever source
• Possible viral source? Sick contacts?
• Uncircumcised
• Immunization status
• One dose of PCV-7 is he protected?

19
Partial Vaccination Evidence

• Efficacy of the vaccine after lt 3 doses is
  unclear at the moment due to lack of sufficient
  data.
• Kaiser study results suggest that immunity
  against invasive disease is good in partially
  immunized infants
• Herd immunity protective
• Two recent studies have demonstrated good
  serotype-specific antibody responses after 2
  doses of the vaccine (Goldblatt, 2006 Huebner
  2002)
• Vaccination against pneumococcus DOES NOT protect
  against UTI, primarily caused by E. Coli

20
Whats your plan?

• Cath U/A
• Negative for LE, nitrites, small blood
• CBC
• WBC 18.7, 75 lymphs
• Blood culture
• Cant obtain blood culture after multiple sticks
• What are your options?
• Try again for blood cultures
• Treat without cx commit to full course of
  antibiotics
• No antibiotics, admit for obs
• No antibiotics, home for obs

21
Another version

• In a similar case, you obtain blood cultures, but
  are unable to obtain spinal fluid after 3 tries
• What are your options?
• Treat without tap
• Commit to full course for presumed meningitis
• Try again tomorrow for cell count
• Dont treat
• Admit for obs without tap (plan to tap and treat
  if ill-appearing)

22
Case 2

• Cheyote is 6 month old girl who just received 3rd
  dose of PCV-7 2 days ago
• She has had a fever for 3 days, has a temp of
  39.8 in clinic, no source for fever on exam or
  history, and is well-appearing
• What studies, if any, would you do on this
  infant?
• How do you obtain urine?

23
Bag vs Cath

• Catheter specimens
• Current gold standard
• For culture Sens 95, spec 99
• Bag
• Less invasive (?)
• BUT results difficult to interpret
• Culture Sens/spec 85

24
Can a bag specimen be used for UA?

• Bottom line No published data compares
  sensitivity and specificity of UA on bag
  specimens to other types of specimens!
• UA from bag may have slightly decreased
  specificity compared to cath specimen
• False positives may result from contamination
  from distal urethra, diaper
• Avoid in patients in whom false positives are
  unacceptable

25
Predictive value of UA

• Predictive value refers to the posterior
  probability of disease, given a positive or
  negative test
• Depends on sensitivity, specificity, and prior
  probability
• Example For a UA positive for LE only
• Prior prob PPV NPV
• 5 20 lt1
• 10 33 1
• 20 53 3
• Which patient is most likely to be impacted?

26
Predictive value of UA

• Predictive value refers to the posterior
  probability of disease, given a positive or
  negative test
• Depends on sensitivity, specificity, and prior
  probability
• Example For a UA positive for LE only
• Prior prob PPV NPV
• 5 20 lt1
• 10 33 1
• 20 53 3
• Which patient is most likely to be impacted?

27
Predictive Value The Bottom Line

• PPV is maximized when PP is high
• NPV is maximized when PP is low
• Best use of UA for
• Low prior prob patient Rule OUT UTI
• High prior prob patient start empiric treatment

28
Can a bag specimen be sent for culture?

• False positives are the major concern
• Contamination rate depends on the population,
  technique, and positive threshold
• Very low in circ boys
• As high as 20 in other populations
• However, false negatives also occur, depending on
  the threshold chosen for positive test
• For gt100,000 org, sens and spec 85

29
Predictive value of bag culture

• NPV of bag cx best in low prior prob patient, PPV
  best in high prior prob pt
• Example
• Prior prob PPV NPV
• 5 23 1
• 10 40 2
• 20 60 4
• The only clinically meaningful use of the bag
  culture is to rule OUT UTI in the low prior
  probability patient

30
Predictive value of bag culture

• NPV of bag cx best in low prior prob patient, PPV
  best in high prior prob pt
• Example
• Prior prob PPV NPV
• 5 23 1
• 10 40 2
• 20 60 4
• The only clinically meaningful use of the bag
  culture is to rule OUT UTI in the low prior
  probability patient

31
Summary Bag specimen

• Characteristics of UA from bag specimen make it
  most useful to rule out UTI in low probability
  patients
• Can also be used to start treatment in high risk
  patient
• Bag culture
• False positive/negative results are a significant
  risk
• Neg results helpful in low-prob patients
• Must weigh the implications of false pos/false
  neg for the patient, against the discomfort of a
  cath

32
Recommendations Collection of Urine Specimen

• High risk infants, or a child who looks sick
  enough to require IV antibiotics/admission
• Obtain a catheter specimen for UA and culture
• Positive UA empiric treatment, confirm with
  culture
• Lower risk patients
• If desired, collect bag specimen for screening
  UA
• Negative UA UTI is unlikely
• Positive UA consider empiric treatment, but
  confirm with a culture
• If you send the bag for culture consider the
  clinical implications before you send the test!

33
Case Three

• Daikon is a 6 week old boy, temp of 101 at home,
  38.7 in clinic
• Its winter, influenza and RSV are rampant
• He is well-appearing, without any URI symptoms on
  exam or history, mom says she has had the flu
  and is wondering if he might have the same thing
• No immunizations yet

34
Key Question

• Would viral testing change your management?

35
Viral Testing - Evidence

• The advent of rapid viral testing has added a new
  option for identifying infants at low risk for
  SBI
• Rapid tests exist for RSV, adeno, paraflu,
  influenza, entero and rotaviruses
• In general, these tests are more specific than
  they are sensitive, which makes false positives
  extremely rare

36
Viral Testing - Evidence

• A number of recent studies, mostly retrospective,
  have evaluated the risk of SBI in infants found
  to have a positive viral test
• Example recent prospective trial (Byington, et
  al 2004) of 1385 febrile infants lt90 days, tested
  for multiple viruses
• Stratified infants into HR/LR by Rochester
  criteria
• Among LR infants, risk of SBI low (1-3)
  regardless of viral test
• Among HR infants, those with viral tests had a
  significantly reduced chance of SBI (16.7 -gt
  5.5)
• Risk of UTI still clinically significant in HR
  infants (4), while bacteremia occurred in lt1,
  and none had meningitis

37
Recommendations

• Bottom LineThe negative predictive value of a
  rapid viral test is best in low probability
  patients!
• Therefore, viral testing is most likely to change
  management in those infants with a low-mod prior
  probability of SBI
• In very young infants or those at high risk, an
  appreciable risk of UTI remains
• Consider testing for UTI in infants at high risk
  of UTI, regardless of viral diagnosis

38
Case 3 - Continued

• You decide to get a CBC and blood culture, a cath
  UA and a rapid viral test for RSV and influenza
• Results
• WBC 18, with 67 lymphs
• Rapid viral test positive for influenza
• Cath U/A negative
• What do you want to do?
• Treat with antibiotics? Admit? Tap?

39
Summary of Recommendations

• 5 questions to ask about child with FWS
• 1. Is this child toxic?
• 2. Is there a source for the fever?
• 3. Has this child been vaccinated against
  pneumococcus?
• 4. If its a boy, is he circumcised?
• 5. Will this child come back if he/she gets
  sick?

40
My Silly Mnemonic

• If the babys smiling at me
• And has had Prevnar X 3
• Skip the CBC
• But dont forget to collect the pee!