That's pretty loud. Is that too loud? I think I think we'll go ahead and and start Before before I actually start, you know going through my presentation I'm a little bit. I'm gonna tell you about a little about myself and then I want to know a little bit about you Okay, so I was asked to come Largely because I'm local so I'm a professor at that Brigham Young University about 40 miles 50 miles south of here and I've spent the last 35 years of my life studying the health effects of air pollution And it happened almost by accident my training my training and backgrounds actually in in economics environmental economics and statistics and There is a local steel mill down in Utah Valley near Provo, Utah it's actually was just outside of Orem and I moved here from Texas A&M at the time and That steel mill actually shut down for 13 months and reopened I'll show you a little bit about this and I was teaching a class that it required a term paper and I had a young woman in my class asked me after class one day You know, what could I do my term paper on and I've been hearing this these stories They were basically anecdotes about little Johnny was sick My little Johnny was sick when the mill was operating got better when the mill shut down and then got Sick again when the mill started operating there was that kind of man I says we could collect the retrospective data from the local hospital Utah Valley Regional Medical Center and just see if there's any evidence of that So that's that's what I did. I said I'll contact and we'll get this D identified data A monthly count data and we'll just look at that. Well, what happened is by the time I got the data She had dropped out of my class Okay, true story So I mean and this was back, you know before electronic transfers or anything It's just that big fat computer paper that they had it's sitting on my desk So I thought well, I was curious one evening. My wife was gone for something. I didn't have any reason to go home So I said I'm gonna spend this evening Putting that data in back in the old desktop, you know the piece of junk computers back then put it in Looked at it just on a graph and I said I've made a mistake. There's some big big air here I'll show you the air here in just a minute and I go I've got I've got to figure this out I ended up publishing the results of that in in American Journal of Public Health and it changed my career I became actually a budding maybe a good economist and now I became a environmental epidemiologist Which which has actually been a very good career and again 30 almost 35 years now So that's that's that's who I am and I'll show you some of the research. I've been involved with here in a minute now Are are how many of you would call yourself? environmental medicine folks Okay, quite a number of you. How many are very strictly occupational medicine folks? That's fair enough, that's fair enough. Okay. Now I have way too many slides to be able to show in the next 55 minutes or something. Okay, so I don't have to get through them all if you ask me questions I won't which is fine. If you don't ask me questions, I still probably won't Okay, which is fine But we're gonna go through them as quickly as we can and you'll see sort of what I'm trying to do With this, so we're let's just talk about air pollution for a minute air pollution is fairly straightforward It's also very complicated this is the straightforward slide All right So you get air pollution from the bad air pollution comes mostly from burning things Burning fossil fuels coal gasoline diesel natural gas It also includes biofuels and from that you get nitrogen dioxide sulfur dioxide carbon monoxide And and and and of course then you get some of these greenhouse gases co2 nitrous oxide methane You have this black carbon that's emitted also from these sources And then what happens is is is these pollutants actually sort of when we when we talk about pollution We're going to talk mostly about PM 2.5 because it's the index of air pollution. That is most Associated with health effects and in fact co it's not really related that much to PM 2.5 although coming from the same sources in o2 s o2 Oxidizes in the in the atmosphere and creates these sulfate nitrate sulfate nitrate particles Then you get this black carbon that's part of the particles and and so we're gonna focus mostly on PM 2.5 Mostly because it's our most important index. Does that make sense? Now we could talk a lot about ozone it's pretty important to Ozone really is is a bit more complex because there's no sort of Primary source what you get is you get the emissions of in o2 you get volatile organic compounds You get sunlight and then you get the creation of ozone Ozone in the upper atmosphere and stratosphere is good filters out ultraviolet rays Ozone in the lower atmosphere we breathe bad harms our lungs, but we're going to mostly ignore ozone So fine particles, what are they? well they're defined as particles less than two and a half micrometers in aerodynamic diameter a Human hair is about about about 60 micrometers 60 to 100 All right If you take the end of a human hair you put a PM 10 particle it'd look like that or smaller PM 2.5 particle is that small or smaller? They're really tiny and and basically the only way you get PM 2.5 particles is through high-temperature Processes and burning things. Okay, you can't you can't mechanically generate these fine particles at least not in sort of a natural way Okay, so so Here here's here are some PM 2.5 particles now the width of that slide notice. Here's our scale That's five micrometers So about half of that Is is is about two and a half micrometers? So these and now now so this is this is obviously magnified So if you're looking, you know, the width of a human hair would be clear over here All right. Now here are that's a PM 2.5 particle My pointer doesn't work real well, but you see that one There's one but all of these are PM 2.5 because PM 2.5 by definitions are those that are two and a half microns or less All right They're very very small and you get a whole bunch of them in the air You can't see any particle with your naked eye, but if they get in the air, they look nasty Okay, a lots of them gives you this Smoggy foggy nasty looking look and and and and I actually have right here This is where we're sitting right now right here in the middle of Salt Lake City in During bad temperature inversions mostly in the winter. We actually get fairly high pollution in in Salt Lake City but never as high as you get in say Beijing or in New Delhi or or other other places This is Agra. That's the Taj Mahal. You can hardly see it right now. All right now How am I going to organize this presentation? by ten different Controversies, we're just going to talk about ten different controversies just to make it more fun All right, and I'm going to ask the controversy and then we're going to ask you to answer it after we talk about it a bit So the first controversy is was London smog romantic or Deadly. Oh lots of novels about from the romance of the London fog and London smog sort of stuff Was it romantic or deadly then you can see, you know, here's these sort of fairly famous Paintings Ellen Shaw paintings. This is Mary Poppins. And you know, yeah, it's pretty cool. Very romantic Monet he loved it. He thought the smogs were romantic as well. All right Others didn't this is this is from a cartoon from the British humor magazine punch clear back in 1880 and Old King Cole and the fog demon you can't see this very well, but this says This says I can't read a bronchitis asthma Pneumonia, you know there was real concern about this air pollution causing Disease primarily respiratory disease. All right The debate about whether or not the smog in London was deadly or romantic Actually was pretty much finished by night 1952 So in in 1930 in Mews Valley, Belgium, this was an industrial valley in Belgium They had a bad temperature inversion And again think of this as sort of like a natural experiment you have you have a whole bunch of folks Living in a you know in a in an exposure chamber a valley and then you have a natural temperature inversion trap the pollution in that valley and That happened just for a few days about four or five days and they had 60 deaths About ten times what was expected and lots of the community was ill and their livestock were dying, too And their livestock were dying to animals were being affected by this Then later a few years later 18 years later in Denora, Pennsylvania Similar sort of thing happened half of the community was ill About half their doctors got so ill they couldn't go around and keep taking care of themselves In fact, the fire department had fire department personnel going around administering oxygen to the people They ran out of coffins All right. It was a very very bad. There's no doubt that at very very high levels of air pollution Contributed to the disease and death in Denora and the question was then is Wow If something like this were to happen in a big city like say London There would be thousands of deaths and then what happened a few years later 1952 in London They had a massive episode and they had thousands of excess deaths Now my wife knows this because she watched the crown Okay, and there is an amazing episode on that and I watched that episode with her it's great But it was a it was a massive episode that The reality is is that the pollution really shot up both both smoke or think of this now as fine particle pollution and sulfur dioxide but that that pollution shot up and was high for about three or four actually about five days and you can see what happened to Deaths it went up from a baseline and and by the way We now know that the baseline was high because of long-term X chronic exposure if you will to the pollution in London But this relatively acute episode you see that the deaths went up from around just 260 or so to 900 or so deaths y'all see that and If you just integrate underneath that excess deaths curve there you get excess deaths of you know a couple thousand but it kept going on for a few more weeks if you integrate longer you get about 12,000 excess deaths and And this there's been a lot published on this it's pretty well established I have a photo here of David Bates. Who's a Friend and colleague he died several years ago, but he was a young pulmonary physician there at st. Bart's and And and had many of his patients were involved with this and he actually went to a cattle show where some of the cattle were dying during this episode and And and and and actually looked at their Their their lungs and their airways that were just clogged up and with with bad inflammation It was it was pretty amazing and and he went on to Canada he was actually the Dean of the med school at University of British Columbia, but has done all sorts of studies and and and and some of them I've worked with him on but that's what's that's 1952 there was no doubt anymore if you call it romantic if you want, but death isn't that romantic Bronchitis isn't that romantic Okay, so what's the answer? We're extreme smog episodes romantic or deadly Deadly we got that down right Okay Second controversy well what about not these great big monster remarkable episodes But what about more moderate levels of pollution can it hurt us now we get to go right here to Utah Let's come right here. All right Now this is this is just a little bit south of us here. So this this is in Utah Valley We we're right now are in Salt Lake Valley You go around the point of the mountain and you drop into another valley called, Utah Valley That's where Provo Orem is and right outside of Orem Here's Provo here's Orem right outside of Orem in a little town they called vineyard there was this big steel mill Now this mill was built in During World War two and they build it here because you had a water source and we're far enough from the west coast They didn't think the Japanese could bomb it Literally, that's why they build it there Now it's a big integrated steel mill and was a big polluter So you have a whole bunch of people living in this valley this exposure chamber Okay, and then you have a source of pollution But then that the mill operated until July of 1986 and then a labor dispute shut it down And it stayed shut down for 13 months before it reopened It's just a cool natural experiment, isn't it? So you just shut down about 60% of the pollution in the valley for 13 months and then reopen it All right, that's what that student would have done I probably would have given her an A All right, but she didn't But I did so there's the steel mill This is what it looks like during a really bad day back in those days the p.m. P.m. 10 nearly all of p.m. 10 was p.m. 2.5 because there's no windblown dust here This is all p.m. 2.5 here, but we only had p.m. 10 monitors at the time. You can see Geneva Steel back there Can you see it? Those are that those are smokestacks look underneath there and those are thermal bubbles I don't know whether they really are I made that up But that seems like that's what they ought to be called that hot you can see that sort of hot You know air coming out of the smokestacks kind of give you those thermal bubbles, but this is a combination This is classic smog a combination of smoke and fog Right and it's trapped on the valley floor and remember there's about 250,000 people at that time it's almost doubled since then but these include asthmatic children and elderly etc. Does this affect their health? well This is Salt Lake City just to kind of give you a feel This is another temperature inversion looking down kind of a canyon into into the valley. It's very common very common Occurrence here here along the Wasatch front when these stationary high-pressure systems come in it traps the pollution on the valley floor All right now what happened in Utah Valley meal operating meal closed meal operating That's the pollution bronchitis and asthma pediatric hospital admissions open closed open pneumonia pleurisy open close I don't even have to be a very good statistician to say I ran regressions and did all sorts of fun stuff, but that's all I have to do It's just stunning it's stunning and I just say oh wow you're a real good researcher. No. This is a crazily good Natural experiment wasn't it? And any economist that loves natural experiments and quasi experimental designs Any econometrician loves this kind of thing and I love it It was that was very neat caused me a lot of problems here remember David Bates that at st. Bart's there. He is there's You know Joel Schwartz and and Bart Ostrow Myself that was a better time in a lot of ways Doug Dockery. There was one time when we were we were having a A Workshop right here in in Provo and in Salt Lake and we stopped to look They all wanted to take a picture in front of Geneva Steel all right, so that's Geneva You can't hardly see it Can you kind of polluted back there somebody argued if somebody would just blow up our van it had pretty much Eliminate all air pollution research at the time We did they didn't we lived Now let's think about again. This is the Wasatch Front, Utah Valley So this is a monitor in Linden, which is just right right next to Oram and You can see this is the air pollution micrograms per cubic meter PM 2.5 You see it spiking up spiking up spiking up. This is another sort of natural experiment you still have all those people living in the valley, but you have the expose you have the Temperature inversions putting a lid on it, so you put the lid on and it gets really high concentrations And then then the the inversion breaks And it gets clean and put it on so you have all of this natural variability and exposure It might be nice to look and ask ourselves If we have that natural Variability and exposure could we see if it's associated with some adverse health outcome? So that's what we're gonna. Do what's a good adverse health outcome to look at how about dead or alive? That's not too bad right, so we're gonna look at dead or alive Now this is dead or alive Now dead or alive in a in a community of about three hundred thousand people looks like this on some days Nobody dies, so this is this is daily counts time and days some days Nobody dies other days one person dies two people die three people die. Do y'all see that? Now most of the variability and deaths are going to be in the area of the population Variability and deaths are going to be due to Poissonian variability even if there's no underlying sort of extra cause of death even if every individual death is is into it is is is Independently associated not associated with any others you're gonna get this exposure variability But I have actually drawn a non parametric smooth through the data. Did you see that? That's sort of showing that taking out the Poissonian variability, and you can see other seasonality in the data more people die in the winter less in the summer more in the winter less in the summer and Then you can also see that if I narrow my smooth I can see a lot of within season variability except for what right here not so much. What's going on? The steel mill was shut down The other thing I can do let's don't do any fancy statistics. Let's simply take these counts and Let's divide them up into quartiles The quartiles will be relative to the quartile of air pollution and That's what you get No, no fancy stats But who wants to publish stuff with no fancy stats? So let's do some fancy stats, shall we? Let's do a Poisson regression because those mortality counts are being generated by a Poisson process. And we're going to hypothesize that that Poisson process is non-stationary. It's affected by seasonality, it's affected by time trends, it's affected by temperature and relative humidity, and it's affected by air pollution. So we are going to take the log of the mean of the Poisson regression in a given day, we're going to regress it on some lag structure of pollution, we're going to do non-parametric smooth to control for temperature, I'm sorry, for long-term time trends, seasonality, et cetera. And that won't go into this. We spent a lot of time trying to fight over what's the best way to do that. The bottom line is, is what did we learn? Air pollution, even over a few days, even at relatively moderate levels, certainly compared to the London smog episode, was correlated or associated with daily mortality counts. Now first off, it was pretty controversial and there's all sorts of re-analyses and that sort of thing going on. Nowadays, there's actually literally hundreds of these studies done in hundreds of cities throughout the world. And then this really cool study that happened just a few years ago. Ambient particulate air pollution and daily mortality in 652 cities around the world. Now New England Journal of Medicine had an editorial with it and the article's entitled, Do We Need Another Daily Time Series Mortality Study? Well, it's pretty close to linear. Oh, supra-linear, that's exactly right. And we're going to come back to that in just a minute because that's going to be important. Now here's another study done right here, right here where we're at. So there's this group of cardiologists here at Intermountain Healthcare. And Intermountain Health, many of you may know, is the biggest sort of healthcare system here in the Intermountain West. It dominates Utah Valley, I'm sorry, Utah along the Wasatch Front. I was actually reading, I was just reading a paper and they were looking at some markers that influenced cardiovascular disease, but I realized, oh wow, they've got this really cool, you know, group of well-defined patients where they have not only when they had their ischemic heart disease event, but they also had, you know, they had also gone coronary angiography so I could even look to see if they had existing coronary artery disease. So I got with these cardiologists and we said, all right, let's do what's called a case crossover design. Now you all know case control design. Well, a case crossover design is similar, only the individual serves as their own control. So what you do is you, I don't have a good slide for it, sorry. But what you do is you have the time when they have their event and then you have times when they didn't have their event, okay, as reference times. So it's sort of quasi-experimental that way because you're controlling by design all of the differences across any individuals because there is no differences across the individual and themselves, especially if your reference periods are real close in time. And if they're real close in time, you're also controlling for time trends and seasonality. And if your reference period is the same day of the week, you're controlling for day of the week, et cetera. At any rate, so we did that analysis and guess what we found? And then you do conditional logistic regression, but what do we find? We found that if you have a day or two or three or four days of higher exposure to air pollution, you're at much higher risk of getting an ischemic heart disease event, heart attacks and serious angina, okay? Then we could stratify by whole bunches of things, nothing made any difference except for one clear thing, and that is the only ones that had an increased risk were those that had existing coronary artery disease, which isn't surprising. How could just a few days of exposure to air pollution really have somebody have a serious acute coronary artery disease or artery event, acute coronary event if they didn't have existing disease? And we're going to talk about long-term exposure in a few minutes, okay? This was amazing. I loved that. These results were just great because, wow, this doesn't appear to be some spurious association. It appears to be real. Now, what's the problem? We don't have enough time to go through all the studies, but short-term exposures, even at moderate levels, have been associated with increased daily death counts, hospitalizations, reduced lung function, increased symptoms of respiratory illness, school absences, ischemic heart disease, and acute lower respiratory infection. Question, does air pollution, even at moderate levels, impact our health? Yes. You got it right. You're right. Short-term exposure to moderate levels of air pollution harm us. Now, third controversy then. Well, if short-term exposures have these kind of impacts, what about long-term exposures? How about like years or decades, even? What would you expect? Yeah, I'd think so. Let's ask ourselves real quick. Now, the earliest studies on this were what are called sort of population-based mortality studies, very simple studies. Lester Lave and Eugene Seskin did the first ones back in the 1970s. Basically, what you could see is that adjusted mortality rates across cities were correlated with fine particle air pollution levels, even when you tried to control for the level of smoking going on in the cities and everything else. And these results were pretty coherent. A lot of other researchers said, we get the same results, and published a whole bunch of them sort of in the 70s and 80s. Critics said what? The controversy was that, ah, no, you got all sorts of problems here. These are population-based studies. You have ecologic bias, and there's no way what you really need is prospective cohort studies where you enroll people and follow them up and get the individual data on everybody and follow them up for long periods of time. Do classic sort of survival studies, right? Okay, fair enough, let's do it. It turned out that Frank Spicer and Ben Ferris, now by the way, what's the problem with those kind of studies? They're expensive, they take a long time, and if you have enough sort of gravitas to get the money to do one, you're going to be dead before you've got an all follow-up, right? It's very frustrating, okay? So Frank Spicer and Ben Ferris, they were these Harvard pulmonologists that had enough gravitas to get the data. They got it, they started a study that we now commonly refer to as the Harvard Sick City Study. Doug Dockery, younger guy, started working with them on that. Doug and I had worked together quite a bit, so he invited me to be part of this group. I was actually at Harvard at the time. Bottom line, what they did is they had this 14 to 16 year perspective follow-up of over 8,000 adults that monitored for TSP, PM10, PM2.5, sulfate, hydrogen ion, etc. And then they actually asked me when I was out there, would you be willing to analyze these data and see what we can learn from them? And of course, what did I say? Oh yeah. They said, are you real familiar with the Cox Proportional Hazard Model? I said, oh yeah, I wasn't. But I figured, it's just a regression model, I can figure it out in a week or two, right? I did, it took me about a week or two to get it all figured out and how to do it and do this stuff. And they know, in fairness to Doug and to Frank, they knew that I'd just learn it. It's not that hard. And guess what we found? Wow, it's so elegant, it was so easy. I mean, simple survival analysis. These are survival curves. So the study was set up so you had basically two really clean cities, Topeka, Kansas, and Portage, Wisconsin. Two really polluted cities, Steubenville, Ohio, and St. Louis, and then two cities kind of in the middle. And then you follow them up and you look at them and guess what happens? These survival curves splay out just like you'd expect. They're dying more rapidly in the more polluted cities than in the clean cities. Well, maybe something else is going on. We should model this more carefully, so let's do a Cox Proportional Hazard Survival Model, which anybody can figure out if they study it a bit, right? It's pretty straightforward. So we're going to do it. Here, Cox Proportional Hazard Regression Model, we're going to control for age, sex, race, smoking history, occupational history, etc., right? We're going to do that and then we're going to look to see, hmm, what about smoking? Wow, smoking's pretty bad for you. Not a surprise, but it roughly doubles your risk of mortality. Now, of course, we're all going to die. I know everybody's, oh, the risk of mortality is 100%. Yeah, sure it is, but that's not what we're measuring here, right? We're measuring the relative risk, relative to your baseline, at any given point in time conditional of living to the beginning of that point in time. That's what we're measuring. So we all know we're going to die, we just don't want to die right now. Or we don't want to die during, you know, this season until the playoffs are over or something, okay? That's the way it works and you can see, if you're a smoker, and 25 pack years is about an average smoker in our cohort, your risk of dying right now is roughly doubled. Now, of course, you might have a pretty low baseline risk, so overall still may not be that high. Lung cancer, 8 times higher. Cardiopulmonary, 2.3 times higher. But now look at air pollution, 1.26. So, whereas smoking increased risk by 100%, living in a more polluted city was 26%. You might say, well, that's not so big a deal. Oh, we were stunned at how big a deal it was. This is big. Think about it. In this community, in Steubenville, for example, how many people smoked? About 25, 26% of the adults, okay? So, if you take the, their risk of dying roughly doubles, but it's only 26% of the adults. Now, what do we get here? We get a 26% excess risk of mortality from just living there, breathing there. How many, what percent of it were breathers? Everyone. So, basically, the attributable mortality to air pollution was as high as to smoking. Now, of course, that's because everybody breathed. But still, that was big. We were stunned. Could it really be true that in some communities, air pollution would be as big or bigger a contributor to mortality risk as smoking? It's hard to believe. So, we said, all right, we can't believe it. And we wouldn't even lick the stamp and send it up. That's back when you lick stamps to submit papers. We said, we've got to make sure we can see this somewhere else. Oh, by the way, this is just, so we did the analysis, and then you can do the same sort of analysis, but put dummy variables in for each city and just plot them. And what do you get? A graph that looks like you cheated. Doesn't it? But that's what really happened. You know, you see, this, as cities with higher air pollution, had higher risk of mortality in a near linear fashion. Super linear, you've got it all. No, we'll come back to that later. We said, all right, we were a little worried about this. If it turns out we're wrong, we're going to get slaughtered. So, we said, we want to do an analysis in another independently collected data set. So, we went to Michael Toon and Clark Heath at the American Cancer Society. They were collecting data, you all know the data probably, the American Cancer Society Cancer Prevention 2 cohort. They had over a million people that they'd been following up. We knew where they lived. We could link that with air pollution data. And for everybody that we had air pollution data for, we could analyze it in a similar way. I would fly down from Boston to Atlanta, would use their CDC mainframe computer back then because this model sort of semi non-parametric, it took a long time to run. We'd analyze these data and guess what we found? We found smoking is still pretty bad for you. And we found that air pollution, measured as sulfates or fine particles, that's not quite as big as 26%, but it was still big enough and statistically significant. So, we actually submitted the results of both of these papers. And what happened? Well, even trying to do that, it still was pretty controversial. It was a rough, it was just a crazy time. It turned out that at that time, the EPA was proposing ambient air quality standards for PM 2.5. There were calls for independent re-analyses of our data. There were legal challenges that ended up going all the way to the Supreme Court. I mean, just for example, you know, in science, Jocelyn Kaiser said, industry and environmental researchers are squaring off over studies linking air pollution and illness and what are calling the biggest environmental fire of the decade. We felt we were getting battered pretty good. We turned our data over to a research team overseen by the Health Effects Institute to re-analyze both the Harvard Six Cities and the ACS study, multi-year study. They spent, there were over 30 of them working on it, had a dedicated computer. They didn't have to fly down and borrow a CDC mainframe computer. Ran, ran, ran, did audits and reproduced the original results and conducted robustness analysis and guess what they found? Basically the same answers. It didn't change hardly a bit. A year later, the Supreme Court ruled unanimously that allowed basically the establishment of PM2.5 standards based on those studies and time series studies. Now after that all happened, it took several years, Rick Burnett happened to be one of the investigators. So after it was all done and they released the report, I called Rick. I said, you did a great job. And they really did. They did a fantastic job. And he's a fantastic mathematical statistician. I said, if we work together, we can do the best job possible. So we said, all right, let's work together. We had a longer follow-up. We were working with, you know, so Rick Burnett, Michael Toon, Eugene Cal, you know, Dan Kruski. We had a real good research team and we did it, we did it again. Published this in JAMA. What did we find? Same sort of thing, although we had cooler. Now we can look at non-parametric smooths and that kind of stuff. But bottom line is increased air pollution, increased the risk of, sorry, my thumb's getting it. Increased air pollution, increased the risk of all-cause mortality, but driven by cardiopulmonary and lung cancer mortality. Now I don't have time to do this, but I will say over the years with the Harvard Six Cities study, there's been that independent reanalysis that we've already talked about by HEI. Then there's been a number of extended analyses of the ACS cohort, a number of extended analyses of the Harvard Six Cities cohort. I won't go into the details here. The next criticism was, we can't use these because they're secret science. Yeah, we'll get to all secret science. We can't use them. I mean, we were publishing them in New England Journal of Medicine, American Journal of Respiratory Critical Care Medicine, JAMA, New England Journal of Medicine, Lancet, you know, et cetera. Hardly secret, right? Why was it secret? Because IRB required that we give confidentiality agreements with all of the participants in the studies. And we never could release the raw data. Now, that issue has passed pretty much, but it took a long time. Here's a fun one. This was a whole bunch of other studies have now been done. It doesn't all rely on these Harvard Six Cities and ACS study. Here's one that was done by Joel Schwartz, Francesca Domenici, and their research team at Harvard. They followed up 60 million, 61 million Medicare beneficiaries. They had almost half a billion person years of follow-up. Guess what they got? Same results we got in the ACS study. But the biggest difference is, look at their confidence intervals. Well, if you have a half a billion person years of follow-up, you're going to have pretty narrow confidence intervals, aren't you? It's pretty amazing. Now, there have been a whole bunch of studies, and I won't go through this. There's now, depending upon the way you count them, but there's been about 30 some odd cohorts and about 75 different studies that have been done. Here's the Harvard Six Cities study with various follow-ups. Here's the ACS with various follow-ups. I'm just showing you the hazard ratios and 95% confidence intervals. Here's the Medicare cohort. Here's one. I won't show you this, but one I've been working on the last four or five years, the National Health Interview Survey data with cases. A big research team with Carnegie Mellon. Again, a bunch of studies coming out of Canada and elsewhere. You put those all together, you do meta-analysis, and guess what results you get? Pretty much the same thing as we saw in the ACS study. Smaller results than what we saw in the Six Cities study, but slightly bigger than what we saw in ACS. All right. Here, this is just for fun. Doing this again with the most recent and modern studies from Harvard, from the Six Cities, and then from the National Health Interview Survey study, and then the big study from Canada. They call it the Maple study. Isn't that cool? The Europe, the ELAP study, studies from Hong Kong, Mainland China, et cetera. So, what's the answer? Can long-term exposures contribute to disease and loss of life? I shouldn't have put it up. You can, don't look. It's cheating. The answer is yes, isn't it? So, next controversy. What is it? Well, does reducing air pollution improve health and reduce mortality? Well, it should. Here's a fun study. This is a big, great study coming out of Southern California, Southern California Children's Health Study. It would take hours just to go through all of this. But the bottom line is, yeah, when you clean up, have less exposure to air pollution, children's lung health is improved, both in terms of development of lung function and development of respiratory disease. Harvard Six Cities study, with a follow-up, a farther follow-up. Francine Layden did a really nice analysis with an eight-year follow-up. So, what happened is, a further follow-up saw a big drop in air pollution in Steubenville and look what happened to mortality risk, dropped, Kingston, dropped, St. Louis, dropped. And even the two cleanest cities, air pollution went down and mortality, adjusted mortality went down. Question? Was there a specific industry that were contributing the most to the air pollution at that point? Yeah, it depends on where you were at, right? But no question, the steel industry, coal-fired power industry, those were the big polluters, although in some places, like, it wasn't in the six-city study, but places like Riverside, California and things, it was mostly traffic. So it depends on where you were at. But the big sources of PM2.5 are high-temperature processes. So think smelters, think, you know, steel mills, think copper smelters, think coal-fired power plants, think diesel vehicles, think gasoline vehicles. Does that make sense? We did a study, in fact, what was going on, I'll show you real quickly here in a second that relates to your comment. But what happened in the, after the Clean Air Act, at least the amendments to the Clean Air Act in 1970, by 1980, there started to be real teeth. They really started to enforce the National Ambient Air Quality Standards. So cities that were out of compliance with the National Ambient Air Quality Standards, they cleaned up a lot more over the next couple of decades. So between about 1980 and 2000. And Doug and I decided, oh, we're going to take that data and we're going to basically test the hypothesis. Were there differential reductions in air pollution that occurred between 1980 and 2000? Were they associated with differential improvements in life expectancy? Again, sort of treating this differential changes in exposure as a natural experiment. And we're going to use differences and differences sort of analysis, that kind of quasi-experimental approach. Well, the bottom line is, is, yep, they did. Every 10 microgram per cubic meter decrease, sorry, my hand keeps hitting the wrong buttons. Every 10 micrograms per cubic meter decrease, depending upon the model, but it would get you about a .6 to one year increase in life expectancy. Which would have been probably the second most effective medical intervention that occurred over those 20 years. Now Stephen Colbert did a comic routine on this. It's kind of fun. He talked about, you know, wow, this study that came out of, you know, Brigham Young University and Harvard showed this. He said, what's the problem? He says, well, it's good you get more life expectancy. You have enough to watch one more season of Ghost Whisperers. He also said, what's the other problem? Well, you have cleaner views of Pittsburgh and Buffalo, you know, that kind of thing. And then he introduced a life bag that would save us all. We could eliminate, we could live forever, right? We put that life bag over our head, scrunch it down, and it protects us from all the particles. Kind of fun though. There, you can look at that. So where was the pollution coming from? Well, Pittsburgh and Buffalo and Cleveland, mostly industrial pollution, okay? But that wasn't true in Portland. That wasn't true in Boise. Those are more, a lot of biofuel as well as vehicle traffic. Okay, I'm going to run out of time. I'm going to skip this one except for to say, ah, there was a copper smelter strike in the 60s. We could use that as a way to look at sharp discontinuities. And what happened? This copper smelter strike resulted in about a two to four percent reduction in mortality. An amazing study out of China. There's the Huai River Valley that runs east and west, and China allowed for free or highly subsidized burning of coal for space heating and things north of the river, but not south of the river. So you have this geographic discontinuity, and you can use a regression discontinuity analysis, and this is what Michael Greenstone and colleagues did, and basically what you found is, wow, there's this big discontinuity right at the river in terms of air pollution and a big discontinuity in terms of life expectancy right at the river. Very fascinating study. So does reducing air pollution improve health and reduce mortality? Yes. Second controversy. Does a safe threshold even exist then? You know, this is getting a little bit to what you're worried about. The Clean Air Act basically assumes that there's a safe threshold, and we should establish national ambient air quality standards so that we're protecting everybody. The problem is, is there is no evidence of such a thing. Here's data that Joel Schwartz and Bart Ostroh used from London looking at the exposure response relationship. This is for short-term exposures, but you can't, do you see a threshold? You see super linear, don't you? What do you see here? Well, these are a whole bunch of time series, London, Detroit, St. Louis, Utah Valley, Sao Paulo, Philadelphia. I'm involved with about half of these studies, and others not. Here's a whole bunch of sort of meta-smoothing. We don't see any evidence of a threshold. If you look at long-term exposure, same sort of thing. We don't see any evidence of threshold. So now we got a problem. Does a safe threshold exist? Well, not in the ranges that we've studied. It appears to be near linear. Then we end up with another controversy. If PM 2.5 effects are this big and near linear, why isn't everybody who smokes dead? Think of it. It's really true. Look at this for just a minute. So here is the excess risk of mortality of about two for smokers, and their exposure is about 240 milligrams of PM 2.5 daily. But for secondhand smoke, that is living with a smoking spouse or working in a smoky environment for secondhand smoke, or air pollution, the dose of the inhaled dose is only, it's actually less than one millimeter, or I'm sorry, milligram. So that's a huge difference. Now if I were to do a linear extrapolation from here down to here, this suggests that these should be almost trivial, the excess risk, right? Or if I were to run a line up through here and extrapolate it out to cigarette smoking exposure, everybody ought to be practically dead, right? How could that be true? Well we said, all right, let's take our American Cancer Society CPS2 cohort data, and let's actually look to see if there's evidence of a linear exposure response relationship that goes through the origin. And what we find is, is that we get a linear exposure response relationship, or pretty close, but it doesn't go through the origin. And it looks like there's something magical about that first puff of cigarette smoke. All right, I'm just going to go, I'm going to skip. So we've spent a little, we've spent several years doing this analysis, did what we call an integrated exposure response function, where we're integrating information from active smoking, secondhand smoking, and air pollution. And what you're doing is you're looking at all that put together, and what we see is, is that the exposure response relationship appears to be really steep early on and then flattens out, just like what you were saying. Or if you, I'm going to plot this same thing over here, but just plot it over a log scale. That looks like pretty coherent results. So why is it that everybody who smokes not dead? Well because the exposure response relationship is not linear. There's this saturation phenomenon. There's these diminishing marginal effects. That happens all the time in biology, right, absolutely, all the time. And so there's no reason to a priori think it has to be linear. There's no reason a priori it has to be log linear either, but that's just reasonable as linear. So the bottom line is, that could be the case. Now I haven't spent any time here, but it might be true that both air pollution and secondhand smoke are more toxic than smoking, than exposures from active smoking, at least on the margin. All right, seventh controversy, we're going to do it. Are these health effects biologically plausible? I should have you answer that right now, but you already know where I'm headed, don't you? You know that. Well the answer is, yeah. So go back to Geneva Steel, that's just down the road. Found out a bunch of toxicologists and medical researchers came and collected from the state of Utah some filters that happened when the mill was operating closed and then operating again and extracted from those filters and then put those in saline solutions and put polluted saline solution in one lung and inserted it in another lung and then did bronchial alveolar lavage and guess what you saw? You had a lot more pulmonary inflammation if you're exposed to the pollution. That was true in vitro and in vivo, human, animal, and in vitro studies. Starting in about the early 2000s, there became this increasingly compelling evidence that inflammation is a major accomplice with LDL cholesterol and the initiation of cardiovascular disease. You all know this. It became very clear. If you have more inflammation as measured by high-sensitive C-reactive protein, more systemic inflammation, you have higher risk of cardiovascular disease. That's also true if you have more cholesterol, but look what's particularly deadly, a combination of the two. This is now well known. There's nothing fancy about this, so what's going on? Well, you breathe fine particles, secondhand smoke, air pollution, or active smoking, and you get more pulmonary and systemic inflammation, more oxidative stress, along with blood lipids, you get more rapid progression and destabilization of atherosclerotic plaques. Cat studies, they find exactly that. Mouse studies. I should show this one, because the developer of this hyperlipidemic mouse got the Nobel Prize, and he's right here at University of Utah, here in Salt Lake. At any rate, very cool study. Which kind of artery would you like, one that looks like that or looks like that? Well, eat high-fat chow and breathe a lot of pollution, and this is what they look like, if you're especially a hyperlipidemic mouse. Could we do these sort of studies with real people? Uh-uh. IRBs won't let us get away with it, but that's what it looks like when you look at real people. We've done a lot of studies trying to figure out, I mean, this happens to be a scientific statement of the American Heart Association that a group of us did, trying to look at these different biological pathways. The simplest one is sort of easy to think of. You all know what the endothelium is. Endothelium, it turns out, do you know what kills most people in the world? Endothelial disease, right? It's endothelial disease that kills us. The endothelium is that inner layer of the blood vessels. As it gets inflamed and as it builds up and gets atherosclerotic lesions and things, what does it cause? Coronary artery disease, ischemic heart disease, eventually heart failure. It also includes ischemic strokes, peripheral arterial disease, which also results in pulmonary embolisms, et cetera. You add all those together and that's the biggest cause of death. And what's the disease? It's the disease of the endothelium that manifests itself in ischemic strokes and ischemic heart disease and peripheral arterial disease and pulmonary embolisms, which is consistent with our epi evidence. Just very quickly. We did a study again right here in Utah, it's kind of fun to use our Utah studies, using Aruni Bhatnagar and Tim O'Toole, this was down in Utah Valley. What we did is we enrolled a bunch of subjects, 72 healthy non-smoking adults from the BYU. Now they don't smoke, they don't live with smokers, this is all air pollution. We took multiple blood draws over a period of three years, we processed the blood, we shipped it to Louisville, University of Louisville, where you did, you know, measured microparticles in immune cells using this multi-laser flow cytometer. We also sent cytokines and markers of endothelial adhesion to Eve Technologies to measure those and then they were sent back to me to do analysis. Just to show you again, remember what we have, we have these inversions that happen, and so I know when an inversion's coming, so I like to get them in. We can't control their exposure, but we can control the timing of the blood draws. So we want to do the timing of the blood draws when the exposure is high and when it's low. And you look at what you see over the period of three years, woo, really high, moderately high, moderately high, low, low, low, moderately high, et cetera. Do you get that? And then we analyze it and what we do is we look at endothelial microparticles. Now endothelial microparticles are an indication of vascular damage, all right? The hypothesis was more pollution, more microparticles, what do we find? Exactly that. We also saw more inflammation. So we have evidence of subclinical damage to blood vessels, evidence of nonspecific immune responses. We also have these cytokines. I won't go into the statistics here, but basically you can line these up and look at those where positive associated, negatively associated, do a Bonferroni correlation for multiple testing and you can say, wow, there's some that are really correlated with air pollution. And then if you do more fancy stuff, you can see, ah, what's happening is that we're getting... Everybody here knows TNF-alpha. Heck, drug companies would be going broke if it wasn't for TNF-alpha inhibitors, right? What do you see? Air pollution contributes to more vascular damage, less vascular repair. All right. So question, are there biologically plausible mechanisms? The answer is yes, but do we understand them all? Not even close. We don't understand with cigarette smoking either, did you know that? How do we know cigarette smoking is bad? Well, we look at cigarette smoking and say, oh, there's lots of nasty stuff in that there smoke, right? And that's true with air pollution too. Now, we're almost there. Aren't air pollution health effects just small compared to other more important risk factors? The answer is clearly, no, they're not. Here's a study of the global burden of disease. Air pollution is basically, for both females and males, the fourth largest risk factor for global burden of disease in the world. And this is a great study. Here's a study of Michael Greenstone. His estimates are even bigger than that. I actually think he's a bit wrong. I think he's overestimating it a bit, but he's arguing that particulate air pollution is actually has a bigger global burden of disease impact than smoking. Why is that? Because everybody breathes. And a lot of people in India, in China, in Pakistan breathe a lot of bad air pollution too. That's what's going on. So I'm not going to go there. We've done analyses trying to figure out how it looks across the world. So doesn't cleaning up air pollution cost too much and hurt our economy? The answer there is easy, no. Why is it? Because think of clean air as one of our economic choices. Clean air is an economic good that contributes to human well-being, human capital. We don't argue about, don't TVs cost us a lot? Well, yeah, they do, but we like them. Doesn't air pollution cost us a lot? It's the fourth largest contributor to global burden of disease. What's more important than living? The production of clean air can contribute to economic prosperity and human well-being. And in fact, this is just in the United States. But since the Clean Air Act Amendments in 1970, what's been happening to our air pollution? It's been declining. Must be destroying our economy, right? Hardly. Most domestic product has gone up by over 250% since then. All right. Last one. How much evidence is needed? This is the last controversy. How much evidence is needed before efforts to clean up our air are no longer controversial? Okay. I'm going to end with a sort of an optimistic thing, because we're actually doing a pretty good job in some places in the world and not so good in others. I'm going to end with a negative thing, because the answer to this question is what? Even if we knew the costs with absolute certainty, it would be controversial. Why? Because a lot of the controversy has nothing to do with the science. It has to do with who gets the benefits and who pays the costs of air pollution. That would be another interesting lecture. Thanks. I'm going to stop now. Appreciate it. I told you. I got through every one of them. That's my last slide. I will take a question or two if you want, but I know we're out of time, so if you raise your question and people throw stuff at you, it's your fault. Question. One of the best talks I've ever heard, so I want to thank you. Oh, thank you. What do you think about, you see the air pollution in Beijing and in China, and you just think that there must be just massive amount of excess mortality, or is it that they're just on the higher end of the slope, or higher end of the curve and there's not as much slope? How come they're not reacting to it? How come their exposures are going down? In fact, just before this lecture, I was looking, I was requested to review another study out of China. The studies are really starting out of China, and they are seeing these effects, just like we'd expect. India, they don't seem to give a hoot. Pakistan, they don't seem to give a hoot. But I mean, who wants to starve? There's other risk factors, too, but air pollution is a big one, so I don't know. I mean, we still fight over it here. I mean, it's sort of nick and tuck. I mean, the last administration basically had his air pollution folks were deniers that air pollution, and deniers that air pollution, either greenhouse gases or these pollutants I'm talking about here, have any impact on the environment, hardly at all. It's a very strange world we live in, okay? Thanks for the question. We're done. Thank you very much. Thank you. Thank you.

air pollution

health effects

evidence

respiratory illness

mortality

hospitalizations

lung function

biological plausibility

inflammation

cleaning up