OBJECTIVE: This study examines the cross-sectional association between prevalent rheumatoid arthritis (RA) and diabetes among noninstitutionalized US civilians aged >or= 60 years between 1988 and 1994.
METHODS: Using National Health and Nutrition Examination Survey III data from the National Center for Health Statistics, RA and diabetes were identified using several classification schemes. In total, 5302 survey participants aged >or= 60 years were included in logistic regression analyses taking survey weights into account. We also conducted sensitivity analyses restricting the study population to participants not recently prescribed glucocorticoids and fasting at least 8 hours prior to blood draw, as well as data incorporated from the Multiple Imputation Project.
RESULTS: Among the 5302 participants aged >or= 60, 144 participants had RA and 24 of these also were found to have prevalent diabetes. The adjusted odds ratios for the cross-sectional association between RA and diabetes ranged from 1.1 to 1.5, but did not reach statistical significance.
CONCLUSION: While this study cannot definitively rule out a modest non-null association, we can conclude that there is no evidence of a strong cross-sectional association between prevalent RA and diabetes in subjects aged >or= 60 years. Future longitudinal studies with more participants with RA are required to further evaluate a possible association between RA and the incidence of diabetes.
