Background/Purpose: Arthritis and activity limitation are risk factors for cardiovascular (CVD) morbidity and mortality. As arthritis is a major cause of activity limitation, the objective of the present study was to estimate the extent to which increased heart disease risks in persons with arthritis are at least partially mediated through activity limitation pathways and potential variations by gender.

Methods: The longitudinal Canadian National Population Health Survey (NPHS) collected information on sociodemographic variables, self-reported physician-diagnosed chronic conditions, activity limitation, and lifestyle/health behaviors every 2 years from 1994/95 through 2010/11. Cause of death records for ischaemic heart disease and heart failure were confirmed against the Canadian Vital Statistics Database. Included variables were repeated measures of arthritis, activity limitation and heart disease lagged to ensure the proper temporal sequence, and the following baseline covariates: age, sex education, race, diabetes, high blood pressure, physical activity, BMI, smoking, alcohol consumption, use of pain medications and non-CVD comorbidity index. The analysis used a novel approach integrating recent advances in causal mediation with analyses of discrete event occurrence. Direct and indirect effects of arthritis on incident heart disease were obtained by combining estimates from a discrete-time survival model of heart disease regressed on arthritis and activity limitation controlling for all covariates measured at baseline with those from logistic regression of the mediator activity limitation regressed on arthritis and the same baseline covariates, stratified by gender.

Results: The sample included 11, 655 (5282 male, 6373 female) persons aged 18+ with no history of, or prevalent heart disease reported in the first 2 cycles of the NPHS (1994/5 and 1996/97). We identified 1442 incident heart disease events reported from 1998/99 to 2010/11. After adjusting for all covariates, arthritis and activity limitation were significantly associated with incident heart disease (arthritis OR (95% CI): 1.31 (1.10-1.56), activity limitation OR (95% CI) 1.61 (1.36-1.90)) in women, while only activity limitation was significant in men (arthritis OR (95% CI): 0.99 (0.81-1.20); activity limitation OR (95% CI): 1.59 (1.30-193)). Arthritis was significantly associated with activity limitation in both men (OR (95% CI): 3.25 (2.99-3.53)) and women (OR (95% CI): 2.85 (2.67-3.05)). Combining results from both regressions yielded significant direct (OR (95% CI): 1.31 (1.06-1.58)) and indirect effects (OR (95% CI): 1.09 (1.06-1.13)) of arthritis on heart disease with a proportion mediated by activity limitation of 28.5% in women. Only the indirect effect of activity limitation was significant in men (OR (95% CI): 1.09 (1.05-1.15)).

Conclusion: Findings from this longitudinal population based study help elucidate differences in the etiology of heart disease in men and women with arthritis, and have policy implications for arthritis prevention programs with priority for targeted interventions towards reducing disability in persons living with arthritis.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/gender-specific-pathways-linking-arthritis-activity-limitation-and-incident-heart-disease-a-causal-mediation-analysis-of-the-canadian-longitudinal-national-population-health-survey/