記住我
Projections indicate that the current number of people living with dementia will triplicate by 2050 [38]. This increase will be mainly due to the rising life expectancy of low- and middle-income countries, however the age-standardized prevalence of dementia is predicted to remain stable in both sexes [38]. Epidemiological studies have estimated a population attributable fraction (PAF) for dementia of 30–50%, suggesting that up to half dementia cases could be prevented if those risk factors were eliminated from the population [7,69,86]. In cancer research, the term “exposome” was coined to describe the cumulative lifelong experiences and exposures that can impact disease risk [115]. An analogous concept has been proposed for dementia, comprised of exogenous (e.g., head trauma, infections) and endogenous (e.g., hypertension) exposures [36]. Here we will critically review new developments and controversies regarding some potentially modifiable risk factors of the dementia exposome, including exogenous such as air pollution, microbial agents, and traumatic brain injury, as well as endogenous such as cardiovascular risk factors and hearing loss. We will use the broader term Alzheimer's disease and related dementias (ADRD) to account for the frequent co-occurrence of multiple brain pathologies contributing to cognitive decline and for the fact that most studies lack biomarker and autopsy data to ascertain the neuropathological substrate(s) of dementia. For each risk factor, we will examine the epidemiological studies supporting the association between the exposure and ADRD risk, the experimental evidence from mouse models supporting a causal pathophysiological link and, whenever available, the results of clinical trials targeting those risk factors.
Box 1: no caption available
CARDIOVASCULAR RISK FACTORS Epidemiological evidenceThe importance of mid-life cardiovascular factors in the risk of developing dementia later in life is underscored by several epidemiological observations. First, cardiovascular risk factors (e.g., hypertension, obesity, and sedentarism) rank at the top of all modifiable risk factors by PAF across all ethno-racial groups [7,14,54,61,69,76,86,99,114]. Second, population-based clinic-pathological studies have revealed that mixed AD and cerebrovascular disease is the most common pathological substrate underlying dementia in community-dwelling individuals [15]. Third, age-adjusted measures of ADRD incidence and prevalence are decreasing or stabilizing in Western countries [69,102], possibly thanks to the expansion of cardiovascular risk screening, prevention, and treatment (e.g., statins, antihypertensive, antidiabetic, and antiplatelet drugs), together with the stricter recommendations to consider hypertension, diabetes mellitus, and hypercholesterolemia adequately controlled. Lastly, and supporting this idea, neuropathological studies have confirmed that the frequency of severe cerebrovascular disease at autopsy has dramatically decreased over the last decades [43▪▪].
Evidence from preclinical studiesA plethora of preclinical studies have shown that, besides their pro-atherosclerosis effects, hypertension and high fat diet can promote the accumulation of Aβ plaques and tau neurofibrillary tangles and worsen cognitive deficits in AD transgenic mouse models, whereas antihypertensive drugs, statins, and exercise improve these AD phenotypes (reviewed in [102]). The importance of exercise in preventing ADRD has been strengthened by new evidence implicating brain derived neurotrophic factor (BDNF) [24] and irisin [49,73] in the exercise-induced amelioration of the cognitive deficits observed in AD mice. Both BDNF and irisin promote hippocampal synaptic plasticity and neurogenesis, and irisin additionally reduces Aβ levels [49,73] through enhancing the secretion of neprilysin – one of the main Aβ-degrading enzymes – by astrocytes [57▪].
Evidence from clinical trialsThis strong epidemiological and preclinical evidence supporting a synergistic effect of cardiovascular risk factors to promote ADRD has led to the design of clinical trials to test the efficacy of multidomain lifestyle interventions (i.e., targeting exercise, diet, cognitive stimulation, and vascular risk control) and cardiovascular drugs at preventing cognitive decline in elderly people at risk for dementia (Table 1). Although the Finnish FINGER trial revealed the benefits of such multidomain lifestyle interventions on cognition [83], the French MAPT trial failed to do so [4]. Moreover, a clinical trial testing the MIND diet has recently failed to slow down cognitive decline, brain atrophy, and white matter hyperintensities in participants without cognitive impairment but at risk of dementia [9▪▪]. Clinical trials with similar design to the FINGER trial are underway worldwide to shed light on these conflicting outcomes [60], including the US POINTER (NCT03688126). In the SPRINT-MIND trial, intensive blood pressure control with antihypertensive drugs (goal systolic < 120 mmHg) significantly reduced the risk of MCI and MCI/probable dementia combined diagnoses over the 5-year follow-up compared to standard control (goal systolic < 140 mmHg) in nondemented individuals who had hypertension and increased cardiovascular risk, but no diabetes mellitus or stroke history [107]. Secondary analyses have shown that intensive blood pressure control increases (rather than reduces) cerebral perfusion [31] and slows down white matter damage [93▪▪,106], however slightly accelerates total brain and AD-like hippocampal volume loss [81,106]. Data on plasma AD biomarkers would be very informative to determine whether this strategy has any impact on the AD pathophysiological process, but are not currently available. Conversely, in the TOMMORROW trial, low dose of the antidiabetic peroxisome proliferator receptor gamma (PPARγ) agonist pioglitazone failed to delay the onset of MCI due to AD relative to placebo in cognitively intact individuals who were deemed to be at high risk of developing AD based on their age as well as APOE and TOMM40 genotypes [19].
Table 1 - Clinical trials testing cardiovascular preventative interventions with cognition as primary outcome Reference Trial name Intervention Trial design Participants Primary endpoints Secondary endpoints Results Ngandu T et al. 2015 [83] Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER)Multidomain : nutritional advice (diet rich in fruits and vegetables, wholegrain cereals, low-fat milk and meat, fish ≥ 2x/wk, limit sucrose <50 g/d, avoid butter), physical exercise program (muscle strength 1–3x/wk, cardio 2–5x/wk, balance), cognitive training (group educational/cognitive skills sessions and individual web-based computer sessions), social activities (group sessions).
Control : regular health advice.
2-year, multicenter, randomized, double-blind, controlled Age 60–77 y, community-dwelling, non-demented but at risk for dementia based on CAIDE score ≥6 and cognitive screening Change in global cognition from baseline over 2 years Change in cognitive domain-specific z-scores Significantly slower cognitive decline in intervention vs. placebo groups, particularly in executive function and processing speed Andrieu S et al. 2017 [4] Multidomain Alzheimer Preventive Trial (MAPT) (NCT00672685) Multidomain + ω3 PUFA (374) vs. Multidomain + placebo (390) vs. ω3 PUFA (381) vs. placebo (380)Multidomain : cognitive training (group reasoning and memory skills sessions), physical exercise (advice on walking ≥30 min 5x/wk and tailored home-based program), nutritional advice (based on France national guidelines).
ω3 PUFA : 2 caps/d, each containing 400 mg DHA and 112.5 mg EPA.
3-year, multicenter, randomized, placebo-controlled superiority (double-blind regarding ω3 PUFA only) Age ≥70 y, community-dwelling, non-demented but at risk of dementia based on spontaneous memory complaint to PCP, limitation in one ADL, or slow gait Change in cognition from baseline over 3 years • Change in cognitive domain-specific z-scoresMIND diet : increase MIND foods (e.g., skinless, not fried chicken/turkey, olive oil, green leafy and other vegetables, fish, whole-grain cereals, bread and pasta, beans/legumes, berries, nuts).
Control diet : focus on portion control, calorie intake, behavioral strategies to lose weight, without changing diet structure.
3-year, 2-center, randomized, controlled Age ≥65 y, with overweight (BMI ≥25) and suboptimal diet (MIND-diet score ≤8), non-demented (MoCA ≥22) but with family history of dementia in 1st degree relative Change in global and cognitive-domain specific cognitive scores from baseline over 3 years MRI-based brain volumes and WMH No difference of MIND diet vs. control diet Williamson JD et al. 2019 [107]; Nasrallah IM et al. 2019 [106] and 2021 [81]; Dolui S et al. 2022 [31] Systolic Blood Pressure Intervention Trial (SPRINT-MIND)The list of clinical trials is not exhaustive. Numbers in parenthesis in the Intervention column indicate the number of participants in each trial arm based on the modified intention-to-treat analysis.ADL, activities of daily living; BMI, body mass index; CAIDE, cardiovascular risk factors, aging and dementia; caps, capsules; CDR-SoB, clinical dementia rating sum of boxes; d, day; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; MIND, Mediterranean-DASH (Dietary Approach to Stop Hypertension) Intervention for Neurodegenerative Delay; MRI, magnetic resonance imaging; PUFA, poly-unsaturated fatty acids; SBP, systolic blood pressure; wk, week; WMH, white matter hyperintensities.
Both oral and intestinal bacterial dysbiosis – a dysregulation of the commensal bacterial flora – have emerged as potential risk factors for the development of dementia. Oral bacterial dysbiosis, such as that occurring in bacterial periodontitis, has been associated with AD through inflammatory mediators [91▪], but whether this association is due to a causal link between the oral microbiome and the AD pathophysiological process or just reflecting reverse causality (i.e., poor oral health as a result of cognitive decline) remains controversial. Biomarkers offer a unique opportunity to resolve the directionality of this association; for example, a cross-sectional study found a higher oral dysbiosis index (measured as a healthy/unhealthy bacteria genome ratio via DNA sequencing) in cognitively unimpaired old individuals positive for Aβ (i.e., with low Aβ CSF levels), suggesting that oral microbial dysbiosis may precede cognitive decline and contribute to AD progression [55]. Similarly, AD has been associated with reduced diversity and altered composition of the fecal microbiome [51▪]. Interestingly, these changes precede cognitive decline [35▪▪], cannot be explained by the changes in diet, caloric intake, and/or nutrition status observed in AD [35▪▪,116], and correlate with CSF AD biomarker levels in both cognitively unimpaired individuals [35▪▪] and patients with AD dementia [116], suggesting a pathophysiological link between gut microbiome dysbiosis and AD. Longitudinal prospective studies with serial AD biomarkers in cognitively unimpaired individuals are needed to confirm this association and unequivocally rule out reverse causality.
Evidence from preclinical studies in mouse modelsPreclinical studies support the idea that gut microbiota may impact Aβ and pTau accumulation. For example, a decrease in Aβ plaque accumulation has been described in AD transgenic mice raised in germ-free vs. conventional conditions [25] or treated with an antibiotic cocktail to deplete the gut microbiome [30]. Similarly, tauopathy mice bred in germ-free conditions or treated with broad spectrum antibiotics exhibit a reduction in pTau levels and pTau-mediated neurodegeneration compared to tauopathy mice raised in conventional conditions or treated with vehicle. Of note, these effects were modulated by sex [30,100▪▪] and in the case of pTau also by the APOE genotype [100▪▪]. Mechanistically, these studies have implicated gut microbiome-induced changes in the peripheral immune system and/or microglial function [25,30,100▪▪,123▪], possibly mediated by secreted short-chain fatty acids (SCFAs) – a major by-product of fermentation [25,100▪▪,123▪]. However, further studies are needed to dissect the mechanisms by which the gut microbiota and their metabolites may interact with the peripheral immune system and/or microglia, and impact ADRD pathophysiology.
Evidence from clinical trialsSeveral randomized, double-blind, placebo-controlled clinical trials have evaluated the efficacy of probiotics in patients with MCI with mixed results [5,10,122]. In addition, the safety and feasibility of oral fecal microbiota transplant is being evaluated [23].
VIRUS Epidemiological evidenceIn a revival of the viral hypothesis of AD [101], the possible implication of certain viral infections in ADRD risk is receiving increasing attention, particularly the reactivation of latent neurotropic viruses of the Herpesviridae family, including herpes simplex virus 1 and 2 (HSV-1/2), varicella-zoster virus (VZV), and Epstein-Barr virus (EBV). Indeed, numerous epidemiological studies in the last few years have tried to address this question but yielded conflicting results (Table 2) [8,22,45,53,66,71,79,97,98,105,111]. Reasons for these mixed findings are likely methodological, including differences in study design (population-based longitudinal cohort vs. electronic health records or claims data), ascertainment of viral exposure (positive IgM or IgG serology vs. ICD codes and/or medical records of antiviral treatment) and of dementia and/or AD diagnosis (ICD codes vs. expert diagnosis), and length of follow-up (a shorter follow-up is prone to reporting bias, thus overestimating the link between viral infection and dementia). Similarly, neuropathological studies examining the frequency of herpesvirus genome detection in postmortem AD vs. control brains have rendered mixed results [3,11,94,121]. The 2019 SARS-CoV2 pandemic has been associated with an increased risk of cognitive decline [68] and ADRD [110▪▪,117], however it is still unclear whether these findings are due to neuroinvasive disease leading to neuropathological changes, reporting bias, or unmasking of a preexisting ADRD caused by the systemic inflammatory milieu; ongoing longitudinal cohort studies will eventually elucidate the long-term impact of SARS-CoV-2 infection on ADRD risk. Studies incorporating imaging and/or fluid biomarkers and APOE genotype (a potential major confounder) are much needed but scarce [66].
Table 2 - Recent epidemiological studies on the association between viral infections and ADRD risk Reference Risk factor/exposure Comparator group Study design Location Outcome Follow-up length (y) HR OR β 95% CI Herpes Simplex Virus (HSV) Linard M et al. 2021 [66] Positive serum HSV IgG Negative serum HSV IgG Population-based longitudinal cohort Bordeaux, Dijon, Montpellier (Southwest France) Incident AD (NINCDS-ADRDA) 6.8 ± 2.6 1.19 N.A. N.A. 0.81, 1.77 Murphy MJ et al. 2021 [79] Positive serum HSV1 IgG Negative serum HSV1 IgG Population-based longitudinal cohort Rotterdam (The Netherlands) Incident dementia (DSM-III-R) 9.1 ± 3.4 1.18 N.A. N.A. 0.83, 1.68 Incident AD (NINCDS-ADRDA) 1.13 N.A. N.A. 0.77, 1.66 Global cognition (MMSE) N.A. N.A. -0.12 -0.24, 0.002 Serum HSV1 IgG antibody titer N.A. Global cognition (MMSE) N.A. N.A. −0.06 −0.11, −0.01 Shim Y et al. 2022 [105] Diagnosis of symptomatic HSV infection (ICD) Controls with no HSV (or VZV) diagnosis National insurance claim data, matched-cohort South Korea Incident dementia (ICD) Up to 10 1.18 N.A. N.A. 1.16, 1.20 Incident AD (ICD) 1.121 N.A. N.A. 1.183, 1.239 Varicella-Zoster Virus (VZV) Chen VC-H et al. 2018 [22] Herpes zoster diagnosis (ICD) Controls with no VZV diagnosis National insurance claim data, matched-cohort Taiwan Incident dementia (ICD) Up to 17 1.11 N.A. N.A. 1.04, 1.17 Johannesdottir Schmidt SA et al. 2022 [53] Incident herpes zoster (ICD) or antiviral treatment No history of herpes zoster or antiviral treatment National EHR data, matched-cohort Denmark Incident dementia (ICD) or antidementia drug 6 (3–11), range 1–21 0.93 N.A. N.A. 0.90, 0.95 Incident AD (ICD) or antidementia drug 0.93 N.A. N.A. 0.90, 0.97 Herpes zoster with cranial nerve involvement (ICD) No history of herpes zoster or antiviral treatment Incident dementia (ICD) or antidementia drug 1.07 N.A. N.A. 0.79, 1.45 Herpes zoster with CNS involvement (ICD) No history of herpes zoster Incident dementia (ICD) or antidementia drug 1.94 N.A. N.A. 0.78, 4.80 Shim Y et al. 2022 [105] Diagnosis of symptomatic VZV infection (ICD) Controls with no VZV (or HSV) diagnosis National insurance claim data, matched-cohort South Korea Incident dementia (ICD) Up to 10 1.09 N.A. N.A. 1.07, 1.11 Incident AD (ICD) 1.106 N.A. N.A. 1.081, 1.131 Epstein-Barr virus (EBV) Torniainen-Holm M et al. 2018 [111] Positive serum EBV IgG Negative serum EBV IgG National health survey Finland Incident dementia (ICD) Up to 13 1.74 N.A. N.A. 0.51, 5.92 Cytomegalovirus (CMV) Barnes LL et al. 2015 [8] Positive serum CMV IgG Negative serum CMV IgG Longitudinal cohort (ROS, MAP, and MARS) Chicago area (USA) Incident AD (NINCDS-ADRDA) 5.0 2.41 N.A. N.A. 1.53–3.78 Torniainen-Holm M et al. 2018 [111] Positive serum CMV IgG Negative serum CMV IgG National health survey Finland Incident dementia (ICD) Up to 13 0.85 N.A. N.A. 0.57, 1.27 Antiviral treatment Chen VC-H et al. 2018 [22] Antiviral drug after VZV diagnosis Controls with no VZV diagnosis National insurance claim data, matched-cohort Taiwan Incident dementia (ICD) Up to 17 0.55 N.A. N.A. 0.40–0.77 Hemmingsson E-S et al. 2021 [45] Positive serum HSV1 IgG with antiviral drug treatment Positive serum HSV1 IgG without antiviral drug treatment Population-based, nested case-control Umea, Sweden (Betula cohort study) Incident AD (DSM-IV) Up to 29 N.A. 0.287 N.A. 0.102, 0.809 Lopatko Lindman K et al. 2021 [71] Antiviral treatment, irrespective of herpes diagnosis (ICD) No history of antiviral treatment or herpes diagnosis (ICD) National EHR and drug prescription data, matched-cohort Umeâ, Sweden Incident dementia (ICD) Up to 12 0.89 N.A. N.A. 0.86, 0.92 Herpes diagnosis (ICD) with antiviral treatment No history of antiviral treatment or herpes diagnosis (ICD) 0.90 N.A. N.A. 0.82, 0.98 Herpes diagnosis (ICD) without antiviral treatment No history of antiviral treatment or herpes diagnosis (ICD) 1.50 N.A. N.A. 1.29, 1.74 Herpes diagnosis (ICD) with antiviral treatment Herpes diagnosis (ICD) without antiviral treatment 0.75 N.A. N.A. 0.68–0.83 Schnier C et al. 2021 [98] History of oral antiherpetic medication No history of oral antiherpetic medication National EHR data Denmark Incident dementia 7.4 (3.8–12.2) × 10 000 person-year 0.91 N.A. N.A. 0.89, 0.93 History of oral antiherpetic medication No history of oral antiherpetic medication National EHR data Scotland Incident dementia 2.7 (1.4–4.2) × 10 000 person-year 0.98 N.A. N.A. 0.64, 1.49 History of herpes treated with oral antiherpetic drugs No history of herpes or oral antiherpetic medication National EHR data Wales Incident dementia 6.7 (3.3–11.3) x 10 000 person-year 0.91 N.A. N.A. 0.86, 0.97 History of herpes treated with oral antiherpetic drugs No history of herpes or oral antiherpetic medication National EHR data Germany Incident dementia 8.8 (4.5–14.5) × 10 000 person-year 1.08 N.A. N.A. 0.98, 1.20 Schnier C et al. 2022 [97] VZV vaccination Non-vaccinated without shingles National EHR data Wales Incident dementia (ICD) Up to 6 0.72 N.A. N.A. 0.69, 0.75 Incident AD (ICD) 0.81 N.A. N.A. 0.77, 0.86 Incident vascular dementia (ICD) 0.66 N.A. N.A. 0.61. 0.71 Severe acute respiratory syndrome coronavirus 2 (SARS-CoV2) Wang L et al. 2022 [117] COVID-19 infection (ICD) Non-infection National EHR data, matched-cohort USA Incident AD (ICD) 1 1.69 N.A. N.A. 1.53, 1.72 Taquet M et al. 2022 [110▪▪] COVID-19 infection (ICD) Other respiratory tract infections (ICD) International EHR data, matched- cohort International Incident dementia (ICD) 0.5 1.33 N.A. N.A. 1.26, 1.41 Liu Y-H et al. 2022 [68] Severe COVID-19 infection (WHO) Non-infection Longitudinal cohort Wuhan, China Early-onset cognitive decline (TICS40, IQCODE) 1 N.A. &#
留言 (0)