Physical Activity Improves Overall Survival in Patients with Hepatocellular Carcinoma

Abstract

Introduction

Physical activity improves survival in various cancers, but its impact in patients with hepatocellular carcinoma (HCC) remains unexplored. We aimed to investigate the association between initiating moderate-to-vigorous physical activity (MVPA) after HCC diagnosis and overall survival among previously inactive patients.

Methods

This retrospective cohort study used Korean National Health Insurance Service data from 21,035 adults diagnosed with HCC between 2010 and 2023 who were physically inactive before diagnosis and completed biennial health examinations pre- and post-diagnosis. MVPA initiation was assessed through standardized questionnaires. Primary outcome was all-cause mortality; secondary outcomes were cancer-specific and non-cancer-specific mortality. Cox models estimated adjusted hazard ratios and 95% confidence intervals.

Results

Among 21,035 patients (mean age 62.1 years; 78.1% male), 5,639 (26.8%) initiated MVPA and 15,396 (73.2%) remained inactive. During follow-up, 7,122 deaths occurred. The 12-year cumulative mortality was lower with MVPA initiation versus remaining inactive (53.9% vs. 61.2%; adjusted HR, 0.86; 95% CI, 0.82–0.92). Consistent associations were observed in patients receiving curative treatment (HR, 0.86; 95% CI, 0.78–0.95), those without curative treatment (HR, 0.87; 95% CI, 0.81–0.93), patients with viral HCC (HR, 0.87; 95% CI, 0.81–0.93), nonviral HCC (HR, 0.86; 95% CI, 0.77–0.96), cirrhotic patients (HR, 0.86; 95% CI, 0.80–0.93), and non-cirrhotic patients (HR, 0.88; 95% CI, 0.81–0.96). Propensity score-matched analysis confirmed these findings (HR, 0.87; 95% CI, 0.82–0.92).

Conclusion

Among physically inactive adults with HCC, initiating MVPA after diagnosis was associated with lower mortality compared with remaining inactive, suggesting that promoting physical activity should be considered in comprehensive HCC care.

Introduction

Hepatocellular carcinoma (HCC) represents a major global health burden, ranking as the sixth most common cancer and the third leading cause of cancer-related mortality worldwide [1]. Despite advances in therapeutic interventions for HCC, the overall prognosis remains poor, with 5-year survival rates of approximately 20% [2]. Growing evidence across various cancer types suggests that physical activity is associated with improved survival and reduced cancer-specific mortality possibly through multiple mechanisms including enhanced immune function [3], reduced systemic inflammation [4], and improved metabolic profiles [5–7]. However, the relationship between physical activity and survival outcomes in HCC patients has been understudied. Although a recent prospective study by Hashida et al. [8] showed that cancer rehabilitation is associated with improved overall survival in HCC patients undergoing transcatheter arterial chemoembolization, large-scale population-based data examining post-diagnosis physical activity initiation in HCC remain scarce.

Unlike other common malignancies, HCC typically develops in the context of chronic liver disease, with many patients presenting with underlying cirrhosis at diagnosis [9]. This unique clinical context raises important questions regarding the feasibility and efficacy of physical activity interventions in this population [10]. Furthermore, most previous research on physical activity and cancer survival has focused on pre-diagnosis activity patterns, whereas the impact of initiating physical activity after cancer diagnosis remains poorly characterized [11]. Evidence from large meta-analyses demonstrates that initiating physical activity after cancer diagnosis is associated with substantial survival benefits across multiple malignancies, with post-diagnosis activity often showing stronger associations than pre-diagnosis activity [11]. However, despite this consistent evidence in other cancers, data on the impact of post-diagnosis physical activity on survival in patients with HCC remain extremely limited. Given that many patients with HCC may be physically inactive prior to diagnosis due to liver disease-related symptoms and functional limitations, understanding whether post-diagnosis physical activity initiation can improve survival has significant clinical implications [10].

We therefore conducted a large-scale, population-based cohort study using data from the Korean National Health Insurance Service (K-NHIS) database to investigate the association between post-diagnosis moderate-to-vigorous physical activity (MVPA) initiation and all-cause mortality in patients with HCC. We specifically examined patients who were physically inactive before diagnosis and evaluated whether initiating MVPA after diagnosis was associated with improved survival compared to remaining inactive. Additionally, we assessed whether this association was consistent across clinically relevant subgroups defined by receipt of curative treatment, etiology, and presence of cirrhosis.

Patients and Methods

Data Sources and Study Population

This retrospective cohort study used data from the Korean K-NHIS database, which includes nearly the entire Korean population. The NHIS database provides comprehensive information on demographics, healthcare utilization, diagnoses coded according to the International Classification of Diseases, 10th Revision (ICD-10), procedures, prescription records, and results from standardized biennial health screening examinations. In addition, the K-NHIS claims database includes data from national health screening examinations, which are part of a standardized health screening program provided to all insured persons every 2 years [12, 13].

HCC was defined as the concurrent presence of the cancer-specific insurance claim code V193 and the ICD-10 code C22.0. In Korea, once an individual is diagnosed with cancer, the patient is registered in the National Cancer Registry with a specific code that indicates the diagnosis and provides eligibility for special insurance benefits. Therefore, the use of this cancer-specific claim code in administrative data has been shown to be highly reliable. In this study, we included all adults diagnosed with HCC between January 1, 2010, and December 31, 2023. To evaluate changes in PA, we identified physically inactive adults with HCC, who were diagnosed between January 1, 2010, and December 31, 2023, and who had completed a health screening within the 2 years prior to diagnosis (exam 1, N = 59,210). Of these, we selected 21,148 individuals with subsequent health screening visit conducted within 2 years following HCC diagnosis (exam 2). Furthermore, to minimize potential reverse causality, we excluded participants who died within 6 months after HCC diagnosis (N = 113). The final analytic cohort included 21,035 patients (online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000550923). The overall study concept is depicted in online supplementary Figure 2.

This study was approved by the Institutional Review Board of Samsung Medical Center (IRB No. SMC 2025-05-018), and the requirement for informed consent was waived due to the use of de-identified administrative data.

Measurement

PA was collected using self-reported structured questionnaires employing a 7-day recall method during national health screening. The questionnaire was similar to the International Physical Activity Questionnaire-Short Form but was simplified by the National Health Examination Committee. The reliability and validity of Korean version were reported in previous studies [14]. The questionnaire consisted of the following three questions: (1) How many days in the past week did you engage in vigorous activities that made you breathe much harder than normal for at least 20 min/day (e.g., running, aerobics, fast biking, or climbing)? (2) How many days in the past week did you engage in moderate activities that made you breathe somewhat harder than normal for at least 30 min/day (e.g., brisk walking, doubles tennis, riding a bicycle at a normal speed, or mopping)? (3) How many days in the past week did you engage in light activities, such as walking for at least 30 min/day, adding up to a total of at least 30 min/day (e.g., walking to and from work or leisure and light household chores)? This questionnaire has been widely used in previous studies [15].

MVPA was defined as ≥20 min/day of vigorous PA on ≥3 days/week or ≥30 min/day of moderate PA on ≥5 days/week [16]. Inactive PA was defined as not meeting the MVPA criteria. According to the study inclusion criteria, all participants were physically inactive adults in exam 1. In exam 2, participants who reported engaging in MVPA were classified as MVPA initiators. As a sensitivity analysis, total physical activity volume was additionally quantified using metabolic equivalents of task (METs). Because the NHIS physical activity questionnaire does not collect exact activity duration per day, the minimum duration thresholds specified in the questionnaire were applied (30 min per day for light and moderate-intensity activity and 20 min per day for vigorous-intensity activity). MET values were assigned according to exercise intensity as follows: 2.9 METs for light-intensity activity, 4.0 METs for moderate-intensity activity, and 7.0 METs for vigorous-intensity activity. Total weekly physical activity volume was calculated as the sum of MET minutes per week derived from reported frequency and minimum duration for each activity category. Among participants who initiated MVPA, individuals were further categorized into <1,000 and ≥1,000 MET-minutes per week based on established guideline thresholds and prior literature [17].

The primary endpoint was all-cause mortality. Secondary endpoints included cancer-specific mortality and non-cancer-specific mortality. Information on vital status and cause of death was obtained from death certification records collected by Statistics Korea, which were linked to the K-NHIS database through unique personal identification numbers. Cause of death was determined based on the underlying cause recorded on death certificates using International Classification of Diseases, 10th Revision (ICD-10) codes. Cancer-specific mortality was defined as deaths with primary cause code C22. Non-cancer-specific mortality was defined as deaths with any primary cause code other than C22. Follow-up was censored on December 31, 2023, or at the time of liver transplantation occurring after initial curative treatment.

Covariables

Data on age, residential area, alcohol consumption, body mass index (BMI), blood pressure, fasting glucose, total cholesterol, high-density lipoprotein, and low-density lipoprotein cholesterol levels were collected during exam 2. Information on the residential area and household income was obtained from insurance eligibility records. Residential areas were classified as metropolitan (Seoul, Busan, Daegu, Daejeon, Gwangju, Incheon, and Ulsan) or nonmetropolitan. Income was categorized as Medical Aid, ≤30th percentile, 31st–70th percentile, or >70th percentile. Hepatitis B virus infection was defined by codes B16, B18.0, B18.1, and Z22.5 and hepatitis C virus infection by codes B17.1, B18.2, and B19.2. Liver cirrhosis was identified using the following ICD-10 codes: K70.2, K70.3, K74.3, K74.4, K74.5, and K74.6. Hepatic decompensation events were defined as composite of clinically significant decompensating conditions identified using inpatient claims and ICD-10 diagnosis codes: ascites (R18), hepatic encephalopathy (K72.90 and K72.91), hepatorenal syndrome (K76.7), variceal bleeding using esophageal varices (I85), or gastric varices (I86.4) in combination with bleeding-related admission codes. To ensure clinical relevance, only events recorded during hospitalization were included. The Charlson Comorbidity Index (CCI) was calculated using diagnosis codes from claims data from the year preceding the study index date [18]. Antiviral therapy for chronic hepatitis B was defined as the prescription of oral nucleos(t)ide analogs during exam 1 and exam 2, including tenofovir disoproxil fumarate (ATC code: J05AF07), tenofovir alafenamide (J05AF13), entecavir (J05AF10), adefovir dipivoxil (J05AF08), telbivudine (J05AF11), lamivudine (J05AF05), clevudine (J05AF09), and besifovir dipivoxil (J05AF15). In this study, curative treatment for HCC was defined as surgical resection (procedure code: Q7221-Q7225, Q7230), radiofrequency ablation (Q7280-Q7285, QZ841, Q0841, M6774, QZ844, and OZ753), or liver transplantation (procedure codes: Q8040-Q8049 and Q8050) within 6 months after diagnosis as initial therapy [19].

Statistical Analysis

Baseline characteristics are reported as mean ± standard deviation, median [interquartile range], or number (%), as appropriate. The standardized mean difference between groups was estimated to compare the distribution of variables. The primary endpoint was overall survival. Follow-up began at the date of the second health examination (exam 2) and continued until death or the administrative end date (December 31, 2023). Because post-treatment liver transplantation could substantially affect survival, patients were censored at the time of transplantation after the initial curative therapy. Cox proportional hazards regression models were used to estimate hazard ratios and 95% confidence intervals. Multivariate models were adjusted for age, sex, income, area, BMI, comorbidities, antiviral of hepatitis B virus, current drinker, smoking status in exam 2 and curative treatment for HCC. The proportional hazards assumption was assessed using log-minus-log plots and Schoenfeld residuals. Pre-specified subgroup analyses were performed to examine whether the association between MVPA initiation and overall survival was consistent across subgroups defined by curative treatment for HCC, hepatitis, cirrhosis, and presence or absence of hepatic decompensation, which were used as a proxy for advanced liver dysfunction.

As a sensitivity analysis, a 1:2 propensity score-matched cohort was constructed using the nearest-neighbor method without replacement. The propensity score was estimated based on age, sex, income level, residential area, BMI, comorbidities, antiviral therapy for hepatitis B, current drinking status, smoking status, and curative treatment for HCC. The balance between groups was assessed using standardized mean differences, with values <0.1 considered acceptable. As an additional sensitivity analysis, we performed an as-treated analysis to account for potential changes in physical activity after exam 2. This analysis was restricted to participants who underwent a third health examination after exam 2. Follow-up was censored at the date of the third examination if physical activity status at the third examination differed from that at exam 2, whereas participants with consistent physical activity status were followed until death or the administrative end date. Furthermore, we also classified participants according to changes in MVPA status between exam 2 and exam 3 into four groups: persistently inactive, persistently active, MVPA initiation, and MVPA discontinuation.

For cause-specific mortality analyses, to account for potential competing risks between cancer-specific and non-cancer mortality, we performed competing risk analyses using the Fine and Gray subdistribution hazard model, treating cancer-specific death and non-cancer death as mutually exclusive events. Subdistribution hazard ratios and 95% confidence intervals were estimated, and results were compared with those from cause-specific models. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) and R version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Study Population and Baseline Characteristics

A total of 21,035 patients with newly diagnosed HCC who had been physically inactive before diagnosis were included: 5,639 initiated MVPA after diagnosis and 15,396 remained inactive (Table 1). The mean age was 62.1 years in the control group and 62.0 years in the MVPA initiation group. Male patients comprised 77.5% and 79.8% of the control and MVPA initiation groups, respectively. Hepatitis B infection was present in 70.9% and 73.0%, hepatitis C in 10.3% and 9.6% and cirrhosis in 53.7% and 50.6% of the control and MVPA initiation groups, respectively. A higher proportion of patients in the MVPA initiation group received curative treatment for HCC compared to the control group (53.3% vs. 46.5%). Baseline characteristics were well balanced between groups, with all standardized mean differences <0.1.

Table 1.

Baseline characteristics of participants (N = 21,035)

​	Control (N = 15,396)	MVPA initiation (N = 5,639)	SMD
Age, years	62.1 (10.3)	62.0 (9.9)	0.003
Sex, male	11,934 (77.5)	4,497 (79.8)	0.055
BMI (n = 21,019), kg/m2	24.3 (3.3)	24.3 (3.0)	0.010
Residential area, metropolitan	8,747 (56.8)	3,368 (59.7)	0.059
Medical aid	420 (2.7)	95 (1.7)	0.066
Comorbidities
Charlson Comorbidity Index	4.9 (2.2)	4.8 (2.2)	0.025
Moderate to severe liver disease	420 (2.7)	95 (1.7)	0.071
Hypertension	7,588 (49.3)	2,774 (49.2)	0.002
Diabetes mellitus	6,019 (39.1)	2,183 (38.7)	0.008
Ischemic heart disease	1,460 (9.5)	493 (8.7)	0.026
COPD or asthma	1,964 (12.8)	647 (11.5)	0.039
CKD	82 (0.5)	10 (0.2)	0.060
Hepatitis B	10,918 (70.9)	4,119 (73.0)	0.047
Hepatitis C	1,589 (10.3)	541 (9.6)	0.024
Liver cirrhosis	8,272 (53.7)	2,854 (50.6)	0.062
Hepatic decompensation eventsa	793 (5.2)	220 (3.9)	0.060
Treatment of hepatitis B	8,324 (54.1)	3,176 (56.3)	0.045
Antidiabetics	4,653 (30.2)	1,669 (29.6)	0.014
Antihyperlipidemic	2,994 (19.5)	1,206 (21.4)	0.048
Antihypertensive	9,904 (64.3)	3,599 (63.8)	0.011
Laboratory findings
Fasting glucose (n = 20,845), mg/dL	110.1 (33.9)	108.1 (28.2)	0.067
Total cholesterol (n = 15,174), mg/dL	162.8 (35.0)	163.6 (34.8)	0.024
HDL cholesterol (n = 15,176), mg/dL	49.4 (13.7)	49.8 (13.2)	0.034
LDL cholesterol (n = 10,810), mg/dL	93.8 (30.1)	95.3 (30.8)	0.048
AST (n = 20,846), U/L	38.6 (25.1)	37.4 (30.6)	0.042
ALT (n = 20,846), U/L	31.2 (26.7)	30.9 (35.1)	0.010
GGT (n = 20,846), U/L	71.1 (129.0)	62.0 (84.3)	0.083
Vital sign
SBP (n = 20,852), mm Hg	124.5 (14.8)	124.8 (14.4)	0.019
DBP (n = 20,852), mm Hg	75.5 (9.9)	75.4 (9.6)	0.010
Other habits
Current drinker, yes (n = 21,024)	2,003 (13.0)	574 (10.2)	0.094
Current smoker, yes (n = 21,034)	2,578 (16.7)	704 (12.5)	0.086
Receipt of curative treatment for HCC	7,166 (46.5)	3,004 (53.3)	0.135

Values are presented as n (%), mean (SD), or median (IQR).

ALT, alanine aminotransferase; AST, aspartate aminotransferase; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DBP, diastolic blood pressure; GGT, gamma-glutamyl transferase; HCC, hepatocellular carcinoma; HDL, high-density lipoprotein; IQR, interquartile range; LDL, low-density lipoprotein; MVPA, moderate-to-vigorous physical activity; SBP, systolic blood pressure; SD, standard deviation; SMD, standardized mean difference.

^aAscites, hepatic encephalopathy, hepatorenal syndrome, or variceal bleeding.

Association between MVPA Initiation and Overall Survival

During follow-up (median 4.11 years, IQR: 1.96–7.01), 5,595 and 1,527 deaths (12-year incidence 61.2% vs. 53.9%) occurred in no MVPA and MVPA initiation groups, respectively (Fig. 1). MVPA initiation was associated with a significant reduction in all-cause mortality (adjusted HR: 0.86; 95% CI: 0.82–0.92).

Fig. 1.

All-cause mortality according to MVPA initiation.

This association was consistent across subgroups, including those receiving curative treatment (adjusted HR: 0.86; 95% CI: 0.78–0.95), no curative treatment (0.87; 0.81–0.93), nonviral HCC (0.86; 0.77–0.96), viral HCC (0.87; 0.81–0.93), non-cirrhotic patients (0.88; 0.81–0.96), cirrhotic patients (0.86; 0.80–0.93), and even in the patients with hepatic decompensation (0.75; 0.60–0.94) (Table 2). In dose-response analyses using METs, a graded association was observed between higher physical activity volume and lower mortality risk (online suppl. Table 1).

Table 2.

Comparison of the risk of all-cause mortality according to change in MVPA

​	Event (12-year cumulative incidence, %)		Crude HR (95% CI)	Adjusted HR (95% CI)
	no MVPA	MVPA initiation
Overall	5,561 (61.2)	1,520 (53.9)	0.80 (0.75–0.84)	0.86 (0.82–0.92)
By receipt of curative treatment for HCC
Curative treatment	1,713 (48.6)	534 (41.9)	0.83 (0.75–0.91)	0.86 (0.78–0.95)
No curative treatment	3,848 (70.8)	985 (64.8)	0.84 (0.78–0.90)	0.87 (0.81–0.93)
Hepatitis
Non-hepatitis B and C	2,249 (55.7)	649 (49.4)	0.77 (0.69–0.86)	0.86 (0.77–0.96)
Hepatitis B or C	3,896 (56.7)	1,103 (49.9)	0.81 (0.76–0.87)	0.87 (0.81–0.93)
Cirrhosis
Non-liver cirrhosis	2,248 (55.4)	648 (48.0)	0.80 (0.73–0.87)	0.88 (0.81–0.96)
Liver cirrhosis	3,312 (65.9)	871 (58.5)	0.80 (0.74–0.86)	0.86 (0.80–0.93)
Hepatic decompensation eventsa
No	5,076 (60.2)	1,422 (53.1)	0.81 (0.77–0.86)	0.88 (0.83–0.93)
Yes	484 (76.0)	97 (62.2)	0.68 (0.55–0.84)	0.75 (0.60–0.94)

Adjusted for age, sex, income, area, BMI, comorbidities, antiviral of HBV, current drinker, smoking status and curative treatment for HCC.

CI, confidence interval; HCC, hepatocellular carcinoma; HR, hazard ratio; MVPA, moderate-to-vigorous physical activity.

^aAscites, hepatic encephalopathy, hepatorenal syndrome, or variceal bleeding.

Cause-Specific Mortality Analysis

MVPA initiation was associated with reduced cancer-specific mortality in the cohort (adjusted subdistribution hazard ratios 0.83; 95% CI: 0.78–0.89), with consistent associations observed across all examined subgroups (Table 3). In contrast, no significant association was observed between MVPA initiation and non-cancer-specific mortality in the cohort (adjusted subdistribution hazard ratios 0.97; 95% CI: 0.87–1.09).

Table 3.

Association between MVPA initiation and cause-specific mortality

​	Event (12-year cumulative incidence %)		Crude subdistribution HRs (95% CI)	Adjusted subdistribution HRs (95% CI)
	no MVPA	MVPA initiation
Cancer-cause mortality
Overall	4,226 (42.2)	1,139 (35.7)	0.78 (0.73–0.84)	0.83 (0.78–0.89)
By receipt of curative treatment for HCC
Curative treatment	1,244 (31.2)	393 (26.2)	0.83 (0.74–0.93)	0.85 (0.76–0.95)
No curative treatment	2,982 (50.8)	746 (45.2)	0.82 (0.76–0.89)	0.83 (0.76–0.90)
Hepatitis
Non-hepatitis B and C	1,168 (48.8)	285 (38.1)	0.75 (0.66–0.86)	0.81 (0.71–0.93)
Hepatitis B or C	3,058 (40.1)	854 (35.2)	0.80 (0.74–0.86)	0.84 (0.78–0.91)
Cirrhosis
Non-liver cirrhosis	1,716 (37.7)	485 (30.1)	0.78 (0.71–0.86)	0.84 (0.75–0.93)
Liver cirrhosis	2,510 (46.0)	654 (40.9)	0.79 (0.73–0.86)	0.84 (0.77–0.91)
Hepatic decompensation eventsa
No	3,857 (41.4)	1,060 (35.1)	0.79 (0.74–0.85)	0.84 (0.78–0.90)
Yes	369 (55.5)	79 (48.7)	0.78 (0.68–0.99)	0.87 (0.68–1.11)
Non-cancer-cause mortality
Overall	1,411 (19.2)	404 (18.0)	0.89 (0.80–0.99)	0.97 (0.87–1.09)
By receipt of curative treatment for HCC
Curative treatment	500 (17.5)	146 (14.6)	0.83 (0.69–0.99)	0.88 (0.73–1.05)
No curative treatment	911 (20.3)	258 (20.6)	0.97 (0.85–1.11)	1.04 (0.91–1.20)
Hepatitis
Non-hepatitis B and C	509 (27.2)	138 (30.8)	0.89 (0.74–1.08)	0.99 (0.82–1.20)
Hepatitis B or C	902 (16.8)	266 (14.4)	0.90 (0.79–1.03)	0.98 (0.85–1.12)
Cirrhosis
Non-liver cirrhosis	562 (18.1)	172 (18.1)	0.90 (0.76–1.06)	1.02 (0.86–1.21)
Liver cirrhosis	849 (20.3)	232 (18.0)	0.89 (0.77–1.02)	0.95 (0.83–1.10)
Hepatic decompensation eventsa
No	1,289 (19.1)	384 (18.3)	0.92 (0.82–1.02)	1.00 (0.89–1.12)
Yes	122 (21.1)	20 (14.0)	0.62 (0.39–0.99)	0.70 (0.43–1.13)

Adjusted for age, sex, income, area, BMI, comorbidities, antiviral of HBV, current drinker, smoking status and curative treatment for HCC.

CI, confidence interval; HCC, hepatocellular carcinoma; HR, hazard ratio; MVPA, moderate-to-vigorous physical activity.

^aAscites, hepatic encephalopathy, hepatorenal syndrome, or variceal bleeding.

Propensity Score-matched Analysis

In the 1:2 propensity score-matched cohort (MVPA initiation: 5,624; control: 11,060), the association between MVPA initiation and overall survival remained consistent (HR: 0.87; 95% CI: 0.82–0.92) across all subgroups (online suppl. Tables 2 and 3). In an as-treated sensitivity analysis that censored follow-up at the third health examination if physical activity status changed after exam 2 (N = 8,492), MVPA initiation remained associated with lower all-cause mortality (adjusted HR: 0.81; 95% CI: 0.69–0.94; online suppl. Table 4). When participants were further classified according to changes in MVPA status between exam 2 and exam 3, the lowest mortality risk was observed among persistently active individuals, whereas the survival benefit was attenuated among those who discontinued MVPA, supporting the robustness of the primary findings (online suppl. Table 5).

Discussion

In this large-scale, nationwide cohort study of 21,035 patients with HCC who were physically inactive before diagnosis, we found that initiating MVPA after diagnosis was associated with a 14% reduction in all-cause mortality compared to remaining inactive. This association remained consistent across clinically relevant subgroups, including patients who received curative treatment, those without curative treatment, viral and nonviral HCC, and both cirrhotic and non-cirrhotic patients. To our knowledge, this is the first study to investigate the impact of post-diagnosis physical activity initiation on survival outcomes specifically in patients with HCC, providing evidence that the survival benefits of physical activity observed in other cancer types extend to this population despite the unique challenges posed by underlying liver disease.

Our findings align with and extend the evidence showing survival benefits of physical activity across various cancer types [11, 20, 21]. A meta-analysis by Friedenreich et al. [11] including 136 studies found that post-diagnosis physical activity was associated with substantially reduced mortality in multiple cancer sites, with hazard ratios ranging from 0.58 to 0.63 for breast and colorectal cancers. However, data on HCC were absent from that analysis, highlighting a critical knowledge gap that our study addresses. The magnitude of benefit observed in our study is comparable to that reported for other solid tumors, suggesting that despite the underlying liver dysfunction characteristic of HCC, patients can still achieve meaningful survival benefits through physical activity. This benefit was observed even among patients with cirrhosis, who have reduced functional reserve due to advanced liver disease.

The observed survival benefit likely results from multiple complementary mechanisms. Physical activity may reduce cancer-specific mortality through direct anti-tumor effects, recurrence prevention, and improved treatment tolerance and adherence. Exercise enhances immune function, including natural killer cell activity and T-cell function, which are critical for tumor surveillance and control [3, 22]. These immune-enhancing effects may create a less permissive environment for tumor growth and progression. Additionally, physical activity may also contribute to recurrence prevention. The survival benefit among patients who received curative treatment, where residual tumor burden is theoretically minimized, suggests that post-diagnosis physical activity may reduce late recurrence risk. This is particularly relevant because late recurrence after curative treatment represents de novo HCC development rather than metastasis from the original tumor. In this context, a meta-analysis showing that physical activity is associated with a 35% reduction in de novo HCC incidence suggests that post-diagnosis physical activity may reduce late recurrence risk in patients who have undergone curative treatment [23]. While systematic reviews in other cancer types have shown approximately 50% reduction in recurrence with exercise [21], specific data on physical activity and HCC recurrence remain limited, representing an important area for future investigation. Furthermore, improved physical fitness may enhance tolerance of anticancer therapies, enabling more complete treatment delivery with fewer dose reductions or interruptions [24]. This improved treatment adherence may contribute to better tumor control and improved survival outcomes, particularly in patients receiving systemic therapies or those with advanced disease.

Beyond direct anti-tumor effects, physical activity may improve survival by preserving liver function and preventing cardiovascular and metabolic complications. Recent studies have shown that muscle status and physical function are important prognostic factors in HCC patients. Yoshio et al. [25] showed that myostatin, a marker of muscle wasting, serves as a fibroblast-activating factor that promotes liver fibrosis and impacts postoperative outcomes, while Kawaguchi et al. [26] showed that decorin, a myokine, is associated with physical function and prognosis in HCC patients. These findings suggest that exercise-induced improvements in muscle mass and function may contribute to better outcomes through both direct metabolic effects and attenuation of liver fibrosis progression. In the context of chronic liver disease, physical activity has been shown to improve hepatic steatosis, reduce hepatic fibrosis progression, and enhance overall liver function [10, 27–31], which is particularly critical given that many HCC patients have compromised hepatic reserve. Additionally, exercise reduces systemic inflammation and improves insulin sensitivity, preventing metabolic and cardiovascular disorders that may contribute substantially to mortality in this population [4, 32]. The metabolic benefits of exercise, including improvements in body composition and reduction of visceral adiposity [33], may further contribute to improved survival in HCC patients with chronic liver disease.

In our cause-specific mortality analysis, the association between physical activity and reduced mortality was observed for cancer-related deaths but not for non-cancer-related deaths. The observed survival benefit may primarily operate through cancer-directed pathways, including enhanced immune function, reduced recurrence risk, and improved tolerance of anticancer therapies. The consistency of findings across subgroups defined by treatment status, etiology, and cirrhosis presence indicates that these beneficial mechanisms operate broadly across diverse HCC populations. While physical activity has well-established benefits for liver function and cardiovascular health in chronic liver disease, these effects may require longer observation periods to manifest as measurable reductions in non-cancer mortality in this population. Additionally, the high rate of cancer-specific deaths in our HCC cohort creates a competing risk scenario that may limit our ability to detect associations with non-cancer mortality, including liver-related deaths from cirrhosis complications.

The feasibility and clinical benefits of exercise interventions in HCC patients have been demonstrated in several prospective studies. Tsuchihashi et al. [34] reported that in-hospital exercise (combination of aerobic and resistance training, 20–40 min/day) significantly improved the Liver Frailty Index in HCC patients without worsening liver function, while Narao et al. [35] showed that physical therapy during hospitalization improved Activities of Daily Living scores as measured by the Functional Independence Measure, particularly in mobility-related domains. These findings support the clinical feasibility of exercise interventions in HCC populations and provide mechanistic insights suggesting that physical activity contributes to survival through maintenance and improvement of physical function and frailty status. The consistency of these short-term functional improvements with our observed long-term survival benefits strengthens the rationale for incorporating physical activity into comprehensive HCC care.

Our study has several notable strengths. The use of nationwide data covering nearly the entire Korean population minimizes selection bias and enhances generalizability. The large sample size enabled robust subgroup analyses to examine consistency of associations across clinically relevant patient characteristics. The prospective assessment of physical activity through standardized health examinations reduces recall bias inherent in retrospective studies. Additionally, the adjustment for potential confounders, including sociodemographic factors, comorbidities, and receipt of curative treatment, strengthens causal inference. The propensity score-matched sensitivity analysis further strengthens the robustness of our findings.

Several limitations warrant consideration. First, our study design required patients to survive until their post-diagnosis health examination to assess physical activity changes, potentially excluding patients with more aggressive disease or poor general health. Therefore, our findings represent results from a population of HCC patients with sufficient health status to undergo health examinations and potentially initiate physical activity. This survivor selection may overestimate the association; however, the landmark-based design was intentionally adopted to minimize reverse causation and to distinguish the effect of post-diagnosis physical activity from underlying disease severity and baseline functional status. Second, as an observational study, residual confounding cannot be entirely excluded despite extensive adjustment. Patients who initiated physical activity after diagnosis may differ from those who remained inactive in unmeasured ways that could influence survival. However, the well-balanced baseline characteristics between groups and consistent findings in propensity score-matched analyses mitigate this concern. Third, physical activity was assessed at a single time point after diagnosis using a self-reported questionnaire and analyzed as a binary variable (MVPA achieved/not achieved), which does not capture quantitative details, or the contribution of light physical activity. Self-reported physical activity may be subject to measurement error, though the use of standardized questionnaires administered during national health examinations likely improves accuracy. To address concerns regarding single-timepoint assessment and exposure misclassification, we conducted complementary sensitivity analyses including as-treated analyses with censoring at exposure reassessment and trajectory analyses comparing persistent, initiating, and discontinuing MVPA patterns. The consistency of results across these analyses suggests our findings are robust to alternative exposure definitions. Nevertheless, more frequent and objective assessment using wearable devices or repeated measurements would provide more precise exposure characterization and strengthen causal inference. Fourth, we lacked information on cancer stage, tumor characteristics, and specific treatment details beyond the broad categorization of curative versus non-curative treatment. Therefore, our findings represent the association between initiating physical activity and survival in real-world HCC population across various disease stages rather than being limited to patients in complete remission, limiting our ability to examine whether associations vary by disease severity or treatment response. We also lacked data on liver function severity as the administrative database does not contain laboratory values necessary to calculate Child-Pugh or Model for End-Stage Liver Disease scores. However, subgroup analyses stratified by cirrhosis status and hepatic decompensation showed consistent beneficial associations across underlying liver status. Finally, the study population was exclusively Asian with predominantly viral etiology, and findings may not be directly generalizable to populations where nonviral etiologies predominate.

In conclusion, initiating MVPA after HCC diagnosis was associated with improved survival among previously inactive patients, regardless of treatment modality, viral etiology, or presence of cirrhosis. These findings suggest that physical activity should be actively promoted as an integral component of comprehensive HCC care. Future research should focus on identifying optimal physical activity prescriptions for HCC patients, including intensity, duration, and timing related to treatment, as well as developing and testing practical implementation strategies to promote physical activity adoption in clinical practice. Randomized controlled trials evaluating structured exercise interventions in HCC patients would provide definitive evidence regarding causality and inform evidence-based clinical guidelines for this population.

Statement of Ethics

This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board (IRB) of Samsung Medical Center (IRB Approval No. 2025-05-018). Written informed consent was not required as the IRB of Samsung Medical Center waived the requirement for informed consent due to the use of de-identified administrative data.

Conflict of Interest Statement

The authors have no conflicts of interest relevant to this study to disclose.

Funding Sources

This research was supported by the Bio&Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (MSIT) (No. RS-2024-00440881) and supported by the National Institute of Health (NIH) research project (project No. 2025-ER1904-00). This research was also supported by the Institution of Quality of Life in Cancer, funded by Samsung Fire & Marine Insurance.

Author Contributions

Study concept and design, data analysis and interpretation, and drafting of the manuscript: Byeong Geun Song, Myeongcheol Lee, Danbee Kang, and Dong Hyun Sinn. Acquisition of data and statistical analysis: Myeongcheol Lee and Danbee Kang. Critical revision of the manuscript: Juhee Cho, Geum-Youn Gwak, Danbee Kang, and Dong Hyun Sinn. Data management and methodology supervision: Juhee Cho and Danbee Kang. Supervision: Danbee Kang and Dong Hyun Sinn. All authors participated in the preparation of the manuscript and have seen and approved the final version.

Funding Statement

This research was supported by the Bio&Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (MSIT) (No. RS-2024-00440881) and supported by the National Institute of Health (NIH) research project (project No. 2025-ER1904-00). This research was also supported by the Institution of Quality of Life in Cancer, funded by Samsung Fire & Marine Insurance.

Data Availability Statement

The data that support the findings of this study are not publicly available due to legal and ethical restrictions related to the use of the Korean National Health Insurance Service (K-NHIS) database but are available from the corresponding author upon reasonable request and with permission from the K-NHIS.