Retrospective study links low handgrip strength to survival in unresectable HCC

In a retrospective study of 154 patients with unresectable hepatocellular carcinoma receiving hepatic arterial infusion chemotherapy-based combination therapy, low baseline handgrip strength was associated with shorter progression-free and overall survival. The association was observed whether patients also had a low skeletal muscle index on CT or retained a muscle mass considered normal. Patients with both low strength and low skeletal muscle index had adjusted hazard ratios of 2.53 and 3.72, respectively.

The findings show an association in this treated cohort, not evidence that weakness caused the poorer outcomes or that handgrip strength alone can determine an individual patient’s prognosis. The analysis was a secondary evaluation of a propensity score-matched subset from a previously reported cohort at Tongji Hospital in Wuhan, China. Patients received treatment between August 2020 and December 2022.

The regimen combined modified oxaliplatin, leucovorin and fluorouracil-based hepatic arterial infusion chemotherapy, known as mFOLFOX-based HAIC, with a targeted agent and an immune checkpoint inhibitor. All 154 participants had pathologically confirmed, surgically unresectable disease and measurable tumors; surgical unresectability meant that curative-intent resection was considered unsafe after multidisciplinary review. Of 265 eligible patients in the source cohort, 1:1 propensity score matching produced 77 patients with low and 77 with preserved handgrip strength.

The present study reused that matched subset without additional matching, and follow-up continued until January 24, 2024. Median follow-up, estimated by the reverse Kaplan–Meier method, was 22.1 months. Handgrip strength was measured before treatment with a digital dynamometer.

Participants sat with the shoulder adducted and neutrally rotated, the elbow flexed at 90 degrees, and the forearm and wrist neutral. Three maximum-effort trials were performed with the dominant hand, separated by one-minute rest intervals, and the highest result was recorded. Low strength was defined before the survival analyses as less than 28 kg for men and less than 18 kg for women, using thresholds from Asian and liver-disease consensus frameworks.

Two investigators reviewed pretreatment CT images acquired within two weeks to calculate skeletal muscle index at the third lumbar vertebral level, with muscle area divided by height squared. Low skeletal muscle index was defined as less than 42 cm²/m² for men and less than 38 cm²/m² for women. The resulting four groups comprised 68 patients with normal strength and normal index, 10 with low index but normal strength, 46 with low strength but normal index, and 30 with both measures low.

Kaplan–Meier curves differed across the four phenotypes for both progression-free survival, defined as the interval to radiological progression or death from any cause, and overall survival; both log-rank P values were below 0.001. Estimated 12-month progression-free survival was 62.2% for patients with normal strength and normal index, 90.0% for the 10-patient low-index group, 23.9% for the low-strength group and 39.0% for the group with both measures low. Corresponding 12-month overall survival estimates were 82.2%, 90.0%, 37.0% and 39.7%.

The multivariable models accounted for age, sex, Eastern Cooperative Oncology Group performance status, Child–Pugh class, maximum tumor diameter, tumor number, portal vein tumor thrombosis, extrahepatic spread and previous HCC treatment. In these models, low handgrip strength and the joint low-strength, low-mass phenotype remained associated with poorer progression-free and overall survival. The strength–mass association was not uniform over time.

Tests of proportional hazards indicated that the overall-survival effect of muscle phenotype varied during follow-up, so the primary Cox hazard ratios represent average associations across the observed period rather than constant effects at every time point. An exploratory time-varying model estimated larger early differences: compared with normal strength and normal index, the adjusted hazard ratio for death was 4.82 at six months for the low-strength group and 5.58 for the both-low group. At 24 months, those estimates had attenuated to 1.24 and 1.03, with wide confidence intervals that included no association.

Restricted mean survival time estimates, which do not require proportional hazards, showed that the low-strength group spent 7.92 fewer months alive over 24 months than the normal group, while the both-low group spent 6.50 fewer months alive. The small, isolated low-index group spent 0.36 months longer alive, but its confidence interval was wide and included shorter survival. These estimates characterize a modest retrospective sample and do not establish a precise timeline for individual prognosis.

Handgrip strength added prognostic information beyond conventional clinical variables in the model comparisons. The clinical model included age, sex, performance status, Child–Pugh class, maximum tumor diameter, tumor number, portal vein tumor thrombosis, extrahepatic spread and prior HCC treatment. Its apparent concordance index was 0.658.

Adding low handgrip strength raised it to 0.731, and bootstrap optimism correction reduced the respective values to 0.625 and 0.705. Adding low skeletal muscle index alone produced only a small change in apparent discrimination, from 0.658 to 0.669, and did not significantly improve model fit. The more elaborate four-phenotype classification yielded an apparent C-index of 0.733 and an optimism-corrected value of 0.701, but it did not improve fit beyond the simpler model containing low handgrip strength; the direct comparison had a likelihood-ratio P value of 0.798.

Thus, within this analysis, most of the incremental discrimination supplied by the joint classification was captured by the simpler low-strength measure. Continuous-variable analyses supported the main finding while highlighting uncertainty in the small categorical low-index subgroup. A one-standard-deviation decrease in handgrip strength was associated with shorter progression-free survival, with a hazard ratio of 2.00, and shorter overall survival, with a hazard ratio of 2.48.

A one-standard-deviation decrease in skeletal muscle index was not significantly associated with progression-free survival, although it was associated with shorter overall survival, with a hazard ratio of 1.29. Handgrip strength and skeletal muscle index were correlated, with a Pearson correlation coefficient of 0.710. When both standardized measures were entered together, lower handgrip strength remained associated with both outcomes.

The skeletal muscle index coefficient changed direction in these mutually adjusted models, but the investigators noted that correlation, statistical suppression and selection from the previously matched cohort could explain the result. It therefore should not be interpreted as evidence that lower muscle mass is protective or that muscle quantity lacks prognostic relevance. The objective response rate was 51.5% in the normal group, 30.0% among patients with low skeletal muscle index but normal strength, 39.1% in the low-strength group and 50.0% in the both-low group, with P = 0.430.

Disease control rates were 92.6%, 90.0%, 95.7% and 96.7%, respectively, with P = 0.623. The investigators proposed that strength may reflect nutritional status, inflammation, neuromuscular factors, inactivity and overall physiological reserve, while also noting that unmeasured differences in treatment tolerance, complications, dose delivery and subsequent therapy could have contributed to the associations. The study did not directly assess muscle quality, global strength, gait speed, chair-rise performance or changes in muscle status over time, so it cannot determine why low handgrip strength marked poorer survival in this cohort.

Several limitations constrain how far the results can be generalized. The study was retrospective, single-center and relatively small; the isolated low-index, normal-strength group contained only 10 patients, and sex was unevenly distributed across phenotypes. Restriction to a subset previously matched according to handgrip status may also have altered the phenotype distribution.

Handgrip strength was measured only in the dominant hand, which was not necessarily the stronger hand, and formal familiarization and standardized encouragement were not documented. Potential influences such as pain, neuropathy, arthritis, edema, cognition and motivation were not systematically assessed. Treatments were heterogeneous, and dose reductions, delays and reasons for discontinuation were not comprehensively recorded.

The handgrip thresholds came from established Asian and liver-disease criteria but have not been independently validated in patients receiving HAIC-based combination therapy, so values near the cutoffs require caution. The overall-survival effect also varied over time, and later estimates were less precise because fewer patients remained at risk. Nevertheless, the associations were consistent in continuous analyses, restricted mean survival analyses, a male-only subgroup and the 127 patients receiving lenvatinib plus an immune checkpoint inhibitor.

External validation in larger prospective multicenter cohorts is needed before these findings can be applied confidently in clinical decision-making.

Subject of Research: Baseline handgrip strength and survival in unresectable hepatocellular carcinoma treated with HAIC-based combination therapy

Article Title: Muscle strength and muscle mass phenotypes and survival in unresectable hepatocellular carcinoma receiving hepatic arterial infusion chemotherapy-based combination therapy

Article References: Su, C., Ji, G., Wang, X., Guo, Q., Lv, X., Jin, X., & Xiao, D. (2026). Muscle strength and muscle mass phenotypes and survival in unresectable hepatocellular carcinoma receiving hepatic arterial infusion chemotherapy-based combination therapy. Frontiers in Oncology, 16, 1974237. https://doi.org/10.3389/fonc.2026.1974237

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1974237

Keywords: hepatocellular carcinoma, handgrip strength, skeletal muscle index, overall survival, progression-free survival, hepatic arterial infusion chemotherapy, prognostic stratification, body composition, retrospective cohort, Muscle, strength, mass