Frequency of attendance at religious services, cardiovascular disease, metamolic risk factors and dietary intake in Americans: An age-stratified exploratory analysis

Background: Few data have been published on the association of attendance at religious services with cardiovascular morbidity and dietary and metabolic risk, factors in representative samples of populations despite a known inverse association with mortality and smoking. Objective: To test the null hypothesis that frequency of attendance at religious services is unrelated to prevalence or levels of cardiovascular disease, dietary and metabolic risk factors. Design: Cross-sectional survey of a large national sample. Participants: American men and women aged 20 years and over with complete data in the Third National Health and Nutrition Examination Survey (N = 14,192). Measurements: Self-reported frequency of attendance at religious services, history of doctor-diagnosed diseases, food intake frequency, 24-hour dietary intake, health status, socio-demographic variables and measured serum lipids and body mass index. Results: Weekly attenders were significantly less likely to report stroke, even after adjusting for multiple variables only in African American women OR = 0.35, 95% Cl 0.19-0.66, p < 0.01. No association was seen for heart attack or diabetes. Fish intake at least weekly was more common in weekly attenders, significantly so only in African American women (odds ratio 1.24, 95% CI 1.01-1.58, p < 0.05) and in older Mexican American men (odds ratio 2.57, 95% Cl 1.45-2.57, p < 0.01). In linear regression analyses, no significant independent associations were seen between attendance frequency and serum lipid levels or dietary intake of energy, or fat in g and % of kcal. Conclusion: Hypotheses generated by these analyses are that in African American women stroke is less prevalent and weekly fish intake more prevalent among weekly attenders than others and that there are no significant independent associations of serum lipids, dietary intake, prevalent CHD, or diabetes with frequency of attendance of religious services. Independent testing of these hypotheses in other samples is needed.