Prognostic value of brain magnetic resonance imaging in neonatal hypoxic-ischemic encephalopathy

Iván Sánchez Fernández, J. Leon Morales-Quezada, Samuel Law, Paggie Kim.  Prognostic Value of Brain Magnetic Resonance Imaging in Neonatal Hypoxic-Ischemic Encephalopathy: A Meta-analysis.  Journal of Child Neurology.  In press.

Abstract

Objective:

To quantify the prognostic value of neonatal brain magnetic resonance imaging (MRI) in neonatal hypoxic-ischemic encephalopathy.

Methods:

Meta-analysis of studies with ≥35-week neonates with hypoxic-ischemic encephalopathy who underwent brain MRI within age 4 weeks and had neurodevelopmental follow-up for at least 12 months.

Results:

An abnormal neonatal brain MRI was more frequent among patients with unfavorable neurodevelopmental outcome: odds ratio = 18.2 (95% confidence interval: 9.4-34.9), P <.0001. The prognostic value of neonatal brain MRI in moderate hypoxic-ischemic encephalopathy had an odds ratio of 17.7 (95% confidence interval: 5.3-59.3) and in severe hypoxic-ischemic encephalopathy, the odds ratio was 125.0 (95% confidence interval: 2.0-7917.1). Therapeutic hypothermia did not change the prognostic value of neonatal brain MRI (odds ratio for hypothermia, 14.0 [95% confidence interval: 3.1-63.6], vs no hypothermia, 18.1 [95% confidence interval: 10.0-33.1], P = .7525).

Conclusion:

Neonatal brain MRI provides prognostic information on outcome beyond early infancy in hypoxic-ischemic encephalopathy and therapeutic hypothermia does not change its prognostic value.

_____________________________________________________________________________

From the article

The current meta-analysis shows that (1) neonatal brain MRI in hypoxic-ischemic encephalopathy is highly prognostic of neurodevelopmental outcome beyond early infancy, (2) the prognostic value of neonatal brain MRI is highest in clinically moderate and severe hypoxic-ischemic encephalopathy (though it did not reach statistical significance in mild hypoxic-ischemic encephalopathy in the current study), and (3) therapeutic hypothermia has not significantly changed the value of brain MRI as a biomarker of neurodevelopmental outcome beyond early infancy…

A subgroup of particular interest is newborns with hypoxic-ischemic encephalopathy who undergo therapeutic hypothermia. Therapeutic hypothermia may have improved outcomes through reduction in brain damage, but also has raised the question of whether MRI has reduced or lost its value as a biomarker of neurodevelopmental outcome. Our results demonstrate that the prognostic value of brain MRI as a test remains largely unchanged for patients who undergo therapeutic hypothermia and those who do not. Future studies will provide finer answers evaluating whether specific MRI biomarkers are more or less affected by therapeutic hypothermia, but MRI as a whole test remains equally prognostic…

The current literature does not provide detailed individual data on the time of neonatal brain MRI. Therefore, it is not possible to correct for time of neonatal brain MRI, a relevant potential confounder in the relationship between neonatal brain MRI findings and outcome. The existing literature does not allow to evaluate whether the predictive power of neonatal brain MRI is higher when MRI is performed early after birth or later in the course near discharge. Similarly, our study cannot answer whether neonatal brain MRI is better than other less expensive examinations such as head ultrasonography. Further, the improvement in outcomes with therapeutic hypothermia may or may not have an anatomic correlate in the neonatal brain MRI. The hypothermia trials on neonatal hypoxic-ischemic encephalopathy found a tendency toward less anatomic injury on brain MRI in the hypothermia group that sometimes reached statistical significance and sometimes did not.  For example, in a series of 128 MRIs, fewer newborns had moderate or severe white matter and cortical gray matter abnormalities on T1- and T2-weighted MRI in the hypothermia-treated compared with the normothermia-treated newborns, and this difference was statistically significant. In the same series, although fewer abnormalities on the posterior limb of the internal capsule were detected on T1- and T2-weighted MRI in the hypothermia-treated group, did not reach statistical significance, and there was also a nonsignificant trend to reduction in diffusion abnormalities in the posterior limb of the internal capsule, basal ganglia, white matter, and cortical gray matter in newborns treated with hypothermia.11 Our study shows that the prognostic value of neonatal brain MRI is essentially unchanged in the hypothermia era.

Furthermore, the current study identifies subgroups where MRI might be particularly prognostic of outcome. Neonatal brain MRI has a high prognostic value for outcome in neonates with moderate hypoxic-ischemic encephalopathy, and this finding has a low variability. Neonatal brain MRI also has a high prognostic value in neonates with severe hypoxic-ischemic encephalopathy, although this finding is subject to a high degree of uncertainty and random error. Finally, the prognostic value of neonatal brain MRI did not reach statistical significance for neonates with mild hypoxic-ischemic encephalopathy, and this may provide a rationale for limiting MRIs for this subgroup if resources are limited. Although some studies provided information on clinical severity, only 5 studies stratified MRI findings and outcome by clinical severity. Future studies may refine our results and may narrow the confidence intervals, but the prognostic value of neonatal brain MRI is unlikely to get better for clinically mild hypoxic-ischemic encephalopathy than for more severe hypoxic-ischemic encephalopathy.