Ultra-processed foods and incident cardiovascular disease and hypertension in middle-aged women

Abstract

Purpose Ultra-processed food (UPF) intake has increased in recent decades, yet limited knowledge of long-term effects on cardiovascular health persists and sex-specific data is scant. We determined the association of UPF intake with incident cardiovascular disease (CVD) and/or hypertension in a population-based cohort of women. Methods In the Australian Longitudinal Study on Women’s Health, women aged 50–55 years were prospectively followed (2001–2016). UPFs were identified using NOVA classification and contribution of these foods to total dietary intake by weight was estimated. Primary endpoint was incident CVD (self-reported heart disease/stroke). Secondary endpoints were self-reported hypertension, all-cause mortality, type 2 diabetes mellitus, and/or obesity. Logistic regression models assessed associations between UPF intake and incident CVD, adjusting for socio-demographic, medical comorbidities, and dietary variables. Results We included 10,006 women (mean age 52.5 ± 1.5; mean UPF intake 26.6 ± 10.2% of total dietary intake), with 1038 (10.8%) incident CVD, 471 (4.7%) deaths, and 4204 (43.8%) hypertension cases over 15 years of follow-up. In multivariable-adjusted models, the highest [mean 42.0% total dietary intake] versus the lowest [mean 14.2% total dietary intake] quintile of UPF intake was associated with higher incident hypertension [odds ratio (OR) 1.39, 95% confidence interval (CI) 1.10–1.74; p = 0.005] with a linear trend (ptrend = 0.02), but not incident CVD [OR 1.22, 95% CI 0.92–1.61; p = 0.16] or all-cause mortality (OR 0.80, 95% CI 0.54–1.20; p = 0.28). Similar results were found after multiple imputations for missing values. Conclusion In women, higher UPF intake was associated with increased hypertension, but not incident CVD. These findings may support minimising UPFs within a healthy diet for women.

Ultra-Processed Foods: A Narrative Review of the Impact on the Human Gut Microbiome and Variations in Classification Methods

“…Of the studies in Table 3 , eight used FFQs [ 81 , 82 , 83 , 84 , 85 , 86 , 87 , 88 ], seven used 24 h dietary recalls [ 77 , 79 , 80 , 89 , 90 , 91 , 92 ], and two used a combination of methods (FFQs and food records) [ 76 , 78 ] to collect food information. Each of these methods have been validated for use in dietary research, but limitations exist that make food processing classification difficult in some cases.…”

“…Each of these methods have been validated for use in dietary research, but limitations exist that make food processing classification difficult in some cases. When there was ambiguity in the classification of foods, studies resolved these items by several methods: group discussion and consensus [ 83 , 86 , 90 , 91 , 92 ], discussion and defaulting to a lower level of processing [ 77 , 81 , 82 , 85 , 86 , 88 , 89 ], expanding discussion to an external group [ 87 ], and utilizing external resources such as grocery store labels [ 84 , 87 , 90 , 91 ]. As shown, some studies used a combination of these methods, while others did not describe how they worked to resolve discrepancies.…”

“… Tomato Sauce: Cordova et al categorized cooked tomato (as an Italian pizza ingredient) as NOVA 1 if it was fresh, but NOVA 4 if on a commercial pizza [ 76 ]. Pant et al counted tomato sauce and tomato paste as NOVA 4 [ 86 ]. Samuthpongtorn et al called tomato sauce without sufficient detail non-ultra-processed in their main analysis [ 87 ].…”

“…Pant, 2023 [86] Referred to Machado, 2019 [96] and Lane, 2023 [85] 101-item FFQ N/A Food items from the FFQ were classified into one of the four NOVA groups and cross-checked between two independent reviewers. If classification was unclear, the NNPAS 2011-12 was consulted or lesser degree of processing was selected.…”