Going Behind the Headlines: Fish Oil Risky for Healthy Hearts?

OmegaMatters: Episode 35

Hosts: Drs. Bill Harris & Kristina Harris Jackson

About this episode:

For this special edition of OmegaMatters, Drs. Bill Harris and Kristina Harris Jackson review a recent BMJ paper that concluded fish oil was bad for healthy hearts. There are several points of contention in this paper that Dr. Bill Harris reviews in detail. For more information on OmegaMatters, visit: https://omegaquant.com/omegamatters-broadcasts/

SHOW TRANSCRIPT

Dr. Kristina Harris Jackson:  Welcome to a special edition of OmegaMatters. We are going to be going over a very recent paper and responding to it, and Dr. Bill Harris is going to take the lead here. So I’ll let you get into it.

Dr. Bill Harris: Thanks, Kristina. Yeah, we thought that this new paper that’s come out in the British Medical Journal, that it’s got a lot of noise over the weekend, should be addressed. And so I’ve got just some headlines on this first slide. “Fish oil supplements linked to first time heart problems.” Not clear what that is. “Regular use of fish oil may increase first time heart disease and stroke.” Both incorrect. “Fish oil supplements may raise your risk for heart disease and stroke if you’re healthy.” So if you’re not doing, take them.

Dr. Kristina Harris Jackson:  That’s wild.

Dr. Bill Harris: All these news stories really raise a lot of concerns and they are really very unfortunate. So here is the study that we’re going to talk about. It’s called. “Regular use of fish oil supplements and course of cardiovascular diseases: prospective cohort.” So it’s unusual to see the word, “course of” in a title like this, but we’ll talk about that. This is a group from China that’s been very active in using the UK Biobank data to look at omega-3 in relationships with disease. The objective to examine the effects of fish oil supplements. Actually it’s not, it’s reported use of fish oil supplements.

Dr. Kristina Harris Jackson:  And what’s the difference between that?

Dr. Bill Harris: If they actually were giving fish oil supplements to people, then they would be studying the effects of fish oil supplements on, but what they did was ask people to report, “Are you taking fish oil supplements or not? Yes or no?”

Dr. Kristina Harris Jackson: It’s probably the lowest level of evidence.

Dr. Bill Harris: Right.

Dr. Kristina Harris Jackson:  When it comes to this.

Dr. Bill Harris: So they’re looking at the clinical course of cardiovascular disease from a healthy state to atrial fibrillation, major adverse cardiovascular events and subsequent death. Prospective cohort study. That’s what we often do. That’s very fine. No problem. UK Biobank, we’ll talk about that in a minute. About 12 years of follow-up. That’s good. They ended up using 415,000 people, data from them. They were 40 to 69 years of age at baseline when they took their survey, and the main outcome was the incident cases of atrial fibrillation and major adverse cardiovascular events, which they call MACE or death, identified by linkage to hospital records, death registries. Then they said the role of fish oil supplements in different progressive stages from healthy status, which is the primary stage, to atrial fibrillation to secondary stage, which is silly to begin with because not all cardiovascular disease moves in this lockstep progression. Afib is a very different disease from a heart attack.

In any event, tertiary stage MACE is major adverse cardiac events and final stage is death. That always makes sense. So that’s what they say they’re doing. In the paper they show these pathways. So this figure on the left shows the number of participants that were in at baseline that were healthy and then 18,000 of them eventually over 12 years developed atrial fibrillation. So the question, what’s the risk of that? And then what’s the risk of going from there onto major adverse cardiac events or to death? So that’s the one, two, three step, the stage groups. So these are considered transition A, transition B, transition, D, E, F. There’s six different transitions, but yet they also have, in their supplementary data, what they call transition patterns. And there are four transition patterns and that’s all that they show data on. They don’t really show the data on this A, B, C, D, E, F pattern, so it’s very confusing.

In any event, the title of this slide says there are 24 different pathways put in this study, each of which examine the difference between the risk between fish oil supplement users and non-users over about 12 years for risk for going down a given pathway. It just gets too complicated to get into every pathway. So what I did was I just looked at the number of statistically significant associations that they found. And when I mean significant, I mean P less than 0.05, which is pretty standard, between a certain CVD outcome pathway and reported fish oil use. And so on the left there are 24 possible outcome pathways. Fish oil supplement users differed significantly from non-users. In 18 times they didn’t differ at all. There was no upside, no downside. There were better outcomes for people using fish oil supplements in five situations and they’re shown here in green, why the bar goes down, because it’s a lower risk for the event.

So going from atrial fibrillation to MACE, going from health to heart failure, going from heart failure to death and going from Afib to MI. Those were all significantly lower risk if you were taking fish oil supplements. There was one worse outcome, where if you were taking fish oil supplements, your increased risk for going from health to Afib was increased. So that’s the picture coming from the study. Of course the problem is, all they talk about is this one and they ignore all these, which is very unfortunate. So if you want to interpret this, it says the risk for going from a healthy state to Afib by reporting fish oil supplement use. Again, we don’t know how much people were taking, we don’t know how regular they were taking them. It’s very vague.

In any event, it was higher in fish oil supplement users than non-users. And those with AF on fish oils who went on to these other events were at lower risk. Interesting, they said in red, the absolute risk for transitioning from health to Afib in fish oil users was 4.8%. Meaning 4.8% of the people who were reporting use of fish oil supplements, developed AF over that time. And the people that were not taking, the 4.2%, took them. So the difference is 0.6% and that’s the 13% relative risk increase. But get real about it. We’re talking about a very, very small amount of increased risk made to look larger when you call it a relative risk. Now we do that all the time. I can’t criticize them for doing that. That’s what we do. That’s the way we play this game. But I think it’s very important to note that they didn’t even mention one other very important study also done in the UK Biobank, also done with reported fish oil supplementation, also looking at cardiovascular outcomes and all-cause mortality.

And it’s a paper by Li and colleagues, again, a group from China, published about four years ago. And what did they find? Well, here’s what they reported. 427,000 people. So it’s virtually all the people in the UK Biobank. Around 10 years of mortality, and if you look here in their most highly adjusted model, we’re looking at hazard ratios, HR, and a hazard ratio means sort of a relative risk. And so if we look at all-cause mortality, the people taking fish oil supplements had a hazard ratio of 0.87 and that means a 13% lower risk for death from any cause. How about cardiovascular mortality? 0.84. That’s a 16% risk reduction if you were taking fish oil. Myocardial infarction, dying from a heart attack, 20% lower risk if you were taking fish oil. Dying from a stroke, 13% lower risk. Cardiovascular event. Now the 13% from stroke was not statistically significant because they had so few people, but obviously the direction is right.

Cardiovascular events, 7% reduction. Myocardial infarction, 8% and stroke, 10%. All of those except the stroke were statistically significantly favorable for fish oil supplement use and the author’s concluded habitual use of fish oil seems to be associated with a lower risk of all cause in cardiovascular mortality and to provide a marginal benefit against CVD events among the general population, i.e. healthy people. So this is exactly the opposite, but it’s done in the same cohort.

Dr. Kristina Harris Jackson:  It’s very confusing.

Dr. Bill Harris: It’s very confusing. And Chen did not mention this paper at all.

Dr. Kristina Harris Jackson:  They didn’t break out AF in this paper?

Dr. Bill Harris: No, this did not have Afib in it.

Dr. Kristina Harris Jackson:  And at some point, the mortality stats should hold more weight than individual CVD outcomes potentially because Afib is coming out as being somewhat problematic in a lot of different studies with omega-3, but at the same time it’s the mortality and the hard endpoints, we’d call them, are in the favorable direction. So this just hits that nerve again, it seems like.

Dr. Bill Harris: Yeah, and there was an increased risk. Let’s look a little bit more at these two papers and again, it’s really unforgivable that the Chen Group would not even refer to and describe why they thought their results differed so much from a different group that studied the same question in the same population and got vastly different results. Didn’t even comment on it. And that’s really not just the fault of the authors, that’s also the fault of the reviewers and the editorial staff at British Medical Journal, who should have asked for that. But there are still some other concerns. You really wonder if the data were extracted properly. If you look at the population characteristics, again, they’re both the UK Biobank, over 400,000 people, and the mean age is the same in both. Each paper describes the population in their table one. It’s sort of the standard way to do it. The number of females was about the same. White people was pretty close to the same. People who never smoked was the same. That was reported right. But in Li et al. they said 16% of the people reported eating more than four servings of fruit per day. So that’s a variable [inaudible 00:11:38]. And Chen, it was 31%. 13% said they ate vegetables more than one once a day. And 65% in the Chen study said they did. There’s no way.

Dr. Kristina Harris Jackson:  There’s no way.

Dr. Bill Harris: Somebody’s got a problem here.

Dr. Kristina Harris Jackson:  That’s not happening.

Dr. Bill Harris: And reporting never using alcohol. In Lee, it was 30% and only 4% in Li et al.

Dr. Kristina Harris Jackson:  In the UK?

Dr. Bill Harris: In the UK. Like, really?

Dr. Kristina Harris Jackson:  Come on.

Dr. Bill Harris: And oily fish consumption, Li et al. said 55% of people said they were consuming oily fish at least twice a week. And in Chen, same data set, it doesn’t make any sense, 18% of them said they were. So somebody’s got a problem in extracting data and this should be addressed by the reviewers.

Dr. Kristina Harris Jackson: How does this work in the journal world? Does somebody see this and write a response to the editor or request for…

Dr. Bill Harris: Yeah, you write a letter to the editor. That’s the old-fashioned way to do it. There’s probably some online way of chatting about these things now, but I’m writing a letter to the editor with a colleague, that it primarily points out how this new study actually supports the use of omega-3 supplements more than it disses it. More importantly, I think if we look at the statistically significant associations between reported fish oil use and disease outcomes in the UK Biobank, these are 10 other studies, separate studies, one of which I just discussed, was the Lee paper, which is the last one in this list. But beyond the Lee paper, there were nine other papers that have looked at the same data set, looking at different outcomes, as a function of fish oil supplement use or not.

And what we see here, somebody looked at incident coronary heart disease, someone looked at kidney stones, someone looked at Afib. There we go. Risk for liver cancer. Total in CVD mortality in people that already have cardiovascular disease. So that’s secondary prevention. Looking at dementia, a couple of people looked at that. Inflammatory bowel disease, fractures. And then Li et al. looked at all these outcomes along with all-cause mortality. If you look at these findings from each study, there were 18 findings that were favorable for omega-3 supplement use. There were three that was not significant, no difference, and one that was adversely affected and it was Afib.

So the looking at the bigger picture of omega-3 supplement use on health writ large in the UK Biobank, it’s strongly in favor of the use of omega-3. So this is something else that was not commented on in the Chen paper, unfortunately. Chen, as we’ve said, used reported fish oil intake, and it’s just a question at the beginning of the study, a touchscreen question, “Do you take fish oil supplements regularly? Yes or no?” Far better it is to look at blood levels of omega-3. So we have looked in the UK Biobank at blood levels of omega-3 and we looked at risk for a variety of things. All-cause mortality in the upper left. You can see these are quintiles of omega-3 levels in the plasma, meaning that higher… Q1 is always the lowest level of DHA in this particular case. And Q5, quintile five, is the highest level and each of these bars contains 20% of the people in the study.

And you can see that the higher the DHA level, the lower risk for dying, total mortality, the lower risk for cardiovascular mortality, the lower the risk for cancer mortality and the lower the risk for all other causes combined. So when you’re looking at not reported fish oil use, but the actual circulating levels of omega-3, which is of course partly driven by fish oil, partly driven by fish intake and other factors, there’s a strong beneficial relationship with the more omega-3 in the blood and these health outcomes. And so that again, something else that was not commented on in this Chen paper. Even beyond this, I’ve been involved for some time with a group called FORCE, the Fatty Acid Outcomes Research Consortium. We have published now 10 or so papers, or maybe it’s six papers, looking at omega-3 levels as predictors of a variety of diseases. Not in one cohort like the UK Biobank, but in 15 to 30 different cohorts all around the world with all new analyses of relationships.

And we’ve seen favorable relationships with omega-3 fatty acids for coronary heart disease, for developing type 2 diabetes. This is all for incident disease. This is not cross-sectional, this is longitudinal. Higher levels of omega-3 are associated with lower risk for coronary disease. Type two diabetes, all-cause mortality, cancer mortality, CVD mortality, lower risk for atrial fibrillation, not higher, lower risk, lower risk for stroke and lower risk for kidney disease. So everywhere we’ve looked so far, the only place that there’s not a favorable is that there’s no significant effect of omega-3 levels on hemorrhagic stroke, which is a good thing, which is fine. Hemorrhagic strokes are much less common. They are bleeding stroke.

So, coming to the end, the problems with Chen and with the press. Some things Chen can’t control. The authors of the paper can’t control what the press does with things. But Chen did find five favorable links to fish oil supplement use but emphasized only one in their paper. They failed to consider prior studies on the same question in the same cohort. They were each very different but favorable conclusions. Fish oil supplement use in the UK Biobank has been associated in 10 other studies with lower risk for 18 different outcomes, unrelated to three, and higher risk for one, atrial fibrillation. Comparing the blood levels of omega-3 is a more objective way than using supplement. You ask no questions. And in the UK Biobank, higher omega-3 levels are associated with lower risk for several major mortality outcomes. None of this was ever mentioned by Chen.

The press, they emphasized only a bad outcome. Completely ignored the good ones. They equated developing Afib with heart disease. True Afib is a heart disease, but it is far from the most common or most dangerous heart disease. So it misrepresents the situation and they failed to use any of the past studies on the same question in their reporting. So it’s really a sad situation when a study like this is done and inflammatory headlines are put out there and it just causes such confusion. So hopefully this little podcast here will help, some people at least, understand what was actually done and what they actually found.

Dr. Kristina Harris Jackson:  Yeah, they actually did a very strange analysis with the pathways, but they did find the same thing that has been found, generally, and it’s just off. There’s nothing here, even when they did it in an interesting way. And a lot of times scientists are just trying to find new ways to look at data but when it just gets put out to the press so quickly and then it wasn’t caught by the peer review process, sometimes… That’s just a volunteer process. So it doesn’t always catch everything and-

Dr. Bill Harris: Crapshoot you get.

Dr. Kristina Harris Jackson: It is.

Dr. Bill Harris: And sometimes they send stuff to me. That’s a crapshoot.

Dr. Kristina Harris Jackson: That’s pretty good. It’s a pretty good shot. But it’s always good to look at the actual paper, look what they actually found and there’s nothing there. Stuff we already know.

Dr. Bill Harris: Nothing to get us afraid about. I think what’s dangerous is not fish oil, it’s just bad reporting.

Dr. Kristina Harris Jackson: Bad information.

Dr. Bill Harris: That’s what dangerous. Bad information is what’s dangerous here.

Dr. Kristina Harris Jackson:  Yeah. Yeah. So that’s great. Thank you so much-

Dr. Bill Harris: You bet.

Dr. Kristina Harris Jackson:  … for putting this together.

Dr. Bill Harris: Okay.

Dr. Kristina Harris Jackson:  All right, thanks.

Dr. Bill Harris: We’ll see you next time.

Dr. Kristina Harris Jackson:  Next time. Bye.