Longevity & Aging · Supplements
coffee decreases all-cause mortality
In plain terms: Do coffee drinkers live longer?
Strong support Longevity & Aging
Part of: • Coffee
RefutedContestedStrong support
consensus score 0.92
On average, moderate coffee drinkers (about 3-4 cups a day) have a modestly lower risk of dying over time than non-drinkers. It's an association, not proof, and piling on many extra cups doesn't add benefit - but a few cups a day looks, if anything, protective rather than harmful.
📅 Last reviewed: 2026-07-15 ⓘ
Evidence ladder
How far up the ladder this claim has climbed. A high consensus on a low rung means "consistent so far," not "proven in people."
Top evidence so far: All trials, pooled (Meta-analysis)
MechanismIn-vitroAnimalObservationalRCTMeta-analysis
How the studies fall
15 support 0 contradict 0 tested null 1 mixed · 16 sources, 11 independent groups
What the evidence shows
Across very large umbrella reviews and dose-response meta-analyses, moderate coffee drinking (about 3–4 cups/day) is associated with the lowest risk of dying from any cause over follow-up — roughly 13–17% lower than non-drinkers, with the curve flattening or reversing at very high intakes. Caffeinated and decaf both track with the benefit.
The evidence (21)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Shi C, Wang YK, Yan MQ, Peng Z, Zhang JL, Tao YM, Yang L, Li L, Ou CQ 2025 · Br J Nutr | observational | supports | moderate | UK Biobank prospective cohort, compositional analysis of self-reported coffee, tea and plain water intake. Lowest mortality came at roughly 7-8 drinks a day in total; above 4 drinks a day, replacing plain water with coffee or tea went with lower mortality, but at 4 or fewer it did not. The best pattern was coffee and tea together in about a 2:3 ratio: hazard ratios 0.55 (95% CI 0.47-0.64) for death from any cause, 0.59 cancer, 0.69 cardiovascular, 0.28 respiratory, 0.35 digestive, versus other combinations. Every estimate is for a coffee-plus-tea mix; no coffee-alone figure is given. |
| Wang X, Ma H, Sun Q, Li J, Heianza Y, Van Dam RM, Hu FB, Rimm E, Manson JE, Qi L 2025 · Eur Heart J | observational | supports | moderate | In 40,725 US adults in NHANES 1999-2018, followed a median 9.8 years with 4,295 deaths, people whose coffee was concentrated in the morning had a lower risk of death from any cause than non-drinkers (hazard ratio 0.88, 95% CI 0.81-0.96; 0.84, 0.74-0.95 after further adjusting for sleep), while those who drank coffee throughout the day did not (0.99, 0.90-1.10). Across the whole cohort, higher coffee intake went with lower all-cause mortality (P for trend <0.01). Intake came from two 24-hour recalls, and the authors say the association cannot be read as cause. |
| Yan G, Dai X, Yan Y, Yan J, Tian W, Jiang R 2025 · Public Health Nutrition | observational | supports | moderate | In 43,114 US adults from NHANES 2001-2018, followed a median 8.7 years with 6,234 deaths, drinking 1 to 2 cups of coffee a day went with a 19% lower risk of death from any cause than drinking none (hazard ratio 0.81, 95% CI 0.75-0.88); more than 3 cups a day, 0.87 (0.78-0.96). Turned into life expectancy at age 50, the 1-2 cup group averaged 2.02 more years (1.17-2.85) than non-drinkers, the 2-3 cup group 1.17 years and the over-3-cup group 1.39 years. About a third of that gain (0.61 years) came from fewer cardiovascular deaths. Cancer mortality was not associated with coffee. |
| Kwapien E, Piszka M, Czarnecki F, Mesyasz M, Owczarska A, Dacyl H, Banach J, Kasprzyk WZ, Luczak JK, Bartkowski J 2025 · Cureus | observational | supports | low | Narrative review (not systematic/meta-analysis) synthesizing cohort literature; concludes moderate coffee intake (3-4 cups/day, ~450-600mL) correlates with reduced all-cause and cardiovascular mortality in a J-shaped dose-response. |
| Li Q, Liu Y, Sun X, Yin Z, Li H, Cheng C, Liu L, Zhang R, Liu F, Zhou Q, Wang C, Li L, Wang B, Zhao Y, Zhang M, Hu D 2019 · J Hum Nutr Diet | meta-analysis | supports | high | Dose-response meta-analysis pooling 21 cohort study articles, 10,103,115 participants and 240,303 deaths, searched to November 2017. The relationship was nonlinear (P for nonlinearity < 0.001): at 3 cups a day the risk of dying from any cause was 13% lower than for no or rare coffee drinking (RR 0.87, 95% CI 0.84-0.89). The authors also report similar inverse associations for caffeinated and for decaffeinated coffee, but give no figures for either and do not say how many of the 21 articles reported decaf. Every pooled study was an observational cohort, so this is an association, not proof. |
| Shahinfar H, Jayedi A, Khan TA, Shab-Bidar S 2021 · Nutr Metab Cardiovasc Dis | meta-analysis | supports | moderate | Dose-response meta-analysis of 10 prospective cohort studies in people with type 2 diabetes (82,270 deaths), searched to November 2020. Compared with drinking no coffee, 4 cups a day was associated with 21% lower all-cause mortality (HR 0.79, 95% CI 0.72-0.87), pooled across all 10 studies, and the association deepened steadily as intake rose. The authors themselves grade the certainty of this all-cause mortality finding moderate, and low or very low for the paper's other outcomes. Every pooled study was an observational cohort, so this is an association, not proof. |
| Kim Y, Je Y, Giovannucci E 2019 · Eur J Epidemiol | meta-analysis | supports | high | Meta-analysis of 40 prospective studies (3,852,651 people; 450,256 deaths counted across all-cause and cause-specific outcomes), searched to March 2019. The dose-response curve was non-linear: all-cause mortality was lowest at 3.5 cups/day (RR 0.85, 95% CI 0.82-0.89) versus no coffee, with no further reduction at higher intakes. The inverse association held across age, overweight status, alcohol, smoking and caffeinated versus decaffeinated coffee, but not across region - Europe and Asia showed stronger associations than the US. Every pooled study was observational. |
| Crippa A, Discacciati A, Larsson SC, Wolk A, Orsini N 2014 · Am J Epidemiol | meta-analysis | supports | high | Dose-response meta-analysis pooling 21 prospective studies, 997,464 participants and 121,915 deaths. The association was nonlinear (P for nonlinearity < 0.001): risk fell to a low point at 4 cups/day, where all-cause mortality was 16% lower (95% CI 13-18) than at the lowest intake, rather than falling steadily with each extra cup. CVD mortality bottomed out at 3 cups/day, 21% lower (95% CI 16-26). Coffee was not associated with cancer mortality. Every pooled study was an observational cohort, so this is an association, not proof coffee causes the lower risk. |
| Liu P, Yao G, Wu Y, Ren H, Zhou Q 2026 · FASEB J | observational | supports | moderate | In 21,553 US adults with cardiovascular-kidney-metabolic syndrome from NHANES 1999-2018, followed a median 9.2 years with 3,565 deaths, a morning-type coffee pattern (47.7% of the group) went with a 21% lower risk of death from any cause (hazard ratio 0.79, 95% CI 0.65-0.97) and a lower risk of cardiovascular death (0.75, 0.50-0.97), independent of how much coffee was drunk. Higher intake tracked with lower mortality only among morning drinkers, not among the 5.3% who drank late in the day (p for interaction 0.035). The abstract does not say what group these were compared against. |
| Yang D, Mao Y, Chen S, Lin W, Li L, Liang J, Huang H, Wen J, Chen G 2025 · Nutr Metab Cardiovasc Dis | observational | supports | low | Among 5,979 US adults with prediabetes in NHANES 2003-2018, 748 died during follow-up. Comparing the highest with the lowest coffee intake, death from any cause was 32% less likely (hazard ratio 0.68, 95% CI 0.55-0.85). Sugar-sweetened, sugar-free, full-fat and fat-free coffee were each inversely associated with all-cause mortality (p<0.05), with no figures given for them. Coffee intake was taken from a single 24-hour dietary recall at baseline. One of the paper's other results is impossible as printed: an odds ratio of 0.0002 with 'p for trend = 1.91'. |
| Chen QY, Keum N, Giovannucci EL 2025 · J Natl Cancer Inst | meta-analysis | supports | moderate | Meta-analysis pooling published studies of diet and lifestyle measured after a colorectal cancer diagnosis (PubMed and Embase searched to September 2024; DerSimonian-Laird random effects). Coffee drinking after diagnosis went with 35% lower death from any cause: summary relative risk 0.65 (95% CI 0.55 to 0.77). Coffee was one of about fifteen exposures pooled, and the paper does not report how many studies or how many patients lie behind that coffee figure. This is a pooled association among cancer survivors, not evidence that coffee caused the lower risk. |
| Poole R, Kennedy OJ, Roderick P, Fallowfield JA, Hayes PC, Parkes J 2017 · BMJ | meta-analysis | supports | moderate | An umbrella review: it gathered 201 meta-analyses of observational coffee research (plus 17 of trials) and re-pooled their published summary numbers, having 'excluded systematic reviews without meta-analysis' and examined no primary studies itself. Its headline figure for death from any cause — largest reduction at three cups a day versus none, relative risk 0.83 — is carried over from an earlier dose-response meta-analysis by Grosso and colleagues. The authors rate the underlying evidence low or very low quality by GRADE, with a median AMSTAR score of 5 out of 11. |
| Schaefer SM, Kaiser A, Behrendt I, Eichner G, Fasshauer M 2022 · Br J Nutr | observational | supports | high | UK Biobank: 354,386 adults who drank alcohol at least occasionally, followed for 4.2 million person-years with 20,201 deaths. Coffee was fitted as a smooth dose curve, mutually adjusted for wine, other alcohol and tea. Death from any cause was lowest at 2 cups a day; people drinking no coffee were 6% more likely to die than that (hazard ratio 1.06, 95% CI 1.03 to 1.09), and intakes above 2 cups differed significantly in neither direction. Coffee showed no significant association with cancer or cardiovascular death. Intake was recorded once, at the baseline visit. |
| Emadi RC, Kamangar F 2025 · Nutrients | observational | supports | low | Narrative review of coffee and health. Large prospective cohorts consistently link moderate intake to lower death rates: in NIH-AARP (400,000 people) two or more cups a day went with a 10-15% reduction in overall mortality, and a 2019 meta-analysis of 40 studies and 3.8 million people put the lowest risk at 3.5 cups a day (RR 0.85, 95% CI 0.82-0.89). Caffeinated and decaffeinated coffee both tracked with the benefit. The senior author consults for the National Coffee Association. |
| Zhao Y, Wu K, Zheng J, Zuo R, Li D 2015 · Public Health Nutr | meta-analysis | supports | moderate | Meta-analysis pooling 17 prospective cohort studies, 1,054,571 participants and 131,212 deaths from any cause. The dose-response curve was U-shaped (P for non-linearity <0.001) rather than falling steadily: against non or occasional drinkers, all-cause mortality was 11% lower at 1 to under 3 cups a day (RR 0.89, 95% CI 0.85-0.93), 13% lower at 3 to under 5 cups (0.87, 0.83-0.91) and 10% lower at 5 or more (0.90, 0.87-0.94). The authors say the association was more marked in females than males but give no sex-specific figures. Every pooled study was an observational cohort. |
| Watanabe D, Yoshida T, Nanri H, Watanabe Y, Yamada Y, Kimura M 2025 · Br J Nutr | observational | supports | moderate | In the Kyoto-Kameoka prospective study of 7708 Japanese adults aged 65 and over, 593 deaths occurred over a median 4.75 years of follow-up. Drinking 3 or more cups of coffee a day was associated with lower all-cause mortality than drinking none (HR 0.62, 95% CI 0.43-0.88), and with less incident disability (HR 0.81, 95% CI 0.65-0.99). Caffeine intake itself tracked with lower disability (200 mg/day or more: HR 0.75, 95% CI 0.63-0.89) but not with mortality, and green tea was associated with neither outcome. Diet was self-reported by food-frequency questionnaire. |
| Shin S, Lee JE, Loftfield E, Shu XO, Abe SK, Rahman MS, Saito E, Islam MR, Tsugane S, Sawada N, Tsuji I, Kanemura S, Sugawara Y, Tomata Y, Sadakane A, Ozasa K, Oze I, Ito H, Shin MH, Ahn YO, Park SK, Shin A, Xiang YB, Cai H, Koh WP, Yuan JM, Yoo KY, Chia KS, Boffetta P, Ahsan H, Zheng W, Inoue M, Kang D, Potter JD, Matsuo K, Qiao YL, Rothman N, Sinha R 2022 · Int J Epidemiol | observational | supports | high | A pooled analysis of 12 prospective cohorts in China, Japan, Korea and Singapore: 248,050 men and 280,454 women, with 94,744 deaths over follow-up averaging 6.5 to 22.7 years. Against coffee non-drinkers, men who drank at least five cups a day had a 24% lower risk of death from any cause (95% CI 17-29%) and women a 28% lower risk (95% CI 19-37%). Inverse associations were also found for death from cardiovascular disease and from cancer in both sexes. Every cohort was observational, so this is an association, not proof that coffee lowers the risk. |
| Grosso G, Godos J, Galvano F, Giovannucci EL 2017 · Annu Rev Nutr | meta-analysis | supports | high | 'we performed an umbrella review of the evidence from meta-analyses of observational studies and randomized controlled trials (RCTs)' — 59 unique outcomes across 112 meta-analyses of observational studies. Coffee showed 'a probable decreased risk of ... cardiovascular disease and mortality'. TWO reasons this cannot count: the paper pools nothing of its own, and the abstract never says whether 'mortality' here is all-cause or cardiovascular, so the endpoint this claim asserts is not settled by the text read. |
| Loftfield E, Cornelis MC, Caporaso N, Yu K, Sinha R, Freedman N 2018 · JAMA Intern Med | observational | supports | high | In the UK Biobank cohort, 498,134 adults were followed about 10 years and 14,225 died. Against non-drinkers, death from any cause was 6% less likely at under 1 cup a day, 8% at 1 cup, 12% at 2-3 and at 4-5 cups, 16% at 6-7 and 14% at 8 or more (hazard ratios 0.94, 0.92, 0.88, 0.88, 0.84, 0.86). The same pattern held for instant, ground and decaffeinated coffee, and it did not differ by a genetic caffeine-metabolism score built from AHR, CYP1A2, CYP2A6 and POR variants (P = 0.17). Coffee intake was recorded once, at the baseline visit. |
| Ribeiro EM, Alves M, Costa J, Ferreira JJ, Pinto FJ, Caldeira D 2020 · Nutr Metab Cardiovasc Dis | meta-analysis | mixed | moderate | Systematic review and meta-analysis of six prospective cohorts of people who had already had a heart attack. Coffee was not associated with death from any cause: pooled hazard ratio 0.85 (95% CI 0.63-1.13), from the 3 cohorts reporting that outcome, with moderate inconsistency between them (I2 = 50%). A separate dose-response analysis did find a significant non-linear inverse association between coffee and all-cause mortality. Cardiovascular mortality was lower (HR 0.70, 95% CI 0.54-0.91; 2 cohorts). Every pooled study was observational. |
| Grosso G, Micek A, Godos J, Sciacca S, Pajak A, Martínez-González MA, Giovannucci EL, Galvano F 2016 · Eur J Epidemiol | meta-analysis | supports | high | Pooled 31 prospective cohort studies covering 1,610,543 people, with 183,991 deaths from any cause. Drinking up to 4 cups a day was associated with 14% lower all-cause mortality (RR 0.86, 95% CI 0.82-0.89); drinking more than that was associated with no further reduction. Among non-smokers the relationship was linear, each extra cup a day associated with 6% lower all-cause mortality (RR 0.94, 95% CI 0.93-0.96). The authors conclude smoking modifies the association. All pooled studies were observational cohorts, so this is an association, not proof. |
Disagree, or know a study we missed?
We grade by evidence, not opinions. The way to weigh in is to point us to a study we haven't cited (check the evidence table above first), or to flag a problem with one we have. Every submission is reviewed; if it holds up, the grade updates and shows in Science Changes Its Mind.
Opens a short form. You'll sign in with Google so submissions are tied to a real account — we don't display your identity, and we only accept a link we can verify (PubMed, DOI, ClinicalTrials.gov).
Educational only, not medical advice. Grades and scores reflect published evidence weighted by study design and quality; see the methodology.