Procalcitonin vs. CRP and IL-6 in Pediatric Bacterial/Viral Differentiation
In a direct head-to-head study of 360 children hospitalized with confirmed infections, procalcitonin (PCT) at a cutoff of 1 µg/L demonstrated superior specificity (93%) and positive predictive value for bacterial infection compared to both C-reactive protein (CRP) and interleukin-6 (IL-6). Specifically, PCT >1 µg/L detected bacterial infection with 83% sensitivity and 93% specificity, whereas CRP ≥20 mg/L showed 86% sensitivity but only 74% specificity due to a 26% false-positive rate in viral infections. IL-6 >100 pg/mL performed worse, with only 53% sensitivity and 86% specificity [1]. This differential performance is critical for antibiotic decision-making in emergency settings, where unnecessary antibiotic exposure must be minimized [2].
| Evidence Dimension | Diagnostic accuracy for bacterial vs. viral infection in children |
|---|---|
| Target Compound Data | PCT >1 µg/L: Sensitivity 83%, Specificity 93% |
| Comparator Or Baseline | CRP ≥20 mg/L: Sensitivity 86%, Specificity 74%; IL-6 >100 pg/mL: Sensitivity 53%, Specificity 86% |
| Quantified Difference | PCT specificity advantage over CRP: +19 percentage points; over IL-6: +7 percentage points |
| Conditions | 360 children, 124 bacterial infections, 236 viral infections; plasma collected at admission; immunoluminometric assay for PCT |
Why This Matters
Procurement decisions must account for PCT's 19% absolute specificity gain over CRP in ruling out viral infection, directly reducing unnecessary antibiotic use in pediatric settings.
- [1] Gendrel D, Raymond J, Coste J, et al. Comparison of procalcitonin with C-reactive protein, interleukin 6 and interferon-alpha for differentiation of bacterial vs. viral infections. Pediatr Infect Dis J. 1999;18(10):875-881. doi:10.1097/00006454-199910000-00008 View Source
- [2] Schuetz P, Wirz Y, Sager R, et al. Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections. Cochrane Database Syst Rev. 2017;10(10):CD007498. doi:10.1002/14651858.CD007498.pub3 View Source
