Method: We examined the prospective association between cumulative average UV-B flux exposure and RA risk among 222,792 women followed in the NHS 1976-2008 (born 1921-46, n=107,116) and NHSII 1989-2007 (born 1947-1964, n=115,676). Residential locations of NHS participants were available for 1976, 1986-2006 and for NHSII participants from 1989-2007. Incident RA cases diagnosed from 1976-2008 for NHS and 1989-2007 for NHSII were confirmed by medical record review. Average annual UV-B flux, a composite measure of mean UV-B radiation level based on latitude, altitude, and cloud cover, was estimated for all nurses according to state of residence during cohort follow-up. The relationship between tertile of cumulative average UV-B exposure and risk of RA was estimated using Cox proportional hazards models separately in each cohort, then combined using meta-analysis random effects models. Age- and multivariable-adjusted models, including age, parity, breast feeding, body mass index, pack-years smoking, physical activity, vitamin D intake, alcohol consumption, race, husband’s education and oral contraceptive use, were used.
Result: We identified and confirmed 1146 incident RA cases. Although no significant heterogeneity between the two cohorts was found, the results were slightly different in the two cohorts. In the older women, NHS, there was a small but significant decrease in RA risk associated with increased UV-B exposure. In the cohort of younger women, however, this effect was not observed. When meta-analytically combined, cumulative UV-B flux exposure was not significantly associated with RA risk overall (Table 1).
Conclusion: Results suggested an inverse association between UV-B exposure and incident RA in NHS, but there was no evidence of an association in NHSII or in combined analyses. Limitations include lack of data on time spent outside and sunscreen use, which could vary with nurse age. The overall results do not support an association between cumulative exposure to ambient UV-B light and RA risk; however, statistical power may have been limited to detect a modest association.
Disclosure: E. V. Arkema, None; K. Bertrand, None; F. Laden, None; A. A. Qureshi, Centers for Disease Control, Novartis, Abbott, 5, Amgen/Pfizer, 2 ; E. W. Karlson, None; K. H. Costenbader, None.