Cohort studyindependent cross-checkabstract checked★ key study I-UPF-105 in-depth review
Ultra-processed food and death in 114,000 US health workers
Association of ultra-processed food consumption with all cause and cause specific mortality: population based cohort study.
Overview
Fang Z et al.. BMJ 2024 · PMID 38719536 · DOI 10.1136/bmj-2023-078476
Over about 30 years, people eating the most ultra-processed food had a 4% higher death rate; processed meat and sweet drinks drove most of it, while cancer and heart deaths were not linked.
Does the study support the specific claim? Partly · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
To link UPF intake with causes of death.
- How was the study done?
Two long-running cohort studies.
- How many people, animals or samples were studied?
114,064 adults.
- What was actually found?
Slightly higher overall mortality; no link for cancer or heart deaths; strong links for processed meat.
- How large was the observed effect?
HR 1.04 (1.01-1.07).
- What do the results mean in everyday life?
Not all ultra-processed foods are equal; processed meat and sugary drinks matter most.
- What are the study's limitations?
Observational; questionnaire-based.
- What does it say about Elawady's specific claim?
Limits the breadth of C-UPF-03.
- Do the results apply to Germany and Europe?
Moderate.
- What can explicitly not be concluded from this study?
Cannot prove that processing causes death.
Scientific in-depth review
1. Full reference
- Authors
- Fang Z et al.
- Journal
- BMJ
- Year
- 2024
- Identifier
- PMID 38719536 DOI 10.1136/bmj-2023-078476 · PMC11077436
- Verification
- verified – PubMed efetch 2026-10-26; Crossref works API 2026-10-26
- Retracted
- no
- Corrections
- none
- Extent checked
- abstract only
- Conflicts of interest and funding
COI statement in record: Competing interests: All authors have completed the ICMJE uniform disclosure form at https://www.icmje.org/disclosure-of-interest/ and declare: support from the National Institutes of Health for the submitted work; NK received a consulting fee from the Pan American Health Organization for three mont; funding listed: NCI NIH HHS
2. Research question
Is UPF intake associated with all-cause and cause-specific mortality?
3. Study design and population
- Study design
- Cohort study – Prospective cohort study (Nurses' Health Study and Health Professionals Follow-up Study)
- Population
US female nurses and male health professionals free of cancer, CVD and diabetes
- Country
USA
- Period
1984/1986-2018; median follow-up 34 and 31 years
- Sample size
74,563 women and 39,501 men; 48,193 deaths
- Inclusion and exclusion criteria
No history of cancer, CVD or diabetes at baseline
4. Methods and endpoints
- Exposure / intervention
UPF intake by FFQ every 4 years
- Comparator
Lowest quarter
- Endpoints
All-cause, cancer, CVD and other mortality
- Statistical methods
Multivariable Cox models; analyses within diet-quality strata
5. Actual results
- Results
Highest vs lowest quarter: all-cause mortality HR 1.04 (1.01-1.07); mortality from causes other than cancer or CVD 1.09 (1.05-1.13); no association with cancer or cardiovascular mortality. Rates 1472 vs 1536 per 100,000 person-years. Meat/poultry/seafood ready-to-eat products HR 1.06-1.43; sugar- and artificially sweetened beverages 1.09 (1.07-1.12). No consistent UPF association within diet-quality quarters; diet quality inversely associated within UPF quarters.
- Absolute risks
All-cause mortality 1472 vs 1536 per 100,000 person-years (difference 64; stats/C-UPF-03_mortality.py)
- Authors' conclusion
Higher UPF intake was associated with slightly higher all-cause mortality, driven by non-cancer, non-CVD causes; associations varied by UPF subgroup.
6. Strengths
- Very long follow-up with repeated diet measures
- Subgroup and diet-quality analyses
7. Weaknesses
- FFQ not designed for Nova
- Health professionals, mostly white
- Residual confounding
8. Risk of bias
- Orienting tool
- ROBINS-E
- Type of assessment
- simplified assessment (abstract only)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Observational; repeated measures reduce misclassification.
9. Assessment of the statistical interpretation
Small HR (1.04); absolute difference about 0.06 percentage points per year; attenuation within diet-quality strata suggests overlap with overall diet quality.
10. Relevance to the specific claim
Limits C-UPF-03: overall mortality association small; no CVD or cancer mortality association; driven by processed meat and sweetened drinks.
11. Independent contrary findings
EPIC (I-UPF-107) and Lane 2024 show CVD mortality associations.
12. Transferability to Germany and Europe
Moderate (US).
13. Overall methodological judgement
High-quality cohort showing that 'UPF' as a whole is heterogeneous; supports targeting specific products.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | high |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Why US and European results differ for CVD mortality.