coffee and decreased mortality, repeatedly affirmed

a new article found that coffee consumption was associated with decreased overall mortality, and this was independent of genetic variations of caffeine metabolism (see coffee and dec mortality jamaintmed2018 in dropbox, or doi:10.1001/jamainternmed.2018.2425).

Details:

--9.2 million individuals from the UK Biobank, a population-based study, including data on baseline demographic, lifestyle, and genetic data, from 2006 to 2016

--498,134 people (non- pregnant) consented to participate

--mean age 57 years, 54% female, 78% coffee drinkers (of whom 56% drank instant coffee, 23% ground, and 19% decaffeinated), 95% white

--coffee drinkers were more likely to be male (range from 42% nondrinkers to 57% if >7 cups/d), white (range from 90% nondrinkers to 98% if >7 cups/d), former smokers (range from 31% nondrinkers to 37% if 4-5 cups/d), and drank alcohol

--those drinking 1-3 cups/d were older (59 yo vs lowest grouping being 55 yo), more likely to have a university degree (42% vs 33% if drink >7/d) and more likely to report "excellent" health (17% vs 14%)

--those drinking >4 cups/d were more likely to be instant coffee drinkers (67% vs 51%) and be current smokers (up to 34% if >7 cups/d vs 10% in nondrinkers)

--primary outcome: relationship between coffee intake and mortality

--also assessed this relationship by caffeine metabolism, defined by a genetic score of previously identified polymorphisms in AHR, CYP1A2, CYP2A6, and POR (these have been shown to have an effect on caffeine metabolism). 

--14225 deaths over 10 years of follow-up

Results:
--with non–coffee drinkers as the reference group, Hazard Ratios (HRs) for all-cause mortality in coffee drinkers (multivariate adjustment for age, sex, smoking status, general health status, BMI, diabetes, history of cancer/heart attack/stroke, physical activity, alcohol, coffee type, genetic risk score):
    -- < 1cup/d: 6% lower risk, HR 0.94 (0.88-1.01)
    -- 1 cup/d: 8% lower risk, HR 0.92 (0.87-0.97)
    -- 2-3 cups/d: 12% lower risk, HR 0.88 (0.84-0.93)
    -- 4-5 cups/d: 12% lower risk, HR 0.88 (0.83-0.93)
    -- 6-7 cups/d: 16% lower risk, HR 0.84 (0.77-0.92)
    -- 8 or more cups/d: 14% lower risk, HR 0.86(0.77-0.95)
    -- trend was statistically significant for inverse dose-dependent trend (p<0.001)
--caffeine benefits were found for all-cancer (58% of deaths), and all-cardiovascular disease mortality (20% of all deaths)
    --there were trends to lower incidence specifically for stroke, coloretal cancer, female breast cancer
    --all-cancer deaths progressively decreased with more coffee (p=0.003), with 13% fewer deaths in those drinking 8 or more cups/d (statistically significant decrease if at least 2-3 cups/d)
    --though both all-cardiovascular disease deaths and stroke deaths had significant trends to benefit with more coffee (both p=0.05), the only significant specific benefit was from drinking 2-3 cup/day, associated with a 14% reduction in all-cardiovascular deaths [HR 0.86 (0.77-0.96)] and 30% decrease in stroke deaths (HR 0.70 (0.55-0.89)]. [though part of the issue with statistical significance is that the number of deaths in each category of coffee consumption is pretty small]
--benefits found for instant, ground, and decaffeinated coffee, though they were stronger for ground coffee
    --for ground coffee, both the significance of the trends was more powerful for ground coffee than the others: the overall benefit at the >=6 cups/d was 26% decrease in all-cause deaths [HR 0.74 (0.60-0.92)], 27% decrease in all-cancer deaths [HR 0.73 (0.55-0.98)], and 49% decrease in all-CVD deaths [HR 0.61 (0.37-1.00)]
    --instant coffee and decaffeinated coffee had similar benefits
--no meaningful difference by sex, age group, BMI, history of diabetes, or previous diagnosis of cancer, heart attack, or stroke
--there was a stronger inverse relationship between coffee and mortality in those reporting worse health, and for former and current smokers than never-smokers (nonsignificant benefit in the latter)
--No difference by genetic coffee metabolism
    --eg: the HR varied from 0.70 (0.53-0.94) to 0.92 (0.78-1.10) for those drinking >= 6 cups/d, varying some with differences in the number of genetic alleles present, but no overall consistent evidence of effect modification across strata of caffeine metabolism score (P=.17 for heterogeneity).
--on subgroup analysis, there was a pretty consistent trend to benefit on all-cause mortality with more coffee consumed for men, women, age younger or older than 55, former and current smokers, different overall health rankings, BMI gradations, and those without history of diabetes

Commentary:
--there have been a slew of studies (and several prior blogs) noting pretty consistent health benefits from coffee consumption
    -- 209K health professionals in the Nurses' Health Studies and men in the Health Professionals Follow-up Study with 332K deaths, finding decreased mortality in all coffee drinkers (if restrict analysis to never smokers; including smokers had association only for those drinking 1-5 cups/d). inverse assocations between coffee drinkers and deaths from cardiovascular disease, neurological diseases, and suicide. no relation with cancer. see coffee and mortality circ2015 in dropbox, or DOI: 10.1161/CIRCULATIONAHA.115.017341; and detailed analysis at http://gmodestmedblogs.blogspot.com/2015/11/drink-coffee-and-live-longer.html
   --186K people in the Multiethnic Cohort, with 59K deaths, followed 16 yrs, finding lower mortality (12% if 1 cup/d; 18% if 2-3 cups/d; 18% if 4 or more cups/d), in African Americans, Japanese Americans, Latinos and whites. see coffee and dec mortality AIM2017 in dropbox, or doi:10.7326/M16-2472
   -- WHO report assessing caffeine as a possible carcinogen, finding that there was a lower risk of liver cancer, also likely lower risk of endometrial cancer, leading to conclusion that “overall coffee drinking was evaluated as unclassifiable as to its carcinogenicity to humans”. see http://gmodestmedblogs.blogspot.com/2016/06/coffee-hot-drinks-and-cancer.html
    --a study in 10 European countries with 521K people, followed 16 yrs, and 42K deaths, finding an association between higher coffee consumption and lower all-cause mortality (12% in men and 7% in women), digestive disease mortality (59% in men and 40% in women), circulatory disease mortality (22% in women), and cerebrovascular disease mortality (30% in women), but a 31% increased risk of ovarian cancer mortality. see coffee and dec multinational mortality AIM2017 in dropbox, or doi:10.7326/M16-2945, see  http://gmodestmedblogs.blogspot.com/2017/08/coffee-and-decreased-mortality.html for details
    -- coffee consumption was associated with decreased colorectal cancer. see http://gmodestmedblogs.blogspot.com/2016/04/coffee-and-decreased-colon-cancer.html for details
    -- a Korean study finding that coffee consumption was associated with a lower coronary artery calcium score on cardiac CTs (coronary artery calcium is highly correlated with clinical cardiovascular events). http://gmodestmedblogs.blogspot.com/2015/03/coffee-and-decreased-coronary-artery.html 
--the 2015 US Dietary Guidelines have concluded that coffee consumption up to five 8-ounce cups/d can be part of a healthy diet
--part of the impetus for this study was reports that some of the above genetic variations in caffeine metabolism might increase cardiovascular risk.  
    --but the finding that those with genetic variations leading to decreased caffeine metabolism/higher blood levels still had mortality benefit is certainly consistent with a conclusion that even high levels of caffeine are not toxic
    --there was an increased benefit (though weak) to those who metabolized caffeine more slowly.  but was this possible benefit diluted by including noncaffeinated beverages in the analysis?? (unclear from the study). Or was this because the real issue is not caffeine but other components in coffee that are still present after decaffeination? (other studies, eg some cited above, also found that decaf had a mortality benefit)
    --and, this conclusion that other non-caffeinated coffee products might really provide benefit reinforces that we should include decaf coffee in future analyses of coffee benefit/harm. it would be interesting to see if caffeinated soft drinks provide any benefit over non-caffeinated ones (thereby isolating the potential role of caffeine itself)
--the relationship between instant and ground coffee: ? distorted by fact that there was a differences in the groups (eg by education level, with those having higher education drinking more ground coffee). Though instant coffee has lower concentrations of bioactive compounds
--proposed mechanism: coffee leading to reduced inflammation, improved insulin sensitivity, improved endothelial function
--limitiations. only 5.5% of the UK Biobank people participated.  ? healthy volunteer bias in those who did agree to participate. also only one population, and ethnically/racially not so diverse

so, pretty consistent associations in these observational studies: coffee seems to be good for you.  which is certainly welcome to me, who has devolved into requiring 1-2 cups every morning to do these blogs (and function in general)

geoff

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

to get access to blogs since 8/15/17:

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

2. click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category

3. or you can just type in a name in the search box and get all the blogs with that name in them

to access older blogs from the BMJ website, from October 2013 until 8/15/17: go to http://blogs.bmj.com/bmjebmspotlight/category/archive/ 

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list