Answer · confidence: partial

Do swap studies give different death risks when counting calories versus grams?

Research question: In multinational and European prospective cohorts, do models that substitute UPF servings with minimally processed foods, energy-matched versus gram-matched, report different all-cause mortality hazard ratios, and what are their confidence intervals?

The records show that in the DRECE Spanish cohort, replacing 10% of energy from UPF with minimally processed foods was associated with a 14% reduction in all-cause mortality ¹. No confidence interval for that estimate appears in the provided facts ¹. In the same cohort, substituting UPF with processed foods, whether matched by energy or by grams, gave hazard ratios below 1 that did not reach statistical significance ². The only gram-based mortality analysis in the facts, NutriNet-Santé, expressed UPF as a percentage of weight, and no gram-matched minimally processed substitution estimate is reported ⁵. The newest dated fact is from April 22, 2026, about five months before the current date of October 2026 ⁴.

Checked 2026-10-21. Only documented and supported facts are used.

Dig deeper on this question

Sources

[1] supported

In the DRECE Spanish cohort, replacing 10% of energy from UPF with 10% of energy from unprocessed or minimally processed foods was associated with a 14% reduction in all-cause mortality.

"A reduction of 14 % (in percentage of energy) and 6 % (in g) in all-cause mortality was estimated ( P -value = 0·018 and P -value = 0·002, respectively)"

Ultra-processed food intake and all-cause mortality: DRECE cohort study - PMC

[2] documented

Substituting UPF with processed foods gave hazard ratios below 1 that did not reach statistical significance, whether matched by energy or by grams.

"the hazard ratio when UPF were substituted by processed foods was less than 1 but did not achieve statistical significance (neither in percentage of energy nor in grams)"

Ultra-processed food intake and all-cause mortality: DRECE cohort study - PMC

[3] supported

Across the study, each 10% increase in energy from UPF was associated with a 15% higher hazard of all-cause mortality (HR 1.15, 95% CI 1.03 to 1.27).

"For every 10 % of the energy intake from UPF consumption, an increase of 15 % in the hazard of all-cause mortality was observed (HR 1·15; (95 % CI 1·03, 1·27); P -value = 0·012)."

Ultra-processed food intake and all-cause mortality: DRECE cohort study - PMC

[4] supported

In the substitution models, the cohort-associated fraction rose to 9.67% for colon cancer and 7.30% for colorectal cancer incidence.

"In substitution models, the CAF increased to 9.67% (4.36; 14.85) for colon cancer and 7.30% (3.05; 11.48) for CRC incidence."

Assessing the role of ultra-processed foods in colorectal cancer incidence: insights from the EPIC cohort (2026)

[5] supported

The NutriNet-Santé mortality analysis expressed UPF intake as a percentage of weight, which is a gram-based metric, rather than as an energy-based percentage.

"NutriNet-Santémortalityanalysisusedpercentageofweightinits"

Beyond the label: a methodological and geographic synthesis of ultra-processed food health outcomes (2026)

[6] supported

Fang et al. reported a modest overall association between ultra-processed food and mortality with a hazard ratio of about 1.04, with the association concentrated in processed meats and sweetened beverages.

"reported that the modest overall association (HR ≈ 1.04)"

Beyond the label: a methodological and geographic synthesis of ultra-processed food health outcomes (2026)