Cholestasis in infants at risk

Abstract

Cholestasis is a condition when bile flow from the liver to the intestines is impaired. Unresolved, it may lead to liver transplantation or death. Overall, it affects 1 out of 2500 term infants, but is far more common in high-risk infants with other conditions such as prematurity and hemolytic disease of the fetus and newborn (HDFN). In preterms and other sick infants in the neonatal intensive care unit (NICU), cholestasis is associated with prolonged parenteral nutrition (PN), where the source of lipids used in PN has been implied in the pathogenesis. Bacterial infection is also a risk factor, but the role of hepatotropic viruses, such as cytomegalovirus (CMV), is less well studied. In infants with HDFN and cholestasis, very few studies are published. Reports on short- and long-term outcomes of cholestasis in specific infant groups at risk are scarce. The aim of this thesis was to study incidence, risk factors, course and outcomes of cholestasis in preterm infants in the NICU and in infants with HDFN.

In two population-based retrospective case-control studies, we included cholestatic very preterm infants (gestational age < 30 weeks) in the Stockholm region during two two-year periods in 2006-2008 and 2010-2011. In paper I, we found an incidence of cholestasis of 14.8% (37/250) in 2006-2008 when a soy-based lipid emulsion was used, and 12.7% (34/268) in 2010-2011 when an olive-based one was used (p=0.52). In a multivariable model we found that the olive-based lipid emulsion might carry a slightly reduced risk for cholestasis when adjusting for the risk factors necrotizing enterocolitis and PN duration. In paper II, outcomes of cholestasis were evaluated in the cohort born 2006-2008. Cholestatic infants had higher rates of bronchopulmonary dysplasia (BPD) and retinopathy of prematurity (ROP) than controls, and mortality was 13.5% versus 2.7% (p=0.040). No significant long-term liver disease was found in those surviving the neonatal period. In paper III, we prospectively analyzed CMV DNA in blood and urine samples from cholestatic preterm infants (<37 weeks) in two NICUs in 2008-2010. Non-cholestatic controls were sampled in 2015-2017 at a similar age. In 69 % (31/45) of cholestatic infants, CMV DNA was detected in at least one sample, versus 13% (3/24) in non-cholestatic controls (p<0.00001). Mortality was 26% among CMV positive cholestatic infants, whereas none of the CMV negative cholestatic infants died (p=0.044). In paper IV, a retrospective population-based cohort study, we evaluated the incidence of cholestasis in infants with HDFN in the Stockholm region in 2004-2015. Cholestasis was found in 7% (11/149), with an early onset. Intrauterine blood transfusions and multiple maternal antibodies including D-, c- or K-antibodies were risk factors for cholestasis.

We conclude that cholestasis is common in very preterm infants and that switching from soybased to olive-based parenteral lipid emulsion affects the incidence only marginally. CMV seems associated with cholestasis in preterms and further scientific attention should be encouraged. Cholestasis is common in infants with HDFN, and this warrants screening for early detection and correct management. Cholestasis may be associated with a more severe short-term outcome in preterms, but liver disease later in childhood is uncommon in both preterms and infants with HDFN.