A clinical pathologic reappraisal of kernicterus.

Kernicterus at autopsy, traditionally associated with hyperbilirubinemia, is now often observed in the absence of markedly elevated levels of serum bilirubin. Attempts to document clinically predictive risk factors for kernicterus have been largely unsuccessful in the current population of sick neonates. In a study of 32 pairs of newborns with and without kernicterus at autopsy and matched for gestational age, weight, length of survival, and year of birth, no differences were found when multiple clinical factors thought to potentiate risk for kernicterus were compared. Microscopic sections of brains from these matched pairs were then evaluated without knowledge of the clinical or gross findings according to prearranged criteria. Statistic analysis of the microscopic findings was done, comparing each observed change in each anatomic site, all changes in each site, and each change in all sites. Although the gross pattern of staining followed that of classic kernicterus, the expected, specific histologic changes characteristic of kernicterus were found in only three patients. Spongy change, or edema, was significantly more common in the grossly stained brains (P less than .0005), but other findings were similar in both groups. Gross bilirubin staining of the neonatal brain with neither hyperbilirubinemia nor specific microscopic changes is probably not the same clinicopathologic entity as classic kernicterus.