Chronic liver disease and cirrhosis mortality following surgery for colon or rectal cancer.

Introduction
Colorectal cancer (CRC), also known as colon or rectal cancer, is a widespread malignant disease worldwide. Adjuvant and neoadjuvant chemotherapy serves as the primary treatment. However, the growing use of chemotherapy is accompanied by various drug-related diverse toxicities, with hepatotoxicity being a primary concern1.
Certain chemotherapeutic regimens used for CRC are linked to liver damage. For example, irinotecan may induce hypertransaminasemia and/or hyperbilirubinemia, hepatic steatosis, and even steatohepatitis2–4, while oxaliplatin primarily affects the sinusoids leading to parenchymal liver damage5,6. Additionally, anti-angiogenic targeted drugs (e.g., Bevacizumab) and immune checkpoint inhibitors (e.g., Pembrolizumab, Nivolumab) are increasingly used in CRC treatment and have also been linked to hepatotoxicity7,8. This hepatotoxicity not only compromises the individual oncologic treatment plans but can also directly contribute to mortality9. However, literature addressing the risk of chronic liver disease and cirrhosis (CLDC) mortality among these patients remains scarce. This research gap limits our understanding of CLDC-related mortality and hinders the development of effective interventions and monitoring strategies.
In this population-based study, we undertook a comprehensive analysis to assess the risk of CLDC-related death among patients with colon or rectal cancer. Our objective was to quantify the overall and long-term risk of CLDC death in CRC patients and to identify high-risk subgroups. The insights derived from this research are instrumental in informing follow-up and management strategies for CRC patients.

Methods

Data collection
Our study utilized data from the Surveillance, Epidemiology, and End Results (SEER) database. The SEER program is a leading global source of high-quality cancer registry data, using standardized methods to ensure consistency and minimize bias10. Specifically, the SEER-12 registry was used, which includes data from 12 regional registries beginning in 1992: San Francisco-Oakland SMSA, Connecticut, Hawaii, Iowa, New Mexico, Seattle (Puget Sound), Utah, Atlanta (Metropolitan), San Jose-Monterey, Los Angeles, Alaska Natives, and Rural Georgia. The SEER-12 covers approximately 12.2% of the U.S. population (based on 2020 census), which collects data from 12 regional cancer registries throughout the USA. Given the use of de-identified data, the requirement for ethics committee approval was waived. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).
Our study encompassed all patients diagnosed with colon or rectal cancer who underwent either endoscopic local excision or surgical resection (e.g., colectomy, proctectomy) from January 1992 to December 2021. We targeted a cohort aged between 50 and 80 years, as this age range is most susceptible to CRC. The inclusion criteria were as follows: (1) case selection based on site and morphology, specifically “Colon and Rectum”; (2) histologic codes corresponding to the International Classification of Diseases for Oncology, 3rd Edition, including codes 8140–8147, 8210–8213, 8255, 8260–8263, 8480–8481, and 8490; (3) active follow-up with a survival period of more than 0 days; and (4) presence of only one primary cancer. Exclusion criteria included cases (1) reported through autopsy or death certificate; and (2) diagnosed in individuals under 50 or over 79 years of age.

Study outcomes
The SEER database records the underlying causes of death using the International Classification of Diseases 10th Revision (ICD-10) codes provided by the National Center for Health Statistics. The primary outcome focused on mortality resulting from CLDC, encompassing specific underlying causes like alcoholic liver disease, chronic hepatitis, and fibrosis and cirrhosis of the liver. Person-years of follow-up were accumulated from the time of initial CRC diagnosis until the patient was lost to follow-up, the date of death, or the end of the study period on December 31, 2023.

Statistical analyses
To determine the risk of CLDC-related mortality among CRC patients, we utilized the standardized mortality ratio (SMR). The SMRs were adjusted for age, sex, and ethnicity to match the general population during the same period and were calculated as the observed to expected death ratio. The observed deaths are the actual number of deaths due to a specific cause within a given timeframe, while the expected deaths are the number of deaths anticipated for the same cause in a demographically similar population over the identical period. The expected death count was derived using the formula: expected deaths = person-years × general population mortality rate for the specific cause. The SMR was considered statistically significant if its 95% confidence intervals (CIs) did not encompass 1. The 95% CIs for SMRs were derived under the assumption that the observed number of deaths followed a Poisson distribution, using exact Poisson limits for the observed counts divided by the expected number of deaths. The absolute excess risk (AER) was then calculated as: AER = 10,000 × ((number observed - number expected)/person-years at risk). The CLDC-related AER indicates the population’s absolute increase in the risk of death from CLDC, essentially representing the burden of CLDC mortality. When reporting AERs, it is customary to present the excess number of deaths without explicitly mentioning the denominator of 10,000 person-years. Additionally, CLDC cumulative mortality was calculated using Fine and Gray’s competing risks model, with the cumulative incidence function serving as the basic descriptive statistic to estimate the absolute risk of endpoint events over time. Other causes of death, such as cardiovascular disease and death from primary CRC, were treated as competing risks in this model.
All statistical analyses were performed using R software (version 4.3.0; Vienna, Austria), Microsoft Excel (version 2016), and SEER*Stat (version 8.4.0, National Cancer Institute, USA).

Results

Patient characteristics
Table 1 presents the baseline characteristics of the study participants. Among patients with colon cancer, 52% were male, 74% were white, 37% were diagnosed between the ages of 70 and 79, and 49% survived for more than 5 years, contributing to 827,016 person-years at risk. The average follow-up period was 7.1 years (ranging from 0.0 to 29.9 years), with a median attained age of 74 years (ranging from 50 to 106 years). Partial proctectomy was performed in 54% of these patients. As of December 2023, a total of 380 CLDC-related deaths were recorded. Among patients with rectal cancer, 60% were male, 76% were white, 36% were diagnosed between the ages of 60 and 69, and 52% survived for more than 5 years, contributing to 298,414.5 person-years at risk. The average follow-up period was 7.5 years (ranging from 0.0 to 29.9 years), with a median attained age of 71 years (ranging from 50 to 105 years). Subtotal colectomy or hemicolectomy was employed in 40.1% of cases, whereas partial colectomy accounted for 32.8%. As of December 2023, a total of 130 CLDC-related deaths were recorded.

Figure 1 shows cumulative CLDC mortality increased steadily with survival time in both patients with colon cancer and patients with rectal cancer.

CLDC mortality among patients with colon cancer
Among patients with colon cancer, the overall observed CLDC mortality to expected (O/E) ratio was 380/215.38, resulting in a SMR of 1.76 (95% CI, 1.59 to 1.95) and an AER of 2.24. Patients diagnosed between the ages of 50–59 had an SMR of 1.68 (95% CI, 1.38 to 2.03) and an AER of 1.98. Those in the 60–69 age group exhibited an SMR of 1.92 (95% CI, 1.62 to 2.25) and an AER of 2.74, while the 70–79 age group showed an SMR of 1.67 (95% CI, 1.38 to 2.00) with an AER of 1.94.
Male patients displayed an SMR of 1.82 (95% CI, 1.60 to 2.06) and an AER of 3.09, compared to female patients who had an SMR of 1.67 (95% CI, 1.39 to 1.98) and an AER of 1.4. The SMR for patients diagnosed between 1992 and 2003 was 1.92 (95% CI, 1.65 to 2.22) with an AER of 2.56. For those diagnosed between 2004 and 2015, the SMR decreased to 1.75 (95% CI, 1.50 to 2.03) and the AER to 2.25. For patients diagnosed from 2016 to 2021, the SMR further declined to 1.08 (95% CI, 0.67 to 1.65) with an AER of 0.28. For patients with follow-up durations of ≤ 12 months, 13–36 months, 37–60 months, and > 60 months, the respective SMRs were 2.99, 1.43, 1.91, and 1.55, with corresponding AERs of 5.71, 1.25, 2.7, and 1.64 (Fig. 2).

Among male patients, SMRs were consistently higher across all age groups compared to the general population, particularly in the 60–69 years age group, which had the highest SMR of 1.96 (95% CI, 1.60 to 2.38) and an AER of 3.66. Temporal trends indicated a decrease in SMR from 2.09 (95% CI, 1.74 to 2.50) for the period 1992–2003 to 0.83 (95% CI, 0.41 to 1.49) for 2016–2021. Among female patients, a downward trend in SMR was observed with increasing age. For both genders, the length of follow-up showed that patients monitored for up to 12 months had the highest SMRs and AERs, with male patients exhibiting an SMR of 2.78 and an AER of 6.63, and female patients an SMR of 3.43 and an AER of 4.73 (Fig. 2C and E). Stratified by disease stage, stages I–III accounted for 364 CLDC deaths, yielding an SMR of 1.82 (95% CI, 1.74–2.50), whereas stage IV patients experienced 16 CLDC deaths with an SMR of 0.62 (95% CI, 0.23–1.35).

CLDC mortality among patients with rectal cancer
Among patients with rectal cancer, the overall observed to expected (O/E) ratio is 130/91.83, resulting in a SMR of 1.42 (95% CI, 1.18 to 1.68) and an AER of 1.28. Patients diagnosed between the ages of 50–59 exhibited an SMR of 1.50 (95% CI, 1.12 to 1.96) and an AER of 1.53. Those in the 60–69 age group had an SMR of 1.31 (95% CI, 0.95 to 1.75) and an AER of 0.96. The 70–79 age group showed an SMR of 1.45 (95% CI, 0.99 to 2.04) with an AER of 1.34.
The SMR for patients diagnosed within the period of 1992 to 2003 stood at 1.56 (95% CI, 1.20 to 1.99) with an AER of 1.61. This figure decreases to an SMR of 1.38 (95% CI, 1.05 to 1.79) and an AER of 1.22 for those diagnosed between 2004 and 2015.
Male patients displayed an SMR of 1.54 (95% CI, 1.26 to 1.87) and an AER of 2.06, compared to female patients who had an SMR of 1.1 (95% CI, 0.74 to 1.58) and an AER of 0.21. For patients with follow-up durations of ≤ 12 months, 13–36 months, 37–60 months, and > 60 months, the respective SMRs were 1.98, 1.23, 1.23, and 1.41, with corresponding AERs of 2.93, 0.71, 0.71, and 1.28 (Fig. 2).
Male patients exhibited an overall SMR of 1.65 across the age groups, with the 50–59 years age group recording the highest SMR of 1.67 (95% CI, 1.22 to 2.23). Temporal trends were observed in the year of diagnosis, with the period from 1992 to 2003 recording the highest SMR of 1.83 (95% CI,1.36 to 2.40) and an AER of 2.93 (Fig. 2D and F). Stratified by disease stage, stages I–III accounted for 116 CLDC deaths, yielding an SMR of 1.42 (95% CI, 1.18–1.69), whereas stage IV patients experienced 14 CLDC deaths with an SMR of 1.26 (95% CI, 0.34–3.24).

CLDC mortality among long-term survivors
A total of 56,936 patients with colon cancer and 20,635 with rectal cancer survived more than 5 years. Among long-term survivors of colon cancer, 51% were male and 35% were diagnosed between the ages of 60–69. Among long-term survivors of rectal cancer, 58% were male and 38% were diagnosed between the ages of 50–59 (Table 2).

Among long-term survivors with colon cancer, the overall observed to expected (O/E) ratio was 173/111.31, resulting in a SMR of 1.55 (95% CI, 1.33 to 1.80) and an AER of 1.64. Males exhibited a higher SMR of 1.72 (95% CI, 1.43 to 2.06) and an AER of 2.75 compared to females, with an SMR of 1.28 (95% CI, 0.95 to 1.67) and an AER of 0.6. White survivors exhibited an SMR of 1.61(95% CI, 1.36 to 1.89) and an AER of 1.97, while Black survivors had a notably higher SMR of 2.02 (95% CI, 1.13 to 3.34) and an AER of 2.12.
Among long-term survivors with rectal cancer, the overall observed to expected (O/E) ratio was 67/47.47, resulting in a SMR of 1.41 (95% CI, 1.09 to 1.79) and an AER of 1.28. Males had a higher SMR of 1.59 (95% CI, 1.19 to 2.08) and an AER of 2.25 compared to females, who had an SMR of 1.02 (95% CI, 0.57 to 1.68) and a minimal AER of 0.04 (Supplementary Fig. 1).

Discussion
This extensive, population-based, long-term follow-up study conducted a thorough evaluation of the risk of CLDC-related mortality among patients diagnosed with colon cancer and rectal cancer. Using nearly three decades of registry data, this is the first comprehensive study to show a higher risk of CLDC-related mortality in colon and rectal cancer patients compared with the general population.
Chemotherapy- and immunotherapy-related hepatotoxicity represent key contributors to long-term CLDC risk among colorectal cancer survivors. Oxaliplatin has been associated with sinusoidal obstruction syndrome and parenchymal liver injury, while irinotecan is linked to steatosis and steatohepatitis. Immune checkpoint inhibitors targeting PD-1/PD-L1 pathways (e.g., pembrolizumab, nivolumab) may induce immune-mediated hepatitis ranging from mild transaminitis to fulminant hepatic failure. Moreover, pre-existing non-alcoholic fatty liver disease (NAFLD) and metabolic comorbidities may act as important confounders, predisposing patients to both treatment-related hepatotoxicity and CLDC development. These factors underscore the multifactorial nature of liver injury in colorectal cancer management and highlight the need for vigilant hepatic monitoring and individualized treatment planning.
Our research revealed that the risk of CLDC-related mortality among patients with colon cancer and rectal cancer was 1.76 and 1.42 times higher, respectively, than that of the general US population. This elevated risk has also been reported in other types of malignant tumors, such as prostate cancer and pancreatic cancer11,12. Conversely, Zhai et al. determined that the risk of short-term CLDC-related mortality does not significantly differ between patients with non-metastatic bladder cancer and the general population13. Treatments for colon, rectal, prostate, and pancreatic cancers often involve systemic therapies, including chemotherapy and hormonal treatments, which can have hepatotoxic side effects14,15. These treatments can lead to a higher incidence of liver complications compared to treatments for non-metastatic bladder cancer, which might be more localized (e.g., transurethral resection, intravesical therapy) and less likely to impact liver function.
Age-stratified analyses revealed distinct patterns: colon cancer survivors aged 60–69 bore the highest burden of CLDC mortality, whereas rectal cancer patients in this age group had comparatively lower risk. These differences may reflect lifestyle or treatment-related factors16,17. Our findings suggest that tailored follow-up strategies might be essential for individuals diagnosed with colon and rectal cancer, particularly those diagnosed between the ages of 60 to 69.
Our study has, for the first time, demonstrated that in the era of targeted therapies and immunotherapy (post-2004), the risk of CLDC-related mortality among patients with colon and rectal cancer persists. This indicates that, even with newer personalized treatments, patients with CRC still face the same CLDC-related problems. Known as immune-mediated hepatitis (IMH), liver injury due to immunotherapy can range from mild, spontaneously resolving elevations in aminotransferases to potentially fatal increases exceeding 20 times the upper limit of normal18–20. Following Pembrolizumab treatment for CRC, the incidence of any grade elevated AST was 10%, with only 1% of cases being grade 3, and up to 3% of patients experienced IMH21. Addressing this challenge requires a comprehensive strategy, including new therapies to protect liver health, improved screening, better management of pre-existing conditions, and personalized treatment plans. Consequently, our study underscores the need for continued research and clinical vigilance to identify and reduce contributors to CLDC-related mortality within the framework of modern cancer treatments22.
Although male patients have higher CLDC-related mortality than female patients, both genders exhibited an elevated risk compared to the matched general population. Interestingly, with attained age, the magnitude of SMR elevation for female patients with colon cancer significantly diminished, a trend not observed in male patients. This difference might be attributed to women being more proactive in seeking healthcare or adhering to follow-up schedules, potentially leading to earlier detection of side effects like IMH and better management of pre-existing liver conditions23,24. In addition, Women may experience a slower progression of certain liver diseases and may respond differently to treatments, which could be partially due to their proactive health behaviors25,26. Therefore, these gender-specific risk profiles and health outcomes underscore the importance of personalized medicine, customized follow-up protocols, and gender-sensitive healthcare policies.
Among stage I–III colon and rectal cancer survivors, SMRs were elevated, corresponding to roughly 40–80% higher CLDC mortality than expected in the general population. These excess risks may, in part, reflect the hepatotoxic effects of adjuvant chemotherapy regimens (e.g., FOLFOX/CAPOX), which are routinely administered in node-positive (stage III) and, in some high-risk stage II cases, to reduce recurrence. Oxaliplatin and 5-fluorouracil have been linked to steatohepatitis and sinusoidal injury; such effects could plausibly contribute to longer-term CLDC mortality, although our data do not permit causal attribution. By comparison, stage IV patients are usually treated with systemic chemotherapy in combination with targeted agents or immunotherapy. Estimates for Stage IV disease should be interpreted with caution. Although patients with metastatic CRC are exposed to more intensive systemic therapies, the few CLDC-related deaths observed in this group are primarily due to the high competing risk of early cancer mortality, which shortens survival and consequently reduces the time at risk for CLDC. This results in imprecise and potentially underestimated SMRs for Stage IV patients.
In our study, approximately half of the CLDC-related mortalities occurred among long-term survivors of CRC. A similar pattern was observed with other treatment-related mortalities, including cardiovascular deaths27. Long-term survivors of cancer face a heightened risk of chronic health conditions, which can stem from their cancer treatment or the natural aging process28,29. After treatment, patients may revert to or maintain lifestyle habits that elevate the risk of CLDC and cardiovascular diseases—these include poor dietary choices, lack of physical activity, and alcohol consumption30–32. Comprehensive survivorship care should address both immediate cancer management and the long-term health effects of the disease and its treatment. The attenuation of SMRs in the 2016–2021 cohort likely reflects limited follow-up for a late-emerging outcome. The apparent decline in SMRs among patients diagnosed during 2016–2021 should therefore be interpreted with caution, as this pattern likely reflects shorter follow-up duration and delayed onset of CLDC-related mortality rather than a true reduction in risk.
The strengths of our study lie in its extensive multicenter population and prolonged follow-up period. It stands as the first and most comprehensive study to evaluate the risk of CLDC-related death in patients with CRC. The substantial size of our study population enabled a detailed, stratified analysis based on patient characteristics, while the extended follow-up period allowed us to accurately assess the long-term risk of CLDC-related death.
This study has several limitations. The SEER database lacked comprehensive treatment data, which precluded evaluation of CLDC-related mortality risk based on specific chemotherapeutic regimens or immunotherapies. In addition, the database did not capture key potential confounders, including comorbidities, alcohol consumption, hepatitis infection, and pre-existing liver disease. The absence of such information, along with missing data on liver function markers such as transaminases, limited our ability to adjust for these important factors and may have introduced unmeasured confounding.

Conclusion
The risk of CLDC-related death in patients with colon or rectal cancer is notably higher than in the general population. Accordingly, a comprehensive approach to care is imperative for CRC patients, particularly among long-term survivors, and should incorporate evidence-based strategies for the prevention, early detection, ongoing surveillance, and management of CLDC-related complications. Specifically, routine liver function assessments and advanced imaging modalities are indispensable for the prompt identification of hepatic abnormalities. Furthermore, lifestyle interventions such as limiting alcohol, adopting a liver-friendly diet, and maintaining regular physical activity can markedly improve liver function and overall outcomes. These findings highlight the need for clinicians to incorporate liver-focused monitoring and preventive strategies into survivorship care plans. Routine liver function assessments, patient education on lifestyle modifications, and proactive surveillance should be prioritized to mitigate long-term hepatotoxic risks.

Supplementary Information
Below is the link to the electronic supplementary material.