APOE Epsilon4 Allele Research Articles

Magnetic Resonance Imaging Traits in Siblings Discordant for Alzheimer Disease

Magnetic resonance imaging (MRI) can aid clinical assessment of brain changes potentially correlated with Alzheimer disease (AD). MRI traits may improve our ability to identify genes associated with AD-outcomes. We evaluated semi-quantitative MRI measures as endophenotypes for genetic studies by assessing their association with AD in families from the Multi-Institutional Research in Alzheimer Genetic Epidemiology (MIRAGE) Study. Discordant siblings from multiple ethnicities were ascertained through a single affected proband. Semi-quantitative MRI measures were obtained for each individual. The association between continuous/ordinal MRI traits and AD were analyzed using generalized estimating equations. Medical history and Apolipoprotein E (APOE)epsilon4 status were evaluated as potential confounders. Comparisons of 214 affected and 234 unaffected subjects from 229 sibships revealed that general cerebral atrophy, white matter hyperintensities (WMH), and mediotemporal atrophy differed significantly between groups (each at P < .0001) and varied by ethnicity. Age at MRI and duration of AD confounded all associations between AD and MRI traits. Among unaffected sibs, the presence of at least one APOEepsilon4 allele and MRI infarction was associated with more WMH after adjusting for age at MRI. The strong association between MRI traits and AD suggests that MRI traits may be informative endophenotypes for basic and clinical studies of AD. In particular, WMH may be a marker of vascular disease that contributes to AD pathogenesis.

Genotype‐adjusted familial correlation analysis using three generalized estimating equations

We analyze familial correlation of a memory score from Caribbean Hispanic families that have multiple family members affected with Alzheimer's disease, adjusting for having at least one APOE-epsilon4 allele, as well as other confounders. To enhance the efficiency of correlation model, this paper proposes an alternative approach for three generalized estimating equations proposed by Yan and Fine (Statist. Med. 2004; 23:859-874). The efficiency of correlation model is evaluated through the asymptotic relative efficiency computation and simulations.

CSF β-amyloid 1–42 and tau in Tunisian patients with Alzheimer's disease: The effect of APOE ɛ4 allele

Alzheimer's disease (AD) is the leading cause of dementia. Currently, no definitive diagnostic test for AD exists. An accurate, convenient and objective test to detect AD is urgently needed for efficient drug development and effective clinical use of emerging therapies. The aim of the present work is to investigate the usefulness of cerebrospinal fluid (CSF) beta-amyloid protein (Abeta1-42) and total tau protein (t-tau) analyses in the diagnosis of AD and whether apolipoprotein E (ApoE) epsilon4 allele is a factor for AD affecting Tunisian people. Abeta1-42 and t-tau levels were measured in CSF from AD patients (n=73), non-Alzheimer dementia (nAD, n=35) and healthy controls (HC, n=38) by sandwich enzyme-linked immunosorbent assay. Abeta1-42 levels were decreased and t-tau increased in AD patients. The combination of Abeta1-42 and t-tau at baseline yielded a sensitivity of 87.4% for detection of AD. The specificities were 97.3% for controls and 82.7% for other dementia. The ApoE epsilon4 allele frequency (29.5%) was significantly higher in the AD patients than in the nAD patients (17.1%) or in the control groups (9.5%). AD patients carrying ApoE epsilon4 allele had lower Abeta1-42 (p<0.001) levels than those without a epsilon4 allele. The combination of t-tau and Abeta1-42 is a robust and reliable assay that may be useful in discriminating cases at risk for AD such as ApoE epsilon4 allele carriers from nAD patients or from age-matched control subjects.

Association analysis of dynamin‐binding protein (DNMBP) on chromosome 10q with late onset Alzheimer's disease in a large caucasian UK sample

A recent scan of single nucleotide polymorphisms (SNPs) in the region 40-107 Mb on chromosome 10q in a large Japanese case-control cohort identified six SNPs in or near the dynamin-binding protein gene (DNMBP) that were associated with late onset Alzheimer's disease (LOAD) in individuals lacking the APOE epsilon4 allele [Kuwano et al. (2006); Hum Mol Genet 15:2170-2182]. We genotyped these six SNPs in 1,212 unrelated Caucasian patients of UK origin with LOAD and 1,389 ethnically, gender and age matched control subjects. We did not observe a statistically significant association with the risk of LOAD for any of the six SNPs in the sample as a whole. When stratifying the sample by APOE one SNP (intergenic SNP rs11190302) was associated with LOAD in individuals lacking the epsilon4 allele (genotypic P = 0.027, allelic P = 0.066). However this association was in the opposite direction to that detected in the Japanese population. It remains to be determined whether DNMBP is associated with LOAD.

Total serum cholesterol levels and Alzheimer's dementia in patients with down syndrome

The risk for dementia in Alzheimer's disease (DAD) in adults with Down syndrome (DS) is higher than in the general adult population. Hypercholesterolaemia has been reported as a risk factor for DAD in the general population. This study investigated the role of serum cholesterol levels in the onset of DAD in the DS population. This study investigated total serum cholesterol levels in 179 DS persons (with and without DAD). The possible association between Apolipoprotein E and amyloid beta1-40 and beta1-42 levels was also investigated. No statistically significant association was found between total serum cholesterol levels and dementia in AD or with amyloid beta levels. However for DS adults with an apoE epsilon4 allele significantly higher serum cholesterol levels were found. Hypercholesterolaemia is not a risk factor for DAD in persons with DS. However, DS persons with an apoE epsilon4 allele are susceptible to high serum cholesterol. Such individuals should be screened on a regular basis.

KIBRA gene variants are associated with episodic memory performance in subjective memory complaints

The KIBRA gene encodes a cytoplasmatic protein, a member of the signal transduction protein family, expressed mainly in the brain. Recent studies have implicated the involvement of a genetic variation in the KIBRA gene (T allele) in human memory in normal subjects and in the risk of developing Alzheimer's disease (AD). We report here the distribution of the KIBRA genetic variant and the Apolipoprotein E (ApoE) epsilon4 allele and their association with neuropsychological measures in older adults reporting problems with everyday memory (subjective memory complaints, SMC). We found that SMC subjects with the CT/TT genotype performed more poorly than those with the CC genotype on long-term memory tests. Thus, in our opinion, these data suggest that the KIBRA genotype could affect memory performance in a different way in those that complain of memory deficits compared to those that do not.

Cortical α-synuclein load is associated with amyloid-β plaque burden in a subset of Parkinson’s disease patients

Amyloid-beta (Abeta) peptide pathology in Alzheimer's disease (AD) comprises extracellular plaques and cerebral amyloid angiopathy (CAA). In Parkinson's disease (PD), alpha-synuclein forms intraneuronal Lewy bodies (LBs), and cortical LBs are thought to play a major role in cognitive decline designated as PD with dementia. As there is increasing evidence that Abeta may also promote alpha-synuclein fibrillization, we assessed the relationship between LB pathology and Abeta deposition in 40 cases of PD and 20 age-matched controls. In five cortical areas, we established the severity of Abeta plaque load using an approach similar to that recommended by CERAD in AD. LB densities were determined using a morphometric approach. CAA was graded using previously described scales. The APOE genotype was established in 38 PD and 19 control cases. We have found that the overall Abeta plaque burden and, in particular, the diffuse plaque load shows a statistically significant 'large' correlation with the overall cortical LB burden. The strength of this correlation further increases in PD cases (about 50% of the cases) with moderate to high Abeta plaque load. The APOE epsilon4 allele is over-represented in this subgroup. Our data indicate a strong association between pathologically identifiable Abeta plaque burden and alpha-synuclein load in cerebral cortex and provide indirect evidence that Abeta pathology is likely to be an important factor contributing to cognitive decline in a subgroup of PD patients.

Traumatic Brain Injury, Apolipoprotein E-ϵ4, and Cognition in Older Adults: A Two-Year Longitudinal Study

Patients with mild-to-moderate traumatic brain injury (TBI) (N=69) were compared with age-, gender-, and education-matched healthy control group subjects (N=79) on performance of neuropsychological tests at one and 2 years following injury, and informant-rated functional abilities. All subjects were assessed for the presence of the Apolipoprotein E-epsilon4 (APOE-epsilon4) allele and rated for "mild cognitive impairment" (MCI) or dementia. Traumatic brain injury patients were no different from the comparison group on measures of cognition or functional impairment. Traumatic brain injury was not associated with higher rates of amnestic mild cognitive impairment or dementia, and APOE-epsilon4 was not associated with cognition.

APOE genotype predicts depression in women with Alzheimer's disease: a retrospective study.

The association between the APOE epsilon4 allele and depression was investigated in a retrospective study of 323 AD patients. Patients were divided into demographically comparable groups based on the presence or absence of depression. Results showed that the frequency of APOE epsilon4 allele was significantly higher in the depressed vs non-depressed AD patients (72% and 58%, respectively), and an interaction revealed that women possessing the APOE epsilon4 allele were almost four times more likely to be depressed than those without the epsilon4 allele. Results are consistent with recent suggestions that the APOE epsilon4 genotype may be over-represented among depressed women with AD and highlight the need for additional research investigating the links between APOE genotype, mood, and gender.

Interaction of angiotensin-converting enzyme and apolipoprotein E gene polymorphisms in ischemic stroke involving large-vessel disease

A relationship between apolipoprotein E (Apo E) genotype and angiotensin-converting enzyme (ACE) insertion-deletion (Ins-Del) mutation and stroke was suggested. We investigated the association of Apo E4 and ACE Ins/Del genotypes with stroke risk and changes in serum lipids in 228 consecutive Tunisian stroke patients, and 323 age-and gender-matched controls. Comparable frequencies of ACE Ins/Del alleles were seen between patients and controls. The prevalence of Apo epsilon3 allele and Apo E3/E3 were lower (P < 0.001), while the frequency of Apo epsilon4 allele and epsilon4-containing genotypes (E3/E4 and E4/E4) were elevated (P < 0.001) among patients. Higher proportion of Apo E4-carrying + ACE Del/Del positive cases were seen in young (<50 years) patients (P = 0.012), and was associated with large vessel stroke (P = 0.035). Mean serum cholesterol, LDL, HDL, and triglycerides were comparable between E4-containing and no E4-containing and ACE Del/Del-positive patients. Apo E4 and ACE Del/Del genotype combination substantially increase stroke risk, supporting the notion that interactions of multiple gene variants influence stroke pathogenesis.

Serotonin Transporter Gene Polymorphism and BPSD in Mild Alzheimer's Disease

The purpose of the present study was to confirm an association of functional polymorphism within the serotonin transporter (5-HTT) gene with Alzheimer's disease (AD) and behavioral and psychological symptoms of dementia (BPSD) in mild AD. Apolipoprotein E (ApoE) gene polymorphism and 2 types of functional polymorphism in the 5-HTT gene, 5-HTT-linked polymorphic region (5-HTTLPR) and a 5-HTT variable number of tandem repeats sequence (5-HTTVNTR) were analyzed longitudinally in outpatients with mild AD to find out whether there was a relation between any such polymorphisms and the occurrence of BPSD. No significant differences in genotype distribution or allele frequencies were identified for 5-HTTLPR or 5-HTTVNTR between AD patients and age- and sex-matched non-demented controls regardless of ApoE epsilon4 allele. No significant differences were noted in 5-HTTLPR genotype or allele distributions between AD patients with or without BPSD. However, significant associations were observed between presence of 5-HTTVNTR allele 10 and BPSD or aggressiveness. This difference was independent of the presence of the ApoE epsilon4 allele. As a result, 5-HTT polymorphisms are unlikely to play any substantial role in susceptibility to AD. Conversely, 5-HTTVNTR influences the risk of developing BPSD or aggressiveness and genetic variations in the 5-HTT gene may be involved in the development of symptomatology for mild AD.

Somatostatin genetic variants modify the risk for Alzheimer's disease among Finnish patients

The levels of somatostatin are consistently decreased in the brain and cerebrospinal fluid of Alzheimer's disease (AD) patients. The somatostatin gene is located on chromosome 3q27.3 close to an association region identified in late-onset AD patients originating from Finland. Since somatostatin is a good candidate on both positional and functional grounds, we studied whether single nucleotide polymorphisms (SNPs) in the somatostatin gene were associated with AD in the Finnish population. We genotyped three SNPs within this gene in Finnish AD patients (n = 424) and non-demented controls (n = 466). AD patients were compared with non-demented control subjects using single-locus and haplotype approaches. In the whole study group, the age, sex and APOE adjusted OR for the risk of AD in C-allele carriers of the SNP rs4988 514 was 1.42 (p <0.05). Interestingly, in APOE epsilon4-allele carriers, the age and sex adjusted OR for the risk of AD in C-allele carriers of the rs4988 514 increased to 2.05 (p <0.01). Additionally, SNP rs4988514 may interact with the APOE epsilon4-allele to increase the risk of AD. Assessment of individual haplotype distributions revealed a 2-fold overrepresentation of the TCG haplotype of SNPs rs3864101, rs4988 514 and rs7624 906 in the AD APOE epsilon4-allele group (p <0.01). Conversely, a major haplotype TTG was significantly underrepresented among all the AD patients as well as APOE epsilon4-allele carrying AD patients. Thus the major haplotype TTG of somatostatin may have a protective effect against AD. This first genetic association study between somatostatin and AD indicates that genetic variations in the somatostatin gene may modify the risk for AD among Finnish AD subjects.

The association between delirium and the apolipoprotein E ε4 allele in the elderly

As not all patients with similar risk factors and eliciting conditions develop delirium; it may be hypothesized that genetic variation may play a role in the risk of delirium. On the basis of the relationship between dementia, respectively reduced cholinergic activity, and the APOE epsilon4-allele, and the similarities between dementia and delirium in reduced cholinergic activity, the APOE epsilon4-allele is a rational candidate-gene for delirium. This study examined the association between APOE epsilon4-allele and delirium in elderly patients. Acutely admitted patients to the Department of Medicine of 65 years and over were included during a 27-month time period. Delirium was scored by the confusion assessment method. Cognitive impairment was diagnosed by Mini Mental State Examination and informant questionnaire on cognitive decline. Genotyping was done with matrix-assisted laser-desorption/ionization time-of flight mass spectrometry. Of 415 included patients, a random sample of 264 patients was genotyped for APOE. The patients who met the criteria for delirium (35%) were significantly older and more frequently had preexisting functional and cognitive impairment. APOE genotype was borderline significantly associated with cognitive impairment in patients below 75 years (P=0.057). The odds ratio for carriers of an APOE epsilon4-allele compared with patients without an APOE epsilon4-allele for developing delirium was 1.17 (95% confidence interval (CI): 0.49-2.78) in the cognitively intact patients and 0.42 (95% CI: 0.14-1.30) in the cognitively impaired patients. No relation existed between the total number of APOE epsilon4-alleles and the different delirium subtypes (P=0.12). We found no convincing evidence that carriers of the APOE epsilon4-allele have a higher risk of delirium.

Glutathione S-Transferase P1 and T1 Gene Polymorphisms Predict Longitudinal Course and Age at Onset of Alzheimer Disease

Oxidative stress has been suggested as a contributor of Alzheimer disease (AD) neurodegeneration, particularly in those patients with late-onset AD (LOAD). Therefore, the authors studied the effect of glutathione S-transferase (GST) P1-M1-T1 gene polymorphisms and their interactions with the apolipoprotein E (ApoE) epsilon4 allelic variant on the three-year longitudinal course of AD. Global cognitive level as measured by the Mini-Mental State Exam, basic activities of daily living (BADLs) as measured by the Physical Self-Maintenance Scale, and behavior as measured by the Neuropsychiatric Inventory, were assessed at baseline and after 1, 2, and 3 years in a sample of 99 LOAD patients. These subjects were drug naive and had undergone the first clinical examination for the diagnosis of AD. A multiple regression analysis indicated that the presence of ApoE epsilon4 allelic variant or GSTT1 null phenotype predicted the faster age at onset of the illness (F = 5.76, df = 2, 96, p = 0.0043). Carriers of GSTP1 *C allelic variant had a faster decline in cognitive functions (repeated measures analysis of variance [ANOVA]: F = 4.00, df = 3, 285, p = 0.008) and in BADLs (repeated measures ANOVA: F = 5.27, df = 3, 285, p = 0.001). This faster decline was independent from ApoE epsilon4 allele possession. No effect of GST P1-M1-T1 polymorphisms was found on behavioral symptom severity. These data are in line with the hypotheses that oxidative damage is a prominent feature in the clinical progression and the age at onset of LOAD.

ApoE4 Allele Is Associated With Late-Life Depression: A Population-Based Study

This study aimed to clarify the relationship between the epsilon4 allele and late-life depression, taking into account lipid profile, vascular diseases, and sociodemographics. Using a multilevel stratified random sampling strategy, a total of 500 subjects aged 65 to 74 years were recruited for this cross-sectional study from the official household records of an entire county in southern Taiwan. Depressive symptoms were assessed by the Taiwanese Depression Questionnaire (TDQ). Cognitive function was assessed by the Short Portable Mental Status Questionnaire. Blood samples were collected for the determination of the apolipoprotein E (ApoE) polymorphism and the lipid profile. A total of 283 subjects (58.7% male, with a mean age of 69.2 +/- 2.7 years) completed all questionnaires and collection of blood samples. Using the chi(2) test, the overall difference for frequency of the presence of the epsilon4 allele was significant among the severe group (TDQ score >18), moderate group (TDQ score 9-18), and mild group (TDQ score <9). The proportion of history of heart disease was significantly higher in the severe group than in the mild or moderate group. Kruskal-Wallis statistics revealed that the mean total and low-density lipoprotein cholesterol levels were significantly higher in the severe group than in the moderate or mild group. With our two-level four-class latent class regression model, the presence of the epsilon4 allele was significantly associated with the severely depressed group as compared to the nondepressed group, adjusting for vascular diseases and lipid profile. The ApoE epsilon4 allele may be correlated with severe depression in the elderly through ways other than the "vascular depression" hypothesis.

Relation between smoking and risk of dementia and Alzheimer disease

Previous studies relating smoking with the risk of dementia have been inconsistent and limited in their validity by short follow-up times, large intervals between baseline and follow-up assessments, and unspecific determination of dementia diagnosis. We re-assessed after longer follow-up time in the large population-based cohort of the Rotterdam Study whether smoking habits and pack-years of smoking are associated with the risk of dementia, Alzheimer disease (AD), and vascular dementia (VaD). Prospective population-based cohort study in 6,868 participants, 55 years or older and free of dementia at baseline. First, Cox proportional hazard models were used to relate smoking status at baseline with the risks of incident dementia, VaD, and AD, using never smokers as the reference category in all analyses. Then Cox proportional hazard models were used to relate pack-years of smoking with the risks of incident dementia, VaD, and AD. To explore the impact of the APOEepsilon4 allele, sex, and age on the association between smoking status and dementia, we repeated all analyses stratifying, in separate models, by APOEepsilon4 genotype, sex, and median of age. After a mean follow-up time of 7.1 years, current smoking at baseline was associated with an increased risk of dementia (HR 1.47, 95% CI 1.18 to 1.86) and AD (HR 1.56, 95% CI 1.21 to 2.02). This increase in disease risk was restricted to persons without the APOEepsilon4 allele. There was no association between current smoking and risk of VaD, and there was no association between past smoking and risk of dementia, AD, or VaD. Current smoking increases the risk of dementia. This effect is more pronounced in persons without the APOEepsilon4 allele than APOEepsilon4 carriers.

Genetic association of CTNNA3 with late-onset Alzheimer's disease in females

Alzheimer's disease (AD), the most common form of dementia in the elderly, was found to exhibit a trend toward a higher risk in females than in males through epidemiological studies. Therefore, we hypothesized that gender-related genetic risks could exist. To reveal the ones for late-onset AD (LOAD), we extended our previous genetic work on chromosome 10q (genomic region, 60-107 Mb), and single nucleotide polymorphism (SNP)-based genetic association analyses were performed on the same chromosomal region, where the existence of genetic risk factors for plasma Abeta42 elevation in LOAD was implied on a linkage analysis. Two-step screening of 1140 SNPs was carried out using a total of 1408 subjects with the APOE-epsilon3*3 genotype: we first genotyped an exploratory sample set (LOAD, 363; control, 337), and then genotyped some associated SNPs in a validation sample set (LOAD, 336; control, 372). Seven SNPs, spanning about 38 kb, in intron 9 of CTNNA3 were found to show multiple-hit association with LOAD in females, and exhibited more significant association on Mantel-Haenszel test (allelic P-values(MH-F) = 0.000005945-0.0007658). Multiple logistic regression analysis of a total of 2762 subjects (LOAD, 1313; controls, 1449) demonstrated that one of the seven SNPs directly interacted with the female gender, but not with the male gender. Furthermore, we found that this SNP exhibited no interaction with the APOE-epsilon4 allele. Our data suggest that CTNNA3 may affect LOAD through a female-specific mechanism independent of the APOE-epsilon4 allele.

The association of the regulatory region of the presenilin-2 gene with Alzheimer's disease in the Northern Han Chinese population

Presenilin-2 is one of the causative genes for familial Alzheimer's disease (FAD). Polymorphism of the promoter region of the presenilin-2 gene (PSEN2) has recently been reported in a Russian population to be associated with sporadic Alzheimer's disease (SAD). The purpose of this case-control study was to determine whether SAD is associated with the PSEN2 gene polymorphism in a Chinese population. We examined PSEN2 and APOE genotypes in 200 SAD patients and an equal number of age- and sex-matched controls from the same community, using the PCR-RFLP method. Allelic and genotypic distributions were performed using the Pearson Chi-square test for homogeneity. The interactions between variables were examined by logistic regression. The results revealed no significant differences in the frequency of the +A/-A polymorphism between AD and controls (chi(2)=3.857, p=0.145). However, in the subgroup of APOE epsilon4 non-carriers, there were significant differences in the distributions of both alleles (chi(2)=6.095, p=0.047) and genotypes (chi(2)=4.433, p=0.035) of the PSEN2 promoter in AD compared with controls. In APOE epsilon4 non-carrier group, with +A/+A as a reference, the -A/-A genotype was associated with a 4.657-fold increased risk for AD (chi(2)=5.783, p=0.016, OR=4.657, 95% CI=1.195-18.152). Using logistic analysis, there were no statistical interactions between PSEN2 and APOE genotypes, or between PSEN2 genotypes and age of onset. It is concluded that in the Northern Han Chinese population, the +A/-A polymorphism of the PSEN2 promoter is a moderate genetic risk factor for developing SAD, independent of the APOE epsilon4 allele.

APOE ε4 allele, cognitive dysfunction, and obstructive sleep apnea in children

Obstructive sleep apnea (OSA) in children is associated with severity-dependent changes in neurocognitive functioning. However, the severity of OSA accounts for only approximately 40% of the variance in cognitive performance. Thus, genetic determinants of individual susceptibility may also contribute to the morbidity of OSA. Considering the unique susceptibility of apolipoprotein E (ApoE) knock-out mice to an experimental model of OSA, we examined whether the APOE epsilon4 allele contributes to increased neurocognitive morbidity in pediatric OSA. Consecutive habitually snoring and nonsnoring 5- to 7-year-old children underwent overnight polysomnography, neurocognitive testing, and a blood draw the next morning. Children were divided into OSA or no OSA, and OSA children were further subdivided into those with > or =2 abnormal cognitive subtest scores and those with normal cognitive scores. The presence of the APOE epsilon4 allele was determined from blood genomic DNA. Among all children without OSA, APOE epsilon4 was present in 3 of 199 children, whereas in those with OSA, APOE epsilon4 was found in 16 of 146 children (p < 0.0002). Furthermore, 16 of 74 children with OSA and cognitive scores <85% had the APOE epsilon4 allele compared with 3 of 72 children with OSA with abnormal cognitive scores (p < 0.002). APOE epsilon4 allele is more frequent in children with obstructive sleep apnea and particularly in those who develop neurocognitive deficits, suggesting that the APOE epsilon4 allele is associated with not only increased odds of having sleep-disordered breathing, but also with an increased risk for neurocognitive dysfunction.

Does NSAID use modify cognitive trajectories in the elderly?

Epidemiologic studies have suggested that nonsteroidal anti-inflammatory drugs (NSAIDs) may be useful for the prevention of Alzheimer disease (AD). By contrast, clinical trials have not supported NSAID use to delay or treat AD. Few studies have evaluated cognitive trajectories of NSAID users over time. Residents of Cache County, UT, aged 65 or older on January 1, 1995, were invited to participate in the study. At baseline, participants provided a detailed inventory of their medications and completed a revised Modified Mini-Mental State Examination (3MS). Participants (n = 3,383) who were cognitively normal at baseline were re-examined after 3 and 8 years. The association between NSAID use and 3MS scores over time was estimated using random effects modeling. Associations depended upon when NSAIDs were started and APOE genotype. In participants who started NSAID use prior to age 65, those with no APOE epsilon4 alleles performed similarly to nonusers (a difference of 0.10 points per year; p = 0.19), while those with one or more epsilon4 allele(s) showed more protection (0.40 points per year; p = 0.0005). Among participants who first used NSAIDs at or after age 65, those with one or more epsilon4 alleles had higher baseline scores (0.95 points; p = 0.03) but did not show subsequent difference in change in score over time (0.06 points per year; p = 0.56). Those without an epsilon4 allele who started NSAID use after age 65 showed greater decline than nonusers (-0.16 points per year; p = 0.02). Nonsteroidal anti-inflammatory drug use may help to prevent cognitive decline in older adults if started in midlife rather than late life. This effect may be more notable in those who have one or more APOE epsilon4 alleles.