Intermittent desaturation episodes are a significant clinical problem in preterm infants as a consequence of immature respiratory control. These alterations in oxygenation are unpredictable and are probably due to sudden, dynamic changes in ventilatory drive. Hypoxia/hyperoxia events are difficult to control using threshold alarms in the neonatal intensive care unit and, although continuous pulse oximetry monitoring has been in use for many years, this technology has yet to be used to characterize the true incidence of intermittent hypoxic episodes in preterm infants and to further assess the relationship of these events to neonatal morbidity. As part of a multicenter study of oxygen supplementation, Rainbow Babies and Children's Hospital has generated two patient data sets that include long-term, high resolution monitoring of oxygen saturation via pulse oximetry. The two infant cohorts, comprising a total of 174 infants, include: 1) infants enrolled in the multicenter SUPPORT trial and 2) infants meeting enrollment criteria but not enrolled in the SUPPORT trial who received normal clinical care. Preliminary data analysis in the second cohort have revealed a significantly higher incidence of desaturation (<80% SaO2) events in infants developing ROP requiring laser therapy. However, there is large inter- and intra-patient variability in the pattern of desaturation episodes over time;additionally, data in animal models has suggested an association between patterns of desaturation events and ROP severity. We hypothesize that repetitive desaturation/resaturation events will result in clustering of hypoxia and hyperoxia events and present a significant risk factor for ROP in extremely low birth weight infants younger than 28 weeks gestational age. To test this hypothesis we will use our existing database of long-term measurements of oxygen saturation obtained via pulse-oximetry for all 174 infants gathered over 8 weeks of NICU care.
In Specific Aim 1, we will use temporal and spectral analysis methods to quantify the linear properties of this patient dataset;additionally we will quantify clustered intermittent hypoxia and hyperoxia events using both linear and non-linear statistical methods to quantify the signal components present in the data.
In Specific Aim 2, we will use a linear mixed model and discriminant analysis to distinguish patterns of desaturation, resaturation, and hypersaturation in infants who do and do not require laser treatment for ROP.
Intermittent desaturation episodes are a significant clinical problem in preterm infants as a consequence of immature respiratory control. Higher incidence of intermittent hypoxic (desaturation) episodes is associated with retinopathy of prematurity (ROP), a condition which can lead to life-long visual impairment and blindness. However, animal models have suggested that variable patterns in the intervals between desaturation episodes are an additional risk factor for ROP. In this application we propose analysis methods to quantify clustered intermittent hypoxia and hyperoxia events in order to minimize risk for ROP in very low birth weight infants. Identification and elimination of these patterns of intermittent hypoxia may lead to reduced incidence of ROP.
| Wang, Katherine; Difiore, Juliann M; Martin, Richard J et al. (2013) Markers for severity of illness associated with decreased snoring in toddlers born ELGA. Acta Paediatr 102:e39-43 |
| Martin, R J; Di Fiore, J M; Macfarlane, P M et al. (2012) Physiologic basis for intermittent hypoxic episodes in preterm infants. Adv Exp Med Biol 758:351-8 |
| Di Fiore, Juliann M; Kaffashi, Farhad; Loparo, Kenneth et al. (2012) The relationship between patterns of intermittent hypoxia and retinopathy of prematurity in preterm infants. Pediatr Res 72:606-12 |
