Associations between metabolic syndrome and cholangiocarcinoma risk: A large-scale population-based cohort study

Abstract

Background and Aims:

This large-scale, population-based cohort study examined the associations between metabolic syndrome and cholangiocarcinoma risk, including its intrahepatic and extrahepatic forms.

Approach and Results:

A total of 4,932,211 adults aged ≥40 years participated in a government-initiated health checkup program (2012–2017), which collected lifestyle data, anthropometric measurements, and biochemical tests. Follow-up continued until 2021, with data linkage to National Cancer and Death Registries to ascertain the occurrence of cholangiocarcinoma and obtain vital status information. Fine and Gray models accounted for competing risks. During 35,879,371 person-years of follow-up, 6117 cholangiocarcinoma cases were identified, with an incidence rate of 17.05 (95% CI: 15.90–18.20) per 100,000 person-years. Individuals with metabolic syndrome had significantly higher incidences of both intrahepatic and extrahepatic cholangiocarcinoma (p<0.0001). The multivariate-adjusted HR for cholangiocarcinoma among those with metabolic syndrome was 1.20 (1.14–1.27). Stratification analyses by age, sex, liver enzyme levels, and comorbidities consistently demonstrated an increased cholangiocarcinoma risk among individuals with metabolic syndrome. A dose-response relationship was observed, with a higher number of metabolic components correlating with an elevated cholangiocarcinoma risk, even after accounting for all-cause mortality as a competing risk. The adjusted subdistribution HRs ranged from 1.16 (95% CI: 1.02–1.32) for individuals with one metabolic component to 1.67 (95% CI: 1.45–1.94) for those with five (p for trend <0.0001).

Conclusions:

The positive association between metabolic syndrome and cholangiocarcinoma risk suggests that managing metabolic risk factors might reduce the occurrence of both intrahepatic and extrahepatic cholangiocarcinoma.

INTRODUCTION

Cholangiocarcinoma encompasses a heterogeneous group of malignancies originating in the bile ducts, either within the liver (intrahepatic) or those connected to it (extrahepatic).¹ Its aggressive nature, nonspecific symptoms, and late-stage diagnosis pose significant clinical challenges, underscoring the need for effective prevention strategies. The incidence and mortality of cholangiocarcinoma have shown a concerning upward trend in recent decades.^2–4 Despite identifying some risk factors, the rising incidence highlights the importance of identifying new risk factors to provide a comprehensive understanding of their etiology and prevention.^5,6

The global rise in obesity has significantly increased the prevalence of metabolic disease across populations and age groups,^7–9 with nearly 40% of adults over 40 years old affected.^7,8 Metabolic syndrome, characterized by elevated waist circumference, blood pressure, triglyceride, and fasting glucose levels, and reduced HDL cholesterol levels,¹⁰ is increasingly recognized as a risk factor for certain gastrointestinal cancers.^11–15 Although some studies have explored the link between metabolic syndrome and cholangiocarcinoma,^{11,12,16–21} the findings have been inconsistent,^16,21,22 with many employing a cross-sectional design,^12,17,18,20 and few studies have comparatively analyzed intrahepatic and extrahepatic cholangiocarcinoma.^16,17,19,20

To address this gap, we conducted a large-scale population-based cohort study to investigate the association between metabolic syndrome and cholangiocarcinoma, including intrahepatic and extrahepatic presentations. Using data from government-initiated Taiwanese adult health check-ups with ~5 million participants, we conducted comprehensive analyses through computerized data linkage with national registries, providing a substantial number of incident cholangiocarcinoma cases for investigation. The study also evaluated the associations of individual metabolic components with cholangiocarcinoma risk.

METHODS

Study design and population

The retrospective cohort study included 5,793,333 individuals who participated in the health check-up program from 2012 to 2017, initiated by the Health Promotion Administration, Ministry of Health and Welfare in Taiwan. The program targeted adults aged 40 years and older and involved triennial examinations, including physical assessments, interviews, and fasting blood collection. With a participation rate of ~30%, the program aimed to manage health proactively. After excluding individuals who had been diagnosed with cholangiocarcinoma or any other cancers prior to the time of study enrollment, as well as those with missing cardiometabolic information or lacking identification numbers for data linkage, the final cohort comprised 4,932,211 individuals. The study was approved by the Institutional Review Board of National Yang Ming Chiao Tung University (YM107033E-6), with informed consent waived due to the use of anonymous and deidentified data. All research procedures were conducted in accordance with the Declarations of Helsinki and Istanbul.

Data collection and linkage

To ascertain the incidence of cholangiocarcinoma, vital status, and health conditions of the study participants, we performed computerized data linkage and integrated 3 additional national health registries provided by the Taiwanese Ministry of Health and Welfare: the National Health Insurance Database, National Cancer Registration, and Death Certification System. These nationwide registries are mandatory and offer near-universal coverage, storing complete, accurate, and up-to-date data, making them valuable resources for biomedical research.^23,24 Participants’ data were linked to these registries using the citizenship identification number and birthdates.

Interview and blood collection

Lifestyle data were collected through standardized questionnaires covering cigarette smoking, alcohol consumption habits, and physical activity. Anthropometric examinations included body weight, height, waist circumference, and blood pressure. Biochemical tests measured serum levels of triglycerides, cholesterol, glucose, and HDL-cholesterol. Serum levels of ALT and AST levels were also measured. Elevated serum ALT levels were defined as ≥40 IU/L for men and ≥31 IU/L for women; elevated AST levels were defined as ≥37 IU/L for men and ≥31 IU/L for women.²⁵ We used anthropometric measurements and serum biochemical profiles at enrollment to define metabolic syndrome and for subsequent analyses.

Definition of metabolic syndrome and potential comorbidities

Metabolic syndrome was defined based on the National Cholesterol Education Program’s Adult Treatment Panel III criteria,²⁶ requiring at least 3 of the following components: central obesity (waist circumference ≥90 cm for men and ≥80 cm for women, adjusted for the Asian population), elevated triglycerides (≥150 mg/dL), low HDL-C (<40 mg/dL in men and <50 mg/dL in women), high blood pressure (systolic blood pressure ≥130 mm Hg or diastolic pressure ≥85 mm Hg), and impaired fasting glucose (≥110 mg/dL). Medications to control these factors were considered, alongside self-reported medical history and previous medical diagnosis. The presence of gallstones, biliary tract diseases, and cirrhosis was identified based on International Classification of Diseases (ICD) codes recorded in the National Health Insurance Database, with detailed information listed in Supplemental Table S1, http://links.lww.com/HEP/J757. The hepatic steatosis index (HSI) was calculated using the formula 8×(ALT/AST ratio) + body mass index (+2, if female; +2, if diabetes), categorized into <30, 30–36, and ≥36, with HSI≥36 considered indicative of steatotic liver disease as reported.²⁷

Follow-up and ascertainment of cholangiocarcinoma

We utilized computerized data linkage with the National Cancer Registration Profiles and the National Death Certification system to determine the occurrence of cholangiocarcinoma and participant vital status from January 1, 2012, to December 31, 2021. Cholangiocarcinoma was identified using ICD-9 code 155.1 and ICD-10 code C22.1 for intrahepatic and ICD-9 code 156.1 and ICD-10 code C24.0 for extrahepatic cholangiocarcinoma. The diagnosis of cholangiocarcinoma was based on elevated tumor markers (CEA and CA19.9), imaging studies such as sonography, CT, and MRI, or histological confirmation via biopsy.

Subcohort included chronic hepatitis B and C virus infections

Previous studies have reported associations between hepatitis B or HBV and HCV infections and cholangiocarcinoma.^24,28–30 Among the 4,932,211 study participants, 727,074 (14.7%) were tested for HBsAg and antibodies against anti-HCV). Additional analyses were conducted to clarify the association between metabolic syndrome and cholangiocarcinoma.

Statistical analysis

Person-years of follow-up were calculated from study entry until cholangiocarcinoma diagnosis, death, or the last computerized data linkage (ie, December 31, 2021), whichever came first. Incidence rates of cholangiocarcinoma, including intrahepatic and extrahepatic forms, were determined by dividing the number of newly diagnosed cases by the person-years of follow-up. Cumulative risks for participants with and without metabolic syndrome were estimated using the Kaplan-Meier method, with statistical significance examined using log-rank tests. Crude and adjusted HRs with 95% CIs for metabolic syndrome were estimated using Cox’s proportional hazards models. The proportionality assumptions of Cox models were examined, and no violation was detected. We also performed additional subgroup analyses to examine metabolic syndrome and the associated risk of cholangiocarcinoma by comparing to those without according to potential confounders. For the subcohort with HBsAg and anti-HCV tests, adjusted HRs were estimated to include chronic HBV and HCV infection status. Individual metabolic syndrome components—central obesity, triglyceride, HDL-cholesterol levels, hypertension, and fasting glucose—and the number of metabolic factors were analyzed for their association with cholangiocarcinoma risk, adjusting for competing events. We used Fine and Gray models to account for competing risks, considering the rate of cholangiocarcinoma among participants who had not yet experienced the outcome or a competing event (all-cause death or any cancer type). Subdistribution hazard ratios (SHRs) with 95% CIs were estimated to assess the associations between metabolic syndrome, its individual components, and cholangiocarcinoma risk. Participants were followed until the occurrence of cholangiocarcinoma or the end of the study period. Those who remained event-free by the end of follow-up were considered censored. The outcome variable was defined as a composite categorical event status (0 = censored, 1 = cholangiocarcinoma, 2 = competing risk), with time (years) measured from study entry to the event or study end. All statistical analyses were conducted with statistical significance defined as a 2-sided p value of <.05. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc.).

RESULTS

The baseline characteristics of 4,932,211 individuals participating in the health examination check-ups are summarized in Table 1. Of these, 2,029,759 (41.1%) were classified as having metabolic syndrome. Individuals with metabolic syndrome were older than those without, with mean ages of 61.1 and 56.3 years, respectively. Participants with metabolic syndrome also showed a higher prevalence of cardiometabolic risk factors, including elevated waist circumferences, triglyceride levels, abnormal HDL-cholesterol levels, high blood pressure, and elevated glucose levels. They also had increased ALT or AST levels and a higher prevalence of gallstones (11.6% vs. 9.0%). Liver steatosis (HSI>36) was observed in 34.5% of the population, with 54.3% of those with metabolic syndrome and 20.7% without metabolic syndrome classified as having steatosis.

TABLE 1.

Baseline characteristics of study participants

Baseline characteristics	Total population (N= 4,932,211) N (%)	Metabolic syndrome (N=2,029,759) N (%)	Nonmetabolic syndrome (N=2,902,452) N (%)
Age (y)
Mean ± SD	58.3±12.0	61.1±11.9	56.3±11.7
40–49	1,329,719 (27.0)	374,573 (18.5)	955,146 (32.9)
50–59	1,502,823 (30.5)	574,233 (28.3)	928,590 (32.0)
60–69	1,180,103 (23.9)	581,832 (28.7)	598,271 (20.6)
≥70	919,566 (18.6)	499,121 (24.6)	420,445 (14.5)
Sex
Male	2,179,218 (44.2)	948,764 (46.7)	1,230,454 (42.4)
Female	2,752,993 (55.8)	1,080,995 (53.3)	1,671,998 (57.6)
Cigarette smoking
No	4,385,294 (88.9)	1,787,644 (88.1)	2,597,650 (89.5)
Yes	546,917 (11.1)	24,215 (11.9)	304,802 (10.5)
Alcohol consumption
No	4,778,560 (96.9)	1,958,034 (96.5)	2,820,526 (97.2)
Yes	153,651 (3.1)	71,725 (3.5)	81,926 (2.8)
Regular physical exercise
No	2,693,404 (54.6)	1,133,922 (55.9)	1,559,482 (53.7)
<2.5 h/wk	1,660,010 (33.7)	673,471 (33.2)	986,539 (34.0)
≥2.5 h/wk	578,797 (11.7)	222,366 (11.0)	356,431 (12.3)
Body mass index (kg/m2)
Mean±SD	24.7±3.9	26.5±3.9	23.5±3.4
<18.5	152,196 (3.1)	19,148 (0.9)	133,048 (4.6)
18.5–22.9	1,527,392 (31.0)	315,439 (15.5)	1,211,953 (41.8)
23–24.9	2,206,166 (44.7)	976,420 (48.1)	1,229,746 (42.4)
≥25	1,046,457 (21.2)	718,752 (35.4)	327,705 (11.3)
Waist circumferences (cm)
Mean±SD	68.4±25.0	76.3±24.8	63.0±23.6
Triglyceride levels (mg/dL)
Mean±SD	142.3±135.7	192.0±177.7	107.6±79.4
<150	3,423,113 (69.4)	926,193 (45.6)	2,496,920 (86.0)
≥150	1,509,098 (30.6)	1,103,566 (54.4)	405,532 (14.0)
HDL-C levels (mg/dL)a
Mean±SD	54.5±17.7	47.7±16.2	59.2±17.2
Normal	3,532,055 (71.6)	981,836 (48.4)	2,550,219 (87.9)
Abnormal	1,400,156 (28.4)	1,047,923 (51.6)	352,233 (12.1)
Blood pressure (mm Hg)
Systolic
Mean±SD	125.2±18.5	136.7±19.3	125.2±18.5
<130	2,536,728 (51.4)	664,634 (32.7)	1,872,094 (64.5)
≥130	2,395,483 (48.6)	1,365,125 (67.3)	1,030,358 (35.5)
Diastolic (mm Hg)
Mean±SD	76.9±12.0	82.5±12.9	76.9±12.0
<85	3,509,168 (71.2)	1,230,075 (60.6)	2,279,093 (78.5)
≥85	1,423,043 (28.9)	799,684 (39.4)	623,359 (21.5)
Fasting glucose (mg/dL)
Mean±SD	107.1±42.9	124.2±54.9	95.2±26.0
<100	2,986,453 (60.6)	663,439 (32.7)	2,323,014 (80.0)
≥100	1,945,758 (39.5)	1,366,320 (67.3)	579,438 (20.0)
Serum ALT levels (IU/L)b
Mean±SD	28.2±35.5	32.1±38.7	25.4±32.8
Men <40 IU/L; women <31 IU/L	3,958,564 (80.3)	1,482,309 (73.0)	2,476,255 (85.3)
Men ≥40 IU/L; women ≥31 IU/L	973,647 (19.7)	547,450 (27.0)	426,197 (14.7)
Serum AST levels (IU/L)c
Mean±SD	27.5±31.1	29.5±35.2	26.0±27.7
Men <37 IU/L; women <31 IU/L	4,136,953 (83.9)	1,607,203 (79.2)	2,529,750 (87.2)
Men ≥37 IU/L; women ≥31 IU/L	795,258 (16.1)	422,556 (20.8)	372,702 (12.8)
Gallstones
No	4,435,891 (89.9)	1,794,729 (88.4)	2,641,162 (91.0)
Yes	496,320 (10.1)	235,030 (11.6)	261,290 (9.0)
Biliary tract disease
No	4,802,307 (97.4)	1,963,326 (96.7)	2,838,981 (97.8)
Yes	129,904 (2.6)	66,433 (3.3)	63,471 (2.2)
Hepatic steatosis indexd
<30	1,109,219 (22.5)	165,739 (8.2)	943,480 (32.5)
30≤ HSI ≤36	2,119,586 (43.0)	761,333 (37.5)	1,358,253 (46.8)
>36	1,703,406 (34.5)	1,102,687 (54.3)	600,719 (20.7)

^a Normal HDL is defined as men ≥40 mg/dL and women ≥50 mg/dL; abnormal defined as men <40 mg/dL and women <50 mg/dL.

^b Elevated ALT levels are defined as ≥40 IU/L for men and ≥31 IU/L for women.

^c Elevated AST levels are defined as ≥37 IU/L for men and ≥31 IU/L for women.

^d Hepatic steatosis index (HSI) = 8×(ALT/AST ratio) + body mass index (+2, if female; +2, if diabetes mellitus).

Abbreviations: HDL-C, High-density lipoprotein cholesterol; HSI, hepatic steatosis index.

Incidence of cholangiocarcinoma according to metabolic syndrome status

During a follow-up period spanning over 35,879,371 person-years, 6117 cases of cholangiocarcinoma occurred, resulting in an incidence rate of 17.05 (95% CI: 15.90–18.20) per 100,000 person-years. As shown in Table 2, the majority of cases (75.6%) were intrahepatic, with 4625 cases of intrahepatic and 1492 cases of extrahepatic cholangiocarcinoma, corresponding to incidences of 12.89 and 4.16 per 100,000 person-years, respectively. Participants with metabolic syndrome exhibited significantly higher incidences of both intrahepatic and extrahepatic cholangiocarcinoma (p<0.0001).

TABLE 2.

Incidence rates of intrahepatic and extrahepatic cholangiocarcinoma by metabolic syndrome status

Outcomes by metabolic syndrome status	Event numbers	Person-years of follow-up	Incidence rates (95% CI) per 100,000 person-years	Crude HR (95%CI)	Adjusted HR (95% CI)
Total cholangiocarcinoma
Nonmetabolic syndrome	2793	21,192,505	13.18 (12.17–14.19)	1.00 (reference)	1.00 (reference)
Metabolic syndrome	3324	14,686,866	22.63 (21.30–23.96)	1.72 (1.63–1.81)	1.20 (1.14–1.27)
Total	6117	35,879,371	17.05 (15.90–18.20)
Intrahepatic cholangiocarcinoma
Nonmetabolic syndrome	2128	21,193,389	10.04 (9.16–10.93)	1.00 (reference)	1.00 (reference)
Metabolic syndrome	2497	14,687,931	17.00 (15.85–18.15)	1.69 (1.60–1.79)	1.18 (1.11–1.25)
Total	4625	35,881,320	12.89 (11.89–13.89)
Extrahepatic cholangiocarcinoma
Nonmetabolic syndrome	665	21,194,921	3.14 (2.64–3.63)	1.00 (reference)	1.00 (reference)
Metabolic syndrome	827	14,689,580	5.63 (4.97–6.29)	1.80 (1.62–1.99)	1.28 (1.16–1.42)
Total	1492	35,884,501	4.16 (3.59–4.73)

Note: Adjusted: age, sex, smoking, drinking, exercise, gallstone, biliary tract disease, elevated liver enzyme (elevated ALT levels defined as ≥40 IU/L for men and ≥31 IU/L for women; elevated AST levels defined as ≥37 IU/L for men and ≥31 IU/L for women).

Cumulative risk analysis illustrated consistently elevated risks of intrahepatic and extrahepatic cholangiocarcinoma in participants with metabolic syndrome (Figures 1A–C). As depicted in Figure 1A, the cumulative risk of cholangiocarcinoma differed by metabolic syndrome status after the median follow-up of 7.6 years. By the end of the follow-up period, 0.22% of participants with metabolic syndrome and 0.13% of those without metabolic syndrome had developed cholangiocarcinoma (p<0.0001). The cumulative risk for intrahepatic and extrahepatic cholangiocarcinoma was 0.17% and 0.05% for those with metabolic syndrome, compared to 0.10% and 0.03% for those without (p<0.0001) (Figures 1B, C).

FIGURE 1.

(A) Cumulative risk of CC by the presence of metabolic syndrome. (B) Cumulative risk of ICC by the presence of metabolic syndrome. (C) Cumulative risk of ECC by the presence of the metabolic syndrome. Abbreviations: CC, cholangiocarcinoma; ECC, extrahepatic cholangiocarcinoma; ICC, intrahepatic cholangiocarcinoma.

Metabolic syndrome and its association with cholangiocarcinoma risk across subgroups

The association between metabolic syndrome and cholangiocarcinoma risk, including both intrahepatic and extrahepatic cases, was examined across various subgroups (Figures 2A–C). As illustrated in Figure 2A, the adjusted HRs for metabolic syndrome and cholangiocarcinoma incidence consistently exceeded one in each group, ranging from 1.04 to 2.09, highlighting a positive association between metabolic syndrome and cholangiocarcinoma risk across different subgroups. We categorized age into 40–49, 50–59, and ≥60 years, allowing for a clear distinction between early and late-onset cases. Notably, metabolic syndrome appeared to have a stronger association with cholangiocarcinoma risk in the 50–59 age group. Furthermore, among individuals without cirrhosis at baseline, the adjusted HRs for metabolic syndrome remained consistent: 1.20 (95% CI: 1.14–1.27) for cholangiocarcinoma overall, 1.18 (95% CI: 1.11–1.26) for intrahepatic cholangiocarcinoma, and 1.27 (95% CI: 1.15–1.41) for extrahepatic cholangiocarcinoma. Figures 2B, C depict the association between metabolic syndrome and intrahepatic or extrahepatic cholangiocarcinoma across subgroups, showing a consistent pattern of elevated risk. Although the association did not consistently reach statistical significance in all subgroups and exhibited borderline significance in some cases, metabolic syndrome overall demonstrated a positive association with intrahepatic or extrahepatic cholangiocarcinoma risks, with HRs above one compared to nonmetabolic syndrome individuals.

FIGURE 2.

(A) Associations of metabolic syndrome and CC, stratified by subgroups. (B) Associations of metabolic syndrome and ICC, stratified by subgroups. (C) Associations of metabolic syndrome and ECC, stratified by subgroups. Abbreviations: CC, cholangiocarcinoma; ECC, extrahepatic cholangiocarcinoma; ICC, intrahepatic cholangiocarcinoma.

Subcohort analyses integrating HBV and HCV tests

In the subcohort analyses, including HBsAg and anti-HCV serostatus as covariates, metabolic syndrome remained significantly associated with cholangiocarcinoma, with an adjusted HR (95% CI) of 1.22 (1.01–1.47). Furthermore, the adjusted HR (95% CI) was 1.25 (1.01–1.55) and 1.13 (0.78–1.64) for intrahepatic and extrahepatic cholangiocarcinoma, respectively. Interestingly, HBsAg and anti-HCV seropositive increased intrahepatic cholangiocarcinoma with adjusted HRs (95% CI) of 2.70 (2.07–3.53) and 1.76 (1.22–2.55), respectively. However, significant associations between HBV and HCV infection on extrahepatic cholangiocarcinoma risk were not observed, with adjusted HRs of 0.83 (0.40–1.73) and 1.66 (0.80–3.44), respectively. Figures 2A–C illustrate the associations between metabolic syndrome and cholangiocarcinoma, stratified by chronic hepatitis B or C virus infection status. Due to the limited number of participants with dual HBsAg and anti-HCV seropositivity, separate analyses for this group were not feasible. However, the multivariable-adjusted HRs consistently showed an increased risk of cholangiocarcinoma with metabolic syndrome across different serostatus categories. In the HBsAg-positive, anti-HCV-negative subgroup, the association with extrahepatic cholangiocarcinoma risk was below one but not statistically significant, likely due to the small sample size.

Metabolic syndrome, metabolic components, and cholangiocarcinoma risks accounting for competing events

In the analysis, we accounted for all-cause mortality and all cancer types as competing risks to evaluate the associations between metabolic syndrome and cholangiocarcinoma. Participants with metabolic syndrome had a significantly increased risk of cholangiocarcinoma, with an SHR of 1.21 (95% CI: 1.15–1.28) (Table 3). The HR estimates for the association between metabolic syndrome and cholangiocarcinoma remained consistent across different modeling approaches, with an adjusted HR of 1.20 (95% CI: 1.14–1.27) using standard Cox regression (Table 2) and a subdistribution HR of 1.21 (95% CI: 1.15–1.28) using the Fine and Gray method (Table 3). A higher number of metabolic components was associated with an increasing risk of cholangiocarcinoma, with adjusted SHRs ranging from 1.16 (95% CI: 1.02–1.32) for one metabolic component to 1.67 (95% CI: 1.45–1.94) for five components, compared to individuals without metabolic syndrome (p for trend<0.0001).

TABLE 3.

Association between metabolic syndrome, its components, and cholangiocarcinoma risk accounting for competing events using Fine and Gray analysis

	All-cause mortality as competing events			All cancer types as competing events
	Cholangiocarcinoma	Intrahepatic cholangiocarcinoma	Extrahepatic cholangiocarcinoma	Cholangiocarcinoma	Intrahepatic cholangiocarcinoma	Extrahepatic cholangiocarcinoma
Metabolic factors	Subdistribution HR (95% CI)	Subdistribution HR (95% CI)	Subdistribution HR (95% CI)	Subdistribution HR (95% CI)	Subdistribution HR (95% CI)	Subdistribution HR (95% CI)
Metabolic syndrome
No	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
Yes	1.21 (1.15–1.28)	1.19 (1.12–1.27)	1.29 (1.16–1.43)	1.21 (1.14–1.27)	1.18 (1.11–1.26)	1.28 (1.15–1.42)
Central obesity
No	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
Yes	1.14 (1.08–1.20)	1.16 (1.09–1.23)	1.08 (0.97–1.21)	1.12 (1.06–1.18)	1.14 (1.07–1.21)	1.06 (0.95–1.18)
Triglycerides levels
Normal	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
Elevated	0.96 (0.90–1.01)	0.90 (0.85–0.97)	1.17 (1.06–1.30)	0.94 (0.89–0.98)	0.88 (0.83–0.93)	1.14 (1.03–1.26)
HDL-C levels
Normal	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
Elevated	1.12 (1.06–1.18)	1.15 (1.08–1.22)	1.01 (0.91–1.13)	1.14 (1.08–1.21)	1.18 (1.11–1.25)	1.04 (0.93–1.17)
Hypertension
No	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
Yes	1.29 (1.20–1.38)	1.30 (1.20–1.40)	1.26 (1.10–1.45)	1.28 (1.19–1.36)	1.28 (1.19–1.39)	1.26 (1.09–1.44)
Fasting glucosea
Normal	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
Prediabetes	1.09 (1.02–1.16)	1.06 (0.98–1.15)	1.18 (1.03–1.35)	1.08 (1.01–1.15)	1.05 (0.97–1.14)	1.16 (1.01–1.33)
Diabetes	1.38 (1.30–1.46)	1.36 (1.27–1.46)	1.43 (1.27–1.61)	1.41 (1.33–1.49)	1.39 (1.30–1.48)	1.47 (1.30–1.65)
No. metabolic risk factors
0	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
1	1.16 (1.02–1.32)	1.14 (.0.99–1.32)	1.23 (0.94–1.60)	1.15 (1.01–1.31)	1.13 (.0.98–1.31)	1.22 (0.94–1.59)
2	1.26 (1.11–1.42)	1.27 (1.11–1.46)	1.20 (0.93–1.56)	1.24 (1.10–1.40)	1.26 (1.10–1.45)	1.19 (0.92–1.54)
3	1.44 (1.27–1.62)	1.43 (1.24–1.64)	1.47 (1.15–1.90)	1.42 (1.25–1.60)	1.41 (1.22–1.61)	1.46 (1.13–1.88)
4	1.38 (1.22–1.57)	1.32 (1.14–1.53)	1.57 (1.21–2.04)	1.36 (1.19–1.54)	1.30 (1.12–1.50)	1.55 (1.19–2.01)
5	1.67 (1.45–1.94)	1.68 (1.42–1.98)	1.65 (1.22–2.24)	1.64 (1.42–1.90)	1.65 (1.39–1.94)	1.63 (1.20–2.21)
	p for trend<0.0001	p for trend<0.0001	p for trend<0.0001	p for trend<0.0001	p for trend<0.0001	p for trend<0.0001

Note: Adjusted: age, sex, smoking, drinking, exercise, gallstone, biliary tract disease, elevated liver enzyme (elevated ALT levels defined as ≥40 IU/L for men and ≥31 IU/L for women; elevated AST levels defined as ≥37 IU/L for men and ≥31 IU/L for women).

^a Normal is defined as fasting glucose <100 mg/dL; prediabetes is defined as fasting glucose ≥100–<125 mg/dL or history of diabetes or medical diagnosis (ICD9=250); Diabetes is defined as fasting glucose ≥125 mg/dL or history of diabetes or medical diagnosis (ICD9=250).

Among the individual components of metabolic syndrome, central obesity, abnormal HDL, hypertension, and impaired fasting glucose, were significantly associated with an elevated risk of cholangiocarcinoma risk (p<0.0001) after accounting for competing risks. Compared to participants with normal fasting glucose, those with prediabetes and diabetes had a higher risk of cholangiocarcinoma, with adjusted SHRs (95% CI) of 1.09 (1.02–1.16) and 1.38 (1.30–1.46), respectively (Table 3). In contrast, elevated triglyceride levels were not significantly associated with cholangiocarcinoma risk but showed a borderline significant association with intrahepatic cholangiocarcinoma. Overall, both metabolic syndrome and an increasing number of metabolic components were consistently associated with higher risks for intrahepatic and extrahepatic cholangiocarcinoma (p<0.001).

DISCUSSION

Our large-scale cohort study found that metabolic syndrome increased the associated risk for cholangiocarcinoma, encompassing intrahepatic and extrahepatic forms. Subgroup analyses consistently reaffirmed these associations across various groups. Furthermore, our study highlighted that most individual components of metabolic risk factors and the cumulative number of these factors were linked to increased cholangiocarcinoma risk. These findings underscore the potential benefits of interventions aimed at weight reduction and managing lipid, glucose, and hypertension levels. Such measures have the dual potential of preventing metabolic syndrome and reducing the risk of cholangiocarcinoma.

In our analyses, we employed the Fine and Gray method to account for competing risks, including all-cause mortality and other cancer types. The minimal discrepancies between these estimates suggest that competing events had a negligible impact on the observed associations. To further reinforce the robustness of our findings, we systematically examined metabolic syndrome, its individual components, and the cumulative number of metabolic components while accounting for competing risks, demonstrating the stability and reliability of our results.

Importantly, metabolic syndrome was associated with cholangiocarcinoma even after adjusting for chronic HBV or HCV infection. Recent global efforts to eliminate these viruses have shifted the etiology of liver cancer beyond chronic viral infections. In the coming decades, obesity and metabolic syndrome are expected to become the primary drivers of several cancers. Our study underscores the importance of metabolic control, which may reduce the incidence of cancers, including cholangiocarcinoma.

Intrahepatic and extrahepatic cholangiocarcinoma may share some risk factors but also exhibit distinct ones,^28,29 suggesting partially different etiologies.³¹ Studies have found positive associations between HBV infections and intrahepatic cholangiocarcinoma,^28,30 while HCV infection has been linked to an increased risk of intrahepatic but not extrahepatic cholangiocarcinoma.²⁹ Our findings on HBsAg and anti-HCV serostatus were in line with previous reports. Similarly, alcohol consumption and NAFLD were more strongly linked to intrahepatic cholangiocarcinoma, whereas biliary tract diseases were more strongly associated with extrahepatic cholangiocarcinoma.^20,28,30,32 These differences may explain the metabolic component associations observed in our study between intrahepatic and extrahepatic cholangiocarcinoma.

In our study, we observed notable discrepancies between intrahepatic and extrahepatic cholangiocarcinoma, although the overall pattern of metabolic syndrome and its components showed similar associations with both intrahepatic and extrahepatic cholangiocarcinoma. For example, central obesity was positively associated with intrahepatic but not with extrahepatic cholangiocarcinoma. Conversely, elevated triglyceride levels were negatively associated with intrahepatic but positively associated with extrahepatic cholangiocarcinoma. The reasons behind the inverse association between triglyceride levels and intrahepatic cholangiocarcinoma are not well understood, but these findings align with recent research indicating that persistently elevated triglyceride levels may decrease cholangiocarcinoma risk.²²

The associations between metabolic syndrome and cholangiocarcinoma have been limitedly studied. Most existing research consists of case-control studies,^12,17,18,20 which do not effectively establish causal relationships. Few studies have adopted prospective designs, and these have yielded inconsistent results,^16,21,22 likely due to an insufficient number of cholangiocarcinoma cases and a consequent lack of statistical power.¹⁶ One large-scale study employed 2 measurements over 2 years, categorizing participants based on changes in metabolic syndrome status.¹⁹ This study found that participants with persistent metabolic syndrome for over 2 years had a 1.07-fold increased risk of developing cholangiocarcinoma compared to those without metabolic syndrome. However, the study involved relatively young participants, who generally had a low prevalence of metabolic syndrome and a lower risk of developing cholangiocarcinoma. Despite adjusting for age, residual confounding may still have been present. Additionally, the study did not differentiate between intrahepatic and extrahepatic cholangiocarcinoma in its analyses, and it did not account for the role of chronic hepatitis B or C infections, which requires further clarification.

In our study population, ~40% of participants had abnormal glucose levels and could be classified as prediabetic. Both prediabetes and diabetes were associated with an increased risk of both intrahepatic and extrahepatic cholangiocarcinoma, showing a significant trend compared to those with normal fasting glucose levels. Our findings align with previous studies indicating that diabetes elevates the risk of cholangiocarcinoma.^33–35 These results suggest that glucose control might be a potential strategy for reducing cancer risk. One biological explanation for hyperglycemia facilitating neoplastic proliferation involves insulin resistance, compensatory hyperinsulinemia, and elevated levels of insulin-like growth factors. Insulin and these growth factors may promote cholangiocyte proliferation and inhibit apoptosis, thereby contributing to cancer development.^36,37

Obesity induces inflammation, characterized by increased accumulation and inflammatory polarization of immune cells in various tissues, including adipose tissue.^38,39 Visceral adipose tissue, in particular, is metabolically active and releases various inflammatory molecules. Chronic low-grade inflammation involving both innate and adaptive immune cells can increase cell cycle rates and decrease tumor suppressor function. IL-6, which is extensively studied, is expressed by cholangiocarcinoma cells and stromal inflammatory cells and may contribute to the development of cholangiocarcinoma by promoting cell proliferation and survival.^40,41

Metabolic syndrome is a complex condition characterized by elevated blood pressure, triglycerides, and glucose levels, as well as reduced HDL cholesterol levels. Previous studies have demonstrated that lipid-lowering or anti-inflammatory drugs, such as statins and aspirin, are associated with reduced cholangiocarcinoma risks^42–44 and improved survival among affected patients.^45–47 These findings provide valuable insights into the role of chemoprevention in cholangiocarcinoma and suggest that effective metabolic control is crucial for cancer prevention. While our study did not account for medication history, we observed that an increasing number of metabolic syndrome components was significantly associated with cholangiocarcinoma risk. This underscores the importance of comprehensive metabolic management and highlights the need for future research to explore the role of specific medications in cholangiocarcinoma prevention.

Our study has several notable strengths. First, it utilized data from government-initiated health check-ups, ensuring the inclusion of individuals with validated and regularly updated information on comorbidities, cholangiocarcinoma incidence, and mortality. We used individual-level measurements and laboratory data collected systematically. This comprehensive data set extends to detailed records of metabolic risk factors and lifestyle habits recorded as part of the health check-up program. The large sample size and rigorous adherence to standard metabolic syndrome definitions provided the estimates of the link between metabolic syndrome and cholangiocarcinoma. Moreover, our study benefits from complete follow-up facilitated by nationwide registries, effectively minimizing the potential for reverse causation. These registries provide consistently updated and thorough data, ensuring high accuracy across all study participants. The large population-based design offers a unique opportunity to explore the incidence of rare cancers such as cholangiocarcinoma. With a substantial number of cholangiocarcinoma cases, spanning both intrahepatic and extrahepatic presentations, we could robustly estimate the associations between metabolic syndrome and these cancers. While not all study participants underwent tests for chronic hepatitis B and C virus infections, our analyses were restricted to those with available information on HBsAg and anti-HCV status. Remarkably, metabolic syndrome consistently emerged as a risk factor for cholangiocarcinoma after adjustment of chronic hepatitis viral infections. Finally, a key strength of our study is the use of Fine and Gray analyses to account for competing risks, including all-cause mortality and other cancer types. This methodological approach further reinforces the robustness of our findings, demonstrating that the observed associations between metabolic syndrome and cholangiocarcinoma persist even after accounting for competing events.

Despite these strengths, there are limitations to consider. Liver fluke infection is a risk factor for cholangiocarcinoma. However, it is rare in the Taiwanese population thus it was not include in the analyses. While our study leveraged data from government-initiated check-ups, it’s crucial to acknowledge that participation in these check-ups was voluntary rather than mandatory. Consequently, individuals who opted to participate may possess a heightened level of health awareness compared to the general population, potentially minimizing generalizability given the “healthy participant” bias. We minimized potential confounders by adjusting covariates, although residual confounding may still exist. Additionally, metabolic syndrome and its components were determined at baseline, without accounting for their potential evolution over time, especially in participants actively managing their lipid, glucose, or blood pressure levels. The use of prescribed drug records from health insurance databases to define metabolic syndrome, rather than individualized seromarker data, introduces further challenges due to the diversity of medications and their potential use for unrelated comorbidities. Nonetheless, the incidence of cholangiocarcinoma was ascertained independently of metabolic syndrome status through computerized data linkage. This independence could lead to nondifferential misclassification, potentially underestimating the true effect. Such a conservative bias reinforces the robustness of our conclusion that metabolic syndrome increases the risk of cholangiocarcinoma, emphasizing the importance of managing these risk factors. Finally, our study lacked key data on the duration of metabolic syndrome and its relationship to cholangiocarcinoma risks. We were also unable to calculate the FIB-4 index, as platelet counts were not available, limiting our ability to account for fibrosis in the analysis. Furthermore, individualized data regarding cholangiocarcinoma histology and tumor size were not accessible, precluding an exploration of whether patients with metabolic syndrome present with larger tumor sizes at diagnosis compared to those without. Nonetheless, previous studies indicate that adenocarcinoma is the predominant histological type of cholangiocarcinoma in Taiwan.

In conclusion, we observed a positive association between metabolic syndrome and the incidence of cholangiocarcinoma, including both intrahepatic and extrahepatic forms, in a large population-based cohort. Effective management of metabolic syndrome and its components may reduce the risk of metabolic syndrome-related comorbidities and cholangiocarcinoma. Further research is warranted to elucidate the underlying mechanisms and to develop targeted prevention strategies.

Supplementary Material

hep-83-261-s001.docx ^{(17.6KB, docx)}