William E King, MS
Introduction
Evidence continues to build that HeRO monitoring improves outcomes of premature infants, including all-cause NICU mortality, (1) mortality after infection, (2) mortality at 18-22 months, (3) mortality-or-severe-cerebral-palsy at 18-22 months, (3) and NICU length of stay. (4) Yet some neonatologists find themselves hesitant to adopt HeRO monitoring for fear that it may lead to higher rates of testing and antibiotic usage. Here, we examine whether those fears are well-founded and the hesitancy justified.
Background
The HeRO Score (aka HRC Index) is calculated every hour and identifies abnormal heart rate patterns of reduced variability and transient decelerations that are associated with cytokines (5-7) and often precede sepsis(8-18) UTI, (19) NEC, (20,21) , meningitis, (19) neuro trauma, (22-25) respiratory decompensation, (26) extubation readiness, (27,28) and death. (23,25, 29-31) HeRO monitoring has been utilized as an early warning system, (32) and Moorman et al. hypothesized that it may lead to early diagnoses, earlier interventions, and improved outcomes. In the largest RCT ever published among premature neonates, 3003 VLBW patients at nine hospitals were randomized to either receive standard of care monitoring, or standard of care monitoring plus HeRO. (1)
While mortality and other outcomes described above were statistically significantly improved for those patients randomized to the HeRO-display group, Moorman et al. described non-significant trends toward increased testing and antibiotics: “Infants whose HRC monitoring results were displayed had 10% more blood cultures drawn for the suspicion of sepsis (1.8 per month compared with 1.6, P = .05) and 5% more days on antibiotics (15.7 compared with 15.0, P = .31, Table).” (1)
Mortality, however, is a competing outcome with both cultures drawn and antibiotic days, and properly accounting for the increase in survival when assessing other outcomes can change the result. Indeed, we have previously reported that length of stay among this cohort was longer among the HeRO-display group when failing to account for the competing outcome of mortality, but shorter when so doing. (4)
We hypothesized that metrics of blood culture rates and antibiotic usage would favor HeRO-monitoring after adjusting for the competing outcome of mortality.
Conclusion
Previous reports have indicated trends toward increased testing and treatment associated with HeRO monitoring, so clinician concern is well-founded. Nevertheless, when examining those concerns after controlling for, and in the context of, the mortality improvement associated with HeRO monitoring, hesitancy in adopting the technology is not justified.