Cohort studyindependent cross-checkfull text checked I-CAN-04 in-depth review
About a third of cancer risk is inherited
Familial Risk and Heritability of Cancer Among Twins in Nordic Countries.
Overview
Mucci LA et al.. JAMA 2016 · PMID 26746459 · DOI 10.1001/jama.2015.17703
Heritability of cancer overall was 33% in 200,000 Nordic twins.
Does the study support the specific claim? No · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Estimate heritability.
- How was the study done?
Twin cohort.
- How many people, animals or samples were studied?
203,691.
- What was actually found?
33% heritability.
- How large was the observed effect?
Melanoma 58%, prostate 57%.
- What do the results mean in everyday life?
Family history matters for screening.
- What are the study's limitations?
Nordic, mainly white population; twin models assume equal shared environments, which cannot be tested.
- What does it say about Elawady's specific claim?
Not all cancer is lifestyle.
- Do the results apply to Germany and Europe?
High.
- What can explicitly not be concluded from this study?
That the remaining two-thirds of risk is due to food; heritability is not the share of cancers preventable by lifestyle (authors); specific genes.
Scientific in-depth review
1. Full reference
- Authors
- Mucci LA et al.
- Journal
- JAMA
- Year
- 2016
- Identifier
- PMID 26746459 DOI 10.1001/jama.2015.17703
- Verification
- verified – PubMed esummary + efetch 2026-10-25
- Retracted
- no
- Corrections
ErratumIn: JAMA 2016;315(8):822
- Extent checked
- full text
- Conflicts of interest and funding
Funding: Ellison Foundation, Nordic Cancer Union, Academy of Finland, BioSHaRE-EU (FP7), Swedish Ministry for Higher Education, Odense University Hospital, Karolinska Institutet, NCI training grants; funders had no role in design/analysis but 'reviewed and approved the manuscript'. COI: one author (Kaprio) personal fees from Pfizer (nicotine dependence consulting 2012-2014); no other disclosures.
2. Research question
Familial risk and heritability of cancer.
3. Study design and population
- Study design
- Cohort study – Prospective population-based twin cohort (Nordic Twin Study of Cancer, NorTwinCan) with registry linkage; biometric (ACE) twin modelling
- Population
Same-sex monozygotic and dizygotic twins from the national twin registries of Denmark, Finland, Norway and Sweden (birth cohorts 1870-2004)
- Country
Denmark, Finland, Norway, Sweden
- Period
Follow-up from 1943 (Denmark), 1961 (Sweden), 1964 (Norway), 1974 (Finland) to end of 2008-2010; median follow-up 32 years
- Sample size
203,691 twins (80,309 monozygotic, 123,382 same-sex dizygotic; 104,251 women); 27,156 incident cancers in 23,980 individuals; 50,990 died, 3804 emigrated
- Inclusion and exclusion criteria
Same-sex twin pairs with known zygosity from population-based twin registries; 5376 twins with missing or inconsistent zygosity excluded; registry coverage 30% (birth cohort 1870-1930), 70% (1931-1952), ~100% later (twins had to survive to age 6 in early cohorts)
4. Methods and endpoints
- Exposure / intervention
Genetic relatedness (monozygotic vs dizygotic) and co-twin cancer status as a proxy for shared genes and shared environment
- Comparator
Cumulative cancer incidence in the whole twin cohort; dizygotic vs monozygotic pairs
- Endpoints
Incident cancer (national cancer registries, ICD-10 grouped by NORDCAN); familial risk; heritability on the liability scale
- Statistical methods
Aalen-Johansen cumulative incidence accounting for left truncation, censoring and competing risk of death; inverse-probability-of-censoring weighted liability-threshold ACE models (R mets 1.1.1); familial risk for 23 cancer types with >1 concordant pair, heritability for types with >=5 concordant pairs
5. Actual results
- Results
Cumulative cancer incidence to age 100: 32%. Familial risk of any cancer if co-twin had cancer: 46% (95% CI 44-48%) in MZ and 37% (36-38%) in DZ twins, i.e. absolute excess 14% (12-16%) and 5% (4-6%). Heritability of cancer overall 33% (30-37%) with no shared-environment component; melanoma 58% (43-73%), prostate 57% (51-63%), nonmelanoma skin 43% (26-59%), ovary 39% (23-55%), kidney 38% (21-55%), breast 31% (11-51%; shared environment 16%), corpus uteri 27% (11-43%). Lower heritability for colon 15% (0-45%), rectum 14% (0-50%), stomach 22% (0-55%); lung shared environment 24% (7-40%). Among concordant pairs >2/3 had different cancer types; median age difference at diagnosis 8.0 (MZ) and 9.3 (DZ) years.
- Absolute risks
Cumulative incidence 8% by age 65, 25% by 80, 32% by 100; familial risk 46% (MZ) and 37% (DZ) if co-twin had cancer; e.g. prostate cumulative risk 10.5% vs familial risk 38.0% (MZ) / 22.0% (DZ); breast 9.4% vs 28.1% / 19.9%
- Authors' conclusion
Significant excess familial risk for cancer overall and for prostate, melanoma, breast, ovary and uterus cancers; useful for patient education and risk counselling. Authors stress that a high heritability does not translate into a low population-attributable risk of lifestyle and environmental factors and does not exclude effective prevention.
6. Strengths
- Largest twin study of cancer; >3 decades follow-up
- Near-complete national cancer registry ascertainment and linkage
- Methods account for censoring, left truncation and competing death
- Cumulative incidence similar to national rates (representativeness)
- Zygosity questionnaire >95% agreement with genetic markers
7. Weaknesses
- Equal-environment assumption of twin models cannot be tested (authors)
- Incomplete registry coverage and survival-to-age-6 requirement in early birth cohorts (selection)
- Mainly white Nordic population; generalisability to multiethnic populations unclear (authors)
- Wide CIs for many cancer types; rare cancers (e.g. many haematological) not estimable
- Heritability is population- and era-specific and includes genetic influences on risk behaviours (e.g. smoking, obesity)
- Funders reviewed and approved the manuscript
8. Risk of bias
- Orienting tool
- ROBINS-E (adapted to a twin heritability cohort)
- Type of assessment
- structured assessment based on the full text by a single reviewer (no formal double review)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Overall downgraded from 'low' (abstract-based) to 'some_concerns' because of the untestable equal-environment assumption and early-cohort selection; outcome ascertainment is excellent. Single-reviewer structured judgement from the full text; a complete formal assessment requires two independent reviewers and the review/study protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| Confounding | some concerns | Heritability estimates rely on the untestable assumption that MZ and DZ pairs share environments equally; violations inflate heritability. |
| Exposure measurement (zygosity) | low | Validated questionnaire zygosity (>95% agreement with genetic markers); 5376 with missing/inconsistent zygosity excluded. |
| Selection of participants | some concerns | Early cohorts had 30-70% coverage and required survival to age 6; mitigated by cumulative incidence comparable to national rates. |
| Post-exposure interventions | not applicable | Exposure is genetic relatedness. |
| Missing data | low | Registry follow-up essentially complete; 3804 emigrants censored; inverse-probability-of-censoring weights. |
| Outcome measurement | low | Mandatory national cancer registration, near 100% complete, verified diagnoses. |
| Selection of reported result | low | Pre-specified thresholds for reporting (>1 / >=5 concordant pairs); all types tabulated. |
Full-text review (25 October 2026): Added exact cohort composition, follow-up by country, exclusions, registry coverage limits, familial risks with CIs, site-specific heritability incl. low GI-cancer heritability, absolute cumulative risks, funding/COI (Pfizer fees one author; funders approved manuscript), and authors' explicit caveat that heritability is not population-attributable risk. ROBINS-E overall changed low -> some_concerns (equal-environment assumption, early-cohort selection).
9. Assessment of the statistical interpretation
About one-third of the variation in liability to cancer in this population is attributed to genetic differences; the remainder is mostly non-shared (individual) environment including chance/stochastic processes. Heritability is not the share of cases caused by genes and not the share preventable by lifestyle (authors explicitly warn against this misreading). Absolute familial excess risks are modest for any cancer (5-14 percentage points).
10. Relevance to the specific claim
Genetics is a substantial non-dietary cause.
11. Independent contrary findings
12. Transferability to Germany and Europe
High for Nordic/northern European populations; unclear for multiethnic populations (authors).
13. Overall methodological judgement
Robust registry-based twin study showing a substantial heritable component of cancer risk alongside a dominant non-shared environmental/chance component; it limits monocausal claims in both directions.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | high |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Which genetic loci explain the heritability; generalisability outside Nordic populations; heritability of rare cancers.
15. Linked fact checks
- Cancer is the body's own plan ('a self plan'): body cells are ordered to transform because this was necessary for the person's life; it is not caused by external agents. – Studies contradicting or limiting the claim
- Chronic disease is mostly a cumulative reaction to what enters the body daily - contaminated food, medicines and cosmetics - not a fate that comes from outside. – Studies contradicting or limiting the claim