Atypical Kawasaki Disease: A Diagnostic Dilemma.

1
Atypical Kawasaki DiseaseA Diagnostic Dilemma.

• Amy Giantris
• Marc Richmond
• Christine Wang
• March 26, 2004

2
Validity of Evidence

• 1. Was there an independent, blind comparison
  with a reference (gold) standard of diagnosis?
• 2. Was the diagnostic test evaluated with an
  appropriate set of patients?
• 3. Was the reference standard applied regardless
  of the diagnostic test result?
• 4. Was the test (or cluster of tests) validated
  in a second, independent group of patients?

3
1. Was There an Independent, Blind Comparison
With a Reference Standard of Diagnosis?

• Did the patients in the study undergo the
  diagnostic test and reference standard
  (confirmation of presence/absence of the target
  disorder)?
• Were those applying and interpreting each
  blinded to avoid bias that might cause the
  reference standard to be over-interpreted when
  the diagnostic test is positive and
  under-interpreted when it is negative?
• Was the selection of the reference standard
  justified?

4
2. Was the Diagnostic Test Evaluated in an
Appropriate Set of Patients ?

• Did the patient sample include an appropriate
  spectrum of patients to whom the diagnostic test
  will be applied in clinical practice (i.e. all
  the common presentations of the
  disorder,including those with its early
  manifestations, and patients with other, commonly
  confused diagnoses)?

5
3. Was the Reference Standard Applied Regardless
of the Diagnostic Test Result?

• Did the results of the test being evaluated
  influence the decision to perform the reference
  standard?
• The properties of a diagnostic test will be
  distorted if its result influences whether
  patients undergo confirmation by the reference
  standard (WORK-UP BIAS).
• The reference standard chosen to confirm absence
  of a disorder must not cause a patient to suffer
  any adverse health outcome during a long-term
  follow-up despite the absence of any definitive
  treatment.

6
4. Was the Test (or Cluster of Tests) Validated
in a Second, Independent Group of Patients?

• In order to distinguish between real diagnostic
  accuracy and chance, similar levels of accuracy
  must be obtained in a second, independent (or
  test) set of patients.
• The methods for performing the test must,
  therefore, be described in sufficient detail to
  permit replication.

7
Determining the Significance of Study Results

• Does the evidence demonstrate the ability of the
  diagnostic test to accurately distinguish
  patients who do and dont have the disorder?
• Is the test accurate in distinguishing patients
  with and without the target disorder?
• Does the test have the ability to change our
  minds from what we thought before the test
  (pre-test probability of target disorder) to
  what we think afterward (post-test probability
  of target disorder)?

8
Diagnosis Worksheet

• Sensitivity a/(ac)
• Specificity d/(bd)
• LR sens/(1 spec)
• LR (1 sens)/spec
• Positive predictive value a/(ab)
• Negative predictive value d/(cd)

• Prevalence (ac)/(abcd)
• Pre-test odds prevalence/(1 prevalence)
• Post-test odds pre-test odds likelihood ratio
• Post-test probability post-test odds/(post-test
  odds 1)

9
Definitions

• Sensitivity
• Specificity
• Likelihood ratio
• Prevalence
• Predictive value

10
SnNout and SpPin

• If a test has very high sensitivity, a negative
  result rules out the diagnosis.
• We apply the mnemonic SnNout to such findings
  (when a sign has a high sensitivity, a negative
  result rules out the diagnosis).
• Similarly, when a test has a very high
  specificity, a positive result rules in the
  diagnosis. Such a finding is called SpPin.

11
Application of Test to a Specific Patient

• 1. Is the diagnostic test available, affordable,
  accurate, and precise in our setting?
• 2. Can we make a clinically sensible estimate of
  our patients pre-test probability?
• 3. Will the resulting post-test probabilities
  affect our management and help our patient?

12
1. Is the Diagnostic Test Available,
Affordable, Accurate, and Precise?

• The test should be performed and interpreted in a
  competent, reproducible way and its potential
  consequences justify its cost.
• Sensitivity analysis- some diagnostic tests based
  on symptoms or signs lose power as patients move
  from primary care to secondary and tertiary care.

13
2. Can We Generate a Clinically Sensible Estimate
of Our Patients Pre-test Probability?

• How can we estimate our patients pre-test
  probability?
• Clinical experience with prior patients
• Regional or national prevalence statistics
• Local, regional or national practice databases
• Applying the pre-test probabilities observed in
  the study we critically appraised for the
  accuracy and importance of the diagnostic test-
  Did they really sample full patient spectrum?
• Research report of a study devoted to documenting
  pre-test probabilities for the array of diagnoses
  that present with a specific set of symptoms and
  signs similar to our patient.

14
3. Will the Resulting Post-test Probabilities
Affect Our Management and Help Our Patient?

• Could its results cause us to stop all further
  testing?
• Test threshold
• Treatment threshold

15
Likelihood Ratio

• Fagan Nomogram

16
C.M.5 Year Old Previously Healthy Male

• Present
• Fever for 9 days- no obvious source
• Papular rash- faded by admission
• Non exudative conjuncivitis
• Myalgias

• Absent
• Cervical lymphadenopathy
• Hand/foot swelling, erythema
• Strawberry tongue or any oral changes

Brother with history of presumptive Kawasakis
Disease
17
Our Question

• How do you differentiate between atypical
  (incomplete) Kawasaki disease and a prolonged
  viral syndrome?

18
Our Search

• We each attempted the search.
• We all found Ovid to be the most useful.

19
Christine

• 1. Kawasaki disease--limit to human,
  English- 775 articles
• 2. Sensitivity and specificity- 29,026 articles
• 3. Then combine the above two to get the one
  article

20
Marc

• Mucocutaneous lymph node syndrome- limit to
  diagnosis.
• Keyword atypical.
• Combine above, limit to English- revealed 23
  articles.

21
Amy

• Atypical Kawasaki only 14 articles identified.
  This was one of the 14.

22
The Article

• The Differentiation of Classic Kawasaki Disease,
  Atypical Kawasaki Disease, and Acute Adenoviral
  Infection

23
Brief Study Overview

• Retrospective
• Compared medical records of kids with complete
  and incomplete Kawasaki disease with those of
  children with acute adenoviral infection
• 43 children studied 23 with complete Kawasaki,
  13 with an atypical presentation and 7 with acute
  adenoviral infection

24
Brief Study Overview, cont

• Objective To compare clinical and laboratory
  features of kids with Kawasaki disease with those
  of acute adenoviral infection
• Results
• 1. Kawasaki kids more likely to have
  conjunctivitis, strawberry tongue, perineal
  peeling, and distal extremity changes. They also
  had pyuria, higher mean WBC/ESR/PLT/ALT
• 2. Adeno kids more likely to have purulent
  conjunctivitis and exudative pharyngitis
• 3. A rapid antigen test for adenovirus had a
  specificity and sensitivity of 100 compared with
  viral culturehuh?

25
Brief Study Overview, cont

• Conclusions
• Kawasaki disease and and acute adenovirus
  infection can present with many of the same
  clinical characteristics
• A DFA for adenovirus may be a helpful adjunctive
  test for distinguishing acute adenoviral
  infection from Kawasaki disease

26
Are the Results of This Diagnostic Study Valid?

The four questions that can help quickly appraise
the article for its proximity to the truth.
27
Was There an Independent, Blind Comparison With a
Reference (Gold) Standard of Diagnosis?
28
Was the Diagnostic Test Evaluated in an
Appropriate Spectrum of Patients?
29
Consider.
30
43 patients enrolled
7 with Adenovirus
18 with classic or incomplete Kawasaki
18 with classic Kawasaki
DFA performed
No DFA performed
DFA performed
31
Was a Reference Standard Applied Regardless of
the Diagnostic Test Result?
32
Was the Test Validated in a Second, Independent
Group of Patients?

• Yes, but in previous studies

33
What Are the Results?

• Kawasaki disease vs. Adenoviral infection
• Significant differences in the following
• Conjunctivitis
• Strawberry tongue
• exudative pharyngitis
• perineal rash and desquamation
• Extremity changes
• Platelet count

34
What Are the Results?

• Atypical Kawasaki vs. Adenoviral infection
• Significant differences in the following
• Conjunctivitis
• perineal rash and desquamation
• Platelet count

35
Sample Calculations

• Extremity changes
• Sensitivity61
• Specificity100
• PPV100
• NPV33
• LR()infinity
• LR(-).39

36
Sample Calculations

• DFA vs. viral culture
• Sensitivity100
• Specificity100
• PPV100
• NPV100
• LR()infinity
• LR(-)0

37
Will This Help Me in My Patient Care?

• Is the diagnostic test available, affordable,
  accurate and precise in your setting
• Yes
• Can you generate a clinically sensible estimate
  of your patients pre-test probability
• Yes
• Will the result affect you management and help
  your patient
• Yes

38
However. . .

• Numbers in this study were small
• Only 13 patients had atypical Kawasaki disease
• Only 7 patients with adenoviral disease
• Retrospective study
• Not every patient underwent a DFA
• Cut off values for lab tests werent used
• Focus of article is unclear

39
Therefore. . .

• Although this study describes trends between
  adenoviral infection and Kawasaki disease, the
  data is not sufficient to provide clinically
  useful conclusions.
• This fact is hidden well among p-values and nice
  looking tables.

40
So. . .

• It is apparent that DFA can accurately diagnose
  adenoviral infection. Perhaps a prospective
  study where all patients being considered for a
  diagnosis of Kawasaki disease have a DFA done
  would be useful in describing the prevalence of
  adenoviral infection in this population

41
Now What. . ?

• Given the difficulty in clinical diagnosis, it
  may be wise to consider a DFA in all patients to
  screen for adenoviral infection before treating
  with IVIG for atypical Kawasaki