Video Abstracts -- brief videos summarizing key findings of new articlesKernicteric Findings at Autopsy in Two Sick Near Term Infants
We present two near term sick infants with unanticipated kernicterus noted at autopsy, which developed in the absence of hemolysis and at serum bilirubin levels that would not have been considered neurotoxic. The kernicterus evolved in the context of known experimental conditions that would have favored the passage of unbound bilirubin across a disrupted blood brain barrier with injury of specific neurons.
The relationship between serum bilirubin concentrations and the development of kernicterus in the term infant is complex. While a distinct link between the maximum recorded level of serum bilirubin and the occurrence of kernicterus is apparent with hemolytic disease of the newborn (most often due to Rh isoimmunization), the risk is less clear in markedly jaundiced babies without hemolytic disease.1-6Moreover, in some instances, clinical and pathologic kernicterus has been observed in premature infants without hemolytic disease and with low bilirubin levels.7,8 The management of jaundice in sick term newborns is a challenge to clinicians vis a vis the risks of undertreatment and overtreatment. To underscore the complexity of this important clinical relationship, we present two patients of unanticipated kernicterus noted at autopsy which evolved in two sick infants, one near term and the other term, in the absence of overt hemolysis and at serum bilirubin concentrations that usually would have not been considered neurotoxic.
CASE HISTORY
Patient 1
BM was a 2210-g, 37-week-old, Latin-American female, born to a 20-year gravida 4, para 3 mother, whose pregnancy was uncomplicated. The infant was delivered vaginally and was meconium stained. The Apgar scores were 9 at 1 and 5 minutes, respectively. Upon delivery the infant was noted to be wasted, with a distended abdomen. No masses could be palpated. An initial hematocrit was 70% and a partial exchange transfusion was performed. The post transfusion hematocrit was 64%. On the second postnatal day, the infant was noted to be lethargic with an erythematous abdomen. A sepsis workup including a spinal tap was undertaken and antibiotics initiated. As part of the initial workup, the infant was noted to have proteinuria on urine dipstix; in addition, a total protein of 2.9 g/dL and an albumin less than 1 g/dL was noted. On the third postnatal day, there was an acute deterioration in the clinical condition characterized by poor perfusion, bradycardia, hypotension, metabolic acidosis, stupor, and the development of seizures, ie, right-sided myoclonic jerking of the upper extremity was observed. Twenty mg/kg of phenobarbital was administered with cessation of the seizure-like activity. The infant was intubated, was started on dobutamine, and an intravenous infusion of albumin 1 g/kg was initiated and repeated every 8 hours in an attempt to raise the low serum albumin levels. The infant was started on hyperalimentation and intralipid infusion on the third postnatal day. The infant remained stuporous over the subsequent four days. An electroencephalogram on the fourth postnatal day revealed abnormal background activity with poor interhemispheric synchrony, and prolonged periods of attenuation. Multifocal sharp waves were also noted without seizure-like discharges. A second electroencephalogram a day later revealed bitemporal seizure activity and interictal left temporal seizure discharges. Laboratory values were pertinent for the following: persistent hypoalbuminemia with a serum albumin less than 1.5 g/dL and hyponatremia with serum sodium values ranging from 122 to 124 mEq/L. Liver enzymes were normal and serum bilirubin levels on the first and third postnatal days were 5.2 mg/dL and 4.8 mg/dL respectively, creatinine kinase on the fourth postnatal day was 988 u/L (normal 38–120), the creatinine kinase-MB cardial fraction 16 ng/mL (normal 0–3); serum ammonia was normal. Blood and spinal cultures were negative. There was no evidence of hemolysis. A cranial ultrasound scan on the third postnatal day was normal; a repeat scan two days later demonstrated increased echogenicity involving the left frontoparietal region within periventricular white matter. Because of a persistent unresponsive state, life support was withdrawn on the eighth postnatal day.
The autopsy was pertinent for evidence of a congenital nephrotic syndrome of the Finnish type with bilateral nephromegaly, tubular cysts and mesangial proliferation. The respiratory system was characterized by edematous alveolar septa and intra-alveolar bright yellow pigment consistent with bilirubin. The liver revealed focal areas of bile stasis within the bile ducts, hepatocytes, and canaliculi. There was edema of the brain, a hematoma involving the left centrum semiovale with contiguous white matter necrosis without cortical involvement (Fig. 1). Macroscopically, marked bilirubin staining within the hippocampus, floccular lobes, inferior cerebellar peduncles and subthalamic nuclei was visible (Fig. 2). Microscopically, the white matter within the frontal cortex adjacent to the hematoma contained patchy areas of karyorrhexis, gliosis, rarefaction, and hemorrhage. Within midbrain and hippocampus, bilirubin pigment was readily identified intracellularly within the neuropil and within the cytoplasm of macrophages and residual viable neurons (Fig.3). Focal neuronal necrosis of the midbrain and pons was also noted. Sections of hippocampus revealed extensive necrosis of the pyramidal cell layer.
Hemorrhage involving left semiovale with contiguous white matter necrosis. Note the sparing of cortical involvement.
Coronal section of cerebrum demonstrating yellow staining in the region of the hippocampus.
Bilirubin pigment evident in the cytoplasm of a histologically viable pyramidal neuron (Hematoxylin and Eosin × 1350).
Patient 2
BG was a 1652-g, 34-week-old, small for gestational age Latin-American female, born to a 17-year gravida 1, para 0 mother. The pregnancy was complicated by the antenatal diagnosis of a gastroschisis. The infant was delivered vaginally, and the Apgar scores were 9 at 1 and 5 minutes, respectively. The infant underwent surgery for reduction of the gastroschisis on the first postnatal day. At surgery, the small intestine was discolored brown and of questionable viability; jejunal and colonic atresias were also noted. The postoperative course was complicated by hypotension requiring fluid replacement therapy on the first postnatal day. However, by the second day, the blood pressure was stable, and the infant had rapidly weaned to room air and low ventilator support. The neurologic examination was unremarkable. Hyperalimentation including intralipid infusions were started on the second postnatal day. The infant remained stable except for a persistent metabolic acidosis with serum bicarbonate levels of 14–15 mEq/L. There was no improvement in the bowel perfusion. On the fourth day, the infant appeared acutely stuporous and required increased ventilator support secondary to progressive respiratory failure with an increase in PCO2 from 40 mm Hg to 75 mm Hg. A total bilirubin level at the time was 14.4 mg/dL with a conjugated fraction of 1.6 mg/dL. Values on the following day were unchanged. There was no evidence of hemolysis. The infant was removed from life support on the fifth day.
At autopsy, thrombosis of the superior mesenteric vein with infarction of the small intestine was noted. Additionally, abundant yellow pigmented material consistent with bilirubin was observed in the distal air spaces of the lung. The brain was characterized by macroscopic yellow-green discoloration of the septum pellucidum, hippocampi, anterior temporal lobes, basal ganglia, mid brain, pons and lateral ventricular walls. Microscopic evaluation revealed extensive neuronal necrosis within the pyramidal cell layer of the hippocampus, associated with the presence of bilirubin pigment within the cytoplasm of macrophages and necrotic neurons. Viable neurons were difficult to identify within the pyradimal cell layer, in contrast to patient 1, precluding evaluation of the question of bilirubin staining of neurons before cell death.
DISCUSSION
In the full-term infant, it has been generally accepted that serum levels of unconjugated bilirubin must be elevated to cause neurotoxicity. Specifically, with hemolytic disease of the newborn (most often due to Rh isoimmunization), a distinct relationship between the maximum recorded level of serum bilirubin and the occurrence of kernicterus is apparent.1,2 However, the neurotoxic risk of marked hyperbilirubinemia in full-term infants in the absence of hemolysis is considered to be far less.3,4 Indeed the aggressive treatment of jaundice in term infants without hemolysis has been questioned, and it has been suggested that such infants are not at risk of mental impairment until bilirubin rises well above 20 mg/dL.5,6,9,10 The autopsy findings of kernicterus in this report are unique in that the brain injury developed in two sick infants, one near term and one term, at serum bilirubin levels that would not have been considered clinically neurotoxic, and in the absence of hemolytic disease. The observations are similar to the bilirubin staining of basal ganglia noted previously in sick extremely very premature infants at low bilirubin levels and in the absence of hemolysis.7,8
The critical factor in the genesis of bilirubin neurotoxicity is the entry of unbound bilirubin into brain and subsequent exposure to neurons.11,12 More specifically, it is the protonized form of the unbound bilirubin, ie, bilirubic acid, that is largely responsible for tissue binding and probably the neuronal injury.13,14 Factors that modulate the potential risk for neurotoxicity include: a) the quantity of unconjugated serum bilirubin, b) the binding of bilirubin to albumin, and c) the status of the blood brain barrier. The causal relationship of one or more of these factors in the genesis of the kernicterus observed in this report is discussed below.
The quantity of unconjugated bilirubin does not appear to be a primary factor in either of the present cases. In patient 1, the maximum serum bilirubin level obtained concurrent with the neurologic deterioration was well within a physiologic range, ie, 5 mg/dL. In patient 2 the level of bilirubin, although elevated, was not at a concentration that would have warranted an exchange transfusion.
Lack of bilirubin binding by albumin would likely have caused a relative increase in the levels of unbound bilirubin in the patient 1. Bilirubin is bound to albumin by a primary high affinity binding site, and one or two weaker lower affinity binding sites.15,16The binding capacity of albumin for bilirubin is influenced by the concentration of albumin and the binding affinity of albumin for bilirubin. With regard to the albumin concentration, at lower values of serum albumin, there is more available unbound bilirubin anion, and ultimately bilirubin acid.11,13,17 Moreover, in experimental conditions, the neurotoxic effects of bilirubin on cultured cells can be reversed by adding albumin.18 The profound hypoalbuminemia in patient 1, which was secondary to a congenital nephrosis of the Finnish type, clearly favors the formation of free bilirubin and potentially increases the risk for neurotoxicity. Congenital nephrosis has been occasionally associated with neurologic manifestations, ie, cerebral edema,19 developmental abnormality,20,21 and kernicterus which has only been described as a complication in a single patient.22 This suggests that additional factors were likely involved in this case (see below). Bilirubin affinity for albumin, can be reduced by endogenous and exogenous competitors such as nonesterified fatty acids, organic acids, and certain medications, eg, sufonamides, salicylates, and ceftriaxone, none of which is likely to have been a factor in either case.23-27
The status of the blood brain barrier would also appear to be an important potential factor in the genesis of bilirubin neurotoxicity in the two patients. Normally the blood brain barrier, comprised of brain capillary endothelial cells with characteristic tight junctions, limits and controls the movement of molecules such as bilirubin between the blood and the brain. Depending on the clinical circumstances, bilirubin may enter the brain across either an intact or disrupted blood-brain barrier. Both situations may have been operative in the two cases. Passage of bilirubin across an intact blood-brain barrier is more likely to occur when there are elevations in free bilirubin as the bilirubin anion readily binds to phospholipid forming lipophilic complexes.13,28-31 Increased bilirubin transport across an intact blood-brain barrier may also occur under conditions of increased cerebral blood flow. Indeed, studies in neonatal piglets subjected to moderate hypercarbia with increases in cerebral blood flow were accompanied by an increase in bilirubin transport into the brain without a concomitant increase in albumin uptake.31 The profound hypoalbuminemia in patient 1 with potential free bilirubin elevation, and the respiratory acidosis with moderate hypercarbia in patient 2 concurrent with the acute neurologic deterioration lends credence to the potential passage of bilirubin across an intact blood brain barrier. Conversely, experimental disruption of the blood brain increases the concentration of bilirubin into brain even when still bound to albumin.28,30-35 The blood brain barrier can be disturbed with asphyxia, meningitis, etc.36-39 The presence of global hypoxic ischemic injury manifested by neuronal neurosis likely contributed to disruption of the blood brain barrier in these cases.
Irrespective of the principal mechanism(s), both patients demonstrated striking bilirubin staining of the brain parenchyma at autopsy. Classic bilirubin staining was observed in both patients within the hippocampus, basal ganglia subthalamic nuclei and cerebellar peduncles.40 There was microscopic evidence of neuronal necrosis accompanied by deposits of intracellular neuronal deposition of bilirubin as well as extra cellular bilirubin pigment in many areas of patient 1. In patient 2, although there was fairly widespread yellow staining grossly, microscopic bilirubin deposits were most conspicuous in the pyradimal cell layer of hippocampus, a region in which extensive neuronal necrosis was also present. In addition, bilirubin pigment was identified within the perikarya of histologically viable neurons in adjacent areas of Ammon's horn, indicating that the hippocampal bilirubin deposition was not exclusively due to passive staining of previously necrotic parenchyma (Fig. 3). The tendency of bilirubin deposits in classic kernicterus to localize in areas vulnerable to hypoxic ischemic injury, such as the pyramidal cell layer of the hippocampus, raises the important question of whether hypoxic-ischemic injury plays a role in the development of the lesions of kernicterus. The consistent coexistence of hypoxic ischemic changes and bilirubin damage suggests that the former may be an important predisposing mechanism for the tissue binding and neuronal injury.7Thus, while yellow staining may reflect opening or disruption of the blood brain barrier, additional events such as hypoxia-ischemia may be necessary for the full expression of the bilirubin neurotoxicity, ie, intracellular bilirubin pigment within neurons, especially in the absence of hyperbilirubinemia. However, an alternative explanation that in some cases the deposition of bilirubin may be an epiphenomenon associated with the hypoxia-ischemia, and not the primary cause of the neuronal injury is particularly relevant in patient 2.
In conclusion, we present the unanticipated finding of kernicterus in two sick infants, one near term and one term, which developed in the absence of hemolysis and at serum bilirubin levels that would not have been considered neurotoxic. The kernicterus evolved in the context of known experimental conditions that would have favored the passage of unbound bilirubin across a disrupted blood brain barrier with resultant injury of specific vulnerable neurons.
REFERENCES
- Copyright © 1997 American Academy of Pediatrics
