#54: Sensitivity and Specificity of Procalcitonin vs C Reactive Protein in Identifying Pediatric Bacterial infections. (28th June 2021)

• Record Type: Journal Article
• Authors: Kim, Sara; Bhatt, Avni; Carr, Silvana; Saccoccio, Frances; Lew, Judy
• Is Part Of: Journal of the Pediatric Infectious Diseases Society. Volume 10(2021)Supplement 2
• Page Start: S16
• Page End: S16
• Publication Date: 2021-06-28
• Subjects: Communicable diseases in children -- Periodicals; Children -- Diseases -- Periodicals; 618.929
• DOI: 10.1093/jpids/piab031.035
• Languages: English
• ISSNs: 2048-7193

Abstract

Abstract: Background: Procalcitonin (PCT) and c-reactive protein (CRP) have been utilized in children to assess risk for serious bacterial infections. However, there have been different cut-offs reported for PCT and CRP, which yield different sensitivity and specificity. This study aims to compare the sensitivity and specificity of PCT and CRP in detecting serious bacterial infections (SBIs), specifically urinary tract infections, bacteremia and meningitis. Methods: In this retrospective, single center cohort study from January 2018 to June 2019, we analyzed children with a fever greater than 38C with both PCT and CRP value within 24 hours of admission. Each patient had a blood, urine and/or cerebrospinal fluid culture collected within 48 hours of admission. No antibiotics were administered from the admitting hospital prior to collection of the PCT or CRP. Our gold standard was a positive culture obtained from blood, cerebrospinal fluid, or urine. The statistical analysis included categorical variables as percentages and compared them using the Fisher exact test. The optimal cutoff values for PCT or CRP were based on ROC curve analysis and Youden Index. Sensitivity and specificity analysis were based on literature review cut offs and ROC curves cut offs. Results: Among 202 children, we had 45 culture positive patients (11 urinary tract infections, 4 meningitis, and 32 bacteremia). The patients with culture positivity had higher PCT levels (7.9 ng/mL vs 2.5 ng/mL, P=0.0111), CRP levels (110.9 mg/L vs 49.6 mg/L, P<0.0001) and temperature (39.2C vs 39C, P<0.0052). The area under the curve (AUC) comparing culture positivity vs negativity for all culture types was 0.72 (p<0.0001) for PCT and 0.66 (p=0.001) for CRP. In Figure 1, the AUC for culture positive bacteremia was 0.68 (p=0.0011) for PCT and 0.70 (p=0.0003). The AUC for culture positive urinary tract infections (UTI) only was 0.86 (p=0.0001) for PCT and 0.70 (p=0.3607). For the cut-off value for PCT at 0.5 ng/mL, the sensitivity and specificity was 64% (95% confidence interval [CI] 0.5–0.77) and 70% (95% CI 0.62–0.77) respectively in identifying children with bacterial infection. For the cut-off value for CRP at 20 mg/L, the sensitivity and specificity was 67% (95% CI 0.52–0.79) and 52% (95% CI 0.44–0.59) respectively in identifying children with bacterial infection. Conclusion: In this study, PCT and CRP are nearly equivalent classifiers for detecting SBIs as a group and bacteremia, but PCT is statistically better for urinary tract infections; however, the clinical utility is unknown.