Abstract
Evidence regarding inpatient rehabilitation for clinically complex patients with chronic liver disease (CLD), including those with advanced cirrhosis, remains limited. We aimed to describe real-world inpatient rehabilitation practice and unadjusted in-hospital changes in hepatic and renal laboratory measures. This single-center retrospective observational study included 200 patients hospitalized for CLD-related conditions at Wakayama Medical University Hospital between January and December 2023. Referral and treatment were determined clinically rather than by standardized criteria. We compared baseline characteristics and unadjusted between-group changes in hepatic and renal laboratory measures, with exploratory analyses stratified by Child–Pugh (CP) class. Rehabilitation-related adverse events were retrospectively assessed using complete inpatient medical records and institutional incident reports. Of 200 hospitalized patients, 57 received inpatient rehabilitation. The rehabilitation group was older, had higher Child–Pugh scores, was more often admitted emergently, and had longer hospital stays than the non-rehabilitation group. Median rehabilitation duration was 11.0 days (interquartile range [IQR], 6.0–20.0), and the median time from hospital admission to rehabilitation referral was 6.0 days (IQR, 3.0–9.0). Across Child-Pugh classes, unadjusted hepatic and renal laboratory changes did not show a consistent pattern of greater deterioration in the rehabilitation group. No serious rehabilitation-related adverse event was identified through review of the complete inpatient medical records or the institutional incident-reporting system. These real-world data characterize inpatient rehabilitation in a clinically complex CLD population, including patients with advanced cirrhosis, and may inform development of liver rehabilitation for hospitalized, medically vulnerable patients. Because referral and treatment were individualized and the comparisons were unadjusted, the observations should be interpreted as descriptive associations rather than evidence of a causal effect, overall safety, or efficacy.
Keywords: Child–Pugh classification, chronic liver disease, hepatic functional reserve, inpatient rehabilitation, renal function, retrospective observational study
1. Introduction
Chronic liver disease (CLD) develops and progresses from diverse etiologies, including viral hepatitis and metabolic dysfunction-associated steatotic liver disease (MASLD), and can lead to cirrhosis and hepatocellular carcinoma (HCC). Liver cancer affected approximately 865,000 people worldwide and caused approximately 758,000 deaths in 2022, ranking third among causes of cancer death.[1] Metabolic dysfunction-associated steatotic liver disease (MASLD), historically termed nonalcoholic fatty liver disease in much of the earlier literature, is a major cause of CLD, and its burden is projected to increase.[2] From the perspective of healthcare resource consumption, CLD imposes a substantial societal burden, necessitating urgent measures.
As CLD progresses, sarcopenia becomes more prevalent and is associated with poorer prognosis in patients with cirrhosis.[3] Sarcopenia is also associated with impaired health-related quality of life in patients with cirrhosis.[4] Traditionally, bed rest was often recommended for patients with liver disease, but concerns have arisen regarding the progression of disuse syndrome due to excessive bed rest, and in recent years, the potential role of rehabilitation has gained attention.
The Japan Society of Hepatology provides liver rehabilitation guidance[5] that describes exercise therapy in patients with relatively preserved hepatic functional reserve, including mild cirrhosis (Child–Pugh class A). Importantly, the guidance states that the usefulness and safety of exercise therapy in patients with decompensated cirrhosis and reduced hepatic functional reserve (Child–Pugh class C) require further investigation.
However, most previous studies have evaluated outpatient or home-based interventions, and few have described rehabilitation during hospitalization, particularly among patients with poor general health, advanced cirrhosis, or severe portal hypertension. Many previous studies have also excluded patients with variceal bleeding or other features of decompensation. Consequently, evidence from the setting in which inpatient rehabilitation is actually considered remains limited.
Accordingly, we conducted a single-center retrospective observational study to characterize real-world inpatient rehabilitation in a clinically complex population of patients hospitalized with CLD, including those with advanced cirrhosis, and to describe unadjusted in-hospital changes in hepatic and renal laboratory measures by rehabilitation status. In this medically vulnerable population, hepatic and renal function may fluctuate or deteriorate during hospitalization; therefore, these measures were selected as clinically meaningful organ-function signals. They were treated as surrogate outcomes rather than comprehensive measures of rehabilitation safety or efficacy. By providing descriptive real-world data on inpatient rehabilitation delivered as part of routine care, we aimed to inform the further development of liver rehabilitation for hospitalized, medically vulnerable patients with CLD.
2. Methods
2.1. Research design & patient groups
This was a retrospective single-center observational study. We screened patients admitted to the Department of Gastroenterology at Wakayama Medical University Hospital for treatment or symptom management of liver disease between January 1 and December 31, 2023. After application of the exclusion criteria, the analytic cohort was restricted to chronic liver disease-related conditions, including chronic hepatitis, cirrhosis, metabolic dysfunction-associated steatotic liver disease (MASLD), metabolic dysfunction-associated steatohepatitis, and hepatocellular carcinoma (HCC), to describe real-world inpatient practice. Acute hepatic and hepatobiliary conditions were excluded to reduce the influence of rapid treatment-related fluctuations in laboratory values. Because the included etiologies and admission indications were heterogeneous, the study was not designed to estimate disease-specific associations. Patients were classified into a rehabilitation group and a non-rehabilitation group.
Patients who declined participation through the opt-out process were excluded. The predefined disease- or procedure-related exclusion criteria included acute hepatitis, drug-induced liver injury, acute conditions not associated with CLD, infectious diseases such as liver abscess and cholangitis, radiofrequency ablation (RFA) or microwave ablation (MWA), liver biopsy, neoplastic lesions other than HCC, congestive hepatopathy due to right-sided heart failure, and traumatic hepatic injury. Six additional patients were excluded because treatment was discontinued (n = 3) or because the electronic medical record contained insufficient information for analysis (n = 3).
2.2. Rehabilitation therapy
Rehabilitation referral was initiated at the discretion of the treating hepatologist; no predefined institutional referral criteria or standardized referral algorithm was used. After referral, a rehabilitation physician assessed the patient’s overall condition and prescribed rehabilitation. The decision to start rehabilitation and its timing remained individualized, based on the attending physician’s clinical judgment and the patient’s condition. Reasons for rehabilitation referral and non-referral were not recorded in a standardized manner and therefore could not be systematically analyzed retrospectively.
Rehabilitation was delivered by licensed physical therapists after assessment by a rehabilitation physician. This collaborative model has previously been described at our institution as physiatrist and registered therapist operating rehabilitation (PROr). Under this approach, therapists aimed to increase workload and session duration as tolerated while tailoring the program to each patient’s clinical condition and motivation.[6] The program could include range-of-motion exercises, resistance training, aerobic exercise, and activities of daily living (ADL) training. The specific content, intensity, timing, and progression were individualized rather than based on a standardized protocol. Before, during, and after each rehabilitation session, therapists routinely assessed the patient’s vital signs, symptoms, and overall clinical condition. Decisions to continue, modify, interrupt, or discontinue rehabilitation were individualized and made under the direction of the treating hepatologist or rehabilitation physician, as appropriate. The duration of inpatient rehabilitation and the time from hospital admission to rehabilitation referral were extracted from routine records. Detailed session-level information, including minutes per session, treatment frequency, progression, adherence, the numbers of prescribed, completed, missed, or interrupted sessions, and reasons for missed or interrupted sessions, was not available retrospectively; therefore, intervention heterogeneity could not be quantified.
2.3. Outcomes
The primary outcomes were in-hospital changes in hepatic functional reserve and renal function measures. Demographic and baseline clinical characteristics (age, sex, height, weight, body mass index [BMI], primary disease, medical history, etiology of cirrhosis, Child–Pugh [CP] class, and CP score) were extracted from medical records. Hospitalization-related data included length of stay, admission type (scheduled, emergency, or interdepartmental transfer), purpose of admission (e.g., transarterial chemoembolization [TACE] and balloon-occluded retrograde transvenous obliteration [BRTO]), whether rehabilitation was performed, the duration of inpatient rehabilitation, the time from hospital admission to rehabilitation referral, and discharge destination (home, transfer to another hospital, or in-hospital death). Potential worsening of hepatic and renal function is a key clinical concern when exercise-based rehabilitation is considered in patients with advanced or decompensated CLD. Therefore, hepatic and renal measures were selected as organ-function signals to describe possible deterioration during hospitalization in this real-world inpatient cohort. They were treated as surrogate outcomes and do not capture functional recovery, health-related quality of life, or overall rehabilitation safety.
Laboratory data were collected at admission (within 24 hours of hospitalization) and at discharge (within 72 hours before discharge, transfer, or death). Laboratory items included albumin, total bilirubin, creatinine, prothrombin activity, estimated glomerular filtration rate (eGFR), blood urea nitrogen, prothrombin time–international normalized ratio (PT–INR), aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyl transpeptidase (γ-GTP), platelet count, ammonia, Model for End-Stage Liver Disease (MELD) score,[7] and albumin-bilirubin (ALBI) score. Changes in laboratory values (Δ) were calculated as discharge minus admission values.
Rehabilitation-related adverse events were retrospectively assessed by reviewing the complete inpatient medical records of patients in the rehabilitation group and cross-checking institutional incident reports submitted by the Rehabilitation Department to the Medical Safety Promotion Department between January 1, 2021 and December 31, 2024. At our hospital, staff are required to report adverse events to the appropriate risk manager. The Medical Safety Promotion Department classifies reported events into 9 severity levels: category A (no error), categories B–D (error, no harm), categories E-H (error, harm), and Category I (error, death). Events were categorized by time of occurrence (during rehabilitation or within 24 hours after rehabilitation) and type (e.g., fall, hemodynamic instability, exacerbated pain, or device removal).
2.4. Ethical considerations
This study involved human participants and was approved by the Ethics Review Committee of Wakayama Medical University (approval No. 4434). The study was conducted in accordance with the Declaration of Helsinki. The requirement for individual written informed consent was waived because of the retrospective observational design. Information regarding study participation was provided via an opt-out notice posted on the hospital website, and participants were given the opportunity to refuse use of their data.
2.5. Statistical analysis
All statistical analyses were performed using GraphPad Prism version 10 (GraphPad Software, San Diego).
The normality of continuous variables was assessed using the Shapiro–Wilk test.
Continuous variables are presented as medians with interquartile ranges.
Between-group comparisons of continuous variables were performed using the Mann–Whitney U test, given the unequal sample sizes and the potential influence of outliers.
Categorical variables are presented as frequencies (%), and Fisher exact test was used for between-group comparisons.
All statistical tests were 2-sided, with P < .05 considered statistically significant.
To explore variation by disease severity, subgroup analyses were stratified by Child–Pugh classification. All between-group comparisons were unadjusted and should be interpreted as descriptive associations; stratification did not eliminate residual confounding from baseline differences or nonstandardized rehabilitation referral.
3. Results
Of 384 patients admitted to the Department of Gastroenterology for treatment or symptom management of liver disease, 178 were excluded for predefined disease- or procedure-related reasons, and 6 were additionally excluded because treatment was discontinued or the electronic medical record contained insufficient information for analysis. The final analytic cohort comprised 200 patients with CLD-related conditions. Of these, 57 received inpatient rehabilitation, and 143 did not (Fig. 1).
Figure 1.
Flow diagram of patient selection. Among 384 patients admitted for treatment of liver disease between January and December 2023, 178 were excluded according to predefined criteria, and 6 were excluded because of treatment discontinuation or incomplete records. Ultimately, 200 patients were included and categorized into a rehabilitation group (n = 57) and a non-rehabilitation group (n = 143). Unadjusted changes in hepatic and renal laboratory measures were described overall and by Child–Pugh classification, whereas rehabilitation-related adverse events were summarized for the rehabilitation group. BRTO = balloon-occluded retrograde transvenous obliteration, MWA = microwave ablation, RFA = radiofrequency ablation, TACE = transarterial chemoembolization.
3.1. Baseline characteristics
Table 1 summarizes the baseline characteristics of the rehabilitation and non-rehabilitation groups.
Table 1.
Baseline characteristics of patients with chronic liver disease.
| Variables | Total (n = 200) | Rehabilitation (n = 57) | Non-rehabilitation (n = 143) | P value |
|---|---|---|---|---|
| Age | 72.0 (63.8–80.0) | 80.0 (70.0–82.0) | 70.0 (60.0–76.0) | <.001 |
| Sex | .009 | |||
| Male | 143 (71.5) | 33 (57.9) | 110 (76.9) | |
| Female | 57 (28.5) | 24 (42.1) | 33 (23.1) | |
| BMI (kg/m2) | 23.7 (21.4–26.7) | 24.1 (22.2–26.1) | 23.2 (21.0–26.7) | .32 |
| Hepatocellular carcinoma | 119 (59.5) | 35 (61.4) | 84 (58.7) | .75 |
| Esophageal varices | 97 (48.5) | 24 (42.1) | 73 (51.0) | .28 |
| Type of admission | .002 | |||
| Scheduled | 143 (71.5) | 31 (54.4) | 112 (78.3) | |
| Emergency | 51 (25.5) | 23 (40.4) | 28 (19.6) | |
| Transfer/referral | 6 (3) | 3 (5.3) | 3 (2.1) | |
| Child–Pugh score | 7.0 (6.0–9.0) | 9.0 (7.0–10.0) | 7.0 (6.0–9.0) | .001 |
| Child–Pugh classification | .008 | |||
| A | 82 (41) | 14 (24.6) | 68 (47.6) | |
| B | 68 (34) | 23 (40.4) | 45 (31.5) | |
| C | 50 (25) | 20 (35.1) | 30 (21.0) | |
| Purpose of admission | .009 | |||
| TACE | 49 (24.5) | 14 (24.6) | 35 (24.5) | |
| Chemotherapy | 35 (17.5) | 8 (14.0) | 27 (18.9) | |
| Variceal treatment | 45 (22.5) | 6 (10.5) | 39 (27.3) | |
| Liver failure treatment | 60 (30.0) | 27 (47.4) | 33 (23.1) | |
| BRTO | 8 (4.0) | 1 (1.8) | 7 (4.9) | |
| Other | 3 (1.5) | 1 (1.8) | 2 (1.4) | |
| Length of hospital stay (d) | 11.0 (8.0–18.3) | 18.0 (12.0–27.0) | 9.0 (7.0–13.0) | <.001 |
| Length of rehabilitation (d) | — | 11.0 (6.0–20.0) | — | — |
| Time from hospital admission to rehabilitation referral (d) | — | 6.0 (3.0–9.0) | — | — |
| Discharge destination | <.001 | |||
| Home | 167 (83.5) | 33 (57.9) | 134 (93.7) | |
| Transfer/referral | 19 (9.5) | 17 (29.8) | 2 (1.4) | |
| Death | 14 (7.0) | 7 (12.3) | 7 (4.9) | |
| At admission | ||||
| Albumin (g/dL) | 3.1 (2.6–3.6) | 2.9 (2.4–3.4) | 3.3 (2.7–3.6) | .005 |
| Total bilirubin (mg/dL) | 1.0 (0.7–2.0) | 1.1 (0.7–2.2) | 1.0 (0.7–1.9) | .69 |
| Creatinine (mg/dL) | 0.96 (0.74–1.32) | 1.13 (0.70–1.66) | 0.88 (0.74–1.23) | .05 |
| Prothrombin activity (%) | 73.6 (55.9–86.6) | 68.9 (52.6–84.9) | 75.7 (60.0–86.6) | .22 |
| eGFR (mL/min/1.73 m2) | 58.5 (39.7–71.3) | 48.9 (31.8–65.6) | 61.1 (44.1–72.2) | .005 |
| Blood urea nitrogen (mg/dL) | 18.4 (13.5–26.5) | 23.0 (16.5–35.0) | 17.2 (12.8–23.7) | .001 |
| PT–INR | 1.18 (1.08–1.36) | 1.21 (1.09–1.44) | 1.17 (1.08–1.33) | .12 |
| AST (U/L) | 38.0 (28.0–62.3) | 35.0 (27.0–59.0) | 39.0 (28.5–62.5) | .50 |
| ALT (U/L) | 25.0 (17.0–38.3) | 26.0 (15.0–35.0) | 25.0 (18.0–38.5) | .42 |
| γ-GTP (U/L) | 55.0 (29.5–118.5) | 40.0 (22.0–90.0) | 63.0 (35.0–130.0) | .008 |
| Platelets (×104/μL) | 10.7 (7.7–15.5) | 10.6 (8.0–14.5) | 10.9 (7.6–15.7) | .79 |
| Ammonia (μmol/L) | 67.5 (53.0–97.3) | 63.0 (51.5–94.5) | 68.0 (54.0–98.0) | .47 |
| MELD | 10.0 (8.0–14.8) | 13.0 (8.8–17.0) | 10.0 (8.0–14.0) | .005 |
Values are expressed as median (interquartile range) or number (%). P values were calculated using the Mann–Whitney U test for continuous variables and Fisher exact test for categorical variables.
An em dash (—) indicates that the value or statistical test was not applicable.
ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, BRTO = balloon-occluded retrograde transvenous obliteration, CP = Child–Pugh classification, eGFR = estimated glomerular filtration rate, MELD = Model for End-Stage Liver Disease, PT–INR = prothrombin time–international normalized ratio, TACE = transarterial chemoembolization, γ-GTP = gamma-glutamyl transpeptidase.
Because rehabilitation was initiated through clinical referral rather than random allocation, substantial baseline differences between groups were expected. The following between-group comparisons were unadjusted and should not be interpreted as causal effects of rehabilitation. Patients with missing data for variables of interest were excluded before the analysis, and all analyses were conducted using a complete-case dataset.
The rehabilitation group was significantly older than the non-rehabilitation group (80.0 vs 70.0 years, P < .001).
The proportion of women was higher in the rehabilitation group compared with the non-rehabilitation group (42.1% vs 23.1%, P = .009).
Regarding admission type, emergency admissions were more common in the rehabilitation group (40.4% vs 19.6%), whereas scheduled admissions were more common in the non-rehabilitation group (78.3% vs 54.4%; P = .002).
The rehabilitation group had significantly higher Child–Pugh scores (9.0 vs 7.0, P = .001), included a larger proportion of patients with Child–Pugh class C disease (35.1% vs 21.0%, P = .008), and had higher MELD scores (13.0 vs 10.0, P = .005).
Length of hospital stay was significantly longer in the rehabilitation group (18.0 vs 9.0 days, P < .001).
Among patients who received rehabilitation, the median duration of inpatient rehabilitation was 11.0 days (interquartile range [IQR], 6.0–20.0), and the median time from hospital admission to rehabilitation referral was 6.0 days (IQR, 3.0–9.0; Table 1). Corresponding values by CP class are reported in Tables 3–5.
Table 3.
Clinical characteristics and laboratory data stratified by rehabilitation status in patients with CP-A.
| Variables | Rehabilitation (n = 14) | Non-rehabilitation (n = 68) | P value |
|---|---|---|---|
| Age | 80.0 (77.3–84.8) | 71.5 (64.8–80.0) | .001 |
| Sex | .75 | ||
| Male | 9 (64.3) | 48 (70.6) | |
| Female | 5 (35.7) | 20 (29.4) | |
| BMI (kg/m2) | 24.1 (24.0–25.2) | 23.2 (21.9–25.9) | .21 |
| Hepatocellular carcinoma | 12 (85.7) | 46 (67.6) | .22 |
| Esophageal varices | 2 (14.3) | 25 (36.8) | .13 |
| Child–Pugh score | 6.0 (5.0–6.0) | 6.0 (5.0–6.0) | .77 |
| Type of admission | .22 | ||
| Scheduled | 12 (85.7) | 63 (92.6) | |
| Emergency | 1 (7.1) | 5 (7.4) | |
| Transfer/referral | 1 (7.1) | 0 (0) | |
| Purpose of admission | .24 | ||
| TACE | 9 (64.3) | 27 (39.7) | |
| Chemotherapy | 2 (14.3) | 14 (20.6) | |
| Variceal treatment | 1 (7.1) | 19 (27.9) | |
| Liver failure treatment | 1 (7.1) | 2 (2.9) | |
| BRTO | 0 (0) | 4 (5.9) | |
| Other | 1 (7.1) | 2 (2.9) | |
| Length of hospital stay (d) | 13.0 (10.0–14.0) | 9.0 (7.0–11.0) | .002 |
| Length of rehabilitation (d) | 6.5 (5.3–7.8) | — | — |
| Time from hospital admission to rehabilitation referral (d) | 6.0 (4.0–7.8) | — | — |
| Discharge destination | .05 | ||
| Home | 11 (78.6) | 66 (97.1) | |
| Transfer/referral | 2 (14.3) | 1 (1.5) | |
| Death | 1 (7.1) | 1 (1.5) | |
| Blood collection upon admission | |||
| At admission | |||
| Albumin (g/dL) | 3.6 (3.3–3.8) | 3.6 (3.4–3.9) | .85 |
| Total bilirubin (mg/dL) | 0.6 (0.5–0.6) | 0.7 (0.6–1.0) | .004 |
| Creatinine (mg/dL) | 0.81 (0.67–1.44) | 0.83 (0.69–1.10) | .82 |
| Prothrombin activity (%) | 90.3 (83.3–97.7) | 85.6 (79.3–92.0) | .16 |
| eGFR (mL/min/1.73 m2) | 65.8 (36.9–70.0) | 63.5 (49.3–74.8) | .73 |
| Blood urea nitrogen (mg/dL) | 18.6 (17.0–26.5) | 16.9 (12.8–22.1) | .12 |
| PT–INR | 1.06 (1.01–1.11) | 1.09 (1.05–1.13) | .17 |
| AST (U/L) | 28.0 (23.5–34.7) | 35.5 (25.7–51.2) | .14 |
| ALT (U/L) | 20.5 (12.5–36.0) | 24.0 (17.5–40.0) | .17 |
| γ-GTP (U/L) | 39.5 (14.2–82.5) | 64.0 (34.5–107.5) | .06 |
| Platelets (×104/μL) | 14.0 (10.4–18.6) | 12.7 (8.9–17.0) | .47 |
| Ammonia (μmol/L) | 54.5 (46.5–67.0) | 60 (51.0–73.0) | .24 |
| MELD | 7.5 (7.0–11.7) | 8.0 (7.0–10.0) | .92 |
| Changes | |||
| Δ Albumin (g/dL) | −0.5 (−0.7 to −0.3) | −0.4 (−0.5 to −0.1) | .18 |
| Δ Total bilirubin (mg/dL) | 0.1 (0.0–0.3) | 0.1 (−0.1 to 0.3) | .83 |
| Δ Creatinine (mg/dL) | −0.06 (−0.11 to 0.02) | 0.02 (−0.07 to 0.04) | .27 |
| Δ Prothrombin activity (%) | −15.8 (−18.5 to −13.6) | −5.2 (−10.2 to 0.0) | .001 |
| Δ eGFR (mL/min/1.73 m2) | 6.6 (0.9–11.2) | 1.0 (−4.5 to 6.6) | .04 |
| Δ Blood urea nitrogen (mg/dL) | −4.6 (−8.0 to 0.3) | −1.2 (−5.1 to 2.2) | .24 |
| Δ PT–INR | 0.11 (0.08–0.13) | 0.03 (0.00–0.08) | .002 |
| Δ AST (U/L) | 1.5 (−4.7 to 31.5) | 1.0 (−8.0 to 5.7) | .06 |
| Δ ALT (U/L) | 17.5 (0.5–25.5) | 1.5 (−3.0 to 19.5) | .15 |
| Δ γ-GTP (U/L) | 3.0 (−1.5 to 15.2) | −3.0 (−13.2 to 3.0) | .02 |
| Δ Platelets (×104/μL) | 0.8 (−0.2 to 3.5) | −0.5 (−1.4 to 0.8) | .07 |
| Δ Ammonia (μmol/L) | −4.5 (−10.5 to 8.5) | 1.0 (−10.5 to 22.0) | .31 |
| Δ MELD | 1.00 (0.00–1.75) | 0.00 (−0.25 to 1.00) | .67 |
Values are expressed as median (interquartile range) or number (%).
P values were calculated using the Mann–Whitney U test for continuous variables and Fisher exact test for categorical variables.
Δ indicates the change from admission to discharge.
An em dash (—) indicates that the value or statistical test was not applicable.
ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, BRTO = balloon-occluded retrograde transvenous obliteration, CP = Child–Pugh classification, eGFR = estimated glomerular filtration rate, MELD = Model for End-Stage Liver Disease, PT–INR = prothrombin time–international normalized ratio, TACE = transarterial chemoembolization, γ-GTP = gamma-glutamyl transpeptidase.
Table 5.
Clinical characteristics and laboratory data stratified by rehabilitation status in patients with CP-C.
| Variables | Rehabilitation (n = 20) | Non-rehabilitation (n = 30) | P value |
|---|---|---|---|
| Age | 69.0 (60.8–82.8) | 54.5 (47.3–70.5) | .003 |
| Sex | .10 | ||
| Male | 12 (60) | 25 (83.3) | |
| Female | 8 (40) | 5 (16.7) | |
| BMI (kg/m2) | 23.5 (21.3–27.3) | 22.7 (21.0–25.9) | .80 |
| Hepatocellular carcinoma | 8 (40.0) | 7 (23.3) | .23 |
| Esophageal varices | 9 (45.0) | 17 (56.7) | .56 |
| Child–Pugh score | 10.0 (10.0–13.0) | 12.0 (10.0–12.0) | .47 |
| Type of admission | .91 | ||
| Scheduled | 7 (35) | 9 (30) | |
| Emergency | 12 (60) | 18 (60) | |
| Transfer/referral | 1 (5) | 3 (10) | |
| Purpose of admission | .13 | ||
| TACE | 0 (0) | 1 (3.3) | |
| Chemotherapy | 2 (10) | 0 (0) | |
| Variceal treatment | 1 (5.0) | 5 (16.7) | |
| Liver failure treatment | 16 (80) | 24 (80) | |
| BRTO | 1 (5) | 0 (0) | |
| Other | 0 (0) | 0 (0) | |
| Length of hospital stay (d) | 25.0 (16.3–32.8) | 11.0 (8.0–19.3) | .001 |
| Length of rehabilitation (d) | 16.5 (9.8–28.0) | — | — |
| Time from hospital admission to rehabilitation referral (d) | 6.5 (3.8–9.8) | — | — |
| Discharge destination | .001 | ||
| Home | 7 (35) | 23 (76.7) | |
| Transfer/referral | 8 (40) | 1 (3.3) | |
| Death | 5 (25) | 6 (20) | |
| Blood collection upon admission | |||
| At admission | |||
| Albumin (g/dL) | 2.6 (2.3–2.8) | 2.3 (2.1–2.6) | .03 |
| Total bilirubin (mg/dL) | 2.5 (1.5–4.8) | 4.8 (2.9–10.6) | .08 |
| Creatinine (mg/dL) | 1.26 (1.05–1.84) | 1.08 (0.80–1.65) | .11 |
| Prothrombin activity (%) | 52.3 (38.4–55.6) | 42.6 (34.3–50.4) | .37 |
| eGFR (mL/min/1.73 m2) | 44.0 (28.1–51.7) | 50.1 (30.9–71.3) | .23 |
| Blood urea nitrogen (mg/dL) | 31.6 (14.7–56.9) | 18.5 (13.0–41.5) | .19 |
| PT–INR | 1.45 (1.37–1.77) | 1.63 (1.44–1.88) | .16 |
| AST (U/L) | 40.5 (30.0–86.0) | 70.0 (36.5–107.0) | .23 |
| ALT (U/L) | 30.5 (18.0–45.8) | 28.0 (19.3–52.5) | .78 |
| γ-GTP (U/L) | 38.5 (26.0–78.8) | 75.0 (35.0–155.0) | .10 |
| Platelets (×104/μL) | 9.3 (6.0–11.6) | 9.3 (5.3–11.6) | .09 |
| Ammonia (μmol/L) | 97.0 (63.3–109.3) | 82.0 (58.8–98.8) | .32 |
| MELD | 19.0 (14.5–23.5) | 18.0 (17.0–24.0) | .42 |
| Changes | |||
| Δ Albumin (g/dL) | −0.2 (−0.3 to 0.0) | 0.2 (−0.1 to 0.5) | .003 |
| Δ Total bilirubin (mg/dL) | −0.2 (−0.8 to 0.1) | 0.1 (−0.6 to 2.2) | .14 |
| Δ Creatinine (mg/dL) | −0.06 (−0.34 to 0.11) | 0.02 (−0.27 to 1.19) | .80 |
| Δ Prothrombin activity (%) | 1.4 (−5.2 to 4.8) | 1.6 (−2.9 to 7.6) | .31 |
| Δ eGFR (mL/min/1.73 m2) | −1.6 (−5.4 to 14.1) | 1.7 (−4.0 to 11.1) | .91 |
| Δ Blood urea nitrogen (mg/dL) | 0.8 (−8.2 to 10.9) | 0.3 (−6.4 to 5.0) | .94 |
| Δ PT–INR | 0.00 (−0.13 to 0.10) | −0.07 (−0.18 to 0.03) | .28 |
| Δ AST (U/L) | −6.0 (−17.5 to 6.5) | −8.0 (−23.3 to −1.0) | .06 |
| Δ ALT (U/L) | −1.5 (−10.3 to 2.8) | −1.5 (−8.8 to 4.5) | .71 |
| Δ γ-GTP (U/L) | −5.5 (−26.3 to 1.0) | 3.0 (−37.8 to 5.8) | .60 |
| Δ Platelets (×104/μL) | −0.9 (−3.3 to 0.7) | −0.3 (−2.0 to 2.8) | .58 |
| Δ Ammonia (μmol/L) | 20.5 (−46.0 to 75.5) | 12.0 (−25.8 to 35.5) | .99 |
| Δ MELD | −1.00 (−4.25 to 0.25) | 0.00 (−2.27 to 11.00) | .15 |
Values are expressed as median (interquartile range) or number (%).
P values were calculated using the Mann–Whitney U test for continuous variables and Fisher exact test for categorical variables.
Δ indicates the change from admission to discharge.
An em dash (—) indicates that the value or statistical test was not applicable.
ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, BRTO = balloon-occluded retrograde transvenous obliteration, CP = Child–Pugh classification, eGFR = estimated glomerular filtration rate, MELD = Model for End-Stage Liver Disease, PT–INR = prothrombin time–international normalized ratio, TACE = transarterial chemoembolization, γ-GTP = gamma-glutamyl transpeptidase.
Discharge to another hospital was also more frequent in the rehabilitation group (29.8% vs 1.4%, P < .001).
At admission, the rehabilitation group had lower albumin levels (2.9 vs 3.3 g/dL, P = .005), lower eGFR (48.9 vs 61.1 mL/min/1.73 m2, P = .005), and higher blood urea nitrogen levels (23.0 vs 17.2 mg/dL, P = .001).
γ-GTP was significantly lower in the rehabilitation group (40.0 vs 63.0 U/L, P = .008).
There were no significant differences in BMI, HCC prevalence, AST, ALT, total bilirubin, platelets, or ammonia.
Table 2 presents unadjusted changes in laboratory values between admission and discharge in the rehabilitation and non-rehabilitation groups. Because the rehabilitation group differed substantially at baseline, these comparisons are descriptive. The groups differed in Δprothrombin activity (−5.2% vs −2.2%, P = .007) and ΔPT–INR (+0.06 vs −0.02, P = .03); these findings cannot be attributed to rehabilitation alone.
Table 2.
Changes in blood test data during hospitalization.
| Variables | Total (n = 200) | Rehabilitation (n = 57) | Non-rehabilitation (n = 143) | P value |
|---|---|---|---|---|
| Δ Albumin (g/dL) | −0.2 (−0.5 to 0.0) | −0.3 (−0.5 to −0.1) | −0.2 (−0.5 to 0.1) | .22 |
| Δ Total bilirubin (mg/dL) | 0.1 (−0.2 to 0.4) | 0.1 (−0.3 to 0.3) | 0.1 (−0.1 to 0.4) | .06 |
| Δ Creatinine (mg/dL) | −0.02 (−0.10 to 0.08) | −0.01 (−0.12 to 0.11) | −0.02 (−0.10 to 0.07) | .75 |
| Δ Prothrombin activity (%) | −3.1 (−10.4 to 3.0) | −5.2 (−15.7 to 1.2) | −2.2 (−7.7 to 4.7) | .007 |
| Δ eGFR (mL/min/1.73 m2) | 0.8 (−4.0 to 7.9) | 0.0 (−4.0 to 11.7) | 0.8 (−4.0 to 7.0) | .43 |
| Δ Blood urea nitrogen (mg/dL) | −0.4 (−5.7 to 2.7) | −1.9 (−8.5 to 3.5) | −0.3 (−4.9 to 2.6) | .49 |
| Δ PT–INR | 0.02 (−0.04 to 0.09) | 0.06 (−0.04 to 0.12) | −0.02 (−0.10 to 0.07) | .03 |
| Δ AST (U/L) | −2.0 (−10.0 to 6.5) | −3.0 (−10.0 to 8.0) | −2.0 (−10.0 to 5.5) | .91 |
| Δ ALT (U/L) | −1.0 (−6.0 to 8.3) | −1.0 (−9.0 to 7.0) | 0.0 (−5.0 to 9.5) | .43 |
| Δ γ-GTP (U/L) | −3.0 (−14.3 to 2.0) | −3.0 (−14.0 to 3.0) | −3.0 (−14.5 to 1.5) | .75 |
| Δ Platelets (×104/μL) | −0.3 (−1.6 to 1.4) | −0.3 (−2.3 to 1.7) | −0.3 (−1.5 to 1.1) | .91 |
| Δ Ammonia (μmol/L) | 2.5 (−20.0 to 30.0) | 1.0 (−20.0 to 27.0) | 3.0 (−21.0 to 31.5) | .97 |
| Δ MELD | 0.00 (−1.00 to 1.00) | 0.00 (−1.00 to 1.00) | 0.00 (−1.00 to 1.00) | .60 |
Values are expressed as median (interquartile range).
P values were calculated using the Mann–Whitney U test.
Δ indicates the change from admission to discharge.
ALT = alanine aminotransferase, AST = aspartate aminotransferase, eGFR = estimated glomerular filtration rate, MELD = Model for End-Stage Liver Disease, PT–INR = prothrombin time–international normalized ratio, γ-GTP = gamma-glutamyl transpeptidase.
No significant differences were observed in ΔAST, ΔALT, Δalbumin, Δbilirubin, Δcreatinine, ΔeGFR, Δγ-GTP, Δplatelets, Δammonia, or ΔMELD.
3.2. Subgroup analysis by Child–Pugh classification (CP-A patients)
Table 3 shows the clinical characteristics and laboratory data of patients with CP class A, stratified by rehabilitation status. The rehabilitation group was significantly older compared with the non-rehabilitation group (80.0 vs 71.5 years, P = .001) and had a longer hospital stay (13.0 vs 9.0 days, P = .002). There were no significant differences in type or purpose of admission.
Regarding discharge outcomes, the proportion of patients discharged home was lower in the rehabilitation group (78.6% vs 97.1%, P = .05).
At admission, total bilirubin was significantly lower in the rehabilitation group (0.6 vs 0.7 mg/dL, P = .004), while no other laboratory parameters differed significantly between groups.
During hospitalization, significant between-group differences were observed in Δprothrombin activity (−15.8% vs −5.2%, P = .001), ΔPT–INR (+0.11 vs +0.03, P = .002), ΔeGFR (+6.6 vs +1.0 mL/min/1.73 m2, P = .04), and Δγ-GTP (+3.0 vs −3.0 U/L, P = .02). These estimates are unadjusted and should be interpreted cautiously in view of the small subgroup size and baseline differences.
The between-group difference in Δplatelets did not reach statistical significance (+0.8 vs −0.5 × 104/μL, P = .07).
No significant differences were observed in other parameters, including changes in albumin, total bilirubin, AST, ALT, blood urea nitrogen, ammonia, or MELD score.
3.3. Subgroup analysis by Child–Pugh classification (CP-B patients)
Table 4 shows the clinical characteristics and laboratory data of CP class B patients. The rehabilitation group was significantly older than the non-rehabilitation group (81.0 vs 70.0 years, P < .001). Although the median Child-Pugh score was 8.0 in both groups, score distributions differed (P = .006). Emergency admissions were more frequent in the rehabilitation group (43.5% vs 11.1%, P = .002), and the length of hospital stay was significantly longer (21.0 vs 9.0 days, P = .001). Discharge destination also differed between groups (P < .001), with fewer patients discharged home in the rehabilitation group (65.2% vs 100.0%) and more requiring transfer to another facility.
Table 4.
Clinical characteristics and laboratory data stratified by rehabilitation status in patients with CP-B.
| Variables | Rehabilitation (n = 23) | Non-rehabilitation (n = 45) | P value |
|---|---|---|---|
| Age | 81.0 (75.5–82.0) | 70.0 (65.0–76.0) | <.001 |
| Sex | .02 | ||
| Male | 12 (52.2) | 37 (82.2) | |
| Female | 11 (47.8) | 8 (17.8) | |
| BMI (kg/m2) | 24.6 (22.2–27.3) | 23.6 (20.7–26.8) | .86 |
| Hepatocellular carcinoma | 15 (65.2) | 31 (68.9) | .79 |
| Esophageal varices | 13 (56.5) | 31 (68.9) | .42 |
| Child–Pugh score | 8.0 (8.0–9.0) | 8.0 (7.0–8.0) | .006 |
| Type of admission | .002 | ||
| Scheduled | 12 (52.2) | 40 (88.9) | |
| Emergency | 10 (43.5) | 5 (11.1) | |
| Transfer/referral | 1 (4.3) | 0 (0) | |
| Purpose of admission | .08 | ||
| TACE | 5 (21.7) | 7 (15.6) | |
| Chemotherapy | 4 (17.4) | 13 (28.9) | |
| Variceal treatment | 4 (17.4) | 15 (33.3) | |
| Liver failure treatment | 10 (43.5) | 7 (15.6) | |
| BRTO | 0 (0) | 3 (6.7) | |
| Other | 0 (0) | 0 (0) | |
| Length of hospital stay (d) | 21.0 (14.5–27.0) | 9.0 (6.0–15.0) | .001 |
| Length of rehabilitation (d) | 16.0 (7.5–20.0) | — | — |
| Time from hospital admission to rehabilitation referral (d) | 5.0 (2.5–8.0) | — | — |
| Discharge destination | <.001 | ||
| Home | 15 (65.2) | 45 (100) | |
| Transfer/referral | 7 (30.4) | 0 (0) | |
| Death | 1 (4.3) | 0 (0) | |
| Blood collection upon admission | |||
| At admission | |||
| Albumin (g/dL) | 2.9 (2.4–3.3) | 3.0 (2.8–3.3) | .07 |
| Total bilirubin (mg/dL) | 1.0 (0.8–1.6) | 1.0 (0.8–1.6) | .96 |
| Creatinine (mg/dL) | 1.03 (0.76–1.37) | 0.87 (0.80–1.20) | .46 |
| Prothrombin activity (%) | 72.2 (61.8–80.0) | 67.8 (61.6–78.1) | .43 |
| eGFR (mL/min/1.73 m2) | 44.6 (33.1–62.8) | 64.7 (45.2–69.6) | .11 |
| Blood urea nitrogen (mg/dL) | 21.2 (17.4–32.1) | 17.1 (12.9–21.1) | .006 |
| PT–INR | 1.19 (1.13–1.31) | 1.23 (1.14–1.29) | .48 |
| AST (U/L) | 38.0 (29.0–48.5) | 36.0 (30.0–57.0) | .80 |
| ALT (U/L) | 24.0 (19.5–32.0) | 23.0 (18.0–32.0) | .99 |
| γ-GTP (U/L) | 42.0 (24.5–96.5) | 58.0 (37.0–135.0) | .23 |
| Platelets (×104/μL) | 10.8 (8.0–16.5) | 10.0 (7.0–15.1) | .50 |
| Ammonia (μmol/L) | 62.0 (52.0–87.0) | 78.5 (63.3–115.3) | .04 |
| MELD | 12.0 (8.5–14.0) | 10.0 (9.0–13.0) | .29 |
| Changes | |||
| Δ Albumin (g/dL) | −0.3 (−0.6 to 0.0) | −0.2 (−0.4 to 0.1) | .29 |
| Δ Total bilirubin (mg/dL) | 0.1 (−0.2 to 0.2) | 0.1 (−0.1 to 0.4) | .52 |
| Δ Creatinine (mg/dL) | 0.01 (−0.09 to 0.14) | 0.00 (−0.09 to 0.07) | .65 |
| Δ Prothrombin activity (%) | −5.2 (−14.0 to 0.4) | −1.6 (−5.8 to 4.8) | .04 |
| Δ eGFR (mL/min/1.73 m2) | 0.0 (−4.7 to 5.9) | 0.0 (−3.0 to 6.0) | .97 |
| Δ Blood urea nitrogen (mg/dL) | −1.5 (−9.0 to 3.5) | 0.3 (−2.1 to 2.6) | .41 |
| Δ PT–INR | 0.05 (−0.01 to 0.12) | 0.02 (−0.03 to 0.04) | .17 |
| Δ AST (U/L) | −3.0 (−9.5 to 5.0) | −2.0 (−9.0 to 6.0) | .93 |
| Δ ALT (U/L) | −4.0 (−9.0 to 1.5) | −2.0 (−7.0 to 5.0) | .29 |
| Δ γ-GTP (U/L) | −5.0 (−19.0 to 1.0) | −3.0 (−17.0 to 0.0) | .60 |
| Δ Platelets (×104/μL) | −0.3 (−1.8 to 1.6) | −0.1 (−1.3 to 1.2) | .76 |
| Δ Ammonia (μmol/L) | 0.0 (−14.0 to 17.5) | 3.0 (−33.0 to 32.0) | .47 |
| Δ MELD | 1.00 (0.00–1.50) | 0.00 (0.00–1.00) | .67 |
Values are expressed as median (interquartile range) or number (%).
P values were calculated using the Mann–Whitney U test for continuous variables and Fisher exact test for categorical variables.
Δ indicates the change from admission to discharge.
An em dash (—) indicates that the value or statistical test was not applicable.
ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, BRTO = balloon-occluded retrograde transvenous obliteration, CP = Child–Pugh classification, eGFR = estimated glomerular filtration rate, MELD = Model for End-Stage Liver Disease, PT–INR = prothrombin time–international normalized ratio, TACE = transarterial chemoembolization, γ-GTP = gamma-glutamyl transpeptidase.
At admission, blood urea nitrogen was significantly higher in the rehabilitation group (21.2 vs 17.1 mg/dL, P = .006), while ammonia was significantly lower (62.0 vs 78.5 μmol/L, P = .04). No other baseline laboratory values differed significantly.
During hospitalization, a significant between-group difference was observed in Δprothrombin activity (−5.2% vs −1.6%, P = .04). No statistically significant between-group differences were observed in ΔeGFR, Δbilirubin, Δalbumin, ΔAST, ΔALT, Δγ-GTP, or ΔMELD.
3.4. Subgroup analysis by Child–Pugh classification (CP-C patients)
Table 5 presents the CP class C subgroup analysis. The rehabilitation group was older than the non-rehabilitation group (69.0 vs 54.5 years, P = .003) and had a longer hospital stay (25.0 vs 11.0 days, P = .001). Child–Pugh score distributions did not differ significantly (10.0 [IQR, 10.0–13.0] vs 12.0 [IQR, 10.0–12.0], P = .47). Admission type (P = .91) and purpose of admission (P = .13) also did not differ significantly.
At admission, albumin was higher in the rehabilitation group (2.6 vs 2.3 g/dL, P = .03). No statistically significant between-group differences were observed in the other baseline laboratory measures, including eGFR (44.0 vs 50.1 mL/min/1.73 m2, P = .23), blood urea nitrogen (31.6 vs 18.5 mg/dL, P = .19), and ammonia (97.0 vs 82.0 μmol/L, P = .32).
During hospitalization, Δalbumin differed between groups (−0.2 [IQR, −0.3 to 0.0] in the rehabilitation group vs 0.2 [IQR, −0.1 to 0.5] in the non-rehabilitation group, P = .003). No statistically significant between-group differences were observed in the other evaluated changes, including Δprothrombin activity (1.4% vs 1.6%, P = .31), bilirubin, creatinine, eGFR, blood urea nitrogen, PT–INR, transaminases, γ-GTP, platelet count, ammonia, or MELD score.
Discharge destination differed significantly (P = .001), with fewer patients discharged home in the rehabilitation group (35.0% vs 76.7%) and more requiring transfer to another medical facility (40.0% vs 3.3%).
3.5. Changes in ALBI score
Figure 2 presents ΔALBI by rehabilitation group. In the overall cohort, the median ΔALBI was 0.24 in the rehabilitation group (n = 57) and 0.21 in the non-rehabilitation group (n = 143), with no significant difference (P = .56). Subgroup analyses stratified by CP class also showed no statistically significant between-group differences: CP class A, 0.46 vs 0.37 (rehabilitation, n = 14; non-rehabilitation, n = 68; P = .10); CP class B, 0.18 vs 0.19 (rehabilitation, n = 23; non-rehabilitation, n = 45; P = .53); and CP class C, 0.17 vs −0.19 (rehabilitation, n = 20; non-rehabilitation, n = 30; P = .053). All Child–Pugh-stratified analyses were exploratory and may have been underpowered to detect between-group differences because of the small subgroup sizes.
Figure 2.
Changes in ALBI score (ΔALBI) by rehabilitation status and Child–Pugh class. Boxes indicate medians and interquartile ranges (IQRs; whiskers omitted). (A) Overall: 0.24 (IQR, −0.03 to 0.51) versus 0.21 (IQR, −0.02 to 0.49), P = .556 (rehabilitation, n = 57; non-rehabilitation, n = 143). (B) CP class A: 0.46 (0.22–0.70) versus 0.37 (0.15–0.56), P = .102 (n = 14 vs 68). (C) CP class B: 0.18 (−0.06 to 0.42) versus 0.19 (−0.05 to 0.41), P = .526 (n = 23 vs 45). (D) CP class C: 0.17 (−0.14 to 0.32) versus −0.19 (−0.43 to 0.08), P = .053 (n = 20 vs. 30). ALBI = albumin-bilirubin score, IQR = interquartile range.
3.6. Adverse events
No serious rehabilitation-related adverse event was identified in the study cohort through review of the complete inpatient medical records or cross-checking of institutional incident reports searched from January 1, 2021 through December 31, 2024.
4. Discussion
Rehabilitation is clinically important in chronic liver disease, in which sarcopenia is associated with poor prognosis and impaired health-related quality of life.[3,4] However, real-world evidence regarding inpatient rehabilitation in medically complex patients with diverse liver disease severity remains limited. In this cohort, patients referred for rehabilitation were older, had more advanced liver disease, and had longer hospital stays. Referral and timing were determined according to the treating hepatologist’s clinical judgment, and rehabilitation was individually tailored to the patient’s condition. These referral patterns may reflect routine clinical practice, in which hepatologists identify patients perceived to have functional needs or clinical vulnerability. However, because reasons for referral and non-referral were not recorded in a standardized manner, this interpretation could not be formally examined. Accordingly, these findings describe the clinically complex patient population and routine-care context in which inpatient rehabilitation is used, rather than establishing a causal treatment effect. They highlight a high-need population for whom future prospective studies should incorporate predefined referral criteria and standardized functional assessments.
Previous studies of exercise in advanced liver disease have reported that structured exercise interventions, including aerobic and resistance components, can be feasible in selected patients.[8,9] A recent meta-analysis also reported a lower incidence of serious events with combined resistance and aerobic exercise in patients with cirrhosis.[10] However, these interventions were largely evaluated in selected outpatient or home-based populations; their findings cannot be directly extrapolated to hospitalized patients with heterogeneous clinical indications.
However, most previous studies have focused on outpatient or home-based interventions and often excluded patients with severe portal hypertension, esophageal varices, or markedly impaired hepatic reserve. Consequently, evidence regarding rehabilitation during hospitalization – when patients may be clinically unstable – has remained limited. Current guidance and expert reviews therefore emphasize the importance of careful pre-intervention assessment and close professional supervision when introducing exercise therapy in patients with liver disease.[5,11]
The present study adds descriptive real-world data by including hospitalized patients across Child–Pugh classes A–C, including patients with advanced liver dysfunction. However, the referral-based design and heterogeneity of both the underlying disease and the rehabilitation program mean that the study cannot determine whether rehabilitation itself caused the observed laboratory changes or adverse-event pattern.
In patients with CP class A, significant between-group differences were observed in Δprothrombin activity, ΔPT–INR, ΔeGFR, and Δγ-GTP. In CP class B, a difference in Δprothrombin activity was observed. These unadjusted differences may reflect periprocedural physiological stress, disease-related factors, baseline characteristics, concurrent medical treatments, or medication use, and cannot be attributed to rehabilitation alone.
Moderate exercise may transiently impair renal function in some patients with cirrhosis and ascites.[12] In the present study, renal-function changes varied across Child-Pugh classes: ΔeGFR differed in CP class A (+6.6 vs +1.0 mL/min/1.73 m2, P = .04), whereas no statistically significant difference in ΔeGFR was observed in CP classes B or C. These unadjusted findings cannot establish a renal effect of rehabilitation, because renal and hepatic measures during hospitalization may also reflect the disease course, treatment of the underlying liver condition, medication use, nutritional management, and other concurrent care.
In patients with CP class C, albumin at admission was higher in the rehabilitation group, whereas eGFR, blood urea nitrogen, and ammonia did not differ significantly. During hospitalization, Δalbumin differed between groups (−0.2 vs 0.2 g/dL, P = .003), while Δprothrombin activity did not (P = .31). These unadjusted findings may reflect baseline nutritional or metabolic status and the overall effects of inpatient medical and nutritional management rather than rehabilitation. The albumin-bilirubin (ALBI) score is an objective measure of hepatic reserve,[13] but laboratory findings should not be interpreted as evidence that rehabilitation itself was safe or effective.
4.1. Interpretation of adverse-event data
No serious rehabilitation-related event was identified through review of the complete inpatient medical records or the institutional incident-reporting system. Because intervention content, dose, and progression were individualized and were not analyzed at the session level, these observations cannot be generalized to all rehabilitation programs or used to define a safe exercise dose. Acute physiological studies in selected patients with cirrhosis and portal hypertension have shown that moderate cycling exercise can increase the hepatic venous pressure gradient.[14,15] In contrast, a small 14-week pilot trial of a supervised program combining aerobic exercise, kinesiotherapy, and nutritional therapy reported a reduction in the hepatic venous pressure gradient.[16] These considerations underscore the need for individualized clinical assessment and monitoring. The absence of reported serious events should not be interpreted as evidence of comprehensive rehabilitation safety.
4.2. Limitations
This study has several limitations. First, it was a retrospective observational study conducted at a single center and, therefore, cannot establish a causal relationship between inpatient rehabilitation and clinical outcomes. Rehabilitation referral was based on the treating physician’s judgment and the patient’s clinical condition without standardized criteria. Reasons for referral and non-referral were not systematically available, introducing substantial selection bias and possible confounding by indication. Patients referred for rehabilitation were older and had more advanced liver disease; therefore, the observed findings may reflect baseline patient characteristics and the overall course of inpatient treatment rather than the effect of rehabilitation itself.
Second, unmeasured factors may have influenced both rehabilitation referral and in-hospital outcomes, including baseline functional status, sarcopenia severity, frailty, cognitive impairment, nutritional status, patient motivation, medication use, and concurrent therapies. The analyses were unadjusted, and Child–Pugh stratification did not remove this residual confounding. The relatively small rehabilitation group (n = 57) and subgroup sizes limited statistical power and precision, particularly in the Child–Pugh-stratified analyses.
Third, rehabilitation was heterogeneous and individually tailored. Although the duration of inpatient rehabilitation and the time from hospital admission to rehabilitation referral were available, detailed session-level information on treatment frequency, minutes per session, progression, adherence, and the numbers of prescribed, completed, missed, or interrupted sessions was not available retrospectively. This limits reproducibility and prevents conclusions about a specific rehabilitation protocol or dose.
Fourth, hepatic and renal laboratory measures were selected as clinically meaningful signals of possible organ-function deterioration in medically vulnerable hospitalized patients and were interpreted as surrogate outcomes. Observed laboratory changes cannot establish the overall safety of rehabilitation. The absence of statistically significant between-group differences should not be interpreted as evidence of equivalence or as definitive evidence of rehabilitation safety, particularly given the unadjusted design and small subgroup sizes. They may reflect the overall effect of inpatient medical management, including disease-specific treatment and supportive care. The heterogeneous underlying liver diseases and admission indications may also have influenced laboratory trajectories and limit generalizability to individual disease subtypes. These measures do not assess functional status, muscle strength, mobility, exercise capacity, activities of daily living, frailty, quality of life, or long-term clinical outcomes. Accordingly, this study cannot assess rehabilitation efficacy.
Finally, although the complete inpatient medical records were reviewed and cross-checked against institutional incident reports, adverse-event ascertainment was retrospective rather than based on prospective session-by-session surveillance. Therefore, minor events that were neither recognized nor documented cannot be completely excluded. Prospective multicenter studies with standardized referral and intervention protocols, systematic adverse-event monitoring, and functional outcomes are required.
5. Conclusion
In this single-center retrospective cohort, inpatient rehabilitation was delivered to a clinically complex population of patients with chronic liver disease, including patients with advanced cirrhosis. Across Child–Pugh classes, unadjusted changes in hepatic and renal laboratory measures did not show a consistent pattern of greater deterioration in the rehabilitation group, and no serious rehabilitation-related adverse event was identified through review of the complete inpatient medical records or the institutional incident-reporting system. These observations provide real-world clinical data on inpatient rehabilitation in hospitalized patients with chronic liver disease and may help inform the further development of liver rehabilitation for medically vulnerable patients. Because referral and treatment were individualized and the comparisons were unadjusted, the findings should be interpreted as descriptive associations rather than evidence of a causal effect, overall safety, or efficacy. Prospective multicenter studies with standardized referral and intervention protocols, systematic adverse-event monitoring, and functional outcomes are warranted.
Acknowledgments
The authors thank all medical staff involved in the care of the patients included in this study. We also acknowledge the staff of the Second Department of Internal Medicine and the Department of Rehabilitation Medicine at Wakayama Medical University for their support throughout the study.
Author contributions
Conceptualization: Rikito Zaiki, Yoshiyuki Ida.
Data curation: Rikito Zaiki, Akiyuki Okamoto.
Formal analysis: Rikito Zaiki.
Investigation: Rikito Zaiki, Akiyuki Okamoto, Ryo Kanehisa.
Methodology: Rikito Zaiki, Yoshiyuki Ida, Ken Kouda.
Resources: Ryo Kanehisa.
Software: Rikito Zaiki.
Supervision: Yoshiyuki Ida, Ken Kouda, Masayuki Kitano.
Validation: Rikito Zaiki.
Visualization: Rikito Zaiki.
Writing – original draft: Rikito Zaiki.
Writing – review & editing: Rikito Zaiki, Yoshiyuki Ida, Akiyuki Okamoto, Ryo Kanehisa, Ken Kouda, Masayuki Kitano.
Abbreviations:
- ALBI
- albumin-bilirubin score
- ALT
- alanine aminotransferase
- AST
- aspartate aminotransferase
- BMI
- body mass index
- CLD
- chronic liver disease
- CP
- Child–Pugh classification
- eGFR
- estimated glomerular filtration rate
- HCC
- hepatocellular carcinoma
- IQR
- interquartile range
- MELD
- Model for End-Stage Liver Disease
- PT–INR
- prothrombin time–international normalized ratio
- γ-GTP
- gamma-glutamyl transpeptidase
This study was supported by internal funding from the Second Department of Internal Medicine and the Department of Rehabilitation Medicine at Wakayama Medical University. The funders had no role in the study design, data collection, data analysis, interpretation of the results, or preparation of the manuscript.
The study protocol was approved by the Ethics Review Committee of Wakayama Medical University (approval No. 4434) and was conducted in accordance with the Declaration of Helsinki. The requirement for individual written informed consent was waived because of the retrospective observational design. An opt-out notice was posted on the hospital website, and patients were given the opportunity to decline use of their data.
The authors have no conflicts of interest to declare.
The datasets generated during and/or analyzed during the current study are not publicly available, but are available from the corresponding author on reasonable request.
How to cite this article: Zaiki R, Ida Y, Okamoto A, Kanehisa R, Kouda K, Kitano M. Changes in hepatic functional reserve and renal function during hospitalization in patients with chronic liver disease: A retrospective observational study of inpatient rehabilitation. Medicine 2026;105:41(e51062).
Contributor Information
Rikito Zaiki, Email: zaiki@wakayama-med.ac.jp.
Akiyuki Okamoto, Email: okkamc@wakayama-med.ac.jp.
Ryo Kanehisa, Email: rkanehis@wakayama-med.ac.jp.
Ken Kouda, Email: kenkouda@wakayama-med.ac.jp.
Masayuki Kitano, Email: kitano@wakayama-med.ac.jp.
References
- [1].Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74:229–63. [DOI] [PubMed] [Google Scholar]
- [2].Estes C, Anstee QM, Arias-Loste MT, et al. Modeling NAFLD disease burden in China, France, Germany, Italy, Japan, Spain, United Kingdom, and United States for the period 2016-2030. J Hepatol. 2018;69:896–904. [DOI] [PubMed] [Google Scholar]
- [3].Tantai X, Liu Y, Yeo YH, et al. Effect of sarcopenia on survival in patients with cirrhosis: a meta-analysis. J Hepatol. 2022;76:588–99. [DOI] [PubMed] [Google Scholar]
- [4].Ando Y, Ishigami M, Ito T, et al. Sarcopenia impairs health-related quality of life in cirrhotic patients. Eur J Gastroenterol Hepatol. 2019;31:1550–6. [DOI] [PubMed] [Google Scholar]
- [5].Social Insurance Committee, The Japan Society of Hepatology. Liver rehabilitation guidance: 4. liver rehabilitation for patients with cirrhosis [in Japanese]. https://www.jsh.or.jp/medical/committeeactivity/shakaihoken/cirrhosis.html. Accessed December 23, 2025.
- [6].Zaiki R, Kamijo Y, Moriki T, et al. Dose-response rehabilitation organized by pror in out-patients with chronic cerebrovascular disorder: a single-center retrospective cohort study. J Stroke Cerebrovasc Dis. 2022;31:106375. [DOI] [PubMed] [Google Scholar]
- [7].Wiesner R, Edwards E, Freeman R, et al. ; United Network for Organ Sharing Liver Disease Severity Score Committee. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology. 2003;124:91–6. [DOI] [PubMed] [Google Scholar]
- [8].Jamali T, Raasikh T, Bustamante G, et al. Outcomes of exercise interventions in patients with advanced liver disease: a systematic review of randomized clinical trials. Am J Gastroenterol. 2022;117:1614–20. [DOI] [PubMed] [Google Scholar]
- [9].Lai JC, Dodge JL, Kappus MR, et al. A multicenter pilot randomized clinical trial of a home-based exercise program for patients with cirrhosis: the Strength Training Intervention (STRIVE). Am J Gastroenterol. 2021;116:717–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Kawaguchi T, Kawaguchi A, Hashida R, et al. Resistance exercise in combination with aerobic exercise reduces the incidence of serious events in patients with liver cirrhosis: a meta-analysis of randomized controlled trials. J Gastroenterol. 2024;59:216–28. [DOI] [PubMed] [Google Scholar]
- [11].Tandon P, Ismond KP, Riess K, et al. Exercise in cirrhosis: translating evidence and experience to practice. J Hepatol. 2018;69:1164–77. [DOI] [PubMed] [Google Scholar]
- [12].Saló J, Guevara M, Fernández-Esparrach G, et al. Impairment of renal function during moderate physical exercise in cirrhotic patients with ascites: relationship with the activity of neurohormonal systems. Hepatology. 1997;25:1338–42. [DOI] [PubMed] [Google Scholar]
- [13].Johnson PJ, Berhane S, Kagebayashi C, et al. Assessment of liver function in patients with hepatocellular carcinoma: a new evidence-based approach-the ALBI grade. J Clin Oncol. 2015;33:550–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].García-Pagán JC, Santos C, Barberá JA, et al. Physical exercise increases portal pressure in patients with cirrhosis and portal hypertension. Gastroenterology. 1996;111:1300–6. [DOI] [PubMed] [Google Scholar]
- [15].Bandi JC, García-Pagán JC, Escorsell A, et al. Effects of propranolol on the hepatic hemodynamic response to physical exercise in patients with cirrhosis. Hepatology. 1998;28:677–82. [DOI] [PubMed] [Google Scholar]
- [16].Macías-Rodríguez RU, Ilarraza-Lomelí H, Ruiz-Margáin A, et al. Changes in hepatic venous pressure gradient induced by physical exercise in cirrhosis: results of a pilot randomized open clinical trial. Clin Transl Gastroenterol. 2016;7:e180. [DOI] [PMC free article] [PubMed] [Google Scholar]