Meta-analysisindependent cross-checkfull text checked★ key study I-INTA-02 in-depth review
Do tea drinkers live shorter or longer?
Tea consumption and risk of all-cause, cardiovascular disease, and cancer mortality: a meta-analysis of thirty-eight prospective cohort data sets.
Overview
Kim Y et al.. Epidemiology and health 2024 · PMID 38938012 · DOI 10.4178/epih.e2024056
Across nearly 2 million people, tea drinkers had about 10% lower death rates than non-drinkers.
Does the study support the specific claim? No · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Tea and risk of death
- How was the study done?
Meta-analysis of 38 cohort data sets
- How many people, animals or samples were studied?
1,956,549 participants
- What was actually found?
Lower all-cause and cardiovascular mortality with tea
- How large was the observed effect?
All-cause 0.90 (0.86-0.95); CVD 0.86 (0.79-0.94)
- What do the results mean in everyday life?
Usual tea drinking is not linked to harm
- What are the study's limitations?
Observational; high heterogeneity; benefit seen mainly in Asian studies, not in European ones
- What does it say about Elawady's specific claim?
Contradicts 'second most dangerous drink'
- Do the results apply to Germany and Europe?
Moderate (no association in European cohorts, but no harm either)
- What can explicitly not be concluded from this study?
That tea itself causes the lower risk
Scientific in-depth review
1. Full reference
- Authors
- Kim Y et al.
- Journal
- Epidemiology and health
- Year
- 2024
- Identifier
- PMID 38938012 DOI 10.4178/epih.e2024056
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
National Research Foundation of Korea (NRF-2021R1F1A1050847); funders had no role; authors declare no conflicts of interest.
2. Research question
Association between tea consumption and all-cause, cardiovascular and cancer mortality
3. Study design and population
- Study design
- Meta-analysis – Systematic review and meta-analysis of prospective cohort studies
- Population
General adult populations in prospective cohorts (studies restricted to patients with a specific disease excluded); cohorts from Asia, Europe, USA and Oceania; mean follow-up 12.3 years (5-23 years)
- Country
Asia (incl. Japan, China, Korea, Singapore), Europe (UK, Netherlands, Finland, Scotland), USA, Australia; no Middle Eastern or African cohorts
- Period
PubMed, Web of Science, Scopus searched to April 2024 (English only); PROSPERO CRD42023400484
- Sample size
38 cohort data sets (27 papers), 1,956,549 participants; 218,948 all-cause, 53,234 CVD and 56,049 cancer deaths
- Inclusion and exclusion criteria
Prospective observational design; tea exposure; all-cause, CVD or cancer mortality; RR/HR with 95% CI per intake group; most extended/largest report per cohort; patient-only cohorts excluded; English-language publications.
4. Methods and endpoints
- Exposure / intervention
Tea consumption (highest vs lowest; dose-response)
- Comparator
Lowest or no tea consumption
- Endpoints
All-cause, CVD and cancer mortality
- Statistical methods
DerSimonian-Laird random effects (highest vs lowest); Q and I2; subgroups by sex, region, follow-up, sample size; leave-one-out; Begg and Egger tests; linear and restricted cubic spline (4 knots) dose-response, 1 cup = 150 mL; Newcastle-Ottawa quality (all studies 7-9).
5. Actual results
- Results
Highest vs lowest: all-cause mortality 0.90 (0.86-0.95; I2 81.6%; 33 data sets); CVD mortality 0.86 (0.79-0.94; I2 74.6%; 31 data sets); cancer mortality 0.90 (0.78-1.03; I2 78.8%; 22 data sets). All-cause by region: Asia 0.84 (0.77-0.91); Europe 1.12 (0.88-1.42), P difference 0.05. Non-linear dose-response: greatest all-cause reduction at 2.0 cups/day (0.91, 0.88-0.94), cancer at 1.5 cups/day (0.92, 0.89-0.96); CVD plateau at 1.5-3.0 cups/day with ES 27% lower up to 8.0 cups/day. No publication bias (all-cause Begg P=0.97, Egger P=0.75).
- Absolute risks
not reported in full text
- Authors' conclusion
Moderate daily tea consumption (1.5-2.0 cups/day) was associated with lower risks of mortality from all causes, CVD and cancer in the general population; well-designed cohorts analysing tea types are needed.
6. Strengths
- Pre-registered (PROSPERO); three databases plus reference lists
- Large number of prospective cohorts and participants incl. UK Biobank and Asia Cohort Consortium
- Dose-response modelling; leave-one-out stable; no publication bias detected
7. Weaknesses
- Observational cohorts: residual confounding and healthy-user effects; adjustment sets vary
- High heterogeneity (I2 75-82%)
- English-only search; selection not explicitly in duplicate (extraction was); no list of excluded studies in main text
- Tea type, strength, additions (milk, sugar) and cup size mostly unreported
- No association in European cohorts (all-cause ES 1.12, 0.88-1.42); no Middle Eastern/African data
- Newcastle-Ottawa only (no domain-based RoB); funding of included studies not reported
8. Risk of bias
- Orienting tool
- AMSTAR 2
- 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 unchanged (approximately low-to-moderate confidence by AMSTAR 2; several critical items only partially met). The main limitation is residual confounding in observational cohorts. Simplified single-reviewer judgement from the full text; a formal AMSTAR 2 assessment needs two independent reviewers.
| Domain | Judgement | Rationale |
|---|---|---|
| Item 1 PICO | low | Clearly defined. |
| Item 2 (critical) protocol | low | Registered on PROSPERO (CRD42023400484). |
| Item 4 (critical) comprehensive search | some concerns | Three databases plus reference lists; restricted to English; no grey literature. |
| Items 5-6 duplicate selection and extraction | some concerns | Extraction in duplicate; duplicate screening not described. |
| Item 7 (critical) excluded studies | some concerns | Flow chart only in main text; no list of excluded full texts visible. |
| Item 9 (critical) RoB of individual studies | some concerns | Newcastle-Ottawa scale (all 7-9); confounding not assessed domain-wise. |
| Item 10 funding of included studies | high | Not reported. |
| Item 11 (critical) meta-analytic methods | low | Random effects, appropriate dose-response methods. |
| Item 13 (critical) RoB in interpretation | some concerns | Limitations listed, but conclusion phrased in terms of recommended intake levels from observational data. |
| Item 14 heterogeneity | some concerns | Subgroups explored; high residual heterogeneity unexplained. |
| Item 15 (critical) publication bias | low | Begg, Egger and funnel plots. |
| Item 16 review COI | low | Funding and COI declared. |
Full-text review (25 October 2026): Added death counts, I2 values, regional subgroups (Europe all-cause ES 1.12, 0.88-1.42 vs Asia 0.84), PROSPERO registration, English-only search, NOS scores, funding (NRF Korea). Bias overall unchanged with itemised AMSTAR 2; consistency changed to mixed; EU transferability refined.
9. Assessment of the statistical interpretation
About 10% lower all-cause and 14% lower CVD mortality in highest vs lowest tea drinkers (relative, observational); cancer mortality CI includes 1. Association driven by Asian cohorts; European cohorts null.
10. Relevance to the specific claim
Tests whether tea is a 'dangerous' drink.
11. Independent contrary findings
Observational design; residual confounding possible.
12. Transferability to Germany and Europe
Moderate: European cohorts showed no significant association (all-cause ES 1.12, 0.88-1.42), so the inverse association is mainly from Asian (largely green tea) populations; no Middle Eastern or Egyptian cohorts.
13. Overall methodological judgement
Consistent observational evidence that habitual tea drinking is not associated with higher mortality and, mainly in Asian cohorts, with lower mortality; causation not established.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | high |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | moderate |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Causality; effects of very strong or very hot tea; sugar added to tea.