Search This Site!
Nutrition Past and Future

The Drivers of the Herd, Part 10

Siri-Tarino’s Meta-Analysis, Part 2 (Saturated Fat and Stroke)

Slide 3                   Siri-Tarino, Patty W., et al. "Meta-analysis of prospective cohort studies evaluating the association of saturated fat with cardiovascular disease." The American journal of clinical nutrition 91.3 (2010): 535-546.


This meta-analysis of prospective cohort studies by Patty Tiri-Sarino and colleagues is widely distributed online. The promoters of unhealthy saturated fats use it as evidence that this stuff is just fine. In the previous video we’ve examined the studies upon which this review is based that seem to excuse saturated fat from any involvement in promoting heart disease. We found that those papers were deeply flawed but usually argued against saturated fat upon closer reading. In this video we’ll carry out the same exercise, but this time we’ll be talking about saturated fats and stroke risk.



Slide 4                   The first favorable study listed for stroke is Goldbourt 1993. We’ve looked at this article already.



Slide 5                   You can see it was also listed under coronary heart disease. You can clearly see in the title it was a study of “coronary heart disease mortality.” I don’t know why this is in the section for strokes. I’d like to tell you what this one said about stroke but it didn’t say anything about it as far as I can tell.



Slide 6                   Goldbourt, U., S. Yaari, and J. H. Medalie. "Factors predictive of long-term coronary heart disease mortality among 10,059 male Israeli civil servants and municipal employees. A 23-year mortality follow-up in the Israeli Ischemic Heart Disease Study." Cardiology 82.2-3 (1993): 100.


There you see in both the title and the conclusion that this was a study of coronary heart disease mortality, not stroke. I don’t get it. I’ll mention now that I do maintain an Errata page on my site. Let me know if I missed something about stroke here. I doubt I did, though.



Slide 7                   The next one that’s favorable is by Gillman and this one is definitely about stroke.



Slide 8                   Gillman, Matthew W., et al. "Inverse association of dietary fat with development of ischemic stroke in men." JAMA: the journal of the American Medical Association 278.24 (1997): 2145-2150.


This one concluded that saturated fat was associated with a reduced risk of stroke. I’ve seen some crazy studies, and I must say, this one ranks up there among the craziest. I’ll need to show you a table so you can see why.



Slide 9                   Here you see they divided their men by quintiles of fat intake. Look at the line for total fat and you can see that indeed, as the quintiles go up so does the fat consumption. No problem there.



Slide 10                 Next we look at age. You see the mean age is more or less the same in each quintile. Again, no problem.



Slide 11                 This one is key. They assigned a number for the level of physical activity for each participant. As you can see, this number increased with fat intake but just by a little bit. There really wasn’t a big difference in physical activity in each quintile. Remember that, please.



Slide 12                 Now you see that body mass index also was more or less the same across the quintiles. Again, there’s nothing too strange about that. Now get up close to your screen.



Slide 13                 You can see that energy intake increases across each quintile. And it doesn’t just increase. It increases radically, so that the top quintile consumed 1.8 times the energy – nearly twice the energy – of the bottom quintile! This is pure nuttiness! They are saying that men who ate more fat weren’t any fatter and they weren’t any younger and they weren’t any more physically active, and yet they just burned off almost twice the energy of the low fat group! Now that’s a metabolic advantage! I looked at this and I thought I must be misunderstanding this somehow.




Slide 14                 Gillman, et al. "Letters." JAMA 279.15 (1998): 1172-1173.


It turns out others noticed the same thing and wrote in to say that something was very wrong here. The authors replied. They said, “The confounding effect of physical activity provides the explanation.” Now you just saw that that can’t possibly be true.



Slide 15                 Physical activity was hardly any different. Their own data says that’s not what happened. Now think about this. Patty Siri-Tarino and her coauthors worked with this data and they found it to be perfectly fine. I said their analysis was superficial. I hope you can see why.



Slide 16                 It gets even more suspicious than that. They calculated their risk ratios based on how many calories they ate. That means that they distorted their data so that the more someone ate, the less they contributed to the overall risk of stroke. Since calories increased so much along with fat, they systematically biased their findings in favor of fat!




Slide 17                 Of course, they found a reduced risk for fat consumption! How could they not!



Slide 18                 There were a couple other weaknesses to this one, not that we need to see any more. This was a 20-year follow-up study. By the end some men would have been 85.



Slide 19                 Their data was based on a single recall of what they ate in a 24-hour period. I hope you realize by now that this is not a precise way of doing things.



Slide 20                 They acknowledged that this was a weakness of their study but they claim that if their data were more precise the associations they found would have been even stronger. But how can they know that? This assumes that there would have been no tendency to underreport fat intake. But many studies have shown this is not true. I’ll talk about that later in the playlist. I have to say, this was not a paper that made a strong case for saturated fat. Low carbers, don’t worry. It gets better.



Slide 21                 A study by He looks pretty favorable for saturated fat.



Slide 22                 He, Ka, et al. "Dietary fat intake and risk of stroke in male US healthcare professionals: 14 year prospective cohort study." BMJ: British Medical Journal327.7418 (2003): 777.


This one was conducted in the US using a cohort of health professionals. It found no association between fat or cholesterol and stroke in men.



Slide 23                 The authors made the decision to take out of their pool of men anyone with heart disease, diabetes, or high cholesterol. They did this because they assumed that such people would be told to cut back on saturated fat and cholesterol and they assumed that these men would actually follow that advice. These are two assumptions for which it seems they provided no statistical basis. It was a questionable decision. Again, saturated fat raises cholesterol. If you remove the men with high cholesterol you may be removing the most instructive cases in the cohort. It would have made more sense to remove men once they verified that they in fact did change their diets.



Slide 24                 This was a nice study in that it provided lots of tables of data. I looked it over and it does appear that associations between fat intake and stroke weren’t apparent. Some trends were interesting though. Let’s look at the associations between intakes of dietary cholesterol and hemorrhagic stroke. Remember what you’re seeing here because it will be relevant to other studies we’ll see. Notice that as dietary cholesterol increased in this cohort of American men, the risk of hemorrhagic stroke also tended to increase. Compare the groups with the lowest and highest intakes of cholesterol in particular.



Slide 25                 You see that there was also an increasing trend for hemorrhagic stroke as red meat consumption increased. Again, make a mental note of this.



Slide 26                 Lastly, look at what happened with increasing egg consumption. Eggs seemed to be powerfully protective for hemorrhagic stroke in the highest category of consumption. And it is possible that eggs really were protective here. But it’s also possible that their removal of men with high cholesterol skewed their findings. I have a video coming up in this playlist, number 20, about the effect that eggs have on serum cholesterol. The most avid egg-eaters may have had especially high cholesterol. They may have been selectively removed here. What seems odd is that there is no other commonly eaten food in America that has a cholesterol content anywhere near what eggs have. You would think that the trend for dietary cholesterol would mirror the trend for eggs. But we just saw it didn’t, and we see that excluding the top quartile, there is no real trend to speak of.



Slide 27                 Here’s that table for dietary cholesterol again. The contrast between the two tables is strange. Moving on.



Slide 28                 We are now up to Iso, the single most favorable study for saturated fat on their list. And it really is favorable for saturated fat consumption as it relates to stroke risk.



Slide 29                 Iso, Hiroyasu, et al. "Fat and protein intakes and risk of intraparenchymal hemorrhage among middle-aged Japanese." American journal of epidemiology157.1 (2003): 32-39.


There is the abstract. Saturated fat and animal protein were found to be protective against hemorrhagic stroke within a Japanese cohort. If you’re remembering our mental note, this is not what we just saw in an American cohort. I’ve mentioned in a past video that findings like this have most commonly come out of Japan, where salt consumption and smoking rates are higher than in most other countries. Both of those tendencies increase the risk of stroke. What’s interesting here is that saturated fat appeared protective whether these people had high blood pressure or not.



Slide 30                 As usual, this was based on a single 24-hour recall. That’s not great but it is the norm.



Slide 31                 Because they were tracking their participants over different periods of time, they made the decision to adjust for expected changes in diet quality. It is true that diets changed rapidly in Japan over this period but this undoubtedly adds an additional layer of uncertainty to their findings. They assumed national trends applied to these folks. We can’t know if that was warranted.



Slide 32                 These researchers did make adjustments for serum cholesterol. This would seem to raise concerns about overadjustment but it turns out that this choice didn’t matter in this study.



Slide 33                 The trends are strong enough and consistent enough that it is likely that there was some sort protection associated with saturated fat here. That adjusted risk ratio only considers age, gender, and location so that’s not a problem. Because this is the first study on our list that seems favorable for saturated fat I think it is wise to stipulate that this finding might apply specifically to the Japanese and their risk of a particular type of stroke. We don’t know have a reason to believe this can be generalized to everyone. Let’s remember this one because it probably has real meaning. You’ll see why soon.



Slide 34                 We have only one favorable study left. This one is by Sauvaget.



Slide 35                 Sauvaget, Catherine, et al. "Animal protein, animal fat, and cholesterol intakes and risk of cerebral infarction mortality in the adult health study." Stroke 35.7 (2004): 1531-1537.


Interestingly, this one also comes to us from Japan. We are told that in Japan high consumption of animal fat and cholesterol is associated with a reduced risk of a type of ischemic stroke. The association was weak for animal protein but saturated fat and dietary cholesterol very clearly were associated with reduced risk.


Let’s step back for a moment. Not only would it be questionable to generalize findings from Japan to everyone else …



Slide 36                 It would be especially questionable to generalize for us all based findings from a cohort of survivors of the atomic blasts in Japan at the end of the second world war. Every single person under study here was exposed to nuclear radiation.



Slide 37                 The authors thought that nuclear radiation wouldn’t be too important a confounder in their study because as far as they knew, radiation exposure doesn’t affect stroke risk. They also thought they could statistically control for the differing locations of these participants during those explosions in their analysis since radiation exposure would have varied by location. They were wrong on both counts.



Slide 38                 First, you see it clearly stated here that they believed that stroke risk was unrelated to radiation exposure.


Slide 39                 Shimizu, Yukiko, et al. "Radiation exposure and circulatory disease risk: Hiroshima and Nagasaki atomic bomb survivor data, 1950-2003." BMJ: British Medical Journal 340 (2010). BMJ 2010;340:b5349


And you see it is clearly stated in this study that they were probably wrong. “This study provides the strongest evidence available to date that radiation may increase the rates of stroke and heart disease at moderate dose levels.”



Slide 40                 Tonda, Tetsuji, et al. "Investigation on circular asymmetry of geographical distribution in cancer mortality of Hiroshima atomic bomb survivors based on risk maps: analysis of spatial survival data." Radiation and environmental biophysics 51.2 (2012): 133-141.


Moreover, you can tell from this study that they did not have the information they would have needed to even attempt to correct for exposures based on location. There was likely an element of risk from radiation not explained by direct exposure. Exposure also occurred through rainfall.


So as you see, underlying assumptions behind this study were faulty, and as long as that is true, their conclusions are suspect. Perhaps saturated fat and cholesterol are protective against strokes among those exposed to nuclear radiation. But we can’t be sure of that based on this study, and we certainly can’t use this study to argue that saturated fat and cholesterol are protective for the rest of us.





Slide 41                 YAMAGISHI, Kazumasa, et al. "Smoking raises the risk of total and ischemic strokes in hypertensive men." Hypertension research 26.3 (2003): 209-217.


As I said earlier, much of the literature which argues for a protective effect from animal foods comes from Japan. This should cause us to wonder if there is something unique about life in Japan that is at play in these studies. I’ve looked into this and I think there are at least a couple unusual factors at work here.


First, as I said earlier, smoking and hypertension are more common in Japan than most other developed nations. These factors synergistically increase stroke risk. Attempts to correct for these factors may not have been adequate in some studies.



Slide 42                 Burger, Joanna, et al. "Sushi consumption rates and mercury levels in sushi: ethnic and demographic differences in exposure." Journal of Risk Research. ahead-of-print (2013): 1-17.


Here are two important factors I have never seen included in discussions about diet and stroke risk. First, the Japanese are avid consumers of fish. We all know at this point that our primary source of mercury exposure is through eating fish. Bluefin tuna are especially polluted by mercury. I only provide one slide to make this point because I will assume we are all aware of this.



Slide 43                 Houston, Mark C. "Role of mercury toxicity in hypertension, cardiovascular disease, and stroke." The Journal of Clinical Hypertension 13.8 (2011): 621-627.


What we may not all realize is that mercury exposure increases our risk of stroke. This is an important confounding variable not considered in the studies examining nutrition and stroke risk in Japan. If you think about this you may understand why it’s important. If a Japanese individual is eating more land-based, high-cholesterol foods like beef and eggs, he is probably eating less fish. Therefore, he would be exposed to less mercury and perhaps have a bit lower risk of stroke. Eggs and land-sourced meats may displace fish in the diet.



Slide 44                 Moon, Katherine A., et al. "Association Between Exposure to Low to Moderate Arsenic Levels and Incident Cardiovascular DiseaseA Prospective Cohort Study." Annals of internal medicine 159.10 (2013): 649-659.


Next, exposure to inorganic arsenic, even at low levels, is associated with a pronounced increase in the risk of stroke, as you see in this recent and important paper. I have never seen arsenic exposure considered as a potential confounder in any study of stroke risk.



Slide 45                 Oguri, Tomoko, et al. "Inorganic Arsenic in the Japanese Diet: Daily Intake and Source." Archives of environmental contamination and toxicology (2013): 1-13.


This is a serious oversight, especially in Japan. The Japanese are exposed to significant and very concerning levels of inorganic arsenic in their diet. This study pinpoints a seaweed called hijiki as well as cereals as the most concentrated sources of arsenic.



Slide 46                 Oguri, Tomoko, et al. "Inorganic Arsenic in the Japanese Diet: Daily Intake and Source." Archives of environmental contamination and toxicology (2013): 1-13.


Take note that the cereal in question here is rice. It is beyond the scope of this project to detail variability in arsenic in rice based on its source but suffice it to say that arsenic concentrations in rice vary widely and they are especially high in Southeast Asian rice. Once again, we can easily understand why this might be a confounding variable. If a Japanese individual were to consume a more western diet like ours, he would be eating less of traditional Japanese foods like seaweed and rice. Diets high in seaweed and rice would tend to be low-cholesterol diets. Therefore, the fact that lower cholesterol levels might be associated with higher hemorrhagic stroke risk in Japan may not imply causation because of the confounding effect of higher arsenic exposure in lower-fat diets there. This is a textbook example of what is called “the third variable problem.”



Slide 47                 Gruber, Joann F., et al. "Associations between toenail arsenic concentration and dietary factors in a New Hampshire population." Nutrition Journal 11.1 (2012): 1-10.


Moreover, nutrients in animal foods, including vitamin B-12 and saturated fat, may actually facilitate the removal of arsenic from the body. So yes, saturated fats may be protective in certain circumstances, even if they are detrimental to overall health.



Slide 48                 This study found evidence that this was in fact the case. Again, I have never seen the confounders of mercury and arsenic exposure considered in the literature about diet and stroke. I think this is an excellent example of why a vegan perspective is necessary in the discussion of diet and health. Meat eaters might take comfort in findings showing benefits to the harmful foods they eat and may lack the interest or motivation to explain those findings. It is not an exaggeration to say that the global medical research community is almost entirely composed of individuals biased in favor of animal foods. These are very important confounders and they were not considered by anyone until now.



Slide 49                 Sasaki, Satoshi, Xin-Hua Zhang, and Hugo Kesteloot. "Dietary sodium, potassium, saturated fat, alcohol, and stroke mortality." Stroke 26.5 (1995): 783-789.


This assumed benefit to fatty diets is dubious. Other evidence contradicts it, especially evidence from other countries beside Japan. A study examining subjects in 17 countries found that saturated fat consumption was strongly correlated with deaths from stroke.



Slide 50                 Shiue, Ivy, et al. "Dietary intake of key nutrients and subarachnoid hemorrhage: a population-based case-control study in Australasia." Cerebrovascular Diseases 31.5 (2011): 464-470.


This study of four cities in Australia and New Zealand found that as the consumption of fat and skin from meats increased, the incidence of subarachnoid hemorrhage increased.



Slide 51                 Collaboration, Asia Pacific Cohort Studies. "Cholesterol, coronary heart disease, and stroke in the Asia Pacific region." International journal of epidemiology 32.4 (2003): 563-572.


This study also extended well beyond Japan into other nations in the Asia Pacific region. Here, total cholesterol was found to strongly and positively associate with both heart disease and ischemic stroke. No association was found for hemorrhagic stroke. This is in line with most findings. It is possible that higher cholesterol only has a benefit with this one type of stroke. That is doubtful but it is not out of the question.



Slide 52                 Chen, Zhengming, et al. "Serum cholesterol concentration and coronary heart disease in population with low cholesterol concentrations." BMJ: British Medical Journal 303.6797 (1991): 276.


Looking elsewhere in Asia, in this study in Shanghai when cholesterol was relatively low, no association was found between low cholesterol concentrations and stroke mortality.



Slide 53                 p.2. Hreib, Kinan K. 100 Questions and Answers About Stroke: A Lahey Clinic Guide. Sudbury, Mass: Jones and Bartlett Publishers, 2009. Google Books.


In the West, hemorrhagic strokes only account for about 10% of all strokes. Those of Asian decent are disproportionately affected. It makes more sense to design your diet to address your most likely killers. If you are a Westerner, heart disease is at the very top of that list of threats. These findings from Japan shouldn’t cause you to lose sight of that. Notice as well that those of African decent are also more at risk. More on that in a moment.



Slide 54                 Amarenco, Pierre, et al. "Statins in stroke prevention and carotid atherosclerosis systematic review and up-to-date meta-analysis." Stroke 35.12 (2004): 2902-2909.


There is also important evidence based on the use of cholesterol-lowering drugs. Statins appear to decrease the incidence of all strokes, and this effect seems tied to their lowering of LDL cholesterol. You may see it argued that this is because statins have other effects beyond what they do for cholesterol, but if lower cholesterol really did cause a stroke risk, we wouldn’t see this relationship. This is where the evidence is now on statins.



Slide 55                 Lee, Jae-Geun, et al. "Characteristics of Subjects with Very Low Serum Low-Density Lipoprotein Cholesterol and the Risk for Intracerebral Hemorrhage." The Korean journal of internal medicine 27.3 (2012): 317-326.


We also see from patients with very low LDL levels that in the cases where low LDL is not an artifact of reverse causation, whether from liver disease, alcohol abuse, or kidney disease, there does not seem to be a relationship with hemorrhagic stroke according to this small study.



Slide 56                 Walker, Alexander RP, and Ulla B. Arvidsson. "Fat intake, serum cholesterol concentration, and atherosclerosis in the South African Bantu. Part I. Low fat intake and the age trend of serum cholesterol concentration in the South African Bantu." Journal of Clinical Investigation 33.10 (1954): 1358.


As usual, it is valuable to look at the early epidemiology of cardiovascular disease so we can see what happened among people with truly different diets rather than our present homogenized and industrial diets. The African Bantu were a population of interest in the early days of diet-heart because of their unusually low cholesterol levels and their low incidence of heart disease. You can see here that their cholesterol numbers were far below those in the US at the time.




Slide 57                 No doubt this was largely attributable to their starch-based, low-fat diets. Grains and beans provided the bulk of their calories.



Slide 58                 Berkson, D. M., and J. Stamler. "Epidemiological findings on cerebrovascular diseases and their implications." Journal of Atherosclerosis Research 5.2 (1965): 189-202.


Did they have high rates of stroke? Not at all. Look how much more of a problem stroke was in Japan at the time. This is more reason we should not generalize based on findings in Japan. This also shows you that at least in this population of African descent, stroke risk was not unusually high.




Slide 59                 Those Bantu did have a more common occurrence of high blood pressure and yet their stroke risk was not high. These authors suspected that because of their tendency toward hypertension, most of their strokes were probably hemorrhagic strokes.



Slide 60                 Lindahl, Olov, et al. "A vegan regimen with reduced medication in the treatment of hypertension." British Journal of Nutrition 52.01 (1984): 11-20.


This is a good opportunity to point out that those who eat only plant foods, commonly known as vegans, have an especially low incidence of high blood pressure. Indeed, a vegan diet has been used successfully to treat high blood pressure. I have found absolutely no evidence that vegans or vegetarians are at an increased risk of any kind of stroke.



Slide 61                 Checkoway, Harvey, et al. "Parkinson's disease risks associated with cigarette smoking, alcohol consumption, and caffeine intake." American Journal of Epidemiology 155.8 (2002): 732-738.


It also seems like the right moment to remind you that hardly anything is entirely good or bad. Even smoking has its benefits. If I told you that there is evidence that smoking is protective against Parkinson’s disease, would you ignore everything else we know about smoking and decide to take up this dirty habit? Just the same, if it turns out to be true that animal fats are protective against a specific type of stroke in some populations, you’d be wise to keep that in some sort of reasonable context.



Slide 62                 Sherzai, Ayesha, et al. "Stroke, food groups, and dietary patterns: a systematic review." Nutrition reviews 70.8 (2012): 423-435.

Focus on what we know and not what is up for debate. We know that plant foods are protective of our health, even when it comes to stroke. I suspect that if we are being honest, nearly every one of us who has not already been indoctrinated by low-carb books and blogs understands that fruits and vegetables help us to be our healthiest. Don’t let the noise from the blogosphere confuse you and cause you to risk your wellness.



Slide 63                 Siri-Tarino, Patty W., et al. "Saturated fatty acids and risk of coronary heart disease: modulation by replacement nutrients." Current atherosclerosis reports12.6 (2010): 384-390.


And don’t let yourself get too impressed with the work of people who receive money from the special interests promoting beef, dairy, and low-carb. All of these authors know enough that they could have spotted most of what I’ve shown you in these two videos. But they didn’t have much incentive to be so critical so they weren’t.


This playlist now shifts focus. Two low-carb promoters, Gary Taubes and Peter Attia, have acquired considerable funding in their efforts to generate biased studies that favor fatty, artery-clogging diets. Before we accept the research they produce, we need to consider the source. We’ll get to know the NuSI guys starting with the next video.