Early instrumental predictors of long term neurodevelopmental impairment in newborns with perinatal asphyxia treated with therapeutic hypothermia

Background. Hypoxic-ischemic encephalopathy (HIE) is a leading cause of disability in full-term newborns. Long-term consequences of HIE, even when treated by hypothermia, are not easily predictable. Aims. To assess the potential role of electroencephalography and neuroimaging parameters as early predictors of neurodevelopmental outcome in HIE newborns treated with hypothermia. Methods. We retrospectively evaluated 13 HIE patients treated with hypothermia in January 2012-September 2014. We reviewed their amplitude-integrated electroencephalography (a-EEG) at 6, 12 and 24 hours (h), cranial ultrasonography (US) at 12, 72 h and >7 days of life (DOL) and brain magnetic resonance (MRI) performed at 7-28 DOL, according to validated scores. aEEG, US and MRI patterns were correlated to neurodevelopmental outcome at 1824 months, considered as negative if one of the following was present: Mental Development Index (MDI)<85, motor, visual or hearing impairment. Results. The severity of a-EEG, US and MRI alterations at each time point was not different according to the outcome. MDI was negatively correlated with aEEG score at 12h (R= -0.571, p=0.04) and with US score at 72h (R= -0.630, p=0.02). A positive correlation was found between aEEG score at 6h and US score at >7DOL (R=0.690, p=0.013). US alterations of the cortical matter at 72h were directly correlated with a-EEG score at 12h (R = 0.606, p=0.028) and 24h (R=0.605, p=0.029). Conclusions. Early instrumental evaluations, in particular aEEG and US, seem to predict neurodevelopmental outcome at 18-24 months in HIE newborns treated with hypothermia.


INTRODUCTION
Hypoxic-ischemic encephalopathy (HIE), following neonatal asphyxia, is a leading cause of neurological disability in newborns: (1) it can result in serious consequences such as death, cerebral palsy, epilepsy, cognitive, behavioural and developmental deficits.(2,3) Brain damage is an evolutionary process characterized by an acute stage of neuronal necrosis followed by a phase of reperfusion injury characterized by neural apoptosis.Hypothermic treatment limits brain damage, reducing mortality and disability at 18-24 months of life.(4) However, long-term consequences of perinatal asphyxia are not easily predictable.(3) Therefore, the aim of this study was to assess the potential role of amplitude-integrated electroencephalography (a-EEG) background patterns, cranial ultrasonogra-phy (US) and cerebral magnetic resonance imaging (MRI) findings as early predictors of neurodevelopmental outcome at 18 -24 months in neonates undergoing therapeutic hypothermia for HIE.

MATERIALS AND METHODS
We retrospectively analysed data of all the patients who underwent hypothermic treatment from 1 January 2012 to 30 September 2014 in the Neonatal Intensive Care Unit of Maggiore della Carità University Hospital, Novara, Italy.Hypothermic treatment was performed according to the current national guidelines for the management of HIE.(5,6) For each subject, a-EEG, cranial US and cerebral MRI, performed at scheduled time-points, were reviewed according to validated scores.(7)(8)(9) a-EEG A trained nurse or technician applied surface electrodes at C3-C4 locations according to the international 10 -20 system.a-EEG was recorded for all the cooling period and for six hours after rewarming.a-EEG recordings were reviewed by a pediatric neurologist (C.P.) skilled in neonatal a-EEG and blinded to clinical factors.A severity score, based on the classification reported by Murray et al. (7)  The cortical gray matter was classified as normal/mildly abnormal in the absence of signal alterations in T1/T2 or diffusion abnormalities in 2 or less areas (including the central sulcus, the inter-hemispheric fissure and the insular region), and moderately/severely abnormal in the case of more extensive involvement.

Neurodevelopmental evaluation
For all patients, a development evaluation was performed between 18 and 24 months of life by a pediatric neurologist (R.P.

RESULTS
The study population consisted of 13 patients (M:F = 6:7 The a-EEG severity scores at 6, 12 and 24 hours were not different in subjects with positive and negative neurodevelopmental outcomes.Globally, there was an improving trend from 6 to 24 hours, with an increase in slight alterations at the expense of severe ones (table 2).
In regard to the cranial US, the two groups of subjects with positive and negative neurodevelopmental outcomes did not differ in severity of alterations in any anatomical site and at any time (table 3).
In regard to the brain MRI, the two groups did not differ in severity of alterations in any anatomical site.Severe anomalies were recorded at all sites in one patient belonging to the group with a negative outcome (table 4).

CORRELATION ANALYSIS
MDI score was negatively correlated with the severity of a-EEG alterations at 12 hours (R= -0.571, p=0.042, df=11 for twotailed test) and with the severity of cortical alterations evaluated by US at 72 hours (R= -0.630, p=0.021, df=11 for two-tailed test).
The severity of US anomalies of the white matter at 7 DOL showed a direct correlation with the alterations of a-EEG recorded at 6 hours (R=0.690,p=0.013, df=11 for two-tailed test).
The severity of US alterations of the corti-cal matter at t=72h was directly correlated with the severity of a-EEG alterations at 12 hours (R=0.606,p =0.028, df=11 for two-tailed test) and at 24 hours (R=0.605,p=0.029, df=11 for two-tailed test).

DISCUSSION
This study aimed to analyse the presence of early anomalies detected by a-EEG, cranial US and brain MRI, in neonates undergoing therapeutic hypothermia for HIE, and to assess their potential prognostic value with respect to neurodevelopmental outcomes at 18 -24 months.Out of 13 newborns treated with therapeutic hypothermia for HIE, five had a negative outcome at 18-24 months on the basis of a pathological MDI and/or the presence of motor, hearing and/or visual impairment.
The first instrumental parameter taken into account was a-EEG recorded at 6, 12 and 24 hours of birth.No difference was found in the a-EEG severity scores at each time point between subjects with positive and negative neurodevelopmental outcomes.However, it could be supposed that it may be due to the small sample size, as at all times the prevalence of subjects with severe/flat a-EEG tended to be higher in subjects with a negative outcome.An improving trend from 6 to 24 hours was also globally found, with an increase of the slight alterations at the expense of the severe ones, that, however, tended to persist into the negative outcome group.Moreover, the severity of a-EEG alterations at 12 hours was indirectly correlated to MDI at 18-24 months, suggesting the potential prognostic value of a-EEG performed at 12 hours of life.Our findings are consistent with the results reported by Murray et al., (7) highlighting the improvement of EEG alterations from 6 to 48 hours and suggesting the importance of performing early EEG assessments as predictors of outcome after perinatal asphyxia, since normal or mildly abnormal EEG results within 6 hours of life were associated with normal neurodevelopmental outcomes at 24 months.Also, in the systematic review by van Laerhoven et al., (3) a-EEG was the most promising test with a sensitivity of 93% and a specificity of 90% in predicting long term neurodevelopmental outcomes.Regarding imaging methods, no significant difference was found in the severity score of US and MRI alterations at any time and in any site according to neurodevelopmental outcome.However, it should be pointed out that only one sub-ject showed severe alterations at all anatomical sites both on US and MRI, and he developed severe neurological sequelae (cerebral palsy with spastic-dystonic quadriplegia, serious impaired motor functions, slurred speech and feeding difficulties).
The severe alterations were not present on US imaging performed at 24 hours of life, and appeared starting from 72 hours, persisting at the following time-points.From a long-term perspective, correlation analysis showed that MDI at 18-24 months was inversely correlated with the severity of US alterations of cortical matter at 72 hours, that was directly related with the severity of a-EEG alterations at 12 and 24 hours.Moreover, later alterations of the white matter, observed by cranial US at 7 DOL, showed a direct correlation with a-EEG alterations at 6 hours.Our data showed a tight interplay between a-EEG and cranial US in detecting anomalies following HIE.MDI, Mental Development Index.
Table 2. Distribution of a-EEG patterns at 6, 12 and 24 hours in all subjects divided according to positive or negative neurodevelopmental outcome.The main limitation of our study was the small sample size, but all newborns who underwent hypothermic treatment in the considered period were included in the analysis.

All
In conclusion, our data confirm that a-EEG is the earliest method capable of identifying alterations with a prognostic neurodevelopmental value in newborns undergoing therapeutic hypothermia for HIE.Moreover, cranial US performed starting from 72 hours, even with the limitation of low accuracy, should be considered a useful imaging technique for identifying which subjects will be deserving of neurological follow-up.

Table 1 .
Global assessment at 18-24 months of all subjects according to neurodevelopmental outcome.

Table 3 .
Distribution of US patterns at 0, 72 hours and >7 days of life in all subjects, divided according to positive or negative neurodevelopmental outcomes.
DOL, days of life; US, ultrasound.Comparisons between subjects with positive and negative neurodevelopmental outcomes were performed by Fischer test.

Table 4 .
Distribution of MRI patterns in all subjects, divided according to positive or negative neurodevelopmental outcomes., magnetic resonance imaging; PLIC, posterior limb of the internal capsule.Comparisons between subjects with positive and negative neurodevelopmental outcomes were performed by Fischer test. MRI