Get Newsletter
AlzRisk Paper Detail

Reference: Launer, 2000
Cohort: Honolulu-Asia Aging Study
Risk Factor: Blood Pressure

Average Follow-up Time Detail
Shortly after the cohort was established, three assessments were completed: 1965-1968, 1968-1970, and 1971-1974. Surviving participants were assessed for dementia status between 1991 and 1993. For the entire cohort, elapsed time between initial blood pressure measurement and dementia assessment was 27 years.

Exposure Detail
Untreated persons were those who were never treated with anti-hypertensive medications. Treated persons were those who reported treatment at any of the four follow-up examinations.

"At each exam, blood pressure was measured 3 times, 5 minutes apart on the left arm of a seated subject with a standard sphygmomanometer with a standard cuff. Diastolic pressure was recorded as the fifth phase."

"As reported previously 20, we created the following categories of mid-life systolic blood pressure history: low (less than 110 mmHg), normal (110–139 mmHg), borderline (140–159 mmHg) and high (160 mmHg or higher). For mid-life diastolic blood pressure the categories were: low (less than 80 mmHg), normal (80–89 mmHg), borderline (90–94 mmHg), and high (95 mmHg or higher). An individual was positive for a given category if his measured blood pressure fell within the categorical limits in two out of three exams. Those who did not meet this criterion were classified as having a mixed history, resulting in five categories of mid-life blood pressure history."

Ethnicity Detail
Inclusion criteria required all participants to be Japanese Americans living on Oahu, Hawaii in 1965-1968.

Age Detail
These age statistics reflect the entire cohort, not only the untreated individuals.

Screening and Diagnosis Detail
Screening Method:
CASICognitive Abilities Screening Instrument (Teng 1994)
IQ-CODEInformant Questionnaire for Cognitive Decline in the Elderly (Jorm 1989)

AD Diagnosis:
NINCDS ADRDA National Institute of Neurological and Communicative Diseases and Stroke/Alzheimer's Disease and Related Disorders Association Criteria (McKhann 1984)

Total dementia definition: Total dementia is defined as as meeting DSM-III-R criteria. However, the authors excluded from analyses all (30) participants who did not have AD or VD; therefore total dementia is effectively defined as AD + VD.

"The three-stage case-finding procedure for dementia has been reported elsewhere 40. Briefly, all men underwent testing for cognitive function using the 100-point Cognitive Abilities Screening Instrument (CASI) [37]. A sub-sample of respondents was randomly selected for assessment within strata of age and CASI score using cell sample fractions proportional to the probability of dementia. All men 85 years and older and those with a CASI score <74 were included. This selection resulted in a 28% (n = 1063) probability sample of the cohort. Phase II included neuropsychological testing, a clinical examination and a test of hearing and vision; the IQCODE questionnaire 19 was administered to an informant to assess changes in cognitive function, autonomy and behavior over a 10 year period. Phase III included all individuals with persistently poor CASI scores, an IQCODE score of > 3.6, plus a stratified random sample of the remaining phase II participants. In total 426 individuals completed the three-step procedure, 226 of who were diagnosed with a sub-type of dementia. CT neuroimaging was available for diagnosis in 85% of the Alzheimer’s disease cases and 88% of the vascular cases. Diagnoses were made in conference; dementia was diagnosed according to DSM-IIIR criteria; 1 probable and possible Alzheimer’s disease according to the NINCDS-ADRDA criteria 22 and vascular dementia according to the criteria proposed by the California Alzheimer’s Disease and Treatment Centers6. Among those diagnosed with Alzheimer’s disease (n = 118), 49 (41%) met the NINCDS-ADRDA criteria for possible Alzheimer’s disease; of these 30 had cerebrovascular disease judged to contribute to the dementia. Among those diagnosed with probable or possible AD with no contributing cerebrovascular disease (n = 88), 24 came to autopsy and 15 (63%) met CERAD neuropathologic criteria for definite or probable AD24."

Covariates & Analysis Detail
Analysis Type:
Logistic regression

"In the analyses, we adjusted for the confounding effects of age (in years) at the fourth examination, education (primary, middle, high), apolipoprotein ε allele phenotype (ε3 allele as reference (ε4 allele present, ε2 allele present, and ε42), smoking through exam 3 (never, quit before exam one, current at exam three, quit between exams one and three, restarted between exams one and three) and alcohol consumption in exam three (one drink/day, one and two drinks/day and > two drinks per day)12....For the analyses, 30 subjects with other dementias (those attributed primarily to Parkinson’s disease, subdural hematoma, B12 deficiency, progressive supranuclear palsy, or trauma) were not included in the analyses because there was no a priori reason to expect that blood pressure contributes to those dementia subtypes. Furthermore, due to the limited number of cases, some cells were too small to perform analyses. This resulted in an analytical sample of 3703, with 197 demented individuals, 118 of who were diagnosed with Alzheimer’s disease and 79 with vascular dementia. All analyses were performed using unweighted data. Persons not selected for evaluation of dementia status were presumed to be not demented. Because misclassification of those persons with dementia who were not evaluated would bias our findings in the direction of the null hypothesis, the use of unweighted data should result in a conservative approach to investigate associations of interest. To account for missing values of the ankle-brachial index a dummy variable (missing yes/no) was entered in the model."

"Multiple logistic regression was used to calculate the odds ratio (95% confidence intervals), as a measure of the risk for dementia at a given blood pressure level, compared to the reference group with normal blood pressure levels. Analyses were stratified by whether or not the men were ever treated with anti-hypertensive medication. We did this to better study the long-term effect of raised levels of blood pressure. In addition, previous analyses on cognitive function suggested the men with untreated high blood pressure had the highest risk for cognitive impairment. Stratification was appropriate as the interaction term between treatment and high systolic blood pressure was marginally significant (p = 0.06) and was significant for borderline high diastolic blood pressure (p = 0.05) and for high diastolic blood pressure (p = 0.006) in models with dementia as the outcome. In separate models we entered mean diastolic (systolic) blood pressure and its squared term to test for a non-linear relation between blood pressure and dementia. The models we present classified over 80% of the subjects correctly as case or non-case and were all significantly better (by the log likelihood deviance test) than reduced models 18."

AD Covariates:
ALCalcohol intake
APOE234APOE e2 e3 e4 genotype
SMsmoking status

TD Covariates:
ALCalcohol intake
APOE234APOE e2 e3 e4 genotype
SMsmoking status