Accurate SARC-F Score in Patients with Liver Disease

SARC-F is a well-accepted screening tool for sarcopenia. A SARC-F value of 1 point is reported to be more discriminating in identifying sarcopenia than 4 points (recommended cutoff point). The prognostic impact of the SARC-F score was investigated in patients with liver disease (LD, n = 269, median age = 71 years, 96 hepatocellular carcinoma (HCC) cases). Factors associated with SARC-F ≥ 4 points and SARC-F ≥ 1 point were also examined. In the multivariate analysis, age (p = 0.048), and Geriatric Nutritional Risk Index (GNRI) score (p = 0.0365) were significant factors linked to SARC-F ≥ 1 point. In our patients with LD, the SARC-F score is well correlated with the GNRI score. The 1-year cumulative overall survival ratio in patients with SARC-F ≥ 1 (n = 159) and SARC-F 0 (n = 110) was 78.3% and 90.1% (p = 0.0181). After excluding 96 HCC cases, similar tendencies were found (p = 0.0289). In the receiver operating curve (ROC) analysis based on the prognosis for the SARC-F score, the area under the ROC was 0.60. The sensitivity was 0.57, the specificity was 0.62, and the optimal cutoff point of the SARC-F score was 1. In conclusion, sarcopenia in LDs can be affected by nutritional conditions. A SARC-F score of ≥1 is more useful than a score of 4 in predicting the prognosis of patients with LD.


Introduction
Sarcopenia is a clinical condition characterized by loss of muscle mass and muscle function and is associated with frailty, cachexia, severe infections, and eventually death [1,2]. Age-related loss of muscle mass and strength is defined as primary sarcopenia, while disease-related loss of muscle mass and strength due to inadequate intake of protein, energy, etc. are classified as secondary sarcopenia [3,4]. Nutrient metabolism is often altered in patients with gastrointestinal diseases, and nutrient requirements and dietary intake differ from those of healthy individuals [5][6][7][8][9]. In other words, gastrointestinal diseases are considered representative of secondary sarcopenia [10,11]. Worldwide, the prevalence of sarcopenia in patients with liver disease (LD) is 5-13% in those in their 60 s and 11-50% in those over 80 years of age, making it a frequent condition [12][13][14]. Sarcopenia in patients with LD can be associated with worse patient quality of life (QOL), a poorer prognosis, and higher healthcare costs [5,[15][16][17]. LDs are representative diseases of secondary sarcopenia caused by the disease burden itself due to protein-energy malnutrition or other metabolic disorders [10]. Sarcopenia in LDs has been on the rise in recent years and can be a major concern in clinical practice.
The SARC-F is a five-question questionnaire and a globally recognized screening tool for sarcopenia [18,19]. Patients are asked to answer on a scale of 0 to 2 from "not at all We are one of the leading high-volume centers for LD in the country. Data on liver patients are continuously accumulated in our database. All hospitalized patients were required to complete the SARC-F questionnaire as a rule upon admission to the hospital, except for those who were unable to complete the questionnaire for reasons such as dementia. Between May 2020 and October 2022, 269 LD patients with SARC-F scores who were admitted to Osaka Medical and Pharmaceutical University can be found in our database. Baseline variables included were: age (years), gender, body mass index (BMI, kg/m 2 ), hemoglobin (g/dL), C reactive protein (CRP, mg/dL), platelet count (×10 4 /mm 3 ), serum albumin level (g/dL), serum bilirubin level (mg/dL), total lymphocyte count (/µL), and estimated glomerular filtration rate (eGFR, mL/min/1.73 m 2 ). All personal information was kept anonymous and carefully protected.

SARC-F Score and Our Analysis
As mentioned earlier, SARC-F in each patient was calculated. SARC-F score and the percentage of patients with SARC-F ≥ 4 points or SARC-F ≥ 1 point were compared according to ALBI grade [24]. Next, the SARC-F score and the percentage of patients with SARC-F ≥ 4 points or SARC-F ≥ 1 point were compared according to the Geriatric Nutritional Risk Index (GNRI) grade (No risk, Mild risk, Moderate risk, and Severe risk [25]). The GNRI is a nutritional risk index that allows classifying patients according to the risk of morbidity and mortality in relation to pathologies in elderly patients that are often associated with malnutrition, as proposed by Bouillanne and colleagues [25]. It is a modified version (for the elderly) of the Nutritional Risk Index, a measure traditionally used to predict the risk of postoperative complications related to nutritional disorders [26,27].
The GNRI evaluates nutritional status using the ideal weight, actual weight, and serum albumin level [25][26][27]. GNRI is calculated by the following formula: GNRI = 1.487 × serum albumin (g/L) + 41.7 × body weight/ideal body weight (kg) [26,27]. GNRI < 82 is defined as severe nutritional risk, 82 < GNRI < 92 as moderate risk, 92 < GNRI < 98 as low risk, and GNRI > 98 as no risk [27,28], and is considered a useful prognostic factor in the elderly, patients with heart failure, dialysis, and malignancies [25,29,30]. Next, factors associated with SARC-F ≥ 4 points or SARC-F ≥ 1 point were examined using univariate and multivariate analyses. Next, a receiver operating characteristic (ROC) analysis of independent factors in the multivariate analysis for SARC-F ≥ 4 points or SARC-F score ≥ 1 point was performed. Next, the overall survival (OS) ratio was compared according to the SARC-F ≥ 4 points or the SARC-F ≥ 1 point. Finally, a ROC curve analysis for the SARC-F based on the prognosis was performed. Ethics approval was obtained from the Ethics Committee of Osaka Medical and Pharmaceutical University Hospital (approval number: 2021-109, approval date: 25 November 2021). The research is shown in detail on our website. Due to the retrospective nature of this study, informed consent from the patient was waived by the ethics committee of our hospital. Alternatively, an opt-out method was adopted for the current study. The protocol of this study strictly adhered to all provisions of the 1975 Declaration of Helsinki.

Statistical Considerations
For continuous parameters, the Student's t-test or Mann-Whitney U test was applied as appropriate after confirming normality for two-group comparisons, and the ANOVA or Kruskal-Wallis test was applied as appropriate after confirming normality for multi-group comparisons. For categorical parameters, the Pearson χ 2 test was applied to estimate group differences. For comparison of survival, the Kaplan-Meier method was applied and tested using the log-rank method. Data for continuous parameters were presented as median values (interquartile range, IQR). Multivariate logistic regression analysis linked to SARC-F ≥ 4 points or SARC-F ≥ 1 point was also performed to identify independent factors. The significance level was 0.05 by using JMP ver. 15 (SAS Institute Inc., Cary, NC, USA).

Patient Baseline Characteristics
Baseline characteristics for all cases (n = 269, 161 males and 108 females, median (IQR) age = 71 (64-78) years) are presented in Table 1 3 . The median (IQR) serum GNRI score was 98.9 (90.9-106.1). No risk of GNRI grade was found in 143 patients; a mild risk of GNRI grade was found in 52; a moderate risk of GNRI grade was found in 55; and a severe risk of GNRI grade was found in 19. The median (IQR) serum ALBI score was −2.38 (−2.03-−2.47). ALBI grade 1 was found in 104 patients, ALBI grade 2 in 152 patients, and ALBI grade 3 in 13 patients. The number of cases according to the SARC-F score is shown in Figure 1. SARC-F 0 points were observed in 159 cases (59.0%), and SARC-F ≥ 1 point in 110 cases (41.0%). SARC-F of < 4 points were observed in 234 cases (87.0%), and SARC-F of ≥ 4 points were observed in 35 cases (13.0%).

Percentage of Patients with SARC-F ≥ 4 Points or SARC-F ≥ 1 Point According to ALBI Grade
The percentage of SARC-F ≥ 4 points in groups of ALBI grade 1 and ALBI grade 2 or

Univariate and Multivariate Analyses of Factors Associated with SARC-F ≥ 4 Points or SARC-F ≥ 1 Point
In the univariate analysis, age (p = 0.0002), eGFR (p = 0.0393), and GNRI score (p = 0.0010) were significant factors related to SARC-F ≥ 4 points ( Table 2). The GNRI score (p = 0.0236) was the only independent factor related to SARC-F ≥ 4 points in the multivariate logistic regression analysis (Table 3). Hazard ratios (HRs) and 95% confidence intervals (CIs) for each factor are shown in Table 3. Likewise, in the univariate analysis, age (p < 0.0001), eGFR (p = 0.0004), GNRI score (p = 0.0035), and ALBI score were significant factors related to SARC-F ≥ 1 point (Table 2). Age (p = 0.048) and GNRI score (p = 0.0365) were independent factors related to SARC-F ≥ 1 point in the multivariate logistic regression analysis (Table 3). HRs and 95% CIs for each factor are shown in Table 3.

Univariate and Multivariate Analyses of Factors Associated with SARC-F ≥ 4 Points or SARC-F ≥ 1 Point
In the univariate analysis, age (p = 0.0002), eGFR (p = 0.0393), and GNRI score (p = 0.0010) were significant factors related to SARC-F ≥ 4 points ( Table 2). The GNRI score (p = 0.0236) was the only independent factor related to SARC-F ≥ 4 points in the multivariate logistic regression analysis (Table 3). Hazard ratios (HRs) and 95% confidence intervals (CIs) for each factor are shown in Table 3. Likewise, in the univariate analysis, age (p < 0.0001), eGFR (p = 0.0004), GNRI score (p = 0.0035), and ALBI score were significant factors related to SARC-F ≥ 1 point (Table 2). Age (p = 0.048) and GNRI score (p = 0.0365) were independent factors related to SARC-F ≥ 1 point in the multivariate logistic regression analysis (Table 3). HRs and 95% CIs for each factor are shown in Table 3.

ROC Analysis of Independent Parameters for the SARC-F Score ≥ 4 Points or SARC-F Score ≥ 1 Point
The ROC analysis of GNRI in the multivariate analysis for SARC-F score ≥ 4 points was performed. The area under the ROC (AUC), the sensitivity, the specificity, and the optimal reference value of the GNRI score are demonstrated in Table 4. The AUC of the GNRI score for the SARC-F score ≥ 4 points was 0.67. Likewise, ROC analysis of independent factors in the multivariate analysis for SARC-F scores ≥ 1 point was conducted. The corresponding AUC, sensitivity, specificity, and best reference point for each factor are demonstrated in Table 4. Age involved the highest AUC for the SARC-F score ≥ 1 point (AUC = 0.66), followed by the GNRI score (AUC = 0.60).

ROC Analysis Based on the Prognosis for the SARC-F Score
In the ROC analysis based on the prognosis for the SARC-F score, the AUC was 0.60 ( Figure 8). Corresponding sensitivity and specificity were 0.57 and 0.62, and the optimal cutoff point of the SARC-F score was 1 (Figure 8). In HCC cases, the 1-year cumulative OS ratio was 74.8%. The 1-year cumulative OS ratio in patients with SARC-F ≥ 4 (n = 19) and SARC-F < 4 (n = 77) was 77.8% and 74.0%, respectively (p = 0.9671, Figure 7C). The 1-year cumulative OS ratio in patients with SARC-F ≥ 1 (n = 39) and SARC-F 0 (n = 57) was 68.4% and 79.3%, respectively (p = 0.1869, Figure 7D).

ROC Analysis Based on the Prognosis for the SARC-F Score
In the ROC analysis based on the prognosis for the SARC-F score, the AUC was 0.60 ( Figure 8). Corresponding sensitivity and specificity were 0.57 and 0.62, and the optimal cutoff point of the SARC-F score was 1 (Figure 8).

Discussion
More than a quarter century has passed since the disease concept of sarcopenia was first proposed, and subsequent research has led to remarkable academic developments in this field [31]. In the field of public health, sarcopenia has received a lot of attention recently because of its close association with clinical outcomes (i.e., falls, fractures, infections, frailty, survival) [1,2,15,17,[32][33][34][35]. In order to diagnose sarcopenia, skeletal muscle mass must be assessed, but imaging tests such as computed tomography and bioelectrical impedance analysis for skeletal muscle mass assessment are often unavailable in smaller clinics. In that sense, SARC-F is convenient and useful, and its use as an initial screening is recommended in the current international guidelines [3,[35][36][37]. Yang et al. reported a very good AUC of 0.86 and 0.90 for men and women, respectively, using the recommended reference point of SARC-F (i.e., 4 points) for the screening of sarcopenia [36]. Thus, SARC-F seems to be very helpful for sarcopenia screening, although the sensitivity of SARC-F for sarcopenia is unsatisfactory low [3,19,36,37]. Additionally, SARC-F has recently been reported to be well correlated with all-cause mortality and all-cause-specific mortality [38].
In recent years, a SARC-F score of 1 point in patients with LD has been shown to have a higher ability to identify sarcopenia than the conventionally used score of 4 points [23]. In our data, the percentage of ALBI grade 1 (104 cases) and ALBI grade 2 or 3 (165 cases) with SARC-F ≥ 4 points was 8.7% for ALBI grade 1 and 15.8% for ALBI grade 2 or 3, which was not significantly different (p = 0.09). While the percentage of patients with SARC-F ≥ 1 point in ALBI grade 1 and ALBI grade 2 or 3 was 31.7% and 46.7%, respectively (p = 0.015), in patients with LD, a SARC-F cutoff of 1 point may clarify the relationship between liver function and sarcopenia. In our results, the SARC-F score and the percentage of patients with a SARC-F score ≥ 4 or a SARC-F score ≥ 1 were well stratified by the nutritional marker as assessed by the GNRI score in all analyses. These findings suggest that the SARC-F score may reflect the nutritional status of LD patients and may provide a clue to

Discussion
More than a quarter century has passed since the disease concept of sarcopenia was first proposed, and subsequent research has led to remarkable academic developments in this field [31]. In the field of public health, sarcopenia has received a lot of attention recently because of its close association with clinical outcomes (i.e., falls, fractures, infections, frailty, survival) [1,2,15,17,[32][33][34][35]. In order to diagnose sarcopenia, skeletal muscle mass must be assessed, but imaging tests such as computed tomography and bioelectrical impedance analysis for skeletal muscle mass assessment are often unavailable in smaller clinics. In that sense, SARC-F is convenient and useful, and its use as an initial screening is recommended in the current international guidelines [3,[35][36][37]. Yang et al. reported a very good AUC of 0.86 and 0.90 for men and women, respectively, using the recommended reference point of SARC-F (i.e., 4 points) for the screening of sarcopenia [36]. Thus, SARC-F seems to be very helpful for sarcopenia screening, although the sensitivity of SARC-F for sarcopenia is unsatisfactory low [3,19,36,37]. Additionally, SARC-F has recently been reported to be well correlated with all-cause mortality and all-cause-specific mortality [38].
In recent years, a SARC-F score of 1 point in patients with LD has been shown to have a higher ability to identify sarcopenia than the conventionally used score of 4 points [23]. In our data, the percentage of ALBI grade 1 (104 cases) and ALBI grade 2 or 3 (165 cases) with SARC-F ≥ 4 points was 8.7% for ALBI grade 1 and 15.8% for ALBI grade 2 or 3, which was not significantly different (p = 0.09). While the percentage of patients with SARC-F ≥ 1 point in ALBI grade 1 and ALBI grade 2 or 3 was 31.7% and 46.7%, respectively (p = 0.015), in patients with LD, a SARC-F cutoff of 1 point may clarify the relationship between liver function and sarcopenia. In our results, the SARC-F score and the percentage of patients with a SARC-F score ≥ 4 or a SARC-F score ≥ 1 were well stratified by the nutritional marker as assessed by the GNRI score in all analyses. These findings suggest that the SARC-F score may reflect the nutritional status of LD patients and may provide a clue to the relationship between malnutrition and sarcopenia in patients with LD. GNRI score (p = 0.0236) was an independent factor related to SARC-F ≥ 4 points in the multivariate logistic regression analysis. Likewise, age (p = 0.048) and GNRI score (p = 0.0365) were independent factors related to SARC-F ≥ 1 point in the multivariate logistic regression analysis. Our results imply that the SARC-F ≥ 1 in LDs is closely associated with not only the primary factor (aging) but also a secondary factor such as the nutritional factor. We have previously reported that patients with advanced gastrointestinal cancers with SARC-F < 4 points have a better prognosis than patients with SARC-F ≥ 4 points [6]. While we found no difference in prognosis in patients with advanced HCC with SARC-F ≥ 4 vs. <4, a SARC-F cutoff score of 4 may not be a sufficiently reliable predictor of prognosis in patients with LD [6]. In terms of survival, the 1-year cumulative OS ratio for all patients in this study was 85.2%. The 1-year cumulative OS ratios for patients with SARC-F ≥ 4 and SARC-F < 4 was 82.0% and 85.7%, respectively, which were not significantly different (p = 0.4348). The 1-year cumulative OS ratio for patients with SARC-F ≥ 1 and SARC-F 0 were 78.3% and 90.1%, respectively, with statistical significance (p = 0.0181). After excluding HCC cases, similar trends were found. In the ROC analysis based on the prognosis for the SARC-F score, the AUC was 0.60, the corresponding sensitivity and specificity were 0.57 and 0.62, and the optimal reference point for the SARC-F score was 1. When the cutoff point of the SARC-F score was set at 4, the corresponding sensitivity and specificity were 0.17 and 0.88, denoting a significant loss of sensitivity. As mentioned above, a cutoff of 4 points on the SARC-F for sarcopenia in patients with LD has relatively low sensitivity, and a cutoff of 1 point may be useful. The recommended SARC-F ≥ 4 points may only identify patients with more advanced sarcopenia, while SARC-F ≥ 1 point may be ideal for the early detection of sarcopenia, especially in patients with LD. Considering these study results, the optimal reference point for the SARC-F score should be carefully considered. It is obvious that lowering the cutoff to 1 from 4 would reduce the number of missed cases of sarcopenia. Ideally, the cutoff point should be at a point where sensitivity and specificity are optimally balanced, but considering SARC-F's position as a screening tool, better sensitivity may be better [39]. Alternatively, a modified version of SARC-F could be used to increase detection sensitivity for sarcopenia [40].
It must be stated that this study has several limitations. First, this is a single-center, cross-sectional Japanese study that, by its nature, is retrospective. We did not examine lifestyle habits in other ethnic groups other than the Japanese. In addition, the SARC-F is a self-report questionnaire that indicates the likelihood of risk of sarcopenia, and data on the number of patients with definite sarcopenia were not included in the present analysis. Finally, in terms of survival analysis, various treatments for background LDs have been given during the course of the disease, which may affect the prognosis, also creating biases. Despite these limitations, the results of this study indicate that a baseline SARC-F score of 1 correlates well with the GNRI score in patients with LD and may be a predictor of prognosis.

Conclusions
Sarcopenia in patients with LD can be affected by nutritional condition. A SARC-F ≥ 1 score is useful in predicting the prognosis of patients with LD. Clinicians should be aware of these. Informed Consent Statement: Patient consent was waived due to the retrospective nature of this study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the personal information.