Association between Dietary Anthocyanidins and Biliary Cancer Risk in 98,458 Participants: Results from a Prospective Study

Biliary cancer refers to a diverse range of tumors that occur in the biliary system, and these tumors are further categorized according to their specific sites of origin (1). A well-known risk factor of gallbladder cancer is gallstones. Although biliary cancer is relatively rare, its prognosis is extremely poor. A previous study has indicated that the survival rate for all subtypes of biliary cancer is notably low, with a reported 5-year survival rate of merely 15.2% for individuals diagnosed with this condition (2). In cases of locally advanced or metastatic biliary cancer, the median survival of patients is less than 1 year (3). The etiology and risk factors of biliary cancer remain poorly understood, highlighting the need for extensive research in this area.

In recent times, there has been a significant focus on diet-based approaches for the prevention and treatment of cancer (4). Anthocyanidins, which are water-soluble flavonoids, have found widespread use as natural food colorants (5). Anthocyanidins possess biological functions, including health-promoting effects and a potential association with reducing the risk of cancer (6, 7). A previous in vitro study showed that anthocyanidins exhibit free radical scavenging and antioxidant properties, along with inhibitory effects on the proliferation of cancer cells (8). Many studies tried to explore the relationship between dietary anthocyanidins and different cancers. For instance, Zhang and colleagues conducted a study and reported a reduced risk of lung cancer associated with dietary anthocyanidins (9). In a study involving 469,008 participants, Sun and colleagues reported a significant association between dietary anthocyanidins and a 28% reduction in the risk of head and neck cancer (10). Cui and colleagues similarly found an association between dietary anthocyanidins and a reduced risk of esophageal cancer (11). Intriguingly, a study by Xu and colleagues also found that dietary anthocyanidins can reduce the risk of kidney cancer (12). However, limited research has been conducted on the correlation between an anthocyanidin-rich diet and biliary cancer. Therefore, the main aim of this study was to examine the potential correlation between the consumption of dietary anthocyanidins and the occurrence of biliary cancer in a substantial cohort of participants.

The data used in this study were collected from the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial, which was sponsored by the NCI. The PLCO Cancer Screening Trial was a multicenter randomized clinical trial specifically designed to evaluate the potential reduction in cancer mortality through screening exams (13). Between 1993 and 2001, the PLCO trial was conducted at 10 different centers located throughout the United States, encompassing a large cohort of 154,887 individuals ages between 55 and 74 years. This trial employed a randomized allocation process, assigning participants to either an intervention group or a control group. In the intervention group, specific screening tests were administered, while the control group received standard or routine care. To collect comprehensive data, participants in the PLCO trial were required to complete two questionnaires: the baseline questionnaire (BQ) and the dietary history questionnaire (DHQ). The BQ aimed to gather relevant baseline information, including general participant characteristics. The questionnaire aimed to capture essential demographic data and other relevant factors at the start of the trial. The DHQ is a self-administered food frequency questionnaire consisting of 137 items. It is designed to assess the participants’ intake of food items and nutrients. Several studies have evaluated the validity of the DHQ and have suggested that it is a reliable tool for nutritional assessment (9, 12). This study received approval from the NCI under Project ID: PLCO-1257.

The exclusion criteria for participants in this study were as follows: Participants were excluded if they failed to return the BQ (n = 4,918). Invalid DHQ were also excluded, which included participants with insufficient frequency responses (less than eight) or extreme energy intake (above the 99th percentile or below the 1st percentile) and with a completion date prior to the date of death (n = 38,462). In addition, participants who were diagnosed with any cancer before completing the DHQ (n = 9,684) were excluded. Participants between randomization and DHQ completion, who either developed biliary cancer, died, or were lost to follow-up, were also excluded (n = 69). Furthermore, participants with extreme dietary intake, characterized by very low or very high caloric intake (below 600 or above 3,500 kcal/day for women and below 800 or above 4,200 kcal/day for men; ref. 14), were excluded (n = 3,296). Ultimately, a total of 98,458 participants were included in the study.

Participants’ information regarding age, gender (male or female), race/ethnicity (White or non-White), education level (college below, college graduate, or postgraduate), baseline body mass index (BMI), family history of biliary cancer (yes, no, or possible), smoking status (never, current, or former), drinking status (no, yes), history of gallbladder stones or inflammation (no, yes), history of diabetes (no, yes), trial arm (intervention or control), alcohol consumption, physical activity level, and energy intake from the diet were collected from the PLCO trial.

The six common anthocyanidins are cyanidin, delphinidin, malvidin, peonidin, petunidin, and pelargonidin (5, 15). The combined intake of cyanidin, delphinidin, malvidin, peonidin, petunidin, and pelargonidin represented the total daily intake of anthocyanidins (15). The participants’ daily intake of the six common anthocyanidins was collected through the aforementioned DHQ questionnaire. To account for losses during processing, it was assumed that the amounts of nutrients in processed foods were 50% of those in raw foods (9). The DietCalc software was used to calculate the DHQ nutrient variables based on questionnaire responses, taking into consideration various factors of food. The software utilized in this study combines the data from the Continuing Survey of Food Intakes by Individuals with the questionnaire responses to assess the daily intake of various nutrients (16). Previous studies have established the reliability and accuracy of the DHQ (17).

In this study, demographic, lifestyle, and medical information were assessed using the BQ questionnaire. This included variables such as sex, race, trial arm (intervention or control group), BMI, smoking status, and family history of biliary cancer. BMI was calculated by dividing weight (in kilograms) by the square of height (in square meters). Age at the completion of the DHQ questionnaire, drinking status, history of gallbladder stones or inflammation, history of diabetes, alcohol consumption, and intake of the six common anthocyanidins were assessed using the DHQ.

The PLCO trial obtained confirmation of biliary cancer diagnosis through study update forms, which were annually mailed to participants to inquire about cancer diagnoses. The primary outcome assessed in this study was the incidence of biliary cancer, encompassing various subtypes including gallbladder cancer (ICD-O-2, C239), extrahepatic bile duct cancer (ICD-O-2, C240), ampulla of Vater cancer (ICD-O-2, C241), overlapping lesions of biliary tract cancer (ICD-O-2, C248), and biliary tract cancer, not otherwise specified (ICD-O-2, C249).

Categorical and continuous covariates with less than 5% missing values were imputed using the mode and median values, respectively. Detailed information regarding the imputation data can be found in Supplementary Table S1.

The relationship between dietary intake of anthocyanidins and the risk of biliary cancer was investigated using a Cox proportional hazards regression model. HRs and their corresponding 95% confidence intervals (CI) were calculated, with the duration of follow-up serving as the time metric. The follow-up period was defined as the time from the completion of the DHQ to the occurrence of biliary cancer, death, loss to follow-up, or the end of the specified follow-up period (Supplementary Fig. S1). To test for a trend between biliary cancer incidence and quartiles of total anthocyanidin intake, each participant was assigned the median value within their designated quartile, which was subsequently treated as a continuous variable in the regression model, using the lowest quartile as the reference group. Potential confounding variables were identified through a comprehensive review of relevant literature and utilizing the clinical expertise of researchers (18–22). The selected covariates were then incorporated into the Cox regression model to mitigate their potential influence on the outcomes. In the multivariate analyses, Model 1 was adjusted for age, sex, and race. Model 2 included additional adjustments for trial arm, BMI, family history of biliary cancer, smoking status, drinking status, and history of diabetes. Model 3 was further adjusted for history of gallbladder stones or inflammation as well as alcohol intake. To aid model convergence, total anthocyanidins were also standardized by subtracting their mean value and dividing it by their SD in Supplementary Table S2. Similar methods were utilized to analyze the associations between intake of anthocyanidin and flavonone subclasses and biliary cancer risk.

To better characterize the dose–response relationship between total anthocyanidin intake and biliary cancer risk, a restricted cubic spline curve with three knots was employed to model the association across the full range of anthocyanidin intake using the “rcssci” R package. The number of knots was chosen to minimize the Akaike information criterion. The reference value (HR = 1) was set at the 10th percentile of anthocyanidin intake. Knots were placed at the 10th, 50th, and 90th percentiles of total anthocyanidin intake. Intake values below the 5th and above the 95th percentiles were excluded.

After stratifying for age (>65 vs. ≤65 years), sex (male vs. female), BMI (≤25 vs. >25 to ≤30 vs. >30 kg/m²), smoking status (never vs. current/former), drinking status (no vs. yes), history of gallbladder stones or inflammation (no, yes), and history of diabetes (no vs. yes), a series of subgroup analyses was conducted. To mitigate the possibility of spurious subgroup effects, a P_interaction was assessed by comparing models with and without multiplicative interaction terms prior to conducting the aforementioned subgroup analyses. To test the robustness of the findings, several sensitivity analyses were performed. (I) participants with a history of diabetes or a history of gallbladder stones or inflammation were excluded; (II) To address concerns regarding reverse causality, cases observed within the first 2 and 4 years of follow-up were excluded from the analysis; (III) the primary analysis was repeated in participants nonmissing data; (IV) Further adjusted for energy intake from diet, physical activity level, and educational level.

The statistical analyses were performed using R 4.2.0 software. A two-tailed P value less than 0.05 was considered significant.

The raw data used in this article are not available because of the NCI's data policy. Access to the dataset should contact the NCI by mail.

In this study, 98,458 participants were included, and the mean (SD) for total anthocyanidins diet was 15.97 mg/day (13.91 mg/day). Participants in the study were divided into quartiles based on their total anthocyanidins diet. Participants in the highest quartile of total anthocyanidins diet were more likely to be female and have a higher educational level. Moreover, they exhibited a higher prevalence of drinkers and had higher energy intake from their diet. Conversely, they had a lower BMI and a higher likelihood of being smokers. The comprehensive baseline characteristics of the study population are presented in Table 1.

During a total of 872,645.3 person-years of follow-up, 95 cases of biliary cancer were documented, resulting in an overall incidence rate of 11 cases per 100,000 person-years. The mean (SD) duration of follow-up was 8.869 (1.899) years. As shown in Table 2, the unadjusted model revealed that participants in the highest quartile of total anthocyanidins intake had a significantly lower risk of biliary cancer compared with those in the lowest quartile (HR _{Q4 vs. Q1}: 0.53; 95% CI: 0.30–0.92; P_trend = 0.044). Even after fully adjusting for potential confounding variables, a similar association remained (Model 3: HR _{Q4 vs. Q1}: 0.52; 95% CI: 0.29–0.91; P_trend = 0.043). As is shown in Supplementary Table S2, the same results were observed when the total anthocyanidins diet was standardized. In the subclass anthocyanidins class analysis, no association was found with biliary cancer, except for Malvidin (P_trend = 0.045) and Peonidin (P_trend = 0.050) in the fully adjusted model (Supplementary Table S3). In addition, as shown in Supplementary Table S4, the subclasses of flavonones were usually not statistically significantly associated with biliary cancer risk [except for Apigenin (P_trend = 0.029)]. It should be noted that we did not find any association between the intake of total anthocyanidins and various anatomic sites of biliary cancer (Supplementary Table S5).

The results from the restricted cubic spline (RCS) analysis demonstrated a linear dose–response relationship between total anthocyanidins diet and the risk of biliary cancer (P = 0.027 for overall association and P_nonlinearity = 0.118), as depicted in Fig. 1. Subgroup analyses revealed that the inverse associations between total anthocyanidins diet and biliary cancer risk were more pronounced in participants who were drinkers compared with those who were nondrinkers (P_interaction = 0.033; Table 3). In sensitivity analyses (Supplementary Table S6), the associations remained consistent even after excluding participants with a history of diabetes, gallbladder stones or inflammation. In addition, when excluding cases observed within the first 2 or 4 years of follow-up, and conducting the analysis using nonmissing data, the associations remained similar. However, it should be noted that when we further adjusted for energy intake from diet, physical activity level, and educational level, the inverse association no longer existed.

This study aimed to explore the relationship between dietary anthocyanidins and the risk of biliary cancer in a prospective analysis of 98,458 American adults. The results revealed a significant inverse association between total anthocyanidins intake and the occurrence of biliary cancer. This association remained consistent even after adjusting for confounding factors and standardizing the total anthocyanidins diet. Subgroup analysis showed a stronger inverse association between anthocyanidins intake and the risk of biliary cancer among participants who reported drinking alcohol compared with nondrinkers. Sensitivity analyses confirmed the robustness of the inverse association, as it persisted when excluding participants with a history of diabetes, gallbladder stones or inflammation and when considering different follow-up durations. Overall, these findings provide evidence supporting a potential protective effect of higher dietary anthocyanidins intake against biliary cancer, with a potential interaction between anthocyanidins intake and drinking status.

Previous research has suggested that dietary anthocyanidins or anthocyanidin-rich fruits and extracts may have protective effects against various chronic diseases (23). Anthocyanidins are known for their strong antioxidant capacity and their ability to inhibit cancer cell growth (24), and have also been shown to inhibit tumor cell invasion and migration (25). Previous study has demonstrated the potential protective effect of anthocyanidins against colorectal cancer (26). In breast cancer, anthocyanidins inhibit phosphorylation of HER2, induce apoptosis, inhibit migration and invasion, and suppress tumor cell growth (27). Bunea and colleagues found that anthocyanidins inhibited the proliferation and induced apoptosis in B16-F10 melanoma mice cells (28). Similarly, in a rat liver cancer model, Bishayee and colleagues found that anthocyanidin-rich black currant has the ability to inhibit abnormal cell proliferation and induce apoptosis (29). However, the association between dietary anthocyanidins and biliary cancer remains understudied, with a limited number of published studies available on this topic. Further research is needed to better understand and establish the potential relationship between dietary anthocyanidins and the risk of biliary cancer.

Hence, this study represents the first analysis examining the association between dietary anthocyanidins and biliary cancer in the U.S. population. Utilizing prospective data from the PLCO cancer screening trial, a total of 98,458 participants were followed for an average of 8.9 years. The study findings revealed that individuals in the highest quartile of total anthocyanidins intake had a 48% lower incidence of biliary cancer compared with those in the lowest quartile. Moreover, the RCS model employed in this study revealed a linear relationship between total anthocyanidins intake and the risk of biliary cancer (P_nonlinear = 0.118). It is worth noting that anthocyanidins have poor bioavailability and the metabolites that can be found in tissues are different from anthocyanidins in foods (30). Further research is needed to characterize the complex effects and biological mechanisms between anthocyanidins and biliary cancer development. In addition, the subgroup analysis results indicated that the inverse association between total anthocyanidins intake and biliary cancer risk was more pronounced in participants who consumed alcohol compared with nondrinkers. A potential explanation is that anthocyanidins may help mitigate the procarcinogenic effects of alcohol on the biliary tract. Alcohol has a carcinogenic effect by producing metabolites such as acetaldehyde (31). Anthocyanidins has potential antioxidant protection effects (32), and it may offset more additional carcinogenic factors of alcoholics. It can also produce effects among nondrinking people, but the results of the subgroup analysis show that the role in the drinker may be more significant. It is worth noting that due to the limited number of biliary cancer cases within each subgroup, we were unable to analyze significant interactions between total anthocyanidins intake and other potentially influential factors on the incidence of biliary cancer.

We tried to explain this inverse correlation between total anthocyanidins diet and biliary cancer by the following mechanism. Dietary intake has been found to have a profound impact on the structure and activity of the vast number of microorganisms that reside in the human gut (33). Previous studies have shown that anthocyanidins have significantly enhanced the growth of Bifidobacterium spp (34), and anthocyanidins and metabolites may play a positive regulatory effect on the intestinal bacterial flora (34, 35). Anatomically, the biliary system is connected to the intestine and therefore the intestinal flora may function in this way (36). However, it is necessary to further research to determine whether the anthocyanidins directly affect the intestinal flora. Previous study have shown that anthocyanidins prevent normal cell transformation by acting on PI3K/Akt and NFκB pathways, inhibit COX-2 and iNOS expression, and achieve antioxidation by regulating the expression of second-stage antioxidant enzymes through the Nrf2/ARE signaling system (37). Then, by targeting the MAPK pathway and AP-1, anthocyanidins regulate the expression of cancer-related genes by inhibiting RTK activity and it signaling cascade pathway, leading to cell-cycle arrest and DNA repair, thereby preventing cancer (37).

Our study has several salient strengths. First, it utilizes data from a large prospective cohort of over 150,000 participants recruited from 10 screening centers nationwide in the United States, with ample follow-up time for ascertainment of outcome events. Second, this represents the first analysis leveraging the PLCO trial to investigate associations between anthocyanidin intake and biliary cancer risk. Third, our findings were corroborated through a series of sensitivity analyses, underscoring the robustness of the results. Finally, we revealed an intriguing inverse association between total anthocyanidin intake and biliary cancer risk that was more pronounced among alcohol consumers compared with nondrinkers.

Several limitations should be acknowledged in our study. First, the assessment of anthocyanidins diet based on a one-time questionnaire may introduce nondifferential bias as participants’ dietary habits could change over the course of follow-up. Second, although subgroup analyses were conducted to explore potential effect modifiers such as age, sex, BMI, smoking status, and history of diabetes mellitus, it is important to note that the limited number of biliary cancer cases might have resulted in insufficient power to detect significant interactions between total anthocyanidins diet and these factors. Third, our study population consisted of individuals ages 55 to 74 years from the United States, who may have different dietary habits and lifestyles compared with other age groups or populations with diverse dietary patterns. Therefore, caution should be exercised when generalizing our findings to other populations or age groups. Fourth, in light of the absence of a statistically significant association between the consumption of total anthocyanidins and various anatomic sites of biliary cancer in this analysis, it is essential to acknowledge that this lack of significance could potentially be due to limitations in sample size, resulting in insufficient statistical power. Therefore, a cautious and judicious approach is advisable when investigating the risk factors associated with different types of biliary cancer. Finally, it seems unlikely that the influence of residual confounders on observed events can be completely avoided. Although we comprehensively adjusted for numerous direct risk factors affecting gallbladder cancer in Model 3, the model lost statistical significance after further incorporating total energy intake, physical activity, and education level in a sensitivity analysis.

In conclusion, the findings from the PLCO trial indicate an inverse association between a diet rich in total anthocyanidins and the risk of biliary cancer. This suggests that increasing the consumption of total anthocyanidins through dietary sources may have a beneficial effect in reducing the risk of biliary cancer. Notably, examining specific food sources high in anthocyanidins could provide insights into the consistent associations observed across different anthocyanidins. Further studies should prioritize analyses of anthocyanidin-rich foods in relation to biliary cancer risk.

No disclosures were reported.

L. Xiang: Conceptualization, resources, writing–original draft, writing–review and editing. D. Wu: Conceptualization, resources, writing–original draft, writing–review and editing. Z. Xu: Data curation, software. Y. Tang: Software, formal analysis. H. He: Formal analysis. Y. Wang: Investigation, visualization. H. Gu: Supervision, methodology, writing–review and editing. L. Peng: Conceptualization, resources, writing–original draft, project administration, writing–review and editing.

We thank the NIH PLCO study group and the NCI for access to NCI's data collected by the PLCO Cancer Screening Trial.

This work was supported by The General Project of Chongqing Natural Science Foundation, Chongqing Science and Technology Commission, China [cstc2021jcyj-msxmX0153 (L. Peng)], [cstc2021jcyj-msxmX0112 (Y. Wang)], and [CSTB2022NSCQ-MSX1005 (H. Gu)], and Kuanren Talents Project of the Second Affiliated Hospital of Chongqing Medical University in China [kryc-yq-2110 (H. Gu)].

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.