Prediction of intraventricular haemorrhage in preterm infants using time series analysis of blood pressure and respiratory signals

Abstract

Despite the decline in mortality rates of extremely preterm infants, intraventricular haemorrhage (IVH) remains common in survivors. The need for resuscitation and cardiorespiratory management, particularly within the first 24 hours of life, are important factors in the incidence and timing of IVH. Variability analyses of heart rate and blood pressure data has demonstrated potential approaches to predictive monitoring. In this study, we investigated the early identification of infants at a high risk of developing IVH, using time series analysis of blood pressure and respiratory data. We also explore approaches to improving model performance, such as the inclusion of multiple variables and signal pre-processing to enhance the results from detrended fluctuation analysis. Of the models we evaluated, the highest area under receiver-operator characteristic curve (5th, 95th percentile) achieved was 0.921 (0.82, 1.00) by mean diastolic blood pressure and the long-term scaling exponent of pulse interval (PI α2), exhibiting a sensitivity of >90% at a specificity of 75%. Following evaluation in a larger population, our approach may be useful in predictive monitoring to identify infants at high risk of developing IVH, offering caregivers more time to adjust intensive care treatment.

Early Pulse Oximetry Data Improves Prediction of Death and Adverse Outcomes in a Two-Center Cohort of Very Low Birth Weight Infants

“…Several recent publications use detrended fluctuation analysis of vital signs to predict IVH. 14,15,16 It is important to note that IVH prediction has its limitations in retrospective studies such as ours due to the fact that the exact timing of the event is unknown. Abnormal HR and SpO 2 patterns in the first 12 hours after birth can represent disturbances caused by the hemorrhage rather than a warning of the event occurring later.…”

The Critical Role of the Central Autonomic Nervous System in Fetal-Neonatal Transition

“…In premature newborns, HRV changes can be seen preceding the development of intraventricular hemorrhage. 83,84 Whether ANS dysfunction in these newborns contributes to brain injury or whether ANS dysfunction is a more sensitive clinical early biomarker, is not exactly known.…”

“…The next step in this continuum of work is to develop real-time ANS monitoring which can be used at the maternal (fetal) or neonatal bedside in both research and in clinical practice settings to identify a fetus (or newborn) at risk of impaired ANS function/development. HRV analysis methods have already been studied and may predict intraventricular hemorrhage, 84 brain injury pattern, 49 developing necrotizing enterocolitis, 64 and risk for apparent life threatening events. 4 This technology therefore has the potential to reduce brain injury and adverse neurologic outcomes in high-risk pregnancies and premature or other at-risk newborns.…”

Impact of delivered tidal volume on the occurrence of intraventricular haemorrhage in preterm infants during positive pressure ventilation in the delivery room

“…Furthermore, several animal studies have demonstrated that excessive V T delivery during PPV in the delivery room causes brain inflammation and injury 20–23. The pathways by which PPV causes brain inflammation and brain injury include (1) altered pulmonary venous return and left ventricular output resulting in rapid abnormal fluctuations to cerebral blood flow, (2) fluctuations in blood pressure due to variable V T during PPV,24 (3) initiation of pulmonary inflammation, which initiates a systemic and subsequent cerebral inflammatory cascade and (4) a sudden rather than gradual increase in oxygen delivery causing a vascular response in the brain 22. Animal studies using premature lambs by Polglase et al 20 22 and Skiӧld et al 21 demonstrate that V T s>8 mL/kg cause brain inflammation and brain injury which is evident within 90 min after initiation of PPV.…”