Get Newsletter
AlzRisk Paper Detail

Reference: Cornelius, 2004
Cohort: Kungsholmen Project
Risk Factor: Non-Steroidal Anti-Inflammatory Drugs

Exposure Detail
Interviewers ascertained the drug use through prescription forms and/or drug containers that the participants provided at baseline and at the first follow-up visit. Information on the name, strength, and dosage of the drug was collected. If the participant could not provide the drug information, researchers interviewed a relative or “health care personnel” and the responses were checked against medical records. The investigators reported results separately for current vs. not using any NSAID, current vs. not using non-ASA NSAID, and current vs. not using ASA. This entry pertains to results on the current vs. not using non-ASA NSAID.

The investigators compared AD risk in two groups: the group of participants who used non-ASA NSAIDs at baseline ("Current use") and the reference group of participants who did not use non-ASA NSAIDs at baseline ("Not using").

Ethnicity Detail
The study cohort are comprised of inhabitants in the Kungsholmen district of Stockholm, Sweden. The distribution of ethnic backgrounds were not reported.

Age Detail
1,020 participants (78.4%) were of ages 74-85 years, and 281 (21.6%) participants were over 85 years old.

Screening and Diagnosis Detail
Screening Method:
MMSEMini-Mental State Examination (Folstein 1975)

AD Diagnosis:
DSM IIIR Diagnostic and Statistical Manual III-Revised
Hachinski's Ischemic Scale (Hachinski, 1975)

Covariates & Analysis Detail
Analysis Type:
Cox proportional hazards regression

"Cox proportional hazards regression models were used to calculate the relative risks (RRs) of dementia and AD among users of aspirin and NSAIDs (adjusted for age, gender, education, underlying diseases), and the 95% confidence intervals (95% CIs).
In a first model, we looked at those who were free of dementia
after the first follow-up (n = 743). If these persons used aspirin or NSAIDs at both baseline and first follow-up (a period of around 3 years), they were considered taking the medication continuously during that time. Incident cases of AD and dementia were assessed at the second follow-up. The time at risk was calculated between the first and second follow-up (around 3 years) and was halved in the event of AD/dementia. Age at first follow-up was used as a continuous variable
whereas the other covariates were treated in a dichotomous way.
A second model was used to prolong the time at risk of developing the disease and increase the cohort. In this model the cohort of dementia-free subjects at baseline (n = 1,301) was used. Regular use of aspirin or NSAIDs at baseline was assessed. Incident cases of AD and dementia were assessed at both first and second follow-up. For those who were diagnosed with AD/dementia at the first follow-up, the time at risk was calculated from baseline to first follow-up (approximately 3 years) and then halved. For the subjects who developed dementia between the first and second follow-up, the time at risk was calculated from the first to the second follow-up (approximately 3 years) and was halved. Age at baseline was used as a continuous variable, other covariates were dichotomised."

AD Covariates:

TD Covariates:

Age was treated as a continuous variable.