Abnormal Heart Rate Variability Precedes the Clinical Diagnosis of Late-Onset Neonatal Sepsis

Abstract 1170 Neonatology: Clinical Studies in the Premature Infant Platform, Tuesday, 5/4

Late-onset neonatal sepsis is a major cause of morbidity and mortality, and an earlier diagnosis should improve therapy of this potentially catastrophic illness. We hypothesized that abnormal HRV would precede clinical diagnosis of this important neonatal illness. In the University of Virginia NICU, we prospectively and continuously monitored HRV in infants who were at risk for developing late-onset sepsis. We collected and electronically stored RR interval data on study infants. We are reporting on 59 infants for whom HRV data were analyzed 7 days before and 3 days after a culture-proven septic event (35 infants with 38 events) or a randomly selected 10 day period in control infants who had no clinical suspicion of illness or cultures obtained (24 infants with 25 "pseudoevents"). Abnormal HRV characterized by reduced baseline variability and subclinical, short-lived decelerations in heart rate preceded the clinical diagnosis of sepsis. Conventional time- and frequency-domain measures of HRV did not change, and we developed novel measures of HRV that detect this abnormality. One of these HRV measures, skewness of the normalized RR intervals, changed 12 to 24 hours before the clinical diagnosis of sepsis. The figure shows the time course of this HRV measure. The time labeled 0 was when the positive blood culture was obtained (event) or was randomly assigned (control). Data points summarize 12 hour epochs ending at the time value on the abscissa. An arrow marks the epoch ending at time 0; bars are S.E.M. Differences were significant over the two 12 hour epochs before the event. Severity of illness as measured by SNAP and NTISS scores were higher for the septic infants but did not change before the clinical diagnosis of sepsis. Thus we have found that the clinical diagnosis of sepsis is preceded by abnormal HRV. We have developed new measures of HRV that detect the abnormality 12-24 hours before the clinical diagnosis of culture-proven neonatal sepsis. Funded by the University of Virginia Children's Medical Center and Virginia Center for Innovative Technology.

Fig 1

No caption