Abstract P218: Quality of Plant-Based Diets is Differently Associated With Cardiovascular Mortality: A Longitudinal Analysis From the Moli-Sani Study

Introduction: Vegetarian diets are characterized by the absence of some animal foods (e.g. red and processed meats), and a high consumption of plant-based foods. However, plant-based foods can include foods with varying nutritional value and divergent health effects. Hypothesis: We assessed the hypothesis that three different pro-vegetarian (PVG) food patterns defined as general (gPVG), healthful (hPVG) and unhealthful (uPVG) are differently associated with the risk of cardiovascular disease (CVD) mortality in adult Italians. Established CVD risk factors ( e.g. blood lipids, inflammatory markers) were analyzed as potential biological mechanisms underpinning these associations. Methods: Longitudinal analysis on 22,895 women and men (mean age 55±12 y) from the Moli-sani Study (2005-2010). Food intake was assessed by a 188-item food frequency questionnaire. A pro-vegetarian food pattern was constructed by assigning positive scores to plant foods and reverse scores to animal foods. Both a healthful and an unhealthful pro-vegetarian food patterns, distinguishing between healthy ( e.g . fruits, vegetables, legumes) and less-healthy plant foods ( e.g . fruit juices, potatoes, sugary beverages), were built up. Cox proportional hazard regression models were used to estimate the association between dietary exposures (quarters of) and mortality. Hazard ratios (HRs) were calculated with their 95% confidence intervals (CIs) for each quarter, considering Q1 as the reference category. Results: After a median follow-up of 12.2 y (272,960 person/years), a total of 792 CVD deaths were ascertained. In multivariable-adjusted analyses controlled for common known risk factors, higher adherence to a gPVG was not associated with CVD mortality (HR=0.92; 0.75-1.13 for Q4 vs. Q1). Increasing adherence to a hPVG was linked to lower risk of CVD mortality (HR=0.78; 0.62-0.98 for Q4 vs. Q1), whereas a higher uPVG led to an increased risk (HR=1.28; 1.04-1.57). Altered levels of subclinical inflammation ( i.e. C-reactive protein) accounted for 13.4% (p=0.012) of the association of the hPVG with CVD mortality. Conclusions: In this large Italian general population, increasing adherence to a general pro-vegetarian food pattern was not associated with CVD mortality. Only participants preferring healthful vegetarian foods experienced a significant protection against CVD mortality risk, possibly through a favorable modulation of the inflammatory pathway. Consistently, a large dietary share of unhealthful vegetarian foods, mostly highly processed, was associated with increased risk of CVD mortality. We conclude that it is not sufficient to adopt a general vegetarian diet, since it is the quality of the plant food consumed that is critical to achieve vegetarian diet-related benefits on CVD mortality.