Table 2.
Meta-analysis results for the association between the presence of gallstones and the risk of BTC by subgroups
| Subgroup | No. of studies | OR (95% CI)1 | I2 value (%) | p for heterogeneity |
|---|---|---|---|---|
| All studies | 26 | 4.38 (3.23, 5.93) | 91.2 | <0.001 |
| Study design | ||||
| Cohort study | 7 | 3.17 (2.28, 4.39) | 79.0 | <0.001 |
| Case-control study | 19 | 5.04 (3.36, 7.56) | 90.5 | <0.001 |
| Sex | ||||
| Male | 9 | 3.40 (2.70, 4.28) | 35.8 | 0.132 |
| Female | 9 | 4.26 (2.75, 6.59) | 84.5 | <0.001 |
| Geographic area | ||||
| Asia | 15 | 5.25 (3.50, 7.86) | 82.4 | <0.001 |
| Non-Asia2 | 11 | 3.58 (2.17, 5.91) | 95.1 | <0.001 |
| Study period3 | ||||
| Before 2000 | 8 | 5.39 (2.57, 11.34) | 95.5 | <0.001 |
| Around 2000 | 7 | 2.67 (2.10, 3.39) | 38.6 | 0.135 |
| After 2000 | 7 | 5.21 (2.13, 12.74) | 85.7 | <0.001 |
| No records | 4 | 5.73 (2.61, 12.61) | 87.3 | <0.001 |
| Measurement of gallstones | ||||
| Medical records with imaging studies | 8 | 7.09 (3.87, 12.98) | 64.5 | 0.004 |
| Medical records without imaging studies | 16 | 3.81 (2.48, 5.85) | 93.9 | <0.001 |
| No records | 2 | 3.47 (2.88, 4.18) | 17.1 | 0.272 |
| Study quality4 | ||||
| High NOS | 14 | 3.99 (2.85, 5.59) | 75.6 | <0.001 |
| Low NOS | 12 | 4.81 (2.87, 8.05) | 94.9 | <0.001 |
| Adjustment for age, yes | 21 | 3.71 (2.66, 5.16) | 90.6 | <0.001 |
| Adjustment for sex, yes | 20 | 3.92 (2.77, 5.55) | 91.2 | <0.001 |
| Adjustment for comorbidities, yes | 7 | 3.05 (1.84, 5.05) | 75.7 | <0.001 |
| Adjustment for lifestyle factors, yes5 | 7 | 4.84 (1.95, 11.98) | 85.5 | <0.001 |
| Adjustment for education, yes | 1 | 9.42 (3.56, 24.91) | - | - |
| Adjustment for geographic areas, yes | 4 | 7.34 (2.28, 23.62) | 87.6 | 0.000 |
BTC, biliary tract cancer; OR, odds ratio; CI, confidence interval; NOS, Newcastle-Ottawa scale.
1
OR refers to a summary estimate of effects based on a random-effects model.
2
Non-Asia including USA and European areas.
3
Study period was defined by the study’s starting point (a) and ending point (b). Before 2000, (a) and (b) are both before 2000; around 2000, (a) is before 2000 but (b) is after 2000; after 2000, (a) and (b) are both after 2000.
4
Quality scores greater than or equal to the median value were judged as a high NOS (≥7).
5
Adjustment for lifestyle factors such as alcohol, smoking, body mass index, etc.