SAN DIEGO—Among healthy men age 60 and older, subjective memory complaints may predict clinical memory impairment eight years later, according to a large, ongoing study. “This [finding] suggests the utility of subjective memory complaints for future prevention trials,” said Erin L. Abner, PhD, at the Clinical Trials Conference on Alzheimer’s Disease.
Dr. Abner, Assistant Professor of Epidemiology at the University of Kentucky College of Public Health in Lexington, and her associates found that hypertension, diabetes, African American ethnicity, and education level were significant risk factors for clinical impairment. The use of antihypertensive agents and statins, however, appeared to be protective factors.
The PREADVISE Trial
In 2002, Dr. Abner and colleagues initiated the Prevention of Alzheimer’s Disease by Vitamin E and Selenium (PREADVISE) trial, which was designed to assess the effect of vitamin E and selenium on the incidence of Alzheimer’s disease and other neurodegenerative disorders. The PREADVISE study was an ancillary to the Selenium and Vitamin E in Preventing Prostate Cancer (SELECT) trial, which was suspended in 2008 because of an interim futility analysis.
In PREADVISE, 7,553 nondemented men age 60 and older were randomized to 400 IU/day of vitamin E, 200 µg/day of selenium, both, or matching placebos. In 2010, 4,246 of the participants consented to centralized follow-up by telephone. Of these patients, 3,701 have been screened to date, said Dr. Abner.
Eligible participants scored 5 or higher on the Memory Impairment Screen (MIS) at baseline. “They received alternating versions of the MIS in subsequent years to minimize learning effects,” said Dr. Abner. “If they failed the MIS, they were given the Consortium to Establish a Registry for Alzheimer’s Disease [CERAD] neuropsychologic test battery. If their age-adjusted CERAD score was 35 or less, they were referred to a physician for a medical work-up and had their medical records forwarded to us,” said Dr. Abner.
If patients failed the MIS over the phone, they were administered the Modified Telephone Interview for Cognitive Status (TICS-M) test. “If they scored 31 or less on TICS-M, they were advised to seek a medical work-up,” Dr. Abner explained. “Although the records were forwarded to us, we were still blinded to treatment status.”
The main outcome of interest was clinical impairment, which was defined as impaired cognition indicated by failing scores on the CERAD or the TICS-M. Impairment was confirmed by a medical work-up, and suspected impairment was defined as impairment that had not been confirmed. Subjective memory complaints were defined as self-perceived changes in memory that may not be reflected by cognitive testing. “Subjective memory complaints are common in older adults and have been previously reported to predict future cognitive decline,” said Dr. Abner. “All PREADVISE participants were asked at baseline if they had noticed any changes in their memory. Positive responders were also asked if they thought they had more problems with their memory than most people.” Thus, participants were classified as having no memory complaint, memory changes, or memory problems at baseline.
The researchers used a stepwise Cox proportional hazards model to identify risk factors associated with the time from baseline to clinical impairment. The model included fixed and time-dependent predictive variables. Fixed variables, which were self-reported, included baseline age, education, African American ethnicity, Hispanic ethnicity, mother’s age at childbirth, family history of dementia, and comorbidities. Time-dependent variables were depression, anxiety, alcohol abuse, high cholesterol level, antihypertensive use, hyperglycemia, thyroid disorder, and sleep apnea.
African American Ethnicity Increased the Risk of Impairment
The mean age of the 3,701 participants was 68, and most (77%) participants had no memory complaints at baseline. Participants underwent an average of eight annual assessments. Hypertension was the most common comorbidity (34%), and 22% of patients used a statin. High cholesterol was the most common time-dependent comorbidity (68%), followed by antihypertensive use (64%), sleep apnea (18%), and depression (11%).
Dr. Abner reported that 436 (12%) of the 3,701 men screened to date have clinical impairments. Of these patients, 332 have suspected mild cognitive impairment (MCI), 85 have confirmed MCI, four have suspected dementia, and 15 have confirmed dementia.
The risk of clinical impairment was significantly increased by an older age at baseline (hazard ratio [HR], 1.12 for every one-year increment); baseline hypertension (HR, 1.96); African American ethnicity (HR, 3.24); a high school education or less (HR, 1.66); a history of diabetes (HR, 1.37); and, in the absence of APOE 4, baseline memory complaint (HR, 1.66 for changes vs no complaint and HR, 4.48 for problems vs no complaint). “It turns out that in the absence of APOE 4, subjective memory complaints are predictive of clinical impairment,” said Dr. Abner. “When APOE 4 is present, the hazard ratios are not significant.” The risk of an observed impairment was significantly reduced by the use of antihypertensives (HR, 0.34) and a report of high cholesterol (HR, 0.75).