
Methods: We screened every patient over the age of 65 who presented to the ED at a large academic medical center from 12/13/13 to 4/11/13 (target enrollment=240). Informed consent was followed by bedside interview, medical record abstraction, and blood draw. PCT levels were obtained using a VIDAS BRAHMS PCT instrument. The treating ED physicians were surveyed about the estimated likelihood of infection. An Infectious Disease physician review of each patient’s full medical record to determine the presence of true infection was the gold standard. Treating physicians and reviewers were blinded from the results of the PCT.
Results: Among 368 eligible patients, 90 were enrolled (ages 65 to 94). Of the enrolled patients, 24 (26.67%) met SIRS criteria in the ED, and 30 (33.33%) were diagnosed with bacterial infection by the gold standard reviewer. Of these bacterial infections, the most common were pneumonia (8 [26.67%]), urinary tract infection/pyelonephritis (6 [20%]), and bacteremia (4 [13.33%]). PCT levels were compared at two cutoffs, 0.2 ng/ml (positive in 21 [23.33%]) and 0.5ng/ml (positive in 11 [12.22%]). Sensitivity and specificity were calculated for the following scenarios: 0.2 ng/ml cutoff for any bacterial infection, 45% and 85%; 0.5 ng/ml cut off 26.7% and 95%; 0.2 ng/ml cut off for SIRS patients, 64.3 % and 60%; 0.5 ng/ml cut off 42.9% and 90%. ROC area under the curve for PCT in all patients was 0.6772 and 0.7857 for those with SIRS.
Conclusion: PCT was specific but not sensitive for identifying true infection in older adults presenting to the ED at a large academic center. ASPs are needed in the ED, and PCT may have a role in determining appropriate use of antimicrobials but further studies are needed.

R. Leininger,
None
J. Caterino, None