Young onset dementia: risk factors

 A prospective cohort study assessed the risk factors for young onset dementia in the UK (see dementia young onset risk factors JAMAneuro2024 in dropbox, or doi:10.1001/jamaneurol.2023.4929)

 

Details:

-- 356,052 participants for England, Scotland, and Wales in the UK Biobank were assessed from 2006-2019

-- all were younger than 65 and without a dementia diagnosis at baseline; those over the 65 or dementia baseline were excluded

-- 39 potential risk factors were identified from systematic reviews of late-onset dementia (LOD) and young-onset dementia (YOD), in the following domains:

    -- sociodemographic: education, social economic status, sex

    -- genetic factors: apolipoprotein E (APOE e4 alleles)

    -- lifestyle factors: physical activity, alcohol use, alcohol use disorder, smoking, diet, cognitive activity, social isolation, marriage

    -- environmental factors: nitrogen oxide, particulate matter, pesticides, diesel

    -- blood marker factors: vitamin D, C-reactive protein, estimated GFR (not stated how this was measured), albumin

    -- cardiometabolic factors: stroke, hypertension, diabetes, hypoglycemia, heart disease, atrial fibrillation, aspirin use

    -- psychiatric factors: depression, anxiety, benzodiazepine use, delirium, and sleep problems

    -- other factors: traumatic brain injury, rheumatoid arthritis, thyroid dysfunction, hearing impairment, and hand grip strength

-- mean age 54.6, 55% women, APOE e4 status (no e4 alleles in 65%/1 in 25%/2 in 4%), education 55% academic/professional, 45% less education, SES by Townsend deprivation index 15% least deprived/53% middle/30% most deprived, regular physical activity 65%, moderate alcohol use 45%/heavy alcohol use 40%/alc use disorder 3%, former smoker 30%/current smoker 15%, healthy diet 45%, social isolation 28%

-- vitamin D level 10-20ng/mL 35%, <10ng/mL 20%, CRP >1 mg/dL 12%

-- medical issues: stroke 4%, hypertension 32%, diabetes 8%, atrial fibrillation 1%, aspirin 15%, overweight 33%, obesity 22%, orthostatic hypotension 0.3%

-- depression 4%, anxiety 7%, benzo use 0.3%, sleep disorder sometimes 42%/usually 30%, hearing impairment 27%

 

Main outcomes: the association between risk factors and incidence of YOD, assessing factors within domains as above, and then across domains, using multivariate Cox proportional hazards regression

 

Results:

-- total of 485 incident cases of YOD were identified: 52% male, incidence rate of 16.8 per 100,000 person-years (15.4-18.3)

 

-- domain analysis: 15 factors were significantly identified with higher YOD risk, in the fully adjusted model (adjusted for sex, APOE e4 status, sociodemographic and lifestyle domains, environmental factors, blood markers, cardiovascular disease, psychiatric factors, and other variables::

    -- higher formal education: 37% decreased risk, HR 0.63 (0.48-0.81)

    -- higher deprivation score: 82% increased risk, HR 1.82 (1.19-2.77)

    -- having two APOE e4 alleles: 87% increased risk, HR 1.87 (1.01-3.44)

    -- alcohol use:

        -- moderate alcohol use: 28% decreased risk, HR 0.72 (0.52-1.00)

        -- heavy alcohol use: 36% decreased risk, HR 0.64 (0.46-0.90)

    -- alcohol use disorder: 2.39-fold risk, HR 2.39 (1.37-4.15)

    -- social isolation: 53% increase risk, HR 1.53 (1.18-1.98)

    -- vitamin D <10 ng/mL: 59% increased risk, HR 1.59 (1.15-2.18)

    -- CRP level >1 mg/dL: 54% increase risk, HR 1.54 (1.03-2.30)

    -- high hand grip strength: 42% decreased risk, HR 0.58 (0.52-0.88)

    -- hearing impairment: 56% increased risk, HR 1.56 (1.22-2.01)

    -- orthostatic hypotension: 4.20-fold risk, HR 4.20 (1.30-13.62)

    -- stroke: 2.07-fold risk, HR 2.07 (1.26-3.41)

    -- diabetes: 65% increased risk, HR 1.65 (1.15-2.36)

    -- heart disease: 61% increased risk, HR 1.61-2.36)

    -- depression: 3.25-fold risk, HR 3.25 (2.08-5.09)

 

-- interactions of individual components with sex and APOE e4 status:

    -- men with diabetes had higher risk of YOD than men without diabetes (but not women)

    -- women with high CRP levels had higher risk of YOD than those with CRP <1mg/dL (but not men)

    -- in the fully adjusted model, there were no interactions between apo E status and the final individual risk factors

 

-- sensitivity analyses:

    -- to account for delays in diagnosis: no difference in results when assessing those 65-70yo vs those <65yo

        -- the associations were somewhat stronger for APOE e4 status, stroke, diabetes, and heart disease; but they were slightly attenuated for SES, education, alcohol use disorder, hearing impairment, orthostatic hypotension, vitamin D, and depression

    -- to minimize the likelihood of reverse causation (eg, did dementia cause some of the risk factors): they restricted analysis to those diagnosed at least 3 years before the baseline assessment

        --  APOE e4 status, alcohol use disorder, CRP levels and orthostatic hypotension ceased to be significantly associated with YOD incidence

    -- when they compared the risk factors for  YOD vs LOD: 

        -- hearing impairment, orthostatic hypotension, and CRP levels were not associated with LOD; there was a significant association with age for APOE e4 status, education and hearing impairment (higher association for education and APOE e4 status with LOD vs YOD; hearing impairment was less significantly associated with LOD.

   

Commentary:

-- importance of chronic inflammation as being associated with many of these potential risk factors:

    -- many of the standard and non-standard risk factors for ASCVD are associated with an increased inflammatory state (as in  diabetes, hypertension, CKD, obesity; but also air pollution, microplastics, stress and depression, etc): https://gmodestmedblogs.blogspot.com/2023/10/update-ascvd-risk-factor-critique.html, and for the potential role of  microplastics: https://gmodestmedblogs.blogspot.com/2024/03/plastics-too-many-and-too-bad-for-our.html

    -- treatment with some potent anti-inflammatories seems to decrease the  ASCVD risk, such as the anti-inflammatory canakinumab in the CANTOS study (cad CANTOS antiinflam helps nejm2017 in dropbox, or 10.1056/NEJMoa1707914) and colchicine (for the latter, see https://gmodestmedblogs.blogspot.com/2019/12/colchicine-benefits-post-mi.html

    -- neuroinflammation has been documented in many studies as being associated with dementia (see https://www.mdpi.com/1422-0067/23/2/616 ).  there has been an argument that atorvastatin (which is lipophilic and crosses the blood-brain barrier) might be more effective in decreasing dementia risk than rosuvastatin (which is hydrophilic), though observational studies to date have not found that to be the case (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7168191/ )

-- perhaps not surprisingly, there are some differences in risk factors for dementia across studies. as with the current study, older people with dementia (mean age 79)  in the ARIC study of 4 communities in the US also found a strong association between hearing impairment and global cognition: https://gmodestmedblogs.blogspot.com/2023/11/dementia-hearing-loss-large-contributor.html

-- overall, this current study found that  there were 15 risk factors associated with YOD, noting that higher formal education, lower physical frailty (as measured by handgrip strength), and alcohol use were associated with lower YOD risk, whereas having 2 APOE e4 alleles, lower SES, higher CRP, orthostatic hypotension, stroke, diabetes, heart disease, hearing impairment, marked vitamin D deficiency, alcohol use disorder and social isolation were associated with increased risk. A few comments from their discussion, with lots of my added ones:

    --higher formal education: hard to know what this means. those with higher formal education may be more inclined to have healthier lifestyles (differences in diet, exercise, smoking, alcohol, lower likelihood of occupational or environmental exposures such as pollution.....)

    --lower physical frailty: hard to know what this really means. in general, frailer people tend to not do so well overall. frailty is associated with inflammation, which is related to decreased physical function (Inflammatory markers and physical frailty: towards clinical application | Immunity & Ageing | Full Text (biomedcentral.com) increasingly so with aging, and likely with YOD. but there are so many potential confounders here: is diabetes-related frailty with diabetes-related inflammation the cause, etc etc etc (for many of the identified risk factors)

    --alcohol use: it is important to stress that the assessment of moderate alcohol use in particular is fraught: it turn out that the demographics/comorbidities for those who drink moderate amounts of alcohol are actually much better than for non-drinkers, largely attributable to the fact that non-drinkers in many studies included those who were prior drinkers who stopped because of alcohol-related illnesses or other illnesses that might be worse with alcohol consumption (hypertension, heart disease, disability from stroke, etc), or the need to take meds that have important interactions with alcohol (ie, they were sicker, became or were nondrinkers, but had worse outcomes despite being classified as current "nondrinkers". For a review of this argument and one stressing the importance of zero alcohol intake, see https://gmodestmedblogs.blogspot.com/2023/11/alcohol-use-disorder-meds.html

    -- having 2 APOE e4 alleles: though there are other genetic risk factors for YOD, APOE confers the largest genetic risk for sporadic YOD cases. However, their findings in this study found that even with 2 alleles for APOE e4, there was not really any increased risk when controlling for several of the other risk factors, as noted above

    -- lower SES: this was assessed by the Townsend deprivation index (based on where people lived, not on the individuals who lived there, though this index correlates well with individual assessment of social deprivation). This index is certainly better than assessing just income, though assessment of wealth inequity may be a more sensitive assessment of the adverse effects of socioeconomic differences and higher mortality risk (https://gmodestmedblogs.blogspot.com/2024/03/mortality-benefits-from-decreasing.html )

    -- orthostatic hypotension: the ARIC study also found that orthostatic hypotension increases the risk of dementia (https://www.ahajournals.org/doi/full/10.1161/HYPERTENSIONAHA.123.21438 ). perhaps there is cerebral hypoperfusion sufficient to cause neural injury from the lower blood pressure. or perhaps there is reverse causation: dementia leading to orthostatic hypotension. but there may well be a significant, if not causal, relationship. of interest, orthostatic hypotension has also been associated with early Parkinson's dementia or Lewy body dementia

    – vitamin D deficiency (if <10ng/mL): a 2014 study found that vitamin D deficiency was associated with cognitive impairment (https://gmodestmedblogs.blogspot.com/2014/08/vitamin-d-deficiency-and-dementia.html ), and another article found that vitamin D supplementation in those with mild cognitive impairment improved cognition (https://gmodestmedblogs.blogspot.com/2018/11/vitamin-d-decreases-cognitive-impairment.html). this is not so surprising since there are vitamin D receptors found in many areas of the brain (Distribution of the Vitamin D receptor and 1α-hydroxylase in human brain - ScienceDirect ). This study found that CRP levels (reflecting systemic inflammation) were associated with YOD, but only when the very low vitamin D was included in their model

    – social isolation: though social isolation might be associated with depression, in this study they found that each independently were associated with YOD. Another study found that social isolation was associated with inflammation (https://www.sciencedirect.com/science/article/pii/S0889159123003562?via%3Dihub ), which could be a factor in an association with dementia. but this is all pretty complex, since there are so many reasons that there may be social isolation (a panoply of psych, social, and medical issues) that it is hard to isolate social isolation (!!) as the actual factor

    – diabetes, stroke, heart disease, and depression (all associated with systemic inflammation) are associated with YOD in other studies as well. of note, this study found that men with diabetes were at higher risk than women, and this may be attributable to the finding in other studies that men tend to have more microvascular complications from diabetes (see Clin Sci (Lond) (2017) 131 (9): 833–846) and perhaps more small-vessel disease leading to dementia (there is some evidence that inflammation itself is a risk factor for microvascular disease in diabetes, since inflammation is associated with endothelial dysfunction and vascular injury: Diabetic vasculopathy: macro and microvascular injury - PMC (nih.gov) 

-- it is surprising that diet and exercise did not bubble up to being significantly protective for dementia:

    -- they do have quite robust but self-reported assessments of these variables:

        -- healthy diet included at least 4 of: at least 3 servings of fruits; at least 3 servings of  veges; fish at least 2x/week; processed meats <2/week; unprocessed red meat <1.5 per week; whole grains at least 3/wk, refined grains <1.5/wk

        -- physical activity: moderate activity >5 days/week with duration >150 min/week; at least 1 day/week of vigorous exercise with duration of >75 min

    -- but there are so many studies finding that diet and exercise are beneficial in terms of dementia:

        -- eg another UK Biobank study found that both genetic risk and lifestyle were both associated with developing dementia: https://gmodestmedblogs.blogspot.com/2019/07/dementia-genetics-and-lifestyle-both.html 

        -- several studies on diet (esp Mediterranean diet) have found an association for having a lower dementia risk: eg, https://pubmed.ncbi.nlm.nih.gov/36915130/

        -- and exercise seems beneficial in both preventing and modifying the course of dementia: eg, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3258000/

        -- that all being said, it could be that diet and exercise do significantly decrease the risk of dementia, but that is mediated through decreasing the risk of medical conditions that increase the risk of dementia (eg, obesity, diabetes, heart disease, stroke, depression, and pretty much everything associated with a chronic inflammatory state) 

 

Limitations:

– this study was an observational one and therefore does not allow for conclusions about causality, just associations

    -- as  noted above, there are likely many confounders with many of their findings.  those who were frail, or socially isolated, or having depression, etc likely had other behaviors or exposures or diseases that put them at higher risk of YOD (eg other substance disorders, family history of dementia, meds that might affect cognition such as most antidepressants or neuroleptics). and perhaps it was those other things that really were related to YOD

-- much of the information about the patients was self-reported, which can be associated with inaccuracies

-- some of the associations had wide confidence intervals because of limited numbers of people (eg orthostatic hypotension had only 5 people having YOD, APOE e4 had only 30 people), which limits our confidence in their conclusions

-- there may have been some cases of YOD not included in their sample (missing some hospitalized or people who died), and the lack of these folks might alter the ultimate statistical analysis

so, this study reinforces a few things:

-- young-onset dementia (<age 65) does have some genetic component, mostly from APOE e4 alleles (though this was not found in the above study for some reason. ??too few people with 2 APOE e4 alleles), but the predominant risk factors were in fact potentially modifiable, largely through lifestyle issues (esp diet, exercise, normal BMI)

-- several of the risk factors (eg pollution, microplastics, wealth inequality, lower levels of education...), however, are big public health and governmental ones, requiring all of us to push these agencies to deal with these important issues

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

if you would the article, please email me.

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list