Great, so welcome to the annual Current Research in Occupational Environmental Medicine and Resident Research presentations. This is an annual event. I am very reluctant to tell you how many years I've been doing this here. I'm very also, even more reluctant to tell you what year I was up in front of this group giving my presentation, my senior year of residency, so I'll let that slide, but we're very grateful to AECOM for their support of this session, their support for travel for the residents, and to let our residents, faculty, and AECOM membership share in what's current, ongoing, and new within the field of research in occupational and environmental medicine. So welcome, and I'll just say what every other speaker says, it's really good to be back and in person and seeing all those faces. So without further ado, I'm going to move on here. I'll just have a couple of ground rule sessions. We've got a podium here for the speakers. We've got a handheld, but remember we're also a virtual session too, so that we may get questions from the virtual audience at home or in their offices, and so the room will actually be wired to pick up your questions from the audience so that you don't have to come forward and pick up in the front mic, and I usually ask people to do that there, and again, guess what? There are no microphones in the aisles, but we moved on in the technology, and this will pick up here. So if you have a question, sort of stand up, state your question clearly, and that will be for the benefit of the recording, the home-off people. Yes, and there is actually a ... Oh, okay, great. So our helpful staff will bring people in the microphone and the questions. So I think without further ado, this is a little unusual. Usually we'd start this at 8.30 in the morning, and we'd go until lunchtime for reasons unknown to me, but probably better for the chance to kind of get refreshed here. We have a split session. We're going to have four presenters in the morning and four presenters in the afternoon with a big break for lunch. Just make sure you don't drowse off after lunch. We've got some really good presentations coming. So this is the preview of coming attractions. We've got four on here, and without further ado, I'll welcome our first speaker, Dr. Nick Blonion from the Health Partners Occupational Residency in the University of Minnesota in Minneapolis, going to speak to us on cardiovascular disease. So welcome, Dr. Blonion. Thanks, Gary. Let's give it a try. Oh, title slide. All right. Thanks for coming to listen, everyone. I'm Nick Blonion, Occ Med resident from Health Partners, like Dr. Meyer said, and I am presenting today on cardiorespiratory fitness in firefighters. So the presentation is going to cover how we might use a model for estimating VO2 max to assess firefighter fitness and cardiovascular disease in this population. Kind of the disclaimer slide. The research I'm presenting was funded by Hologic, which was the manufacturer for the DEXA scans used to assess the body composition in firefighters. Authors have no financial disclosures with the company, and the research has been approved by Health Partners IRB. So why are we talking about firefighters, or why are we talking about fitness in firefighters? Well, they have strenuous job demands, obviously. And the high levels of fitness required to meet these demands, it needs to be done safely and effectively. So despite this, many firefighters are not models of high cardiorespiratory fitness. You can picture the calendars of firefighters. They're probably not as healthy as they look. So in fact, nearly half of firefighter line-of-duty deaths are due to sudden cardiac death. In medicine, we often use BMI as a quick or crude indicator of overall fitness. And while we know that this is a mediocre at best measure of body composition in the general population, in individual firefighters, it may be even less useful, given that firefighters may maintain high levels of muscle mass relative to the general population. Using BMI, 77 to 90% of firefighters are classified as overweight or obese, and this is relative to 73.6% for the general population. So BMI may not be useful at all in assessing firefighter overall exertional capacity or fitness for duty. And using this metric might ultimately result in misclassification as unfit, which could cost a fit firefighter his or her job. Alternatively, normal BMI should not be reassuring of adequate cardiorespiratory fitness. And if it is used in this way, it may expose an unfit firefighter to the exertional stress that might increase risk for sudden cardiac event. We need to find a better way to estimate cardiorespiratory fitness in firefighters. Cardiorespiratory fitness has been proven a useful predictor of work performance and health. And the best metric for measuring this is VO2 max, which is defined as the body's ability to take up and utilize oxygen at high levels of exertion. The National Fire Protection Association recognizes the importance of fitness in firefighters and recommends that they maintain a capacity for 12 METs or a VO2 max of 42, which is about the effort required to walk five miles an hour at 5% incline with 20 kilograms of weight for the average person. Note that there's no age or gender adjustment built into this recommendation. And I've made a table of ballpark estimates of VO2 max, which are based on the American College of Sports Medicine's classification. I'll note that this table is my own kind of summary display of very rough averages for a 20 to 40-year-old age group and not a replication of ACSM's published numbers. So key questions our study aimed to address were, are other body composition measures more useful indicators of cardiorespiratory fitness in firefighters than BMI, such as fat mass percentage? Can we develop a model using fat mass percentage or other body composition measures to cost-effectively estimate fitness in firefighters? And ultimately, would this model be useful in estimating cardiovascular disease risk? So we took a phased approach to addressing these questions. In phase one, we sought to identify which body composition metrics best correlate with BMI and VO2 max, and we developed a multivariate model to better estimate cardiorespiratory fitness than traditional methods. In phase two, we explored how well our model explained variability in cardiovascular disease risk and with additional enrollees and therefore larger sample size, we looked to improve the goodness of fit of our phase one model for VO2 max. This is a cross-sectional correlational study that looks at career firefighters employed by a fire department in the Twin Cities metro area. In phase one, firefighters were assessed during three separate sessions. Session one included enrollment, a survey to collect demographic information, and a subjective report of physical activity level. There was a physical assessment to collect traditional body composition measures and a 2,000-meter timed row on a Concept 2 rower, which some of you may have, to estimate VO2 max, which used a calculation established by the manufacturer to calculate VO2 max. Session two was treadmill-based testing of VO2 max using Bruce protocol, which is the gold standard for measuring VO2 max. Session three consisted of DEXA scanning, which is the gold standard for body composition measurement. At this point, data from phase one was analyzed and published as initial results from our phase one in 2021 in JOEM. Then in phase two, we enrolled additional volunteers to complete phase one evaluation, so repeated phase one evaluation on new enrollees. We collected blood pressure and lipid measurements on all study participants so that we could ultimately calculate Framingham risk score to represent cardiovascular disease risk. Phase one had 49 male and four female volunteers. Body composition and VO2 max results are shown. Worth noting on the slide is that mean measured VO2 max for both men and women were in the average to above average range for the smaller sample size. Also worth noting is that the rower-based estimate underestimated measured VO2 max. You'll see that there are estimations for trained versus not highly trained estimations, and this is because the concept two calculation actually factored this in based on participant report of their own fitness level. The table on the left shows that fat mass percentage was not strongly correlated with BMI, yet waist circumference was. Our table on the right shows results from a univariate linear regression demonstrating that fat mass accounts for most VO2 max variability of body composition measures we looked at, and that it actually outperforms BMI, which wasn't surprising. From these phase one results, we concluded that BMI is indeed not a useful indicator for cardiorespiratory fitness in firefighters. And of the body composition measures, fat mass percentage has the strongest correlation with VO2 max. After running multiple multivariate regressions using variables that demonstrated high correlation with VO2 max and individually accounted for significant amounts of variability in VO2 max, we found that a multivariate model using age, gender, bone mineral density, fat mass percentage, and estimated VO2 max using the not highly trained calculation achieved an R squared of 0.7. Seventy percent of the variability in measured VO2 max can be explained by our model. And these results, again, like I said, were published in JOEM in early 2021 with the title of the paper listed there. So as a reminder, the intent of phase two consisted of enrolling additional firefighters in the study to hopefully validate or even improve the goodness of fit of our phase one model in estimating cardiorespiratory fitness. Again, this phase consisted of repeating phase one's analysis on the new enrollees, collecting blood pressure measurements and lipid profiles for all participants and calculating the Framingham risk score. The overall goal of phase two was to see if our phase one model stood strong in explaining variability in VO2 max with the increased sample size and if the model could be used to explain variability in cardiovascular disease risk. Finally, we were hoping to assess whether DEXA and Rohwer-based VO2 max estimates are clinically useful when evaluating fitness and cardiovascular disease risk in firefighters. Of note, not all study participants completed all measurements, and this table shows that. COVID was disruptive to the data collection process and it fragmented our assessments a little bit. So things to note on this slide, we had 90 participants who completed all assessments needed for a multivariate regression with Framingham risk score as the dependent variable, and we had 106 participants who completed all assessments needed for a multivariate regression with VO2 max as the dependent variable. So our increase or our assessment of the predictiveness of model one for VO2 max from phase one increased from 52 to 106. The results of the phase two assessment are displayed here. Most important item to note from this slide is the inclusion of the SAFE and HUNT-3 VO2 max prediction scores, which are based on questionnaires assessing physical activity level. Note that these are both estimates of VO2 max based on participants' subjective report of their own physical activity level. They are non-exercise estimates. The table on the left displays results of a univariate linear regression using various body composition measures as independent variables and Framingham risk score as the dependent variable. The take-home point here is that waist circumference accounts for the most variability in Framingham risk. The table on the right displays results from a univariate regression using various biometrics as independent variables and Framingham risk again as the dependent variable. The take-home point here is that cholesterol ratio accounts for the most variability in Framingham risk score and that estimated VO2 max actually outperformed measured VO2 max. Using our phase one model in a multivariate regression with Framingham risk score again as the dependent variable, we found that our model accounted for 73% of the variability of Framingham risk. We ran several regressions with additional covariates and found that we could not improve the model. Our R-squared remained the same. We again used phase one model, now with a sample size of 106 versus 52 like we discussed, in a multivariate regression with true VO2 max as the dependent variable, and this is the display on the right. We found that our R-squared actually decreased from 0.7 to 0.64 as our sample size increased, which was disappointing. In addition, other variables improved the goodness of fit only slightly, and it's worth noting that these are unadjusted R-squareds, so the trivial increase might actually be expected. So concluding, a multivariate model that includes age, gender, fat mass percentage, bone mineral density, and a rower-based VO2 estimate might be used to help estimate cardiorespiratory fitness and cardiovascular disease risk in firefighters. Next steps will be to explore whether firefighters find that having this additional info is helpful and whether or not it helps drive behavioral change like exercising more, eating better, losing weight. We might also assess whether the model might predict firefighter performance in skills testing such as tunnel crawls, stair climbs with gear, or even mask mazes like they have in the Twin Cities area. Finally, we might research the relationship between cardiorespiratory fitness and other health factors such as sleep quality, eating habits, toxicological exposures, and more. Thanks a lot. Thanks very much. It's always nice to have both an excellent presentation as the first one, but one that finishes on time as an example to everyone here. That's why we picked you. Exactly. Probably Zeke told me. We've got time for a question or two. There's a circulating mic for the questions, sir. Test. All right. Just a quick question. Is fat mass percentage the same as body fat percentage? Yes. Okay. I've just never seen it written that way. Thank you. That was easy. I told you they'd be softball questions, Dr. Hines. Thank you. Great talk. I actually have two questions. Hopefully, I can answer them. The first question is, why do you think the estimated VO2 outperformed the measured VO2 in predicting Framingham risk? Great question. I actually asked our lead faculty about that on the paper. Like me, he didn't know. We're using cardiorespiratory fitness as a marker of cardiovascular disease risk. If we're hoping to estimate that metric, it probably should agree most with cardiovascular disease risk, right? Our estimate from concept two ROAR also factored in additional variables. If that could simulate maybe a multivariate regression of estimating or accounting for variability in Framingham risk because of the additional variables that that calculation included, that may be why. Thank you. Can I ask a second question? Okay. Second question. How do we incorporate this fat mass variable into future assessment of fitness, like in cardiovascular risk? What do we do with this? Fantastic question. We talk about this all the time at HealthPartners. Our intent actually was to assess whether or not this is cost effective. We would, using DEXA to find fat mass percentage and using that in our model, including the ROAR-based estimate, would all that be more cost effective than actually having someone measure VO2 max by a treadmill? Goals for our next phase will be whether or not a firefighter knowing his or her body fat percentage or knowing their actual VO2 max drives some behavioral change. Clinically, if that results in firefighters getting more motivated to make changes and then ultimately becoming healthier, perhaps that could drive a change in cardiovascular disease risk. Great. Thank you. Did I see one more question? Just a comment that it's time to bury BMI. Totally agree with that. If somebody is six feet tall and weighs 185 pounds, they're overweight, I don't think you want firefighters with those dimensions. It was thought up by a mathematician over 200 years ago. Biology doesn't happen in 20, 25, 30, 35, and 40. It happens more confluently, in my opinion. What do you think of that? 100% agree. Our faculty talk about this all the time at HealthPartners. One of them in particular is always pretty fired up about it. We all agree. BMI is kind of an archaic measure. In DEXA scan in the Twin Cities, you can get for $100 and even cheaper for research purposes. That might be helpful if it ultimately provides us a more valuable metric for body composition and fitness. Absolutely. All right. Thank you. Thanks. Awesome. Great. Always good to be off to a good start there. We have one comment, which is Zeke McKinney saying, yay, go Dr. Blanian. Your PD. He's your PD. No, he doesn't have to. All right. Thank you. Moving on. We've got one. There were several presentations that were done with the collaboration of our friends at OSHA and the Office of Medicine and Nursing there. I forgot to mention at the outset, and I'll probably say it again, we're grateful to them for the resident experience. Our first presenter from there is Dr. Romero Santiago from the Yale Occupational and Environmental Medicine Fellowship Program. He's going to talk just about carbon monoxide exposures and back extrapolation. Come on up, Dr. Santiago. Thank you all, everyone, for being here. As Dr. Meyer mentioned, I had the opportunity last summer to work with Federal OSHA, the Office of Medicine and Nursing. Even though it was a virtual experience, it was truly phenomenal. In terms of conflicts of interest, no conflicts, I just wanted to mention my training grant that I have through my fellowship for this experience. In terms of introducing this topic, carbon monoxide, it's one that is unfortunately common in terms of inhalational hazards that leads to morbidity and mortality, non-occupational as well as occupational. The issue really with carbon monoxide is the way that it disrupts our human body's ability to transport oxygen, where carbon monoxide, you know, 200 to 250 times, you know, higher affinity than oxygen in terms of the red blood cell binding. And so that presents a lot of issues just on a biochemical level, and then you extend it so many different levels. And so the reason that this topic really, you know, comes to fruition is not only does it still remain a major public health issue, but the other component is that within OSHA, it's serendipitous that I'm giving this presentation here in Salt Lake City because OSHA's Salt Lake Technical Center that has been providing support for over 20 years in doing these back extrapolations, really there was not as much documentation of what that experience has been, how that experience has led to helping OSHA field investigators and field offices offer citations for employers that are above the PEL, the permissible exposure limit. So this is a multidisciplinary team that OSHA engages in for these investigations, and they have this long history. So we wanted to take both of those reasons together and really look closer into what the data looks like. So in terms of the variables, this is the list of variables that the Salt Lake Area Technical Center utilizes. And without getting into the weeds, they utilize the Coburn-Forster-Cain equation, and this is one that even other branches such as the Army have utilized to – they've done it the forward way to see, like, from carbon monoxide exposures what the carboxyhemoglobin blood level would be, and we were doing it the opposite of back extrapolating. So this is the list of all the variables that are utilized. And so in addition to analyzing this data, we wanted to take the opportunity to really see what the common causes of carbon monoxide exposure are and its characteristics in multiple different angles, which I'll walk through. And we also then wanted to take it a step further to document the results of OSHA inspections, where the Salt Lake Technical Center's methodology was utilized to see how it's on the ground helping area officers enforce citations. And then on a holistic level, see how we can guide prevention efforts and inform the way that OSHA's website has this information and ways that we increase awareness and education on this topic. And so for the methodology, so we had to do a couple of things. One is the Salt Lake Technical Center houses this data. Like I mentioned, they've done this for over 20 years of back extrapolation. So they have a data set. And what we utilize internally is OSHA's information system called OIS, which is the next generation of their information management system. You may ask why it changed. And one of the things is that OIS, as a newer system, includes other programs such as the Voluntary Protection Program that OSHA has for recognizing employers who actually do a good job of maintaining healthy workplaces in the context of carbon monoxide as an example. And so utilizing this new version that's really trying to bring all the elements together, we had to see what Salt Lake had and see what the internal database had and try to blend the two together. So if we look at the years 2007 to 2020, there's 270 total inspections that Salt Lake had provided us that they have information from. I had to look in the OIS search to see in terms of, say, region, employer, things like that, what the comprehensive information looked like. Unfortunately, that comprehensive information only started in 2011. So that narrowed my search to 2011 to 2020. And after looking through the database to see, you know, which inspections have more complete information to really analyze all these different dimensions, the final count was 147. So that was the final count that I utilized for the rest of the talk. And so what I'll walk through is the different angles that we took in terms of analyzing this data, you know, starting with the distribution, and I'll walk through each of these. So just at the outset, this was the distribution from 2011 to 2020 in terms of the year that the inspection was filed. And you can see the trend there. So there's a peak in 2015, 35 carbon monoxide-related inspections. And this is nationwide. So in terms of carbon monoxide, so the enforceable limit in terms of the permissible exposure limit is 50 parts per million, 8-hour time with an average. So I wanted to take the chance to see what it looks like in terms of how many inspections in terms of citation, you know, how many are under the Pell versus over the Pell. And you can see the majority are over the Pell by twice, even four times the Pell. So it was a total of 335 blood samples that from the data and the inspections that I narrowed down to that Salt Lake City had done back extrapolations from the blood carboxyhemoglobin level. So the other angle we took was to see the source of carbon monoxide and the rank list. And there are papers that, for example, Washington State has done a paper, Reeb Whitaker in 2005. We have other papers that are looking at what are the common sources. And here from this data that we have, we can still see that devices such as the forklift used in construction and manufacturing still remain at the top. But you also see other gas-powered saws and concrete saws that are used, pressure washers. One interesting one that I'll mention later is the propane-powered floor machines that are used for maintenance of floors, removing asbestos-containing tiles. That was the source that was listed in the top. Then from a seasonality standpoint, this has been discussed in other papers, we wanted to see what it looked like with this data. We utilized the standard definitions for winter, spring, summer, and fall. And we can see that the cold climates in the winter still has the most number of inspections. And we could think, obviously, there are various reasons for that that we can discuss later. But this was the rank order by season. And then also geographically, we want to look at what this looks like. And so I included two different graphics. The table on the left includes just the federal regions and where that lies. But we can see here, colder climates, the Midwest and the Northeast are at the top in terms of number of carbon monoxide-related inspections that receive citations. And then from an industry standpoint, we also wanted to look by NAICS code, the two-digit NAICS code, to see which industries are producing most of these carbon monoxide inspections. So we see that construction, it's one of the ones that's most known for carbon monoxide-related fatalities, morbidity, but still remains at the top when we look at these years. But we see other industries, manufacturing and transportation and warehousing, are ranked pretty high here. And so in addition to construction, there's just many more industries that this public health topic really touches. The other angle we wanted to look at is, like I was mentioning with this data, how our OSHA area office is able to act on it, and in terms of standards that OSHA has for citation. And out of these, you can see we kind of did it by broader category, like general industry construction. But we then took a microscopic look at, for example, the air contaminant standard. That's an important one in terms of directly it mentions, you know, it needs to be below the PEL of 50 parts per million for the eight-hour time we averaged. So that was one we looked at closely. But we also saw that the construction safety and respiratory protection programs, there were citations given for that, too. That was not just about the airborne level, but also the protections in place. The other challenge with carbon monoxide is that the symptoms are very nonspecific. And so we wanted to see if, in this data, we noticed any kind of systematic differences there. And really, it was consistent with other studies that the reported symptoms were typical, headache, dizziness, loss of consciousness, nausea, vomiting, chest pain. So if really someone is not thinking broader, they might miss that, you know, in terms of if they really don't ask about work and what happened when a physician is seeing this individual in the emergency room. And so with this, we wanted to look at, in terms of the number of inspections, you know, how many had fatalities versus hospitalizations versus someone who went to the emergency room or an outpatient clinic, and then there were some that had missing information. Because this OIS system really wasn't designed to really look into this particular question, and so we took the opportunity to really see what the data has, and then we can talk about next steps moving forward. And the reason I have greater than equal sign is that some of these inspections were not specific of exactly how many numbers of people were hospitalized versus in the ED setting, but fatalities was an exact number. And so from the work fatality standpoint, I wanted to also look at, for example, of the fatalities, what are the common sources of carbon monoxide exposure that were represented? So you can see here the pressure washer, you know, gas powered was top one, forklifts, heaters, and auto vehicles, truck, auto, bus were also included. And you can see the industries wise, it's a wide range from manufacturing to food services. And then in terms of the fatalities, the back extrapolated carbon monoxide level, it's, as you can see, 91.3, which is almost twice the PEL, that was the low point, and it went all the way to 1,163. And then duration of exposure here was actually similar when you look at the non-fatalities versus fatalities, it was a very similar range. And so putting all this together, as I had mentioned earlier, we saw how colder climates, northeast, midwest, captured most of the fatalities. In addition to the number of inspections, fatalities wise too, they ranked the highest, and these are the four states that were among the top. And we see that forklifts are still the most common source of carbon monoxide exposure. And like I mentioned earlier, the propane powered floor machines, that was one specific one that was very interesting in terms of a source of carbon monoxide exposure. And it also had that same origination from the midwest and northeast, where they're using these floor machines for maintenance of older floors that have asbestos, helping with the concrete operations, for example. But there are limitations to this initial steps that we've taken, that in terms of carbon monoxide, it's not specific in terms of symptoms, and could be so many different things when you look at just the symptom profile. And like I mentioned, missing data is present, and hopefully this kind of initial steps can really help to make it more and more user friendly for this kind of approach. And obviously I didn't look at the full sample size for that reason. So in conclusion, the air contaminant standard that I had cited earlier, we see that Salt Lake's back calculations have enabled OSHA to cite 84 different employers with that air contaminant standard. But then beyond that, really looking at in terms of the industry, it's not only important in construction, but we saw even from food services, retail trade, that this is an issue in other industries as well that needs to be addressed. And then looking at from an exposure control standpoint, really when I was looking through these inspections, what happened, so many different things from inadequate ventilation to the actual air level, as well as the machines themselves. So like the propane powered gas machines that are utilized, they were not serviced at the right times. And so that was another reason why they were faulty and led to increased elevations of carbon monoxide exposure. And then obviously from a seasonality standpoint, the winter months still is important, but it's not zero in terms of inspections during the summer months too. So we have to be aware year round. And I'd really like to thank Scott Jones from the Salt Lake Technical Center for providing that raw data that we then utilized. And with the mentorship of Dr. Tustin, Andre Taylor, Dr. Hodgson, Dr. Cannon, I had the opportunity to do a particular inspection, really we're looking more at the long-term effects of carbon monoxide exposure. And so really in terms of next steps of this kind of intervention is really we looked at the acute effects, and hopefully we can look at the long-term effects going forward. And Dr. Weaver for facilitating this overall opportunity. So with that, thank you all so much. Thanks, Dr. Santiago. For the residents in the room, that was sort of a near-perfect board review for the occupational exposure to carbon dioxide. So just take that and you'll be good, and there will be questions on it. Any questions from the audience here? So use the microphone. Dr. Russi. Romero, thanks. Nice, nice presentation. So I would guess that factors like age and comorbidity and just the sort of heterogeneity across the country in different hospitals for hospital admission figure prominently in predicting who is admitted, et cetera. But did you look at a correlation between the estimated level of exposure and whether someone ended up in the ED, ended up getting admitted, ended up dying? Peripherally, we looked at that in terms of looking at the overall Pell and the carboxyhemoglobin levels with the model that they utilize. They have certain values for like age, smoking status, and so forth that are inputted. So on the back end, that was done, but definitely in terms of knowing exactly the comorbidities of that individual, that was a lack in terms of the OIS system where you couldn't really see some things of like one particular individual that had that blood sample age-wise and how that correlated. That's a great question, Dr. Russi. Pass the mic. Thank you. Very nice presentation. I don't know if we have mentioned, but did you look at the carboxyhemoglobin level of the blood of the workers? And also, did you look at the smoking history, their personal history? Yes. And also, did you look at the smoking history? Yes. And also, did you look at the smoking history? Yes. And also, did you look at the smoking history? Yes. And also, did you look at the smoking history? Yes. And also, did you look at the smoking history? Yes. And also, did you look at the smoking history? Yes. And also, did you look at the smoking history, their personal history? Yes. Yes. So they were factored into the model in terms of the carboxyhemoglobin levels that one is measuring in the emergency room. It's kind of afterthought where you have the exposure, but we're now looking at the ED values of the carboxyhemoglobin. So yes, those were the values we looked at. And then smoking status was part of the Coburn-Foster-Cain equation to factor that in. Because you're right. The smoking status increases for every pack per day and can increase the carboxyhemoglobin level by at least 2.5%, so. Thank you so much. Thank you. Thank you. A couple more questions. We're keeping our mic staff busy. Thank you, Dr. Santiago. My question is, do you know how many workers were treated with hyperbaric oxygen or what the latest literature states on the efficacy of hyperbaric oxygen with carbon monoxide treatment? That's a wonderful question, Dr. Nakasi. So that is something that actually there was an inspection that I had mentioned that I worked with Dr. Kainon separate from this long-term project where we were looking specifically at that. What is the standard of care for in the emergency room to the hyperbaric oxygen versus normal nasal cannula? And the evidence is mixed. Really, it's not conclusive one way or the other. Health systems are doing one of both things. For this particular project, we really didn't have that data to see what they had done for that particular individual if they were in the emergency room or admitted. Great question. Great. Dr. Harber. That was a great presentation. Can I ask about the drunk and the lamppost? I think some people know what I'm getting at. You only find what you look for, our priority. So do you know anything about the cases that are not captured in the data system or for which they're incomplete information? In other words, the difference between the 147 you looked at and the ones that had incomplete information or any mortality data from any other case reports of occupational mortality data from any other of the federal databases that are available and see if they came in there so that, again, Dr. Meyer will explain the drunk and the lamppost. Basically, the question is, do you know how representative these are or does this represent only what was captured by the OSHA database? Thank you, Dr. Harber. I think when I look at the other census, other studies and look like, for example, what Washington State had done or Reed Whitaker, there was definitely alignment in terms of the top sources of carbon monoxide exposure, winter months. There was certainly alignment there in terms of the sources of carbon monoxide exposure. I think the limitation was in terms of the ones I wasn't able to include, it really didn't have as much on the source information, you know, what was the source, as well as how many people, even rough estimates of people, had to go to the ED for care or were admitted. So fatality data was the most exact. That was the most prevalent. But other data was missing because it's not just about the fatalities, it's also the morbidity aspect, too. I had seen a question here, but it seemed to have disappeared, so we may, sort of, if they pop up, we'll have them. There's one about the last presentation, which I won't drag him up for, but let's just make sure we're on the right track here. Dr. Don Cannon says, go, Dr. Santiago. And while we're doing that, we'll give a shout out to Dr. Virginia Weaver in the audience here, who's, I don't have to embarrass you, but thank you for, you know, I think we have one other presentation from OSHA here, so thanks to the staff for that. Yes. Yeah. Yeah. Yeah. Thank you. Thank you. Thank you so much. Great. All right. Switching subjects a little drastically, we'll welcome Dr. Dorian Kenley from the University of Washington Occupational and Environmental Medicine Residency, who's going to talk to us about probably something we've never thought very much about, which is cannabis allergy, because we're now seeing occupationally exposed individuals as well as recreational users. So come on up, Dr. Kenley, and we'll be interested to hear on that. How's the gizmo work? I know this is a... It's just, yeah, that's the pointer, and this is your, that's your forward button, and your backward button. Okay. Great. Thank you so much. So we're good to go? Morning, everybody. I'm Dorian Kenley from the University of Washington, and I'm going to present, I hope, some compelling findings with short and long-term implications from the work that we've been doing with the adolescent cannabis industry in Washington state. So the work I'm going to present today has been very generously supported by a variety of research and training grants. We do use human subjects. All of our work with human subjects has been reviewed by the Institutional Review Board of the University of Washington. As far as I know, everybody involved in this project has no financial conflicts of interest. Here we go. So cannabis is now legal in 37 U.S. states and territories, and recreational use is available in 20 of those states. And this has birthed a new agricultural industry that has also formed a very fast-growing group of agricultural process and retail workers. Because cannabis remains officially illegal at the federal level, it's a Schedule I substance, that also has created a very murky regulatory landscape, and that has real implications for consumption, cultivation, business, occupational health and safety regulation, as well as research. So our studies in the state of Washington focus primarily on workers in indoor cannabis growth facilities. And so these workers are subject to a variety of occupational exposures. You've got chemicals from things like fertilizers. You've got UV radiations. You've got toxic gases. For anybody following along online, I'm looking at the diagram on the left-hand side there. The infographic that I've got here is from 3M, and that summarizes very nicely some of those common exposures. One thing I do take issue with is this idea that you can get drug exposures just from handling simple plants. But the other thing I do question is the difference between diagrams and reality inside facilities. As all of us know as occupational medicine physicians, the reality of the workplace usually doesn't match the diagram. So what we tend to focus on in our group, and the prior work I'm going to show you focuses on just how relevant these exposures can be. We tend to look at indoor air quality, respiratory exposures, and atopic health symptoms in these workers. And so if it'll click for me, maybe. There we go. In our previous cross-sectional study, I'm sorry. I think I'm having some clicker issues, guys. If you'll look at the top left, we did a previous study led by my mentors, Dr. Sack and Simpson. We found a very high prevalence of respiratory symptoms among cannabis workers. And so about half of those workers demonstrated some abnormal fractional exhaled nitric oxide and airflow obstruction. More than half of those workers demonstrated sensitization to cannabis on skin prick testing. Those workers demonstrated changes in their FEV1 and FEC. Those are measures of spirometry. And those were found to change over the course of the work week. So they'd come in from the weekend. And over the course of the work week, those measures would decrease. Phenol would increase over the work week to a statistically significant degree. And that work has already been published in some of the pulmonary literature. There we go. So what do we know about occupational cannabis allergy symptoms? And what might actually be happening? Well, it turns out that there isn't actually much data about the prevalence of occupational cannabis allergy. That's really what motivated our original study. We do know that, as with any allergy, occupational or otherwise, there are numerous manifestations. And we suspect that in the workplace, it's likely that symptoms that patients report are probably a combination of IgE processes and irritant processes. You saw some of the pictures. There are plenty of bad habits running around. And so in indoor growth facilities, you've got workers with exposures that could lead to irritant-type symptoms. And we've got previous studies from the European literature that show, even in exposed workers, they can report atopic-type symptoms without evidence of sensitization or IgE-mediated processes to a high degree. But there is also a lot of similarity, if my clicker will work for me. There we go, with other members of the cambiaceae family, things like hemp, as well as cross-reactivity to things like certain fruits and vegetables, to, again, a very high degree, as well as certain airborne sensitizers in these growth facilities, molds and so forth. But we do have a lot of tools that we can use to analyze these. And so what we aimed to do, the overarching goal of our project at the University of Washington, is really to characterize the prevalence of these work-related symptoms in cannabis workers. But our initial studies revealed a problem. Turns out that 97% of people who work in the cannabis industry use these products outside of work. So you can see how that could be a confounder, those of you with an MPH. So we aimed to develop some control groups. And our project that I'm presenting today looked at developing some of those control groups. We wanted to develop control groups of recreational users and non-users, leveraging some of our prior study data. And so our aims were to construct those control groups and determine the prevalence of allergic and irritant symptoms in cannabis employees compared to users without occupational exposure. And we hypothesized that those occupationally exposed workers would demonstrate higher prevalence of allergic and irritant symptoms compared to our controls. So our study population was obtained from workers at two of our indoor cannabis growth facilities. And our control population was a demographically similar group recruited from the community in Seattle, Washington. They were asked to complete a detailed questionnaire of their atopic symptoms as well as their medical history. And we also looked at their spirometry measures, their pheno and their skin prick testing response to common allergens, cannabis and hemp. Our health measures and outcomes were actually very straightforward. So pheno we obtained directly using a handheld device and we handled that as a continuous variable. Spirometry we assessed using NIOSH certified clinicians. FEV1, FEC and the ratio were again just continuous variables. We looked at that using ATS criteria. We didn't do post bronchodilator measurements in that group. The main reason for that is because in our occupational group we didn't have a very robust set of post bronchodilator data to compare to. So we just used the pre bronchodilator data to do our comparison. Skin prick testing, again, we administered tests, four types of molds, two strains of cannabis grown locally, hemp, some common Pacific Northwest allergens, grasses, histamines, so on and so forth. Very straightforward skin prick testing. Health measurements questionnaire, again, asked individuals if they had any allergic or irritant manifestations, thermal, ocular, nasal, respiratory symptoms within the last 12 months following exposure to cannabis either in the workplace or at home with handling and so forth. That image on the left, there it goes, is a representative sample of the kind of questionnaire that we use. These are representative questions. Our questionnaire setup was a lot more detailed than this. I'm trying to just give you guys a little bit of a taste. We also collected just general health demographic data, information on smoking history, chronic medical conditions, what meds are you on, so on and so forth. Primary outcome here really is the report of irritant or allergic manifestations over the last 12 months. So to look at our continuous variables, we also did apply linear regression. In the case of pheno, this was log pheno to normalize that data. We defined our model equations a priori based on clinical knowledge and just our past published reports. There's a lot of information out there about how you can model that. For the sake of the talk today, I haven't presented that. I do have some slides if people are interested about how you can go about modeling that or establishing those regression models. Calculated adjusted odds ratios for reporting atopic symptoms based on exposure groups. So we adjusted for those demographic factors as well as the medical history of our patients. So let's take a look at some results. You know, the skin prick test, really striking. Nobody among non-users demonstrated cannabis sensitivity, but over 31% of your cannabis workers got sensitized. Occupationally exposed workers compared to only a single recreational user demonstrated cannabis or hemp sensitivity on skin prick testing. Fractional exhaled nitric oxide. There's a significant difference, statistically significant difference between occupationally exposed and unexposed groups with an increase in mean pheno with occupational exposure. That holds up even if you split out the recreational users from the non-users. I think it's better if you take a look at this in terms of our modeling work that we did. The regression results, I think, really put this into better perspective. So geometric mean pheno in cannabis workers, 1.5 times higher regardless of your outside use status, whether or not you're a recreational smoker. If you work in a cannabis grow facility, you're going to have a geometric mean pheno 1.5 times higher, ignoring anything else. We did look at FEV1 percent predicted. I'm showing this here really for context. This result is not statistically significant, but you do establish a trend of decreasing FEV1 predicted with occupational exposure. The regression model that we did indicates that there would be about an 8% decrease in FEV1 percent predicted if you assume a smoking interaction. Again, I just want to emphasize in showing this, it's not a statistically significant result, but there is a trend that you can see there. And that's been a consistent trend with our studies that we've done and a consistent trend that has sort of been established with other work. So I think there's something there. And when you consider that this is small sample size data, it's something that merits worth further exploration. But let's go back to statistically significant results. Let's have some bigger numbers. Allergic and irritant symptoms with exposure. Cannabis workers, if you look at our adjusted odds ratios, 30 times more likely to experience atopic symptoms when you adjust for demographics and medical history. You know, that increased odds ratio, that holds up even if you ignore recreational use status. 9.5 times more likely to experience symptoms. So I kept it short today because I really want to try to drive robust discussion of this topic. This is a new industry, a new field. People are kind of polarized with these discussions. And I think there's a lot of meat here. So I really just want to roll this up. Occupationally exposed cannabis workers have significantly increased odds of respiratory and allergic symptoms. They have significantly higher phenos. They have a higher prevalence of sensitization on skin prick testing. And they have a trend toward decreased measures of pulmonary function. Now, you know, I've shown some big numbers. But I concede I do have some limitations. So one, I have small sample sizes, admittedly. I think that explains some of my wide confidence intervals. You know, that explains some of the difficulty that I've had with modeling. And that does limit some of my inferential utility of my regression models that I've done. We had a lot of difficulty getting IRB approvals with this. The University of Washington is a publicly funded institution. I mentioned early on, cannabis is a Schedule I substance. We get federal money. Boy, oh, boy, like 40% of my project was just wrangling with different muckety mucks. Sorry if any muckety mucks are listening from UW. But I'm graduating in two months. You can't do nothing to me to get approval to do some of this work. And I think in our profession, that's a big challenge. Because what it does is it forces a lot of this work to be done in a private setting, in a private lab setting, in a private research setting. And it challenges us because it makes it difficult. And it disincentivizes us to do this in a public setting where we can present results. You know, the other issue is that our atopic symptom report might not be the best proxy for IGE sensitivity. And similarly, because we had to go about our cannabis sensitivity testing different ways in different groups using different strains in different groups, maybe that's not the best representation of sensitization between an occupational group A, occupational group B, control group A, control group B. We're trying to address that. We already have a, if I can point with this, a larger study, including some of these preliminary findings with a lot more power, a lot more patients. We do intend to utilize the blood testing this time. And one of the things that I really want to address is a better exploration of the association of cigarette smoking, as well as smoking cannabis as the method of recreational use, particularly with that phenomeasure. And we can leave that for a little bit further discussion if anybody's interested. But first, discussion points. I want to, again, provoke some thoughts here. You know, we've got this adolescent, rapidly growing industry in an agricultural workforce. These guys all have exposures that create risks for irritant and allergic occupational illnesses. Cannabis operations right now are regulated by states. So that means that you've got inconsistent and limited protections for workers. We don't have a lot of information in the US literature on cannabis. You don't have really much to go on in terms of establishing standards, because the standards are different in every state. You don't have TLVs. You don't have your quality data. You only really have to meet a PEL for dust at the federal level. And that's not really low enough for a sensitizer, if anybody deals with sensitizers in their clinics. Our best proxy might be the existing agricultural knowledge for hemp or the grain industry. So what's the long-term implication of that? What's the long-term implication of occupational asthmas due to sensitizers in an industry that isn't really considered valid? What happens to these guys after 10 years when they come into your clinic and they have a disease or illness that's not even recognized? And federal regulations make this research really, really hard to do at public institutions. So it's all private. And how do we deal with that? How do we cope with that? I really have to acknowledge my mentors, Dr. Sack, who is an occupational pulmonologist at the University of Washington. She's been really formative for me, as well as Dr. Chris Simpson, who is one of the chairs at the Department of Environmental Occupational Health Sciences at the University of Washington. Dr. Cherry, who has really given me a lot of opportunities in my career. She's sitting there, so everybody, say nice things to her. The Cannabis Alliance, who was a great partner to us in Washington state to do this. Canyonlands, which is in the southern part of the state. That's me there in 2015, after the conference. I'm going on vacation there. And I'm a little bit heavier now, but I'm going to take another picture down there, and it's going to be great. And I really just thank you all for your time. Thanks so much. Hopefully we have a great discussion. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Questions from the audience? All right. We're going to run out of time there, so all right. Thank you. And then I think Dr. Weaver after you. Hello, Captain Smith from Madigan Army Medical Center Preventive Medicine Residency. Being that this is a growing adolescent exposure, have you thought about comparing this to models of other recent sensitized occupational allergic reactions similar to, I don't know, latex, something that's been studied extensively, done multiple models on it, on how it grows over the course of occupation? So we have. And I think that our best idea right now is to make a comparison to the grain industry, as well as some of the things that we've seen with hemp or perhaps hops. Both of those have basically an aerosolized dust. Both of those probably have a plant protein or perhaps a mold epitope. Both of those have a GI or upper respiratory tract route of exposure. And both of those have similar clinical presentations. And so that's just kind of your starting point. So that's right now the thought. But this is kind of new stuff. Thank you very much for a very insightful presentation. My name is Alexey Karenev. I'm from the University of Cincinnati. I just wanted to ask if you had thought, in terms of future directions, next steps, incorporating some work with bioaerosols in terms of looking at exposures for the workers. Because it's definitely hard to really drill down and nail whether this is a plant dust, a plant component, or is this a bioaerosol. Thank you. I would love to partner with a chemist who has an interest to do that. Absolutely. I would love to find somebody who has an interest in the knowledge base, pardon me, and skill set to do that. Our institution, we've done a lot of work with bioaerosols in terms of enclosing. Sorry. Sorry about that. I'm going to sort of cut it short. It's not a question because we're running out of time. Just quick, two points I wanted to commend you for comparing occupational and recreational users. They're totally different exposures. The occupational setting is way, way more. And the second thing is great questions that you're asking in your discussion. We clearly have to grapple with those as a field, moving forward. Dr. Sommel, last question. Three real quick questions. First of all, did you bring samples? We got one. Secondly, did you compare between the gardeners and the processors as different types of work? And for your future work, are you looking to see about engineering controls, PPE, and how that could affect it? So to answer your first question, no, Utah is not a legal state in any capacity, as far as I'm aware. To answer your second question, we do have that data. I did not incorporate it for this study. But in some of our other studies, we did look at the different jobs and what's involved with them and their exposure levels. And to answer your third question, we did look at some of the engineering controls. Again, I didn't address it here for our purposes. But the engineering controls do come into play for the different job titles. And again, whenever you talk about engineering controls, you talk about individuals' adoption of said engineering controls. You can tell a guy you need to use a respirator, but does he use that respirator perfectly every time? Is he fit-tested for it? And what are the standards that you're applying? Thank you. Sorry, we ran out of time for questions. I'll encourage you to buttonhole Dr. Kenley after the presentation is over with more informal questions. But I want to try to keep on time. We'll welcome our last speaker for the morning, remembering that there's an afternoon session. Dr. Tran Le, who has come all the way from the University of Utah Occupational Medicine Program to talk to us about crash rates and sudden incapacitation risk in female truck drivers. So come on up here. So this is just the forward and back. There's a slide there. This one? Yeah. Yeah, just sort of pointed at that stuff. OK. All right. Thank you. OK. Thank you for the introduction, Dr. Meyer. As he said, I'm Tran Le, a first-year resident from the University of Utah. And I'll be presenting our research project on crash rates and sudden incapacitation risk factors among female truck drivers. And here are my disclosures of our funding sources and IRB exemption status. So female commercial motor vehicle drivers are a growing minority of workers in a male-dominated occupation. Looking at data from the Bureau of Labor Statistics that combines drivers, sales workers, and truck drivers, from 2003 to 2021, the percentage of females increased from about 5% to 8%. And upon review of the current literature on gender differences in crash rates, we found that male drivers in the general population have a greater likelihood of crash and engagement in risky driving practices. They also have higher fatality rates. Male truck drivers, although not as well studied as the general population of drivers, have been found to have a greater likelihood of crash and higher fatality rates as well. So truck crashes continue to be a serious problem in the US. They are overrepresented in fatal crash frequency and serious property damage compared to passenger cars. FMCSA data on large truck crashes demonstrate that in 2019 alone, there were about 5,000 fatal crashes. Further research is needed to analyze risk factors for crashes in order to help target prevention strategies to driver groups that are most in need. To our knowledge, sex differences in crash rates with adjustment for risk factors for sudden incapacitation risk factors have not been studied. So our project focused on two main study questions. One, is there a difference between male and female truck drivers' rate of any crash or any preventable crash? And two, is the difference, if any, affected by adjusting for sudden incapacitation risk factors? So we hypothesized that female truck drivers will have lower rates of total and preventable crashes, and that this difference is due to males having higher rates of risk factors for sudden incapacitation. So this was a retrospective cohort study where we compared sudden incapacitation risk factors and crash data between male and female truck drivers. Oh, there we go. OK. Sorry. So we started off with a large population of truck drivers and divided them into male and female truck drivers where females were the exposed group. And we controlled for covariates, including sudden incapacitation risk factors and employment duration. And then we analyzed our primary outcomes of any crash and preventable crash incidences. So our project combined data from two separate data sources, commercial driver medical exam data and a large US truck company's data. The data collection period ranged from January 5, 2005 to October, 2012. And the medical exam data was obtained from RoadReady, which is a large commercially maintained data set of CDMEs. From this, we obtained measured values of BMI, blood pressure, as well as self-reported medical history and substance use history. The company data was accessed via a nondisclosure agreement. From this, we obtained our primary crash outcomes, some secondary crash outcomes of whether the crash was DOT reportable, whether there were injuries associated with the crash. We had crash causes and crash types. And we also looked at driver employment data from this data set as well. So we performed descriptive statistics on all variables, compared covariates using chi-square for categorical variables, and then Wilcoxon-Ringsom tests for continuous variables. We performed logistic regression to calculate the relative risk for crash outcomes and proportionate hazard regression to calculate the hazard ratio for crash outcomes. So as far as our cohort selection, we started off with a population of truck drivers, 78,000. We excluded drivers who were determined to be medically unqualified or medically-pended. The reason for disqualifying the medically-pended group was their employment durations tended to be short. About 75% of these drivers were employed for less than one week. And we ultimately ended up with an analytic cohort of 74,000 drivers, with about 70% of these qualified for a two-year card and 30% qualified for less than a two-year card. So here is a table summarizing the descriptive statistics performed on our covariates and primary crash outcomes. About 4% of drivers were female in our population, and females were significantly older and had higher average body mass index. Aside from BMI, males had higher rates of medical risk factors for sudden incapacitation, including hypertension, cardiovascular disease, and diabetes, whereas females had higher drug and medication risk factors for sudden incapacitation, including drug use, opioid medication use, and benzodiazepine medication use. And then in the last two rows, you can see our primary crash outcomes of any crash and preventable crash, with females having statistically significantly higher rates. And here are the relative risk values of our primary outcomes. The figure on the left shows relative risk values of our crude data, and the figure on the right shows the relative risk values adjusted for age, BMI, hypertension, cardiovascular disease, diabetes, and company tenure. So both figures depict that females have a higher risk of being involved in any crash, as well as a preventable crash. And all of our relative risk values were statistically significant. Even with adjusting for the covariates listed, the magnitude of the relative risk did not change much. As you can see, for any crash, the relative risk increased from 1.2 to 1.26. Preventable crash, relative risk increased from 1.15 to 1.18. That leads us to believe there is another factor explaining the sex differences. So when looking at the hazard ratio, females were more likely to be involved in a crash sooner after their medical exam than male drivers. And this applied to both any crash and preventable crash. The any crash relative risk was statistically significant, whereas the preventable crash was approaching statistical significance. So moving on to our secondary outcomes, when looking at DOT reportable crashes and DOT reportable crashes with injuries, none of the relative risk values were statistically significant. Interestingly, females had a lower risk of being involved in a DOT reportable crash that was determined to be preventable. But this was not statistically significant. It was borderline significant. It's possible that we were statistically underpowered in capturing differences in accident severity between male and female drivers. And looking at the top seven crash types, which made up about 75% of the total crashes, the top three crash types were hitting a stationary object, backing, or being hit by another vehicle. And when comparing males and females for each crash type, the only significant difference was found with being hit by another vehicle with females having a higher risk of being hit by another vehicle, or higher rate of being hit by another vehicle. And looking at the crash causes, males had a significantly higher rate of crashing due to inattention, whereas female drivers had a significantly higher rate of crashing due to committing a driving error. And these included things like failure to yield, failure to maintain control, failure to maintain proper distance. So in summary, after adjusting for risk factors for sudden incapacitation, female truck drivers are still at higher risk of our primary crash outcomes. Our initial hypothesis that females would have lower risk of crash is unsupported. And our findings differ from the current literature on crash rates in both truck drivers and the general population. Reasons for this could include differences in lifetime tenure and lifetime miles driven between males and females in our population. We did have company tenure data, but not lifetime experience. So we just didn't have that data available to us. There may have been differences in reporting accidents to the company. Most of the accidents were not serious crashes, with only 9% of total crashes categorized as DOT reportable. Although vehicles were inspected by the company on a regular basis, very minor crashes could have had differences in reporting. And potentially, female drivers were reporting minor crashes to the company more often than male drivers. And finally, female truck drivers may possess different individual factors and or face different environmental factors compared to the general population of female drivers. These include individual level of risk taking, facing adverse work conditions, such as low social support, isolation from family, and poor ergonomic fit of vehicles. So the strength of our study, one, we had a very large population size of 74,000 truck drivers. We did have sufficient statistical power to assess for our primary outcomes of any crash or preventable crashes. And the crash data was company collected data. So this was not data obtained from a driver survey, where we asked their crash history. The driver data was individualized, where we linked their medical exam data to the crash data through their CDL number. And we were able to demonstrate temporality for the relationship between covariates and crash outcomes. For example, we knew that someone had high blood pressure diagnosed before they had a crash. And some of the limitations I've mentioned, we didn't have data on lifetime tenure or miles of drivers. We might not have sufficient power to assess for crash severity, and there could have been differences in reporting minor crashes. The medical data was limited to what was documented on the CDME form. So of course, this is subject to driver reporting and quality of provider documentation. So given the differences identified in crash rates between males and females, future research is needed to identify the causal pathway. We suggest the next steps would be a prospective cohort study that addresses our study's limitations and considers the other factors that could be at play that I've mentioned. And our findings can inform targeted driver training. Drivers should be trained on the most common causes of crash and crash types among all drivers. They should also be taught the sex-specific differences in crash causes, crash types, and crash rates. This is essential as more workers enter the trucking profession to fill ongoing shortages. So in conclusion, female truck drivers are at higher risk of crash, even after adjusting for risk factors for sudden incapacitation. Further research is needed. And our most common crash types and accident causes were identified in this study with some differences found between male and female drivers. Well, I would like to thank my research mentors, Dr. Theis. Big thanks to him. He's sitting in the back. And Dr. Chen and Dr. Hegman, as well as my University of Utah PGY-2 co-residents, Dr. Jones and Dr. Salo, for their input and support. And here are my references. And thank you for your time. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you.

occupational health

cannabis industry

legalization

US states

agricultural workers

retail workers

regulations

health risks

chemical exposure

respiratory symptoms

research

worker protection