S1 Database -

Background
C reactive protein (CRP), a marker for the presence of inflammation, has been extensively studied for distinguishing bacterial from non-bacterial infection in febrile patients, but its role in excluding malaria in the febrile child has not been thoroughly evaluated.
Method
This was a cross-sectional study at the Douala Gyneco-Obstetric and Pediatric Hospital which included all patients between the ages of one month and 16 years presenting with fever. Consenting patients received complete clinical examinations, then venous blood samples were collected and tested for CRP values, bacterial infection and malaria.
Results
Samples of 220 children were analyzed. 142/220 had viral infections, 50/220 had malaria and 49/220 had bacterial infections. 7/220 had both malaria and bacterial infection. There was no significant difference between mean CRP values in malaria and bacterial infection (p = 1), but CRP means were significantly higher in malaria/bacterial infection than in viral infection (p<0.0001). Area Under the Receiver Operating Characteristics Curve (AUROC) values were 0.94 for malaria and 0.86 for bacterial infection, with a calculated cut-off of 23.6mg/L for malaria and 36.2mg/L for bacterial infection. At these cut-offs, CRP had a Positive Predictive Value (PPV) of 68.75% and 85.00% for malaria and bacterial infection respectively, with a Negative Predictive Value (NPV) of 94.74% and 89.05% respectively.
Conclusion
CRP can effectively exclude malaria and bacterial infection in febrile children in low-resource settings without the need for additional tests.