Table 3.
Pooled ORs and corresponding 95% CIs of pancreatic cancer according to alcohol consumption in strata of selected covariates in the International Pancreatic Cancer Case–Control Consortium (PanC4)
| Frequency, drinks per daya |
P value for interaction | |||||
| 0 to <1 (referent) |
≥6 |
ORb (95% CI) | ||||
| Cases, n (%) | Controls, n (%) | Cases, n (%) | Controls, n (%) | |||
| Overall | 3587 (64.4) | 7044 (59.6) | 308 (5.5) | 622 (5.3) | 1.46 (1.16–1.83) | |
| Sex | ||||||
| Men | 1675 (53.4) | 3077 (45.2) | 283 (9.0) | 590 (8.7) | 1.43 (1.07–1.91) | 0.708 |
| Women | 1912 (78.7) | 3967 (79.2) | 25 (1.0) | 32 (0.6) | 1.64 (0.86–3.11) | |
| Age (years) | ||||||
| <65 | 1676 (60.8) | 3638 (56.8) | 208 (7.5) | 423 (6.6) | 1.55 (1.21–2.00) | 0.600 |
| ≥65 | 1911 (68.0) | 3406 (63.0) | 100 (3.6) | 199 (3.7) | 1.37 (0.92–2.04) | |
| Cigarette smoking | ||||||
| Never smokers | 1600 (78.6) | 3890 (76.0) | 32 (1.6) | 105 (2.1) | 1.31 (0.68–2.54) | 0.677 |
| Ex-smokers | 1094 (58.4) | 1785 (48.6) | 117 (6.2) | 214 (5.8) | 1.67 (1.23–2.26) | |
| Current smokers | ||||||
| <20 cigarettes/day | 560 (56.2) | 958 (46.9) | 76 (7.6) | 172 (8.4) | 1.46 (0.94–2.27) | |
| ≥20 cigarettes/day | 209 (47.2) | 193 (34.6) | 72 (16.3) | 109 (19.5) | 0.99 (0.42–2.34) | |
| Race/ethnicity | ||||||
| Non-Hispanic White | 2877 (62.8) | 4747 (55.8) | 259 (6.0) | 475 (5.6) | 1.48 (1.15–1.92) | 0.523 |
| Non-Hispanic Black | 174 (57.6) | 619 (57.4) | 32 (10.8) | 104 (9.6) | 1.83 (1.02–3.28) | |
| Study area | ||||||
| North America | 2788 (68.8) | 4523 (63.4) | 191 (4.7) | 323 (4.5) | 1.32 (0.97–1.80) | 0.157 |
| Europe | 374 (38.7) | 1124 (39.0) | 107 (11.1) | 272 (9.4) | 1.96 (1.25–3.09) | |
| Sources of controls | ||||||
| Hospital | 1173 (64.5) | 1600 (45.8) | 110 (6.1) | 269 (7.7) | 1.66 (0.98–2.82) | 0.519 |
| Population | 2414 (64.4) | 5444 (65.4) | 198 (5.3) | 353 (4.3) | 1.37 (1.07–1.75) | |
a
Derived from cumulative lifetime consumption for study providing information (see ‘Methods’).
b
Pooled ORs were computed using random effects models. Study-specific ORs were adjusted for age, sex, race/ethnicity, education, body mass index, history of diabetes, tobacco smoking (in categories, plus a continuous term), and center for multicentric studies.
OR, odds ratio; CI, confidence interval.