Abstract Body: Introduction: Greater religiosity, or degree of involvement in religious activities, associates with better CV health profiles. Yet, studies of the association between religiosity and incident CV events have yielded mixed results. The southeastern US is characterized by strong religious practice but has among the highest rates of incident heart failure (HF). The relationship between religiosity and incident HF is not well characterized, particularly among low-income populations.

Methods: We studied 23,893 Black or White Southern Community Cohort Study participants (median age 53 years, 70% Black, 64% women, 70% annual household income <$15k) without prevalent HF or CAD at enrollment (2002-2009). Religiosity, defined by religious service attendance, was ascertained at enrollment via a single item adapted from the Duke University Religion Index. Incident HF was determined using Centers for Medicare and Medicaid Services Research Identifiable Files (ICD-9/-10 codes 428.x or I50, I110) through December 31, 2016. We quantified the association of religious attendance with incident HF via Cox models adjusted for demographics, HF risk factors (HTN, HLD, DM, BMI, CVA), health behaviors (smoking, alcohol, physical activity, DASH diet), socioeconomic (income, education, employment, neighborhood deprivation, rurality) and psychosocial factors (marital status, social support, depressive symptoms, anti-depressant use).

Results: Religious service attendance was common: >1/week (24%), 1/week (27%), <1/week (33%), and never (15%). Participants attending >1/week were more likely to be older and female, had relatively higher SES and more comorbidities, but better health behaviors and psychosocial support. HF developed in 6081 participants (25%; median 10-yr follow-up). Adjusting for demographics and HF risk factors, attendance >1/week associated with lower HF risk (HR 0.88 [0.82,0.94] and HR 0.92 [0.84,1.00] compared to <1/week or never, respectively). Attendance 1/week compared to <1/week showed similar association (HR 0.91 [0.85, 0.97]). Further adjustment for health behaviors, socioeconomic and psychosocial factors attenuated the association to non-significance. There was no interaction between religiosity and race.

Conclusion: Greater frequency of religious service attendance associated with lower incident HF risk adjusting for demographics and HF risk factors; however, this association attenuated with adjustment for health behaviors, socioeconomic and psychosocial factors.