Transient cholestasis in newborn infants with perinatal asphyxia

1Division of Gastroenterology and Nutrition and Division of Neonatology, Department of Pediatrics, Sainte-Justine Hospital, Montreal; 2Division of Neonatology, Children’s & Women’s Health Centre of British Columbia, Vancouver, British Columbia Correspondence: Professor F Alvarez, Division of Gastroenterology and Nutrition, Department of Pediatrics, Sainte-Justine Hospital, University of Montreal, 3175 Côte Sainte-Catherine, Montréal, Quebec H3T 1C5. Telephone 514-345-4931, fax 514-345-4999, e-mail alvarezf@ere.umontreal.ca Received for publication January 15, 2002. Accepted January 17 2003 D Herzog, P Chessex, S Martin, F Alvarez. Transient cholestasis in newborn infants with perinatal asphyxia. Can J Gastroenterol 2003;17(3):179-182.

P athological neonatal cholestasis is diagnosed as often as one in 500 newborn infants (1).It is a nonspecific feature of a liver disorder that occurs with several congenital diseases such as biliary atresia, α1-antitrypsin deficiency and other metabolic, endocrine or perinatal infectious maladies.Newborn infants with transient neonatal cholestasis have recently been described, most of whom suffered from asphyxia and/or prematurity (2)(3)(4).Though transient, neonatal cholestasis presents in a way that is indistinguishable from that occurring with more severe causes; it is self-limiting and needs neither further investigation nor therapy other than the substitution of lipid-soluble vitamins.Therefore, this condition deserves more precise characterization.Until now, its definition relied exclusively on the serum level of conjugated bilirubin and on the exclusion of the aforementioned diseases.The aim of the present study was to look at the incidence of transient neonatal cholestasis in children with perinatal asphyxia, and to search for associated risk factors.

METHODS
To select charts for review, a database was used that registered each patient admitted to the neonatology department between 1989 and 1993 as well as definitive diagnoses and complications.Asphyxia was defined by the presence of postanoxic encephalopathy defined by an abnormal Apgar score (a one-minute Apgar score of less than four, a five-minute Apgar score of less than seven, and/or the need for immediate resuscitation or intubation); clinical signs of encephalopathy; and its confirmation by appropriate examinations such as electroencephalogram and brain imaging techniques (echography, computed tomography).Cholestasis was defined as conjugated bilirubin of more than 18 µmol/L and/or greater than 20% of the total bilirubin concentration (1,5,6).Bilirubin was measured in the clinical biochemistry laboratory at Sainte-Justine Hospital, Montreal, Quebec by automated colorimetric methods (Beckman Synchron CX-5, Beckman Coulter Canada Limited, Missisauga, Ontario).Cholestatic newborns with proven inborn errors of metabolism, infections, malformations and hypothyroidism were excluded from the study.
The results were statistically evaluated by the Mann-Whitney U test for comparision of groups with different sizes and without normal distribution, by the χ 2 test for comparison of group sizes and by logistic regression for multivariate analysis.

RESULTS
Of a total of 6737 patients admitted to the neonatology department of Sainte-Justine Hospital between 1989 and 1993, and registered in a neonatology database, 256 (3.8%) were classified as asphyxiated, 659 (9.78%) as small neonates for their gestational age (SGA), and 24 (0.36%) as a combination of both.Of all those registered cases, 212 charts of asphyxiated neonates with appropriate weight for their gestational age (AGA) (90 females), and 22 (12 female) charts of SGA neonates were available for study (after exclusion of one with polycystic kidney disease, one with generalized cytomegalovious infection, and one with urea cycle disease).Of the AGA and SGA cases, 49 (20 female) and four (two female), respectively, died within the first 48 h of life.Thus, data were analyzed from a total of 181 (163 AGA and 18 SGA) asphyxiated infants surviving the first 48 h of their life.
A total of 14 (8.5%, four term and 10 preterm infants) asphyxiated AGA infants and six (33%, one term, five preterm) SGA infants had cholestasis.In both groups, cholestasis developed more frequently in premature newborns.The characteristics of infants with cholestasis are given in more detail in Table 1.

Further characteristics
Alanine aminotransferase and aspartate aminotransferase were measured in eight of 20 cholestatic infants and normalized in all after the end of total parenteral nutrition (TPN) (Table 1).Bilirubin values were controlled in 13 of 20 cholestatic patients until normalization occured.
Emergency cesarean section was performed in 55.2% of pregnancies with asphyxiated neonates (82 in AGA and 18 in SGA), and in 12 of 20 (60%) children with additional transient neonatal cholestasis.
Intubation and ventilation were necessary in 113 of 163 (69.0%) of the asphyxiated AGA infants and in 16 of 18 (89%) of the asphyxiated SGA neonates.The average duration of ventilation in AGA children without cholestasis was 5.4 days (median one day, range zero to 91 days) versus 20.5 days (median 13 days, range five to 114 days) in those with   Antibiotic therapy was initiated in 124 of 163 AGA infants (75.6%), in 100% of SGA asphyxiated neonates, and in 100% of infants with cholestasis.Antiepileptics were used in an equal proportion of asphyxiated newborns of each group (six of 18 (33%) SGA, 48 of 163 (29%) AGA and in five of 20 (25%) infants with cholestasis.All infants required nasogastric tube feeding for enteral nutrition with Similac (Abbot Laboratories, St Laurent, Quebec) or mother's milk.If oral feeding was not tolerated by day 4, TPN was initiated.TPN was started in 44% (41 of 93) of premature asphyxiated AGA infants, in 8% (seven of 90) of term asphyxiated AGA neonates, and in 14 of 18 (78%) SGA infants, compared with 18 of 20 (90%) infants with cholestasis at diverse moments after birth.

Follow-up
Seven of 18 cholestatic children had only one measurement of direct bilirubin before starting TPN; there was no further increase in six of 18 infants, whereas in five of 18 infants direct bilirubin was elevated after starting TPN.The latter group had controls until normalization after the end of TPN.Three children died during the first year of life, and follow-up time of the remaining children was between one and 12 years.Follow-up was necessary for neurological, behavioural, visual, auditive and infectious problems, requiring liver function tests in three patients (one growth failure, two antiepileptic treatments), which were normal.No overt chronic liver disease developed.
The control group consisted of SGA neonates without asphyxia (n=635).Cholestasis was found in 25 (3.94%) children.It was associated in one case with Nieman-Pick disease and in another with biliary atresia, in seven with intestinal diseases needing surgical intervention, in four with complex cardiopathies, in nine with infections, in all infants with TPN and in three patients without TPN, but with ventilation and antibiotic therapy.Five patients died of suspected infection, and three were autopsied.These revealed cirrhosis in the patient with biliary atresia and giant cell hepatitis in another with congenital infection.Follow-up disclosed chronic hepatitis B in one patient.

DISCUSSION
Neonatal cholestasis is a condition that demands extensive work-up to exclude inborn errors of metabolism, intrauterine infections, endocrine dysfunction or malformation.Frequently, these investigations do not yield positive findings, and if the cholestasis resolves spontaneously, the condition is called transient neonatal cholestasis.Asphyxiated infants seem to be at risk for this type of transient cholestasis (7).The present study defines the incidence of transient neonatal cholestasis in asphyxiated infants.
Transient neonatal cholestasis was found in 8.5% of AGA and in 33% of SGA asphyxiated neonates.Cholestatic neonates of both groups were severely affected, as shown by longer periods of mechanical ventilation and a larger number of patients requiring antibiotic therapy and TPN, when compared with asphyxiated newborns without cholestasis.It therefore seems that severe illness is an additional risk factor for the development of cholestasis in asphyxiated neonates.Other factors, such as family or pregnancy history, could not explain the severe perinatal illness or the occurrence of cholestasis in these neonates.Prematurely-born infants were more frequently cholestatic than children born at term, but the difference by univariate analysis was significant only in AGA infants.Multivariate analysis showed an increased risk for all infants born before the age of 35 weeks.Using data from the same registry, we found that elevated direct bilirubin, including all etiologies, occurred in 3.9% of nonasphyxiated SGA neonates, and transient neonatal cholestasis in three (4‰).The mechanism leading to cholestasis in asphyxia is not well understood.It is thought to be of multifactorial origin.Inefficient enterohepatic circulation, immaturity of bile secretion (8) and lack of enteral nutrition (7) may combine with fetal distress and reduced splanchnic blood flow, inducing poor oxygenation of the liver and bowel.The subsequent decreased expression of bile acid transport proteins of basolateral and apical hepatocyte membranes may further impair the immature secretion of bile acids (9)(10)(11) and enhance physiological cholestasis in the neonate.
Most of the infants (13 of 20) qualifying for the diagnosis of transient neonatal cholestasis had levels of conjugated bilirubin that slightly exceeded the upper normal limit.None of them had acholic stools, and liver function tests were not done routinely.However, none of the results obtained called for further investigations.None of these children had liver biopsy, and none of them developed liver disease.Thus, the population analyzed in the present study may differ not only in the severity of cholestasis from that described by Jacquemin et al (2), but also in its duration, in the degree of neonatal distress and in comorbidity; we can assume that patients seen at Bicêtre (Paris, France) (2) represent those referred with more severe neonatal cholestasis whereas the patients described in the present study represent the patient population of a mother-child centre.In spite of advances in the molecular mechanisms of cholestasis, the exact etiology of asphyxia and, therefore, of transient neonatal cholestasis in asphyxiated neonates remains unknown, and the reason for cholestasis may not be the same in the patients from Bicêtre and in our population.

CONCLUSIONS
We found that asphyxia, duration of ventilatory support, gestational age of less than 35 weeks and male sex are risk factors for transient neonatal cholestasis.This, in conjunction with intrauterine growth retardation, increases the risk for newborn infants to develop transient neonatal cholestasis by 33%.Transient neonatal cholestasis is thus by far the most frequent etiology of cholestasis in asphyxiated SGA newborn children.
Herzog et al Can J Gastroenterol Vol 17 No 3 March 2003 180