| Alcohol consumption | Never smoking | Ever smoking | ||
| Exp ca/co* | OR† (95% CI) | Exp ca/co* | OR† (95% CI) | |
| EIRA (ACPA-positive RA) | ||||
| Drinkers | 159/269 | 1.0‡ (Referent) | 456/492 | 1.7 (1.3 to 2.1) |
| Non-drinkers | 47/76 | 1.1 (0.7 to 1.7) | 73/33 | 4.2 (2.6 to 6.6) |
| p interaction¶ | <0.001 | |||
| CACORA (ACPA-positive RA) | ||||
| Drinkers | 63/181 | 1.0‡ (Referent) | 183/298 | 1.9 (1.3 to 2.7) |
| Non-drinkers | 20/29 | 1.9 (1.0 to 3.6) | 42/25 | 4.8 (2.7 to 8.5) |
| p interaction¶ | <0.001 | |||
*Number of exposed (Exp) cases (ca) and controls (co). †Adjusted for sex, age and residential area (EIRA only). ‡Reference category. ¶The proportion attributable to interaction between smoking status and lack of alcohol consumption was 0.6 (95% CI 0.5 to 0.7) for EIRA and 0.4 (95% CI 0.3 to 0.5) for CACORA.
ACPA, antibodies to citrullinated peptide antigens; CACORA, Case–Control Study on Rheumatoid Arthritis; EIRA, Epidemiological Investigation of Rheumatoid Arthritis; RA, rheumatoid arthritis.