Fish consumption, fish oil supplements and risk of atherosclerosis in the Tromsø study

Abstract

Background: <br>Whether long-chain n–3 PUFAs of marine origin have an anti-atherogenic effect in the general population has hardly been studied. In this population-based study, we hypothesized that fatty fish and fish oil intake protect against development of novel atherosclerotic plaques and is associated with reduced plaque size.<br> Methods: <br>We obtained questionnaire-based information on fish consumption and carotid ultrasonography from 3900 persons aged 45–74 years. The questionnaires were validated by measuring serum concentrations of PUFAs and triglycerides in a subgroup. At follow-up seven years later, 2983 (76%) went through a second ultrasound scanning. Logistic regression and general linear models were used to analyze the outcome (plaque presence and plaque area) as a function of fish consumption, including analyses stratified on fish oil supplements. <br>Results: <br>At baseline, lean fish intake < 1 time/week vs. 1–1.9 times/week was associated with risk of plaque (OR 1.34, 95% CI 1.03–1.76). Fatty fish intake and use of fish oil supplements were not statistically significantly associated with atherosclerosis at baseline. In persons without plaque at baseline, total fish consumption ≥3 times/week vs. 1–1.9 times/week was associated with risk of novel plaque (OR 1.32, 95% CI 1.01–1.73) and larger plaque area (1.76 mm2 vs. 1.46 mm2 , p = 0.02) at follow-up. Adjustments for use of fish oil supplements had no impact on the associations, and no interactions were seen between total, fatty or lean fish consumption and fish oil intake.<br> Conclusions: <br>We found no protective effect of fatty fish eating or fish oil supplements on atherosclerotic plaque formation or plaque area in a general population. Lean fish consumption was associated with a reduced risk for plaque in cross-sectional analysis, suggesting that the beneficial effects of fish consumption on atherosclerosis may be mediated through other mechanisms than n-3 PUFAs.