Simple method for predicting muscle volume loss using geriatric nutritional risk index in hepatocellular carcinoma patients

The liver is a central organ that controls metabolic nutrition, whereas tumour burden and hepatic function are well-known major prognostic factors in hepatocellular carcinoma (HCC) patients.^{1, 2} Nutritional status generally becomes worse with progression of hepatic function decline and conditions such as protein-energy malnutrition (PEM) often complicated in liver cirrhosis (LC) patients.³ As a result of such a worsened status, muscle volume loss (MVL) often develops in chronic liver disease (CLD) patients.⁴ MVL has been recognized as an important prognostic factor in HCC patients treated either curatively or palliatively.⁵ However, special technologies, such as computer software for use with computed tomography (CT) or devices for bioelectrical impedance analysis (BIA), are generally needed for assessment of muscle volume; thus, many institutions have difficulties accessing such methods because of their expense. Previously, a nutritional assessment index termed geriatric nutritional risk index (GNRI),⁶ which is calculated with use of only serum albumin level, height and body weight, was developed.

The present study aimed to elucidate the clinical usefulness of GNRI as an easy nutritional assessment method using well-known clinical factors to predict a high risk of MVL in CLD patients with HCC.

Four hundred forty two HCC patients, who underwent CT examinations performed at our hospital from January 2017 to June 2022 and within 1 month before starting treatment for HCC, were enrolled. None had a past history of HCC. Their records were kept in an institutional database and analysed in a retrospective manner.

The present results showed that the frequency of MVL, which has been defined as pre-sarcopenia,¹⁶ increased as nutritional status (GNRI) worsened (P < 0.001). Although the GNRI was originally created for assessing geriatric nutritional status, the present study was conducted under the consideration that it also reflects the effects of muscle loss. When the cut-off GNRI score for predicting MVL was analysed according to gender, those values were approximated (males 99.7, females 99.4). The GNRI uses different formulas for calculating standard weight for males and females, which may have contributed to those results. Thus, the cut-off GNRI score for MVL was 99.7 (approximately equal to the cut-off value for GNRI mild decline) in all patients, with the same score found in patients without ascites. For the GNRI normal status patients with MVL (28/283: 9.9%), that was thought to be mainly due to aging, because those with MVL were older (77 vs. 72 years, P = 0.006).

Recently, decreased muscle has been commonly reported as a complication in CLD patients.¹⁷ Hanai et al. noted a hazard ratio (HR) of mortality from sarcopenia of 3.03 (95% CI: 1.42 to 6.94)¹⁸ and, in another study, found that LC patients showed a muscle volume decline of −2.2%/year.¹⁹ It is important to assess sarcopenia, especially in cases of LC, because the HR for mortality of LC patients in accordance with muscle mass was found to be 0.78 (95% CI: 0.68 to 0.89, P < 0.001), implying that mortality decreases at a rate of 22% in cases with higher muscle mass.²⁰ Moreover, MVL has also been described as a prognostic factor for recurrence after curative treatments (HR 1.77, P < 0.001), as well as overall survival (OS) in HCC patients treated with either curative (HR 2.152, P < 0.001) or palliative (HR 2.358, P < 0.001) procedures.⁵

As noted above, an evaluation of MVL has clinical importance, though an important issue is that the assessment requires special expensive equipment, such as BIA or CT, and/or subjecting the patient to X-ray exposure. Previously, a finger-circle (yubi-wakka) test using the patient's own fingers was reported as an easy to perform tool for assessment of the early stage of MVL in CLD patients,²¹ though it is thought to be difficult for evaluation of relative changes in nutritional status. Therefore, the results presented here indicate that GNRI might be a predictive tool for MVL in CLD patients that is easy to use in clinical situations. When GNRI assessment of a CLD patient shows a decline that is mild or greater, the clinician should keep in mind the assessment of muscle volume along with routine nutritional intervention²² with a goal to maintain daily activities of the patient²³ to prevent progression of sarcopenia.

Immune checkpoint inhibitors (ICIs) have recently been developed and shown to have a great role in cancer treatment. Meta-analysis findings of patients treated with an ICI showed that MVL was related with poor objective response rate (ORR) (OR 0.46, 95% CI: 0.28 to 0.74, P = 0.001), disease control rate (DCR) (OR 0.44, 95% CI: 0.31 to 0.64, P < 0.0001), progression-free survival (PFS) (HR 1.46, 95% CI: 1.20 to 1.78, P = 0.0001) and OS (HR 1.73, 95% CI: 1.36 to 2.19, P < 0.0001).²⁴ Furthermore, also in patients who received atezolizumab plus bevacizumab treatment for unresectable HCC, MVL was found to be a prognostic factor related to PFS (HR 1.479, 95% CI: 1.020 to 2.144, P = 0.039) and OS (HR 2.119, 95% CI: 1.150 to 3.904, P = 0.016).²⁵ These results indicate that MVL is also an important prognostic factor in the current treatment of HCC. Therefore, it is important to evaluate MVL in HCC patients. In our results, the cut-off value of GNRI for MVL was 99.7, which was approximate for the cut-off for GNRI mild decline, and a cut-off ALBI score for predicting GNRI mild decline was −2.478 (AUC 0.892, 95% CI: 0.863 to 0.921). That cut-off value for GNRI mild decline status is near the middle of mALBI grade 2a, whereas the cut-off value for GNRI for predicting mALBI grade 2b was 96.7 (AUC 0.867, 95% CI: 0.831 to 0.903), approximating that for the upper range of mild decline status (GNRI 98). Together, these results suggest that hepatic reserve function and nutritional status are closely related in CLD patients with HCC. Thus, it is suggested that nutritional status begins to deteriorate when ALBI grade 1 changes to 2a, whereas GNRI mild decline status likely has become established by the time the patient reaches mALBI grade 2b.

Based on our results, GNRI is considered to be a useful predictor for MVL in CLD patients. However, this study has some limitations. First, this was a single-centre study conducted in a retrospective manner. Second, all the subjects were HCC patients. Third, there were no data related to muscle strength such as handgrip strength available for the present cohort. Finally, the relationship between relative changes in GNRI score and muscle volume in each patient was not assessed. To obtain concrete conclusions, a multicentre study is needed with a larger number of CLD patients without HCC.

In conclusion, the present findings show GNRI to be an easy and possibly effective prediction tool for MVL in CLD patients. To maintain a normal GNRI, nutritional intervention is thought to be important and muscle volume should be assessed when an abnormal GNRI value is demonstrated.