Meta-analysisindependent cross-checkfull text checked★ key study I-EGG-01 in-depth review
Eggs and heart disease: 1.7 million people
Egg consumption and risk of cardiovascular disease: three large prospective US cohort studies, systematic review, and updated meta-analysis.
Overview
Drouin-Chartier JP et al.. BMJ 2020 · PMID 32132002 · DOI 10.1136/bmj.m513
Eating up to one egg a day was not linked to heart disease or stroke.
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?
Is egg intake associated with cardiovascular disease risk?
- How was the study done?
Three prospective US cohorts (NHS, NHS II, HPFS) plus systematic review and dose-response meta-analysis of prospective cohorts (PROSPERO CRD42019129650)
- How many people, animals or samples were studied?
215,618 cohort participants (14,806 CVD events); meta-analysis 33 estimates, 1,720,108 participants, 139,195 CVD events; US women and men free of CVD, diabetes and cancer; meta-analysis of global cohorts
- What was actually found?
Cohorts: an egg a day or more vs <1 per month HR 0.93 (0.82-1.05). Meta-analysis per egg/day: CVD 0.98 (0.93-1.03); Asia 0.92; in people with type 2 diabetes possibly higher risk (1.25, 0.99-1.59).
- How large was the observed effect?
Pooled Cox models; random-effects meta-analysis
- What do the results mean in everyday life?
For most people, an egg a day is not linked to heart risk.
- What are the study's limitations?
Observational; egg eaters had less healthy profiles (residual confounding possible); heterogeneity; few people above one egg a day; possible higher risk in type 2 diabetes is uncertain.
- What does it say about Elawady's specific claim?
Contradicts the archive claim.
- Do the results apply to Germany and Europe?
European cohorts: no association.
- What can explicitly not be concluded from this study?
Cannot prove absence of any effect at high intakes or in people with type 2 diabetes.
Scientific in-depth review
1. Full reference
- Authors
- Drouin-Chartier JP et al.
- Journal
- BMJ
- Year
- 2020
- Identifier
- PMID 32132002 DOI 10.1136/bmj.m513
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
NIH grants for the cohorts and NIH/CIHR support for the work; outside the submitted work JPDC received honoraria from Dairy Farmers of Canada, YL grants from the California Walnut Commission, FBH research support from the California Walnut Commission and lecture honoraria from Metagenics and others. No egg-industry funding reported.
2. Research question
Is egg intake associated with cardiovascular disease risk?
3. Study design and population
- Study design
- Meta-analysis – Three prospective US cohorts (NHS, NHS II, HPFS) plus systematic review and dose-response meta-analysis of prospective cohorts (PROSPERO CRD42019129650)
- Population
US women and men free of CVD, diabetes and cancer; meta-analysis of global cohorts
- Country
USA; meta-analysis multinational (US, Europe, Asia)
- Period
Cohorts: NHS 1980-2012, NHS II 1991-2013, HPFS 1986-2012 (>5.54 million person-years); meta-analysis search of PubMed, Embase, Web of Science to 6 August 2019
- Sample size
Cohorts: 215,618 (83,349 NHS, 90,214 NHS II, 42,055 HPFS), 14,806 CVD events; meta-analysis: 28 studies / 33 risk estimates, 1,720,108 participants, 139,195 CVD events
- Inclusion and exclusion criteria
Cohorts: excluded baseline CVD, cancer, type 2 diabetes, coronary surgery, implausible energy (<500/>3500 kcal women; <800/>4200 men), >70 FFQ items blank, missing egg intake. Meta-analysis: prospective cohorts of egg intake and incident CVD in general population or people with T2D; PROSPERO CRD42019129650.
4. Methods and endpoints
- Exposure / intervention
Egg intake
- Comparator
<1 egg/month (cohorts); per 1 egg/day (meta-analysis)
- Endpoints
Incident CVD (non-fatal MI, fatal CHD, stroke)
- Statistical methods
Cohorts: Cox models with cumulative-average egg intake (FFQ validity r=0.77 women, 0.80 men); Model 1 age; Model 2 lifestyle/clinical; Model 3 + energy and egg-associated foods (red/processed meat, refined grains, potatoes, etc.); substitution analyses; E-values; sensitivity analyses stopping diet updating after intermediate diagnoses. Meta-analysis: random-effects dose-response (per 1 egg/day) and high-vs-low; Newcastle-Ottawa scale; Begg and Egger tests; duplicate screening/extraction/RoB.
5. Actual results
- Results
Cohorts: >=1 egg/day vs <1/month, age-adjusted HR 1.10 (0.97-1.23); fully adjusted HR 0.93 (0.82-1.05); CHD 0.90 (0.77-1.05); stroke 0.99 (0.81-1.22); >=2 eggs/day HR 0.91 (0.72-1.15); no association in T2D participants (HR 1.06) or without T2D; similar when diet updating was stopped after intermediate diagnoses. Meta-analysis per 1 egg/day: CVD RR 0.98 (0.93-1.03), I2 62.3%; CHD 0.96 (0.91-1.03), I2 38.2%; stroke 0.99 (0.91-1.07), I2 71.5%; US 1.01 (0.96-1.06); Europe 1.05 (0.92-1.19); Asia 0.92 (0.85-0.99). People with type 2 diabetes (9 estimates, 28,608 participants): RR 1.25 (0.99-1.59) per egg/day; high vs low intake RR 1.40 (1.00-1.97), with considerable heterogeneity. No evidence of publication bias for the main analysis.
- Absolute risks
not reported in full text (relative estimates only)
- Authors' conclusion
Moderate egg consumption (up to one egg per day) is not associated with CVD risk overall and associated with potentially lower risk in Asian populations.
6. Strengths
- Three large cohorts with up to 32 years follow-up and repeated diet measures
- Validated FFQ and physician-confirmed outcomes
- Extensive dietary confounder control, E-values and sensitivity analyses for reverse causation
- Registered meta-analysis protocol, duplicate screening and NOS risk-of-bias assessment, regional stratification
7. Weaknesses
- Observational data; higher egg eaters were less healthy (higher BMI, more red meat, more T2D), so residual confounding possible
- Health-professional cohorts may limit generalisability
- Moderate-to-high heterogeneity in the meta-analysis (I2 62.3%; stroke 71.5%)
- Few participants consumed more than 1 egg/day
- Possible higher risk among people with type 2 diabetes (borderline, heterogeneous)
- Egg cooking method and accompanying foods not captured
8. Risk of bias
- Orienting tool
- AMSTAR 2 (meta-analysis component); cohort component ROBINS-E-oriented
- 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
- low
- Rationale
AMSTAR 2 overall confidence: moderate to high (no critical flaws; non-critical issues). Overall unchanged (low). Cohort component: some concerns for residual confounding. Single-reviewer structured judgement; a formal assessment needs two independent reviewers and the protocol.
| Domain | Judgement | Rationale |
|---|---|---|
| 1 PICO components | low | Clearly defined. |
| 2 Protocol established a priori (critical) | low | PROSPERO CRD42019129650. |
| 3 Explanation of included designs | low | Prospective cohorts only, justified. |
| 4 Comprehensive search (critical) | some concerns | PubMed, Embase, Web of Science to August 2019; grey literature and language limits not clear in retrieved text. |
| 5 Duplicate selection | low | Two authors independently. |
| 6 Duplicate extraction | low | Two authors independently. |
| 7 List of excluded studies with reasons (critical) | some concerns | Presumably in supplementary flow chart; not verifiable from retrieved text. |
| 8 Description of included studies | low | Supplementary characteristics tables. |
| 9 Risk of bias of individual studies (critical) | low | Newcastle-Ottawa scale in duplicate; confounder control documented (15/28 controlled all primary confounders). |
| 10 Funding of included studies | high | Not reported. |
| 11 Appropriate meta-analytic methods (critical) | low | Random-effects dose-response with stratification. |
| 12 Impact of RoB on results | some concerns | Stratified analyses by region and population; RoB-stratified results not visible in retrieved text. |
| 13 RoB considered in interpretation (critical) | low | Residual confounding and heterogeneity discussed. |
| 14 Heterogeneity explained | some concerns | Partly explained by region; residual heterogeneity. |
| 15 Publication bias (critical) | low | Begg/Egger and funnel plot; no evidence of bias. |
| 16 Conflicts of interest of review | low | ICMJE disclosures; industry ties outside the work (dairy, walnut) reported. |
Full-text review (25 October 2026): Added cohort exclusions, FFQ validity, model sequence (age-adjusted HR 1.10 vs adjusted 0.93), search databases/date, NOS RoB, T2D meta-analysis signal (RR 1.25, 0.99-1.59; high vs low 1.40), funding/COI details; AMSTAR 2 itemised; bias unchanged (low).
9. Assessment of the statistical interpretation
Tight CIs around 1 for up to one egg per day exclude more than small effects in the general population. The age-adjusted cohort HR (1.10) shifted to 0.93 after adjustment for diet and lifestyle, illustrating strong confounding by the dietary pattern accompanying eggs. The diabetes subgroup signal is borderline and imprecise.
10. Relevance to the specific claim
Main counter-evidence for C-EGG-04; supports C-EGG-02/11 being unsupported.
11. Independent contrary findings
A-EGG-12 (US pooled cohorts) found positive associations.
12. Transferability to Germany and Europe
European cohorts: no association.
13. Overall methodological judgement
Strongest single body of evidence that moderate egg intake (up to one per day) is not associated with cardiovascular disease in the general population; a possible higher risk in people with type 2 diabetes remains uncertain.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | low |
| Precision of results | high |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Effects above one egg/day; people with type 2 diabetes; role of cooking method and dietary context.
15. Linked fact checks
- All eggs (chicken, quail, pigeon, ostrich, etc.) are harmful because an egg is an embryo / contains the 'full components of a living organism'. – Studies contradicting or limiting the claim
- Higher egg (and dietary cholesterol) intake increases the risk of cardiovascular disease and all-cause mortality (each additional half egg per day: +6% CVD, +8% mortality). – Studies contradicting or limiting the claim
- Egg-yolk phosphatidylcholine is converted by gut bacteria into TMAO, and therefore eggs raise the risk of major cardiovascular events (about 2.5-fold). – Studies contradicting or limiting the claim
- Eggs should be completely avoided (all kinds, all cooking methods). – Studies contradicting or limiting the claim