Introduction
Biliary tract cancers (BTCs), including bile duct cancer (BDC), gallbladder cancer (GC), and ampullary cancer (AC), are associated with poor prognosis and rising incidence. Identifying modifiable and demographic risk factors for each subtype is essential for early detection. However, comprehensive cohort studies evaluating multiple BTC subtypes simultaneously remain limited, especially for AC. This study identifies risk factors for each BTC simultaneously using data from a large Japanese regional database.
Aims & Methods
Using data from the Shizuoka Kokuho Database (SKDB), a large longitudinal insurance claims dataset covering over 2.5 million individuals, we analyzed risk factors for each BTC subtype. Adults undergoing health checkups between April 2012 and September 2021 were eligible. We excluded those with <1-year follow-up or a prior cancer diagnosis. Incident BTCs were identified via ICD-10 codes. Multivariable Cox regression models included age, sex, lifestyle factors, laboratory data, and comorbidities. Predictive performance was evaluated using Harrell’s C-index with sensitivity analyses conducted by removing key variables.
Results
Among 625,513 eligible individuals, 1,433 developed BDC, 838 GC, and 205 AC over a median 6.2-year follow-up. Annual incidence per 100,000 person-years was 41.89 (BDC), 24.50 (GC), and 5.90 (AC).
In BDC, significant risk factors included older age (age categories and corresponding hazard ratios are shown in Table 1), male sex (HR: 1.84 [1.64–2.07]), elevated γ-GTP ≥100 IU/L (HR: 2.94 [2.38–3.64]), elevated AST ≥50 IU/L (HR: 1.65 [1.26–2.15]), diabetes (HR: 1.31 [1.12–1.54]), hepatitis B (HR: 1.73 [1.05–2.86]), hepatitis C (HR: 1.58 [1.08–2.29]), and common bile duct stones (HR: 3.40 [2.21–5.23]).
In GC, older age (detailed in Table 1), male sex (HR: 1.25 [1.06–1.48]), elevated AST (HR: 1.32 [1.07–1.62]), physical inactivity (HR: 1.23 [1.06–1.43]), gallstones (HR: 3.85 [2.96–5.01]), and choledocholithiasis (HR: 1.94 [1.02–3.71]) were significant. Although GC incidence was higher in women (n=503) than men (n=335), the adjusted analysis identified male sex as a risk. Notably, sex differences disappeared in participants without gallstones but were pronounced in those with gallstones.
In AC, older age (see Table 1), male sex (HR: 1.59 [1.17–2.15]), γ-GTP ≥100 IU/L (HR: 3.25 [1.97–5.35]), diabetes (HR: 1.83 [1.25–2.67]), rheumatic disease (HR: 2.09 [1.10–3.98]), and common bile duct stones (HR: 3.81 [1.21–11.94]) were significant.
Model Performance: In BDC, the full model’s C-index was 0.752; removing age reduced it most (−0.0563). In GC, the C-index was 0.702; age again contributed most. In AC, the C-index was 0.732; exclusion of age or rheumatic disease slightly reduced predictive accuracy.
| Table 1. |
|
|
|
|
|
|
|
|
|
|---|
| Bile duct cancer |
|
| Gallbladder cancer |
|
| Ampullary cancer
|
|
|
Category
| HR
| 95% CI
| p-value
| HR
| 95% CI
| p-value
| HR
| 95% CI
| p-value
|
<50 years
| 1 | - | <0.001
| 1
| - | <0.001
| 1
| -
| <0.001
|
50-<60 years
| 2.99
| 1.30-6.89 | - | 3.21
| 1.32-7.84
| -
| 0.79
| 0.11-5.61
| -
|
60-<70 years
| 8.39
| 3.96-17.8 | - | 5.26
| 2.32-11.9
| -
| 5.46
| 1.33-22.5 | -
|
70-<80 years
| 18.65
| 8.79-39.6 | - | 10.22
| 4.51-23.2
| -
| 9.97
| 2.41-41.3 | -
|
≥80 years
| 34.19
| 16.0-72.9 | - | 15.55
| 6.78-35.6
| -
| 10.38
| 2.44-44.1 | -
|
Conclusion
This study demonstrates distinct risk profiles for BDC, GC, and AC using a single population-based cohort. Novel associations—such as the link between rheumatic disease and AC—emerged. These findings may guide future preventive strategies and inform risk-stratified screening approaches.