Standard Blood Parameters Indicating Infection
C-reactive protein (CRP), procalcitonin (PCT), and white blood cell count with differential (particularly absolute neutrophil count, lymphocyte count, and eosinophil count) are the primary standard blood parameters that indicate infection, with CRP and PCT being the most clinically useful biomarkers for distinguishing infectious from non-infectious inflammation.
Primary Biomarkers for Infection Detection
C-Reactive Protein (CRP)
CRP is an acute-phase protein that rises 12-24 hours after an infectious or inflammatory insult, peaking at 48 hours 1. For ruling out bacterial infection when clinical probability is low to intermediate, measuring CRP is recommended 1. Normal values are typically below 5 mg/L, with a diagnostic cutoff of 10 mg/L 1. CRP demonstrates moderate diagnostic accuracy with an area under the curve of 0.73 for sepsis diagnosis 1. Importantly, CRP performs better than PCT for identifying clinical infection overall (0.75 vs 0.71 ROC-AUC) 2.
Common pitfall: CRP is non-specific and can be elevated in any inflammatory condition, not just infection. It can also be affected by neutropenia, immunodeficiency, and NSAID use 1.
Procalcitonin (PCT)
PCT rises within 4 hours of bacterial exposure, peaking at 6-8 hours 1. Normal values are less than 0.05 ng/mL. PCT is superior to CRP for predicting bacteremia specifically, with a ROC-AUC of 0.78 versus 0.67 for CRP 2. A PCT cutoff of 0.5 ng/mL provides 72.6% sensitivity and 69.5% specificity for bacteremia, with an excellent negative predictive value of 92.9% 3.
Critical caveat: Recent evidence shows PCT can be elevated in severe viral infections including influenza and COVID-19, reducing its discriminatory power for bacterial versus viral etiology 1.
White Blood Cell Count and Differential
The complete blood count with differential provides multiple infection indicators:
- Leukocytosis (elevated WBC): Sensitivity of only 66% and specificity of 56% for infection—poor as a standalone marker 4
- Absolute neutrophil count: Values below 7,000/mm³ make infection unlikely 4
- Lymphopenia (< 1,200/mm³): 58% sensitivity and 73% specificity for infection 4
- Eosinopenia (< 100/mm³): 91% sensitivity but only 38% specificity—highly sensitive but not specific 4
The combination of normal neutrophils (< 7,000/mm³), normal eosinophils (> 100/mm³), and normal lymphocytes (> 1,200/mm³) reduces infection probability to less than 3% 4.
Algorithmic Approach to Interpretation
When Infection Probability is Low to Intermediate:
- Measure CRP or PCT alongside clinical evaluation 1
- If CRP > 10 mg/L or PCT > 0.5 ng/mL → pursue infection workup
- Check WBC differential for eosinopenia and lymphopenia as supporting evidence
When Infection Probability is High:
Do not rely on CRP to rule out bacterial infection—proceed with appropriate cultures and empiric treatment 1
For Suspected Bacteremia Specifically:
- PCT is the best single parameter with 0.78 ROC-AUC 2
- PCT > 0.5 ng/mL strongly supports bacteremia
- Consider bilirubin elevation as an additional marker (0.65 ROC-AUC for bacteremia) 3
Additional Useful Parameters
Platelet Indices
- Mean platelet volume (MPV) and platelet count (PLT) combined with other parameters improve discrimination between different pathogens 5
- The combination of CRP, hemoglobin, and platelets can distinguish between viral, bacterial, and atypical infections in children 6
Hemoglobin
Lower hemoglobin correlates with bacterial infection, with a ROC-AUC of 0.72 for distinguishing bacterial from viral infection 6
Key Limitations and Clinical Context
No single biomarker is sufficiently discriminating to be used independently as a diagnostic tool 2. The 2023 Critical Care Medicine guidelines explicitly recommend against routine use of biomarkers in established sepsis and septic shock due to uncertain benefit and cost issues 1. However, these markers have value when infection probability is uncertain and can guide antimicrobial discontinuation decisions.
The most effective approach combines multiple parameters: A logistic regression model incorporating CRP, hemoglobin, platelets, eosinophils, monocytes, and lymphocytes significantly outperforms any single marker 6, 5.