Ultra-Processed Food Consumption and Hypertension Risk in the REGARDS Cohort Study.

BACKGROUND

This study examined the longitudinal association between ultraprocessed food (UPF) consumption and overall hypertension risk and explored the contribution of UPF to racial disparities in hypertension risk.

METHODS

We analyzed data from 5957 participants from the REGARDS cohort study (Reasons for Geographic and Racial Disparities in Stroke) who were free from hypertension during visit 1 (2003-2007), had complete dietary information at visit 1, and completed visit 2 (2013-2016). UPF consumption was measured using the Nova classification system and operationalized as percent calories and grams. The main outcome was incident hypertension. Logistic regression was used for analysis.

RESULTS

Thirty-six percent of participants developed hypertension at visit 2. Results showed a positive linear relationship between UPF and hypertension incidence. Aggregate model results showed that participants in the highest UPF consumption quartile had 23% greater odds of incident hypertension compared with the lowest quartile. Multivariable results showed that Black and White participants in the highest consumption quartile had 1.26 (95% CI, 0.92-1.74) and 1.22 (95% CI, 1.01-1.47) greater odds of hypertension compared with those in the lowest quartile, respectively. Analyses using UPF consumption as percent grams showed similar aggregate results; however, race-stratified results differed. Findings were not statistically significant among White participants (odds ratio, 1.09 [95% CI, 0.89-1.33]) but showed higher risk of hypertension among Black participants in the highest versus lowest UPF quartiles (odds ratio, 1.43 [95% CI, 1.01-2.02]).

CONCLUSIONS

This study demonstrated that high consumption of UPF is associated with increased hypertension risk. Further research is warranted to better understand differences in the intakes of UPF subgroups that may underpin the racial differences in hypertension incidence observed with different UPF metrics.