(P13-045-24) Ultra-Processed Food Intake and Mortality in the NIH-AARP Diet and Health Study

Objectives: We investigated the association of ultra-processed food (UPF) intake with mortality risk using alternative methods of assigning Nova classification.

Methods: Our study population included 318,889 men and 221,607 women in the US NIH-AARP Diet and Health Study who were 50 to 71 years of age at baseline in 1995-1996. To increase study rigor and reproducibility, we used two methods to classify food frequency questionnaire (FFQ) items according to the Nova system. Our first approach relied on expert consensus. Our second, novel approach involved disaggregating FFQ items into FNDDS food codes, which were further broken down into ingredient codes, and assigned a Nova classification. We used multivariable Cox regression models, adjusted for mortality risk factors including smoking, to estimate hazard ratios (HR) and 95% confidence intervals (CI) for UPF intake (continuous, g/day) and mortality risk. Finally, we used two non-consecutive 24-hour dietary recalls, collected in a subgroup (n=1908), as the reference to calibrate FFQ-based risk estimates in the full cohort.

Results: During follow-up (median 22.9 years) a total of 178,111 men and 101,766 women died. Those who consumed more UPF tended to be younger and have higher body mass index (BMI) but lower Healthy Eating Index (HEI)-2015 scores, indicating worse diet quality. Using the food code-based approach to define UPF intake (g/day), multivariable adjusted HR for death among men and women were 1.10 (95% CI: 1.09 – 1.11) and 1.09 (95% CI: 1.08 – 1.10) comparing the 90^th to the 10^th percentile of UPF intake, respectively. Risk estimates adjusted for BMI and HEI-2015 and those based on the expert consensus approach were similar (all P-values < 0.01). HR estimates using regression calibration to account measurement error in the FFQ were 1.14 (95% CI: 1.13 – 1.15) and 1.15 (95% CI: 1.14 – 1.17) comparing the 90^th to 10^th percentile of UPF intake, for men and women, respectively. Stronger positive associations were observed for deaths due to heart disease and diabetes, but no association was observed for deaths due to cancer.

Conclusions: In this large prospective study, UPF intake was associated with higher risk of all-cause as well heart disease and diabetes-specific mortality. Future research is needed to understand how food processing impacts health and longevity.

Funding Sources: NIH Intramural Research Program