Hi. I’m Dr. Chris Masterjohn of chrismasterjohnphd.com, and you are watching Masterclass with Masterjohn. And today we are in our thirteenth in a series of lessons on the antioxidant defense system. We’ve completed our discussion of how that system works, and we’re in the process of applying it to the problem of heart disease. And from the beginning of this series we’ve talked about how the antioxidant defense system is all converging on, at least in one of its core operating principles, trying to prevent lipid peroxidation and it’s polyunsaturated fatty acids or PUFAs that are uniquely vulnerable to lipid peroxidation. And in the last lesson we talked about how oxidative stress, especially peroxidation of the LDL membrane, is what drives atherosclerosis. And this seems to introduce a conflict because for decades we’ve been told that saturated fats are the bad fats and replacing them with polyunsaturated fat will prevent heart disease because doing so would lower cholesterol levels. And it’s true that replacing saturated fat with PUFAs does lower cholesterol levels. This has been known and clearly established since the 1950’s. So what we’re going to look at today is what are the two sides of this argument, and why should it be so controversial? Shown on the screen is a meta-analysis of observational studies comparing higher and lower PUFA intakes, and how that affects the risk of heart disease. A meta-analysis pools together the results of many different studies to get a look at the big picture and to achieve more statistical power by increasing the total sample size. In the figure we have individual studies given an abbreviation as their name. And then we have this visual display of their statistical power and the risk that they suggest. One is where we suggest there’s no benefit and no harm. If we move to the right of one, that suggests that PUFAs are bad for heart disease. If we move to the left of one, it suggests that there’s a benefit. The numerical significance is that if we get two that means that PUFAs double the risk. If we get to .5, that means they cut the risk in half. For each of these studies they have different sizes. And so they contribute to the pooled overall result, which is shown in the bottom with the diamond, to different degrees. The greater their weight, which is driven by a greater size, the greater the shaded box around this point for each study is. The other thing that’s displayed is the 95% confidence interval, which is the line that goes through the dot. So we call this point the point estimate, that’s saying if we estimate a single point where we think the risk is, it’s right here, but the 95 % confidence interval is saying that were 95% confident that the true risk is somewhere between here and way over here. You can see this particular study, because it was small, It had very little weight. You don’t even see the shaded box around it. And it had a huge confidence interval indicating that we have a very imprecise idea about what the true result is from that study. So some of these studies that have really big shaded boxes they carry a lot of weight, they have very good statistical precision, because that line is so small that means we’re 95% confident that the true risk is somewhere very close to the point estimate. The overall results are shown in this diamond, and they suggest a 15% decrease in risk of heart disease from the highest PUFA intakes versus the lowest. That’s actually a pretty small decrease in risk considering that we’d have to go from the lowest PUFA intake to the highest to obtain it. But my confidence that this represents a true cause and effect relationship is really, really low. And the reason is that not a single one of these studies was done before 1990. The American Heart Association began a public health campaign to tell people that they should protect their hearts by replacing saturated fat with PUFAs in 1961. Every single one of these studies had three decades of a consistent public health message telling people that if you care about your health you’ll eat more PUFA and less saturated fat. So it seems impossible to tease out the effect of PUFAs from the fact that the people eating the PUFAs are the people who care more about their health by and large, and who buy into the mainstream paradigm about how to be healthy and who act on what they believe. So let’s take a look at the randomized controlled trials which give us higher confidence in cause-and-effect relationships. Shown here is a meta-analysis of RCTs suggesting a very similar benefit in the overall results. And the diagram is patterned in the same way. So all the same rules about the 95 percent confidence interval, and the shaded area indicating the weight, and the diamond indicating the overall result, all of it’s the same. Now although pooling these studies together suggests a similar benefit of PUFAs, we should look at some of these studies one by one to get a truer sense of what they’re really telling us. And one way that we can decide what to focus on is looking at this figure which of these trials are making a big contribution to the net result, and which aren’t really that important. For example, if you look at the STARS trial, this out of every other trial showed the biggest benefit, but it had the widest 95% confidence interval. And because it was small it had the least weight. By contrast, the DART trial had a fairly considerable weight and had a much narrower confidence interval. But it didn’t show hardly any benefit as indicated by being so close to one. The Minnesota Coronary Survey also made a significant contribution. It’s actually on the other side of one. So what are the trials that had considerable weight, and also showed considerable benefit? Well the biggest out of all of them is the Finnish Mental Hospitals Study. If you add, they separated it in this figure between men and women, that’s one study. So if you put them together, that’s 16% of of the weight. And that’s almost as significant as any of the other weights. You can see one is 20%, one’s 18%. So it’s holding its own in terms of contributing to the net result. And out of all the trials that make some significant contribution to the net results, it’s the one that shows the most benefit. So let’s take a look at that one first. The Finnish Mental Hospital Study wasn’t a randomized controlled trial. It shouldn’t even be in this meta-analys. The Finnish Mental Hospital Study followed a model where there were two hospitals. And for a period of about five years, they had Hospital A eat a lot of PUFA and Hospital B eat a lot of saturated fat. And then after that they switched them for another five years so that the hospital that was getting PUFAs is now getting saturated fat and vice versa. Well, that’s not a randomized controlled trial, because the unit of randomization was the hospital. But if that’s the case, then each group had one hospital in it. How do you do statistics with an N of one? You can’t. The people in the hospitals weren’t randomized to anything. So you can’t analyze the risk of the individuals and call it a randomized controlled trial if what you randomized was one hospital to each group of hospitals. Sorry. And this could be really significant. Because, for example, all of the excess risk of heart disease, as Christopher Ramsden’s group has pointed out, was found at one time point in one of the hospitals where at that time they were administering a cardiotoxic anti-psychotic medication that we now know to cause heart disease. So how do we know it wasn’t all due to that medication? Had they randomized the people in the hospital to the two different diets, then if the randomization was successful the distribution of that cardiotoxic medication would have been the same between the different groups. Since this is in no way, shape, or form, a randomized controlled trial and has no place in any meta-analysis of randomized controlled trials, that wasn’t the case and we really can’t get a good handle on the cause-and-effect interpretation. Three of the studies that are in the meta-analysis fall into a specific problem of being multifactorial trials. These are the Oslo Diet-Heart Study, the DART Study and the STARS Study. In these three trials, the STARS made a very small contribution to the net result and had a big effect. The DART trial had a significant contribution to the result, but had hardly any effect. And the real contributor is the Oslo Diet-Heart Study because it made a meaningful contribution to the net results. And it also had a pretty significant benefit shown. If we look at these trials, keeping in mind that the Oslo Diet-Heart, is the biggest one contributing to the net results. The STARS trial involved increased fruits and vegetables, nutritional counseling, all on top of increased PUFA intake. The DART trial replaced total fat with carbohydrate in addition to replacing saturated fat with PUFA. And actually the PUFA intake wasn’t really different between the two groups. So that’s not really giving us much of a sense of what PUFA is doing. The Oslo Diet-Heart trial, is the one that was most successful. And it’s the one that made the biggest contribution to the net results out of these three. It’s intervention included restricting margarine, increasing fruits and vegetables, increasing fish, distributing free sardines canned in cod liver oil probably containing glycine rich bones and probably rich in vitamins A and D. The control group also had more over weight, and older subjects and by the end of the trial had twice as many heavy smokers. But those things they didn’t try to do. The intervention that they did purposefully contained all these things that we would expect, for example, to have vitamin E and vitamin C increasing. To have polyphenols that increase the synthesis of the entire antioxidant defense system, glutathione and all the related enzymes. Glycine that would contribute to glutathione status. Vitamin C and copper from the vegetables and glycine from the bones, they contribute to collagen synthesis to secure the fibrous cap. Although we haven’t covered this in this series on the antioxidant defense system, the vitamins A and D are likely to prevent calcification of the fibrous cap, and prevent it from breaking. So how do we know if in the Oslo Diet-Heart Study, the benefit was a result of PUFAs or was a result of all these other aspects of the intervention. What if the ideal diet to prevent heart disease restricts margarine, increases fruits and vegetables, increases fish, contains sardines canned in cod liver oil, but has lower total PUFA, and actually has some butter? If you look at this meta-analysis it also excludes two important trials shown over here. The Rose, 1965, and Sydney Diet-Heart. In Rose 1965 this was the first such trial. And they had one group eat what they normally ate, they had another group that cut out a lot of fat from their diet and added corn oil, and they had a third group that did the same but with olive oil. All of them were people who had established heart disease. And what they found after two years was that corn oil doubled the risk of heart disease, olive oil was intermediate, and their statistics achieved borderline statistical significance. And they concluded, under the circumstances of this trial, corn oil cannot be recommended as a treatment of ischemic heart disease. It is most unlikely to be beneficial, and it’s possibly harmful. The Sydney Diet Heart Study was a larger study also again with men who had heart disease and this time it lasted 2-7 years instead of just two years. You can see on the screen that the survival is plotted over time the dotted line is the PUFA group, and the solid line is the saturated fat group. And you can see that throughout the entire study the PUFA group had lower survival. On the whole, replacing saturated fat with PUFAs increased mortality by 39%. Oddly enough they concluded that “Men who have had myocardial infarction are not a good choice for testing the lipid hypothesis.” In this meta-analysis there are two double-blind studies: the Minnesota Coronary survey and the LA Veterans Administration Hospital study. Double-blind, randomized controlled trials, are the gold standard. Of these, the Minnesota Coronary Survey makes no contribution to the overall net estimate of a protective effect, because it actually found a non statistically significant harm of PUFAs. So let’s zero in on the LA Veterans Administration Hospital Study. In the LA Veterans Administration Hospital Study, they took just under 850 patients and randomized them to one of two dining halls. These people were long-term patients. They were allowed to leave the dining halls, but they ate most of their meals there. And they kept track of how many meals they ate using meal tickets that had to be punched. Dining hall A had traditional animal fats, and it included some hydrogenated oils. Dining hall B, restricted eggs to one egg per day and otherwise they ate the exact same foods, but they replaced all the traditional fats with corn, soy, safflower and cotton seed oils, which are high in PUFAs. The double blinding was maintained with color-coded meal tickets and the trial is terminated after eight years. Among the unique properties of the LA VA study, it was one of the only two double-blind randomized controlled trials. The only trial where the mean age was over 60, and because the subjects were older, that allows us to have greater insight into the effects of these oils on cancer. And it was the longest trial lasting over eight years with most of the subjects following the diets for about six years. If we simply analyze the data on the surface, what we find is that the vegetable oil group had a lower risk of atherosclerotic death shown on the left. But a higher risk of non cardiovascular death shown on the right. Altogether there was no difference in total mortality. You can see the experimental group is the PUFA group, the control group is the traditional fats group. Their survival is plotted over time. You can see that these are elderly people over eight years, so the majority of them are dying, but there’s no difference between either group. However, when they looked at cancer, the experimental group started to diverge with a higher cancer risk than the control group, meaning PUFAs were resulting in higher risk after two years. And you can see that it’s really after five years that that gap seems to widen. And this is disconcerting for a couple reasons. First of all it suggests that PUFAs cause cancer. But second of all, it suggests that we can’t see that result at all until at least two years. And we don’t see the true effects until after five years. That’s disconcerting because almost every other trial was five years or shorter. Shown on the screen is total mortality among people who didn’t die of heart disease. This allows us to look at survival over time where all of the deaths are due to an non cardiovascular cause. This is even more disconcerting because despite the rise in cancer at two years, the two groups are running head-to-head. Despite the fact that cancer really takes off at five years, you just barely see a gap beginning to emerge, that’s hardly perceptible between these two groups. And then at seven years is where you see this big divergence. That suggests that even though cancer risk is increasing in these early years it’s not becoming a predominant cause of death until later and its importance doesn’t take over until seven years. Additionally, as people age, heart disease becomes less and less important of a risk and cancer increases in importance. So when we look at the difference in total mortality, we have to wonder that had these subjects been even older, and cancer thus even more important relative to heart disease, or had the trial gone on for longer, would we have seen a difference in total mortality. One thing to note, is that after eight years we see these lines emerge in a way that would favor greater survival in the control group, but you really can’t tell that that’s anything more than noise here. But just given the other data in the study, it makes us wonder: What would have happened if this was nine and ten years long. Nevertheless, let’s look at this beyond the surface. Even though this was a randomized controlled trial, it wasn’t big enough to fully equalize the potential confounding factors between the two groups. There were 60% more smokers of one to two packs a day in the control group versus the vegetable oil group, And there were twice as many heavy smokers defined as two or more packs per day in the control group versus the vegetable oil group. Furthermore don’t know exactly what they were eating in the control group but it sure enough wasn’t pastured butter. Shown on the screen is the vitamin E content normalized to the PUFA content of the control diet and of the vegetable oil diet. You can see first of all that the vegetable oil diet is more or less adequate in the amount of vitamin E that is needed to protect the PUFA, but the control diet is not. By any estimate of how much vitamin E you would need to protect that PUFA, it doesn’t have enough. And it’s almost three times lower than in the vegetable oil diet. Furthermore, let’s say they included commercial butter as would be sold in the store. That would actually have vitamin E that would be similar to the vegetable oil diet. But if it was grass-fed butter, from the rapidly growing grass in the spring, that would have the most vitamin E out of all. After the study was originally published, they published an analysis separating cardiovascular death by cigarette smoking. You see the control group on the left and the vegetable oil group on the right; in each panel you go from fewer than 10 cigarettes per day to 10 to 20 cigarettes per day, to over 20 cigarettes per day. And what you can see is that there’s not really any more heart disease, and we don’t even have non-smokers plotted here and they didn’t report that data, but just in the smokers of fewer than 10 cigarettes there’s no excess of heart disease over the vegetable oil group. All of the excess atherosclerotic death occurred in people in the control group smoking more than 10 cigarettes per day. Well if we go back to the vitamin E lesson from this series, what does cigarette smoking do to vitamin E? It increases your turnover and makes vitamin E disappear from the body at a higher rate. So the control group had more smokers and had three times less vitamin E with that smoking increasing the need for vitamin E. And all of the excess death occurs in the people who are smoking 10, 20 cigarettes or more. A simple explanation for this is that there was more oxidative stress and lipid peroxidation and LDL oxidation in these people because they were consuming vitamin E-deficient diets and throwing cigarettes into the mix. I just wish that we had the same kind of analysis for cancer. Because we should expect cancer to be greater in the people who smoke more cigarettes. And yet cancer was greater in the vegetable oil group. One of the key takeaways from the LA VA study is the need for a long study. It’s possible that even in this context with cigarette smoking and vitamin E deficiency contributing to heart disease in the control group, that if the study had gone on long enough the increase in cancer risk in the PUFA group would have become large enough to move the needle, not only in non cardiovascular mortality, but in total mortality over time. But don’t take my words for it, listen to these words from the authors of the LA VA study themselves: “This small excess of nonatherosclerotic mortality in the late years of the study raises the very important and difficult question of whether future clinical trials of diets rich in unsaturated fat must be planned for periods well in excess of eight years rather than for the five-year period that have been the usual goal. Trials that long were never done. If we look at these trials, they don’t clearly support really any conclusion, but on the balance they don’t support replacing saturated fats with PUFAs to prevent heart disease. They suggest that maybe we want to do the opposite and we certainly don’t want to be consuming a lot of vegetable oil if we’re concerned about cancer. This demonstrates one of the problems with meta-analyses. If you have a lot of good data, a meta-analysis can be super helpful in looking at the big picture. And if you have a lot of data, you can also use a meta-analysis to compare the results of higher quality trials to lower quality trials. Or the results of omega-3 versus omega 6; or the results of males versus females. Or the results of anything that you want to pick if you have enough studies that fit those characteristics to try to answer the question. But if you just have a small heterogeneous group of trials with totally different approaches sometimes just pooling the results together into an average, can actually decrease the clarity that you need to analyze those studies with, and can mislead you to the wrong conclusions. So maybe what we want to do is find ways to maximize LDL receptor activity without the risk of increasing LDL membrane PUFA content. So we can focus on thyroid hormone and on insulin status, and on inflammation as areas in which to optimize instead of relying on vegetable oils. At least that’s my take on it. But now I’ve given you a wealth of data of what the antioxidant defense system looks like. And a wealth of information about the mechanisms underlying heart disease, so you can come to your own opinion. I hope you enjoyed this. Signing off, this is Chris Masterjohn of chrismasterjohnphd.com, You have been watching Masterclass with Masterjohn and I will see you in the next lesson.