Characterization of the degree of food processing in the European Prospective Investigation into Cancer and Nutrition: Application of the Nova classification and validation using selected biomarkers of food processing.
Level 4 - case-series / case-control
Cross-sectional biomarker validation and dietary characterization study within a prospective cohort
PubMed 36590209 · doi:10.3389/fnut.2022.1035580
What was done
The Nova classification system (grouping foods into unprocessed/minimally processed, processed culinary ingredients, processed foods, and ultra-processed foods [UPF]) was applied to dietary data from 476,768 participants (71.5% women, mean age 51 ± 9.93 years) in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort. To validate the classification against objective markers of industrial processing, consumption of each Nova category was compared to circulating industrial elaidic acid (an industrial trans-fatty acid) and urinary 4-methyl syringol sulfate (a marker of smoked food and liquid smoke flavoring) in cohort subsamples (n = 417 to 9,460) using partial correlation analyses adjusted for age, sex, BMI, and country.
What was found
UPF contribution to daily diet ranged across countries from 7% (France) to 23% (Norway) by food weight, and from 16% (Spain and Italy) to >45% (UK and Norway) by total energy intake. Participants in the highest quartile of UPF intake were younger, taller, less educated, current smokers, more physically active, had higher total energy intake, and consumed less alcohol. Nova-defined UPF intake (% kcal/day) was positively correlated with circulating industrial elaidic acid (r = 0.54) and urinary 4-methyl syringol sulfate (r = 0.43). Unprocessed and minimally processed food intake was inversely correlated with both biomarkers (r = -0.07 for elaidic acid; r = -0.37 for 4-methyl syringol sulfate), with other Nova categories showing inverse or non-significant associations.
Why it matters
This study validates the Nova dietary classification system across European populations by demonstrating that higher self-reported ultra-processed food intake corresponds to higher circulating and urinary levels of objective industrial processing biomarkers.
Limits
The design is cross-sectional. Biomarker validation was restricted to small subsamples (n = 417 to 9,460) rather than the full cohort. Only two specific chemical markers of processing were assessed, which do not reflect the full spectrum of additives, processing contaminants, or structural alterations found across all ultra-processed foods. Dietary intake relied on self-report, which carries inherent measurement error.
Cited by
- contradicts Grinding whole wheat berries at home renders the flour ultra-processed.