Pages Navigation Menu

Meet 'em, greet 'em, treat 'em and street 'em

SGEM#166: Which febrile child with sickle cell disease should get a chest x-ray?

SGEM#166: Which febrile child with sickle cell disease should get a chest x-ray?

Podcast Link: SGEM166

Date: November 21st, 2016

Reference: Eisenbrown et al. Which Febrile Children with Sickle Cell Disease Need a Chest X-Ray? AEM November 2016

Guest Skeptic: Dr. Corey Heitz is an associate professor of emergency medicine at the Virginia Tech Carilion School of Medicine in Roanoke Virginia. He is also the CME editor for Academic Emergency Medicine and the associate editor for emergency medicine simulation at the AAEM MedEdPORTAL.

Case: You are working in the Emergency Department on an overnight pediatric coverage shift. A worried mother brings her 2-year-old child in with a fever of 38.6C (that’s 101.5F). The female child’s medical history is significant for sickle cell disease. On exam, the child is uncomfortable appearing, tachycardic, tachypnic and febrile. Mom says the child has had a runny nose and a mild cough along with the fever.

Background: Children with sickle cell disease who develop fever are at higher risk of severe bacterial infection than children without sickle cell disease. The National Heart, Lung, and Blood Institute (NHLBI) suggest a routine workup that includes a CBC, blood cultures, and empiric antibiotics (NHLBI Expert Panel Report 2014).

One of the life-threatening infections for which these children are most at risk is acute chest syndrome (ACS). The NHLBI recommend a chest x-ray (CXR) for children with respiratory signs or symptoms (shortness of breath, tachypnea, cough, and/or rales).

Controversy exists as to whether the history and physical exam are sensitive enough to determine which febrile children need a CXR.

Clinical Question: Which febrile children with sickle cell disease presenting to the emergency department should get a CXR to help diagnose acute chest syndrome?

Reference: Eisenbrown et al. Which Febrile Children with Sickle Cell Disease Need a Chest X-Ray? AEM November 2016. 

  • Population: Children age three months to 21 years with sickle cell disease presenting to one of two children’s emergency departments with a fever of 38.4C or greater.
  • Intervention: Accuracy of white blood cell count, history and physical exam findings to rule-in or rule-out acute chest syndrome.
  • Comparison: None
  • Outcome:
    • Primary Outcome: Presence of acute chest syndrome.
    • Secondary Outcomes: Classification and regression tree (CART) analysis, sensitivity, specificity, positive and negative likelihood ratios of constellations of WBC, history/physical exam findings for acute chest syndrome.

The SGEM HOP episodes always have one of the authors on the show. Dr. David Brousseau is a Pediatric Emergency medicine physician at the Children’s Hospital of Wisconsin.

Authors Conclusions: Children with SCD presenting to the ED with fever and shortness of breath, tachypnea, cough, rales, or chest pain should receive a CXR due to high ACS rates. A higher WBC count or history of ACS in a child without one of those symptoms may suggest the need for a CXR. Prospective validation of these criteria is needed.

checklistQuality Checklist for Clinical Decision Tools:

  1. The study population included or focused on those in the ED. Yes
  2. The patients were representative of those with the problem. Yes
  3. All important predictor variables and outcomes were explicitly specified. Yes
  4. This is a prospective, multicenter study including a broad spectrum of patients and clinicians (level II). No, this was a retrospective chart review at two children’s hospitals
  5. Clinicians interpret individual predictor variables and score the clinical decision rule reliably and accurately. Yes
  6. This is an impact analysis of a previously validated CDR (level I). No
  7. For Level I studies, impact on clinician behavior and patient-centric outcomes is reported. Not applicable
  8. The follow-up was sufficiently long and complete Yes
  9. The effect was large enough and precise enough to be clinically significant. Yes

checklistQuality Checklist for A Chart Review:

  1. Abstract Training – Were the abstractors trained before the data collection? Yes, but unclear how
  2. Case Selection Criteria –Were the inclusion and exclusion criteria for case selection defined? Yes
  3. Variable Definition – Were the variables defined? Yes
  4. Abstraction Forms – Did the abstractors use data abstraction forms? Yes
  5. Performance Monitored – Was the abstractors’ performance monitored? Yes
  6. Binding to Hypothesis– Were the abstractors aware of the hypothesis/study objectives? No
  7. Inter Rater Reliability (IRR) Mentioned – Was the interobserver reliability discussed? Yes, for the identification of fever
  8. IRR Tested – Was the interobserver reliability tested or measured? Yes, for the identification of fever
  9. Medical Record Identified – Was the medical record database identified or described? Yes, if by that you mean the population (2 hospitals, Jan 1 2010-Dec 31 2012, detailed inclusion criteria)
  10. Sampling Method – Was the method of sampling described? Yes, assuming this means the description of the sample size calculation.
  11. Missing Data Management Plan – Was the statistical management of missing data described? No
  12. Institutional Review Board Approved – Was the study approved by the institutional or ethics review board? Yes

Key Results: There were 1,837 febrile emergency department visits made by 697 children with sickle cell over two years. The median age was 3.5 years and it was a 50/50 male/female

Primary Outcome: 10% (185/1,837) of the febrile sickle cell children presenting to the emergency department met acute chest syndrome criteria.

  • Secondary Outcomes: CART Model
    • Using NHLBI guidelines alone, 27 cases of ACS would have been missed if no CXR was done but avoided CXRs in 45% (825/1,838) of children
    • Using NHLBI or CP, 23 cases of ACS would have been missed (3%), increased sensitivity to 88% and would avoid CXRs in 43% (781/1,1837) of children
    • Using NHLBI or CP or WBC>18.75, 12 cases of ACS would have missed (2), increased sensitivity to 94% and avoid CXRs in 32% (593/1,837) of children
    • Using NHLBI or CP or WBC> 18.75 or history of ACS, 4 cases of ACS would have been missed, increased sensitivity to 98% and would avoid CXRs in 23% (430/1,837) of children

Test Characteristics Sickle Cell Disease

Screen Shot 2015-04-25 at 3.11.12 PM 

Dr. David C. Brousseau

Dr. David C. Brousseau

Listen to the podcast to hear Dr. Brousseau’s responses to our questions.

  1. Use of ICD-9 Codes: You used ICD-9 codes to identify children with sickle cell disease. Is this a validated method?
  2. Relying on Charting: You considered a patient having shortness of breath if something was documented on the chart but the absence of documentation of respiratory symptoms was treated as a negative. How do you think that could have influenced your results?
  3. Normal Range: We were impressed you used Fleming et al Lancet 2011 to determine tachypnea, a paper we have covered on SGEM#68.
  4. White Blood Cell (WBC) Count: Why did you pick WBC as the laboratory test when other tests also had significant differences?
  5. WBC Cut-Off: How did you decide >18.75 would be the cut off?
  6. Classification and Regression Tree (CART): You used CART analysis to determine predictive factors for a diagnosis of ACS. Can you explain the CART process?
  7. Acceptable Miss Rate: What is an acceptable miss rate for acute chest syndrome?
  8. Retrospective Chart Review: This was a retrospective chart review. While providing some information you do need to prospectively validate the results.

Comment on authors conclusion compared to SGEM Conclusion: We generally agree with the authors’ conclusions.

SGEM Bottom Line: Adding clinical features such as chest pain, WBC count >18.75 or history of ACS may improve sensitivity of criteria directing the decision to order a CXR in febrile sickle cell children presenting to the ED. A prospective validation study in febrile children with sickle cell disease presenting to the ED using these criteria is needed to determine which children should get a CXR to help diagnose acute chest syndrome.

Case Resolution: You obtain a CXR in your febrile patient with sickle cell disease, along with a complete blood count and blood cultures, and start empiric antibiotics. The CXR is negative for an acute infiltrate, and the patient is admitted to the hospital for further care.

Dr. Corey Heitz

Dr. Corey Heitz

Clinically Application: Getting a CXR in pediatric sickle cell disease patients presenting to the emergency department with NHLBI consensus criteria or chest pain will identify most cases of acute chest syndrome. However, getting a CXR in patients with WBC >18.75 count or history of acute chest syndrome is associated with an increased sensitivity while decreased specificity.

What do I tell my patient? During your discussion with the patient’s mother, you explain that the minimal respiratory symptoms her daughter has combined with a negative CXR make you feel comfortable that the patient does not have acute chest syndrome. However, febrile patients with sickle cell disease are still at high risk of serious bacterial illness, and you would like to admit her daughter to the hospital for treatment. 

Keener Contest: Last weeks’ winner was Jaroslaw Gucwa from Poland. He knew the medial term for the ketamine tiger is emergence delirium.

Listen to the podcast for this weeks’ keener contest question. If you know the answer then email it to me at and the first correct answer will receive a cool skeptical prize.

Continuing Medical Education (CMD) Credits: SGEMers can claim their CME credits by following the process listed below.

  1. Go to the Wiley Health Learning website
  2. Register and create a log in
  3. Search for “Academic Emergency Medicine – November”
  4. Complete the five questions and submit your answers

This is a new feature and there may be some glitches. Please email Corey ( with any questions or difficulties.

FOAM logoOther FOAMed Resources on Sickle Cell Disease:

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.

SkiBEEM 2017

  • Anthony Crocco

    Great SGEM HOP! Looking forward to the prospective validation – has potential to help direct management and minimize resource utilization.

    For those who need more info on development of clinical decision instruments should see the 3 part series:

    • TheSGem

      Thanks Anthony for the positive feedback on this #SGEMHOP episode.

      Keep those SketchyEBM videos coming.

  • Manrique Umana

    Although I don’t see kids, we do see quite often SCD patients and the diagnosis of acute chest syndrome is one frequently missed or delayed. This post helps us analyse the utility of several clinical key features.

  • David Brousseau

    We just hope that this helps further the conversation about how to best care for febrile patients with sickle cell disease.

  • Pingback: Sickle Cell Acute Chest Syndrome()

  • Salim R. Rezaie

    Great topic and discussion…my take away is this improves sensitivity, but still needs prospective validation #SGEMHOP

  • Justin Morgenstern

    Thanks for joining us on the SGEM HOP Dr. Brousseau. I’m looking forward to your future work validating this prospectively.

    I wonder about the accuracy of chest x-ray as a gold standard test. Do we know what proportion of patients without an infiltrate on chest x-ray ultimately are diagnosed with acute chest syndrome? Conversely, are there children with what appears like an infiltrate on chest x-ray who might do fine without aggressive treatment? Finally, are you aware of any work looking at chest ultrasound for the diagnosis of acute chest syndrome?

    I have some expanded thoughts on this paper shared in a seperate blog post:

  • Nikki Abela

    Great, useable data. Although it still needs prospective validation, it is a good starting point to direct decision making for imaging in SCD.

    I.e. with features of chest pain, Wcc >18.75 or history of ACS, a CXR is probably a good idea in febrile sickle cell children presenting to the ED.

  • Simon Carley

    Interesting paper and discussion in comments from #FOAMed friends. In Virchester we have a large sickle cell population but because of the amazing work of our red cell team (sub speciality haematologists) presentations to the ED are really quite uncommon (in adults and kids) as compared to when I originally trained. My take on that is that there will be significant population differences between different centres as a result, and as we nerdy people know although prevalence does not affect the performance of a test statistically it sure as hell does if populations are different.

    This means that prospective assessment and local confirmation may well be required.

    I’m also minded that ACS is a pretty significant diagnosis and CXR is a pretty low risk intervention and so the risk benefit ratio needs careful consideration.

    Lastly. Although there is little data our there in kids, chest USS is fabulous and you get great images. As others have commented it may be a better standard than CXR.

    tvb and stay fabulous


    • David Brousseau

      Agree with your thoughts. One question about CXRs being low risk – if xrays are read as infiltrate versus CXR – then there may be harm in obtaining the CXR and calling someone ACS.

    • Seth Trueger

      I agree with Simon — CXR is a remarkably benign test and ACS is a terrible (but treatable!) disease. My rule: sickle + anything* = CXR [*eg chest pain, SOB, fever, cough, URI symptoms, hypoxia]. The false positives for infiltrate are basically pneumonia, which is probably something we should treat.

  • rklaassen

    Nice study with good face validity.

    I am a pediatric hematologist who follows children with sickle cell disease. Probably the one predictor that I have an issue with is the WBC of 18.75. The landmark NEJM study from 1993 looking a selecting low risk patients for outpatient therapy used a WBC cut-off of 30 – what happened to your analysis if you altered the WBC threshold?
    When putting together a prediction rule it is nice to have a validation set – did you not have adequate numbers to split the population into a derivation set (typically first 2/3rd and validation set last 1/3). Obviously not possible to do at this point now that you have already done your analysis.

    Nice work!

    • Corey Heitz

      Great questions and thanks for joining the discussion! I can start to take the first one; I’ll leave the second for David.

      The analysis performed was a computerized one (the CART) analysis; the WBC of 18.75 came from the computer-generated “ideal number” that gave the best balance of sensitivity and specificity. Presumably, at higher WBC counts, you’d lose sensitivity and gain specificity.

      David can you confirm, and speak to his second question re: a validation set?

    • David Brousseau

      Thanks for the great points. We are not attempting to alter the WBC cut-off for low risk criteria for discharge, which is 30K. That WBC criteria assumes no ACS and that the child looks well. The question is whether every febrile child with SCD needs a CXR – that is different than discharge criteria although certainly more CXRs probably means more ACS diagnoses. Unfortunately, we did not have enough for an internal validation.

  • Kirsty Challen

    A great HOP, thanks. SCD is very rare in my part of the world – so very helpful to have a reference point like this in case.