Predicting motor outcome and death in term hypoxic-ischemic encephalopathy

Abstract

Objectives: Central gray matter damage, the hallmark of term acute perinatal hypoxia-ischemia, frequently leads to severe cerebral palsy and sometimes death. The precision with which these outcomes can be determined from neonatal imaging has not been fully explored. We evaluated the accuracy of early brain MRI for predicting death, the presence and severity of motor impairment, and ability to walk at 2 years in term infants with hypoxic-ischemic encephalopathy (HIE) and basal ganglia-thalamic (BGT) lesions.

Methods: From 1993 to 2007, 175 term infants with evidence of perinatal asphyxia, HIE, and BGT injury seen on early MRI scans were studied. BGT, white matter, posterior limb of the internal capsule (PLIC), and cortex and brainstem abnormality were classified by severity. Motor impairment was staged using the Gross Motor Function Classification System.

Results: The severity of BGT lesions was strongly associated with the severity of motor impairment (Spearman rank correlation 0.77; p < 0.001). The association between white matter, cortical, and brainstem injury and motor impairment was less strong and only BGT injury correlated significantly in a logistic regression model. The predictive accuracy of severe BGT lesions for severe motor impairment was 0.89 (95% confidence interval 0.83-0.96). Abnormal PLIC signal intensity predicted the inability to walk independently by 2 years (sensitivity 0.92, specificity 0.77, positive predictive value 0.88, negative predictive value 0.85). Brainstem injury was the only factor with an independent association with death.

Conclusion: We have shown that in term newborns with HIE and BGT injury, early MRI can be used to predict death and specific motor outcomes.

Figure 1. Basal ganglia and thalami (BGT) injury and motor outcome

Axial T1-weighted images showing (A) mild, focal abnormal signal intensity (SI) in the BGT (arrow) with normal symmetric high SI from the myelinated posterior limb of the internal capsule (PLIC) (short arrows); (B) moderate, multifocal abnormal SI in the BGT (arrows); note the absence of high SI from myelin in the PLIC; and (C) severe, widespread abnormal SI in the BGT. There is no high SI from myelin in the PLIC. The relation between the severity of BGT injury and the severity of motor impairment is shown in table 1 (Spearman rank correlation 0.77; p < 0.001).

Figure 2. Signal intensity (SI) in the posterior limb of the internal capsule (PLIC) and ability to walk at 2 years

Axial T1-weighted images showing (A) normal SI from myelin in the PLIC (arrow); (B) equivocal (reduced) SI in the PLIC (arrow); and (C) abnormal (absent) SI from myelin in the PLIC. The relation between SI in the PLIC and ability to walk at 2 years is shown in table 2 (p < 0.001).

Figure 3. Brainstem injury and death

Axial T2-weighted images at the level of the mesencephalon (top row) and the pons (bottom row) showing (A) normal brainstem (56 infants, none died); (B) moderate brainstem injury (40 infants, 25% died), loss of anatomic details and focal high signal intensity (SI) in mesencephalon (arrows), and excessive differentiation between anterior and posterior pons (double arrow); and (C) severe injury (79 infants, 49% died); widespread abnormal SI in the mesencephalon and the pons (arrows). The difference among the 3 groups was statistically significant (p < 0.001).