Background Cholelithiasis and cholecystectomy have been proposed as risk factors for liver cancer, but findings have been inconsistent. We assessed this association using data from the Shanghai Women's and Men's Health Studies.
Methods History of cholelithiasis and cholecystectomy were reported at baseline and follow-up interviews, and liver cancer diagnoses were ascertained from the Shanghai Cancer Registry and Vital Statistics Unit. Adjusted hazard ratios (aHRs) and 95% CIs were calculated after adjustment for potential confounders.
Results A history of cholelithiasis and cholecystectomy was reported by 9.5% and 3.6% of participants at baseline, respectively. After a total of 859 882 person-years of follow-up for women and 391 093 for men, incident liver cancer was detected in 160 women and 252 men. A positive association was observed between a history of cholelithiasis or cholecystectomy and liver cancer in men (aHR 1.46; 95% CI 1.02 to 2.07) and women (aHR 1.55; 95% CI 1.06 to 2.26). Similar results were observed for cholelithiasis only, but cholecystectomy did not reach statistical significance. There was no strong evidence for detection bias of liver cancer due to cholelithiasis or cholecystectomy.
Conclusions Our study suggests that cholelithiasis and possibly cholecystectomy may increase the risk of liver cancer.
- Cancer epidemiology
- Cohort studies
- Epidemiology of chronic diseases
Statistics from Altmetric.com
Worldwide, liver cancer ranks as the fifth incident cancer among men and seventh among women. Owing to the high fatality rate, it is the second most common cause of cancer death among men and sixth among women.1 The burden of liver cancer resides largely in China, accounting for more than 50% of both incident liver cancer cases and deaths worldwide.1 Currently, there are no standard screening methods for liver cancer, which indicates that primary prevention of liver cancer is a priority to decrease the incidence and mortality from this malignancy.
Cholelithiasis (ie, gallstones) has been previously considered as a disease of the West due to its association with diet and obesity.2–4 Symptomatic cholelithiasis is often treated by cholecystectomy,5 removal of the gallbladder, with varying rates of elective and non-elective cholecystectomies worldwide.6–8 In developed countries, gallstone disease has been estimated to affect anywhere from 10% to 15% of adults and cost approximately $6.2 billion annually in the USA.9 Both cholelithiasis and cholecystectomy have been proposed as risk factors for liver cancer. Previous epidemiological studies have detected inconsistent associations between cholelithiasis, cholecystectomy and liver cancer, and many have relied on data from registries, which reduces the ability to adjust for important confounders.10–21 Inconsistent results could be related to a number of factors, including differences in study design, exposure assessment or population, and lack of confounder adjustment. Therefore, we evaluated the association of cholelithiasis and cholecystectomy with the risk of liver cancer using data from the Shanghai Women's Health Study (SWHS) and the Shanghai Men's Health Study (SMHS).
The designs for the SWHS and SMHS have been published in detail previously.22 ,23 From 1996 to 2000, 74 941 women aged 40 to 70 years old from seven urban communities in Shanghai were recruited for the SWHS with an overall participation rate of 92.7%. From 2002 to 2006, 61 483 men aged 40 to 74 years old with no previous cancer were recruited for the SMHS from eight urban communities in Shanghai with an overall participation rate of 74.1%. Surveys and anthropometric measurements were collected by trained retired medical professionals. Follow-up interviews were conducted every 2 to 3 years for the SWHS and SMHS. Written informed consent was obtained from all participants, and both cohorts were approved by all relevant Institutional Review Boards.
History of cholelithiasis was assessed at baseline and in a subsequent follow-up questionnaire. At baseline, participants were asked whether they had ever been diagnosed with cholelithiasis by a physician and, if so, the age at diagnosis. In the third follow-up interview for the SWHS and first follow-up interview for the SMHS, participants were asked whether they had ever been diagnosed with cholelithiasis or gallstones by a physician and whether the diagnosis was confirmed by ultrasound or X-ray examination. At baseline, 8161 women (11.15%) and 4614 men (7.52%) reported having a history of cholelithiasis. Of participants who reported a history of cholelithiasis at baseline, 6047 women (74.10%) and 3244 men (70.31%) reported at the follow-up interview to have a history of cholelithiasis. Of participants who consistently reported having a diagnosis of cholelithiasis (N=6047 for SWHS and N=3244 for SMHS), 5863 women (96.96%) and 3174 men (97.84%) reported that the diagnosis was confirmed by ultrasound or X-ray examination.
Due to the discrepancy in reporting of cholelithiasis, we assessed the effect of different self-reported cholelithiasis exposures: (1) baseline-reported cholelithiasis; (2) baseline-reported and follow-up-reported cholelithiasis using repeated measures; and (3) concordant reporting of cholelithiasis at baseline and follow-up (ie, yes/yes or no/no for history of cholelithiasis).
History of cholecystectomy was assessed at baseline and updated in follow-up questionnaires. At baseline, participants in the SWHS and SMHS were asked whether they had ever had a cholecystectomy and, if so, at what age it occurred. In subsequent follow-up, the participants were only asked about new cholecystectomy procedures since baseline, so we were unable to assess the consistency of reporting. Participants were categorised as having a history of cholecystectomy if they reported a previous cholecystectomy at baseline.
We additionally categorised participants as having either a history of cholelithiasis or cholecystectomy at baseline. We calculated the time since cholelithiasis and cholecystectomy as the difference between the reported age at the time of diagnosis of cholelithiasis or the receipt of cholecystectomy and the age that the participant developed cancer or was censored due to death, loss to follow-up or end of study.
Participants in the SWHS and SMHS were followed up by annual record linkage with the population-based Shanghai Cancer Registry for incident cancer diagnoses and the Vital Statistics Unit of the Shanghai Center of Disease Prevention and Control for mortality. Incident cancer cases identified through linkage with the Shanghai Cancer Registry were verified through home visits. Medical charts were reviewed for verification and to document cancer diagnosis and treatment information. Incident liver cancer was defined as a primary tumour with an International Classification of Diseases (ICD)-9 code of 155 (malignant neoplasm of liver and intrahepatic bile ducts). Specific sites, primary neoplasms of the liver (ICD-9 155.0) and of the intrahepatic bile ducts (ICD-9 155.1) were also considered. TNM staging data were available for some incident liver cancer cases (53.6%). Follow-up data were available through 31 December 2010.
Potential confounding variables
Information about potential confounders was collected at baseline. Demographic variables included age, sex, education level and family income. Body mass index was calculated from the height and weight of each participant measured at baseline (kg/m2). Behavioural characteristics included cigarette smoking, alcohol consumption, menopausal status (for women only) and amount of leisure time physical activity per day in metabolic equivalent hours per day (MET hours/day). Total energy intake was assessed in a baseline food frequency questionnaire. Participants also reported family history of liver cancer, history of diabetes and history of hepatitis or chronic liver disease. Since missing data were minimal, missing responses were included in the most common category.
Women diagnosed with cancer prior to baseline (N=1598) were excluded from analysis. We excluded participants from both studies who were lost to follow-up shortly after enrolment, diagnosed with cancers of unknown origin or had missing data on baseline cholelithiasis or cholecystectomy (N=146 for SMHS and N=134 for SWHS). After exclusions, 61 337 men and 73 209 women were available for analysis.
Analyses were conducted separately for each cohort. We calculated age-adjusted descriptive statistics by baseline-reported cholelithiasis and/or cholecystectomy and used the Cochran–Mantel–Haenszel test for categorical variables and the test of differences in least-squared means for continuous variables except for physical activity, which was tested using ranks in a general linear model due to non-normality. We determined age-adjusted hazard ratios (HRs) and 95% CIs for incident liver cancer by categories of cholelithiasis and cholecystectomy using the Cox proportional hazards regression model. Entry time was defined as the age that the participant started in the SMHS or SWHS, and exit time was defined as the age when the participant developed incident liver cancer or was censored due to death, loss to follow-up or end of study on 31 December 2010. We also calculated adjusted HRs (aHRs), including potential confounders in the models. We used a random-effects model to pool the summary estimates for cholelithiasis and cholecystectomy for men and women when the Cochrane Q statistic had a p value less than 0.1 for the test of heterogeneity; otherwise fixed-effects models were used. We evaluated the proportional hazards assumption by including an interaction between the cholelithiasis and cholecystectomy categories with the logarithm of time. The interaction was not statistically significant, so we assumed that there were no violations of proportional hazards.
A number of additional analyses were conducted. We evaluated the associations by liver cancer type. We additionally tested the exclusion of participants with a history of hepatitis or chronic liver disease and participants with a history of diabetes. We assessed the number of liver cancer cases identified within a short interval (ie, less than 5 years) since diagnosis of cholelithiasis or cholecystectomy at baseline to evaluate potential detection bias introduced by treatment for cholelithiasis. As an estimate of potential early detection of liver cancer due to increased surveillance, we also assessed the association between baseline-reported cholelithiasis and/or cholecystectomy with stage of liver cancer at diagnosis and with liver cancer mortality using the Pearson χ2 test among participants who developed liver cancer during follow-up. Statistical analyses were conducted using SAS V.9.3 (SAS Institute, Cary, North Carolina, USA), and a two-sided p value of 0.05 was considered statistically significant.
At baseline, a history of cholelithiasis was reported by 11.1% of women and 7.5% of men while 4.3% of women and 2.7% of men reported a previous cholecystectomy. The age-adjusted descriptive statistics according to baseline history of cholelithiasis and cholecystectomy are presented in table 1. Participants with a history of cholelithiasis or cholecystectomy differed from those without a history on some factors, including age, education, smoking history and leisure time physical activity (table 1).
After a total of 859 882 and 391 093 person-years of follow-up for women and men, respectively, incident liver cancer was detected in 160 women and 252 men. Among the liver cancer cases in women, 97 were primary liver, 45 were intrahepatic bile duct and 18 were unspecified. In men, 155 cases were primary liver, 20 were intrahepatic bile duct and 97 were unspecified. Similar associations were observed between a history of either cholelithiasis or cholecystectomy and incident liver cancer among women (HR 1.74; 95% CI 1.20 to 2.53) and men (HR 1.74; 95% CI 1.22 to 2.46) in age-adjusted analyses. After adjustment for confounders, the associations were attenuated, but remained statistically significant. When the multivariable adjusted estimates were pooled for men and women, the aHR for incident liver cancer was 1.50 (95% CI 1.16 to 1.94) for participants with a history of cholelithiasis or cholecystectomy at baseline. A similar association estimate was observed for a history of cholelithiasis at baseline (pooled aHR 1.49; 95% CI 1.15 to 1.94), and the association estimate was strengthened for the model that updated the cholelithiasis exposure based on follow-up reporting (pooled aHR 1.74; 95% CI 1.36 to 2.22). The concordantly reported cholelithiasis pooled estimate was weaker than both the baseline only and updated exposure model (pooled aHR 1.39; 95% CI 0.98 to 1.96). Although a history of cholecystectomy increased the risks of liver cancer in men and women, none of the observed associations reached statistical significance (table 2).
When primary liver cancer (ICD-9 155.0) was the outcome, the results for women were generally similar to the overall analyses. For men, the associations with primary liver cancer were generally strengthened from the overall analyses. There were insufficient cases of intrahepatic bile duct cancer to conduct analyses (results not shown). When participants with a self-reported history of hepatitis or chronic liver disease were excluded from analysis, all observed associations for women were strengthened and were generally weakened for men. When participants with a self-reported history of diabetes were excluded, observed associations for women and men were strengthened (results not shown).
Among participants with a self-reported history of cholelithiasis at baseline who developed liver cancer during follow-up, only four cases (5.5%) were identified within 5 years of diagnosis. Similarly, among participants with a self-reported history of cholecystectomy at baseline who developed liver cancer during follow-up, only one case (3.8%) was identified within 5 years of cholecystectomy. Baseline-reported cholelithiasis and cholecystectomy were not associated with stage of diagnosis of liver cancer or liver cancer death among participants who developed liver cancer. A higher proportion of participants with cholelithiasis presented with stage IV liver cancer (40.48% vs 29.05%), although the difference was not statistically significant (results not shown).
In this large cohort study, we found that having a history of either cholelithiasis or cholecystectomy increased the risk of liver cancer by about 55% for women and 46% for men. When considered individually, only the association between cholelithiasis and liver cancer reached statistical significance. Both baseline-reported cholelithiasis and analyses restricted to participants who consistently reported their history of cholelithiasis at baseline and follow-up yielded similar estimates. The association for the concordant reporting of cholelithiasis with liver cancer did not reach statistical significance; however, this is likely to be related to sample size since after exclusions only 296 cases of liver cancer remained compared with 412 in the other analyses. When primary liver cancer was considered, results were generally similar for women, but strengthened for men.
A number of studies also found that cholelithiasis was a significant risk factor for liver cancer; however, the prospective studies were mostly based on registry data and were therefore unable to adjust for potentially important confounders, such as alcohol consumption.10–16 For example, in Denmark, a standardised incidence ratio (SIR) of 1.4 (95% CI 1.2 to 1.8) was found for the association between gallstone disease and liver cancer, but when stratified by cholecystectomy status, the association was only significant among patients who had not had a previous cholecystectomy (SIR 1.9; 95% CI 1.3 to 2.6).12 A recent case–control study in the USA found an increase in the odds of liver cancer by 2.35 (95% CI 2.18 to 2.54) or people with a history of gallstones17 while a large network of case–control studies conducted in Italy and Switzerland found no statistically significant association with an OR of 1.17 (95% CI 0.83 to 1.65).18
The literature is similarly conflicting about whether previous cholecystectomy is a risk factor for liver cancer.15–17 19–21 For example, a cohort in Taiwan estimated a SIR of 3.29 (95% CI 2.55 to 4.18) for liver cancer when comparing the incidence for people with a cholecystectomy to the general population.19 A prospective study conducted in the UK identified an association between cholecystectomy and liver cancer (incidence rate ratio (IRR) 1.45; 95% CI 1.09 to 1.90), but when cancers identified less than 2 years prior to the cholecystectomy were excluded, no association was observed (IRR 0.91; 95% CI 0.64 to 1.25).21 Since we had few participants with a previous cholecystectomy who developed liver cancer (15 women and 11 men), we may have been underpowered to detect a statistically significant association, although we did estimate a 38% increased risk of liver cancer in pooled analyses. Similarly, many of these studies were based on registry data and confounder adjustment was limited or not done, which could lead to residual confounding.
Chronic inflammation is a commonly proposed pathway for carcinogenesis for numerous cancer sites, including the liver.24 Responses to chronic inflammation, such as oxidative stress and liver regeneration, may be related to the development of liver cancer.25 Gallbladder inflammation, increased mucin secretion and oxidative stress are all common features of gallstone formation that may in turn affect the liver.26 Treatment with cholecystectomy may decrease these risk factors for carcinogenesis by removing the affected organ. However, in an animal study,27 cholecystectomy increased accumulation of fat in the liver, which may lead to non-alcoholic fatty liver disease and inflammation of the liver.28 Our study may suggest that chronic inflammation related to cholelithiasis increases the risk of liver cancer; however, future research should address the mechanism through which cholelithiasis and cholecystectomy may be related to liver carcinogenesis.
Misclassification of cholelithiasis and cholecystectomy was possible due to the self-reported nature of the data. However, we assessed the consistency of report and additional diagnostic information obtained through subsequent questionnaires. The estimates from the baseline-reported cholelithiasis were assumed to have some measurement error, but no missing data, while the concordant-reported estimates should have less measurement error, but more missing data and therefore less power to detect the association. The estimates were similar between analyses, which suggests that misclassification of the exposure may not have strongly influenced our results. However, even if the exposures were misclassified, misclassification would likely be non-differential due to the prospective nature of the study, which would bias the results towards the null and underestimate the strength of the association.
Detection bias may have also influenced our results. Participants diagnosed with cholelithiasis or received a cholecystectomy may have increased surveillance due to their condition, which could lead to incidental detection of liver cancer. Increased risks of cancer have been observed shortly after diabetes diagnoses, a likely result of detection bias related to diabetes treatment.29 When we considered the interval of time since diagnosis with cholelithiasis or cholecystectomy, less than 6% of liver cancer cases were identified within 5 years of diagnosis. Similarly, if detection bias were present, participants with cholelithiasis or cholecystectomy would likely be diagnosed at earlier stages or have decreased mortality; however, we did not identify any differences in stage at diagnosis or liver cancer mortality by history of cholelithiasis or cholecystectomy. Given these findings, there is little evidence that the associations were due to detection bias.
Other limitations of this study include that data on chronic hepatitis and liver disease were only available by self-report, so underreporting due to asymptomatic chronic liver disease or fibrosis was likely. In the general population in Shanghai, the prevalence of current hepatitis B and C infection has been estimated to be about 5.5% and 0.4%, respectively, with a previous hepatitis B infection detected in about 27.6% for anti-HBc marker of the population.30 We were also unable to ascertain whether cholelithiasis or cholecystectomy occurred prior to or after hepatitis infection, so future work should assess whether cholelithiasis is associated with liver cancer independent of hepatitis infection.
This study also has important strengths. The SWHS and SMHS are both rigorously designed cohort studies with high participation and retention rates. All of the covariates were assessed prior to the development of cancer, which should decrease misclassification bias. Also, we had a relatively large number of incident liver cancer cases and long follow-up on which to base this analysis.
In conclusion, cholelithiasis and cholecystectomy may be risk factors for liver cancer. People diagnosed with cholelithiasis may need increased surveillance for liver cancer. Future research should focus on the potential mechanisms of liver carcinogenesis related to cholelithiasis and cholecystectomy.
What is already known on this subject?
Worldwide, liver cancer ranks as the fifth incident cancer among men and seventh among women and it has a high fatality rate.
Cholelithiasis and cholecystectomy have been proposed as risk factors for liver cancer.
Previous results have been inconsistent and many studies have not adjusted for potentially important confounders.
What this study adds?
In this large, prospective study of men and women in Shanghai, China, a history of cholelithiasis statistically significantly increased the risk of liver cancer.
The association between cholecystectomy and liver cancer was weaker and did not reach statistical significance.
Increased surveillance of patients with cholelithiasis for liver cancer may be warranted although research is needed to confirm these findings.
We would like to thank the participants and the staff from the Shanghai Women's and Men's Health Studies for their contribution to this research.
Contributors Y-BX contributed to the conception and design of the study. Y-BX, H-LL and Y-TG acquired data. EV, CY, X-OS and Y-BX performed the statistical analysis and the interpretation of results. EV wrote the first draft. All authors contributed to the critical review of the manuscript and approved the final manuscript. Y-BX had full access to all of the data and the final responsibility for the decision to submit for publication.
Funding This work was supported by funds from the State Key Project Specialized for Infectious Diseases of China (No. 2008ZX10002-015 and 2012ZX10002008-002 to Y-BX); the United States National Institutes of Health (R37 CA070867 to WZ, and R01 CA082729 to X-OS); the Fogarty International Clinical Research Scholars and Fellows Program at Vanderbilt University (R24 TW007988-5 to EV and H-LL); and the Cancer Prevention and Control Training Program at the University of Alabama at Birmingham funded through the National Institutes of Health (5R25 CA047888 to EV). The funding organisations had no role in the design and conduct of the study; the collection, analysis and interpretation of the data; or the preparation, review or approval of the manuscript.
Competing interests None.
Ethics approval Shanghai Cancer Institute (China), Vanderbilt University (USA) and National Cancer Institute (USA).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.