Impaired Bone Health in Children With Biliary Atresia

Abstract

Objectives: 

The aim of the study was to examine the frequency of rickets and bone fractures and to assess areal bone mineral density (aBMD) in childhood among patients with biliary atresia (BA).

Methods: 

We gathered data on all patients diagnosed with BA in Finland that survived to ≥1 year of age between 1 January 2000 to 30 June 2018. Data on gestational age, birth weight, postsurgical medications, and history of rickets and bone fractures were collected retrospectively. Serum levels of 25-hydroxyvitamin D [25(OH)D] postportoenterostomy (PE) were collected. Plain radiographs and dual energy X-ray absorptiometry (DXA) measurements of study subjects were reviewed.

Results: 

Out of 49 patients, 7 (14%) were diagnosed with rickets during infancy. Clearance of jaundice [odds ratio 0.055, 95% confidence interval [CI] 0.00266–0.393; P < 0.01] was a protective factor against rickets. Sufficient 25(OH)D levels were reached 3 months post-PE. Eleven (22%) patients suffered at least one bone fracture (range 1–9) during childhood and adolescence. In DXA measurements, median lumbar spine aBMD anthropometrically adjusted z-scores were as follows: in native liver survivors 0.8 (interquartile range [IQR] −1.9 to 1.4) at 5 and −0.3 (IQR −1.3 to 0.8) at 10 years and for liver transplanted patients 0.4 (IQR −0.2 to 1.1) at 5 and 0.6 (IQR −0.1 to 1.3) at 10 years.

Conclusions: 

BA patients have an increased risk for rickets and bone fractures compared with the normal population. Most BA patients have aBMD within normal range between 5 and 10 years of age irrespective of liver transplantation status.

What Is Known/What Is New

What Is Known

• Cholestatic liver diseases can cause metabolic bone disease in childhood.
• The frequency of rickets and bone fractures in biliary atresia patients is poorly known.

What Is New

• Biliary atresia patients have an increased risk for rickets and bone fractures in infancy.
• The bone mineral content and areal bone mineral density of biliary atresia patients are within normal range between the ages of 5 and 10 years regardless of liver transplantation status.

Biliary atresia (BA), a rare inflammatory cholangiopathy, is one of the most common causes of cholestatic liver disease in infancy ^(1,2). Metabolic bone disease manifesting as decreased bone density, increased bone fracture rates, and osteoporosis, is a known complication of advanced cholestatic liver diseases in adults ^(3–5). The underlying mechanisms of metabolic bone disease that are associated with cholestatic liver diseases remain poorly understood, although impaired osteoblast function leading to a decreased rate of bone formation is likely to play a role ^(6,7). Similar to adults, metabolic bone disease has been reported in children with end-stage liver disease and cholestatic liver diseases of childhood, such as BA and Alagille syndrome ^(8–11). Current data on the impact of metabolic bone disease on the health of children with BA are, however, very limited ^(9,12–15).

In this study, we investigated the frequency of decreased bone mineral density, bone fractures, and rickets in a cohort of children diagnosed with BA between the ages of 1 and 18 years.

METHODS

Patients

Patients diagnosed with BA in Finland between 2000 and 2017 who had undergone portoenterostomy (PE) or primary liver transplantation (LT) and had survived to ≥1 year of age as of June 30, 2018 were included. Since 2005, the care of BA patients born in Finland has been centralized from all 5 university hospitals to Helsinki University Children's Hospital, which is a tertiary center for pediatric liver diseases with a LT program ⁽¹⁶⁾. We retrospectively collected data on gestational age at birth, birth weight, age at PE, time to clearance of jaundice, usage of intramuscular or intravenous vitamin D after PE, dosage of vitamin D supplementation at discharge after PE and at 3 and 6 months post-PE, usage of phenobarbital, cumulative dosage of post-PE glucocorticoid treatment, and rickets diagnoses from medical records. The cumulative dosage of postsurgical glucocorticoids was converted to prednisolone as described earlier ⁽¹⁷⁾. Clearance of jaundice was defined as serum total bilirubin level <20 μmol/L. Fracture history of study subjects and trauma mechanisms were collected. Low-energy fractures were defined as fractures resulting from falls at standing height or less.

Radiological Examinations

Plain Radiographs

We reviewed all radiological examinations performed in our tertiary hospital between

January 1, 2000 and June 30, 2018 on included patients for diagnoses of rickets and bone fractures. Data on possible diagnoses of rickets and fractures in other tertiary and central hospitals participating in the follow-up of BA patients were collected from medical records. Bone ages from left-hand radiograph were assessed. Bone age was evaluated by a pediatric radiologist based on the Greulich and Pyle ⁽¹⁸⁾ radiographic atlas until 2013 and after 2013 by the automated software BoneXpert ⁽¹⁹⁾.

Dual-energy X-ray Absorptiometry Measurements

All dual-energy X-ray absorptiometry (DXA) measurements performed on patients during the study period were collected from medical records. Bone mineral content (BMC, g) and areal bone mineral density (aBMD, g/cm²) were assessed for lumbar spine (LS) with DXA (DXA; pediatric software, Discovery A, versions 12.1 to 13.5.1; Hologic, Marlborough, MA). DXA measurements were converted to standard z-scores based on age- and sex-adjusted reference data for Caucasian American children ⁽²⁰⁾. If bone age differed ≥1 year from chronological age, the bone age adjusted z-score was used. To account for the possible effect of short stature, LS BMC and aBMD anthropometrically adjusted z-scores were also calculated based on age, sex, race, height, and weight ⁽²¹⁾. Previous studies have shown that end-stage liver disease has a significant impact on bone mineral density in children, but aBMD improves to normal levels at 1 year after transplantation ^(22,23). To compare BMC and aBMD of patients with native or transplanted livers at different ages, patients were only included in one of the groups. As the age at transplantation differed between patients, all DXA measurements before LT and 1 year after LT were excluded from the analyses. To detect vertebral fractures (VF) of the thoracic and lumbar spine, vertebral fracture assessment (VFA) images were obtained with the same DXA equipment ⁽²⁴⁾. All images were analyzed according to the method by Mäkitie et al ⁽²⁵⁾. Assessment of VF were performed by two independent experienced examiners, a pediatrician (S.L.) and a pediatric radiologist (O.L.). Discrepancies between the examiners were resolved by consensus.

Biochemistry

For patients with radiologically confirmed diagnosis of rickets, levels of serum 25-hydroxvitamin D [25(OH)D] and plasma ionized calcium (Ca-ion), parathormone (PTH), alkaline phosphatase (ALP), and phosphate (P) were collected from medical records. Measurements taken within a 2-week period of the date of the radiograph confirming rickets were accepted. 25(OH)D levels for all patients included in the study treated with primary PE were gathered at discharge after PE and at 3, 6, and 12 months (±1 month) post-PE. Vitamin D deficiency was defined as 25(OH)D <30 nmol/L and insufficiency as 25(OH)D 30 to 50 nmol/L. A PTH level of 8.5 pmol/L and ALP level >7.7 μkat/L were considered elevated. P level <1.3 mmol/L was considered decreased. Ca-ion level of 1.16 to 1.39 mmol/L was considered normal.

Definitions of Rickets and Osteoporosis

Rickets was defined as cupping and coarse trabecular pattern of metaphyses, widening of the growth plate in a plain radiograph of any anatomical site, or both ⁽²⁶⁾. Osteoporosis was defined according to the criteria by the International Society for Clinical Densitometry in 2013 ⁽²⁷⁾. Osteoporosis diagnosis required both a clinically significant fracture history and BMC or aBMD z-score ≤−2.0. A significant fracture history consisted of a vertebral compression fracture or 2 or more long bone fractures by ages 10 years, or 3 or more long bone fractures at any age up to age 19 years.

Statistical Analyses

Data are expressed as median with interquartile range (IQR) unless otherwise stated. The relation between patient- and treatment-related factors to the risk of rickets or fractures were analyzed with simple logistic regression. Odds ratios (OR) with 95% confidence intervals (95% CI) were calculated. For BMC and aBMD z-scores, the difference to expected median (0) at different ages was analyzed with Wilcoxon signed-ranked test. Comparisons between NL survivors and LT recipients z-scores were assessed with Mann-Whitney test. For 25(OH)D values, the differences between timepoints were evaluated with Mann-Whitney test. P <0.05 was considered statistically significant. All statistical analyses were performed with GraphPad Prism 8.3.1 (GraphPAD Software, San Diego, CA).

Ethics

This study was approved by the Ethical Committee of Helsinki University Hospital. No informed consent was required for this type of register-based study in Finland.

RESULTS

Patient Characteristics

We traced 63 patients with BA. The initial study inclusion criteria were fulfilled by 50 patients (Figure, Supplemental Digital Content 1, https://links.lww.com/MPG/B919). One patient was excluded based on diagnosis affecting bone maturation (Sotos syndrome). Forty-eight patients underwent PE while 1 patient underwent primary LT. Patient characteristics are shown in Table 1. At the end of follow-up, 28 (57%) patients were alive with NL whereas 19 (39%) had been transplanted at median 1.5 (range 0.4–8.8) years. The median age for patients at the end of follow-up was 8.8 (range 1.7–18.4) years.

TABLE 1 - Patient and treatment characteristics in relation to risk of rickets and risk of fractures

	All patients	Patients with rickets	OR (with 95% CI), P value	Patients with fractures	OR (with 95% CI), P value
Number of patients	49	7		11
Gestational age, wk, median (IQR)	39 (38–40)	39 (38–41)	0.96 (0.56–1.66)	40 (38–42)	1.44 (0.90–2.47)
Females, n (%)	25 (50)	4 (57)	1.33 (0.263–7.5)	5 (45)	0.75 (0.187–2.90)
Birth weight, kg, median (IQR)	3.350 (3.010–3.620)	3.470 (2.560–3.920)	0.89 (0.163–6.0)	3.590 (3.320–3.660)	1.46 (0.344–7.6)
Biliary atresia splenic malformation, n (%)	5 (10)	0 (0)		0 (0)
Age at portoenterostomy (PE), days, median (IQR)	64 (39–85)	64 (49–81)	1.0 (0.97–1.02)	65 (44–87)	1.0 (0.98 to 1.02)
Clearance of jaundice, months, n (%)∗
Yes	34 (70)	1 (14)	0.055 (0.00266–0.393), P < 0.01	4 (36)	0.178 (0.0370–0.76), P < 0.05
No	14 (29)	5 (71)		6 (55)
<3	25 (51)	1 (14)	0.142 (0.0071–0.98), P < 0.05	2 (18)	0.152 (0.0209–0.71), P < 0.05
≥3	23 (47)	5 (71)		8 (73)
Phenobarbital usage, n (%)	18 (37)	3 (43)	2.00 (0.333–12.1)	4 (36)	1.28 (0.285–5.3)
Vitamin D
At discharge, in use, n (%)	28 (58)	2 (29)	0.231 (0.029–1.35)	4 (36)	0.367 (0.077–1.64)
Dosage, μg, median (range)†	40 (10 – 60)	40 (20 – 60)	0.98 (0.921–1.02)	1 (9)	0.985 (0.945–1.02)
3 months postsurgery, in use, n (%)	26 (54)	2 (29)	0.306 (0.036–2.09)	3 (27)	0.235 (0.041–1.15)
Dosage, μg, median (range)†	30 (10 – 60)	20 (10 –30)	0.95 (0.87–1.01)	30 (12 – 40)	0.96 (0.91–1.01)
6 months postsurgery, in use, n (%)	31 (65)	4 (57)	0.52 (0.082–4.3)	5 (45)	0.24 (0.045–1.26)
Dosage, μg, median (range)†	40 (10 – 120)	45 (10 – 100)	1.00 (0.98– 1.04)	50 (10 – 120)	1.00 (0.98–1.04)
Postsurgical prednisolone dosage, n (%) = in use	39 (80)	5 (71)	1.03 (0.135– 21.3)	8 (73)	0.77 (0.144–6.0)
Cumulative per weight, mg/kg, median (IQR)	48 (32 – 74)	42 (17 – 46)	0.98 (0.95–1.01)	45 (17 – 77)	1.00 (0.97–1.02)
Age at the end of follow-up, years, median, (range) ‡	8.8 (1.7–18.4)	8.8 (7.4–13.6)		11 (1.9–16.7)
Status at the end of follow-up, n (%)
Dead§	2 (4)	2 (29)		2 (18)
Alive with native livers	28 (57)	1 (14)		4 (36)
Alive with liver transplant	19 (39)	4 (57)		5 (45)
Liver transplantation before age of 2, n (%)	13 (27)	5 (71)	8.5 (1.41–69.8), P < 0.05	6 (55)	8.0 (1.79–41), P < 0.01

All n (%), odds ratios (with 95% CI) for postsurgical medications calculated only for patients with primary portoenterostomy (total number of patients 48). CI = confidence interval; IQR = interquartile range; OR = odds ratio. Odds ratios with significant P values (defined as P ≤ 0.05) typed in bold.

^∗One patient with rickets and fractures treated with primary LT.

^†Median (range) dosage for patients on vitamin D supplementation.

^‡For patients alive at the end of follow-up with native or transplanted liver.

^§Dead: ages 1.1 and 1.7 years.

Vitamin D

For the 48 patients treated with PE, 25(OH)D levels were available for 21 (44%), 20 (42%), 20 (42%), and 20 (42%) at discharge and at 3, 6, and 12 months post-PE, respectively (7 patients transplanted before 12 months post-PE were excluded). As the median age at PE was 64 days, 25(OH)D levels were recorded between the ages of 3 months (at discharge after PE) to 16 months (12-months post-PE [± 1 month]). Analyses of vitamin D levels were restricted to this timeline. At discharge (median 4 weeks, range 2–8 weeks), most measured values were insufficient (median 21 nmol/L, IQR 0–43 nmol/L) (Fig. 1). Three months post-PE, sufficient 25(OH)D levels (median 69 nmol/L, IQR 19.5–100 nmol/L) were reached. Only 5 (10%) patients received intramuscular or intravenous vitamin D after PE. At discharge, 28 (58%) patients received oral vitamin D and 31 (65%) at 6 months post-PE (Table 1).

FIGURE 1: Serum levels of 25-hydroxyvitamin D at discharge for patients alive with native liver after portoenterostomy and at 3, 6, and 12 months post portoenterostomy. Most measured values were insufficient at discharge with a median of 21.0 (IQR 0–43.0) nmol/L. Sufficient 25(OH)D levels (median 69.0 nmol/L, IQR 19.5–101) were reached at 3 months post-PE with significant increase from discharge levels (P < 0.01). Grey area with dots under 30 nmol/L: 25(OH)D deficiency, grey area under 50 nmol/L: 25(OH)D insufficiency. [25(OH)D] = 25-hydroxyvitamin D; PE = portoenterostomy.

Rickets

Seven (14%) patients were diagnosed with rickets (Table 1). Median age at diagnosis was 0.63 (range 0.60–0.93) years. Six patients had biochemistry measurements available at diagnosis. As shown in Supplemental Digital Content 2, https://links.lww.com/MPG/B920 (Table), 6/6 patients had insufficient 25(OH)D levels, 4/6 had elevated PTH levels, 5/6 had elevated ALP levels, and 4/6 had decreased P levels. Ca-ion levels were within the normal range for age (Table, Supplemental Digital Content 2, https://links.lww.com/MPG/B920). Six out of 7 (85%) patients diagnosed with rickets also had bone fractures at the time of diagnosis; the 1 patient that did not have a fracture was diagnosed based on bone-age radiograph revealing highly transparent bone structure of wrist and hand. In logistic regression, both clearance of jaundice at any stage post-PE (OR 0.055, 95% CI 0.00266–0.393; P < 0.01) and clearance of jaundice before 3 months post-PE (OR 0.142, 95% CI 0.0071–0.98; P < 0.05) were inversely associated with rickets. Ricket diagnosis was directly associated with LT before ages 2 years (OR 8.5, 95% CI 1.41–69.8; P < 0.05).

Bone Fractures

There were 11 patients with at least 1 fracture; 6 (55%) of these were diagnosed with rickets. In the entire cohort, there were a total of 36 fractures with a median of 3 (range 1– 9) per patient. There were altogether 27 fractures during the study period for patients with rickets; the number of fractures per patient ranged from 0 to 9. All fractures in the rickets group were low-energy fractures and none of these patients suffered a fracture after 1 year of age. The other 5 patients had altogether 9 fractures, with a median number of 1 (range 1–3) fracture per patient. The median age for first fracture for nonrickets patients was 1.55 (range 1.40–15.0) years. The most common fracture sites for the entire group were costal (11/36, 31%), fibula (6/36, 17%), tibia (5/36, 14%), and femur (4/36, 11%). All femur fractures were diagnosed in patients with rickets in infancy.

VFA images were available for 29 patients; these patients had 96 VFA images taken at a median age of 7.9 (range 3.4–17.9) years. Thoracic vertebra number 8 was visible in 62% of VFA images, whereas all vertebrae between thoracic vertebra 9 to lumbar vertebra 5 were reliably assessable in over 80% of images. We did not find a single VF in the whole series.

Dual Energy X-ray Absorptiometry Measurements Between Ages 5 and 10 Years

DXA measurements were available for 34 patients between ages 5 and 15 years, including 5 patients with history of rickets. Of these, all were alive at the end of follow-up (17 with NL and 17 with LT). Due to the small number of measurements after the age of 10 years in the NL survivors group, data from ages 5 to 10 years were analyzed. Bone-age adjusted z-scores for LS aBMD (38 of 170 measurements, 22%) were compared with anthropometrically adjusted z-scores. As there was no significant difference (data not shown) between the 2 methods, we used the anthropometrically adjusted z-scores.

At 5 years, median LS aBMD anthropometrically adjusted z-score for NL survivors was 0.8 (IQR −1.9 to 1.4) compared with 0.4 (IQR −0.2 to 1.1) for LT patients (U = 21, n = 5,9, P NS). Between the ages of 5 and 10 years, there were no significant differences in LS aBMD z-scores between the 2 groups (Fig. 2). For NL survivors and LT patients, median LS BMC anthropometrically adjusted z-scores at 5 years were −0.1 (IQR −1.5 to 0.9) and 0.3 (IQR −1.4 to 1.4), respectively (U = 19, n = 5,9, P NS). The median z-scores for LS aBMD or BMC did not differ significantly from the expected median (0) between ages 5 and 10 years (Fig. 2) except for BMC for LT patients at ages 10 years (Fig. 2). Five patients (10%), including 1 LT patient, had LS BMC, an aBMD anthropometrically adjusted z-score ≤−2.0 at least once during follow-up, or both. As none of these patients had any fractures, none of the patients met the diagnostic criteria for osteoporosis.

FIGURE 2: Lumbar spine areal bone mineral density and bone mineral content anthropometrically adjusted z-scores compared for patients surviving with native and transplant livers between ages 5 and 10 years. ^∗Comparison to expected median (0) at different ages (P = NS at all time points for both groups, except for LS BMC for LT patients at 10 years; P < 0.01) Comparisons between the 2 groups at different ages with P = NS at all ages.

DISCUSSION

We assessed the bone health of BA patients surviving with native or transplanted liver between the ages of 1 and 18 years in a national cohort that covered 18 years. Using a rigid definition of rickets that required bone structure changes, we found 7 (14%) patients with this diagnosis. Clearance of jaundice was a protective factor against the risk of rickets, and rickets patients had an elevated risk for LT before 2 years of age. Patients were not at high risk for low aBMD at LS according to anthropometrically adjusted z-scores between the ages of 5 and 10 years.

There is no current data available on the incidence of rickets in the pediatric Finnish population. A recent study from Norway ⁽²⁸⁾ reported an estimated incidence rate of 0.3/10,000 person-years for children under 5 years, whereas a study from Sweden ⁽²⁹⁾ reported an incidence of 14.7/100,000 live births between 1997 and 2014. Kobayashi et al ⁽³⁰⁾ found rickets in 4/15 patients with surgically corrected BA compared with 11/21 patients diagnosed with neonatal hepatitis and 2/4 patients with intrahepatic cholestasis. Heubi et al ⁽³¹⁾ reported 3/6 patients with rickets in infancy for patients with failed PE whereas 0/5 BA patients with successful PE were diagnosed with rickets. In our cohort, the frequency of rickets was 14%, which was considerably higher than in neighboring Scandinavian countries and closer to figures (16%–40%) reported in preterm infants ⁽³²⁾. It is likely that the diagnosis of rickets is associated with the severity of cholestasis and reduction in vitamin D absorption.

Studies evaluating bone fractures in childhood in patients with BA mainly consist of case reports ^(33–35). Katayama et al ⁽³³⁾ reported varying bone structural changes from generalized demineralization to fractures in 8 infants in a series of 38 cases; the timeline was not specified. Derusso et al ⁽³⁴⁾ reported 3 cases of infants between 4 and 18 months of age who had undergone a PE procedure and suffered several low-energy bone fractures. In their cross-sectional study of 219 BA patients with NL ages 5 to 18 years, Ng et al ⁽³⁵⁾ found 15% patients who had suffered a bone fracture since PE. This is consistent with our finding of 11% patients with NL without rickets suffering at least one fracture. Altogether, 22% of our patients experienced at least one fracture, but 75% of all fractures occurred in infancy in connection with rickets. In the general population, in Finland's capital area Helsinki in 2005, Mäyränpää et al ⁽³⁶⁾ reported an overall annual incidence of 163 fractures per 10,000 for children ages 0 to 15 years.

We assessed aBMD at LS, which is considered to reliably describe bone mineral density in child populations with high reproducibility ^(37,38). Chen et al ⁽⁹⁾ reported osteoporotic (1/29) or osteopenic (6/29) LS aBMD values for 7/29 (24%) patients in their cross-sectional study of BA patients surviving with NL. In our study, 10% of our patients had BMC, aBMD z-score ≤−2.0 at least once during follow-up, or both. None of these patients, however, had significant fracture history required for the diagnosis of secondary osteoporosis in children. Kramer et al ⁽¹⁵⁾ observed in a cohort of 16 BA patients surviving with NL, the LS BMC was on average 12% and whole-body-less-head BMC 9% lower than in controls. In their study, bone mineral deficits were greater in older subjects. Consistent with this, in our study median, LS BMC was slightly below average in NL survivors of age 7 to 9 years. In their cross-sectional study of liver-transplanted patients, Guthery et al ⁽³⁹⁾ observed that 2/61 patients with BA had an aBMD z-score for LS lower than −2.0. Consistent with this, only 1/17 of our transplanted patients had an aBMD z-score for LS lower than −2.0 at least once during follow-up.

Systemic corticosteroids increase the risk of osteoporosis in adults and have an impact on aBMD in children ^(40–42). Likewise, long-term use of phenobarbital increases the risk of rickets ^(43,44). Low birth weight and prematurity are other well-known risk factors for metabolic bone disease ⁽³²⁾. In our cohort of 49 patients, we did not find an association between glucocorticoid dosage, use of phenobarbital, gestational age, or birth weight and the risk of rickets or fractures. This is not surprising, as the vast majority of our patients were born term with normal weight. In our cohort, the median time for phenobarbital treatment was 9 months and 18 days for postsurgical glucocorticoid treatment. As bone deficits associated with phenobarbital or glucocorticoids increase with prolonged exposure times, the relatively short treatment times could explain our observations.

Current BA postsurgical treatment strategies include supplementation of fat-soluble vitamins. Despite this, there is evidence of insufficient levels of fat-soluble vitamins even after successful PE ^(45,46). In a retrospective study of 129 BA patients, Ng et al ⁽⁴⁵⁾ reported insufficient 25(OH)D levels (defined as <50 nmol/L) at 1,4, and 6 months post-PE with sufficient levels being reached at 12 months post-PE. In contrast with these findings, in our cohort, sufficient 25(OH)D levels were reached already at 3 months post-PE. Although we did not observe an increased risk between vitamin D supplementation dosage and risk of rickets or fractures in our cohort, it is noteworthy that only 5 (10%) patients received intramuscular vitamin D supplementation after PE and 29 (59%) were taking oral vitamin D supplementation at discharge. Adequate fat-soluble vitamin supplementation and robust follow-up of 25(OH)D levels after PE could theoretically be a viable means of prevention of bone deficits in BA patients.

Our study has its limitations. Despite a span of 18 years and national coverage, our cohort is relatively small with 49 patients. We were unable to analyze DXA measurements in teenage years because of the limited number of measurements available. As our study is retrospective by design, data collection was not systematic and was dependent on clinical indications for measurements. Also because of the retrospective design, we could not gather data on the nutrition patients received during infancy. VFA images offered the possibility for screening vertebral deformities, although they are not as accurate for detecting VF as X-ray images ⁽²⁴⁾.

CONCLUSIONS

In conclusion, we observed that BA patients are at an increased risk for rickets and fractures compared with the normal population. Early clearance of jaundice was a protective factor against rickets and fractures. Reassuringly, aBMD was normal regardless of LT status between the ages of 5 and 10 years. As the overall survival rate for BA patients has improved globally over recent decades, the focus of care has shifted towards maintaining good long-term health. Further work is needed to identify early observation and treatment options for deteriorating bone health in this vulnerable patient group.