The value of different CT‐based methods for diagnosing low muscle mass and predicting mortality in patients with cirrhosis

Abstract Background & Aims Low muscle mass impacts on morbidity and mortality in cirrhosis. The skeletal‐muscle index (SMI) is a well‐validated tool to diagnose muscle wasting, but requires specialized radiologic software and expertise. Thus, we compared different Computed tomography (CT)‐based evaluation methods for muscle wasting and their prognostic value in cirrhosis. Methods Consecutive cirrhotic patients included in a prospective registry undergoing abdominal CT scans were analysed. SMI, transversal psoas muscle thickness (TPMT), total psoas volume (TPV) and paraspinal muscle index (PSMI) were measured. Sarcopenia was defined using SMI as a reference method by applying sex‐specific cut‐offs (males: <52.4 cm2/m2; females: <38.5 cm2/m2). Results One hundred and nine patients (71.6% male) of age 57 ± 11 years, MELD 16 (8‐26) and alcoholic liver disease (63.3%) as the main aetiology were included. According to established SMI cut‐offs, low muscle mass was present in 69 patients (63.3%) who also presented with higher MELD (17 vs 14 points; P = .025). The following optimal sex‐specific cut‐offs (men/women) for diagnosing low muscle mass were determined: TPMT: <10.7/ <7.8 mm/m, TPV: <194.9/ <99.2 cm3 and PSMI <26.3/ <20.8 cm2/m2. Thirty (27.5%) patients died during a follow‐up of 15 (0.3‐45.7) months. Univariate competing risks analyses showed a significant risk for mortality according to SMI (aSHR:2.52, 95% CI: 1.03‐6.21, P = .043), TPMT (aSHR: 3.87, 95% CI: 1.4‐8.09, P = .007) and PSMI (aSHR: 2.7, 95% CI: 1.17‐6.23, P = .02), but not TPV (P = .18) derived low muscle mass cut‐offs. In multivariate analysis only TPMT (aSHR: 2.82, 95% CI: 1.20‐6.67, P = .018) was associated with mortality, SMI (aSHR: 1.93, 95% CI: 0.72‐5.16, P = .19) and PSMI (aSHR: 1.93, 95% CI: 0.79‐4.75, P = .15) were not. Conclusion Low muscle mass was highly prevalent in our cohort of patients with cirrhosis. Gender‐specific TPMT, SMI and PSMI cut‐offs for low muscle mass can help identify patients with an increased risk for mortality. Importantly, only TPMT emerged as an independent risk factor for mortality in patients with cirrhosis.


| INTRODUC TI ON
Sarcopenia is highly prevalent in advanced chronic liver disease (ACLD) with reported prevalence rates ranging between 22% and 70%. 1,2 Once sarcopenia develops, prognosis is significantly impaired and patients are at increased risk for liver-related morbidity and mortality. 1,[3][4][5][6] Based on these findings, the 'MELD-Sarcopenia'-Score has been developed and has shown a higher accuracy in predicting mortality within 3 months compared to MELD alone. 7 However, a recent study did not confirm an increase in the prognostic value of sarcopenia in addition to MELD. 4 Computed tomography (CT)-based methods represent the gold standard for diagnosing sarcopenia, 8 with skeletal-muscle index (SMI) being the most commonly used parameter. SMI is calculated from the cross-sectional area of abdominal skeletal muscles at the third lumbar vertebrae, normalized by body height. 8,9 Although the SMI was reported to be an independent risk factor for mortality, 5,10 it has two main limitations: firstly, a specific software is needed to measure the cross-sectional area of abdominal skeletal muscle, and secondly, this measurement requires the expertise of an experienced radiologist.
Hence, several studies have used different CT-based methods to investigate the prognostic value of sarcopenia in cirrhosis. The total psoas muscle area and the corresponding psoas muscle-index have been widely studied 1 but also require volumetry, and thus, specific software. Total psoas muscle volume (TPV) predicted post-operative complications following hepatic surgery. 11 More recently, the paraspinal muscle index (PSMI) has been shown to predict mortality in a large cohort of patients with cirrhosis. 12 Transversal psoas muscle thickness (TPMT) is an easy-to-use and readily available parameter in clinical practice since it is just based on the psoas diameter. Importantly, the TPMT was an independent risk factor for mortality on top of MELD or MELD-Na. 3 The detrimental impact of sarcopenia on survival has subsequently been confirmed by several studies. 4,6,7,13,14 Thus, strategies to improve muscle wasting are now widely accepted as important treatment goals in ACLD patients, 9,15,16 however standardized and clinically feasible assessment of muscle mass is still lacking and randomized controlled trials on specific interventions are rare. Intramuscular testosterone injection has been shown to increase muscle mass in male patients with ACLD, however, no beneficial effects on liver-related outcomes were observed. 17 Therefore, we aimed to compare four different CT-based methods for diagnosing low muscle mass (SMI, TPMT, TPV and PSMI) in regard to their prognostic value for liver-related mortality.

| Patients
Patients from a prospective registry undergoing a CT scan within ±200 days of inclusion were analysed ( Figure 1). Inclusion criteria were: available CT scan, diagnosis of cirrhosis (based on clinical, laboratory or radiological findings) 18 and available information on standard laboratory parameters. Exclusion criteria were: missing CT scan and/or impossibility to calculate SMI and TPMT, missing laboratory parameters or unavailability of clinical follow-up and hepatocellular carcinoma at baseline. Inclusion of patients started in 2012, liver-related complications (ie decompensation) were recorded and patients were followed until transplantation, death or date of last clinical visit. Survival time was assessed from the time of the CT scan until censoring/the first event, as defined above. During this study, patients were treated according to the guidelines in effect during the time. [18][19][20] Out of 338 patients included in the prospective database 219 were excluded because of missing CT scans within ±200 days of study inclusion. Further 10 patients were excluded due impossible calculations of SMI and TPMT (incomplete abdominal CT scan with missing third lumbar vertebrae layer).

| Image analysis
All measurements were obtained on axial CT scans of the abdomen performed on a multidetector CT scanner with a patient size-adapted tube voltage (80-120 kVp) an active tube current modulation. Seventy to one hundred and twenty millilitres (depending on the body weight) of iodinated contrast agents (300-400 mg/mL Iodine concentration) was given intravenously at a peripheral vein at a flow rate of 4-5 mL/s, followed by a saline flush of 20 mL using a power injector. All imaging data were acquired on transverse venous phase images using a soft tissue kernel (B30F), with a section thickness of 3 mm and a reconstruction interval of 2 mm.
The SMI, TPMT, PSMI and psoas volume were calculated in all patients at the level of the third lumbar vertebrae (L3) using OsiriX medical imaging software for iOs (Pixmeo, Version 7.5) and syngo.via software (Siemens Healthcare GmbH, Version VB30).
Total psoas volume (TPV) of the entire psoas muscle was calculated in 79 patients using syngo.via software (Siemens Healthcare

Key points
• Low muscle mass is a risk factor for early hepatic decompensation and death in patients with cirrhosis. Computed tomography (CT) based methods are the gold standard to diagnose low muscle mass.
• In this study we compared four different CT-based methods for the diagnosis of low muscle mass and found the transversal psoas muscle thickness (TPMT) as an independent risk factor for mortality.
• The TPMT is easy to calculate, does not require specific radiologic software and could therefore emerge as a feasible tool for the clinical hepatologist to diagnose low muscle mass.
GmbH, Version VB30). Thirty patients had only upper abdominal CT scans and measurement of TPV was therefore not possible.
Transverse venous phase images of the abdomen were loaded.
As previously described, 21 the third lumbar vertebral body, where both transverse processes were depictable, was identified.
The SMI was defined as the total cross-sectional area of all abdominal muscles at the level of L3 on a single scan image normalized by height: semi-automated demarcation of the muscle tissue was based on Hounsfield unit (HU) thresholds from −29 to +150 with manual correction by the reader. The included muscles are the psoas muscle, erector spinae, quadratus lumborum, transversus abdominis, external and internal obliques and rectus abdominis. The calculated area (=total muscle area = TMA), corrected by height, contributes to the final SMI formula, calculated as followed: TPMT-L3 was defined as the transversal diameter of the right psoas muscle perpendicular to the largest axial psoas muscle diameter at the L3 endplate. The results were normalized to body height and shown as mm/m.
TPMT-umbilical was defined as the transversal diameter of the right psoas muscle perpendicular to the largest axial psoas muscle diameter at the level of the umbilicus. Results were normalized to body height and shown as mm/m. In 11 patients TPMT-umbilical could not measured because of missing umbilicus layers as a result of either CT scans limited to the splenoportal axis or massive ascites.
Paraspinal muscle index was defined as the bilateral, total paraspinal muscle area (psoas major and minor muscles, quadratus lumborum muscles, transversospinal muscles and erector spinae muscles) at the L3 endplate and results were normalized by height and are shown in cm 2 /m 2 .
Total psoas volume of the right psoas muscle was calculated semi-automatically, by manual outlining of the boarders of the muscle, starting at the level of the last thoracic or first lumbar vertebra continuing until the psoas muscle becomes indistinguishable from the iliopsoas muscle. Results are shown in cm 3 .
An independent reader, instructed by a senior board-certified radiologist analysed all variables. In case of SMI and TPMT a second independent reader, instructed by a senior board-certified radiologist additionally analysed the variables, and mean values of both measurements where then taken into account for statistical analysis.

| Statistical analysis
Continuous variables were reported as mean ± SD or median (95% confidence interval, CI) and categorical variables were reported as number (n) of patients with certain characteristic (proportion of patients with certain characteristics, %

| Patient population
One hundred and nine patients were included in this study. Low muscle mass as defined using SMI was present in 69 patients (63.3%; Table 1). The main patient characteristics stratified according to presence or absence of low muscle mass are summarized in

| Establishing low muscle mass cut-offs in cirrhosis for TPMT, PSMI and TPV
Using SMI as the reference method to define muscle wasting, diag-

| Defining cut-offs for muscle mass parameters that correlate with survival
Using mortality as the endpoint, cut-offs were calculated based on AUROC analysis and Youden index. The SMI-AUC was 0.63

| D ISCUSS I ON
Giving the detrimental effects of sarcopenia on liver-related outcomes and mortality in patients with ACLD, reversing muscle-loss represents a management priority in these patients. 9,22 Sarcopenia in ACLD is multifactorial and the pathophysiology include proteinmalnutrition, increased proteolysis from skeletal muscles, accelerated starvation response, physical inactivity or humoral factors such as upregulation of myostatin as a result of hyperammonaemia or hypotestosteronism. 8,9,[22][23][24] Nutritional supplementation with branched chain amino acids (BCAAs) has been shown to improve liver-related outcomes 25 and muscle mass. 26   In our study, we could show that low muscle mass in ACLD can be identified using different CT-based diagnostic algorithms, thereby we found a high prevalence of low muscle mass in up to two thirds of ACLD patients. Importantly, lower mortality-derived muscle mass cut-offs for SMI and similar ones for TPMT and PSMI were found to already be associated with increased mortality in our ACLD cohort. Most strikingly, the TPMT was the only independent predictor of mortality, irrespective of the chosen cut-off (muscle mass-or mortality-derived), on multivariate analyses and therefore even outperforming the SMI that requires much more infrastructure and re- in 10%), which may add substantial variability to the measurement of TPMT that is related to the measurement position, rather than psoas muscle itself. We also measured TPMT at the umbilicus level and also observed considerable heterogeneity regarding the relation between the axial skeleton and the umbilicus and were not able to measure TPMT-umbilical in 11 patients because of missing umbilicus on the CT scan (mostly because of CT scans limited to the splenoportal axis and massive ascites). Nevertheless we found a good prognostic capability of gender specific, mortality-derived TPMT-umbilical cut-offs in the 98 remaining patients ( Figure S2).Thus, we propose to use gender-specific mortality-derived TPMT-L3 cut-offs for men and women at <12 mm/m and at <8 mm/m respectively, to identify low muscle mass already at a level that impacts on mortality in ACLD.
The In conclusion, low muscle mass is highly prevalent in cirrhosis.
The TPMT is a valuable tool to diagnose muscle wasting and to identify patients at risk for increased mortality.  Ultimately, timely diagnosis of low muscle mass at gender-specific TPMT cut-offs might accelerate the initiation of specific treatment strategies against muscle wasting in ACLD patients.

CO N FLI C T O F I NTE R E S T
The authors report no real or potential conflict of interest related to this study. Present key elements of study design early in the paper 5-7 Setting 5 Describe the setting, locations and relevant dates, including periods of recruitment, exposure, follow-up and data collection 5-7 Participants 6 (a) Cohort study-Give the eligibility criteria, and the sources and methods of selection of participants. Describe methods of follow-up Case-control study-Give the eligibility criteria, and the sources and methods of case ascertainment and control selection. Give the rationale for the choice of cases and controls Cross-sectional study-Give the eligibility criteria, and the sources and methods of selection of participants