Avoiding Ultra-Processed Foods: A Deep Dive into Health Impacts and Research Insights

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Introducer (00:02):
Welcome to the Dr Journal Club podcast, the show
that goes under the hood ofevidence-based integrative
medicine.
We review recent researcharticles, interview
evidence-based medicine thoughtleaders and discuss the
challenges and opportunities ofintegrating evidence-based and
integrative medicine.
Continue your learning afterthe show at www.

(00:23):
d rjournalclub.
com.
Please bear in mind that thisis for educational and
entertainment purposes o nly.
Talk to your doctor beforemaking any medical decisions,
changes, etc.
Everything we're talking aboutthat's to teach you guys stuff

(00:43):
and have fun.
We are not your doctors.
Also, we would love to answeryour specific questions on
drjournalclubcom.
You can post questions andcomments for specific videos,
but go ahead and email usdirectly at josh at
drjournalclub.
com.
That's josh at drjournalclub.
com.
Send us your listener questionsand we will discuss it on our

(01:05):
pod.

Dr. Joshua Goldenberg (01:32):
Well, as they say, the second you're born
, you start dying.
That's right.
So this is just a way to kindof speed it up a little bit
there.
Yeah, it's like come on, comeon, we got clock is ticking here
.
It's like come on, come on, wegot clock is ticking here, let's
get a little bit faster.
What are some delicious thingsthat I can get in my belly to?
So, speaking of which,yesterday we were talking about

(01:57):
how you have to create anenvironment to minimize like
triggers for like eating garbage, in this case, highly processed
foods.
No, what's that?

Introducer (02:04):
No, not garbage, not garbage, not garbage, just
calorically, calorically densefood okay, and um, I noted, and
I think you, you had a similaridea, which was that essentially
, yeah, you just have to be inone one environment.
You have to, like, control yourenvironment, type of thing.
Anyway, um, I went up to get acup of coffee and I immediately

(02:24):
grabbed like a handful of trailmix, ultra-processed trail mix
and it's just like, as I waseating it, I was like, oh, this
is what we were talking about.
I should have kept.

But is it ultra-processed?
Is it like the planters withM&Ms and whatnot?

Introducer (02:40):
Oh it's garbage.
Oh it's garbage.
It just got back from it thekirkland?

is it the kirkland?

Introducer (02:46):
like yeah, nice oh it's yeah, it's costco garb,
like it's like everything andit's like mass.
I told theo I was like this isa hell, like this is garbage.
I didn't say garbage.
This is junk food masqueradingas health food.
It's like trail mix and it'sjust straight up, highly
processed, like reese Pieces,essentially.

When you see the serving sizes on that
and like you actually weigh itout, it's like why even freaking
bother?

Introducer (03:13):
Two tablespoons.
I did.
I looked it up because I feltguilty and I had to input it in
my log and it's not quite twotablespoons.
I think it's three tablespoons,which is essentially like three
of these little Reese's Piecescups, which is, of course you
have more of those and that'sbasically your breakfast for the
day.

What I've really enjoyed is now a lot of
the ice cream companies will doper serving and then per pint,
uh-huh, oh yeah.
Yeah.
For the nutritional labels,because we all know the pint of
ice cream is the actual servingsize.
Like no one's like when thepint says it serves four people,
you're like yeah, no, no, yeah,it's like.

Introducer (03:53):
It's like having two tablespoons worth of ice cream.
Yeah, no, even like bags ofchips.
Now, it's like you know servingsize and like, oh, if you
actually ate the bag of chips itwould be, you know, xyz if you
ate the bag of chips and youactually realize it, you're like
all right, I'm gonna go cry ina hole now yeah, yeah, I was
very tense.
I just I had finished a run andI was starving and I was like at

(04:15):
a gas station.
I was like, oh my gosh, I'mgonna get these cheetos highly
processed, ultra processedterrible, it's gonna kill you
which we're gonna talk about ina second.
Um, but I was like, okay, justkind of curious, like how many
calories are we dealing withhere?
And it was like, oh, like okay,500 calories, like per serving.
And there was like, basically,when you do the math, there's
essentially like 1000 caloriesin this tiny little bag.

(04:35):
And I'm just like, oh, dear God, you literally just adjusted
the serving size so that thecalorie count would be
equivalent for a snack, which isjust this absurd.
Like what is?
How do you determine a servingsize?
It must literally be well, asnack is 150 calories.
Therefore, when you have twofingernails width worth of this

(04:58):
snack, we're going to call thata serving size.
Anyway, you know what this isspeaking to.
Yeah, this is speaking to ourability to do tangents, which we
just got called out on lastweek.
We got to hold our stufftogether here, man.
We got to stay on task.

Hey, if you're going to call out our
tangents and you have a problem,you can just go listen to a
different podcast or you cancreate your own.
How about that?

Introducer (05:21):
Yeah, I was like.
So we're speaking, of course,of my father, who is our
greatest fan and also ourfiercest critic.

Also I'd like to know how you're the one
who's always going on tangentslike actually I know I did have
something to say that's relevantto like our podcast today, but
now all of your tangents haveactually made me forget it.
Yeah, well I've just you know oh, now I remember what it was.
Now I remember if you actuallythink about it, right, right, if
you think about it we used tojust have three, three meals a

(05:51):
day, or like two meals a day,breakfast and dinner right, like
breakfast.
It would be huge.
You'd go work the fields, youcome back dinner and then you're
exhausted, you go to bed okay,for the sake of argument, I will
agree and not ask for citations.

Introducer (06:03):
Don't know that that's completely true, but
let's just roll with it for now.

And then, you know, then you also have
like breakfast, lunch and dinner.
But, like, as we became moremodern, we then had this whole
like snacking industry and we'vekind of come up with our like
in our minds that, oh, I need asnack, like to bridge me over.
And then, or you see the snack,you know, we have this idea

(06:28):
like, oh, it's, it's just asnack, like it's just a little
little tiny thing, yeah, Ideserve it, but they're like
these gigantic calorie bombs.
So a snack went from hey, justlike, go munch on an apple or a
couple carrot sticks, to hey,let's go have you know, two
Reese's Pieces, king-size cups,which are, you know, ends up
being like 400 thousand caloriein a bag nuts yeah, and then

(06:53):
that's, and then that's.
That's how it becomes a problem.
And then so, like, our entirelike expectation is like what a
what a snack is changes from youknow, we, we have completely
lost the idea as to like howmuch a serving size of a snack
is.
So like a handful of nuts, evenif they're unprocessed.

Introducer (07:12):
Yeah, I know, ever count out the cashews in a
handful.
It's like seven or eight andyou're over.

You know, the actual handful is like why
even bother?
It's like four nuts is likethree, is like 300 calories,
like I'm not even kidding, yeah,it's, it's, it's terrible.
That's why everyone was like,oh, I can't lose weight, all I
have is a handful of nuts.
I'm like, okay, stop eatingnuts.
And then next thing you knowtwo weeks later that yeah, it's
like chipmunk fistfuls.
Yeah, because that's what it is.

(07:38):
We eat them like chipmunks.
You have like a whole handfulof nuts.

Introducer (07:42):
It's not to demonize nuts, but but it's like I got a
handful in each cheek and ineach hand, yeah.

And then you know you tell someone, hey,
all right, well, stop eatingnuts.
And then like two weeks laterthey're down 10 pounds.

Introducer (07:56):
It's like okay yeah well, 20 000 calories a day
worth of uh healthy nuts becausenuts are so calorically dense.

yes, they're good're good for you All
, right now we got to go.

Introducer (08:05):
This is six and a half minutes.
We're talking about nuts andchipmunks.
I blame you and I blame you.
I may have started it, but youare continuing it, sir All right
, so end it.
All right, fine, I'll thinkabout it.
All right, so wait, no, here's,we need the transition.
Here's the transition, as wespeak about ultra-processed
foods.
Now we are going to go intothis very interesting BMJ

(08:28):
article about howultra-processed food consumption
will kill you Slowly.
Well, maybe not slowly.
Well, we'll talk about all that.
Okay, do you want to start usoff with any background or do
you want me to jump into allthat jazz?

Yeah, let's actually talk about it for
a little bit.
So I think it's reallyimportant first to like get the
ideas like what is thedefinition of ultra processed
foods and basically theirdefinition that they use in this
paper?
Is ready to eat or heatindustrial formulations made

(09:02):
mostly or entirely fromsubstances derived from foods,
including flavors, colors,texturizers and other additives,
with little, if any, intactwhole food.

Introducer (09:13):
After you read that, did you just cringe inside
you're like, oh my gosh, that'swhat I've been eating not really
.

I'll be honest, I don't really like eat
a lot of processed food.
So, like I don't know, yeah, Ijust don't really eat it.
So I didn't feel bad about itat all.

Introducer (09:29):
Well, but describing it that way, which is such a
fair way to describe it, it'slike wow, this really just is
this industrialized productthat's peripherally related to
food.

That's also kind of dumb because, if
you think about it, the milk isprocessed food, because that's
industrialized.

Introducer (09:47):
Well, I don't know that that would meet these
definitions, made entirely from.

Yeah, but yes, it doesn't meet that
definition, but from the idea ofindustrialized food is what I'm
saying.

Introducer (09:56):
Yeah, so this is different, and so I think that's
their point is that there'sthis idea that there is this
Okay, there's processed food,maybe there's highly processed
food, but we're talking aboutwhat they call ultra processed
food, which they have a veryspecific, and I think we all
kind of know what we mean.
It's the Doritos.

Basically it's a yeah, it's the fancy
name for junk food lownutritional quality and high
energy density.
Um, and in middle incomecountries, ultra processed foods
account for 57 percent of dailyenergy intake among adults and
67 among youths I just aboutfell off my chair when I read

(10:35):
that.

Introducer (10:36):
I couldn't believe it.

That's nuts, I mean I could believe it
if you.

Introducer (10:39):
I mean just like take a look outside I know, but
like more than 50%, two thirdsof all calories, dude, nobody
cooks, nobody cooks.
I mean, this is, this is I meanI was, I was surprised.

Nobody cooks Like look, go, go onto a
college campus Like we're likedowning ramen noodles.

Introducer (10:58):
Well, it does skew lower, that's true, like the
younger, younger people areeating more and more of this.
But anyway, yeah, I didn't know.
I mean, I knew it was a problem, but to two-thirds of your
calories coming fromultra-processed foods, that
seemed a little scary to me.

I mean it's sad, but that's the reality
.
I wasn't like, oh my God, I waslike, yeah, that's probably
correct, sad.
But I mean, you know, it's kindof the state of things, but
anyway, with it's kind of the,the state of of things, but
anyway, with the definition too,they usually disproportionately
contribute added sugars, sodium, saturated fats and trans fats
and refined carbs to the diet,together with low fiber.

(11:31):
So I mean, basically, this islike you're.
You know, just walk down anaisle in the, in the supermarket
, and it's all those bagged likefrozen meals, it's, it's the
healthy, or the quote-unquotehealthy organic food that's
still in a bag and you just popit in a microwave and in two
minutes, there you go, chatamasala, and now you think you're
eating healthy.
But really, if you turn on theback, it's, it's the same

(11:54):
garbage.
Like, at the end of the day,like, stop trying to deem, like
healthify it or whatever, justbecause it says it's organic or
has a label.
If it's taking you two minutesto make and there's a 400
ingredients, it's, it's probablyultra processed.
And then, uh, basically, um,there's a a lot of growing
evidence from various cohortstudies that are showing that

(12:18):
these, these diets that are highin ultra processed foods.
Um, or just eating processedfoods are contributing to a lot
of our chronic comorbidities, soobesity, cardiovascular disease
, diabetes, cancer etc.
Etc.
And there was a recentsystematic review that showed
high ultra-processed foodconsumption was associated with

(12:39):
an increased risk of all-causemortality, cardiovascular
disease, metabolic syndrome,depression and postmenopausal
breast cancer.
But the issue with some ofthese studies is that they lack
long-term follow-up in theirdefinition for that, meaning
they didn't really follow up forlonger than 20 years, and so

(12:59):
really what this study wastrying to do was look at
durations of at least 30 yearsand also using data that was a
little bit stronger.
A lot of the prior cohorts usedweaker data, so they kind of
just wanted to expand upon this,both from a temporal standpoint
and to see if there's astronger association with

(13:22):
improved data.

Introducer (13:24):
Yeah, fair enough.
Okay, so let's jump.
Should we jump into the designthen?

Yeah.
So what they did was they useddata from two really large
prospective cohort studies thatanyone in who or who does a lot
of evidence-based medicine isreally aware of these cohorts,
because just so much data isderived from them and lots and
lots of cohort studies basicallyuse these two data sets.
We have the Nurses' HealthStudy, and that started in 1976

(13:51):
and included basically 120,000female registered nurses between
the age of 30 to 55 from 11states.
And then we have the HealthProfessional Follow-Up Study,
which is essentially the samething but in male health
professionals, and that startedin 1986 between the ages of 40

(14:12):
to 75, from all 50 states.
And basically with both ofthese cohorts it's like every
year essentially they keepfollowing up with these
participants to ask them abouttheir health status, about their
habits, keep following up withthese participants to ask them
about their health status, abouttheir habits, so diet, exercise
, smoking status, marital status, socioeconomic status, et

(14:33):
cetera, et cetera, and just tonsof data is just churned out on
a yearly basis using data fromthese two cohorts.

Introducer (14:42):
Yeah.
So what do you think about it?
I agree, like it's an amazingcohort or cohorts and long, long
follow up period, but just kindof curious.
And then maybe we laugh at thatbecause, like we know, because
we are healthcare professionalsand we know healthcare
professionals, but like thisidea that this is not a normal
sample of humanity, right, likewe're looking now, we're going

(15:04):
to ask this question about theassociation between mortality
and ultra processed foods andand their health habits and
we're having a population ofnurses and doctors.

Well, I would say I mean it was one
limitation that they that theycertainly did address and they
talked about it.
I mean it's a it's a giganticcohort, and I mean some of the
things you have to think aboutis you're working with a cohort
of people who, compared to theaverage, are likely of higher
socioeconomic status and so canafford to do healthier things,

(15:37):
afford more doctor visits,afford more or live in zip codes
that are generally healthierless food, deserts perhaps, less
exposure to chemicals andwhatnot and then you know they
just have a higher education andtypically people who have more
education tend to be healthier.
But I mean, it is a very largesample, so it's somewhat

(16:02):
reflective of the US, not ashocker.
Healthcare professionals alsoaren't very healthy.

Introducer (16:12):
Yeah, that's what I was more curious about.
I think you bring up reallygood points about socioeconomic
piece, because that of coursehas such an impact on health and
mortality.
I was thinking more about, like, their health behaviors.
But then I was thinking, well,maybe they're not all that
different than the standardAmericans.

Ask a healthcare professional how much
sleep they get.

Introducer (16:32):
Yeah, yeah, no, totally.

You know how much sleep and how much
exercise they get, or how manyprofessionals do you know smoke.
So I mean, yeah, obviously it'snot reflective of a global
population, obviously it's notreflective of the general
American population, but it'sdata that we have and data that
we can work with.
And it's a limitation that theauthors addressed and I think

(16:55):
that we all have to just kind ofrecognize too, is that, hey,
this is data from healthcareprofessionals in America.

Introducer (17:02):
Yeah, so totally true, I was just trying to think
it through for a bit.
Though.
It's a great data set thatmakes sense.
It's maybe not completelyanalogous to standard American
population.
However, if you're stilllooking for the associations,
they should still be there and,if anything, I would think

(17:23):
health professionals would bebetter about filling out the
data about what they eat, orfilling out these questionnaires
right Like.
At the very least, you couldmake that assumption perhaps,
and so maybe the accuracy or theprecision of these instruments
might actually be better in thispopulation and we might have a
better approximation of like.
Okay, if we do see anassociation, it might actually

(17:46):
be real.

So I don't know, I'm I am curious,
but I mean, I think there's acouple of things that we also
have to take into account.
Um, when it comes to like longcohort studies, people just suck
at answering the questionnaires.

Introducer (17:57):
That's one thing too yeah, yeah, well, and
questionnaires suck.
That's the other.
That's the other thing I wantto talk about.
Yeah, questionnaires suck, yeah, and questionnaires suck.
That's the other.
Yes, that's the other thing Iwant to talk about.

Yeah questionnaires suck.
Yeah, so questionnaires suck.

Introducer (18:08):
People suck at answering questionnaires and
therefore the data derived fromthose two things is sketchy.

Yeah, it's sketchy, but it's what we
have.
Um, and especially when itcomes to things like food and
exercise, consistently peopleunder report how much they eat
and over report how much theyexercise.
So if anyone's like, oh, Ican't lose weight, I'm only I'm
only eating 1600 calories, I'mlike you're probably eating 25,

(18:35):
like 2500, like you're probablyway over underestimating and in
this case it would be, theywould probably be
underestimating their ultraprocessed food intake.

Introducer (18:46):
Sorry, there'd be.
Yeah, they'd be underestimatingtheir ultra processed food
intake.
Yeah, and so if they see anassociation, you would argue
that would.
That would even be moreimpressive, because if anything,
that should bias towards thenull.

Right.

Introducer (18:59):
OK, interesting, yeah, All right, I was just kind
of curious about that.
And then the other things, justreal quick.
On study population I just wantto touch on was you know, these
questionnaires are notoriousand so they would throw out
questionnaires if, like, therewas implausible calorie intakes,
like you said.
So they tried to have somecontrols on that.
But I guess we're more worriedabout it wasn't obviously off,

(19:20):
like they didn't obviously skipit, or like, no, they didn't
have 100 calories that day, likethat's unrealistic and no, they
didn't have 50,000 caloriesthat day.
That's realistic.
But it's more the ones that arelike, well, maybe they're 30%
off or 40% off, or things thatlike could plausibly be true and
would have a huge impact on thedata, which is what I worry
about.
But again, I guess they try tobalance that across large

(19:42):
population sets.

And then also something that's really
important to highlight is thatthey excluded and this makes
sense because you're trying tofind an association with these
things but they excludedparticipants at baseline if they
already reported a history ofcancer, cardiovascular disease,
diabetes.
I thought that was good Becausewe're trying again to find the
association really with deathfrom those things, so you don't
want them to already have itBecause it's an incidence.

(20:11):
We're trying to find new cases,not existing, which would be
prevalence.

Introducer (20:16):
Excellent, totally agree.
And then also there's thisother interesting benefit of
that, which they talk aboutlater, which is that reverse
causation thing, where it's likeyou get diabetes, you start
eating healthy.
And then if you measure peoplethat you know and you're like,
oh well, healthy eating isassociated with diabetes, well,
it is, but because the diabetescaused them to eat healthy, not

(20:37):
the other way around, right,yeah.
And so by sort of excludingthat from the in from the get go
, and then later with theirsensitivity analyses, they kind
of um, deal with that as well.
But, yeah, perfect, okay, Ithink that's good for me.

on study population, and then when we
looked at the at the food intake, uh, they basically broke it
down into four categories tobegin with, so they have
unprocessed or minimallyprocessed food.
I think that's easy like thinkapple right uh processed
culinary ingredients.
So I kind of think of likedried spices okay processed food

(21:10):
.
So I would, I would considerthat as like butter, right,
because butter comes from milkand you kind of have to break it
down okay and then you haveultra processed food.
So you went from milk to butterto now butter product.

Introducer (21:26):
I don't even know what that would be.

I mean, they they do have like some
examples uh in the um, uh in thein the supplement, uh, so like
for ultra process, like I said,dairy, so uh, let's see what
they have here so like dairybased desserts, so artificially
sweetened yogurt, flavoredyogurt, ice cream, sherbert yeah
yeah, like candy bars, right.

(21:51):
So if you went from cocoa beanto cocoa powder to like raw coke
, I don't know, twix, yeah, yeahto them on twix, yeah, there
you go, boom.
And then it's interestingbecause, like we, have non-dairy
.

Introducer (22:02):
Yeah to the Nutwix.
Yeah, there you go, boom.
And then it's interesting,because we have non-dairy quote
unquote butter for Theo, who'slactose intolerant too, I guess.

That's ultra processed, I'm sorry.

Introducer (22:12):
That would be ultra processed, right.
So that's an interesting thing.
You got your milk to yourbutter.
But it's organic but it'sorganic and healthy because it's
non-dairy.
Yeah, no, that's a good example.

And it's more expensive and it's in the
healthy aisle.

Introducer (22:26):
It has a little leaf on it, which naturally inspired
food.
So, that's a great example ofultra processed foods, now I
want to highlight that.
So I think the buckets makesense, but the way they got this
right, so they sent out thissemi-quantitative food frequency
questionnaire with all thecaveats we just talked about,
every four years.
I just want to underline thisonce more, the data that we're

(22:51):
basing everything we're about totalk about is.
They mailed out this reallycrummy by design questionnaire.
I know it's validated, butthese things have so many
problems every four years.
That's going to give us anassumption about how much, how
truthfully, how much ultraprocessed food is in their diet.
Now the counter argument iswell, josh, if everybody is

(23:13):
measuring it wrong and then youstratify by quartiles and you're
looking at that difference,which is what they're doing,
maybe it balances out.
But then I would say no,because the people that maybe
are having less of that may bemeasuring that differently than
the people that have more thanthat.
Right, like these, populationsmay differ in other ways.

(23:34):
If you're more health conscious, you may eat less ultra
processed foods, but you alsomay be more careful.
You may be more comfortablewith logging right, for example,
and be more accurate in thisreporting.
So I just have so many issues.
I don't have a better solution.
Like I know, you can't have150,000 person study where
you're going in and weighingeveryone's food in like a

(23:56):
science kitchen.
I get that.
And like this is just like alleverything we talk about in
nutrition is based on stuff likethis.
I'm like it's garbage.
It's like, well, we sent out aquestionnaire that's highly
flawed every four years and nowwe're looking for associations.
I just OK.
End of rant Moving on.
Hey, josh.

Yeah, you don't remember.
We ate on on Friday, may 31stof 2019.

Introducer (24:21):
I know it's crazy and then tell me how many and
what the serving size was, andit's just like okay, so, okay, a
continuation of rant.
Um, so actually, I think thiswill be interesting now that we
have because you and I are intothis, we're tracking our food
and stuff and like you've gotthese, these apps now and
they're not quite good enough totake a picture and like, get
serving size, but now with, likeyou just scan your barcode and

(24:45):
you know you have a really goodability to log.
Now, granted, even that isgoing to be flawed, but think
about how much better that isreal time logging compared to
asking someone to remember howthey ate over a four year period
, or or asking them how to eatover a one week period and
assuming that's an approximationof the past four years.
So it could very well be thatwe get better.

(25:07):
So counter.
Yeah.

Slight counter argument to that and
this is like this is again endof tangent Is that if you were
to then ask people to track foodand they're like actually
measuring things out, then thatdoes change their dietary
patterns, because they're goingto.

Introducer (25:26):
Then of course, eat less take healthier choices.

And this goes back to what we were
talking about earlier, where, ifyou actually knew how much a
serving size of trail mix was,you're like you wouldn't eat it.
You'd be like this is this isstupid like there's like three,
three almonds and two m&ms.

Introducer (25:39):
I'm not gonna bother , that's 200 calories yeah, or
you would put it in the garageuntil it's brought back in by
your son or your wife, and then,yeah, that's there's that issue
too right and then, whereas ifyou're just asking kind of
people in general, like what doyou eat, you know you're looking
again, you're looking for thatpattern, you're not looking yeah
, that's a good point for theexact amount, because if you ask

(26:00):
for the exact, people willchange their dietary patterns
and the amount that they'reeating, so you're getting
completely different data.

So that's just something to think about.

Introducer (26:10):
Yeah, okay, yeah, that's a very good point.
No, okay, no, hold on.
Okay, I'm taking you seriouslynow.
So that's a super, super validpoint.
So you're saying these areactually very different measures
, because, while I'm saying youmight get better accuracy,
you're saying okay, yeah, butyou're sacrificing precision for
bias, because you're now goingto change the way that you eat

(26:31):
and, even worse, you're going tochange the way that you eat in
that moment and it's not goingto be reflective of the rest of
the period that's driving thedisease, supposedly.
Yeah, that's a really excellentpoint.
So that's a challenge, right?
So how do you get accuratereporting that isn't influencing
the what you're trying tomeasure?

that's a hard one yeah, the audacity is
not taking me seriously too.
Um, okay, so then these fourcategories of unprocessed uh,
processed, culinary processedfoods and ultra processed the
ultra process were then, um,subdivided into nine categories
from there, which makes sense.

(27:09):
Um, they then divided it tookay, you have ultra processed
foods and then you have ultraprocessed breads and breakfast
foods, fats, condiments andsauces, fats, condiments and
sauces being one category.
Packaged sweet snacks anddesserts.
Sugar, sweetened andartificially sweetened beverages
.
I kind of wish they separatedthat out.

Introducer (27:30):
They did later, I think.

They did later Okay.
Yeah, I think there's asensitivity there uh, ready to
eat slash, heat mixed dishes andthen meat, poultry, seafood
based, ready to eat products,for example processed meat.
Um, I don't think that thatcounts as like ground beef, like
just 100 ground beef.

Introducer (27:48):
I think that that is more of that's, more of like
salami bologna or or chickennuggets like if you've got
littles, like half of their dietis like chicken freaking
nuggets and I'm just like, oh mygosh, what the hell are you
eating, dude?
I know no.
I read this and I'm like weneed to change everything that
we're doing like now you would.

You would not survive at my, at my
apartment.
Dude, I have like, I have likeno.

Introducer (28:14):
I think we should.
We should have, like a, aintervention where we go to
Adam's apartment and learn howto eat properly, or remember how
to eat properly.
All right, Um, okay.
So what else we got?
Oh, the, the.

But I do want to talk about um, I just
have a pop-up on my screenTangent alert.

Introducer (28:39):
It said the Taco Bell brought back the cheesy
crunch gordita, as we'repodcasting, okay, a, that speaks
to your history of searchingand now you're outed, and B,
clearly the microphone isreporting this to whatever it is
Facebook or whatever that isnow serving you ads.

AI.

Introducer (28:54):
Although they're clearly not good enough to know
the context in which we speak.
Okay, did we talk about Ireally want to talk about the
Alternative Healthy Eating Index.
Do you want to touch on that?

Honestly, I don't want to talk about that
because it's just kind of likea measure of your diet quality.
I don't want to get lost in theweeds with it.

Introducer (29:13):
Okay, we won to get lost in the weeds with it.
Okay, we won't get lost, but Ithink this is huge.
I think this is huge as far asa flaw in the study, which I
want to, not a flaw in the study, but a flaw in the
interpretation of the study.

So just super briefly so can we, can we
hold off on that?
Actually, let's, let's wrap upthe food, sure okay and then
they also said that becausealcohol is well studied for
premature death and a distinctfactor, they did not consider
alcohol in the ultra processedfoods.
They kind of had it as its ownseparate category altogether and

(29:43):
then kind of looked at the datawith that in it and without.
Makes sense.
And then they also did the samething with whole grains, because
whole grains are just so wellestablished from a mortality
standpoint that they looked atit from.
If you removed whole grainsfrom ultra processed and then
also put them back in, how thatwould change things.
Ok, you can.

(30:03):
Now you can talk about yourstupid index thing or whatever.

Introducer (30:07):
So one of the other things they measured is this
index of healthy eating, whichis a sort of um, somewhat
simplified you know, how much ofthis food do you eat, how much
of that food do you eat?
But it's going to be reallyimportant because I think and
tell me that as we read throughthis, you tell me if you buy my
hypothesis or not but I thinkthat you could make an argument

(30:28):
that essentially, ultraprocessed eating, ultra
processed foods, is a surrogatemarker for this quality index
and that essentially, what we'remeasuring is the mortality
associated with the quality ofyour food, not necessarily
ultra-processed food.
Now, ultra-processed food ispart of that index, but it's not
all of that index.
So something to consider as wemove forward.

Yeah, I would say again and again, it's
not the food item, it's thedietary pattern component to it.
So I think that if you look atlike eating ultra processed
foods as the food product andthen the AHEI, which is the
Alternative Healthy Eating Index, as your dietary patterns
overall like if you eat 90%whole foods and like you're very

(31:15):
minimally processed and thenyou add in you know some sweets
here and there versus onlyeating processed foods the
dietary patterns is your dietarypattern is mostly healthy.
With a couple you know thingshere and there because you're a
human and you enjoy life.
With a couple you know thingshere and there because you're a

(31:35):
human and you enjoy life.
The other one is a very lowAHEI because you're just eating
I don't know donuts and muffinsall day long.

Introducer (31:41):
Well, I think that's my point, which is, like you
can have people so like they'retrying to show that it's
specifically thisultra-processed food and then
they try to go further, althoughthey weren't super successful
in saying it's this specifictype of ultra-processed food
that's most associated, and whatyou're saying, and I think what
I'm saying, or what I'm saying,and I think what you're saying

(32:02):
is no, it's the pattern ofpeople that eat this type of
food that is more important.
And when we look at thesensitivity analysis, I think
that bears out our position morethan their position.
And you can just imaginesomeone like a health food nut
or like a health food nut, butlike someone who works out a lot
.
You know mostly it's healthyfood, but it also has, like
that's highly processed, likeprotein bar, you know
pre-workout type of thing andpost-workout right, that's ultra

(32:25):
processed food, multipleservings a day, but yet super
health mindset Otherwiseeverything else is excellent
Would probably score very highlyoverall on this healthy eating
index, but would also score highin a bad way on the ultra
process.
So anyway, so just a case studyto kind of think about there.
Okay, so let's.

(32:45):
What else did we want to talkabout?
Oh, you wanted to talk aboutthe missing data points.
Should we touch on that now?

I think in one second we can.
When they looked at theoutcomes, the study
investigators were blinded tothe exposure status when they
were looking at deathcertificates.
So when they were analyzing thedata, they didn't actually know
if people were eating a healthydiet or not.
They were just confirming howthey died and they basically
used death certificates.

(33:13):
Okay, this person died fromcancer, this person died from a
heart attack, etc.
Etc.
And they were just confirmingthat and then looking at the
data.
So not not to skew things,which is important- yeah, that
was good.

Introducer (33:28):
That's detection outcome.
Detection bias issues yeah,very good.

And then some of the covariates that they
looked at body weight, maritalstatus, smoking status, smoking
status, both in pack years anddo you smoke or not, which I
think is important, because ifyou're someone who's like, oh, I
socially smoke when I go out todrink and maybe I have one or
two cigarettes and I do that onon a monthly basis, like you're

(33:54):
gonna have a very low pack yearhistory, but technically you
answer yes.
Versus someone who's you know,I chain smoke two packs at two,
three packs a day is may be alittle different.

Introducer (34:06):
uh, when we're looking at this kind of data,
yeah, well, and they, they pointthat out later that when they
compare their results to otherstudies, that just bucketed it
like that very different results.
And the fact that they wereable to sort of stratify in a
more nuanced way, I think isimportant.
Which.

I thought was actually really really cool
, because usually what I'm usedto seeing is it's as a yes, no.

Introducer (34:25):
Yeah, ever smoker question yeah.

Yeah, so I do think that this was kind of
novel in that standpoint.

Dr. Adam Sadowski (34:31):
Yeah, so I do think that this was kind of
novel in that standpoint.

They also looked at physical activity
status, family history of cancer, cardiovascular disease,
diabetes.
They looked at basically, youknow, are people who are
basically getting annual exams,are they healthier than those
who don't?
And then looking at menopausalstatus and post-menopausal

(34:59):
hormone use for women alcohol,like I said earlier.
And then, as you said earlier,using the AHEI, when we looked
at the statistical analysis,they basically followed people.
I think they started.
They didn't start at the likeofficial start date of the
cohort, they kind of startedlike when when ultra processed
foods came out.

(35:20):
Yeah, yeah, yeah.
And then they follow them, um,until they they had the event or
until they were followed for 30years, whichever came first.
And that was their cutoff pointto to analyze the data, cause
at some point you have to stop.
And then the primary analysiswas basically they pooled
everything together, looking atboth cohorts together, and then

(35:43):
sort of like just ultraprocessed foods as a broad
definition, and then they lookedat the data at the second, at
each cohort on their own.
So they looked at how did thisaffect basically men and how did
this affect women?
And then they also looked atspecific food trends within the
ultra-processed food.

(36:04):
So does processed meat have aneffect that would be different
than processed bread products,et cetera, et cetera.

Introducer (36:20):
Yeah, and one of the sensitivity analyses or yeah, I
think they're calling it asensitivity analysis which is
around that A-H-E-I, thathealthy eating index.
So they stratified everyoneinto these four buckets, these
quartiles, with their healthyeating index, and then looked at
the relationship within eachquartile with the lowest highly
processed food intake versus thehighest.
And that's sort of a way ofcontrolling for the healthy

(36:42):
index.
So people that are super, superhealthy index eaters, you take
the people and say, okay, theseare healthy eaters, but let's
take that group and compare thepeople that have the most ultra
processed food to the least andsee if there's an association
there.
And we take the people with theunhealthiest diet and look at
the association there.
So that's one way to kind ofcontrol for this issue that I'm

(37:02):
worried about, which is likemaybe we're just measuring the
quality of the food and it'snothing to do with the or very
little to do with the ultraprocessed food.
And then the other thing I hadabout all these sensitivity
analyses that you're touching onis did you see anywhere where
they said this was like a priori?
Did they publish their protocolon their statistical plan?

(37:23):
I mean, that's a lot ofsubgroups, a lot of specific
calls here.
I didn't see anything aboutpublishing that ahead of time.
Look, the thing is we don't dothis for money.
This is pro bono and, quitehonestly, the mothership kind of
ekes it out every month or so,right?
So we do this because we careabout this, we think it's

(37:45):
important, we think thatintegrating evidence-based
medicine and integrativemedicine is essential and there
just aren't other resources outthere.
The moment we find somethingthat does it better, we'll
probably drop it.
We're busy folks, but right nowthis is what's out there.
Unfortunately, that's it, andso we're going to keep on
fighting that good fight.
And if you believe in that, ifyou believe in intellectual

(38:07):
honesty in the profession andintegrative medicine and being
an integrative provider andbringing that into the
integrative space, please helpus, and you can help us by
becoming a member on dr journalclub.
If you're in need of continuingeducation credits, take our
nancyak approved courses.
We have ethics courses,pharmacy courses, general
courses.
Interact with us on socialmedia, listen to the podcast,

(38:28):
rate our podcast, tell yourfriends.
These are all ways that you cansort of help support the cause.

And they didn't.
Yeah, there's no likestatistical method that they
said that they use to controlfor all of these additional
testing.
And that's important, becauseif you just test for 100 things,
you're bound to find something,and so just by chance you'll
probably find something bychance.
And for every P statistic thatyou run or you know analysis

(39:00):
that you run, the chance offinding something that's
significant incrementally goesup unless you control for that.
So your P result would have tobe smaller with each analysis
that you ran, like when you'recontrolling for it.
So you would want, like, let'ssay, instead of the typical P
equals 0.05 would bestatistically significant.

(39:23):
When you're controlling formultiple analyses, it then may
be like 0.0025 is thenconsidered statistically
significant.

Introducer (39:32):
Yeah, and I think that what you're talking about
will address the issue of justlike, let's just go fishing and
see if we can find something.
And that is very true.
And I am particularly worriedthat the way they chose to do
things was influenced by apresumed bias of ultra processed
foods are bad.
Not that there's a huge lobbyof people that think that ultra

(39:53):
processed foods are bad, notthat there's like a huge lobby
of people that think that ultraprocessed foods are good, but,
like, if you look at the authorson this paper and where they
come from in their previous work, like, and also just reading
the introduction, is very, veryclear that they believe that
ultra processed foods is bad.
And I just worry about going inwith those sort of intellectual
predispositions, even if theymake sense.

(40:13):
You know, did we say ahead oftime how we were going to do
these sensitivity analyses andsubgroups and statistical
analysis and all the things?
Because, yeah, I don't know.
I just I worry about that sortof thing.

Yeah, and then on the comment that you
were alluding to earlier, ontheir analysis, they carried
forward non-missing values fromthe previous survey to replace
any missing data, which can beproblematic, because you know.
Basically what they're sayingis oh, they missed this answer,

(40:48):
let's just copy and paste fromthe previous four years, which
is kind of assuming that withinfour years nobody changed their
dietary habits.
So perhaps someone actuallywent from, you know, eating
nothing but ultra processed foodto then having some sort of
intervention where they're like,oh, now I'm going to stop that
and only eat healthy.
You know, you know how dietarytrends happen and you know, post

(41:12):
holidayiday people always wantto lose 10 pounds or something
like that, and so they changetheir eating habits.
And so, on this assumption,it's hey, we're assuming that
people are not changing theirhabits and we're just going to
kind of carry forward.
Whether they're alwaysreporting a healthy diet, we're
just going to kind of move thatforward.
And then, if they're alwayseating an ultra-processed diet,

(41:34):
we're moving that forward, whichat first I kind of had an issue
with.
But I also get why they did thatBecause it would be so hard to,
and you don't want them to belike, well, we lost all this
data, so now we can't doanything with it.
So now we can't do anythingwith it.
If you think about it from likea temporal standpoint, it kind

(41:55):
of strengthens the data in a wayof like okay, well, if they
were super ultra processed foodeaters in the past and we're
continuing to move this along,then they're temporally exposed
to this risk factor that we'reinterested in for a longer
standpoint and you would assumethat that would then have, you
know, more of an effect on theresults that we're seeing.

(42:17):
And similarly, then we wouldalso see more of a protective
factor if they're eatinghealthier for longer.

Introducer (42:24):
Yeah, it's a big question.
Like in the methods world, it'shighly debated how we deal with
missing outcome data.
And to your point, you don'twant to just throw everything
out because it's highly debatedhow we deal with missing outcome
data.
And to your point, you don'twant to just throw everything
out because it's importantinformation.
But the question is like theassumptions behind.
So basically what they're doingis like I think they sometimes
call it like last measurementcarried forward in other

(42:45):
instruments, right when,basically just take whatever
they said last time and youassume it's the same.
And you know there are issueswith that, because, A, it could
be wrong, like you're saying,but B, if the reason for the
missing data is connected toeither the thing that the
exposure so ultra food, ultraprocessed foods or the outcome,

(43:06):
like death or disease, then wehave a bias issue, not just a
precision issue.
And so you can imaginescenarios where it's like I just
got diabetes and I'm too guiltyto say how much Twickers
Snickers I eat and so I'm goingto just not answer that one or
whatever, and so you can imagineenvironments where it could be

(43:28):
associated, and then you get abias issue, so anyway.
So it is a big issue.
Traditionally that's whateveryone did, Last measurement
carried forward.
These days, I feel like thereare, you know, people do
multiple imputation, likethere's.
You can do sensitivity analyses, which is how we've dealt with
this issue.
You say, okay, assuming theworst, or assuming worst
plausible, you know, would theresults survive, right?

(43:51):
So I think, anyway, there'sjust different ways to do it.
But I think your points arevalid.
We don't want to just throw outthe data, but I don't know that
less measurement carriedforward is the best way to
handle that.

And two quick tangents on that.
One is this is why, when youwhen you're because stuff like
this is always in the news andit's why you can't just buy the
headline you really got to readthe study and, like, actually be
able to critically evaluate it,to understand there's a lot of
nuance going on, and then beable to you know there's data

(44:22):
and then it's okay.
Well, what do we do with thisdata?
How do we interpret it in thecontext of all these things
biases and study flaws andwhatnot, and best available
evidence that we have.
So you always you kind of haveto tie it all together.
You can't just say, oh see,this, this is in line with my
bias.
Therefore, I'm going to agreewith it.
And it's why, in like in ourprevious episode, I kind of am

(44:45):
critical about things and wantpeople to be critical about
things, so that means thatyou're actually thinking about
things.

Introducer (44:50):
Um, and then the other tangent is well, I forgot
it, so it doesn't matter well, Iagree with tangent one, and
because the the headline here islike ultra processed foods kill
you, right?
So that's the sexy headline andyou're right.
There's a lot and we haven'teven got into the results yet,
and there's a lot of nuance hereyeah, yeah.

So, that being said, let's get into the
results okay, cool, so let'smove.

Introducer (45:12):
I I don't know that we have to perseverate too much
on the results.
I think they were prettysomewhat straightforward, but so
, basically, we're talking abouta median of 34 years of
follow-up, so that's outstanding, right?
So they have, like I don't know, 120, 130, 140,000 people total
studied over three, almostthree and a half decades, with
median levels.
So that is the real advantageof this type of study, as well

(45:35):
as, to your point, a lot ofnuanced data, even with the
caveats that we talked about.
And OK, so I want to kind ofjump right into the results and
then talk about it in line ofthis healthy eating index.
Is there anything you wanted totalk about before we?

we do that um, I do want to kind of
give a just a brief breakdown ofour, our baseline cohort, like
characteristically what we'relooking at.
So sure, remember, um, onecohort included 122 000 people,
the other one included, uh, 51500 um.
In total I think there was what48 000,000 deaths or something.

(46:13):
Yeah, 48,000 deaths Um, total.
Uh, of that we had 13 and ahalf from cancer, 11 and a half
from cardiovascular disease,4,000 from respiratory diseases,
so things COPD um, and then6,000 from neurodegenerative
diseases, um, so we started offwith what did I say like 160 and

(46:33):
50 died for easy numbers.
And then they said thatparticipants with higher
ultra-processed food consumptionwere younger, more physically
inactive and more likely tosmoke and had higher BMI, lower
consumption of alcohol, wholefruits and vegetables, whole

(46:54):
grains and a lower AHEI score.
But then when we actually lookat the data, when we're looking
at our quartiles, so quartileone being less ultra-processed
food consumption, so quartileone is your healthier diet, if
you will.
And then quartile four is yourmore processed, is your least

(47:17):
healthy?
Yeah, so most processed foods.
When we actually look at theage difference there, that's
really, it's really like notthat different 66 versus 65 from
a quartile one to quartile twoand that was in the health
professionals follow-up study.
So that was the male study.
In the nurse's health study, 66versus 64.
So really, from a clinicalstandpoint, like, the age

(47:38):
difference there is like ispretty negligible.
The majority of participantswere white 89% in quartile one
versus 92% in quartile four.
From a marital standpoint, 67percent in quartile one versus

(47:59):
68 percent quartile four.
When we look at bmi, 25 versus26 and a half total activity met
hours 33 in in quartile one.
So healthier people justexercising a little bit more and
more vigorously, versus 25 inquartile four.
That one was pretty notable.
Notable, uh, from an alcoholstandpoint.
Or again, it's based on alcoholunits.
So how many grams of alcohol umare are within your your drink,

(48:23):
um, so 13, uh in quartile oneversus nine in quartile four.
Um, so that that was prettydramatic.
You can kind of see that peoplewho are eating worse are
actually drinking less alcohol.
Just something that'sinteresting.
And then from an AHEI score Idon't know what the max score is
, but in people who are eating ahealthier diet they had a 53 on

(48:45):
average versus a 44 in quartilefour, and this is kind of
consistent across both groups,meaning the nurse's health study
and then the healthprofessional's follow-up study.
When we look at those who neversmoked, those who ate healthier
, 55% were never smokers versus45% who were in quartile four.

(49:08):
And then the mean pack years ofsmoking.
Those who did smoke or wereformer smokers in the healthier
dietary pattern standpoint had amean pack year history of nine
years versus 14 and a half inquartile four.

(49:29):
Um, and then otherwise, mostthings otherwise were pretty, uh
, pretty, um, um, equal acrossthe board.
When we looked at mean ultraprocessed food intake based on
servings per week, the people inthe healthier dietary habit
pattern, um, on average, uh, ate22 servings of ultra-processed

(49:53):
food per week versus 56 inquartile four.
So I think that that's prettyremarkable there.
Something to note that was forthe health professionals
follow-up study, so in men andthen for women it was 21
servings versus 50 servings.

Introducer (50:14):
So a nice spread there to look for an association
for sure.
So if you take okay, drum rollplease If you take the lowest
quartile of people overall thathad the lowest amount of ultra
processed foods, which is amedian of about three servings a
day, and you compare those tothe highest quartile, which is

(50:34):
about seven and a half servingsa day of ultra-processed foods,
which is a lot, there was a 4%higher risk of total death.
And that's after multiply,adjusting for all the different
variables that we talked about,like BMI and smoking and things
like that.
So basically we're talkingabout for those, if you compare

(50:56):
those that had the fewestamounts of ultra processed foods
to those with the most, overlike a 35 year period, you're
going to see four extra peopledead per hundred in the highest
ultra processed group than thelowest.
One way to think about thatsort of and what's the
confidence interval on that?
1.01.

(51:17):
So like barely uh barely therestatistically significant, and
1.07 now to as the as the upperbound there.
So it's interesting.
They say it's a strong positiveassociation.
That's speaking to the p value.
But actually the magnitude ofthe effect?
I mean it's a strong positiveassociation, that's speaking to
the p value, but actually themagnitude of the effect I mean
it's quite modest.
Right, we're talking about afour percent difference over

(51:39):
from the highest group to thelowest group over a 35 year
period.
So that's, that's one thingthat they found.
So that's their main kind oftake on here increased risk of
death, and we're talking about afour percent difference over
five years and if we look atsort of like the baseline risk,
sorry, 35 years, yeah, if welook at the baseline risk.

Uh, they did an incidence rate per 100
000 person years, which whichmakes sense because you're
having these high numbers, whichis essentially the same thing
as like okay.
Well, if you were to then lookat the numbers, like, you can
still convert it to a percentagestandpoint.
So in quartile one, so thosewho are eating healthier.
Basically, your baseline riskof total mortality was 1.47%, so

(52:21):
let's say 1.5%.
Okay.
Versus quartile four, which was1.536%, so versus 1.56%, versus
so versus 1.56.
So then to that baseline riskyou add that four percent.
You went from a 1.5 risk ofmortality to like 1.6 that's a

(52:42):
really good way to think aboutit.

Introducer (52:44):
That's a really that's a much better way to
think about it that's why youhave me on this podcast.

This is why you tune in people, I know,
I know.
So get that, go get that trailmix, joss, and go chomp down.
You're not.

Introducer (52:58):
Well, that's what I was saying.
I was like, hey, I'm not, I'mlike I'm totally on the
bandwagon, like ultra processedfood is bad.
I was like we're talking abouta 4% difference here over 35
years, from the highest tolowest quartiles, and then okay,
quartiles, and then okay, andthen get this on top of that.
Once they controlled for thishealthy eating index.

There was zero effect.
Oh, hold on, hold on.
My dominoes order just came in.
Let me go get that from thedoor while I go crush this bag
of doritos and get a tub of icecream I mean, it's just like
it's crazy, it's and, um, yeah,so we have this headline ultra
processed foods kills everybody.

Introducer (53:37):
And again, just to underline this, if this is real
and there's all the questions of, like the measurements and all
this, we're talking about thebest to the worst comparison.
You know, 1.5% risk to a 1.54%risk over 35 years.
This is what we're seeing.
And then, if you control forhealthy eating, it's actually
not a thing, and so I don't know.

(54:00):
It's just like, yeah, likedon't eat ultra processed foods,
like I don't want you to die,and we're talking about a really
minuscule effect size here.
And actually the ultraprocessed foodsed foods
themselves may not be the issue.
It might be the overall qualityof the food, and again that
we're measuring a surrogate foroverall dietary quality.
Like you said earlier, thedietary pattern is the one that

(54:24):
matters there.

Yeah, and then when?
So that was total mortality.
What about mortality fromcancer?
Yeah, so I don't think they sawan association with cancer
right yeah, it was 0.96, soprotective, but if we look at
the confidence interval, andfrom 0.91 to 1.0, so yeah, this
is really interesting.

Introducer (54:46):
No association with cardiovascular death, cancer
death or respiratory death.
That's so interesting.
Yeah, so there's something else.
So their argument is there'ssomething else that this is
doing.
It's killing us in other ways.
And what I thought wasinteresting we had to get
Heather on the podcast for thisis they saw the signal with

(55:07):
neurologic disease and she'salways talking about ultra
processed foods and themicrobiome and cognition and
neuro, you know,neurodegeneration, which they
did see an association with, andso you know that might be the
signal here, which is, you know,quite fascinating.
I think it's a small signal,but that might be the signal.

And a couple of things too, that I
thought were interesting, wasall cause mortality seemed to be
stronger in participantsconsuming less alcohol and not
smoking say that once more allcause mortality had a stronger
interaction oh, excuse me, itseemed to be stronger um in

(55:52):
participants consuming lessalcohol.

Introducer (55:55):
So an inverse relationship, and not currently
smoking I think that makes senseagain in that, like you don't
have those other risks for death, and so then the association
would be stronger.
That's how I make sense of thatsentence.
Am I thinking about that wrong?

um, I kind of I guess maybe I read it
wrong and I kind of interpretedit as they kind of found the
opposite of what you wouldexpect, like people who are
drinking less and not smokinghave a higher incidence of
mortality.

Introducer (56:22):
Oh, I was reading it as like.
Once they took that out of themultivariable.

But then there's also that, yeah, which
is probably more accurate, whichwould then mean that, hey,
there's a U-shaped curve toalcohol, which is kind of
something that has been sort ofestablished of like.
There is sort of this healtheven though alcohol is a toxin.
It is a U-shape where too muchis bad for you.
But we're also seeing that noconsumption, or very little

(56:48):
consumption, seems to beassociated with more health,
like adverse health outcomes.
But there's a sweet spot therewhere it's kind of neutral or
maybe beneficial, and so is BMI.
Bmi is also well known to havea U-shaped curve to it, where
people who are very low on theBMI in this study they looked at

(57:10):
less than 25, there's actually,you know, an increased risk and
those who are like morbidlyobese, increased risk.
And those who are like morbidlyobese there's also an increased
risk.
But there is this healthy rangeof like, of like, you know,
normal to slightly overweight,where there is actually a
neutrality or possible healthbeneficial effect there.

Introducer (57:33):
Yep, yep, yep, which totally, which I think makes
sense A lot.
I think we're learning a lot ofthings in in medicine and
health are have U-shapes, whichwhich totally makes I think
makes sense a lot.
I think we're learning a lot ofthings in in medicine and
health are have u shapes, whichwhich totally makes sense to me,
although maybe not ultraprocessed foods, I don't know um
, you know interesting whichgoes to that really dumb saying
that I hate of everything inmoderation, because then you
have to find what is moderationyeah, everything to the left of

(57:57):
the curve in the U.
So okay, I just want to pull up.
Oh, we talked about that.
We talked about that.
We talked about that, okay.

Do you want to break it down?

Introducer (58:10):
real quick by the ultra-processed food type you
can Sure.

Go ahead.
Basically the main one thatthey looked at that did have the
strongest association, and onethat I did think was actually
kind of important was the ultraprocessed meat, um, with a
hazard ratio of 1.13 in thatconfidence interval.
The chicken nuggets, yeah, itwas 1.10 to 1.16, so it's smack
dab in a narrow confidenceinterval in that that's pretty

(58:35):
high amount, like 13, whichranges from 10 to 16 increased
risk yeah, and again, these arerelative numbers like, if you
were to like make one change getrid of the chicken nuggets like
.
What would you do, instead ofsaying, oh, you can't eat
anything.
Get rid of the chicken nuggets,get rid of, like you know, know

(58:56):
, your sausages, your yourbologna.

Introducer (58:58):
Yeah, totally true.
Okay, can I get back on thishealthy eating index hypothesis
thing?
So I called that up, right?
So I'm convinced that whatwe're really talking about here
is we're just it's just asurrogate for this healthy
eating pattern, right, it's notnecessarily the chicken nuggets,
although, um, anyway.
So when you look at the whatgoes into this, I just pulled up

(59:19):
, like the constituents of thisindex.
So it's like asking people howmany fruits and vegetables,
whole grains, nuts and legumes,omega-3s, pufas, things like
that.
And then on the negative side,you know, sugar-sweetened
beverages, red and processedmeat, trans fats, sodium and
alcohol, right?
So these are kind of the thingsthat we all kind of assume by

(59:40):
consensus are part of a healthyor unhealthy diet.
And so, yeah, there's someoverlap, which is, when you say
red and processed meat, that'sgoing to overlap with the ultra
processed stuff.
Oh well, the processed meatversion, rather not the red meat
, which is the unprocessedversion, but not totally
overlapping.
And again, I really think it'sjust a freaking surrogate for

(01:00:01):
this, which is the overallhealthy eating pattern.
And, yeah, if you eat healthier, you won't die as much,
although the difference appearsto be pretty darn small.
I don't know, I'm really.
I'm like.
The more I look at this Adam,like the more I'm convincing
myself, like, more I'mconvincing myself, like it's not
the nuggets, it might be thenuggets, but I'm definitely

(01:00:24):
getting rid of the nuggets.
I'm just putting that out there.
So the gluten-free dino nuggetsthat is like a mainstay of our
kids diet it's out.
It's out today.
Um, anyway, I think it's, Ithink it's really a surrogate
for this.
I'm just like, more and moreconvinced, and I think if
there's anything novel in thispaper, it's this
neurodegenerative piece which Ithink is very interesting and

(01:00:45):
makes sense through themicrobiome.

Dr. Joshua Goldenberg (01:00:48):
And then the last thing I would say Well,
before, you get into that onthe context of excluding diet or
not diet, but excluding foodswithin your diet, because I hear
this all the time of like, oh,diet sodas and like these diet
drinks are terrible for you andthat you should just drink the
real thing.
And then, like you know, quoteunquote mexican coke is

(01:01:09):
healthier.
Um, when?
When they actually furtherseparate out sugar sweetened and
artificially sweetenedbeverages, they found a stronger
association for sugar sweetenedthan artificially sweetened
beverages.
They found a strongerassociation for sugar sweetened
than artificially sweetened.

Introducer (01:01:21):
Wasn't that interesting.

I mean, but like duh right, Because
again it kind of goes back downto this like no nutritional
value in calorically dense food.
If you have something withoutthe calories, you're not going
to be getting sort of like theexcess caloric intake from.
You know, if you're drinkinglike five actual cokes a day
versus five diet cokes, you'regetting like an increase yeah,

(01:01:46):
but remember they're controllingfor bmi though, right yeah, so
that, but yeah, but it's notjust a weight thing.
Well, there is that.
I mean there's that componentwhich which they did control for
, and then when they controlledfor that, they still found that
the that the artificiallysweetened diet cokes or diet
beverages were not the actualfood of the non-diet had an

(01:02:13):
increased risk versus the diet.

Yeah, I thought that was interesting.
Just because I constantly hearthat it's like chatter of like
oh to how terrible dietary orhow they're totally counter to
the standard yeah, and I meanit's it's so consistent, it's so
consistent in the literatureabout that yeah, it's very
interesting, but um, so I wantto get, we should get have
heather, have like her thirdappearance.

Heather's wiki, dr's wiki she gets an
unlimited amount of appearancesyeah, she can make as many darn
appearances as she wants.
Like that she, she's a standingorder.
If if she was like hey, I wantto be on the podcast, we'll just
be like okay, cool, like yeah,standing orders, stat she can
get on, that's right.

Um, she's got that stat card, um, and so, yeah, I
think the neurodegenerationthing is actually in my mind,
although they only mentioned itas like a side note, like the
most interesting thing to mehere and the other quick aside.
I wonder why.
I wonder if one of the reasonswhy they didn't see the
cardiovascular and cancerassociation is because they

(01:03:16):
controlled for weight and so you, the whole calorie issue, which
of course is associated, issort of um, netted out in that.
Right, they're just lookingspecifically, not necessarily at
the calories, but at the ultraprocessed foods specifically.
But now here's the thing that Iwanted to say.
This is spin on like good stuff, it's not industry sponsored or

(01:03:36):
anything but like, just look athow I mean the way this was
sold, the way they talk aboutthis in some places, like strong
association, um, you know well,you know they, actually, they,
they controlled it pretty well.
Like, if you look at theirconclusion in the abstract, they
say at the end they said theysimmered down.

But I wonder if that was from peer
review.

They simmered down because I think it's because the
yeah, I think it's bmj forcedthem to simmer down because
they're very strict about thatsort of thing and I know, you
know, some of these authors arejust very much into this, which
is good, I mean, I agree.
Like you know, um, you know,eat whole foods.
Totally agree, that's what I do.
However, like the associationthat we're talking about here is

(01:04:19):
minuscule and it goes away assoon as you talk about this
index and so, like you know,this idea that it's this massive
association, but actually theabsolute numbers are like
laughably small, the slightincreased risk and yada, yada,
yada.
Anyway, I just find it spin isso interesting, but I think they

(01:04:40):
, the authors, either theauthors and or the journal
editors, did a good job intempering that language in the
abstract and the conclusion.

Yeah, and I mean it goes back to like
yeah, like eat mostly a wholefood, like a whole foods diet,
but don't ever nutritional shamepeople for what they eat.
First off, you're a jerk if youdo that too.
Right, like no seriously youare, and then too like it also,
like you're allowed to eat theseother processed food things and

(01:05:11):
, in the grand scheme of things,it's not going to move the
needle not by much, not by much.
Like focus on the bigger things.
Like don't be someone who, who,chain smokes two packs a day,
but then be like oh, I'm notgonna eat doritos because
doritos are bad for you, likeit's, you'd probably be better
off not smoking did you do a um,did you do?

I didn't do like a numbers needed to treat thing
here because you had you werebetter with the absolute numbers
here.
I am curious, like over 35years, like how many people
would have to eat the lowestlevel of ultra processed food
versus the highest to preventone death?

I think it would be like really a large
numbers needed to treat herewell, while I talk about, um,
you know, the great benefits ofof dr journal club.
Why don't you calculate that?
And while we run a quick ad,yeah, so where are your?

where are the absolute numbers that you came
from?
You said 1.5 percent, right,where did that come from?

I think it was table two okay, so table
two yeah so if you go to Table 2and then go to Total Mortality,
the incidence rate per 100,000person.

Years in Quartile 1 is 1,000.
Well, that's the problem.
It's person years yeah.
So we can't really.
I mean, that's the issue is,they're not giving us absolute
numbers, like we can't calculatethat based on person years.
It's probably small, it'sprobably really freaking small.
And I'm suspicious that theydon't have some comment about an
approximation of what we'retalking about here.

Um, but no, we can't actually do that I
think I think if you sent theman email or press them or sent
like a letter to the editorasking them for that, they
probably would be able to dothat.
Yeah, I feel like it letter tothe editor asking them for that,
they probably would be able todo that.

Yeah, I feel like it should be required.
You know, like, for example,for Cochrane reviews, you must
have an absolute risk statement.
And I think one of the biggestissues with nutrition research
is you talk about relativepercentages all the time, and
even here the relative percentdifference of 4% is very small,
but I can't even imagine whatthat would be in absolute terms.

(01:07:22):
Anyway, I was just kind ofcurious about that.
But no, you can't do that fromperson-year calculation, and I
would be interested to know,because it's like they have the
number of.
Yeah, maybe we could.
Well, I don't know, becauseit's challenging technically
because they count them overdifferent periods of time.
That's why they do the.
Anyway, we're on a statisticaltangent, but the take-home is,

(01:07:44):
it's probably pretty stinkingsmall, but it's not medical
advice.
Don't go out and eat dinonuggets, or if you do, don't
blame us.
Okay, anything else you want tosay last minute before we wrap
up, sir?

No, I mean, I still thought it was an
interesting study, it was agreat study yeah.
I thought it was.
Yeah, no, I thought it was agreat study.
I thought it was, I thoughtthere was some stuff here that
was really interesting.

You know what the issue is.

Nothing really matters, Well it doesn't
matter, it's not that it doesn'tmatter.
It doesn't matter.
It's not that it doesn't matter.
It doesn't matter to themagnitude we want it to.

We think we nothing really matters.
That's the really word therefrom metallica I mean, I think,
like even this over 35 years.

Are you sure it's metallica and not
queen?

isn't?
Oh, is that queen or metallica?

nothing, it's Queen, it's Queen, it is
Queen.
Yeah, dude.
Nothing really matters whatsong is that it's Bohemian
Rhapsody.

That is Bohemian Rhapsody.
What am I?
Doing Come on dog, I know, Iknow, I know.
Should we keep this or shouldwe call this a?

B-roll.
No, we should definitely keepthis, so people know we should
probably keep this.

We should just out me.
You're a fraud.

Dr. Joshua Goldenber (01:08:59):
Forgetting Bohemian Rhapsody.
You're a fraud.
You don't know anything, josh.
You don't even knowevidence-based medicine.
It's all made up.
It's like whose line is itanyway?
It's all made up.

Man, I just conflated Queen and Metallica.
I feel like someone just rolledover in their grave.
I feel like that is a major.
We've lost our entire audienceCommandments.

Faux pas.

If they're still listening.
This is a particularly ramblingepisode.
All right.
Should we close it up at thispoint?

Should we ?
Let's go on a tangent about Idon't know?

This is terrible.

Let's tangent about closing out.
Yeah, about closing out podcast.
Yeah, it's closing time.

Okay, now, what's that?
Now, is that okay?
Is that what's close?
Who sings closing time?

I don't know who that is.
Is it Creed?

I don't know, I don't think so I want to.
I want to say Blink 182, but Iknow it's not blink 182.
Don't jump on me.
It's um, what is it it was?
Weren't they like one hitwonders?
Closing time?
No idea, closing time song,semi-sonic?
Yeah, they were, I don't know.
Maybe people will say they'renot one hit wonders and that was
like their major, their majorsong semi-sonic, like the

(01:10:17):
semi-quality or semi-qualitative.
Okay, I'm cutting it.
This is ridiculous.
We need to end the podcast.

Okay.

Bye everybody.
Thank you for listening, ifyou're still listening.
If you enjoyed this podcast,chances are that one of your
colleagues and friends probablywould as well.
Please do us a favor and letthem know about the podcast and,
if you have a little bit ofextra time, even just a few
seconds, if you could rate usand review us on Apple Podcast

(01:10:43):
or any other distributor, itwould be greatly appreciated.
It would mean a lot to us andhelp get the word out to other
people that would really enjoyour content.
Thank you, hey y'all.
This is Josh.
You know we talked about somereally interesting stuff today.
I think one of the things we'regoing to do that's relevant.
There is a course we have on DrJournal Club called the EBM
Boot Camp.
That's really meant forclinicians to sort of help them

(01:11:05):
understand how to criticallyevaluate the literature, et
cetera, et cetera Some of thethings that we've been talking
about today.
Go ahead and check out the shownotes link.
We're going to link to itdirectly.
I think it might be of interest.
Don't forget to follow us onsocial and interact with us on
social media at DrJournalClub.
Drjournalclub on Twitter.
We're on Facebook, we're onLinkedIn, et cetera, et cetera.

(01:11:26):
So please reach out to us.
We always love to talk to ourfans and our listeners.
If you have any specificquestions you'd like to ask us
about research, evidence, beinga clinician, et cetera, don't
hesitate to ask.
And then, of course, if youhave any topics that you'd like
us to cover on the pod, pleaselet us know as well.
Thank you for listening to theDoctor Journal Club podcast, the

(01:11:49):
show that goes under the hoodof evidence-based integrative
medicine.
We review recent researcharticles, interview
evidence-based medicine thoughtleaders and discuss the
challenges and opportunities ofintegrating evidence-based and
integrative medicine.
Be sure to visitwwwdoctorjournalclubcom to learn
more.

All Episodes

Episode Transcript

Popular Podcasts

Fudd Around And Find Out

Crime Junkie

The Breakfast Club

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Avoiding Ultra-Processed Foods: A Deep Dive into Health Impacts and Research Insights

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}Fudd Around And Find Out

Crime Junkie

The Breakfast Club

All Episodes

Avoiding Ultra-Processed Foods: A Deep Dive into Health Impacts and Research Insights

Fudd Around And Find Out