Table 1 Study design

Population

AD and mixed dementia patients with hyperlipidaemia

Outcome

MMSE points

Statistical model

Linear mixed-effects regression models with inverse probability of drop out weighting

Medication use

Time-updated exposure

• Primary analysis: average DDD during the 6-month period preceding index date

• Secondary analysis: yes/no use of medication during the 6-month period preceding index date

Analysis

Primary:

1. Statin users vs non-users of statins

Secondary:

2. Simvastatin vs atorvastatin users

3. Simvastatin vs rosuvastatin users

4. ¹Lipophilic vs ²hydrophilic statins users

5. ³Fungal vs ⁴synthetic statins users

6. Non-statin lipid lowering medication users vs statin users

Covariates

Age, gender, diagnosis year, residency, care unit, comedications, comorbidities

Stratified analyses

1) Gender, 2) age (split at mean age at index date)

Sensitivity analyses

1) Multiple imputations of MMSE, 2) incident users