Three Studies, Three Different Answers on Supplements and the Brain

WEBVTT

00:00.000 --> 00:06.300
Hey everybody, it's Mikki here. You're listening to Mini-Mikkipedia on a Monday.

00:06.300 --> 00:13.760
I hope you had a great weekend. And in fact, last week, cracking week, as they all are,

00:14.320 --> 00:20.240
I got a text in a WhatsApp group that I'm part of just with a couple of my very smart friends.

00:20.700 --> 00:27.960
And one of them put a link to an article that was in the NZ Herald. And the headline reading,

00:27.960 --> 00:34.400
popular supplements could harm the aging brain. And I'm like, oh yeah, that's interesting. I

00:34.400 --> 00:40.200
wonder what's going on here. And she had a note in there saying, is this just healthy user bias

00:40.200 --> 00:46.120
that people who are concerned with their brain health might already have a small amount of

00:46.120 --> 00:50.880
sort of cognitive decline. They start taking supplements and then this sort of reverse

00:50.880 --> 00:56.540
association relationship occurs. And I'm like, I'd actually love to dig a little bit deeper

00:56.540 --> 01:01.140
into the article, see what it's saying. And then where are they even getting their information

01:01.140 --> 01:06.960
from? So the article itself was originally a Washington Post article that was published

01:06.960 --> 01:14.580
last week as well. And it was one article discussing vitamin D, omega-3 and glucosamine.

01:15.100 --> 01:21.260
However, these come from three different recent studies. So it wasn't just one study.

01:21.880 --> 01:26.180
And the caveat to all of this that I'm about to go into,

01:26.540 --> 01:31.380
that even though the original article acknowledges that the findings are preliminary,

01:32.220 --> 01:38.500
mostly observational, sometimes contradictory, and does not establish that these supplements

01:38.500 --> 01:44.640
cause cognitive decline. The thing with the headline is it's such clickbait for so many people

01:44.640 --> 01:50.060
who won't really read the finer details and will just look at some of the commentary that

01:50.060 --> 01:55.360
article provided. And of course, I will put the article in the show notes for you to have a look

01:55.360 --> 01:56.180
at yourself.

01:56.540 --> 02:03.140
Because as per usual, that headline asks a much stronger question than the evidence can currently

02:03.140 --> 02:11.080
answer. But humans being humans, we're going to think that it already has. So before getting into

02:11.080 --> 02:17.640
the individual supplements and studies, just sort of clarify what I mean by observational study

02:17.640 --> 02:23.720
versus a randomized controlled trial, which is the gold standard in this type of research when

02:23.720 --> 02:26.020
we're looking at one particular intervention.

02:26.540 --> 02:32.820
So an observational study is, as the name suggests, the researchers observe what people

02:32.820 --> 02:39.120
already do. So people taking a supplement can differ from non-users of supplements in dozens

02:39.120 --> 02:45.600
of ways. And this is something called the healthy user bias, like people who are interested in

02:45.600 --> 02:51.520
taking supplements are probably also interested in being active, in monitoring their drinking,

02:51.520 --> 02:55.920
and not smoking, and watching their sleep, and all of those sorts of things, which,

02:55.920 --> 03:02.680
of course, researchers can statistically adjust in their analyses of whatever results they find.

03:03.060 --> 03:09.220
But you cannot adjust away that healthy user lifestyle. And you also can't adjust away for

03:09.220 --> 03:15.980
things that aren't measured. So there'll be maybe these other underlying factors related to

03:15.980 --> 03:20.080
supplement use that we aren't even aware of. So we don't know how to adjust them.

03:20.080 --> 03:25.840
So this associations that are found in observational studies, as I'll go on,

03:25.920 --> 03:32.820
to explain, they don't necessarily, or they certainly don't, suggest causation. But often

03:32.820 --> 03:38.920
that's where our brain goes. And often that's where newspaper headlines tend to lead us.

03:39.520 --> 03:44.920
And a really great way to think about an observational study is that they set the scene,

03:44.920 --> 03:52.600
or they are hypothesis producing, in that the results of an association study should then be

03:52.600 --> 03:55.900
taken into that randomized controlled environment to test the results. And that's where the

03:55.900 --> 04:00.520
data comes from. And that's where the data comes from. And that is, in fact, a very good contrast,

04:00.680 --> 04:06.820
that randomized controlled trial, to an observational study. So researchers randomly

04:06.820 --> 04:13.580
assign a particular supplement or intervention to a group that there are no differences,

04:13.860 --> 04:19.200
no indifferences between the different groups in the study. So all groups, all people within the

04:19.200 --> 04:23.660
groups, there are sort of selection criteria for which you would be involved in a study.

04:23.660 --> 04:25.880
And then everyone is randomized to,

04:25.900 --> 04:32.040
a particular group. And because of that randomization process, no one is, as at that

04:32.040 --> 04:36.980
group level, no one is different from one group to the next. And it means that these groups are

04:36.980 --> 04:45.060
much more comparable. So if observational studies suggest harm, but a randomized controlled trial

04:45.060 --> 04:52.480
on that particular observation wasn't able to reproduce the harm or reproduce the finding,

04:52.480 --> 04:55.260
we have to be quite cautious about,

04:55.900 --> 05:01.580
including that the observational research was correct. Because literally that randomized

05:01.580 --> 05:06.900
controlled trial was designed to test the mechanism that the observational trial sort

05:06.900 --> 05:13.440
of brought up. Anyway, and that really is, has to be an underlying theme as I go through some of what

05:13.440 --> 05:18.800
this article was suggesting, and then actually what the studies associated with the article

05:18.800 --> 05:24.840
actually found. So let's think about vitamin D, for example, and this is a really important

05:24.840 --> 05:25.620
nutrient.

05:25.900 --> 05:31.480
Some people call it a hormone, and it is involved in many, many different pathways in the body. And

05:31.480 --> 05:36.320
of course, we talk about vitamin D a lot over here in New Zealand, particularly in winter,

05:36.320 --> 05:43.260
because we synthesize the majority of our vitamin D from the sun. But for many months of the year,

05:43.420 --> 05:49.300
the sun's UVB rays aren't strong enough for our body to synthesize vitamin D. And therefore,

05:49.300 --> 05:55.740
we need to take it as a supplement. Now, in the British Medical Journal Open study in

05:55.740 --> 06:02.820
2026, a database from the US Health and Retirement Study was used to look at vitamin D and cognitive

06:02.820 --> 06:11.160
decline. So there were over 5,000 adults, their average age was 67 and a half years. 40% of those

06:11.160 --> 06:18.680
users took vitamin D, and they followed these people for six years. And they assessed the

06:18.680 --> 06:25.140
population for global cognition, memory, and executive function. And this was not a randomized

06:25.140 --> 06:25.720
vitamin D study. It was a randomized vitamin D study. And they assessed the population for global

06:25.720 --> 06:26.260
cognition. And they assessed the population for global cognition. And they assessed the population

06:26.260 --> 06:26.480
for global cognition. And they assessed the population for global cognition. And so a lot

06:26.480 --> 06:27.180
of the countries were not following the data. And they assessed the population for global

06:27.180 --> 06:27.620
cognition. And they assessed the population for global murmurs, and calculates their average

06:27.740 --> 06:31.660
People simply reported whether they took vitamin D or not, which is how they came up with the number

06:31.660 --> 06:36.680
of people that actually took vitamin D. And what they found across that six years was that

06:36.680 --> 06:45.220
vitamin D users had a statistically faster decline in global cognition and executive function. Memory

06:45.220 --> 06:50.740
decline was not statistically significantly different. So the association appeared primarily

06:50.740 --> 06:55.700
in people who started with what the researchers classified as adequate vitamin D status,

06:55.700 --> 07:02.500
rather than those that were deficient or insufficient so let's have a look at what

07:02.500 --> 07:11.520
they actually found so global cognition was 0.052 points a year lower in vitamin d users

07:11.520 --> 07:19.340
and executive function was 0.021 points per year lower and to put that effect size into context

07:19.340 --> 07:27.000
the global cognition scale went up to 27 so the excess decline associated with supplement use

07:27.000 --> 07:37.740
across those six years was essentially 0.31 points so 0.31 out of a 27 point scale so whilst

07:37.740 --> 07:44.480
it was statistically significant can you really tell me that one third of a point across 27 points

07:44.480 --> 07:49.100
is clinically important i don't know i would probably argue not

07:49.100 --> 07:49.320
you

07:49.340 --> 07:54.120
and you've got to remember that sometimes a real strength of a study is the fact that they are

07:54.120 --> 07:59.640
looking at thousands of participants right so that is you know obviously a strength however

07:59.640 --> 08:04.880
the reverse is actually true as well is that when you get large databases like that you can detect

08:04.880 --> 08:09.780
very small differences and they can be statistically significant even if they're

08:09.780 --> 08:14.600
not clinically important and i think that's a really worthwhile thing to sort of bear in mind

08:14.600 --> 08:19.100
there were some big limitations to the study so they did not know the actual

08:19.100 --> 08:19.320
actual value of the study and they didn't know the actual value of the study and they didn't

08:19.340 --> 08:21.640
know it to a great extent also because of the fact that they were normally taking vitamin d

08:21.640 --> 08:27.300
dose in sufficient detail so you could have been taking 400 international units or iu you could be

08:27.300 --> 08:33.760
taking 2 000 iu or 5 000 iu and all of these people were sort of lumped into that broad category of

08:33.760 --> 08:40.460
vitamin d user you can see that it was such a vast difference between the dosing and also of course

08:40.460 --> 08:46.340
as with a lot of nutrition supplement use was self-reported in addition to that vitamin d users

08:46.340 --> 08:49.100
were already actually different from the non-users so vitamin d users were actually the ones that were

08:49.100 --> 08:54.520
For example, their baseline global cognitive scores were slightly higher.

08:54.800 --> 08:59.180
They were 16.3 out of that 27 versus 15.8.

08:59.500 --> 09:02.160
So residual confounding also remains possible.

09:02.760 --> 09:09.800
And when vitamin D users were compared specifically with people taking neither a vitamin D or

09:09.800 --> 09:19.280
a multivitamin, the difference fell to 0.044 points per year, which was no longer statistically

09:19.280 --> 09:19.980
significant.

09:20.580 --> 09:27.000
So on observational level, there was a slight difference in that global cognition and that

09:27.000 --> 09:34.200
executive functioning by 0.3 of a point across six years, but that was pretty much it.

09:34.500 --> 09:36.500
And this was just an observational trial.

09:36.500 --> 09:39.340
So what actually happens when we randomize vitamin D?

09:39.340 --> 09:42.820
And I think this is a really important point and an important comparison.

09:43.260 --> 09:51.120
There was a study called VITAL, and it randomized older adults, similar age, 60 plus, to 2,000

09:51.120 --> 09:56.180
international units of vitamin D3 a day versus taking nothing at all.

09:56.640 --> 10:03.660
And there was a cognitive sub-study within that that had 4,000 adults who were, as I

10:03.660 --> 10:08.880
said, aged over 60, and they found no significant difference in cognitive decline between these

10:08.880 --> 10:09.300
groups.

10:09.860 --> 10:13.140
And there was no interaction with baseline vitamin D status.

10:13.320 --> 10:14.600
So that's interesting as well.

10:14.720 --> 10:20.460
Like in the previous observational study, those people taking vitamin D tended to have

10:20.460 --> 10:24.800
an adequate status, whereas here, there was no sort of interaction with that vitamin D

10:24.800 --> 10:25.200
status.

10:25.860 --> 10:31.220
So the major take-home probably is that there isn't good evidence that taking vitamin D

10:31.220 --> 10:37.460
when you already have adequate vitamin D stores prevents cognitive decline.

10:38.260 --> 10:39.220
And this new...

10:39.340 --> 10:44.760
This new observational study, as highlighted by this newspaper report, it also doesn't

10:44.760 --> 10:49.160
establish that ordinary vitamin D supplementation causes cognitive decline.

10:49.520 --> 10:52.500
Yet that is sort of the message that's pervading that article.

10:53.160 --> 10:57.380
So the take-home for me that I think all of us should be thinking about is we need to

10:57.380 --> 11:03.420
correct a deficiency, we should aim for adequacy, and also don't automatically assume that pushing

11:03.420 --> 11:06.120
a nutrient higher and higher creates benefit.

11:06.640 --> 11:09.320
But if you are taking vitamin D, you don't really need to be worried about that.

11:09.340 --> 11:10.960
You don't need to be worried about your cognitive function.

11:11.420 --> 11:16.780
Now, this one really surprised me, the next supplement that they looked at, and that was

11:16.780 --> 11:18.580
omega-3 or fish oil.

11:19.040 --> 11:23.880
So there was a 2026 paper in the Journal of Prevention of Alzheimer's Disease, and they

11:23.880 --> 11:29.340
used the data from the Alzheimer's Disease Neuroimaging Initiative, or ADNI.

11:30.000 --> 11:34.360
And so that's an important context because this wasn't just a random sample of healthy

11:34.360 --> 11:35.200
older people.

11:36.100 --> 11:39.200
ADNI specifically follows people across that cognitive system.

11:39.340 --> 11:45.380
So we're looking at cognitively normal, subjective memory concerns, mild cognitive

11:45.380 --> 11:48.220
impairment, right through to Alzheimer's disease.

11:48.800 --> 11:59.040
And researchers compared 273 omega-3 users with 546 matched non-users, and they followed

11:59.040 --> 12:02.280
them for around or a medium of about five years.

12:02.820 --> 12:08.780
And what they found was that omega-3 use was associated with faster decline across three

12:08.780 --> 12:09.320
cognitive systems.

12:09.340 --> 12:18.660
And these are all indexes or questionnaires looking at cognitive function, MMSE, ADAS,

12:18.920 --> 12:21.780
COG-13, and CDR-SB.

12:22.180 --> 12:28.620
And they found that in the MMSE, or mini mental state examination, that declined by an additional

12:28.620 --> 12:33.740
0.266 points per year amongst omega-3 users.

12:33.740 --> 12:39.000
The ADAS, COG-13, or Alzheimer's Disease Assessment Scale.

12:39.340 --> 12:44.920
That worsened by 0.823 points per year.

12:45.420 --> 12:55.920
And the CDR-SB, or Clinical Dementia Rating Sum of Boxes Scale, that increased by 0.205

12:55.920 --> 12:57.540
points per year faster.

12:58.020 --> 13:03.560
So different to vitamin D, this isn't just some sort of numerical difference.

13:03.560 --> 13:07.560
There is a signal here, and it is worth investigating.

13:08.160 --> 13:08.960
But it is still...

13:09.340 --> 13:10.960
An observational signal.

13:11.060 --> 13:12.500
And I think that's really important to remember.

13:12.580 --> 13:14.580
This is still just observing individuals.

13:15.060 --> 13:22.900
And what they were able to do was use an imaging technology to examine the metabolic activity

13:22.900 --> 13:24.240
of the brain.

13:24.780 --> 13:30.060
And so what they used was something called FDG PET, P-E-T.

13:30.360 --> 13:32.680
And I'm probably going to get this wrong, but I'll give it a go.

13:34.920 --> 13:37.360
Fluorodeoxyglucose positron emission topography.

13:37.860 --> 13:39.320
That was actually easier than I thought.

13:39.340 --> 13:42.380
So fluorodeoxyglucose PET.

13:43.200 --> 13:46.820
And essentially what they do is they inject a radioactive glucose-like tracer.

13:47.180 --> 13:51.800
They see how much glucose that the different brain regions actually take up.

13:52.020 --> 13:55.340
And lower uptake can indicate lower metabolic activity.

13:56.540 --> 14:03.340
And researchers found greater FDG hypometabolism in Alzheimer's vulnerable brain regions among

14:04.220 --> 14:05.400
those omega-3 users.

14:06.060 --> 14:08.240
So that is really interesting.

14:08.240 --> 14:09.240
And they said...

14:09.340 --> 14:14.860
So they thought, looking, using another statistical analysis, that this could explain anywhere

14:14.860 --> 14:22.120
between 19 and 41% of the relationships that we found in those scales that I just walked

14:22.120 --> 14:22.680
you through.

14:23.060 --> 14:29.220
Now, they didn't find any amyloid or tau or gray matter atrophy, which could explain that

14:29.220 --> 14:29.780
association.

14:29.960 --> 14:34.180
And that's worth mentioning because these are conditions which are definitely associated

14:34.180 --> 14:36.760
with that sort of progression to Alzheimer's.

14:36.760 --> 14:38.460
But I guess the thing is...

14:38.460 --> 14:44.160
Even with the FDG PET results, is that this just measures glucose metabolism.

14:44.160 --> 14:50.340
It does not directly measure neuronal death, mitochondrial damage, lipid peroxidation,

14:50.800 --> 14:56.780
DHA oxidation, DHA is a type of an omega-3, and synaptic destruction.

14:57.480 --> 15:00.160
So it is worth noting that as well.

15:00.720 --> 15:06.720
And I guess some of the sort of key criticisms and objections to the study that I've seen

15:06.720 --> 15:08.160
online since it's come...

15:08.460 --> 15:11.080
Since it was published is...

15:11.080 --> 15:14.160
Well, the biggest one is that the omega-3 intake wasn't randomized.

15:14.940 --> 15:24.280
So the matching was principally around factors such as age, sex, APOE4, genetic sort of presence,

15:24.640 --> 15:26.460
and also, of course, diagnosis.

15:26.960 --> 15:30.160
So this wasn't a randomized controlled trial.

15:30.420 --> 15:38.440
Again, it was just an associative trial looking at people that were already in that age,

15:38.780 --> 15:46.080
And so even when you match participants according to different age and sex,

15:46.580 --> 15:50.800
and even diagnosis, you can't account perfectly for things like

15:50.800 --> 15:54.820
why somebody started taking a fish oil, their overall diet health,

15:55.220 --> 16:00.240
their health consciousness, cardiovascular disease, supplement use in general,

16:00.680 --> 16:07.000
any sort of other cognitive concerns, socioeconomic factors, exercise wasn't accounted for,

16:07.000 --> 16:08.000
overall metabolic...

16:08.460 --> 16:13.280
Health and their ability to utilize glucose, and other unmeasured differences.

16:14.080 --> 16:19.660
And when I said that, you know, the study has been sort of challenged by other scientists,

16:20.100 --> 16:25.340
there were letters that were published subsequently raising concerns about exposure classification.

16:25.940 --> 16:28.080
You know, how long had they been taking the fish oil?

16:28.640 --> 16:30.560
Residual confounding from other factors.

16:30.560 --> 16:37.780
The fact that the FDG hypometabolism was interpreted as synaptic dysfunction,

16:37.780 --> 16:43.060
not adequately accounting for omega-3 status, or the formulation of the omega-3,

16:43.060 --> 16:48.420
how much EPA, how much DHA that these omega-3s are contained.

16:48.420 --> 16:53.460
And also, this is super important, the inconsistency with previous evidence

16:53.460 --> 16:56.660
showing that omega-3 was beneficial for the brain.

16:56.660 --> 17:02.220
And then when we compare it to a randomized trial, again vital,

17:02.220 --> 17:07.400
vital actually randomized older adults to receive one gram of marine omega-3,

17:07.400 --> 17:14.780
omega-3 a day, that provided 460 milligrams of EPA, 380 milligrams of DHA.

17:14.780 --> 17:16.960
And that was versus a placebo.

17:16.960 --> 17:22.840
And they had more than 4,200 older adults included in their cognitive analyses.

17:22.840 --> 17:26.660
This was for around two to three years, and they found no significant difference in cognitive

17:26.660 --> 17:28.000
decline.

17:28.000 --> 17:33.540
And interestingly, they found no benefit from taking the omega-3 with cognitive change.

17:33.540 --> 17:37.020
And in fact, vital found that fish oil really didn't change.

17:37.020 --> 17:39.080
It didn't change cognition.

17:39.080 --> 17:45.380
So it did not demonstrate that it improved cognition, but importantly, it does not demonstrate

17:45.380 --> 17:51.520
substantial cognitive harm suggested by that observational ADNI study.

17:51.520 --> 17:53.240
And that is a big distinction.

17:53.240 --> 17:58.360
So if you have, I guess, these two studies sitting together, and you're wondering where

17:58.360 --> 18:03.520
the sort of strength of evidence is, it has to be in that randomized control trial where

18:03.520 --> 18:06.620
they gave them set doses of omega-3s.

18:06.620 --> 18:09.460
Rather than what we see in observational studies.

18:09.460 --> 18:17.120
We know that omega-3 is so helpful for inflammation, cardiovascular disease, improving lipid profiles,

18:17.120 --> 18:22.880
vascular function, endothelial function, blood pressure, so much.

18:22.880 --> 18:30.120
And having dietary omega-3 rich foods, salmon, mackerel, sardines, other fatty fish, super

18:30.120 --> 18:31.120
important.

18:31.120 --> 18:33.620
Or taking a supplement also can be super beneficial.

18:33.620 --> 18:35.680
And I certainly take an omega-3 supplement.

18:35.680 --> 18:36.480
Pillar performance.

18:36.480 --> 18:39.420
Not that it's an ad, but that's what I take.

18:39.420 --> 18:43.880
What I think we know right now is that taking an omega-3 supplement purely because you think

18:43.880 --> 18:48.080
it will prevent dementia, that's just not that strongly supported right now.

18:48.080 --> 18:53.800
But one observational analysis is nowhere near enough to tell someone taking fish oil

18:53.800 --> 18:55.820
that it's damaging their brain.

18:55.820 --> 19:01.500
And one last supplement that they looked at in the article was glucosamine.

19:01.500 --> 19:05.420
And in fact, I am very familiar with glucosamine.

19:05.420 --> 19:11.040
It was sort of all the rage maybe 15 to 20 years ago in relation to joint health and

19:11.040 --> 19:17.820
helping sort of lubricate in and around the joints, knee osteoarthritis, hip osteoarthritis,

19:17.820 --> 19:19.280
along with chondroitin.

19:19.280 --> 19:26.540
Now this one, in this particular study, which appeared in Nature Metabolism 2026, looked

19:26.540 --> 19:30.780
at glucosamine and its role in Alzheimer's disease.

19:30.780 --> 19:33.460
And so this wasn't an observational supplement study, actually.

19:33.460 --> 19:34.460
The reason it wasn't an observational supplement study, actually, is that it was a study that

19:34.460 --> 19:35.420
looked at glucosamine and its role in Alzheimer's disease.

19:35.420 --> 19:41.300
And the researchers, elegantly really, they combined post-mortem human brains, Alzheimer's

19:41.300 --> 19:48.880
mouse models, experimental manipulation of glycosylation, and human electronic health

19:48.880 --> 19:54.600
records to look at all of these different mechanisms and the impact of glucosamine on

19:54.600 --> 20:01.360
that hyperglycosylation process that may drive Alzheimer's disease.

20:01.360 --> 20:04.160
So glycosylation, what even is it?

20:04.160 --> 20:09.000
Glycols attach carbohydrate structures called glycans to proteins and other molecules.

20:09.000 --> 20:12.280
And that's just like normal biology.

20:12.280 --> 20:17.640
Researchers have found that excess glycan production, hyperglycosylation, has been found

20:17.640 --> 20:19.800
in Alzheimer's brains.

20:19.800 --> 20:25.680
And when researchers reduced glycan synthesis in Alzheimer's mice, cognitive performance

20:25.680 --> 20:26.940
improved.

20:26.940 --> 20:31.500
And when they gave glucosamine, cognitive impairment worsened.

20:31.500 --> 20:34.040
So this is looking at that preclinical diagnosis.

20:34.040 --> 20:40.100
This is a preclinical trial in mice showing that glucosamine worsened cognitive impairment.

20:40.100 --> 20:42.240
But what happens in humans?

20:42.240 --> 20:45.340
So researchers then looked at electronic health records.

20:45.340 --> 20:51.980
They had a sample of 24,000 or over 24,000 people with Alzheimer's-related dementias.

20:51.980 --> 20:57.080
And they had 41,000, close to 42,000 people with mild cognitive impairment.

20:57.080 --> 21:01.820
And they followed these people up, or they had records for them for approximately five

21:01.820 --> 21:02.820
years.

21:02.820 --> 21:06.920
And they found that the glycosamine use was associated with approximately 25% greater

21:06.920 --> 21:12.400
progression from mild cognitive impairment to Alzheimer's-related dementia.

21:12.400 --> 21:17.620
And among people who already had Alzheimer's or a related dementia, there was an increased

21:17.620 --> 21:20.500
mortality risk by 25%.

21:20.500 --> 21:23.400
That sounds certainly concerning, right?

21:23.400 --> 21:28.120
But I guess I think the real thing I want you to remember is that you don't want to

21:28.120 --> 21:30.880
extrapolate it to everybody taking glucosamine.

21:30.880 --> 21:31.880
The human component is still the real thing.

21:31.880 --> 21:32.600
So that's the first thing.

21:32.600 --> 21:33.600
The second thing I want you to remember is that you don't want to extrapolate it to

21:33.600 --> 21:34.600
the retrospective observational data.

21:34.600 --> 21:38.860
So it does not prove glucosamine caused dementia progression.

21:38.860 --> 21:43.840
And critically, the signal was particularly relevant to people who may already have existing

21:43.840 --> 21:48.040
mild cognitive impairment or established neurodegenerative disease.

21:48.040 --> 21:52.460
So if you are one of these healthy 50-year-old people who's taking glucosamine for your

21:52.460 --> 21:58.700
knee, there's no evidence to suggest that glucosamine is increasing your risk of Alzheimer's

21:58.700 --> 21:59.700
disease.

21:59.700 --> 22:01.660
So to sort of tie all this together, what was that herald?

22:01.660 --> 22:04.780
Well, what was that herald article really telling us?

22:04.780 --> 22:08.060
We've got three studies, but they're not showing the same thing.

22:08.060 --> 22:14.680
With the vitamin D, we saw observational association with a very small excess decline that was

22:14.680 --> 22:19.720
found, but there is good randomized controlled evidence that it does not demonstrate harm.

22:19.720 --> 22:24.820
And then with omega-3, there was this interesting observational signal of a faster cognitive

22:24.820 --> 22:30.420
decline and that hypometabolism of glucose in the brain, but that directly conflicts

22:30.420 --> 22:31.460
with a randomized control.

22:31.460 --> 22:41.620
And with glucosamine, this was probably the most interesting signal because there is experimental

22:41.620 --> 22:47.700
biology supporting it, plus observational human evidence, but there is still insufficient

22:47.700 --> 22:51.440
evidence to conclude that glucosamine causes dementia in humans.

22:51.440 --> 22:55.380
And really, I mean, supplements are biologically active.

22:55.380 --> 23:00.460
And this is much broader than just these three supplements, this just means that even though

23:00.460 --> 23:04.860
they're not inert and more isn't always better.

23:04.860 --> 23:10.260
And I guess it's also important to again, state that one observational study showing

23:10.260 --> 23:15.320
harm doesn't suddenly turn an established supplement into a toxin.

23:15.320 --> 23:20.660
You always want to ask who was studied, were they deficient in that particular nutrient?

23:20.660 --> 23:21.900
How old were they?

23:21.900 --> 23:25.200
Did they already have cognitive impairment in this case?

23:25.200 --> 23:26.600
What dose was used?

23:26.600 --> 23:28.620
Was supplementation actually measured properly?

23:28.620 --> 23:29.620
Was the exposure of the supplement actually measured properly?

23:29.620 --> 23:29.820
Was the exposure of the supplement actually measured properly?

23:29.820 --> 23:31.760
Was the exposure randomized?

23:31.760 --> 23:33.360
How long were they exposed?

23:33.360 --> 23:35.060
How big was the effect?

23:35.060 --> 23:40.060
Does the finding here agree with the randomized control trials, which really are gold standard

23:40.060 --> 23:42.980
when we're thinking about these one supplement studies?

23:42.980 --> 23:47.080
Is there a plausible mechanism with which this could even be an issue?

23:47.080 --> 23:49.780
And has the study been replicated?

23:49.780 --> 23:54.060
And that is how you interpret the study rather than the headline that the Herald is giving

23:54.060 --> 23:55.060
you.

23:55.060 --> 23:59.220
So vitamin D, you do want to supplement to correct or prevent deficiency where appropriate.

23:59.220 --> 24:02.220
And it is appropriate here in New Zealand, particularly during winter.

24:02.220 --> 24:07.460
There is not convincing evidence that higher vitamin D protects cognition once you're replete,

24:07.460 --> 24:13.280
but neither is there persuasive evidence from these data that usual supplementation damages

24:13.280 --> 24:14.280
the brain.

24:14.280 --> 24:18.360
And with omega-3, fish and seafood remain excellent foods.

24:18.360 --> 24:19.360
Obviously.

24:19.360 --> 24:23.880
Omega-3 supplementation hasn't convincingly prevented cognitive decline in randomized

24:23.880 --> 24:29.220
controlled trials, but current evidence also doesn't justify telling people that fish or

24:29.220 --> 24:30.220
vegetable oil is neurotoxic.

24:30.220 --> 24:35.400
And for what it's worth, it is one of my go-to supplements for a range of different reasons.

24:35.400 --> 24:38.400
And glucosamine, this is particularly interesting.

24:38.400 --> 24:43.040
If you are an individual with mild cognitive impairment or established dementia, you do

24:43.040 --> 24:44.880
want to watch this one.

24:44.880 --> 24:49.700
And I think what we have to wait for is replication of study findings.

24:49.700 --> 24:52.960
And ideally you'd have some intervention data here as well.

24:52.960 --> 24:55.080
So there's mini Mikkipedia for you.

24:55.080 --> 24:58.280
If you saw that article last week and were a little bit panicked by what it told you,

24:58.280 --> 25:02.620
hopefully this just alleviates some of your fears.

25:02.620 --> 25:06.160
And if you didn't see it, I've got the link in the show notes for you as well.

25:06.160 --> 25:07.160
Yeah.

25:07.160 --> 25:08.940
Keen to hear your thoughts.

25:08.940 --> 25:14.560
Hit me up on Instagram threads or X at Mikki Williden Facebook at Mikki Williden nutrition

25:14.560 --> 25:16.980
or head to my website, Mikki Williden.com.

25:16.980 --> 25:21.200
I've got this recipe portal with over a thousand recipes.

25:21.200 --> 25:22.560
It's 12 bucks a month.

25:22.560 --> 25:26.660
I'm constantly adding to it and they are macro friendly.

25:26.660 --> 25:27.660
They are carb appropriate.

25:27.660 --> 25:29.420
They are higher protein.

25:29.420 --> 25:30.680
They are higher fat.

25:30.680 --> 25:33.760
Some of them in the early days, there's something for everyone there.

25:33.760 --> 25:36.540
If you want some quick, easy, tasty meals to add to your repertoire.

25:36.540 --> 25:37.540
All right, guys, you have the best week.

25:37.540 --> 25:37.580
See you later.