Prevalence of hepatic steatosis and fibrosis in Turner syndrome: A prospective case–control study

Abnormal liver chemistries are common in Turner syndrome (TS). Guidelines suggest that TS patients undergo annual screening of liver enzymes, but the role of non‐invasive screening for steatosis and fibrosis is not clearly defined. We compared the prevalence of hepatic steatosis and fibrosis among TS patients to healthy controls using ultrasound with shear‐wave elastography (SWE) and assessed for risk factors associated with steatosis and fibrosis in TS.


| INTRODUC TI ON
Turner syndrome (TS) is a genetic condition in which females are missing the second sex chromosome. 1,2TS affects 1 in approximately 2000-2500 phenotypic females.Common medical complications associated with TS include cardiac conditions, short stature, autoimmune diseases and infertility. 2Abnormal liver enzymes have been estimated to occur in 20%-80% of TS patients, depending on age. 3,4The most common cause of abnormal liver enzymes in TS is steatosis and steatohepatitis (steatotic liver disease, SLD), which can create architectural changes within the liver parenchyma including fibrosis and cirrhosis. 3[3][4] SLD is estimated to occur in 21% of women in the United States and is the fastest growing cause of liver transplantation in the general population. 5,6[3][4] Due to the wide spectrum of potential liver injury in TS, it is suggested that patients undergo annual serologic monitoring of liver enzymes starting at 10 years of age as a means of early detection and treatment. 2,3However, the frequency at which these guidelines are adhered to and their impact on clinical outcomes is unknown.Furthermore, abnormal liver enzymes have been shown to inconsistently correlate with the degree of underlying chronic liver injury and fibrosis in TS.
The prognosis and management of chronic liver disease depends on the degree of liver fibrosis.This was traditionally assessed with liver biopsy, but newer tests such as serum screening scores (e.g.0][11] Additionally, ultrasound with SWE analyses liver volumes over 100 times bigger than a biopsy sample, providing a result that is more representative of the total hepatic parenchyma. 12evious studies have evaluated the use of non-invasive scores (e.g.FIB-4) and Fibroscan™ in TS, suggesting that these could be helpful risk stratification tools. 1,13One limitation of studies performed to date is the small population sizes used, which has created conflicting information regarding the best clinical approach to liver disease screening in TS.As a result, the application of non-invasive screening methods such as serologic scoring systems, Fibroscan™ or ultrasound with SWE in TS is not clearly defined.
The primary aim of this study was to compare the prevalence of hepatic steatosis and fibrosis between TS patients and healthy controls using ultrasound and SWE.Our secondary aim was to assess for risk factors associated with steatosis and fibrosis in TS.

| Study design and population definitions
We performed a prospective single centre case-control study of

| Assessing fibrosis and steatosis
All patients were placed supine or left semi-lateral decubitus with the 2D ultrasound convex probe placed at the right mid-clavicular line or anterior axillary line in between intercostal spaces until an adequate window of the liver was obtained.All images were blinded to the reviewers.

| Fibrosis: Defined using shear-wave elastography
SWE values are obtained by generating a focused acoustic beam that creates shear wave within the liver, with the degree of fibrosis reflected upon the speed of this wave within the liver parenchyma. 9,10is then creates a calculated speed and pressure to machine software measured in metres per second, and kilopascals (kPA), respectively. 10,11Measurements are obtained at a distance of at least 2 cm from the capsule and away from any large blood vessels.An average of 12 measurements is taken to ensure a representative sample is obtained.An interquartile range (IQR) of less than 25% was used to ensure accurate results and readings.The average of all SWE values obtained was then correlated to the respective Metavir scoring system of fibrosis (Table 1).

Key points
TS is associated with an increased prevalence of hepatic steatosis and fibrosis compared to age-matched controls, even after adjusting for body mass index (BMI).GGT is a more sensitive marker than AST and ALT for early liver disease in TS.Serum GGT and ultrasound with shear-wave elastography could be considered a component of annual health screening in TS patients.

| Steatosis: Image interpretation
To define steatosis, all ultrasound images were assessed separately by all three authors using commonly applied radiologic definitions for mild, moderate and severe hepatic steatosis (Table 2).If conflicting interpretations existed, consensus among all three authors was used to allocate a uniform assessment of steatosis for all patients.

| Covariates of interest
To evaluate risk factors, data were divided into four parts: presence of steatosis, severity of steatosis, presence of fibrosis and severity of fibrosis.Risk factors that were compared between the age-matched TS and control groups included, height, weight, body surface area, body mass index (BMI) and the presence/absence of autoimmune disease.For TS patients only, karyotype, history of growth hormone therapy, cumulative oestrogen deficiency (defined as the number of years a patient had not been using hormone replacement therapy beyond the age of 16 years), history of cardiac disease, history of cardiac surgery and hypertension were assessed for their relationship to hepatic steatosis and fibrosis.Additionally, TS patients underwent fasting laboratories on the same day as their imaging studies including a fasting lipid profile, haemoglobin A1c, gamma glutamyl transferase (GGT), alanine transaminase (ALT), aspartate transaminase (AST), alkaline phosphatase (ALT) and platelet count.From these laboratories, we calculated the AST to platelet ratio (APRI), and FIB-4 score, both validated non-invasive tools for assessing underling liver fibrosis. 8,9

| Statistical analysis
Descriptive statistics were used to compare TS cases and healthy controls.Categorical variables were compared using chi-square analysis, and continuous variables were compared using either t-test or the Wilcoxon rank sum test as appropriate.Multiple regression models were performed within the TS group for each of the covariates independently followed by a combined model for all those that were individually significant at p-value <.05.Analysis was performed using SAS 9.0 (Cary, North Carolina).

| RE SULTS
A total of 55 patients and 50 controls were included.One TS patient and one control were excluded due to use of potentially hepatotoxic medications, and one TS patient was excluded for a prior liver transplant at 10 years of age for hepatic adenomatosis.There was no difference in age, weight or ethnicity between the groups.Although there was no significant difference in body weight between the groups (65.4 kg in TS vs. 60.1 kg in control, p = .16), due to TS patients being significantly shorter, as expected (147.4cm vs. 161.9cm, p < .0001)TS patients had a higher BMI (29.3 vs. 22.7, p < .0001;Table 2).
Within the TS group, there were 21 girls aged 6-18 years with the remaining women aged 19-54 years.The percentage of girls to women in both patient and control groups was identical at 41%.Karyotypes revealed mosaicism with a normal cell line (45,X/46,XX) in 7 (12%), 45,X in 33 (60%), X isochromosome in 8 (14%) and other chromosomal derangements in 7 (14%).

| Risk factors associated with steatosis and fibrosis
Risk factors associated with the presence and severity of steatosis and fibrosis are listed in Table 4. Older age, heavier weight, larger body surface area and higher BMI all had statistically significant TA B L E 1 Shear-wave elastography (SWE) results were recorded in metres per second in the imaging software used for this study.The Metavir scoring system was then used to translate these recordings into the patient's fibrosis score.In TS patients, GGT was better than ALT (F = 6.13, p = .01)and CRP (F = 4.21, p = .04)for detecting the presence/absence of steatosis.There was no difference in predicting the presence of steatosis between Fib-4 and GGT (p = .10)or APRI and GGT (p = .61).GGT was better than APRI (p = .02)and ALT (p = .02)for predicting the presence or absence of fibrosis.GGT was no different than CRP (p = .05) or Fib-4 (p = .08)for predicting presence/absence of fibrosis.
After correcting for age, weight and body surface area, oestrogen deficiency, c-reactive protein, APRI, Fib-4 and hypertension were not significant predictors of the severity of steatosis or fibrosis.

| DISCUSS ION
We present the first known study of ultrasound with SWE in patients with TS.We also present the largest prospective case-control study directly comparing the prevalence of hepatic fibrosis and steatosis in TS to healthy control subjects.
Our findings are consistent with some prior studies which also showed that increasing age, elevated triglycerides and elevated BMI are risk factors for liver disease. 12Additionally, prior studies have found an association between ALT, GGT, BMI and dyslipidaemia, which is also consistent with our study. 13significant finding in our study is that the risk of hepatic steatosis and fibrosis in TS is independent of BMI.Previous studies have suggested that hepatic changes in TS are secondary to many TS patients being overweight and/or having elevated BMI due to the shorter stature in this population.However, our findings suggest that mechanisms outside of weight alone contribute to the hepatic architectural changes in TS.Although the pathophysiologic mechanisms of SLD in TS continue to be explored, we hypothesize that genetic mutations due to TS itself may impact the natural history of SLD, like PNPLA3 or other genetic polymorphisms which predispose to SLD regardless of BMI.Progressive architectural changes to the hepatic parenchyma from these mutations could also explain why older age is a predictor of both steatosis and fibrosis in TS.

TA B L E 3 Summary statistics between TS patients and controls
included in the study.We found no correlation between degree of steatosis or fibrosis and AST, or alkaline phosphatase.Both GGT and ALT correlated with the degree of steatosis and severity of fibrosis.GGT was significantly different between stages 0 and 3 steatosis and stage 2 and 3 steatosis.In addition, increased GGT correlated with severity of fibrosis with correlation coefficient of 0.46.When comparing the two tests, GGT was a significantly better predictor of steatosis and fibrosis than ALT.There are few studies that have evaluated the role of GGT in diagnosing SLD.However, multiple population-based studies suggest that GGT is an independent risk factor for SLD, and that increased GGT levels may be associated with the severity of steatosis and fibrosis. 14,157][18] GGT has low specificity, and its role in TS will need continued study, but our results suggest that it may assist in SLD risk assessment in TS.
Current guidelines recommend annual serologic evaluation. 2 However, assessing for liver fibrosis, the current prognostic marker of chronic liver disease has not been well-established in TS.Although serologic evaluation and the use of non-invasive serum scoring systems such as FIB-4 and NFS can be useful in patients with hepatic steatosis, prior studies have shown that even patients with normal liver enzymes and low serum scores may have advanced fibrosis. 13 a result, serologic scores do not tell the whole story from a prognostic standpoint.
One prior study found that Fibroscan™ can be a useful diagnostic and risk stratification tool for TS as well. 13Although Fibroscan™ has become an increasingly utilized non-invasive tool for staging liver disease, the device is expensive and requires training to perform an accurate study and interpret results.As a result, its utilization as a large-scale screening tool in all TS without evaluation by a hepatologist is limited.On the other hand, ultrasound with SWE can be ordered by any provider and is interpreted by a radiologist.As a result, it is widely available, cost-effective, and could be paired with annual serologic evaluation as a standard screening tool in patients with TS.[21] Our findings suggest that continuing annual serologic evaluation and calculating fibrosis risk scores based on serologic testing is a helpful screening tool, but given its low specificity, may not diagnose steatosis or fibrosis in a subset of TS patients.Our study also shows that GGT is more sensitive than AST or ALT in identifying early liver changes.
Although a prior study 1 included over 2000 TS patients, this cohort did not evaluate GGT or karyotype in TS patients, and was entirely dependent on ICD codes to identify patients which could introduce significant bias.
Although oestrogen therapy is typically associated with cholestatic liver injury, deficiency is also known to be detrimental to liver health. 22Maintaining a steady state of oestrogen should help balance the hepatic effects of deficiency while also avoiding toxicity from excess use.
We suggest the use of ultrasound with SWE to screen for liver disease in TS and risk stratify patients.For TS patients with significant steatosis and/or fibrosis, referral to a hepatologist should be considered to exclude alternative liver diseases common in TS including autoimmune hepatitis and celiac disease, and to provide individualized treatment and surveillance.Early risk factor mitigation including timely oestrogen replacement, growth hormone utilization, weight control and normalization of lipids should be addressed.
TS versus control patients from 2019 to 2021 at the University of Nebraska Medical Center.Patients with genetically confirmed TS living in a three-state region were recruited into the prospective Turner syndrome heart, liver, vascular and endothelial function study through the local TS support group and TS clinics.Healthy, aged-matched controls were recruited from the general local population for portions of the study including liver imaging.No attempt was made to match for height or body mass index to avoid potential inclusion of undiagnosed females with Turner syndrome, or other confounding systemic disorders associated with short stature or childhood obesity.Participants were excluded if they were receiving any potentially hepatotoxic medications, were too young to comply with testing procedures and had other medical conditions associated with altered hepatic architecture (inflammatory bowel disease, celiac disease, type I diabetes, known liver pathology apart from SLD or congestive heart failure).
positive correlations with the presence and severity of both steatosis and fibrosis between TS patients and controls.Within the TS cohort, a history of hypertension and serum GGT and ALT levels were the only other variables significantly positively associated with the presence and severity of steatosis and fibrosis.Absence of growth hormone treatment was related to the presence of steatosis.

1
Results of shear-wave elastography and corresponding fibrosis stage for patients included in this study.Dyslipidaemia and diabetes are established risk factors for SLD in the general population.Interestingly, in our TS study cohort, these risk factors were associated with both the presence and severity of steatosis, but had no significant increase in the presence or severity of fibrosis.These findings add further to the notion that physiologic mechanisms outside of metabolic syndrome may contribute to SLD in TS.
Variables or risk factors associated with the presence and severity of hepatic steatosis and fibrosis.
TA B L E 4