Objective: Suggested predictors of rheumatoid arthritis (RA) include environmental exposure, such as smoking. Our purpose was to investigate potential predictors of RA in a nested case-control study based on a prospective cohort.
Method: Between 1991 and 1996, 30,447 persons were included in the Malmö Diet and Cancer Study (MDCS). Individuals who developed RA after inclusion up to 31 December 2004 were identified by linking the database to different registers. Four controls were selected for every case. Data on lifestyle factors were collected in the MDCS.
Results: We identified 172 incident cases of RA [36 men/136 women, mean age at diagnosis 63 years, 69% rheumatoid factor (RF) positive, median time from inclusion to diagnosis 5 (range 1-13) years]. In bivariate analyses, baseline smoking [odds ratio (OR) 2.02, 95% confidence interval (CI) 1.31-3.12] and a low level of formal education (i.e. ≤ 8 years; OR 2.42, 95% CI 1.18-4.93 vs. University degree) predicted subsequent development of RA. Infrequent baseline alcohol consumption was a predictor of RA (OR 3.47, 95% CI 1.91-6.30) compared to recent use (within the past month), and individuals with moderate baseline alcohol consumption (3.5-15.2 g/day vs. < 3.5 g/day) tended to have a reduced risk of RA (OR 0.48, 95% CI 0.22-1.05) in multivariate analyses, adjusted for smoking and level of education.
Conclusions: Smoking and a low level of formal education were found to be independent predictors of RA. Moderate alcohol consumption may also be associated with a reduced risk.