coffee consumption timing assoc with with total and cardiovascular mortality

 a recent retrospective observational study found that the pattern of coffee consumption (eg morning vs all day) and the quantity of coffee consumed were associated with lower risk of all-cause mortality and cardiovascular deaths (see coffee dec heart disease and mortality ESC2025 in dropbox, or  doi.org/10.1093/eurheartj/ehae871). this blog is just after one finding that caffeinated coffee and tea drinkers had lower risk of dementia and some improvement in cognitive function (https://gmodestmedblogs.blogspot.com/2026/03/coffee-and-tea-help-cognitive-function.html)

Details:

-- 40,725 adults were assessed, with 14,643 from the National Health and Nutrition Examination Survey (NHANES) from 1999–2018 who had complete information on dietary data and 772 women from the Women’s and 691 men from the Men’s Lifestyle Validation Study: WLVS and MLVS were 2 substudies to validate lifestyle assessments for the women in the Nurse's Health Study (NHS)  and the men in the Health Professionals Follow-up Study (HPFS) who had complete data on 7-day dietary record, in these observational studies

    -- WLVS was conducted from 2010 to 2012 and consisted of 796 women aged 45–80 years randomly sampled from the NHS and NHS II, two ongoing, nationwide, prospective cohorts of US female registered nurses established in 1976 and 1989, respectively

    -- MLVS was conducted from 2011 to 2013 and consisted of 914 men aged 46–82 years randomly recruited from the HPFS and members of Harvard Pilgrim Healthcare, a Boston-area health plan

        -- WLVS and MLVS substudies were designed to investigate the validity of self-reported dietary and other lifestyle assessments over a period of ∼15 months

    -- individuals with cardiovascular disease or cancer at baseline were excluded, as well as pregnant women

-- NHANES collected dietary data using 24-h dietary recalls; participants reported the timing, types, and amounts of all food and beverages consumed midnight-to-midnight the day before the interview

    -- All participants completed one dietary recall by in-person interview when they underwent a physical examination

    -- A second dietary interview was added to the survey starting from 2003, collected 3–10 days later by telephone

    -- a single dietary recall was used from 1999 to 2002 and two non-consecutive dietary recalls were used from 2003 onward

    -- Collected dietary recall data were analyzed using the US Department of Agriculture (USDA) food composition sources

    -- Coffee consumption was determined by searching USDA food code descriptions, which included all appropriate beverages containing the term ‘coffee’. Both caffeinated and decaffeinated coffee were selected, and the average amounts of consumption were calculated separately

    -- coffee drinking timing was categorized in three time periods, including morning (from 4 a.m. to 11:59 a.m.), afternoon (from 12 p.m. to 4:59 p.m.), and evening (from 5 p.m. to 3:59 a.m.).

-- the participant characteristics varied between the coffee drinkers; the following compared non-drinkers vs morning drinkers vs all-day drinkers (though the latter 2 categories were pretty similar):

    -- age 39 vs 51 vs 51

    -- race: non-Hispanic white 35% vs 53% vs 52%; non-Hispanic Black 30% vs 16% vs 10%;Mexican-American 19% vs 21% vs 21%

    -- family income: low 33% vs 24% vs 28%; intermediate: 34% vs 35% vs 35%; high 25% vs 32% vs 29%

    -- education level: less than high school 21% vs 26% vs 28%; high school 20% v 23% vs 20%; at least some college 44% vs 49% vs 49%

    -- BMI: all about 29

    -- diabetes: 11% vs 15% vs 16%

    -- hypertension: 29% vs 42% vs 39%

    -- high cholesterol: 20% vs 33% vs 34%

    -- smoking: never 58% vs 49% vs 50%; former 13% vs 30% vs 27%; current 17% vs 23% vs 21%

    -- regular exercise: 39% s 34% vs 33%

    -- AHEI (alternative healthy eating index, score 0-110, the higher the healthier): 39 vs 42 vs 43

    -- short sleep duration ( <7 hrs): 35% vs 34% vs 35%

    -- trouble sleeping: 21% vs 25% vs 22%

    -- total calorie intake, in kcal: 2064 vs 2001 vs 1980

    -- tea intake, in grams: 206 vs 178 vs 156

    -- caffeinated soda intake, in g/day: 314 vs 239 vs 188

    -- mean total coffee intake, in g/day: NA vs 459 vs 598

    -- caffeinated coffee intake in g/d: NA vs 407 vs 515

    -- decaffeinated coffee intake, in g/d: NA vs 52 vs 83

    -- percent of decaffeinated coffee: NA vs 13% v 15%

-- brief summation:

    -- compared with non-coffee drinkers, participants with morning-type and all-day-type pattern were older, more likely to be white, had higher family income, had higher prevalence of diabetes, hypertension, and high cholesterol; were more likely to smoke, less likely to drink tea, and less likely to have caffeinated beverages

    -- among coffee drinkers, participants with morning-type pattern were more likely to consume tea and caffeinated soda but consume less coffee (both caffeinated and decaffeinated coffee) compared with those with all-day patterns

-- Several covariates were adjusted in the models, including age (continuous), sex (male, female), self-reported race and ethnicity (Non-Hispanic White, Non-Hispanic Black, Mexican American, Other Hispanic, and other race), family income [Ratio of family income to poverty <1.3 (low), ≥1.3 and <3.5 (intermediate), and ≥3.5 (high)], educational attainment (less than high school, high school, and some college or above), diabetes (yes, no), hypertension (yes, no), high cholesterol (yes, no), regular physical activity [yes (≥150 min of moderate-intensity activity per week or ≥75 min of vigorous-intensity activity per week or an equivalent combination), no], smoking status [never, former, current (1 to <5 cigarettes, 5 to <10 cigarettes, 10 to <15 cigarettes, 15 to <20 cigarettes, 20 to <25 cigarettes, ≥25 cigarettes per day)], time of smoking cessation among former smokers (<1 year, 1 to <5 years, 5 to <10 years, or ≥10 years before baseline), the Alternative Healthy Eating Index diet score (quintiles), total energy intake (quintiles), caffeinated coffee intake (continuous), decaffeinated coffee intake (continuous), tea intake (continuous), caffeinated soda intake (continuous), short sleep duration [yes (sleep duration <7 h), no] and trouble sleeping (yes, no).

-- Information on death and death date was obtained through linkage to the National Death Index through 31 December 2019.

-- primary endpoint: to assess the relationship between the patterns of coffee drinking timing in the US population and their association with all-cause and cause-specific mortality

Results:

-- over the 9.8 years of follow-up, there were a total of 4295 deaths, 1268 from cardiovascular disease (CVD) and 934 from cancer

-- all-cause mortality, as compared with non-coffee drinking, adjusting for age, sex, race, NHANES cycles, family income, education levels, diabetes, hypertension, high cholesterol, smoking, physical activity, Alternative Healthy Eating Index score, total calorie intake, and the amount of caffeinated coffee and decaffeinated coffee:

    -- morning-type coffee drinking pattern had a significant 12% lower risk, HR 0.88 (0.81–0.96)

    -- all-day-type coffee drinking pattern had no association, HR 0.99 (0.90–1.10)

-- the results did not change after further adjusting for tea consumed and caffeinated soda consumed, or for further adjustment for short sleep duration and trouble sleeping:

    -- morning-type pattern had a significant 16% lower risk, HR 0.84 (0.74–0.95)

    -- all-day-type pattern had no association, HR 0.96 (0.83–1.12)

--CVD-specific mortality, compared with non-coffee drinking:

    -- morning-type pattern had a significant 31% lower risk of CVD-specific mortality, HR 0.69 (0.55–0.87)

-- cancer-specific mortality, compared with non-coffee drinking:

    -- morning-type pattern had no significant association, HR 0.97 (0.75–1.25)

-- the all-day-type pattern was not associated with risk of CVD-specific mortality or cancer mortality as compared with non-coffee drinking

-- these associations were not materially changed by considering competing risks from other causes for the analyses of CVD-specific and cancer-specific mortality

-- similar associations were observed when the patterns of caffeinated coffee drinking timing or decaffeinated coffee drinking timing were evaluated separately

-- all-cause mortality relationship between the amounts of coffee consumed vs non-coffee drinking:

    -- morning-type pattern, with coffee consumption:

        --  >0 to 1 cup/day: almost significant 15% reduction, HR 0.85 (0.71–1.01)

        -- >1 to 2 cups/day significant 16% reduction, HR 0.84 (0.73–0.96)

        -- >2 to 3 cups/day significant 28% reduction, HR 0.72 (0.60–0.86)

        -- >3 cups/day significant 21% reduction, HR 0.79 (0.65–0.97)

            -- this quantitative association with all-cause mortality displayed a significant linear trend, with P <.001

    -- all-day-type pattern: coffee consumption was not significantly associated with all-cause mortality risk

-- similar interaction pattern was observed for CVD-specific mortality, but the interaction term was not significant (fewer people in this category than in the all-cause mortality one)

--The graph below is a summation of the results:

Commentary:

-- coffee is one of the most commonly consumed beverages globally

    -- for those who consume caffeine, 25% of the total caffeine in those 15-19yo and 65% of those 35-49yo are from coffee, and 28% in those 15-19yo and 16% of those 35-49yo are from tea

-- several prospective studies have found that moderate coffee consumption is associated with decreased risk of type-2 diabetes, cardiovascular diseases and death (see https://gmodestmedblogs.blogspot.com/2019/01/coffee-and-decreased-mortality.html and https://gmodestmedblogs.blogspot.com/2020/11/diabetes-coffee-and-green-tea-decrease.html ); decreased recurrence of atrial fibrillation (https://gmodestmedblogs.blogspot.com/2026/03/atrial-fib-caffeine-decreases.html); and perhaps decreased colon cancer risk in a case-controlled study (https://gmodestmedblogs.blogspot.com/2016/04/coffee-and-decreased-colon-cancer.html). this latter blog also cites studies on coffee having a favorable effect on the gut microbiome

    -- there is a CYP1A2 genetic variant that is associated with slow caffeine metabolism and is associated with albuminuria, and hypertension; this was not taken into account in this study: https://gmodestmedblogs.blogspot.com/2023/02/coffee-consumption-chronic-kidney-dz.html

-- there is, of course, the issue that coffee consumption may well be associated with consumption of some bad health actors: smoking, sweeteners, and even some eviddence that the coffee bean used, the degree of roasting and the brewing method matters

    -- for example, there are differences in the health benefits of coffee depending on how the coffee was brewed: brewed coffee in a coffee shop has 235 mg caffeine in 12 ounces of coffee, vs 150 mg if Americano in a coffee shop, vs 138 mg if brewed elsewhere, vs 94 mg if instant coffee, vs 42 mg if brewed green tea, vs 70 mg if brewed black tea....: see coffee preparation caffeine level NEJM2020 in dropbox, or DOI: 10.1056/NEJMra1816604.

        --- perhaps the finding in this study that tea did not statistically affect the coffee consumption groups is that the caffeine content is much lower than coffee (and, besides, we have no information about the type of tea or the associated amount of caffeine consumed)

    

-- one observation from other studies is that the timing of food intake can modify the association of food with health outcomes: https://link.springer.com/article/10.1007/s00125-020-05238-w or diabetes timing of food effects Diabetologia2020 in dropbox

    -- the incentive for this study was that drinking coffee later in the day could disrupt circadian cycles and modify its effect on health outcomes

-- this observational study found that as compared to non-coffee drinkers, those consuming coffee in the morning had significantly lower risks of all-cause mortality and CVD-specific mortality, unlike those individuals consuming coffee during the day

    -- this coffee benefit peaked with consumption of >2 to 3 cups/day 

    -- the benefit of coffee for all-cause mortality in this study was largely attributed to the benefit for cardiovascular mortality

        -- especially since there are potentially devastating non-fatal cardiovascular events, it makes sense (to me) to also assess the benefits for cardiovascular morbidity as well as mortality

    -- other studies supporting the conclusions of this current one include those accessing a UK Biobank, the European Prospective Investigation into Cancer and Nutrition study, and the National Institutes of Health--AARP Diet and Heart Study, where even heavy coffee consumption was significantly associated with lower risk of all-cause and CVD-specific mortality. there are, however, some negative studies. Perhaps this can be explained by some heavy coffee drinkers who also smoke or have lots of sugar/artificial sweeteners (https://gmodestmedblogs.blogspot.com/2023/06/health-effects-of-non-sugar-sweeteners.html and https://gmodestmedblogs.blogspot.com/2020/11/cardiovasc-disease-inc-with-sweet.html) and perhaps lots of saturated fat creamers that increase CVD risk

    -- one evident benefit of this study is that the databases used are so well developed with frequent 24-hour food recalls as well as the most accurate food intake measure, the 7DDR (7-day Dietary Record)

-- potential mechanisms to explain this divergent finding from coffee intake timing:

    -- later coffee consumption is association with disruption of circadian rhythms, as suggested by evidence that there is a 30% decrease of melatonin in those with heavy coffee consumption in the afternoon or evening vs those drinking decaffeinated coffee, along with decreases in total amount of sleep, quality of sleep and 6-sulphoxymelatonin (main metabolite of melatonin) throughout the following night: https://www.sciencedirect.com/science/article/pii/S1389945702000151?via%3Dihub

    -- there is some evidence that low levels of melatonin are associated with high levels of oxidative stress, blood pressure and CVD risk. Melatonin has been found to have direct interactions with the nervous system and indirect interactions with blood vessels in the heart

    -- coffee's health benefits may well be associated with its well-documented anti-inflammatory effects, increases in adiponectin, and decreases in LDL cholesterol levels (the latter mostly associated with green tea consumption): https://www.sciencedirect.com/science/article/pii/S1043661822005424?via%3Dihub

         -- some of the anti-inflammatory markers and cytokines have internal circadian patterns and have the highest anti-inflammatory effects in the morning and then decline to the lowest level around 5PM, hence being more beneficial with morning coffee consumption

            -- this is true from both caffeinated and decaffeinated coffee consumption

Limitations:

-- this retrospective cohort study allows conclusions about associations but in the absence of a randomized controlled trial, does not allow for determining causality, since there is potential unmeasured  confounding

-- as noted in the patient characteristics, there were pretty significant differences between the non-coffee drinkers and the coffee drinkers. the statistical analysis did mathematically control for these factors, but with so many of them, there is increased potential for inaccuracy

-- there is an issue with relying on death records, since the cause of death may not be accurate (https://www.ahajournals.org/doi/10.1161/CIR.0000000000000156). My guess from my experience over the years is that there may well be a bias to labeling the cause of death to cardiovascular causes: the phone call from a funeral home to me to provide the necessary cause of about a death at home is basically guess work (?MI, ?pulmonary embolism, ? the underlying cancer or, ???); no autopsy to confirm the cause leads to the reasonable guess of it being a cardiac cause...

-- the covariates assessed were in broad categories, potentially obscuring the reality that there are large differences between buckets analyzed: eg, there are large differences between those who did not exercise and those who exercised at least 150 minutes per week. and we know that smoking cessation can have health effect up to 20-30years after cessation (https://gmodestmedblogs.blogspot.com/2026/01/smoking-cessation-20-30yrs-to-normalize.html), though their buckets were up to 10 years only, and there was no distinction about whether the diabetes or hypertension was controlled and what meds were taken (and some meds actually increase cardiovascular risk)

-- there could be a recall bias in the NHANES study, since individuals need to document their exposures and covariates in a self-reported questionnaire

    -- there could also be a bias if the individuals drinking coffee in the morning were different than those who chose to drink coffee later in the day. were there unmeasured differences in those who drank coffee in the morning? did they do more exercise or eat differently in ways not measured and perhaps those differences improved affected their cardiovascular health (and the measures of exercise, for example, were not continuous but were binary: [yes (≥150 min of moderate-intensity activity per week or ≥75 min of vigorous-intensity activity per week vs no)? were they on medications that improved cardiovascular health (there was no information in the study of what medications people were on)?

-- there is a clear genetic difference in the metabolism of caffeine. did those who drank morning coffee happen to have slower metabolism and therefore had longer systemic exposure to the beneficial aspects of caffeine on the cardiovascular system; perhaps they actually had a similar caffeine level in their blood as those who drank coffee later in the day??

-- there are significant differences in foods eaten per the AHEI, perhaps times they are eaten, what beverages with food, etc in the US vs other countries (and even within different parts of the US), that could affect the generalizability of the results in this study

-- as per above, there is no information about how the coffee was brewed. this information very well could affect the results of the study. perhaps those individuals drinking coffee in the morning went to a coffee shop on the way to work and consumed lots of caffeine in the process, yet those drinking it later may have had instant coffee or coffee brewed at work with significantly less caffeine content??? 

    -- also, there w\was no information about the types or quantities of tea consumed that would add to the blood caffeine levels

so,

-- coffee drinking in the morning in this study was associated with decreased all-cause mortality and cardiovascular-specific mortality

-- this was true independent of the amount of coffee ingested

-- however, the best results for mortality were in those drinking >2 to 3 cups/day in the morning

-- as with all studies, there are limitations to the conclusions as noted above, including how the coffee was made; these types of limitations are more evident in retrospective observational studies than well-designed randomized controlled trials

    -- so in this study, though they analyzed some of the best dietary information of any studies, we can only conclude that there were associations of coffee consumption timing and quantity and not a clear causal relationship....

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

to get access to the dropbox: https://www.dropbox.com/scl/fo/pcf4vymtpo77f49wd2ljr/ANyJqB40CuBd-DMEr67XDoI?rlkey=5dsalmuhf8ekd0svsgvn93xm1&dl=0

or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them