 Now, today's webinar is Age of Metapasms in Relation to Frailty and Biological Age in the CLSA Conference of Cohort. Let me introduce our distinguished speaker, Dr. Chris Vershore. Chris is an assistant professor in the Department of Health Research Methods Evidence and Impact here at McMaster University. His research-focused features a comprehensive approach that combines epidemiological techniques to identify, inherent and modifiable risk factors of age-related disease and morbidity in in vitro and ex vivo molecular methods to examine potential biological mechanisms. This approach often features immune and inflammation-related soluble and cell-associated biomarkers that not only correlate with aging and or age-related disease, but also modify relevant cellular innate immune processes. Hello, everyone. Welcome and thank you for joining me today. Well, I give a bit of a talk of what I've been doing for the last year or two on my project with frailty and its relation to age at menopause and some preliminary work on biological aging. Thank you for the introduction, Carol. I've never been called distinguished before, so that was wonderful. And if everyone can bear with me, this is my first webinar and I find it absolutely disorienting. So just, yeah, bear with me while I get my bearings. So let's continue. Okay, off to a good start. So I think this is a figure that many of you, if not most of you, have seen before. Our demographics and the demographics of many developed nations and as a whole many nations across the globe are getting older. So what we have here is Canada and Canada in the 70s and Canada about 10 years from now. And we can see that there is and has been a large shift. The baby boomer generation is in the blue. So in the 70s and earlier, a lot of our population was made up by this group. So the less than 30, less than 20 percentage. And above 65 was a very, very small proportion. If we fast forward to around now, it's 2011. It's not quite now, but a few years back, we'll see that those baby boomers were, you know, either retiring or they've already entered retirement and the demographic is starting to even out. If we go about 10 years from now, and this is what we're going to see for the next 10, 20, 30 years and so on, is that things have shifted dramatically. We actually have a lot more individuals that are 65 and older than the 20 and younger than we did in the 70s. And as a whole, we're pretty much equal, more like a sausage and a pyramid, if you will. So that is going to be a bit of an issue because now we have a lot more individuals that require a lot more care. So this is not any surprise. We spend more money on older individuals. So once you hit about 60, 65 years old, there's a sort of an exponential increase in how much healthcare is being spent on you. But it's not just the healthcare dollies that is obviously going to increase as the years move on. Things like CPP, Canadian Penetration Plan and old age security, that isn't going to increase as well quite dramatically beyond 2030 and up to 2050. But it's not just about money. It's not just the economic impact. Whereas also sort of the human impact, we know, and it's no surprise that individuals as they get older, they're more likely to obtain a chronic condition. So the prevalence of a number of diseases, arthritis, heart disease, cancer, diabetes, etc., they all increase as we get older. So there are going to be a lot more people living with a higher number of chronic conditions. So I think as researchers, the big question here is what the heck do we do about it? So in recent years, there's been a focus on a concept called healthy aging. What exactly is healthy aging? Well, I think in its simplest, there we go, in its simplest form, it's really just maintaining good health as we age. But good health is a bit of a can of worms, isn't it? There's a lot more to this term than just what we would think of as health. So really, if we're on a rapid, it's maintaining good physical health, mental health, social health, quality of life is just really a wrapper for all these different components. And then independence is really one of the most important areas of healthy aging. And what to consider here is that we have all these different components that are all very important and they all interact. So really, it's not just one thing that we are trying to improve. It's many things that all interact. So in order to do something, what we need to do is embrace strategies that prevent or mitigate the root cause of age-related decline or really unhealthy aging, if you will. It's not just a one-problem solution. It's going to be many problems that we need to consider. So if we look at a simple aging trajectory, you know, we have somebody getting older, they pick up some comorbidity, some chronic conditions, and there is a loss of independence and not far after there is a loss of life. Some would live longer in a dependent state than others. Well, healthy aging really looks to sort of stomp this out, minimize the number of chronic conditions that are acquired, lengthen the number of independent years. And, well, we're not going to live forever. We can't stop mortality, but at least lengthening those years and lengthening healthy years. Now, one of the areas I've been focusing on is a bit of the antithesis to healthy aging. So measuring healthy aging is one thing, but we can also measure sort of the opposite. And this is a syndrome called frailty, which also increases with age. And with it comes our loss of independence and mortality. So there has been talks on frailty, previous webinars, but I'll go over it again for anyone who is not familiar. So frailty is a clinical state. It's geriatric syndrome in which there is an increase in an individual's vulnerability for developing increased dependency and or mortality when exposed to a stressor. So the important parts there is vulnerability when exposed to a stressor. So somebody who is frail, when they're exposed to an acute or some sort of chronic event, some sort of trauma, their ability to get back to that healthy state that they were in or at that level of health that they were in is less. So they're never able to really get back to that homeostatic state. And that continues to get worse. If we look at a figure here, they have two examples, two trajectories of individuals. One is what they consider, they're calling the normal aging trajectory. And the other is they call accelerate aging. But we can kind of look at them as healthy and unhealthy aging. And that healthy and aging individual will live a long life of dependence. They will still see a decrease in a global measure of performance of health, if you will, but they won't enter sort of a frailty state until later on in life. An unhealthy individual, an unhealthy aging individual, they'll have a much more steep decline. They'll enter these frailty years earlier, and this will put them at a much, much higher risk of being disabled, that loss of independence, and then, unfortunately, as well, the higher likelihood of mortality. So I'm talking a bit about frailty and its importance. Well, how exactly do we measure frailty? Well, frailty is measured, can be measured a number of ways. There's a lot of great work on frailty that has been published. Now, I'm going to talk about two different methods that are a bit more popular. First is Linda Freed's phenotype model. So what Freed did is she conceptualized five components that she felt was at the core of the frailty syndrome, those being exhaustion, weakness, weight loss, slowness, and low physical activity. She devised thresholds in order to categorize people as having any of those components or not having those components. And then, furthermore, she summed up these components to further classify people as either healthy or robust, having none of these components, pre-fail having one or two of any, and then being frail, so having any three of these components or more. So another way that we can measure frailty is Ken Rockwood's frailty index. And what he did with this index is, instead of looking at frailty sort of as this categorical model, these categorical states, he's looking at as more as a continuous state, a continuous model that is really increasing and is made up of deficits. And what are the deficits? Well, this is where I kind of consider frailty index as sort of the kitchen sink method, because a deficit is anything that you could say is adverse to health, if you will. So perception of health, satisfaction with life, chronic conditions, or diseases, depression, physical activity, poor nutrition. I mean, the risk list really goes on and on. Now, mind you, there is a more appropriate way of defining what a deficit is. I'm being very generalizing here, but that's sort of the gist of it. So what you do for the frailty index is you identify your list of deficits. You dichotomize them as zero or one. So one person has it, zero. They do not have it. You sum them up, and you take the proportion, and what you end up with is with a nice continuous variable between zero and one. And that is the frailty index. So we have two ways of going about it here, these nice categories, and as well as sort of a continuous variable. So why measure frailty? Well, if I bring up an example here, so these are three individuals that I made up with three different trajectories of frailty, and we can sort of categorize them as being healthy aging, unhealthy aging, or you know, very unhealthy aging in an individual is a very steep increase of frailty, and whose life expectancy is shorter than these other two individuals. And the importance of being able to measure frailty, the importance of being able to actually create these trajectories and find out where a person is, is because we can then look for determinants. We can find out why it's happening. So if we ever want to optimize healthy aging, if we ever want to come up with strategies that will help us optimize and maintain, we need to find out what causes unhealthy aging or frailty. So it really intrigued me when I was starting to learn about frailty was that one of the strongest determinants that we know of, and we've known for a while, is sex. So this is a result from a systematic review published a number of years ago, a very good systematic review. And what they looked at was prevalence of frailty across a number of studies. And what I want to bring focus to is this, is that women have much higher rates of frailty than men, their estimate was twice as much. And this was really, really intriguing to me. And it led me to try to understand why exactly this is. Now, when I bring this up, the first thing that people say is like, well, that's obvious, because men don't live as long as women. And we've known this for, you know, hundreds of years. Well, recent work. So this is came from a recent study from Ruth Hubbard's group in Dalhousie, is that that's not necessarily the case. So what they did is they did a large meta analysis of a number of studies, and they looked at the frailty index. And what they showed is that it doesn't matter which age group you look at, women are almost or are always higher than men when it comes to their score for frailty. So what this shows that it's not actually a matter of this that women are living longer than men. It's actually showing that women are able to deal with frailty. They have higher rates of frailty than men. And what this is really interesting to me is this is very much a representation of the well-known male-female health survival paradox. So this is something that we've known for a very, very long time, hundreds of years, that like I said, men live longer than women. But with that, is women live longer than men, but they do so with poor health. Well, it really looks like frailty, or at least measured by the frailty index, is really recapitulates this poor health that women sort of deal with as they get older. So my training is not in women's health, if you haven't noticed already, or if you didn't know me, my training is as an immunologist. But this is very intriguing to me. Why the heck would women have higher rates of frailty than men? Why would it be more prevalent than women than men? There must be at least a biological mechanism for it. So if we look at trajectory of a woman's life, it's a heck of a lot of interesting than men's, that's for sure. There's a lot going on that happens for them that do not happen for us, I mean men, one of which obviously being pregnancy, but that's probably not something that's affecting older adults too much. But one thing that older women do go through, many of you have already probably said it to yourselves, is menopause. So this is a condition that happens in women. So this is a cessation of menstruation for at least a full year. And there are a lot of symptoms that come along with it, and it's irreversible. So when menopause happens, there's no going back from it. And obviously it doesn't happen in men. So what exactly is happening on the inside? So we'll just look at the levels of two different hormones, female sex hormones, estrogen as well as progesterone. And we can see here in the fertile years, so after puberty, we have this cycling of estrogen and progesterone. When a woman reaches perimenopause, so this can be anywhere between five to 10 years from the time of menopause for menstruation secedes and essentially the time of reproduction is over. There is a decrease in progesterone and it gets to essentially its minimal levels. And then once menopause hits, estrogen essentially is bananas. It's all over the place until the time of menopausal transition is over. And then estrogen levels are lower and they continue to decrease. Now this is a very, very large difference than what happens in men. With men, testosterone levels do decrease as we get older, but it does so very gradually. And there are men who can have substantial levels of testosterone while into their 80s and 90s, although I'm sure that's not very common. With women, they have this, I want to say it's almost an acute event that they go through, substantial changes that occur and are not reversible. So when this happens, this is from the CLSA. So this is Canadian women and in Canada, the average age of menopause is 50 years old with standard deviation of about five years. And this is very common for most developed nations that menopause occurs around 50 years old. So it appears to be very physiologically impactful, seems to be very important, but I wanted to bring up some evidence. So there are a lot of great studies out there to show different types of health-related outcomes related to menopause and age of menopause. But this is one that I think sums it up nicely. So this is a systematic review of that analysis and what they looked at here is the relationship between the age of onset of menopause. So that's when a woman reaches menopause and its relation with cardiovascular disease and other types of outcomes and all-cause and cardiovascular disease related to mortality. And what they showed in this meta-analysis looking in almost 50,000 individuals is that when you compare women who've reached menopause before 45 years old, so this is about a full standard deviation below what we see in the CLSA, you compare them to women who've reached menopause at or above 45 years old, there's a pretty consistent increase in the relative risk of coronary heart disease. So now if we look at all-cause mortality, which is kind of the be-all and end-all of health-related code comes, you see again, this is nearly 100,000 individuals and the relative risk is not as high, but we do see a significant increase in the relative risk for age of menopause. So women who've reached menopause early tend to have higher relative risk compared to women that have reached it at around the sort of the normal time. So that is a great meta-analysis that is human data at its best, but really what else is going on sort of on the the person level, the physiological level? Well to do that we can't do a lot of experimental studies on people obviously, but we can turn to our little furry friends here. Now mice do not go through menopause, they cycle until they they pass away, but we can induce menopause surgically in mice by removing their ovaries. In doing so, there's again a number of studies that have shown these types of effects, but I've shown a few here and what we have here is lifespan, vaccine responses, tumor burden, they are all negatively related to OVX or overectomy and this is surgical menopause, so mice don't live as long when they've lost female sex hormones. Vaccine responsiveness is lower and they also have higher tumor burden. So really as a whole, menopause or surgical menopause when we're looking at our nice furry little experiment models here, it's all really negative really related and this is what I found interesting by reading through all this literature is there's nothing been shown on frailty though, although it seems pretty obvious that there should be an association. So that brings me to research question and hypothesis for my study. So the primary research question is a very simple one. What is the relationship between natural or surgically induced menopause and frailty? And with that, I hypothesize that an early age of menopause or having had a hysterectomy will be associated with higher levels of frailty later in life for community dwelling older women. If we bring back the aging trajectories example that I showed, this is more or less what I was expecting that early menopause hysterectomy, there would be these higher levels of frailty and then normal agent menopause 45 and above would be lower levels of frailty. So for this study, I use the Canadian Long-Tunnel Study of Aging, Comprehensive Baseline Dataset version 3.0. So this is about 30,000 little over 30,000 community dwelling adults age 45 to 85 and half of which 15,000 and change being women. So from this cohort, I removed the men, put them aside because we're going to look at them, but they weren't involved in any of the major analyses. And then of the women that were remaining, we removed anyone who had reported breast ovarian or other genital cancer diagnosis. The reason being here is that these types of cancers and their treatments can mask the true age at menopause. So they were left out. Anyone who refused had missing or did not know their menopause classification or their agent menopause, as well as women that reached menopause before 30 or after 62, they were also removed. And then any women that had not reached menopause or were within five years of the menopause transitional period. So within five years of their reported age at menopause, they were also removed. And the reason for this is that we didn't want to have any non-postmenopausal women and women within a menopausal transition period. This can be a pretty arach time women dealing with different magnitudes of menopausal symptoms. So we didn't want this to confound our frailty measures at all. So we left them out. And what we were left with was our five groups, premature menopause group. And these are women that reached menopause between 30 and 39 years. Early menopause group, and this was 40 to 45 years, normal menopause was within one standard deviation, a little less than one standard deviation of the agent menopause. And late menopause, which was 55 to 62. And then of course any women who reported having a hysterectomy. And then what we see here is the total sample within each group. And then the number of individuals that had frailty index and freed frailty measure scores that I could derive. So as I said, I classified frailty two ways. I used freed frailty-freedotype models. So I categorized it as healthy, pre-frail, or frail. And then the frailty index, I'm not going to talk about all the components that went into it, but I'm happy to share over email or for the discussion outside the webinar. It was a 93 component index spanning chronic diseases, functional status, ADLs, depression, satisfaction with life, nutritional risk, physical activity, and perceived health. And this is what they looked like in the CLSA Comprehensive Cohort. The frailty index averaged at 0.101. So that ends up being about nine deficits on average for each person with a standard deviation of about 0.07. And that's pretty common or that has also been shown in the literature. I had a minimum of 0 and maximum of 0.53 and increased nice and linearly with age, which has also been previously shown. Freed frailty, the percentages are pretty much exactly what we would think. The healthy group dominated the earlier years about 70% versus 30% of the frail groups at around 40 to 45. That drops with age, and the pre-frail group and the frail group slowly increase. But the frail group didn't really increase till about 70 years old. But again, this is probably what we would expect. So for the analysis, we treated menopause two ways. We treated it as a continuous variable, so that was just the straight agent menopause. And then we also treated it as a categorical variable. And that would be those five categories that I mentioned previously. We did two types of analyses. For freed frailty, we compared the frail versus a combination of the healthy and pre-frail groups using binomial logistic regression. And for the frailty index, we used linear regression. The covariates were chosen a priori, so that was based on previous studies with frailty. And those include age, marital status, ethnicity, co-residents, smoking, alcohol consumption, annual income education, social support, which was defined using the social support survey. So that is an ordinal variable between zero and five, and hormone replacement therapy use ever. So that's yes or no whether they've ever used hormone replacement therapy. So these are the averages for the free and the frailty indexes. And what we see here, I've included male as well as the other categories, and this is exactly what I was hypothesizing. This is exactly what I was expecting, is that males were consistently the lowest, followed by late and normal menopause group. The late menopause group seems to dip a little bit lower, and that might just be because of lower numbers relative to the other groups. Then above the normal menopause group, we have the early menopause group, and above that we have the hysterectomy group, as well as the premature menopause group. And you can see the premature menopause group has a bit of a weird pattern, and the reason for that probably most definitely is because of the relative lower numbers. I think there were only about 300 individuals that fell in that group, so likelihood it was just difficult to get an accurate estimation of the mean. So for Fritz model, we're looking here at just the percentage of frail individuals, and we really don't see a heck of a lot of discrimination until we reach about 65-70 years old, and this is when we know that the frail group percentage actually increases for the entire cohort, and all we see there is that the hysterectomy group starts to separate out from the other groups. So here we have results from the association analysis. So we have unadjusted and then fully adjusted models. Those were adjusted using the covariates that I previously mentioned. So for the frailty index, what we saw was exactly what we hypothesized. So for women who reached menopause prematurely, their frailty index scores were .024, so that's about 24% of the main higher than women who reached menopause normally. Early menopause was about 8% higher than normal, and women who reported having hysterectomy were about 21% higher than the normal menopause group. If we look at Fritz frailty, again, this really is similar to what was shown in the average plots in the previous slide. We don't see a heck of a lot of significance here except for the hysterectomy group, which showed about 1.5 odds of being frail versus being healthier pre-frail as compared to the normal menopause group. Now, one thing you might be asking, and it's an obvious assumption, is what's the relationship with hormone replacement therapy use, especially since my hypothesis for this, at least the biological mechanism, is really rooted in the loss of female sex hormones. So it would make a lot of sense that hormone replacement therapy should be associated with frailty. So we also looked at that as a secondary outcome. Oops, I'm getting ahead of myself. This is the continuous measure of age at menopause. So this is the regression coefficient and 95% confidence interval in the fully adjusted model, and what we see here is that there's an inverse relationship as women reach menopause later. Their frailty index scores are lower, and that's about 75.5% of the mean per five years difference in age at menopause. So not a big, big difference here, not a big effect, but it is significant. And I think it's quite validating that both the continuous measure as well as when we looked at menopause categorically both associated with frailty. So to the hormone replacement therapy now. So this is what we see when we look at the different variables that are available in the CLSA. So available in the data set is HRT use ever, yes or no, and those were used, that was used as a covariate for the previous models. But then we also have a length of HRT use in years, age at onset of HRT in years, and then the type of HRT used. And we have our unadjusted and then fully adjusted models. And the length, age, and type of HRT use, those were all included as fixed effects in the same model. And HRT use was in a separate model. We don't see a lot here. We see that HRT use is associated with 5% of the mean increase in frailty index. We don't see much for any of the other variables. We don't see much for freeze frailty. I want to caution conclusions from these results, really take them with a grain of salt, because the HRT use variables and questions in the CLSA are excellent. The data is excellent, but I think there's a lot of latent complexities to this data that I may not be giving full justification. So for example, when you look at age of HRT onset, one thing that you see is there are a number of women who report their HRT onset was about 30 years prior to their age at monopause. So it's unlikely that women are taking hormone replacement therapy similar to say what they would for menopausal symptoms in their 20s. And most likely they may have confused it with hormone contraceptives, so birth control. So I think that is an aspect that can be teased out in this data. And with that too, if you're familiar with the patterns of HRT use over the years, you'll know that in the early 2000s results from the large RCT, the Women's Health Initiative, came out showing that hormone replacement therapy use was actually related significantly related to negative outcomes, namely cardiovascular disease and ovarian cancer, breast cancer, women's cancers. So that changed the patterns of prescription of HRT quite dramatically. So there's a lot going on with this data here. So I don't think from this we can really say whether or not HRT is beneficial or not beneficial. Okay, so to summarize that data, what I showed is frailty significantly hired women that reported having hysterectomy or reached menopause prior to 46 years old. And it's about 7.5% lower with every five years of age at natural menopause. And there's few association with HRT related variables. But as I said, the complexities of this data I think are worth another look. This is a good point. I'll plug the paper for this data, which was recently accepted. So this was done by myself and Halitimim at York University, who's also published a couple papers on age at menopause in the CLSA, as well as hormone replacement therapy use in the CLSA. So thank you to her for that. So in the next part of this talk, I'm going to bring back this figure that showed two trajectories, and we talked about them as being healthy and unhealthy aging and how they relate to frailty and so on. And what exactly these authors described as the unhealthy aging, they described it as accelerated aging. So I think a lot of people can conceptualize what accelerated aging is and what it means. But let's talk about it a little bit further, actually define it. There's my little question mark guy. So accelerated aging simply enough is the difference between somebody's biological age and their chronological age. And this is really where the important point is, is biological age and how that differs from someone's chronological age, which is based on their birth year. So biological age is actually based on biological factors, so biomarkers, if you will, that change with age. It's really how old your body is telling you that you are, at least different measures that we can take from somebody's body, how old someone is based on that. Now is it important though? Well, there is a lot of data being produced that says it's very important. So here's a study that was very recently published, and it is from a large Korean longitudinal study. And they looked at 17-year survival for community-dwelling adults. And they measured biological age in these individuals and then used that to calculate accelerated aging. But what they called is the age diff. Again, just the difference between the biological age and chronological age. And what they show is for 17-year survival, individuals who have an age diff or an age acceleration less than two years, so that means their biological age is less than two years different from their chronological age. They have the highest survival probability, you know, around 95% over 17 years. Individuals that are between two and five years from their chronological age have a bit lower survival probability. It drops only a few percent. But it's the group that has an age diff of greater than or equal to five years. So this would mean that somebody who's 75 years old, their biological age is at least 80, somebody who's 65, their biological age is at age is at least 70. They have the lowest survival probability, about 80%, which is quite substantially lower than the estimates for the other two groups. So again, this is mortality and this is kind of the be all and end all. But what I'm really showing you here is that, yes, this seems to be important. So how exactly do you measure biological age? Well, it's actually quite simple. So you take some biomarkers. So I have a few examples here, neutrophil count, lung function, cognitive score. And we know that they go up with age and they go down with age. And each of the plots are on the x-axis is age between 40 and 90 years old. And then the y-axis is just arbitrary numbers that represent each biomarker. And then what you can do is you can just take three examples. So we have three individuals here, 55 year old, that is, we'll call them normal. So they're not too healthy. They're not too unhealthy. They're doing pretty good. We have an 85 year old who's very healthy. So maybe this man or woman exercises regularly, eats well, has a lot of friends, has a pretty happy life. And then we have a 65 year old, maybe they're frail, maybe they're sick, have a multi-morbidity, maybe they're depressed, whatever the reason might be, they're not doing well. And then we would do these measurements in these individuals. And this is where we'd expect them. The fall, the 55 year old would be right on this line of best fit, this regression line for all the markers. The 85 year old is a healthy individual. They'd be a bit lower for their neutrophils. They'd be a bit higher. They'd have good lung function, great cognitive score for their age. And then the 65 year old would be pretty much the opposite. Now there are a number of ways to estimate biological age, but using any of those methods, this is what you'd end up with. The 55 year old would have biological age about the same as their chronological age. And they're even Steven, they're right on the line. The 85 year old would therefore look younger. They'd actually look about 18 years younger. And then you have the 65 year old who'd actually look about the same age as the 85 year old. They'd look about 21 years older than they actually are. And we could say that these individuals are undergoing age acceleration. But the main point being here is that they'd be at a much higher risk of mortality based on the data I showed in the previous slide, as well as a number of other conditions. So this has led me to think, led me to theorize or hypothesize. Well, you know, that figure said that frailty, accelerated aging, healthy aging, they're kind of similar concepts. Then wouldn't early menopause and hysterectomy be related to accelerated aging when one possibly promotes the other? And possibly even early menopause or hysterectomy would lead to accelerated aging, which would promote frailty. Now, this is completely theoretical. I have no data to back this up, but I think this is very plausible. So it turns out I was right. Unfortunately, somebody else beat me to the punch. Indeed, early menopause is related to accelerated aging. So this was published just a couple years ago by Morgan Levine and Steve Horvath. And the way they estimated biological aging is they didn't use any of the biomarkers similar to as I showed before. What they used were epigenetic markers. And so these are little genomic markers based on, markers that are based on essentially people's genome. And you can use that to estimate age in very similar ways. You'd use any other biomarkers and what they showed is indeed using a number of different studies that women who reach menopause really are likely to have higher biological age relative to their chronological age. So nonetheless, this data came out, but I was very interested in seeing if we could use the markers that are available from the CLSA right now and see if they are related to early menopause and hysterectomy. So to do this, again, I use CLSA comprehensive baseline data set, same exclusion and classification criteria as previous. And to estimate biological age, I use an equation developed by Clemera in Dubal in 2006. So this has been shown by quite a few studies that it is one of the optimal ways of estimating biological age. And I estimate biological age by training on a random sample of 80% of the comprehensive data set. So what you'll see from my sample sizes of data that I'll be showing is only on a few thousand individuals, and that would have been my test data set. I defined accelerated aging as Delta BA. So that's the difference between biological age and chronological age. That would be the same as the age diff that I showed in the previous Korean study. And I used 27 biomarkers to estimate biological age. So I use hematology markers. So these are the complete blood count data that we have available. What I categorize as physiological markers, so blood pressure, pulse, perometry data, lean mass, performance markers. So these are our physical function tests like gate speed, the time to get up and go, grip strength, and the cognitive tests. So the battery of cognitive tests that we have available, the mat, the coad, the strupe, etc. So what does biological look like? Biological age look like in the CLSA. Well, using these markers and estimate biological age, it looks exactly what I was hoping it would look like. So there's a nice linear association between biological age and chronological age, correlation of nearly 0.7, minimum age of 27, and a maximum estimation of 109. If we look only at the Delta BA, so this is the difference between the biological and the chronological age, what we see is the average is a little less than zero. So there's about minus one year average for the population. And I don't know if that has to do with Canadians being maybe a little bit healthier than most. Not really sure. And the minimum and maximum was minus 29 years and plus 38 years. And then the mean absolute error, so that is the average absolute Delta BA was about 7.2 years for the test data set, which was about 2,860 individuals. Okay. So firstly, the major assumption here is that biological age and the Delta BA should be related to frailty. And indeed it is. So for every year of difference between Delta BA or between biological age and chronological age, the frailty index increases about 1.5%. So for every 10 years difference, frailty increases about 15% of the mean. And then there were no associations between freed frailty. So the Delta BA was about the same for every single group we looked at. I don't really know why this is right now. Probably something to do with the number of components in freed frailty, potentially that it's more focused on physical frailty than many of the other dimensions that frailty index does look at. But nonetheless, nothing showed up for between the different categories for freed frailty. So if we look at biological age and age at menopause as a continuous variable, what we see that it is indeed significantly associated, there's an inverse relationship between the two. So using the fully adjusted model, so this is incorporating those covariates that I had mentioned previously, it looks like for every year of age at menopause, there's about a one and a half month reduction in biological age. This is not a huge difference. It's significant, but it's a pretty minute impact it looks like. Now when we look at the categories though, it becomes a little bit more significant. So this is again, similar to the previous results I had shown, these are fully adjusted, these are regression coefficients from fully adjusted models. And what we see here is that for the premature menopause group, the age at menopause, sorry, the biological age difference was about 2.8 years. So that means that biological age is about 2.8 years higher for the premature aging group compared to chronological age as compared to what was seen for the normal group. The early menopause group had a delta BA of about 1.4 higher than the normal group. And then the hysterectomy group was about the same, about 1.5 years of a difference between biological and chronological age as compared to the normal menopause group. So we're seeing pretty much the same as what the previous study from the Horovath lab showed. We're adding to it that hysterectomy also has an effect on the estimate for biological age. Okay, so I'll summarize here. This is it. Similar to frailty age at menopause or menopause classification associated with increased biological age, so accelerated aging. And then premature menopause seemed to have the most substantial effect followed by early menopause or having had a hysterectomy. Okay, so that's it for me. I have some thanks to give here. Collaborators on these studies, Hallett Mim from Orc University who helped me out a lot with the age at menopause study. Thank you to Halla. Dan Belsky from Duke who's helping me out a lot with the biological age estimations. Of course, individuals from the CLSA, David who's helped me out greatly generating the frailty index, but understanding the frailty index. We've had a lot of great shots about that. Jin-Wi who's helped me out a ton with statistics for both of these studies. And of course, I couldn't forget Lauren Griffith and Perman Durena who's helped me out with frailty with aging with epidemiological methods. I mean, you name it. They've been great help, so thank you to them. And more recently, Stacey Vol from University of Victoria who's helped me out a lot with understanding the cognitive measures which I incorporated into the biological age estimation. Lastly, funding for the CLSA data that was provided by the McMaster University Institute for Research on Aging. And don't be poaching yourself or any of your friends, but I think that's a really funny figure. So thank you. Thanks, Chris. I think that was a really great presentation. Thanks for being here. Thanks, girl. I'd now like to open it up for questions for Dr. Bashore who I think we all now agree is the distinguished Dr. Bashore. So first question is from Linda Strobel. Did the participants with hysterectomy also have bilateral euphorectomy? So do they have their ovaries removed as well? Yeah, that I don't know. I don't believe the question. First of all, it is not a question that's directly asked. Someone can correct me if I'm wrong. The hysterectomy question is not one of those directly asked. It's a if the participant mentions that they did hysterectomy that are classified as such. So it doesn't differentiate by bilateral euphorectomy or complete hysterectomy or anything, any other variant of a hysterectomy. And also it doesn't, unfortunately, we don't have the age at hysterectomy. I think that would have been very, very telling and would have added a lot to the association analysis if we'd known exactly when they'd had a hysterectomy. And that would probably have made the HRT-related data make a little bit more sense as well. And a follow-up question real quick. I guess we don't know. The participants with HRT, were they using it because of that? We don't have much more data on the reason why people were using the HRT. Yeah, it doesn't seem like it, unless I'm missing something with that data. It's a straight question of when you took it and how long you took it and what you took. The length of taking it didn't seem to make a huge effect. It didn't look like it. No, it didn't look like the length of time taking it. But I think I put them in the model as simple of age and length. And there might be other ways that you can derive maybe a time since taken, since menopause, or something along those lines. Like I said, I think there's a lot more that can be done with those variables that I didn't do quite in justice. And I made it pretty clear in the paper. So Rosemary Clark, specifically, did you adjust for comorbidities and lifestyle habits? So because the frailty index included comorbidities, most comorbidities, sorry, that's probably a bad thing to say, but it glued a number of comorbidities. I did not include it in the model as a fixed effect only because they would be inherently related to each other. But I think there's probably some sensitive analysis that could be performed to look at that. I mean, I did include BMI. My reviewer had asked for that to have that in the model regardless, even though it was in the frailty index. And it didn't change the estimates appreciably. As for lifestyle habits, while we had smoking, we had alcohol consumption, what else was in there? I'd have to go back through the slides, and I think social support, which I guess whether or not you'd argue that's a lifestyle habit. Kind of a follow-up from Don Bowdish, the largest factor that influenced biological age. Did you look at any of the other direct impacts of those things that you were putting into that? What is driving at all? I guess the question, Don, would be, are you asking what factors, person factors that don't make up biological age, which one of those are the strongest? Is sex the biggest thing that's associated with it, or is it smoking, or a BMI? Or do you mean within the biological age estimate, which components, say, lung function or something else, what makes drives the biological age estimate the most? I don't know if you can type that fast. I meant within the biology. Yeah, that's actually really interesting. And that is something I'm working on actually right now is sort of a paper. It's not really a follow-up paper to this work. It won't be on menopause, but it will be on relates to frailty and hematology measures. And actually, and if I can quickly pull it up here on my computer without ruining anything, I have done that. What it looks like, at least in the biomarkers that I'm using right now, I don't have access to all the biomarkers, but it looks like for both men and women, spirometry, so whether you use force vital capacity or the forced expiratory volume in one second, it seems to have the strongest association or drives the biological age estimate the most followed by the balance test, which drops quite a bit, but followed by balance, the strupe test, so that's a cognitive measure. And then in men and women, it changes a little bit, but the top markers for both men and women end up being diastolic and systolic blood pressure, lean body mass, and what is common between the two. While we also have the timed up and go, we have choice reaction time, which I think that has some limitations to it's grip strength, but really it's lung function balance and the strupe test seem to be really the strongest markers that drive the biological age estimation. Now mind you though, I mean I'll add to that without trying to take up too much more time, is that they are the strongest drivers of the biological age estimate and by that I mean that when you use them you get a much tighter association with chronological age. The error is much lower when you have them included, but just because they lead to the most accurate or the least error biological age estimate, it doesn't necessarily mean that they are associated with every adverse health outcome equally. If we look at different diseases or other types of negative related health outcomes, we see things that if you weight a biological age clock, if you will, we can call it that more on the performance measures, so that would be grip strength, gate speed, time get up and go, chair raise. You actually have a stronger association to disability, so that's having at least one or more activity, basic or instrumental activity, daily living as compared to any other clock. For the same and more healthy and more obvious example would be people with asthma. If you have a clock that includes any of the spirometry measures, that clock is going to have a higher association for people, for the prevalence of asthma compared to other diseases. So anyways, what I want to get at is that we know some heavy duty drivers of this estimate, but the actual health outcome that they're related to can change depending on what you're looking at. So I have a lot of other questions, but I think we need to go ahead and finish up here. So one last question from Sheldzowski. Can you comment on the HRT supplementation? Did you know actually what kind of HRT? And I'll put a follow-up question in there as well, which is, do you think that there could be an intervention for, is there an intervention for increasing age to menopause, and would you think that that would improve healthy aging if people actually took that up as kind of a public health message? Okay, so for that latter part, I don't know, and I don't think there is. I think when it comes to exogenous hormones being prescribed since the Women's Health Initiative, that's, I think, a very tenuous area. A lot of what's come out since then has shown that, okay, it's not flat out that hormone replacement therapy is bad for your free health. It's really about when it's prescribed and the type that is prescribed and background, a women's background, and whether or not it's going to have a negative health outcome. Whether or not there's any interventions to increase the age of menopause, that I don't know. I pulled up the slide here. We do know what types of HRT was used. Okay, well thank you again. I really enjoyed that and I thought it was great. I'd like to remind everyone that CLSA data access request applications are ongoing. The next deadline for applications is September 24th, 2018. Please visit the CLSA website under Data Access to review available data for their information and details about the application process. Also, a reminder that today is the last of our 2017-2018 webinar series. We'll have a short summer break before starting our 2018-2019 series in September. In September, we'll be welcoming Dr. Daryl Leong, Assistant Professor of Division of Cardiology, Department of Medicine, and an Investigator Population Health Research Institute to discuss global importance of frailty and pre-frailty and middle-aged adults of pure study. So enjoy your summer and we'll see you in September for a brand new webinar series from the CLSA. Please go to our CLSA website anytime to familiarize yourself with our platform and to register for our webinar series soon and join us for our upcoming webinar series. Thank you again for joining today's presentation and thank you again, Dr. Chris Bashore.