Criteria for the discontinuation of antibiotic therapy during presumptive treatment of suspected neonatal infection.

Unstable newborns are often subjected to multiple diagnostic tests and then treated presumptively for bacterial infection. Inconclusive test results may perpetuate unnecessary therapy. One hundred twenty-three neonates were prospectively evaluated for infection. Complete physical examinations and chest radiographs were performed. Of ten screening tests only the white blood cell count, absolute band count, absolute band/neutrophil ratio and C-reactive protein showed statistical differences (P less than 0.05) among 32 patients with positive "nonpermissive" (blood, cerebrospinal fluid, suprapubic or catheter urine, needle aspirate, tracheal aspirate) cultures and 50 with negative cultures who had antibiotic therapy discontinued within 72 hours. Forty-one additional patients were continued on therapy despite negative cultures. Incomplete bacteriologic evaluation resulted in unconfirmed "pneumonia" in 16 of these children. No statistical differences between the culture negative groups existed regardless of treatment status, suggesting that patients in the latter group may not have required continued treatment. A definitive, bacteriologic evaluation emphasizing "nonpermissive" cultures should be completed in newborns suspected of infection which, when negative, should allow discontinuation of antibiotics. In equivocal cases, a negative C-reactive protein best supports cessation of therapy. This approach can reduce the total duration of newborn exposure to antibiotics in a high risk nursery by 20%. A prospective, definitive, diagnostic evaluation (emphasizing non-permissive cultures) avoids uninterpretable but frequently used tests and reduces cost while adding confidence in the termination of therapy.