Genes & Fish Oil — What’s the REAL Story Behind the Headlines?
OmegaMatters: Episode 3
Hosts: Drs. Bill Harris & Kristina Harris Jackson
About this Episode:
A recent study implies that one’s genetic makeup may influence how fish oil impacts triglycerides and that this might put someone at increased risk of heart attacks. In this episode, Drs. Harris and Jackson deconstruct this fatally flawed study and recommend multiple retractions. Making the leap to an increase in heart attacks by these authors is completely irresponsible, according to Drs. Harris and Jackson. Listen to what else they have to say about this “uncorrected proof” of a study.
Transcript:
Kristina Harris Jackson: Hello, and welcome to a special April Fool’s Day edition of the Omega Matters. I’m Kristina Jackson. This is Bill Harris, and we’re here to break down the most recent omega-3 paper that’s making the headlines these days. You may have seen these in the last few days, and we’ve been asked about them at OmegaQuant and FARI.
Kristina Harris Jackson: Okay. So these are some headlines that we were sent — Genes May Determine if Fish Oil Supplements Are Good or Bad for You; Do Fish Oil Supplements Support Heart Health? Study Says it Depends on Your Genes. One Major Side Effect of Taking Fish Oils, Says a New Study. Are Fish Oil Actually a Waste of Money? That’s always a favorite. And finally, If You Take This Popular Supplement, Your Heart May be at Risk, Study Says. So these are, you know, all over the place. And also vaguely terrifying, as headlines typically are. So we wanted to take a look at what study they’re actually talking about.
Bill Harris: One study looked at the genome-wide association of fish oil supplementation on lipid traits in 81,000 people. And it’s published in this journal called PLoS Genetics. There’s a lot of words there that most people don’t understand. But in any way, the question they’re asking is to what extent do genetic variations between people affect the response of serum lipids to fish oil supplementation? So we all know fish oil… taking fish oil lowers triglycerides. It’s been known for 40 years that if you take large doses, you get large reductions. And, and if you’ve ever have very high triglyceride levels, you get 30, 40% reductions in triglycerides. Those are huge.
Bill Harris: If you have very low or normal triglycerides, fish oils don’t do much to your triglyceride cells. They genotyped, meaning that they analyzed the DNA for over 80,000 people and, and analyzed all of the individual, what we call nucleotides. There are 3.1 billion nucleotides in a DNA strand in every cell in your body. And so they analyzed the sequence, and they organized them into genes and they looked for genetic variance. What we call mutations, small mutations that may or may not have any effect on anything.
Bill Harris: Anyway, they did this in a large group of people, over 80,000 people from two different cohorts. And we won’t go into the different cohorts it’s not that important. And in all of these people, they had data on their cholesterol level, their triglyceride level, their LDL and their HDL levels. And they also had some information about whether the people were using supplements or not. They answered yes or no to a question “I’m using fish oil supplements”. They don’t say how much, or how many, or how frequently, or which product, any of that stuff, just yes or no. Then they looked for what we call gene variants — called SNPs, which is just the shorthand word for a variant or a mutation in a gene. Primarily triglycerides is really what they focused on.
They associated with differences in triglyceride levels between those reporting and not reporting use of fish oil supplements. So that’s the general idea. Is there some certain SNP that is, if you have it, then you have a different triglyceride level than if you don’t have it—after you take fish oil.
Bill Harris: What they found was a statistically significant interaction between a specific SNP. We won’t go into what it is, it’s not important. And the difference in triglyceride levels in people reporting or not reporting fish oil supplement use. The found something statistically significant, which will be explained in a moment. Because when you’re dealing with sample sizes of over 80,000 people, you can find a statistically significant difference between group A and B and a completely irrelevant, clinically, difference in whatever you’re measuring. That’s what’s happened in this study. Unfortunately.
Bill Harris: To quote the conclusion of the researchers: “This research suggests that interpersonal variation in triglyceride response to fish oil supplements is in part.” They don’t say what part, unfortunately, because it’s probably less than a quarter of a percent… “In part explained by genotype.” This means your genetic makeup. “And that fish oil dose adjustment based on genotype should be investigated.” They didn’t study this at all. “As a means to protect against cardiovascular disease risk.” And they didn’t study cardiovascular disease risk at all.
Bill Harris: I think the take this really too far. And the biggest problem with this paper actually is the interpretation by the authors of their findings. That’s the problem. The data itself are fine. So that’s the study. Now let’s look at the abstract.
Kristina Harris Jackson: We read the journal articles first. And then came to the paper and we’re having a heck of a time trying to figure out how they were connected. So the abstract says the minor allele is significantly associated with decreased triglycerides, triacylglycerols in individuals with fish oil supplementation. But with increased TAGs in those without supplementation. So that’s what was written in the paper. We assume they’re comparing this to the major allele. So we go over here, the AA, AG and GG, these are the different genetic SNPs and alleles.
Kristina Harris Jackson: So the AA has the most people. This is the major allele; it’s what most people have. The AG is a heterozygous. There’s two different kinds of nucleotides there. It’s less common. And GG is very rare. So we have these two main alleles, major and minor. Now let’s just try and orient ourselves to the graph so we can try and figure out what they’re trying to say. The AA is the most people, whether or not they took fish oil, didn’t really affect their triglyceride level. If you look at these two bars.
Bill Harris: So if we’re talking about the green versus the green, this group here is not… says they’re not taking fish oil. This group says they’re taking fish oil. So look at the triglyceride levels, uh, between green and green, right?
Kristina Harris Jackson: Yeah. Just to try and orient yourself.
Bill Harris: There’s definitely no difference.
Kristina Harris Jackson: The difference here is technically 1.7 milligrams per deciliter. It’s 149 in the non-fish oil people and 150 in the fish oil people. This is actually what they’re saying. When they’re saying there’s an increase in triglycerides when you take fish oil with this genetic variant. The other genetic variant you have in the red, and you can see with people not taking fish oil versus people taking fish oil. There’s a little bit more of a difference about 12 milligrams per deciliter. So the minor allele is significantly associated with decreased TAGs in the individuals with fish oil supplementation. That’s that red bar 140. But with increased TAGs in those without supplementation, that’s the red bar 152. And we assume they’re actually comparing against the major allele. So 152 versus 149 is increased, 140 versus 150 is decreased. That’s what that statement means, right?
Bill Harris: Kristina and I know a little bit about science and how to read papers. And we had a terrible time trying to understand because they’ve written this incorrectly is part of the problem. For example, in this just here, in the abstract, the first sentence, the minor allele, which is the AG, the red is significantly associated with decreased triglycerides and individual fish oil supplementation. But they didn’t say compared to who.
Kristina Harris Jackson: Right.
Bill Harris: They didn’t even mention who you’re talking about, who their comparison is with.
Bill Harris: If it’s compared to the major, the green bar, this one here, then that is significant. It is completely irrelevant. Clinically, 149 to 152 is virtually the same number. But when you have sample size of 80,000, these becomes statistically significant and very easy to, to misunderstand. Then it says that the minor allele, I’m just, I’m, forgive me for just repeating this, I’m just trying to understand it.
Kristina Harris Jackson: We’re still trying.
Bill Harris: The minor allele is associated with increased triglycerides and those with supplementation, the minor allele the… nothing there makes sense. Here’s the problem. It says the minor alleles blah, blah, blah, but it, the minor allele is associated with increased triglycerides in those without supplementation.
Bill Harris: Without supplementation. So there is a slight, a three milligrams per deciliter difference and they’re calling that significant, it’s statistically but not clinically significant.
Kristina Harris Jackson: Yeah.
Bill Harris: The minor allele is significantly associated with a decrease in triglycerides in individuals with supplementation. That’s this, if you got the minor allele the red, then your triglycerides are lower than if you have the green, fair enough. There’s a 10 milligram difference, 140, 150, still completely irrelevant in the clinic. Um, so at 10 milligrams, so this is pointing at this one here, the 10 milligram per difference between fish oil supplementation in these two groups.
Bill Harris: All this really means is that if you’re taking fish oil and you have this genetic characteristic and you’re taking fish oil supplements, your triglycerides are not going to be affected as much as they’re affected if you have the genetic characteristic in the red bar. And the difference is only 10 milligrams, which is virtually nothing. So this is the problem. And then another comment that we have here at the bottom about the extrapolation of these findings. I’d say some minor difference, very small difference in triglyceride levels, but to say: “To increased risk for heart attacks is completely irresponsible.” Um, the authors have gone way, way beyond the data in order to scare people and get headlines apparently. By interpreting this very, very tiny difference in triglyceride levels as increased risk for heart disease is ridiculous and maddening.
Kristina Harris Jackson: So for the triglyceride level context, we’ve put this into milligrams per deciliter, cause that’s usually what we see it at, in the US expressed as. So 150 is kind of a cutoff, um, for moderately elevated triglycerides, essentially. Being below 150 is like normal, healthy, better.
Bill Harris: Well, 200, 250, or over 500 veer into the very high triglyceride level.
Kristina Harris Jackson: Right. So these are numbers that are maybe borderline high, but they are in the normal range. So, you know, one, you assume that the triglyceride effect for fish oil is most obvious in people with high levels of triglycerides, which there’s going to be some in here, but it’s mostly not. And you also assume people are taking a therapeutic dose of fish oil to have that effect. There are issue with self-report and nutrition. No one’s surprised to hear that. So you take two assumptions there, and then I think the wording of increased and decreased is problematic because it’s really just a higher or lower level.
Kristina Harris Jackson: But in the context of clinical use, this pretty much doesn’t matter. It’s almost irrelevant because the numbers are so close. But they’re not between 300 and 100. They’re between 140 and 152. So there’s some context to these numbers that is definitely missing. And it’s nowhere in the paper that I saw. Where they actually put into context what these triglyceride values meant and then how those are actually do relate to heart disease risk. Because that’s not the clearest line either.
Bill Harris: So here’s their actual report, not the abstract, but this is in the body of the paper. They say, “Triglyceride levels were significantly higher in the heterozygotes.” That’s the AG, the red bars, um, who did not take fish oil. And then they have this thing… difference in triglycerides that this line should be out, it’s your, it’s wrong. Shouldn’t be in the paper.
Bill Harris: And it said, “Those who did not take fish oil versus those who did take fish oil.” So that comparison is right here, red bar, the AG, uh, the difference is 11.7. So this red bar to that red bar. So that’s about 12 milligrams per deciliter. Hopefully they’ll correct this. They called it a fish oil effect. It’s not an effect. It’s a difference. You can only see effects when you actually randomize people and put them on fish oil or not. And they didn’t do that. Then going on in their writing, in the section of the paper with respect to, and this is just a long-term for this AG red gene variant. “While fish oil supplementation is associated with lower triglycerides in heterozygotes.”
Bill Harris: In the reds, that, which they stated up here, actually, this here say, it’s higher. Now we’re talking about lower. So they’ve backwards stated, no wonder this is confusing. It has a slight opposite effect. It doesn’t have any effect at all. There’s no effect here, this is a difference in the AA homozygotes, the green ones. Where we have the group in green here, and now here, we’re talking about this 1.7 difference, where’s that? Right here, the green one. So green to green, no SNP, no fish oil, fish oil. And if they’re going to call that an increase in triglycerides of less than two milligrams per deciliter, then this paper should be pulled.
Kristina Harris Jackson: Yeah.
Bill Harris: It’s absolutely wrong. Then we just made the comment here, no way, can this be construed as an increase in triglycerides with fish oil supplements, without a controlled intervention, a difference does not mean a change. And certainly this difference here with fish oil supplementation on these, these two different genetic types, this is not an increase. It’s a failure to decrease.
Bill Harris: So this is not elevated. If you have this, this particular SNP, it’s just, you don’t really get a treatment effect. You didn’t get an effect on triglycerides, if you had this, if you were in this green group. Which is fine… but that’s different than raising triglycerides, which is what they’re trying to say.
Kristina Harris Jackson: Right.
Bill Harris: And that’s a total misunderstanding here. Let me go onto this. Tina, you want to take this one now? This is from the press release.
Kristina Harris Jackson: The quote from the press release was: “What we found is that fish oil supplementation is not good for everyone.” And they’re basing this on… from what we can tell, that difference in triglyceride levels, um, that can’t be called bad and does not mean people are at increased risk for heart disease. Being at 149 versus 150 is basically the same thing. So interpreting that slight increase, as they say, as being at increased risk, increased triglyceride levels, is pretty off. It just does not have the right context. It’s just talking about statistics and it doesn’t take it to the clinical, like common sense realm.
Kristina Harris Jackson: Then they go on to say, “It depends on your genotype. If you have a specific genetic background then fish oil supplementation will help lower your triglycerides. But if you don’t have the right genotype taking a fish oil supplement actually increases your triglycerides.” This is a very extreme interpretation of what they found here.
Bill Harris: What they should have said was it didn’t lower triglycerides.
Kristina Harris Jackson: Yes.
Bill Harris: It does not increase them. That’s completely wrong.
Kristina Harris Jackson: Yeah. The increased side is just flat out wrong. I’m not seeing the effect, um, in the dominant phenotype with people who reported taking fish oil, it is more of what you would focus on in this study. It’s still making a lot of assumptions about whether or not that fish oil was therapeutic or even needed. So it’s just like all of these triglyceride differences between the groups are so close that it really should not be making big statements about what that triglyceride level means for people’s risk. Because they’re nowhere in that ballpark. If you’re starting to talk about really high risk levels. And so if you did, I think maybe we already said that, but if you wanted to answer this question, which some people have attempted to do this, we just found a 2019 paper where they did this. They recruited people, measured their SNPs, and put everyone on two grams of fish oil.
Kristina Harris Jackson: And then measuring triglyceride response to that fish oil supplement. And then you can see, “Oh, if you have this SNP and you are on this dose of fish oil, you decrease by 50%, if you had the other SNP, you only decreased by 10 or you didn’t decrease at all.” That would be the type of study you’d have to do to say these things.
Bill Harris: Just to wrap this up, this is the actual paper as it’s been presented, and it’s obviously an uncorrected proof. And I can’t think of a better term for this paper than an uncorrected proof. It vastly needs correcting. It leaves out a lot of context and this discussion needs to be much muted down. And hopefully they can retract some of these statements that they’ve made. They probably won’t or they, if they do, won’t be picked up by the press.
Kristina Harris Jackson: Right.
Bill Harris: The bottom line is this study does not show that fish oils are harmful by any means. And it does not show that fish oils are dependent upon your genetic makeup for triglyceride lowering. We do know that there are varying response of triglycerides to fish oil. And it’s not impossible to imagine that that is there, there’s some genetic component to it, but we don’t know. And this paper does not tell us how much genetic variation there is.
Kristina Harris Jackson: I think that wraps it up really well. Hope this is helpful.