Background: Inappropriate antibiotic use is associated with bacterial resistance, adverse events, and an increased risk of Clostridium difficileinfection. However, providers often prescribe antibiotics for viral respiratory infections and treat infections such as bacterial pneumonia for unnecessarily prolonged durations.  Serum procalcitonin (PCT)-guided treatment is known to safely reduce antibiotic use and duration of treatment in pneumonia. Yet, adoption of PCT testing is limited and patterns of use are not well described. Using an ongoing prospective study, we analyzed use of PCT and its effect on antibiotic duration in patients hospitalized with pneumonia.

Methods:

Trained abstractors collected data between November 2015 and October 2016 on 1,995 non-ICU patients hospitalized with pneumonia (discharge diagnosis of pneumonia plus symptoms within 48 hours of admission) at 10 Michigan hospitals. Documented signs and symptoms of pneumonia, diagnostic tests, and antibiotic treatment were collected. Data were analyzed using logistic or linear regression, with adjustment for hospital clustering.

Results:
Half (5/10, 50%) of hospitals had PCT testing available. Within those hospitals, use varied widely (Figure 1). When PCT testing was available, patients were more likely to be tested if they were admitted to a hospitalist service (OR 2.44, 95% CI: 1.98-3.00) or had more severe illness (OR 1.14 [95% CI: 1.06-1.23] per unit increase in pneumonia severity index). Most patients (71.2% [316/444]) had only one PCT test ordered (77.0 % [227/295] for initially negative tests; 59.7 % [89/149] for initially positive tests). A negative PCT was not associated with a shorter duration of antibiotic therapy when compared to patients in hospitals without PCT testing (8.46 days [95% CI: 6.82, 10.10] vs 8.91 days [95% CI: 8.27, 9.55], p=0.567, Figure 2).

Conclusions:

PCT use in patients hospitalized with pneumonia in Michigan hospitals was not associated with reduced antibiotic duration.  Improper use of the test, including lack of repeat testing, is a possible causal factor needing further study.