Who Is at Risk of Liver Injury from Immune Checkpoint Inhibitors? A Comprehensive Review for Cancer Patients

Table of Contents

• Key Points
• Background: Why This Research Matters
• How the Research Was Conducted
• Study Selection and Patient Characteristics
• Quality Assessment of the Included Studies
• Key Findings: Risk Factors for Immune-Mediated Hepatotoxicity
• Key Findings: How Often Does Liver Injury Occur?
• Clinical Implications: What This Means for Patients
• Limitations: What the Study Could Not Prove
• Recommendations for Patients and Clinicians
• Frequently Asked Questions
• Source Information

Key Points

• In a pooled analysis of 24 studies with 9,076 patients, any-grade immune-mediated hepatotoxicity occurred in 14% and severe (grade 3 or higher) in 7%.
• Higher risk was linked to liver metastasis, prior immune checkpoint inhibitor treatment, diabetes, chronic hepatitis B, or combination CTLA-4 and PD-(L)1 inhibitor therapy.
• Lower risk was associated with older age, higher body mass index, male sex, and lung cancer, based on the pooled data.
• Liver injury typically appeared about 2.16 months after starting treatment and resolved in a pooled mean of about 1.22 months.
• Most cases are asymptomatic and found by routine liver function tests; report jaundice, dark urine, abdominal pain, or fatigue promptly.

Background: Why This Research Matters

Immune checkpoint inhibitors (ICIs) are a type of immunotherapy. They are monoclonal antibodies, which are laboratory-made proteins that act like natural antibodies. These drugs block specific proteins on the surface of T cells, which are immune cells that fight cancer.

Specifically, ICIs target two proteins: the anti-programmed death receptor 1 (PD-1) and the cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4). By blocking these proteins, the drugs prevent them from interacting with their partner molecules, called PD-L1/PD-L2 and CD80/CD86. This blockade restores T-cell activation and strengthens the immune system's ability to find and destroy tumor cells.

ICIs have become a major anticancer treatment, alongside surgery, chemotherapy, radiotherapy, and small-molecule targeted therapy. They have considerably improved survival rates and prognosis in patients with hard-to-treat tumors. These tumors include melanoma (MM, a type of skin cancer), non-small-cell lung cancer (NSCLC), renal cell carcinoma (RCC, kidney cancer), and hepatocellular carcinoma (HCC, liver cancer).

The field of immunotherapy is evolving. Recent research indicates that the anti-tumor effectiveness of ICIs is significantly increased by innovative antibody–drug conjugates (ADCs). A fundamental study indicates that ADC drugs, combined with immunotherapy, can enhance PD-L1 expression. This combination may reduce resistance to either treatment, providing a new approach in cancer treatment.

Despite these durable anti-tumor effects, ICIs occasionally cause a range of unintended adverse events (AEs). These are called immune-related adverse events (irAEs). They result from an excessive immune response that targets normal tissues or organs, and they can undermine the effectiveness of ICI therapy.

Among individual ICIs, the most common irAEs affect the skin (cutaneous), hormone glands (endocrine), heart and blood vessels (cardiovascular), and digestive system (gastrointestinal), including the liver. Immune-mediated hepatotoxicity (IMH) caused by ICIs is generally asymptomatic, meaning it produces no symptoms. Its incidence rate ranges from 1% to 15% in randomized trials.

Typically, IMH is identified incidentally through routine liver function tests during monitoring after treatment. Moreover, IMH has emerged as a significant concern within drug-induced liver injury (DILI), since it is considered a distinct and specific subtype of DILI. Ongoing research continues to reveal additional clinical features related to IMH. A deeper understanding of how IMH develops and how it presents can guide evidence-based treatment strategies for clinicians.

Despite the low severity of most IMH cases, life-threatening cases related to acute liver failure and even death have been reported during phase III clinical trials and post-marketing surveillance. The overall estimated mortality rate for irAEs ranges from 0.3% to 1.3%.

Data from the VigiLyze–VigiBase database of the World Health Organization indicate that 20.2% (124 out of 613) of patients who experienced fatal ICI-related toxic events died from IMH. This finding underscores the urgent need for early detection and effective management of these events.

Numerous studies have explored the risk factors of IMH in patients receiving ICIs, and a comprehensive treatment approach has been proposed based on these findings. However, substantial disagreements remain regarding the risk factors that occur after ICI treatment. This inconsistency arises from differences in data collected from clinical trials and real-world patient groups, and from variations in the included populations. Therefore, this systematic review and meta-analysis aimed to determine the incidence, prevalence, and risk factors of IMH following immunotherapy. The goal is to provide treatment options for cancer patients at potential risk.

How the Research Was Conducted

This work followed the PRISMA statement, a widely accepted guideline for reporting systematic reviews and meta-analyses. The study was registered in the PROSPERO database (registration number CRD42024531146).

Two independent authors searched four major medical databases: PubMed, Embase, Web of Science, and the Cochrane Library. They searched for articles published before 30 March 2024. They used Medical Subject Heading (MeSH) terms, including "Immunotherapy," "Risk factors," and "Chemical and Drug Induced Liver Injury," along with corresponding free-text words. They also used customized search terms such as "pembrolizumab," "Keytruda," "MK-3475," and "tislelizumab." There were no limitations on study design or publication language.

Articles were included if they met four criteria:

• The study included individuals diagnosed with cancer who received treatment with ICIs, either alone or in combination.
• The study provided information on the number of patients with any grade of IMH and those who did not develop IMH.
• The study examined the risk factors associated with IMH.
• The full text of the article was available.

Publications were excluded for several reasons. These included duplication of research, guidelines, consensus statements, letters, comments, notes, reports, case reports, conference abstracts, reviews, meta-analyses, experimental studies, and studies that were irrelevant or had insufficient data. Studies without full text available were also excluded.

Two researchers independently reviewed the titles, abstracts, and full texts to screen for eligible articles. For any discrepancies in study selection, discussions and consultations were conducted with a third researcher. The following information was extracted from the included articles:

• Basic details included first author, publication year, region, data source, study design, and sample size. Basic details also included criteria for evaluating adverse events, tools for assessing the causal relationship between ICIs and IMH, and the definition of IMH.
• Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) of potential risk factors. These factors included age, gender, combination therapies of anti-PD-(L)1 and anti-CTLA-4, other medical conditions (comorbidities), liver metastasis, liver cirrhosis, and HCC.
• Additional characteristics of both groups included demographic features, cases in each group, and number of IMH patients for all grades. Additional characteristics also included ICI types, presence of liver metastasis, history of ICI treatment, and underlying liver disease. Additional characteristics also included treatment regimen (monotherapy or combined therapy), performance status score, and laboratory tests.

Data were documented using a standardized extraction sheet. The quality of each article was appraised with the Newcastle–Ottawa scale (NOS) by three authors. The NOS uses 8 questions covering three areas: selection (0–4 scores), comparability (0–2 scores), and outcome (0–3 scores). A NOS score of 1–3 indicated low quality, 4–6 indicated moderate quality, and 7–9 indicated high quality.

All calculations were performed and all graphs were created using Stata Statistical Software version 17.0. The reported sample size and IMH events were used to calculate incidence. The inverse variance method was applied to pool IMH incidence. Subgroup analyses were conducted to determine the influence of region, data source, cancer types, ICI types, and the criteria used to define IMH.

The risk factors for any grade of IMH were also investigated. Extracted continuous data, expressed as median with range or interquartile range, were converted to mean with standard deviation for pooling. Heterogeneity across studies was assessed using the Cochrane Q test and the inconsistency index (I²) statistic. An I² statistic below 50% or a p value above 0.05 implied low heterogeneity, which warranted the use of a fixed-effect model. A higher I² statistic required the adoption of a random-effects model. Egger's tests were used to evaluate publication bias, with p < 0.05 indicating notable publication bias. Additionally, sensitivity analyses were performed. All analyses were considered statistically significant at p < 0.05.

Study Selection and Patient Characteristics

A total of 1,217 articles were retrieved after a comprehensive search conducted up to 31 March 2024. This included 149 records from PubMed, 549 from Embase, 186 from Cochrane, and 333 from Web of Science.

After removing 378 duplicates, 839 studies were retained for the initial title and abstract screening. Among these, 814 articles were excluded for the following reasons:

• Conference abstracts (n = 15)
• Guidelines and consensus statements (n = 9)
• Experimental studies (n = 15)
• Letters, comments, notes, case reports, and editorials (n = 140)
• Reviews and meta-analyses (n = 205)
• Irrelevant studies (n = 342)
• Clinical trial registrations (n = 84)
• Inability to access the full text (n = 4)
• Insufficient data (n = 2)

Furthermore, one study was added by searching for eligible articles from the reference lists. Moreover, Celsa et al. compared the characteristics and outcomes of immunotherapy-related liver injury in patients with HCC and other advanced solid tumors. To enhance the accuracy of patient information extraction, the meta-analysis divided that research into two parts.

Finally, 24 studies comprised the meta-analysis, resulting in a pooled cohort of 9,076 participants. Only one study was conducted prospectively, while the rest were retrospective. Among the included studies, 23 reported IMH of any grade, and 15 reported grade 3 or higher IMH.

The study covered various cancer types, including HCC, MM, RCC, urothelial cancer (UC), head and neck cancer, and lung cancer. Several studies reported immune-related adverse events in addition to IMH. One study reported adverse events in all patients, including gastrointestinal, endocrine, skin, lung, neuromuscular, and rheumatologic events. Nine studies documented concurrent irAEs in patients with IMH. The most prevalent concurrent irAEs were endocrine, gastrointestinal, skin, and colitis events.

The included studies came from many regions, including China, the UK, the USA, France, Switzerland, Canada, Japan, the Netherlands, Korea, Denmark, and multicenter collaborations across Europe, the USA, and Asia. Sample sizes ranged from 63 patients (Smith 2022, melanoma) to 1,096 patients (Miah 2023, multiple cancer types).

Quality Assessment of the Included Studies

Of the 24 included studies, 17 were classified as high quality, with scores ranging from 7 to 9 on the Newcastle–Ottawa scale. The other studies were categorized as moderate quality. These moderate-quality studies had deficiencies in providing follow-up information and in adequately addressing confounding factors, which are other variables that could influence the results.

Key Findings: Risk Factors for Immune-Mediated Hepatotoxicity

The baseline characteristics of patients in the IMH group and the non-IMH group were compared across all 24 studies. The significant factors were extracted and illustrated in a forest plot, a type of graph that shows the size of an effect for each factor.

In the figures, red dots signify that a factor reduces the risk of IMH, while blue dots indicate that a factor increases the risk of IMH.

Age, Body Mass Index, and Gender

The risk of IMH after ICI therapy is generally associated with a younger age. The standardized mean difference (SMD) was −0.18, with a 95% confidence interval (CI) of −0.33 to −0.04. The I² was 54.0%, and the p value was 0.012, based on 15 studies.

A lower body mass index (BMI) was also associated with IMH risk. The weighted mean difference (WMD) was −2.15, with a 95% CI of −3.92 to −0.38. The I² was 77.8%, and the p value was 0.017, based on 7 studies.

No significant publication bias was detected for either factor, according to Egger's test. Sensitivity analyses regarding the sources of heterogeneity for age and BMI were also performed.

Moreover, male gender was related to a reduced risk of IMH. The pooled adjusted odds ratio (OR) was 0.44, with a 95% CI of 0.27 to 0.72. The I² was 0%, and the p value was 0.001, based on 3 studies.

Liver Metastasis

A total of 17 studies reporting liver metastasis were pooled in a meta-analysis. The pooled OR was 1.80, with a 95% CI of 1.47 to 2.20. The I² was 47.7%, and the p value was less than 0.000. There was no significant publication bias (p = 0.451). This indicates a significantly elevated risk of IMH in patients with liver metastasis who were treated with ICIs.

History of ICI Treatment

The number of patients who had received prior ICI treatment was reported in five trials. These trials showed statistical differences between patients with IMH and those without. The pooled OR was 3.09, with a 95% CI of 1.21 to 7.89. Of note, substantial heterogeneity was observed across these five studies (I² = 70.3%). Sensitivity analyses addressing the source of heterogeneity were performed.

Other Medical Conditions (Comorbidities)

Within the pooled cohort, the risk of IMH after ICI treatment was notably higher in cancer patients with diabetes. The pooled OR was 2.19, with a 95% CI of 1.36 to 3.51. The I² was 0.0%, and the p value was 0.001, based on 4 studies.

The risk was also higher in patients with chronic hepatitis B virus (HBV) infection. The pooled OR was 2.46, with a 95% CI of 1.04 to 5.81. The I² was 69.9%, and the p value was 0.039, based on 7 studies.

However, the risk was not significantly higher in patients with several other liver conditions:

• Non-alcoholic fatty liver disease (NAFLD): pooled OR 1.84, 95% CI 0.55 to 6.11, I² = 51.3%, p = 0.321, based on 3 studies
• Hepatic steatosis (fatty liver): pooled OR 1.18, 95% CI 0.68 to 2.04, I² = 7.4%, p = 0.562, based on 3 studies
• Hepatitis C virus (HCV) infection: pooled OR 1.29, 95% CI 0.73 to 2.30, I² = 28.8%, p = 0.384, based on 6 studies

Consistent with the results from the pooled binary variable on chronic HBV, chronic HBV was also a significant risk factor for IMH after pooling the adjusted ORs from four studies. The pooled adjusted OR was 3.06, with a 95% CI of 1.11 to 8.46, and a p value of 0.031. There was no marked heterogeneity (I² = 41.1%) and no publication bias (p = 0.370).

Cancer Types

Immunotherapy with ICIs was used for various types of cancer, such as lung cancer, RCC, UC, MM, head and neck cancer, and HCC. Among them, lung cancer (564 out of 3,793 patients) and MM (383 out of 3,477 patients) were the most frequently reported.

The meta-analysis further investigated the risk factors for IMH, with a specific emphasis on various cancer types. Lung cancer was associated with a reduced risk of IMH development compared to other cancer types. The OR was 0.58, with a 95% CI of 0.41 to 0.83. The I² was 47.9%, and the p value was 0.003, based on 10 studies. There was no substantial difference in heterogeneity.

Concomitant Medications with Immunotherapy

In some patients, immunotherapy was combined with additional medications. These included tyrosine kinase inhibitors (TKIs), other immune checkpoint inhibitors, proton pump inhibitors (PPIs), and nonsteroidal anti-inflammatory drugs (NSAIDs).

The impact of concomitant ICIs was analyzed. When pooling the adjusted ORs, three studies indicated that the combination of ICIs was significantly associated with an increased risk of any grade IMH. The pooled adjusted OR was 8.73, with a 95% CI of 2.41 to 31.59. The I² was 71.9%, and the p value was 0.001. This combination specifically referred to CTLA-4 and PD-(L)1 inhibitors used together.

Key Findings: How Often Does Liver Injury Occur?

The mean time to onset of any grade IMH within the pooled cohort was 2.16 months, with a 95% CI of 1.18 to 3.13.

The estimated incidence of IMH following ICI treatment was 14% for any grade, with a 95% CI of 11% to 17% and an I² of 94.9%. For grade 3 or higher IMH, the estimated incidence was 7%, with a 95% CI of 5% to 8% and an I² of 91.0%. This analysis was based on 22 studies with 1,004 IMH patients.

Given the substantial heterogeneity, subgroup analyses were conducted regarding the region, data source, cancer types, ICI types, and criteria for adverse events. None of these factors were revealed to be independent contributors.

In terms of clinical patterns, the cholestatic pattern (52%, 95% CI 34% to 70%) tended to be marginally more prevalent than the hepatocellular pattern (20.0%, 95% CI 14% to 26%) and the mixed pattern (19%, 95% CI 9% to 30%). The cholestatic pattern involves reduced bile flow, the hepatocellular pattern involves direct liver cell damage, and the mixed pattern has features of both.

In addition, three studies investigated the median time from IMH onset to resolution. This time ranged from 7 to 514 days, with an overall pooled mean time of 1.22 months (95% CI 0.08 to 2.36).

Clinical Implications: What This Means for Patients

IMH is frequently observed as an adverse event in cancer patients undergoing ICI therapy. Despite its mild severity in most cases, there have been recorded cases of immune-related acute liver failure and even death.

Additionally, patients with underlying liver diseases were typically excluded from many studies. This exclusion happened because it is challenging to accurately diagnose IMH in these patients and to identify specific biomarkers or symptoms that might distinguish IMH from the cancer itself. Consequently, the occurrence and associated risk factors might be underestimated.

This meta-analysis included 24 studies with 9,076 patients. It found the pooled incidence of any grade of IMH was 14% (95% CI 11–17%), and the pooled incidence of grade 3 or higher IMH was 7% (95% CI 5–8%) in cancer patients. IMH typically developed approximately 2.16 months after the initial ICI treatment and resolved approximately 1.22 months from the onset.

This study will provide clinicians with information on potentially high-risk groups for IMH. These patients need to be cautiously monitored for liver function when receiving immunotherapy.

The findings suggest that certain patient characteristics should prompt closer monitoring. Patients with liver metastasis, a history of prior ICI treatment, diabetes, chronic hepatitis B, or those receiving combination ICI therapy face a higher risk. Conversely, older patients, patients with higher BMI, males, and patients with lung cancer appear to have a lower risk.

Limitations: What the Study Could Not Prove

This study has several limitations that patients and clinicians should consider.

First, only one of the 24 included studies was conducted prospectively. The remaining 23 studies were retrospective, meaning they looked back at existing data. Retrospective studies can be affected by incomplete records and selection bias.

Second, substantial heterogeneity was observed across studies for several outcomes. The incidence analysis had an I² of 94.9% for any grade IMH and 91.0% for grade 3 or higher IMH. This high heterogeneity means the studies varied considerably in their results. Subgroup analyses by region, data source, cancer types, ICI types, and adverse event criteria did not identify independent contributors to this variation.

Third, patients with underlying liver diseases were typically excluded from many of the original studies. This exclusion makes it difficult to accurately diagnose IMH in these patients and to identify specific biomarkers or symptoms that might distinguish IMH from the cancer itself. As a result, the occurrence and associated risk factors might be underestimated.

Fourth, some risk factors were analyzed in only a small number of studies. For example, male gender was analyzed in only 3 studies, diabetes in 4 studies, and concomitant ICI use in 3 studies. These small numbers limit the strength of the conclusions.

Fifth, the study could not determine cause and effect. It identified associations between certain factors and IMH risk, but it cannot prove that these factors directly cause liver injury.

Recommendations for Patients and Clinicians

Based on the findings of this systematic review and meta-analysis, the following recommendations can be made:

1. Patients with liver metastasis, prior ICI treatment, diabetes, chronic hepatitis B, or those receiving combination ICI therapy should be considered high-risk for IMH. These patients need cautious monitoring of liver function during immunotherapy.
2. Routine liver function tests should be performed during post-treatment monitoring, since IMH is generally asymptomatic and is typically identified incidentally through these tests.
3. Clinicians should be aware that IMH typically develops approximately 2.16 months after the initial ICI treatment. Monitoring should be most intensive during this period.
4. Patients should be informed that the cholestatic pattern of liver injury is the most common clinical pattern, occurring in 52% of IMH cases. The hepatocellular pattern occurs in 20%, and the mixed pattern occurs in 19%.
5. Patients with diabetes or chronic hepatitis B should discuss their liver monitoring plan with their care team before starting ICI therapy.
6. Patients receiving combination ICI therapy (CTLA-4 and PD-(L)1 inhibitors together) should be aware that this combination is associated with a significantly increased risk of any grade IMH.
7. Patients should report any symptoms that could indicate liver problems to their care team promptly. These may include jaundice (yellowing of the skin or eyes), dark urine, abdominal pain, or unexplained fatigue.

This study provides clinicians with information on potentially high-risk groups for IMH. These patients need to be cautiously monitored for liver function when receiving immunotherapy. Early detection and effective management of IMH are essential, given that 20.2% of patients who experienced fatal ICI-related toxic events died from IMH.

Frequently Asked Questions

How common is liver injury from immune checkpoint inhibitors?

In a pooled analysis of 24 studies covering 9,076 cancer patients, the incidence of any-grade immune-mediated hepatotoxicity was 14%, and severe (grade 3 or higher) liver injury occurred in 7%. Most cases are mild and produce no symptoms, but life-threatening cases have been reported. Routine liver function tests during treatment help detect it early.

Who is at higher risk of liver injury from immunotherapy?

Risk was higher in patients with liver metastasis, prior immune checkpoint inhibitor treatment, diabetes, chronic hepatitis B, or those receiving combination immunotherapy with CTLA-4 and PD-(L)1 inhibitors together. In the pooled analysis, these factors were associated with increased risk. If you have any of these, discuss closer liver monitoring with your care team before starting treatment.

Who is at lower risk of liver injury from immunotherapy?

In the pooled analysis, older age, higher body mass index, male sex, and having lung cancer were associated with a lower risk of immune-mediated hepatotoxicity. These findings come from multiple studies but do not mean no risk. Everyone receiving immune checkpoint inhibitors should still have routine liver function monitoring as recommended by their care team.

When does liver injury usually appear after starting immunotherapy?

In the pooled analysis, immune-mediated hepatotoxicity typically developed about 2.16 months after the first immune checkpoint inhibitor treatment. Monitoring should be most intensive during this period. Because it is usually asymptomatic, it is often found incidentally through routine liver function tests. Report any symptoms like jaundice, dark urine, abdominal pain, or unexplained fatigue promptly.

How long does it take for liver injury to resolve?

In three studies that looked at this, the time from onset of immune-mediated hepatotoxicity to resolution ranged from 7 to 514 days, with an overall pooled mean of about 1.22 months. Resolution times varied widely. Your care team will monitor your liver function and adjust treatment as needed. Always follow their guidance for your individual situation.

What does the 14% incidence of liver injury mean for me?

It means that in a pooled analysis of 9,076 patients, about 14 out of every 100 developed any-grade immune-mediated hepatotoxicity after immune checkpoint inhibitor treatment. This is an average across many studies; your personal risk may be higher or lower depending on factors like liver metastasis, diabetes, or combination therapy. Discuss your individual risk with your care team.

What are the symptoms of liver injury from immunotherapy?

Immune-mediated hepatotoxicity is generally asymptomatic, meaning it produces no symptoms, and is usually found through routine liver function tests. However, patients should report any symptoms that could indicate liver problems promptly, such as jaundice (yellowing of the skin or eyes), dark urine, abdominal pain, or unexplained fatigue. Early detection and management are important.

I'm starting immune checkpoint inhibitor therapy and have liver metastasis and diabetes — when should I get a second opinion on my liver monitoring plan?

A second opinion is reasonable before starting immune checkpoint inhibitor therapy if you have liver metastasis, prior ICI treatment, diabetes, or chronic hepatitis B, since these raise the risk of immune-mediated hepatotoxicity. Combination CTLA-4 and PD-(L)1 therapy also increases risk. Liver injury usually appears about 2.16 months after the first treatment and is often asymptomatic, so monitoring matters. A review can confirm your risk factors and follow-up plan. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: Risk Factors of Immune-Mediated Hepatotoxicity Induced by Immune Checkpoint Inhibitors in Cancer Patients: A Systematic Review and Meta-Analysis.

Authors: Jiang Y, Li R, Li X, Zhang N.

Author affiliations: Department of Pharmacy, Zhongshan Hospital Fudan University, Shanghai, China; and Department of Gastroenterology, Zhongshan Hospital Fudan University, Shanghai, China

Publication details: Current Oncology, 2024, Volume 31, pages 7129–7143. Published 13 November 2024. DOI: 10.3390/curroncol31110525

Registration: PROSPERO (CRD42024531146)

Note: This patient-friendly article is based on peer-reviewed research. The original study was a systematic review and meta-analysis of 24 studies including 9,076 cancer patients. This article is for informational purposes only and does not replace professional medical advice. Patients should consult their healthcare providers about their individual risk factors and monitoring plans.