Clinical background [return to contents]

Dementia is an increasing public health and economic burden for the United States as the population ages.¹ The total number of adults who will develop dementia each year is projected to approach about 1 million in 2060 compared with just over half that number in 2020.¹ Women are at higher risk than men, and Black adults are at higher risk than White adults.¹

Mitigation of some of the actionable risk factors for dementia may attenuate these projections. Dementia incidence (eg, per 1,000 people tracked for 1 year) in the United States is actually declining—possibly due to reduction of population-attributable risk factors over time, including higher education attainment and less smoking.² Prevalence, in terms of the percentage of people with dementia for any given age group, appears to be similarly decreasing.³ Compared to preceding generations, younger cohorts have improved access to medication to control other dementia risk factors, which include high low-density lipoprotein levels, depression, and hypertension.³

Diabetes and obesity are dementia risk factors that continue to increase in the US population that may have dampened declines in dementia in the past.³ Population-level lifestyle changes and potentially new medications, such as glucagon type-1 receptor agonists, may help mitigate these risk factors in the future.⁴

Dementia is a syndrome with a collection of symptoms that can be grouped into different etiologies, the most common of which is Alzheimer’s disease (AD). Some conditions have dementia-like symptoms but are treatable and in some cases reversible; thus, identifying the cause of dementia is crucial for providing appropriate patient care.⁵

This Clinical Focus describes the role that laboratory testing can play in diagnosing mild cognitive impairment (MCI), dementia, and dementia-like symptoms caused by AD (vs other causes), excluding neuroimmune causes. Once a diagnosis is made, laboratory testing also plays a role in assessing prognosis and drug and clinical trial eligibility. This information is not intended as medical advice. Test selection and interpretation, diagnosis, and patient management decisions should be based on the physician’s education, clinical expertise, and assessment of the patient.

Dementia causes

AD causes 60% to 80% of dementia cases; according to the Alzheimer’s Association, an estimated 7.4 million people 65 years and older currently have AD, and the number is expected to rise to nearly 14 million by 2060.⁵ Other causes of dementia include, but are not limited to,⁵

• Cerebrovascular disease (5%-10% of cases)
• Lewy body disease (LBD, 5%)
• Parkinson disease (PD, 4%)
• Frontotemporal degeneration (FTD, 3% ≥65 years; 10% <65 years)
• Hippocampal sclerosis (HS, <2%)
• Limbic-predominant age-related TDP-43 encephalopathy (LATE, percentage unknown but autopsy studies suggest that it could be a common cause in patients ≥75 years)

Mixed pathologies cause >50% of dementia cases and are common in patients ≥85 years (eg, HS prevalence, which increases to 3%-13% as a mixed etiology).⁵ Although progress is being made in slowing disease progression with the approval and use of the first disease-modifying therapies in AD, most of these dementias are irreversible.

Causes of other non-AD dementias include Huntington disease, HIV infection, traumatic brain injury, substance use disorder, prion disease, and stroke.⁶ Neurocognitive symptoms may emerge after disease onset in most of these conditions as well as in PD.⁷ In contrast, cognitive and behavioral symptoms are the first manifestation of the disease in AD, cerebrovascular disease, LBD, FTD, and LATE.

Some conditions can cause dementia-like symptoms but are not considered dementia (see “Non-AD dementia and dementia-like symptoms” section).⁵ These include depression, sleep apnea, delirium (eg, caused by drug side effects, substance intoxication or withdrawal), infection (eg, syphilis), brain blood clots or tumors, thyroid disease, head injury, or certain vitamin deficiencies.^5,8-11 Importantly, dementia-like symptoms may often be reversed with treatment once these causes are identified; untreated, these causes may result in irreversible dementia (eg, Wernicke-Korsakoff syndrome¹²).

Dementia diagnosis

Dementia is diagnosed based on evidence of significant cognitive decline reported by the patient, an individual close to the patient, or a clinician, preferably documented through objective quantifiable clinical assessment.^6,7 This decline is relative to the patient’s previous abilities in complex attention, executive function, learning and memory, or perceptual-motor or social cognition. For a diagnosis, these cognitive impairments must interfere with the patient’s independence in everyday activities, such that they minimally require assistance (eg, with paying bills and managing medications). Additionally, impairments cannot be attributed to delirium or other mental disorder (eg, major depressive disorder or schizophrenia).

Dementia is also defined as “a major neurocognitive disorder” by the Diagnostic and Statistical Manual of Mental Disorders DSM-5-TR^™.⁶ The DSM-5-TR definition differs from the previous DSM-IV definition in that impairments in learning and memory are not necessary for diagnosis and it is more inclusive for younger symptomatic patients.⁷

Mild cognitive impairment (MCI) diagnosis

DSM-5-TR was also revised to define “a mild neurocognitive disorder” with a very similar definition to “a major neurocognitive disorder” but with cognitive declines being qualitatively “modest.”⁶ DSM-5-TR defines these cognitive declines as not interfering in everyday activities (although more effort or accommodation is required) and being quantitively less (ie, 1-2 standard deviations [SD] below normative means for standardized cognitive tests vs >2 SD for dementia).^6,7

Despite the DSM-5-TR definitions, the terms “dementia” and “MCI” are favored in the literature. Knowing whether a person with a mild neurocognitive disorder/MCI is at risk for developing AD can help patients and their families to better prepare for future care needs¹³ and enable clinical trial eligibility. Ruling out AD can prompt investigation of non-AD causes of MCI or dementia.^8-10,13

Clinical trials for AD therapeutics have used various criteria to identify eligible patients with MCI.^7,14-16 Bondi et al proposed neuropsychological criteria that (1) defined impairment as 1 SD below normative means to capture more individuals with suspected MCI, (2) required 2 impaired scores for a cognitive test to support a reliable observation, and (3) incorporated activities of daily living scores.¹⁷ The last criterion has been incorporated into the Clinical Dementia Rating (CDR^®) sum of boxes score, which is commonly used as a primary endpoint in many clinical trials.¹⁸ In head-to-head comparisons with other MCI criteria, a higher percentage of study participants who fulfilled the Bondi criteria progressed to dementia.^17,19

Importantly, the Bondi criteria were also supported by significant associations with cerebrospinal fluid (CSF) AD biomarkers including beta-amyloid (1-42) (Aβ42), total tau, and hyperphosphorylated tau (see “CSF-based biomarkers” section),¹⁷ indicating that they may be better suited to identifying MCI-AD.

Given that different phenotypic presentations of AD are well known, biomarkers are important in diagnosis and monitoring.^20,21 Accordingly, the Alzheimer’s Association Workgroup revised the definition of AD from one based on clinical presentation to “a biological process that begins with the appearance of AD neuropathologic change.”²⁰

Blood-based biomarkers

Compared with CSF and imaging biomarkers, blood-based biomarkers have emerging roles as less invasive and less expensive methods for characterizing MCI as a prodromal state of AD (Table 1), but they have been challenging to incorporate into clinical practice until recent years. Blood is a more complex matrix, and biomarker levels are much lower than in CSF.⁴¹ Fortunately, ultrasensitive methods have been developed that better quantify the blood-based biomarkers.²⁴ Depending on the care setting and clinical use, different performance thresholds are proposed for assessing amyloid pathology (Table 2).^42,43

According to Alzheimer’s Association guidelines and the Global CEO Initiative (CEOi) consensus statement, these thresholds should only be interpreted in clinical context that includes the predictive ability of the test within the population examined, which depends on the prevalence of amyloid pathology in that population.^42,43 For example, prevalence in a primary care setting is likely to be closer to 20%,⁴⁴ whereas it is likely to be 50% to 80% in specialized care; thus, follow-up on a positive result is better supported in specialized care (eg, with positron emission tomography (PET) and CSF testing capabilities).⁴³

The prevalence in combination with the sensitivity and specificity of the test determine the positive predictive value (PPV) and the negative predictive value (NPV). High PPV can be used to help “rule in” and high NPV can be used to help “rule out” a diagnosis based on a test result. For example, in a population with a 50% prevalence of PET positivity⁴³

• For a confirmatory test, PPV and NPV should both be approximately 90%
• For a high-specificity triaging test, the PPV should be 86% and NPV should be 89%
• For a low-specificity triaging test, the PPV should be 78% and NPV should be 88%

To confirm or rule out amyloid pathology, these performance metrics can be achieved in 2 ways⁴³:

• A single static threshold for the biomarker (ie, cutpoint value, such as a ratio, concentration, or score, above or below which the required sensitivity and specificity is attained)
• A 2-value, dynamic high-low threshold approach (the lower cutpoint is set to achieve the desired sensitivity and the upper cutpoint is set to achieve the desired specificity; ideally, this approach requires no more than 15% to 20% of patients obtaining an intermediate/indeterminate result with values between the high-low thresholds)

Revised criteria for diagnosis and staging of AD provided by the Alzheimer’s Association Workgroup²⁰ cite the CEOi consensus statement in evaluating plasma assays. “Core 1” biomarkers are those that have sufficient performance to be used for early AD detection of amyloid pathology and tau hyperphosphorylation. Abnormal results for a Core 1 biomarker can be considered equivalent to CSF assays in detecting abnormality in PET in the intended use population (MCI and AD). As a single biomarker, only p-tau217 meets these clinical performance standards in plasma.²⁰ Hybrid ratios of biomarkers may be interpreted similarly but represent the associated assay rather than the single biomarker.²⁰

The Workgroup also provides 2 other categories of biomarkers. These include “Core 2” biomarkers, which can be used for evaluating later stages of the disease and providing prognostic information in AD assessment, but are not stand-alone tests for AD diagnosis.²⁰ When abnormal, Core 2 biomarkers, which currently include selected tau protein fragments, help assess the contribution of AD to patient symptoms. (These are not further discussed in this Clinical Focus.) The third category includes nonspecific biomarkers that are important in AD pathology but are not specific to AD (Table 1).

CSF–based biomarkers

CSF-based testing is considered a benchmark against which blood-based biomarker testing can be compared.²⁰ Concentrations of biomarkers are much higher in CSF than in blood, and results correlate well with PET imaging analysis for Aβ.^34,35 CSF biomarkers also provide cutpoints that better separate PET-positive vs PET-negative individuals and tend to be more robust, as assay imprecision and bias affect test results and patient classification to a lesser extent than with blood-based biomarkers.³⁵ CSF Aβ42, p-tau217, p-tau181, and p-tau2³¹ are considered Core 1 biomarkers for AD (Aβ and tau) pathology by the Workgroup (Table 1).

Blood-based tests

In this section, combinations of biomarkers in panels or as hybrid ratios are first introduced before their individual components—since combined biomarkers can better achieve the desired performance characteristics^39,47 defined by guidelines and the CEOi consensus statement.^42,43

AD-Detect panels

AD-Detect^® ABeta 42/40 and p-tau217 Evaluation, Plasma (test code 14258) panel combines results for Aβ42/40 and p-tau217 into a model that generates a likelihood score for PET positivity in individuals ≥50 years.⁴⁷ This 2-marker model achieves the performance characteristics detailed by the CEOi statement for a confirmatory Core 1 diagnostic test for Aβ pathology.⁴³ The model was optimized by using 2 cutpoints, 1 for high and 1 for low risk of amyloid PET positivity, evaluated at ≥90% specificity and sensitivity, and it exceeds the recommended test specifications of ≤20% of patients having an intermediate (indeterminate) result.⁴³

The model was developed based on data from an ethnically diverse (54% Hispanic) population of 215 participants from a memory clinic (1Florida AD Research Center [ADRC]) with a diagnosis of MCI or AD and a moderate prevalence of PET positivity.⁴⁷ Importantly, PET-positive and PET-negative patients were matched based on age, sex, ethnicity, and years of education in contrast to previous studies.^48,49

The model achieved an area under the receiver operating characteristic curve (AUC-ROC) of 0.929 with dual cutpoints at ≥90% sensitivity and specificity providing the following: high cutpoint, 87% PPV; low cutpoint, 91% NPV; 83% accuracy (both cutpoints) at the 46% prevalence of PET-positivity in the intended use population (MCI/AD). The dual cutpoint results in most patients receiving a “high” or “low” likelihood score for PET positivity with relatively few having an “indeterminate” result (15% in the ADRC cohort).⁴⁷ Cutpoints for interpretation of results are provided in Table 4.^39,47

Table 4. Plasma Biomarkers and Likelihood of Having Amyloid Pathology [return to contents]

AD-Detect^® ABeta 42/40, p-tau217, and ApoE Evaluation, Plasma (test code 16493) panel is another confirmatory test for amyloid pathology without the need for follow-up testing (although patient history and clinical and cognitive testing are still needed per guidelines⁴²). The combined results are used in a 3-marker model with a likelihood score (Table 4) improved by adding APOE4 allele count (assessed by analysis of apolipoprotein E [apoE] proteotypes; see “APOE genotypes and apoE proteotypes”). Performance is similar to the 2-marker model but with slightly higher accuracy (85%-87%) and fewer indeterminate results (10% vs 15%, 7% in >4,000 patient specimens submitted for the 3 component tests).⁴⁷

Beta-amyloid 42/40 ratio

Quest AD-Detect^®, Beta-Amyloid 42/40 Ratio, Plasma test (test code 11786) is an LC-MS/MS assay that helps initially assess whether AD pathology is present in the context of dementia. Measuring Aβ42/40 provides greater assessment accuracy than measuring Aβ42 alone because it controls for the natural variance of β-amyloid production between patients.²²

LC-MS/MS is a preferred method for measuring Aβ42/40. A cross-sectional study assessed the diagnostic accuracy of 8 ultrasensitive Aβ42/40 plasma methods (4 immunoassays and 4 MS assays) among 286 individuals. This study included 104 with and 182 without MCI with presence/absence of Aβ pathology previously determined using CSF Aβ42/40 by MS.24 Of these assays, an LC-MS/MS analysis-based assay demonstrated higher accuracy (AUC-ROC ≥0.84) than other assays that were standardized for high throughput (AUC-ROCs ranging from 0.60 to 0.80; P<.05).²⁴

The Quest test achieves ≥91% sensitivity and ≥74% specificity (AUC-ROC ≥0.84) at a single cutpoint of Aβ42/40=0.160.⁵⁰ The test was developed using LC-MS/MS data for specimens with associated data for amyloid PET imaging, diagnosis, and demographics.

High sensitivity and negative predictive value (NPV) for Aβ-PET positivity were observed at a prevalence of 40% of PET positivity in the cohort, consistent with the clinical performance of other plasma LC-MS/MS assays, but with greater separation between Aβ42/40 values for individuals with positive vs negative Aβ-PET results. At the study prevalence of Aβ-PET positivity, a cutpoint (0.170) was identified with 99% NPV, which could help predict that AD is likely not the cause of a patient’s cognitive impairment and help reduce PET evaluations by about 40%.⁵⁰

P-Tau181 and p-tau217

The Quest AD-Detect^® Phosphorylated tau181 (p-tau181), Plasma (test code 13690) and Quest AD-Detect^® Phosphorylated tau217 (p-tau217), Plasma (test code 13825) tests help assess whether MCI or dementia is consistent with the presence of AD pathology in patients ≥55 years. Plasma levels of p-tau181 and p-tau217 are elevated in the MCI and dementia stages of AD and are associated with the presence of Aβ and tau pathology.^25-29 Both are considered Core 1 biomarkers in AD assessment.²⁰

Of the 2 biomarkers, p-tau181 has been more extensively studied and the assay has been FDA cleared.⁵¹ A meta-analysis of 18 studies demonstrated that measured p-tau181 blood levels reliably differentiate between Aβ-PET–positive vs Aβ-PET–negative individuals.⁵² However, plasma p-tau217 has a particularly strong association with Aβ pathology, which increases early in AD progression.⁵³ In addition, p-tau217 typically outperforms p-tau181 in head-to-head comparisons according to the Alzheimer’s Association Workgroup.²⁰ Increased baseline plasma p-tau217 can predict progression of cognitive impairment, and longitudinal increases correlate with declining cognition.^54,55 Increased plasma p-tau181 levels have been shown to predict clinical progression to more severe cognitive impairment better than measures of amyloid pathology.^56,57

Plasma levels of both p-tau181 and p-tau217 are dynamic and highly correlative biomarkers of AD pathology as assessed by PET and CSF.^26,56-58 Levels increase longitudinally in Aβ-positive, but not Aβ-negative, individuals.^27,54 Levels also increase across advancing stages of tau pathology.^27,29 The markers can differentiate AD from many other neurodegenerative diseases, such as FTD, progressive supranuclear palsy, vascular dementia, and PD.^25,28,56,58 In addition, both plasma p-tau217 and p-tau181 are leading markers in monitoring AD pathology in response to anti-amyloid therapies targeting protofibrils (lecanemab), insoluble fibrils (donanemab), and plaques (aducanumab).³⁰

The following results are consistent with MCI and dementia caused by AD pathology, and follow-up assessment using PET or CSF analysis should be considered:

• P-tau217 levels⁴⁷ >0.23 pg/mL
• P-tau181 levels⁵¹ >0.72 pg/mL

Importantly, both plasma p-tau181 and p-tau217 can also increase in those with chronic kidney disease or a history of myocardial infarction or stroke.²⁶ Plasma p-tau181 can also be increased in patients with amyotrophic lateral sclerosis (ALS) and is associated with lower motor neuron disease.⁵⁹ Consequently, the results of these assays should be considered in conjunction with the findings from medical and family history, nutritional deficiency biomarkers (see “Dementia-like symptoms” section below), neuroimaging, and physical, neurological, and neuropsychological examination.

Neurofilament light chain (NfL)

The Neurofilament Light Chain (NfL), Plasma test (test code 13979) provides an accessible, minimally invasive option for assessing neurodegeneration or neuronal injury; NfL is considered a biomarker of nonspecific processes resulting from AD pathology.²⁰

NfL is a structural protein expressed exclusively in neurons. Upon neuronal degeneration or injury, NfL is released into extracellular space, resulting in increased concentrations in CSF and blood.⁶⁰ Though NfL levels are lower in blood than in CSF, advancements in assay development have enabled the use of blood-based NfL as a more accessible biomarker of neuronal injury and/or neurodegeneration.⁶⁰ NfL is generally useful for informing prognosis and monitoring disease progression, most commonly for multiple sclerosis but also for neurodegeneration in AD. Because elevated levels are associated with AD progression, NfL has emerged as a candidate biomarker in the biological definition of AD.^31,32

Although lacking specificity for any given disease, NfL measurements can be useful for certain differential diagnoses. Levels are elevated in FTD vs psychiatric disorders, ALS vs motor neuron disease mimics, and atypical parkinsonian syndromes vs PD.^60-64

NfL levels can be influenced by many factors, including age, body mass index, kidney disease, and a history of diabetes or cardiovascular conditions.⁶⁰

Accordingly, interpretation of plasma NfL test results depends on the NfL levels relative to age-specific reference ranges and the clinical context:

• Normal plasma NfL levels are generally consistent with MCI or dementia not caused by neuronal injury or neurodegeneration.
• Higher-than-normal NfL levels are consistent with MCI or dementia caused by clinically relevant neuronal injury or neurodegeneration. In patients with a known neurologic condition, elevated NfL or increased levels from an established, patient-specific baseline may indicate poorer prognosis and/or disease progression.⁶⁰ In patients with diagnostic uncertainty, elevated NfL may support differential diagnosis of a suspected neurologic condition.⁶⁰

The results of this assay should be considered in conjunction with the findings from medical and family history, neuroimaging, and physical, neurological, and neuropsychological examination.

Glial fibrillary acidic protein (GFAP)

The Glial Fibrillary Acidic Protein (GFAP), Plasma test (test code 14319) is a tool for assessing AD progression and anti-amyloid therapy treatment response; GFAP is considered another biomarker of nonspecific processes involved in and resulting from AD pathology.²⁰

GFAP is an intermediate filament protein found only in astrocytes and is a biomarker of neuronal injury (reactive astrogliosis) when released from the central nervous system.^30,65 In contrast to most other AD biomarkers, GFAP performs better in plasma than CSF for distinguishing Aβ-PET–positive vs Aβ-PET–negative individuals.³⁰ A meta-analysis of over 2,400 patients on the AD continuum (from cognitively unimpaired to MCI to AD) found that plasma GFAP concentrations increased with disease progression.³³ Furthermore, GFAP has been found to be a consistently relevant biomarker in clinical trials for anti-amyloid therapies: compared with baseline, treated patients had a 10% to 20% decrease in plasma GFAP levels compared with a 10% to 15% increase in patients receiving a placebo.³⁰ The placebo-drug difference and effect size was similar to those observed for clinical trials that used plasma p-tau181.³⁰

Interpretation of plasma GFAP test results depends on the clinical context:

• Normal plasma GFAP levels indicate that MCI or dementia does not involve active astrogliosis.
• Higher-than-normal plasma GFAP levels indicate that MCI or dementia involves active astrogliosis.

GFAP may also be elevated in mild traumatic brain injury, stroke, multiple sclerosis, neuromyelitis optica spectrum disorder, cardiovascular disease (cardiac arrest and atrial fibrillation), infection (West Nile and SARS-CoV-2), and sepsis-related encephalopathy.⁶⁵ The results of this assay should be considered in conjunction with the findings from medical and family history, neuroimaging, and physical, neurological, and neuropsychological examination.

APOE genotypes and apolipoprotein E proteotypes

The Quest AD-Detect^® Apolipoprotein E (ApoE) Isoform, Plasma test (test code 12563) is a tool for AD risk assessment. ApoE phenotypes are identified based on the detection of proteoform-specific peptide(s). Results of phenotyping are 100% concordant with APOE genotyping results.⁴⁰

ApoE is the primary brain apolipoprotein and the most studied blood biomarker for AD. ApoE has 3 common proteoforms, E2, E3, and E4, which are encoded by the APOE geneε2, ε3, or ε4 alleles, respectively. The 6 combinations of APOE alleles are ε2/ε2, ε2/ε3, ε2/ε4, ε3/ε3, ε3/ε4, and ε4/ε4; frequencies vary among racial and ethnic groups.⁵

Specific APOE alleles are associated with increased AD risk: the presence of an ε4 allele confers higher risk of AD compared with ε3 (the most common allele), whereas the ε2 allele confers a protective effect.³⁷ However, patient sex, environment, race, and ethnicity as well as the presence of other risk alleles also contribute to AD risk associated with the APOE genotype.^36,37

Establishing apoE proteoform status (APOE genotype) is recommended for patients with early AD who are candidates for anti-amyloid monoclonal antibody therapy.⁶⁶ Lecanemab, donanemab, and aducanumab are 3 monoclonal antibodies that target aggregated forms of beta-amyloid and have been approved by the FDA⁵—although aducanumab was withdrawn from the market as a “business decision” not related to safety or efficacy.⁶⁷ Individuals receiving these therapies are at risk of developing amyloid-related imaging abnormalities (ARIAs) detected as edema (ARIA-E) or hemorrhagic changes (ARIA-H); ARIAs are mostly asymptomatic but occasionally may result in life-threatening symptoms.⁶⁶ Risk of ARIA is increased in APOE ε4 carriers, especially APOE ε4 homozygotes.⁶⁸ Therefore, testing for APOE ε4 status, either with genotyping or apoE phenotyping, helps inform the discussion of treatment risk between the prescriber and patient.^66,68

The presence of the ε4 allele is not necessary or sufficient to cause AD. People who are ε4 allele carriers may never develop AD, and over 30% of patients with AD are not carriers of the ε4 allele.⁵ Therefore, routine testing for apoE proteoform status using this test or APOE genotype (test code 10642) alone is not recommended to predict AD risk.^69,70 These tests can be ordered alongside other AD-Detect tests.

Compared to the most common combination of apoE proteoforms (E3/E3 phenotype)

• E2/E2 and E2/E3 phenotypes (ε2/ε2 and ε2/ε3 genotypes) suggest lower-than-average AD risk
• E2/E4 and E3/E4 phenotypes (ε2/ε4 and ε3/ε4 genotypes) suggest higher-than-average AD risk

Compared with all other phenotypes, an E4/E4 (homozygous) phenotype (ε4/ ε4 genotype) may confer the highest AD risk; homozygosity has been suggested to be a genetically distinct form of AD.⁷¹

Risk for ARIA is higher in individuals who are receiving anti-amyloid monoclonal antibody therapy and have the apoE4 proteoform-specific peptide, especially the E4/E4 phenotype.⁶⁶

Phenotyping results should be considered, knowing that patient sex, environment, race, and ethnicity as well as the presence of other risk alleles also contribute to AD risk associated with the apoE phenotype/APOE genotype.^36,37

CSF–based tests

Beta-amyloid 42/40 ratio and apoE

The Beta-Amyloid 42/40 Ratio and Apolipoprotein E (ApoE) Isoform Panel, CSF (test code 94628) combines results for these biomarkers into a model to assess the risk for AD in adults exhibiting signs of MCI or dementia. Panel components can be ordered separately: Beta-Amyloid 42/40 Ratio, CSF (test code 94627) or Apolipoprotein E (APOE) Isoform, CSF (test code 94626). If only phenotyping or genotyping is required, consider plasma tests for less invasive specimen collection.

An Aβ42/40 ratio cutoff of <0.16 had a clinical sensitivity of 78% for distinguishing patients with AD from those with non-AD dementia (clinical specificity, 91%) or normal cognition (clinical specificity, 81%). The Aβ42/40 ratio decreased significantly (P <.001) with increasing dose of APOE4 alleles.⁷²

Based on these data, a model was developed (test code 94628) to examine the biomarkers separately and combined into a risk score to determine the risk that a patient has AD (Table 5).

Table 5. CSF Biomarkers and Risk of Having AD [return to contents]

P-tau/total-tau/Aβ42

ADmark^® Phospho-Tau/Total-Tau/A Beta 42, Interp, CSF (Symptomatic) (test code 92433) uses immunoassay results from 3 FDA-cleared tests to compute total tau/Aβ42 and p-tau181/Aβ42 ratios using the CSF concentrations (pg/mL) of Aβ42, total tau, and p-tau181. This test is for adult patients ≥55 years being evaluated for AD and other causes of cognitive impairment.

The following ratios are consistent with MCI and dementia caused by AD pathology:

• Total-tau181/Aβ42 >0.28
• P-tau181/Aβ42 >0.023

Similar ratios demonstrated 90% agreement with PET imaging results for patients with mild or subjective cognitive symptoms⁷³ and could distinguish AD from FTD with 89% sensitivity and 75% to 77% specificity.⁷⁴

Non-AD dementia and dementia-like symptoms

Non-AD dementia and dementia-like symptoms can have several etiologies, some of which are treatable and reversible. US and European guidelines recommend laboratory testing to assess these conditions as well as comorbidities that often accompany dementia.^9,10 Quest offers Dementia Panel, Secondary Causes (test code 34541) that includes recommended testing for folate, vitamin B12, thyroid-stimulating hormone, calcium, glucose, complete blood cell count, and renal and liver function tests.

Infectious disease testing has also been recommended for atypical presentation or clinical features suggestive of HIV infection (test code 91431) or syphilis (test code 36126).¹⁰ An algorithmic approach has been published for incorporating most of these tests in evaluation of suspected dementia.⁸ Tests are also available for other possible infectious disease causes of dementia, including prion disease (test code 37989),¹¹ as described in Table 3.