diet and cardiovasc or diabetes outcomes

 Given my recent remarkably upsetting blog about dark chocolate and lead poisoning, here is a recent review of the evidence of an array of foods and nutrients on cardiovascular disease and diabetes (see diet cardiovasc diabbenefit jama2022 in dropbox, or doi:10.1001/jamanetworkopen.2021.46705)

 

Details:

-- PubMed was assessed for articles from May 2015 to February 2021 for systematic reviews of randomized controlled trials and prospective cohort studies dealing with consumption of one or more dietary factors of interest, with the quality of evidence being assessed rigorously

-- inclusion criteria included 43 food, beverage, or nutrient groups that the authors thought were of interest: fruits; fruit juices; vegetables; potatoes; beans or legumes; nuts or seeds; whole grains; refined grains; milk; yogurt; cheese; unprocessed red meats; processed meats; fish or seafoods; lean fish; fatty fish; eggs; sugar sweetened beverages; nonnutritive sweetened beverages; coffee; tea; chocolate; saturated fatty acids; monounsaturated fatty acids; polyunsaturated fatty acids; seafood based omega-3 fatty acids; plant based omega-3 fatty acids; trans fatty acids; total protein; animal protein; plant protein; dietary cholesterol; dietary fiber; cereal fiber; fruit fiber; vegetable fiber; legume fiber; glycemic index; glycemic load; dietary sodium; dietary potassium; dietary calcium; and total energy

 

Results:

-- 2058 potentially relevant reports were identified, and the final selection included 28 meta-analyses representing 62 associations between diet and cardiometabolic disease (CMD) 

-- the following items did have significant associations between food groups and CMD:

 

    -- for cardiovascular outcomes:

   --nutrients (all statistically significant)    

    -- dietary fiber: 3 studies with an average of 23%-46% decreased risk for 20g/d consumed, one study assessed cardiovascular disease, finding a 46% reduction per 20 g per day; another coronary heart disease finding a 41% reduction per 20 g per day; and another stroke, finding a 23% reduction per 20 g per day

         -- fruit fiber: 2.2 million person-years, 5% decreased risk of stroke for 2 g/d

         -- vegetable fiber: 2.2 million person-years, 9% decreased risk of stroke for 2 g/d

         -- glycemic index: 300,000 people, 24% increased risk of coronary heart disease for 10 units

         -- glycemic load: 300,000 people, 57% increased risk of coronary heart disease for 80 g/2000 kcal

         -- polyunsaturated fatty acids replacing carbohydrates: 10% decreased risk of coronary heart disease per 5% energy/d

         -- polyunsaturated fatty acids replacing saturated fatty acids: 9% decreased risk of coronary heart disease by replacing saturated fatty acids for 5% energy/d

         -- transfatty acids: 140,000 people, 23% increased risk of coronary heart disease by replacing trans fats for 2% energy/d

         -- sodium: 250,000 people, 6% increased risk of stroke for1 g/d

         -- potassium: 230,000 people, 13% decreased risk of stroke for 1g/day

 

      -- foods (all statistically significant):

        -- fruits: 5 studies with total of more than 4 million people, average of 10% decreased risk for 100 g/d

        -- vegetables: 4 studies with 3 million people, average of 7% decreased risk for 100 g/d

        -- nuts and seeds: 2 studies with 1 million people, average of 48% decreased risk for 28 g/d

        -- whole grains: 3 studies with 1.1 million people, average of 15% decreased risk for 50g/d

        -- red meats: 3 studies with 1.1 million people, average of 10% increased risk for 300g/wk

        -- fish and seafood: 4 studies with 1.2 million people, average of 10% decreased risk for 300g/wk

        -- chocolate: 5 studies with 1.3 million people, average of 8% decreased risk for 10g/d

 

    -- per diabetes outcomes: 

      --nutrients (all statistically significant):

        -- cereal fiber: 2 studies, 470,000 people, 13%-17% reduction for 5 g/d in the first study and 20 g/d in the second

         -- glycemic index: 350,000 people, 27% increased risk for 10 units

         -- glycemic load: 625,000 people, 26% increased for 80 g/d per 2000 kcal

         -- polyunsaturated fatty acids replacing carbohydrates: 80,000 people, 10% decreased risk by replacing carbs for 5% energy/d

         -- total protein: 400,000 people, 9% increased risk by replacing total protein for 2% energy/d

         -- animal protein: 360,000 people, 12% increased risk by replacing animal protein by 2% energy/d

 

       -- foods (all statistically significant):

        -- potatoes: 360,000 people, 5% increased risk for 100g/d

        -- whole grains: 460,000 people, 21% decreased risk for 50 g/d

        -- red meats:  530,000 people, 15% increased risk for 300g/week

        -- processed meats: 50,000 people, 16% increased risk per 150 g/week

        -- yogurt: 36,000 people, 26% decreased risk per 244 g/d

     

Commentary: 

-- cardiometabolic disease (CMD) has been increasing internationally for the last three decades, with an estimated 523 million adults living with cardiovascular disease and 463 million with diabetes worldwide in 2019

-- it has been clear for a long time that CMD diseases are associated with potentially modifiable risk factors, including diet and exercise, as well as normal body weight, stress, etc, as well as some meds (eg statins and GLP-1/SGLT-2 meds for diabetes)

-- in the above analysis, there was no evidence of an association found for fruit juice, beans or legumes, refined grains, cheese, lean fish, fatty fish, eggs, nonnutritive sweetened beverages, coffee, saturated fatty acids, monounsaturated fatty acids, dietary seafood or plant omega-3 fatty acids, plant protein, dietary cholesterol, legume fiber, dietary calcium, and total energy: though in many of the observational studies, there were associations with CMD outcomes, but these associations involved either too few studies or not high enough quality data (per Bradford-Hill criteria)

    -- coffee (with caffeine, a xanthine similar to the finding in chocolate) was notably absent int benefit. As mentioned in previous blog, there are lots of studies on caffeine which do find mortality and other benefits, eg:

        -- http://gmodestmedblogs.blogspot.com/2019/01/coffee-and-decreased-mortality.html is a pretty recent blog on studies on coffee/caffeine and the associated multiple health benefits

-- and some analyses had discordant results for cardiovascular and diabetes: eg fruits and vegetables, as well as various forms of fiber (total dietary, fruit and vegetable fiber) were only associated with cardiovasc disease and not diabetes outcomes

 

Limitations:

-- the major limitation of a systematic review (as is the case with the USPSTF and some other recommendations) is that they are based on high quality studies. And, many important and clinically useful findings/interventions are dismissed just because the necessary studies have not been done. i.e., the lack of a recommendation does not mean that the intervention is not justified...

    -- on the other hand, some of our expected observations may not be true when tested rigorously

-- these were pretty much all observational studies (the sodium one did have an RCT), and there is undoubtedly much confounding that was not measured and might have had significant effects: eg, those who ate more healthy foods or nutrients were likely also to be doing other healthy things that might have explained some of their better outcomes, such as higher intensity of cardiovascular exercise, or higher general high quality foods consumed. and this brings up a few related issues:

    -- these analyses tend to look at single foods and not the overall diet, and though some people may have eaten more fruits and vegetables, they may also have eating fewer carbs at the same time, or taken certain vitamins, or...

    -- even measures of fruits and vegetable or other ingredients in some studies do not have the fine gradations necessary, such as types of fruits/vegetables or quantities (the latter often being condensed to an either binary/yes or no, or in large buckets such as 1-2 vs >2).  

    -- stress and psychosocial variables are often not measured, and we know that stress (and the availabillity of social supports to moderate the effect of stress), depression, access to healthy foods, access to sites to exercise, etc are important aspects of diet and also of CMD

        -- for a recent blog on stress and cardiovascular disease, see http://gmodestmedblogs.blogspot.com/2022/01/stress-induced-cardiovascular-disease.html

    -- also, some of the studies used biomarkers of disease, and these studies were typically not included (eg some of the studies on polyunsaturated fatty acids and diabetes had biomarkers, which themselves were associated with lower diabetes risk). But some of these studies do not bubble up to being rigorous enough for this analysis.

    -- another basic issue with many dietary studies is reductionism: we (and other species) have evolved to have metabolic systems that function well on the available foods (fruits, vegetables, etc with their associated micro and macronutrients). And the benefits are likely related to the complex interactions of the multiple components of these whole foods with our bodies. Studies that isolate specific micro or macronutrients may miss the boat on these perhaps essential interactions in the natural foods that promote health. So, perhaps it is not fiber that is the healthy component, but fiber along with the vitamins and other micronutrients in apples that matter. And just taking manufactured fiber (eg psyllium) just may not the same as eating an apple….

-- So, in these types of observational studies, we can only make associations and not assess causality.  

-- another issue with systematic reviews is which articles are included and which are excluded from the analysis; i.e., it is not so uncommon that two different systematic reviews/meta-analyses may come out pretty much simultaneously in the medial literature and have different conclusions based on which articles they included in their analyses

 

So, lots of data, with an eye to assessing only the highest quality data available. But as a result, several very likely healthy foods were not found to be so healthy, and this is likely because we just do not have sufficiently robust studies on many of them (and, of course, some of them actually might not be as healthy as we think….). and, unfortunately, we will never really have the opportunity to conduct very large randomized controlled trials with people segregated into 2 or more groups, eating only specified foods, and followed for many years. in this light, and since both cardiovascular disease and diabetes have some important genetic components, there may be utility in utilizing the technique of mendelian randomization, which approximates randomized controlled trials though with some limitations, and might provide further insight into causality (eg see mendelian randomization bmj2019 in dropbox, or doi.org/10.1136/bmj.k601) . for prior blogs on mendelian randomization, see http://gmodestmedblogs.blogspot.com/2019/05/stroke-risk-lowest-if-zero-alcohol.html

 

Addendum: since some of you asked, I did check my lead level as a dark chocolate overconsumer, and it was 2.6 mcg/dL: this is below the “normal” level of <5 mcg/dL, but above what is necessary for the human body (0 mcg/dL). Though, as mentioned in the blog on chocolate and lead/cadmium, there are other veges as well that contribute to the lead burden, including sweet potatoes, spinach and carrots (which I also eat). and, it does seem reasonable that the FDA actually check all consumables that might be exposed to lead and intervene when necessary......

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 (2 options):

1. go to https://www.bucommunitymedicine.org/ , a website from the Community Medicine section at Boston Medical Center.  This site does have a very searchable and accessible list of my blogs and is the easiest to view blogs and displays more at a time.

2. go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

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

  -- 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

  

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