Mean liver density independently predicts therapeutic response and liver-related mortality in patients with primary biliary cholangitis Authors: Zilong Wang, Shuo Yang, Juan Zhong, Xinyue Liu, Jiajun Feng, Qian Jin, Wenhui Ren, Jian Wang, Danli Ma, Huiying Rao, Rui Huang Published: 2024 ABSTRACT BackgroundSarcopenia, myosteatosis, and sarcopenic obesity are known to adversely affect nutritional status and quality of life in cirrhotic patients. However, there is limited research on these conditions in individuals with primary biliary cholangitis (PBC). This study aims to identify risk factors for sarcopenia, myosteatosis, and sarcopenic obesity in PBC patients and to investigate their association with therapeutic outcomes and liver-related mortality. Furthermore, this research aims to uncover novel imaging-based risk factors that may influence biochemical response rates and mortality risk.MethodsIn this retrospective study, patients with PBC were enrolled. Sarcopenia and myosteatosis were defined using the third lumbar vertebra (L3) skeletal muscle index (SMI) and skeletal muscle density (SMD) obtained from computed tomography (CT) imaging. The prognosis of PBC patients can be accurately evaluated using the PARIS I criteria.ResultsThe study included 117 PBC patients, of whom 73 (62.4%) exhibited sarcopenia and 51 (43.6%) presented with myosteatosis. Seventeen (14.5%) patients were with sarcopenic obesity. The rate of biochemical response was 38.5%. During a mean follow-up period of 68 months, seven patients died. Patients with sarcopenia were older, with a lower body mass index (BMI) and a higher prevalence of myosteatosis, as well as lower visceral adipose tissue index (VATI) and subcutaneous adipose tissue index (SATI) levels (P<0.05). Patients with myosteatosis were older and had higher levels of albumin, SATI, and VATI (P<0.05). They also had a higher prevalence of sarcopenia and sarcopenia obesity (P<0.05). Multivariate analysis revealed that neither sarcopenia nor myosteatosis impacted the biochemical response rates or liver-related mortality in PBC patients. Conversely, mean liver density was identified as a significant risk factor for biochemical response and liver-related mortality (P<0.05).ConclusionsMean liver density is an independent predictor for biochemical response and liver-related mortality in patients with PBC. Introduction Primary biliary cholangitis (PBC) is a chronic progressive autoimmune cholestatic liver disease that predominantly affects middle-aged women and is often marked by the presence of anti-mitochondrial autoantibodies (AMAs) (1). PBC is characterized by heterogeneity in clinical manifestations, disease trajectories, and therapeutic responses (2,3). Typically, PBC has an insidious onset and is frequently devoid of specific symptoms, which can result in delayed diagnosis and may culminate in end-stage liver disease (4,5). Recent epidemiological data suggest an increase in the prevalence of PBC, with estimates ranging from 21.7 to 39.2 per 100,000 people from 2004 to 2014 (6). Sarcopenia, the decline in muscle mass and function associated with aging, is linked to numerous adverse outcomes, including increased risks of falls, depression, social dysfunction, and mortality (7). Based on this, the progressive increase in obesity worldwide has spawned a new public health problem: sarcopenic obesity, a condition in which sarcopenia and obesity exist at the same time (8). Not only that, the updated European Working Group on Sarcopenia in Older People (EWGSOP) has recently underscored the importance of muscle quality as well as muscle mass in the diagnosis of sarcopenia (9). Muscle quality refers to the micro and macro structure and composition alterations, such as fat infiltration termed myosteatosis. Computed tomography (CT) is the premier non-invasive modality for measuring muscle mass and quantity clinically, with CT assessments at the third lumbar vertebra (L3) level significantly correlated with the muscle of the whole body (10). In patients with PBC, the predominant symptoms of fatigue and asthenia can significantly impair quality of life and increase the risk of falls (11,12). Recent investigations revealed a 23.1% prevalence of sarcopenia among 117 PBC patients, a rate twice as high as anticipated (13,14). Mechanistically, PBC is associated with a spectrum of metabolic and nutritional abnormalities, including impaired vitamin D absorption (15), disrupted bile acid metabolism (16), intestinal microbiota alterations (17), all of which may precipitate changes in body composition. Furthermore, chronic liver inflammation, which is a hallmark of PBC, can also promote progressive muscle wasting and accelerate the progression of sarcopenia (18). Sarcopenia, myosteatosis, and sarcopenic obesity are known to impact nutritional status and quality of life in cirrhotic patients, however, there is a scarcity of research focusing on these conditions in individuals with PBC. Consequently, this retrospective study aims to identify risk factors associated with sarcopenia, myosteatosis, and sarcopenic obesity in PBC patients and to explore their potential relationship with biochemical responses and liver-related mortality risk in this population. Additionally, we seek to uncover novel image-related risk factors influencing biochemical response rates and mortality risk. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-24-553/rc). Methods Data source In this single-center retrospective study, we included all patients diagnosed with PBC in accordance with the 2017 European Association for the Study of the Liver PBC diagnosis guidelines in Peking University People’s Hospital between January 2009 and December 2022 (1). All PBC patients received ursodeoxycholic acid (UDCA) usually at a standard dose of 13–15 mg/kg/day once daily. The inclusion criteria were: (I) aged 18 years or older; (II) complete clinical, laboratory data; (III) CT images at the L3 were available; (IV) availability of follow-up information. Patients were excluded from the study if they met any of the following exclusion criteria: (I) incomplete clinical and imaging information; (II) presence of other liver disease, such as viral hepatitis [the 10th edition of the International Classification of Diseases (ICD-10): B15-19] and alcoholic liver disease (ICD-10: K70); (III) presence of incurable malignancy other than hepatocellular carcinoma (HCC), severe chronic respiratory or heart disease, nephrotic syndrome; (IV) less than 3 years of follow-up. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the ethics committee of Peking University People’s Hospital (No. 2022PHB231-001). Due to its retrospective nature, informed consent was waived. Baseline characteristics We extracted patient hospitalization information from the clinical data repository. The following patient characteristics were collected: age, sex, body mass index (BMI), laboratory indicators [alkaline phosphatase (ALP), gamma-glutamyl transpeptidase (GGT), alanine transaminase (ALT), aspartate aminotransferase (AST), total bilirubin (TBIL), albumin white blood cell (WBC), platelet (PLT), and international normalized ratio (INR)], co-morbidities [autoimmune hepatitis (AIH), primary sclerosing cholangitis (PSC), hypercholesteremia, Sjogren syndrome, and other immune diseases (systemic lupus erythematosus and rheumatoid)]. Model for End-Stage Liver Disease (MELD) score was also calculated. Analysis of CT imaging parameters All CT images were conducted using multi-detector row CT systems (Philips 256-slice iCT scanner, Eindhoven, Netherlands; GE Lightspeed VCT 64 and GE revolution 256 layers, Milwaukee, WI, USA) with the following acquisition parameters (tube voltage 120 kV, tube current 140–190 mAs, slice thickness 5 mm). In our study, the CT scans were obtained within a 6-month window around the time of PBC diagnosis, selected randomly either before or after the diagnosis. We collected the CT-based measures of skeletal muscle areas (SMA), including psoas, erector spinae, quadratus lumborum, transversus abdominis, rectus abdominis, and external and internal obliques, subcutaneous adipose tissue area (SATA) and visceral adipose tissue area (VATA) at the level of the L3 using SliceOmatic V5.0 software (Tomovision, Montreal, Canada) which enabled specific tissue demarcation with standard Hounsfield unit (HU) thresholds. The CT thresholds were −29 to 150 HU for skeletal muscle −190 to −30 HU for subcutaneous adipose tissue (SAT), and −150 to −50 HU for visceral adipose tissue. The experienced radiologist was blinded to all the clinical information. The skeletal muscle index (SMI; cm2/m2), visceral adipose tissue index (VATI; cm2/m2), and subcutaneous adipose tissue index (SATI; cm2/m2) were calculated by dividing SMA, VATA, and SATA with the square of height (m2). Visceral fat to subcutaneous fat ratio (VSR) was calculated by dividing VATI with SATI. The mean skeletal muscle density (SMD) was determined as the average density of the SMA at the level of L3. Measurement of mean liver density was performed using the upper abdomen CT and was reported as the average of nine measurements on three slices using circular regions in the right lobe of the liver. Fatty liver was defined as a liver attenuation (LA) value ≤40 HU (19). Sarcopenia was defined as L3 SMI <38 cm2/m2 for females and <42 cm2/m2 for males in patients with liver disease (20). Myosteatosis was defined as the mean SMD at L3 <41 HU for BMI <25 kg/m2 and <33 HU for BMI ≥25 kg/m2 (21). Sarcopenic obesity was characterized by the combination of sarcopenia and obesity in patients with VATA ≥100 cm2 in both sexes (22-24). Study outcomes The primary endpoint was transplant-free survival. The secondary outcome was biochemical response rates after 1 year from UDCA treatment. Incomplete biochemical responders were defined according to the PARIS I criteria, which is ALP ≤3× upper limit of normal (ULN), AST ≤2× ULN, and TBIL ≤1 mg/dL (25). Statistical analysis Continuous variables were expressed as mean ± standard deviation, and categorical variables as frequency (%). The t-test and Chi-squared test were utilized for variable comparison. Multivariate analysis was used to adjust for potential confounders. Multivariate logistic regression analysis was employed to evaluate the risk factors associated with sarcopenia, myosteatosis, sarcopenic obesity, and biochemical response rates. Additionally, multivariable Cox proportional hazards regression analysis was performed, presenting results as hazard ratios (HRs) with 95% confidence intervals (CIs). All statistical analysis was conducted by SPSS 26.0 (Chicago, IL, USA) and R 4.1.2 (Vienna, Austria), with a significance threshold set at P<0.05. Results Baseline characteristics of patients with PBC stratified by sarcopenia A total of 117 patients diagnosed with PBC were enrolled in this study. Among these, 101 were female with a mean age of 62.0±11.1 years. Sarcopenia was identified in 73 patients. Those with sarcopenia were predominantly female, with a higher mean age (64.2±11.2 years) compared to non-sarcopenic counterparts (58.4±9.97 years, P<0.01). They also had a significantly lower BMI (21.8±2.84 vs. 25.3±3.09, P<0.01). Additionally, there was lower proportion of PSC in the sarcopenia group vs. those with no sarcopenia (0% vs. 11.4%, P=0.04). Radiographic assessments demonstrated significant differences, with those with sarcopenia showing higher rates of myosteatosis (53.4% vs. 27.3%, P=0.01) and sarcopenic obesity (23.3% vs. 0%, P<0.01). Furthermore, sarcopenic patients exhibited markedly lower SATI (38.7±16.2 vs. 56.8±21.4, P<0.01) and VATI (27.7±20.8 vs. 42.8±21.3, P<0.01) (Table 1). Multivariate logistic regression analysis identified male gender [odds ratio (OR) =0.07; 95% CI: 0.01–0.37; P<0.01], BMI (OR =0.71; 95% CI: 0.54–0.91; P<0.01), SATI (OR =0.92; 95% CI: 0.87–0.97; P<0.01), and myosteatosis (OR =20.00; 95% CI: 4.67–119.00; P<0.01) as significant predictors of sarcopenia (Table S1). Figure 1 depicts the CT images utilized to assess the muscularity of patients with PBC. Figure 1A,1B shows transverse sections of the abdomen from PBC patients with and without sarcopenia, respectively. Table 1Baseline characteristics by sarcopenia, sarcopenic obesity, and myosteatosis in PBC patientsCharacteristicsTotal (n=117)SarcopeniaMyosteatosisSarcopenic obesityYes (n=73)Non (n=44)P value†Yes (n=51)Non (n=66)P value‡Yes (n=17)Non (n=100)P value§Age (years)62.0±11.164.2±11.258.4±9.97<0.0166.8±11.158.3±10.7<0.0169.3±10.660.8±10.7<0.01Male16 (13.7)7 (9.6)9 (20.5)0.209 (17.6)7 (10.6)0.403 (17.6)13 (13.0)0.90BMI (kg/m2)23.1±3.3621.8±2.8425.3±3.09<0.0123.1±3.2823.1±3.45>0.9924.0±2.2322.9±3.500.10Laboratory parameters ALP (U/L)207±191211±203201±1710.80176±91.4231±2400.09157±75.8216±2040.04 ALT (U/L)49.3±75.850.3±89.547.6±45.80.8033.5±20.861.5±97.90.0336.0±22.451.6±81.40.10 TBIL (μmol/L)36.2±61.834.0±54.439.7±73.00.7030.8±56.240.3±66.00.4018.1±9.9839.2±66.3<0.01 Albumin (G/L) 36.7±6.3537.0±6.2636.3±6.540.6038.0±6.3635.7±6.200.0539.7±5.3536.2±6.380.02 Leukocyte (×109/L)4.69±1.934.54±2.054.94±1.690.305.07±2.124.41±1.730.085.89±2.254.48±1.800.03 Thrombocyte (×109/L)145±83.5141±79.4151±90.60.50149±69.4142±92.80.70169±74.5140±84.60.20 INR1.14±0.231.12±0.171.18±0.300.301.16±0.271.13±0.200.401.09±0.091.15±0.250.10Co-morbidities AIH23 (19.7)12 (16.4)11 (25.0)0.405 (9.8)18 (27.3)0.032 (11.8)21 (21.0)0.60 PSC4 (3.4)0 (0.0)5 (11.4)0.041 (2.0)3 (4.5)0.800 (0.0)4 (4.0)0.90 Hypercholesteremia19 (16.2)8 (11.0)11 (25.0)0.206 (11.8)13 (19.7)0.202 (11.8)17 (17.0)0.90 Sjogren syndrome22 (18.8)17 (23.3)5 (11.4)0.2011 (21.6)11 (16.7)0.703 (17.6)19 (19)>0.99 Other immune diseases12 (10.3)10 (13.7)2 (4.5)0.206 (11.8)6 (9.1)0.902 (11.8)10 (10.0)>0.99MELD score3.41±3.773.19±3.283.75±4.450.503.22±3.933.55±3.670.701.63±1.283.70±3.96<0.01Liver cirrhosis87 (74.4)56 (76.7)31 (70.5)<0.0137 (72.5)50 (75.8)0.8511 (64.7)76 (76.0)0.49Decompensated cirrhosis56 (47.9)37 (50.7)19 (43.2)0.5521 (41.2)35 (53.0)0.284 (23.5)52 (52.0)0.06Ascites51 (43.6)34 (46.6)17 (38.6)0.5220 (39.2)31 (47.0)0.524 (23.5)47 (47.0)0.12Hepatic encephalopathy16 (13.7)8 (11.0)8 (18.2)0.419 (17.6)7 (10.6)0.411 (5.9)15 (15.0)0.53Radiographic analysis L3 SMI (cm2/m2)37.3±7.033.1±4.644.1±4.5<0.0136.0±7.438.2±6.60.1134.0±3.537.8±7.3<0.01 L3 SMD (HU)32.1±7.530.1±4.635.4±6.1<0.0126.4±6.336.5±4.9<0.0125.8±5.833.2±7.2<0.01 SATI45.5±20.338.7±16.256.8±21.4<0.0149.8±17.442.2±21.80.0424.6±12.843.9±20.9<0.01 VATI33.4±22.127.7±20.842.8±21.3<0.0141.8±23.026.9±19.3<0.0159.3±14.729.0±20.1<0.01 Mean liver density (HU)54.6±6.7355.0±6.3354.1±7.380.5053.5±5.8855.5±7.240.1053.7±7.4554.8±6.620.60 Liver density (HU)0.650.820.92 ≤403 (2.6)1 (1.4)2 (4.5)2 (3.9)1 (1.5)1 (5.9)2 (2.0) >40114 (97.4)72 (98.6)42 (95.5)49 (96.1)65 (98.5)16 (94.1)98 (98.0) Sarcopenia––––39 (70.6)34 (51.5)0.0117 (100.0)56 (56.0)<0.01 Myosteatosis51 (43.6)39 (53.4)12 (27.3)0.01–––15 (88.2)36 (36.0)<0.01 Sarcopenic obesity17 (14.5)17 (23.3)0 (0.0)<0.0115 (29.4)2 (3.0)<0.01–––Categorical values are shown as n (%); continuous variables are shown as mean ± standard deviation. PBC, primary biliary cholangitis; BMI, body mass index; ALP, alkaline phosphatase; ALT, alanine aminotransferase; TBIL, total bilirubin; INR, international normalized ratio; AIH, autoimmune hepatitis; PSC, primary sclerosing cholangitis; MELD, Model for End-Stage Liver Disease; L3, third lumbar vertebra; SMI, skeletal muscle index; SMD, skeletal muscle density; HU, Hounsfield unit; SATI, subcutaneous adipose tissue index; VATI, visceral adipose tissue index. Baseline characteristics by sarcopenia, sarcopenic obesity, and myosteatosis in PBC patients Categorical values are shown as n (%); continuous variables are shown as mean ± standard deviation. PBC, primary biliary cholangitis; BMI, body mass index; ALP, alkaline phosphatase; ALT, alanine aminotransferase; TBIL, total bilirubin; INR, international normalized ratio; AIH, autoimmune hepatitis; PSC, primary sclerosing cholangitis; MELD, Model for End-Stage Liver Disease; L3, third lumbar vertebra; SMI, skeletal muscle index; SMD, skeletal muscle density; HU, Hounsfield unit; SATI, subcutaneous adipose tissue index; VATI, visceral adipose tissue index. CT images used for the muscularity assessment of patients with primary biliary cirrhosis. Comparison of two patients without (A) and with (B) sarcopenia (SMI: 60.83 vs. 33.02 cm2/m2). Comparison of two patients without (C) and with (D) sarcopenia obesity (SMI: 50.76 vs. 28.86 cm2/m2; VATA: 36.18 vs. 153.6 cm2). SMAs are highlighted in red. Subcutaneous adipose area is highlighted in green. The visceral adipose area is highlighted in dark blue. Comparison of two patients without [E; patient has normal mean muscle attenuation (44 HU)] and with [F; patient has low mean muscle attenuation (7 HU)] myosteatosis. SMA with high radiodensity (33 to 150 HU) is highlighted in red, and low radiodensity muscle (−29 to 32 HU) is highlighted in yellow. Intramuscular adipose area (−190 to −30 HU) is highlighted in light blue. CT, computed tomography; SMI, skeletal muscle index; VATA, visceral adipose tissue area; SMA, skeletal muscle area; HU, Hounsfield unit. Baseline characteristics of patients with PBC stratified by myostatosis Out of the total cohort, 51 (43.6%) PBC patients presented with myosteatosis. Compared to those without, patients with myosteatosis were older (66.8±11.1 vs. 58.3±10.7 years, P<0.01) and exhibited higher albumin levels (38.0±6.36 vs. 35.7±6.20, P=0.05). Conversely, they had lower ALT levels (33.5±20.8 vs. 61.5±97.9, P=0.03) and fewer comorbid AIH cases (9.8% vs. 27.3%, P=0.03) in the myosteatosis group. Radiographic analysis revealed significant difference, with myosteatosis patients showing higher SATI (49.8±17.4 vs. 42.2±21.8, P=0.04) and VATI (41.8±23.0 vs. 26.9±19.3, P<0.01). Myosteatosis was also associated with a higher prevalence of sarcopenia (70.6% vs. 51.5%, P=0.01) and sarcopenic obesity (29.4% vs. 3.0%, P<0.01) (Table 1). Multivariate logistic regression analysis revealed age (OR =1.08; 95% CI: 1.03–1.15; P<0.01), SATI (OR =1.08; 95% CI: 1.04–1.14; P<0.01), and sarcopenia (OR =18.45; 95% CI: 4.09–113.02; P<0.01) as independent risk factors for myosteatosis (Table S2). Figure 1C,1D depicts transverse sections of the abdomen from PBC patients, illustrating those with and without myosteatosis. Baseline characteristics of patients with PBC stratified by sarcopenic obesity Of all enrolled participants, 17 (14.5%) PBC patients were categorized as having sarcopenic obesity. These patients were generally older (69.3±10.6 vs. 60.8±10.7 years, P<0.01) and exhibited elevated WBC (5.89±2.25 vs. 4.48±1.80, P=0.03), and high albumin levels (39.7±5.35 vs. 36.2±6.38, P=0.02) compared to those without sarcopenic obesity. Conversely, they had a lower ALP level (157±75.8 vs. 216±204, P=0.04), and TBIL (18.1±9.98 vs. 39.2±66.3, P<0.01), as well as lower MELD score (1.63±1.28 vs. 3.70±3.96, P<0.01) in the sarcopenic obesity group. Radiographic analysis uncovered pronounced differences, with the sarcopenic obesity group showing lower SATI (24.6±12.8 vs. 43.9±20.9, P<0.01) and higher VATI (59.3±14.7 vs. 29.0±20.1, P<0.01). The prevalence of sarcopenia (100.0% vs. 56.0%, P<0.01) and myosteatosis (88.2% vs. 36.0%, P<0.01) was notably higher in the sarcopenic obesity group compared to the non-sarcopenic obesity group (Table 1). Multivariate logistic regression analysis pinpointed VATI (OR =1.05; 95% CI: 1.01–1.11; P=0.02) and myosteatosis (OR =8.46; 95% CI: 1.50–79.29; P=0.03) as significant risk factors for sarcopenic obesity (Table S3). Transverse sections of the abdomen depicted in Figure 1E,1F, illustrate the disparities between patients with and without sarcopenic obesity. Predictors of biochemical response after 1 year of therapy Albumin, WBC, PLT, INR, VATI, MELD score, and mean liver density were identified were identified as predictors of biochemical response in the univariate logistic regression analysis, and were subsequently included in the multivariate logistic regression analysis. However, multivariable logistic analysis indicated that sarcopenia and myosteatosis were not associated with biochemical response. Only mean liver density (OR =1.53; 95% CI: 1.12–2.62; P=0.04) was the independent predictor for biochemical responses (Table 2). Table 2Univariate and multivariate OR of predictors for the biochemical response rates (PARIS I criterion)CharacteristicsUnivariableMultivariableOR (95% CI)P valueOR (95% CI)P valueAge0.98 (0.94–1.03)0.430.98 (0.83–1.14)0.75Male2.25 (0.58–9.68)0.253.69 (0.209–139)0.40BMI1.13 (0.98–1.31)0.100.90 (0.56–1.40)0.62Laboratory parameters ALP1.00 (0.98–1.01)0.58 Albumin1.24 (1.12–1.40)<0.011.33 (0.98–2.17)0.15 Leukocyte1.50 (1.13–2.07)<0.010.55 (0.12–1.50)0.31 Thrombocyte1.02 (1.01–1.03)<0.011.00 (0.98–1.03)0.92 INR0.00 (0.00–0.09)0.0118.20 (0.00–1.4e+07)0.68Co-morbidities AIH0.71 (0.38–10.33)0.43521.0 (0.15–3.6e+07)0.17 PSC1.33 (0.05–34.74)0.84 Hypercholesteremia1.47 (0.43–5.23)0.54 Sjogren syndrome0.13 (0.01–0.80)0.07 Other immune diseases0.19 (0.01–1.22)0.14MELD score0.44 (0.24–0.67)<0.010.508 (0.17–1.10)0.13Liver cirrhosis0.04 (0.00–0.25)<0.01Radiographic analysis SATI1.02 (0.99–1.05)0.21 VATI1.03 (1.00–1.05)0.031.06 (0.97–1.21)0.27 Mean liver density1.16 (1.06–1.30)<0.011.53 (1.12–2.62)0.04 Sarcopenia0.54 (0.20–1.47)0.230.004 (0.00–0.874)0.12 Myosteatosis1.38 (0.30–6.35)0.67 Sarcopenic obesity1.27 (0.47–3.45)0.6360.90 (0.58–1.0e+05)0.15OR, odds ratio; CI, confidence interval; BMI, body mass index; ALP, alkaline phosphatase; INR, international normalized ratio; AIH, autoimmune hepatitis; PSC, primary sclerosing cholangitis; MELD, Model for End-Stage Liver Disease; SATI, subcutaneous adipose tissue index; VATI, visceral adipose tissue index. Univariate and multivariate OR of predictors for the biochemical response rates (PARIS I criterion) OR, odds ratio; CI, confidence interval; BMI, body mass index; ALP, alkaline phosphatase; INR, international normalized ratio; AIH, autoimmune hepatitis; PSC, primary sclerosing cholangitis; MELD, Model for End-Stage Liver Disease; SATI, subcutaneous adipose tissue index; VATI, visceral adipose tissue index. Predictors of liver-related mortality of PBC patients Univariate Cox regression analysis showed that high TBIL (HR =1.01; 95% CI: 1.00–1.01; P=0.02) and low mean liver density (HR =0.88; 95% CI: 0.78–1.00; P=0.05) were significantly associated with increased liver-related mortality. The multivariate regression model including male gender (HR =12.70; 95% CI: 1.02–158.21; P=0.05), BMI (HR =0.54, 95% CI: 0.36–0.80; P<0.01), TBIL (HR =1.01; 95% CI: 1.00–1.02; P=0.03) and mean liver density (HR =0.82; 95% CI: 0.69–0.98, P=0.03) as independent predictors of liver-related mortality (Table 3). Nomograms (Figure 2) were constructed based on the significant prognosis factors identified by multivariate Cox regression. These nomograms indicate that mean liver density markedly impacts liver mortality. Table 3Predictors of liver-related mortality of PBC patients in univariable and multivariable Cox regression analysisCharacteristicsUnivariableMultivariableHR (95% CI)P valueHR (95% CI)P valueAge1.02 (0.95–1.10)0.580.99 (0.88–1.11)0.90Male3.39 (0.62–18.49)0.1612.70 (1.02–158.21)0.05BMI0.83 (0.64–1.07)0.150.54 (0.36–0.80)<0.01Laboratory parameters ALP1.00 (0.99–1.00)0.610.99 (0.98–1.00)0.21 ALT0.97 (0.93–1.02)0.26 TBIL1.01 (1.00–1.01)0.021.01 (1.00–1.02)0.03 Albumin1.05 (0.92–1.19)0.46 Leukocyte1.11 (0.72–1.69)0.64 Thrombocyte0.99 (0.98–1.01)0.40 INR7.18 (0.73–70.23)0.09Co-morbidities AIHNA PSCNA HypercholesteremiaNA Sjogren syndromeNA Other immune diseasesNAMELD score1.16 (0.97–1.37)0.10Liver cirrhosisNANARadiographic analysis L3 SMI L3 SMD IMAC SATI0.95 (0.89–1.00)0.10 VATI1.00 (0.97–1.04)0.87 VSR2.95 (0.88–9.87)0.08 Mean liver density0.88 (0.78–1.00)0.050.82 (0.69–0.98)0.03 Sarcopenic0.63 (0.13–3.13)0.570.14 (0.01–1.51)0.11 Myosteatosis 2.89 (0.53–15.84)0.226.15 (0.61–62.48)0.12 Sarcopenic obesity1.38 (0.16–11.95)0.77PBC, primary biliary cholangitis; HR, hazard ratio; CI, confidence interval; BMI, body mass index; ALP, alkaline phosphatase; ALT, alanine aminotransferase; TBIL, total bilirubin; INR, international normalized ratio; AIH, autoimmune hepatitis; PSC, primary sclerosing cholangitis; MELD, Model for End-Stage Liver Disease; L3, third lumbar vertebra; SMI, skeletal muscle index; SMD, skeletal muscle density; IMAC, intramuscular adipose tissue content; SATI, subcutaneous adipose tissue index; VATI, visceral adipose tissue index; VSR, visceral fat to subcutaneous fat ratio. Predictors of liver-related mortality of PBC patients in univariable and multivariable Cox regression analysis PBC, primary biliary cholangitis; HR, hazard ratio; CI, confidence interval; BMI, body mass index; ALP, alkaline phosphatase; ALT, alanine aminotransferase; TBIL, total bilirubin; INR, international normalized ratio; AIH, autoimmune hepatitis; PSC, primary sclerosing cholangitis; MELD, Model for End-Stage Liver Disease; L3, third lumbar vertebra; SMI, skeletal muscle index; SMD, skeletal muscle density; IMAC, intramuscular adipose tissue content; SATI, subcutaneous adipose tissue index; VATI, visceral adipose tissue index; VSR, visceral fat to subcutaneous fat ratio. Visualization of risk factors associated with liver mortality in patients with PBC. Prediction nomogram derived from multivariate Cox regression analysis. Asterisks (*) indicates statistical significance at P<0.05. TB, total bilirubin; BMI, body mass index; PBC, primary biliary cholangitis. Discussion In the present study, we observed that about half of the patients with PBC were diagnosed with sarcopenia or myosteatosis. However, sarcopenic obesity was not so common in PBC. Sarcopenia, myosteatosis and sarcopenic obesity were found to be significantly associated with each other in PBC. Female gender, low BMI, and SATI were predictors of sarcopenia, while aging and high SATI can predict myosteatosis independently. Furthermore, we found a correlation between mean liver density and both the biochemical response to UDCA therapy and liver-related mortality in PBC patients. These findings suggested that imaging-based clinical data may be predictive of prognosis in PBC. In our study, a significant difference was observed in age difference between PBC patients with sarcopenia and non-sarcopenia, myosteatosis and non-myosteatosis, as well as between those with sarcopenic obesity and non-sarcopenic obesity, which is consistent with their definition. Moreover, the distinctly change SATI and VATI were observed among the three groups. Specifically, patients with sarcopenia exhibited reduced SATI and VATI levels. The decline in muscle mass characteristic of sarcopenic individuals may precipitate a decrease in metabolic rate, potentially influencing the distribution and accrual of body fat. Contrasting with sarcopenia, myosteatosis is characterized by elevated SATI and VATI. Visceral fat is metabolically active and can lead to the release of pro-inflammatory cytokines, which in turn may contribute to muscle inflammation and further fat infiltration into the muscle, perpetuating the cycle of myosteatosis (26). In terms of predicting biochemical response, we found that mean liver density was linked to the rates of biochemical response to UDCA. Quantifiable through CT in HU (27), a normal liver typically exhibits smooth contours and uniform density, with attenuation between 50 and 60 HU. PBC patients often display signal abnormalities in the liver parenchyma, indicative of necroinflammatory changes or fibrosis (28). Our findings imply that CT-derived mean liver density could hold clinical significance for indicating biochemical response rates in PBC patients, a relationship previously unexplored. In general, a mean liver density below 40 HU may indicate the presence of a fatty liver. The literature suggests that PBC patients with concomitant non-alcoholic steatohepatitis (NASH) may experience more severe bile duct damage and fibrosis, potentially due to additional hepatic injury from NASH, which could be a factor in the reduced biochemical response to UDCA (29). However, a large sample size is needed to draw a cut-off value or formula, which would greatly benefit its potential clinical application. Moreover, our study demonstrated that the mean liver density is an independent predictor of liver-related death. Changes in liver density may be a manifestation of aggravated liver inflammation, a known feature of PBC (28). A variety of immune cells are involved in the process of liver damage and can result in more severe progression of the disease (30). Thus, clinicians should consider incorporating liver density measurements into their routine assessment of PBC patients. Warrants active screening of PBC patients with liver density is needed, which can help us to identify patients with high-risk of death. Additionally, male gender, elevate BMI, and TBIL levels emerged as independent risk factors for mortality, corroborating the results of other studies. A study of 532 PBC patients indicated that male gender is associated with an increased risk of liver-related mortality. Another study, involving 2,555 participants, found that TBIL levels ≥0.6 times the ULN were linked to the highest mortality risk (31,32). Markers of body composition are known to forecast outcomes across various liver diseases. Current research extends beyond muscle mass to encompass muscle quality, particularly the recognition of myosteatosis, characterized by fat infiltration within muscle tissue. Despite this growing interest, our study found no correlation between either myosteatosis or sarcopenia and the rate of biochemical response to UDCA therapy. The mechanisms underlying the variable response to UDCA are multifaceted and not yet fully elucidated, potentially involving demographic characteristics and immunological factors (2). While sarcopenia and myosteatosis, conversely, are influenced by genetic, nutritional, and age-related factors, as well as physical activity levels (33). Hence it can be considered that body composition markers have a minimal or non-existent relationship with biochemical response. Additionally, no link was observed between sarcopenia or myosteatosis and an increased liver-related mortality risk, which may contradict some literature in other liver diseases including autoimmune liver disease. Recent research involving 116 patients with PSC demonstrated a significant association between myosteatosis and transplant-free survival. In another study, a study of 480 cirrhotic patients showed that the co-occurrence of sarcopenia and myosteatosis could diminish both overall survival and liver transplantation-free survival. Moreover, sarcopenic obesity has been identified as an independent predictor of increased long-term mortality in patients with cirrhosis (34-36). This discrepancy can be attributed to prior research focusing on end-stage liver disease patients, who typically exhibit severe metabolic disturbances, in contrast to the relatively mild metabolic impairment observed in our PBC cohort. Moreover, no association between biochemical response rates and fatigue was identified in a prospective study involving 140 PBC patients (37). There are several limitations in this retrospective study. First, as a single-center study with a small sample size, the results may not be generalizable; thus, multicenter studies with larger populations are warranted to reduce the risk of selection bias. Second, the reported sarcopenia prevalence of 62.4% among PBC patients surpasses figures from existing literature, indicating potential selection bias. This could be due to patients with absent or mild symptoms being less likely to seek hospital care. Third, there was unintended selection bias in this study due to the inclusion and exclusion criteria such as randomness of CT acquisition time. And then, the minimum follow-up duration of 3 years may be insufficient to comprehensively track long-term outcomes in PBC. Lastly, the absence of standardized definitions and diagnostic criteria for body composition markers could have influenced our conclusions. Conclusions In conclusion, our study investigation revealed that mean liver density serves as an independent predictor of both biochemical response and liver-related mortality in patients with PBC. This evidence indicated that markers derived from CT imaging could represent a non-invasive and convenient approach for prognostication in the future clinical practice. Supplementary The article’s supplementary files as