We appreciate the thoughtful commentary by Dr Burchfield. His concerns are commonly encountered by neonatal clinicians and were considered carefully in the preparation of the revised Clinical Report.

His first concern addresses the decision to continue or stop antibiotics among infants receiving empirical treatment when blood cultures are sterile after maternal intrapartum antibiotic treatment. Culture media with antibiotic neutralization material are widely used in clinical microbiology laboratories to mitigate the impact of circulating antibiotics. If appropriate blood volumes and cultures methods are used, a sterile blood culture provides strong evidence that the infant is not bacteremic. Dr Burchfield also raises concern that an infant who is not bacteremic still may have a focal bacterial infection that will recrudesce if antibiotics are stopped prematurely. This theoretical concern cannot be definitively resolved without a controlled trial of antibiotic administration in the absence of bacteremia. Nonetheless, available observational data (cited in the Clinical Report) suggest that extended antibiotic administration in the absence of a positive blood culture does not result in better outcomes among preterm infants, whether measured by the incidence of death or significant morbidities.

We do not agree that the data of Barton et al support his concern. In that autopsy case series of 111 infants with birth weight 300–1000 g who died in 1990–1993, the authors identified infection as the primary cause of death if white blood cell infiltrates were observed histologically in autopsied organs. “Congenital infection” was diagnosed in 30 cases, among which only 11 infants had blood cultures obtained before death. Pathogenic bacteria were isolated from 9 of 11 (no data are provided for 2 cultures), demonstrating that neonatal blood cultures confirmed the authors’ histologic diagnosis in at least 82% of the cases.

Finally, Dr Burchfield suggests that clinical decision-making will be enhanced by obtaining serial CRP values from infants born in the setting of intrapartum antibiotic administration. While it is true that consistently low CRP levels over the first 48 hours imply an approximately sixfold reduction in the probability of culture-proven sepsis,1  the positive predictive value of CRP levels up to 6 mg/dL was <5% in the cited study and even levels >6 mg/dL had poor positive predictive values for culture-proven sepsis. Nonetheless, the threefold increase in the posterior probability of infection implied by an elevated CRP level may convince some clinicians to extend antibiotic treatments in specific culture-negative cases. The Report states, “Consistently normal values of CRP and procalcitonin over the first 48 hours of age are associated with the absence of EOS, but serial abnormal values alone should not be used to extend antibiotic therapy in the absence of culture-confirmed infection.” The key word is alone; context is important.

Dr Burchfield’s commentary highlights the difficulty of uncertainty. The revised Clinical Report discusses the relative merits of different approaches to risk assessment among term and preterm infants. In the end, no predictive model, clinical algorithm, or laboratory test result will perform perfectly, and we agree with Dr Burchfield that individual cases will always require some degree of clinical judgment.

1
Benitz
WE
,
Han
MY
,
Madan
A
,
Ramachandra
P
.
Serial serum C-reactive protein levels in the diagnosis of neonatal infection
.
Pediatrics
.
1998
;
102
(
4
):
E41

Competing Interests

CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.