Affiliate disclosure: Some links in this article are affiliate links. We may earn a commission at no additional cost to you.

Medical disclaimer: This article reviews epidemiological research on green tea consumption. It is not medical advice, diagnosis, or treatment. Consult a qualified healthcare professional before making changes to your diet, supplementation, or lifestyle.

TL;DR

• The Ohsaki cohort (Kuriyama et al., JAMA 2006): roughly 40,500 Japanese adults followed 11 years. Drinking 5+ cups of green tea daily was associated with 16% lower all-cause mortality in men and 23% lower in women, compared to fewer than 1 cup per day.
• The JPHC (Japan Public Health Center-based Prospective Study): multiple analyses across 80,000+ adults. Cardiovascular mortality shows the strongest and most consistent association; all-cause mortality signal is present but somewhat smaller.
• Cancer mortality was not significantly associated with green tea consumption in the Ohsaki study. JPHC cancer results are mixed across cancer type and sex.
• Both studies are observational. The associations are dose-dependent and biologically plausible, but confounding cannot be fully eliminated from either dataset.

The Ohsaki cohort

The Ohsaki National Health Insurance Cohort enrolled adults aged 40-79 in Miyagi Prefecture in northeastern Japan. Kuriyama and colleagues published the mortality analysis in JAMA in 2006 (PubMed 16968850). Design details:

• Enrollment: 40,530 adults, 1994
• Follow-up: 11 years (through 2005)
• Exposure measurement: self-reported green tea cups per day at baseline, categorized as less than 1, 1-2, 3-4, 5-6, or 7+ cups
• Primary outcomes: all-cause mortality, cardiovascular disease mortality, cancer mortality

Adjusted mortality findings (controlling for age, sex, smoking status, alcohol, BMI, dietary variables):

Green tea intake	All-cause mortality — men	All-cause mortality — women
5+ cups/day vs < 1 cup/day	HR 0.84 (approx. 16% reduction)	HR 0.77 (approx. 23% reduction)
3-4 cups/day	Intermediate, statistically significant	Intermediate, statistically significant
Dose-response trend	Present across all categories	Present across all categories

Cardiovascular mortality showed a stronger association: 5+ cups/day was linked to roughly 26% lower cardiovascular mortality in the fully adjusted model. Cancer mortality, by contrast, showed no significant association at the whole-cohort level, though certain cancer-specific subgroup analyses showed sex-dependent trends that were not consistently replicated.

The JPHC study

The Japan Public Health Center-based Prospective Study draws from 45 public health center areas across Japan, with enrollment primarily in the early 1990s and follow-up extending into the 2010s and beyond. Multiple research groups have published analyses from this cohort, making it the more varied and contested dataset of the two.

Representative findings on green tea and mortality:

• Cardiovascular disease mortality: multiple JPHC papers report 15-25% lower risk at 5+ cups/day, with consistent dose-response curves across analyses.
• Stroke: a widely cited JPHC analysis of approximately 40,000 adults aged 40-69 found that 4+ cups/day was associated with roughly 20% lower stroke risk.
• All-cause mortality: positive associations present across analyses, with effect sizes ranging from roughly 10-20% at the highest consumption tier depending on sex, age group, and follow-up duration.

On cancer: JPHC publications have examined gastric, colorectal, lung, and breast cancer outcomes separately. Results are inconsistent — some analyses show associations in specific sex-age subgroups; others show no association. The JPHC cancer data does not support a general statement linking green tea to lower cancer mortality. Some subgroup associations remain in the literature but require independent replication to interpret.

What the mortality signal actually says

Taking the Ohsaki and JPHC data together:

What is consistent: a dose-dependent, statistically significant association between daily green tea consumption at 5+ cups and lower cardiovascular and all-cause mortality in Japanese adults. The association has been replicated across independent cohorts drawing from different Japanese regions and different enrollment periods — one of the standard markers of credibility in observational epidemiology.

What is not consistent: any specific anti-cancer effect. The cardiovascular signal is the durable finding; cancer mortality is the weaker, more mixed, and less replicated signal. Popular summaries frequently overextend the Ohsaki and JPHC findings to include cancer claims that the data does not robustly support.

Biological plausibility: catechins — particularly EGCG — have cardiovascular-relevant mechanisms studied in smaller controlled trials, including modest blood pressure reduction, endothelial function effects, and lipid metabolism. These provide a mechanistic pathway for the cardiovascular mortality association, which is why researchers treat it as more credible than the cancer associations, where the mechanism from cell studies has not translated reliably to human cohort outcomes.

What the cohorts cannot tell you

Neither the Ohsaki nor the JPHC cohort was a randomized intervention. Both observed people who already drank green tea as part of their lives — and in Japan, heavy green tea drinkers differ from non-drinkers in ways that extend well beyond the cup count.

Common confounders in this literature:

• Dietary pattern: heavy green tea drinkers in Japan are more likely to follow traditional Japanese eating patterns broadly — fish, vegetables, fermented foods, lower processed food consumption. Controlling statistically for individual dietary variables does not fully resolve a confounding that operates across the whole pattern.
• Socioeconomic factors: tea consumption correlates with income and education in Japan; higher income correlates with better diet quality and better healthcare access, both independently associated with lower mortality.
• Smoking: adjustment for smoking is standard in both cohorts, but non-smoking and green tea consumption are negatively correlated at the population level — meaning non-smokers drink more tea. Residual confounding from this overlap is difficult to fully eliminate.
• Physical activity: traditional lifestyle patterns that include habitual tea-drinking also tend to include higher daily physical activity (walking, farming, traditional occupation) in the cohort-era Japanese population.

Both the Ohsaki and JPHC analyses apply standard covariate adjustments. Statistical adjustment narrows the confounding but does not close it.

Methodological limits specific to these cohorts

• Self-reported intake, measured once: cup-per-day categories are self-reported at baseline and used as the fixed exposure variable. Green tea consumption likely shifted over an 11-year follow-up, and the analysis cannot capture within-person changes.
• Cup size and preparation not standardized: a “cup” is not defined by catechin content. Catechin extraction varies substantially with preparation temperature, steeping time, and leaf quality. The cohort studies measure cups reported, not catechin dose.
• Population transferability: the associations come from a Japanese population where green tea is embedded in a broader cultural diet and lifestyle. Whether adding green tea consumption to a Western dietary pattern replicates the same associations is unknown and cannot be read from Ohsaki or JPHC data.
• Reverse causation: people who report drinking almost no green tea may disproportionately include those already in poor health or with limited mobility at baseline. Standard adjustments for baseline health status reduce but do not eliminate this risk.

Dose and practical context

The dose-response in both cohorts concentrates the association in the higher consumption tiers — 5+ cups/day. At 1-2 cups/day, the mortality reduction is smaller and in some analyses falls short of statistical significance. This matters for anyone trying to translate cohort data into practice: the observed association comes from people drinking green tea as a throughout-the-day habit, not from one morning cup.

For catechin context: high-quality loose-leaf sencha at 5 cups/day provides roughly 300-500 mg of EGCG and related catechins daily (see sencha and catechin content). Teabag green tea at the same cup count provides substantially less. The cohorts did not distinguish leaf quality, but the “5 cups” reported by rural Miyagi and Iwate residents in the mid-1990s almost certainly reflects freshly brewed loose-leaf sencha — the standard preparation in that region and era, not the teabag product now available globally.

For sourcing loose-leaf Japanese sencha, specialty importers and iHerb both carry midrange options that approximate the preparation quality relevant to the cohort exposure.

Where the evidence stands

The Ohsaki and JPHC cohorts offer some of the most consistent dietary epidemiology linking a specific beverage to reduced all-cause and cardiovascular mortality. The associations are large by observational standards, dose-dependent, and biologically plausible — and they have been replicated rather than contradicted as more analyses have been published.

They are not causal proof. No randomized controlled trial has tested whether adopting green tea consumption in a non-Japanese diet reduces 11-year mortality — and such a trial is unlikely to exist given the practical constraints. The cancer mortality signal is meaningfully weaker and should not be used as a basis for strong claims.

For someone evaluating whether the evidence supports making daily sencha a habit: the cardiovascular mortality association is credible enough to be informative. The relevant dose is 5+ cups/day of properly brewed loose-leaf tea. Whether that benefit transfers outside the Japanese dietary context remains an open question.

---

Part of our research series. See also: Matcha vs Sencha vs Hojicha: which has the strongest health evidence?, 5 Japanese longevity habits backed by research.