1 Introduction
Gallstone is a major kind of benign gastrointestinal disease. In the past decades, the prevalence of gallstone disease has been growing rapidly, mainly due to the aging populations and the shift toward a high-fat diet [
1]. The prevalence of gallstone disease varies from 20% to 30% among western adult populations [
2] and reaching 11% in China according to the latest meta-analysis [
3]. Gallstones can lead to severe complications, such as acute abdomen pain, and often require surgical interventions. Currently, cholecystectomy is the standard treatment for symptomatic gallstones [
4].
Although a medical history of gallstones has been associated with several gastrointestinal cancers according to previous literature [
5–
7], the accurate impact of gallstones on other types of cancers remains controversial across studies and ethnicities [
8,
9]. For cholecystectomy, the results are even more conflicting [
10,
11]. For example, some studies reported a positive association of cholecystectomy with liver, stomach, and breast cancer, whereas other research observed either non-significant or negative associations [
12,
13]. Furthermore, evidence of the effect of gallstone disease on cancer risk in Chinese adults is especially limited, with existing literature mainly focusing on gastrointestinal cancers [
14–
16]. These findings might be biased due to the interplay of common risk factors shared by cancer and gallstones, such as sedentary lifestyles, oxidative stress, and chronic inflammation [
17]. Therefore, Mendelian randomization (MR) is suitable to address these issues. With the use of genetic variants as proxies for gallstones and cholecystectomy, MR study can effectively filter unmeasured confounders and avoid reverse causality.
Given the high prevalence of gallstones in China and many other countries, the association of gallstones with major types of cancer should be determined. Here, we first performed a nationwide, large sample size, cross-sectional study, followed by a well-defined two-sample MR analysis. With the help of these two methods, we aimed to investigate the possible causal effect of gallstones and cholecystectomy on cancers.
2 Materials and methods
The architectural design of the study is illustrated in Fig.1. The population-based observational study was conducted with data from the baseline of the Risk Evaluation of cAncers in Chinese diabeTic Individuals: a lONgitudinal (REACTION) study [
18]. In brief, REACTION study is an ongoing, multicenter, prospective cohort study conducted in 25 communities between 2011 and 2012 across Chinese mainland. At the baseline, 259 657 participants aged 40 or older were identified from local registration records and approached by trained staff. No restriction on gender or ethnicity was applied. The study has been approved by the Ethic Committee of Ruijin Hospital Affiliated to Shanghai Jiao Tong University School of Medicine.
Two-sample MR was performed according to the following principles: (1) the genetic variants should be significantly associated with gallstones and cholecystectomy; (2) the genetic variants should not be associated with any confounders; and (3) the genetic variants should influence the cancers of interest completely due to gallstones and cholecystectomy [
19]. Results of MR analysis were reported according to the STROBE of MR studies [
20]. Institutional review board approval was not applicable for MR analysis because only de-identified data were used.
2.1 Observational study
2.1.1 Study population and data collection
In-person interviews were administered, and questionnaires were completed by the participants. Information on medical history, lifestyle factors (smoking and drinking behaviors, diet, and physical activity), menopausal variables, and family history of tumors were collected through a validated questionnaire [
18,
21]. Data on dietary habits and physical activity were obtained by the semiquantitative food frequency questionnaire suggested by the Chinese Center for Disease Control and Prevention [
22]. Information on physical activity was accessed by the short form of International Physical Activity Questionnaire [
23]. Body weight and height were measured by experienced nurses on site according to standard protocols. Individuals with complete information on medical history of gallstones and cancers were included in the current analysis.
2.1.2 Assessment of exposures, covariates, and outcomes
Medical history of gallstones, cholecystectomy, and cancer was self-reported via questionnaires as mentioned above. Participants were asked about the whether they had a definite diagnosis of gallstones or had undergone a cholecystectomy procedure because of gallstones. They were further asked to provide the accurate date of surgery if cholecystectomy was performed. Medical records were required from the participants to verify the diagnosis. Our primary exposure was a medical history of gallstones, either with or without cholecystectomy. Participants were divided into three groups: no history of gallstones, a history of gallstones without cholecystectomy, or gallstones with cholecystectomy.
Dietary habits were assessed subsequently in these four parts: fish of more than 3.5-oz servings twice a week; fruits and vegetables of more than 4.5 cups/day; sweets or sugar-sweetened beverages of less than 450 kcal/week; and soy protein of more than 25 g/day. Each part was assigned 1 score, and a diet score higher than 2 was defined as a healthy diet [
24]. The intensity of physical activities was evaluated according to the 2008 Physical Activity Guidelines for Americans. Moderate and vigorous physical activity of ≥ 150 min/week or vigorous-intensity physical activity ≥ 75 min/week was regarded as physically active [
25]. Body mass index (BMI) was calculated as body weight in kilograms divided by body height squared in meters (kg/m
2).
Outcomes of interest were site-specific cancers, which were verified by cross-validation with medical records, tumor registry data, and diagnosis certificate from doctors. In the general population, gender-neutral cancers (e.g., stomach cancer, colorectal cancer, liver and bile duct cancer, pancreatic cancer, thyroid cancer, lung cancer, kidney cancer, and bladder cancer) were analyzed. For women, female breast cancer, cervical cancer, and endometrial cancer were additionally included. For men, prostate cancer was further discussed.
2.1.3 Statistical analysis
One-way ANOVA and CHISQ square test were used to compare the differences in continuous and categorical data across groups, respectively. Multivariable logistic regression models were used to examine the associations of a history of gallstones, cholecystectomy, and the risk of certain cancers. Conventional confounders were adjusted in the regression models, including baseline age, sex (for gender-neutral cancers only), smoker (yes/no), drinker (yes/no), healthy diet (yes/no), physically active (yes/no), high school education (yes/no), marital status (married or single), and BMI (kg/m
2). For female-specific cancers (female breast cancers, ovarian cancers, endometrial cancers, cervical cancers, and ovarian cancers), age at menarche and menopause (yes/no) were further adjusted considering the well-established association in between [
26,
27]. All statistical analyses were performed using SAS version 9.4 (SAS Institute). Two-tailed
P < 0.05 was considered statistically significant.
2.2 Mendelian randomization study
2.2.1 Instrumental variable selection for gallstones and cholecystectomy
Genetic proxies for gallstones (general and gender-specific) were obtained from the Neale laboratory analysis of UKB round 2 GWAS [
28]. The total cases/controls were 7157/353 984 for general, 5575/188 578 for female-specific, and 1582/165 406 for male-specific gallstones. Proxies for cholecystectomy were retrieved from “ukb-b-6235” (
n case = 26 145;
n control = 436 788). No single nucleotide polymorphisms (SNPs) of gender-specific cholecystectomy were available.
SNPs strongly associated with gallstones and cholecystectomy were used as instrumental variables (IVs), for which a genome-wide significance threshold of P < 5 × 10−8 was applied. The default clumping threshold (r2 < 0.001) was used to avoid any linkage disequilibrium. Furthermore, the strength of the IVs was appraised using the F-statistic through the formula below
“beta” was the effect size of each IV, and “SE” referred to the corresponding standard error of “beta” . An F-statistic > 10 indicated sufficient statistical strength. Finally, 22 SNPs for general gallstones, 34 SNPs for female-specific gallstones, and 45 SNPs for male-specific gallstones were retained as the IVs. No gender-specific SNP information on cholecystectomy is available at present. Furthermore, 45 SNPs to proxy genetic cholecystectomy (general) remained after clumping with European ancestry. Detailed information on SNPs proxied for exposures is presented in Tables S1 and S2.
2.2.2 GWAS of cancers of interest
GWAS summary data for the cancers were downloaded from the most recent GWAS within European ancestry in Open GWAS API [
29]. Specifically, IVs of stomach cancer, colorectal cancer, liver and bile duct cancer, pancreatic cancer, thyroid cancer, female breast cancer, endometrial cancer, cervical cancer, and ovarian cancer were obtained from pan-cancer GWAS [
30]. IVs of bladder cancer and prostate cancer were extracted from two other GWASs [
31,
32]. Additionally, proxies for lung and kidney cancer were acquired from TRICL consortium and FinnGen Biobank, respectively. Detailed information of the data resources for exposures and outcomes is listed in Table S3.
2.2.3 Statistical analysis
The inverse-variance weighted (IVW) method was utilized as the primary analytic method, and the Wald ratio method was instead when only one SNP was used for the analysis. The “harmonization” statement was used to reconcile the exposure and outcome data. OR (95% CI) and
P values without multiple testing were presented. Sensitivity analyses included pleiotropy assessment and outlier tests. MR–Egger was used to identify any potential directional pleiotropy, and MR–PRESSO was used to detect horizontal pleiotropy and outlier SNPs [
33]. Leave-one-out analyses were performed to assure that the analysis did not depend on a particular variant [
34]. The MR Steiger approach was used to determine the direction of a possible causal effect between gallstones and cholecystectomy on cancers [
35]. In the current study, no SNPs with false causal directions were identified. All statistical analyses were performed using R software (version 4.0).
3 Results
3.1 Baseline characteristics of the observational study
Overall, a total of 259 657 individuals were included for the observational study. After excluding those without information on the medical history of gallstones, 239 799 remained in the final analysis. Baseline characteristics of the participants are shown in Tab.1. The average age was 57.4 (standard error: 9.6) years old, and 65.3% of them were female. Individuals with a history of gallstones had higher BMI, were older, and more likely to be physically active than those without. However, they were less likely to be drinkers or smokers, and few of them had optimal diet patterns. Additionally, no difference was observed in terms of education and marital status. The ratio of having a family history of tumor was higher among those with a history of gallstones, especially among those experiencing a cholecystectomy procedure, compared with their counterparts.
3.2 Observational associations of gallstones and cholecystectomy with cancer risk
The risk estimates of cancers associated with gallstones and cholecystectomy are shown in Fig.2. Compared with those without a medical history of gallstone disease, a history of gallstones without cholecystectomy was associated with a higher risk of stomach cancer (adjusted odds ratio [aOR] = 2.54, 95% confidence interval [CI]: 1.50–4.28), liver and bile duct cancer (aOR [95% CI]: 2.46 [1.17–5.16]), kidney cancer (aOR [95% CI]: 2.04 [1.05–3.94]), and bladder cancer (aOR [95% CI]: 2.23 [1.01–5.13]). Additionally, a history of gallstones with cholecystectomy was positively associated with stomach cancer (aOR [95% CI]: 2.41 [1.29–4.49]), liver and bile duct cancer (aOR [95% CI]: 2.58 [1.11–6.02]), and colorectal cancer (aOR [95% CI]: 1.83 [1.18–2.85]). For gender-specific cancers, a positive relationship was only observed between gallstones and cervical cancer (aOR [95% CI]: 1.69 [1.12–2.56]) among women, and cholecystectomy was not significantly related to any gender-specific cancer in the observational study.
3.3 Associations of genetically proxied gallstones and cholecystectomy with cancer risk
All the IVs proxied for gallstones and cholecystectomy showed strong associations with the exposures (all F-statistics > 10), indicating that the overlap between exposure and outcome was unlikely to cause bias [
36]. Moreover, no outliers of SNPs were detected with MR–PRESSO in the analysis for gallstones. Results of the MR analyses are shown in Fig.3 and Fig.4. Genetically proxied gallstones were causally associated with stomach cancer (OR = 2.67, 95% CI: 2.53–2.81,
P < 0.001), liver and bile duct cancer (OR = 1.02, 95% CI: 1.01–1.03,
P < 0.001), kidney cancer (OR = 67.49, 95% CI: 6.91–659.56,
P < 0.001), and bladder cancer (OR = 1.01, 95% CI: 1.00–1.02,
P = 0.004). However, the present analysis showed no causal effect of cholecystectomy on cancers of any site (Fig.4). The results of the leave-one-out analysis confirmed the findings of the main analysis (Figs. S1 and S2).
4 Discussion
Previous studies have suggested an inconsistent association between gallstones and cancers, whereas no definitive evidence for the causality is available. Conventional logistic regression analysis and MR method were applied in the current study to evaluate the association of gallstones, cholecystectomy, and cancers. In the observational study, a history of gallstone without cholecystectomy was associated with a heightened risk of stomach cancer, liver and bile duct cancer, kidney cancer, and bladder cancer in the overall population, as well as cervical cancer risk in women. Meanwhile, gallstone with cholecystectomy was associated with elevated odds of stomach cancer, liver and bile duct cancer, and colorectal cancer. Further MR analysis supported the causal effect of gallstones on stomach, liver and bile duct, kidney, and bladder cancer. However, no genetic associations between cholecystectomy and any types of cancer were detected.
Research exploring the relationships between gallstones and cancers has yielded inconsistent results. In addition to some negative results [
8,
37], several studies have reported that gallstones significantly increase the risk of gastrointestinal cancers, including stomach, liver, and colorectal cancers [
14,
38,
39]. Other investigators also reported a relationship between gallstones and kidney cancer [
11]. In the present observational study, we similarly detected a significant association between gallstones and cancers of the stomach, liver and bile duct, colorectal, and kidney. Furthermore, we identified the effect of gallstones on bladder cancer. Notably, these associations have been validated in the subsequent MR analysis. These findings not only demonstrated that gallstones are causally connected to these gastrointestinal cancers but also established a causal association between gallstones and bladder cancer for the first time. Regarding gender-specific cancers, previous studies suggested that gallstones increase the risk of breast cancer in women [
40,
41] and prostate cancer in men [
7]. Nonetheless, we found no association between gallstones and breast or prostate cancer. Instead, gallstones exhibited a significant correlation with cervical cancer in the current analysis, which has not been reported in previous studies. Unfortunately, MR analysis did not support the causality of the gallstones–cervical cancer relationship, so further prospective cohort study is warranted to address this issue.
The effect of cholecystectomy on cancer risk is controversial. Cholecystectomy is traditionally associated with colorectal cancer [
37,
42]; recently, it was found to be associated with an elevated risk of stomach and liver cancer [
8]. These findings might cause concern among surgeons before performing the removal of the gallbladder. Despite the similarly positive associations in the observational study, the subsequent MR study failed to ascertain these findings. Other studies reported that cholecystectomy might decrease the risk of several cancers [
8,
13], possibly because the inflammation status improves after the removal of infectious gallbladder. Previous studies showed that the detrimental effects of gallstones gradually vanish over time, indicating that conclusions from the observational studies could be merely due to surveillance bias, or reverse causation caused by pre-existing neoplasms [
43,
44]. We speculated that a history of cholecystectomy in the observational data setting might represent a history of having gallstone disease, which differed from the genetically determined cholecystectomy in MR analysis. MR is particularly beneficial in neutralizing such biases. Previous MR studies have already postulated the cause–effect relationship between gallstones and kidney cancer [
11], but the cholecystectomy–cancer relationship was not validated in MR analysis [
45,
46]. Our study was highly consistent with these findings. We also addressed the causal relevance of gallstones with stomach, liver, and bladder cancer for the first time.
Several mechanisms were proposed to explain the possible carcinogenesis caused by gallstones. Dysregulated bile acid flow, metabolic and hormonal changes inside the adjuvant structures of gallbladder, and chronic inflammation status were all the possible biological process [
42,
47]. Given that cholecystectomy changes the natural anatomy of bile tracts and results in increased bile flow secretion, it may cause dysbiosis with primary bile acids transforming to secondary bile acids [
48]. Therefore, gallstones and cholecystectomy are closely related to cancers nearby biliary tracts. Unfortunately, the effect of gallstones on kidney and bladder cancers is not well understood. Future studies are needed to determine the carcinogenic mechanism on cancers of the urinary tract.
This study had several strengths, including the comprehensive analysis toward major kinds of cancer, the large sample size of the observational study, and the use of gender-specific MR analysis. However, some limitations have to be addressed. First, information on gallstones and cholecystectomy was obtained retrospectively by questionnaires and medical records in the population-based study, which might introduce recall bias. However, the validity of self-reported gallstone disease has already been confirmed in previous studies [
7]. Second, gallstones and cancers were assessed simultaneously, for which a reverse causation might exist. Nevertheless, the consecutive MR study added definitive evidence. Third, all the cancers included in the analysis were solid tumors, so no analysis was performed regarding hematological cancers. Fourth, the observational study was conducted in Chinese individuals, whereas the summary data used for MR study were obtained from GWAS among European populations. Although results from the two methods demonstrated mutual verification, further research in diverse populations is warranted to obtain cross-population generalization, considering racial/ethnic disparities in the formation and etiology of gallstones.
In conclusion, observational and MR studies supported significant associations between gallstones and cancers of the stomach, liver and bile duct, kidney, and bladder. The observed association between gallstones and cervical cancer, as well as cholecystectomy and colorectal cancer, might not be causal. Considering the potential cancer risk related to gallstones, special attention should still be paid toward those diagnosed with gallstone disease.
PDF
(3156KB)
