View  Transcript

The population of older adults living with dementia is expected to swell to nearly 14 million by 2050 and is estimated to cost the US economy more than 500 billion each year. In the absence of a cure for Alzheimer's disease, the primary cause of dementia, there's interest in understanding modifiable risk factors. In theory, getting a handle on the modifiable risk factors for dementia, would enable public health efforts to reduce cognitive decline in dementia at the population level. We've come a long way in understanding the risk factors for Alzheimer's disease and other related dementias. However, there's still work to be done. In this episode, we'll speak with Dr. Josh Ehrlich, a researcher at the University of Michigan, who has examined vision impairment as a risk factor for dementia.

More resources

• Joshua Ehrlich Faculty Profile: https://medicine.umich.edu/dept/ophthalmology/joshua-r-ehrlich-md-mph
• Articles Referenced in the Podcast: 
• Ehrlich JR, Goldstein J, Swenor BK, Whitson H, Langa KM, Veliz P. Addition of Vision Impairment to a Life-Course Model of Potentially Modifiable Dementia Risk Factors in the US. JAMA Neurol. 2022 Jun 1;79(6):623-626. doi: 10.1001/jamaneurol.2022.0723. Erratum in: JAMA Neurol. 2022 Jun 1;79(6):634. PMID: 35467745; PMCID: PMC9039828.
• New York Times Article, July 2022: New Dementia Prevention Method May Be Behavioral, Not Prescribed
• Lancet Commission on Dementia Prevention, Intervention, and Care (2020) 
• Livingston G, Huntley J, Sommerlad A, Ames D, Ballard C, Banerjee S, Brayne C, Burns A, Cohen-Mansfield J, Cooper C, Costafreda SG, Dias A, Fox N, Gitlin LN, Howard R, Kales HC, Kivimäki M, Larson EB, Ogunniyi A, Orgeta V, Ritchie K, Rockwood K, Sampson EL, Samus Q, Schneider LS, Selbæk G, Teri L, Mukadam N. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020 Aug 8;396(10248):413-446. doi: 10.1016/S0140-6736(20)30367-6. Epub 2020 Jul 30. Erratum in: Lancet. 2023 Sep 30;402(10408):1132. PMID: 32738937; PMCID: PMC7392084.
• CAPRA Website: http://capra.med.umich.edu/

Transcript

Matt Davis:

The population of older adults living with dementia is expected to swell to nearly 14 million by 2050, and is estimated to cost the US economy more than 500 billion each year. In the absence of a cure for Alzheimer's disease, the primary cause of dementia, there's interest in understanding modifiable risk factors. In theory, getting a handle on the modifiable risk factors for dementia, in other words, the risk factors that could potentially be mitigated throughout a person's life, would enable public health efforts to reduce cognitive decline in dementia at the population level. We've come a long way in understanding the risk factors for Alzheimer's disease and other related dementias. However, there's still work to be done. In this episode, we'll speak with a researcher who has examined vision impairment as a risk factor for dementia. I'm Matt Davis.

Donovan Maust:

And I'm Donovan Maust.

Matt Davis:

You're listening to Minding Memory, a podcast devoted to exploring research on Alzheimer's disease and other related dementias.

Today we're joined by Dr. Joshua Ehrlich. Dr. Ehrlich is an ophthalmologist and assistant professor in the Department of Ophthalmology and Visual Sciences at the Kellogg Eye Center, as well as a research assistant professor at the Institute for Social Research at the University of Michigan. He's the co-director of the Kellogg Eye Center for International Ophthalmology that fosters international collaborations to increase research capacity and strengthen eyecare delivery systems around the world.

Dr. Ehrlich is a population health and health services researcher whose work focuses on understanding and addressing the influence of vision on health, disability, and quality of life, particularly among older adults, including the role of visual decline in cognitive psychosocial and physical wellbeing. His work has been continually supported by the National Institutes of Health, and he's here today to speak with us about his recent research that investigated vision impairment as a risk factor for dementia over the life course. Josh, welcome to the podcast.

Josh Ehrlich:

Hi, Matt. Thank you.

Matt Davis:

Dr. Ehrlich was the lead author of the study titled Addition of Vision Impairment to a Life Course Model of Potentially Modifiable Dementia Risk Factors in the US that was published in the Journal JAMA Neurology. The link to his article is attached to this episode.

So from what I gather, the motivation of your study was to explore vision impairment as a dementia risk factor, specifically included in models such as the Lancet Commission Life course model. Just to start things off, could you tell us a little bit about the 2020 Lancet Commission report?

Josh Ehrlich:

Yeah, Matt. The Lancet Commission on Dementia Prevention Intervention and Care has published two iterations of its seminal report. This report really serves as a blueprint for the clinical and research communities around key issues in dementia prevention, intervention and care, including an understanding of risk factors that are potentially modifiable across the life course. This report is highly influential, both in research circles, but also as far as funding goes, and in informing really the larger understanding of the key factors associated with, and important for, dementia prevention. So when we saw the second version of the Lancet Commission report, the first having been published in 2017, we were actually quite surprised to find vision impairment was not among the 12 potentially modifiable risk factors highlighted in the report.

Matt Davis:

What percent of total dementia cases are actually caused by modifiable risk factors according to the report?

Josh Ehrlich:

Well, the report that was published in 2020 cites a figure of around 40% of cases worldwide being potentially attributable to these 12 risk factors that were included in the commission's model.

Matt Davis:

What factors are most important, would you say?

Josh Ehrlich:

Yeah, so according to the risk factor model that the commission published, the number one risk factor was, in fact, hearing loss. We can talk more about why hearing loss and vision loss sort of are similar in many ways, as far as the risk that we think that they engender for dementia. But hearing loss, largely because of its very high prevalence worldwide, came out as the number one modifiable risk factor in the model, with the number two closely behind hearing loss being less education.

Matt Davis:

I think I read somewhere too that, or maybe you mentioned that air pollution is one of the factors. Is that right?

Josh Ehrlich:

Yeah, that's right. Air pollution was added. It was one of three risk factors added, between 2017 and 2020, to the model, and according to the commission's calculations, air pollution may account for somewhere around two to two and a half percent of worldwide dementia cases.

Matt Davis:

Seems like an odd one to include right off the bat, but-

Josh Ehrlich:

Well, I think there's been a growing interest in recognition of the potential effects of air pollution on a variety of systems, the least of which is not cognitive health.

Donovan Maust:

Do you think should vision impairment have been on the list?

Josh Ehrlich:

Yeah, so in my opinion, vision impairment should have been on the list. Vision impairment is, there's considerable both cross-sectional and longitudinal evidence showing that vision impairment increases the rate of cognitive decline, as well as the risk of incident dementia. And according to the commission report, when they selected their potentially modifiable risk factors, these were the kinds of things that they were looking for, was a body of evidence supporting this association, this kind of longitudinal association with incident dementia, and the evidence is out there. So it was actually quite surprising for us to observe that it was missing from the list.

Donovan Maust:

Asking you to speculate, is it maybe because the prevalence of visual impairment isn't as high, or is it that the science, relatively, is newer? If you had to speculate, do you think there's any reason it wasn't included?

Josh Ehrlich:

I think it's a great question, and of course I've pondered the same thing myself. I think a big part of it may be inadequate advocacy and visibility of the research to the larger dementia research community. I think that we, as vision researchers, have largely, to our own peril, siloed ourselves in the vision research community. And I think there's a real need to disseminate the research that we're doing, particularly as it relates to health and wellbeing, beyond just the eyes, to other research communities, to the gerontology community, to the dementia community, et cetera. And I think that lack of advocacy and visibility has played a major role.

Donovan Maust:

Really simple question. Is the Lancet report just about Alzheimer's or all different types of cases of dementia?

Josh Ehrlich:

It's about dementia more broadly.

Donovan Maust:

So this is probably kind of emerging, but I'm curious, what do we know about, and what's sort of the theory that underpins the connection between a sensory loss or sensory sort of deficit and the risk of dementia?

Josh Ehrlich:

Yeah, so another great question, and before I answer the question, I'll note that the theories, the hypotheses that are out there around the connection between vision impairment and dementia, are really identical to those same pathways and hypotheses that are out there for the connection around hearing loss and dementia, identical in fact. So one possibility, and I should preface this by saying there's probably no one right answer, no one pathway that fits all individuals, but the different pathways are likely at play for different people with different risk factors or profiles. One possibility though, is that there's a common etiology underlying the neurodegeneration involved in dementia and neurodegeneration or vasculopathy involved in eye disease. The eye is a direct extension of the brain. It contains neural tissue, and there's no reason to think that neurodegeneration couldn't, or vasculopathy could not affect both the brain and the eye simultaneously.

Another possibility, another hypothesis is that as we have impaired sensory function, we of course get less information sent to the brain through the visual pathways, or in the case of hearing, via the auditory pathways. And this decreased afferent input into the brain may over time cause a direct alteration of brain structure that could be pathological.

Another hypothesis is that because older adults with a sensory impairment need to do many of the same things that an older adult who's well sighted or can hear fine needs to do, that they're facing a higher cognitive load in trying to do these day-to-day tasks, balance a checkbook, go grocery shopping, et cetera, and that this increased cognitive load, over time, may come at a cost.

And then finally, there's the mediator hypothesis. So we know that vision impairment is a risk factor for social isolation, physical inactivity, and depression. We also know that all three of those things, social isolation, physical inactivity, depression, are themselves risk factors for dementia. So it is possible that that relationship between vision impairment and hearing impairment and dementia may be mediated by one or more of those factors.

Matt Davis:

It makes me wonder, has anybody ever thought about or looked at individuals who lost their hearing or their vision early in life, and is that different than losing it later in life?

Josh Ehrlich:

Yeah, that's a terrific question. And the truth is that we've not had great data at scale to really answer those questions yet. Fortunately, loss of hearing and loss of vision in early life are rare conditions, and the research on that has really been limited to very small sample sizes, and the data just really hasn't existed to answer in any meaningful way to date. That said, there is evidence around adaptation and sensory substitution and things of this nature, that suggests that loss of a sense earlier in life is quite different than loss of a sense later in life because of the adaptations and because of the opportunities to develop sensory substitution that evolve in individuals who lose sensory function early in life, whereas in later life, these same opportunities may not exist. And so I think there is good reason to believe that these things likely play out differently among people who lose sensory function later in life and those who lose sensory function earlier in life. Although, I would emphasize again that it's the numbers are quite small among those who lose it early in life.

Matt Davis:

It seems like it'd be a potential mechanism to tease out the biological theories from the social isolation. I can just imagine people that learn how to communicate despite their deficit. But I guess, like you said, you just don't have the numbers in terms of the-

Josh Ehrlich:

Absolutely, though.

Matt Davis:

Yeah.

Josh Ehrlich:

No, absolutely. It's something, once again, that I've given quite a bit of thought to and hope that in the future we'll be able to answer more meaningfully.

Donovan Maust:

So vision impairment is not really something that we've talked about before on the podcast. I was wondering, just at a very high level, if you could help us sort of think through what are the different causes of vision impairment and to what extent are they preventable or treatable?

Josh Ehrlich:

Yeah, so the causes do vary from place to place in population to population, not surprisingly, but there are some universals, and I can talk a little bit also about the United States in particular. But thinking globally by far and away, the two major causes of vision impairment, they account for more than 75% of cases, are cataract and uncorrected refractive error, meaning the unmet need for eyeglasses. These are both extremely addressable with cost-effective, clinically effective interventions, namely cataract surgery and eyeglasses. Now in the United States, the number one cause of vision impairment is still cataract, but certain other diseases play a more significant role, including three age-related conditions, age-related macular degeneration, glaucoma, and diabetic retinopathy, all three of which become much, much more prevalent with each decade of life. Now, I would also note that not only do these conditions become more common with each decade of life, but so too does the prevalence of vision impairment and blindness increases appreciably in later decades of life, and is unfortunately more common amongst certain groups within the US population, namely black and Hispanic older adults.

Donovan Maust:

So considering so much of it's treatable, there must be sort of a health disparities and equity component to your work, right?

Josh Ehrlich:

Oh, very much so. Very much so. It's a big, big issue. And one of the things that actually drew me to wanting to go into ophthalmology, frankly, and to an interest in vision research, was that I always had a background in an interest in public health and population health. And I saw this incredible disparity and incredible opportunity to right the disparity that is vision impairment globally, and here in the United States. An estimated 80% of vision impairment, on a global scale, is preventable or has simply yet to be addressed. And we know that all of those 80% of cases is not equitably distributed, hardly anywhere in the world. And so yes, there is a major health equity lens here, and that does comprise a large part of our work because this really is a fixable problem, but it's a matter of getting the right resources to the right people.

Donovan Maust:

All right, let's talk about your study. Could you tell us a little bit about what data you used and in broad strokes what you did?

Josh Ehrlich:

Yeah, so in the broadest strokes, what we aim to do is to recreate the Lancet Commission Risk Factor model, adding vision impairment to the model in order to understand what the incremental benefit of including that additional risk factor of vision impairment would be. Underlying this was really a desire to say to the research community and to say to the Lancet Commission, "Hey, look, there's an important risk factor that's not being thought about adequately, that really ought to receive more attention." So in order to achieve that goal, we tried to follow as closely as possible, the methodology that the Lancet Commission used to build their own life course model and to calculate their own statistics around the potential for dementia prevention. The statistic that the Lancet Commission calculates is one that's been in the literature for many decades and is called the population attributable fraction or PAG.

The PAF, theoretically, corresponds to the percentage of cases that would be eliminated if the risk factor were itself eliminated. So in other words, if you eliminated vision impairment, what fraction of dementia cases would be eliminated? And this is the statistic the Lancet Commission uses to report the relative importance of the risk factors denominated in the report. So again, we sought to replicate that, but we sought to do it through a US lens. And so we used the Health and Retirement study, a large ongoing panel study of over 20,000 older adults that started in the early 1990s and collects detailed cognitive data. The HRS, or the Health and Retirement Study, collects only self-reported data on vision. And so we sought to right that through a method that I'll delve into more detail in on just a moment. But to calculate the PAF, we needed to get a few pieces of data.

We needed to have relative risks of dementia for each risk factor. We needed prevalence, and we needed to know how different risk factors were shared or clustered within the population because of course, an individual may have more than one risk factor. An individual could have hypertension, depression, and vision impairment, and we need to figure out how to attribute risk in that individual with multiple risk factors. So those are the three pieces of data that we needed. So the relative risks, we took largely from the Lancet Commission to make sure we were using the same relative risks that they used. In the case of vision impairment, we looked at the literature and there were three meta-analyses in the literature on the association between vision impairment and dementia. So we looked at those three meta-analyses and we actually did the analysis with each of the three meta-analyses, but I believe that it was the one with the middle effect size, not the highest, not the lowest, but right in the middle. And it was not so different from either of the others that we chose to be our primary relative risk.

We then calculated prevalence of each risk factor from the Health and Retirement study, and we calculated how the risk factors are clustered from the Health and Retirement study. We then use statistical methods to calculate the PAF, the population attributable fraction, and to adjust that PAF for the clustering of risk factors. Now, I mentioned in the case of vision impairment, that in the health and retirement study, and this is the case in most large population-based studies, unfortunately, sensory health is self-reported. How well do you see? How well do you hear?

So in order to do sensitivity analysis and understand how our calculated PAF for vision impairment might vary with other data on the prevalence of vision impairment, we looked to the literature and we drew estimates from a recently conducted meta-analysis done in the United States that estimated a prevalence of vision impairment based on objectively measure data. And so we essentially redid this analysis in many different ways, substituting different data in order to get a range of possible values, but there was not a huge amount of variance among the values, which was somewhat reassuring as well.

Donovan Maust:

And so sort of at a high level, can you summarize the findings and sort of put vision impairment into context relative to the other risk factors you looked at?

Josh Ehrlich:

So what we found was that vision impairment was strongly associated with about 2%, or a hundred thousand, prevalent cases of dementia in the United States. This ranked vision impairment, certainly, within the range of other risk factors that were included in the Lancet Commission model in 2020. It ranked it above excessive alcohol consumption. It ranked it in the same ballpark as social isolation, but less than other risk factors.

But nonetheless, accounting for about 2%, about a hundred thousand cases in the United States. I would note that we've since done similar work, that remains unpublished, in some other countries with a considerably higher prevalence of vision impairment and have seen vision impairment really rise to the top of the risk factor list when the prevalence of vision impairment is very high. Prevalence, of course, drives the population attributable fraction along with relative risk. And so in places where the prevalence is particularly high, or even in subgroups in the United States where the prevalence of vision impairment is particularly high, we may even see a higher population attributable fraction.

Donovan Maust:

Do you think the effect of vision impairment would at all be different for different underlying etiologies of dementia?

Josh Ehrlich:

Yeah, I mean, that's a really great question, and it's not really one that we can answer yet. I think that if there are, what is likely to vary is potentially relative risk, right? So the relative risk of dementia might be higher for individuals with, let's say, a certain condition, maybe macular degeneration. And if that's the case and that relative risk is higher, certainly that's going to affect the PAF, but the prevalence is going to be lower relative to vision impairment more broadly. So how that would shake out in terms of the actual PAF, I'm not quite sure. Certainly an interesting question though.

Matt Davis:

You mentioned some other analyses looking at this in other countries, I think it's easy to forget that uncorrected error, mainly not having eyeglasses, is actually incredibly common in some parts of the world. Can you just give us a sense for how common is it and where is it particularly common?

Josh Ehrlich:

So, certainly the contribution of under corrected and uncorrected refractive error varies considerably from region to region. If we look globally, the uncorrected refractive error seems to be a particularly important issue in sub-Saharan Africa, south Asia, Southeast Asia, and North Africa, perhaps not surprisingly. We see the greatest number of cases per population in these places, but when we look at the total pie of vision loss, it's also not surprising that in less affluent regions of the world, some of these other causes are also more prevalent. For example, un-operated cataract. And so the number of cases is quite high in those places. But we also see, concomitantly, that the number of cases of other causes of vision impairment, like cataract, is high in those places as well. But taken together, I typically, I will often think about cataract and uncorrected refractive error in the same breath because giving people glasses and operating on their cataracts actually, together, represent two of the most cost-effective interventions in all of medical care.

Matt Davis:

I'm just curious so far, I mean, your article came out not that long ago. Has your work gotten any traction, specifically among the folks who worked on the Lancet report?

Josh Ehrlich:

So our article, we knew we were publishing our article in a high impact journal, and we were very pleased about that, but our article immediately got more attention than we ever could have bargained for. In fact, the New York Times called shortly before the publication came out and wound up writing an article in the Science Times that featured our research and highlighted our research as potentially a new way of thinking about dementia prevention as opposed to some of the other drugs that have been, certainly, talked about recently in the press. So yeah, we were very pleased to be noticed in the New York Times, and in fact, the head of the Lancet Commission was interviewed for that article as well. So I do believe that the article got the attention of the commission, and in that article, the head of the Lancet Commission mentioned that they would be evaluating vision impairment for consideration of inclusion in the next iteration of the report.

Donovan Maust:

So what comes next in this work?

Josh Ehrlich:

A couple of things that our team is working on. One is extending this exact same type of work to other contexts. In particular, India is a place where we're very interested in replicating some of this work, for a number of reasons. One is that we have very high quality population level data in India, but the reason that that sort of data in India is important is that it's number one, the most populous country on earth, but number two has a rapidly aging population, and a medical and public health system that will be forced, with relative resource limitation, to quickly adapt to this aging population.

So it's an important place to be thinking about these issues. So that's on the one hand, some of the extension of this work that we're carrying out. We also were fortunate to receive a grant from the National Eye Institute that will allow us to leverage some population-based data in the United States that includes very rigorous testing of various visual functions, to allow us to better understand the relationship between social determinants of health, vision impairment, and a variety of adverse late-life health outcomes, among them cognitive decline and dementia. And so we're excited to really have more rigorous data that's longitudinal and that it allows us to link rigorous data, not only on vision, but also on cognitive outcomes and on social determinants of health. So keep an eye out for that as we'll be working on that for the next five years or so.

Donovan Maust:

So Josh, if you think about these preventable causes of visual impairment, if someone experiences visual impairment, say if you correct that in someone's thirties versus their forties versus their fifties, do you think the length of time with that uncorrected impairment matters in terms of developing their dementia?

Josh Ehrlich:

Yeah, that's a really good question. We don't really have the data to answer that, if length of time with vision impairment matters. My sense is that when in the life course it occurs is probably an important factor for some of the reasons we discussed before around adaptation. One study that we did carry out recently was motivated by the question, when in the course of cognitive decline does vision seem to matter? So in other words, does vision seem to matter for increasing the risk of going from normal to mild cognitive impairment? Is that what's driving the association? Does it seem to matter for transitions from cognitive impairment to dementia, or all of the above? And so we used data from a study called the Aging Demographics and Memory Study, which is a substudy of the Health and Retirement study, that actually did measure visual function objectively.

And we asked this very question, and what we found was that, as we've seen in other studies, there was a summary association, adjusted for appropriate co-variates, that showed that vision impairment increased the hazard, or the longitudinal risk, of developing incident dementia. We then broke things down and we said, "Okay, well, what is the hazard of developing mild cognitive impairment if you're visually impaired at the time that you're still normal, and what is the hazard of moving from mild cognitive impairment to dementia if you're visually impaired?"

And what we found was that the association between vision impairment and dementia seems to be driven by these early transitions. That is an individual that has impaired vision and is cognitively normal, appears to be at risk of going on to develop mild cognitive impairment. However, once you have mild cognitive impairment, vision doesn't seem to matter as much anymore. The risk of going on to dementia is really high at that point, and in some ways it seems that the cat is already out of the bag, but for those early transitions, vision seems to play a really important role, at least for some people.

Matt Davis:

Yeah, it's so interesting to think about the extent to which the pathways and mechanism of visual impairment and hearing impairment are similar, and if you could do that similar analysis, looking at hearing impairment, there was just, I think, in the past couple of weeks, a study in Lancet where they looked at correcting hearing impairment and incident dementia. It's just thinking about all the ways you can go back and forth between those two types of sensory impairment to think about mechanisms and interventions, and the ideal time point of intervention is pretty fascinating and lots of exciting science ahead of you, I think.

Josh Ehrlich:

Yeah, no, I completely agree. That paper actually came out just yesterday or Tuesday, excuse me. It coincided with unveiling at AAIC in Amsterdam, and certainly it's exciting. It's the first interventional evidence that sensory correction may in fact slow cognitive decline. There's non-randomized control evidence from the vision literature around cataract surgery slowing cognitive decline and preventing dementia. A very well-designed study, but certainly not the level one randomized control trial type evidence that we see from the hearing worlds just this week.

Matt Davis:

I think you're going to find yourself doing more and more sophisticated methods as you go too, in terms of trying to tease out cause and effect from all these various factors.

Josh Ehrlich:

Yes, definitely. Definitely.

Donovan Maust:

And trials too.

Josh Ehrlich:

Yeah, yeah. No, it's funny you say that. I think a week from today, today's Thursday, right? Yeah. So I'm actually very close with a group at Hopkins that did that trial. I work pretty closely with many of them, but the lead author, Frank Wynn, myself and a colleague of mine at University of Washington, have a call next Thursday to talk about the potential of doing something similar in the vision world.

Donovan Maust:

Awesome. We'll have you back in a couple seasons and you can tell us all about it.

Matt Davis:

This is all really interesting stuff, and we're looking forward to seeing where your work goes from here. Josh, thanks so much for joining us, and thanks to all of you who listened in.

Josh Ehrlich:

No, thank you very much. I appreciate you having me.

Matt Davis:

If you enjoyed our discussion today, please consider subscribing to our podcast. Other episodes can be found on Apple Podcasts, Spotify, and SoundCloud, as well as directly from us at capra.med.umesh.edu, where a full transcript of this episode is also available. On our website, you'll also find links to our seminar series and data products we've created for dementia research. Music and engineering for this podcast was provided by Dan Langa. More information available at www.danlanga.com. Minding Memory is part of the Michigan Medicine podcast network. Find more shows at UofMhealth.org/podcast.

Support for this podcast comes from the National Institute on Aging at the National Institutes of Health, as well as the Institute for Healthcare Policy and Innovation at the University of Michigan. The views expressed in this podcast do not necessarily represent the views of the NIH or the University of Michigan. Thanks for joining us, and we'll be back soon.