A Practical Approach to Omega Fats

Hunter-gatherers and healthy non-industrial cultures didn't know what omega-6 and omega-3 fats were. They didn't balance nutrients precisely; they stayed healthy by eating foods that they knew were available and nourishing. Therefore, I don't think it's necessary to bean count omega fats, and I don't think there's likely to be a single ideal ratio of n-6 to n-3. However, I do think there's evidence for an optimal range. To find out what it is, let's look at what's been done by healthy cultures in the past:

• Hunter-gatherers living mostly on land animals: 2:1 to 4:1


• Pacific islanders getting most of their fat from coconut and fish: 1:2


• Inuit and other Pacific coast Americans: 1:4 or less


• Dairy-based cultures: 1:1 to 2:1


• Cultures eating fish and grains: 1:2 or less

It looks like a healthy ratio is between 4:1 and 1:4 n-6 to n-3. Some of these cultures ate a good amount of n-3 polyunsaturated fat, but none of them ate much n-6 [One rare exception is the !Kung. SJG 2011]. There are three basic patterns that I've seen: 1) low fat with low total n-6 and n-3, and a ratio of less than 2:1; 2) high fat with low total n-6 and n-3 and a ratio of 2:1 or less; 3) high fat with low n-6 and high n-3, and a low carbohydrate intake.

I think there's a simple way to interpret all this. Number one, don't eat vegetable oils high in n-6 fats. They are mostly industrial creations that have never supported human health. Number two, find a source of n-3 fats that can approximately balance your n-6 intake. In practical terms, this means minimizing sources of n-6 and eating modest amounts of n-3 to balance it. Some foods are naturally balanced, such as grass-fed dairy and pastured lamb. Others, like coconut oil, have so little n-6 it doesn't take much n-3 to create a proper balance.

Animal sources of n-3 are the best because they provide pre-formed long-chain fats like DHA, which some people have difficulty producing themselves. Flax oil may have some benefits as well. Fish oil and cod liver oil can be a convenient source of n-3; take them in doses of one teaspoon or less. As usual, whole foods are probably better than isolated oils. Weston Price noted that cultures throughout the world went to great lengths to obtain fresh and dried marine foods. Choose shellfish and wild fish that are low on the food chain so they aren't excessively polluted.

I don't think adding gobs of fish oil on top of the standard American diet to correct a poor n-6:n-3 ratio is optimal. It may be better than no fish oil, but it's probably not the best approach. I just read a study, hot off the presses, that examines this very issue in young pigs. Pigs are similar to humans in many ways, including aspects of their fat metabolism. They were fed three diets: a "deficient" diet containing some n-6 but very little n-3; a "contemporary" diet containing a lot of n-6 and some n-3; an "evolutionary" diet containing a modest, balanced amount of n-6 and n-3; and a "supplemented" diet, which is the contemporary diet plus DHA and arachidonic acid (AA).

Using the evolutionary diet as a benchmark, none of the other diets were able to achieve the same fatty acid profile in the young pigs' brains, blood, liver or heart. They also showed that neurons in culture require DHA for proper development, and excess n-6 interferes with the process.

With that said, here are a few graphs of the proportion of n-6 in common foods. These numbers all come from nutrition data. They reflect the percentage n-6 out of the total fat content. First, animal fats:


Except salmon oil, these are traditional fats suitable for cooking. Except schmaltz (chicken fat), they are relatively low in n-6. Next, vegetable oils:


These range from very low in n-6 to very high. Most of the modern, industrially processed oils are on the right, while most traditional oils are on the left. I don't recommend using anything to the right of olive oil on a regular basis. "HO" sunflower oil is high-oleic, which means it has been bred for a high monounsaturated fat content at the expense of n-6. Here are the meats and eggs:

n-3 eggs are from hens fed flax or seaweed, while the other bar refers to conventional eggs.

A few of these foods are good sources of n-3. At the top of the list is fish oil, followed by n-3 eggs, grass-fed butter, and the fat of grass-fed ruminants. It is possible to keep a good balance without seafood, it just requires keeping n-6 fats to an absolute minimum. It's also possible to overdo n-3 fats. The traditional Inuit, despite their good overall health, did not clot well. They commonly developed nosebleeds that would last for three days, for example. This is thought to be due to the effect of n-3 on blood clotting. But keep in mind that their n-3 intake was so high it would be difficult to achieve today without drinking wine glasses full of fish oil.

Omega Fats and Cardiovascular Disease

I noticed something strange when I was poring over data about the Inuit last month. Modern Inuit who have adopted Western food habits get fat, they get diabetes... but they don't get heart attacks. This was a paradox to me at the time, because heart disease mortality typically comes along with the cluster of modern, non-communicable diseases I call the "diseases of civilization".

One of the interesting things about the modern Inuit diet is it's most often a combination of Western and traditional foods. For example, they typically use white flour and sugar, but continue to eat seal oil and fish. Both seal oil and fish are a concentrated source of long-chain omega-3 (n-3) fatty acids.The 'paradox' makes much more sense to me now that I've seen this:

It's from the same paper as the graphs in the last post. Note that it doesn't take much n-3 to get you to the asymptote. Here's another one that might interest you:

The finding in this graph is supported by the Lyon diet heart study, which I'll describe below. One more graph from a presentation by Dr. Lands, since I began by talking about the Inuit:


Cardiovascular disease mortality tracks well with the n-6 content of blood plasma, both across populations and within them. You can see modern Quebec Inuit have the same low rate of CVD mortality as the Japanese. The five red triangles are from MRFIT, a large American intervention trial. They represent the study participants divided into five groups based on their plasma n-6. Note that the average percentage of n-6 fatty acids is very high, even though the trial occurred in the 1970s! Since n-3 and n-6 fats compete for space in human tissue, it makes sense that the Inuit are protected from CVD by their high n-3 intake.  [Update: I don't read too much into this graph because there are so may confounding variables.  It's an interesting observation, but take it with a grain of salt.. SJG 2011].

Now for a little mechanism. Dr. Lands' hypothesis is that a high n-6 intake promotes a general state of inflammation in the body. The term 'inflammation' refers to the chronic activation of the innate immune system. The reason is that n-3 and n-6 fats are precursors to longer-chain signaling molecules called eicosanoids. In a nutshell, eicosanoids produced from n-6 fatty acids are more inflammatory and promote thrombosis (clotting) more than those produced from n-3 fatty acids. Dr. Lands is in a position to know this, since he was one of the main researchers involved in discovering these mechanisms. He points out that taking aspirin to 'thin' the blood and reduce inflammation (by inhibiting inflammatory eicosanoids) basically puts a band-aid over the problem caused by excess n-6 fats to begin with.  [Update- this mechanism turns out not to be so straightforward. SJG 2011]

The Lyon Diet Heart Study assessed the effect of n-3 fat supplementation on CVD risk. The four-year intervention involved a number of diet changes designed to mimic the American Heart Association's concept of a "Mediterranean diet". The participants were counseled to eat a special margarine that was high in n-3 from alpha-linolenic acid. Overall PUFA intake decreased, mostly due to n-6 reduction, and n-3 intake increased relative to controls. The intervention caused a 70% reduction in cardiac mortality and a large reduction in all-cause mortality, a smashing success by any measure.

In a large five-year intervention trial in Japan, JELIS, patients who took EPA (a long-chain n-3 fatty acid) plus statins had 19% fewer cardiac events than patients taking statins alone. I don't know why you would give EPA by itself when it occurs with DHA and alpha-linolenic acid in nature, but it did nevertheless have a significant effect. Keep in mind that this trial was in Japan, where they already have a much better n-6/n-3 ratio than in Western nations.

In my opinion, what all the data (including a lot that I haven't included) point to is that a good n-6 to n-3 ratio may be important for vibrant health and proper development. In the next post, I'll talk about practical considerations for achieving a good ratio.

Omega-3 Fats and Brain Development

Another interesting study that Dr. Hibbeln sent me is about the link between maternal seafood consumption and neurodevelopmental outcomes in children. The study is about as powerful as epidemiology gets, with an enrollment of 11,875 mothers.

The bottom line is short and sweet: compared to the children of mothers who ate 340 grams or more of fish per week, children whose mothers ate very little fish had an increased risk of low verbal intelligence, poor social behavior, poor motor skills, poor communication skills, and poor social development. These associations remained after adjusting for 28 potential confounders, including social status, level of education, stressful life events, smoking, alcohol, and several others.

In support of this association, in another study the four-year-old children of mothers who were given DHA and arachidonic acid supplements had higher IQs than those given "placebo" (corn oil). There have been a number of trials of varying quality that have shown varying results with n-3 supplementation, so I'll leave you to decide what you think of this. A 2007 review I found on n-3 supplementation and brain development states that "the evidence for potential benefits of LCPUFA [long-chain polyunsaturated fatty acid] supplementation is promising but yet inconclusive".

I do think it's interesting to note that the brain has the highest concentration of long-chain n-3 fats of any organ, and eating n-3 fats in the form of fish, fish oil or cod liver oil increases the amount in tissues. Eating too much n-6 depletes the brain of DHA and adversely affects neuron development in piglets. n-3 deficiency affects the release of serotonin (a neurotransmitter) in rat brains.

Put it all together, including the data from the last two posts, and I think there's some evidence that a good balance of n-3 to n-6 fatty acids is important for optimal brain function and perhaps development.

The Omega Ratio

The theory advanced by Dr. Lands and Hibbeln is that the ratio of omega-6 to omega-3 fatty acids in the diet is the crucial factor for mental and physical health, rather than the absolute amount of each. Omega-6 and -3 fats are essential long-chain polyunsaturated fatty acids. The number refers to the position of the double bond near the methyl end of the carbon chain.

The ratio of omega-6 to omega-3 (hereafter, n-6 and n-3) in the diet determines the ratio in tissues. Since these molecules have many cellular roles, it doesn't stretch the imagination that they could have an effect on health. Hunter-gatherer and healthy non-industrial cultures typically have a favorable ratio of n-6 to n-3, 2:1 or less. In the US, the ratio is currently about 17:1 due to modern vegetable oils.

DHA, a long-chain n-3 fatty acid, is concentrated in neuronal synapses (the connections between neurons) and is required for the normal functioning of neurons. n-6 fats compete with n-3 fats for space in cellular membranes (which have a fixed amount of total polyunsaturated fat), so a high intake of n-6 fats, particularly linoleic acid, displaces n-3 fatty acids. Lower tissue levels of DHA and total n-3 correlate with measures of hostility in cocaine addicts. Feeding mice a diet high in linoleic acid increases aggressive behavior, and increses the likelihood of rats to kill mice.

If the ratio of n-6 to n-3 in the diet predicts psychiatric problems, we'd expect to see an association with n-3 intake as well. Let's take a look:

This is homicide mortality vs. n-3 intake for 24 countries, published here. The association is significant (p> 0.001) even without correcting for n-6 intake. Of course, one could see this as a cloud of points with a few well-placed outliers. Here are some closer associations from the same paper:

It's clear that both a high n-6 intake and a low n-3 intake correlate with negative psychiatric outcomes. Together, the data are consistent with the hypothesis that the ratio of n-6 to n-3 impacts brain function. Dr. Hibbeln and Dr. Lands do not claim that this ratio is the sole determinant of psychiatric problems, only that it is a factor.

Now to address the big criticism that was brought up by very astute readers of the last post, namely, that the data were purely correlative. Believe me, I wouldn't even have posted on this topic if I didn't have intervention data to back it up. In addition to the animal data I mentioned above, here are more studies that support a causal role of fatty acid balance in psychiatric problems:

• EPA supplementation (a long-chain n-3 fatty acid) reduced aggression and depression in a group of women with borderline personality disorder.
• DHA supplementation improved aggression in young Thai students.
• n-3 supplementation improved symptoms of bipolar disorder.
• EPA supplementation improved symptoms of unipolar depression.
• n-3 supplementation improved perinatal depression.
• n-3 supplementation improved symptoms of major depression.
• DHA and EPA supplementation reduced suicidal behaviors and improved markers of well-being in patients with recurrent self-harm.
• n-3 supplementation decreased anger and anxiety in substance abusers.
  

Most of those were placebo-controlled trials. If we can see a significant effect of n-3 supplementation in short-term trials, imagine how well it would work as a long-term preventive measure.

Vegetable Oil and Homicide

One of the major dietary changes that has accompanied the downward slide of American health is the replacement of animal fats with industrially processed vegetable oils. Soybean oil, corn oil, cottonseed oil and other industrial creations have replaced milk and meat fat in our diet, while total fat consumption has remained relatively constant. The result is that we're eating a lot more polyunsaturated fat than we were just 30 years ago, most of it linoleic acid (omega-6). Corn oil may taste so bad it inspires you to violence, but its insidiousness goes beyond the flavor. Take a look:


This figure is from a paper that Dr. Joe Hibbeln sent me recently, of which he is the first author. This followed an interesting e-mail conversation with Robert Brown, author of Omega Six: the Devil's Fat. He put me in touch with Dr. Hibbeln and Dr. William Lands (NIH, NIAAA), both of whom warn of the dangers of excessive linoleic acid consumption from modern vegetable oils. Dr. Lands has been researching the relationship between dietary fat and inflammation since the 1970s, and has been a critic of modern vegetable oils for just as long. Both Dr. Hibbeln and Dr. Lands were very generous in sending me a number of their papers. The figure above shows the homicide rate vs. linoleic acid consumption of five countries over the course of 40 years. Each point represents one year in one country. The U.S. has the distinction of being in the upper right.

I asked Dr. Hibbeln how he selected the five countries, and he told me the selection criteria were 1) available homicide and linoleic acid consumption statistics, 2) "first world" countries, and 3) countries representing a diversity of linoleic acid intakes. I'm satisfied that there was probably not a significant selection bias.

What's interesting about the graph is that not only does the homicide rate track with linoleic acid consumption across countries, but it also tracks within countries over time. For example, here is the same graph of the US alone:


And here is the UK, which doesn't suffer as much from the confounding factor of firearm availability:


I don't think we can draw any solid conclusions from this, but it is worth noting that epidemiological associations don't get much better. In the next few posts, I'll explore the data from intervention trials that support the hypothesis that excessive omega-6 consumption, and insufficient omega-3 consumption, cause serious problems for psychiatric and physical health.