Episode 11 - Looking Your Age

What does your perceived age say about you? In a special 'deep dive' into one of the BJD's recent studies, we talk to Prof Tamar Nijsten from Erasmus MC. Prof Nijsten recently led a team publishing a study on perceived ageing and its association with internal diseases of ageing. In this episode, we discuss ageing and epidemiology within dermatology.

Transcript

SPEAKER_01 0:18

Hello there and welcome to BJD Talks, where the official podcast of the British Journal of Dermatology. In our podcast, we delve into publish studies and explore real world implications of dermatology research in a way that we hope is accessible. Our podcast is for anyone with an interest in skin health research. So that's whether you're a dermatology professor, researcher, registrar, patient, or even simply a skin enthusiast. We hope you'll join us as we build on our world-leading research with friendly discussion. My name is Dr. Johnny Guccian, and I'm a dermatology registrar in Leeds, in Yorkshire in the UK, as well as the BJD's Podcast Associate Editor. Together we'll look at a huge range of issues which are important in dermatology. These include areas like social media in dermatology, artificial intelligence, and a variety of subspecialties within dermatology scholarship. In this season of BJD Talks, we will, from time to time, focus in on some key, interesting papers published in our journal. We'll speak to the authors who undertake the incredible work which impacts outcomes for real patients and pose important questions for future research. So this time we're going to kick off with Professor Tamar Neisten. Prof. Neichten is a dermatology professor at Erasmus MC in Rotterdam, where he heads a dermatoepidemiology group. He led a team of researchers who investigated the association between perceived age, or younger looks, and age-associated morbidities. The paper is called Younger Facial Looks Are Associate with a Lower Likelihood of Several Age-related morbidities in the middle age to elderly. And you can check it out on the BJD website alongside this podcast right now. The findings were fascinating. So welcome to the podcast, Prof Nashton. Thanks so much for coming on. You're welcome. So could you just tell us a little bit about the study that you and your team undertook?

SPEAKER_00 2:20

Well, it all started about 10 years ago, I guess, where we were involved in this population-based cohort study, it was what we call the Rotterdam study. It's basically they follow up about 15,000 people now for approximately 25 to 30 years. And when we embarked in the Rotterdam study, we were most interested in skin cancer. And to look at skin cancer, you know, you have to ask questions about UV exposure, etc. But they're very difficult. So we thought, okay, well, let's just take a picture of the face, and maybe that can help us to differentiate who has high UV exposure or who hasn't. And this was something we thought of years ago. And all of a sudden, this this data collection of photographs, they kind of started to live their own lives in a way, scientifically. And we hooked up with people across the Netherlands, but also with uh with people in the UK to look at skin aging or healthy aging or skin aging. And honestly, in the beginning that was not really my cup of tea, and I'm not gonna consider myself an expert in skin aging, and I have no aspiration in cosmetic dermatology, etc. But it was just good fun, and the data was so nice to look at. But as a physician, of course, the question was always nagging in the back of my mind what is the clinical relevance of looking at skin aging? And thus the facial skin mirrors some kind of internal aging as well. So after, I don't know, four or five PhD students working on skin aging, we finally got the work done to say, okay, what does the skin reveal about the internal organs?

SPEAKER_01 4:08

I think that what's fascinating there is that that commonality between all good research is that so what question. And that that as you say, that nagging feeling at the back of your mind, what what is the clinical relevance? And it's interesting that you can take something that you know you yourself might not have been all that interested in, you know, and with the parallels to cosmetics, as you say, and find that real internal medicine so what, which is which is really interesting. Um what I really like is that your team very eloquently starts the paper by saying, Since time immemorial, mankind has been on a quest for the source of eternal youth. Uh love that. Um and I think that's something that a lot of our listeners might empathize with. So why did you think at the start that it's so important to study and understand this perceived aging process?

SPEAKER_00 4:55

Well, if you think about perceived age, it's easy to kind of downgrade it as a very subjective measurement. When you look in the mirror every morning, there are differences across the week or months, or you do see change over time. And the thing is, what triggered me is can we kind of objectify that and quantify that in a way that it makes scientifically sense? And that by itself is a challenge. And just getting your thoughts around perceived aging was a was a very nice challenge, and we work together with the people from Unilever in the UK, and they helped us tremendously with that. And then, of course, you know, we are physicians in the end of the day. So if you look bad, are you bad? And what is the relationship? So I thought that was just yeah, there are two interesting parts of of the study. So then the first part of what how do you quantify perceived age is what we basically did is we had 3D images of 2,500 people, lots of images, and they were presented to a panel of 25 people from the general population, and they were asked, How old do you think this person is? And then we compared the average of that 25 people panel with their actual calendar years, and that basically creates a delta, and that delta tells you something, you know, do I look younger, better than I actually am, or is it the other way around? And it's funny that there are really differences, and you can also kind of objectify why some people are scored older than they are actually are and younger than they actually are. So we had a research line around this, and when I initially in the postcard said, you know, I'm not into cosmetics, that's what that's basically what the first reaction of people is. But the the whole concept of healthy aging is also telling you that if you understand how something works normally, that can really reveal about what goes wrong in the disease. And we as physicians we are so used to think about this issue the other way around, where we say, okay, we start with the disease, and if we know what the gene is that causes the disease, then we know what the protein is that from the gene results, and then we understand the disease. And by understanding the disease, we understand physiology. And in this case, we turn it around and say, okay, let's start with physiology and see whether we can explain what is normal. So that's just coming back to my first point. The other point was about objectifying perceived AIDS, which also has a fun factor to it in a way, and so we're able to do that, and it was actually associated with genes, and we could, you know, we did GWAS analysis and conventional EPI approaches, and it all made some kind of sense. And then the question, of course, as as a physician, you just want to know it is intriguing to think if facial appearance actually tells you something about other aging processes. So, what we did is we went to the PIs of all the different organ blocks in in the Rotterdam study and say, Look, of your organ, get me one or two or three diseases which you think are really age-related. For example, we have done to the ophthalmology people, and they said, Well, you know, for us, cataract is really the aging disease. And you went to the pulmonology people and they said, Well, you know, if you want to look at aging, you know, maybe COPD is something of interest to you, osteoporosis. So we collected a whole bunch of diseases which were not selected by us, because there's always this danger of that you kind of start cherry-picking, but we let other people decide what they thought were interesting AIDS-related diseases, and then we just started comparing.

SPEAKER_01 8:40

Fantastic. And but so before we talk about those, because I we'll need to talk about your results because it's they're really interesting. I'm I'm fascinated by the methodology, and you know, you could do a whole episode just on that, a whole paper seemingly just on the process, because it's quite novel, isn't it? And I had a couple of questions. Firstly, how did you pick your 25 people?

SPEAKER_00 8:58

Yeah, so this this was a a panel from the UK, which was already established, and they had done something similar before. So these people, not all 25, looked at the 2,500 pictures, but they looked at several hundreds or thousands of pictures, and we made sure that every image was at least scored by 25, and they were from the general population, and they were paid for it.

SPEAKER_01 9:24

Yeah, I guess I'm gonna say it sounds like a lot of work. In my head, I was imagining like those TV shows where you have people like it with with microphones in the street and saying, How old do you think this person is? So I was sure it was more scientifically robust than that, but but um at the same time, I think that would be interesting.

SPEAKER_00 9:41

That would be interesting to see. In contrast to the microphone, this image would not kick back or respond back to you. So you are uh you know, we're a little bit more secure. And the and the thing is, of course, if you ask this to one person, then the variation or the the the reliability of that observation has some kind of limits. But if you ask these questions to a bunch of people, and the more people you ask it, the more likely it is to be quite accurate, actually. And somewhere in statistics, uh 25 is a golden number. Uh we often think that if you do 25, then you can start doing and applying statistics to comparisons, etc.

SPEAKER_01 10:24

Absolutely. Uh and I feel like you'd have to be a bit brave to give up your photographs to be judged by by all the people. But then I think people would people would do it, people do these things. And so I I I mean saying that, I I'm I'm quite glad that this podcast is recorded um uh without video because uh um there's a reason there's a reason I've done this by audio rather rather rather than video, because I wouldn't want to give myself up right now. I've I I've spent six months doing on calls and I feel like I made a hundred years. Um another question I had was just you know talking about the data and the and the statistics and things of things, we're we're living in an age of of AI and big data. And do you see bots on algorithms, etc., online which claim to be able to guess your age? And some of them are just like those BuzzFeed quizzes which are asking what kind of bread are you, but some of them claim to have proper methodology behind them. Do you see that that that um artificial intelligence might might be able to undertake this this kind of future?

SPEAKER_00 11:25

Yeah, absolutely, Johnny. That's an excellent question because that's what we are doing. I mean, and and this is also so opportunistic approach to it because we have over 10,000 pictures, but we can't report them to have them scoreboard 25 people all over again. And we have now collected multiple images of people in time as well. So we have about I don't know, 15,000, 20,000 images. So we need a more automated approach to kind of yeah, to objectify perceived aids. So we're actually working on that with an image group working on the AI, and can we kind of figure out how to do that? A very clever bioinformatician from Newcastle actually flew over and worked with us for two or three weeks to kind of help us with this issue. Unfortunately, not perfect yet. Keep on working on it, and I think that is, of course, the future for lots of image-related analysis because otherwise it's just too cumbersome, too much work. Yeah, so in the future we will go that way, absolutely.

SPEAKER_01 12:24

Fantastic. And and for listeners, we we do have an episode of BJD Talks on Artificial Intelligence and Big Data Um earlier in the series. So do check that out. Now we mentioned your results, and I stopped you before you gave us any spoilers. You mentioned some conditions. Um, which conditions are associated with this uh well impacted by perceived age?

SPEAKER_00 12:46

This is it's some almost some kind of hypothesis-free approach. So we say, okay, these are the diseases, just very exploratory, let's have a look. And about half of them are actually associated. Associated means that you have this significant p-value and everybody starts to be happy, but not necessarily tells you something about causation or about its effect size. So some of the things that popped up and I thought were really intriguing, was like osteoporosis, cataract, COBD, and also cognitive level of cognition. So they seem to be for their age somewhat more clever if you look younger. But some of the other things we looked at, like atrial fibrillation or renal impairments, or macular degeneration of the eye, they didn't pop up. So it's always interesting to see what you know what wasn't associated. Can it help? Can it explain something? Is there some kind of common factor or denominator behind it? And I think, especially, for example, if you look at osteoporosis, is a very nice example, I think. And the biology is a long way back, but fibroblasts and osteoblasts, they share common uh several letters at the end. So there seems to be something going on about you know, might it be true that if you adjust for all the lifestyle factors, etc., is there a biological explanation for this? That some of your blasts actually go into senescence early for some people, and that basically means that if your fibroblast becomes less active or starts to hibernate, that might actually also be true and be correlated to the same senescence states for osteoblasts for us. So that I think was a very intriguing finding. And I said, you know, the B-value of 0.05 is very arbitrary, so we're always interested in in effect sizes, and here you could actually see that we had like 30 or 35 percent protective effects, which in Ebby is actually quite high. Often we're talking about five or ten percent changes, so something seems to be going on at least. Fascinating.

SPEAKER_01 15:04

And just in terms of kind of next steps for this work, obviously, this, as with all research, that's it's a continuous process. And you know, you said this followed on from work in the Rotterdam study. What do you think is next in the world of perceived aging?

SPEAKER_00 15:17

For me, the pleasure in epidemiology is that it is hypothesis generating. We're not necessarily proving anything, but we're just you know You're asking questions. We're asking questions and we give some possible answers, and then they need to be proven further onwards. So in this design, for example, it's really a cross-sectional design. So we basically look at one point in time and say, okay, if we take this picture at this age, is that associated with some other diseases or health states? And in in any Epi research, and you have you would like to have a cohort effect where you have longitudinal data. So where you could say, okay, well, we have this picture at age 60, and then we have a picture at 65 or 70, and see whether that holds true. And you want to focus on incident cases. So, for example, you know, you could think about what we call um confounding by indication, is where you say, Well, if I have renal impairments, I look older. So if you judge my face, yeah, do not be surprised if you find renal impairment. So these are all kind of epi issues that need to be solved, and for that we need, like everything in life, you need more time, more data, more analysis. And that's something within the realm or or of what we can be, we should be able to do within the Rotterdam study. Like other Epi studies, it needs to be validated in different cohorts. So this might be very well true for people in Rotterdam, and of which the majority actually was Caucasian. They were selected 30 years ago because for genetic reasons, because they were now more homogeneous, but this needs to be validated in different countries, different more diverse populations, etc. And then the other part is that, of course, ideally we would have this confirmed in more laboratory studies where they start to compare aging processes of different organs or different cells within organs and see whether they are related. But that's a whole different ball game.

SPEAKER_01 17:27

Oh, yeah. I mean, there's there's you could almost take a different subspecialty within dermatological and scientific research for each of those studies, couldn't you? Yeah, absolutely. Um and one one thing I find interesting, you you you mentioned it was diverse populations, as you see you put in your limitations, you've discussed it here, is about the makeup of the population in this study. And I guess when you're dealing with people, you deal with people's biases and population biases, and you rarely get biases which are as sometimes as as strong as as racial factors and ethnic diversity. And I imagine that not only would it be interesting to see this replicate in different populations, but also judging perceived age across population groups, are there any issues with that? I mean, call me cynical, maybe I would imagine there might be, but would that be replicated in a scientific study? That that would that would be fascinating to see.

SPEAKER_00 18:20

Well, the perceived age is actually very variable across skin colors. So, you know, we we see that people with a darker skin actually really age differently. And and just think about taking a character like Morgan Freeman, you know, relatively few wrinkles, but what you do see is hyper-pigmented spot on the face, and so it it is very different. So it's gonna be quite challenging to do this in different populations because the way we kind of guess the age of other people's background is very challenging. For me, as a very wide person, it is quite challenging to estimate correctly the age of people with an Asian background, for example, or of a more African descent. Simply not trained for it yet. And it is different. So different populations, yeah, that's gonna be challenging because perceived age is gonna be different. Underlying diseases might be different because of different exposures, different genetic backgrounds, etc. But I do know that for many companies, for example, in personal hygiene care products are very interested in changes of Indian skin, Asian skin, etc. And we have to learn still quite a lot about that actually.

SPEAKER_01 19:39

Yeah, I I think it's one to get the sociologists and social scientists in in to help answer some of those questions. Those who really get the people before the numbers. Yeah. Absolutely fascinating. So I do recommend that listeners have a read of the study. Um, and there's an associate kind of information released from the BAD too, which is usefully translates this to the general population, because I think certainly the general population would be very interested in this. I'm certainly going to go consider just looking at getting some Botox from my face, looking at looking at myself on the screen right now. What I wanted to talk about just before we finish is we've talked a bit about epidemiology, and without going into the you know great, great depths of of of epidemiology, because again, that's a whole episode in itself. You are the section editor for epidemiology for the BJD. Can you tell us a bit about some of the work that's happening there at the moment, some of the uh interesting studies that have come out recently? What should our listeners be looking out for?

SPEAKER_00 20:34

Let me first respond to your your you know epidemiology is a bit kind of uh tricky, and I it's interesting that for for a lot of physicians, epidemiology is kind of uh is not very popular in a way. It is somewhat difficult on the one side, and statistics are not necessarily a hobby of many physicians either, but then again, good physicians are epidemiologists all the time in their daily practice because they observe, they look for associations and do risk assessments. So every good doctor is an epidemiologist, even if it's NS1 cases, or and then you reflect it to your other patient population. So for me, it is just more numbers, but it's the same way of thinking as any good physician would do. From a section point of view, I think that we started as a section editor about I don't know, six, seven, eight years ago, and I think what we try to do is learn from other sciences, so incorporate new methodology, new approaches, also learn from other specialties that replicating or validating your findings within one study is very much appreciated. So, yeah, we often talk about uh not slicing the salami too thin, where you say, Okay, I'm gonna look at this disease and this comorbidity, and then do it ten times over again for different. Comorbidities, that's really not the way forward. We're not helping patients, we're not helping anybody by this approach. So we we need a broader perspective and think about more collaborations across people, across specialties, across nations to validate each other's findings, and that's a really scientific approach. The other approach is that we see we have a wish and a tendency for more clinical relevant studies. And this sounds like so obvious as as clinicians, but sometimes we're just well, we're kind of data dredging and looking at all different kinds of stuff. So where you think, okay, does this really matter? And if it does matter, how would this affect patient care? And and also thinking about translating your findings to such a way that physicians can use them in practice. So if I present you a table with all hazard ratios or odds ratios, and no way you can apply that to an individual patient. So, you know, you can say it is a bit higher or lower risk, uh, it doesn't look good for you, maybe, but I'm not sure. Uh this are numbers working for a group, but not for individual people. So I think there the transition from risk prediction to prediction models is I think something very interesting and intriguing. Because that is something we can use in clinical practice as well. Yeah, I think these are are some of the well the changes we see over time. And I think that what what is also happening is that, and this is true for I think dermatology in in a very broad sense, is that we are being more and more recognized as a medical specialist in the scientific field. So we talked about the Rottenham study earlier on, you know, it the Rottenham study was already ongoing for 15 years, and nobody included a dermatologist in the beginning. So we kind of basically had to fight our way in. And this was a cohort study focused on aging in elderly people, and nobody thought about the skin. And then this is just telling you something about how we have evolved as a specialty as well, where I think the whole drug development in psoriasis has put us on the map. The increase of skin cancer, which is not going down, has made it a bigger public health issue. What is happening right now with the topic dermatitis and the rise of it, of course, for decades, but also the new treatments, is really making Durham a more appreciated specialty, which then means that we are incorporated or invited to participate in bigger consortia, in the UK Biobank, in the Nurses Health Study in Boston, in different resources, skin is in the game now. And that is something very, very nice.

SPEAKER_01 24:55

Play on words, skin in the skin in the game. I love that. And one thing that some listeners might recall from what I've talked about in previous episodes is that I like about your answer and throughout this chat about the study is that dermatological relationship with internal medicine and with you know with it with the so what that the lay person or the non-dermatologist may be interested in. And I feel that as you say, uh more recently dermatology has become feeling more legitimized, yes, but also just becoming more of an integral part of the community, you know, the the medical community, and engaging more rather than being off an asilo and viewing the skin as being in isolation when actually it's so connected to everything that we do. I think that's really interesting. I I like that. And you might be encouraged to hear that when I started medical school back, I think it was 2009, the first specialty I wanted to do was epidemiology. Then I realized I was I was terrible at maths. Um so uh whilst I've always had an acquired appreciation for it, I have a lot of respect uh and perhaps professional envy um regarding uh um the work the work that you you guys do. With with that in mind, it so let's say you have someone who's just getting started in dermatology or you know, maybe thinking about doing a bit of epidemiology work or research. Is there anywhere that you could point them in terms of resources, papers to have a look at um to get started?

SPEAKER_00 26:23

Yeah, I think if you if you I don't know whether people still read books, uh but when I started, I you know, you buy a couple of Epi books, and they are basically the best way of not getting interested in the subject. Because they are so dry and so difficult that you you know you can finish them. So I think Epi is something you start to appreciate while doing. And of course, there are you know nowadays they're on the internet and many introductions uh in in Epi. Um but the best way is just start doing it, and you know many medical students or or young residents you know they will do a case report, and then it's a case series, and then maybe uh a small cohort, and all of a sudden you're an epidemiology. And even a case series is in fact the beginning of epidemiology and and clinical science. What worked for me when I was younger, I actually never wanted to be an epidemiologist. I was actually one of the things in med school where I thought this is not for me. So when I started writing my research grants, when I figured out that I wanted to do research, you know, was very much on molecular biology, you know. I think that was sexy and difficult, and that's where I needed to go. And then my mentor basically was an epidemiologist, he he kind of gently directed me into the direction of epidemiology, and what I learned from it is that just good sound reasoning is the basis of epidemiology, and trying to figure out what is going on, and I think that's that that's great fun. And I don't have to work with mice, and I don't have to work with fibroblasts and cells in a battery dish. You know, I can work with actual people. I mean, that's just great, you know, and and and that's it. Yeah, so I think finding a good mentor is basically the beginning of a a successful career in anything.

SPEAKER_01 28:23

Absolutely, and I love that because it's it's it's so similar to to other areas and other answers that guests I've had in this podcast have shared. It's mentorship and role modelling are so important in career formation. And I also think that with an epidemiology, you've talked a lot about how you guys like data and uh you know lots of lots of data as well. So I guess that would lend itself to a lot of collaboration with other centres and other groups, and there are echoes of that in getting started in these subspecialties because I always recommend to, as I say, medical students or um early career doctors who I mentor that never start something by yourself or never see that you have to start something by yourself because that can seem so intimidating, especially in a new completely new field. So collaboration is is your friend and you can learn so much in so many unexpected ways. So I think that's a that was uh some some of my mentors have always have always taught me. Get get make friends and get help. Yeah. I I guess in in an epi, knowing one or two statisticians is always gonna be very is always gonna be very helpful if you're like me and allergic to maths.

SPEAKER_00 29:29

We have a saying in touch that basically goes something like uh better well stolen than uh badly thought of. So don't be afraid to copycat, you know? It's always a brilliant idea and they did a very excellent job. Look at how they did it. Look at how they structured their paper, what type of analysis they've done, and just start with doing it over again. And and and it's just learning by stealing in a way.

SPEAKER_01 29:56

Well, there you go. So we have this formal advice from Prof. Neison uh in order to make epidemiology sexy again, the answer is a bit of light plagiarism. Uh um and and I think with that optimistic note um and a bit a bit of excellent wisdom, we'll we'll draw this episode to to an end. Thank you so much um for for joining. We've talked a bit about delving into this uh perceived aging study and had a chat uh about epidemiology within dermatology more widely. Um so thank you very much, Prof Niceon, for joining. You're very much welcome. It was a pleasure. Brilliant. And so uh to listeners, we look forward to sharing our next episode of BJD Talks. But in the meantime, please do get in touch with us and let us know if there are any hot topics within dermatology or even outside dermatology that you reckon we should discuss. Uh we're always on social media, uh and you can grab us on at brjdermatol on Twitter and at brjdermatology on Instagram, or by using the hashtag hashtag BJD Talks. Bye for now.