Caudle, Susan E.*; Katzenstein, Jennifer M.†; Karpen, Saul‡; McLin, Valérie§
Biliary atresia (BA) is the most common form of chronic liver disease leading to liver transplantation (LT) in young children (1–3). Although life threatening, BA's survival rates have increased significantly in the last 25 years, largely owing to the advent of whole and partial LT (1,4–6). With increased survival, attention has turned to long-term outcomes, including cognitive functioning and quality of life. It has been demonstrated that even before liver transplant, children with BA experience neurocognitive weaknesses in comparison with their healthy peers. Cognitive measures have revealed reduced global cognitive functioning in children with liver disease following liver transplant (7–9), as well as specific weaknesses in areas such as motor skills and receptive language (RL) abilities (10–12); however, most studies have examined cognitive functioning with measures that offer only gross estimates of global functioning, include participants of widely varying disease/transplant status, and include participants varying widely in age. Recently, our group used the Mullen Scales of Early Learning (MSEL) (13) in infants with BA at the time of assessment for LT and showed noticeable deficits before transplantation affecting expressive language (EL) and gross motor (GM) skills (14). We found that all of the indices fell significantly below normative expectation; however, the study population was mostly girls, limiting the generalizability of our findings.
Although several growth- and disease-related factors have been found to be associated with later cognitive outcomes, the precise relation between these factors and neurocognitive development is unclear. Poorer nutritional status early in life and reduced head circumference have shown particularly strong relations with cognitive functioning, both before and after transplantation (10,15–19). Other variables that have been correlated with negative cognitive outcomes in children with liver disease include growth parameters, hypoalbuminemia (19), and low serum vitamin E levels (18). Other factors potentially contributing to poor neurocognitive outcomes include repetitive hospitalizations (12,19), increased abdominal fatigue (20), and feeding difficulties and the emphasis placed on supplemental nutrition. Indeed, to meet their metabolic and growth requirements, these infants typically require supplemental nutrition, possibly limiting their movement or field of vision secondary to medical equipment. Additionally, Krull et al suggested that serum bilirubin at the time of transplant was correlated with adverse neurocognitive outcomes following liver transplant (12), although this was not statistically significant in a small cohort of patients with BA (19). Thus, there are several possible contributors to reduced neurocognitive functioning that are likely already in process pretransplantation. In fact, studies have shown that cognitive deficits may be present and measureable before transplantation (21–23). Even extremely early in life, aspects of cognitive development have been found to be affected, with poorer scores reliably correlated with international normalized ratio (INR), growth parameters, and age at Kasai (14).
To our knowledge, however, there has been no investigation to date examining the effects of sex in young children with BA or any liver disease. Girls are diagnosed as having BA slightly more often than boys, with a male:female ratio estimated to fall in the range of 1:0.89 to 1:0.59 (24,25). There is ample evidence of sex differences in brain development and some aspects of cognitive functioning in healthy individuals (26,27), as well as data supporting differential response to diseases and even treatments according to sex (28).
Therefore, the purpose of the present study was to assess sex effects on the cognitive development in a group of children younger than 2 years who were being considered for LT for end-stage liver disease secondary to BA. The secondary aim was to clarify the relation between growth parameters and disease-related variables and cognitive performance across specific domains within each group. We hypothesized that we would continue to see the areas of deficit that we initially identified (14), but that consistent with research being conducted in the area of pediatric cancer (28), girls with BA will demonstrate weaker cognitive development across domains in comparison to boys, placing them at greater risk for long-term cognitive dysfunction.
The study was approved by the institutional review board of the Baylor College of Medicine. Patients were recruited among all-comers being evaluated for LT. The indications for LT were either a failed Kasai (worsening jaundice following portoenterostomy) or a functional Kasai in children with severe portal hypertension (gastrointestinal bleed, ascites, failure to thrive). The recommendation for cognitive testing as part of their transplant evaluation was made by the attending physician. As such, the referral depended on several clinical severity parameters that would affect testing results artifactually. Two patients were excluded from analyses due to known chromosomal anomalies likely to result in delayed development. During the time of the study, a total of 46 patients were evaluated for transplantation secondary to BA from August 2006 to November 2011. Of these, 33 underwent MSEL evaluation and thus were included in this retrospective study.
Because the association of nutritional status and growth parameters with delayed cognitive functioning in otherwise typically developing infants is well established (15–17), mode of energy intake was determined at the time of evaluation for each participant. Accordingly, weight and height at the time of liver transplant evaluation were recorded, and z scores were calculated using the Emmes online calculator (http://spitfire.emmes.com/study/ped/resources/htwtcalc.htm) based on Centers for Disease Control and Prevention growth charts. These data are summarized in Table 1. Reliable data were not available regarding the total and supplemental amount of energy received.
The following biochemical and clinical variables were collected at the time of liver transplant evaluation or as close as possible to the MSEL administration: serum conjugated bilirubin (normal <0.0 mg/dL) (C-bilirubin), prealbumin levels (normal range 18–44 mg/dL), INR (normal range 1.0–1.2), and age at Kasai procedure. An accurate count for patients referred from outside institutions was difficult to obtain; therefore, a best estimate was made upon detailed review of the medical record.
The MSEL is a measure designed for young children to estimate functioning across areas thought to predict later intellectual development. The age range for the MSEL is birth to 5 years, 8 months. Items administered for this test load onto 4 or 5 scales (depending on the child's age), including GM (large motor movements such as sitting and walking), fine motor (FM; distal control such as pincer grasp), visual reception (VR; nonverbal problem-solving skills), EL (using sounds/words to express thoughts to another), and RL (comprehension of the words of others). All of the children who participated in the present study were of an age that all scales were computed, including the GM scale. Each scale comprised items that are developmentally appropriate and use engaging materials. Because cognitive domains are assessed separately, this measure is particularly useful in the assessment of young children to determine specific areas of weakness. Normative scores for each scale are provided in T-score format (mean 50, standard deviation [SD] 10). An overall Early Learning Composite is calculated for each patient, including all subdomains except for GM skills, and is presented in a Standard Score format (mean 100, SD 15). Data were analyzed using SPSS software version 18.0 (SPSS Inc, Chicago, IL).
The participant group consisted of 33 pediatric patients with BA (21 girls, 12 boys) who were undergoing comprehensive evaluation for liver transplant between March 2006 and May 2011. Participants ranged in age from 3 months to 20 months. Girls were slightly younger at the time of evaluation (mean 7.78 months, SD 4.06; boys mean 9.33 months, SD 5.06), but the difference was not statistically significant (t = −0.97; P = 0.34). Of the participants, 21 were non-Hispanic and 12 were Hispanic. English was the primary language spoken in the home for 24 of the children, and Spanish was the dominant language for the remaining 9 children. In these cases, a Spanish-language interpreter was present. English- and Spanish-speaking children did not differ significantly for expressive (t = −0.14, P = 0.76) or receptive (t = −0.31, P = 0.89) language skills, nor was there a difference between groups for number of children exposed to a language other than English (P = 0.72). Parental educational levels were available for all of the biological parents. Maternal education levels ranged from 3 to 18 years of education (mean 13.07, SD 3.29), and paternal educations ranged from 6 to 19 years of education (mean 12.55, SD 3.15), with 12 years of education being equivalent to a high school diploma. There were no differences between the groups in terms of maternal and paternal education levels.
Of the 33 children tested, 20 had undergone Kasai procedure (Table 1) and there was no difference in age between the girls who underwent Kasai procedure and the boys (t = 0.58; P = 0.57). Of those children, 13 are enrolled in the National Institutes of Health–sponsored Childhood Liver Disease Research and Education Network study. Twenty-two of the children underwent LT following developmental evaluation (see Table 1 for additional demographic information). Data on 16 of the 33 children were presented previously (14).
A total of 2 patients (both boys) were missing INR values and were not included in analyses in which INR was included. There were no other missing data. All of the assumptions of statistical analyses were met, and all of the medical variables of interest met requirements for skewness and kurtosis. It is notable that our ages were somewhat skewed, with 50% of the sample falling below 6.5 months of age (19 children), whereas only 16% were above 12 months of age (5 children). Pearson correlations between the MSEL subscales and medical variables of interest for the entire sample were initially completed, and data results indicated a replication of our previous findings (14). Specifically, INR was significantly correlated with GM skills (P < 0.05) and a trend was noted for the correlation between INR and FM skills (P < 0.10). Weight z score was correlated with EL skills (P < 0.01), as was height z score (P < 0.05). Age at Kasai again significantly predicted RL skills (P < 0.01) in this sample (Table 2).
When comparing medical variables of interest between the 2 groups, divided by sex, girls had higher C-bilirubin levels than did boys at the time of evaluation (t = 2.03, P = 0.05), and were of shorter length at time of evaluation than boys (P = 0.009; Table 1). No other differences were seen between the 2 groups, including hospitalization or intensive care unit (ICU) admission before transplant, although gross numbers did suggest that more girls were either inpatient or in ICU. Demographic data for each group are presented in Table 1.
Both boys and girls were compared with the norms for the MSEL and with the full sample to determine whether the findings from our previous study were replicated. Findings indicate that girls performed more poorly on the GM, VR, RL, and EL subdomains when compared with norms, whereas boys only performed significantly more poorly when compared with norms on the GM and EL subdomains (Table 3). When the entire sample was compared with the test norms, the full sample performed more poorly on the GM, RL, and EL subdomains, with the RL findings likely driven by the girls’ performance.
Independent samples t tests were then used to compare the male and female groups on all of the MSEL subdomains and medical variables of interest. Girls scored significantly worse on the VR subscale (P = 0.05) and on the composite measure of development when compared with boys (P = 0.05). Additionally, a trend was observed for girls to earn lower scores on the EL subscale (P = 0.08) than boys. No differences were found between girls and boys in the areas of GM skills, FM skills, or RL skills (Fig. 1). Given that C-bilirubin levels were different between the boys and girls groups, between-subjects analysis of covariane analyses were completed to compare boys and girls on all of the cognitive subdomain scores. The 2 groups did not differ on any of the cognitive subdomain scores when C-bilirubin levels were controlled for, although trends remained for the VR and EL domains (GM, P = 0.15; FM, P = 0.36; VR, P = 0.06; RL, P = 0.80; and EL, P = 0.06). In addition, the groups were compared on developmental outcomes based upon whether they had received a Kasai procedure. A 2 (sex) × 2 (Kasai) analysis of variance revealed no significant differences between the groups on any of the developmental outcomes.
The purpose of the present study was to determine whether sex differences in cognitive development are apparent in extremely young children diagnosed as having BA who have not yet undergone LT. Indeed, in our relatively small sample, significant sex effects were apparent, consistent with our hypotheses that girls would perform more poorly than boys. Among the disease-related factors investigated in the present study, only C-bilirubin at the time of the evaluation was found to differ significantly between boys and girls. In contrast, INR, age at Kasai, time on the transplant list, prealbumin levels, pediatric end-stage liver disease (PELD) score at listing, and PELD score at transplant were roughly equal between groups; however, height z scores, a variable that can be postulated to represent long-term nutritional status, were found to differentiate our groups, with girls showing reduced linear growth compared with boys. The significance of this finding is unclear, but it may suggest more severely compromised nutrition. Alternatively, it is intriguing to consider whether there may be an association with the higher C-bilirubin in the female group. The extrahepatic effects of bile acids are an area of increasing interest and investigation, and conjugated bilirubin has been shown to affect osteoblast function. It is therefore tempting to consider an effect on bone growth in cholestatic infants (32–34). Importantly, we found EL to be highly correlated with both weight and height z scores, suggesting that growth retardation (and possibly underlying nutritional status) renders young children particularly vulnerable to language delays in this population.
Serum C-bilirubin levels may explain group differences in cognitive development because sex effects were reduced to nonsignificant levels when this variable was covaried; however, both VR and EL scores trended toward significance, with girls earning lower scores in both domains when C-bilirubin was covaried. In fact, P values for between-sex group comparisons changed from <0.05 to <0.06 in both cases. It appears likely that this finding is related to reduced power and findings would perhaps remain significant in a larger sample. Regardless, it is unclear why girls may have higher C-bilirubin levels to begin with, or whether this finding is spurious in our sample. There is no evidence that human girls experience higher C-bilirubin levels than boys at birth (29) or in the midst of liver disease, although 2 points distinguish girls from boys when it comes to cholestasis: girls are subject to hormonally induced cholestasis (intrahepatic cholestasis of pregnancy) and are more subject to certain cholestastic diseases (BA and primary biliary cirrhosis). Of note, studies in rats have demonstrated sex differences in bilirubin levels (30,31), rendering this area open to further study. Furthermore, the difference in C-bilirubin between boys and girls may be ascribable to differences in serum estrogen levels between the sexes, something that has been shown in young infants (35–38). It is well accepted that estrogen promotes cholestasis. High estrogen states have been shown to be associated with cholestasis in a number of clinical and laboratory conditions. First, girls with cholestatic diseases experience a worsening of their clinical condition during pregnancy (39). Second, the use of oral contraceptives has been shown to induce or exacerbate cholestasis (40,41). Therefore, C-bilirubin may be a surrogate for higher serum estrogen levels in girls.
The next question is why this increased C-bilirubin may be associated with differences in neurocognitive performance before transplantation. This is an area where less is known. Most research to date has focused on the effects of unconjugated bilirubin on the developing brain or its components. The conjecture may be that there are some overlapping mechanisms of neurotoxicity between conjugated and unconjugated bilirubin. Alternatively, the mechanisms are different, and the field is open for exploration.
Finally, could elevation of C-bilirubin be a surrogate for severity of disease? Similar PELD scores between boys and girls at the time of listing speak against this possibility; however, at the time of transplant, there was a trend for higher PELD scores among the female group, and more girls were inpatients or admitted to the ICU than were boys, although the numbers did not reach statistical significance. Alternatively, C-bilirubin elevation is often associated with Gram-negative bacteremia and sepsis, but we did not have the ability to show that girls before LT had a higher incidence of infection.
It is encouraging that despite the fact that both boys and girls in our sample exhibited cognitive weaknesses in comparison with the normative group, most of the children remained broadly within the average range (defined here as falling within 1 standard deviation of the mean; T score 40–59) across domains of functioning. The exception was the domain of GM, where 13 of 21 girls (62%) and 7 of 12 boys (58%) exhibited performance >1 SD below the mean. Furthermore, whereas more boys remained at age expectation for the domain of EL (8 of 12, or 67% fell within 1 SD of the mean), the girls in our group fared less well. Only 9 of 21 (or 43%) girls in our sample earned EL scores that fell in the average range. Less encouraging is the trend for girls to exhibit lower scores in the domain of VR in comparison to the boys. This area has been conceived of as the domain of the MSEL most related to nonverbal intellectual functioning later in life (42), and our data suggest that girls may experience increased vulnerability in area that is much less amenable to rehabilitative efforts. Although both physical/occupational therapy and speech-language therapy are available to address developmental lags in these skill areas, there is little that has been shown to be effective in the building of nonverbal problem-solving skills beyond exposure of the young child to a stimulating and thought-provoking environment. This finding is consistent with studies of other illness populations, suggesting that girls are more at risk for hormonally induced cognitive deficits than boys (43–45). It follows that the data from this cohort suggest that female sex may be a risk factor in patients with BA for later cognitive deficits.
Although our sample size has increased significantly from our first study (14), we continue to be likely underpowered when comparing boys and girls, and the strength of the differences between groups may increase with greater sample size. Furthermore, we are limited by the slight skewness of our sample, with more children being younger (younger than 6 months) compared with older children (older than 12 months). Additional limitations include that not all of the patients with BA at our hospital were referred for neuropsychological evaluation, and the referred group from our previous study tended to be healthier than the nonreferred group.
Longitudinal evaluation of this group of children will assist in determining how these findings predict long-term cognitive outcomes and whether sex differences remain over time and, in particular, following transplant. In addition, prospective examination of ongoing changes in biomedical factors, including nutritional status, in relation to cognitive outcomes will be important, especially in determination of needs for rehabilitation/habilitation and compensatory strategies. Transplant is lifesaving in these children, but efforts need to focus on neurodevelopmental and nutritional care before transplant if our patients are to fully benefit from their transplant in the long term.
The authors thank the following members of the Liver Center and transplant team for assistance with recruitment: Beth A. Carter, MD; Paula M. Hertel, MD; Douglas S. Fishman, MD; Ryan Himes, MD; Daniel Leung, MD; Julie Economides, RN; Diesa Stephens, RN; Alison Skelton, RN; and Carie Reid, RN. We are also indebted to Jessica Sigers and Vanessa Suarez for logistical help and to Myra Grant, Kristin Adkins, Monica Thomas, and Laura Yaffee for their assistance with evaluation.
1. Goldman M, Pranikoff T. Biliary disease in children. Curr Gastroenterol Rep 2011; 13:193–201.
2. Kelly DA, Davenport M. Current management of biliary atresia. Arch Dis Child 2007; 92:1132–1135.
3. Sokol RJ, Mach C, Narkewicz MR, et al. Pathogenesis and outcome of biliary atresia: current concepts. J Pediatr Gastroenterol Nutr 2003; 37:4–21.
4. McDiarmid SV. Liver transplantation: the pediatric challenge. Clin Liver Dis 2000; 4:879–927.
5. Chardot C. Biliary atresia. Orphanet J Rare Dis 2006; 1:28.
6. Grabhorn E, Ganschow R, Helmke K, et al. Liver transplantation in infants younger than 6 months old. Transplant Proc 2002; 34:1964–1965.
7. Alonso EM, Sorensen LG. Cognitive development following pediatric solid organ transplantation. Curr Opin Organ Transplant 2009; 14:522–525.
8. Kaller T, Boeck A, Sander K, et al. Cognitive abilities, behaviour and quality of life in children after liver transplantation. Transplantation 2005; 79:1252–1256.
9. Sorensen LG, Neighbors K, Martz K, et al. Studies of Pediatric Liver Transplantation (SPLIT) and Functional Outcomes Group (FOG). Am J Transplant 2011; 11:303–311.
10. Hobbs SA, Sexson SB. Cognitive development and learning in the pediatric organ transplant recipient. J Learn Disabil 1993; 26:104–113.
11. Thevenin DM, Baker A, Kato T, et al. Neurodevelopmental outcomes of infant multivisceral transplant recipients: a longitudinal study. Transplant Proc 2006; 38:1694–1695.
12. Krull K, Fuchs C, Yurk H, et al. Neurocognitive outcome in pediatric liver transplant recipients. Pediatr Transplant 2003; 7:111–118.
13. Mullen EM. Mullen Scales of Early Learning. Circle Pines, MN: American Guidance Service; 1995.
14. Caudle SE, Katzenstein JM, Karpen SJ, et al. Language and motor skills are impaired in infants with biliary atresia before transplantation. J Pediatr 2010; 156:936–940.
15. Stewart SM, Silver CH, Nici J, et al. Neuropsychological function in young children who have undergone liver transplantation. J Pediatr Psychol 1991; 16:569–583.
16. Schultz KH, Wein C, Boeck A, et al. Cognitive performance of children who have undergone liver transplantation. Transplantation 2003; 75:1236–1240.
17. Kaller T, Schulz KH, Sander K, et al. Cognitive abilities in children after liver transplantation. Transplantation 2005; 79:1252–1256.
18. Stewart SM, Uauy R, Waller DA, et al. Mental and motor development correlates in patients with end-stage biliary atresia awaiting liver transplantation. Pediatrics 1987; 79:882–888.
19. Wayman KI, Cox KL, Esquivel CO. Neurodevelopmental outcome of young children with extrahepatic biliary atresia 1 year after liver transplantation. J Pediatr 1997; 131:894–898.
20. Swain MG. Fatigue in liver disease: pathophysiology and clinical management. Can J Gastroenterol 2006; 20:181–188.
21. Lacerda SS, Guimaro MS, Prade CV, et al. Neuropsychological assessment in kidney and liver transplant candidates. Transplant Proc 2008; 40:729–731.
22. Meyer T, Eshelman A, Abouljoud M. Neuropsychological changes in a large sample of liver transplant candidates. Transplant Proc 2006; 38:3559–3560.
23. Stewart SM, Campbell RA, McCallon D, et al. Cognitive patterns in school-age children with end-stage liver disease. J Dev Behav Pediatr 1992; 13:331–338.
24. Yoon PW, Bresee JS, Olney RS, et al. Epidemiology of biliary atresia: a population-based study. Pediatrics 1997; 99:376–382.
25. Nio M, Ohi R, Miyano T, et al. Five- and ten-year survival rates after surgery for biliary atresia: a report from the Japanese biliary atresia register. J Pediatr Surg 2003; 38:997–1000.
26. Hines M. Sex-related variation in human behavior and the brain. Trends Cog Sci 2010; 14:448–456.
27. Kinsbourne M. Development of cerebral lateralization in children. In: Reynolds CR, Fletcher-Janzen E, eds. Handbook of Clinical Child Neuropsychology. New York: Springer; 2009: pp. 47–66.
28. Buizer AI, deSonneville LMJ, Veerman AJP. Effects of chemotherapy on neurocognitive function in children with acute lymphoblastic leukemia: a critical review of the literature. Pediatr Blood Cancer 2009; 52:447–454.
29. Grantham-McGregor S. A review of studies of the effect of severe malnutrition on mental development. J Nutr 1995; 125:2233S–2238S.
30. Grantham-McGregor S, Ani C. Cognition and undernutrition: evidence for vulnerable period. Forum Nutr 2003; 56:272–275.
31. Gorman KS. Malnutrition and cognitive development: evidence from experimental/quasi-experimental studies among the mild-to-moderately malnourished. J Nutr 1995; 125:2239S–2244S.
32. Ruiz-Gaspà S, Martinez-Ferrer A, Guañabens N, et al. Effects of bilirubin and sera from jaundiced patients on osteoblasts: contribution to the development of osteoporosis in liver diseases. Hepatology 2011;54:2104–13.
33. Janes CH, Dickson ER, Okazaki R, et al. Role of hyperbilirubinemia in the impairment of osteoblast proliferation associated with cholestatic jaundice. J Clin Invest 1995; 95:2581–2586.
34. Weinreb M, Pollak RD, Ackerman Z. Experimental cholestatic liver disease through bile-duct ligation in rats results in skeletal fragility and impaired osteoblastogenesis. J Hepatol 2004; 40:385–390.
35. Tioseco JA, Aly H, Milner J, et al. Does gender affect neonatal hyperbilirubinemia in low-birth-weight infants? Pediatr Crit Care Med 2005; 6:171–174.
36. Kayali R, Aydin S, Cakatay U. Effect of gender on main clinical chemistry parameters in aged rats. Curr Aging Sci 2009; 2:67–71.
37. Całkosiński I, Dobrzyński M, Kobierska-Brzoza J, et al. The influence of strain, sex and age on selected biochemical parameters in blood serum of Buffalo and Wistar rats. Pol J Vet Sci 2010; 13:293–299.
38. Bidlingmaier F, Versmold H, Knorr D. Sex differences in plasma estrogen concentrations in infancy. 21 Symp. Dtsch. Ges. Endokrin Abstract 103. Acta Endoc 1975; 193 (Suppl):103.
39. Mistilis SP. Liver disease in pregnancy, with particular emphasis on the cholestatic syndromes. Australas Ann Med 1968;17:248–60.
40. Von Oldershausenh, Eggsteun M, Dold U, et al. Icterus in intrahepatic cholestasis following the administration of contraceptive steroids. Dtsch Med Wochenschr 1965; 90:1290–1294.
41. Orellana-Alcalde JM, Dominguez JP. Jaundice and oral contraceptive drugs. Lancet 1966; 2:1279–1280.
42. Caudle SE, Katzenstein JM, Oghalai JS, et al. Nonverbal cognitive development in children with cochlear implants: relationship between the Mullen Scales of Early Learning and later performance on the Leiter International Performance Scales—Revised. Assess 2012; February 20 [epub ahead of print].
43. McEwen BS, Alves SE, Bulloch K, et al. Clinically relevant basic science studies of gender differences and sex hormone effects. Psychopharmacol Bull 1998; 34:251–259.
44. Luine V. Sex differences in chronic stress effects on memory in rats. Stress 2002;5:205–16.
45. Genazzani AR, Pluchino N, Luisi S, et al. Estrogen, cognition and female ageing. Hum Reprod Update 2007; 13:175–187.