Cytomegalovirus in biliary atresia is associated with increased pretransplant death, but not decreased native liver survival

Background:

Biliary atresia (BA) is likely caused by a common phenotypic response to various triggers; one proposed trigger, cytomegalovirus (CMV), may lead to worse outcomes. The aim of this study was to determine the severity of disease and pretransplant outcomes of infants with BA, who have evidence of CMV (CMV+) at diagnosis compared with CMV-negative (CMV−) infants.

Methods:

The study used data and biospecimens from the Childhood Liver Disease Research Network PROBE study of cholestatic infants. Plasma obtained at the time of hepatic portoenterostomy (HPE) of 249 infants with BA was tested for CMV by DNA-PCR and CMV-IgM. Comparisons between CMV+ and CMV− infants were made using Wilcoxon rank sum, Student t test, chi-square, or Fisher exact test. Native liver survival (NLS) outcomes were analyzed using Kaplan-Meier and Cox regression adjusting for age at HPE; pretransplant patient survival outcomes were analyzed using a competing risk model and adjusting for age at HPE.

Results:

CMV+ infants (n = 29, 12%) underwent HPE later (67.8±13.6 d vs. 55.1±18.5 d, p = 0.0005) and had higher baseline alkaline phosphatase and aminotransferases. There was no difference between groups in jaundice clearance or NLS. The subdistribution HR of pretransplant death for CMV+ infants adjusted for age at HPE was 3.8 (p = 0.034).

Conclusions:

CMV infection at the time of HPE in infants with BA is not associated with worse NLS despite the association with worse liver injury, older age at HPE, and increased risk of pretransplant death adjusted for age at HPE. Continued evaluation of the consequences of CMV infection and the effects of antiviral treatment should be explored.

Biliary atresia (BA) is a progressive fibroinflammatory obliterative cholangiopathy occurring only in infants. While the exact pathogenesis of BA remains unknown, it likely represents a common phenotypic response to various prenatal and perinatal insults, including genetic mutations, infections, vascular, or toxic insults.1 Infectious triggers may not only result in direct viral damage to the cholangiocyte but may also trigger an autoimmune proinflammatory process that causes ongoing liver damage, even after the initial infection has been cleared.1 One infectious agent that has been detected in 10%–38% of infants with BA at the time of diagnosis is cytomegalovirus (CMV).2–11 There is evidence to suggest that the immune response triggered by CMV may perpetuate the inflammatory obstruction of extrahepatic and intrahepatic bile ducts in BA.4,6,9,10,12–14

Whether the presence of CMV infection at the time of diagnosis of BA holds a prognostic value for post-hepatic portoenterostomy (HPE) outcomes is uncertain. A single center report of 121 infants with BA from the United Kingdom suggested that infants with BA who were CMV IgM+ at presentation were older at diagnosis of BA, had more significant liver inflammation and advanced scarring on histology, and markedly poorer outcome after HPE with decreased native liver survival (NLS) and increased mortality compared with those infants who were CMV− at presentation.8 A more recent meta-analysis of 9 studies, including 694 evaluable infants with BA, reported that patients with CMV exhibited significantly lower jaundice clearance rates following HPE but had no significant difference in NLS.15 The aim of this study was to expand upon prior studies to determine the relationship between CMV status at the time of diagnosis of BA and clinical outcomes among a large multicentered cohort of infants from North America. Specifically, we sought to determine whether CMV+ infants formed a prognostic subgroup amongst children with BA and, if so, if that group had worse baseline liver disease and/or outcomes, including clearance of jaundice, NLS, and pretransplant patient survival.

EXPERIMENTAL PROCEDURES

Participant selection

Participants enrolled in the NIDDK-supported Childhood Liver Disease Research Network (ChiLDReN, http://childrennetwork.org) longitudinal study entitled Prospective Database of Infants with Cholestasis (PROBE: NCT00061828) were considered for inclusion in the present study. ChiLDReN includes 13 clinical sites in the United States and 1 in Canada. In PROBE, cholestatic infants <6 months old are enrolled prospectively before undergoing liver biopsy and cholangiography and then followed longitudinally for up to 20 years. Inclusion criteria for this study included the diagnosis of BA (confirmed by intraoperative cholangiogram and bile duct histology) in patients who underwent HPE before 90 days of life and had study follow-up until at least 24 months of age if they survived with the native liver to 24 months.

Participants eligible for this study were enrolled in PROBE between June 2004 and September 2019 at the time of identification of neonatal cholestasis before HPE. As part of PROBE, plasma specimens were collected immediately before or during the HPE procedure, processed, frozen, and stored at −70 °C at the NIDDK Biorepository. Written informed consent was obtained from the parent/legal guardian of each participant, and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki. Each ChiLDReN site underwent approval by their respective Institutional Review Boards or equivalent. The current ancillary study was approved by the ChiLDReN Ancillary Studies Committee and ChiLDReN Steering Committee (August 11, 2020).

Demographic and clinical information

Demographic and clinical information was collected prospectively for each participant. Demographic information included sex, ethnicity, and race. Clinical information included age at HPE, the presence of extrahepatic anomalies/malformations, complications of BA (including infections, portal hypertension, and poor growth), laboratory data, tissue pathology when available, and clinical outcomes. Length, weight, head circumference, mid-arm circumference, and triceps skinfold thickness were recorded at baseline and at annual follow-up visits. Laboratory data were collected prior to or at the time of HPE, at 3 (±2 wk) and 6 (±1 mo) months post-HPE, at age 12 (±1 mo) months, and annually (±6 mo) thereafter. Jaundice clearance was defined as serum total bilirubin <2 mg/dL at any time before 3 months post-HPE or at subsequent study visits. When available, the central ChiLDReN Pathology Committee reviewed slides of liver tissue obtained at HPE and assigned an Ishak fibrosis stage (0–6). Outcome data included age at the listing for liver transplantation, age at transplantation, age at death, and cause of death where applicable. Participants remain in the study until 20 years of age, liver transplantation, or death.

Plasma testing for CMV status

Plasma samples obtained at the time of PROBE enrollment and stored at the NIDDK Central Biorepository at −70 °C were subsequently shipped on dry ice to the University of Colorado Anschutz Medical Campus. Both CMV PCR and IgM testing were done in order to improve the specificity of results.16 Real-time quantitative PCR testing for CMV DNA was performed on plasma using the Light Cycler 2.0 instrument (Roche Diagnostics GmbH, Sandhofer Strasse 116, Mannheim, Germany) as described.17 Positivity for CMV was defined by any qualitatively positive results. CMV IgM testing was also performed by ELISA using a Gold Standard Diagnostics ELISA kit (Davis, CA, USA) as per the manufacturer’s instructions; positive was defined as an index value between 1.1 and 3.4. When IgM and PCR testing were discordant, CMV IgG testing was performed using a Gold Standard Diagnostics ELISA kit (Davis, CA, USA) as per the manufacturer’s instructions, and positive was defined as index value >1.1. CMV positivity for a given participant was defined as either CMV DNA PCR+ or IgM+. CMV negativity was defined as both PCR− and IgM−. If results between PCR and IgM were discrepant, IgG was run, and if negative, that participant was considered CMV−.

Statistical analysis

Based on previous studies, we estimated that 249 participants would allow for at least 80% power to detect significant differences between the CMV+ and CMV− groups for clearance of jaundice, NLS, and pretransplant patient survival. The ChiLDReN Scientific Data Coordinating Center (SDCC) chose 249 participants at random out of 354 potential PROBE participants who met the enrollment criteria for this study and had adequate plasma available from the PROBE study. Comparisons between CMV+ and CMV− participants at baseline and follow-up time points were made using the Wilcoxon rank sum test or Student t test for continuous variables and the chi-square test or Fisher exact test for categorical variables. NLS outcomes were analyzed using the Kaplan-Meier method univariately and Cox regression while adjusting for age at HPE. For Kaplan-Meier analyses, participants were first grouped solely based on CMV status (positive or negative) and then were grouped based both on CMV status (positive or negative) along with age at HPE (< or ≥60 d of life). NLS was defined as the period from birth until liver transplantation, death, or last visit date for censored observation (whichever came first). Pretransplant patient survival outcomes were analyzed using a competing risk model to avoid the potential biases inherent to the Kaplan-Meier and Cox proportional hazard analyses. Nonparametric competing risk analysis technique (Gray18; Fine and Gray19) was used to assess the cumulative incidence function (CIF) for death before transplantation. In this analysis, the occurrence of transplantation was considered a competing risk for death before the transplant or vice versa. The outcome measure is months from HPE to one of the 3 events, whichever comes first: death before transplant, liver transplantation, and being seen alive before transplantation (ie, censored observation). Patients were first grouped solely based on CMV status (positive or negative) and then were grouped based both on CMV status (positive or negative) along with age at HPE (< or ≥60 d of life). The between-group difference in CIF was tested using Gray test.18 Moreover, Fine and Gray19 survival analysis was used to model the outcome while controlling age at HPE as a continuous variable.

RESULTS

Participant CMV status

At the time of diagnosis of BA, 29 of 249 children (12%) were classified as CMV+: 15 were PCR+/IgM+, 12 were PCR+/IgM−/IgG+, and 2 were PCR−/IgM+/IgG+. A total of 220 study participants were classified as CMV−.

Participant demographics and clinical information

There were no differences in enrollment in PROBE between CMV+ and CMV− groups by sex, ethnicity, or race (Table 1). Baseline weight and head circumference were slightly, but significantly, higher in CMV+ participants; however, these differences were not deemed to be clinically meaningful (Table 1). Length, mid-arm circumference, and triceps skinfold thickness thickness were similar in both groups. There were no differences between the groups in the presence of congenital anomalies or malformations (Table 2); specifically, no differences were reported in the presence of cardiovascular, gastrointestinal, or immunologic anomalies, including asplenia or polysplenia. There was also no difference in groups for the use of corticosteroids at any time during the study follow-up (Table 2). At the baseline, CMV+ participants had higher median alkaline phosphatase (617 vs. 466 U/L, p = 0.001), Alanine aminotransferase (ALT) (164 vs. 109 U/L, p = 0.01), Aspartate aminotransferase (AST) (233 vs. 169 U/L, p = 0.001), AST-to-platelet ratio index (1.2 vs. 0.9, p < 0.001), and International normalized ratio (INR) (1.1 vs. 1.0, p < 0.001) compared with CMV− participants (Table 3). Serum total and direct bilirubin, albumin, gamma-glutamyl transpeptidase (GGT), white blood counts, and platelet counts were similar between groups. Participants were significantly older at HPE for those who were CMV+ (67.8 ± 13.6 d) compared with CMV− (55.1 ± 18.5 d, p = 0.0005). There were no differences in Ishak fibrosis scores from available liver biopsies obtained at the time of HPE (Table 3).

TABLE 1.

Demographics and anthropometrics at baseline exam for biliary atresia participants in PROBE study

	CMV positive 29 (12%)	CMV negative 220 (88%)	p
Female [n (%)]	16 (55)	109 (50)	0.69
Hispanic ethnicity [n/N (%)]	8/28 (29)	54/220 (25)	0.65
Race [n (%)]	26 (90)	212 (96)	0.19
White	13 (48)	136 (65)
Black	3 (11)	21 (10)
Asian	4 (15)	11 (5)
Other/multiracial	7 (24)	43 (20)
Length (cm) (mean ± SD)	56.1 ± 4.5	55.5 ± 7.1	0.52
Weight (kg) (mean ± SD)	4.7 ± 0.6	4.4 ± 0.7	0.03
Head circumference (cm) (mean ± SD)	37.8 ± 0.7	37.4 ± 1.9	0.02
Right mid-arm circumference (cm) (mean ± SD)	11.5 ± 1.3	11.2 ± 1.4	0.49
Right triceps skinfold thickness (mm) (mean ± SD)	5.8 ± 1.7	5.0 ± 1.5	0.10

Note: Bolded values are those that are statistically significant.

Abbreviation: CMV, cytomegalovirus.

TABLE 2.

Anomalies/congenital malformations and corticosteroid treatment in participants with biliary atresia

n (% of participantsa)	CMV positive, n = 26	CMV negative, n = 213	p
Cardiovascular	7/26 (27)	43/213 (20)	0.45
Aortic coarctation	0	3	1.00
Interrupted IVC	1	11	1.00
Complete AV canal	0	2	1.00
Dextrocardia	1	3	0.37
Other cardiovascular anomaly	14	79	0.28
Central nervous system	0/26 (0)	2/213 (1)	1.00
Ear, nose, and throat	0/26 (0)	3/213 (1)	1.00
Eye	0/26 (0)	2/213 (1)	1.00
Gastrointestinal	5/26 (19)	32/213 (15)	0.57
Abdominal heterotaxy	2	13	0.67
Duodenal/jejunal atresia	0	2	1.00
Intestinal malrotation	1	8	1.00
Midline liver	0	12	0.37
Right-sided stomach	1	3	0.37
Genitourinary	2/26 (8)	9/213 (4)	0.34
Immunologic	3/26 (12)	23/213 (11)	1.00
Asplenia	0	2	1.00
Polysplenia	1	15	1.00
Musculoskeletal	0/26 (0)	4/213 (2)	1.00
Respiratory	0/26 (0)	2/213 (1)	1.00
Corticosteroid treatmentb	3/22 (14)	16/160 (10)	0.72

Note: Bolded values are those that are statistically significant.

^aPercentage refers to the percentage of participants with any anomaly in each category. In each category is also listed the number of participants with specific anomalies for which a participant may have multiple anomalies. Thus, the total number of anomalies may exceed the percentage of participants with anomalies.

^bCorticosteroid treatment used at any time in study follow-up. Information on steroid use was not included for all participants.

Abbreviations: AV, Atrioventricular; CMV, cytomegalovirus.

TABLE 3.

Laboratory results and clinical features at baseline/time of HPE

	CMV positive (n= 29)	CMV negative (n = 220)	p
Age at HPE (d) [mean (SD)]	67.8 ± 13.6	55.1 ± 18.5	0.0005
Liver biopsy fibrosis staginga [n (%)]	16 (55)	134 (61)	0.56
Stage 0	0	3 (2)
Stage 1	0	11 (8)
Stage 2	5 (31)	39 (29)
Stage 3	7 (44)	48 (36)
Stage 4	1 (6)	16 (12)
Stage 5	1 (6)	13 (10)
Stage 6	2 (13)	4 (3)
Albuminb (g/dL)	3.7 (2.6–4.3)	3.6 (1.7–5.7)	0.23
Alkaline phosphataseb (U/L)	617 (332–1804)	466 (95–1190)	0.001
ALTb (U/L)	164 (16–522)	109 (11–890)	0.01
ASTb (U/L)	233 (63–738)	169 (39–1777)	0.001
Bilirubin (total)b (mg/dL)	7.6 (4.2–15)	8.3 (1.1–19.8)	0.91
Bilirubin (direct)b (mg/dL)	5.3 (2.6–11)	5.6 (1.4–12)	0.99
GGTb (U/L)	738 (130–2491)	544 (60–3087)	0.15
INRb	1.1 (0.9–6.9)	1.0 (0.8–8.0)	<0.001
Plateletsb (×103/μL)	420 (220–818)	460 (28–1160)	0.43
WBCb (×103/μL)	12.5 (5.7–31.1)	12.9 (4.1–611.5)	0.74
APRib	1.2 (0.5–5.0)	0.9 (0.2–15.8)	<0.001

Note: Bolded values are those that are statistically significant.

^aPer Ishak grading.

^bLaboratory values presented as median (range).

Abbreviations: ALT, alanine aminotransferase; APRi, AST-to-platelet ratio index; AST, aspartate aminotransferase; CMV, cytomegalovirus; GGT, gamma-glutamyl transpeptidase; HPE, hepatic portoenterostomy; INR, international normalized ratio; WBC, white blood count.

Participant outcomes

Clearance of jaundice

There was no difference between the CMV+ and CMV− groups in percent of participants who cleared jaundice at three months post-HPE [CMV+: 9/20 (45%) vs. CMV−: 83/184 (45%), p = 0.834] or at 6 months post-HPE [CMV+: 10/17 (59%) vs. CMV−: 93/155 (60%), p = 0.908]. In addition, there was no difference between groups in percent of participants who cleared jaundice who survived with the native liver at ages 12 months [CMV+: 9/13 (69%) vs. CMV−: 97/128 (76%), p = 0.631] or 24 months [CMV+: 8/9 (89%) vs. CMV−: 82/90 (91%), p = 0.871].

Native liver survival

There was no significant difference between the CMV+ and CMV− groups in the percent of participants with NLS through at least 2 years of age [CMV+: 11/29 (38%) vs. CMV−: 97/220 (44%), p = 0.108] (Figure 1A). When NLS was analyzed, including both CMV status and age at HPE, there was no difference between the CMV+ and CMV− groups with HPE at <60 days of life or at ≥60 days of life (p = 0.622 and 0.448) (Figures 1B, C). In a Cox proportional hazard regression model adjusting for age at HPE, there was similar NLS in CMV+ versus CMV− participants (HR: 1.27, 95% CI: 0.79–2.05, p = 0.320) (Figure 2A). In this model, a 1-month increase in age at HPE led to significantly lower survival with native liver (HR: 1.42, 95% CI: 1.07–1.88, p = 0.014) (Figure 2B).

FIGURE 1.

Kaplan-Meier analysis showing the proportion of participants with biliary atresia with native liver survival in CMV positive versus CMV negative groups, including (A) any age at HPE, (B) only those with HPE at younger than 60 days of life, and (C) only those with HPE at 60 days of life or older. No differences between CMV positive and negative groups were found. Abbreviations: CMV, cytomegalovirus; HPE, hepatic portoenterostomy.

FIGURE 2.

Cox proportional hazard regression analysis modeling of native liver survival in (A) CMV positive versus CMV negative participants, adjusted for age at HPE, and (B) participants with HPE at 1.4 versus 1.9 versus 2.4 months of age, adjusted for CMV status. Abbreviations: CMV, cytomegalovirus; HPE, hepatic portoenterostomy.

Pretransplant patient probability of death

Pretransplant patient probability of death was significantly higher in the CMV+ group compared with the CMV− group (p = 0.013) (Figure 3A). Four of 29 (14%) CMV+ participants compared with 8 of 220 (4%) CMV− participants died prior to receiving a liver transplant (see Supplementary Table 1, http://links.lww.com/HC9/A369 for known details of the cause of death). Two of the 4 CMV+ participants who died and 6 of the 8 CMV− participants who died had complex congenital cardiac disease in addition to BA. Two of the 4 CMV+ who died and 3 of the 8 CMV− who died had splenic malformations. More detailed data on the cause of death than that provided in Supplemental Table S1 (http://links.lww.com/HC9/A369) were not available through ChiLDReN.

FIGURE 3.

Gray’s competing risk analysis with liver transplantation as a competing risk showing CIF of death in participants with biliary atresia before liver transplant in CMV positive versus CMV negative groups, including (A) any age of HPE, (B) only those with HPE at younger than 60 days of life, and (C) only those with HPE at 60 days of life or older. Abbreviations: CIF, cumulative incidence function; CMV, cytomegalovirus; HPE, hepatic portoenterostomy.

When analyzed by both CMV status and age at HPE, there was a trend toward the worse cumulative probability of death in the CMV+ compared with CMV− groups in both those with HPE at <60 days of life and at ≥60 days of life (p = 0.121 and 0.091) (Figures 3B, C). In a Fine-Gray subdistribution hazard model adjusting for age at HPE, there was a statistically significant difference in pretransplant patient probability of death in CMV+ versus CMV− participants within 40 months post-HPE (subdistribution HR: 3.8, 95% CI: 1.1–13.4, p = 0.034) (Figures 4A, B). In this model, age at HPE had no effect, as a 1-month increase in age at HPE did not lead to a higher pretransplant probability of death (subdistribution HR: 1.2, 95% CI: 0.5–2.8, p = 0.71) (Figure 4C).

FIGURE 4.

Fine-Gray competing risk subdistribution hazard modeling predicted values of CIF of death until 40 months post- HPE in (A) CMV positive versus CMV negative participants, adjusted for age at HPE, and (B) participants with HPE at 1.4 versus 1.9 versus 2.4 months of age, adjusted for CMV status. Abbreviations: CIF, cumulative incidence function; CMV, cytomegalovirus; HPE, hepatic portoenterostomy.

DISCUSSION

In this multicenter study from North America, we found that 12% of infants with BA were CMV+ at the time of HPE, similar to the prevalence in other reports and higher than that seen in healthy controls (<1%).10,11,20 Demographics and comorbidities were similar between the CMV+ and CMV− groups. Despite these similarities, CMV+ BA participants had biochemical evidence of worse liver injury and a later age at presentation and HPE than CMV− participants. Jaundice clearance rates were similar, and there was no difference in NLS, which is in contrast to prior reports and a meta-analysis derived from European and Asian centers.8,15 However, CMV+ participants were found to have a higher pretransplant probability of death than CMV− participants after adjusting for age at HPE.

At the time of HPE, CMV+ participants had evidence of more aggressive hepatocellular liver injury based on serum biochemistries. This may be related to a more proinflammatory phenotype caused by the virus itself or as a host immune reaction to the viral infection, which may, in part, lead to the increased risk of pretransplant patient death in CMV+ infants with BA. We also found that infants who were CMV+ presented at a later age for HPE, which supports previously published findings.8,10,21 This delay in diagnosis and treatment could be because CMV infection was diagnosed clinically and presumed to be the cause of cholestasis, leading to a delay in further diagnostic evaluation for cholestasis until the cholestasis failed to improve with time or while receiving antiviral therapy. Another possible explanation for the relationship between CMV positivity and later presentation with more biochemical evidence of liver injury is that CMV+ infants may represent an endotype of BA characterized by those clinical features compared with CMV− infants. Finally, we could not ascertain when the infants were infected with CMV (prenatally, immediate postnatal, or during the first 1–3 mo of life), so it is possible that some of the CMV+ infants had an acute CMV infection superimposed on or causative of the diagnosis of BA at the time of HPE. Regardless of the cause of delayed diagnosis, there is extensive literature showing that a later age at HPE is associated with a poorer prognosis, with lower jaundice clearance in those whose HPE was performed after 60 days of life compared with an earlier age and worse survival with native liver for older infants.22–29 To our knowledge, no prior studies that reported an older age at presentation for CMV+ BA infants have adjusted for older age at HPE in evaluating the cause of worse clinical outcomes in CMV+ infants.

Knowing that age at HPE is a prognostic factor for BA outcome, we adjusted for age at HPE in multiple ways. In doing so, we found that CMV infection had no impact on NLS regardless of age at HPE, which is similar to the findings of a meta-analysis, including 694 patients with BA.14 However, after including age at HPE as a variable in competing risk analyses, we found that CMV+ infants with BA had a higher pretransplant probability of death within 40 months post-HPE compared with the CMV− infants. By adjusting for the later age at HPE in CMV+ participants, these data suggest that differences in patient pretransplant survival outcomes are not due to age at HPE alone, but rather these findings suggest that the CMV infection may be associated with this poorer prognosis. It is important to note that many of the infants who died had severe comorbidities, which could have contributed to mortality irrespective of the degree of liver disease. However, there was no difference between CMV status groups in the frequency of severe comorbidities. Prospective studies will be required to address this issue more definitively.

Small studies and audits of clinical practice have suggested that treatment with antiviral therapy at least partially improves outcomes in CMV+ infants with BA.30–32 Recently, Fischler et al10 published the clinical experience of CMV in BA infants from 4 UK and European centers in which 1 center reported improved NLS in CMV+ infants who received antiviral therapy (decisions for treatment were made on clinical grounds) compared with those CMV+ infants who did not receive this therapy, with the outcomes of those treated being similar to those who were CMV−. Our study did not replicate the worse NLS in CMV+ infants reported by the UK center; however, we did find a worse pretransplant probability of death in CMV+ infants. In order to determine if antiviral therapy would be of benefit in post-HPE care of CMV+ infants, a carefully conducted, double-blind, placebo-controlled, randomized multicentered clinical trial would need to be conducted.

There are several limitations to this study, which should be noted. First, inherent to observational studies is a potential for missingness of data dependent on each center’s practice and data reporting. ChiLDReN has protocols and rigorous data quality control in place, but it is challenging to account for incomplete or inaccurate data. Second, we are unaware of when CMV transmission occurred in our participants, and exposure could have been prenatal, perinatal, or postnatal. The relationship of CMV positivity to older age at HPE could point to postnatal acquisition of CMV, which would be more common as infants aged. Third, liver immunohistochemistry for the presence of CMV antigens was not performed on the liver biopsies obtained at the time of HPE; however, prior reports did not identify CMV in the liver of CMV+ BA patients.14 Finally, we also acknowledge that the differences in the pretransplant probability of death that we identified were based on relatively small numbers of patients, raising caution in overinterpreting our results.

In conclusion, unlike previous reports, evidence of CMV infection at the time of HPE in infants with BA was not associated with worse NLS despite being associated with increased biochemical evidence of liver injury and older age at HPE. However, increased pretransplant probability of death was associated with CMV infection, regardless of age at HPE. Continued evaluation of the consequences of CMV infection at BA diagnosis and the possible effects of antiviral treatment need to be explored.

Supplementary Material

hc9-7-e0175-s001.docx ^{(22KB, docx)}