Radiographic evaluation for suspected cerebrospinal fluid shunt obstruction

Objective: To measure the predictive value of plain radiographs (shunt series) and computed tomography (CT) scans in a group of children undergoing evaluation for suspected shunt obstruction in a tertiary care pediatric emergency department (ED). Methods: Radiology reports were reviewed for all ED patients who underwent a shunt series over an 18-month period. Two investigators categorized all reports as normal, possibly abnormal (eg, kink in shunt tubing, no prior CT scan for comparison), or abnormal (with definite evidence of shunt dysfunction, such as shunt tubing disconnection and increase in ventricular size since prior CT scan). Studies for which there was disagreement were re-read independently by a pediatric radiologist. Medical records were reviewed to determine outcomes. Results: A total of 233 patients had shunt series and CT scans ordered. Of these, 60 patients subsequently required surgery for shunt obstruction. The shunt series revealed abnormalities in 12 patients (sensitivity, 20%; negative predictive value, 22%), whereas CT scans showed definite or possible abnormalities in 50 patients (sensitivity, 83%; negative predictive value, 93%). Combined, the two tests detected 53 shunt obstructions (sensitivity, 88%; negative predictive value, 95%). Two obstructed patients had abnormalities on shunt series that would not have been suspected after physical examination or CT scan. Conclusions: Over one quarter of pediatric ED patients evaluated radiographically for suspected shunt obstruction required surgical management. One in eight obstructed patients had normal radiographic studies. Routine performance of shunt series had a low overall yield but on rare occasions detected abnormalities that were missed by CT. Prospective studies are needed to improve the use of radiographic tests for shunt evaluation and determine clinical indications for further workup when studies are normal.