00:20:21.780Like this just isn't something that like you need to worry about on like this micro of a scale.0.99
00:20:28.740It sounds a little bit like cholesterol in that very low cholesterol is like has the same risks, essentially, as very, very high cholesterol, right? The tenor of the conversation around these things make it seem like the more antioxidants you can eat, the better, the lower your cholesterol can be the better. And like, that's not actually borne out in the research. Is that a fair assessment of what's happening here?
00:20:55.000Oh, totally. I mean, this is interesting for researchers and I'm really glad that people
00:20:58.900are looking into it and maybe they will discover something about this system that becomes useful
00:21:03.240to consumers in 20 or 30 years or whatever. But right now, all we can really say is that like,
00:21:09.680yeah, try to eat a wide variety of food. Try to eat fruits and vegetables. If you're super worried
00:21:14.920about your free radicals antioxidant balance, like maybe you're neglecting your blood pressure
00:21:19.080balance or maybe you're neglecting your cholesterol balance. It's like, and there's probably 50 other
00:21:23.000of these balances inside of your system that like we don't know about, like haven't become
00:21:27.180marketing yet. This taps into two things that are particular bees in my bonnet. One is this idea
00:21:34.980that every new scientific or nutritional finding has to immediately translate into individual
00:21:41.080practices. Yeah. Yeah. And the other thing that feels challenging about this is we have gotten
00:21:46.240so dead set on a conversation that focuses on weight as the end all be all of health indicators
00:21:53.880and i'm like there's like so many systems in your body there's so much that's happening right the
00:22:01.320idea that you could focus on any one single number whether that's your weight or your cholesterol or
00:22:08.100your blood pressure or your mood evaluations or whatever the idea that that could be the sole
00:22:14.960reflector of an entire picture of your health and the functioning of your body is bananas it's like
00:22:22.020oh i have 75 health points today you're not a sim you know so basically in the 1990s this this
00:22:31.620becomes the conventional wisdom that there is something about alcohol that is protective
00:22:36.900of the cardiovascular system and the most likely explanation is that it's resveratrol
00:22:42.540some other antioxidant inside of red wine specifically. So, you know, I love like a
00:22:49.640try hard structure for an episode. You know, like having little chapters.
00:22:54.960Are we going to be joined by a little skeleton?
00:22:56.520It's going to be a little skeleton. I wanted to sort of go through the assumptions that were
00:23:03.220going into this conventional wisdom. If you were interested in actually investigating what's going
00:23:09.680on here what are the things that you have to actually establish first before you can give
00:23:16.400anybody individual health advice one a direct connection between red wine and heart attacks
00:23:23.400well it goes even further back tell me the first thing you have to establish is does france actually
00:23:32.580have a low rate of heart attacks yay let's do this this is the part of like the research for
00:23:40.400the episode i was like i don't know if there's like an episode here and then i found this and
00:23:43.520i was like okay there's like four episodes here i think this is like such a perfect metaphor for
00:23:48.100like so many things that we talk about on the show so obviously there's death statistics for
00:23:52.880every country like if you want to know how many car accidents there were in romania last year you
00:23:56.860can look it up and it's something that has like the weight of statistics right you're like well
00:24:01.320these are just numbers and this is just science. And obviously there's just statistics and you
00:24:05.440can't really complain about them or whatever, right? The main thing to know about death
00:24:10.020statistics is that they all go back to an individual death certificate. When somebody
00:24:15.780dies at a hospital, a doctor fills out a death certificate and then that's like entered into a
00:24:20.080database and then that eventually flows upward into national death statistics, right?
00:24:25.220I think I've mentioned on the show before that when I was doing grad school in London,
00:24:29.120my boyfriend was an er doctor what in britain they call it arts and entertainment yeah so i this was
00:24:35.980like me doing journalism for this episode i like opened a facebook messenger window i was like
00:24:40.080excuse me tell me about your relationship with death certificates and so he sent me the guidance
00:24:47.560that the uk gives to doctors of like how to fill out death certificates and like this is all based
00:24:53.340there's standardized kind of formats for death certificates that are based on WHO. There's codes
00:24:59.280for like what people die of. So like all of this, they're trying to standardize this as much as
00:25:05.120possible so that you can look at different causes of death between countries. So the way the death
00:25:11.640certificates work is there's like the direct cause of death, and then there's all of the
00:25:16.340underlying causes of death. So the example that they use in the UK guidance is the immediate
00:25:22.840cause of death is ruptured liver. And then the underlying cause of death is pedestrian knocked
00:25:27.240over by car. So it's like you'll have something like cranial fracture, car accident. Gotcha.
00:25:32.440There's also, you know, you think about long-term diseases, right? You think about like something
00:25:35.600like HIV. HIV doesn't kill you, right? It just makes your immune system so weak that oftentimes
00:25:40.460you get pneumonia or you get some sort of opportunistic infection. So you'd say like
00:25:44.080this person died of pneumonia, but the underlying cause of death was HIV. This is also one of the
00:25:49.120things they say in the guidance is this is actually how we get you know the term died of old age
00:25:53.220yeah in the guidance they just say like if you don't know and somebody's over 80 just put died
00:25:58.300of old age as the underlying cause so it's just like uh old age yeah this feels like in mental
00:26:06.120health world there's like often a classification that's like such and such not otherwise specified
00:26:13.140right so for a long time there was eating disorder not otherwise specified it's just like
00:26:17.360hey man your relationship to food is real weird but it's not one of the big ones that we already
00:26:21.860figured out you got a special thing going on this is all like a completely normal system right you
00:26:26.040you want to have these different layers of causes so that you can do statistics at the end of the
00:26:30.840year right you want to be able to say like there were this many car accidents this year not like
00:26:34.580how many fractured craniums were there this year right like you have to have a couple of different
00:26:39.460types of information where the system gets tricky is when you have older people who have more than
00:26:47.200one underlying condition. So this is from a study of death certificates. It says,
00:26:52.640the WHO disseminates standard death certificates, coding rules for determining underlying cause of
00:26:57.720death, and standardized tables for comparison. However, it cannot control what is written on
00:27:03.020individual death certificates or whether the diagnosis is evidence-based or merely an expression
00:27:08.240of one person's opinion. So what researchers have found when they look at these kind of overall
00:27:13.760death statistics, and then they go back and they look at individual cases, so they double check
00:27:18.940the national statistics against actual individuals, is that it's kind of a judgment call what gets
00:27:26.600written on the death certificate as an underlying cause of death. Because only around 10% of people
00:27:34.020get medical autopsies. Most of the time, it's like, okay, this person's 85 years old, and they
00:27:38.640went to sleep, and they just kind of never woke up. We're not going to do like a whole big
00:27:42.720investigation. Sure. And also, like, despite what we've all seen on Law and Order, it's also
00:27:47.360unusual to have cops show up when you die. And so when researchers go back and check things,
00:27:54.240what they find is that a huge number of people who had, you know, heart attack or diabetes or
00:28:00.600whatever listed on their death certificate didn't actually have that. There's no actual evidence of
00:28:04.640that. Wait, wait, wait. You're telling me that people had diabetes or heart attack or whatever
00:28:10.940listed on their death certificate as a cause of death, but they don't really know if that person
00:28:16.300had diabetes or had a heart attack? Yes. So I found this really interesting study called
00:28:20.560Death Certificates Are Not Reliable, where they go to a single hospital and they go back and they
00:28:24.400find 223 autopsy reports. So they find 25 cases where an autopsy later finds a heart attack,
00:28:32.680but the doctor didn't write it on the death certificate. And they also find nine cases
00:28:37.360where the doctor wrote heart attack on the death certificate but the person died of something else
00:28:42.040so around half of the errors are not finding a heart attack that's there and 25 of the errors
00:28:48.860are finding a heart attack that wasn't there the study says death certificates were frequently
00:28:54.520inaccurate and in 21.5 of cases were of no value because of an inadequate diagnosis uh i'm sure
00:29:02.680that this is not part of the data but what do you want to fucking bet that a bunch of that 25 was0.85
00:29:08.620fat people well this is this is what's so fascinating to me there was an interesting0.90
00:29:12.800study in italy that looked at diabetes deaths and found that basically people with the same
00:29:18.100symptoms women and people who were found in their home as opposed to dying in the hospital
00:29:25.240were much more likely to have diabetes written as an underlying cause of their death huh it's
00:29:30.700basically a judgment call based on the medical examiner or whatever doctor is looking at you
00:29:35.000to determine like what are the factors that contributed to this person's death.
00:32:37.940But, like, I'm 80% sure I'm just going to write heart attack on the death certificate.
00:32:41.500So when people go back and look at the death statistics that were informing all this French paradox stuff, they find that France has a really low rate of heart attacks, but a very high level of unclassified cardiac events.
00:33:22.440One of the best papers I read on this was in 2008, a French doctor does this like overall
00:33:28.260analysis where he's like, okay, let's just like look at the French paradox.
00:33:31.640Like, what are we actually looking at here?
00:33:34.580And he goes over all the death data and, like, the data they were using to formulate the French paradox.
00:33:40.340And he says, strictly defined coronary death rates in French national statistics, both in the past and nowadays, should be considered as negatively biased estimates at the population level and cannot be used validly in ecological correlation studies.
00:33:54.720the 60 minutes episode says over and over again that if you're an american you are three times
00:34:01.800more likely to have a heart attack than if you're a french person and when they go back and they
00:34:06.560kind of rerun the numbers and they classify everything correctly if you're an american
00:34:09.920you're about 30 percent more likely to have a heart attack than if you're french so is that
00:34:14.600nothing like no that's actually a pretty big difference but 30 percent more likely and 200
00:34:20.040percent more likely are like very different sociological phenomena and imply very different
00:34:26.680like public policy responses basically yes totally if you're trying to solve a problem
00:34:31.400step one is you got to be able to wrap your arms around the magnitude of the problem and who it's
00:34:36.660exactly and it sounds like we're not currently able to do that just based on like the baseline
00:34:44.040data that's available and also it's not really the french paradox because france has roughly the
00:34:48.660same heart attack rates as like italy and spain oh my god how long until we get reporting on a
00:34:54.980spanish paradox and everyone starts eating like sardines yeah yeah eat at 10 p.m yeah that's right
00:35:01.740okay so that was try hard structure that was like galaxy brain level one is like our heart attacks
00:35:09.540low in france right galaxy brain level two is does drinking prevent heart attacks like is this thing
00:35:16.840of like moderate drinking. Good for you. This is a part of sort of health and wellness science
00:35:23.280world where I have spent almost no time is the booze end of things. And also like, I'm just not
00:35:29.260a big drinker. It's not my thing. Totally. Wait, so you're so you're a moderate drinker.
00:35:34.500I'm a social drinker pretty much exclusively. Like I have like a bottle of whiskey at home
00:35:40.900for like if something big happens but that's like once every year or two that i like a drink for
00:35:48.480myself by myself at home right everything else is when i'm out with people and i haven't been
00:35:53.000out with people for the most part because of covid so i just don't really drink much at all
00:35:57.660i mean this is like an interesting sort of self-reported data thing because on the surveys
00:36:01.920you would show up as a moderate drinker even though in reality you're a non-drinker basically
00:36:08.060right i just don't love it the feeling of a hangover is such a fucking bummer also like0.97
00:36:11.980once you're in your 30s it's a whole other fucking ball game i thought people in their 30s were like0.95
00:36:16.880overblowing stuff when i was in my 20s and now i'm here and i'm like no i understand back pain0.97
00:36:22.920and hangovers i got it oh my god yes i just realized the other day that everything that
00:36:27.280used to bother me about my parents i am now doing i'm like oh why do you have to wear earplugs to
00:36:32.740sleep and now i'm googling like most powerful earplugs yeah for night noise yeah that's right
00:36:38.680but then okay here we get methodology queeniness i'm a non-drinker many of my friends are like
00:36:45.560moderate drinkers because i'm fairly similar to them i always kind of assumed that when you do
00:36:51.800these studies that compare non-drinkers to moderate drinkers you're basically comparing
00:36:56.520roughly the same group, it turns out that non-drinkers, or at least people who say that
00:37:03.620they do not drink on surveys, are like a really distinct group. The main thing that differentiates
00:37:10.720non-drinkers from moderate drinkers is that non-drinkers includes people who have been told
00:37:17.560not to drink by their doctor because they have pre-existing health stuff. A huge number of people
00:37:23.440with diabetes do not drink. A huge number of people who like have high cholesterol or at risk
00:37:28.980of things like heart attacks have been told not to drink. Those people have stopped drinking,
00:37:34.660but they still have all of the health risks, right? So if you're like someone who's been a
00:37:39.580really severe alcoholic your entire life, at age 60, you finally quit alcohol. You're now a non
00:37:45.340drinker. Well, you still have all of the liver damage and all of the other health effects that
00:37:49.980the drinking caused for your entire life but you show up on the surveys as a non-drinker because
00:37:55.240they're essentially taking a snapshot of what is your relationship to alcohol now not what has it
00:38:00.180ever been in your life exactly yeah and so these these people in the studies are called sick
00:38:06.320quitters which just seems like like a kickball team or something god that's like the biggest0.98
00:38:12.480loser like what a weird mean thing to name that like fuck dude but then even people who drink0.99
00:38:18.540are called healthy users. It also doesn't sound that cool. I know it's all bad. But then another1.00
00:38:23.660thing about non-drinkers that I didn't know is that non-drinkers are like distinct from moderate
00:38:30.160drinkers in a million other ways too. So black people are twice as likely to be non-drinkers
00:38:35.840as white people. People without health insurance are 1.5 times more likely to be non-drinkers.0.74
00:38:42.080unemployed people 25% more likely to be non drinkers people who earn less than $25,000 a year
00:38:50.020are three times more likely to be non drinkers I think because alcohol is expensive yeah totally
00:38:55.920alcohol is expensive and also this is like a fascinating moment of like what an interesting
00:39:01.560thing that so many people who are sort of on the downside of power and privilege yeah are doing
00:39:06.740this thing that broadly is understood to be a health promoting behavior and we are not talking0.86
00:39:11.760about it one fucking bit yeah the things that we were talking about as health promoting behaviors0.71
00:39:16.040are the things that are expensive and the things that rich people do the end right and it's also0.98
00:39:20.600really hard to make comparisons between these two groups because when you zoom out non-drinkers and
00:39:27.340moderate drinkers aren't just groups that differ in this one behavior non-drinkers it turns out
00:39:33.020includes a disproportionate number of people with fairly serious pre-existing conditions
00:39:39.820and a lot of poor people, ethnic minorities, unemployed people, and also older people.
00:39:46.800All of those groups have higher mortality rates, but they don't have anything to do with drinking.
00:39:50.820Sure. I'm thinking particularly of the uninsured people. That feels like a pretty straight line.
00:39:57.680They've tried to solve this problem by only comparing moderate drinkers to people who have
00:40:04.820never drunk in their lives. But then that's not a remotely representative sample either,
00:40:09.920right? That's mostly religious people. So you can't really get around it that way either.
00:40:15.480There's also the fact that, I mean, we talk about this all the time, so I don't want to
00:40:17.840belabor it. But like, there's also the fact that all of these surveys of drinking behavior are
00:40:22.080based on self-reported data. Moderate drinkers encompasses people who have one drink a year
00:40:30.180and people who have two drinks a night. What? Right. A lot of people who like drink very rarely
00:40:35.120still call themselves moderate drinkers, right? Or like seldom drinkers. The idea of like me as
00:40:41.680someone who's had, I don't know, two drinks in the last year as a moderate drinker seems bonkers to
00:40:47.620me. Well, one of the things I think is so weird that it doesn't really come up in the literature
00:40:51.940about the French paradox too is that like there's something about the self-reporting of red wine
00:40:56.920consumption that also seems off to me because red wine is typically associated with like I'm a
00:41:04.020sophisticated person. It's like asking people how much they go to the opera. Do people who go to the
00:41:10.640opera have lower rates of heart attacks? Probably. But it's not because of the opera. That's seen as
00:41:16.500kind of virtuous behavior for a certain kind of person who probably has a low death rate anyway
00:41:22.400i just don't know how much you can actually really say about this stuff just because it's so
00:41:27.240difficult to get clear answers from people on these like really culturally loaded activities
00:41:34.500yeah i mean this is a thing i uh have had a number of conversations with pals from the uk who will
00:41:39.960talk about like oh my god we had our like weekly drinks out with the staff yeah on friday night
00:41:46.320like our whole office went out for drinks and i got totally plastered and i'm like yeah that would
00:41:52.940be a fireable offense in the u.s right if you went to a staff function and you got completely0.69
00:41:59.580shit-faced dude when i worked in london people used to call in hungover i can't make it in today0.99
00:42:06.820i just got hammered last night i as an american was like what the fuck totally and it's just this0.99
00:42:13.100fascinating moment of like contradictory sort of cultural expectations and judgments of social0.98
00:42:18.600experiences yeah yeah it's all fake um okay so the next sort of debunkery is the antioxidants
00:42:26.180stuff you can't really synthesize these things there's been various attempts to like create
00:42:33.740antioxidants in a lab and turn them into vitamin supplements there's no evidence that they do
00:42:39.040anything right this is the americans have the most expensive pee in the world stuff right well
00:42:44.520i mean speaking of expensive pee basically all of the studies on resveratrol like since the 1990s
00:42:53.080have found like you don't really absorb it you you get a spike of it for roughly 30 minutes and then
00:42:59.780it just like goes out in your pee or it kind of goes through your system and is turned into some
00:43:04.400sort of like other chemical thing like through organic chemistry through your body it basically
00:43:08.820becomes this other compound so you're not like you're not able to get the amount of resveratrol
00:43:14.200that would give you any kind of benefit take that mall kiosk yeah your skin's gonna look amazing for0.80
00:43:20.340a half an hour and then it's gonna look like garbage and also like as with most things the
00:43:24.960only real evidence that resveratrol could be isolated and could have these health effects
00:43:31.120is on studies of yeast fruit flies and mice so like not a great start and then the one that they
00:43:39.780gave mice the dose was so high that to get that amount of resveratrol you'd have to drink 1000.93
00:43:46.440liters of wine so like not recommended also this is the shit that i was like about to throw my0.94
00:43:54.180fucking laptop out the window this is from a 1992 new york times article so like pretty early in the0.98
00:44:00.720French paradox. This was being reported. While the amount of resveratrol in wine varies widely0.99
00:44:06.740depending on the type of wine, purple grape juice has an amazingly constant amount no matter where
00:44:12.260in the country it comes from, reported Dr. Leroy Creasy, a specialist in fruit cultivation at
00:44:17.180Cornell University. He found that grape juice had more of the heart-protecting chemical than 60%
00:44:22.980of the wines he analyzed. What? Get out the juicy juice. It's such a tell to me. If you
00:44:30.520believed all this like resveratrol has this magical effect on the body thing there's like
00:44:37.160there's no reason you have to get it from wine there's other sources of reserve you can also0.85
00:44:41.780just eat fucking grapes yeah totally so it really feels like this whole thing was just like a0.82
00:44:46.780permission structure to drink red wine which is fine i think drinking red wine is totally legitimate0.99
00:44:51.740but it's like it feels like nobody actually gave a shit about what they were saying for like a0.52
00:44:58.140decade and a half yeah yeah it feels a little bit like the don't worry darling discourse which is0.97
00:45:03.700just like oh we're just gonna get all real fixated on this one question did this guy spit on this
00:45:08.680other guy yeah yeah let's dig in on this thing that seems fun and interesting right and i think
00:45:13.620there is something a little bit um not quite salacious but sort of enticing about the french
00:45:18.820paradox yeah it is the ultimate claim of any diet which is you can eat whatever you want and still
00:45:27.740lose weight. And I do think it plays into a bunch of weird cultural expectations that Americans
00:45:33.540have about French people and the sort of Europe envy that happens amongst like white Americans
00:45:39.840in particular. Right. Like it just plays into a bunch of things that we're already sort of primed
00:45:45.600to think. Right. Right. So can we say anything definitively about the effects of drinking or of
00:45:52.920red wine. I read a really interesting meta-analysis of sort of what we know about mortality and
00:45:59.960alcohol and J-shaped and all this kind of stuff. And they basically say that like there is very
00:46:06.020good evidence that not drinking is the healthiest way to live. And there is very good evidence that
00:46:13.100moderate drinking is the healthiest way to live. Basically, if you're someone who doesn't drink