|
|
4 Effect
Background The APOE
Methods Using local advertisements, we recruited cognitively normal subjects between the ages of 21 and 97 years, who were grouped according to their APOE
Results We analyzed 815 subjects: 317 APOE
Conclusions Age-related memory decline in APOE
4 allele is associated with the risk of late-onset Alzheimer's disease. The age at which memory decline diverges among persons who are homozygous for the APOE 4 allele, those who are heterozygous for the allele, and noncarriers is unknown. 4 status. We then followed the subjects with longitudinal neuropsychological testing. Anyone in whom mild cognitive impairment or dementia developed during follow-up was excluded. We compared the rates of decline in predetermined cognitive measures between carriers and noncarriers of the APOE 4 allele, using a mixed model for longitudinal change with age. 4 carriers (79 who were homozygous for the APOE 4 allele and 238 who were heterozygous) and 498 noncarriers. Carriers, as compared with noncarriers, were generally younger (mean age, 58.0 vs. 61.4 years; P<0.001) and were followed for a longer period (5.3 vs. 4.7 years, P=0.01), with an equivalent duration of formal education (15.4 years) and proportion of women (69%). Longitudinal decline in memory in carriers began before the age of 60 years and showed greater acceleration than in noncarriers (P=0.03), with a possible allele–dose effect (P=0.008). We observed similar although weaker effects on measures of visuospatial awareness and general mental status. 4 carriers diverges from that of noncarriers before the age of 60 years, despite ongoing normal clinical status.
4 allele, the most prevalent known genetic risk factor for Alzheimer's disease, may account for up to half of all sporadic and familial late-onset cases of Alzheimer's disease.12,13 Carriers of the APOE 4 allele have more rapidly progressive memory loss and reduced learning efficiency in their 50s and 60s than do those who do not carry the APOE 4 allele (noncarriers).14,15,16 Such decline correlates with reduced cerebral metabolism as much as 5 to 10 years before the onset of cognitive symptoms.17 The point of transition from normal aging to Alzheimer's disease has been sought in population-based studies, but cross-sectional designs are limited by demographic differences among subjects that influence neuropsychological test results. Longitudinal studies may be limited by loose entry criteria and attrition of subjects.18 Studies that focus on the elderly19,20,21 are ill designed to determine the age at which the rate of memory decline in clinically healthy APOE e4 carriers diverges from that of memory decline in noncarriers.
We therefore performed longitudinal growth modeling on a genetically enriched cohort, using a mixed-model approach for cross-sectional and longitudinal data to compare the age-related trajectories of memory loss in APOE 4 carriers and in noncarriers in the absence of mild cognitive impairment or Alzheimer's disease.
Methods
Study Subjects
From January 1, 1994, through August 6, 2007, using local media advertisements, we recruited cognitively normal residents of Maricopa County who were 21 years of age or older into the Arizona APOE cohort, a longitudinal study of cognitive aging.15 In addition, from January 1, 2000, through August 6, 2007, cognitively normal residents of Maricopa and Pima Counties over the age of 65 years were enrolled in either the Arizona APOE cohort or the Arizona Alzheimer's Disease Center cohort. With the exception of age, the cohorts were similar to each other. We gathered data on demographic characteristics and family and medical history for each subject undergoing APOE genotyping, and identity was coded by a study assistant. Race or ethnic background was determined by self-report. All subjects provided written informed consent to participate in the study, which was approved by the institutional review board at each center. The subjects agreed to have the results of APOE testing withheld from them as a precondition of their participation in the study. Genetic determination of APOE allelic status was performed with the use of polymerase-chain-reaction assays.22
The recruitment strategy for the Arizona APOE cohort involved recruiting, in groups of four, two 4 carriers and two noncarriers, matched according to age, sex, and educational level. The original protocol called for the inclusion in each group of either one homozygous subject and one heterozygous subject or two heterozygous subjects; all heterozygous subjects carried 3 on the other allele, and we excluded heterozygous subjects carrying the protective 2 allele. However, we identified many more heterozygous subjects and noncarriers than homozygous subjects, which precluded the inclusion of a homozygous subject in each group, especially those comprising subjects over the age of 70 years.
Each subject was then invited to return for screening tests that included a neurologic examination, the Folstein Mini–Mental State Examination (MMSE),23 with scores ranging from 0 to 30 and scores of 20 to 26 indicating some cognitive impairment; the Hamilton Depression (Ham-D) Rating Scale,24 with scores ranging from 0 to 49 and higher scores indicating more severe depression; the Functional Activities Questionnaire (FAQ); the Instrumental Activities of Daily Living (IADL) scale; and the Structured Clinical Interview from the third revised edition of the Diagnostic and Statistical Manual of Mental Disorders.25 Entry criteria included a score of at least 27 on the MMSE (and a score of at least 1 out of 3 on the recall subtest),23 a score of 10 or less on the Ham-D rating scale24 at the time of their first visit, and no indication of loss of function according to the FAQ and the IADL scale. The entry criteria for the Arizona Alzheimer's Disease Center cohort (which did not match the patients in groups of four) were identical. None of the subjects had potentially confounding medical, neurologic, or psychiatric problems, nor did they meet the published criteria for mild cognitive impairment,6 Alzheimer's disease,26 any other form of dementia,27 or major depressive disorder.27
During follow-up, we excluded 16 subjects (4 noncarriers, 4 heterozygotes, and 8 homozygotes) who met published criteria for mild cognitive impairment,6 Alzheimer's disease,26 or any other form of dementia.27 A consensus panel of behavioral neurologists determined the diagnostic status at entry and follow-up.
Neuropsychological Testing
On the basis of previous experience,18,19,20,24,28 as the primary end point, we selected a single measure of long-term memory loss, the long-term memory score on the Auditory–Verbal Learning Test (AVLT-LTM), which ranges from 0 to 15, with higher scores indicating better performance.29 The AVLT-LTM was administered to subjects as part of a standardized battery of neuropsychological tests15 at baseline and then at intervals of 1 to 2 years. (The duration of participation in the ongoing trial is until death, the onset of mild cognitive impairment or dementia, or voluntary withdrawal from the study.) Single measures were also selected a priori for the evaluation of general cognition and nonmemory domains. These instruments included the Folstein MMSE; the Controlled Oral Word Association Test (COWAT), which tests executive and language skills, on a scale with a lower limit of 0 and no upper limit, with higher scores indicating better performance24; and the Judgment of Line Orientation (JLO) test for evaluation of visuospatial function, with scores ranging from 0 to 30 and higher scores indicating better performance.24
Statistical Analysis
To isolate the longitudinal effect of age on cognitive measures in cross-sectional and longitudinal data, we used a statistical method that simultaneously modeled the cross-sectional and longitudinal effects of age on each cognitive measure.30 We focused on the longitudinal effect because the cross-sectional effect is potentially masked by selection bias (i.e., subjects were required to be cognitively normal at baseline, which limited the cross-sectional effect of age on each cognitive measure). We compared the acceleration in the rate of decline in each of the predetermined measures for carriers (collectively and also separately for each of the subgroups of homozygous and heterozygous subjects) with the acceleration in the rate of decline in noncarriers, using a mixed-model approach for cross-sectional and longitudinal data (for details, see the Supplementary Appendix, available with the full text of this article at NEJM.org).30,31
A quadratic model was selected to allow for comparison of the acceleration in the rate of decline between groups. Age was centered (i.e., the mean age was subtracted from each age) in analyses to reduce the correlation between the age and age-squared terms and to aid in the interpretation of coefficients. From these models, a test of significance of the interaction term between the variable for carrier status (1 for carriers and 0 for noncarriers) and the modified age-squared variable was used to assess the difference between carriers and noncarriers in the quadratic longitudinal effect of aging on the outcome measure being modeled. Modeling was performed with the use of the PROC MIXED procedure in SAS software, version 9 (SAS Institute).
In subsequent analyses, the model was modified to replace the carrier variable with a continuous variable equal to 0 for noncarriers, 1 for heterozygotes, and 2 for homozygotes; this model was used only to test for linear trend associated with allele dose. The model was also modified with two indicator variables to assess differences between noncarriers and heterozygotes and between noncarriers and homozygotes; this model was used for all other reported results. These analyses were prespecified but were considered exploratory, given the small number of homozygotes. Baseline characteristics and those recorded during follow-up were compared among groups with the use of the two-sample t-test, the analysis-of-variance F-test, or the Pearson chi-square test.
Results
Subjects
We analyzed 815 subjects: 317 APOE 4 carriers (79 homozygous subjects and 238 heterozygous subjects) and 498 noncarriers (Table 1). There were fewer carriers than noncarriers because we identified fewer healthy APOE 4 carriers over the age of 75 years. Overall, carriers were younger than noncarriers (mean age, 58.0 vs. 61.4 years; P<0.001) and had a higher reported rate of having a first-degree relative with dementia (73.5% vs. 52.8%, P<0.001). However, adjustment for the presence of a first-degree relative with dementia did not significantly alter the results for any of the measures. With respect to race or ethnic background, 85% of the subjects identified themselves as white non-Latino, 12% as Latino, and 3% as other. There were no significant differences between carriers and noncarriers in sex (female sex, 68.8% and 69.1%, respectively; P=0.99), educational level (mean years, 15.4 in the two groups; P=0.98), or the number of subjects who underwent more than one session of testing (76.0% and 73.1%, respectively; P=0.08). However, among subjects who underwent more than one session of testing, carriers had slightly more years of follow-up than noncarriers (5.3 vs. 4.7 years, P=0.01). There were more subjects in two age groups — 50 to 59 years and 60 to 69 years — than in younger age groups, including a higher proportion of 4 carriers (especially homozygotes).
|
4 carriers than among noncarriers (P=0.03) (Figure 1, and Table S1 in the Supplementary Appendix). The mixed model for the AVLT-LTM predicted a decline in scores for APOE 4 carriers beginning in their 50s (on the basis of a comparison of a predicted annual increase at age 50 with a predicted annual decline at age 60) and a decline in noncarriers beginning in their 70s (Table 2). There were no significant differences in the quadratic longitudinal effects of aging between carriers and noncarriers on the MMSE (P=0.75), the COWAT (P=0.57), or the JLO test (P=0.78). However, there were significant differences in the linear longitudinal effects of aging between carriers and noncarriers on the MMSE (P=0.03) and the JLO test (P=0.009).
|
|
4 allele–dose effect with quadratic age on the AVLT-LTM (P=0.008 for trend). However, the difference was significant only for the comparison between homozygotes and noncarriers (P=0.009) and not for the comparison between heterozygotes and noncarriers (P=0.18). In addition, the small size of the homozygous subgroup warrants caution in interpreting these results. The mixed model for the AVLT-LTM predicts a decline in scores for APOE 4 homozygotes starting in their 50s, a decline in scores for APOE 4 heterozygotes starting in their 60s, and a decline in scores for noncarriers starting in their 70s (Figure 2, and Table S3 in the Supplementary Appendix). For the other measures, there was no significant association between the linear APOE 4 allele–dose effect with quadratic age and the MMSE score (P=0.12 for trend), the COWAT score (P=0.35 for trend), or the JLO score (P=0.22 for trend), although significance was reached for homozygotes, as compared with noncarriers, for the MMSE score (P=0.01) and the JLO score (P=0.04) (Fig. S1 in the Supplementary Appendix).
|
Our findings suggest that the APOE 4 allele affects age-related memory performance independently of mild cognitive impairment and dementia. They also suggest that accelerated memory decline among persons with the APOE 4 allele may be caused by subclinical Alzheimer's disease, since age-related memory decline was increased and visuospatial function was decreased in carriers of the APOE 4 allele. In addition, our model-predicted memory decline coincided with the predictions of Corder et al.12 for the age of onset of Alzheimer's disease on the basis of the APOE 4 genotype.
Other lines of evidence support this explanation. We previously found that subjects in their early 60s who were homozygous for the APOE 4 allele had accelerated neuropsychological decline in a pattern that was specific to the cognitive domain (termed pre–mild cognitive impairment), which anticipated the onset of clinically symptomatic memory loss by several years.16 A study of subjects with such a pattern of memory decline showed a reduced cerebral metabolic rate for glucose patterns in cortical regions that overlap those known to be affected by Alzheimer's disease.17 Moreover, studies of asymptomatic subjects in their early 60s who were homozygous for the APOE 4 allele showed early amyloid deposition in frontal cortexes and in other regions that are affected in Alzheimer's disease.32 The findings we report here temporally coincide with, and so may be the cognitive correlate of, this presymptomatic disease stage. It is less clear, however, whether the memory decline we observed was a direct reflection of this pattern of amyloid deposition, of the formation of neurofibrillary tangles in the entorhinal cortex,33 or another pathogenic event. Finally, neuropathology that resembles Alzheimer's disease in subjects in their 30s and 40s34,35 and in elderly subjects without dementia36,37,38 is more prevalent and severe in APOE 4 carriers than in noncarriers.
APOE 4 may be a genetic cause of dementia before the age of 60 years,39 but there is no consensus on the potential effect of APOE 4 on cognitive functioning in healthy younger adults. In a previous study,40 we found no correlation between APOE genotype and intellectual achievement, as measured by educational and occupational outcomes, and our current data show that age-related cognitive decline before the age of 50 years is essentially identical in APOE 4 carriers and noncarriers.
Among the potential limitations of our study, we focused on carriers of the APOE 4 allele rather than on the general population. Subjects who are homozygous for APOE 4 represent the single largest source of persons whose risk of Alzheimer's disease is nearly that of autosomal dominant mutation carriers. This small but important subgroup provides an opportunity to study the changes that may occur before the clinical onset of mild cognitive impairment or Alzheimer's disease, and the compilation of such a cohort is not practical in a population-based study in which subjects are randomly recruited. It is possible that we recruited persons concerned about their own cognitive health, possibly because they have Alzheimer's disease. We therefore eliminated anyone who was found to have clinically symptomatic mild cognitive impairment or dementia at any point. Although this factor might raise the possibility of survivor bias in our study, the number of clinical "converters" was small, and if anything this would have reduced the sensitivity to an effect of APOE 4 on the rate of memory loss. Another potential limitation was the unbalanced distribution of age and APOE 4 status with respect to age. A majority of subjects were over the age of 50 years, and the noncarrier group was slightly older than the group carrying the risk allele. Possibly the study of a larger cohort of younger subjects or the use of an instrument that detected memory loss with greater sensitivity than the AVLT-LTM would implicate the effect of APOE 4 at an earlier age than we have reported here.
Supported by the National Institute on Aging (P30-AG19610 and R01-AG031581), the National Institute of Mental Health (R01-MH057899), the Alzheimer's Association (IIRG-98-078), the Arizona Alzheimer's Consortium, and the State of Arizona.
Dr. Caselli reports receiving consulting fees from Myriad Pharmaceuticals and Medivation and having an equity interest in Pfizer; Dr. Sabbagh, receiving consulting or advisory board fees from Amerisciences, Eli Lilly, and Elan/Wyeth, lecture fees from Novartis, Eisai, Pfizer, and Forest Laboratories, and grant support from Eli Lilly, Medivation, Abbott, and Bristol-Myers Squibb; Dr. Ahern, receiving research support from GlaxoSmithKline, Elan Pharmaceuticals, and Avid Radiopharmaceuticals; Dr. Shi, receiving lecture fees from Novartis; and Dr. Woodruff, receiving grant support from GlaxoSmithKline. No other potential conflict of interest relevant to this article was reported.
We thank Sandra Yee Benedetto, Bruce Henslin, Jessie Jacobsen, Anita Prouty, Yi Zhuang, Marci Zomok, and Bonnie Schimek for technical assistance and Joseph Hentz for statistical guidance during the initial stages of this project.
Source Information
From the Mayo Clinic Arizona, Scottsdale (R.J.C., A.C.D., D.O., B.K.W., D.E.C.L., C.H.S.); Sun Health Research Institute, Sun City (M.N.S., D.J.C.); the University of Arizona, Tucson (G.L.A., S.Z.R., G.E.A., E.M.R.); Barrow Neurological Institute, Phoenix (L.C.B., J.S.); and Banner Alzheimer's Institute and Translational Genomics Research Institute, Phoenix (E.M.R.) — all in Arizona; and the Mayo Clinic, Jacksonville, FL (R.R.).
Address reprint requests to Dr. Caselli at the Department of Neurology, Mayo Clinic Arizona, 13400 E. Shea Blvd., Scottsdale, AZ 85259, or at caselli.richard{at}mayo.edu.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Age-Related Memory Decline and the APOE 4 Effect
Chuang J.-Y., Ma L., Zhao G., Xia F., Caselli R. J., Dueck A. C., Reiman E. M.
Extract |
Full Text |
PDF
N Engl J Med 2009;
361:1996-1997, Nov 12, 2009.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
