Clinician assessment for acute chest syndrome in febrile patients with sickle cell disease: is it accurate enough?

STUDY OBJECTIVE: To determine whether the use of empiric chest radiography (CXR) is of significant value in detecting clinically unsuspected acute chest syndrome (ACS) in febrile patients with sickle cell disease (SCD).

METHODS: Patients with SCD presenting to the emergency department and hematology clinic with temperature greater than or equal to 38 degrees C were prospectively evaluated using a physician-completed questionnaire. The questionnaire included inquiries into the patient's physical signs and symptoms and the physician's clinical impression for the presence of ACS. The questionnaire was completed before obtaining CXR results in all patients.

RESULTS: Seventy-three patients with SCD with 96 febrile events were evaluated over a 1-year period. Twenty-four percent (23/96) of the patients had CXR evidence of ACS. On the basis of the questionnaire data, 61% (14/23) of ACS cases were not clinically suspected by the evaluating physician before obtaining CXR. Comparing the patients with and without ACS revealed that, with the exception of splinting (4/23 [17%] versus 0/73 [0%]), no symptom or physical examination finding helped to identify which patients had ACS. Fifty-seven percent of patients with ACS had completely normal findings on physical examination. The presentation of patients with clinically detected versus clinically unsuspected ACS also did not differ significantly. Length of hospitalization, oxygen use, and need for transfusion were the same in both the unsuspected and detected ACS groups. Overall physician sensitivity for predicting ACS was only 39%, and diagnostic accuracy did not improve significantly with increasing levels of pediatric training.

CONCLUSION: ACS is common in patients with SCD who present with fever and was grossly underestimated by evaluating physicians. History and physical examination appear to be of little value in defining which febrile patients require CXR. In view of the mortality and morbidity associated with ACS, empiric CXR should be considered when evaluating a febrile patient with SCD.