1. Critical Analysis
•ALICIA CHANG •CLAY COHEN •DAVID STEFFEN
•ERIN DOHERTY •EVA ZIMMERMAN •EVAN LI
•MATT MILLER •MELISSA JONES •NETTA SCHNELLER
•SHAWKI QASIM •TAYLOR OLMSTAD

2. Critical Questions
1. What are the hypothesis and aims of the study?
• Null Hypothesis (H0): “Serum Procalcitonin levels in children with central lines, presenting to the ED
with fever is the same for bacteremic and non-bacteremic patients”
• Alternative Hypothesis (HI): “There is an association between fever due to bacteremia in children with
central lines and their serum procalcitonin level at their first ED presentation”
• Aim: “To evaluate the clinical use of procalcitonin (PCT) as a rapid marker for the identification of
bacteremia in the emergency department (ED) population of children with fever and a central venous
catheter (CVC)”
2. What gap in knowledge was meant to be filled?
• Absence of a normal cut-off definition for serum procalcitonin level with respect to identifying
bacteremia in children with fever in the presence of CVC.
• Absence of literature on using serum procalcitonin level as a triage tool for children with central line who
present to the ED with fever.

3. Critical Questions
3. What were the methods used (including study design, study population, power calculations,
sample size determination, approach and techniques)?
• Design: Prospective Cohort Observational Study.
• Study Population: Patients younger than 18 years, febrile, with no focus of infection, and with CVC
presenting to a single institution ED with no antibiotics in the 24 preceding hours.
• Power/Sample Size Calculations: Based on an assumed risk of SBI of 17.9 from another source, authors
calculated that a sample of 40 patients would be sufficient to generate a power of 80% with alpha at 0.05;
they used a sample of 62 to make sure they have enough bacteremic patients.
• Approach: there was no intervention in this study; all eligible patients had central blood cultures, serum
procalcitonin level drawn and antibiotics administered. Admissions were decided based on clinical status
and ANC for patients with oncologic diagnoses.

4. Critical Questions
3. What were the methods used (including study design, study population, power
calculations, sample size determination, approach and techniques)?
• Statistical Design/Techniques:
• Summary statistics, including mean and SD, range, and percentages, as necessary, was used for data
presentation.
• Categorical data were compared by Fisher exact test.
• The independent variables were compared with the Mann-Whitney U test and the dependent
variables with the Wilcoxon test.
• The optimal cutoff of PCT level for a maximum sensitivity and specificity of a bacterial infection was
assessed by conducting a receiver operating characteristic (ROC) analysis.
• The statistical significance level was defined as a value of > = 0.05, with 2-sided alternative
hypotheses and no adjustment for the multiple comparisons.
• Analysis was performed with PASW Statistics Version 18.

5. Critical Questions
4. What are the confounders?
• Confounders adjusted for: Age, oncologic versus non-oncologic diagnoses and presence or absence of
neutropenia.
• Unadjusted Confounders:
• Variation in temperature measurement technique.
• Onset and Duration of fever.
• Exposure to antibiotics before 24 hours from presentation.
• Recent history of bacteremia.
• Degree of immunosuppression.
• Differences in underlying oncologic diagnoses (Solid vs liquid, Hemolymphatic vs Non-Hemolymphatic
and Localized vs Metastatic malignancies).

6. Critical Questions
5-6. What are the results and How are they interpreted?
• Sixty-two patients were enrolled, and 14 (23%) had a positive culture.
• Mean PCT value in bacteremic patients was 18.47 ± 31.6 ng/mL and 0.65 ± 1.2 ng/mL in nonbacteremic
patients (P <0.001).
• Median PCT for negative blood culture was 0.23 ng/mL (interquartile range, 0.11-0.61) and 1.15 ng/mL
for a positive blood culture (interquartile range, 0.45-29.16).
• The receiver operating characteristic analysis identified a level of PCT of 0.3 ng/mL as the best cutoff
point that produced a sensitivity of 93% and a specificity of 63% (area under the curve, 0.82).
7. What are the conclusions of the study and are they justified?
• Conclusion: “Serum PCT is useful in identifying CLABSI in children who present to the ED with fever and
a CVC”.
• Justification: presuming an accurate calculation of the study power, the findings are justified for this
particular population in this particular setting, but we are concerned about the use of a definite SBI
prevalence of 17.9% to calculate power as this is derived from a study in infants < 3 months of age.

7. Critical Questions
8. Are there alternate explanations for the results?
Not necessarily full alternative explanations, but unadjusted potential confounders:
• Cancer patients in general have higher procalcitonin levels
• Leukemia patients generally have higher PCT levels than solid tumor cancer patients (Carnino et al,
2010)
• Advanced stage tumors are associated with higher PCT levels (Showali, 2012)
9. What is the quality of the discussion, and does it lead to other investigations that expand
the field?
Discussion is of good quality and suggests:
• Larger patient sample over a longer observation duration.
• Evaluation of PCT as a follow up marker.
• Evaluation of the utility of PCT after antibiotic therapy.