UK Biobank Studies: Fish Oil Use

OmegaMatters: Episode 19

Hosts: Drs. Bill Harris & Kristina Harris Jackson

Background and Key Takeaways:

The UK Biobank is a large-scale biomedical database and research resource that is enabling new scientific discoveries to be made that improve public health. The resource provides accredited researchers access to medical and genetic data from half a million volunteer participants to improve our understanding of the prevention, diagnosis and treatment of a wide range of serious and life-threatening illnesses. Through the long-term commitment of 500,000 participants, together with the support of its funders, the UK Biobank is enabling new scientific discoveries that benefit people’s health. The UK Biobank has been used in several recent studies examining the relationship between fish oil use and several health outcomes. In this episode, Drs. Harris and Jackson discuss 8 of these studies. For more information on OmegaMatters, visit:


Dr. Kristina Jackson: Welcome back everyone. It’s OmegaMatters. I’m Dr. Kristina Harris Jackson and this is Dr. Bill Harris. And we are going to do a special version of OmegaMatters this week. Bill has a lot of interesting research around the UK Biobank to share, and we wanted to do a kind of overview of what that is and what kind of information we can give from it. And so we’re going to have slides this time, and it’s just going to be us chatting about it. So take it away.

Dr. Bill Harris: All right. Thank you. Thank you, Tina. We at the Fatty Acid Research Institute, FARI have been, generally interested in studies looking at the relationship between omega-3 levels in the blood and risk for a variety of diseases. We’ve recently been able to access the UK Biobank data set through the good work of our primary bio-statistician, Nathan Tintle. So what is the UK Biobank? This is a  prospective, observational, population-based cohort of about a half a million people that was assembled between 2007 and 2010 of people between the ages of 40 and 69 at several centers throughout the UK.

Dr. Bill Harris: The baseline data that they have in this database were from touchscreen questionnaires that collected a socio-demographic characteristics, diet, lifestyle, as well as biological samples and physical measurements. But as we’ll see in a minute, not everybody got measured for blood,   metabolites. Longitudinal recording of what happened to the people that after they got enrolled was through a mix of primarily in person and also electronic medical record. Because in the UK everybody’s kind of in the system. And so you can track sort of in real time what’s happening with anybody in this UK Biobank and create data sets.

Dr. Bill Harris: So the question that we’re going to talk about today is: What do we know about the relationship between taking fish oil and a variety of outcomes? I’m not going talk about any of our research at FARI on this particular podcast. I want to talk about what other people have done. And, and virtually all of these studies I’m going to talk about have been done by Chinese investigators who have been through UK Biobank date and are asking these good questions. So the question we’re going to look at today is one that was asked of all the participants. Do you regularly take any of the following? And you can select one or more. What they found at baseline with about 500,000 people whether you take fish oil or cod liver oil. And about 31% of people said yes they did.

Dr. Kristina Harris Jackson: Is that pretty high?

Dr. Bill Harris:I thought that was kind of high. But then we talked to Ellen Shutt at GOED, and she said, “Well, you know, in Great Britain there’s a lot of people,” and their data was around 35, 31, and 25%. So this is not that unusual.

Dr. Bill Harris: There was a re-ask of this question between 2012 and 2013, they found around 33% said they were taking. So it’s probably true, that’s right around that one third of people reported taking fish oil supplements. And so the rest of this presentation is going to be people who have taken the information that I’m showing you right here and asked the question, “Did people who were taking fish oil supplements develop disease X, Y, and Z any differently than people that did not take fish oil supplements or did not report it?” So that’s where we’re going. And as you can see, we’ve got quite a few people taking supplements. So there’s a lot of data here that can be mined.

Dr. Bill Harris: So I’ll start with the first one. I am going to go through about seven or eight papers fairly quickly. This first one, and this is in a fairly random order, was called the “Association Fish Oil Supplementation with Risk of Incident Dementia.” And this is a prospective study that had 215,000 people. So these people that they used in this study certainly couldn’t have dementia at baseline. And they had an average age 64. So they picked a group of people that was kind of older. They didn’t include the 40-year-olds who didn’t have time to develop dementia. So about 215,000 were involved with this one. Again, they got the fish oil supplementation data from the questionnaire. They used standard ICD-9, ICD-10 codes for the development of dementia. And then they did some statistical analysis.

Dr. Bill Harris: They found 2000 incident cases of dementia over a 7.9 year period. So that’s the question: Was developing dementia related to taking fish oil supplements or not? Here are the data they report in the paper. This is their primary outcome paper. What I have in the red box is the fish oil supplement users with 84,000 people and non-supplement users on the left hand side, 130,000. And what this shows different statistical models where more and more things are adjusted for. The first model adjusted for 15 variables and model two adjusted for 12 more dietary variables. And then model three adjusted for a healthy diet score.

Dr. Bill Harris: And what you see here, is that the fish oil supplement use on the most conservative model, one that’s adjusted for the most things, there was a 13% lower risk for developing dementia if you were taking fish oil supplements. 

Dr. Bill Harris: The next paper was kind of a similar disease, except now we are looking at people with distinct APOE e4 level. APOE-e4 is gene variant and people who have this APOE e4 variant are at higher risk for developing dementia, Alzheimer’s disease. So the question was: Is there a relationship with fish oil supplement use in people who didn’t have any APOE-e4, some people who had one e4 allele and, and one e- non e4. And then people had two APOE-e4 alleles.

Dr. Bill Harris: In any event, there’s increasing doses as you were of genetic risk for developing Alzheimer’s. There were many more people, 446,000, involved with this one, free of dementia baseline complete data and supplement use. Again, they looked at hazard ratios and for people who did and did not use fish oil. And what they found here was… now what we’re showing here is panel A is for all cause dementia, panel B is for Alzheimer’s disease, and panel C is for vascular dementia. Different ways you can,  your brain can get worn out getting older. And if we’re looking at just at Alzheimer’s disease, what they showed here was that taking fish oil was associated actually with a reduced risk. Well, there was a trend toward reduced risk.

Dr. Bill Harris: So there’s no significant relationship here in Alzheimer’s,   for all cause dementia. There was a benefit that’s shown by this bar, this dot on the left hand side of the red line, there was a significant reduction. That’s 14% reduction in risk if you didn’t have any APOE-e4 alleles. If you had one APOE e4 allele, then you were still significantly lower, but less of an effect. And if you had two APOE e4 alleles, taking fist oil supplements didn’t help at all. Didn’t change anything. That make sense?

Dr. Kristina Harris Jackson: Mm-hmm.

Dr. Bill Harris: Good.

Dr. Kristina Harris Jackson: So when they say fish oil benefits and non-fish oil benefits, that just means the fish oil is benefiting or it’s not?

Dr. Bill Harris: Means if the dot is on this side of the fish oil benefit line, that means there was fish oil benefit. If there was a dot on the other side of the line, kind of like this one over here for vascular dementia.

Dr. Kristina Harris Jackson: Right.

Dr. Bill Harris: If it’s over here, it means that actually the people taking fish oil did worse.

Dr. Kristina Harris Jackson: Okay.

Dr. Bill Harris (11:55):

Not taking fish oil, was better if you had two alleles of the APOE e4. But if you had zero or one, so about 25% of people have this. So it would be, two of these together would be the 25%.

Dr. Bill Harris: This, this could be viewed as an adverse effect. Like whereas these are beneficial effects. So taking fish oil, if you, for like 80% of the population that doesn’t have APOE-e4 doesn’t have e4, taking fish oil is beneficial for all cause dementia… and a trend for Alzheimer’s disease.

Dr. Bill Harris: So they concluded,  fish oil supplement use was associated with reduced risk for dementia, more in those without APOE-e4 than those with it.

Dr. Kristina Harris Jackson: Yeah.

Dr. Bill Harris: This kind of fits with some stuff that we’ve been doing here at FARI as well.

Dr. Kristina Harris Jackson: Nice.

Dr. Bill Harris: All right. What’s the next one? These are all over the map. Here, we’re looking at primary liver cancer. These researchers were interested in cancer and they asked the same question everybody else is asking. Started with 430 some patients free of cancer at baseline. And again, they were computing hazard ratios for developing liver cancer associated with fish oil use. And they adjusted for a variety of other things like they always do. And so what we have here on the right hand side is their primary results. This top box is liver cancer. And then this second box is hepatocellular carcinoma. So that’s a certain subset of liver cancer. And then there’s intrahepatic cholangios sarcoma. That’s a different kind of liver cancer. So these are subsets of the two.

Dr. Bill Harris: As you see, there’s 205 cases of total liver cancer, which is about a little bit more than the sum of those two. So what they find for total liver cancer, when you look at multi-variable adjusted hazard ratio where people who were not taking fish oil are the reference group and people who were taking fish oil, the hazard ratio’s 0.56. So that’s a 44% lower risk of developing liver cancer if you were taking fish oil supplements. Impressive.

Dr. Kristina Harris Jackson: Mm-hmm.

Dr. Bill Harris: And there’s two subsets. 0.48 hazard ratio for hepatocellular carcinoma, 52% reduced risk and then 0.6 for this cancer, a 40% reduction. All statistically significant.

Dr. Bill Harris (14:45): … pretty interesting. There was, as we see here, fish oil supplement use associated with 40 to 52% lower risk of developing liver cancers.

Dr. Kristina Harris Jackson: Yeah. And it looks like they adjusted for a whole heck of a lot of things.

Dr. Bill Harris: Yeah. It’s all the fine print down here that they adjusted for. Right. Of course. All this, everything that we’re going over today is, or I guess you’d call it hypothesis generating data because you can’t prove cause and effect.

Dr. Kristina Harris Jackson: Mm-hmm.

Dr. Bill Harris: This is actually reported fish oil use based intake. So this is interesting information that supports a beneficial effect so far but doesn’t prove that the fish oil’s doing it.

Dr. Bill Harris: What’s the next disease? Next have inflammatory bowel diseases. And here we’re looking again at 450,000 touchscreen questionnaires. They had 1600 incident cases of inflammatory bowel diseases, broken into two categories — Crohn’s disease and ulcerative colitis.

Dr. Kristina Harris Jackson: Mm-hmm.

Dr. Bill Harris: And about twice as many had ulcerative colitis as had Crohn’s. So they’re going to look at both of these. And what they found here was, again, looking at our hazard ratio here, is the non-fish oil supplement users, 308,000, and 140,000 people using them. Looking at the multi-variable adjusted hazard ratio with the non-fish supplement oil users as the reference group, 0.85. So that’s a 15% reduction in risk for ulcerative colitis in people that were taking fish supplements.

Dr. Kristina Harris Jackson: (16:26):


Dr. Bill Harris (16:27):

Crohn’s disease 0.88, so that’s fundamentally a 12% lower risk, but it was not statistically significant.  0.22 was the p value, so we can’t say there was any effect. There certainly is a trend in that direction. This is a much smaller number of people. So it’s more difficult to get a significant p value. But the authors concluded habitual use of fish oil supplements was significantly associated with lower incidence of inflammatory bowel disease and its subtype ulcerative colitis not in Crohn’s. Which, and altogether these are probably due to the anti-inflammatory effects of omega-3.

Dr. Kristina Harris Jackson: (17:04): Yeah. And it is also is important, and maybe obvious to us, but just when we talk about lower risk for these multifaceted diseases, to keep in mind this is not a medicine, this is a nutritional component. This is a nutritional supplement. And so it’s, it’s kind of important to not think, “Oh, if I take this supplement, I won’t get ulcer of colitis.” That’s many steps beyond what this says. This just means there are a lot of people who had ulcerative colitis in both groups, but a lot more of the non-fish oil supplement users had it compared to the fish oil users. Sometimes I think we get sucked into thinking that if I take fish oil, I’ll never get these things. It’s only potentially lowering risk.

Dr. Bill Harris: Lowering risk. Right.

Dr. Kristina Harris Jackson: We need more evidence. But just in the wild, you’re just looking at people living their lives. Which is what’s cool about the UK Biobank is that it’s fairly representative.

Dr. Bill Harris: Right.

Dr. Kristina Harris Jackson: For example, an omega quant cohort is not representative, because people who work with us choose to work with us and buy their own test kits — that’s not representative. But  this is just anybody going to their doctor. So that’s the power of this cohort, and it’s huge. And of course some weaknesses of this kind of data are that we don’t know really how often they were taking these supplements, how much they were taking, what brand they were taking, any of that stuff. Usually the bigger data gets, the harder it is to know the quality.

Dr. Bill Harris: Yeah.

Dr. Kristina Harris Jackson: You’re hoping the signal comes through the noise.

Dr. Bill Harris: Next one. Now we’re looking at type 2 diabetes. This is a common problem and they’re looking here now around 393,000 middle-aged older adults who did not have diabetes or major cardiovascular disease or cancer at baseline. So that’s part of why there’s not as many people in this study, because they eliminated a bunch of folks who already had some type of disease.

Dr. Kristina Harris Jackson: Mm-hmm.

Dr. Bill Harris: There was some 24-hour recalls. Okay, 10 years of follow up here, 7,200 incident cases of type 2 diabetes. So there’s a lot of people who got diabetes in this time. Let’s see how it was related to fish oil supplement use. Once again, looking at the red box, now we’re looking at this, the same story as we’ve been looking at before here. If we just look at the top part, fish oil supplement use zero top line is zero. Meaning that’s the reference group. And number two is, yes they were. That’s the omega-3 group,   0.91 hazard ratio. So that is a 9% lower risk for developing diabetes and it’s statistically significant. Interestingly, this group went a little bit further with this using data; there was some follow up data from a food frequency questionnaire. Out of 163,000, they looked at constant non-use of fish oil, occasional use of fish oil, modestly constant, kind of more constant, highly constant.

Dr. Bill Harris: So people answered the question somewhere, indicating “Yeah, I’m pretty good at that. I do fairly good. …at taking fish oil.” So there was kind of a dose response thing going on here. So if you look at the odds ratios, again, the people that say they didn’t take it at all, that’s their reference group. And then people were occasional, so their risk was 0.93— a 7% reduction. And then little more fish oil, a little more regular so .85 hazard ratio — a 15% reduction. Then a little more regular and now we’re talking a 22% reduction, .78 hazard ratio. And then highly constant, which was a 23% reduction. So we’re very similar here at the end. But when they say “constant” use that is a 0.82 hazard ratio and overall 18% lower risk. So in any event, that’s a very nice dose response relationship that we see.

Dr. Bill Harris: People who are more regular taking fish oil supplements seem to have lower risk of developing diabetes. And if you could read the small print at the bottom, you see all the things that they adjusted for trying to control for that whole healthy lifestyle stuff that we always bring up.

Dr. Bill Harris: You hear about omega-3s being a surrogate for…

Dr. Kristina Harris Jackson: So can you explain that a little bit more?

Dr. Bill Harris: Right. Sometimes you hear the concern about whether it’s a biomarker base or if I use fish oil, it’s just maybe people who use fish oil are just living healthier lives than several other categories of life. Maybe they exercise more, don’t smoke, eat more fruits and vegetables, drink moderately, things like that, lifestyle. And that it’s not the omega-3s and the fish oil that’s doing anything — it’s just this other lifestyle and the fact that you’re taking omega-3 is just a marker for that.

Dr. Kristina Harris Jackson: Yeah.

Dr. Bill Harris: The only way to really test that is with a randomized trial. Just randomize people in two groups and make sure every lifestyle factor is evenly distributed in both groups. So that controls for it. But what we try and do statistically is we control for, try to adjust for many of these factors and say, “Well, independent of whether you’re smoking or not, independent of whether you’re fat or not, independent of whether you eat vegetables or not, there’s still this relationship.”  so that it, it works toward trying to get more of a cause and effect relationship. But it’s not perfect.

Dr. Kristina Harris Jackson: Right.

Dr. Bill Harris: So that’s the lifestyle issue. So this is a nice study for looking at a dose response. Okay, what was this one? I’ll tell you. Well this is oily fish. This is also kind of interesting. People also answer questions in the UK Biobank about how often they eat oily fish or non-oily fish. And so what they found here was, again, if you just look at the oily fish intake versus here’s non oily fish you see the same kind of thing. People who are eating more and more oily fish have lower and lower risk of developing diabetes.

Dr. Kristina Harris Jackson: Yes.

Dr. Bill Harris: But that’s only true for the oily fish. It’s not true for non-oily fish intake levels.

Dr. Kristina Harris Jackson: And oily fish definition is cold water fish that usually are higher in omega-3s — higher than non-oily fish. So oily fish, including salmon and mackerel and tuna herring and non-oily is more tilapia, white fish, things like that, that have lower omega-3 content.

Dr. Bill Harris: This next study looked at 427,678 people who reported fish oil supplement use and they looked at total mortality now, not any given disease, just dying of anything. So we got the blue box on this side. It’s the most conservative model. On the first line, it’s risk for all-cause mortality or dying,   from any cause. 0.87. So if you haven’t figureed it out by now, you just subtract that from 100 and you get 13% lower risk of death from any cause. Cardiovascular mortality is the next line. 0.84, 0.80 is myocardial infarction mortality, dying of a heart attack. That’s 20% lower risk. Stroke is not statistically significant probably because there’s not many strokes — thousands over here and only 400 for stroke mortality. Cardiovascular events, myocardial infarction (MI) and stroke, etc, were lower with people who were taking fish oil supplements. So there’s another good sign there that things were happening again.

Dr. Kristina Harris Jackson: And so there was death from those things and just having one of those events and the fish oil was stronger at preventing death from those events, but less strong about whether or not you’d actually get a heart attack. Like you could get a heart attack and be more likely to survive…

Dr. Bill Harris: So that’s this example here. So just having an MI, whether you’ve survived or not, there’s 5,300 people that had an MI and 1000 people died of MI.

Dr. Kristina Harris Jackson: Yeah.

Dr. Bill Harris: So the odds ratio is 0.92, that’s 8% lower risk, but the odds of dying are 20% lower.

Dr. Bill Harris: Okay. What’s next? We’re getting close to the end here.

Dr. Kristina Harris Jackson: Let’s see.

Dr. Bill Harris (26:23): This one is bone fractures. Again, there were around 500,000 people in this and they looked at over a period of eight years, who broke what and what they found. And what they found is kind of tiny. In fact, it’s so, so tiny I’m going to blow it up here. This is the column we’re looking at.

Dr. Kristina Harris Jackson: Yeah.

Dr. Bill Harris: So this is the most conservative model. Total fractures 0.93, statistically significant, 7% lower. Hip fractures, 17% lower. Vertebral fractures, 15% lower and other fractures still significant, but only 5%. So primarily hip and vertebral. The same kind of story in total recurrent fractures, significantly lower, significantly lower for vertebral, not for hip. So the conclusion of that one is, it’s just a lower incidence and recurrence of total fractures associated with taking fish oil supplements.

Dr. Kristina Harris Jackson: Have you ever seen anything in that area of research before? Bone health?

Dr. Bill Harris: I’m sure it’s in the paper. I have to read the discussion in the paper, but off top of my head, there’s not been a lot of connection of omega-3 with bone health. It’s just not been studied as much.  There’s some data out there, and I’m sure in their discussion they just combed the literature to find what’s out there.

Dr. Kristina Harris Jackson: There is research about muscle retention as you get older and fish oil, so it could lead to better gait, better muscle retention…

Dr. Bill Harris: Good point. You know, and I don’t know if in this paper I have to look at it again, whether they actually looked at rates of falling.

Dr. Kristina Harris Jackson: Right.

Dr. Bill Harris: I just saw this study today, so this is brand new, published September 1st, but I didn’t see it until today. Here they’re looking at habitual fish oil supplementation and risk for incident atrial fibrillation. So atrial fibrillation is where the upper chambers of the heart beat a little more irregularly than they should. Typically people who have atrial fibrillation, or a-fib, are at higher risk for stroke because blood clots can form in the upper chambers of the heart and go up to the brain. And there’s been some evidence from randomized trials with fish oil recently that perhaps there’s an increased risk for atrial fibrillation. So this is a very topical question. So these investigators looked at it in the UK Biobank.

Dr. Bill Harris: And it’s maybe not as easy to see. Here’s the overall result. They had about 470,000 people that did not have atrial fibrillation at baseline. So we’re looking at developing new a-fib.   There were fish oil users and fish oil non-users. What they found was 6.2% of the people who reported taking fish oil developed a-fib and 5.2% developed it if they were not taking fish oil. That’s a 1% risk, absolute risk difference. But the relative risk is 10% because hazard ratio is 0.1, 1.1. So that means 10% higher risk of developing a-fib if you were on fish oil supplements. And here’s kind of the curves you can see over time that the fish, the reddish line is the people who reported taking supplements and the green is the ones who did not — we won’t worry about the subgroup analysis here. But the point of this one, of course, is that habitual fish oil supplementation was associated with about a 10% increase in the risk of incident atrial fibrillation, regardless of whether you had this genetic predisposition to it or not.

Dr. Bill Harris: This was observed only in individuals that did not have cardiovascular disease at baseline. People who did have CVD at baseline, there was no relationship.

Dr. Kristina Harris Jackson: Yeah.

Dr. Bill Harris: So to wrap up —in the UK Biobank fish oil supplement studies to date,  7 of the 8 studies examining the relationship between supplement use and disease outcome were statistically significant, showing benefit in developing type 2 diabetes, fractures, dementia, primary liver cancer or total mortality. The one exception, as we just said was atrial fibrillation, where risk was about 10% higher in users. And most of these studies did control for other healthy lifestyle factors that might otherwise explain these associations. So I will stop there.

Dr. Kristina Harris Jackson: You’re not supposed to take your diet and health advice from large epidemiology studies, but how do you balance the good news with the a-fib news?

Dr. Bill Harris: When we spoke with  Dr. James O’Keefe, a cardiologist who is part of our FARI,   research team, he said, this probably is right because it increases vagal tone, which is part of why you don’t have fatal heart attacks, why MI death is down.”

Dr. Kristina Harris Jackson: Mm-hmm.

Dr. Bill Harris: But on the other side of that coin is an increased risk for atrial fibrillation. So it’s not surprising. It’s consistent with what other people see. Is it a worry? It’s something that not being a medical doctor or a cardiologist, I can’t give advice on … It’s nobody really wants to have, but if it protects you about against 15 other adverse outcomes, then you take 1% increase in absolute risk. It might be worth the balance.

Dr. Kristina Harris Jackson: Yeah. And there’s maybe more to be learned about dose and fish versus fish oil supplements.

Dr. Bill Harris: Mm-hmm.

Dr. Kristina Harris Jackson: I mean, a lot of the studies we saw were pharmaceutical level.

Dr. Bill Harris: Yeah. They’re high dose. Right.

Dr. Kristina Harris Jackson: … these people are probably not taking that much.

Dr. Bill Harris: Not taking that much. No. So it was small. But we have a study in the Force Consortium that’s looked at blood levels of omega-3 and risk for a-fib and we find a lower risk for a-fib…

Dr. Kristina Harris Jackson: Yeah.

Dr. Bill Harris: …and people that have higher blood levels. So this is not a clear story yet. That’s why we continue to do research on this to try to understand how to balance the benefits and risks of omega-3.

Dr. Kristina Harris Jackson: And so, and hopefully soon we’ll be getting into do blood levels in this population, in this UK Biobank.

Dr. Bill Harris: Yeah. That’s something we want to look at is, you know, about one-fourth of the population, they randomly took blood and we have blood levels. So we’ll be able to see if there’s the same relationship they saw with reported supplement use compared to actual blood levels of omega-3.

Dr. Kristina Harris Jackson: Yeah. Cool.

Dr. Kristina Harris Jackson: Well thank you for sharing.

Dr. Bill Harris: Yes.

Dr. Kristina Harris Jackson: We’ll wrap and we’ll be following up on it, I’m sure.

Dr. Bill Harris: We will. Great.

Dr. Kristina Harris Jackson: Okay. Thanks.

Dr. Bill Harris: Good to be with you. Bye-bye.

Dr. Kristina Harris Jackson: Bye.

These statements have not been evaluated by the Food and Drug Administration. This test is not intended to diagnose, treat, cure, prevent or mitigate any disease. This site does not offer medical advice, and nothing contained herein is intended to establish a doctor/patient relationship. OmegaQuant, LLC is regulated under the Clinical Laboratory improvement Amendments of 1988 (CLIA) and is qualified to perform high complexity clinical testing. The performance characteristics of this test were determined by OmegaQuant, LLC. It has not been cleared or approved by the U.S. Food and Drug Administration.