Objective: To determine whether quantitative C-reactive protein (CRP) level less than 10 mg/L could be used as a reliable index for discontinuation of antibiotic treatment in Thai neonates with clinical sepsis.

Material and Method: All neonates with birth weight greater than 1,500 grams, diagnosed as clinical sepsis, were enrolled to the study. Serum CRP was measured at 24 to 48 hours after the first dose of antibiotics. If CRP level was less than 10 mg/L, infants were randomly divided to groups Ia and Ib. If CRP level was 10 mg/L or more, infants were randomly divided to groups IIa and IIb. Antibiotics were discontinued promptly after the CRP level was reported in group Ia, while CRP level was measured daily and antibiotics were discontinued after it returned to less than 10 mg/L in group IIa. In controlled groups (Ib and IIb), antibiotics were continued until all bacterial cultures were negative. The outcome measurements were the number of patients who required retreatment for clinical sepsis within three days and 28 days after discontinuing antibiotics.

Results: Of 98 neonates with clinical sepsis, 76 (77.6%) were in group I. The duration of antibiotic treatment in group Ia was shorter than group Ib significantly, 1.68 vs. 5.47 days (p<0.01). One neonate in group Ia was retreated on the third day after discontinuing antibiotics due to positive blood and urine cultures. The negative predictive value of CRP for discontinuation of antibiotics in group I was 97.4%. The durations of antibiotic treatments were 5.27 and 7.09 days in group IIa and IIb, respectively. One neonate in group IIa was retreated on the second day after discontinuing antibiotics since the patient’s clinical and laboratory results suggested severe sepsis although all bacterial cultures were negative. No patient was readmitted for treatment of sepsis within 28 days after discontinuing antibiotics.

Conclusion: CRP levels were less than 10 mg/L in the majority of neonates with clinical sepsis. The negative predictive value for using this level as a guide for antibiotic discontinuation was 97.4%.

Keywords: Antibiotics, C-reactive protein, Neonatal sepsis