Impact of a Procalcitonin Testing and Treatment Algorithm on Antibiotic Use and Outcomes in the Pediatric Intensive Care Unit

The timely use of effective antibiotics can markedly reduce the morbidity and mortality
associated with bacterial infections, particularly in the intensive care unit (ICU). However,
in this setting, much antibiotic use is empiric, and administered to patients with
non-bacterial or non-infectious causes of inflammation that do not respond to antibiotics.
This widespread empiric use of antibiotics drives the emergence of multi-drug resistant
organisms and antibiotic-associated adverse events, such as C. difficile infections.
De-escalation of broad-spectrum empiric antibiotics for ICU patients without proven bacterial
infections can reduce unnecessary antibiotic use, slow the development of antibiotic
resistance, and reduce complications associated with antibiotic therapy. Thus, antibiotic
de-escalation is an important goal of antimicrobial stewardship programs. Specific tests and
pathways to predict which patients have bacterial infections and those that would benefit
from antibiotic therapy would accelerate de-escalation and greatly facilitate antimicrobial
stewardship efforts.

Procalcitonin (PCT) has been investigated as a biomarker for critically ill adult patients
with bacterial infection, particularly pneumonia and sepsis. Following bacteria-induced
activation of monocytes and adherence of monocytes to endothelial surfaces, procalcitonin is
expressed and secreted. PCT levels have been shown to rise rapidly and remain elevated during
ongoing bacterial infections, and PCT levels are more specific for bacterial infections than
CRP or total white blood cell count. PCT rises approximately 4 hours after bacterial
exposure, peaks between 12-24 hours, and has a half-life of 24 hours once the infectious
stimulus is removed.

In many adult trials investigating PCT-guided algorithms for antibiotic cessation (refer to
section 3.0), a high proportion of providers (up to 50%) chose not to follow algorithm
guidance for subjects randomized to the PCT-guided group. Thus, although PCT appears to be a
useful guide for safe antibiotic de-escalation in the ICU, the ideal method for implementing
the test and integrating it into clinical care in order to maximize its impact in the
pediatric population is unclear. Notably, none of the prior trials evaluated PCT-associated
outcomes in critically ill children nor integrated PCT testing into antimicrobial stewardship
activities.

The investigators propose the evaluation of a PCT testing and treatment algorithm on patient
outcomes in the pediatric ICU, a setting in which PCT-guided antibiotic de-escalation has not
been previously studied.

The proposed project will evaluate whether a procalcitonin (PCT) testing and treatment
algorithm, implemented through daily antimicrobial stewardship audit and feedback, can
promote early and safe antibiotic de-escalation in the pediatric ICU. The investigators will
conduct a pragmatic, prospective randomized controlled trial comparing antimicrobial use and
outcomes among children admitted to the ICU who receive either: 1) Routine laboratory testing
and treatment with antimicrobial stewardship review (control), or 2) PCT testing and
treatment with antimicrobial stewardship review (intervention). In both arms, baseline daily
review of antimicrobial management by the stewardship team will occur. In the intervention
arm, the stewardship provider also will recommend PCT testing and antibiotic modifications
using a PCT-based treatment algorithm. PCT levels will be measured a total of four times in
the intervention arm - on enrollment, then daily through day 3 post-randomization and on day
5 post-randomization. This research is not to determine if PCT is a good test; this has
already been established and evaluated as part of the FDA approval process. This pragmatic
outcomes trial is evaluating if use of the PCT, implemented together with antimicrobial
stewardship program oversight, improves the quality of care the investigators can provide for
children at Vanderbilt Children's Hospital. The investigators hypothesize that patients in
the intervention arm will have shorter duration of antibiotic therapy and similar outcomes,
as compared to patients in the control arm.

Specific Aims

1. Compare antimicrobial utilization among children in the ICU who receive standard-of-care
testing plus stewardship vs. PCT-based treatment plus stewardship. The investigators
will compare days of antibiotic therapy in the first 14 days following randomization
between the study arms. The investigators will test the hypothesis that duration of
antibiotic therapy will be 2 days shorter in the group with PCT-guided management vs.
the group with standard of care testing and treatment.

2. Compare clinical outcomes and safety among children in the ICU who receive
standard-of-care testing plus stewardship vs. PCT-based treatment plus stewardship. The
investigators will compare mortality, length of stay, recurrence of infection, and
antibiotic-associated adverse events (rash, myelosuppression, renal impairment,
hepatotoxicity, C. difficile infection) between the study arms. The investigators will
test the hypothesis that outcomes and safety will be comparable between the study arms.

Inclusion Criteria:

- 18 years of age or younger

- Prescribed or administered antibiotics in the hospital less than or equal to 24 hours
prior to enrollment

- Have parents or legal guardians who provide informed consent

- Provide assent (if > 7 years of age)

Exclusion Criteria:

- Are not prescribed antibiotics in the hospital

- Receive intravenous antibiotics within 7 days prior to identification for study
enrollment

- Primary or secondary immune deficiency

- History of malignancy, bone marrow transplant or solid organ transplant

- A diagnosis of cystic fibrosis

- Neonates < 34 weeks gestation

- Patients receiving treatment for endocarditis, osteomyelitis, meningitis,
mediastinitis or other invasive infection, for which long duration of antibiotics is
needed

- Do not provide informed consent/assent