Abstract
Osteopenia and bone fractures (fx) are significant causes of morbidity in children with cholestatic liver disease. Dual-energy X-ray absorptiometry (DXA) analysis was performed in children with intrahepatic cholestatic diseases who were enrolled in the Longitudinal Study of Genetic Causes of Intrahepatic Cholestasis (LOGIC) in the Childhood Liver Disease Research Network (ChiLDReN). DXA was performed on participants age > 5 years (with native liver) diagnosed with bile acid synthetic disorder (BASD), alpha-1 antitrypsin deficiency (A1AT), chronic intrahepatic cholestasis (CIC), and Alagille syndrome (ALGS). Weight, height, and body mass index Z-scores were lowest in CIC and ALGS. Total bilirubin (TB) and serum bile acids (SBA) were highest in ALGS. Bone mineral density (BMD) and bone mineral content (BMC) Z-scores were significantly lower in CIC and ALGS than in BASD and A1AT (p<0.001). After anthropometric adjustment, bone deficits persisted in CIC but were no longer noted in ALGS. In ALGS, height- and weight-adjusted subtotal BMD and BMC Z-scores were negatively correlated with TB (p<0.001) and SBA (p=0.02). Mean height- and weight-adjusted subtotal BMC Z-scores were lower in ALGS participants with a history of fx. DXA measures did not correlate significantly with biliary diversion status.
Conclusion:
CIC patients had significant bone deficits that persisted after adjustment for height and weight and generally did not correlate with degree of cholestasis. In ALGS, low BMD and BMC reference Z-scores were explained by poor growth. Anthropometrically-adjusted DXA measures in ALGS correlate with markers of cholestasis and fx history. Reduced bone density in this population is multifactorial and related to growth, degree of cholestasis, fracture vulnerability, and contribution of underlying genetic etiology.
Keywords: dual-energy X-ray absorptiometry (DXA), osteoporosis, osteopenia, Alagille syndrome, alpha-1 antritrypsin deficiency, progressive familial intrahepatic cholestasis
Reduced bone density is a common complication of chronic liver disease in both adults and children. Factors contributing to bone mineral deficits are dependent on the physiology and severity of the underlying liver disease and may include chronic nutrient and calcium malabsorption leading to malnutrition and growth failure, deficiencies of fat-soluble vitamins, level of physical activity, and circulating inflammatory cytokines. High rates of osteoporosis and bone fracture have been observed in adults with chronic cholestatic liver diseases, such as primary sclerosing cholangitis (PSC) and primary biliary cholangitis (PBC), and cirrhosis of any cause (1–3). Metabolic bone disease is also common in children with chronic cholestasis or end-stage liver disease, but fewer studies have been conducted to characterize the bone mineral deficits in this vulnerable population.
Osteopenia and pathologic fractures have been reported in children with chronic cholestatic liver disease, including Alagille syndrome (ALGS), progressive familial intrahepatic cholestasis (PFIC), and biliary atresia (4–7). In some cases, bone disease can be severe enough to be an indication for liver transplantation. In contrast to what is seen in adults, children have remarkable improvement in bone mineral density (BMD) by 1 year after liver transplantation, often accompanied by significant catch-up growth (8,9). Despite the impact of metabolic bone disease on morbidity and quality of life in pediatric patients with chronic liver disease, few detailed studies have been done to quantify bone deficits and identify risk factors for osteopenia and fracture. In one published study, ALGS patients (n=31) and 80 healthy control participants underwent evaluation by dual-energy X-ray absorptiometry (DXA). The ALGS children were small for age and had reduced bone mineral content (BMC) for age and BMC for height, with bone mineralization positively related to coefficient of fat absorption, but not to dietary intake (10). In another study of 16 non-jaundiced children with biliary atresia, whole body and lumbar spine BMC were found to decline with age, indicating that children surviving with native liver are at risk of compromised bone health, even in the absence of elevated bilirubin levels (11). To date, there have been no published studies of DXA analysis in children with other inherited chronic liver diseases, such as alpha-1 antitrypsin deficiency (A1AT) or PFIC.
The Childhood Liver Disease Research Network (ChiLDReN) is a National Institutes of Health-funded consortium of pediatric centers in North America focused on the study of rare pediatric liver diseases. The Longitudinal Study of Genetic Causes of Intrahepatic Cholestasis (LOGIC - NCT00571272) enrolled children with one of four diagnoses: ALGS, PFIC, A1AT, or bile acid synthetic disorder (BASD). As part of the LOGIC study protocol, participants age > 5 years underwent DXA scanning, the results of which are reported here. The objectives of this study were to address an important knowledge gap by investigating the prevalence of bone mineral deficits in this cohort of children with chronic liver disease and to determine factors associated with lower bone density, including growth parameters, laboratory values, or clinical events, such as fracture and biliary diversion.
EXPERIMENTAL PROCEDURES
Study Design and Participants
Participants in this study were enrolled in the LOGIC study (NCT00571272) through the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)-funded multicenter ChiLDReN consortium. This study protocol was approved by the Institutional Review Boards at each participating center and conformed to the ethical guidelines of the 1975 Declaration of Helsinki. Informed consent was obtained from parents/guardians or participants 18 years or older, and assent was obtained from participants > 7 years of age, per local guidelines.
LOGIC is a longitudinal study designed to investigate the natural history of several genetic causes of cholestasis, including ALGS, A1AT, BASD, and PFIC. Children and young adults with a confirmed diagnosis of ALGS, A1AT, BASD, or PFIC were eligible for enrollment in LOGIC from birth through 25 years of age. Detailed enrollment criteria for each disease group are presented in Supplementary Table S1. In the PFIC group, 20 of 41 participants had a confirmed genetic diagnosis: 10 with mutations in ATP8B1, 4 with ABCB11 and 6 with ABCB4. Since not all of the participants in the PFIC group have a documented genetic diagnosis, for the purpose of this study, we have categorized them as chronic intrahepatic cholestasis (CIC).
In the LOGIC protocol, defined data elements, including clinical events, laboratory values, growth parameters, physical examination findings, and history and location of bone fractures were collected at enrollment and at yearly visits for 10 years, or until liver transplantation or death. LOGIC participants age > 5 years with their native livers were eligible to undergo DXA once during the course of the study. DXA scans were performed at 12 ChiLDReN centers from 2008 through 2013 (2008-2013 for ALGS and CIC participants; 2008-2011 for A1AT; and 2010 for BASD). Interim analysis in 2011 revealed minimal DXA abnormalities in A1AT and BASD participants, so DXA scanning was discontinued in those cohorts at that time. DXA results reported here include those for participants with ALGS (n=49), A1AT (n=44), BASD (n=14), or CIC (n=41). Supplementary Figure S1 shows the derivation of the sample population.
DXA Measurements
DXA scanning of whole body, spine, and femur was performed using Hologic (Newark, DE) or Lunar (GE Health Sciences, Pittsburgh, PA) equipment. DXA equipment for each brand was standardized using BMIL phantoms provided by the DXA Coordinating Center (12). The same phantom was used for longitudinal drift correction where necessary. Adjustment between the two brands of scanners was performed with patient-based correction formulas (13). Standard Z-scores were calculated based on age, sex, and race reference data (14). In addition to these analyses, adjusted Z-scores were calculated based on age, sex, race, height, and weight (15). For a special, supplementary analysis, adjusted Z-scores with all the above listed parameters and % body fat were also calculated (15). In most cases, anthropometric measurements were obtained on the same day as the DXA scan; otherwise, the measurements closest to the date of scan were used.
Statistical Methods
The pre-planned analyses of the DXA data include all LOGIC participants with their native liver for whom a DXA was measured. Descriptive statistics are displayed as means and standard deviations or median with first and third quartiles for continuous variables and as frequencies and percentages for categorical variables. Winsorizing refers to a statistical technique in which extreme values are limited in order to reduce the effect of possibly spurious outliers (16). To reduce the influence of outliers, all DXA Z-scores were Winsorized using a lower limit of −3 and an upper limit of 3. All DXA Z-scores presented are adjusted for age, sex, race (black vs. non-black), weight, and height, unless otherwise indicated.
We characterized DXA Z-scores in two ways --- as a continuous variable and as a dichotomous variable based on the proportion of participants < −1.5 vs. > −1.5. Comparisons of DXA Z-scores and other continuous variables among diagnosis groups were tested using analysis of variance (ANOVA) (or by nonparametric Kruskall-Wallis tests if the distributional assumptions of ANOVA were not met), and comparisons of dichotomous DXA Z-scores and other categorical variables, by diagnosis group, were tested with chi-square tests. Within diagnosis groups, differences in laboratory values between participants with DXA Z-scores < −1.5 vs. > −1.5 and differences in DXA Z-scores for participants with and without bone fractures and with and without biliary diversion were compared using two sample t-tests.
Pearson correlation coefficients and their corresponding p-values were used to examine associations between participant anthropometric parameters (height, weight, and body mass index [BMI] Z-scores) and reference DXA Z-scores (defined as the standard method to adjust for age, sex, and race) and between DXA Z-score and laboratory values. Laboratory values were obtained within 1 year of DXA scan. Height, weight, and BMI Z-scores were calculated based on CDC growth charts (reference year 2000; https://www.cdc.gov/nccdphp/dnpao/growthcharts/resources/sas.htm). Statistical analyses were carried out using SAS version 9.4 (SAS Institute; Cary, NC). Due to the large number of comparisons being made, we considered p-values <0.01 to be statistically significant.
RESULTS
Characteristics of study cohort
Our study cohort consisted of a total of 148 participants enrolled in the LOGIC study as part of the ChiLDReN consortium. Supplementary Figure S1 shows the derivation of the study population. Participants had ALGS (n=49), CIC (n=41), A1AT (n=44), and BASD (n=14). Table 1 displays characteristics of participants in the study by disease group. The A1AT and BASD groups had higher percentages of male participants than the ALGS and CIC groups: 70% and 79% versus 57% and 44%, respectively. Participants were similar across disease groups for race, ethnicity, and age. The rate of bone fracture occurrence was not statistically different among disease groups. Anthropomorphic measurements were significantly different across disease groups (p<0.001), with ALGS and CIC participants having lower mean height, weight, and BMI Z-scores than participants with A1AT or BASD. Non-parametric Kruskall-Wallis tests confirmed results from ANOVA (data not shown). Across all disease groups, few participants had a BMI Z-score < −2. The mean percentages of lean body mass were similar among disease groups, ranging from 74% to 76% (Table 1).
Table 1:
Demographic and Clinical Characteristics* of Participant Population by Disease Group
| Liver Disease | Alagille Syndrome (n = 49a) | Chronic Intrahepatic Cholestasis (n = 41b) | Alpha-1 Antitrypsin Deficiency (n = 44c) | Bile Acid Synthetic Disorder (n = 14d) | p-value^ |
|---|---|---|---|---|---|
| Number (%) of Participants | |||||
| Male sex | 28 (57%) | 18 (44%) | 31 (70%) | 11 (79%) | 0.04 |
| Race | 0.16 | ||||
| White | 38 (78%) | 34 (83%) | 43 (98%) | 11 (79%) | |
| Black | 5 (10%) | 2 (5%) | 0 (0%) | 1 (7%) | |
| Other | 6 (12%) | 5 (12%) | 1 (2%) | 2 (14%) | |
| Ethnicity | 0.62 | ||||
| Hispanic | 4 (8%) | 7 (17%) | 3 (7%) | 3 (21%) | |
| Non-Hispanic | 44 (90%) | 33 (80%) | 40 (91%) | 11 (79%) | |
| Not reported | 1 (2%) | 1 (2%) | 1 (2%) | 0 (0%) | |
| Any bone fractures** | 0.20 | ||||
| No | 30 (61%) | 31 (76%) | 35 (80%) | 11 (79%) | |
| Yes | 19 (39%) | 10 (24%) | 9 (20%) | 3 (21%) | |
| Frequency of bone fractures | 0.06 | ||||
| 0 or 1 | 40 (82%) | 38 (93%) | 43 (98%) | 13 (93%) | |
| 2 or more | 9 (18%) | 3 (7%) | 1 (2%) | 1 (7%) | |
| Spleen size >2cm and platelet count < 150×103/μL*** | 6 (16%) | 4 (13%) | 5 (12%) | 0 (0%) | 0.59 |
| BMI Z-score < −2*** | 3 (6%) | 2 (5%) | 1 (2%) | 1 (7%) | 0.73 |
| Mean (SD) | |||||
| Median (Q1, Q3) | |||||
| Age (years) | 10.8 (4.6) | 11.0 (5.0) | 11.0 (5.3) | 12.0 (6.7) | 0.90 |
| 9.8 (7.7, 13.1) | 10.9 (6.6, 15.0) | 8.7 (6.5, 14.5) | 8.4 (6.8, 20.0) | ||
| Weight Z-score | −1.3 (1.3) | −1.1 (1.5) | 0.6 (0.9) | 0.9 (1.4) | <0.001 |
| −1.1 (−2.3, −0.4) | −1.1 (−2.0, 0.0) | 0.7 (0.1, 1.2) | 1.1 (−0.2, 1.7) | ||
| Height Z-score | −1.5 (1.1) | −1.3 (1.8) | 0.5 (1.2) | −0.1 (0.7) | <0.001 |
| −1.6 (−2.1, −0.7) | −1.2 (−2.3, −0.3) | 0.5 (−0.1, 1.1) | 0.1 (−0.5, 0.2) | ||
| BMI Z-score | −0.6 (1.2) | −0.2 (1.0) | 0.4 (1.0) | 1.0 (1.4) | <0.001 |
| −0.3 (−1.3, 0.3) | −0.1 (−0.8, 0.4) | 0.6 (−0.2, 0.9) | 1.3 (0.4, 1.9) | ||
| Mean (SD) | |||||
| Lean body mass (%)**** | 73.8 (5.6) | 76.4 (7.4) | 75.8 (7.7) | 74.6 (8.5) | 0.31 |
| Total bilirubin, mg/dL | 3.7 (5.1) | 1.6 (1.9) | 0.8 (1.2) | 0.3 (0.3) | <0.001 |
| Direct bilirubin, mg/dL | 2.4 (3.2) | 0.5 (0.9) | 0.1 (0.1) | ----- | <0.001 |
| GGT, U/L | 353.6 (309.4) | 97.9 (172.5) | 57.0 (66.9) | 23.6 (4.9) | <0.001 |
| Alkaline phosphatase, U/L | 501.9 (345.1) | 457.0 (229.8) | 234.9 (114.4) | 215.6 (122.7) | <0.001 |
| ALT, U/L | 164.9 (120.4) | 76.2 (53.8) | 59.0 (39.2) | 26.1 (22.4) | <0.001 |
| AST, U/L | 151.8 (110.8) | 78.5 (51.5) | 61.1 (48.6) | 51.8 (17.0) | <0.001 |
| Serum bile acids (µmol/L) | 117.3 (106.3) | 69.8 (102.2) | 16.6 (19.0) | 27.5 (43.4) | <0.001 |
| 25-OH vitamin D level (ng/mL) | 34.1 (19.0) | 37.8 (16.4) | 37.8 (10.6) | 25.9 (9.9) | 0.12 |
| INR | 1.08 (0.16) | 1.07 (0.17) | 1.06 (0.10) | 1.03 (0.10) | 0.67 |
| Albumin, g/dL | 4.2 (0.6) | 4.2 (0.5) | 4.5 (0.4) | 4.6 (0.4) | 0.005 |
| WBC, × 103/μL | 6.5 (2.8) | 5.9 (2.2) | 5.9 (2.5) | 6.8 (2.2) | 0.44 |
| Hemoglobin, g/dL | 12.9 (1.7) | 13.0 (1.4) | 13.8 (1.1) | 14.4 (1.3) | <0.001 |
| Platelet count, × 103/μL | 232.5 (105.0) | 272.7 (142.2) | 232.1 (114.8) | 249.1 (109.4) | 0.39 |
For Alagille syndrome, the number of non-missing values for continuous variables ranged from 38 to 49, except for direct bilirubin (n=22).
For chronic intrahepatic cholestasis, the number of non-missing values for continuous variables ranged from 35 to 41, except for direct bilirubin (n=22) and vitamin D (n=33).
For alpha-1 antitrypsin deficiency, the number of non-missing values for continuous variables ranged from 38 to 44, except for direct bilirubin (n=24) and vitamin D (n=13).
For bile acid synthetic disorder, the number of non-missing values for continuous variables ranged from 10 to 14, except for direct bilirubin (n=0).
p-value based on ANOVA reported for mean rows and from Chi-square or Fisher’s Exact test for discrete outcomes.
Values closest to the time of the DXA scan for each participant were used in the analysis. Measurements were required to be within 1 year of the scan.
Bone fractures could have occurred before or after the DXA scan was performed.
Percentages are among participants with non-missing data.
Lean body mass was measured by DXA.
Note: Non-parametric Kruskal-Wallis tests confirmed results from ANOVA.
Abbreviations: ALT, alanine transaminase; ANOVA, analysis of variance; AST, aspartate transaminase; BMI, body mass index; DXA, dual energy x-ray absorptiometry; GGT, gamma-glutamyl transpeptidase; INR, international normalized ratio; OH, hydroxyvitamin; SD, standard deviation; WBC, white blood cell.
Mean laboratory measurements of bilirubin, gamma-glutamyl transpeptidase (GGT), alkaline phosphatase, alanine transaminase (ALT), aspartate transaminase (AST), and serum bile acids (SBAs) were highest in participants with ALGS, intermediate for CIC, and lowest in participants with A1AT and BASD. Mean albumin and hemoglobin were lower in ALGS and CIC compared with A1AT and BASD. There were no statistical differences in international normalized ratio (INR), white blood cell (WBC), and platelet counts between groups. Few participants met criteria for portal hypertension as defined by spleen size > 2 cm below the left costal martin and platelet count < 150,000/microliter (17), with no significant differences across disease groups (Table 1).
Overall, LOGIC participants who underwent DXA scanning were similar to those who did not, with a few exceptions. Supplementary Table S2 shows demographic and clinical characteristics of participants who were eligible but did not undergo DXA scan. In the BASD group, some differences were identified in sex, race, and ethnicity between participants who did (n=14) and did not (n=7) undergo DXA. The mean age tended to be younger in the non-DXA participants. In the CIC group, participants who underwent DXA had lower weight and BMI Z-scores. In the ALGS and CIC groups, DXA study participants had higher 25-hydroxyvitamin (OH) vitamin D levels than those who did not undergo DXA scanning. Minor differences hemoglobin were also identified in the BASD and A1AT groups.
DXA scores
The A1AT and BASD participants did not have significant bone mineral deficits as measured by DXA (Supplementary Table S3); therefore the results are focused on the ALGS and CIC groups. Table 2 shows summary measures of reference as well as height- and weight-adjusted BMD and BMC Z-scores resulting from DXA scans by disease group for ALGS and CIC participants. The reference scores take into account the participant’s age, sex, and race. Mean DXA reference Z-scores were similar between ALGS and CIC participants, except for spine BMD, which was lower in the CIC group (Table 2). Since some of the children in the study cohort have significant growth deficits, we examined correlations between DXA reference Z-scores and height, weight, and BMI Z-scores by disease group (Table 3). All DXA reference Z-scores correlated strongly with height Z-scores in the ALGS participants (p<0.0001). In the ALGS group, weight Z-scores also correlated strongly with total body and total body minus head BMD and BMC DXA reference Z-scores (p<0.0001), and to a lesser extent with spine, total hip, and hip neck DXA reference Z-scores (Table 3). Height and weight Z-scores for the CIC participants correlated with total body and total body minus head BMD and BMC DXA reference Z-scores. Fewer correlations were identified in A1AT and BASD (Supplementary Table S4), but these participants also did not have significant growth deficits at the time of DXA (Table 1). No correlations were found between DXA reference Z-scores and BMI Z-scores for any disease group. Because DXA reference Z-scores were so tightly correlated with growth parameters, we elected to adjust all DXA Z-scores for weight and height according to the method of Short et al (15).
Table 2:
DXA Reference and Height- and Weight-Adjusted Z-Scores by Disease Group
| Mean (SD) % with Z < −1.5 | Alagille Syndrome (n=46-49) | Chronic Intrahepatic Cholestasis (n=39-41) | p-value* | Winsorized** values n (%) |
|---|---|---|---|---|
| BMD Measures | ||||
| Total body minus head | ||||
| Reference Z-score | −1.10 (1.57) 46.9% |
−1.58 (1.11) 56.1% |
0.10 | 13 (14%) |
| Adjusted Z-score | 0.23 (1.65) 20.8% |
−0.37 (1.45) 20.0% |
0.07 | 9 (10%) |
| Spine | ||||
| Reference Z-score | −0.08 (1.66) 20.4% |
−1.01 (1.36) 43.9% |
0.005 | 9 (10%) |
| Adjusted Z-score | 0.13 (1.65) 20.8% |
−1.15 (0.98) 40.0% |
<0.001 | 4 (5%) |
| Total hip | ||||
| Reference Z-score | −0.78 (1.38) 23.4% |
−1.17 (1.23) 47.5% |
0.16 | 10 (11%) |
| Adjusted Z-score | −0.46 (1.41) 21.7% |
−1.17 (1.30) 43.6% |
0.02 | 7 (8%) |
| BMC Measures | ||||
| Total body minus head | ||||
| Reference Z-score | −1.23 (1.45) 44.9% |
−1.80 (1.33) 63.4% |
0.06 | 25 (28%) |
| Adjusted Z-score | 0.36 (1.34) 10.4% |
−0.76 (1.16) 27.5% |
<0.001 | 7 (8%) |
Tests for differences DXA Z-scores between disease groups were performed using two sample t-tests
Values less than −3 were set to −3 and values greater than 3 were set to 3.
Abbreviations: BMC, bone mineral content; BMD, bone mineral density; DXA, dual energy x-ray absorptiometry; SD, standard deviation.
Table 3:
Pearson Correlations between DXA Reference Z-Scoresa and Height, Weight, and BMI Z-Scores for ALGS and CIC
| Alagille Syndrome (n = 47-49) | Chronic Intrahepatic Cholestasis (n = 40-41) | |||||
|---|---|---|---|---|---|---|
| Height | Weight | BMI | Height | Weight | BMI | |
| BMD Measures | ||||||
| Total body | 0.62*** | 0.53*** | 0.18 | 0.56** | 0.55** | 0.23 |
| Total body minus head | 0.67*** | 0.59*** | 0.24 | 0.59*** | 0.55** | 0.19 |
| Spine | 0.58*** | 0.44* | 0.09 | 0.53** | 0.54** | 0.29 |
| Total hip | 0.57*** | 0.48** | 0.19 | 0.33 | 0.32 | 0.18 |
| Hip neck | 0.56*** | 0.42* | 0.16 | 0.33 | 0.35 | 0.24 |
| BMC Measures | ||||||
| Total body | 0.76*** | 0.69*** | 0.34 | 0.73*** | 0.68*** | 0.28 |
| Total body minus head | 0.81*** | 0.73*** | 0.34 | 0.74*** | 0.68*** | 0.27 |
Reference Z-score is adjusted for age, sex, and race (Black vs. non-Black).
p<0.01
p<0.001
p<0.0001
Abbreviations: BMC, bone mineral content; BMD, bone mineral density; BMI, body mass index; DXA, dual energy x-ray absorptiometry.
After adjustment for height and weight, mean total body minus head BMD and BMC Z-scores were higher (than when calculated with the reference methods) in both ALGS and CIC groups, indicating that these bone deficits are at least in part due to small size (Table 2). Interestingly, the CIC group had deficits in total spine and total hip BMD reference Z-scores that persisted after adjustment for height and weight, with total spine and total hip BMD Z-scores in the CIC group of −1.15 and −1.17 respectively. Over 40% of the CIC cohort had a Z-score < −1.5, in contrast to the expected distribution of 7% for the reference population.
Figure 1 displays the reference and adjusted Z-scores for total body minus head BMD (Figure 1.1) and BMC (Figure 1.2) for the ALGS and CIC groups (results for A1-AT and BASD disease groups are shown in Supplementary Figure S2). For both the ALGS and CIC disease groups, the data indicate that low weight and height for age are contributing to observed bone mineral deficits. Supplementary Table S3 shows data for additional DXA measurements (total body and hip neck) and all DXA measurements for the A1-AT and BASD disease groups. All DXA measures are also presented with adjustment for weight, height, and % body fat (Supplementary Table S3). Of note, differences in DXA Z-scores among body areas are not unexpected due to intrinsic differences in cortical and trabecular bone composition.
Figure 1: Comparison between DXA Reference and Height- and Weight-Adjusted Z-scores for Total Body minus Head Measurements (Figure 1.1 BMD; Figure 1.2 BMC) for ALGS and CIC disease groups.
The reference line indicates equality between the reference and adjusted Z-scores. Points above the line indicate observations where the Z-score adjusted for height and weight is higher than the reference Z-score. Points below the line indicate observations where the adjusted Z-score is lower than the reference Z-score.
[Footnote] *Non-Winsorized BMD values presented
Abbreviations: BMC, bone mineral content; BMD, bone mineral density; DXA, dual-energy X-ray absorptiometry.
Associations between DXA and laboratory measurements
We examined potential associations between DXA height- and weight-adjusted Z-scores and laboratory measurements using Pearson correlation coefficients. The correlations for ALGS and CIC are shown in Table 4. In ALGS participants, all of the height- and weight-adjusted BMD and BMC measures were negatively correlated with serum total bilirubin (TB; p<0.001). In the ALGS group, many of the DXA measurements were also correlated with INR and albumin. There were no significant correlations found in the CIC (Table 4), A1AT or BASD groups (data not shown).
Table 4:
Pearson Correlations between DXA Height- and Weight-Adjusted Z-Scores and Laboratory Measures of Cholestasis and Liver Synthetic Function
| Alagille Syndrome (n=37-47α) | Chronic Intrahepatic Cholestasis (n=32-40α) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Total bilirubin (mg/dL) | Serum bile acids (µmol/L) | Vitamin D (ng/mL) | INR | Albumin (g/dL) | Total bilirubin (mg/dL) | Serum bile acids (µmol/L) | Vitamin D (ng/mL) | INR | Albumin (g/dL) | |
| BMD | ||||||||||
| Total body minus head | −0.54** | −0.37 | 0.07 | −0.41* | 0.35 | −0.12 | −0.29 | 0.19 | 0.08 | −0.08 |
| Spine | −0.59*** | −0.31 | 0.27 | −0.33 | 0.44* | −0.36 | −0.29 | 0.09 | 0.00 | 0.20 |
| Total hip | −0.66*** | −0.36 | 0.14 | −0.44* | 0.53** | −0.24 | −0.09 | 0.03 | 0.13 | 0.22 |
| BMC | ||||||||||
| Total body minus head | −0.63*** | −0.39 | 0.33 | −0.31 | 0.48** | −0.39 | −0.23 | −0.09 | 0.01 | 0.29 |
number of participants varies due to missing values for DXA or lab measurements
p<0.01
p<0.001
p<0.0001
Abbreviations: BMC, bone mineral content; BMD, bone mineral density; DXA, dual energy x-ray absorptiometry; INR, international normalized ratio.
Supplementary Table S5 compares mean lab values for ALGS and CIC participants with DXA Z-scores above and below −1.5. There were no significant differences found in the CIC group. In ALGS, albumin was lower in participants with height- and weight-adjusted BMD Z-scores < −1.5 compared with those ≥ −1.5.
Differences in DXA Z-scores by bone fracture and biliary diversion
Table 5 shows differences in DXA height- and weight-adjusted Z-scores, TB, and SBAs by whether or not the participant had a bone fracture. There were no significant differences seen for any of the DXA measurements or the lab values in the CIC group. In ALGS, adjusted spine BMD and total body minus head BMC Z-scores were lower in fracture group (−0.6 vs. 0.6, p=0.02 and −0.2 and 0.7, p=0.02). Serum TB and SBAs were both higher in the fracture group (6.1 vs. 2.0, p=0.03 and 157 vs. 80, p=0.02) for ALGS participants. However, the p-value of 0.02 for these comparisons does not meet our cutoff for statistical significance of p<0.01.
Table 5:
Differences in DXA Height- and Weight-Adjusted Z-Score Measures and Lab Values by Occurrence of Bone Fracture*
| Alagille Syndrome (n=49) | Chronic Intrahepatic Cholestasis (n=41) | |||||
|---|---|---|---|---|---|---|
| n | Mean (SD) | p-value** | n | Mean (SD) | p-value** | |
| BMD | ||||||
| Total body minus head | 0.07 | 0.33 | ||||
| Bone fracture | 19 | −0.3 (1.8) | 10 | −0.8 (1.7) | ||
| No bone fracture | 29 | 0.6 (1.5) | 30 | −0.2 (1.4) | ||
| Spine | 0.02 | 0.12 | ||||
| Bone fracture | 19 | −0.6 (1.7) | 10 | −1.6 (1.0) | ||
| No bone fracture | 29 | 0.6 (1.4) | 30 | −1.0 (0.9) | ||
| Total hip | 0.21 | |||||
| Bone fracture | 18 | −0.8 (1.6) | 10 | −1.5 (1.5) | 0.41 | |
| No bone fracture | 28 | −0.2 (1.3) | 29 | −1.1 (1.2) | ||
| BMC | ||||||
| Total body minus head | 0.02 | 0.96 | ||||
| Bone fracture | 19 | −0.2 (1.4) | 10 | −0.8 (1.2) | ||
| No bone fracture | 29 | 0.7 (1.2) | 30 | −0.8 (1.2) | ||
| Lab measures | ||||||
| Total bilirubin | 0.03 | 0.89 | ||||
| Bone fracture | 17 | 6.1 (6.8) | 9 | 1.5 (1.4) | ||
| No bone fracture | 25 | 2.0 (2.7) | 29 | 1.6 (2.1) | ||
| Serum bile acids | 0.02 | 0.58 | ||||
| Bone fracture | 19 | 156.9 (119.2) | 8 | 87.6 (95.7) | ||
| No bone fracture | 20 | 79.7 (77.9) | 26 | 64.3 (105.3) | ||
Bone fracture could have occurred before or after the DXA scan.
p-value based on two-sample t-test.
Abbreviations: BMC, bone mineral content; BMD, bone mineral density; DXA, dual energy x-ray absorptiometry; SD, standard deviation.
We also examined DXA scores by biliary diversion status. There were nine (18%) ALGS participants with a biliary diversion performed prior to the DXA scan. The median time between diversion and the DXA scan was 7 years. The CIC group had 23 (56%) participants with a biliary diversion performed prior to the DXA scan. The median time between diversion and DXA was 6 years. Table 6 shows mean DXA height- and weight-adjusted Z-scores for ALGS and CIC by the diversion status. The spine DXA Z-score was lower in ALGS participants with biliary diversion compared with those without (−1.1 vs. 0.4, p=0.013). In participants with CIC, the mean total body minus head BMD Z-score was higher in the diversion group (0.1 vs. −1.0, p=0.03). However, these differences do not reach the threshold of p < 0.01 for statistical significance in this study.
Table 6:
Differences in DXA Height- and Weight-Adjusted Z-Score Measures by Biliary Diversion*
| Alagille Syndrome (n=49) | Chronic Intrahepatic Cholestasis (n=41) | |||||
|---|---|---|---|---|---|---|
| N | Mean (SD) | p-value** | N | Mean (SD) | p-value** | |
| BMD | ||||||
| Total body minus head | 0.34 | 0.03 | ||||
| Diversion | 9 | −0.2 (1. 9) | 23 | 0.1 (1.5) | ||
| No diversion | 39 | 0.4 (1.6) | 17 | −1.0 (1.2) | ||
| Spine | 0.013 | 0.90 | ||||
| Diversion | 9 | −1.1 (1. 5) | 23 | −1.1 (1.0) | ||
| No diversion | 39 | 0.4 (1.6) | 17 | −1.2 (1.0) | ||
| Total hip | 0.10 | |||||
| Diversion | 9 | −1.1 (1.1) | 22 | −1.0 (1.4) | 0.40 | |
| No diversion | 37 | −0.3 (1.4) | 17 | −1.4 (1.1) | ||
| BMC | ||||||
| Total body minus head | 0.13 | 0.65 | ||||
| Diversion | 9 | −0.2 (1.5) | 23 | −0.7 (1.2) | ||
| No diversion | 39 | 0.5 (1.3) | 17 | −0.9 (1.2) | ||
All diversions were performed prior to DXA scans. The median (IQR) times between diversion and scan for ALGS and CIC were 7 (7-10) and 6 (5-10) years, respectively.
p-value based on two-sample t-test
Abbreviations: ALGS, Alagille syndrome; BMC, bone mineral content; BMD, bone mineral density; CIC, chronic intrahepatic cholestatis; DXA, dual energy x-ray absorptiometry; IQR, interquartile range; SD, standard deviation.
DISCUSSION
In this multicenter study, we report the results of DXA analysis in children with four types of intrahepatic cholestasis – ALGS, CIC (including genetically determined PFIC), A1AT, and BASD. Importantly, participants with A1AT or BASD did not have meaningful bone mineral deficits detected by DXA, but these participants were also only mildly cholestatic at the time of the DXA scan (Table 1). In contrast, study participants with ALGS and CIC had significant bone mineral deficits that correlated strongly with growth parameters. Adjustment for height and weight normalized DXA BMD and BMC Z-scores in ALGS but to a lesser extent in CIC. In ALGS, weight- and height-adjusted DXA Z-scores correlated with laboratory indicators of cholestasis and hepatic synthetic function. Bone fracture occurrence correlated with DXA parameters and lab values in the ALGS group, who were prone to fractures. In this cohort, biliary diversion did not have a major impact on DXA Z-scores.
DXA is a two-dimensional projection technique that summarizes total bone mass within the projected bone area. This measurement of areal BMD (g/cm2) is known to systematically underestimate volumetric BMD (g/cm3) in children with poor growth (18), and various methods have been used to adjust for body size (19). In light of this intrinsic limitation of DXA analysis, and the fact that in our study cohort, we observed a strong correlation between DXA measurements and anthropometrics (Table 3), we elected to adjust for height and weight using the method of Short et al (15). Although the ALGS and CIC participants were similar with respect to weight, height, and BMI Z-scores, only the ALGS group showed a significant normalization of DXA Z-scores after anthropometric adjustment. The reasons underlying this divergent phenotype are not entirely clear, but are likely to be multifactorial. Individuals with ALGS are at risk of growth delay associated with the multisystem involvement of the disease and the underlying genetic defect in Notch pathway signaling. Both ALGS and CIC patients can have profound cholestasis that would be expected to impact absorption of nutrients to a similar extent. A subset of CIC patients, specifically those with mutations in ATPase phospholipid transporting 8B1 (ATP8B1), may have intestinal manifestations that could further contribute to malabsorption. Cytokine and immune mediated alterations may also contribute to bone mineral deficits in patients with chronic cholestatic liver disease. One study of 16 pediatric patients with biliary atresia, ALGS, and PFIC found elevated circulating levels of interleukin (IL)-8, which correlated with severity of hepatic fibrosis (20). In addition, immune dysregulation has been reported in ALGS as a result of altered cluster of differentiation (CD)46-Notch pathway crosstalk (21,22). In our study, we did not measure levels of inflammatory cytokines that could have a potential impact on bone health in this patient population. Further studies will be required to determine the etiology of persistent bone mineral deficits in the CIC cohort after adjustment for patient size.
One of the key findings in this study is the strong correlation between weight- and height-adjusted DXA Z-scores and laboratory measures of cholestasis and liver synthetic function, specifically in ALGS (Table 4). Similar correlations were not identified in the CIC group. Notably, 25-OH vitamin D levels were not correlated with DXA Z-scores in this cholestatic population. The effect of cholestasis on bone metabolism has been studied previously both in vitro and in vivo. Ruiz-Gaspa et al treated human primary osteoblasts with bilirubin or serum from jaundiced patients and found decreased osteoblast viability and differentiation and reduced expression of osteogenic transcription factors and up-regulation of factors inducing osteoclastogenesis (23). Histomorphometric analysis of 50 patients with PBC and PSC at the time of liver transplantation showed decreased bone formation in both males and females and increased bone resorption only in females (24). These studies support the hypothesis that cholestasis has direct effects on bone remodeling. The divergent results for the ALGS and CIC groups with respect to correlations between DXA measures and cholestasis are not well-understood, but may be explained to some extent by the fact that ALGS participants in this study had a higher mean TB and SBA levels as compared with the CIC group. However, the majority of the CIC participants (56%) had undergone biliary diversion, indicating a history of longstanding and profound cholestasis at some point in the past, which is unlikely to be reflected by current laboratory values. Varying clinical response to biliary diversion, ranging from complete resolution of cholestasis to ongoing issues with fat-soluble vitamin deficiencies, complicates interpretation of these results.
ALGS is an autosomal dominant disorder caused by mutations in the Notch ligand JAG1 gene in over 90% of cases, with mutations in NOTCH2 found in a small percentage of patients (25,26). JAG1 encodes the Jagged1 ligand in the Notch signaling pathway, which is essential for cell fate decisions in multiple tissues and cell types, accounting for the multi-system involvement in ALGS (27). Skeletal anomalies, primarily facial dysmorphism and butterfly vertebrae, are well-recognized clinical features of ALGS. Interestingly, a genome-wide association study in Chinese women identified a single nucleotide polymorphism (SNP) in the JAG1 gene to have a protective effect on BMD and osteoporotic fractures (28). The risk SNP was found to increase JAG1 expression in human bone-derived cells. This finding, which has now been replicated in other ethnic cohorts, indicates a role for JAG1 and Notch signaling in bone homeostasis. Additional studies support an important function for Notch signaling in skeletal development, as well as postnatal bone maintenance and remodeling. Data derived from animal models suggest that the underlying genetic defect in ALGS may lead to structural abnormalities of the cortical and trabecular bone, which may not be adequately characterized by DXA analysis (29–31).
In this study, weight- and height-adjusted DXA Z-scores were lower in ALGS and CIC participants who had fractures as compared with those who did not, and total bilirubin and serum bile acids were higher in the ALGS fracture group (Table 5). In ALGS but not in CIC, differences in DXA measures in participants with a history of fracture reached a marginal statistical significance. It is unclear why DXA measures would be lower in ALGS patients with a fracture history and not in CIC; it is possible that Notch signaling deficits play a prominent role in fracture development separate from the severity of cholestasis in ALGS. In this case, the CIC cohort would be an ideal cholestatic control population for study of this relationship. However, variable timing of bone fracture and biliary diversion complicates interpretation of the DXA data in the CIC group. In healthy children, there is evidence that bone mass is lower in those with a fracture history. BMC and areal BMD measures were lower at all body sites in normal children who had sustained fractures compared with those who had not (32). Modest decreases in bone mass have been observed in children with a fracture history (33). Previous studies have shown increased fracture vulnerability in ALGS. In a cohort of ALGS patients, 12 of 42 (28%) reported 27 fractures (4) which occurred at a young age, predominantly in the lower extremities and with minimal trauma. Femur fractures were 50 times more likely than expected for the general population (4). The modest differences in DXA Z-scores identified in our ALGS cohort do not provide an explanation for predisposition to pathologic fracture in ALGS. The genetic underpinnings of ALGS may lead to intrinsic abnormalities of bone strength and composition that are underestimated by DXA imaging techniques.
In this study, we found no statistically significant differences in height- and weight-adjusted DXA Z-scores between participants with and without biliary diversion. Variable clinical response to biliary diversion significantly complicates the interpretation of the results of these analyses (34). Some patients may respond to biliary diversion with normalization of bilirubin and bile acids, but low intestinal luminal bile acid levels may lead to profound fat soluble vitamin deficiencies. Limited numbers of participants who underwent biliary diversion precluded sub-analysis of bile flow after the procedure with regard to clinical response and vitamin D status. After successful drainage procedures, some patients develop bouts of cholestasis that can last months, akin to BRIC (35). As such, the timing of DXA scanning relative to intermittent cholestasis events could affect the DXA findings.
Our study has some limitations. Since our study is cross-sectional in design, it was not possible to correlate bone mineral deficits with progression or improvement of underlying liver disease. Given that these inherited pediatric liver disorders are all quite rare, population sizes were limited and did not allow for multiple comparisons. In addition, only 40% of eligible ALGS participants and 51% of eligible CIC participants underwent DXA analysis. Within the CIC cohort, small numbers precluded sub-analysis of specific genetic etiology of liver disease. This was especially true in the analysis of those individuals with biliary diversion. In this group, it would have been ideal to subdivide the patients by level of cholestasis and vitamin D sufficiency, but this was not possible due to limited numbers of participants.
In conclusion, our results represent the largest study, to date, of DXA analysis in a well-characterized cohort of children with intrahepatic cholestatic liver diseases. In particular, children with ALGS and CIC have significant bone mineral deficits that can impact clinical outcomes. There are clearly complex and poorly understood multifactorial influences on bone mineralization in this population, with varying effects of growth, degree of cholestasis, fracture vulnerability, and direct contribution of underlying genetic etiology. Fractures in this population are unlikely to be simple manifestations of vitamin D deficiency, but are the result of a much more complex pathophysiology. Further studies will be required to fully characterize bone health in this unique patient population, to develop methods to predict patients at high risk for fracture, and to generate new approaches to prevent those fractures.
Supplementary Material
ACKNOWLEDGEMENTS
The authors would like to acknowledge the ChiLDReN study coordinators and all of the patients and families who participated in the study.
Heather Van Doren, MFA, senior medical editor with Arbor Research Collaborative for Health, provided editorial assistance on this manuscript.
Financial Support
This work was supported by U01 grants from the National Institute of Diabetes and Digestive and Kidney Diseases, (DK 62497 [Cincinnati Children’s Hospital Medical Center], DK 62481 [The Children’s Hospital of Philadelphia], DK 62456 [The University of Michigan], DK 84536 [Indiana University Hospital for Children], DK 62500 [UCSF Children’s Hospital], DK 62503 [Johns Hopkins School of Medicine], DK 62466 [Children’s Hospital of Pittsburgh of UPMC], DK 62453 [Children’s Hospital Colorado and University of Colorado], DK 62452 [Washington University School of Medicine], DK 62436 [Ann & Robert H Lurie Children’s Hospital of Chicago], DK103149 [Baylor College of Medicine], DK103135 [The Hospital for Sick Children], DK 84575 [Seattle Children’s Hospital], DK 62470 [Children’s Healthcare of Atlanta]). Note: Saint Louis University School of Medicine (LOGIC) is a sub‐site included via the parent grant of Colorado.
In addition, the project described was supported by National Institutes of Health, National Center for Advancing Translational Sciences, Clinical and Translational Sciences Award (NCATS CTSA) grants: University of Colorado UL1 TR001082, Cincinnati Children’s Hospital Medical Center UL1 TR000077, UCSF Children’s Hospital UL1 TR001872, Indiana University Hospital for Children UL1 TR001108, Children’s Hospital of Pittsburgh of UPMC UL1 TR001857, The Children’s Hospital of Philadelphia UL1 TR001878, Seattle Children’s Hospital UL1 TR000423 and UL1 RR025014, Children’s Healthcare of Atlanta UL1 TR000454.
List of Abbreviations
- A1AT
alpha-1 antitrypsin deficiency
- ALGS
Alagille syndrome
- ALT
alanine transaminase
- ANOVA
analysis of variance
- AST
aspartate transaminase
- ATP8B1
ATPase phospholipid transporting 8B1
- BASD
bile acid synthetic disorder
- BMC
bone mineral content
- BMD
bone mineral density
- BMI
body mass index
- BRIC
benign recurrent intrahepatic cholestasis
- CD46
cluster of differentiation 46
- ChiLDReN
Childhood Liver Disease Research Network
- CIC
chronic intrahepatic cholestasis
- DXA
dual-energy X-ray absorptiometry
- FAB-MS
fast atom bombardment-mass spectrometry
- fx
bone fractures
- GC-MS
gas chromatography-mass spectrometry
- GGT
gamma-glutamyl transpeptidase
- IL-8
interleukin-8
- INR
international normalized ratio
- Jag1
Jagged1
- JAG1
JAGGED1
- LOGIC
Longitudinal Study of Genetic Causes of Intrahepatic Cholestasis
- NIDDK
National Institute of Diabetes and Digestive and Kidney Diseases
- PBC
primary biliary cholangitis
- PFIC
progressive familial intrahepatic cholestasis
- PI
protease inhibitor
- PISZ
protease inhibitor phenotype SZ
- PIZZ
protease inhibitor phenotype ZZ
- PSC
primary sclerosing cholangitis
- SBA
serum bile acids
- SNP
single nucleotide polymorphism
- TB
total bilirubin
- WBC
white blood cell
REFERENCES
- 1.Mahmoudi A, Sellier N, Reboul-Marty J, Chalès G, Lalatonne Y, Bourcier V, et al. Bone mineral density assessed by dual-energy X-ray absorptiometry in patients with viral or alcoholic compensated cirrhosis. A prospective study. Clinics and Research in Hepatology and Gastroenterology. 2011;35:731–737. [DOI] [PubMed] [Google Scholar]
- 2.Vargas AA, Acevedo JMP, Domingo IG, Jiménez RG, Martín JMS, Ríos MTF, et al. Prevalence and Characteristics of Bone Disease in Cirrhotic Patients Under Evaluation for Liver Transplantation. TPS. 2012;44:1496–1498. [DOI] [PubMed] [Google Scholar]
- 3.Guichelaar MMJ, Kendall R, Malinchoc M, Hay JE. Bone mineral density before and after OLT: Long-term follow-up and predictive factors. Liver Transpl. 2006;12:1390–1402. [DOI] [PubMed] [Google Scholar]
- 4.Bales CB, Kamath BM, Munoz PS, Nguyen A, Piccoli DA, Spinner NB, et al. Pathologic Lower Extremity Fractures in Children With Alagille Syndrome. Journal of Pediatric Gastroenterology and Nutrition. 2010;51:66–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hoffenberg EJ, Narkewicz MR, Sondheimer JM, Smith DJ, Silverman A, Sokol RJ. Outcome of syndromic paucity of interlobular bile ducts (Alagille syndrome) with onset of cholestasis in infancy. The Journal of Pediatrics. 1995;127:220–224. [DOI] [PubMed] [Google Scholar]
- 6.Chongsrisawat V, Ruttanamongkol P, Chaiwatanarat T, Chandrakamol B, Poovorawan Y. Bone density and 25-hydroxy vitamin D in extrahepatic biliary atresia. Pediatr Surg Int. 2001;17:604–608. [DOI] [PubMed] [Google Scholar]
- 7.Chen HL, Chang MH. Growth failure and metabolic bone disease in progressive familial intrahepatic cholestasis. Journal of Pediatric Gastroenterology and Nutrition. 2004;39:328–330. [DOI] [PubMed] [Google Scholar]
- 8.Guthery SL, Pohl JF, Bucuvalas JC, Alonso MH, Ryckman FC, Balistreri WF, et al. Bone Mineral Density in Long-Term Survivors Following Pediatric Liver Transplantation. Liver Transpl. 2003;9:365–370. [DOI] [PubMed] [Google Scholar]
- 9.Kryskiewicz E, Pawlowska J, Pludowski P, Ismail H, Karczmarewicz E, Teisseyre M, et al. Bone Metabolism in Cholestatic Children Before and After Living-Related Liver Transplantation—a Long-Term Prospective Study. Journal of Clinical Densitometry. 2012;15:233–240. [DOI] [PubMed] [Google Scholar]
- 10.Olsen IE, Ittenbach RF, Rovner AJ, Leonard MB, Mulberg AE, Stallings VA, et al. Deficits in Size-Adjusted Bone Mass in Children with Alagille Syndrome. Journal of Pediatric Gastroenterology and Nutrition. 2005;40:76–82. [DOI] [PubMed] [Google Scholar]
- 11.Kramer RA, Zemel BS, Arvay-Nezu JL, Stallings VA, Leonard MB, Haber BA. Bone Health in a Nonjaundiced Population of Children with Biliary Atresia. Gastroenterology Research and Practice. 2009;2009:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hangartner TN. A study of the long-term precision of dual-energy X-ray absorptiometry bone densitometers and implications for the validity of the least-significant-change calculation. Osteoporos Int. 2006;18:513–523. [DOI] [PubMed] [Google Scholar]
- 13.Hanson J Standardization of femur BMD. J Bone Miner Res. 1997;12:1316–1317. [DOI] [PubMed] [Google Scholar]
- 14.Zemel BS, Kalkwarf HJ, Gilsanz V, Lappe JM, Oberfield S, Shepherd JA, et al. Revised Reference Curves for Bone Mineral Content and Areal Bone Mineral Density According to Age and Sex for Black and Non-Black Children: Results of the Bone Mineral Density in Childhood Study. The Journal of Clinical Endocrinology & Metabolism. 2011;96:3160–3169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Short DF, Gilsanz V, Kalkwarf HJ, Lappe JM, Oberfield S, Shepherd JA, et al. Anthropometric models of bone mineral content and areal bone mineral density based on the bone mineral density in childhood study. Osteoporos Int. 2014;26:1099–1108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wilcox R, Keselman HJ. Modern robust data analysis methods: measures of central tendency. Psychological Methods. 2003;8:254–274. [DOI] [PubMed] [Google Scholar]
- 17.Shneider BL, Abel B, Haber B, Karpen SJ, Magee JC, Romero R, et al. Portal hypertension in children and young adults with biliary atresia. Journal of Pediatric Gastroenterology and Nutrition. 2012;55:567–573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Zemel BS, Leonard MB, Kelly A, Lappe JM, Gilsanz V, Oberfield S, et al. Height Adjustment in Assessing Dual Energy X-Ray Absorptiometry Measurements of Bone Mass and Density in Children. The Journal of Clinical Endocrinology & Metabolism. 2010;95:1265–1273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wasserman H, O’Donnell JM, Gordon CM. Use of dual energy X-ray absorptiometry in pediatric patients. Bone. 2017;:1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Nobili V, Marcellini M, Giovannelli L, Girolami E, Muratori F, Giannone G, et al. Association of Serum Interleukin-8 Levels with the Degree of Fibrosis in Infants with Chronic Liver Disease. Journal of Pediatric Gastroenterology and Nutrition. 2004;39:540–544. [DOI] [PubMed] [Google Scholar]
- 21.Shamoun ST, Le Friec G, Spinner N, Kemper C, Baker AJ. Immune dysregulation in Alagille syndrome: A new feature of the evolving phenotype. Clinics and Research in Hepatology and Gastroenterology. 2015;39:566–569. [DOI] [PubMed] [Google Scholar]
- 22.Le Friec G, Sheppard D, Whiteman P, Karsten CM, Shamoun SA- T, Laing A, et al. The CD46-Jagged1 interaction is critical for human TH1 immunity. Nat Immunol. 2012;13:1213–1221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ruiz-Gaspà S, Martinez-Ferrer A, Guañabens N, Dubreuil M, Peris P, Enjuanes A, et al. Effects of bilirubin and sera from jaundiced patients on osteoblasts: Contribution to the development of osteoporosis in liver diseases. Hepatology. 2011;54:2104–2113. [DOI] [PubMed] [Google Scholar]
- 24.Guichelaar MMJ, Malinchoc M, Sibonga JD, Clarke BL, Hay EJ. Bone Histomorphometric Changes After Liver Transplantation for Chronic Cholestatic Liver Disease. J Bone Miner Res. 2003;18:2190–2199. [DOI] [PubMed] [Google Scholar]
- 25.Warthen DMD, Moore ECE, Kamath BMB, Morrissette JJDJ, Sanchez-Lara PAP, Sanchez PP, et al. Jagged1 (JAG1) mutations in Alagille syndrome: increasing the mutation detection rate. Hum Mutat. 2006;27:436–443. [DOI] [PubMed] [Google Scholar]
- 26.Kamath BMB, Bauer RCR, Loomes KMK, Chao GG, Gerfen JJ, Hutchinson AA, et al. NOTCH2 mutations in Alagille syndrome. J Med Genet. 2012;49:138–144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Penton AL, Leonard LD, Spinner NB. Notch signaling in human development and disease. Seminars in Cell and Developmental Biology. 2012;23:450–457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Kung AWC, Xiao S- M, Cherny S, Li GHY, Gao Y, Tso G, et al. REPOR T Association of JAG1 with Bone Mineral Density and Osteoporotic Fractures: A Genome-wide Association Study and Follow-up Replication Studies. Am J Hum Genet. 2010;86:229–239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Hilton MJ, Tu X, Wu X, Bai S, Zhao H, Kobayashi T, et al. Notch signaling maintains bone marrow mesenchymal progenitors by suppressing osteoblast differentiation. Nature Medicine. 2008;14:306–314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Youngstrom DW, Dishowitz MI, Bales CB, Carr E, Mutyaba PL, Kozloff KM, et al. Jagged1 expression by osteoblast-lineage cells regulates trabecular bone mass and periosteal expansion in mice. Bone. 2016;91:64–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Lawal RA, Zhou X, Batey K, Hoffman CM, Georger MA, Radtke F, et al. The Notch Ligand Jagged1 Regulates the Osteoblastic Lineage by Maintaining the Osteoprogenitor Pool. J Bone Miner Res. 2017;31:266–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Manias K, McCabe D, Bishop N. Fractures and recurrent fractures in children; varying effects of environmental factors as well as bone size and mass. Bone. 2006;39:652–657. [DOI] [PubMed] [Google Scholar]
- 33.Clark EM. Association Between Bone Density and Fractures in Children: A Systematic Review and Meta-analysis. PEDIATRICS. 2006;117:e291–e297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Wang KS. Analysis of Surgical Interruption of the Enterohepatic Circulation as a Treatment for Pediatric Cholestasis. Hepatology. 2017;:1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Squires JE, Celik N, Morris A, Soltys K, Mazariegos G, Shneider B, et al. Clinical Variability After Partial External Biliary Diversion in Familial Intrahepatic Cholestasis 1 Deficiency. Journal of Pediatric Gastroenterology and Nutrition. 2017;64:425–430. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.