Abstract P219: An Elevated Consumption of Ultra-Processed Food Modifies the Association of Body Mass Index With All-Cause Mortality: Prospective Analyses From the Moli-Sani Study

Abstract

Introduction: Body mass index (BMI) has J-shaped associations with all-cause mortality, but it is unclear whether diet can modify the impact of BMI on mortality. Hypothesis: We tested the hypothesis that a large dietary share of ultra-processed food (UPF) can act as an effect modifier of this association. Methods: Longitudinal analysis on 22,836 women and men (mean age 55±12 y) from the Moli-sani Study (2005-2010) with BMI values ≥18.5 kg/m 2 . Food intake was assessed by a 188-item FFQ. UPF was defined by the NOVA classification, and calculated as the proportion (%) of UPF in the total weight of food eaten (g/d). The Food Standards Agency nutrient profiling system (FSAm-NPS) dietary index was used to evaluate participants’ diet quality. To test for a non-linear, continuous relationship of BMI with mortality, we used multivariable Cox regression analysis with BMI modelled as restricted cubic splines and used the value of 25 kg/m 2 as the reference value. Main analyses were stratified by baseline low or high UPF consumption (i.e. below or above our population sex-specific median of UPF consumption, respectively). Results: Over a median follow-up of 12.2 y, 2,230 all-cause deaths occurred. In multivariable-adjusted models also including the FSAm-NPS dietary index and UPF, BMI had a J-shaped association with mortality (p-value<0.001; p for non-linearity<0.001); amongst lower UPF consumers, BMI values ≥33.6 kg/m 2 were associated with increased death risk, while BMI values ≤25 kg/m 2 were not ( Figure ). Higher UPF consumers experienced greater mortality risk both at higher and lower BMI values ( Figure ; p for interaction across UPF intake groups=0.016). Conclusions: In this cohort of adult Italians, UPF consumption acts as an effect modifier of the association between BMI and mortality, regardless of the overall diet quality. These findings suggest that a body weight considered to be healthy, when associated with a large dietary share of highly processed foods, is unlikely to provide substantial long-term health advantages.