Multiparametric analysis of body position effects in heart rate variability in premature babies

Preterm infants are at greater risk for a variety of neurodevelopmental disorders, including Sudden Infant Death Syndrome (SIDS), due to altered autonomic function. Sleeping in prone position is one of the risk factors for SIDS and has been related to altered physiological markers of autonomic control, including heart rate variability (HRV). Accurate, prospective identification of babies likely to succumb is not yet possible. More sensitive and quantitative measures of HRV may allow improved assessment of vulnerability due to position. The aim of this thesis is to employ parameters proposed for adult and fetal analysis and to add non-linear analysis of HR to standard assessments of HRV in order to measure differences in a high risk population as a function of sleep position. Continuous measurement of HR was done in 35 premature babies before hospital discharge (Birth weight 807-1485 g, gestational age 25 to 33 weeks, post-conceptional age at study 32-43 weeks). During a 6 hour sleep study, data were collected for 20 minutes in supine and prone positions. Time domain measures, frequency domain, and nonlinear were calculated. Data were compared for both positions and for quiet and active sleep state. Results: Significant Measures in Active Sleep comparing Sleep Positions Time Domain: SDNN p-value=0.015, LTI p-value=0.007 Frequency Domain: HF p-value= 0.033 Non Linear: SampEn p-value=0.0047, PRSA decelerations p-value=0.008 Significant Measures in Supine Position comparing Sleep States Time Domain: SDNN p-value=3.9E-05, LTI p-value=5.5E-06 Frequency Domain: HF p-value=0.0124 Non Linear: SampEn p-value=3E-05, PRSA decelerations p-value=2E-04 Preterm infants sleeping in the prone position have higher values of SampEn lower values of SDNN, LTI, HF power. These differences are more precisely and comprehensively characterized using multiple approaches to quantitating HRV and may serve as robust markers of altered autonomic regulation in this population. These results could be the basis for a personalized monitoring system able tp alert increase risk conditions in preterm infants