Content area
Abstract
Objectives
Low muscle mass has been found to be associated with adverse outcomes in patients with acute-on-chronic liver failure. However, data regarding the prognostic role of low muscle function are limited. Therefore, we aimed to investigate the predictive effect of low muscle function on 90-d mortality in patients with acute-on-chronic liver failure.
MethodsThis prospective study consecutively enrolled acute-on-chronic liver failure patients from March 2021 to October 2022. Muscle function was assessed using the liver frailty index, and the time-dependent receiver operating characteristic curve with the highest Youden index was used to determine the optimal cutoff values of liver frailty index for diagnosing low muscle function.
ResultsThe study included 126 acute-on-chronic liver failure patients. The median liver frailty index was 3.89 (0.83), with 51 (40.5) patients classified as having low muscle function. Multivariate Cox analysis identified low muscle function (hazard ratio = 4.309; 95% CI, 1.795–10.345; P = 0.001) and number of organ failures (hazard ratio = 4.202; 95% CI, 2.040–8.656; P < 0.001) as independent risk factors for 90-d mortality. However, the multivariate analysis did not retain the significant effect of low muscle mass. Furthermore, multivariable logistic analysis revealed that age (odds ratio = 1.042; 95% CI, 1.002–1.083; P = 0.038), organ failures (odds ratio = 2.572; 95% CI, 1.331–4.968; P = 0.005), and low muscle mass (odds ratio = 6.607; 95% CI, 2.579–16.927; P < 0.001) were independent risk factors for low muscle function.
ConclusionsThe prognostic value of low muscle function was found superior to that of low muscle mass in patients with acute-on-chronic liver failure. Therefore, it is important to assess the muscle function and develop potential targeted treatment strategies in this population.
Full text
Acute-on-chronic liver failure (ACLF) is a clinical syndrome characterized by acute deterioration of liver function in chronic liver disease with or without cirrhosis and accompanied by hepatic and extrahepatic organ failure (OF) [1]. Although many efforts have been made recently to ameliorate the prognosis of ACLF, this condition remains associated with a high risk of short-term mortality. Therefore, enhancing risk stratification for high mortality is important to provide specific interventions to improve these patients’ prognoses. At present, many scoring systems can be used to evaluate the prognosis of ACLF, such as the Model for End-stage Liver Disease for sodium (MELD-Na) score, the Asian-Pacific Association for the Study of the Liver ACLF–Research Consortium score, the Chronic Liver Failure Consortium ACLF score, and the Chinese Group on the Study of Severe Hepatitis B ACLF score. However, these laboratory indicators do not include nutritional or functional status parameters. Because the liver is an important organ involved in protein, fat, and carbohydrate metabolism and energy production, numerous studies have found that malnutrition is an independent predictor of adverse outcomes in patients with ACLF [2–4].
Loss of muscle mass and impaired muscle function are both significant clinical manifestations of malnutrition [5]. Several studies have investigated the high prevalence of low muscle mass and its independent association with adverse outcomes in patients with ACLF [6,7]. Nevertheless, data on the low muscle function in patients with ACLF are limited. Additionally, Some studies suggest that combining assessment of muscle mass and muscle function can enhance the predictive capacity for clinical outcomes in patients with hepatocellular carcinoma [8,9], although it remains unclear whether the combination of low muscle mass and impaired muscle function will predict the worst clinical outcomes in patients with ACLF.
The liver frailty index (LFI), which includes assessment of grip strength, chair stands, and balance, was recently developed by Lai et al. [10] as an objective and simple for measure muscle function and predicting mortality in cirrhotic patients awaiting liver transplantation. In this study, therefore, we aimed to 1) investigate the prognostic value of the low muscle function and the coexistence of low muscle mass and low muscle function on 90-d mortality and 2) determine the potential risk factors of low muscle function, as measured by LFI, in patients with ACLF.
Materials and methods Study populationConsecutively hospitalized patients diagnosed with ACLF at Beijing Youan Hospital, Capital Medical University (Beijing, China), between March 2021 and October 2022, were prospectively enrolled. ACLF was defined as a total bilirubin level ≥5 mg/dL and international normalized ratio ≥ 1.5 for patients with non-cirrhotic liver disease or cirrhosis, according to the Asian-Pacific Association for the Study of the Liver consensus. The exclusion criteria were patients with 1) hepatocellular carcinoma or other malignancy tumors; 2) serious extrahepatic disease, such as respiratory failure, heart failure, or chronic renal insufficiency; 3) neuromuscular diseases or long-term bedridden status; and 4) stage Ⅲ or Ⅳ of hepatic encephalopathy.
The procedures of this prospective study conformed to the Declaration of Helsinki and were approved by the Ethics Committee of the Beijing Youan Hospital, Beijing, China. All participants provided written informed consent (LL-2020–178-K).
Clinical data collectionDemographic and clinical data, including age, sex, height, weight, occurrence of liver cirrhosis, etiology of disease, ascites, hepatic encephalopathy, acute kidney injury (AKI), levels of total bilirubin, albumin, sodium, creatinine, international normalized ratio, white blood cell count, and platelet count, were abstracted from the electronic medical records. The MELD-Na score was calculated to assess the clinical severity of each patient. OFs at admission were evaluated by European Association for the Study of the Liver–Chronic Liver Failure Consortium system [11].
Body mass index (BMI) was calculated as dry weight (kg) divided by height squared (m2) and BMI ≥25 kg/m2 was defined as overweight. The dry weight was measured considering fluid retention by subtracting a certain amount of body weight (mild, 5%; moderate, 10%; severe, 15%; and 5%, if there was peripheral edema) [12]. Patients were followed up by means of a medical record system or telephone. Outcomes were death in ≤90 d of the listing.
Assessment of skeletal muscle areaAbdominal non–contrast-enhanced computed tomography (CT) scans were performed in the supine position using a 64-row CT scanner (GE LightSpeed). CT protocols were followed according to these parameters: tube voltage, 120 kV; tube current, 380 mA; detector collimation, 0.625 mm; slice thickness, 5 mm; reconstruction thickness, 0.625 mm; and pitch, 5 mm. Skeletal muscle area was quantified at the third lumbar (L3) vertebral level using a Hounsfield unit range of –29 to +150 by the image analysis software, sliceOmatic version 5.0 (Tomovision, Magog, Canada). The cross-section of the skeletal muscle area was automatically calculated and normalized for height squared to obtain the L3 skeletal muscle index (SMI) (cm2/m2). We defined low muscle mass as SMI <40.2 cm2/m2 for men and <31.6 cm2/m2 for women, as previously reported [13].
Evaluation of muscle functionMuscle function was evaluated using the LFI based on previous studies [10]. The LFI consists of thrtee components: grip strength, balance testing, and chair stands. Grip strength was measured in the patient's dominant hand using a hand dynamometer (in kilograms), and the mean of three trials was recorded. Chair stands were measured as the number of seconds it takes to stand up and sit down in a chair 5 times with the patient's arms folded across the chest. Balance testing was measured as the number of seconds that a patient can balance in three positions (feet side to side, semitandem, and tandem) for a maximum of 10 s each. The LFI was calculated as (−0.330 × sex-adjusted grip strength) + (−2.529 × number of chair stands per second) + (−0.040 × balance time) + 6.
Statistical analysisNormally and non-normally distributed data were presented as mean ± SD or median (IQR). Categorical variables were presented as numbers (percentages). Significant differences between two groups were analyzed using independent sample t test or Mann-Whitney U test, whereas the differences in multiple groups were analyzed using one-way analysis of variance or Kruskal-Wallis test. Categorical data were compared using χ2 test. The correlations between LFI and age, BMI, MELD-Na, and L3-SMI were visualized with scatterplots and assessed using the Pearson correlation coefficient. We identified the optimal cutoff values of LFI for predicting 90-d mortality based on the time-dependent receiver operating characteristic (ROC) curve with the highest Youden index and divided all patients into two groups stratified by LFI categories (low muscle function versus normal muscle function). The survival curve was generated by the Kaplan-Meier method and compared with the log-rank test. Univariate and multivariate Cox regression were used to analyze the hazard ratio for mortality. Univariable and multivariable logistic regression were used to identify risk factors of low muscle function. Statistical analyses were performed using SPSS version 24.0 (IBM Corp., Armonk, NY) and R × 64 version 4.1.2 (http://www.r-project.org/).
Results Study population and baseline characteristicsOverall, 169 patients were diagnosed with ACLF based on the Asian-Pacific Association for the Study of the Liver criteria, and 21 patients were excluded for hepatocellular carcinoma or other malignant tumors (n = 12), serious extrahepatic disease (n = 5), myasthenia gravis (n = 1), and stage III of hepatic encephalopathy (n = 3). The remaining 148 patients underwent an evaluation of muscle function on admission. Of them, 12 patients did not have a CT scan in ≤2 wk of admission, and 10 (7.4%) patients received liver transplant in ≤90 d. Finally, a total of 126 patients with ACLF were included for analysis.
The baseline characteristics of the study population are listed in Table 1. The cohort, with a mean age of 46.1 ± 10.9 y, comprised 108 (85.7) men. The median BMI was 22.7 (4.9) kg/m2, and 44 (34.9) patients were overweight. The etiology of ACLF was hepatitis B virus in 68 (54), alcoholic liver disease in 37 (29.4), hepatitis B virus combined with alcoholic liver disease in 15 (11.9), and other reasons in 6 (4.8). In total, 99 (78.6), 44 (34.9), and 15 (11.9) patients had ascites, hepatic encephalopathy, and AKI, respectively. According to the European Association for the Study of the Liver–Chronic Liver Failure Consortium OF assessment system, 30 (23.8), 73 (57.9), and 23 (15.1) patients were classified as 0 OF, 1 OF, and ≥ 2 OFs groups, respectively. The median time between the CT imaging used for segmentation and admission was 2 d (IQR = 1–5). The mean L3-SMI was 44.6 ± 11.6 cm2/m2. Of the 126 ACLF patients, 33 (26.2) were diagnosed with low muscle mass. By the end of follow-up, 93 (73.8) patients survived without liver transplantation.
Establish the optimal cutoff value of LFI for defining low muscle functionThe median LFI was 3.89 (0.83). LFI had a weak positive correlation with age (ρ = 0.300; P = 0.001) and MELD-Na (ρ = 0.228; P = 0.010). However, a moderate negative correlation existed between LFI and BMI (ρ = –0.432; P < 0.001) as well as L3-SMI (ρ = –0.518; P < 0.001). Patients with ≥ 2 OFs had significantly higher LFI values than those without OFs (4.09 [1.26] versus 3.72 [0.54]; P = 0.019). Additionally, LFI in patients with non-survivors was significantly higher than that in survivors (4.48 [1.91] versus 3.75 [0.65]; P = 0.001) (Fig. 1).
Based on the time-dependent ROC curves and Youden index, the optimal cutoff value for LFI for predicting 90-d mortality was 4.0, with sensitivity and specificity values of 0.720 and 0.758, respectively (area under the curve = 0.760) (Fig. 2). Patients were divided into two groups based on their LFI scores: low muscle function and normal muscle function. As indicated in Table 1, compared with patients with normal muscle function, patients with low muscle function were more likely to be older (48.8 ± 10.9 versus 44.3 ± 10.5; P = 0.021) and have a higher MELD-Na score (24.3 [9.8] versus 22.2 [6.7]; P = 0.008). Moreover, patients with low muscle function had higher incidence of ascites (90.2% versus 70.7%; P = 0.009), hepatic encephalopathy (45.1% versus 28%; P = 0.048), AKI (19.6% versus 6.7%; P = 0.028), and low muscle mass (45.1% versus 13.3%; P < 0.001) than those with normal muscle function.
Prognostic value of low muscle function for 90-d mortality in patients with ACLFPatients with low muscle function exhibited a higher mortality at 90 d compared with those with normal muscle function (49% versus 10.7%; P < 0.001) (Fig. 3A). As indicated in Table 2, univariate Cox analysis found that ascites, hepatic encephalopathy, acute renal injury, MELD-Na score, number of OFs, low muscle function, and low muscle mass were risk factors for 90-d mortality. Variables with P < 0.01 in the univariate analysis and low muscle mass were included in the multivariate analysis. Multivariate Cox analysis identified number of OFs (hazard ratio = 4.202; 95% CI, 2.040–8.656; P < 0.001) and low muscle function (hazard ratio = 4.309; 95% CI, 1.795–10.345; P = 0.001) as independent risk factors for 90-d mortality.
We further determined whether the coexistence of muscle mass loss and impaired muscle function resulted in a worse 90-d prognosis compared with only impaired muscle function. The study population was divided into four groups: normal mass/normal function, low mass/normal function, normal mass/low function, and low mass/low function. Survival analysis found no significant difference in the 90-d survival probability between the normal mass/low function and low mass/low function groups (Fig. 3B).
Evaluate the risk factors associated with low muscle function in patients with ACLFUnivariate logistic analysis found that age, overweight, ascites, AKI, MELD-Na, number of OFs, and low muscle mass were associated with low muscle function. Variables with P < 0.05 in the univariate analysis were included in the multivariate analysis. It was found that advanced age (odds ratio [OR] = 1.042; 95% CI, 1.002–1.083; P = 0.038), increased numbers of OFs (OR = 2.572; 95% CI, 1.331–4.968; P = 0.005), and low muscle mass (OR = 6.607; 95% CI, 2.579–16.927; P < 0.001) were independent risk factors associated with low muscle function (Table 3).
DiscussionIn this prospective cohort study, we were the first to illustrate the influencing factors and prognostic role of low muscle function, as assessed by LFI, in patients with ACLF. The study had two novel findings. 1) LFI had a weak positive correlation with age and MELD-Na score and a moderate negative correlation with BMI and L3-SMI. As a categorical variable, advanced age, increased number of OFs, and low muscle mass were independent risk factors for low muscle function. 2) Low muscle function was identified as an independent risk factor for 90-d mortality, whereas low muscle mass was not. Our research enhanced the understanding of the relationship between low muscle function and low muscle mass, two aspects of malnutrition, and found that the prognostic values of impaired muscle function was superior than that of low muscle mass in patients with ACLF.
As is well known, cross-sectional imaging assessment by CT is the most accurate and objective tool for diagnosing low muscle mass [14]. However, there is no gold standard for diagnosing low muscle function. Several assessment tools have been established in the cohorts of community-dwelling adults, and further studies are needed to determine their applicability to patients with liver disease [15]. LFI is a novel liver-specific frailty index and Lai et al. [10] establish the cutoff at the 80th percentile (>4.5) of the derivation cohort to diagnose low muscle function in the previous study. In order to stratify the greatest risk of waitlist mortality in liver transplant candidates, Kardashian et al. [16] identified an optimal LFI cutoff of 4.4 for 3-mo, 4.2 for 6-mo, and 4.2 for 12-mo mortality using the area-under-the-curve approach . Due to different races and the application setting, in this study, we performed a time-dependent ROC curve with the highest Youden index in our population to establish the LFI cutoff value for diagnosing low muscle function and found the optimal value for predicting 90-d mortality to be 4.0, which should be further verified in the future studies.
As a continuous variable, we found a moderate negative correlation between LFI and BMI. As a dichotomous variable, the incidence of obesity in patients with low muscle function was significantly lower than those with normal muscle function. However, a multicenter prospective study of outpatients waiting for liver transplantation found no difference in the incidence rate of low muscle function in different obesity grades. In contrast, the mortality associated with low muscle function is higher in patients with class ≥ 2 obesity (BMI ≥35 kg/m2) than non-obese patients and those with class 1 obesity. In obese patients, higher adipose tissue may result in a relative decrease in skeletal muscle mass. However, to maintain daily physical activity, the absolute skeletal muscle mass may be comparable or even higher than that of non-obese individuals. Therefore, evaluation of muscle mass may be less reliable in patients with obese, and muscle function assessments may be more helpful [17].
LFI is positively correlated with MELD score, indicating that muscle function is influenced by disease severity. Patients with ACLF often experience multiple OFs, which can exacerbate the decline of body function. An increased number of OFs is associated with a 2.572-fold increase in the risk of low muscle function. Other factors, such as age, nutritional status, and non–liver-related complications (diabetes, cardiovascular and cerebrovascular diseases, etc.), can also contribute to the progression of low muscle function [18]. Both decreased muscle mass and impaired muscle function are associated with muscle abnormalities. We further assessed the relationship between low muscle mass and low muscle function in patients with ACLF. LFI exhibited a moderate negative correlation with L3-SMI, and low muscle mass increased the risk of low muscle function by 6.607 times. Similarly, in a previous study conducted by Fozouni et al. [19], low muscle mass was significantly associated with the risk of low muscle function (OR = 2.81; 95% CI, 1.19–6.67) in male cirrhosis patients after adjusting for MELD-Na.
Although a previous study recognized low muscle mass as an independent risk factor for adverse outcomes in patients with ACLF [6], our study found that low muscle function was a risk factor, whereas low muscle mass was not. Despite a strong correlation between low muscle mass and low muscle function, they represent distinct conditions. Simultaneous presence of low muscle function may account for the poor prognosis observed in patients with low muscle mass [20]. In 2019, the European Working Group on Sarcopenia in Older People defined sarcopenia as a syndrome characterized by low muscle strength, low muscle mass, and low muscle quality. When combined with poor physical performance, it is diagnosed as severe sarcopenia [21]. Therefore, we speculate that a decline in muscle function resulting from a decrease in muscle mass will have a negative effect on prognosis. As for the cumulative effects of low muscle mass and low muscle function, the coexistence of low muscle function and low muscle mass did not increase the 90-d mortality, compared with only having low muscle function. However, different findings have been reported in patients with liver cirrhosis. Guo et al. [22] revealed that low muscle mass and low muscle function, when present alone, had similar mortality, whereas their coexistence was associated with a higher risk of mortality than either condition in isolation in patients with decompensated cirrhosis.
The pathogenesis of low muscle mass and low muscle function is multifactorial, including malnutrition, insufficient physical activity, and liver disease–related factors (such as synthesis dysfunction, fluid retention, hyperammonemia, and liver disease etiology) as well as other systemic factors (systemic inflammation, metabolic disorders, visceral fat accumulation, insulin resistance, advanced age, endocrine diseases, etc.) [5]. At present, only a few randomized controlled trials have explored the effects of nutritional intervention [23,24] and exercise [25] on improving muscle mass and muscle function indicators in patients with end-stage liver disease. However, no study has investigated the effects of nutritional intervention and exercise on improving muscle mass and function in patients with ACLF. The primary goal of nutritional support treatment for ACLF patients is to achieve the target levels of energy and protein intake to reduce muscle consumption. Regarding exercise interventions, the specific type, intensity, timing, and safety of exercise remain unclear, and further research is needed to address this issue.
Our study presents several limitations. First, the absence of a universally accepted diagnostic criterion for low muscle function poses challenges. The LFI, integrating three performance-based indicators, was designed to gauge the muscle function of cirrhotic patients. In our research, we used the LFI and identified an optimized cutoff value of LFI ≥ 4 to categorize patients with low muscle function using a time-dependent ROC curve. Nonetheless, further research is imperative to confirm the generalizability and applicability of our findings. There is also a pressing need for an index tailored to assess low muscle function specifically for patients with ACLF. Lastly, given the small sample size of our study, it is crucial to undertake larger, multicenter studies in the future to substantiate our conclusions.
ConclusionsPatients with low muscle mass had higher mortality, and the prognostic value of low muscle function plays a more important role than that of low muscle mass. Advanced age, increased number of OFs, and low muscle mass are risk factors for low muscle function. These findings emphasize the importance of muscle function assessment and targeted therapeutic strategies in patients with ACLF.
Declaration of competing interestThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
| Variables | Total ( | Low muscle function ( | Normal muscle function ( | |
| Age, y* | 46.1 ± 10.9 | 48.8 ± 10.9 | 44.3 ± 10.5 | 0.021 |
| BMI, kg/m2† | 22.7 (4.9) | 21.1 (3.5) | 24.1 (5.0) | < 0.001 |
| Obesity, | 44 (34.9) | 10 (19.6) | 34 (45.3) | 0.003 |
| Sex, male, | 108 (85.7) | 40 (78.4) | 68 (90.7) | 0.054 |
| Liver cirrhosis, | 91 (72.2) | 38 (74.5) | 53 (70.7) | 0.636 |
| Etiology, | 0.374 | |||
| CHB | 68 (54.0) | 24 (47.1) | 44 (58.7) | |
| ALD | 37 (29.4) | 19 (37.3) | 18 (24.0) | |
| CHB + ALD | 15 (11.9) | 5 (9.8) | 10 (13.3) | |
| Others | 6 (4.8) | 3 (5.9) | 3 (4.0) | |
| Ascites, | 99 (78.6) | 46 (90.2) | 53 (70.7) | 0.009 |
| HE, | 44 (34.9) | 23 (45.1) | 21 (28.0) | 0.048 |
| AKI, | 15 (11.9) | 10 (19.6%) | 5 (6.7) | 0.028 |
| MELD-Na score† | 23.0 (6.9) | 24.3 (9.8) | 22.2 (6.7) | 0.008 |
| OFs, | 0.072 | |||
| 0 | 30 (23.8) | 7 (13.7) | 23 (30.7) | |
| 1 | 73 (57.9) | 31 (60.8) | 42 (56.0) | |
| ≥ 2 | 23 (15.1) | 13 (25.5) | 10 (13.3) | |
| Total bilirubin (μmol/L)† | 315.9 (314.3) | 425.1 (340.4) | 269.3 (254.9) | 0.006 |
| Albumin (g/L)* | 30.4 ± 5.1 | 29.3 ± 5.0 | 31.2 ± 5.0 | 0.036 |
| Sodium (mmol/L)† | 136.8 (6.1) | 134.8 (6.5) | 137.4 (4.5) | 0.004 |
| Creatinine (μmol/L)† | 63 [25] | 61 [32] | 64 (18) | 0.621 |
| INR† | 1.99 (0.75) | 1.99 (0.88) | 1.98 (0.58) | 0.362 |
| WBC count (× 109/L)† | 6.7 (4.6) | 7.7 (6.6) | 6.0 (3.0) | < 0.001 |
| Platelet count (× 109/L)† | 102.5 (83) | 108 (85) | 91 (82) | 0.509 |
| L3-SMI, cm2/m2* | 44.6 ± 11.6 | 39.8 ± 10.2 | 47.8 ± 11.4 | < 0.001 |
| Low muscle mass, | 33 (26.2) | 23 (45.1) | 10 (13.3) | < 0.001 |
| Liver frailty index† | 3.89 (0.83) | 4.51 (1.38) | 3.60 (0.38) | < 0.001 |
| 90-d mortality, | 33 (26.2) | 25 (49.0) | 8 (10.7) | < 0.001 |
| Variables | Univariable analysis | Multivariable analysis | ||
| HR (95% CI) | HR (95% CI) | |||
| Age | 1.019 (0.987–1.052) | 0.252 | ||
| Obesity | 0.474 (0.206–1.092) | 0.079 | ||
| Liver cirrhosis | 0.795 (0.378–1.671) | 0.545 | ||
| Ascites | 4.673 (1.118–19.533) | 0.035 | ||
| HE | 2.617 (1.318–5.196) | 0.006 | 1.269 (0.623–2.586) | 0.511 |
| AKI | 2.396 (1.039–5.529) | 0.040 | ||
| MELD-Na | 1.129 (1.079–1.181) | < 0.001 | 1.042 (0.998–1.088) | 0.059 |
| OFs | 5.194 (2.846–9.477) | < 0.001 | 4.202 (2.040–8.656) | < 0.001 |
| Low muscle function | 6.012 (2.706–13.355) | < 0.001 | 4.309 (1.795–10.345) | 0.001 |
| Low muscle mass | 2.356 (1.181–4.702) | 0.015 | 1.580 (0.721–3.463) | 0.253 |
| Variables | Univariable analysis | Multivariable analysis | ||
| OR (95% CI) | OR (95% CI) | |||
| Age | 1.040 (1.006–1.076) | 0.023 | 1.042 (1.002–1.083) | 0.038 |
| Obesity | 0.294 (0.129–0.673) | 0.004 | ||
| Liver cirrhosis | 1.213 (0.544–2.707) | 0.637 | ||
| Ascites | 3.819 (1.339–10.894) | 0.012 | ||
| HE | 2.112 (1.001–4.459) | 0.050 | ||
| AKI | 3.415 (1.091–10.683) | 0.035 | ||
| MELD-Na | 1.060 (1.011–1.110) | 0.015 | ||
| OFs | 2.065 (1.150–3.707) | 0.015 | 2.572 (1.331–4.968) | 0.005 |
| Low muscle mass | 5.339 (2.249–12.675) | < 0.001 | 6.607 (2.579–16.927) | < 0.001 |
©2023. Elsevier Inc.