The PUFA Investigation: An Expert Interview

Linda Brookes, MSc; Christopher E. Ramsden, MD


March 18, 2013

In This Article

The Interview

Medscape: Why did you decide to revisit the SDHS data? What additional insight did you think they might provide into the omega-6 PUFA debate?

Dr. Ramsden: Over the past several decades, there has been much more recognition and appreciation that individual PUFAs -- not only the omega-6 and the omega-3 families, but also the individual nutrients within those families -- are likely to have different biochemical effects and possibly different clinical effects as well. Once we identified that the SDHS investigators provided safflower oil -- which is a concentrated source of one of those polyunsaturates, the omega-6 PUFA linoleic acid -- we realized that it provided a unique opportunity to evaluate the cardiovascular effects of linoleic acid itself rather that a combination of different types of PUFAs.

We were very fortunate to collaborate with one of the original SDHS investigators, Dr. Boonseng Leelarthaepin, now retired from the University of New South Wales in Sydney, Australia, who is a coauthor on our study. He was able to validate the extracted data and provide us with other key information, such as ethical considerations. We have had evidence from only 7 previous data sets from randomized controlled trials of PUFA interventions or interventions that substituted PUFAs in place of SFAs, so this trial provided a unique opportunity to evaluate omega-6 linoleic acid itself.

Medscape: Are there other trials that used only omega-6 PUFAs rather than a mixture with omega-3 PUFA?

Dr. Ramsden: There are 3 trials altogether,[2,11,12] and the SDHS had the longest follow-up (median, 39 months). The original research team estimated nutrient intakes from 7-day food records throughout the trial, so the study had some nice advantages compared with some of the others.

Medscape: What did you know about the PUFA content of the "Miracle" brand margarine that was used in the SDHS? One of the online responders to your study suggested that it might contain considerably less PUFAs than the oil, as well as up to 40% trans fatty acids, which could have been the cause of the increased CHD risk in your study.[13]

Dr. Ramsden: Of course, the SDHS is a 40-year-old trial, and the association of trans fatty acids with increased cardiovascular risk was not known at that time, so they were not recorded. In our article, we acknowledged the possibility that trans fat intake may have been altered in both groups. The primary intervention, however, was the liquid safflower oil, which is an omega-6 specific polyunsaturated fat source that actually replaced a substantial amount of trans fat. Compared with other common table margarines at the time, the polyunsaturated margarine provided in the SDHS was likely to contain a lower amount of trans fats, and it did replace butter and some of the other margarines. Many patients in the control group also began to replace butter with commercially available polyunsaturated margarines.

I think if you take the whole study design into consideration, as we did in our online response,[14] it is unlikely that trans fats were altered much in either group. But the margarine certainly provided some trans fats.

Medscape: What did you know about what other interventions were being prescribed for these secondary prevention patients? Obviously at the time the SDHS was run, less was known about the causes of cardiovascular disease, and there was little that could be prescribed in secondary prevention, for example after MI, beyond painkillers and bed rest. Many medications, such as beta-blockers, angiotensin-converting enzyme (ACE) inhibitors, and statins, were not yet available. Also, the people in the study were relatively young, their average blood pressure was 136/88 mm Hg, and their body mass index (25 kg/m2) wasn't high. Do you know how they were being treated apart from the study intervention?

Dr. Ramsden: Unfortunately, we were unable to recover data related to what medications they were on, but there were many fewer options. Of course, in a randomized clinical trial, you would expect -- at least ideally -- that both groups would be similar, including in whatever pharmaceuticals they were taking. However, because the medication data were not recovered, we cannot be sure.

Medscape: Were there other limitations of the study?

Dr. Ramsden: The really important limitations were (1) it was a patient population with established heart disease, so they all had a coronary event -- MI in most cases (86%) -- before randomization, and there is always a question of generalizability in randomized controlled trials; and (2) the median intake of omega-6 linoleic acid was almost 15% of calories. In the United States, consumption of omega-6 linoleic acid has increased from about 2% of calories in the early 20th century to about 7% of calories currently.[15] So in the SDHS it ended up approximately twice as much as the average American consumes nowadays.

A third limitation was that only relatively young men aged 30-59 years were studied; there were no women or older persons in the study.

Medscape: What were the effects of smoking and alcohol? Almost 70% of SDHS participants were smokers at baseline, probably more than there would be today, and about 36% were moderate to heavy drinkers.

Dr. Ramsden: Everyone received the standard medical care at the time, which for smokers included advice to quit or at least to cut back and for heavy drinkers to cut back. At baseline, the 2 groups were similar in both, but it seemed that those who were smokers or moderate to heavy drinkers may have been at even higher risk for mortality with the high linoleic acid intervention.

We previously proposed that the combination of high omega-6 linoleic acid diets and an endogenous source of oxidative stress, such as smoking or heavy drinking, facilitated OXLAM-mediated atherosclerotic progression and increased cardiovascular disease mortality. The magnitude of the link between the increase in linoleic acid and mortality was robust in smokers and drinkers, supporting our proposed model.

Medscape: Although increasing dietary omega-6 linoleic acid was associated with an increase in cardiovascular mortality, it was also associated with a highly significant reduction in total cholesterol, consistent with the aim of the original trial. This was presumably mainly due to decreased low-density lipoprotein (LDL) cholesterol, because several studies have demonstrated that omega-6 fatty acids lower LDL cholesterol levels. Because this should presumably reduce cardiovascular risk and improve survival, how do you explain the apparent contradiction?

Dr. Ramsden: It certainly was an unexpected finding for the original team of investigators. One way to interpret this would be that some things that lower LDL cholesterol reduce cardiovascular risk, some don't seem to have much of an effect, and others possibly increase risk. I am not a lipidologist, but it seems that these data might be consistent with the idea that linoleic acid falls into one of these latter 2 categories. We already had that knowledge about the cholesterol reduction and increased mortality in the SDHS, and because the study was published it in 1978,[2] we included a proposed mechanistic model for the link between high dietary linoleic acid and increased mortality with our report.

Medscape: Could you explain more about this? You implicate OXLAMs, the most abundant oxidized fatty acids in oxidized LDL, as leading to the increased risk for cardiovascular death in the linoleic acid intervention group. Many investigators have proposed that higher consumption of omega-6 PUFAs results in the increased production of eicosanoids/prostaglandins with high proinflammatory, oxidative, and carcinogenic potential.

Dr. Ramsden: This is where the concept of emphasizing only 1 mechanism can become problematic, and why it is important to have randomized clinical trials with clinical outcomes rather than just biochemical or intermediate endpoints. OXLAMs are very bioactive in multiple organ systems and cell types, and so the physiology becomes very complicated. OXLAMs have also been associated with nonalcoholic steatohepatitis and are mechanistically linked to physical pain, so the potential implications of high intakes of vegetable oils enriched in linoleic acid, particularly processed or heated oils, may extend beyond cardiovascular disease.

Medscape: What seems to be in dispute at present is whether high intake of omega-6 PUFAs can overwhelm the beneficial/anti-inflammatory actions of linolenic acid (omega-3) fatty acids. Does the ratio of dietary omega-6 to omega-3 PUFAs matter at all? The 2009 AHA advisory said that "The focus on ratios, rather than on levels of intake of each type of PUFA, has many conceptual and biological limitations."[1]

Dr. Ramsden: I think there has been much oversimplification of the omega-6 to omega-3 PUFA ratios and inflammation; it is very complex. There are many different compounds, and they do not have the same biochemical effects. For instance, linoleic acid, by far the most abundant dietary omega-6 PUFA, is very different from arachidonic acid, the second most abundant.

Medscape: Do we ingest much arachidonic acid?

Dr. Ramsden: Only very small amounts -- 100 mg or so a day, from animal food sources -- versus, for most people, 15 g or so of linoleic acid. And only a very small amount of linoleic acid is converted into arachidonic acid. However, some recommendations treat them as though they are the same in terms of being proinflammatory, whereas their biochemistry is much more complicated.

The omega-6 PUFAs are very bioactive compounds. They are precursors for several different families of bioactive mediators, everything from endocannabinoids to eicosanoids. We are very interested in these compounds. Some appear, at least in animal trials and with some indications in humans, to regulate all sorts of states, from hunger and addictive behaviors to chronic pain and other conditions. Certainly the OXLAMs, especially arachidonic acid-derived mediators, have been implicated in various disease processes. It is likely that dietary PUFAs play a role not only in heart disease but also a variety of other conditions, including chronic pain, addiction, and psychiatric illness.

I don't think we yet know enough. But we need to look at the individual fatty acids, not just the families. There is a lot that needs to be teased out mechanistically.

Medscape: How will this be done? It would seem difficult to demonstrate an effect in clinical trials after controlling for a lot of confounding factors.

Dr. Ramsden: It is possible. In this particular study, the increased risk for death was 60%-70%, depending on the outcome. Given those effects, it might be possible to do a randomized controlled trial comparing the clinical effects of omega-6 linoleic acid to the major dietary monounsaturated fat, oleic acid. By providing study vegetable oils that are either rich in linoleic acid or oleic acid, and keeping other aspects of the intervention equivalent in both groups, the effects of linoleic acid vs oleic acid can be determined. In order to have enough coronary events, it probably would have to be a secondary prevention trial. It is going to be hard to get it done, and it would cost a lot of money.

Medscape: In the meantime, how do you think clinical practitioners and other health professionals should approach this topic? I have read some of your statements about the 2009 AHA Scientific Advisory on dietary omega-6 PUFA intake, and I gather that you do not support its recommendations.

Dr. Ramsden: I think you can see that in this article, that we took a very cautious tone. We did not single out any health promotion agencies, and we did not give any diet advice. We just said that basically, our findings might have important implications.

For a little perspective, if you eat a diet from foods that are raised naturally -- it doesn't matter if it is predominantly vegetable-based or meat-based -- as everybody would have done up until about 100 years ago, you are going to consume something like 2%-3% of calories from linoleic acid. So with that perspective, the burden of proof would seem to be on the group advocating changes from what you would get in a natural diet, rather than vice versa.

Medscape: So is the best advice a physician might give would be to eat as natural diet as possible, because it would be very difficult to determine your omega-6 PUFA intake from reading food packets?

Dr. Ramsden: Most of the vegetable oils that are currently added to packaged and processed foods, at least in the United States, are high in linoleic acid, and they vary considerably among different oils. So the observational trials that asked people every few years to estimate or self-report how often they ate certain foods have a very hard time separating out their intake of linoleic acid from other PUFAs, because the oils are labeled such things as "vegetable oil," or "safflower and/or sunflower and/or canola oil" on our products. It would be extremely challenging to evaluate linoleic acid in a setting outside of a randomized controlled clinical trial.

The data from the SDHS aren't going to tell you how much you should consume, because this study was designed to lower cholesterol by increasing omega-6 PUFAs markedly. It is just by chance that the investigators happened to select a concentrated source of linoleic acid that didn't contain any other PUFAs. But I think it probably has important implications, and it gives you a red flag showing the potential ill effects of excess consumption of linoleic acid.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.