Clinical manifestations

Presentation and course

Subjective cognitive decline. Patients with pre-mild cognitive impairment and subjective cognitive decline usually complain of memory loss, but some might notice problems with other cognitive domains such as organizational skills, language, or visuospatial function (44; 110; 100). The term “subjective” implies that validation of the subjective experience is not needed on a cognitive test. In the context of establishing the diagnosis of pre-mild cognitive impairment, it is important that cognitive testing is performed to exclude cases of true mild cognitive impairment (44; 110). Patients with subjective cognitive decline often exhibit symptoms of mild behavioral impairment (with affective or psychotic symptoms) in addition to their cognitive complaints (89; 100). In one study, the strongest predictor for subjective memory complaints was depressed mood, whereas reduced hippocampal thickness was associated with objective memory impairment but not subjective complaints (53).

Age of onset. Most studies of subjective cognitive decline unrelated to the autosomal dominant type of Alzheimer disease restrict the age of symptom onset to 60 years or older (44). Including younger ages of onset increases the risk of enrolling too many cases that will go on to develop non-Alzheimer types of dementia.

Sex preponderance. More women than men have Alzheimer disease, but in a metaanalysis of 56 different studies, there were no significant sex differences in the incidence or prevalence of amnestic mild cognitive impairment (04). In one multimodality brain imaging study, cognitively normal older women and men were compared with respect to sex-based differences in brain biomarkers; women were shown to develop the Alzheimer phenotype earlier than men with higher beta-amyloid accumulation and lower levels of glucose metabolism (83).

Racial disparities. Several studies have shown higher incidence rates for Alzheimer disease among African Americans than for white Americans. A meta-analysis demonstrated that the Alzheimer incidence rate for African Americans was 64% higher than for Caucasians (95). The authors concluded that the reasons for these disparities in incidence rates were unknown but may include both biological and socioeconomic factors. A longitudinal study based in Chicago also demonstrated higher incidence rates for Alzheimer disease among black Americans, compared to white Americans (108). Similar racial disparities were also observed when subjective cognitive decline patients were followed over 20 years in a multiracial community sample in New York City (15).

Studies have demonstrated racial disparities in molecular biomarkers for Alzheimer disease, especially in the prodromal stages, such as in patients with pre-mild cognitive impairment. At the Knight Alzheimer Research Center in St. Louis, for example, a total of 1255 participants (173 African Americans) were enrolled in biomarker studies from 2004 to 2015 (69). These authors found that in participants with a positive family history of dementia, hippocampal volumes were lower in African Americans, compared to Caucasian participants. Mean CSF concentrations of total tau and p-tau were also lower in African Americans, even when data were adjusted for sex, educational level, and APOE-e4 status. These findings of lower levels of CSF tau-based biomarkers in African Americans with prodromal Alzheimer disease were replicated in a case-control study from Emory University in Atlanta (28).

In the Atherosclerosis Risk in Communities (ARIC) study, analyses of Black American participants showed associations between the presence of ventricular arrhythmias and dementia, whereas no such associations were found among white participants (73). Another observation that helps to explain racial disparities is that living in a disadvantaged neighborhood appears to have an effect on accelerated degeneration of the brain in signature regions that are affected by Alzheimer disease (37). Those living in disadvantaged neighborhoods in later life are also less likely to perform well on standard cognitive tests.

Heredity. Cognitively normal older adults who carry at least one APOE-e4 allele are at higher risk for developing pre-mild cognitive impairment (and at an earlier age) than noncarriers (13). The APOE-e4 allele is the strongest genetic risk factor for late-onset Alzheimer disease. Those carrying the APOE-e4 allele (compared to those carrying e2 or e3) are more likely to progress more rapidly from the pre-mild cognitive impairment state to the Alzheimer dementia state (82).

Anosognosia. There is a subset of pre-dementia Alzheimer disease patients who have no subjective awareness of their cognitive decline in the very early stages (13; 44). They represent a group with early anosognosia, or denial of cognitive deficits. When one group of investigators examined how awareness of diagnosis affected quality of life among older adults with mild cognitive impairment, they learned that those who were aware of their diagnosis had lower satisfaction with daily life and lower physical well-being than those who were unaware (96).

Prognosis and complications

Subjective cognitive decline. Several studies with subjective cognitive decline patients have shown that the rate of cognitive and functional decline is about 7% per year. When patients with global deterioration scale stage 2 (subjective cognitive decline with subtle cognitive impairment) were followed for seven years, they were shown to be 4.5 times more likely to progress compared to those with GDS stage 1 (no subjective or objective cognitive decline). The rate of decline in GDS scores was 6.7% per annum (85). Besides progression of GDS scores, these subjective cognitive decline patients also experienced decline in their global cognitive scores, remote memory scores, and functional/self-care scores. This rate of progression matched the 7% per year that had been described earlier by others (65). When subjective cognitive decline patients with and without depression were compared in an Austrian study, those with depression were shown to have more difficulties with activities of daily living than those without depression (97). When Chinese patients with subjective cognitive decline were divided into those with and without worry, those with worry were noted to have poorer sleep quality and reduced sleep efficiency, which was the reason put forward for their higher risk of developing Alzheimer disease (16). “Dementia worry” is a term that refers to the fear of developing dementia; it can have positive outcomes (patients may adopt more positive lifestyle habits), or it can have negative outcomes (patients may begin to underutilize the healthcare system or mistrust their providers) (66). In a longitudinal Italian study of subjective cognitive decline, those who converted to Alzheimer disease within 2 years were more likely to have problems at baseline with long-term verbal memory (07).

Mild behavioral impairment. This is a construct describing the emergence in older adults (> 50) of neuropsychiatric symptoms that are precursors to cognitive decline and dementia. The symptoms of mild behavioral impairment must last at least six months, but not be severe enough to be captured into a traditional psychiatric diagnosis. There is a 34-item mild behavioral impairment checklist that can be completed by the patient, caregiver, or clinician (40). Older adults with subjective cognitive impairment experience increasing affective symptoms, reduced engagement in their usual activities, and lower quality of life than age-matched controls (36). Mild behavioral impairment is more prevalent among those with mild cognitive impairment (85.3%) than among those with subjective cognitive decline (76.5%) (89). Pre-mild cognitive impairment patients who have symptoms of mild behavioral impairment have a more rapid decline in attention and working memory, suggesting that mild behavioral impairment might be a novel target for dementia clinical trials (17). Other investigators have shown that baseline neuropsychiatric inventory anxiety scores are associated with decline in recall on the symbol digit memory task over a period of 5.7 years (11). Baseline neuropsychiatric inventory total scores are associated with a more rapid rate of decline on word list memory, praxis recall, and animal frequency tasks among clinically normal older adults and are predictors for pre-mild cognitive impairment.

Clinical vignette

Case. This 62-year-old woman was referred for memory problems, anxiety, and insomnia for the last eight months. The memory problems began when she started taking steroids for sudden onset of hearing loss in her right ear, but they persisted after she stopped taking the steroids a month later. She had an MRI scan for the hearing loss, but it showed no signs of stroke, hydrocephalus, or hippocampal atrophy. She had a family history of dementia (her mother died at age 85 of Alzheimer disease). She lived on her own and denied functional impairment in performing everyday tasks. Her other medications included lorazepam 1 mg bid and citalopram 20 mg/d for anxiety and diphenhydramine 50 mg qhs for insomnia. She scored 28 out of 30 (normal range) on the Mini-Mental State Examination, losing 2 points on attention tasks. The rest of her neurologic examination was unremarkable. Her complete blood count, lipid panel, and blood chemistries showed no signs of anemia, diabetes, or renal or liver disease.

Advice. This patient was advised to stop the diphenhydramine and to start trazodone 50 mg qhs. The dose of trazodone was eventually moved up to 100 mg qhs because it helped her with both anxiety and insomnia. Diphenhydramine has both anticholinergic and antihistaminergic properties and is known to worsen memory complaints, so it needs to be stopped in those who have symptoms of pre-mild cognitive impairment (30). She was advised to walk for 30 minutes a day and to eat a Mediterranean diet to prevent further progression of her memory problems. Exercise and diet have been shown to slow progression along the continuum from healthy aging to Alzheimer disease (70; 72; 86; 22; 114; 01).

Biological basis

Genetic risk factors for Alzheimer disease. First-degree relatives of patients with Alzheimer disease are at increased risk for developing the disease themselves (RR=1.73) (12). Older adults who carry at least one APOE-e4 allele are at higher risk for memory decline to the pre-mild cognitive impairment state compared to noncarriers (13). The APOE-e4 allele is the strongest genetic risk factor for late-onset Alzheimer disease. APOE-e4 carriers have been observed to have increased accumulation of beta-amyloid in the brain, but this phenomenon has been shown to be attenuated in those who are heterozygous for the KLOTHO gene variant KL-VS (25). The APOE genotype contributes to the heterogeneity in the rate of transition from pre-mild cognitive impairment to Alzheimer dementia as MMSE scores in those carrying the APOE-e4 allele have been noted to fall faster than in those carrying APOE-e2 or e3 alleles (82). New data suggest that the blood brain barrier is affected by the APOE genotype (05). In APOE-e4 mice (compared to e2 or e3 mice), basement membrane thickness is reduced as a function of decreased phosphorylation of occludin (an enzyme that oxidizes NADH) and decreased levels of collagen IV. The APOE genotype has an influence on the obesity paradox in dementia (92). Obesity was noted to be associated with more cognitive decline in cognitively normal older individuals without the APOE-e4 allele, especially in those with the APOE-e2 allele (these patients have more microinfarcts), but obesity is also associated with less cognitive impairment among those with mild cognitive impairment, especially those who carry the APOE-e4 allele. This is called the obesity paradox.

Imaging and CSF biomarkers for Alzheimer disease. In pre-mild cognitive impairment and subjective cognitive decline, several neuroimaging and CSF biomarkers for Alzheimer disease become abnormal before patients develop clinical symptoms of dementia. In the Baltimore Longitudinal Study of Aging, 171 cognitively normal participants 56 to 95 years of age had neuroimaging to measure mean cortical amyloid levels and MRI to assess hippocampal volume (08). Those with signs of amyloidosis or hippocampal atrophy were more likely to experience cognitive progression over time, even though they met criteria for being in the preclinical stages of Alzheimer disease. This was also true in the Ginkgo biloba Memory Study because several of the cognitively normal older participants (over 80 years) had abnormal amyloid PET scans at baseline in that clinical trial (55). In the Swedish Bio FINDER study, for example, increased isotope uptake on tau PET scans was associated with worse clinical performance on a variety of cognitive tests in those who began with pre-mild cognitive impairment (76). These findings were similar to those of the large (n=1343) Mayo Clinic study, where abnormal amyloid PET scans appeared during the pre-mild cognitive impairment stage of pre-Alzheimer disease, but tau PET scans did not become abnormal until patients developed signs of dementia (42). In a smaller longitudinal study of 60 clinically normal participants who were 73 years of age at study onset, Hanseeuw and others showed that an initial rise in amyloid burden on PET was usually followed by tau changes, then cognitive decline (34). Other investigators found in a large group (n=814) of cognitively normal older individuals that three CSF biomarkers were significantly associated with greater cognitive decline after seven years, compared to those who were free of abnormal biomarkers (beta-amyloid, phosphorylated-tau, and total-tau) at baseline (94). These findings supported earlier observations demonstrating the clinical utility of the Erlangen score algorithm as a useful tool in identifying predementia subjects in two large longitudinal study cohorts (the U.S. Alzheimer’s Disease Neuroimaging Initiative and the German Dementia Competence Network) (51). In that algorithm, if CSF beta-amyloid or tau were abnormal, then Alzheimer disease was a possible or probable future diagnosis. Wolfsgruber and colleagues showed that patients with subjective cognitive decline had worse performance on tests of global cognition, as well as lower concentrations of CSF beta-amyloid 42/40 and beta-amyloid 42/p-tau 181 (110).

Imaging and plasma biomarkers for Alzheimer disease. In France, older adults with subjective cognitive decline (n=276) were recruited into a longitudinal study of plasma biomarkers and neuroimaging (14). Cavedo and colleagues found that plasma concentrations of total tau were associated with faster rates of brain atrophy in the basal forebrain. This association was independent of the patient’s beta-amyloid status and APOE genotype. In the Mayo Clinic Study of Aging (n=1343), participants who were recruited were either cognitively unimpaired, had mild cognitive impairment, or had probable dementia (42). The authors learned from this study that a positive amyloid PET scan could be associated with normal cognition, whereas an abnormal tau PET scan was never associated with a cognitively unimpaired person. In the ADNI study, plasma levels of p-tau 181 were increased in the preclinical stages of Alzheimer disease, and they increased further when patients reached the mild cognitive impairment stage (46). In the BioFINDER study, plasma levels of both p-tau 181 and A-beta 42/40 were both good predictors for Alzheimer dementia within six years in those who had mild cognitive impairment at baseline (77). Of these biomarkers, plasma p-tau 181 has emerged as the most sensitive and specific to detect Alzheimer disease in older adults (02).

Imaging risks for mild cognitive impairment and dementia. In the United States, older adults with unimpaired cognition (n=1553) were followed in a multicenter longitudinal imaging study (111). The MRI factors that increased the risk of developing mild cognitive impairment were decreased hippocampal volume (HR=1.27) and any microhemorrhages (HR=1.79). Lacunar infarcts increased the risk for dementia (HR=1.66), as did higher volume of white matter hyperintensities (HR=1.44) and lobar microhemorrhages (HR=1.90). In a French study of 318 older adults with subjective memory complaints and a mean age of 76, the mean MMSE score at baseline was 28.7 (24). Four participants (all positive for amyloid-PET at baseline) progressed to prodromal Alzheimer disease (they were more likely to be older, to have more highly positive PET scans, and to be carriers of the APOE-e4 allele).

Etiology and pathogenesis

Neurodegeneration. In most cases of pre-mild cognitive impairment, the etiology is presumed to be neurodegeneration due to early Alzheimer disease (13; 42; 76). In patients who were imaged in various stages of the Alzheimer disease continuum, the accumulation of amyloid (on amyloid-PET) was shown to occur in the pre-MCI stage of Alzheimer disease (41). This preceded the deposition of tau (on tau-PET) and the development of neurodegeneration on MRI. In a large longitudinal study (n=1843), patients with cognitive complaints about one year prior to autopsy were classified according to no complaint, self-only, informant-only, or mutual (both self and informant) complaint. The mutual complaint group had higher odds (OR=5.82) of having autopsy evidence of Alzheimer disease (29). Longitudinal imaging studies with tau PET have shown that baseline tau uptake patterns map better onto the clinical phenotype of Alzheimer disease than longitudinal tau uptake patterns, supporting the important causal role of tau as a driver of clinical dysfunction in Alzheimer disease (93; 105).

Neuroinflammation. Alzheimer disease is also associated with neuroinflammation as well as cerebrovascular changes. The authors of the Swedish Bio FINDER cohort followed 256 patients with mild cognitive impairment, 57 with Alzheimer dementia, and 508 cognitively unimpaired elderly persons for about six years, monitoring 11 neuroinflammatory and cerebrovascular biomarkers, as well as conventional biomarkers of Alzheimer disease (43). They found that elevated levels of five inflammatory and cerebrovascular biomarkers were associated with Alzheimer pathology. Neuroinflammatory biomarkers were detected in the pre-mild cognitive impairment stage, as well as the mild cognitive impairment and dementia stage of Alzheimer disease.

Blood-brain barrier breakdown. Animal studies have suggested that both amyloid and tau can lead to blood vessel abnormalities and blood brain barrier breakdown. Nation and others examined capillary damage in humans who were in the early stages of cognitive dysfunction on the Alzheimer continuum (71). They showed evidence of blood brain barrier breakdown in the hippocampus that was independent of levels of amyloid or tau.

Aortic stiffness. In the Vanderbilt Memory and Aging Project, data on aortic stiffness, biomarkers of neuroinflammation, and Alzheimer biomarkers were collected from 146 stroke-free and dementia-free participants (68). Among those who were 74 years of age and older, greater aortic stiffness was associated with biomarkers of neuroinflammation, phosphorylated tau, synaptic dysfunction, and neurodegeneration, but not with amyloidosis. The authors proposed that neuroinflammation and synaptic dysfunction might be the common two pathways by which aortic stiffness exerted its injurious effects on the brain.

Intermittent hypoxia. In a 2-year prospective longitudinal study of nondepressed healthy older adults, the severity of obstructive sleep apnea was shown to be associated with the annual rate of change in CSF beta-amyloid indices after adjustment for age, sex, BMI, and APOE-e4 (88). The authors attributed the mechanism to be due to either sleep fragmentation or to intermittent hypoxia.

Epidemiology

Progression from pre-mild cognitive impairment to mild cognitive impairment or dementia. Pre-mild cognitive impairment patients with clinical complaints but no neuropsychological deficits were observed to progress to mild cognitive impairment or dementia at a rate of 22% over a period of 2 to 3 years, compared to a rate of 3.7% in an age-matched group of normal controls (54). The progression rate was faster (38.9%) in those who also had objective neuropsychological deficits in one or more memory measure. Progression was best predicted by the baseline performance on certain cognitive tests (at least two memory domains). When a meta-analysis was performed of studies of older individuals with and without subjective memory complaints, the annual risk of developing mild cognitive impairment in those with memory complaints was 6.67% (65).

Progression from healthy control to mild cognitive impairment. When pre-mild cognitive impairment patients were compared in a longitudinal study to those without cognitive impairment, progression to mild cognitive impairment was faster in those with pre-mild cognitive impairment (23). The speed of progression to dementia was even faster among those who had mild cognitive impairment. The time required for clinically meaningful cognitive decline in amyloid positive patients with normal cognition at baseline to mild cognitive impairment was about six years (39). The amyloid positive patients had abnormal amyloid PET scans or positive results on CSF studies. This study was done to determine that a clinical trial with a preventive agent in pre-clinical Alzheimer patients would need 2000 subjects at baseline to achieve 80% power in a 4-year trial to see a treatment effect size of 25%. Van Dyck and others enrolled a somewhat smaller number of mild cognitive impairment or mild Alzheimer dementia patients into the lecanemab (n=898) and placebo (n=897) arms of their clinical trial (103). By the end of 18 weeks, they found 27% slowing in the severity of disease progression among those treated with lecanemab, compared to those assigned to placebo.

Progression from healthy control to Alzheimer dementia. In the Pittsburgh participants from the Ginkgo biloba Memory Study, the time required to decline from normal cognition (n=145) to dementia (n=60) among a group of older than 80-year-old participants was about five years (55). The time required for significant cognitive decline in two groups (65 years and older) of African Americans and non-Hispanic whites in Chicago was about six years (22). After that period of time, 20.5% to 24.9% had developed Alzheimer dementia. In a national multiethnic longitudinal study (Health and Retirement study), the risk for dementia by age 76 was 10%, whereas the risk for mild cognitive impairment was 22% (58).

Prevention

Diet. High adherence to three different diets (MIND, Mediterranean, and DASH diets) has been shown to reduce the risk of Alzheimer disease in healthy adults aged 58 to 98 years who were followed for an average of 4.5 years (70). Current alcohol use was associated with reduced risk of progression to incident dementia in those younger than 87 years (27). Moderate adherence to the MIND and the Mediterranean diets may also reduce Alzheimer risk (22; 01). In one randomized controlled trial of older adults who were not yet cognitively impaired (but who had a family history of dementia), the MIND diet with mild caloric restriction improved cognitive scores after three years to a similar extent as the control diet with mild caloric restriction (06). It is important to note that enrollment criteria for this study included a BMI greater than 25 and a baseline “suboptimal diet”, as determined by a 14-item questionnaire. This trial is likely to have been under-powered to detect cognitive differences between the two treatment groups at three years, since only 301 participants were assigned to the MIND diet, and 303 to the control diet.

Exercise. Physical activity in older adults may improve cognitive reserve so that the brain can better withstand various neurodegenerative and vascular insults. A meta-analysis of 19 exercise studies comparing exercise with nonexercise control groups (1145 subjects with a mean of 77 years) concluded that exercise training may delay the decline in cognitive function that occurs in individuals at risk of Alzheimer disease, with aerobic exercise being the most effective intervention (79). The Rush Memory and Aging Project reviewed data from 454 decedents in the Chicago area who had had actigraphic data; they showed that higher total daily physical activity was independently associated with better global cognitive scores (10). A meta-analysis of 41 exercise intervention studies demonstrated that adherence to exercise interventions in older people with mild cognitive impairment can be limited by adverse events, such as mild joint pain, injuries, and falls (21). In a prospective examination of 501,376 nondemented United Kingdom Biobank participants, 5185 cases of dementia were identified at 10.7 years (114). Frequent engagement in exercise (household-related activity or leisure-time activity) was shown to be associated with reduced risk of dementia at 10.7 years (HR=0.65). In the Southern Community Cohort Study of 17,209 nondemented older persons, 1694 cases of dementia were identified over a median period of four years (113). Engaging in greater than 150 minutes of moderate exercise per week was shown to reduce the risk of all-cause dementia (HR=0.89). This beneficial effect was found to persist regardless of age, sex, race, or income level.

Cognitive reserve and brain reserve. Compared to those in the predementia stages, those with dementia have less education and a greater amount of baseline cerebral atrophy (31). This study suggested that educational levels provided a proxy for cognitive reserve (slowing the development of cognitive decline), whereas intracranial volume as measured by MRI was a proxy for brain reserve (slowing the thinning of cortical areas involved in Alzheimer disease). In a nation-wide longitudinal study of aging (Health and Retirement study), investigators found that each year of education produced a measurable decrease in the risk of both dementia and mild cognitive impairment (58). Living in an advantaged neighborhood proved to be another factor that was correlated with both cognitive reserve and brain reserve (37). Those who live in better neighborhoods are less likely to have cognitive decline in their later years, and they also have less cortical thinning in areas of the brain that are affected early on by Alzheimer disease.

Multidomain interventions. In Europe, a multidomain intervention of diet, exercise, cognitive training, and vascular risk monitoring was shown to produce a small amount of cognitive improvement at the end of two years (n=631), compared to a control intervention (n=629) in an at-risk elderly population (72). In the United States, a multidomain intervention study has been designed for those who are at various stages of cognitive impairment, including mild dementia, mild cognitive impairment, and subjective cognitive decline (86). The intervention consists of weekly phone calls and three visits with the patients’ “health care champion” to reinforce adherence to the Mediterranean diet, regular physical activity, and cognitive activity. A United States randomized trial of aerobic exercise and the DASH diet in 160 sedentary older men and women who had cognitive impairment with no dementia (CIND) demonstrated that those assigned to the aerobic exercise arm (with no dietary change) experienced significant improvement in executive function (09). The largest cognitive improvements, however, were seen in those assigned to the exercise plus diet arm of the study. Multidomain intervention has also been demonstrated to be beneficial in two other United States studies, where those who adhered to at least four of five lifestyle changes had a 60% reduction of Alzheimer risk (at least 150 min/wk of moderate exercise, smoking avoidance, light-moderate alcohol intake, MIND diet, and cognitive activities) (22). The Lancet Commission added three new modifiable risk factors to the “standard” list of nine modifiable risk factors for dementia (poor education, hypertension, hearing impairment, smoking, obesity, depression, physical inactivity, diabetes, and insufficient social contact). The “new” risk factors added by the Lancet Commission are excessive alcohol consumption, traumatic brain injury, and air pollution (52).

Controlling blood pressure. Elevated blood pressure (> 140/90) during midlife and the persistence of high blood pressure into late life have both been associated with an increased risk of incident dementia in a community-based cohort (59). The risk of dementia increased with increasing systolic blood pressure, with an additional 20% increase in risk for every 10 mm Hg increment in midlife systolic blood pressure. Persistence of high blood pressure into late life results in a 25% increase in risk of dementia above the risk attributable to midlife hypertension alone. The SPRINT MIND Research Group conducted a randomized clinical trial at 102 sites among those over 50 years of age with hypertension, but without diabetes or stroke, with the aim of reducing mild cognitive impairment and dementia (109). They showed that intensive blood pressure control (systolic less than 120 mm), compared to standard control (systolic less than 140 mm), significantly reduced the risk of mild cognitive impairment (HR=0.81; p=0.007), but there was no significant difference for probable dementia.

Avoiding air pollution. Long-term exposure to ambient air pollution is known to be associated with increased risk of cardiovascular disease and stroke (48). Studies have shown that long-term exposure to air pollution in urban areas (especially nitrogen dioxide and fine particulate matter) is also a risk factor for cognitive decline (47; 52).

Avoiding anticholinergic medications. Risk factors for mild cognitive impairment were investigated in 688 cognitively normal participants in the Alzheimer Disease Neuroimaging Initiative (ADNI) who were 73.5 years of age at baseline and were followed for over 10 years (107). Those who were chronically taking drugs with anticholinergic activity had increased risk for progression to mild cognitive impairment (HR=1.47); those who were taking anticholinergic drugs and carried an APOE-e4 allele showed even greater risk (HR=2.69).

Avoiding disruptive sleep habits. In the A4 study of cognitively unimpaired older adults, data were collected on sleep duration and amyloid PET on 4425 participants (mean age was 71.3 years) (38). Each additional hour of nighttime sleep was shown to be associated with reduced global amyloid uptake on PET. Daytime sleep could not compensate for loss of nighttime sleep in terms of the protection that sleep offered with respect to the accumulation of brain amyloid. Obstructive sleep apnea is a potential modifiable risk factor for cognitive disorders in the elderly (88; 38). Positive airway pressure has been demonstrated to be associated with a lower incidence of cognitive disorders and cognitive decline among adults who have obstructive sleep apnea (91).

Differential diagnosis

Confusing conditions

Anticholinergic medications. Impaired cognition is a known risk of using anticholinergic medications, especially in those who are over the age of 65 years. In a large cohort study of the elderly (n=3434) with a mean follow-up of 7.3 years, 797 participants developed dementia; higher cumulative anticholinergic use was associated with an increased risk of dementia (30). When we hear our patients telling us about memory complaints, we must be sure to discontinue antihistamines, gastrointestinal antispasmodic agents, antivertigo agents, bladder antimuscarinic drugs, and antidepressants with anticholinergic effects.

Ischemic amnesia. Amnesia as the main complaint of an ischemic event is rare, and it can be difficult to distinguish ischemic amnesia from transient global amnesia (63). Most amnestic strokes are cardioembolic in origin, involving multiple territories, including the posterior circulation and the limbic system. About a third of amnestic stroke patients have persistent memory problems.

Associated or underlying disorders

Sleep disorders. Sleep-disordered breathing, characterized by recurrent arousals and intermittent hypoxemia, has been shown to be a prelude to cognitive impairment and accumulation of brain amyloid (45; 33). In a large 5-year study of community-dwelling men and women, the presence of sleep-disordered breathing was associated with a younger age of onset of mild cognitive impairment (mean age = 72.6 years), compared to a group with no sleep disordered breathing (mean age = 83.7 years) (75). The link between poor sleep and Alzheimer disease has been demonstrated as an association between disruptions of non-REM slow wave sleep and CSF levels of beta-amyloid (57). This overexpression of CSF beta-amyloid with sleep disruption can be seen both in animal models of Alzheimer disease as well as in nondemented humans (88; 38).

Psychiatric conditions. Depression can be a factor associated with symptoms of pre-mild cognitive impairment and mild cognitive impairment. Among 183 mild cognitive impairment patients, the presence of significant depressive symptoms on the Cornell Scale for Depression in Dementia appeared to be a predictor of progression to Alzheimer disease (HR = 2.06) (102). Those who had signs of severe depression were excluded in this study. The mild behavioral impairment checklist is useful in separating patients with mild behavioral impairment from those who might have traditional psychiatric conditions (40). Mild behavioral impairment patients, for example, might have decreased motivation, but they may not score as high as someone with severe bipolar depressive disorder. Referral to geriatric psychiatry is always recommended whenever this distinction is in doubt. Late-onset psychosis and REM sleep behavior disorder may predate clinical signs of dementia by many years in patients with prodromal dementia with Lewy bodies (101).

Medical illnesses. Hearing and visual loss, as well as anemia and hepatic and renal failure, may all lead to symptoms similar to those of pre-mild cognitive impairment (26; 52; 112). In a large study of elderly Stockholm residents, 5.8% developed dementia during a 5-year follow-up period (112). Those who had lower glomerular filtration rates had a greater likelihood of developing dementia. In the Atherosclerosis Risk in Communities Study, there was a higher rate of dementia among older adults who had evidence on ECG of ventricular arrhythmias than among controls (73). These factors need to be considered (and appropriate laboratory levels drawn and ECG performed) whenever pre-mild cognitive impairment patients are being evaluated for memory complaints.

Diagnostic workup

Medication history. It is important to take a thorough medication history in patients who have pre-mild cognitive impairment and discontinue or reduce the dose of drugs that have anticholinergic effects (26). These drugs can have serious cognitive effects, and in some cases, they can even lead to a diagnosis of dementia (30; 107).

Laboratory tests. Pre-mild cognitive impairment patients need a complete blood count to exclude anemia, complete chemistry panel to exclude renal and hepatic failure, vitamin B12 and D levels to exclude deficiencies in those vitamins, and thyroid stimulating hormone level to exclude hypothyroidism as a cause of memory loss (26).

Alzheimer disease biomarkers. Three pathologic processes are key to making an accurate diagnosis of Alzheimer disease: deposition of extracellular amyloid plaques, accumulation of intracellular tau-based neuro-fibrillary tangles, and evidence of neurocortical neurodegeneration. During the pre-mild cognitive impairment stage, many patients will be curious about their risk of Alzheimer disease and will request plasma or CSF tests of Alzheimer disease biomarkers. Recent guidelines have been published to help clinicians to determine: a) the appropriateness of CSF or plasma biomarker testing, b) what pretest educational/consent information to implement, and c) what follow-up procedures to use to best promote a person’s well-being (50). One commercially available blood test for Alzheimer disease calculated an amyloid probability score after measuring plasma amyloid-beta 42/40 ratio, APOE-e4, and age (02). The cost of this test in 2022 was $1250. The test cannot make the diagnosis of Alzheimer disease, but it can increase confidence that the patient has an increased number of amyloid plaques in the brain.

Magnetic resonance imaging (MRI). Most neurologists recommend an MRI for patients with pre-mild cognitive impairment to exclude treatable conditions such as tumor or hydrocephalus. In the Rotterdam Study (a large population-based longitudinal study with n=4198), investigators asked which MRI correlates were most likely to be associated with prognosis of mild cognitive impairment; they found that the best correlates were larger white matter lesion volume and higher prevalence of lacunes (19). Measurements of hippocampal volume have been used to predict which cognitively unimpaired individuals will be most likely to progress to mild cognitive impairment or dementia (55; 111). The combination of the Mini-Mental State Examination and hippocampal volume, normalized by the cerebellar volume, can distinguish which mild cognitive impairment patients will progress to Alzheimer disease within five years from those who will not progress (98). In the Atherosclerosis Risk in Communities Study, 1553 participants who were dementia-free at baseline were examined with MRI and then tested cognitively about five years later (111). The MRI features that best predicted the onset of mild cognitive impairment were reduced hippocampal volume (HR=1.27) and any microhemorrhages (HR=1.79). In the Mayo Clinic Study of Aging, cortical thickness on MRI was measured in 192 patients with various mild cognitive impairment subtypes and 1257 cognitively normal individuals (56). The subtle cognitive impairment cluster (pre-mild cognitive impairment patients) showed signs of atrophy in the entorhinal and parahippocampal cortices, whereas other mild cognitive impairment subtypes demonstrated atrophy in these areas and others, depending on their clinical presentation. In the Swedish National study on Aging and Care in Kungsholmen (SNAC-K), 436 cognitively normal elderly participants were imaged with MRI at baseline and followed at 3, 6, and 9 years (106). Incident dementia was found in 46 participants and was associated with the baseline microvascular lesion score (including lacunes, white matter hyperintensities, and perivascular spaces) as well as the neurodegeneration load (enlarged ventricles and reduced hippocampal volume). Advanced MRI techniques are capable of revealing the “cingulate island sign” that is pathognomonic for prodromal dementia with Lewy bodies (101).

FDG-PET scans. These scans are not routine, but they are intended for research studies of pre-mild cognitive impairment patients. When pre-mild cognitive impairment and “non-decliner” control groups were compared in longitudinal studies of APOE-e4 carriers, the pre-mild cognitive impairment patients were shown to have lower baseline cerebral metabolic rates on FDG-PET scans in posterior cingulate gyri, bilateral parietal cortices, and left prefrontal cortex, and these lower rates correlated with subsequent decline in verbal memory (13). FDG-PET, neuropsychology tests, and CSF biomarkers were examined from 85 amnestic mild cognitive impairment patients from the Alzheimer’s Disease Neuroimaging Initiative study cohort to see which would be the best predictors of conversion from mild cognitive impairment to Alzheimer dementia, and abnormalities on baseline FDG-PET (HR= 2.72) and an episodic memory test (HR=4.30) were demonstrated to be the best predictors of conversion (49). When stable mild cognitive impairment patients (n=46) were compared with those who progressed over time (n=33), the dynamic features of FDG-PET at baseline appeared to be the best predictors of progression, compared to the static features (99). In the pre-mild cognitive impairment FDG-PET study of Tondo and others, 105 participants (45%) had normal scans, whereas 58 (55%) showed evidence of hypometabolism (100). The most frequent abnormal pattern was frontal hypometabolism (28%), followed by an Alzheimer-like pattern (11%).

Amyloid PET scans. These scans have been used in identifying prodromal or mild Alzheimer patients in studies of antibody-based immunotherapies such as aducanumab, donanemab, and lecanemab (87; 78; 64; 103). These studies demonstrated that monthly infusions of aducanumab were capable of reducing brain levels of beta-amyloid in a dose-dependent and time-dependent manner. Besides being useful as endpoint measures in clinical trials, amyloid PET scans are also useful predictors of disease progression. When preclinical Alzheimer patients with and without evidence of abnormal beta-amyloid were assessed and followed with cognitive tests, it was demonstrated that those with abnormal beta-amyloid were the ones who deteriorated cognitively over a 5- to 6-year period of time and eventually met criteria for dementia or mild cognitive impairment (24; 55; 39). Jack and colleagues made the point that abnormal amyloid PET scans could be recorded in cognitively normal individuals, whereas that was never true of abnormal tau PET scans (42).

Functional MRI. These scans are for research. The default mode network (measured by functional MRI) is active when people are daydreaming or not focused on the outside world. When 124 presymptomatic individuals with a family history of Alzheimer disease were studied with default mode network MRI imaging, higher connectivity between the medial temporal memory system and the rest of the brain was associated with better baseline memory, attention, and global cognition scores (104). Those 55% with subjective cognitive decline were found to be more likely to have an Alzheimer-like brain connectivity pattern.

Tau PET scans. These scans are used for research. In preclinical Alzheimer patients, increased uptake of the PET tracer F-18 flortaucipir in various regions of the brain was associated with worse performance on several cognitive tests at a stage when there was no apparent association between cognitive function and change in cortical thickness measures on MRI (76; 60). Jack and others pointed out that abnormal tau PET rarely occurred in the absence of abnormal amyloid PET, but that the reverse was common (42; 34). In their dynamic biomarker model of Alzheimer disease, amyloidosis was shown to develop first, then tauopathy appears later and was associated with the onset of clinical symptoms. Tau PET scans are useful in being able to distinguish the typical amnestic Alzheimer disease phenotypes from the atypical Alzheimer variants, such as posterior cortical atrophy and the logopenic variant of primary progressive aphasia (93; 105). In preclinical Alzheimer disease patients, the presence of tau deposition in the entorhinal cortex on tau-PET appears to mediate the association between novelty-related activity in the locus coeruleus and memory decline (81).

Management

Cognitive interventions. The theory behind cognitive training in those with pre-mild cognitive impairment or mild cognitive impairment is to prompt neuronal reorganization to meet the demands of new tasks, but the results of existing systematic reviews have been mixed. Sherman and colleagues conducted a meta-analysis of 26 randomized controlled trials published on this topic and showed that memory-based approaches were more successful than multidomain methods (90). The effect size measured in the Sherman study (Hedges g=0.39) was in the same range as that (g=0.36) reported in the meta-analysis of Mewborn and associates (61). Working memory interventions were also noted by Mewborn’s group to be the most effective (g=0.48). Cognitive training has been part of some multidomain intervention trials (72).

Continuous positive airway pressure (CPAP). If patients with pre-mild cognitive impairment have signs on polysomnography of sleep-disordered breathing, they are likely to benefit from CPAP. In a large prospective study of older adults with sleep-disordered breathing, CPAP was shown to significantly delay onset of mild cognitive impairment by about 10 years (75). Treating sleep apnea with CPAP also improves quality of life in those with predementia and dementia (74; 91).

Healthy lifestyle. Older at-risk participants in two large urban-based studies were followed for 5 to 6 years, and 20% to 25% developed incident Alzheimer dementia (22). Those who followed at least four elements of a healthy lifestyle (nonsmoking, moderate exercise for greater than 150 min/week, light-moderate alcohol use, high quality diet, sleeping for 7 to 9 hours per night, and engagement in cognitive activities) had at least a 60% reduction in risk of incident Alzheimer dementia.

Anti-amyloid therapies. There are now two anti-amyloid therapies that have been approved by the FDA for the treatment of patients with mild cognitive impairment and mild dementia: aducanumab (18) and lecanemab (103). A third anti-amyloid agent, donanemab (64) is likely to be approved by the FDA by the end of 2023. The cognitive treatment effects of these agents are small (lecanemab reduces disease progression by 27% over 18 months, compared to placebo), but both approved drugs are effective at removing amyloid from the brain. The downside risk of lecanemab therapy was that 12.6% of treated patients developed a side effect known as amyloid related imaging abnormality - edema (ARIA-E) (103), somewhat less frequently than the 35.2% figure recorded in the aducanumab trials (18). Not all of these patients with ARIA-E were symptomatic, and most of the symptoms resolved with discontinuation of the drug. Simulation models have calculated that if anti-amyloid therapy is started in the pre-dementia stages, then it is possible for the patient to have an additional year to spend in the community, rather than in institutional care (67).

Outcomes

Nonsteroidal antiinflammatory drugs (NSAIDs). The hypothesis that NSAIDs might protect against the development of Alzheimer disease has been supported by a large number of cohort studies in older adults (35). There was a biologically plausible role for NSAIDs in pre-mild cognitive impairment because there is evidence for inflammatory activation in the early interaction of microglia and astrocytes with beta-amyloid and hyperphosphorylated tau. Nevertheless, Meyer and associates reported a negative result when they performed a 2-year, double-blind, placebo-controlled prevention trial examining the use of naproxen (vs. placebo) to slow the progression of disease in 195 presymptomatic Alzheimer disease patients (62). The primary outcome measure was rate of change on a multimodal presymptomatic Alzheimer Progression Score, which included tests of cognition, olfaction, hippocampal volume, and CSF levels of tau and beta-amyloid.