Abstract
Background:
Sorafenib, an oral multikinase inhibitor, was the first systemic therapy to demonstrate an overall survival benefit in advanced hepatocellular carcinoma (HCC). Gastrointestinal (GI) adverse events (AE), particularly diarrhea, nausea, and abdominal pain, are commonly reported toxicities and often drive dose modifications.
Objectives:
To map the prevalence of GI AE in patients with HCC receiving sorafenib.
Design:
Systematic review and meta-analysis conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.
Data sources and methods:
Clinical trials and relevant observational studies from PubMed reporting on GI side effects with sorafenib were included. Exclusion criteria were studies not involving Sorafenib, studies not involving HCC, previous meta-analyses or systematic reviews, and inaccessible publications. Prevalence of reported GI complications, including diarrhea, nausea, GI toxicity, weight loss/anorexia, constipation, and abdominal pain, were extracted from the included studies and analyzed using GraphPad Prism.
Results:
The analysis included 136 studies (178 study divisions; total 14,416 patients). Diarrhea was reported in 132 studies, with a mean prevalence of 40.44% (95% CI: 37.31–43.58) and a weighted prevalence of 42.21%. Nausea was reported in 92 studies (mean 24.75%, 95% CI: 21.19–28.28; weighted 20.35%). Abdominal pain was reported in 62 studies (mean 25.85%, 95% CI: 21.52–30.18; weighted 22.30%). GI toxicity was reported in 69 studies (mean 16.96%, 95% CI: 13.07–20.85; weighted 16.11%). Weight loss and anorexia was reported in 66 studies (mean 24.79%, 95% CI: 20.58–29.00; weighted 22.67%). Constipation was reported in 33 studies (mean 16.93%, 95% CI: 13.97–19.89; weighted 14.13%).
Conclusion:
This meta-analysis synthesizes current evidence on GI AEs with sorafenib in HCC, highlighting diarrhea as the most frequent toxicity and underscoring the need for standardized AE reporting and proactive management strategies to maintain adherence and outcomes.
Plain language summary
Sorafenib is a used to treat people with advanced liver cancer (hepatocellular carcinoma, HCC). It was the first treatment shown to help patients with liver cancer live longer. However, many people taking sorafenib experience stomach and gut-related side effects such as diarrhea, nausea, and stomach pain. These side effects can be serious enough to require lowering the dose or stopping treatment. In this study, we looked at how common these side effects are by reviewing results from 136 clinical studies that together included more than 14,000 patients treated with sorafenib. We found that diarrhea was the most commonly reported GI-related side effect, affecting about 4 out of 10 patients. Both nausea and abdominal pain occurred in about 1 in 4 patients, while weight loss affected about 1 in 5 patients. Constipation and general gut-related toxicity were less common, affecting about 1 in 6 patients. These results show that stomach and gut problems, especially diarrhea, are frequent in people treated with sorafenib. Better ways to monitor and manage these side effects are needed to help patients continue treatment and get the most benefit from it.
Keywords
Introduction
Sorafenib, an oral multikinase inhibitor, was the first systemic therapy to demonstrate an overall survival benefit in patients with advanced hepatocellular carcinoma (HCC) and subsequently established itself as the global standard of care for over a decade. 1 By targeting multiple signaling pathways involved in tumor proliferation and angiogenesis, including RAF, vascular endothelial growth factor receptor (VEGFR), and platelet-derived growth factor receptors (PDGFR), sorafenib slows tumor growth and progression. 2 Unlike locoregional therapies such as transarterial chemoembolization (TACE), sorafenib acts systemically, which contributes to a distinct and broader spectrum of adverse events (AEs). 3
Among these common toxicities are upper and lower gastrointestinal (GI) disturbances, which can substantially impact treatment adherence and quality of life.4–7 These symptoms include diarrhea, nausea, vomiting, abdominal pain,6,8 and in some cases gastrointestinal bleeding or mucositis.9–11 These AEs range from mild and self-limiting to severe, often requiring dose reduction or treatment discontinuation.3–5,7 Diarrhea is particularly notable, frequently reported as a GI side effect and a common reason for sorafenib dose modification.6,12
The mechanisms behind sorafenib-related GI toxicity are likely multifactorial.8,13 Inhibition of VEGFRs and PDGFRs can disrupt angiogenesis and reduce mucosal perfusion, impairing epithelial repair and promoting ischemic or inflammatory injury to the gastrointestinal lining. Inhibition of RAF kinases may also interfere with epithelial cell proliferation and turnover, further compromising mucosal barrier function. In addition, VEGF pathway inhibition has been implicated in altered intestinal permeability and dysregulated fluid transport, which may contribute to diarrhea and abdominal discomfort observed in treated patients. 14 These mechanisms likely act concurrently, leading to impaired absorption, increased luminal secretion, and heightened susceptibility to mucosal injury. Emerging evidence also suggests that tyrosine kinase inhibitors, including sorafenib, may induce alterations in the gut microbiome, potentially exacerbating mucosal inflammation and gastrointestinal symptoms through dysbiosis-related immune and metabolic effects.8,15 Nausea and abdominal pain may reflect a combination of systemic drug exposure, local mucosal irritation, and treatment-related metabolic disturbances.11,16 Patient-specific factors, including baseline hepatic function, comorbid conditions, and prior therapeutic exposures may further increase susceptibility to these AEs. 8 Despite their frequency and clinical importance, gastrointestinal complications associated with cancer therapies are inconsistently reported across studies, resulting in substantial heterogeneity in outcome assessment. 11 This heterogeneity limits cross-study comparability and complicates the development of standardized management strategies.
The aim of this meta-analysis is to comprehensively map the prevalence of gastrointestinal complications in HCC patients treated with sorafenib, focusing on diarrhea, nausea, abdominal pain, weight loss, constipation, and GI toxicity (including mucositis, intestinal bleeding, and ulcerations). Although gastrointestinal toxicities associated with sorafenib are well recognized and summarized in the prescribing information, these data are largely derived from individual clinical trials with heterogeneous designs, patient populations, and reporting standards. The rationale for the present meta-analysis was not to establish the existence of these AEs, rather to provide a systematic, pooled quantification of their incidence across diverse study settings, thereby offering more precise and generalizable estimates than those available from single trials. Pooled estimates are particularly relevant given evolving treatment paradigms, differences between trial populations and real-world practice, and variability in the reporting of gastrointestinal AEs across studies.
Importantly, quantifying the prevalence and distribution of gastrointestinal toxicities at the meta-analytic level has implications beyond routine clinical awareness. This data can inform anticipatory counseling, risk–benefit discussions, and the design of monitoring and supportive care strategies, particularly for patients at higher risk of treatment intolerance. From a research perspective, the findings highlight inconsistencies in AE reporting and underscore the need for standardized toxicity endpoints and time-to-onset analyses in future trials. By synthesizing available evidence in this way, this work seeks to provide a clearer understanding of the burden of these side effects, highlight gaps in knowledge, and inform clinical practice to improve patient management and outcomes.
Methods
Eligibility criteria
Studies were considered eligible if they reported gastrointestinal AEs in patients with HCC treated with sorafenib, specifically diarrhea, nausea, abdominal pain, anorexia/weight loss, constipation, or gastrointestinal toxicities (mucositis, ulceration, hemorrhage). Only clinical trials (including both nonrandomized and randomized controlled trials (RCTs)), and prospective or retrospective studies were included. Exclusion criteria were absence of GI AE reporting, nonsorafenib studies, meta-analyses, systematic reviews, case reports, secondary analyses of prior trials, and inaccessible publications. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed (Figure 1, Supplemental Table 1). The review protocol was not registered.
PRISMA diagram illustrating the progression of information through the phases of the systematic review.
Information sources and search strategy
The literature search was performed in PubMed using the terms “sorafenib” and “hepatocellular carcinoma,” limited to clinical trials and relevant observational studies. Data extraction was conducted by six independent reviewers, with each entry verified by two others; therefore, all data points were independently assessed by three reviewers. All methodological procedures were predefined prior to data collection.
Data collection
Data extraction included study identifier (PMID), author, publication year, and study design, as well as geographical location. Participant demographics included sample size and median age. Treatment variables recorded were sorafenib dose, treatment duration, and the occurrence of dose reductions or interruptions. The primary outcomes were the prevalence of gastrointestinal AEs—diarrhea, nausea, abdominal pain, anorexia/weight loss, constipation, and gastrointestinal toxicities (mucositis, ulceration, gastrointestinal hemorrhage)—reported as percentages. AEs not assessed or reported in a study were coded as missing rather than zero and excluded from prevalence estimates to minimize bias. Study quality and risk of bias were assessed using the Mixed Methods Appraisal Tool (MMAT). 17 Following the MMAT framework, studies were first classified by study design (RCTs, nonrandomized quantitative studies, or descriptive quantitative studies) and then appraised using the five design-specific MMAT criteria. Assessments were performed independently by three reviewers, with each MMAT criterion rated as “Yes,” “No,” or “Can’t tell,” and discrepancies resolved by majority decision (2/3 agreement). Results are presented as criterion-level ratings in accordance with MMAT standards (Supplemental Table 2), without calculating an overall numeric score.
Statistical analyses
Data were managed in Microsoft Excel and imported into GraphPad Prism version 10.2.3 for analysis. Prevalence of each gastrointestinal AE was summarized across study arms using both mean prevalence (±standard error of the mean (SEM)) and weighted prevalence based on sample size. Bar charts were plotted to show mean prevalences values within the included studies, with error bars indicating SEM. To explore potential causes of variability among study results, subgroup analyses were conducted (e.g., by sorafenib dose, treatment setting (monotherapy vs combination), and type of combination regimen). Unpaired Student’s t-tests were used to assess between-group comparisons and a two-sided p value < 0.05 was considered statistically significant. Principal component analysis (PCA) was performed on standardized study-level variables (z-scored; mean = 0, SD = 1) to explore clustering patterns. Components were retained using the Kaiser criterion (eigenvalues > 1). No variance-based fixed- or random-effects meta-analysis model was applied.
Results
Study selection
A systematic PubMed search was performed to identify studies reporting GI AEs in patients with HCC treated with sorafenib. The search strategy combined keywords and MeSH terms for “sorafenib” and “hepatocellular carcinoma,” covering publications from 2005 to 2025. Filters were applied to include clinical trials and RCTs, while excluding systematic reviews, meta-analyses, study protocols, and case reports.
The initial search identified 368 records (Supplemental Table 3). All titles and abstracts were screened according to the predefined inclusion and exclusion criteria. Studies were excluded if they were unrelated to sorafenib treatment, did not focus on HCC, or lacked data on gastrointestinal AEs. Following full-text review, 85 studies were excluded for not involving sorafenib, 62 for not reporting GI AEs, 60 as duplicate analyses of previously published trials, and 25 for other reasons, including inaccessibility, language limitations, or ineligible study design (Supplemental Table 4). A total of 136 studies met inclusion criteria and were retained for final analysis (Table 1 and Supplemental Table 5). Study quality was evaluated using the MMAT tool, with detailed ratings presented in Supplemental Table 2.
Included studies.
Characteristics of included studies
The included clinical studies showed heterogeneity in design, geographical location, and sample size. Most (62 studies) were RCTs, including Phase III trials, followed by 39 Phase II trials and 3 prospective cohort studies. Studies including multiple treatments were divided per treatment arm to allow assessment of differences in sorafenib treatment settings. Altogether, the included studies covered 14,416 patients, with individual study sizes ranging from 4 participants 100 to 717 participants 124 (Figure 2(a)). The geographical distribution spanned 20 countries, with the majority of studies conducted internationally across multiple regions, followed by single-country studies predominantly from China and Japan (Figure 2(b)).
Overview of included studies and treatment modalities. (a) Distribution of sample sizes among studies investigating Sorafenib treatment. The bar chart presents the sample size for each study. (b) Geographic distribution of the studies, shown as a donut chart. The chart depicts the proportion of studies conducted in different countries, with multinational, China, and Japan representing the largest contributions. (c) Proportion of studies using Sorafenib alone versus in combination with another treatment modality. (d) Proportion of studies investigating Sorafenib at a starting dose of 800 mg versus another dosage.
This systematic review included studies evaluating diverse sorafenib treatment modalities. Particular focus was given to studies comparing different dosing regimens (Figure 2(c)) or assessing sorafenib as monotherapy versus combination therapy with systemic or locoregional interventions (Figure 2(d)). Each treatment arm including sorafenib was analyzed independently, yielding a total of 178 study arms. Among these, 83% reported standard-dose sorafenib (800 mg daily), while 17% investigated alternative dosing strategies.28,30,40,45,53,110,128,132,140 Sorafenib was administered in combination with other treatments in 41% of study arms and as monotherapy in 69%. The most common combination regimen involved TACE.7,18,24,28,39,53,64,69,71,87,89,92,103,107–109,113,115,118–120,129,130,132,133,143,147,152
Prevalence and reporting of gastrointestinal side-effects
In our meta-analysis of 136 studies (178 study arms; 14,416 patients), we found variability in GI side effect reporting and documentation (Figure 3(a) and Table 1). Diarrhea was the most consistently reported symptom, documented in 132 studies (174 study arms; 6053 cases), corresponding to 97.8% of all included studies. The mean prevalence of diarrhea was 40.44% (SEM 1.59; 95% CI: 37.31–43.58), with a weighted prevalence of 42.21%. Nausea was reported in 92 studies (118 study arms; 2298 cases), representing 66.3% of studies, with a mean prevalence of 24.75% (SEM 1.79; 95% CI: 21.19–28.28) and a weighted prevalence of 20.35%. Abdominal pain was reported in 62 studies (81 study arms; 2105 cases), accounting for 45.5% of studies, with a mean prevalence of 25.85% (SEM 2.18; 95% CI: 21.52–30.18) and a weighted prevalence of 22.30%. GI toxicity, including ulcerations, intestinal hemorrhages, and mucositis, was documented in 69 studies (85 study arms; 867 cases), corresponding to 47.8% of studies, with a mean prevalence of 16.96% (SEM 1.95; 95% CI: 13.07–20.85) and a weighted prevalence of 16.11%. Weight loss or anorexia was reported in 66 studies (84 study arms; 2047 cases), representing 47.2% of studies, with a mean prevalence of 24.79% (SEM 2.12; 95% CI: 20.58–29.00) and a weighted prevalence of 22.67%. Finally, constipation was reported in 33 studies (47 study arms; 813 cases), corresponding to 26.4% of studies, with a mean prevalence of 16.93% (SEM 1.47; 95% CI: 13.97–19.89) and a weighted prevalence of 14.13%. When stratified by Barcelona-Clinic Liver Cancer (BCLC) stage, the majority of patients across the included trials were in stage C, consistent with the clinical use of sorafenib in intermediate to advanced HCC. 1
Documentation and prevalence of GI side effects and BCLC staging. (a) Percentage of included studies reporting each GI-related adverse effect. (b) Percentage of total patients reporting each GI-related adverse effect. (c) Percentage of total patients by BCLC stage.
Diarrhea
Our meta-analysis identified diarrhea as the most frequently reported gastrointestinal side effect of sorafenib, documented in 132 studies (174 study arms; 6053 cases; Table 2). The mean prevalence across these studies was 40.44% (SEM 1.59; 95% CI: 37.31–43.58), with a weighted prevalence of 42.21% (Figure 3(b)). When stratified by daily dose, patients receiving standard-dose sorafenib (800 mg/day) experienced diarrhea at similar rates compared with those receiving reduced doses (400 mg/day), although the scatter plots suggest considerable inter-study variability (Figure 4(a) and (b)). No clear association was observed between diarrhea prevalence and stage distribution; trials with higher or lower rates of diarrhea included similar proportions of BCLC B and C patients. Stage A patients were rarely represented, and when present, diarrhea rates did not differ markedly from trials dominated by stage B or C cohorts. In treatment setting comparisons, sorafenib monotherapy and combination therapies both showed substantial rates of diarrhea, with no differences in combination regimens (Figure 4(c)). Among combination strategies, the highest proportions of diarrhea were observed when sorafenib was combined with resminostat, TACE, and tigatuzumab54,110,117 (Figure 4(d)).
Prevalence of gastrointestinal symptoms in HCC-patients receiving Sorafenib.
GI, gastrointestinal; SEM, standard error of the mean.
Prevalence of diarrhea in patients undergoing Sorafenib treatment. (a) Comparison of percentage of patients experiencing diarrhea between those treated with 400 mg daily dose Sorafenib versus 800 mg daily dose. (b) Prevalence of patients experiencing diarrhea, cross-examined with prevalence of patients within each Barcelona-Clinic Liver Cancer classification. (c) Comparison of percentage of patients experiencing diarrhea between those treated with Sorafenib alone versus Sorafenib combined with another treatment modality. (d) Bar chart depicting the prevalence of patients experiencing diarrhea while undergoing Sorafenib and a combinational treatment, categorized by combinational treatment modality.
Nausea
Nausea was the second most frequently reported gastrointestinal side effect in sorafenib-treated patients, documented in 92 studies (118 study arms; 2298 cases; Table 2). The mean prevalence across studies was 24.75% (SEM 1.79; 95% CI: 21.19–28.28), with a weighted prevalence of 20.35% (Figure 3(b)). When stratified by daily dose, patients receiving standard-dose sorafenib (800 mg/day) showed nausea rates comparable to those receiving reduced doses (400 mg/day), although inter-study variability was evident (Figure 5(a)). Scatter plots stratified by BCLC stage showed that most trials primarily included stage B and C patients, as expected for sorafenib therapy. No clear correlation was observed between the proportion of patients at a given stage and the prevalence of nausea in each trial (Figure 5(b)). When comparing treatment settings, nausea was significantly more common in combination regimens than in sorafenib monotherapy (Figure 5(c)). Among the combination strategies, higher rates of nausea were observed when sorafenib was combined with TACE, resminostat, or refametinib, with mean prevalence exceeding 40% in these studies (Figure 5(d)).54,58,110,122,127 In contrast, combinations with HAIC-based regimens or radioembolization reported lower prevalence rates, generally below 20%.
Prevalence of nausea in patients undergoing Sorafenib treatment. (a) Comparison of percentage of patients experiencing nausea between those treated with 400 mg daily dose Sorafenib versus 800 mg daily dose. (b) Prevalence of patients experiencing nausea, cross-examined with prevalence of patients within each Barcelona-Clinic Liver Cancer classification. (c) Comparison of percentage of patients experiencing nausea between those treated with Sorafenib alone versus Sorafenib combined with another treatment modality. (d) Bar chart depicting the prevalence of patients experiencing nausea while undergoing Sorafenib and a combinational treatment, categorized by combinational treatment modality.
GI toxicity
GI toxicity, defined as gastrointestinal ulcerations, hemorrhages, or mucositis, was reported in 69 studies (85 study arms; 867 cases; Table 2). The mean prevalence across these studies was 16.96% (SEM 1.95; 95% CI: 13.07–20.85), with a weighted prevalence of 16.11% (Figure 3(b)). When stratified by daily dose, there was no clear difference in the prevalence of GI toxicity between patients receiving standard-dose sorafenib (800 mg/day) and those on reduced doses (400 mg/day), although the overall prevalence remained relatively low compared with other GI AEs (Figure 6(a)). Scatter plots stratified by BCLC stage indicated that most trials included predominantly stage B and C patients, with no obvious correlation between the proportion of patients at each stage and reported prevalence of GI toxicity (Figure 6(b)). Comparisons between treatment settings revealed that GI toxicity was significantly more common in combination regimens compared to sorafenib monotherapy (Figure 6(c)). Among the combination strategies, the highest rates of GI toxicity were observed with sorafenib combined with 5-FU,55,73,106 where prevalence exceeded 40% in some studies (Figure 6(d)). Other combinations, including TACE, MTL-CEBPA, 37 and galunisertib,56,151 showed intermediate prevalence rates while DEB-TACE combinations were associated with the lowest rates.
Prevalence of GI toxicity in patients undergoing Sorafenib treatment. (a) Comparison of percentage of patients experiencing GI toxicity between those treated with 400 mg daily dose Sorafenib versus 800 mg daily dose. (b) Prevalence of patients experiencing GI toxicity, cross-examined with prevalence of patients within each Barcelona-Clinic Liver Cancer classification. (c) Comparison of percentage of patients experiencing GI toxicity between those treated with Sorafenib alone versus Sorafenib combined with another treatment modality. (d) Bar chart depicting the prevalence of patients experiencing GI toxicity while undergoing Sorafenib and a combinational treatment, categorized by combinational treatment modality.
Abdominal pain
Abdominal pain was reported in 62 studies (81 study arms; 2105 cases; Table 2). The mean prevalence across these studies was 25.85% (SEM 2.18; 95% CI: 21.52–30.18), with a weighted prevalence of 22.30% (Figure 3(b)). When stratified by daily dose, abdominal pain was observed at comparable frequencies in patients receiving standard-dose sorafenib (800 mg/day) and those on reduced doses (400 mg/day), with wide variability between trials (Figure 7(a)). Scatter plots stratified by BCLC stage showed that most trials enrolled predominantly stage B and C patients, with no apparent correlation between stage distribution and abdominal pain prevalence (Figure 7(b)). Comparisons between treatment settings demonstrated that abdominal pain was significantly more common in combination regimens than in sorafenib monotherapy (Figure 7(c)). Among combination strategies, the highest rates were observed in studies combining sorafenib with TACE, where mean prevalence exceeded 40% (Figure 7(d)). Elevated prevalence was also reported in combinations with refametinib and tigatuzumab,58,117,122,127 while combinations with DEB-TACE87,113 or MTL-CEBPA 37 were associated with lower prevalence.
Prevalence of abdominal pain in patients undergoing Sorafenib treatment. (a) Comparison of percentage of patients experiencing abdominal pain between those treated with 400 mg daily dose Sorafenib versus 800 mg daily dose. (b) Prevalence of patients experiencing abdominal pain, cross-examined with prevalence of patients within each Barcelona-Clinic Liver Cancer classification. (c) Comparison of percentage of patients experiencing abdominal pain between those treated with Sorafenib alone versus Sorafenib combined with another treatment modality. (d) Bar chart depicting the prevalence of patients experiencing abdominal pain while undergoing Sorafenib and a combinational treatment, categorized by combinational treatment modality.
Weight loss or anorexia
Weight loss or anorexia was reported in 66 studies (84 study arms; 2047 cases; Table 2). The mean prevalence across studies was 24.79% (SEM 2.12; 95% CI: 20.58–29.00), with a weighted prevalence of 22.67% (Figure 3(b)). When stratified by daily dose, prevalence of weight loss and anorexia was similar in patients receiving standard-dose sorafenib (800 mg/day) and those on reduced doses (400 mg/day), though variability between studies was considerable (Figure 8(a)). Scatter plots stratified by BCLC stage showed that most trials enrolled predominantly stage B and C patients, with no apparent correlation between stage distribution and abdominal pain prevalence (Figure 8(b)). Unlike several other gastrointestinal side effects, no significant difference was observed between sorafenib monotherapy and combination regimens (Figure 8(c)). Within combination arms, prevalence varied according to the partner treatment: higher rates were observed with galunisertib, TACE, and resminostat, while lower rates were reported in combinations with DEB-TACE, HAIC, or MTL-CEBPA (Figure 8(d)).
Prevalence of weight loss/anorexia in patients undergoing Sorafenib treatment. (a) Comparison of percentage of patients experiencing weight loss/anorexia between those treated with 400 mg daily dose Sorafenib versus 800 mg daily dose. (b) Prevalence of patients experiencing weight loss/anorexia, cross examined with prevalence of patients within each Barcelona-Clinic Liver Cancer classification. (c) Comparison of percentage of patients experiencing weight loss/anorexia between those treated with Sorafenib alone versus Sorafenib combined with another treatment modality. (d) Bar chart depicting the prevalence of patients experiencing weight loss/anorexia while undergoing Sorafenib and a combinational treatment, categorized by combinational treatment modality.
Constipation
Constipation was the least frequently reported gastrointestinal side effect in sorafenib-treated patients, documented in 33 studies (47 study arms; 813 cases; Table 2). The mean prevalence across studies was 16.93% (SEM 1.47; 95% CI: 13.97–19.89), with a weighted prevalence of 14.13% (Figure 3(b)). When stratified by daily dose, interpretation was limited because only one study using a reduced dose of sorafenib (400 mg/day) reported constipation. As such, no reliable conclusions can be drawn about dose–response patterns, and the overall prevalence estimates are primarily driven by standard-dose sorafenib studies (Figure 9(a)). Scatter plots stratified by BCLC stage suggested a slight tendency toward higher prevalence in trials enrolling larger proportions of stage C patients, although no strong correlation was evident (Figure 9(b)). In contrast to diarrhea and weight loss, constipation was significantly more common in combination regimens compared with sorafenib monotherapy (Figure 9(c)). Among combination strategies, constipation was most frequently reported with tigatuzumab and DEB-TACE,87,113,117 while lower rates were observed in sorafenib combined with TACE (Figure 9(d)).
Prevalence of constipation in patients undergoing sorafenib treatment. (a) Comparison of percentage of patients experiencing constipation between those treated with 400 mg daily dose sorafenib versus 800 mg daily dose. (b) Prevalence of patients experiencing constipation, cross-examined with prevalence of patients within each Barcelona-Clinic Liver Cancer classification. (c) Comparison of percentage of patients experiencing constipation between those treated with sorafenib alone versus sorafenib combined with another treatment modality. (d) Bar chart depicting the prevalence of patients experiencing constipation while undergoing sorafenib and a combinational treatment, categorized by combinational treatment modality.
Overall GI-adverse effects
To further identify the relationships between gastrointestinal side effects, we performed PCA across all study arms. The first two components explained 71.96% of the variance (Figure 10(a)). The biplot revealed that diarrhea, weight loss/anorexia, and GI toxicity clustered together, indicating partial overlap in reporting and prevalence patterns across studies (Figure 10(b)). In contrast, nausea and abdominal pain loaded more independently, suggesting they may represent distinct AE profiles.
PCA and study-level prevalence of gastrointestinal toxicities in sorafenib based treatments. (a) Screen plot showing the proportion of variance explained by each principal component. PC1 and PC2 together capture ~70% of the variability across study arms. (b) PCA biplot of the first two components, with loadings (purple) and study arms (gray). GI toxicity, diarrhea, nausea, abdominal pain, and weight loss cluster along PC1, indicating they are the major drivers of variation. Constipation and proportion of stage C patients load separately along PC2. (c–g) Bubble plots showing the relationship between overall GI toxicity (x-axis) and other adverse events (y-axis), including weight loss (c), diarrhea (d), constipation (e), abdominal pain (f), and nausea (g). Each dot represents a study arm, with bubble size proportional to sample size and color indicating sorafenib monotherapy (blue) or combination therapy (purple).
Dimensionality reduction via PCA was applied to identify clusters within the dataset based on treatment modalities (Figure 10(c)–(g)). When stratifying bubble plots by treatment setting, combination regimens (purple) were consistently associated with greater heterogeneity in GI toxicity prevalence compared with sorafenib monotherapy (blue; Figure 10(c)–(h)). Higher GI toxicity prevalence tended to coincide with increased rates of diarrhea and anorexia/weight loss (Figure 10(c) and (d)), aligning with the clustering patterns observed in the PCA. For nausea, sorafenib monotherapy studies clustered at low prevalence, while combination regimens were more dispersed, with some reporting concurrent elevations of nausea and GI toxicity (Figure 10(e)).
Discussion
This meta-analysis of 136 clinical trials including 14,416 patients provides a comprehensive summary of GI AE associated with sorafenib treatment in HCC. Our results confirm that GI side effects are frequent, with diarrhea as the most consistently reported symptom, affecting over 40% of patients, similar to what has been noted in other studies.4,5,12 Nausea, abdominal pain, weight loss/anorexia, GI toxicity (including ulcerations and hemorrhage), and constipation were also observed at meaningful prevalence levels, though with considerable variability across studies. Importantly, adverse effect profiles differed between monotherapy and combination regimens, and multivariate analyses highlighted symptom clusters that may reflect shared pathophysiological mechanisms.
When compared with our previous meta-analysis of GI adverse effects in TACE, 11 sorafenib demonstrates a different side effect profile. While abdominal pain was the most common complication of TACE (reflecting the embolization procedure itself), sorafenib treatment was dominated by systemic toxicities, with diarrhea as the leading symptom. 11 Nausea and anorexia/weight loss were also more prominent in sorafenib, whereas GI ulcerations and hemorrhage, though less frequent, carry potential clinical significance due to their impact on patient safety and treatment discontinuation. These differences underscore the distinct mechanisms underlying locoregional and systemic treatments: ischemia-driven postembolization syndromes in TACE11,154 versus off-target kinase inhibition and epithelial stress in sorafenib.11,13,16
Our stratified analyses revealed that most gastrointestinal side effects occurred at similar rates in standard-dose (800 mg/day) and reduced-dose (400 mg/day) sorafenib, although inter-study variability was wide. This finding suggests that toxicity may not increase in a strictly dose-dependent manner, but instead be influenced by patient-specific susceptibility, duration of exposure, and the use of supportive care interventions. Notably, fewer studies were conducted with a reduced-dose, raising the possibility of reporting bias in more complex trial settings. However, real-world data support the clinical relevance of dose adjustments. A recent retrospective evaluated 80 patients treated with reduced-dose sorafenib (400 mg/day) and reported a clinical benefit rate of 56%, with median progression-free survival of 3.7 months and overall survival of 5.3 months. 155 Grade 3 toxicities occurred in only 11% of patients, and just 7.5% discontinued treatment due to AEs despite the lower dose. 155 Similarly, a large Veterans Health Administration cohort of 4903 patients demonstrated that reduced starting doses (<800 mg/day) were not associated with inferior overall survival after propensity matching, but were linked to lower discontinuation rates due to gastrointestinal toxicity and reduced cumulative treatment costs. 156 Together, these findings support dose reduction as a clinically reasonable option in selected patients, particularly those with poorer baseline liver function, higher comorbidity burden, or limited access to alternative therapies.
Combination therapies consistently showed higher and more variable prevalence of several adverse effects compared with monotherapy. Diarrhea and nausea, for example, were particularly common in regimens combining sorafenib with TACE, resminostat, or refametinib. In contrast, combinations with DEB-TACE or HAIC were associated with lower rates, suggesting that treatment-specific interactions may shape the GI toxicity profile. These findings highlight the importance of carefully selecting drug combination and monitoring additive toxicities, particularly in patients with limited hepatic reserve.
PCA and scatter plot analyses revealed that gastrointestinal symptoms did not occur independently but instead grouped into distinct patterns. Diarrhea, weight loss/anorexia, and GI toxicity clustered together, suggesting a shared pathway, possibly related to mucosal injury, epithelial turnover, and altered nutrient absorption. In contrast, nausea and abdominal pain loaded more independently, reflecting different mechanisms such as central nervous system effects or treatment-related abdominal stress. Constipation emerged as a distinct phenomenon, largely confined to combination settings.
These patterns provide insights into the pathophysiology of sorafenib-induced toxicity and may guide targeted management strategies. For instance, clustering of diarrhea with weight loss and GI toxicity suggests that addressing diarrhea proactively may also mitigate downstream nutritional compromise and mucosal injury. Conversely, nausea and abdominal pain may require separate supportive interventions.
From a clinical perspective, our findings emphasize the need for systematic monitoring and proactive management of GI AEs in patients receiving sorafenib. Diarrhea and anorexia, in particular, have major implications for patient quality of life, nutritional status, treatment adherence, and overall clinical outcomes, and may necessitate dose modification or treatment interruption if inadequately managed. Early recognition through structured monitoring protocols may allow early supportive interventions—this, in turn, may reduce symptom severity and prevent downstream complications such as dehydration, weight loss, and overall functional health decline.
While diarrhea was the most frequent gastrointestinal AE in our meta-analysis, increasing evidence suggests that not all events are clinically equivalent. Beyond its base symptomatic burden, the timing of diarrhea onset may carry some prognostic significance. In a cohort of 344 patients, Díaz-González et al. reported that early diarrhea (within 60 days of treatment initiation) was associated with significantly worse survival and may serve as a clinical signal of suboptimal therapeutic benefit, prompting consideration of transition into second-line therapies. These observations highlight diarrhea not only as more than a toxicity requiring management in and of itself but also as a potential biomarker of treatment nonresponse. This distinction is particularly relevant in modern clinical settings, where multiple effective alternative agents are available and early treatment optimization may meaningfully influence long-term patient outcomes. 12
From a research perspective, the study highlights gaps in the reporting of GI side effects. Although diarrhea was nearly universally documented, other important symptoms such as constipation and anorexia were inconsistently reported, potentially underestimating their clinical impact. Standardization of AE reporting in HCC trials remains an urgent need. Furthermore, the observed differences between monotherapy and combination regimens underscore the importance of including robust safety endpoints in trials testing sorafenib in multimodality contexts.
This meta-analysis has several limitations. First, heterogeneity across studies in design, patient populations, and reporting standards complicates direct comparisons. In particular, grading of AEs (e.g., CTCAE severity categories) was inconsistently reported and subject to discrepancies in definitions and reporting thresholds, limiting the feasibility of comparative analyses of toxicity severity; therefore, we focused on toxicity prevalence as the most consistently reported outcome across trials. Second, selective reporting remains a challenge: some adverse effects may be underrepresented if not actively monitored. Third, the lack of individual patient-level data prevents stratification by duration of treatment, prior therapies, or liver function, which may influence toxicity profiles. Finally, prevalence outcomes were summarized using mean and weighted prevalence estimates across study arms, but no variance-based random-effects meta-analysis was performed; therefore, I² was not calculated. Also, the reliance on published studies introduces the possibility of publication bias, particularly for underreported side effects.
Although this study primarily reports the prevalence of gastrointestinal AEs associated with sorafenib, these findings have several implications for clinical practice and future research. The high frequency of these toxicities supports the need for routine and systematic monitoring, as well as timely supportive care and dose adjustment when indicated, to limit treatment-related morbidity and preserve treatment continuity. In the research setting, the results underscore the value of standardized AE reporting, inclusion of time-to-onset analyses, and integration of patient-reported outcomes to improve the interpretability and clinical relevance of safety data. Taken together, these considerations provide a rationale for extending descriptive prevalence data toward more structured approaches to toxicity management and study design.
In summary, gastrointestinal side effects are common and clinically significant in sorafenib-treated HCC patients, with diarrhea as the predominant symptom, followed by nausea, abdominal pain, and anorexia/weight loss. Multivariate analyses reveal clustering of diarrhea, weight loss, and GI toxicity, while constipation appears confined to combination regimens. Compared with TACE, sorafenib demonstrates a distinct toxicity profile reflective of systemic rather than locoregional mechanisms. These findings underline the importance of proactive GI toxicity management, standardized AE reporting, and careful evaluation of combination regimens to optimize patient outcomes.
Conclusion
In conclusion, this meta-analysis synthesizes the available evidence on gastrointestinal AEs associated with sorafenib in patients with HCC, confirming diarrhea as the most frequently reported toxicity. Beyond quantifying incidence, these findings highlight the need for consistent AE reporting and the implementation of systematic, proactive management strategies aimed at minimizing treatment-related morbidity, preserving treatment adherence, and ultimately optimizing clinical outcomes.
Supplemental Material
sj-docx-5-tam-10.1177_17588359261442686 – Supplemental material for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients
Supplemental material, sj-docx-5-tam-10.1177_17588359261442686 for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients by Nathalie Arendt, Tania Payo-Serafín, Anna Kolosenko, Maria Kopsida, Markus Sjöblom and Femke Heindryckx in Therapeutic Advances in Medical Oncology
Supplemental Material
sj-pdf-1-tam-10.1177_17588359261442686 – Supplemental material for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients
Supplemental material, sj-pdf-1-tam-10.1177_17588359261442686 for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients by Nathalie Arendt, Tania Payo-Serafín, Anna Kolosenko, Maria Kopsida, Markus Sjöblom and Femke Heindryckx in Therapeutic Advances in Medical Oncology
Supplemental Material
sj-pdf-2-tam-10.1177_17588359261442686 – Supplemental material for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients
Supplemental material, sj-pdf-2-tam-10.1177_17588359261442686 for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients by Nathalie Arendt, Tania Payo-Serafín, Anna Kolosenko, Maria Kopsida, Markus Sjöblom and Femke Heindryckx in Therapeutic Advances in Medical Oncology
Supplemental Material
sj-pdf-3-tam-10.1177_17588359261442686 – Supplemental material for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients
Supplemental material, sj-pdf-3-tam-10.1177_17588359261442686 for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients by Nathalie Arendt, Tania Payo-Serafín, Anna Kolosenko, Maria Kopsida, Markus Sjöblom and Femke Heindryckx in Therapeutic Advances in Medical Oncology
Supplemental Material
sj-pdf-4-tam-10.1177_17588359261442686 – Supplemental material for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients
Supplemental material, sj-pdf-4-tam-10.1177_17588359261442686 for Prevalence of gastrointestinal side effects in hepatocellular carcinoma patients receiving sorafenib: a meta-analysis of 136 studies and 14,416 patients by Nathalie Arendt, Tania Payo-Serafín, Anna Kolosenko, Maria Kopsida, Markus Sjöblom and Femke Heindryckx in Therapeutic Advances in Medical Oncology
Footnotes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
